key: cord-0005036-ss1d834y
authors: nan
title: Nucleic acids structure and gene expression
date: 1993
journal: Experientia
DOI: 10.1007/bf02073406
sha: 317d8d3f6e1d5181db422854e0a84c1f2732752f
doc_id: 5036
cord_uid: ss1d834y

nan

Low doses of radiation or genotoxic chemicals reduce cytogenetic damage in cells upon subsequent treatment with larger doses of these agents, a phenomenon known as the adaptive response. We were able to detect distinct biochemical changes in quiescent human keratinocytes (HaCaT cell line) adapted to a very low concentration of the alkylating agent Nmethyl-N'-nitro-N-nitrosoguanidine (MNNG). One hour adaptation of the cells with 2.5 nM MNNG decreased the levels of constitutive DNA strand breaks, NAD and pely(ADPribose). Subsequent challenging of HaCaT cells with 2.5 ~tM MNNG reduced the amount of DNA strand breaks and NAD levels, and increased poly(ADP-ribose) levels, as compared to cells exposed to the challenging treatment alone. The response was dependent on the length of the challenge treatment and was abolished by 2 mM benzamide (an inhibitor of poly(ADP-ribosyl)ation). This is the first demonstration of early biochemical events accompanying the adaptive response of human cells to an alkylating agent.

Bednar, J., Furrer, P., Stasiak, A. and Dubochet, J., Laboratoire d'Analyse UItrastructurale, Universite de Lausanne, CH-1015 Lausanne-Dorigny. A cryo-electron microscopy approach to study the shape of supercoiIed DNA molecules freely suspended in vitrified buffer was combined with Monte Carlo simulations. We investigated the influence of the neutralization of intersegmental electrostatic repulsion on the writhe of supercoiled DNA molecules. We provide experimental evidence that the decrease of the effective diameter of DNA causes a higher fraction of the linking number deficit to be absorbed by the writhe. We show that under ionic conditions typical for in vitro reactions with DNA, supercoiled DNA adopts a so called tight intenNound structure. In this structure the two opposing segments of interwound superhelix are in direct longitudinal contact with each other. It is likely that under in vivo conditions where the charge of DNA is greatly neutralized, supercoiled DNA also adopts this tight form. We propose that properties of the tight form (direct longitudinal contacts of two DNA helices) can have important consequences for cis acting regulatory mechanisms like transcriptional and recombinational enhancement. Fanconi'8 anemia (FA) is an autosomal recessive disease associated with chromosomal instability and a predisposition to cancer. A defect in DNA repair appears to be involved, however, studies concerning the ability of FA cells to repair DNA damage have led to conflicting results. To study ONA repair, we treated control and FA cells in culture with mitomycin C (MMC) which results in DNA-interstrand crosslinks that can be quantified before and after repair. Quantification was performed at the level of the actively transcribed rRNA genes using a denaturationrenaturation gel electrophoresis procedure. It was apparent for all cell lines studied that MMCcrosslinks were induced rapidly and that their level remained unchanged after 12 hours, indicating that MMC lesions are more stable in the genome than previously reported. After 48 hours, repair efficiencies varied greatly in the two control cell lines suggesting heterogeneity in crosslink repair of rRNA genes. In two FA cell lines, orosslink repair was not detectable over 48 hours supporting the idea that FA cells have an impaired DNA~crosslink repair mechanism. The group I intron located within the chloroplast 23S ribosomal RNA genes of the green unicellular algae Chlamydomonas reinhardfii has been shown to have the ability to move to an intronless version of the 23S rRNA gene (23S cDNA). This intron encodes a DNA endonuclease (I-Crel) which specifically recognizes and cleaves the 23S cDNA close to the intron-insertion site. Deletion of the I-Crel gene from the chloroplast DNA showed that the I-CreI endonuclease is required for the in vivo mobility of this intron. The I-Crelstrain was crossed to a wild-type strain such that the modified chloroplast DNA (I-CreI-, 23S cDNA) was inherited by all the progeny. Shortly after fusion of the cens, double-stranded DNA breaks were detected in vivG close to the intron-insertion site. Analysis of the progeny of these crosses revealed that the ribosomal intron invades with high frequency (~70%) the intronless allele of the 23S rRNA gene, thus demonstrating the genetic mobility of this intron. Xeroderma pigmentosum is a rare autosomal recessive disorder characterized by hypersensitivity to ultraviolet light and predisposition to skin cancer. Seven genetic complementation groups (XP-A to XP-G) are thought to be due to mutations in distinct genes from the nucleotide excision repair pathway. We have isolated frog and human cDNAs that encode proteins resembling RAD2, a yeast protein involved in this pathway. The human protein expressed from an EBV-based vector is able to reduce to normal levels the UV sensitivity of lymphoblastoid cells from XP group G but not C. The -4 kb XPGC mRNA is present in normal amounts in XP-G cells, suggesting that the defect may be due to a small deletion, insertion or point mutation in a normally active allele.

D0bbeling, U., Hobi, R., Berchtold, M.W. and Kuenzle, C.C. Institute of Veterinary Biochemistry, University of Ztirich, Winterthurerstrasse 190, CH-8057 Ztirich.

Rearrangement of immunoglobulin genes by V(D)J recombination normally occurs only in pre B and pre T-ceils, but experimentally also in ceils which have been stably transfected with the recombination activating genes rag-1 and rag-2. Using the transfected fibroblast cell line IA we show that V(D)J recombination is stimulated by the protein kinase A (PKA) pathway and the transcription factor Oct2A. PKA probably exerts its effect on V(D)J recombination through the transcription factor CREB, as CREB AI33S a mutant which cannot transactivate but binds well to DNA blocks efficiently the stimulation of V(D)J recombination by the cotransfected catalytic subunit of PKA. V(D)J recombination is repressed by TPA an activator of protein kinase C. This effect is probably mediated by the transcription factors AP-1 and jun B. Thapsigargin that increases the intracellular Ca 2+ concentration also reduces the rate of V(D)J recombination. TPA and thapsigargin overcome the stimulating effect of the PKA activators Br-cAMP and caffeine, but not that of Oct 2A.

Naegeli, H., Bardwell, L. and Friedberg, E.C., Department of Pathology, The University of Texas Southwestern Medical Center at Dallas, Dallas Tx 75235 NER is a genetically and biochemically complex DNA repair pathway which processes a wide diversity of DNA lesions. In S. cerevisiae RAD3 is one of at least six genes that are required for early events in NER, including damage recognition and damage-specific incision of DNA. The RAD3 gene (and presumably its human homolog ERCC2) encodes a DNA-dependent ATPase and DNA helicase. We have demonstrated that the Rad3 DNAhelicase activity is profolmdly inhibited by DNA damage caused by UVradiation, by small or bulky DNA adducts, and by sites of base loss. In all cases the inhibition of helicase activity is strictly DNA strandspecific. The Rad3 helicase is not inhibited by drugs that bind to DNA non-covalently, suggesting that Rad3 protein is exclusively sensitive to the chemistry of DNA. Using filter binding and competition assays we have also show~ that Rad3 protein forms stable complexes with damaged DNA presumably at sites of covalent base lesions. Based on these results we propose that the DNA helicase activity of Rad3 protein may serve to strand-specifically target sites of base damage for further processing by the NERmachinery. In plants, the nuclear PsaFgene encodes the PSI-F subunit of the photosystem I (PSI) complex. This subunit is located on the luminal side of the thylakoids and although its precise role is not known, it is generally accepted that it functions as a docking protein for plastocyanin, an electron carrier protein. The PsaFgene has been cloned from the green alga C. reinhardtii and we have devised a strategy to disrupt its resident counterpart by using a mutated version of it. Briefly a mutated copy of the cloned PsaF gone was cut by a restriction endonuclease to create a small gap n the middle of the coding sequence. The gene was then cotransformed along with the ASL gone (conferring arginine prototrophy) in the double mutant alga strain cw15(cell wall less) ; arg7A (arg. auxotroph), by vortexing the cells in the presence of glass beads and DNA. Disrupted PsaFmutants were detected by their characteristic PSI-fluorescence phenotype. We are currently analysing several such transformed mutants for the targetted disruption of PsaFat the DNA level.

yeast S.pombe. The product of the start gene cdclO is a component of this factor in S. pombe. In S. cerevisiae the level of activity of the DSC1 factor during the cell cycle has been subject of contradictory reports. We present results showing that the S. pombe DSCl-like factor binds to the promoter of the cdc22 gene coding for the large subanit of ribonucleotide reductase in a cell cycle dependent manner. The binding activity disappears as a cell exits from M phase and reappears in late G1 just before S-phase, as cells progress through the START control and become committed to the mitotic cell cycle with respect to other fates. Interestingly the gene coding for ribonucleotide reductase is regulated at its expression level in .many eukaryotes and may be a rate limiting step for entering into S-phase. It is not yet clear if a DSCl-like activity is present in all those organisms, but budding and fission yeasts are widely divergent evolutionarily so that systems conserved in both organisms are frequently widespread in nature. We will also present results showing which part of the p85 cdclO protein are important for the DSCl-like activity.

Martin Hug; Sandro Rusconi; Institut fiir Molekularbiologie H der Universitdt, UZ lrchel, Winterthurerstrasse 190, 8057 Ziirich, Switzerland Monotonous aminoacid repeats can have a strong influence on the activity of the host protein and are often encoded by triplet nucleofides whose number is subject to marked polymorphism. Slippage during DNA replication has been proposed among the mechanisms which could explain the remarkable incidence of length polymorphism. Slippage replication of repeated motifs has been documented so far only in vitro or in bacterial systems. In order to assess the extent of such phenomena in a eucaryotic system we have constructed plasmids containing different triplet nucleotide repeats of progressive length along with an SV40 origin of replication. In a typical assay, plasmids are transfected into FleLa cells together with an SV40 T-antigen expression vector. After two days low MW DNA is harvested and digested with Dpnl to eliminate unreplicated material. The digested DNA is transformed into competent bacteria which can be grown in a pool. The DNA prepared from the transformants serves either for another transfection cycle or is analyzed by endlabelling of a fragment spanning the repeated motif. The endlabelled fragment is analyzed in high resolution gels. The resuits suggest that slippage replication occurs in both eucaryotically and procaryotically replicating systems although the extent may be different, and that the phenomenon is dependent on both sequence composition and length as reported with in vitro systems. We speculate that slippage replication mechanisms may play an important role important during cell ageing.

Malatesta, M.~ 2, Zancanaro, C.~ 3, Vogel~ P.* and Fakan, S. ~ ~ of Electron Microscopy, and *Institute of Zoology and Animal Ecology, University of Lausanne, 21nstitute of Histology and Laboratory Analyses, University of Urbino, 31nstitute of Human Anatomy and Histology,University of Verona Modifications of the cell nucleus during rapid and intensive changes in cell activity of the brown adipose tissue (BAT) and liver from Muscardinus avellanarius were studied. Morphometric analyses of nuclear and nucleoplasmic surface as well as of perichromatin granule frequency, carried out on ultrathin sections, indicated significant tendency for all parameters to rank according to the sequence hibernating ~ arousing ~ euthermic. Moreover, ultrastructural cytochemistry allowed us to describe and characterize unusual nuclear structures such as coiled body-like constituents, amorphous bodies and bundles of nucleoplasmic fibrils in hibernating animals. Finally, a series of specific antibodies recognizing different nuclear proteins and ribonucleoproteins were used as probes, in order to better approach possible functional role of the above structural constituents. The pSC101 origin encodes a protein essential for replication, RepA, and contains a cluster of directly repeated sequences RS1, RS2, and RS3 and two palindromes, IR2 and IR1, partially homologous to the direct repeats and overlapping the repA promoter. We found that RepA binds to directly repeated sequences and initiate replication in monomeric form, but that it binds to inversely repeated sequences and autoregulates its own transcription in dimeric form (Msnen et al. PNAS 89: 8923-8927, 1992 ). Here, we study in more details the influence on the regulation of the plasmid copy number of a region which comprises the IR1 inverted repeated sequences and straddles the -35 region of the repA promoter. We show that the integrity of IR1 greatly influences pSC101 copy number. IR1 is separated from the nearest repeated sequence, RS3, by approximately four turns of the DNA helix. Copy number is preserved if this distance is increased by one whole turn but not if it is increased by a fraction of a turn. These results suggest interactions between RepA dimers binding at IR1 and RepA monomers binding at the RS sequences in the initiation of replication.

yon Schack, M.L. and Fakan, S., Centre de Microscopie Electronique de l'Universit~, CH-1005 Lausanne Cryofixation followed by freeze-substitution in an organic solvent on biological materials has been used as an alternative to conventional chemical fixation for ultrastructural studies. Mouse cell culture line and liver tissue were cryofixed without the use of any cryoprotectant by impact freezing onto a cooled copper mirror, or frozen under high pressure. They were then substituted in pure acetone without any chemical fixative and embedded in different acrylic or epoxy resins. Fine structural studies on the cell nuclei comparing different cryofixation and embedding techniques showed a nicely preserved morphology similar to conventionally prepared material but sometimes wiLh more detail regarding e.g. the nuclear membrane or nuclear RNP constituents. For larger tissue samples, high pressure freezing gave better structural preservation than slam freezing. Considering that no chemical fixatives were used, these results provide excellent morphology and are very promising for further cytochemical and immunecytochemical applications. Using cryo-electron microscopy we studied the structure of chrornatin freely suspended in a thin layer of vitrified solution. We utilized purified minichromosomes of SV40 in low salt buffers for the observations. Under these conditions there are no internucleosomal interactions and the path of DNA between the nucleosomal cores can be followed throughout the circular minichromosome. In this structure the DNA path is uniquely determined by wrapping around the nucleosome cores and by the length and number of helical turns of the linkers. When linkers are approximated as straight segments the angle between consecutive linkers should be uniquely determined by the primary structure of the nucleosome core or chromatosome, while the torsion angle for any three consecutive linkers is determined by the twist of the middle linker. By comparing the observed images with computer simulations of chromatin structure m which the effect of variable linker length and linker entry/exit angle are incorporated we are able to test and verify various models of the arrangement of nucleosomes and linker DNA in the folding of chromatin.

In order to test the structural and functional properties of histone H1 in vivo, the sea urchin histone t-tl was expressed from the inducible GALl promoter in a 21~ vector (strain YCL7). H1 copurifyies with crude NP-40 washed nuclei in roughly stoichiometric amounts compared with core histones. Fractionation of soluble chromatin on sucrose gradients showed that fractions containing DNA and core histones also contained HI. No soluble H1 was found. We conclude that H1 was bound to chromatin. However, no obvious change in the nucleosome spacingwas observed. Expression of H1 inhibits growth of YCL7 on plates and in liquid media. Furthermore, incorporation of 3H-uracil into RNA was reduced, suggesting that H1 might repress transcription by binding to chromatin. Pattern formation is generated by a spatially and temporally controlled expression of regulatory genes (i.e.homeotic genes). Early established patterns need to be faithfully maintained through development. In Drosophila the genes of the Polycomb group (Pc-G) are part of such a cellular memory mechanism, by keeping developmental regulators permanently and stably repressed in appropriate domains. We have evidence that the products of the Pe-G exert their repressory function by changing the higher order chromafin structure into a heterochromatin-like state. Proteins of the Pc-G form multimeric complexes, associated with particular target genes. So far we have found four members of the Pc-G (Pc, ph, Psc and Pcl) to be part of the complex. We are testing whether the complexes change the local chromatin organization by compaction and/or by compartmentalization within the nucleus. The Polycomb (Pc) protein contains the highly conserved chromo domain. We find that a particular modification of the domain is necessary for building the Pc-G complexes. Using in vivo cross-linking methods we have been able to monitor the DNA sequences on the homeotic BX-C interacting with the Pc-G complexes. The finding of homologous Pc-G proteins in other systems suggest that the stable maintenance of gene repression imprinted into the higher order chromatin organization is a general mechanism of gene regulation. The number of copies of the Drosophila gene Suvar(3) 7 is a dose-limiting factor in this phenomenon, and seems from its sequence to encode a large protein with six widely spaced zincfingers.

This novel arrangement of zinc-fingers could help in packaging the chromatin fibre into heterochromatin, and could reflect a novel method of controlling the expression from DNA domains. The transcript and protein distribution have been determined and show a prevalent maternal contribution.

An homolog of the gene was isolated and sequenced in D. teissieri. The sequence is notably divergent (70% identity) for a species very close to D. melanogaster, but the fingers, their spacing and other motives are well conserved. Dusserre,Y and Mermod, N., Institut de Biologie animale, Universit~ de Lausanne, CH-1015 Lausanne. TO dissect the mechanisms of gene transcription control by human CTF/NF-I DNA binding factors, purified full length or truncated proteins were covalently coupled to a chromatography resin, which was then used to fractionate cell nuclear extracts by protein affinity chromatography. Two polypeptides which specifically interact with CTF/NF-I species were identified. In vitro reconstitution experiments indicated that these polypeptide co-factors are involved in CTF/NF-I mediated induction of transcription. For instance, these cofactors restore regulation when added to a depleted extract, they mediate histone HI antirepression, and they alleviate squelching by an excess of activator. Functional and physical interaction of these purified cofactors with activators and with the basal transcription machinery will be discussed.

Martinez, E., Roy, A.L., Gregor, P., Carruthers, C., Gutjahr, T. and Roeder, R.G., Laboratory of Biochemistry and Molecular Biology, The Rockefeller University, 1230 York Ave., New York, N.Y. 10021, USA.

TFII-I is a novel basal transcription initiation factor that interacts with the initiator element(s) in the core promoter of several cellular and viral RNA polymerase II(Pol II)transcribed genes and can functionally substitute for TFIIA in an in vitro transcription system reconstituted with purified basal transcription initiation factors and Pol II. Unlike the other basal initiation factors, TFII-I also interacts with the upstream USF site found in the Adenovirus major late promoter. In order to further characterize the function of TFII-I, the human TFII-I cDNA has been cloned. TFII-I mRNA has been detected in all tissues analyzed, suggesting that the protein is ubiquitously expressed. TFII-I is a proline-rich protein containing 6 tandemly repeated domains and has the potential to form repeated a-helices and turns in alternation. Recombinant TFII-[ wild-type and mutant proteins have been produced and are currently been tested functionally. The results of these functional analyses will be presented.

Tiziano Tallone, Franca Baggi, Astrid Rtifli, Sandro Ruseoni; Institutftir Molekularbiologie H der Universitiit, UZ Irchel, Winterthurerstrasse 190, 8057 Ziirich, Switzerland In general enhancer and promoter regions are composed of clustered bindhag sites for transcription factors whose geomelrical arrangement appears to be important in determining the degree of cooperation or interference among the binding factors. We have systematically studied the influence of site arrangernent in model promoters by constructing different reporter genes with a system which allows quantitative measurement of the transcription rate. Our model clusters consist of glucocortieoid receptor response elements (GRE) intermingled with other sites (e.g. Spl or Oetarner). The spacing between elements has been varied ha different constructs and the clusters have been placed either in proximity of the TATA-box (promoter position) or at a 2 kb distance from a minimal promoter (enhancer position). Care has been taken to avoid cryptic GREs generated by bacterial dam methylation. The results confirm and extend our previous observations that the full length GR has different geometry optima than the carboxy-tnmeated GR, and further suggest that it is possible to "bridge" the cooperation among GREs by appropriately placing intervening binding sites for unrelated factors. We are testing whether enhancers can be composed of binding sites for "true" enhancer activating factors (like GR) along with sites for rather "enhancer inert" factors (like Spl) which albeit can function as key intermediates for local cooperation required for maximum enhancer activity.

Institut fiir Molekularbiologie II der Universitgt Ziirich, CH 8057 ZiJrich Katja Seipel, Hans-Peter Gerber, Oleg Georgiev, Licen Xu, Kostadin Gramatikov and Walter Schaffner

We have previously characterized activation domains of known transcription factors according to their ability to activate transcription in ceil culture (in vivo) from proximal (promoter) and/or remote (enhancer) positions (Seipel et al.,EMBO J,1992; I 1No.13) . We also tested the same chimeric proteins produced in mammalian ceils in cell-free extracts (in vitro). Activation domains rich in acidic residues stimulated Iranscription in vivo from remote and close by position, but failed to work in vitro. By contrast, glutamine-rich activation domains stimulated transcription in vitro, while in vivo they were unable to stimulate transcription from remote positions and dependent on the presence of an enhancer to work from proximal positions. Proline-rich domains exhibited intermediate features: they did not stimulate transcription in vitro, while they were weak activators in vivo from both proximal and remote positions. We also tested the chimeric proteins produced in E.coli in frog oocytes. Furthermore we found that N-terminal subdomains of the human TATA binding protein (TBP) functioned like the "proximal" activation domain of Oct-l, while the C-terminal domain of RNA polymerase II functioned like the VP16 activation domain. These findings imply that TBP and RNA pol II CTD are directly involved in promoter and enhancer function, respectively.

Mahfoudi, A. I, Roulet, E. I, Parker, M.G. 2 and Wahli, W. 1 l.lnst, de Biol. Animale, Uni Lausanne,CH-1015 Dorigny,2.Mol. End. Lab., Imperial center research fund, Lincoln's Inn field, London WC2A 3PX,UK.

The oestrogen receptor is made of two activation functions AF-I in the N-terminal part of the protein, which is constitutively active, and AF-2, C-terminally located, which is ligand dependant. Cotransfection experiments in Hela cells using the wild type M0R together with the human CTF/NF-1 Proline-rich domain fused to the Gal4 DNA binding domain resulted in a synergistic transcriptional activation.Deletion of AF-2 decreases significantly the synergistic transcriptional activation whereas removal of AF-I produces a strlckingly higher synergism when compared to the wild type HeR receptor. Moreover, synergistic activation increases gradually when the HeR is progresssively deleted in AF-I. These results suggest that AF-2 is mainly responsible for the synergism observed with the CTF/NF-I Prolinerich domain and that AF-I could act rather as an inhibitor of this interaction. The use of deletion and point mutants in the AF-2 domain will help us to define the part of AF-2, which is implicated in the MOR/CTF/NF-I synergistic transcriptional activation.

The octamer transcription factor Oct-2a is mostly expressed in lymphoid ceils and belongs to the POU domain protein family. In previous work two major transcriptional activation domains at the Nterminus and the C-terminus were identified by testing N-and C-terminal deletions of Oct-2a in transcriptional activation assays done in heterologous cells (HeLa) . The N-terminal activation domain is rich in glutamine whereas the C-terminus contains a high amount of proline, serine and threonine residues. At the moment it is not clear which role the C-terminal region plays in transcriptional activation and which amino acids are involved. In addition it is also unknown whether the N-and C-terminal activation domains are interchangeable or interdependent. To further address these open questions and to find out whether duplication or multimerization of single activation domains create a more potent activator, we have constructed expression vectors containing different parts of the Oct-2a activation domains and the DNA binding domain of the yeast transcription factor GAL4 in varying arrangements. These chimeric proteins were tested in transient transfection assays performed in HeLa cells as well as in the B cell lines X63Ag8 and Namalwa to see whether there is a B cell-specific factor involved in transcriptional activation by Oct-2a. The conventional assay for transcriptional activation domains is based on testing the activity of fusion proteins, consisting of a DNA binding domain and different segments of the protein to be analyzed. We have developed a new and fast method, which allows selection in mammalian cells and subsequent cloning of transcriptional activation domains from a mixture of DNA segments. The DNA to be analyzed is fragmented by sonication and fused to the GAL4 binding domain (aa 1-93) in an expression vector containing the SV40 origin of replication. These randomly generated pools of fusion proteins are transfected into monkey ceils, which are stably transformed with the SV40 T-antigen gene under the control of a promoter with five copies of the GAL4 binding site. Replication of the transfected plasrnids occurs only in ceils expressing a transactivating fusion protein, which in turn stimulates T-antigen expression. Replicated plasmids are selected by DpnI digest, individual clones are isolated and finally analyzed by sequencing. Using this selection scheme, we have isolated from a large excess of unrelated DNA the activation domains of the herpesvirus VP16 and pseudorabies IE proteins.

The Mapping of RNA polymerase on Mammalian Genes in Cells and Nuclei. Jovan Mirkoviteh* and James E. Darnell, Jr. The Rockefeller University, USA. *Present address: ISREC, 1066 Epalinges The assembly of an RNA polymerase II initiation complex at a promoter is associated with the melting of the DNA template. Using the specific single-stranded modifying reagent KMn04 and a new genomic sequencing technique, we have explored the melted regions of specific genes in genomic DNA of whole cells or of isolated nuclei. We have demonstrated for the first time in vivo the melting in the promoter proximal transcribed region associated with the presence of RNA polymerase II complexes. An interferon-inducible gene exhibited KMn04 sensitivity downstream of the RNA initiation site in interferon treated cells when the gene was actively transcribed, but not in untreated cells where the gene was not transcribed. The extent of KMn04 modification w%s proportional to transcription levels. The KMnO 4 sensitivity was retained when nuclei where isolated from induced cells, but was lost if the engaged polymereses were further allowed to elongate the nascent RNA chains. The sensitivity to KMn04 in isolated nuclei was retained if the run-on incubation was performed in the presence of a-amanitin which blocks progress of engaged po[ymerases, A similar analysis identified an open sequence of only about 30 bases just downstream of the start site el the transthyretin (TTR) gene in nuclei isolated from mouse liver, a tissue where TTR is actively transcribed. This abrupt boundary of KMnO 4 sensitivity, which was removed by allowing engaged polymerases to elongate RNA chains suggests that most polymerases paused at about position +20. The possibility of mapping at the nuclsotide level the position of actively transcribing RNA polymsrases in whole cells or isolated nuclei opens new prospects in the study of transcription initiation and elongation. The gene for the cell wall glycine-rich protein OEP 1.8 of French bean is specifically expressed in young xylem elements and phloem fibers. A 205 bp promoter fragment of the GRP 1.8 gene directed vascular-specific expression of a reporter gene in transgenic tobacco. Deletions of the promoter to bp -186 resulted in loss of specific gene expression: Vascular expression was lower in transgenic plants with this deletion but the gene was also expressed in the cortex and other non-vascular tissue. Nuclear extracts from tobacco and tomato cell cultures and leaves were found to contain a protein binding close to the specificity element defined by the in vivo experiments. This protein binding site was~bp in length. A gene encoding the binding protein was cloned by screening a cDNA expression library of tomato. It encodes a basic leucine zipper (bZip) transcription factor. The 28 bp specificity element was able to activate the hetero!o~los 55S CsMV specifically in vascular cells.

Stefan Wieland; Rainer Lanz; Martin Hug; Manuela I-I6fferer; Sandro Rusconi; Institut fiir Molekularbiologie II der Universitdt, UZ Irchel, Winterthurerstrasse 190, 8057 Zfirich, Switzerland In the amino terminal portion of the rat glucocorticoid receptor (GR) cDNA there is a stretch of CAG triplets which encodes a string of Gin residues. This repeat is not phylogenetically conserved and is probably not very relevant for the function of the GR itself. Nevertheless, these repeats are frequently found in the coding portion of many transcription factors and it is conceivable that in some cases they may have acquired a significant role in the function of the host factor. Database searches suggest that the translational reading frame is subject to some constraints in natural sequences. We can show that in the rat GR there are different effects depending on the frame by which the repeated (CAG)I9 segment is translated (Poly-Gln, Poly-Ser or Poly-Ala). The N-terminal GR segment encoding poly-Gln, poly-Ser or Poly-Ala has been linked to other transcription factors (Octamer, GALA-and GAIA-VP16 chimeras). In these recombinant factors we consistently observed that Poly Alanine exerts the strongest intracistronie dominant negative effect. However, inter-cistronic dominance (trans dominance) is observed only when Poly-Ala bearing GR is used to challenge wild type GR and has not been so far observed in the chimeric trans-activators. We are extending our studies to other types of repeats, such as the Poly-(GlnAla) motif found in some yeast and Drosophila regulatory factors.

Department of Genetics and Microbiology, CMU, 121l Geneva 4 HLA class II deficiency is a regulatory defect affecting all HLA class II genes. At least three different complementation groups have been identified by cell fusion experiments. We present here the first identification of a gene which is defective in one complementation r the genetic complementation we constructed a eucaryotic cDNA expression system based on plasmid vectors carrying the EBV origin of replication. Size selected eDNA libraries of HLA class II positive cells were transfected into the class II negative mutant Burkitt lymphoma cell line RJ2.25. Transfected RJ2.25 cells were selected for re-expression of HLA DR molecules. Plasmid pools rescued from several independent sortings displayed the same preferential insert type. Retransfection of this CIITA ("class II transactivator") cDNA into RJ2.25 cells restored the expression of all HLA class II types RjDR, DQ, DP) to wild type levels. Analysis of the CIITA gene in 2.25 revealed an internal deletion spanning 1.8 kb of cDNA which leads to the absence of a full length CIITA messenger in RJ2.25. CIITA is a new gene without homology to known genes. Preliminary results indicate a crucial role for CIITA in HLA class I1 expression.

Siegrist, C.A., Reith, W. and Mach, B.,Department of Genetics and Microbiology, CMU, 1211 Geneva 4

Demonstration that regulatory factor RFX-I is identical to nuclear factor EF-C (Reith et al, see accompanying abstract) allowed to study the functional role of this transcription factor on the hepatitis B virus enhancer. First, B cells transfectants with the RFX-I binding site of the HLA-DRA promoter replaced by the inverted repeats of HBV EF-C sites upstream of a CAT reporter gene were shown to conserve the RFX-I dependant gammainterferon inducibility of the HLA-DRA promoter. Second, cotransfection of HBV-CAT with RFX-I in hepatic cells resulted in a dose dependant transactivation of the viral enhancer.Last, RFX-I antisense oligonucleotides were added to hepatic cells stably transfected by a plasmid coding for the HBV surface envelope under control of its own promoter-enhancer. RFX-I antisense oligomers were shown to inhibit the synthesis of the HBV envelope measured by HBsAg ELISA in the culture supernatant. These data demonstrate a new function for regulatory factor RFX-I.

Reith, W., Durand, B., Ucla, C., Siegrist, C.A. and Math, B., Department of Genetics and Microbiology, CMU, 1211 Geneva 4 RFX1, RFX2 and RFX3 are members of a novel family of DNA binding proteins binding to the X box motif of MHC class II gene promoters. RFX proteins show over 90% homology in their DNA binding and dimerization domains but are divergent elsewhere. They have identical DNA binding specificity and can bind as both homo-and heterodimers. RFX1 and RFX3 are expressed constitutively in most cells and tissues while expression of RFX2 is more restricted and variable in level. RFX2 and RFX3 are expressed as two alternatively spliced variants. We have now shown that RFX1 is identical to a nuclear complex called either EF-C (an HBV enhancer binding complex) or MDBP (a site-specific methylation-dependent DNA binding complex). This has allowed us to show that 1) RFX1 is a transactivator of the HBV enhancer as well as of MHC class II genes, and 2) RFX factors bind to certain sites only when these contain methylated CpG dinucleotides. This observation suggests that RFX proteins may also mediate the inhibitory effect of cytosine methylation on gene expressiOn. Human papillomaviruses (HPV) infect keratinocyte but not fibroblast cells. Several factors, including AP1 (Jun/Fos), contribute to the cell-type specific transcription of HPV genes. Here, we show that AP1 levels are low in early passage human fibroblast extracts. In contrast, human keratinocyte extracts contain high levels of AP1. In vine, Jan and Fos are relatively stable in human keratinocyte cells after serum induction whereas they are much less stable in human fibroblasts. Thus we think that the low level of API found in fibroblasts is due to its rapid breakdown. In vitro, AP1 is degraded by a proteinase in fibroblast extracts. The proteinase is blocked by an inhibitor of cysteine proteinases but not serine proteinases. Furthermore, the activity of a cathepsin B-like cysteine proteinase is elevated in these human fibroblast extracts relative to other cell types. Interestingly, when the fibroblast ceils are kept in culture for more than 20 passages they lose this proteinase activity and the amount of AP1 increases to a level similar to that of the keratinocytes. We conclude that the quantitative difference in AP1 proteins between human keratinocytes and early passage fibroblasts is due to the more rapid breakdown of AP1 proteins in the fibroblasts. This may be catalyzed in vine by a cathepsin Blike proteinase.

While the CpG dinucleotide is poorly represented in bulk mammalian DNA, there are interspersed sequences of a peculiar composition every 20-100 kb, so-called CpG islands, originally discovered by Adrian Bird. CpG islands are DNA segments of 0.5-2 kb length, with a high (G+C) content and high CpG density, often overlapping with the promoter-leader region of genes. Most CpG islands remain unmethylated at all stages of differentiation. The most remarkable exception is found in female mammals where the CpG islands of the inactivated X chromosome are cytosine-methylated, Aberrant methylation of a CpG island is also correlated with a form of mental retardation (fragile X syndrome). We have analyzed by computer the CpG islands era number of corresponding genes in mouse and man, including CpG islands of housekeeping and cell type-specific genes from many chromosomal locations, but avoiding members of repeated gene families. We extend the findings of Aissani and Bernardi (Gene 106, 1991,173-183; 185-195) and establish that the human CpG islands are generally longer and contain a higher CpG density than mouse CpG islands. It seems that the human CpG islands are the "better" CpG islands, which should allow the associated genes to be more stably maintained either in the active or in the repressed state, depending on CpG methylation. We also propose a mechanism for promoter methylation associated with fragile X syndrome. Four distinct oDNAs, xCTFI, xNFXI, xNFX2, xNFX3 were isolated from a Xenopus cell line library.

Sequence analyses revealed that all these cDNAs have a conserved N-terminal binding domain and variable C-terminal domains: xCTFI is the homologue of the human hCTFI subtype whereas the three distinct xNFXs share homology with the NF/ X subtype and are more likely derived from a common precursor by alternatively spliced mRNAs.

In vitro expression and DNA binding experiments show that the activity of the identical N-terminal DNA binding domain is negatively regulated by a novel type of C-terminal domain present only in the NFX proteins. This negative active Cterminal domain does not prevent the DNA binding of heterodimers made of xNFX and xCTFI.

However cotransfection experiments performed in Schneider cells show that all these four proteins could activate transcription even as full length products although they are very weak DNA binding proteins. Using a serie of deletion mutants we were able to confirm in vivo the localisation of this NFX " binding inhibitory region" together with the domains responsible for transcriptional activation.

Two genes encoding liver enriched transcription factors, lap and c/ebp, produce multiple polypeptides by differential initiation of translation at two in-phase AUGs. Lap has been shown to produce LAP, a 36 Kd transcriptional activator protein and LIP, a 21 Kd transcritional inhibitor protein (Descombes P. and U.S. Cell fiJ_ 5691991). Likewise, the c/ebp gene produces a 43 Kd transcriptional activator protein as well as a 30 Kd protein that activates transcription poorly or not at all. The different translation of two proteins from one LAP or C/EBP mRNA is conserved among vertebrates, including Xenopus, chicken, mouse, rat and man. LAP, LIP, C/EBP 43 and C/EBP 30 all accumulate to higher level in adult as compared to newborn rats. However. while the ratio of the two CfEBP proteins remains constant, the LAP/LIP ratio increases about five fold during this time period. As shown by immunostaining of tissue sections and western blot analysis of nuclear proteins, the developmental increase of LAP and C/EBP levels is due to both an increase in LAP -C/EBP expressing cells and a higher accumulation of these proteins in expressing cells. As LAP and C/EBP are maximally expressed in terminally differentiated, non-dividing hepatocytes, we examined whether these proteins may themselves be involved in the proliferation arrest of parenchymal liver cells. In contrast to differentiating adipocytes in which such a function has been documented for C/EBP (Umek. R. et a11991 Sciences 52A!. 288), neither C/EBP nor LAP appear to be sufficient for hepatic growth arrest,

Philippe Douville and Martin E. Schwab, Brain Research Institute, University of Zurich, August-Forel-Str. 1, 8029 Ztirich A variety of POU-box transcription factor genes have been identified in the mammalian nervous system and several of these have been shown to have crucial roles in cellular differentiation. We sought to assess the composition and to identify novel POU-box genes in developing rat CNS and differentiating oligodendrocytes. Our approach was to PCR amplify POU-box genes from postnatal day 0 (P0) rat spinal cord and P16 rat oligodendrocytes with degenerate primers from both the POUspecific and POU-homeodomains. We sequenced a sample of 23 clones from the PCR amplification of the oligodendrocyte template and identified a previously known POU-box, SCIP. No other POU-box genes were found in oligodendrocytes. Sequence information from a sample of 37 clones of the spinal cord template showed amplification of several POU-box genes including Oct-2 (51% of sample), Brn-4 (22%), Brn-3 (16%), and SCIP (2.5%). Some of these clones showed nucleic acid differences which could result in amino acid substitutions. We are currently assessing these differences using several assays to see if they represent novel POU-box gene products.

Abt. Zellbiologie, Biozentrum der Univ. Basel, 4056 Basel

We have identified by PCR and eDNA cloning five POU genes expressed during early embryogenesis of zebrafish. Four of these genes show extended homology to the brn-1 class of POU genes identified before in mammals. Northern blot analysis and in situ hybridizations indicate that the expression of these genes begins shortly after gastrulation in the neural tube. In the 24 hour old embryo various structures in the brain and in the spinal cord express these genes. We are currently investigating how the expression patterns of the four brn-l-like genes differ from each other and how far they overlap. A fifth gene analyzed contains a POU domain that forms the prototype for a novel subclass of these DNA binding domains. This particular gene seems to be expressed already during the first cleavages in addition to a large maternal RNA pool. In early gastrulae the transcript is predominantly distributed in a ring around the equator of the egg. Later during gastrnlation the expression domain is narrowing to two strong stripes and the posterior half of the rostrocaudal axis. After 24 hours the transcript can only be found in the tip of the tail. The early expression and the spatial arrangement of the transcript strongly argue that this POU geue is involved in early developmental decisions determining the body plan of the zebrafish.

Schubart, D., Abbey, NL., Hoffman, M. and Matthias P. Friedrich Miescher-lnstitut, 4002 Basel

To identify cDNAs coding for new POU proteins we designed degenerate amplimers for PCR amplification of DNA segments encoding POU domains. We were particularly interested in identifying new T cell-or brain-specific POU proteins. We prepared cDNA from activated Jurkat T-ceUs and used it as template for PCR reactions. The resulting PCR products were cloned and analysed by sequencing. This approach allowed us to identify a new clone having a very high homology to rat Bm-3 (a POU protein for which only a small part of the POU domain was known so far). Until now expression of that gene had not been reported in T cells. By screening a Jurkat cDNA library we then isolated overlapping clones covering most of the cDNA. We are presently further characterising these clones and a functional analysis will be presented. Metallothioneins (MTs) are small cysteine-rich proteins whose structure is conserved frOm fungi to man. MTs strongly bind heavy metals, notably zinc, copper and cadmium. Upon cellular exposure to heavy metal and other adverse treatments, MT gene transcription is strongly enhanced. Metal induction is mediated by several copies of a 15 bp consensus sequence (metal-responsive-element = MRE) present in the promoter region of MT genes. We and others have demonstrated the presence of an MRE-binding factor in HeLa cell nuclear extracts. This factor, termed MTF-I (MRE-binding transcription factor) is inactivated/reactivated by zinc withdrawal/zinc addition. Here we report the cloning of the eDNA of mouse MTF-1, a 72.5 kD protein. MTF.1 contains 6 zinc fingers (C2H2) and separate transcriptional activation domains with high contents of acidic and proline residues. Ectopic expression of MTF-1 in primate or rodent cells strongly enhances transcription of a reporter gene that is driven by 4 consensus MREd sites, or by the complete mouse MT-I promoter, even at normal zinc levels. The concentration of the factor is unchanged by metal treatment, which is in accordance with previous findings that transcriptional induction does not depend on de hove protein synthesis. Indications that other factors than those belonging to the HLH family (MyoD, Myogenin, Myf5, MRF4) might be involved in vertebrate myogenesis made us search for homeobox-containing proteins in human muscle tissues. Using degenerated primers several homeobox-containing sequences belonging to the POU-domain family were found. One of them, called BDOM130, which shows a unique linker sequence was further analyzed. Its 3'end and at least part of the 5'end have been cloned using the RACE protocoll. We now have a eDNA that contains an open reading frame of 906 bp giving rise to a protein product with an apparent weight of 40 kD. The expression pattern of BDOM130 in adult human tissues and during mouse development has been studied by Northern blot analysis and in situ hybridization, respectively. The bacterially produced fusion protein GST-BDOM130 binds the octamer motif (ATTTGCAT) to which all POU-domain proteins show some affinity. We therefore conclude to have found a new member of the POU-domain family constituting a novel subclass. Currently we are investigating the influence of BDOM130 on the transcription of reporter gene constructs.

Motejlek K., Leitgeb S., Rtllicke T., l-I,~uselmann R. and B. Lfischer Abteilang fOr Molekalare Nearobiologie, Pharmakologisches Institat der Universitat Zfirich, Gloriastrasse 32, CH-8006 ZOrich Gene promoters driving differential neuron-specific expression of three ~taminobutyric acidA (GABAA) receptor subanits have been isolated and mapped. Analysis of the otl gene indicates that >80% of its mRNA is Wanscribed from a single promoter that features a TATA box. The 5' end of the "[2 subunit gene is A/T-rich and the major start site also has a TATA-llke element. In contrast, the 5' end of the ~ subunit gene shows paradoxically all the features known to be typical of so-called "house-keeping" genes: a cluster of initiation sites in the center of a CpG-rich island, a prototype initiator sequence and no TATA box. Thus, GABAA receptor genes show significant divergence in their promoter slructure. Nevertheless, comparison of the 5' flanking sequences of these three GABA A receptor genes revealed a novel conserved sequence element. A neuron-specific DNA-binding protein has been identified, that binds to this element and may be important for cell type-specific expression of GABA A receptor genes. Experiments to further characterize and purify this protein are under way. To determine functional elements for neuron type-specific expression we have generated several wansgenic mouse lines carrying a laeZ gene driven by ~ subunit gene sequences. 5' flanking sequences allow faithful neuron-specifc expression in hippr~campus, thalamie nuclei, and in the cerebral cortex, bat not in the cerebellar granule cells. Data on transgenie mice designed to test the function(s) of downstream sequences will be presented.

Erne H., K~ry P. and Monard D. Friedrich Miescher-Institute, Basle, Switzerland.

Through analysis of rat genomic clones the three first exons and the promoter of rat glia derived nexin (GDN) were identified. The promoter sequence and the first exon were found to be extremely GC-rich thus localizing the 5' part of the rat GDN gene within a GC-island. A TATA box like sequence (TGATAAA) but no CAAT box was found. Primer extension and RNase protection assays identified one transcriptional start site. A 1600 bpputative promoter fragment cloned in a reporter plasmid was shown to induce firefly luciferase expression after transient transfection in COS-7 cells and several other cell-lines. GDN gene expression in C6 glioma cells and Rat-t fibroblasts was found to be regulated by a number of GC binding transcription factors, among which are Spt and a family of early growth response transcription factors (NGFI-A, Krox 20, Egr3, WTt and NGFI-C) which have been shown to be important regulators of gene expression during development.

A negative regulator was found to bind to the GDN promoter sequence between -377 and -238. This factor recognizes the E box consensus site (CAXXTG). In cultures of primary Schwann cells GDN mRNA increases with time while the amount of the negative regulator decreases. This suggests that this factor is regulating GDN expression in Schwann cells.

Schwarzenbach, H., Newell, J. and Matthias P. Friedrich

Miescher-lnstitut, 4002 Basel Some immunoglobulin variable region promoters contain, in addition to the highly conserved octamer motif, a pyrimidine-dch sequence:

-CTTCCTTA-, or a variation of it. This motif has previously been shown to be important for full promoter activity, in particular for those promoters which have only an imperfect octamer site. We have analysed the factors binding specifically to this pyrimidine-rich motif, By electrophoretic mobility shift assays with extracts from B cells at least two sets of complexes can be detected. One represents the binding of a B-cell specific factor of the ETS family. The other is due to the binding of a ubiquitous, but lymphoid cells-enriched factor, The interplay between these factors and octamer factors is being studied and a functional analysis will be presented. Most studies on the regulation of gene expression in human papillomaviruses have focussed on the promoter for the early genes E6 and E7 located in the long control region (LCR). We are interested in knowing whether or not there are other promoters outside the LCR. By in vitro transcription and primer extension analysis we found promoter activities near the start of the E2 gene and at the end of the L2 ORF in the genome of HPV-18. These promoters are active in HeLa cells as shown by CAT assay. Co-transfection experiments with a BPV-1 E2-expressing vector and reporter plasmids containing HPV-18 DNA fragments with the promoter activities showed that the promoter located within the E2 ORF is transactivated by E2 even in the absence of binding sites for the E2 protein. Transactivation was also observed when a plasmid expressing only the transactivation domain and not the DNA-binding domain of E2 was used. This suggests that E2 mediates transactivation through binding to another factor. Spl being a potential target for E2, we investigated the role of a putative Spl site (AGGTGG) in the promoter activity. Deletion of this site completely abolished transcription. Bandshift analysis showed a weak binding of Spl to the AGGTGG site. These results were in agreement with the model in which the E2 protein is targeted to the promoter region by physical interaction with Spl The trans-active factors regulating IL-2 expression have been analyzed by bandshift assays and functional transcription studies in the Xenopus oocyte, In human primary T lymphocytes, IL-2 regulation turns out to be strikingly different from the models elaborated using immortalized T cell lines: T cells in different states of maturity were isolated through immune-labeling of their surface antigens and cell sorting on FACS. In resting naive T helper cells, the IL-2 gene is repressed by a silencer binding to the Pu-rich promoter elements. Upon first antigenic stimulation, an activator appears, displaces the silencer, and derepresses IL-2 transcription. Post-stimulated, rested memory cells do not acquire a silencer, but their activator looses its functional activity and is trans-Iocated to the cytoplasm. New stimulation of such memory cells leads again to nuclear targeting of the activator and to its reactivation without requiring de novo protein synthesis. By contrast, immortalized tumor cells, such as Jurkat, have no silencer and their activator (NF-AT) is not capable of de-repressing genes arrested with silencer from primary T cells. Tumor cells thus may not be used as a model to study IL-2 regulation in vivo.

The promoters of the ctl(VI) and cr (VI) collagen genes exhibit features characteristic of housekeeping genes and of proto-oncogenes: They have a high G/C content and lack a typical TATAA box. Transcription is initiated at multiple start sites. Previously, we have analyzed the cd and c~2 promoter by footprinting analyses and found multiple DNA elements interacting with nuclear factors. Now we have further characterized the two promoters by gel retardation assays. The ul promoter contains three binding sites for nuclear proteins showing the following pattern: AP1 -Spl -Spl. The second Spl site is located just upstream of the major transcription initiation site. The cr promoter exhibits four binding sites for nuclear factors: Spl -Spl -X -Spl (X stands for a novel, not further characterized factor binding specifically to a DNA element of the ft.2 promoter). Again, the last Spl site is located just upstream of the major transcription initiation site. To analyze the biological relevance of these factors we constructed artificial mini-promoters utilizing the DNA elements of the cd and ct2 promoter recognized by transcription factors. These constructs were tested for their promoter activity in transient expression vector assays. We found that the promoter constructs had an activity comparable to that of the corresponding promoter fragments. Steroid 17ct-hydroxylase cytochrome P450, the product of the CYP17 gene, is required for steroid hormone biosynthesis in the adrenal cortex and the gonads. CYP17 gene expression is restricted to these steroidogenic tissues and regulated by pituitary peptide hormones via cAMP. Transcription of the bovine CYP17 gene involves at least two distinct cAMP-responsive sequence elements (CRS) which are located in its Y-flanking sequence (Lurid et al., JBC 265:3304) . Our studies on the distal element CRS1 show that it activates reporter gene transcription only in steroidogenic cells suggesting that it may play a dual role, namely in cAMP-inducible as well as in cen-specific gene transcription. By in vitro transcription and other evidence we demonstrate that it interacts with a factor which is unrelated to the cAMP response element (CRE) binding protein CREB. We are presently isolating CRS1 binding proteins from bovine adrenal cortex. The proximal element CRS2 apparently interacts with several different factors one of which is only detectable in steroidogenic cell-types. Mutational analysis is being performed to elucidate their sites of interaction as well as their functional significance for cAMP-dependent and cell-specific transcription. UZ Irchel, Winterthurerstrasse 190, 8057 Ziirich, Switzerland One of the Ilc residues in the second zinc finger is highly conserved among the members of the steroid receptor superfamily. We have analyzed the effect of substitution by every possible aminoacid at the corresponding position of the rat glucocorticoid receptor (rGR map position 484). The results indicate that only hydrophobic aminoacids can functionally substitute for I1e484. Interestingly, (and opposite to wild type GR and other mutations within the zinc finger region) the mutant Cys484 fails to transactivate when tested as a small GR fragment 407-556 but is permissive for transactivation when extended to include the entire Nterminus. Furthermore, the transactivation of the Cys-substitution of I1e484 is similar to the corresponding GR fragment lacking the putative transactivation domain tau-2 located C-terminally to the second zinc finger. Mobility shift and in vivo trans-footprint assays confirmed the DNA binding properties of the mutants. Crystallographic analysis by Luisi (ct al., Nature 352) of the DBD has revealed a hydrophobic core involving also the ultra conserved residue Cys500. Mutation of this residue to Ala yields in a fully functional receptor while its substitution by Phe destroys receptor functions. We have tried to construct compensatory mutants by simultaneous substitution of Cys500 by Phe and Phe464 by Ala.

Louvion, J.-F., Ghika A., Fankhauser, C., P. and Picard, D. D~partement de Biologie CelIulaire, Universit~ de Gen~ve, CH-1211 Gen~ve 4

The hormone binding domain (HBD) of steroid receptors is an autonomous regulatory cassette which can subject activities of heterologous proteins to hormonal control probably as an effect of HSP90 binding to this portion of steroid receptors in the absence of ligand. We now show that heterologous proteins can also be regulated by HBDs in yeast. We fused the estrogen receptor HBD to a GAL4-VP16 activator and found that transcriptional activity of the chimera is completely estrogen dependent. This provides an efficient tool to express proteins of interest in yeast in a regulated fashion. We are currently exploring the possible role of HSP82 (HSP90 yeast homologue) in this regulation. The unliganded HBD acts as a negative regulator of various steroid receptor functions including nuclear localization signals (NLS). While the HBD of the glucocorticoid receptor (GR) contains a NLS, the HBD of the mineralocorticoid receptor (MR) does not. To further define this hormone-regulated function, the NLS activity of a set of chimeric GR-MR HBDs, fused to 8-galactosidase as a tag, was tested.

We investigated the role of a tandem binding site for the ubiquitous oct-1/NF-III factor, located upstream of the TATA box and adjacent to a CTF/NF-I site in the MMTV promoter. In mouse L cells stably transfected with a mutant of both oct-1 sites we observed a 50-fold reduction of the base level of transcription; the glucocorticoidstimulated level was unaffected. Binding of NF-I and oct-1 on wildtype and mutant DNA was studied by DNAse I footprinting with nuclear extracts and oligonucleotide competition experiments. We conclude that in L cells oct-1 is mainly involved in basal promoter activity, and NF-1 in glucocorticoid induction. In the distal promoter region, we analyzed DNA-binding nuclear proteins from mouse tissues (spleen, permissive for MMTV expression, and liver, nonpermissive) by DNAse I footprinting, methylation interference, gel retardation and oligonucleotide competition, and UV-crosslinking. 5' of the glucocorticoid receptor binding site, a sequence (DRa, -189 to -206) bound tissue-specific factors. Next to it, a sequence (DRc, -206 to -223) bound a protein present with different abundance in the tissues. We are presently testing in transfection assays if these sequences can modify the level of transcription from the MMTV promoter. The nuclear hormone receptors PPARs (peroxisome proliferatot-activated receptors) control the peroxisomal ~-oxidation of fatty acids by induction of the acylwCoA oxidase gene which encodes the rate limiting enzyme of the pathway. Gel retardation and cotransfectlon assays revealed that PPAR~ heterodimerizes with retinoid X receptor ~ (RXR~) and that the two receptors cooperete for the activation of the acyl-CoA oxidase gene promoter. The strongest stimulation of this promoter was obtained when both receptors were exposed simultaneously to their cognate activators. Furthermore, we show that natural fatty acids, and especially polyunsaturated fatty acids, activate PPARs as potently as does the hypolipidemic drug Wy 14'643, the most effective activator known so far. Moreover, we discovered that the synthetic arachidonic acid analogue 5,8,11,14eioosatetraynoie acid (ETYA) is 100 times more effective in the activation of PPARd than Wy 14'643. In conclusion, our data demonstrate a convergence of the PPAR and RXR signalling pathways in the regulation of the peroxisomal fatty acid ~-oxidation system by fatty acids and retinoids. Rainer B. Lanz and Sandro Rusconi; lnstitutfiir Molekularbiologie II der Universit~t UZI, Winterthurerstrasse 190, 8057 Ziirich, Switzerland. We have studied the behavior of several mutations in the glucocorticoid receptor (GR) by expressing GR-eDNAs along with appropriate reporter genes in mammalian cells. Here we report experiments involving mutations in the hormone binding domain (HBD) of the receptor. Others have shown that mutageaesis in the most carboxy terminal portion of the GR can abolish ligand dependent transcriptional activation without apparent effect on steroid or DNA binding (Danielian et al., 1992) . On the other hand, a progesterone receptor (PR) deletion mutant lacking the earboxy terminal 42 aa has been reported to be not responsive to progesterone, but efficiently activated by the potent antiprogestin RU486 (Vcgcto ct al., 1992) . We reconstructed these earhoxy terminal mutations by either substitution (GR aa 770, 771) or by deletion of the C-terminal 29 na of the HBD and tested them for response to agonist/antagonist and for trans-dominance. We found that the substitution mutants have a differential response to dcxamethasone/RU486, while the truncated form failed to react in the way anticipated from work with PR. Moreover, deletion of two residues -which are conserved in the GR/PR/MR/AR-cIass of the steroid receptor superfamily-, generated a mutant which can respond positively to RU486, but not to the agonist. None of these GR mutants appears to have trans-dominant negative properties. Further C-terminal deletions have been made and tested for responsiveness to RU486. The Ah receptor as a bHLH transcription factor binds a variety of environmentally important carcinogens such as chlorinated polycyclic aromatic hydrocarbons (eg. TCDD) as well as polychlorinated biphenyls (PCBs) and activates transcription of several genes involved in xenobiotic metabolism, such as cytochrome P450 CYPL41 and CYPL42, as well as ghicuronyl-transferase (GT) and ghitathion-S-transferase (GST). The unliganded receptor is generally found in cy~osol as a complex of the ligand binding subunit (LBSI, heat shock protein 90 (HSP90) and other as yet unidentified proteins. After ligand binding, a dimerie complex of LBS and the Ah receptor nuclear transloeator (ARNT) protein translocates into the nucleus and binds to eis-aeting regulatory sequences, termed Xenobiotie Responsive Elements (XRE), which control transcription of the CYP/A1 gene by the Iiganded Ah-receptor. We investigated the role of ARNT in mediating CYP1A1 transcription in the mouse hepatoma cell line Hepa-1 and provide evidence for a TCDD dependent dirnerization of Am, rr with LBS after release of HsPg0, nuclear translocation and direct interaction of both subudits of the heterodimer with the 6 bp XRE core sequence. Functional domains within the ARNT molecule are presemIy being investigated by means of site directed mutagenesis. We have cloned a cDNA sequence of Chironomus coding for a protein homologous to the Drosophila ecdysteroid receptor. Its N-terminal domain, the hinge domain, and the hormone binding domain were separately overexpressed in E. coil The resulting polypeptides were injected into rabbits after gel purification for antisera production. An analysis of the antisera by Western blotting revealed the presence of antibodies specific for each of the protein domains. On Western blots of protein extracts of Chironomus tissue culture ceils, several protein bands were detected. One band of an approximate molecular weight of 65 kDa was detected with each of the three antisera. This corresponds with the size of the protein predicted by the amino acid composition. Using these sera, immunohistoehemistry of squash preparations of Chironomus prepupal salivary gland polytene chromosomes was performed. About ten bands showing strong immunofluorescent signals, including the known early ecdysteroid inducible puff sites 1-18C and IV-2B, were detected. These results agree with the notion that the protein against which the antisera were raised represents an ecdysteroid receptor.

Zygotic transcription in Ascaris lumbricoides A. Spicher, H. Tobler and F. Mueller, Institute of Zoology, University of Fribourg, P6rolles, CH-1700 Fribourg The onset of zygotic transcription is an interesting and challenging problem in developmental biology. Organisms differ in the timing and abruptness of this process. In sea urchins, both pronuclei may already transcribe before fusion, and the zygotic nuclei are clearly transcriptionaly active. Transcription in mammals, ascidians, gastropods and molluscs, however, only starts within one or two cell divisions after fertilization. Xenopus and Drosophila, in contrast, appear to use maternal gene products until the mid-blastula transition, when a marked increase of zygotic transcription occurs.

Here we describe a gene (Fert-1) of the nematode Ascaris lumbricoides whose transcription begins just after the end of meiosis of the female pronucleus. Fert-1 generates different polyA+ and polyA-transcripts. One of the polyA+ transcript is 560nt long, composed of two exons and transspliced at its 5'end. In contrast to the early zygotic activity of Fert-l, general zygotic transcription is switched on only at the 4-cell stage embryo.

Hihi, A., Keller, H.-J., Dreyer, C. and Wahii, W. Institut de Biologi% Animale,Universit6 de Lausanne, CH-1015 LAUSANNE Max Planck Institut fir Entwicklungsbiologie, D-7400 TOBINGEN A cDNA has been isolated which codes for the Xenopus homolog of the Drosophila FTZ-F1 protein (dFTZ-FI) and the Mouse ELP protein. FTZ-Fl and ELP are members of the nuclear hormone receptor superfamily. The Xenopus FTZ-FI (xFTZ-FI) was overproduced using the Vaceinia Virus expression system, and production of xFTZ-FI was assayed by gel retardation analysis.As probe we used the dPTZ-FI response element (Ftz-RE) which is located in the Zebra element of the Fushi-tarazu (FTZ) gene promoter, and contains the core AGGTCG sequence.A specific bandshift was only observed with extracts of cells infected with the recombinant virus containing the xFTZ-FI cDNA, and we did not detect binding to an Estrogen Response Element,containg a core AGGTCA sequence.We observed a less efficient binding to a newly described Hormone Receptor Response Element(RE) containing the AGGTCA sequence, the Peroxisome Proliferator Activated Receptor RE .Finally the results suggest that the recombinant Vaceinia virus produces the xFTZ-FI protein which is able to bind to the Ftz-RE. A 47 kd human autoantigen called the La protein transiently binds to the 3' end of nascent RNA polymerase III (Pol III) transcripts and has been implicated in their synthesis, termination and release. Using cloned reagents from Xenopus laevis (La cDNAs, proteins, and polyelonal antibodies), we have examined the effects of La depletion of cell-free extracts on the stability of pre-synthesized RNA, and on the efficiency of Pol III transcription under conditions that distinguish single from multiple rounds of synthesis. The results demonstrate that La is able to protect nascent Pol III transcripts from 3' exonucleolytic attack, that it can slow the rate of normal 3' processing of tRNA precursors, and they suggest that it is not required either to complete the synthesis of the primary Psi III transcripts or to act as a release factor. The steroid hormone 20-OH-ecdysone triggers coordinate changes in Chironomus larval development that result in metamorphosis to the adult insect. In an effort to understand the molecular mechanisms of this regulatory network we set out to clone candidate genes that specify proteins of the steroid hormone receptor superfamily. A eDNA library that was generated from rnRNA isolated from a developmental stage of larvae (prepupae) of a high ecdysteroid titre was screened with an oligonucleotide probe specific for a conserved region encoding part of the zinc finger DNA binding domain of steroid hormone receptors. Beside a clone carrying the coding sequences for a putative ecdysone receptor (eEcR; M. O. Imhof, Ph.D. thesis ETH No. 9847) several other clones that appear to specify members of the supeffamily were detected. Amino acid sequence comparison of the identified ORFs suggests that these clones code for the Chironomus homologs of the Drosophila ultraspiracle (usp), E75, and DHR3. While the temporal developmental profile for the expression of all three Drosophila genes closely parallels that for the ecdysteroid titre, E75 and DHR3 were additionally mapped to loci of eedysteroid-inducible puffs in polytene chromosomes. The developmental expression pattern, the response to ecdysteroids, and the chromosomal localization of the various cloned ChironomuS genes are discussed. 

Standke, G., Groner, B., Friedrich Miescher-lnstitut, 0H-4002 Basel Differentiated mammary epithelial cells express high levels of milk proteins. 90% of the total milk protein comprises almost six tissue specific proteins, the caseins (o~, IB,'y); c~-Iactalbumin; tB-lactoglobulin and whey acidic protein (1}. Induced by lactogenic hormones cell cultures of mammary epithelial cells, HC11, express milk proteins in a transcriptional regulated manner (2). The milk protein genes are highly divergent except for a few consensus regions in the promoter regions of the genes. It is reasonable to assume that the milk protein genes are subjected to common regulatory mechanisms and that these regulatory signals are conferred by elements of the promoter region. These promoter regions are recognized by nuclear proteins from mammary epithelial cells. Biochemical and functional studies are carried out to identify the role of common sequence elements and their interaction with individual transcription factors in the hormone, tissue and differentiation specific regulation of milk protein gene transcription.

(1)Clark, A,J., 1992 , J.OelI.Biochem.49, 121-127 (2)Schmitt-Ney,M et al., 1992 76 LI-FRAUMENI SYNDROME AND TP53 GERMLINE MUTATIONS Scott, R., Mary, J.~L., Weber, W., Spycher, Departement Forschung, Kantonsspital Basel, CH 4051 There exist a wide spectrum of diseases encompassing soft tissue sarcoma, brain cancer, adrenocortical carcinoma and leukaemia which segregate in rare families. This association of neoplasias initially i described by Li and Fraumeni in 1969 is also atypical as the onset of malignancy occurs at an unusually young age. Recently, the underlying genetic mutation giving rise to this association of neoplasias has been shown to be mutations in the turnout suppressor gene0 TP53. The usual function of tumour suppressor genes is to control normal cell proliferation, however, when this control is disrupted (for instance, by carcinogens or radiation) the development of neoplasias appears to be inevitable. We identify a family, where two brothers have developed different types of malignancy yet carry the same germline mutation. The identification of germline p53 mutations has important implications for the handling of patients belonging to the Li-Fraumeni syndrome. Due to the different approaches towards handling the two patients who carry the same germline p53 mutation, very different and unpredictable outcom~ may occur.

Werlen, G., Belin, D., Conne, B., Roche, E., Lew, D. P. and Prentki, M., Faculty of Medicine, University of Geneva, CH-1211 Geneva 4. Various Ca 2+ agonists, including growth factors and Ca2+-ionophores cause rapid induction of early response genes such as c-fos and z/f268. To gain more direct insight into the action of Ca 2 § on transcriptional regulation, we have developed an intact cell model in which cytosolic free Ca 2+ concentration can be measured and fixed at any level for various times in parallel with the assessment of early gene expression. Using fura-2 loaded HL-6O cells, we have observed maximal c-fos and z/f268 mRNA accumulation at -200 nM Ca 2+. In addition, full transcript accumulation of both genes measured at 30 min was obtained with a Ca 2+ perturbation of one minute. Run-on transcription assays indicated a small effect of Ca 2+ on transcriptional initiation and a pronounced Ca2+-modulated relief of a block to transcriptional elongation in intron 1 of c-fos. This extreme sensitivity to Ca 2+ in term of both time and dose of Ca 2+ required for full gene induction, demonstrates that Ca 2+ is a major regulator in its own right of early response gene expression. The results in addition confirm the presence of an intragenic Ca 2+ response element in the c-fos gene. We are currently investigating the signal transductioa pathway implicated in the Ca2+-regulated c-fos gene expression. Preliminary experiments indicate that the C-kinase and the Ca2+/ calmodulin pathways do not mediate the action of Ca 2+ on c-fos.

CHARACTERISATION OF P55 DELETIONS IN COLON CARCINOMAS. Billotte, J., Voutravers, P. and Show, P., Jnstitut universitdre de pathologic, division d'oncologie e• 1011 Lausanne. P55 mutations are one of the most common genetic Jes]ons appearing in virtually every type of human cancer. Both alleles of p55 are altered in 80% of colon carcinomas, one allele being usually mutated (missense point mutation) while the otherone is completely deleted. Such observations suggest that mutations in Lhe p55 gene ore recessive to the wild type ulrele. ]he ex~en~ and mechanism of detetion has not ye~ been characterized. We have established a detailed restriction map of the 17p chromosomal region encompassing the p53 gene. This was accomplished by analyzing a YAC (Yeast Artificial Chromosomel con~aining ~he p55 gene. This done contains o human insert of 620 Kb. Lambda BASH II was used to subclone the haman sequences, and these subdones were used to screen the presence of microsatellites surrounding the p55 gene. in an average of 80% of the cases the microsaM[ites ore heterozygotes and offer therefore a very powerful tool to screen primary tumors for deMions. Analysis of the p55 chromosomal region will be presented and discussed. The presence and extent of deletions of p~,3 locus in colon carcinomas will be discussed in view of experiments using m]crosateil]~es dnalys]s, as well as results obtained wi~h more ~radi~ionat probes derived from the p55 chromosomal region. The T1 mRNA was initially found after expression of the oncoproteins Ras or Mos in mouse fibroblasts (NIH 3T3). Transcription of T1 can also be stimulated by serum growth factors. Peak levels of T1 mRNA are reached within 4 to 6 hrs following serum stimulation. The induction depends on protein synthesis. We identified an 80 bp long sequence, 3.5 kb upstream of the transcription initiation site, which mediates stimulation of T1 expression. It contains one TRE (TTAGTCA) and two E-Boxes (CACATG). This 80 bp fragment was cloned in front of a TK-minimal-promoter CAT-construct (pBLcat2). The 80 bp fragment strongly stimulates transcription in transient transfection assays. Point mutations were introduced into the TRE or into one or both E-boxes. All mutated forms showed a significantly decreased activation of transcription in transient transtection assays. We stably transfected NIH 3T3 with an expression plasrnid directing the synthesis of a fusion protein consisting of FosB and the estrogen receptor. In these cells T1 expression can be stimulated by the addition of estrogen. To test the effect of c-Myc on the T1 promoter, we stably transfected NIH 3T3 cells with expression plasmids coding for a fusion protein consisting of the estrogen receptor fused either to c-Myc (tranfectants are called NME) or fused to a dominant negative mutant of c-Myc (NMAE). Addition of estrogen stimulates T1 transcription in NME and represses the expression of T1 in NMAE. The human genome contains large regions that are highly structured.

Sequence related members of multigene families are often found in a clustered organization. Here we describe a new gene cluster composed of genes coding for calcium binding proteins of the $100 family. The linkage of six genes was established by pulse field gel electrophoresis, and a contigous DNA sequence of 15 kb contains the full coding regions of four different $100 genes. This is the tightest gene cluster discovered so far. Two novel $100 genes are located within the cluster which both exhibit unique structural features when compared to other $100 genes. St00E is cysteine-rieh, whereas STOOD contains a long hydrophobio N-terminal tail. The gene cluster was assigned to chromosome 1q21, one of the bands showing rearrangements in neoplasms at high frequency. The deregulated expression of some $100 genes in the cluster during tumor progression suggests that chromosomal abnormalities may influence the expression of $100 genes in 9 late stages of cancer, particularly associated with the formation of metastases. The human calmodulin (CAM) gene family consists of three members coding for an identical protein. In addition, at least one intronless gene for a CaM-like protein (CLP) and numerous pseudogenes exist in the human genome. The complete structures of all human CaM genes have been elucidated by a combination cf genomic screening and walking techniques as well as by direct PCR-based amplification of genomic DNA. Aft ~ntron focatfons are precisely conserved in the human genes; in contrast, the rat CaM[I gene appears to have lost intron 3. The hCaMII gene can be distinguished from all other CaM genes by its over 50% larger size (due mainly to an expanded intron 2). The putative promoter regions are highly divergent in the three CaM genes indicating separate mechanisms of regulation. The genes for CaMI, II, III and for CLP are widely dispersed in the genome; they have been mapped to human chromosomes 14, 2, 19 and 10, respectively, by a combination of Southern blotting, PCR and in situ hybridization methodology. The high level of occurrence of point mutations in plasmid-cloned PCR products from the mitochondrial cytochrome B gene of several birds led us to ask what proportion of these mutations was due to misincorporation by the ~ polymerase during the PCR process versus heterogeneity in the DNA template. We therefore cloned and sequenced 20 copies of a 0.7kb CyIB sequence amplified from a single clone in order to estimate the Taa polymerase error rate. We found a significantl~ lower level o1 point mutations, yielding an empirical rate of 2.2 xl0 b errors per nucleotide polymerized by the ~ polymerase. We concluded that up to 74% of the original variation observed was due to heteroplasmy of the mtDNA template. This microheteroplasmy may arise by somatic mutation due to a lower fidelity of the mtDNA polymerase or because of a high number of duplications of the mitochondrial genome. These results show that the sequencing of cloned mtDNA can reveal a significant amount of variation which is not phylogenetically informative and which could have a significant effect in certain types of studies.

Lindberg R L P, Grandchamp B, Percher (7, Ledermann B, BOrki K, and Tetrahydrobiopterin (BH 4) is the essential cotactor for the hepatic phenylalanine hydroxylase. A cofactor deficiency causes malignant hyperphenylalaninemia, an autosomal recessive disease. BH 4 is synthesized by at least three consecutive enzymatic steps, and most of the patients suffering from BH4 deficiency have been shown to bear a defect in the second enzymatic step, the 6-pyruvoyltet rahyd r opt erin syathase (PTr~). As a first step towards the understanding of the molecular nature of the deficiencies, we h~ve cloned the human liver cDNA encoding the PTPS, The coding region predicts a polypeptide of 145 amine acids with a 82% identity when compared to the rat liver enzyme. Expression of the cDNA in E. coil fielded the active enzyme. Fibroblasts originating from three different patients exhibiting a enzymatic deficiency in PTPS activity were the source for total RNA isolation and subsequent eDNA production. PCR amplification and DNA sequence analysis of the PTPS cDNA yielded three different types of mutations in these patients: (1) a C to A and an A to G transition, causing a change in Arg25 to Gin, and Tyr128 to Cys, respectively; (2) a triplet deletion resulting in Va157 deletion; and (3) a deletion of 23 nucleotides causing a truncated protein of 55 amino acids due to a frame shift mutation. Investigations are under way to confirm whether these alterations are the cause for the reduced enzymatic activity, and whether the mutations are homozygous m the different patients. Rapid isolation of highly purified enzyme was achieved by overexpressing PTPS in E. coli as a fusion protein with MalE. Subsequent affinity chromatography purification followed by enzymatic cleavage and size fractionation to separate the PTPS from the fusion partner yielded the enzyme in mg quantities, This will allow us to investigate the biophysical and structural properties of the protein.

WITH TERMINAL LIVER FAILURE Schnelder-Yin, X., Hungereeker, G,, Seh/~fer, B., Burg. G. and Minder, E. I. Institute of Clinical Chemistry, University Hospital, CH-8091 Zurich Human erythropoietic protoporphyria {EPP} is an inherited metabolic disorder and characterized bioehemically by accumulation of protoporphyrin as a result of deficiency in ferroehelatase activity. A common clirdeal symptom is photosensitivity. However, in rare instance, EPP patients develop terminal liver failure. We report here the molecular defect in a EPP patient who had severe liver disease and was treated by liver transplantation. Sequencing of ferroehelatase cDNAs from this patient revealed a single point mutation, C to T at nueleotide 185, This nonsense mutation converted gtutaminv'to a stop coden near the amino-terminus of the enzyme, Q-59stop eodon. The patient was heterozygous since a normal ferrochelatase sequence was also identified besides the mutated one. We investigated two other EPP patients without liver involvement. The molecular defect in the patient with liver failure was different from the mutations in those two patients and in 3 published EPP eases, who were all lack of hepatic complication. We conclude that EPP is heterogenous on molecular level since 6 different mutations have been identified so far. Further studies are to be done to investigate ff there is any correlation between the molecular defect and the severity of the disease, as it has been established in cystic fibrosis. Oppliger, E., Wermuth, B. and *Liechti-Gallati, S, Chemisches Zentrallabor, Inselspital and *Institut for Reehtsmedizin, Universit~t Bern, CH-3010 Bern Ornithine transcarbamylase (OTC) deficiency, the most common inborn error of the urea cycle, shows X-linked inheritance with frequent new mutations. Symptoms range from lethal neonatal hyperammonemia to cases with mild clinical symptoms. Recently. Hata and coworkers reported a G to T mutation in the OTC gene leading to the substitution of Trp for Arg in two Japanese boys with mild hyperammonemia (Hum. Genet. 87, 28-32, 1991) . Using the polymerase chain reaction followed by sequence analysis, we screened 19 male patients with OTC deficiency for the presence of the E to T base exchange. The mutation was present in four apparently unrelated patients from Austria, Italy (2) and Turkey. Liver biopsy was performed on three patients. In agreement with the findings from the Japanese patients, two liver samples exhibited residual engyme activities of about 20% of control values and higher enzyme activity at pH 9.5 than at the usual pH optimum of 7.7. OTC activity in the liver sample of the third patient was 6% of control values and no shift in pH optimum was detectable. Chronic granulomatous diseases (CGD) are rare immunodeficiencies characterized by a predisposition to recurrent bacterial and fungal infections. The underlying cause are defects in phagocytic NADPH oxidase, leading to absence or malfunction of microbicidal killing. Close to 70% of all CGD patients suffers from the X-linked form of the disease resulting from mutations in the gene encoding the gp91-phox subunit of the flavocytochrome b558, the terminal redox center of the oxidase. Gp91phox mRNAs from two related patients (uncle and nephew) were analyzed by RNA PCR. Using overlapping primer pairs sequence parts in between primers at sequence positions 179 and 816 respectively were shown to be missing in both patients. PCR using these two primers detected two shortened products in both boys, while no products with normal length were detectable. Amplifications of both mothers' mRNAs revealed a mixture of normal and truncated PCR products. Sequencing of the PCR products showed the deletion breakpoints to correspond to intron/exon boundaries of exons 5 and 6. Both patients have mixed populations of truncated gp91-phox trancripts lacking either exons 5 and 6 or only 6. Genomic PCRs are currently performed to find out about the mutation at the genomic level that caused this unusual type of splicing defect.

The response to xenobiotics of cultured rat hepatocytes is affected by physiological oxygen tension. P. Maier, B. Saad and H.P. Schawalder, Institute of Toxicology, Swiss Federal Institute of Technology and University of Z~rich, CH-8603 Schwerzenbach,. The role of oxygen tension in the site specific toxicity of xenobiotics within liver Iobules was analyzed in male rat hepatocytes cultured under peripcrtal-(incubator 13%O2) or perivenous-like (4%02) oxygen tension. Cells were exposed to phenobarbital (PB) or to 3methylcholanthrene (3-MC) from day 1 to 4 or 6 to 9. The inducibility of 6 different P450 isoenzymes Was analyzed with ELISA using monoclonar antibodies.

Enzyme activities were determined in cell homogenates (ethoy-resorufin-O-deethylase) or in situ (testosterone hydroxylation). In untreated cultures, oxygen tension did not affect xenobiotic metabolism. After exposure to PB or 3-MC, a dose dependent pattern of alterations was detectable similar to that found in vivo. CYP1A1/2 and CYP2C6 were more strongly inducible in 13% 02 compared to 4% 02 cultures and CYP3A in 4%02 cultures only. Similar differences were found in enzyme activities.

It is concluded that oxygen tension in rat hepatecyte cultures affect gene expression during the cellular adaptive response.

Genetic correction by viral transfeetion of a cystic fibrosis pancreatic adenocarcinoma cell line modifies interleukin-lct stimulated interleukin-6 production and activities of free radical scavenging enzymes Y. Cao, W. Liu, S.P. Inaebnit, R.A. Frizzell*, P. Birrer and U.N. Wiesmann, Dept. of Pediatrics, University of Berne, Switzerland and *Dept of Physiology and Biophysics, University of Alabama, Birmingham, USA Genetic correction of inherited genetic defects by viral genomic vectors in cultured cells has raised hope for patients. We have compared gene corrected cystic fibrosis (CF) pancreatic adenocarcinoma cells expressing chloride channel regulating protein (P+) with uncorrected CF (P-) and non CF ((2) pancreatic careinoma ceils for interleukin-l~ 0L-1) induced interleukin-6 (IL-6) production and 02" radical scavenging enzymes. In P-and C cells no response to stimulation with IL-1 was noted while gene corrected P+ cells showed greatly increased IL-6 production (+1000-2000%). Glutathionedisulfide reductase was 300%, glutathione-S-U'ansferase was 25% and glutathione peroxidase was slightly increased above control cells. Cu/Zn-Superoxide dismutase activities and reduced glutathione were same in all cultures. Inhibition of cell growth with mitomycin C had no effect on the observed cellular characteristics. We conclude that complementation of defective gone products after retrovirus mediated gene wansfer to the genome with random insertion of genes may possibly modify yet other biological characteristics of cells and may lead to unexpected and possibly undesired effects in vitro and in vivo. In order to study the physiological significance of parvalbumin, we produced transgenic mice overexpressing this Ca2+-bindingprotein. Several hybrid constructs containing tissue-specific (neuron-specific enolase), or general promoters (SV 40 and metallothionein), coupled to the rat parvalbumin coding sequence were used. One transgenic line containing the human metallothionein IIB promoter was analyzed by RNA-PCR with oligonucleotides specific for ectopic parvalbumin transcripts as well as by immunohistochemistry using a polyclonal antibody. Both methods revealed expression of the transgene in several tissues, such as liver, brain, kidney, pancreas and organs of the male genital tract. A detailed analysis of both the constitutive and heavy metal induced expression of this protein in cellular subpopulations, and on its subcellular distribution, will be presented. The urokinase-type plasminogen activator (uPA) is an extracellular protease which converts plasminogen to plasmin, a protease involved in many biological processes, uPA is highly expressed in different transformed murine endothelial cell lines, namely five mT-transformed cell lines (e.g, eEnd2) and a cell line F2 derived from a UV light-induced t~umour. Another SV40transformed endothelial cell line, TME, expresses only low levels of uPA. The difference of uPA expression between these cells, as shown in a protein activity assay, is correlated to the uPA mRNA concentrations. The uPA mRNA levels in these cells are 80:40:1 for eEnd2, F2 and TME, respectively.

To determine the mechanisms underlying the uPA mRNA up-regulation, we investigated both transcriptional and posttranscriptional events. The stability of uPA mRNA was modest in eEnd2 (T1/2 = 15h) and TME (T1/2 = 9h) but very high in F2 (Tva >> 24h). The template activity of the uPA gene in eEnd2, F2 and TME is 8:3:1. These results suggest that transcriptional events are more important for the different uPA expression in these cells although the influence of the mRNA stability can not be excluded, especially for the uPA regulation in F2 cells. Functional analysis of the uPA promoter euggests that a PEA3 site at -2.4kb is essential for uPA expression in eEnd2 and F2. Chromatin diminution takes place in the early development of the nematode Ascaris lumbricoides. The eliminated material comains mainly satellite DNA, but also a gene encoding a ribosomal protein (S19) (Etter et al., 1991 , Prec. Natl. Acad. Sci. USA, 88, 1593 .

We looked for other ribosomal protein genes. An Ascaris eDNA library was screened and four cDNAs (4, 61, 66, and 69) were cloned. Their sequences were found to have homologies with ribosomal proteins from other species, including yeast, chicken, rat, mouse and human. In Southern hybridizations, S19 is eliminated in the genomie larval DNA, but cDNAs 4, 61, 66, and 69 hybridize to both oocyte and larval genomic DNA. These results show that the genes coding for 4, 61, 66 and 69 are not eliminated during early development of Ascaris. Northern blots on germ-line and somatic RNA revealed that the four ribosomal proteins are transetibed in all tissues analysed. Ribosomal proteins are known to have housekeeping functions and the expression of their genes are coordinately regulated. Current experiments are in progress to clone the genes of these four noneliminated ribosomal proteins, to test by in vitro transcription the housekeeping functions of the five promoters, and to prove that chromatin diminution is an alternative way to silence housekeeping promoters. Nutrients must be metabolized by the pancreatic g-cell to cause insulin release. We have proposed that malonyl-CoA serves as a metabolic coupling factor when I~-cells are stimulated with nutrient secretagogues (Prentki et al. JBC. 267:5802, 1992) . The formation of malonyl-CoA is catalised by acotyl-CoA carboxylase (ACC). We have investigated the long-term regulation of ACC by nutrients using the g-cell line INS-1. Glucose caused a marked (20 fold) accumulation of ACC mRNA. The threshold glucose concentration was 5raM; half maximal and maximal effects occured at 15 and 20raM, respectively. The lag time of the induction was 3h and maximal transcript accumulation occured at 24h glucose stimulation. ACC mRNA induction was associated with ACC protein accumulation. The action of the sugar is most likely transcriptional since glucose'did not change the stability of the ACC transcript. Non-metabolizable glucose analogues mimicked the action of glucose. The result suggest that glucose is a major physiological regulator of the ACC gene and that in contrast to its action on insulin release, the sugar does not need to be metabolized to induce the ACC gene. The thiamine molecule consists of a pyrimidine and a thiazole moiety. The two halfs are synthezised in separate, still unknown pathways and condense to thiamine monophosphate. We defined structural genes responsible for thiazole (thi2) and pyrimidine (thi3) synthesis and for the condensation reaction (thi4).

The expression of these 3 genes is strongly repressed by thiamine. Th e regulatory mutants tnrl, tnr2, tnr3 and thil, which affect the expression of thiamine repressible acid phosphatase (pho4) also affect the 3 structural genes. All tnr mutants are derepressed for the genes thi2, thi3 and thi4. Thil mutants are repressed for thi2, thi3 and thi4. These results suggest that the thiamine biosynthetic pathway is under complex regulation.

A newly formed somatic telomere in Ascaris iumbricoides has no influence on the transcription of a nearby located gene Y4. Huang, H. Toblcr and F. MUller, Institute of Zoology, University of Fribourg, P&-oUes, CH-1700 Fribeerg

Telomeric position effects were first demonstrated in Saccharomyces cervisiae: if a gena is placed near a telomere, its transcription is reversibly repressed. This effect can be eliminated or alleviated through cell~ar.transand eis-meehanisms. Moreover, telomefic effects on the transcnptaon of different genes exist in other organisms as well. Preliminary data indicated that a gene encoding a putative GTP-binding protein is located within only 10kb of a de novo formed somatic telomere in Ascaris lumbricoides. Tiffs gene provided us with a good opportunity to see whether the newly formed telomeres of somatic chromosomes during the process of ehromatin diminution exert a telomeric position effect on nearby located genes. Our recent experiments show that this gene is expressed in all investigated stages before and after the elimination process, suggesting ~ the addition of the new telomere has no influence on the transcription ot this pamcmar gene. Does the gene simply escape from the .position effectorare there more complicated mechanisms invowea in tins pnenomenon, to answer this question, we are currently searching for additional genes being located in the chromosomal breakage region. Their expression pattern before and after chromatin diminution will he compared.

Moss transformation with potato phyA gene Bisztray, Gy. 1, Huges, j2, Schaefer, D 1, G~tz, C. 3 and Zry'd, j.p The aquatic plant Spirodela polyrrhiza, can be induced to form dormant buds (turions) by both environruental (cold) and chemical (abscisic acid, ABA) triggers. By differential screening of a eDNA library we have isolated a full-length eDNA for an ABA-induced peroxidase (tur4). The transcript level is increased within 2 hours of ABA treatment and is also induced by low temperature. The low temperature induction of tur4 is reversible by cytokinin (which i~ahibits ABA-and cold-induced turion formation). The deduced sequence of tur4 has a maximal 48% identity with those of previously published plant peroxidase sequences, and thus represents a new peroxidase family. A signal peptide is found at the amino-terminal end of the peroxidase, which suggests its association with ABA-induced processes in the cell wall. Chloroplasts have their own set of translation elongation factors (EF-G and EF-Tu ) which in higher plants are nucleus encoded. A partial soybean genomic library (3 to 4 kb EcorI fragments, lambda gtl0) was screened with a 200 bp DNA probe of the N-terminal part of the pea EF-G gene (gift of Dr. C. Breitenberger, Ohio State U.) to the end of retrieving the nuclear gene(s) for the chloroplast specific EF-G which so far has never been isolated. We isolated and sequenced a clone (3.6 kb insert) containing the entire coding part for the chloroplast EF-G. The gene is split having two introns in the core part. A third intron separates the putative transit peptide from the core part. There are two EF-G genes per haploid geuome as verfied by Southern experiments and eDNA analysis. We currently study the question of EF-G expression during leaf development trying to understand the role of elongation factors in the regulation of chloroplast protein synthesis (see Maurer F. et al, this poster session).

Andrew Fleming, Therese Mandel, Isabelle Roth, and Cris Kuhlemeier Institute of Plant Physiology, Bern University, CH-3013 Bern

The shoot apical meristem plays a vital role in plant development. Not only is it the ultimate source of all the cells of the aerial part of the plant, the cell divisions that occur in this area are organised so as to define the initial steps in leaf and stem morphogenesis. Various models of meristem structm'e and function have been proposed identifying distinct cellular compartments within the apex. Using in situ hybridisation, we have carded out an analysis of the expression patterns of a number of genes obtained from a meristem cDNA library. This analysis has revealed a variety of patterns within the apical meristem.which, when put into the context of the proposed models, necessitates a re-interpretation of how the apical meristem is organised at the cellular level.

COEXIST IN 3-DAY OLD SOYBEAN COTYLEDONS.

Guex, N., Widmer, F., and Richter, H., Institut de Biologic et de Physiologic VEgEtales de l'Universit6, CH-1015 Lausanne.

Polyclonal antibodies raised against isocitrate lyase (ICL; EC 4.1.3.1) were used for immunoscreening a X-ZAP II expression library constructed from 3-day old dark germinated soybean cotyledons (Glycine max. L.). Four independent clones were purified, revealing that at least two isoenzymes of ICL coexist in cotyledons at this stage of development. The cDNA clones are nearly full length and show 95% identity in the coding region and 80% identity in the 3' untranslated region. While most differences between the two isoforms are found to be silent, 11 amino acids are different (including one gap of one amino acid), and determine slightly different pls for the two isoenzymes. While the isoenzyme containing the gap is more similar to all plant ICL sequences reported so far in databases, the other one appears more unique (or not yet identified in other plant species were used for immunoscreening a ~.-ZAP II expression library constructed from 3-day old dark germinated soybean cotyledons (Glycine max. L.) Ten independent clones were purified and partially or totally sequenced. All of them proved to be MS. Eight showed approximately the same size (1,8 kb) on a 1% agarose gel electrophoresis, whereas one was shorter (1.6 kb) and one considerably longer (ca 3 kb). The hypothesis of a chimeric eDNA can be excluded since the size of the insert was determined after digestion of the plasmid with the same restriction enzyme used to clone the eDNA fragments and that a single insert of 3 kb was detected. As the ORF of MS is about 1.7 kb and knowing that splicing in plants is less efficient compared to the animal system, the 3 kb clone could reflect an unprocessed mRNA. While partial sequencing data on the 5' end identifies MS, the 3' end is totally different, which suggests an uncommonly long 3' untranslated region with possibly interesting regulatory aspects. Further studies are necessary to identify the real nature of this 3 kb clone. Chloroplast development and function is tightly controlled by the nucleus. According tn several reports the translation elongation step may be one of the control points and the expression of the respective genes in the nucleus could be crucially involved. We have identified an EF-Tu gene family (two sub-families with two members each) and so far sequenced two complete genes including the upstream parts with promoter elements. The promoter region of one gene is studied in detail. Upstream sequences of defined length ( deletions ) were linked to the GUS-reporter gene. Promoter activity is tested in transformed tobaco. We are currently testing gene activities at various developmental stages (F1 generation). We try to define cis -elements required for control; we test possible differential expression of members of the EF-Tu gene family; we correlate gene expression with the elongation factor activity within chloroplasts. These studies are complemented by the experiments of Hernandez Torres (see this poster session).

We previously showed that three glycolytic marker genes are coordinately expressed in leaf and rhizome of the amphibious plant Acorus calamus in the laboratory during extremely long periods of anoxia. Here we investigated how the same genes were expressed under natural conditions in the lake. We observed marked differences to our laboratory experiments. In the leaves high mHNA steady-state levels of pyruvate decarboxylase and alcohol dehydrogenase, the two enzymes in alcoholic fermentation, coincided with leaf submergence (hypoxia) in winter and fall. In the permanently flooded rhizome, transcript levels of the same genes rapidly increased with the onset of growth in March. The mRNA levels of fructose 1,6 bisphosphate aldolnse, an enzyme in the glycolytio stem, were high in the leaves throughout the year. In the rhizome aldolase transcript levels correlated with leaf submergence and with the generation of young rhizome in summer. The results indicate that under natural conditions multiple interacting factors differentially regulate glycolytic gene expression in leaf and rhizome and that oxygen deprivation and cold may be two of them.

Reto Brosi, Karsten Grtning, Diana Blank* and Angela Kramer;, Dtpartement de Biologic Cellulaire, Universit6 de Gen%ve, CH-1211 Gentve; *Biozentrum, Universit~t Basel, CH-4056 Basel.

The splicing of inb'ons from nuclear pre-mRNA occurs in large complexes (spliceosomes) which are assembled by transacting factors (snRNPs and proteins) in a stepwise fashion. The first ATP-dependent intermediate in spliceosome assembly is pre-splicing complex A. We have previously shown that at least three protein factors (SF1, SF3 and U2AF) are required for its formation in addition to U1 and U2 snRNPs. We have separated SF3 into two activities, both of which are essential for the formation of complex A. Purified SF3a consists of three polypeptides of 60, 66 and 120 kDa and interacts with U2 snRNP in the presence of SF3b. The 60-kDa subunit of SF3a crossreacts with an antibody directed against the yeast splicing factor PRP9 that is essential for the incorporation of U2 snRNP into the spliceosome. Moreover, a eDNA encoding the 60 kDa-" subunit of SF3a shows homology to PRP9. In addition, a human homologue of PRP9 has recently been detected as a U2 snRNP-specific polypeptide of 60 kDa. Taken together, these data support the hypothesis that SF3a represents a loosely associated subunit of U2snRNP which is essential for the U2snRNP/pre-mRNA interaction. Since SF3b by itself binds to U2 snRNP and is required for the subsequent association of SF3a, it is possible that it represents other U2 snRNP-specific polypeptides.

ANATOMY AND PROCESSING OF THE psbD-psbC GENE TRANSCRIPT OF EUGLENA GRACILIS Spielmann A., Orsat B., Marc-Martin S. and Stutz, E. Laboratoire de Biochimie vtgttale, Universit6 de Neuchfitel

The chloroplast genes psbD and psbC code for PSII reaction center proteins of 39 anf 51 kb, respectively. The two genes together extend over better than 20 kb containing 22 introns. The 3'terminal part ofpsbD overlaps with the 5'terminal part ofpsbC, and the last intron ofpsbD corresponds to the first intron of psbC. Exon 2 of psbC is part of the 3'tail of the psbD mRNA. The final processed psbD and psbC mRNAs are 1.4 and 1.5 kb, i.e. a common primary transcript including the entire psbD and psbC gene must undergo multiple splicing and processing steps, including alternate splicing in the tail region. The psbC exons 2 and 3 are separated by a gap of 4.1 kb. In previous EM studies no intron was seen in this gap region while, e.g. each of the 11 introns was visible in the psbD region . This could mean that two smaller precursor RNAs exist and exon 2 and 3 of psbC are trans-spliced. Experiments are under way to settle this question. The U7 RNAs from mouse, human, Xenopus and sea urchin have an atypical Sm binding site: AAUUUGUCU. In transient expression studies in Hela cells the mouse U7 RNA accumulates to 3-4 fold lower level than the U1 RNA. When its Sm binding site is mutated to the canonical sequence: AAUUUUUGG (U7 Sm CAN), nuclear accumulation and anti-Sin precipitability reaches that of U1 RNA. Transcription of the mouse U7 and U1 gane is equally efficient in Hela cells, thus the atypical Sm site must be responsible for poor assembly of the U7 snRNP, resulting in degradation of the free RNA. We have developed a fimetional assay for the U7 snRNI' based on injection of the mouse gene into Xenopus oocytes. We show that the mouse U7 RNA is functional in Xenopus oocytes. However, the U7 Sm CAN RNA is assembled into a nonfunctional particle, although the particle does aeeumalate at 3-4 fold higher amount in the undans. The U7 Sm CAN RNI? is inactive regardless of the amount of RNA transcribed from the injected gene, excluding a possibility that overexpression of the RNA titrates out a limiting assembly factor(s). UV crossllnking experiments have revealed a difference in the pattern of proteins that can be crnsslinked to U7 RNA and U7 Sm CAN RNA. This suggests that positioning, conformation or compositon of proteins complexcd with the U7 Sm CAN has been altered, rendering the particle inactive. Histone mRNA 3' ends are formed by a unique endonucleolytic cleavage resulting in poly(A)" mRNAs with a highly conserved 3'-terminal hairpin loop structure. Mutations of this structure do not abofish processing but have variable quantitative effects depending on the gene and on the biochemical environment (extract) used. In contrast, mutations abolishing the potential base-pairing between a downstream pre-mRNA spacer element and the 5' end of U7 RNA present in U7 small unclear ribonucleoproteins (snRNPs) completely inhibit processing. We have observed that this base-pairing potential extends further than previously recognised in either direction. In particular, the region surrounding the actual processing she could i~teract with a U7 RNA segment involved in binding of structural proteins of the U7 snRNP. Mutations in this region of the pre-mRNA do not abolish processing but have variable quantitative effects and sometimes affect the specificity of cleavage. We are currently determining which parts of the pre-mRNA are involved in bona fide and productive base-pairing with U7 RNA by further mutagenesis and by psoralen photo-crosslinking experiments. The finding that a mutation in the spacer element which increases the base-pairing potential with U7 RNA has no effect on processing by itseff but can rescue the partial processing deficiency of a stem-loop mutation suggests that the hairpin (and factors binding to it) may serve to stabilise the pre-mRNA:U7 interaction.

Silke Backes and Angela KrOner; Dtpartement de Biologie CeUulaire, Universit6 de Gen~ve, CH-1211 Gen~ve.

We are investigating the splicing of introns from nuclear pre-mRNA with components fractionated from HeLa cell nuclear extracts. At least four protein factors (SF1, SF3a, SF3b and U2AF) and U1 and U2 snRNPs are required at the onset of the reaction for the assembly of a pre-splicing complex. SF1 was purified to homogeneity. It consists of a single polypeptide of 75 kDa which is completely heat-resistant. Tryptic peptides of SF1 were sequenced and degenerate oligonucleotides were used to screen a HeLa eDNA library. Thus far, we obtained a partial eDNA that encodes an ORF of 369 amino acids containing nine of the twelve SF1 peptides sequenced. No significant homology to proteins in current data bases was found. Northern blot analysis with HeLa poly A § RNA or RNA from different human cell lines revealed two major mRNAs of 3.0 and 3.5 kb.

The exact function of SF1 during pre-splicing complex assembly is still unclear. Neither in vitro assays nor the derived amino acid sequence revealed intrinsic activities (RNA binding, ATPase, RNA helicase, RNA annealing) that may play auxiliary roles during the splicing reaction. In a search for characteristic protein motifs a putative leucine-zipper and several possible phosphorylation sites were found within the SF1 sequence. Taken together, our results suggest that SF1 does not interact with the pre-mRNA substrate directly, but may associate with other components of the splicing machinery by protein-protein contacts.

Irmgard Haussmann, Renu Mital, Jean-Claude Schaer* and Daniel Sch0mperli, Abteilung for Entwicklungsbiologie, Battzerstr. 4, CH-3012 Bern and * Pathologisches Institut, Murtenstr. 31, CH-3010 Bern Histone mRNA is generated by a unique 3' processing reaction which is regulated during the cell cycle and in response to changes in cell proliferation. U7 small nuclear ribonucleoprotein (snRNP) participates in this reaction by basepairlng of the U7 RNA 5' end with a conserved 3' spacer element of histune pre-mRNA Additional transacting factors involved in the reaction are a heat labile factor (HLF) and a factor binding to the hairpin present at the 3 ~ end of mature histone mRNA (FIBF). It was previously shown that growth-arrested cells are deficient in HLF and the accessibility of the 157 srtRNA 5' end to micrococcal nuclease (MN) undergoes a marked change, although these two events were reported for different cellular systems. We are currently analysing several parameters of histone RNA processing in a variety of cellular systems. Cells to be analysed include a ts mouse mastocytoma cell cycle mutant (21-Tb), mouse Iibroblasts that can be arrested by serum starvation (C127), and Chinese hamster ovary cells synchronised by mitotic selection. In time course experiments, we are assaying histone RNA 3' processing, the separate activities of HLF and HBF, the abundance and mobility on native gels of U7 snRNPs, snRNP structural proteins that can be UV--crosslinked to U7 snRNA and the accessibility of the U7 RNA 5' end to MN. Preliminary results with Gl-arrested 21-Tb cells indicate that HLF and H]3F activities are lost at an early stage, whereas structural changes in the U7 snRNP occur with more prolonged proliferation arrest. Chromatin diminution in the nematode A. lumbricoides leads to the formarion of somatic cells containing less DNA than cells of the germ-line. We have identified a gene which is cis-and trans-spliced and which is located in the germ-line specific material. The following evidences indicate that this gene, called ALEP-1 (Ascaris lumbricoides eliminated protein 1), codes for a ribosomal protein. The predicted translation product shows significant homologies to proteins which are considered to be part of tlae small subunit of ribosomes in different species. In agreement with this finding, antibodies against ALEP-1 demonstrate that the protein is associated with the small subunit of the A. lumbricoides ribosome under salt conditions which are thought to be diagnostic for ribosomal proteins. The fact that tiffs gene is included in the eliminated portion of the genome indicates a difference in somatic and germ-line ribosomes and may suggest an implication in the regulation of translation. Furthermore, the elimination process itself may be used by the organism to specifically shut down transcription of the ALEP-1 gene in cells which are not capable to do so by normal cellular mechanisms designed for gene regulation. The yeast PRP20 protein is highly homologous in structure and function to the RCC1 protein of higher eukaryotes. The RCC1 protein is involved in the regulation of the onset of mitosis, whereas the PRP20 protein was shown to be required for accurate and efficient mRNA metabolism. Here we report the isolation of new temperature-sensitive alleles of the PRP20 locus. Most of the mutations affect conserved amino acid residues in the C-terminal region of the PRP20 protein. The first observable phenotype in mutant prp20 cells when shifted from permissive to nonpermissive temperature is a loss of nuclear PRP20 protein.

Concomitantly, an accumulation of polyA + RNA in the nucleus is observed. The temperature-sensitive RCC1 allele in the mutant hamster cell line tsBN2 leads to a similar accumulation of mRNA in the nucleus. The synthesis of nitric oxide (NO) from L-arginine by mammalian cells is a pathway involved in many biological functions including macrophage cytotoxicity. Previous studies have shown that one target of NO-action are iron-sulfur-dependent enzymes like mitochondrial aconitase. Furthermore, in murine macrophages, which produce NO in response to stimulation with IFNy, nitrosyl-iron can be detected supporting the idea that NO can interact with [Fe-S]-clustcrs. Iron regulatory factor (IRF), an RNA-binding protein modulating the expression of different proteins involved in iron metabolism, has sequence homology to mitochondrial aconitase. Recently it could be shown that IRF by itself possesses aconitase activity which is dependent on the presence of a [4Fe-4S]-cluster. Here we found that in the murine macrophage cell line RAW 264.7 the RNA-binding activity of IRF is activated after stimulation with IFNT, due to the production of NO. IRF is also activated when incubated with the molsidomine derivative SIN-l, which spontaneously liberates NO. The eukaryotic translation initiation factor elF-4A is an RNA dependent ATPase which, together with elF-4B, unwinds ds RNA in vitro. In higher eukaryotic cells the initiation factor 4A can also be found in the cap binding complex elF-4F, i.e. together with the cap binding protein elF-4E and a high molecular weight protein of 220kDa. In the yeast S. cerevisiae two genes, T1F1 and T/F2, code for exactly the same 4A protein. We have also isolated the S. pombe genes encoding elF-4A. To understand better the function of elF-4A we carried out a mutational analysis of the T/F genes from S.c. genes encoding this factor. We have used some of these mutations to isolate suppressors in order to find genes coding for proteins interacting with elF-4A. In this analysis we have isolated several intragenic suppressors, which restore elF-4A function. We also have isolated gene dosage dependent suppressors using a wild type library on a multicopy plasmid. These suppressors are currently under investigation. Iron regulatory factor (IRF) is a cytoplasmic RNA-binding protein involved in regulating iron homeostasis. The IRF controls expression of ferritin and transferrin receptor mRNA post-ti~anscriptionally, via specific binding to stem-loop iron-responsive elements (IREs) located in the untranslated regions of these mRNAs. We have observed that formation of a second RNA-protein complex in rodent cell extracts is (like IRF) regulated by iron levels. This faster migrating complex represents a specific interaction between the IRE and an as yet under'reed iron-regnlated protein, that is distinct from IRF, as concluded from the following evidence: a) cross-linking and V8 protease digestion reveals different peptide patterns for the two RNA-protein complexes, b) antisera raised against two different IRE pepfldes immunoprecipitate only the tRF bandshift complex, c) IRF could be separated from the second IRE-binding protein by ion-exchange chromatography, and oltly the fraction containing 1RF reacted with the antibodies in Western blots, and d) the affinity of the faster migrating bandshift complex for the IRE is similar to that of IRF, as demonsWated by competition assays with related RNA stem-loop structures.

Chamot, D. and Kuhlemeier, C., Pflanzenphysiologlsches Institut der Unlversith't Bern, CH-3013 Bern

Translation initiation factor eIF-5A is a protein involved in the formation of the first peptide bond of protein synthesis, eIF-5A is also the only known protein to contain the unique posttranslationally modified amino acid, hypnsine, which is derived from the polyamlne spermidlne. In order to study the role of erF-5A and its hypusine modification in plants, we have characterized in detail 3 members of a mnltigene family representing eIF-5A in tobacco. At least 2 of the genes (NeIF-5A1 and 2) are differentially expressed in various tissues at the mRNA level. Transgenlc plants expressing sense or antisense constructs using NeIF-5A1 and NeIF-5A2 are being used to study the possible differential functions of these 2 genes. NefF-5A3, a genomic done, is virtually identical to the NeIF-5A1 eDNA. Transgenle plants expressing NelF-5A3 promoter-GUS fusions reveal tissue specificity for this third gene member. IEF western blots also reveal a differential expression pattern of NeIF-SA polypeptides between tissues. Taken together, these data suggest that specific NeIF-5A isoforms perform different functions. The possible roles of different NeIF-5A isoforms are being investigated further in terms of the translational regulation of tissue-specific processes in tobacco.

Owttrim, G., Brander, K., Mandel, T., Lutziger, I., and Knltlemeier. C., Pflanzenphysiologisches Institut, Universit~it Bern, CH-3013 Bern

Eukaryotic translation initiation factor 4A (eIF-4A) is an RNAdependent ATPase which functions as an RNA helicase, removing secondary structure from the 5' UTR region of mlLNA during translation initiation in enkaryotic systems, eIF-4A is also the prototype member of a large gene family, the D-E-A-D box family, whose members function in a diverse range of processes including RNA splicing, cell growth, and cell development.

We have characterized two divergent eIF-4A geue families in tobacco, NeIF-4A2 and NeIF-4A3, which are coordinately expressed in all tobacco organs. The NeIF-4A2 gene family consists of at least eight members which are expressed in leaves. An anti-plant eIF-4A antibody detects numerous eIF-4A isoforms whose presence changes both quantitatively and qualitatively in various organs. Furthermore at least one eIF-4A isoform is chloroplastically localized. Promoter analysis indicates that one NeIF-4A2 family member is expressed only in mature pollen. These results suggest that the expression of specific eIF-4A genes may be developmentally regulated and possibly even regulating in plants. Transgenic plants which either over-(sense) or under-(anti-sense) express NeIF-4A2 and NeIF-4A3 are currently being produced to further analyze eIF-4A function in vivo.

Meins, Jr, The Friedrich Miescher Institute, Box 2543 4002 Basel Transgenes introduced into plants can interact with homologous host genes leading to a decreased expression of both genes. We found a similar phenomenon in N. sylvestris containing a tobacco class I chitinase transgene. This effect, called "silencing," only occurs in plants homozygous for the transgene and is developmentally regulated. Although all young seedlings showed the high-ohitinase phenotype, three patterns of expression were observed in mature plants: most plants showed uniformly high or low chitinase levels in different leaves. A few plants showed position-dependent expression which varied in different leaves. The incidence of silencing is also environmentally regulated. Approx. 20-40% of plantlets germinated from seed in closed culture vessels showed the silent phenotype at maturity in the greenhouse, whereas, none of the plants germinated and raised to maturity in the greenhouse showed this phenotype. The implications of silencing for gene technology will be discussed.

Beffa, R., Neuhaus, J.-M. and Meins, F., Jr., Friedrich Miescher-lnatitut, P.O. Box 2543, CH-4002 Basel Vacuolar class I 8-1,3-glucanses (EC 3.2.1.39) have antifungal activity and are believed to be important in defending plants against infection by pathogens. We used antisense transformation to test this hypothesis and identify other possible functions of this enzyme. Here, we report that N. sylvestris and tobacco plants infected with viruses, Tobacco Necrosis Virus and Tobacco Mosaic Virus, respectively, can compensate physiologically at the level of enzyme activity for deficiencies generated by antisense transformation. Virus induction of class I 8-1,3-glucanase antigen was markedly inhibited in leaves of antisense transformants of both plant species obtained with different expression vectors. In both experimental systems a serologicafly distinct B-1,3-glucanase activity was induced in antisense transformants by virus infection. This serotype did not accumulate in comparable healthy or infected tissues of empty vector transformants nor was it induced in antisense transformants by stress hormone ethylene. The compensatory 6-1,3-glucanase activity is an intracelblar enzyme distinct from the isoforms of tobacco 8-1,3-glucanase described previously. We conclude that plants infected by viral pathogens can specifically compensate for a deficiency in the class I isoform by producing a functionally redundant protein. This argues strongly for an imporlant function of 8-1,3-glucanases in pathogenesis. 

The function of a small GTP-binding protein in the regulation of cell cycle in higher plants. The life of a dividing cell is an orderly progression of events associated with growths, DNA replication and mitosis. Stringent dontrols prevent late events from occurring before completion of earlier events. Incompletely replicated DNA activates a regulatory mechanism, in fungi and mammalian cells, to ensure that mitosis strictly follows S phase. Unreplicated DNA is probably detected by a complex of RCCI, a DNA binding protein, and Ran, a Ras-related small GTP-binding protein in those organisms. We cloned the tobacco homoloque of the Ran gene. The Nt-Ranl protein shows high (>80 % ) homology to fungal and mammalian Ran proteins. The Nt-Ranl gene expressed predominantly in actively dividing tissues (root tip, apex). Moreover, overexpression of the Nt-Ranl cDNA results in the complementation of the relevant fungal mutation. Therefore we propose, that the same control system also operates in the cell cycle in plants.

Emery, G., Jaussi, R. and Crompton, N. We are investigating radiation induced cell cycle arrests. The S phase cells in populations of exponentially growing V79 hamster fibroblasts were pulse labelled with BUdR. Then, the cells were exposed to x-rays (dose = 6 Gy} and the average velocity of the labelled cells through the S phase was monitored by flow cytometry. The cells do not all arrest immediately but over a period of 2 hours. The cells were arrested for an average of 0.6 hours. This arrest is independent of the delay to cells entering S phase, which can be observed in the unlabeled cell fraction.

Caffeine abrogates the arrest induced by the radiation, and there is even an indication that it causes progression through S phase to be faster than in the control. Breast tissue biopsies of patients with ductal invasive carcinoma (82 cases, age range 36 to 89 years old) were homogenized and nuclei were isolated by means of two different procedures (lysis or gradient centrifugation after nitrogen pulverization). DNA index and proliferative activity (% S-phase) were estimated by four mathematical algorithms after flow cytometry analysis. The remaining tumor powder was homogenized, centrifuged at 100000 g and the high speed cytosol fraction assayed for estrogen (ER) and progesteron (PR) receptors. The activity of the methyltransferase-I (PMT-I), an enzyme catalyzing the methylation of the membrane phosphatidylethanolamine to phosphatidylcholine was determined in the mierosomal pellet. 70 % of samples showed aneuploid DNA index and % S-phase ranging between 5-40 % without any correlation with both ER, PR and PMT-I activity. %S-phase from nuclei prepared with the two procedures gave similar values when evaluated with three algorithms (RFIT, SOBR and POLY), whilst the SFIT model yielded significantly higher values. We have used the polymerase chain reaction (PCR) to isolate and characterize cDNA fragments encoding human protein kinases.

This method has been succesful in that several novel protein kinases have been isolated.

One of these isolates has been cloned and has 52% sequence identity to the protein kinase "Polo", characterised in the fruit fly Drosophila.

The precise function of this protein is not known, but analysis of Polo mutants reveals that the gene product is essential for chromosome condensation and microtubule spindle organisation during mitosis.

Experiments to examine the subcellular localization and biochemical activity of the putative human homologue of Polo are in progress.

Bochaton-Piallat M-L, Gabbiani F, Ropraz P and Gabbiani (3. Department of Pathology, University of Geneva, CMU, CH-1211 Geneva 4.

Artedal smooth muscle cells (8MC) implicated in the formation of atherornatous plaque ((3abbiani et al., J Clin invest 73:148, 1984; Kocher et al., Circ Res 56:829, 1985) or cultured from adult rats (Skalli et al., J Submicrosc Cytol 18:481, 1986) show decreased levels of specific cytoskeletal proteins such as oL-emooth muecle (SM) actin, desmin and SM-myosin, assuming phenotypie features of fetal SMC. We have studied phenotypic features of populations and cloned arterial SMC cultured from rats of different ages. Newborn 8MC cultured in the presence of fetal calf serum maintained differentiated features up to the 5th passage whereas SMC cultured from adult and old SMC did not. Old SMC showed the most important replicative activity. Newborn and adult 8MC, cloned in primary culture, continued to express (z-SM actin, SM myosin and in some cases desmin at the 5th passage.

Compared to their parental populations, these clones express differentiated features in culture. In contrast, clones of old 8MCs were poorly differentiated and expressed ~-SM actin only in 80% of the cases. Thus, the capacity of SMC to differentiate in culture decreases when animals become old. Clonal populations may be useful to study the mecanisms of SMC differentiation and of athemmatous plaque formation. ( The cell cycle is controlled by cyclin-dependent protein kinases. While B-type cyclins associate with p34 cdc2 to trigger entry into mitosis, progression through S phase requires cyclin A, presumably in association with p33 c~2. A-and B-type cyclins display strikingly different subcellular localizations, suggesting that they may target different cdk catalytic subunits to appropriate substrates. Here, we have used N-and C-terminal deletion mutants of cyclin A to determine the structural requirements for nuclear localization, as well as for complex formation with cdk catalytic subunits. We show that deletion of residues within the centrally located "cyclin-box", or deletion of as few as 15 residues from the C-terminus of cyclin A abolishes both nuclear localization and kinase association, whereas deletion of more than 100 residues from the N-terminus is without effect on either parameter. We conclude that nuclear transport of cyclin A is not mediated by a classical intramolecular nuclear localization signal, but instead depends on the formation of multiprotein complexes involving cdk catalytic subunits. Protein phosphatase 2A (PP2A) encompasses a family of holoenzymes with a core structure consisting of a 36-kDa catalytic subunit and a 65-kDa regulatory subunit which can be associated with a third subunit of either 54, 55 or 72 kDa. We analyzed the spatial and temporal expression of several PP2A subunits during Drosophila development. The transcripts encoding the two core subunits are expressed ubiquitously, but the 55-kDa subunit (PR55) transcripts are mainly found in early embryos and at lower levels in the nervous system and in the gonads. PR55 expression in these highly proliferative tissues occurs in parallel with elevated levels of the two core subunit transcripts. A mutant strain, termed aad for abnormal anaphase resolution, was found to contain a P-element insertion within the PR55 gene. Larval neuroblasts show an increased mitotic index, overcondensed chromosomes and defects in anaphase where either chromatin is bridging the two poles or single lagging chromosomes fail to migrate to the poles. The mutant phenotype can be rescued by the wild type PR55 gone, suggesting that the PR55 containing PP2A holoenzyme plays a specific role in cell cycle regulation. We have used a PeR-based strategy to identify human protein kinases involved in cell proliferation. Degenerate oligonucleotide primers derived from conserved amino acid motifs shared between serine/threonine -specific protein ldnases were used to amplify sequences from a leukemic human cell line, HL60. Partial sequences of over 40 protein kinases were identified and two of these, HuPK21 and HuPK36, showed homology to nimA of Aspergillus nidulans. The nimA protein kinase regulates entry into mitosis and is activated independently cclc2 of the mitotic regulator p34

. A full length cDNA for HuPK21 and a partial cDNA for HuPK36 have been obtained. The amino-terminal catalytic domains of HuPK21 and HuPK36 show 48 % and approximately 40 % identity to that of nimA, respectively. Studies to characterize the substrate specificities, subcellular localizations, and cell cycle regulation of these protein kinases are in progress. and strong immunohistochemical staining is observed at particular moments of the cell cycle: in G1 and in mitotic cells. We postulated that this protein could intervene in some particular phenomena .controlllng the cell cycle. We therefore studied the effect of down-regulating CR expression by antisense oligo nucleotides (AS, ollgo thioderivatiyes, 3gM), taking as controls parallel cultures containing nonsense oligonueleotides (NS) or non-treated cultures. In AS-treated cultures, cells grow as compact clusters that are neither present in the NS-treated cultures nor in the controls; calretinin immune-reactivity is strongly decreased and the CR level, as determined by a Western Blot is lowered. G 1 cells accumulate and the mitotic index approaches zero along with the duration of the culture, as established by cytophotometric mesurements. These results demonstrate that calretinin antisense oligonucleotides down-regulate CR synthesis and this manipulation strongly affects the proliferative cycle of WiDr cells.

Interdepartmental Electron Microscopy and M.E.-MfiUer-Institute, Bioeenter, University of Basel. Klingelbergstrasse 70, CH-4056 Basel

Most CLSM provide only two or three wavelength maxima for excitation of fluorescently labelled probes. The commonly used DNA-specifie fluorochrome DAPI is excited at UV-wavelength thus requiring a CLSM equipped with an expensive UV-laser. For some time now, Molecular Probes, Inc. has offered DNA-and RNA-specine fluorescent probes that can be excited by an Argon-ion or Argon-lraypton laser.

Here we have investigated nuclei of dividing flbroblast cell lines in an CLSM equipped with an Argon-ion laser after fluorescent double labeUing the spindle apparatus (tubulin) and the chromosomes (DNA). For this purpose, the spindle apparatus was labelled with anti-tubulin antibodies coupled to TRITC (excitation 540 rim, emission 570nm) and the chromosomes with TOTO-1 (excitation 509 nm, emission 533 rim). By this labelling protocol we got excellent separation of the two fluorescent signals one coming from the spindle apparatus and the other from the chromosomes of eells in anaphase. To explore the 3-D organization of these two mitotic structures in greater detail, serial optical sections (0.3 mm apart) were recorded from which li) projections were calculated and (ii) 3-D data stacks computed and volume-rendered.

Identification of a vertebrate cdc2 mutant which is unable to complete the GI/S transition. Nicole Sc~tz, E.A.Nigg and Viesturs Simanis.. ISREC, Chemin Boveresses 155, 1066 Epalinges.

S. pombe strains have been constructed which should permit the genetic and molecular analysis of the events which occur in late G1 when cells become committed to the mitotic cell division cycle and the initiallon of DNA synthesis. A number of mutants of the chicken cdc2 gene were constructed during the course of analysis of plmsphorylation sites some of which lead to the interesting phenotype of cold sensitivity when expressed in a cdc2 null background, when the only source of p34 cdc2 is the chicken homologue of the gene, expressed from a multicopy plasmid. In order to create a genetically tractable strain, the wild-type chicken cdc2 gene and one mutant into the S. pombe genome to create a strain in which cell cycle progression is dependent upon the chicken cdc2 gene. Analysis of this strain indicates that the cells block predominantly before replication of DNA. In order to determine whether the cells were arrested before or after the execution of the start control their ability to conjugate at the arrest point was assessed. Consistent with a block before the traverse of start and commitment to S-phase, cells were able to conjugate with hig~a efficiency. Further support for the view that this mutant is defective only for the G1-S transition is provided by the observation that it is able to complement a cdc2A21 mutation in trans. This mulant is defective only for G2 function and not traverse of start. The double mutant is viable at the restrictive temperature of either of the parents and is no longer cold sensitive. Further work will concentrate on cloning genes involved in the traverse of GI into S phase in fission yeast. 

Dual role of the S. pombe cdcl6 gene: requirement for regulation of septum cdc2 formation and maintenance of p34 kinase activity in mitosis. Christian ~ John Marks, Alexandre Reymond and Viesturs Simanis. Unit6 de recherches sur le cycle celinlaire, ISREC, 1066 Epalinges.

In the fission yeast S. pombe a number of mutants have been identified in which Sphase and mitosis continue in .the absence of septum formation and cytokinesis resulting in the formation of highly elongated multinucleate ceils (cdc7, cdc11, and cde14) or in which septum formation is deregulated leading to the formation of multiple septa without cytokinesis (cdc16). Genetic studies have suggested lhat the products of these four genes may interact in regulating the formation of the septum. In order to study the mechanisms responsible for regulation of septurn formation and its coordination with mitosis, we have isolated these genes. This paper describes the cloning and analysis of the cdc16 gene. The sequence of the predicted gene product (p34 cdc16) shows homology to the BUB2 gene of S. cerevisiae. When cdc16-116 cells were shifted to the non" permissive temperature mitotic p34 cdc2 kinase activity decayed rapidly and cells failed to respond normally to the absence of the mitotic spindle, initiating septum formation even when mitosis had not been completed. Deletion of the gene demonstrated that it is essential for cell proliferation: spores lacking a functional cdc16 gene germinated~ completed mitosis and then formed multiple septa without undergoing ceil cleavage. The cdc16 gene product may be part of a feedback mechanism prevent premature exit from mitosis by maintaining high levels of p34 cdc2 activity until the initiation of anaphase thereby coordinating mitosis with septation, cytokinesis and entry into the next cycle.

Meili, R., Kaech, S. and Ballmer-Hofer, K., Friedrich Miescher-Institute, box 2543, CH-4002 Basel, Switzerland pp60 e'sre, a cellular tyrosine kinase, becomes transiently activated during mitosis. Activation is accompanied by phosphorylation of three sites in the amino-terminal regulatory domain of the protein, pp6() c'src is negatively regulated by phosphoryladon at tyroslne 527 and it has been shown that this site is transiently dephosphorylated in mitotic cells. Here we report that a non-myristylated mutant of pp60 e'sre is not activated and only partially phosphorylated at the amino-terminus in mitotic cells. Additional mutants lacking one (TTAc-src), two (AASc-src) and three (AAAc-src) cdc2 phosphorylation sites had slightly higher kinase activity than wt pp60 c'sre in interphase cells yet were not activated during mitosis. However, all four mutant proteins were still transiently dephosphorylated at tyrosine 527 during mitosis suggesting that myristylation and amino-terminal phosphorylafion may be necessary but are clearly not sufficient for mitosis-specific activation.

Messi, F., Department of Biotechnology, ETH Zurich, CH-8093 Zurich

Chinese Hamster Ovary (CHO) cells can be cultivated in both static and agitated culture without any supplementation of serum, proteins or complex additives such as hydrolysates and extracts. In mechanically agitated cultures, the cell growth kinetics are comparable to those observed with serum-dependent culture processes. These results are achieved by: Using a convenient culture medium which satisfies both physicochemical and nutritional requirements of the selected cells.

Ensuring a homogeneous dispersion of cells and media. Furthermore, serum-independent CHO cells grown in the chemically defined medium FMX-8 demonstrate to be a very promising culture system for expression and production of recombinant proteins in biotechnological processes. Amounts up to 25 mg ml "1 urokinase plasmlnogen activator were recovered in batch cultures of transfected serum-free growing CHO cells cultivated in spinner flasks. We have shown that the subcutaneous administration of transforming growth factor-I~l (TGFIM) to rats resulted in the formation of a granulation tissue in which cr actin expressing MF were particularly abundant. Other cytokines and growth factors such as platelet-dedved growth factor and tumor necrosis factor-~, despite their profibrotic activity, did not induce mSM actin in MF. In situ hybridization with an cr actin probe showed a high level of (z-,-qM actin mRNA expression in MF of TGF[~l-induced granulation tissue. Moreover, TGFIM induced ec-SM antin protein and mRNA expression in growing and quiescent cultured fibroblasts. These results suggest that TGF]51 plays an important role in .MF differentiation during wound healing and fibrocontractive diseases.

Influence of thyroid hormones on hepatic protein kinase Cis o forms S. Dotti-Sigrist, D. Fabbro* andA. Jakob, Biochemisches Institut, Universit~it, CH-4051 Basel, *Ciba AG, Pharmaceutical Dept., CH-4002 Basel, Switzerland.

It was previously shown that hypothyroid rats express higher levels of hepatic a-form of protein kinase C (PKC). In the present study we investigated PKC-subspecies in primary cultures of hepatocytes of normal and hypothyroid adult rats. Measurements of serum levels of free thyroxine (T4) and total trijodothyronine (T3) showed about 10% of normal levels in hypothyroid rats, Ceils were cultured for 72 h and treated either with vehicle (0.1 mM NaOH) or T 3 (1 uM) for 48 h. Cytosolic proteins were separated by SDS-gelelectrophoresis. Immunoblotting was performed with monoclonal antibodies against et, 61, B2, % 6, ~, v 1 and ~ PKC-isoforms and showed the presence of PKC-subspecics ct, 132, 6, e, and ~. Hypothyroid rats had higher levels of immunodetectable ct-PKC than normal rats and lower levels of the other isoforms. Treatment of cultured hypothyroid cells with T 3 reversed or-, B 2-and 6-PKC levels towards normal. Using differentiable (6C8) and nondifferenfiable (G3) subclones of an erythroleukemia cell line (RED-1), we tested whether factors promoting differentiation or prohferation affected G protein levels and G protein-mediated cellular signalling. Erythroid differentiation initiated by human recombinant erythropoietin (rhEPO, 1U/ml) and dimethylsulfoxide (DMSO, I%) in 6C8 was associated with a 80% decrease in antibody-reactive or ADP-ribosylated membrane levels of Gui and G[~3,, wlrile Gets remained unchanged. No loss of God or GI~'I was observed in G3. Upon initiation of differentiation in 6C8 cells, basal and forskolin-stimulated adenylate cyclase activity (AC) decreased by 35% and 50% respectively, but maximal isoprenalinestimulated cAMP formation remained unchanged, rhEPO alone, promoting cell proliferation but no differentiation, caused a 3-fold rise in B-receptor density, but no change in Gcd and basal AC activity. DMSO alone inhibited cell growth, reduced basal and forskolin-stimulated cAMP formation, but had no effect on G protein levels. In G3, containing twice as much Gi than 6C8, thrombine caused a PTX-inhibitable 25% decrease in isoprenalinestimulated cAMP accumulation. This effect was not observed in 6C8. Our results suggest that significant changes in G protein-mediated signalling occur during erythroid differentiation. The drop in G~ti may help to initiate differentiation in committed erythroid cells.

DD PK2 ACT AS AN UNUSUAL CATALYTIC SUBUNIT OF cAMP DEPENDENT PROTEIN KINASE Anjard, C., Etchebehere*, L., Pinaud, S., Veron*, M. and Reymond, C. D., Intitut d'Histologie de l'Universit6, CH-1005 Lausanne, * Institut Pasteur, 75724 Paris Cedex 15 cAMP dependent protein kinase (cAPK) plays an essential role during the development of Dictyostelium discoideum . We have described the Dd PK2 gene which encodes a putative protein of 648 aa. The C-terminal half of Dd PK2 present 54% indentity with mammalian cAPKs. We previously showed that overexpression of Dd PK2 in Dictyostelium results in higher PKA activity and leads to rapid development. We had conclude that Dd PK2 is a catalytic subunit of cAMP dependent protein kinase since it is physically associated with the regulatory subunit of Dd cAPK. The PKA activity is inhibitable by purified R subunit and is fully recorvered in presence of cAMP. Futhermore, Dd PK2 co-eluate with cAPK activity upon partial purification in DEAE column. Biotin is a water-soluble vitamin that belongs to the vitamin B complex. Biotin deficiency has been associated with skin diseases, such as squamous, seborrhoic dermatoses and skin depigmentation. We have studied the effects of pharmacological doses of biotin on the growth and differentiation of cultured epithelial cells, namely outer root sheath cells (ORSc) isolated from anagen hair follicles of both humans and farm animals (bovine, swine). A slight, but reproducible, stimulation of cell proliferation has been observed at concentrations of this vitamin in the range 10-8M -10-6M. The effect of biotin on the differentation of isolated ORSc has been examined both on monolayer and organotypic cultures. The latter were made possible by the use of a special device called Combi-ring-dish (CRD). The microscopic analysis of specifically stained specimens showed that cultured ORSc can be stimulated in vitro to produce an orderly structured and differentiated tissue comparable to normal epidermis. Calcitonin gene-related peptide (CGRP) is issued from alternative splicing of calcitonin mRNA and mainly produced in nervous tissues. This neuropeptide is present in perivascular innervation and can induce vasodilatation in vivo and in vitro. However, this peptide was recently shown to have trophic effects, e.g. it induces a dopaminergic phenotype in neurons of the olfactory bulb or increases cell-surface acetylcholine receptor numbers. This peptide also stimulates proliferation of a vas deferens smooth muscle cell line or of cultured endothelial cells. We decided to test the effects of CGRP on rat aorta vascular smooth muscle cells (VSMC) used at the 5th passage. When the density of 104 ceUs/cm 2 is used, we observed that 10-7M CGRP added every two days since seeding on cells grown in 10% FCS medium decreases significantly cell proliferation. We studied the expression of a specific marker of smooth muscle differentiation, the a-smooth actin isoform. CGRP decreases the level of asmooth actin in the cells 4 and 8 days after the beginning of treatment, as shown on immunoblots. Thus, although CGRP reduces cell proliferation, it does not seem to induce differentiation of the cultured VSMC. In vivo, CGRP innervation occurs very early during the differentiation of the vessel wall, when the cells are not yet fully differentiated. It could be possible that at this moment, it stabilizes the growth of the vessel wall.

Krebs,J. 1 and Honegger,P. 2 1institute of Biochemistry III, ETH Zentrum, CH-8092 Z0rich; 2Institute of Physiology, University of Lausanne, CH-1005 Lausanne.

In the rat brain, tdiodothyronine (T3) regulates the expression of a number of neuronal and glial membrane markers dudng rat brain development. Here we report the specific induction of a CaM-binding protein (p64) at a very eady stage of differentiation of a fetal rat lelencephalon cell culture system which can grow under chemically defined conditions. A number of CaM-binding proteins in the range of 40 -160 kDa could be identified which bind 1251-CAM in a Ca 2+dependent manner but only one was specifically induced by T 3. It is a soluble protein of 64 kDa which is expressed only in those cell cultures which have been exposed to T 3. The protein is visible already at very low T 3 concentrations (3xl 0" tOM). It seems to be specifically expressed in neurons since it is also visible in the presence of Ara-C which suppresses the growth ot gila cells. In addition, it can be induced by retinoic acid which is typical for proteins containing a T3-receptor responsive element in their promoter region. It is also interesting to note that p64 can be induced by T 3 already after 6 hours at which time a differentiation of the cells is not yet morphologically visible. The protein has been partially purified using a CaM affinity chromatography and its properties will be characterized. In this respect it might be ol interest that a CaM-kinase which seems to be specifically expressed in brain and in T-lymphocytes has a M r of 64 kDa.

Caldesmon and calponin expression in cultured smooth muscle cells from human airways.

The aim of the study was to explore in cultured smooth muscle cells (SMC) from human bronchi the expression of caldesmon and calponin, two proteins able to play a role in the regulation of contraction. The cells were found to express m-smooth muscle actin (eSMA), caldesmon and calponin after several passages. This was verified by immunocytochemistry, using the avidin-biotinperoxidase technique with monoclonal antibodies and also by immunoblotting. Between passages 3 and 7, more than 95% of cells were positive histologically for each protein and specific bands corresponding to cLSMA, to the heavy isoform of caldesmon (120 KDa) and to calponin were characterized on immunoblots. The expression of the heavy isoform of caldesmon and of calponin has been reported by others I to disappear in cultured SMC from the first passage. In sharp contrast, the present results show for the first time that cultured SMC from human airways continue to express these proteins. of Cell Biology, ETH-H6nggerberg, 8093 ZOrich Insulin and type I IGF receptors were demonstrated and characterized in long-term cultures of ARC. Type I IGF receptor number/surface area increased 8-fold during 9 days, whereas insulin receptor number/surface area remained unchanged. Thus, specific upregulation of the type I IGF I receptor occurs during redifferentiation of ARC in culture. No specific binding of GH was found. IGF I, but not GH, added to the cultures, increased granular density and spreading of cells after 7 days. After 16 days, IGF I-treated cultures showed, as compared to controls, a dramatic change in the extent of cross-striation as detected with immunofluorescence staining for myomesin. At the same time, ~sm-actin accumulation decreased. GH had no effect. Obviously, IGF I specifically enhances myofibril development in ARC. Furthermore, insulin and IGF I stimulated glycogen synthesis over the same concentration range despite a lO0-fold difference in the affinity for each other's receptor. Most likely, each of the two hormones acts on glucose metabolism in ARC via its own receptor. Transferrin (Tf) has strong growth-promoting-activity on a large variety of cell types in culture.

In order to test in how far nerve cells and glia cells are selectively influenced by Tf isolated from certain species and of different degree of iron occupancy, and in how far iron can replace Tf in the culture medium, the effects of holo and apo human and chick ovo Tf, and bovine apo-Tf, on viability and differentiation in serum-free cell cultures of embryonic chick brain and neuronal retina were determined.

The effects were dependent on the Tf used, the origin of the cells in the culture and the parameter measured. Peak stimulation of viability and differention was observed after the addition of chick ovo apo-Tf at concentrations close to 40 mg/l.

The effects of ovo-Tf could not be mimicked by the addition of iron.

Our data suggest that Tf may interfere with processes other than iron uptake.

The neurogenesis in the olfactory epithelium may involve two types of neuronal precursors: an immediate neuronal precursor which generates two postmitotic daughter neurons; and a neuroepithelial stem cell which is presumed to be the progenitor of the immediate neuronal precursor. Neurogenesis is ongoing even in adulthood. Ablation of the olfactory bulb causes neuronal death in the olfactory epithelium.

We cultured the olfactory epithelium neurons with olfactory bulb-, olfactory epitheliumand pituitary-extracts.

Only the olfactory bulb extracts were able to increase the differentiation of the olfactory neurons suggesting that their development is regulated by an unknown soluble factor released by their CNS target. We tried to determine whether any of the already known growth factor (NGF, FGF, TGF~, LIF, ...) had an effects on olfactory epi-thelium neurogenesis in vitro. None of the tested factor was specific for the survival or the differentiation of olfactory neurons.

The Amino-Terminus of Nerve Growth Factor Is Involved in the Interaction with the Receptor Tyrosine Kinase p140 IrkA *Kahle, P., #Burton, L. E., #Schmelzer, C. H. and *Hertel, C.; *F. Hoffmann-La Roche Ltd., Pharma Division, Preclinical Research, CH-4002 Basel, and #Genentech, Inc., South San Francisco, CA 94080, U. S. A.

The amino-terminus of nerve growth factor (NGF) is susceptible to proteolytic cleavage. A comparison of the bioactivity of highly purified fulllength recombinant human (1-118)rhNGF and N-terminal truncated (10-118)rhNGF revealed lower potency of (10-118)rhNGF with regard to early NGF responses in neuron-like PCI2 cells. Approximately fifty times higher concentrations of (10-118)rhNGF than (1-118)rhNGF were required to elicit the same extent of tyrosine phosphorylation of the NGF receptor tyrosine kinase p140 trkA, phospholipase C-ft. 1, and the extracellular signal-regulated kinase ERK1. A similar reduced potency for induction of the transcription factor c-Fos was observed with (10-118)rhNGF compared to (1-118)rhNGF. The lower potency of (10-118)rhNGF in triggering early responses correlated with its 50-fold lower affinity for PC12 cells, compared with full-length rhNGF (produced in CHO cells or E. colO. Whereas (10-118)rhNGF had a more than 300-fold lower affinity for the high-affinity receptor p140 trkA than (1-118)rhNGF, amino-terminal truncation of NGF changed its affinity for the low-affinity receptor p75 NGFR only slightly. These observations suggest that amino acids 1-9 of NGF are important for binding to the signal transducing receptor p140 trkA. Proteolyfic cleavage of the NGF amino-terminus, therefore, reduces its potency in starting several second messenger pathways leading to neuronal differentiation of PC12 ceils. Receptor-type tyrosine kinases are activated by growth factors and control processes such as cell proliferation, cell survival and differentiation. Tyrosine kinases are also important components of signal transduction processes occurring during the establishment of the vertebrate body plan. In Xenopus, expression of a dominant-negative mutant of the FGFR-I causes mmcation of posterior structures in the embryo (Amaya et al., Cell 66, (257) (258) (259) (260) (261) (262) (263) (264) (265) (266) (267) (268) (269) (270) . We have used RT-PCR to determine the repertoire of receptor-type t.yrosine kinases expressed in the early Xenopus embryo. Three different tissues were examined: unfertilized eggs to determine the maternal contribution, gastrulae (st. 10.25) for early zygotic expression, and dorsal blastopore lips representing nascent mesoderm. We have identified ten tyrosine kinase. Three of the clones encode presumably amphibian homologs of FGFR-4, PDGFR-a, and ECK. The remaining clones represent novel members of the EPH/ECK, PDGFR, TYK2/JAK1 and CSK subfamilies. We are currently characterizing the expression of these tyrosine kinases in the early embryo by RNase protection assays, Northern blots and in situ hybridizations. The pathogenicity of the c-erbB-2 oncogene was evaluated in transgenic mice by introducing a DNA sequence comprising the promoter-enhancer region of the MMTV LTR and an activated allele of the human c-erbB-2 growth factor receptor gene into the germ line of mice. Transgene expression was observed in kidney, lung, mammary gland, salivary gland Harderian gland and in epithelial cells of the male reproductive tract. All transgenie mice expressing the c-erbB-2 receptor died within four months. Histopathological analysis showed severe preneoplastic lesions, especially in kidney and lung. The renal tubules displayed focal dilatation and proliferation of tubular epithelial cells. Widespread hyperplasia of bronchial and bronchiolar epithelium was found in the lung, where pseudopapillary proliferation of epithelial cells narrowed the bronchial lumen. The mammary gland was the most dramatically affected organ. The glands of two parous mice were underdeveloped, lacking lobular-alveolar structures and were lactation deficient. A virgin mouse developed a focal adenocarcinoma infiltrating the mammary fat pad. The expression of the c-erbB-2 protein was always rcstricled to the proliferating epithelial cells. Transgenic males were sterile with epithelial hyperplasia in the epididymis, vas defereas and seminal vesicles. The transgene is not uniformly expressed in the tissues where thc MMTV LTR is transcriptionally active. Scattered transgene expression coincides with epithelial hyperplasia. Preneoptastic lesions in kidney and lung most likely cause organ failure and the early death of the transgenic mice. Serum-free aggregating rat brain cell cultures provide sufficient cell surface and paracrine interactions between neurons and glial cells for compact myelination. We are interested in the part played in these signalling pathways by protein kinases and have used a PCR cDNA cloning appro&ch to catalogue the protein kinase genes expressed by these cultures. Seven previously described receptor protein kinases were identified: IGF1-R, trk B, bFGF-R, cmet, Tyro2, Tyrol, Tyro4. We have now proceeded to characterize a novel gene which is highly expressed in the brain cultures during differentiation and myelination. Like Tyro 1 and 4 it has all the features characteristic of the eck gene family and the protein has been demonstrated to have protein tyrosine kinase activity.

We have used an affinity purified antiserum to a recombinant fusion protein to identify in the cultures a prominent phosphoprotein of 110 k molecular weight. This is in agreement with the open reading frame of 873 amino acids and the two possibilities for N-glycosylation sites. Expression studies are now under way to assess whether this gene has a role in myelination. We measured AP-1 and Oct-1 transcription factor activities throughout mouse mammary gland development. Very low AP-1 levels were found at all times. Oct-1 activity was high in the virgin gland, during pregnancy and lactation. An alarming decrease of Oct-1 activity was seen at day 2 to 3 of involution and Oct-1 activity was virtually undetectable between day 3 to day 12 of involution. Both, Oct-I mRNA measurements and in vivo labelling experiments showed that Oct-1 mRNA is present and Oct-1 protein is synthesised in the epithelial compartment of the mammary gland during involution. We found that the DNA-binding activity was reduced when Oct-1 prepared from lactating mammary gland or synthesised in vitro was phosphorylated by PKA. In vivo PKA enzyme activity in the mammary gland was found to be considerably high at all stages of involution. Experiments are in progress to evaluate whether proteases are also involved in the observed changes in Oct-1 DNA-binding activities. Data will be presented on the correlation between the expression of milk protein genes (g-casein, WAP) and possible markers for apoptosis (sulfated glycoprotein, TGF-13, Clone 54, c-jun and junD) in the involuting mouse mammary gland and the observed loss of Oct-1 DNA-hinding activity. In order to investigate the influence of biological factors on cell maturation and cartilage mineralization in the process of endochondral ossification it is necessary to isolate and separate the growth plate chondrocytes from their different zones, Using Percoll gradient centrifugation, we have successfully developed methods to separate five different viable subpopulations (A,B,C,D, and E) of chondrocytes from fetal bovine growth plates representing different in viv0 maturational stages. They can each synthesize type X collagen, an extraeellular matrix and, can subsequently calcify this matrix in a time dependent manner different for each subpopulation and depending upon their maturational stage at isolation. In this study we have compared the effect of T 4 in serum-free medium, with the effect of 10% FCS on in vitro maturation of the different chondroeyte subpopulations leading to expression of the hypertrophic phenotype. Thyroxine alone induced the synthesis of type X coUagen and matrix calcification (45Ca+2 incorporation) effectively as well as 10% FCS. Moreover, the hypertrophic phenotype was expressed more rapidly with T 4 when compared to FCS. These preliminary results indicate that T 4 alone is sufficient to induce the maturation of mammalian growth plate chondrocytes leading to expression of the hypertrophic phenotype which culminates in the calcification of the matrix.

Karin Sadoul, Philippe A. Halban and Dominique Rouiller

The four main cell types in the islet of Langerhans are A-, B-, D-and PP-cells, secreting glueagon, insulin, somatostatin and pancreatic polypeptide respectively. Islet architecture is highly ordered and both cellcell and cell-extr~icellular matrix (ECM) interactions must be important for the organization and function of the islets of Langerhans.

To establish techniques for characterizing the latter interactions we have investigated adhesion and morphology of RIN-2A cells, a transformed B-cell line, plated on different substrates. Cells adhered best to poly-l-lysine (44%) and laminin (41%) compared to fibronectin (9%) and collagen (8%). Cells stayed rounded up on fibronectin, poly-l-lysine and collagen, whereas on laminin cells were spread out.

In order to identify cell surface receptors for molecules of the ECM, we used homology PCR amplification for integrin subunits. By comparison with the GenBank, our preliminary sequences could be identified as the previously unsequenced rat homologues of the integrin subunits 131, c~3 and C~v.

Applying these methods to primary B-and non-B-cells will allow the characterization of islet cell integrins. Comparison of primary and transformed B-cells may give further insight into cell surface changes possibly typical of insulinoma tissue.

Sabine Koch-Schneidemarm, Peter Gehr*, Hans M. Eppenberger Instiute of Cell Biology, ETH H6nggerberg, 8093 Z~irich; * Department of Anatomy, University of Berne, 3000 Berne Adult rat cardiomyocytes (ARC) were cultivated on five different substrates: gelatine, fibronectin, laminin-nidogen complex (laminin), E8 lamininfragment, E1 laminin-fragment. Comparitive cell attachment assays have shown that ARC prefer adhesion to E8 laminin-fragment and laminin. It was shown by video time lapse (VTL) studies that, during the redifferentiation process of ARC in culture, the morphology of ARC grown on laminin, fibronectin and gelatine is undistinguishible, whereas the size of ARC grown on ES-fragment is larger, and when grown on El-fragment definigly smaller than ARC on the whole laminin protein. Immunostaining for vinculin combined with reflection contrast microscopy were used to visualize the focal contacts of ARC on these substrates. Quantitative measurements, done with the help of a test line system, show that the number of adhesion plaques/gm 2 on gelatine, fibronectin and laminin are about the same. On the E8 fragment more attachment sitesAtm 2 and on the E1 fragment less attachment sitesAtm2were counted than on whole laminin protein. This suggests that substrates influence the number of focal contacts. Correlating these results with the observations made in the VTL recording system, one can suggest that the number of adhesion sites/gm 2 increases in very flat and large cells.

In the mammary gland, the structural organization changes drastically depending on the functional state and requires a repeated degradation and reconstruction of the extracellular Matrix (ECM). Massive changes in tissue architecture are also hallmarks of tumor development. We are analyzing histochemically the expression of the ECM protein Tenascin and of the metalloproteinase styomelysin 1 in various stages of murine mammary differentiation and carcinogenesis. Both proteins are of stromal origin and surround growing epithelium at puberty and postlactational involution. Similar expression and distribution was found in early premalignant lesions of mammary glands from Wap-ras transgenic animals. In contrast to the well defined distribution and the transient presence during normal growth, the expression of both, Tenascin and Stromelysin 1 increases and expands with tumor progression. Thus, the unscheduled activation of a normal cellular program may establish aggressive tumor growth already in the earliest stages of neoplastic development.

Cell-substrate interactions regulate differentially cytochrome P-450 isoenzymes In cultured rat hepatocytes.

B. Saad, F. A. Scholl, and P. Maier Institute of Toxicology, ETH and University of Z0dch, 8603 Schwerzenbach

The influence of cell-substrate interactions on the preservation and inducibility of microsomal cytochrome P-450 isoenzymes in cultured rat hepatocytes was investigated. Hepatocytes were cultured on collagen type I (COL), laminin (LM), fibronectin (FN) or liver crude membrane fractions/collagen type I (CMF/COL). The relative contents of P-450 isoenzymes were measured in ELISA using monoclonal antibodies. Hepatocytes cultured for up to nine days on CMF/COL, retained their relative cytochrome P-450 contents at 1.5-3 fold higher than cells cultured on COL, FN or LM. After exposure of hepatocytes cultured on CMF to PB from day 3-6, CYP3A proteins were enhanced more than two-fold and, depending on the exposure level, CYP2B1/B2 increased 1.3 -6fold. After exposure to 3-MG, a 3-fold increase of CYP1A proteins was found in CMF/COL-but also in LM-cultures. These results indicate that CMF/COL as substrate in rat hepatocyte cultures enables the cells to respond qualitatively similar to that observed in the liver. Adult rat cardiomyocytes (ARC) in culture represent an ideal system to explore key questions of cell differentiation and cell contact. These cells do not undergoe cell division and grow by cellular hypertrophy. It is assumed that extending pseudopodia establish cell-cell contact, subsequently, tile intimate contact structure, called intercalated discs, is regenerated and finally electric coupling is restored, The N-cadherin isoform plays a key role in this adhesion process dependent of Ca 2+ in ARC. Cadherins are also involved in morphogenesis and it is believed that the precise regulation of cadherin expression at the quantitative as well as qualitative level is crucial. A major step forward to reveal the adhesion process was to adapt microinjection into ARC of cDNA cloned in an appropriate expression vector. In the present study, the expressed chicken N-cadherin is followed by immunostaining with a specific antibody which discriminates between homologous isoforms of different species. We report here that exogenous chicken N-cadherin is indeed expressed in ARC and specifically localize in the sites of contact. Cadherin overexpression does not disturb the precise localization and function of the endogenous homologue, moreover it is involved in contact and newly synthesized adherens junctions in ARC cocultivated with fetal rat cardiomyocytes. Finally, eDNA structure modification experiments are performed in order to elucidate as how cadherin localization and function are coordinated in heart muscle.

Koch, M., and Chiquet, M., Abt. Biophysikalische Chemic, Biozentmm der Universiffit, CH-4056 Basel Type XII collagen is an extracellular matrix protein associated with collagen fibrils in vivo. The molecule has three very large, extended noncollagenous domains and a short collagen tail. Its non-collagenous domains are characterized by von WiUebrand factor A (vWF A) domains and fibronectin type IN (F!N III) repeats. While the protein isolated from embryonic tendon has subnnlts of 220 kDa, we have recently purified from chick embryo skin fibroblasts a very large splicing variant with subunits of 350 kDa. These contain two additional vWF A and eight more FN HI repeats, as well as chondroitin sulfate chain(s). We isolated a mAb which recognizes the additional domain and used it to separate the large variant from the small variant of type XII collagen. By comparison, we found that the small form binds more strongly to fibronectin in solid phase assays than the large one. Conversely, the large form binds to heparin-Sepharose in 0.15 M NaCI while the small one does not. Both variants affect collagen fibril formation in vitro, but in a distinct manner. In embryonic chick skin, small type XII coUagen is localized to a defined layer everywhere in the dermis. In constrast, the large form is confmed to the basis of feather buds. Our results indicate that the two forms of type XII collagen have distinct functions in morphogenesis.

Anatomy, University of Berne, CH-3000 Berne 9, Switzerland, and +Institute of Physiological Chemistry and Pathobiochemistry, University of Mlinster, D-4400 Mtinster, Germany.

A novel 760 kDa dermatan sulfate proteoglycan has recently been isolated. It consists of two disulfide cross-linked core proteins (460 kDa and 300 kDa), each carrying one or very few dermatan sulfate chains of -20 kDa. (Breuer et al. 1991, J. Biol. Chem. 266: 13224-13232) . Three different antisera (anti-300, -460, 760 kDa) were raised and used to study its light and electron microscopical tissue distribution. So far, immunostaining of the proteoglycan (immunofluorescence and immunogold) could be observed in some (e.g. corneal epithelium and Descemet's membranes), but not all basement membranes (e.g. capillaries, striated muscle, and liver); and also in corneal stroma, skin, smooth muscles, and in EHS-tumor. In Descemet's membrane the proteoglycan specific staining was predominately observed next to the cell-surface of the endothelium.

Adhesion is essential for invasion of wounded soybean leaf tissue by the fungal pathogen Phytophthora megasperma f. sp. glycinea

Instltut fi~r Pflanzenbiologle, Zollikerstr.107, 8008 ZOMch Phytophthora megasperma f. sp. glycinea (Ping), race 1, was used to study the role of adhesion in colonization of soybean ev. Harosoy leaf disks through their cut edges. Invasion and colonization of the tissue was inhibited in the presence of Con A, IgG or D-mannose. Light microscopy revealed that only very few germ tubes attached to the ceils and attempted penetration of the tissue. Of several sugars tested only glucose reversed the effect of ConA or IgG. Neither of the three substances, Con A, IgG or D-mannose, inhibited germination or germ tube growth of the pathogen substantially. The results suggest that inhibition of infection and colonization is due to the inhibition of adhesion between the fungal parasite and the host cell wails, and that adhesion is an essential prerequisite for these processes. branches. We tested biotin-and digoxigenin-conjugated PHA-L for lectin blotting and histochemistry. Techniques using digoxigenin-conjugated PHA-L proved to be of higher specificity for lectin blotting and both more sensitive and specific for labeling of cell and tissue sections. In lectin blots of the parental HCT 116 line and the low invasive line HCT l16b a single reactive band of =162 kI)a was detectable. In the highly invasive line HCT 116a the intensity of this band was greatly increased and a second major band of =170 kDa occurred. Morphometric analysis of cell surface staining by PHA-L in electron microscopy revealed a significantly higher density of labeling in the highly invasiv cell fine. In sections of sporadic haman colonic carcinoma, invasiv parts of the tumor showed higher PHA-L staining compared to the normal colonic epithelium. Collectively, these data indicate that differences in PHA-L staining seem to be correlated with differences in metastatic potential and may be an adjunct technique in the prognostic evaluation of tumors. Supported by the Sassella Stifung Ztitich, Switzerland.

Brandenberger, R., Brubacher, D., and Chiquet, M., Abt. Biophysikalische Chemic, Biozentmm, CH-4056 Basel

Laminins comprise a family of multidomain extracellular matrix glycoproteins. The classical laminin isolated from mouse EHS tumor is composed of three polypeptides designated A (400 kD), B1 and B2 (220 kD). Apart from classical laminin, other isoforms have been found where the A chain is replaced by a related M chain or the B1 chain by an S chain.We isolated laminins from chick heart and gizzard. By rotary shadowing, cross-shaped and T-shaped particles could be observed in both preparations. On SDS-PAGE under nonreducing conditions, a pattern of three laminin bands at 700, 800 and over 10OO kD could be detected in the heart preparation. In contrast to the gizzard p.repara.tion: tOe heart lanai. "nin, after reduction, contained a 350 kD polypeplade winch is lmmunologxcally related to the M chain. We generated monoclonal antibodies against chick heart laminin. One mAb called 8113-3 selectively labeled the myotendinous junctions of the dermamyotome in the 6-day-old chick embryo and similar structures in the legs of older embryos. This mAb is directed against a polypeptide present in heart but not in gizzard laminin. By affinity chromatography using this mAb, we separated from other heart laminin variants an 800 kD isoform with the likely subunit composition M/S/B2. Interestingly, this variant promoted neurite outgrowth in vitro, although the S chain has been postulated to contain a stop signal to the growing neurite. Members of the S-100 calcium-binding protein family are good candidates as markers defining specific stages in mammary tumour progression. Their human genes are located in a cluster on chromosome lq21, a region which is often involved in rearrangements observed in breast cancer. Additionally CAPL, a member of this protein family, has been associated with metastatic behaviour of mouse mammary tumour cell lines.

We found a correlation between the expression pattern of CAPL and the pathogenesis of human breast cancer cell lines. In particular MDA-MB-231, a cell line for which translocations in lq had been found cytogenetically, showed a very high expression level. The expression of other $100 proteins of the cluster, however, did not vary dramatically, which suggests individual control mechanisms for the different members of the gene cluster. We will examine additional mammary mmour cell lines and will try to verify these results for clinically well studied primary turnouts. We are testing those cells, in which we see an altered expression pattern, for genomic aberrations in the gene cluster region by the use of Southern blots and Pulse field electrophoresis. Our preliminary study shows that S-too proteins and CAPL in particular are possible markers for a high transformation level of human breast cancer cells. Solid primary tumors can be removed by surgery and radiotherapy. However, recurring t~mors or metastases are often not eliminated by these treatments. We have designed chimeric molecules which are expected to fulfill therapy requirements imposed by small secondary tumors which may be spread throughout the body: A targeting part of the chimeric molecule which is called "DOG" (e.g. Epidermal Growth Factor) is crosslinked to a "SLEDGE" (e.g. a synthetic DNA) which carries a "LOAD" (e.g. a radioactive bisbenzimide derivative). An example of such a "DOG-SLEDGE-LOAD" vehicle is presented and its DNA "SLEDGE" is characterized.

Chantal W~lchli, Judith Trueb and Beat Trueb Biochemie I, ETH Zentrum, 8092 Zurich

We have isolated and characterized several overlapping eDNA clones for chicken collagen XW which span a total of 6.5 kbp. These clones contain an open reading frame of 5571 bp encoding the entire collagen X1V polypeptide. The predicted polypeptide has an estimated molecular mass of 205 kDa in its glyeosylated form. It is composed of 1857 amino acids which are arranged in 16 individual subdomains, including a signal peptide of 28 residues. The large amino-terminal globular domain of collagen XIV (NC3) comprises 1l of these 16 subdomains. Two of them are related to the A modules of yon Willebrand factor, eight show some relationship to the type Ig repeats of fibroneclin and one is similar to the NC4 domain of collagen IX. The carboxy-terrninal triple helical domain is composed of two collagenous segments (COL1 and COL2), which make up less than 14% of the entire molecular mass, and of two short noncollagenous domains (NCI and NC2). A detailed analysis of our eDNA clones indicates that colIagen XW exists in two alternatively spliced forms which differ by 31 amino acids in their NC1 domain. The variant form of the polypeptide contains therefore 1888 residues with a total molecular mass of 208 kDa. Our results demonstrate that collagen XIV displays a complex multidomain structure resembling the one proposed for collagen XII. We are interested in the early development of C. elegans. The mainly invariant celI lineage led people to believe that maternal cytoplasmic components were the major strategy to bring about specification in the early embryo. However. zygotic RNA transcription begins as early as at the 8 to 16-cell stage and is appreciable one or two cell cycles later. Such a stepwise activation of transcription is observed and well studied in D. melanogaster where pattern formation genes, among them the HOM-C (homoobox-complex) genes, are expressed before general transcription begins. The early activation of transcription in C. elegans and the aiscevery of a large number of C. elegans homologs to genes important in other developmental systems, provide growing evidences for postmaternal contribution to cell fate detemahaation.

We are working on ceh-13, a C. elegans homeobox conmirting gene with labiallike smtetore. It belongs to a cluster of four genes with sequence homologies to genes in the same relative order in the HOM-C in Drosophila and vertebrates. We are analyzing its pattern of expession by monitoring 13-galactosidase activity in stable transgenic lines. Our preliminary results suggest that ceh-13 plays an important poatmatemal role in the C. elegans development. Ftrrthennore, we are raising a polyclonal antibody to confh'm the lacZ staining pattern and to possibly identify a maternal product. During early Drosophila development, the midgut fomas from two different endodermal primordia located near the embryonic poles, outside the expression domains of most segmentation genes. Despite this apparent unsegmented origin of the endoderm, the larval midgut epitheliurn shows striking functional and morphological differentiation along its anteroposterior axis. Recent experiments have provided evidence that region specific gone expression in the midgnt epithelium is dependent on the expression, in the adhering visceral mesoderm, of homeotie selector genes that trigger expression of growth factor homologues. The midgut therefore represents an ideal system to study inductive events across germ layers in the Drosophila embryo.

We have recently found that the expression of the Drosophila POU-box gone pdm-I is repressed m two regions of the midgut, an anterior and a central region. Repression iu the central domain is dependent on induction trough the visceral mesoderm. However, pdra-t repression uses a different genetic cascade then that previously reported to be involved in the induction of the homeodc gene lab in the midgut. Interestingly, the anterior repression domain ofpdm-1 is independent of the expression of known homeotic genes and characterized genes encoding secreted signalling molecules. To study in more detail the spatial regulation of pdm-1 and lab expression through cell-cell communication, we have recently initiated a novel genetic screen in order to identify additional component involved in this process. These studies should help to elucidate the molecular basis underlying the transmission of positional information across germ layers, and might also provide a key to the understanding of induction mechanisms in ver~bmte development. Expression of gab, wg, and en is activated by pair-rule proteins but soon thereafter maintained by mutual regulatory interactions.

Expression of the gab gene, which encodes a paired-domain as well as a homeodomain, completely depends on wg, the homolog of the murine Wnt-i gene, and is reciprocally required for maintaining wg, but not an.

We further present evidence that gab specifies the ventral denticle pattern, the most conspicuous metameric feature of the larva, by suppressing denticle formation in gab-expressing cells and their neighboring anterior cells. The neurogenic locus E(spl) is defined as a genomic region containig at least 15 embryonically active transorition units. Only one of them, the lethal(3)groucho gene is vital, whilst none of the others appear to be essential for fly survival. Seven of this genes are highly similar and share a basic helix-loop-helix motif of transcriptional regulators. These E(spl) bHLH genes are thought to be functionally redundant. Transformation and complementation analyses have shwon that beth, the l(3)gro and the bHLH genes are indeed involved in Drosophila neurogenesis. For proper function the bHLH genes require a full maternal complement of l(3)gro. Maternal l(3)gro transcripts are distributed ubiquitously in early embryos, whereas zygotic l(3)gro expression becomes restricted to the peripheral and central nervous system and its primordia, However, the nuclear l(3)gro protein can be found ubiquitously throughout embryogenesis.

Analysing the transcriptional regulation of the l(3)gro gene in lacZ reportergene constructs we found zygotic regulatory elements located around the transcription start and additionally within parts of the first intron. Interestingly, the putative regulatory intron sequences contain target sites for bHLH proteins, rendering an involvement of E(spl) bHLH proteins in l(3)gro regulation possible.

B. Wemer, A. Preiss, Abt. Zellbiologie, Biozentrum der Universit~it Basel, Klingelbergstrasse 70, 4056 Basel

The segregation of neural from epidermal precursor ceils requires a cell interaction mechanism that distinguishes both cell types and directs their respective differentiation. Involved in this postulated signal transduction pathway is the neurogenic Enhancer of split (E(spl)) gene complex. Its neurogenie function is composed of at least seven very similar.genes encoding proteins with a basic helix-loop-helix (bHLH) motif sharing homology with mammalian transcription factors. They are candidates for repressors of the proneural genes thereby specifying epidermal fate in the expressing cells. Ectodermal differentiation might well be dependent on the collective activity of the bHLH genes, despite their supposedly redundant fi.mction. Therefore, we have started a detailed study of the transcriptional activity of the individual E(spl) bHLH genes. The spatial and temporal expression pattern is highly dynamic and a direct comparison of the staining patterns is extremly difficult.

Preliminary data suggest that the spatial activity of the various E(spl) bHLH genes is only partially overlapping. In order to increase resolution both spatially and temporally, the embryos were accurately staged and folded open by dissection. A newly developed double labeling technique might enable us to superimpose expression patterns of two E(spl) bHLH genes simultaneously and relate it to that of e.g. engrailed as positional marker. The segmental organization of the Drosophila head is achieved by a flow of positional information from maternal gone products to the zygotic gap genes. Recent genetic analysis has identified three new gap genes involved in head development. One of these genes is the empty spiracles (eras) gone.

Mutations in empty spiracles cause severe defects in the head and the Filzk6rper at the posterior end are missing. The eros gene has a DNA binding domain, the homeodomain, and two activating domains, a proline-rich domain and an acidic domain consistent with the role of the eros gone as a transcription factor. A 2.4 kb RNA is expressed in two phases of embryonic development. First expression is seen at the blastoderm stage in a single anterior band, correlating with its function as an anterior gap gene. The gone is expressed in the preantennal, antennal and intercalary segments and is required for the development of the antennal sense organ, the optic lobe and parts of the head skeleton. Later during embryogenesis eros is expressed in the posterior spiracles as well as lateral regions of each segment where the tracheal pits form and lateral neuroblasts originate. Using g-gal fusions we could identify at least five different regulatory elements in the eros promotor region responsible for tissue specific expression of the gone. Since eros expression is dependent on the anterior and the terminal system, we used our g-gal fusions to identify the target sites for this regulation. A 300 bp element responsible for the early expression which is dependent on the maternal gene bicoid, the key gene of the anterior system, was identified and studied in detail. This analysis should allow us to elucidate the molecular mechanisms by which the morphogen bicoid regulates subordinate target genes like eros in the Drosophila embryo. In order to understand cell proliferation in imaginal discs of Drosophila rnelanogaster, we are studying the gene lethal(3)discs overgrown(doG).

Depending on the combination of dOG alleles, disc anlagen degenerate or develop to hyperplastic imaginal discs. Genetic and molecular analysis allowed us to restrict the locus to a region of 55kb, which covers at least tree genes. Four EMS alleles of dco were analyzed on genomic Southern blots, but showed no obvious mutation. Recently, we got a fifth X-ray allele (from John Merriam, UCLA), whose phenotypc is not analyzed yet; but it allowed us to exclude one candidate gene. A second candidate contains a homeobox. A corresponding eDNA was cloned from a total disc eDNA library. In situ hybridization with this gene on whole-mount embryos revealed transcripts in the developing nervous system, mesoderm and endoderm. The third transcript will be analyzed as well. We have studied the effects of 22 lozenge (lz) mutations on the antennal sensilla. The antenna of strong lz alleles lacks basiconic sensilla (BS), but shows a significantly increased density of coeloconic sensilla (CS). Intermediate alleles have few BS, they exhibit a highly increased density of trichoid sensilla (TS) but a normal CS density. BS on the maxillary palps are weakly affected even by strong Iz alleles. Most of the strong and intermediate mutations are partially dominant for the antennal phenotype. None of 12 selected alleles complement each other indicating that they define a single cisu:on. Temperature shifts of the lztsl allele showed that gene activity is crucial around the larval-pupal transition period. Applying restrictive temperature before pupariation results in a 'novel' phenotype that is characterized by a dramatic decrease of the TS density, whereas restrictive temperature after pupariation leads to a 'normal' intermediate pheno- Zebrafish embryos with a degenerate oligonucleotide led to the discovery of two new members of the DVR subfamily. One of these is conserved throughout evolution and differentially expressed during embryonic development. Whole mount in situ analysis detected the message uniformly in 4-cell embryos. During rate gastrulation the message disappears from the prospective dorsal region of the embryo. In later stages a specific signal is found in distinct parts of the brain. Using a PCR based approach we could isolate three activin clones in Medaka two of which display high homology to known activins in higher vertebrates. The three clones represent three distinct genes that are expressed in ovary and testis. The Zebrafish activin 8B homologue was isolated from a Zebrafish cDNA library using the corresponding Medaka probe. When expressed in COS cells the protein exhibits a strong mesoderrn inducing potential in the animal cap assay. Variants with the ability to interfere with the wildtype activin were generated. The dominant suppression of the mesoderm inducing potential of the wild type activin rib with these mutants could be demonstrated in the animal cap assay. These mutants will allow us to investigate the role of activins as well as of the newly discovered members of the TGF8 family in transgenic fish in vivo, We describe the constraction and characterization of a polyoma-based vector that is maintained extrachromosomally in mouse embryonic stem (ES) cells. Vector pMGD20neo contains the polyoma (Py) virus origin of replication harboring a mutated enhancer (PyF101), a modified gene for Py large T antigen and a neomycin resistance gene. After electroporation, the vector is replicated in embryonic stem (ES) ceils and is found as an extrachromosomal element in 15% of G418 resistant clones. The size and restriction patterns of the extrachromosomal DNA were identical to the transfected DNA. However, the vector or parts of it had integrated into the genome, suggesting that the extrachromosomal DNA might derive from one or more integrated copies. In one clone (1.19), the vector was maintained at 10-30 copies per cell for at least 22 passages (74 days) in the presence of G418. After supertransfection, clone 1.19 was able to replicate and maintain a second plasmid carrying the Py origin of replication but lacking a functional gene for Py large T antigen. Two vectorcontaining cell lines (1.19 and 1.24) have yielded several chimeric animals upon microinjection into host blastocysts and reimplantation into the uterus of pseudopregnant mice, indicating that the expression of Py large T antigen and the presence of extrachromosomal elements do not prevent the ES cells from populating an embryo. This system offers an useful tool to manipulate and analyze gene expression in ES cells without altering their pluripotential status. Ultrastructural distribution of hnRNPs, snRNPs and of ribosomal proteins PI/P2 and L7 was studied during mouse spermatogenesis and spermiogenesis using specific antibodies and immunoelectron microscopy, hnRNPs and snRNPs were identified until the spermatid elongation phase when RNA synthesis is known to cease. Labeling for ribosomal proteins was no longer detectable during the cap phase. Nucleoplasmic RNPs first occurred in nuclear structural components comparable with those described in somatic cell nuclei. In later spermiogenic stages a new type of nuclear component,resembling perichromatin granules, was observed, often in clusters. Since these components were labeled with probes recognizing nucleoplasmic snRNPs, a non-nucleolar origin of these granules is suggested.

The expression of ealretinin in mesenchymal cells during development. E. Kiraly, D.M. Vogt, A. Poncino, M.R. Celio, Institut for Histologie und allgemeine Embryologie, Universitat Perolles, CH-1700 Fribourg Calretinin (OR), an EF-hand Ca2+-binding protein, is known to be a neuronal marker in adult mammals. During development we observed CR-like immunoreactivity in mesenchymal cells located in different parts of the growing embryo. The staining was strong around somites, vessels and in the serous membranes of both the chick and the rat. Furthermore in the rat it was present in the condensed mesenchymal tissue of the growing limb buds, nipple, vibrissae and the regions of body-bending. CR positive mesenchymal cells were mainly found in regions which are involved in rapid growing and remodelling processes. In these regions the cells express the cellular binding proteins for retinoic acid. Recently it was shown that CR occurs in the mitotic spindle in a cell line in culture (WiDr), indicating a function for this protein during the cell cycle. Furthermore, CR expression in these cells was shown to be dependent on factors in the serum which are removed by the stripping of the serum with activated charcoal (e.g. steroids, thyroids, retinoids). These observations led us to the conclusion that CR is specifically expressed in mesenehymal cells which are involved in the growth of the body along certain axes. It could be one of the intracellular proteins which control cell proliferation under the control of steroids.

Intussusceptive microvascular growth in the chicken chorio-allantoic membrane may be implemented by capillary fusion Patan, S., Haenni, B. and Barri, P.H. Institute of Anatomy, University of Berne, Berne, Switzerland Intussusceptive capillary growth is a new principle of microvascular growth originally described in the lungs of rats (Caduff et at., Anat. Rec. 216, 1986; Barri and Tarek, Anat. Rec. 228, 1990) and relevant for various organ systems @aLan et al., Arch. Histol. Cytol. 55, Suppl, 1992) . Instead of sprouting the capillary network expands by insertion of slender transcapillary tissue pillars that give rise to full-size intercapillary meshes. In the lung, pillar birth was induced by disk-like interendothelial contacts between opposite capillary walls. Thin pillars were also observed in the chicken chorio-allantoic membrane (CAM) microvasculature by ia vivo and transmission eleclxon microscopy (TEM; Patan et al., Anat. Embyol., in press). TEM-analysis of horizontal serial sections of the CAM aged 7 days suggests nowa mode of pillar formation related to capillary fusion: The intervening wall between two juxtapposed capillaries is successively thinned out on beth sides of a centrally located core of a tissue pillar that is formed by collagen fibrils ensheared by extensions of endothelial-like cells. Following fusion of the endothelial leaflets the wall opens up on both sides of the pillars core which results in a free intraluminal tissue pillar. These findings suggest that there are many different modes of implementation of intussusceptive growth. The results also indicate that pillar formation is closely related to growth and remodelling of the vascular system. Tyrosinase (EC 1.14.18.1) is regarded as the key enzyme in melanin synthesis, and is expressed in the retinal pigment epithelium, a cell layer derived from the optic cup, and in neural crest-derived melanocytes of skin, hair follicle, choroid and iris. The tymsinase gene has been cloned and shown to map to the well-characterized c-locus (albino locus) of the mouse. Fusions of tyrosinase 5' sequences to a CAT reporter gene defined the regulatory region of the gene to 270bp 5' of the transcriptional start site.Further studies demonstrated that a functional tyrosinase minigene containing either 5.5kb or 270bp of 5' sequence was able to rescue the albino phenotype in transgenic mice. The transgene was expressed in a cell type-specific manner in skin and eye. During development of the mouse, the tyrosinase gene is expressed in the pigment epithelium at day 10.5 of gestation, and, in the hair follicle, from day 16.5 onwards. This developmentally regulated expression is largely reproduced in transgenic mice. Further studies will concentrate on the promoter region to identify cis-regulatory elements and the possible importance of further upstream sequences. Programmed Ceil Death is involved in many developmental and tissue organizational processes of cells which have fulfilled their biological roles. This is true for over 90% of the secretory epithelial ceils in the mammary gland which die a few days after weaning 9 Similarly, most prostatic epithelial ceils die 2-3 days after castration of the mature male. To define the early processes leading to apoptosis, we took advantage of a well-described subtraction protocol in combination with a novel, so called "coincidence" screening method. 1st strand cDNAs deriving from a) 2-3 days involuting mouse mammary glands and from b) 2-3 days post-castration prostates were hybridized in solution with mRNAs derived from the respective pre-apoptotic organs. Apoptosis-relevant cDNAs from novel transcribed genes remain single stranded and can be separated from the excess of cDNA/mRNA hybrids via hydroxylapatite chromatography. Single stranded fractions from "a" and "b" were individually used as probes for the coincidence screening of double lifts from a mouse involuting mammary gland library. By isolating only plaques which could be detected with both probes, we identified several genes which are strong candidates for participating in programmed cell death. The cDNA's of full length human BI,4 galactosyltransferase (gal-T) from HeLa cells and =2,6 sialyltransferase (sialyl-T) from HepG2 cells has been cloned by PCR, ligated into a yeast/E.coli shuttle vector and placed under the control of a constitutive PHO5 promoter. Both enzymes were expressed in yeast strain and shown to be membrane-bound and enzymatically active. Immunoprecipitation of both enzymes from 35Smetabolically labeled yeast cells revealed correct translocation, N-and probably O-glycosylation. GaI-T was purified from detergent yeast extracts and shown to be kinetically similar to the human milk enzyme. Product analysis of both gal-T and sialyl-T activity by IH-NMR revealed gal Bi-4glcNAc and Neu5Aa2-6LaeNAc, respectively. Thus, we conclude that yeast is a suitable host for the heterologous expression of mammalian glycosyltransferases. Supported by grant 2305.1 to EGB of the KWF.

Chantal Wtilchli and Beat Trueb, Biochemie I, ETH Zentrum, 8092 Ziirich.

Interstitial collagen fibrils form the supporting scaffold of most tissues and organs. A new group of collagens, unable to build fibrillar structures by themselves but intimately associated with collagen fibrils, has recently been described and termed FACITs (fibril-associated collagens with interrupted triple-helices). This group includes, beside type IX collagen, the two closely related collagen types XII and XW. These collagens seem to occur only in tissues containing type I collagen fibrils to which they may bind through their conserved COL1 domains.

Using three specific cDNA probes we have begun to investigate by in situ hybridization the tissue distribution of type XII and XW collagen relative to that of type I collagen. In tissue sections of 17-day-old chick embryos we observed expression of collagen XII and XW in the perichondrium and periosteum, whereas the bone matrix was negative. Like collagen I, collagen XIV is present in the connective tissue structures associated with skeletal and cardiac muscle and in the walls of small blood vessels; these tissues, however, seem to be devoid of collagen XII. Because of its more restricted tissue distribution, we conclude that collagen XII may serve a more specialized function than collagen XW which shows the same distribution pattern as type I collagen.

HETEROLOGOUS EXPRESSION OF SPECTRALLY ACTIVE HUMAN CYTOCHROME P450 2D6 Kempf, A. and Meyer, UA., Department of Pharmacology, Biocenter of the University of Basel, CH-4056 Basel Cytochrornes P450 are hemoproteins which catalyze the oxidation of a large variety of endogenous and exogenous chemicals. Numerous isozymes have been isolated and characterized by various means. The 3-dimensional structure is only known for bacterial (soluble) forms of P450, but not for the membrane-bound form of higher organisms, The human P450 2D6 is one of the best characterized isozymes and is involved in hepatic metabolism of over 30 drugs. Although it's functional and catalytic properties are well known, the 3-dimensional structure of the protein and the exact reaction-mechanisms remain unknown. --We are presently developing expression systems with the aim to produce large amounts of P450's in order to study their structure and function. --Transient expression in COS-I cells is conveniant and well established for functional assays, but the amount of expressed protein is so low, that it usually cannot be detected spectrally. --We therefore tested bacterial expression of CYP2D6. In an attempt to produce a soluble P450 which can easily be purified, we replaced the amino-terminal 30 amino acids, which presumably are involved in membrane anchoring of the protein, by a [His]6tag. Spectral analysis revealed, that up to 3 mg of home-containing P450 protein per I of bacterial culture could be produced, but only in the presence c! ~-,~LA (a heine precursor), Preliminary data indicate, that the protein was not associated with the membrane fraction but could be purified from the soluble fraction using NF § affinity chromatography and DEAE-anion exchange chromatography. The major problems encountered in this system were protein aggregation and a high proportion of apoprotein. --We are presently testing insect cell culture (baculovirus), a eucaryotic systern, that has the potential for producing large amounts of protein.

Ruepp, B., and Wermuth, B., Universit~t Bern, Chef~isches Zentrallabor, Inselspital, OH-3010 Bern

The genes of various detoxication enzymes, e.g. NAD(P)H:quinone reductase (NQR), contain transcription regulatory elements in their 5'-untranslated regions. These DNA-sequences exhibit enhancer activity in response to endogenous and foreign compounds such as polycyclie aromatic hydrocarbons and phenolic antioxidants. In addition to NQR, human hepatoblastoma (HepG2) cells contain a second quinone reductase: carbonyl reductase (OR). To investigate whether xenobiotics modulate the transcription of the CR gene similarly to the NQR gene, HepG2 cells were exposed to 3-methylcholanthrene (3-MC), 5-naphthoflavone (~-Nf) and t-butylhydroxyanisole (BHA). The transcription rates were estimated indirectly from the amount of the expressed proteins on Western blots and from the menadione reducing activity of the cell extracts using turin and dicoumarol as specific inhibitors of CR and NQR, respectively. After 3-MC and 5-Nf treatment NQR protein and enzyme activity increased time-and dose-dependent up to 2fold whereas CR was hardly affected. The results indicate that the expression of CR is regulated by mechanisms different from NQR and other detoxieation enzymes. In most mammals expression of lactase activity is highest during the suckling period and falls off to a low residual level after weaning. The regulatory mechanisms of surface expression of lactase are unknown. Earlier studies have suggested that posttranslational events may be involved (Sterchi et al. J. Clin. Invest. 86:1329 , 1990 . We have studied synthesis and processing of lactase in rabbits of different age using an organ-culture system. Enzymatic activity and lactase-specific mRNA were also assayed. Enzymatic activity was highest during the suckling period and declined sharply after weaning. Concommitantly with this decline intracellular maturation of lactase was retarded, leading to the presence of an increased amount of precursor lactase. In addition, the ratio of lactase-specific mRNA/specific activity increased 4-fold after weaning. This suggests that relatively large amounts of RNA were still synthesized in post-weaned rabbits with low lactase activity. These data are further evidence that post-translational events may have a regulatory role in lactase expression. In M. lam. the apcC gen encodes for the Lc 8-9. A 282 bp fragment was obtained by directed PCR-amplification from the apcABC operon. Primers were synthesized to introduce an NcoI site at the ATGstart codon, thus inserting the fragment into the expression vector pET3d. The construct was grown in the expression strain BL21(DE3)pLysE and induced by the addition of IPTG. Inclusion bodies were produced after induction, which were shown to consist of the Lc so9 by N-terminal sequencing over 31 steps and by SDS-PAGE. Reconstitution of the (~)3APCLc 8"9 complex was performed both with LC 8"9, ~APC, ~APC isolated from M. lam. as with overexpressed Lc 8-9 The two complexes have been shown to be identical with the native complex by means of sequence, SDS-PAGE, UV/VIS-, fluorescence-and CD-spectra. NMRexperiments did not show significant structural elements for the Lc s.9 .

Sidler, W.,R~egsegger, U. and Zuber, H., Institut ffir Molekularbiologie und Biophysik, ETH-H6nggerberg, CH-8093 Zfirich, Schweiz.

The cpcG2 gene encoding for one of the four rodcore linker polypeptides from the phycobilisome of the cyanobacterium M. laminosus PCC 7603 was overexpressed in E. coli BL21(DE3) cells, using the T7 RNA-polymerase system. LCR 29"5 was successfully reconstituted with (u~) 3 PC to the (u~) 3PC'LRc 29'5 and with (a~) 3 AP 9 Lc 8 -9 to the the rod-core complex (u~)6PC.LRc 29-5. (U~)3AP.Lc 8.9. A redshift of 16.4 nm in the absorption maximum, characteristic for a correctly reconstituted (u~)6PC-LRc 29.5 complex was observed. The reconstitution of (u~) 6PC. LRC29.5. (U~)3AP-Lc 8"9 resulted in a blueshift of the emission maximum at 659 nm and was only 3 nm minor than (~) 3 AP 9 LC8.9.

Coulin, F., A/let, B., Losberger, C. and Wells, T., Glaxo Institute for Molecular Biology S.A., CH-1228 Plan-LesK)uates/Geneva.

We have constxucted a librca-y of peptides to find sequences which specifically bind to antibodies and other receptors. A peptide cassette consisting of ten random amino acids flanked by two cysteines was expressed at the amino ten'ninus of the adsorption protein (pm) of the M]3 bacteriophage. The two flanking cysteines were added such that under o:ddising condition the peptide expressed would have a cyclic conforma~on. It is therefore possible to model its three dimensional structure. In addition, since the conformation is restricted it is possible to determine the three dimensional s~uctu.re of the epitope or receptor binding sequence using multidimensional NMR techniques, Results using an antibody against the reverse franscriptase enzyme of the virus HIV will be discussed. Phosphomannose isomerase (PMD catalyses the interconversion of mannose-6-phosphate and ~ructose-6-phosphate, which is the first step in the synthesis of glycoproteins, The enzyme therefore plays a key role in fungat cell wall biosynthesis, The enzyme is a ~nc dependent metalloenzyrne, but it is also inhibited by the addtion of excess zinc. Renaturation of the enzyme by removal of the denaturant, either by in.finite dilution, dialysis or stepwise dilution, does not result in any regain of enzyme activity. Re-activation has been found to depend on the inclusion of two factors, the reducing agent, dithiothreitol, and the zinc metal ion. Incl~L, don of the substrate has no effect. The effects of temperature, pH and ionic strength have been investigated. The basal level of in v/fro refolding is increased by the inclusion of the bactenal chaperones. NO effect is observed from the inclusion of a non-specific protein such as BSA. Ferredoxin-NADP+-oxidoreductase (FNR), a chloroplast enzyme, was used as a model to study the cross-reactivity of anti~peptide polyclonal antibodies with the parent native protein, Here we report results concerning the highly mobile and exposed N4erminal segment of the protein, which is invisible in the crystal structure of FNR. Three partially" overlapping peptides covering 8, 10 and 20 Nterminal residues of FNR ($8, St0, $20, respectively) were synthesized, Rabbit antisera against these peptides were tested by solid-phase and solution-phase immunoassays. Of the three antisera only the serum against $20 reacted with native FNR to a significant degree. This same serum did not react with FNR lacking 8 N-terminal residues, and the reaction with $8 and $10 was very weak. The data suggest the following: 1. Peptide $20 may possess a folded structure and, therefore, give rise to conformation specific antibodies. To understand the relaOdonship between s~racture card function of C-proteins. Moleculcu modelling techniques were used to construct a model by homology and compututional techniques to understund the specificity of the ctc~dve site for GDP, GTP gnd other unologues. We have constructed a pur~al model of the alpha subunit of C-q from mouse using homology modelling -with the 1.35/~ H-ROS 21 COordinates (Pod et cd., 1990) . No attempt wgs made to model those regions present in C--proteins but ubsent in ras. AMBER was used to t:~rgrnetize both ~ and Mg 2+ and deck them into Gq. A few steps of energy ~'on was then cgnied out to !field a low energy structure. ~ from mouse brQi~ has been cloned (Sfr~ culd Simon. 1990) (rod well expressed in E. co// (in which the protein is present in the inclusion bodies) and in insect cells with bac,/lovirus (in which it is in the soluble phase). C.~ was extracted from insect cells and purified on DEAE, hydroxyapatite, MonoQ and gnfibody af~_m'ty columns. The active subur/t was measured using a 3H-GTP binding assay. Retinol-binding protein (RBP) is the specific blood carrier for the transport of vitamin A (retinol) from the liver stores to the target tissues, Serum of patients with chronic renal failure (CRF) shows increased level of RBP compared to normal human serum (NHS). The amino acid sequence of the RBP described until now differs from gene's sequence by one amino acid residue, We purified RBP species from both NHS and CRF sera and we compared their physical properties (retinoid binding. PAGE-immunoblotting and masses). The level of apo-RBP in CRF as analyzed by PAGE-immunoblotting was strongly increased and its retinol affinity was altered in comparison with NHS. SDS-PAGE-immunoblotting showed two major forms of RBP in patients with CRF whereas only one was detected in NHS. Mass spectrometry of RBP showed the presence of three molecular forms which expression varies significantly between NHS and CRF. One of these forms is in accordance with the amino acids sequence deduced from the gene, suggesting that the two others forms are post-translationally processed. Increased levels of RBP concomitantly to altered expression of RBP species in CRF might interfer with retinol metabolism and could explain a defective delivery of vitamin A to the target cells. It has always been postulated that, for sterical reasons, substitutions of a glycine of the repeating Gly-X-Y triplets of the collagen e-chains would interfere with collagen triple helix formation. We therefore measured the kinetics of triple helix formation of procollagen I in fibroblast cultures from controls and 7 patients with osteogenesis imperfecta (OIL a heritable connective tissue disorder caused by structurally abnormal collagen I. After a 4-mln pulse-labelling with 3~S-methionine, the appearance of proteo~yticaHy stable and thus helical collagen molecules was followed for variable chase times; in controls, 50% of the molecules were fv(lyhefical after 14 rain, In 6 of the 7 el cell strains, all having a single Gly~Cys substitution at positions 94, 223, 526, 691, 748, and 988 of the helical domain of the ol (l)-chain, respectively, triple helix formation of abnormal collagen containing two mutant el(])-chains was determined by the appearance of el(I}dimers, which appeared later than the normal chains, Their delay ranged from 5 to 60 rain and correlated inversely with the thermal stability of abnormal collagen containing o1(I)-dimers. Folding time and melting temperature of procollagen I in the seventh cell strain with a GIy~Cys substitution at position 1017, located outside of the helical domain in the C-terminal telodeptide, were normal. Here, we directly demonstrated the hitherto postulated delay in the zipper-like folding of abnormal collagen caused by structural defects affecting the helical part of the molecule, which is responsible for the overmodification of the o-chains N-terminal to the site of the mutation. The stability of coiled-coil (x-helix dimers dependens on hydrophobic interactions between leucine side-chains in the inner core of the the dimer structure. We used synthetic peptides, based on a 29-residue sequence [1} to study the influence of sequence changes on coiled-coil stability. Exchange of two leucines by alanine produced a less stable coiled-coil. The structure collapsed to random coil when two residues of the solvent-exposed surface were replaced by proline. Stability was tested by CD spectroscopy in the far UV region. Fluorescent donor (coumarin) and accepter (fluorescein) groups were linked to the N-terminus to follow association and dissociation of the coiled-coil by fluorescence energy transfer. The fluorescent peptides can be used to study the influence of sequence variations and solvent parameters on the dynamics of coiled-coil formation. The photosynthetic apparatus of Rp. marina consists of the light-harvesting core complex (B880) where the light energy is trapped and the reaction center (RC) where the energy conversion to chemical energy takes place. We purified the B880-complex by ammonium sulfate precipitation and subsequent ion exchange chromatography. The resulting RC free protein complex was crystallised in two dimensions by microdialysis. The appearing crystals exhibited a hexagonal lattice with a lattice constant of 102 + 3/~. The optical difraction pattern showed spots up to 17/~ resolution. Image processing of these highly ordered areas revealed a sixfold symmetry of the ring like structure of the B880-complex.

The photosynthetic apparatus of the purple non-sulfur bacterium Rhodocyclus tenuis DSM 109 was found to be composed of two types of light-harvesting pigment-protein complexes, i) a core (B890) complex that surrounds and interconnects the membranebound reaction center, it) a pedpheral (B800-868) complex that is associated with the core complex. Experiments were performed to optimize conditions for the solubilization of both types of complexes by applying a combination of several different detergents like octylglucoside(OG), LDAO and SDS to the photosynthetic membrane. The two types of pigment-protein complexes were separated by subsequent sucrose gradient centfifugation, and, were analyzed by means of absorption and NIR circular dichroism spectroscopy. Separation and assignment of the polypeptide components of the peripheral (B800-868) and the core (B890-RC) complexes were performed by SDS gel electrophoresis, reverse phase HPLC and protein chemical analyses. The determination of the amino acid sequences of the apoproteins is in progress. Lehmann, R.P., Brunisholz , R.A. and Zubor , H. Institut for Molekularbiologie und Biophysik, Switzerland The green photosynthetic bacterium Chloroflexus aurantiacus has a strong near infra-red absorption maximum at 740 am. This absorption band is caused by the antenna system, the chlorosome, that comprises several thousands of bacteriochlorophyll c molecules and four polypeptides with the Mr 18, 11, 5.8, 5.7 kDa. We purified the 18 and 11 kDa polypeptides on reverse phase chromatography and started with the sequence determination. The role of the different polypeptides for the organisation of the photosynthetic pigments in the entire chlorosome is not understood. With the use of limited proteolysis on isolated chlorosomes we obtained changes of the circular dichroism characteristics. The alterations of the polypeptide pattern where analysed on SDS-PAGE and reverse phase HPLC.

Monovalent cation binding sites in the crystal structure of dialkylglycine deearboxylase: a structural explanation for the effect of K + and Na+ on enzyme activity Erhard Hohenester, Michael D. Toney, Sandra W, Cowan & Johan N. dansonius Biezentrum der Universit&t Basel, Abt Strukturbiologie, CH-4056 Basel Dialkylglycine decarboxylase (DGD) is a bacterial PLP-dependent enzyme that catalyses the oxidative decarboxylation of (z-dia]kylamino acids such as 2methylalanine or isovaline. DGD activity depends on the presence of K + ions, whereas Na § ions show an inhibitory effect. Recently, we have solved the crystal structure of DGD by multiple isomorphous replacement and solvent flattening. The structure is now refined (R = 0.178, good stereochemistry) to a resolution of 2.1 A. in spite of low sequence homology the overall architecture of DGD is very similar to that of aspartate aminotransferase (AAT). We have identified two putative metal binding sites in the crystal structure of DGD that are not present in AAT. Site 1 is located in close vicinity to the active site and can be occupied by either a K + or a Na + ion depending on the buffer composition. Replacement of K + by Na + affects the conformations of the active site residues $80 and Y301" suggesting a structural explanation for the effects of these ions on enzyme activity. Site 2 is found on the periphery of the molecule and is located in a tight turn at the C-terminus of an c~helix. Although we cannot completely rule out the possibility that this site is a highly unusual Ca ++ site, we believe that it is occupied by a Na + ion in the crystal structure of DGD thus representing a novel type of metal binding site in a protein.

Keller, T.A., Lustig,A. and Rosenbusch,J.P. Biozentrum der Universit~t, Abt.Mikrobiologie, CH-4056 Basel

The outer menlbrane of Klebsiella pneumoniae contains a maltose-inducible protein homologous to maltoporin (LanuB) from Escherichia coli KI2. Its gene (lamB) encodes the precursor, a 429 amino acid polypeptide, which was placed under the IPTG-inducible tac promoter control and expressed from the plasmid pCW2 (Mol.Gen. Genet.(1992) 233, 372-378) in the E. coli K12 strain pop6510 (dex-). The protein was extracted from the outer men~orane fraction using the non-ionic detergent octyl-polyoxyethylene (oPOE) and purified by co-lu~un chromatography.

The pure protein fully retained its maltodextrin-binding properties as tested by affinity-chromatography on a Starch-Sepharose column. The N-terminal l0 amino acid residues of the mature protein were sequenced (P.Jen6, Abt. Biochem• and found to be identical with the E. coli protein. The sedimentation coefficient (S2o.w) was determined by analytical ultracentrifugation to be 6.3 • S, sedimentation equilibrium centrifugation gave a mass of 136 kDa. These results are in agreement with the proposed trimeric structure (mature monomer with 404 amino acid residues, calculated M r 45300). A preliminary screening of crystallization conditions has already yielded crystals of bipyramidal morphology.

Protein-Proteln and Protein-Drug Interactions in a decameric Cyclophilin Cyclosporin Complex. *Gaston M. Pfiuegl~ Michel Sanner, "Tilman Schirmer, Joerg Kallen *Johan N. Jansonlus and Malcolm D. Walkinshaw *Biocenter, University Basel and Sandoz Pharma AG, Basel Human cyclophilin (Cyp) an ubiquitous intracellular protein of 165 amino acids is the likely target for the cyclic undecapeptide immunosuppressant drug cyclosporin A (CsA). Cyclophilins also have a peptidyl-prolyl isomerase (PHase) activity and a chaperone activity by which they enhance the rate of protein folding. The crystal structure of a decameric cyclophilin cyclosporin complex has been recently solved by molecular replacement I. This structure provides the first detailed picture of the protein-drug interaction between Cyp and CsA. Cyp has an eight stranded anti-parallel fl-barrel structure. The PPIase active site lsiocated in a cleft on the surface of the barrel. CsA binds in this cleft like a coin in a slot. Two mouomeric Cyp-CsA complexes form an intimate twofold-related dimer. A local fivefold axis builts up a decamer from such a dimer. The decamer has a doughnut-like structure in which the CsA molecules, largely excluded from contact with solvent, form a double inner ring, sandwiched above and below by Cyp p.entamers. A detailed analysis of the monomeric Cyp-CsA receptor/drug interaction will be presented along with the other Cyp-Cyp protein/protein and Cyp-CsA protein/drug interactions found in the supramoleeular "decameric Cyp-CsA sandwich". The cDNA encoding the 40 kDa carboxy-terminal domain of human poly(ADP-ribose) polymerase was inserted in an expression vector, the recombinant protein was overproduced in E. coil, and purified to homogeneity. The 40 kDa catalytic domain has the same affinity (Kin) for NAD + as the full-length enzyme, expresses abortive NAD + glycohydrolase activity, catalyzes the initiation, elongation, and branching of ADP-ribose polymers, but exhibits no DNA dependency. Its specific activity is approximately 500 fold lower than that of the whole enzyme, fully activated by DNA strand breaks. Surprisingly, the catalytic domain exhibits the processive mode of polymer attachment typical of full length poly(ADP-ribose)polymerase and is able to modify histones HI and H2B. Finally, the polymer sizes formed by the 40 kDa fragment are influenced by histone HI. Mitochondrial creatine kinase (Mi-CK) forms octamers consisting of four stable homodimers and is supposed to participate in the formation of contact sites between the mitochondrial membranes, Previously, some evidence for limited active site accessibility in the octamer and slight differences in enzyme activity between octameric and dimerie Mi-CK have been reported by others. In the present study a 25% decrease in tryptophan fluorescence during the time course of octamer dissociation by the transition state-analogue complex (Cr+Mg2++nitrate+ADP) is demonstrated, and is shown to be due to enhanced quenching of an active site trp residue by the nucleotide substrate, suggesting a higher substrate binding capacity for the Mi-CK (timer in comparison to the octamer. Upon enzymatic removal of ADP, reoctamerization can also be followed on-line by monitoring a biphasic fluorescence increase. The individual dissociation and reassociation rate constants represent useful parameters for mechanistic studies on the octamer-dimer transition and for the systematic comparison of Mi-CK mutants. Dissociation rates of mutated proteins are presented and correlated with enzyme kinctical parameters. Both slower and faster decaying mutants have been investigated.

Aspartate Aminotransferase Y70H: Catalytic Activity and Substrate Specificity Peng PAN, Roll JAUSSI, Heinz GEHRING, w GIANNATI'ASIO and Philipp CHR/STEN. Biochemisches Institut der Universit~,t Z0rich, CH-8057 ZOrich, Switzerland Y70 of chicken mitochondrial aspartate aminotransferase (AspAT) was replaced with a histidine residue by oligonucleotide-directed mutagenesis. AspAT Y70H retained 13% of the activity of the wild type enzyme toward dicarboxylic substrates, whereas the activities toward aromatic amino acids were only about 0.6% of that of the wild type enzyme, corresponding to a 22-fold increase in the ratio of the activities toward these two types of substrates. The kcat/K' m values for aspartate and 2-oxoglutarate were decreased 40-and 70-fold, respectively. The spectrophotometrically determined pK a value of the internal aldimine formed between pyridoxal 5'-P and K258 in the mutant enzyme was similar to that in the wild type enzyme. The dissociation rate constant of pyridoxamine 5'-P in AspAT Y70H proved to be increased only 3 times in contrast to the 80-fold increase in E. coil AspAT Y70F [Toney, M. D. & Kirsch, J. F.(1987) JBC 262, 12403] . These results as well as CD spectra of the coenzyrne moiety indicate that the replacement of Y70 with an unprotonated histidine residue leaves the active site topochemistry essentially unchanged. However, upon protonation of H70 (pK a' = 6.3) the catalytic activity of the mutant enzyme is reversibly abolished. Mitochondrial creatine kinase transfers a phosphate from ATP to creatine, yielding phosphocreatine. Previous investigations have shown that cysteine 278 can be chemically modified with sulfhydryl-specific reagents or creatine analogues, leading to an inactive enzyme.

We have taken a molecular approach to further elucidate the role of cya27a in the chicken cardiac mitochondrial creatine kinase (Mib-CK). Using site directed mutagenesis and heterologous expression of the protein in E.coli, the cys27a was replaced by a glycine and a serine. Under standard conditions C278G and C2788 were reduced 40 to 250 fold in their enzymatic activities. The activity of both mutant proteins could be activated 5 to 75 fold by Cl-anions. In contrast, the wild type protein is slightly inhibited by CI-anions. In addition, the pH optima of the mutant proteins are about 1 unit lower. Under optimal reaction conditions the mutant proteins show a Vmax that is only about twofold lower than for the wildtype, but the Km for creatine and phosphocreatine is about 10 to 20 fold increased. The mutant proteins are 5 to 10-fold more sensitive to inhibition by free ADP and ATP than the wild type protein. Our data suggest that cys27a plays an important role in maintaining the correct shape of the active site in Mib-CK for substrate binding, but the residue is not important for catalysis.

Gupta, S., Hotienstein, R :H, Kochhar, S., and Christen P. B ochemisches Institut and :i:Organisch-chemisches 'lnstitut der Universit~tt Z0rich, CH-8057 Zfidch.

Oxidation of carbanionic enzyme-substrate intermediates by extrinsic oxidants may result in irreversible inactivation of certain enzymes (EJB 63, 223, 1976) . Fru-l,6-P^ aldolase was paracatalytically modified with [U-I~c] fru-l,6-P 2 p~us KaFe(CN)8 and digested with pronase. The isolated radioac2ive peptide contained 1 mol of a phosphorylated C 3 moiety per mol. Amino acid sequencing of a chymotryptic fragment had previously shown that the polypeptide chain was covalently crosslinked at active-site K229 and K146 (PNAS 76, 2527,1979) . The structure of the crosslink was now investigated by MS, 'H-and 13C -NMR. Apparently, an amidine group connects the e-amino groups of the two lysine residues: Based on the highly similar metal binding properties of rPVwT and rPVFlo: w the latter protein was used to study cation dependent conformational changes. Trpfluorescence emission and UV-difference spectra of PVFI02 w indicate that the Trp-residue at position 102 is confmed to a hydrophobic core. Upon Ca 2 § or Mg2+-binding the structural organization of the region around the Trp is reinforeed and an electron-withdrawal effect takes place. Spectral studies indicate that in the Mg2+-conformation the Trp-envimnment is less hydropbobic than in the Ca2+-confonnation. The conformational change upon binding Ca 2+ increases linearly from 0 to 2 cations bound, indicating that the binding of both Ca2+-ions contribute equally to the structural organization in this protein.

*Vacca, R.A., Christen, P. and Sandmeier, E., Biochemisches Institut der Universit&t ZQrich, CH-8057 ZQrich

Aspartate aminotransferase (AspAT) catalyzes the slow racemization of amino acids which results from protonation of the coenzyme-substrate adduct at C a from the re side. Ionizable groups being absent on the re side, this protonation is effected by a water molecule, the diffusion of which into the interior of the enzyme is rate-limiting [Eur. J. Biochem. 203, 563, 1992] . In an attempt to enhance the rate of racemization, W140, 117, on the re side, and V37, in the plane of the coenzyme pyridine ring, were separately replaced by histidine residues. AspAT W140H racemizes alanine one order of magnitude faster (kca t 3 x 10 2 min q) than the wild type enzyme. AspAT I17H slowly decarboxylates aspartate to alanine (kea t 2 x 10 -3 min1). AspAT W140H, 117H and V37H retain 20, 20, and 60 %, respectively, of the wild type activity toward dicarboxylic subatrates, whereas in all three mutants activity toward aromatic amino acids is decreased by two orders of magnitude. Pyridoxamine 5'-phosphate dissociates from AspAT W140H 52 times faster than from the wild type enzyme. The reaction specificity of pyridoxal 5'-phosphate-dependent enzymes is thought to be determined in part by stereoelectronic control. Orientation of a bond at Ca orthogonal to the coenzymesubstrate = system facilites its cleavage. Indeed, in aspartate aminolransferase (AspAT) the Cct-H bond of the substrate moiety is positioned in this way by interaction of its cc-carboxylate group with R386, one of the four residues invariant in all aminotransferases. In order to change the orientation of the enzymebound substrate relative to the coenzyme ring, R386 was replaced by alanine and a new arginine residue was introduced at position 225 by oligonucleotide-directed mutagenesis. With this altered substrate binding site, AspAT R386A/Y225R decarboxylates Laspartate to L-alanine 100 times faster (kcat = 1.4 rain -t) than the wild type enzyme, whereas the transaminase activity towards dicarboxytic amino acids is decreased by 3 orders of magnitude (kcat = 11 rain4). Two reaction paths are followed: decarboxylation of L-aspartate to L-alanine and transamination-decarboxylation of L-aspartate to pyruvate. Reduced thioredoxin f is an essential and specific activator protein of chloroplast fructose-l,6-bisphosphatase. Compared to the other plant thioredoxins it contains in the G-terminal part a unique third Cys residue which is surface exposed and easily modified by chemical modification (e.g.N-ethylmaleimide) and which may, in part, be responsible for the high specificity. The chemical modification results in reduced efficiency of activation.

We have now replaced this Cys-73 by three different amino acids, Gly, Ser or Ala and studied the effects of these mutations on the properties of thioredoxin f. Whereas the Cys-73-Gly mutation produces a very unstable protein that may be partially uncorrectly folded, the other mutations, Cys-73-Ser and Cys-73-Ala, yield apparently correctly folded, stable proteins. However, there activation efficiencies are considerably reduced when compared to the wild type protein suggesting that Cys-73 is an imoortant residue for the protein-protein interaction. consists of about 15% hydrophilic, salt-soluble enzyme (SS-AChE), and of 83% amphiphilic, detergent-soluble form (DS-AChE). Sucrose density gradient centrifugarion showed that SS-AChE was composed of about 85% tetramer (10.3 S) and 15% monomer (3.3 S). In amphiphilic DS-AChE, 85% tetramer (9.7 S), 10% dimer (5.7 S), and 5% monomer (3.2 S) were seen. The enzyme is N-glycosylated, no O-linked carbohydrate could be detected. Use of two monoclonal antibodies, one directed against the catalyric subunit, the other against the hydrophobic anchor, gave new insights on the subunit assembly of brain ACHE. It is shown that in tetrameric ACHE, not all of the subunits are disulfide bonded, and that two populations of tetramers exist, one carrying one, the other two hydrophobic membrane anchors. The functional form of the hydrophilic subunit III m~n is a dimer. Each monomer of III m~n consists of two structurally and functionally distinct domains, P13 and P20, which are both transiently phosphorylated. There are four tryptophan residues per monomer in positions 12, 33, 69 (P13 domain) and 182 (P20 domain). Intrinsic fluorescence studies were performed with III m~, the P13 and P20 domains, as well as with the following tryptophan to phenylalanine mutants: W12F, W69F and W12,69F. Results from fluorescence emission studies and fluorescence quenching by acrylamide indicate that Trpl2 is a surface residue, Trp182 and Trp69 are buried, and Trp33 is intermediate in location. Analyses of fluorescence intensity measurement suggest that energy transfer occurs between Trp69 and Trpl2. Trpl2 is near Hisl0 which is transiently phosphorylated. Trpl2 probably plays an important role in the phosphoryl transfer reaction since its substitution by Pbe results in a drastic decrease of enzyme activity. Three monoclonal antibodies, mAb 132-4 (IgG1), 132-5 (IgGl), and 132-6 (IgG3), specific for amphiphilic AChE from mammalian brain, were shown to mainly react with native tetrameric forms of AChE but not with dimeric and monomeric forms indicating that assembly of the tetrameric form of AChE by the hydrophobic anchors is essential for the epitope. Western blots after SDS-PAGE in non-reducing conditions showed that the mAbs reacted with tetramers, trimers, and heavy dimers carrying the hydrophobic anchors but not with light dimers or monomers carrying no anchor. Since in reducing conditions the mAbs did not recognize the hydrophobic anchor, it is concluded that disulfide bonding contributes to the epitope. The mAbs crossreacted with brain AChE from several sources indicating that the structure of the hydrophobic anchor and the assembly of mammalian brain AChE are similar.

Stadelmann, B., Zurbriggen, A. and Brodbeck, U. Institute of Biochemistry and Molecular Biology and Institute of Animal Neurology, University of Bern, CH-3012 Bern It has been reported that mammalian serum, and to a lower extent mammalian liver, brain, pancreas, udder, and milk, contain glycosyl-phosphatidylinositol-specific phospholipase D (GPI-PLD) activity. However, the sites of synthesis have not been determined. In order to study in which cell(s) of the organism synthesis of GPI-PLD takes place, we undertook a systematic screening of 12 different bovine tissues. In situ hybridization experiments with a specific anti-sense RNA probe, derived from a bovine liver cDNA, revealed that GPI-PLD mRNA is present in mast cells of adrenal gland, lung, and liver. On the other hand, our probe detected no GPI-PLD mRNA in bovine pancreas, brain, and udder, although enzyme activity has been reported in these tissues. Northern blot analysis of total bovine liver RNA demonstrated two distinct GPI-PLD mRNAs of approximately 3.3 kb and 4 kb length suggesting that two forms of the enzyme may exist. The first step in the betalain pathway, the conversion of tyrosine to DOPA, is catalyzed by the enzyme tyrosine hydroxylase. Subsequently, DOPA is converted, through the action of DOPA-4,5-dioxygenase, to betalamic acid, the betalain chromophore. Betalamic acid may condense with any amino acid or amine, giving rise to betaxanthins, or with cyclo-DOPA, to form betacyanins.

In the present study, tyrosine hydroxylase was isolated from the fly agaric mushroom (Amanita muscaria), purified and characterized. Purification was also attempted using beet root (Beta vulgaris L.) celt line cultures as the tissue source, but the enzyme proved to be very unstable. Some notable differences and some common features of the two enzymes will be discussed. In bovine serum, GPI-PLD occurs in unusually high amounts of about 40 ~tg/ml. The purified enzyme is an ampbiphilic protein the majority of which appears to be associated with the high density lipoprotein (HI)L) fraction. The amphiphilic properties of GPI-PLD and its interaction with apolipoprotein A I (apo A), the major constituent of HI)L, was investigated by sucrose density gradient eentrifugation. In presence of 0.1% TX-100 in the gradient, the enzyme sedimented with an apparent sedimentation constant of 6.0 S and formed aggregates up to 14.5 S in absence of detergent. In presence of apo A, the enzyme became desaggregated and formed a complex with apolipoproteln which sedimented at 7.6 S. In the aggregated state, the enzyme is virtually inactive; activation was seen by increasing amounts of apo A. Half maximal stimulation was obtained at a 10 fold molar excess of apo A over GPI-PLD. Desaggregation and stimulation of GPI-PLD was also obtained with TX-100 with optimal activity at 0.018%, the CMC of the detergent. From our data we conclude that in serum, GPI-PLD is in association with apo A which could provide a vehicle for the uptake of GPI-PLD into ceils.

The prostate is an accessory sex gland exclusivelyfound in mammals.

The motivation for the extensive study of this organ stems from the many pathological complications affecting this gland, the most important being benign prostatic hyperplasia (BPH). All forms of prostatic growth, be it normal or neoplastic exhibit androgen dependency. It is now clear that dihydro-testosterone (DHT) and not testosterone (T) is the important androgen in the prostate. Inhibition of 5AR, the enzyme which converts Tto DHT, seems a promising approach for a pharmaceutical therapy of B PH. This task has become more challenging bythe recent report of two cDNAs encoding different, homologous 5AR isozymes. Here we present kinetic evidence for the presence of two catalytically active isozymes (5AR1 and 5AR2) in human prostatic microsom es. The Km values for T were determined to be 19 #mol/I and 20 nmol/I for 5AR 1 and 5AR 2, respectively. These values are close to those determined for 5AR 1 and 5AR 2 expressed in yeast (25 #mol/ I and 23 nmol/I). Under saturating T concentrations in human prostatic microsomes 55% of the specific activity (mole/min/mg of protein contained in microsomes) is related to 5AR 1 and 45% to 5AR 2.

Alessandro Puoti and Andreas Conzclmann* From the Institute of Biochemistry, University of Lausanne and the *Institute of Biochemistry, University of Fribourg (CH)

We have structurally characterized two very polar glycolipids (CPa and CPb) which can be observed after metabolic labeling of mouse lymphoma cell lines S1A and EL-4 with either tritiated myo-inositol, mannose or ethanolamine. The glycosyl-phosphatidylinositol (GPI) protein anchor of SIA cells has also been characterized and it appears that the structure of its carbohydrate moiety is identical to the one of the CPb lipid, namely: (X-)Manctl,2Manctl,6(Y-)Manc~-GlcN-Inositol, X and Y being hydrofluoric acid-sensitive substituents (most likely phosphoethanolamine). Mutant lymphoma S1A -b, EL-4-f and BW -e cell lines do not synthesize CP lipids, but accumulate abnormal inositol-containing tipids which are not present in the corresponding wild type cell lines. These lipids have been characterized and their structure suggests they are potential precursors of GPI anchor biosynthesis. As mutant lymphoma S1A -b, EL-4-f and BW -e accumulate GPI-related glycolipids having 3, 2 or 0 mannose residues respectively, we propose that these mutants are deficient in GPI anchor biosynthesis and that the accumulation of these incomplete GPI structures is due to the lack of the specific processing enzymes.

Nine HCC cell lines (Co112, HT29, SW480, SW620, ISRECol, CaCo2, CoSut, COL0205, SW1116) were treated with 20 ng/rel TNFec or 5 ng/ml ILl~, for 24 h in serum-free culture conditions. PAs and PNs were analyzed in sopernatants by zymegrephy using SDS-PAGE with copolymenzed plasminogen-fich fibdnegen and by enzyme immunoassaye for tissue-type PA (t-PA), u-PA, PAl-1 and PAl-2.. 8/9 and 9/9 cell lines increased their production of u-PA respectively from 4 to 50 fold after TNF treatment and from 2 to 100 told after ILl treatment. PAl-1 preduction was also augmented after TNF treatment from 2 to 500 fold in 5/9 ceil lines and from 3 to 90 fold in 5/9 cell lines after ILl treatment. There was no .correlation between increased production of u-PA and that of PN-1. The cytokine stimulator/eftect was shown to be time-and concentration -dependent. Northern analysis of polyA RNA extracts after cytokine treatment confirmed the stimulatory effect of TNF and ILl on .of o-PA and PAM expression. 2/9 and 1/9 cell lines also produced more t-PA after respectively TNF or IU treatment. No PN-2 was detected in supematants of the 9 cell lines examined. Using a glutaraldehyde immobilized 1251-1aminin degradation in vitro assay, we show that the [ncreased u-PA produc~on correlates with higher proteol~o degradation induced by the cytokine -treated ceils. ISREC, CH-1066 Epalinges S/Lausanne. We have shown recently that human colon carcinoma cell lines secreted both procathepsin B and cystatln C into culture media. Expression by colonic cells of the proteinase and the inhibitor also occurred at the mRNA level. We have now investigated wether the concomitant expression of these two antagonists by SW480 cells could be modulated by phorbol esters (PMA). In SW480 ceils, secreted procathepsin B represents 17 to 19% of newly synthesized prneathepsin B; the major part is targeted to lysosomes.

Without PMA treatment, cycloheximide (CHX), an inhibitor of protein synthesis, drastically decreased secretion of prneathepsin B, but decreased only by 16 to 17% the intracellular pool of eathepsin B. PMA induced a 5.5 fold increase in total cathepsin B synthesis upon 24-h of treatment. Both secreted procathepsin B as well as lysosomal cathepsin B increased after PMA treatment. Both increases were completely blocked by CHX. PMA had no effect on the 24-11 accumulation of eystatin C in media, but CHX strongly inhibited its de nova synthesis. These results were confirmed on the mRNA levels for cathepsi, B. The transcript of cyslatin C however, decreased 1.3 fold upon 8-h of PMA stimulation, and at 2,1-h increased 1.4 fold the initial level, which explains the lack of effect of PMA observed above on the net cyslatin C protein levels at 24-h. The expression levels of cathepsin B being increased in many human cancers, it is of importance to understand the dynamic of the balance between the proteinase and its endogencous inhibitors, The present work shows that the tumor-promoter PMA induces, at 8-h, a switch in the expressions of eathepsin B and cystatin C.

Supported by the Swiss Science Foundation.

Alanine aminopeptidase activities in poppy leaves Aminopeptidases contribute to the complete degradation of polypeptides to free amino acids. Several forms with different affinities to N-terminal amino acids were identified previously in various plants, Alanine aminopeptidase extracted from poppy leaves (Papaver somniferum L.) was investigated with p-nitroanilides (pNA) of L-ala, of D-ala and of the dipeptide L-ala-L-ala. The activity against D-ala-pNA was below 15% of that against L-ala-pNA. The initial liberation of p-nitroaniline from L-ala-pNA was linear with time, while a delay was detected with L-ala-L-ala-pNA. This suggests that the liberation of the two amino acids from the latter substrate was caused by two independent events. The hydrolysis of L-ala-pNA was slightly inhibited (about 20%) by the addition of 8 ram L-ala (free amino acid) or of 8 ram L-ala-L-ala (dipeptide), whereas D-ala or dipeptides containing D-ala were not or much less effective, The inhibitory effect of L-ala-L-ala may be caused by the dipeptide itself or by free L-ala liberated from its cleavage. In addition, it is known that adult flukes secrete enzymes that can break down immunoglobulins.

Degenerate primers designed specifically for cysteine-and serine-proteases were used in the polymerase chain reaction using as starting material either DNA or RNA isolated from adult F. hepatica. Clones containing eDNA fragments of eight different cysteine-proteases (FCP1 -FCPS) and one clone encoding part of a serine-protease eDNA (FSP1) were generated. The homology at the nucleotide level among the different cysteine-protease clones varied from 50 to 80 %. A comparison of the predicted amino acid sequences with proteases of other organisms revealed homology to proteases of the cathepsin-L and cathepsin-B type.

Northern blot analysis of RNA isolated from adult F. hepatica revealed abundant protease transcripts of different length. FCP1 and FCP2 transcripts were particularly abundant. One protease eDNA was cloned in an expression vector and antibodies against the fusion protgein were produced in rabbits. Western blot analysis confirmed the abundance of proteases in adult F. hepatica.

Run,or, S. and Cerletti, N., Biotechnology Department. Pharmaceutical Division, CIBA Ltd., CH-4002 Basel.

Transforming Growth Factor Beta-2(TGF-S2)is a disulfidelinked homodimer of Mr 25 kDa released from many cell types. Among numerous pleiotropic effects, it also influences the rate of proliferation of different cells, controls processes of osteogenesis, epithelial cell differentiation and immune cell function. In order to characterize the stability of rh-TGF-~2, we have followed the unfolding transitions induced either by heat, urea or guanidinium-HCl (Gu-HCI) by using Near, or Far-UV Circular Dichroism Spectroscopy. Under acidic and neutral conditions, the secondary and tertiary structures are lost within a narrow range of chaotrop concentration (mid-point of transitions around 3.5M Gu-HCI and 5M urea). The protein exhibits a good heat-stability as the native tertiary (Near-UV CD) and backbone structures (Far-UV CD) are lost only at temperatures, respectively, higher than 70 oC and 80oC. After removal of the chaotrop(s) by dialysis or lowering the temperature, respectively, TGF-~2 exhibits the original CD-spectrum as well as the initial bioactivity (growth inhibition of Mv-l-Lu Cells), demonstrating the reversibility of the unfolding transitions and the remarkable stability of this protein. Allomyces, an aquatic fungus, has two predominant phases in the life cycle. The growth phase, which starts with the spore germination and mycelial growth occurs in balanced medium. Morphogenetic phase is induced by nutrient deprivation and involves the cessation of hyphal elongation and the transformation of the hyphal apex into sporangium in which the flagellated zoospores differentiate. Superimposed on this two developmental pathways is the activation and inactivation of two main intracellular neutral proteases. A cysteine-calpain-like protease characterizes the vegetative growth phase and a serine protease which characterizes the morphogentic pathway 9 Biochemical properties of the calpain-like protease has been adequately characterized (I, 2). We present our results on the correlation between appearance and disappearance of the enzyme activity and the calpain-like antigen by immunobiotting. For example the antigen diasappears completely after 1 hour of induction of differentiation which correlates with the appearance of serine protease. A partial characterization of the serine protease is presented. References: 1. Ojha & Wallace, J. Bact. 170, 1254 Bact. 170, -1260 Bact. 170, (1988 

MCP-1 is a human monocyte chemotactic protein produced by several cell types such as keratinocytes, mesangial ceils, fibroblasts, PMA stimulated PBMC and human glioma cell-line U-105MG. Presence of MCP-1 mRNA is demonstrated in permanent human glioma cell Iines with or without stimulation by inflammatory cytokines. MCP-1 secretion in cell lines supernatants is analyzed by immunoprecipitation using a specific monoclonal antibody (Ell). These in vitro data show that glioblastoma ceils are able to release MCP-1 in the extracellular environment.

To prove that this could also occur in vivo we first examined the presence of MCP-1 mRNA in biopsies of human low and high grade astroeytomas. Secondly, by immunohistochemistry, we are currently analyzing the presence of MCP-1 protein on frozen tissue sections using a polyclonal antibody. Finally, the in vivo secretion of MCP-1 will be tested in CSF and tumor cyst fluids of patients.

These results suggests that MCP-1 may play an important role in vivo in the recruitment of monocytes in tumor tissues. the Krebs and glyoxylate cycles. The latter is localized in the glyoxysomes, which appear at the beginning of oilseed germination and are converted into peroxisomes when fat reserves are depleted. The CS primary structure from various organisms has been determined, but at present no plant glyoxysomal CS sequence is known. Antibodies raised against mitochondrial and glyoxysomal CS are now necessary to screen a cDNA library. Previous studies demonstrated that CS isoforms cannot be separated from crude extract, the mitochondrial form(s) being much more abundant than the glyoxysomal form. Consequently, a more appropriate purification method is based on density centrifugation (organelle isolation on sucrose gradients), membrane sedimentation and phase partition (CS is weekly adsorbed on membranes), as well as standard purification procedures. CS isoforms from the cotyledons of 5-day old soybean seedlings (Glycine max. L.) were thus obtained from SDS-PAGE, and used to raise antibodies. Most purification schemes of calpain (CANP) involve a number of chromatographic steps. The final preparations invariably contain impurities, including degradation fragments. To obtain a pure CAMP preparation suitable for biochemical and structural studies, a new peptide-affinity column was developed, using a 27 amino acid peptide corresponding to the product of exon ib of the endogenous inhibitor calpastatin. Crude preparations of CAMP, isolated by a modification of the procedure described by Melloni et al. (1982) were incubated with a highly specific, irreversible, active site-directed synthetic inhibitor [(ZLLYCHN2) , Anagli et al., (1991) ] which blocks the enzyme in the inactive 80 kDa form. The crude enzyme preparation was then loaded on the peptlde affinity column in the presence of calcium.

Pure CANP was eluted with an EDTA-containing buffer. The procedure allowed the isolation of CAMP with a high grade of purity, i.e., a 80 kDa band wich accounted for more than 95% of the protein eluted from the column with the EDTA buffer. Cyclic AMP-specific phosphodiesterase &DE-IV) has been purified from htma~m mononuclear leukocytes in four-step purification procedure. The MNL suspension was sonieated, centrifuged (120'000xg), and applied to Q Sepharose (anion exchange). Active fractions eluted at 0.23-0.32 M NaC1. The affinity column for the next purification step was prepared from 4-[3-(carboxypropyloxy)-4-methoxyphenyl]-2-pyrrolidinone (J. Demnitz, Sandoz Basel) immobilised to EAH-Sepharose 4B (S. Fougier at. al, BBRC, 138, 205) . After washing with 0.5 M MaC1 PDE-IV was Muted with 0.5 M MaC1 and 1 mM cAMP. Cyclic AMP was separated from PDE-IV on Mono Q (anion exchange). Activity Muted at 0.32 M NaC1. The overall recovery was 17% (30 Bg protein, purification 1868x) from 1.1 -10 TM cells. PDE-IV was concentrated and separated on gel filtration (Superose 12). The single peak of enzyme activity was associated with a single protein peak, which Muted at a Mr of 100-140 kDa. Under non-denaturing conditions, Gel electrophoresis revealed one major band corresponding to -120 kDa. Under denaturing conditions this band disappeared and several bands with lower apparent M r appeared, probably representing ptoteolytic fragments of PDE-IV. Glycoproteins altered in the glycan moiety may be markers for diseases, e.g. in cancer, inflammations and alcoholism. In this study we combined the high protein resolution capacity of twodimensional gel electrophoresis with the high sensitivity of glycoprotein detection by specific lectins. Human plasma proteins were separated by mini two-dimensional gel electrophoresis (7x9 cm), transferred to polyvinylidene difluoride membranes and incubated with biotinylated lectins. We focused on lectins reacting with sialic acid residues (Wheat germ and Sambucus nigra), ~-D-galactose (Datura stramonium and Ricinus communis), e-fucose (Aleuria aurantia) and Nacetyl-glucosamine (succinylated wheat germ). Known plasma glycoproteins such as ~l-antichymotrypsin, ~l-antitrypsin, HSglycoprotein, el-acid glycoprotein, #-haptoglobin and transferrin were easily detected in ng amounts. This technique is suitable for studying changes in glycoproteins of body fluid and tissue extracts with high resolution, sensitivity and specificity. The most abundant protein of the vaccinia virus envelope is the palmitated P37K protein. This protein is required for viral envelopment and subsequent release from infected cells. Little is known about the nature of the association of P37K with the envelope, whether P37K interacts with other proteins of viral or cellular origin, or the functional role played by the attached palmitate moiety. Therefore, a variety of experiments were performed in order to better characterize the protein and its interactions. Triton X-114 partitioning experiments and sodium carbonate treatment indicated that P37K has a hydrophobic nature and is tightly associated with the membrane. This might be due to the attached palmitic acid since computer analysis indicates that the protein is fairly hydrophilic and does not contain a well defined transmembrane domain. In addition, limited proteolysis of intact virions demonstrated that the P37K is localized to the inner surface of the viral membrane. Experiments are currently in progress in an attempt to identify other viral or cellular proteins which interact with P37K, in addition to studies involving proteolytic digestion of purified P37K to determine which portion(s) of the protein are subject to palmitation.

During automatic Edman-degradation, the N-terminus of a polypeptide is first coupled with phenylisothiocyanat (PITC) then specifically cleaved and extracted from the support. After conversion to its phenylthiohydantoin (PTH-) derivate, the amino acid is characterized by HPLC. Ethylacetate (S2reagent), whose function is to wash excess PITC from the carrier, is known to represent a critical parameter. Small impurities may react with the coupling product or interfere with the PTH-amino acid detection. When small amounts of polypeptide were degraded we observed a up to two-fold detection enhancement by reducing the S2-delivery. Since this effect seems to depend on the support-type the presence of impurities (perhaps water) might influence in a not well understood way the extraction yield of the ATZ-amino acid.

Lthhy, R., and Bucher, P., Swiss Inst, for Exp Cancer Research. Epalinges

The sequence profile method (Gribskov et al.. PNAS 84:4355-4358) is a powerful tool to detect distant relationships between amino acid sequences. A profile is a table of position-specific scores and gap penalties, providing a generalized description of a protein motif, which can be used for sequence alignments and database searches instead of an individual sequence. A sequence profile is derived from a multiple sequence alignment. We have found several possibilities to improve the sensitivity of sequence profiles: 1, In the originally described method all sequences in the alignment are given the same weight. Usage of individual weights for each sequence avoids bias towards closely related sequences. These weights are automatically assigned based on the similarity of the sequences. 2. For the construction of the profile an amino acid mutation table is used. We have found that in some cases the original Dayhoff matrix works better, than the modified matrix used in the 'current implementation. 3. Gap penalty multipliers are parameters used to align profiles with sequence, Current implementations recommend constant values for all profiles. We developed a heuristic procedure to obtain profile specific gap penally multipliers, Profiles derived by the new method are more sensitive and selective in a number of cases where previous methods have failed to completely separate true members from false positives.

High pressure freezing increases the' vitrification depth. N. Sartori, K. Richter, J. Dubochet, Laboratoire d'analyse ultrastructurale, Universit6 de Lausanne, CH-1015 Dorigny Freezing is a well known process to observe biological materiel in their aqueous environment. The vitrification depth which can be obtained with standard freezing techniques does not exceed 20 gtm. Using crystals of catalase suspended in 15 % sucrose, we show that by high pressure freezing, the depth of vitrification can reach at least 700 ~tm.

Li, X. and Benz, G., Entomologisches Institut, ETH-Zentrum, 8092 Zfirich

High resolution 2-dimensional. (2-D) gel electrophoresis was used to analyse an insect virus, Adoxophyes orana granulosis virus (AoGV). By SDS-PAGE 15-20 protein bands are resolved in the enveloped nucleocapsids (ENC), while by 2-D gel electrophoresis more than 40 proteins can be separated. Among them 8 protein spots appear repeatedly. They are V35a, b, c, d; V34; V33a, b; and V31. The advantage of 2-D electrophoresis is that those proteins with the same molecular weight but with different isoelectric points (IEP), can be well separated (V35a, b, c, d; V33a, b) . They are possibly subunits of a compound protein in the ENC structure. These viral proteins separated by 2-D gel electrophoresis can be demonstrated again without doubt by immunoblot using three anti-virus antisera. Most viral proteins of the ENC have their IEPs ranging from pH 4.5-6.6, but V34 has its IEP at pH 8.5. The granulin has its IEP at pH 5.5. Two different commercial standards were used as monitor for the pH gradients, so that our 2-D electrophoresis profdes can be compared with those from other laboratories. A primary index of the 8 structural proteins of AoGV was established.

N.A.Bersinger, Dept. of Obstetrics and Gynaecology, University of Berne, Switzerland.

The risk of a pregnancy to be affected by Down syndrome (DS, Trisomy 21) and other chromosomal abnormalities increases sharply with maternal age. Conclusive prenatal diagnosis of such a pathology is only possible by karyotyping cells from the fetoplaeental unit obtained by amniocentesis or chorionic villas sampling (CVS). These invasive procedures themselves carry some risk of abortion for a normal pregnancy; it is therefore necessary to detect as many abnormalities with as few amniocenteses/CVS as possible. When selection is based on maternal age only (usually 35 yr), less than half of all Down syndrome are detected since the vast majority of pregnancies occur at a maternal age of less than 35 years. Determination of alpha-fetoprotein (AFP) in maternal serum was found to be useful in D$ screening. Tests for HCG and uneonjugated oestriol have been added ("Triple Screen", "AFP-plus"), but 20-35% of D$ still escape detection. PAPP-A has been found to be a good prognosticator in the ist trimester. We have developed a time-resolved fluoroimmunometric assay for PAPP-A and an ELISA for PAPP-B to test in DS.

Andrea Huwiled'2, Doriano FabbrC and Josef Pfeilschifter 2, 1 : Ciba-Geigy Ltd., Basel; 2: Dept. Pharmacology, Biocenter, University of Basel, CH-4056 Basel

The isoforms of PKC present in glomerular mesangial cells were identified by immunoblot analysis. Mesangial cells were found to express 4 PKC isoenzymes, PKC-a, -15, -5 and r. No PKC-7 and -rl isoforms were detected. On exposure to five structurally different activators of PKC (PMA, bryostatin 1, thymeleatoxin, dihydroteleocidin B, debromoaplysiatoxin, 100 nM each), a translocation and marked down-regulation of PKC-a and -5 is observed in response to all compounds investigated. In contrast, PKC-e is only down-regulated in response to bryostatin 1 and thymeleatoxin. Moreover, PKC-~ is resistant to down-regulation by all PKC activators examined. These data suggest, that it is possible to selectively activate and downregulate PKC-isoenzymes and such compounds could be used to study the physiological functions of a particular PKC isoenzyme. The T lymphocyte cell sulface receptors CD4 and CD8 and the T cell receptor (TCR)-CD3 complex are crucially involved in T cell differentiation and activation. One of the earliest events that follows the triggering of T lymphocytes via these receptors is the tyrosine phosphorylation of a distinct set of cellular proteins leading ultimately to alterations in gene expression and the onset of replication. In our studies we addressed four major points. First, the involvement of src-like, nonreceptor protein tyrosine kinases in these early receptor-mediated signalling events. Second, the possible regulation of these protein tyrosine kinases by the CD45 phosphotyrosine phosphatase and the p50 csk protein tyrosine kinase. Third, SH2 and SH3 domain-mediated intramolecular and intermolecular interactions of these protein tyrosine kinases. Fourth, a potential link between receptor-mediated signalling pathways involving p21 ras and protein tyrosine kinases. We recently reported the cloning of a new subfamily of serine/threonine protein kinases termed RAC , of which two closely related isofroms RACc( and RAC/3 have been identified. RAC appears to represent the phylogenetic mid-point between protein kinase C and A. The N-terminal domain of RAC has linfited homology with SH2 domains. More recently, RAC has been identified as the cellular homologue of v-akt, an oncogene encoding a 105 kDa phosphoprotein containing h truncated gag N-terminal of a wildtype RACo: In order to gain insights into its function and regulation we have over expressed RAC in COS cells and in the baculovirus system. To determine the functions of the various domains of RAC, N-and C-terminal deletions of RAC were expressed in COS cells, looking at their effects on suhcellular localization and activity. RAC purified from the baculovirus system has been examined bioehemieally, including determining the sites on which it is phosphorylated in this system and its substrate site specificity. Specific properties of the various RA C domains, namely the SH2-1ike, kinase, and C-terminal regions, have also been investigated when expressed as GST fusions in bacteria. The 50 kDa CD2 transmembrane receptor molecule mediates adhesion as well as signal transduction function in T lymphocytes and natural killer (NK) cells. One of the earliest events following CD2 engagement is phosphorylation on tyrosine residues of a distinct set of cellular proteins including the k-chain of the T cell receptor (TCR)]CD3 complex. Here, we investigated potential associations of the CD2 molecule with the src-family protein tyrosine kinases p56/ck and p59fY n and the k-chain. Results from coimmunoprecipitation and co-capping experiments provided evidence for a specific association of CD2 with k-chain and p59fy n in vitro and in viva. They suggest that CD2-mediated signal transduction involves a multimolecular complex of at least three components: CD2, ~-chain and p59fyn. A number of proteins contain non-catalytical domains called SH2 and SH3. They are supposed to be important in mediating proteinprotein interactions. We have used recombinant proteins containing the SH2 domain of the src-kinase p56 lck to study the binding of this region to tyrosine phosphorylated polypeptides. First, we demonstrated that recombinant /ck-SH2 bound specifically to a synthetic phosphopeptide representing the tyrosine phosphorylated C-terminus of p56 tck. Second, we showed that a specific subset of tyrosine phosphorylated proteins present in tr~asformed NIH3T3 cells bound to recombinant lck-SH2. Two of the bound proteins were identified as PI 3-kinase and pl20-GAP. In summary our results suggest that the lck-SH2 domain has a dual function: it can bind the phosphorylated C-terminus ofp56tck an thereby restrict kinase activity or it can mediate intermolecular interactions with cellular signaling proteins. Mammary carcinoma is the most frequent malignancy and a major cause of death in women. In the Western world, approximately 10% of women become afflicted and the incidence rate is still increasing. Protein tyrosine kinases (PTKs) play a fundamental role in normal cell growth and differentiation and there is compelling evidence for their involvement in neoplastic disease. Using a PCR-aided gene family cloning strategy we have determined the constellation of PTKs expressed at various stages of mammary gland development, particularly during the estrus cycle. Our interest in the estrus cycle reflects the idea that perturbations in PTK expression contributing to uncontrolled epithelial cell growth may accumulate during the repetitive cycles. To identify possible candidate PTKs we want to know which PTKs are expressed during the normal estrus cycle and how their expression is controlled. We have isolated several known and unknown PTK clones expressed during the estrus cycle some of which show differential expression in the phases of the cycle. Blaschke, RJ., Andres, Ziircher, G., Reid, H.H. and Ziemiecld, Universit[it Bern, Tiefenanstrasse 120, 3004 Bern The activation of receptor protein tyrosine kinases (PTKs) by external signals represent the initial step in a complex cascade that eventually result in differentiation phenomena observed in different states of normal and neoplastie development. We have isolated partial cDNAs for three novel putativ growth factor receptors (myk I to myk III) expressed in a differentiation dependent manner in the mouse mammary gland. Sequence analysis confirmed their affiliation to the eph/eck/elk subfamily of growth factor receptors for yet unknown ligands. Comparison of the earboxyterminal region of the predicted proteins revealed a conserved tyrosine residue that might be a phosphorylation target and therefore play a role in the regulation of interactions between the members of this FIK-subfamily and their prospective substrates. One of the novel growth factor receptors has a MW of approx. 120 kD and was shown to have an in vitro kinase activity. Expression analysis revealed an elevated expression in rastransgenie tumors for allthree putative receptors. In contrast more differentiated mye-transgenic tumors did not show this overexpression. In addition, one of the identified PTKs (myk III) is differentialy expressed in different stages of the hormonaly controlled oestrus cycle. To understand the role of PTKs specifically in mammary biology we employed a PCR based cloning strategy to identify PTKs important for growth and differentiation of mammary epithelial cells. By this method EF11, a member of the eph/eck/elk family of receptor PTKs for unknown ligands, was isolated from the mammary epithelial cell line 31E. RNA analyses revealed that EF11 is expressed as a 4.7 kb transcript in a variety of organs. I n mammary glands EF11 expression is at least partially of epithelial origin and can only be detected in mature mammary glands after puberty. Expression is downregulated during pregnancy induced differentiation of the mammary epithelium. This indicates a hormonal control of EF11 expression which will be verified using a cell culture model system. Most importantly, mammary tumors of transgenic mice bearing an activated human Haras oncogene showed a highly increased expression, suggesting that EF11 may play an important role in mammary carcinogenesis as well as in its differentiation. Two calcium-binding proteins of 8 kDa (Calgranulin A or p8) and 14 kDa (Calgranulin B or p14) have been shown to translocate from the cytosol to the membrane of human neutrophils upon calcium-dependent stimuli (&Biol.-Chem.267:27,19379-19382,1992) . When phorbol myristate acetate (PMA), a calcium-independent stimulus was used, no translocation was observed. Calcium was found to bind to both proteins on nitrocellulose blots after SDS-PAGE electrophoresis of eytosol or membrane preparations of resting and stimulated neutrophils. Interestingly, activation of the cells using both calcium-dependent and independent (e.g.PMA) stimuli strongly increased the level of phosphorylalion of p14 compared with unstimulated neutrophils. On the contrary, p8 was phosphorylated in PMA-aetivated cells only, correlating with the absence of translocation of the protein to the membrane. To determine if protein kinase C was responsible of this phosphorylation, staurosporine was incubated with PMA-stimulated cells. The phosphorylation of pS, but not of p14, was completely inhibited. These results suggest that 138 and p14 are independently regulated by phosphorytation and that the absence of translocation of to8 with PMA could be a consequence of its phosphorylation by PKC. The trivalent arsenical phenylarsenoxyde (PAO) that interacts with vicinal sulfhydryls was used to inhibit the tyrosine phosphatase (PTPase) activity of CD45, the major transmembrane glycoprotein of murine T cells. In an attempt to understand the mechanisms of CD45 regulation, the effects of the inhibitor on the following CD45 parameters were investigated: 1} metabolic incorporation of phosphate, 2) oligomerisation and 3) lateral mobility in intact cell membranes. Low concentrations of PAO (1-5 pM) slightly increase (5-10%) the amount of phosphate incorporated into serines, without affecting the oligomeric state or lateral mobility of CD45. On the contrary, PAO concentrations above 25/~M result in a 40% decrease in phosphate incorporation into serines and increases the phosphate content in higher order CD45 oligomers, which no longer undergo patching and capping after antibody-mediated cross-linking. Our results show that CD45 exists in a monomer-oligomer equilibrium that may be controlled by serine phosphorylation. We also suggest that CD45 PTPase activity is different in different order oligomers and may reflect distinct regulatory mechanisms and interactive capacities of CD45 with membrane proteins.

MOLECULAR CLONING OF cDNA ENCODING CALMODULIN FROM Neurospora crassa. Capelli, N., van Tuinen, D., Ortega Perez, R., Arrighi, J-F. and Turian, G. Laboratoire de Microbiologie g~n~rale, Universitd de Gendve, Sciences II1, CH-1211 GENEVE 4. A full-length eDNA encoding Neurospora crassa calmodulin was isolated from a lambda ZAP II eDNA expression library by immunoselection using highly specific polyclonal antibodies purified as described previously (l) . The open reading frame (ORF) encodes a protein of 148 amino acid residues with a calculated Mr of 16854 dahons. In addition, alignment of the predicted aa sequence showed 84.5 % analogy with vertebrate calmodulin. A single substitution (Phel3/Tyr) differentiates calmodulin from Neurospora crassa and an other Ascomycete Aspergillus nidulans.

Using site-directed mutagenesis, the complete cDNA was ligated into a pTrc promoter-regulated bacterial expression vector to allow expression of Neurospora crassa calmodulin in Escherichia coll. The expressed protein was purified from bacterial lysates by phenyl-Sepharose chromatography and exhibited an electrophoretic shift, in the presence of Ca 2+, identical to that of the mitive protein. Southern analysis of restriction digests of genomie DNA indicates that calmodulin is encoded by a single-copy gene.

(1) van Tuinen, D., Barja Protein phosphatase 2A (PP2A) consists of multiple different heteromeric enzymes which have a similar core, composed of the catalytic subonit, a 36 kDa protein, associated with a regulatory protein of 65 kDa, called PR65. PR65 was overexpressed both in insect Sf9 ceils and in bacteria. In both expression systems, the PR65 was soluble and purified to apparent homogeneity. Gel filtration and non-denaturing polyacrylamide gel electrophoresis analyses revealed that purified PR65 could associate with the catalytic subunit of PP2A. This association led to an inhibition of the catalytic subunit towards various phosphorylated peptide substrates but stimulated its activity towards para-nitrophenyl phosphate. Studies with substrates phosphorylated on different sites by different protein kinases, cyclic AMPdependent protein kinase, protein kinase C and casein kinase 1I, indicated that the degree of inhibition apparently depended mainly on the substrate structure but not so much on the sequence surrounding the phosphorylatiun site. Finally, association of the PR65 with the catalytic subunit conferred polycation-sensilivity to the heterodimeric form of PP2A. Tetrandrine, an alkaloid extracted from a Chinese medicinal herb traditionally used in hypertension treatment, inhibited aldosterone production induced in bovine adrenal glomerulosa cells by either K + ion, angiotensin II (AnglO, or adrenocorticotropin in a concentration dependent manner (IC50 = 10 uM). The inhibition of the response to K + by tetrandrine had a pattern very similar to that of Ni 2+, which is known to preferentially block T-type Ca 2+ channels. In addition, tetrandrine dose-dependently prevented Ca 2+ influx induced by K + or Angll, without affecting the Ca 2+ release phase stimulated by the hormone, The effect of tetrandrine on voltage-activated Ca 2+ channels was investigated, under voltage clamp, using the whole cell configuration of the patch clamp technique. T-type Ba 2+ currents were isolated by recording the slowly deactivating currents elicited during repolarization of the cell to -65 mV after various depolarizing pulses. These currents were blocked by micromolar concentrations of tetrandrine. Their sensitivity to vottage activation was not affected by the drug, however, tetrandrine significantly decreased their deactivation kinetics. Tetrandrine also affected L-type currents, as assessed after T channels inactivation for 100 ms, but at higher concentrations of the drug.

In conclusion, tetrandrine affects with a similar efficacy aldosterone production, calcium influx and T-type calcium channel activity. This finding strongly suggests a role for these channels in calcium signalling and control of steroidogenesis in adrenal glomerulosa cells.

Mehlen, P., R~my C., Chareyron, P., Briolay, J., and Arrigo, A,-P. Molecular and Cellular Genetics, CNRS UMR-106, Claude Bernard University, Lyon-I, France.

At normal temperature, the monokine tumor necrosis factor-alpha (TNF) was found to rapidly increase the level of phosphorylation of the unique alpha-crystallin-related stress protein hsp28, The phosphoryMtion of this chaperonin represents one of the earliest cellular effects described for TNF and was found to be C-kinase independent. In cell made thermotolerant to heat shock, the TNF-mediated hsp28 phosphorylation as well as the eytotoxicity to this monokine was greatly reduced, A similar decrease in TNFinduced-hsp28 phosphorylation was observed in stable transformants expressing high levels of the detoxifiant enzyme glutathion peroxidase. These different levels of hsp28 phosphorylation correlated with modulations in the activation of the transcription factor NF-kB. Our results indicate that the TNFmediated hsp28 phosphorylation is linked to the presence in the cell of oxidizing agents, such as free oxygen radicals, known as potential seco[~d messengers. Hsp28 phosphorylation may therefore be linked to TNF-alpha signal transduction and/or to the cellular resistance to these oxidizing agents, We studied the interaction between bradykinin-and caffeine-induced Ca 2+ release from intracenular Ca 2+ stores in a clone of the rat pheochromocytoma cell line (PC12) treated with nerve growth factor (NGF) for 4-6 days. The bradykinin-induced Ca 2+ release is mediated by inositol 1,4,5 trisphosphate tiP3), while the caffeinesensitive store can be depleted by Ca2+-induced Ca 2+ release (CICR). The effect of Ca 2+ release from these Ca 2+ stores on eytosolic free Ca 2+ ([Ca2+]c) was measured by means of Fura-2 single cell microfluorimetry. Caffeine application caused no or only a small Ca 2+ release in ceils kept at rest in normal culture medium. The caffeine-sensitive pool could be filled by Ca 2+ entry into the ceils through either voltage-activated Ca 2+ channels or ligand-gated cation channels. Bradykinin application produced substantial Ca 2+ release in cells kept at rest in normal culture medium. The response was enhanced after K+-depolarizafion of the cells. The bradykinin-induced release of Ca 2+ also caused depletion of the caffeine-sensitive pool by CICR. However, Ca 2-~ released from the IP3-sensitive store was not sequestered into the caffeine-sensiti\,e Ca 2+ store. The caffeine-induced rise in [Ca2+]c was blocked by ryanodine in a usedependent manner. In addition, a substantial use-dependent ryanodine block resulted from the bradykinin-induced rise of [Ca2*]c and subsequent CICR. By contrast, the K +induced rise of [Ca2+]c caused only a marginal use-dependent rynnedine inhibition of Ca 2+ release. Our results suggest an enhancement of the IP3-induced [Ca2+]c rise in the cytoplasm by CICR from the caffeine-sensitive pool. A mathematical model adequately simulates our experimental data. The fluorescence photnbleaching recovery technique has been applied to measure directly the lateral mobility of plasma membrane-localized hormone receptors and examine the role of receptor lateral mobility in signal transduction. Receptors for insulin and EGF are known to be largely immobile at physiological temperatures. This probably relates to their signal transducdon mechanism, which appears to require intermolecnlar autophosphorylation (receptor aggregration) for activation. In contrast, G-protein coupled receptors must interact with other membrane components to effect signal transduction, and it is noteworthy in this regard that the adenylate cydase activating vasopressin (VP) V2-receptor is highly laterally mobile at 37~ We have been able to reversibly modulate the V2-receptor mobile fraction (f) to largely varying extents, and to demonstrate thereby a direct effect on the maximal rate of i~vJxo cAMP production at 37~ in response to VP. A direct correlation between [ and maximal cAMP production indicates that f may be a key parameter in hormone signal transduction Ltt xiv_o, especially at sub-KD (physiological) hormone concentrations, consitent with mobile receptors being required to effect G-protein activation. The effects of adrenergic stimulation were, for the first time, directly measured on single brown adipoeytes by monitoring the intracellular calcium levels. Preadipocytes from mouse brown adipose tissue were cultured according to conditions known to elicit the expression of the uncoupling protein and the /53-adrenoreceptor. Cells cultured for seven to ten days on cover-slips were loaded with the fluorescent Ca+L indicator fura-2AM. Changes of [Ca=*]i were continuously monitored at 340 and 380 nm excitation and expressed as the ratio between these two wavelengths. The results obtained show that norepinephrine (10-rM) induced, in most cells, oscillatory increases in [Ca2+]i, the tXl-agonist phenylephrine (10-6M) induced a rapid [Ca+2] i increase, whereas both /~l-selective agonist tazolol (10-8M) and /3-selective agonist BRL 37344 (10-PM) induced delayed and sustained increases in [Ca2 § Parallel experiments performed on single cells from rat cultured brown adipocytes show similar results for 0t-and /3-agonists; in this model however, norepinephrine induced [Ca2+]i increases without oscillatory patterns. These results will be discussed in relation to adrenergicmediated metabolic and membrane potential changes in this target tissue which possesses all the known adrenoreceptor subtypes. Metformin and phenformin lowered the frequency of the [Ca2+]cyt oscillations in a concentration-dependent manner with IC50-values of 0.3mM and 3p.M, respectively. Parallel application of the biguanides and insulin lead to an synergistic inhibition of the oscillations. By contrast, agents which cause an increase of the cellular cAMP level (glucagon, forskolin, di-butyryl-cAMP) reversed this inhibition. Measurement of the Mn2+-influx into the cell via the quench of the fura-2 fluorescence at 360nm revealed that 0.5raM metformin and 3p.M phenformin inhibited the hormone-stimulated Ca2+-influx across the plasma membrane. In addition, by assessing the mitochondrial membrane potential in hepatocytes in situ with Rhodamine 123 we found no or only a weak depression of this potential by the biguanides in a concentration range sufficient to inhibit the oscillations.

Moritz, W., B0ni-Schnetzler, M., and Froesch, E.R. Universi~tsspital, Abteilung for Endokrinologie und Stoffwechsel, R/imistr. 100, 8091 Z~rich Insulin receptors from 4 patients with the syndrome of extreme, type A insulin resistance were analysed for the presence of defects using EBV-transformed lymphoblast cell lines. The following parameters were examined: (1) insulin receptor number/cell, (2) affinity of insulin and (3) insulin regulated tyrosine kinase activity. Results: All patient receptors displayed normal ligand affinities (Kd of 1.3-2.6 x 10 -t~ M). P2 and P3 expressed markedly reduced receptor numbers (15% of controls) and P6 had virtually no insulin receptors (3% of controls). P1 had normal receptor numbers but impaired autophosphorylation activity. In the presence of insulin the number of 32p incorporated/receptor was 0.35+0.08 for controls (n=8) and 0.53-1-0.16 for P1 receptors (n=4) and in the presence of insulin 2.72_+0.77 for controls and 0.81 __+ 0.26 for P1. Thus, insulin receptors of P1 carry a defects that specifically impairs the activation of the kinase by insulin. The molecular localisation of the mutation should give us insight into the receptor structures involved in the transmembrane activation of the receptor kinase by insulin. In addition we compared 13-adrenergic transmission in control and Zellweger cells with that of the same cell strains after exposure to hexadecylglyeerol a precursor-compound which is known to nearly normalize plasmalogen levels in deficient cells. The presence of introns in CTR gene is in contrast to other members of G protein-coupled receptors genes such as for 0~ z, ~t, [32adrenergic and Ml-muscarinic receptors that are intronless. Interestingly, the seven putative transmembrane domains are encoded by seven separate exons (exons 6 to 12). The 542 bp fragment of the CTR gene promoter containing putative binding sites for the transcription factor Spl and the enhancer-binding protein AP-2 was cloned in front of the luciferase gene. Comparable levels of the luciferase activity were measured after transfection into LLC-PKt, HeLa, COS-I, MCF-7 and T47D cell lines. The CTR gene was mapped to chromosome 9ql 1-12 by in site hybridization using a tritiated eDNA probe and fluorescence in situ hybridization using a genomic probe.

Glucagon-like peptide 1 receptor is coupled to adenylyl cyclase but not to the formation of inositol phosphates.

Glucagon-like peptide 1 (GLP-1) is a 30 amino acids long hormone which is secreted postprandially by intestine L cells. This hormone potentiates glucoseinduced insulin secretion by pancreatic beta cells, thus defining it as a glucoincretin. GLP-1 interacts with beta cells via a specific membrane receptor (the OLP-1 receptor) that has recently been cloned, This receptor belongs to the seven lransmembrane segment receptor superfamily, members of which are coupled to G proteins. Previous studies have demonstrated that ligand occupied GLP-1 receptor activates adenylyl cyclase. This is also the case for the peptidic receptors which are the most closely related in sequence to the GLP-1 receptor. Some of these related receptors (i.e. the calcitonin receptor) are however also coupled to phospholipase C and thus lead to formation of inositol phosphates when stimulated by ligand binding. The presence and role of transduction mechanisms other than the one mediated by the formation of cAMP have not been clearly established in the case of the GLP-1 receptor, We therefore tested whether GLP-1 can activate, via its receptor, the formation of inositol phosphates and a consequent rise in free cytoplasmic calcium concentration. We show here that GLP-t receptor-transfected fibroblasts produce cAMP with an ECS0 close to the dissociation constant of GLP-1 from its receptor (about 1 nM). Similar findings were obtained using several insulinoma celt lines. In contrast, in no situation inositol phosphate produclion was observed. These results will contribute to our understanding of the interactions between GLP-t transduction mechanisms and the glucose sensing machinery of the pancreatic beta cell. In the presence of 100 ~tM gaanosine-5'-O-(3-thiotriphosphate), the K d of t25I-o:-bCGRP binding remained uncha~ged in the cerebellum, but was increased 3-fold in the liver and spleen suggesting interaction with GTP-binding proteins. In accordance, adenylate cyclase was stimulated by CGRP in the liver and spleen, but not in the cerebellum and braiustem. Furthermore, the linear analogue [acetamidomethyl-Cys2,7]ct-hCGRP enhanced cAMP formation in the liver, but not in the spleen. In conclusion, rat CGRP receptors with tissue-specific Nglycosylation but indistinguishable protein molecular mass have been identified in the cerebellum, brainstem, spinal cord. liver and spleen. Activation of adenylate cyclase by CGRP in the liver and spleen, but not in the central nervous system, and by the linear analogue in the liver alone provide evidence for CGRP receptor subtypes. The platelet glycoprotein (GP) Ib (140kDa) composed of the disulphide linked Ib~ (140kDa) and Ib6 (27kDa) chains, forms a noncevalent complex with GPIX (22kDa). GPIb~ can be divided into 2 regions: The N-terminal 45kDa domain, which contains the binding sites for yon Willebrand Factor (vWF) and thrombin and the macroglycopeptide domain, which is highly O-glycosylated. The 45kDa domain contains 7 leucine rich motifs (Ibp and IX have each one) and between Ala200 and Lys297 a hinge region, containing a double loop linked by 4 Cys residues and a region rich in charged amino acids. This region may participate in a shear stress induced conformation change exposing the binding site for vWF. Cross linking was performed with bis(sulphosuccinimidyl)suberate and 3'3-dithiolbis-(sulphosuccinimidylpropionate).

A 200kDa band was detected in cross-linked platelets using gel electrophoresis and Western blotting. After tryptic cleavage some fragments still containing Ib~ and the 45kDa domain of Ib~ were detected indicating that these regions are in contact.

The principle of this new type of sensor is the change of effective refractive index of an optical waveguide, which can be measured and recorded, caused by some change in the waveguide's environment, such as increased concentration of analyte.

Selectivity and sensitivity may be conferred upon the basic sensor optical chip by depositing a phospholipid membrane on its surface.

Many drugs partition favourably into bilayer lipid membranes.

In this contribution~ it is shown how the partition coefficient (PC) of any molecule for which the structural formula is known can be determined.

If the PC is known~ then the coated chip can be used as a sensor for the drug~ with a sensitivity greatly enhanced compared with the uncoated chip. We have recently made two striking observations in toad skins exposed to Hg : (a) an exocytic effect leading to the appearance of long-lasting apical aggregates of intramembrane particles ; (b) anion-induced changes in osmotic water permeability (Pf), such that Pf is low in C1-Ringer (CI-R) and high in SO4-R. Since these Pf changes appear to be related to the presence of apical water channels, we investigated if similar anion effects were found in skins stimulated with a hydrosmotic agent before exposure to Hg. To ensure a stable, high Pf, skins were first stimulated with isoproterenol, fixed with glutaraldehyde and then exposed to Hg added to the outer medium. In a typical example, representative of 6 skins, Pr (/~m/sec) was 51 after exposure to isoproterenol and glutaraldehyde. Despite tissue fixation, Hg caused Pf to fall down to 28. After removal of Hg, Pf rose to 52 in SO4-R and fell to 40 in C1-R. These Pf changes could be repeated several times and, in each instance, Pf values in SO4-R were close to those found prior to Hg exposure. In fixed, non-stimulated skins, anions had no appreciable effect on Pf, The results strongly suggest that : (a) the anion-induced changes in Pf reported here require the presence of apical water channels ; (b) glutaraldehyde fixation very likely abolishes the exocytic effect of Hg. Min~gard, F., Hausel, P., Horisberger, J.-D. and Diezi J., lnstitot de Pharmacolegie et de Toxicologie de l"Universit6, 1005 Lausanne The kidney is a target organ of heavy metal toxicity. Two renal epithelial cell lines, with proximal (LLC-PK 1) or distal (MDCK-I) properties grown on collagen coated filters were used to study the shortterm Na + transport modifications induced by Cd 2+, p-chloromercuribenzoate (PCMB) and p-chloromercuriphenylsulfonate (PCMBS). Shortcircuit current (Isc) and transepithelial resistance (Rt) were measured using a modified Ussing chamber allowing perfusion of the apical and basolateral sides. In Hg-treated toad skins, anions induce striking changes in osmotic water permeability (Pf) : the latter is low in C1-Ringer (CI-R), whilst it increases 4-to 5-fold in SO4-R. Since cells loose Cl and shrink in SO4-R, the question arose as to whether cell volume and/or intracellular C1 concentration modulate the permeance of the water channel. To further investigate this problem, we looked at anion-induced Pf changes in anisosmotic conditions of the inner medium. When isotonic SO4-R was replaced by hypertonic CI-R, Pt fell to its basal value, as previously found with isotonic CI-R. Conversely, when isotonic CI-R was replaced by hypotonic SO4-R, Pf rose dramatically. Since serosal hypertonicity is known to cause, by itself, an exocytosis of water channels in amphibian epithelia, we focused our attention on hypotonic media. In a group of 6 skins, Pf (#m/sec) was as follows : The blastoderm of the gastrulating chick embryo transports sodium ions in dorso-ventral direction. In this study, the blastoderms were mounted in Ussing conditions, allowing the measurement of shortcircuit current (Isc) and total electrical conductance (Gtot). In order to test the influence of the extracellular pH, the dorsal and ventral compartments' (with HEPES pH 7.3 in control conditions) were superfused with pulses of HEPES pH 6.6 of variable durations. The control values were : Isc = 19.2 + 7.6 ~A/cm2and Gtot = 2.4 _+ 0.6 mS/cm 2 (n=1l). When the replacement was done on the dorsal side, the Isc decreased by 13 + 4 % and the Gtot by about 0.2-0.4 mS/cm 2 . However, the decrease of Isc was transient : Isc progressively returned towards control values with a mean slope of 5.6 #A/cm 2/h. Upon the return to pH 7.3, a rebound increase of Isc by 18 _+ 4 % was observed. For each embryo, this rebound was linearly related to the slope of the Isc return and the pulse duration. On the contrary, the Gtot decrease was stable and returned to the control value upon pH 7.3. No such effects were observed on the ventral side. EIPA (1 #mol/1) had no influence on these effects. These results suggest an active proton extrusion mechanism in the embryonic cells.

H.A. is supported by a "UNIL-4506me" grant Cells of an insect cell line (Aedes albopictus, clone C6/36) were grown at low density (0.2 -1.106 cells/ml). Two cells in close vicinity were pushed together to allow formation of gap junction channels. Flow of intercellular current was measured using the dual voltage-clamp method.

We were able to study the conductance of newly formed gap junction channels. The channels exhibited multiple conductance states. The conductance of a fully open channel was 365 pS. The subconductance increment was 1/7 to 1/6 of the maximal conductance. Conceivably, the lowest conductance level reflects a residual conductance. The voltage gradient across the gap junction, 4' had no effect on the conductance of a fully open channel, but affected the dwell time spent at particular conductance states. Small gradients were associated with fully open channels, large ~ gradients with low subconductance states. These results offer an explanation for the ~-dependent conductance of gap junction membranes (bell-shaped, non-zero for large Vj). (supported by SNSF # 31-25333.88).

Voltage-gated Na + channels can be pharmacologically distinguished with respect to their blocking affinities for saxitoxin (STX5 and divalent cations. To understand the molecular mechanisms underlying channel block by these ligands and to localize their specific binding sites on the channel we combined site-directed mutagenesis and electrophysiological studies of channel function. We identified a Cys between the transmembrane segments $5 and S6 of the first homologous domain of the heart Na + channel (SKM2) responsible for its resistance to STX and its high affinity for Zn 2~. The corresponding Cys for Tyr mutation on the STX-sensit&ve Na + channel (SKMI) lowered its affinity for STX by more than 100 times, and resulted in 300 fold increase in Zn 2+ affinity, suggesting a common receptor for STX and divalent cations, and similar mechanisms of channel block. Based on additional mutations we will present a detailed structural and functional analysis of this receptor, presumably located at the external mouth of the channel pore, which modulates the channel conductance.

Krause, R. M., 8ernheim, L., and Bader, C. -R. Division de Neurophysiologie Clinique, HdpitalCantonal Universitaire, CH-1211 Geneva 4 and D~partement de Physiologic, Centre M6dical Universitaire, CH-1211 Geneva 4 During exercise, large quantities of protons can be released in the cytosol of skeletal muscle fibers due to lactic acid generation. Intracellular pH can fall even below 6.5 during intense exercises. So far, three ion transporters have been suggested to participate in proton extrusion in skeletal muscle: the Ha+/H +o and Cl/ HCO s" exchangers and the lactate transporter. Here, we describe the presence of a voltage-dependent proton current (IH) in clonaUy cultured human myotubes using the whole-cen patch-clamp technique. We find that, with pH in = 6.3 and pH out = 8.0, I H is activated at voltages more depolarized than -50 mV, the maximal conductance being observed at voltages more depolarized than +10 mY. I H can also be activated at pH in and pH out = 7.3. The reversal potential of the current depends on internal and external pH, and shifts of the reversal potential can be predicted by the Nernst equation. No or minute contributions of Na +, K +, Ca 2*, and CI" as charge carriers are found, although activation kinetics of I H are influenced by physiological concentrations of external C~-. Various divalent cations block IH in a voltage-dependent manner, being more efficient at hyperpolarized than at depolarized voltages. Cd 2+ is the most efficient blocker, Physiological concentrations of extracellular Ca z+ also block I H with Km of 0.9 mM at-38 mV and 8.1 mM at -8 inV. Variance analysis of the current noise indicates that the elementary event is below resolution of available techniques. A model of I H aCtivation suggests that under extreme conditions, the conductance can reach 40% of the maximum value after 10 action potentials.

Widmer, H.*, Hamann, M.*, Aubry, J;P.#, Baroftio, A.*, Bijlenga, P.*, and Bader C.R . Departement of Physiology , CMU, and GLAXO Institute for Molecular Biology, Geneva #.

Physiological properties of human muscle satellite cells (SC) are difficult to study in situ. In addition, SC cannot be easily distinguished from other cells in the suspension obtained from dissociated muscle. Therefore, we used a new method based on cytometdo flow to obtain a purified suspension of SC (Baroffio et el., in press). Whole cell recording was then performed before the cells divided (i.e. during the first 24 hours after sorting). The mean capacitance of these small cells was 5 pF (n=55). Depolarization beyond a threshold of -45 mV elicited a transient Na + current which was abolished by removing external Na + and was inhibited by high concentration of tetrodotoxin (KD= 3 gM). This current was expressed in 62 % of the cells and maximal conductance was 250 pS/pF _+ 340 (S.D., n=445. An outward K + current was induced by depolarization from -100 mV to levels above -20 mV and was abolished by 20 mM external TEA, This current was expressed in 60 % of the cells and conductance at +40 mV was 63 + 80 pS/pF (S.D., n=21). Thus it is possible to study SC before they begin to divide and we observe that they already express Na + and K + currents. Comparison with proliferating SC showed that in cLonal culture conditions, 99% of the cells expressed the K + current. Furthermore, the K § conductance nom-tatized to the cell capacitance increased by 5 to 10 times. By contrast, neither the percentage of cells which express the Na + current nor the Na + conductance normalized to capacitance are significantly affected by cell proliferation.

Daniel Markovich, Juika For~o, Gerti Stange~ Jfirg Biber and Heini Muter. Institute of Physiology, University of Z6rieh, CH-8057 Ziirich, Switzerland.

Brush border membrane Ha/SO 4 cotransport is an essential step in proximal tubniar sulfate reabsorption in the mammalian kidney. In order to structurally identify this transport system, we have used a functional cloning strategy based on expression cloning in Xenopus laevis oocytes. Injection of rat kidney cortex mRNA into stage VI oocytes showed a 3-fold stimulation in Ha-dependent sulfate transport (Ha/SO 4 cotransport), as compared to water injected controls. Size fraetionation of this mRNA through a sucrose density gradient, gave rise to an even higher stimulation (6-fold over water) of Ha/SO 4 cotransport by an mRNA fraction of 1.8-2.4kb in size. A eDNA library was constructed using the positive mRNA fraction and was screened, by the injection of in-vitro transcribed cRNA into oeeytes and measurement of Ha/SO 4 cotransport. Using sib selection to subdivide pools of colonies, the library was screened until a single colony was identified inducing Ha/SO 4 cotransport. This clone (named NeSt-l) encodes a protein which specifically stimulates Ha-dependent sulfate transport in X.laevis oocytes. NeSt-1 eDNA shows hybridization with RNAs from rat kidney cortex, small intestine and colon, and also some cross-species hybridization with mouse and rabbit kidney cortex RNA. We believe to have cloned an epithelial specific transcript that most likely is a rat renal brush border Ha/SO 4 cotransporter.

S. Dotti-Sigrist and A. Jakob, Biochemisches Institut, Universitgt Basel, CH-4051 Basel, Switzerland Several hormones enhance K+-efflux from liver cells, possibly by opening K+-channels in response to cell volume changes. We investigated the mechanisms of regulation of K+-permeability of isolated perifused hepatocytes by measuring 86Rb-efflux from preloaded cells. Under basal conditions 86Rb-efflux remained stable for 90 min. Addition of 10 mM alanine resulted in cell swelling of 12% and an increase in 86Rb-efflux of 20-25%. Cell swelling induced by hypotonic buffer also resulted in increased of flux whereas cell shrinking caused a decrease. Rising the concentration of Ca2+ from 1.27 mM to 2.54 mM in the medium decreased the 86Rb-efflux and removing Ca2+ by 5 mM EGTA enhanced it independently of cell volume. Ca2+ and EGTA effects could be abolished by 0.5 mM 4-aminopyridine (4-AP) which had no influence on alanine induced 86Rb-efflux. Vasopressin (10 nM) mimicked the effect of Ca2+. In rat hepatoeytes more than one type of K+-channels appear to be operative since extracellular Ca2+, EGTA, 4-AP and possibly vasoprcssin affect 86Rb-efflux independently of cell volume changes. We have recently shown that the activity of the rat hepatocyte NHA is increased in BC, i.e. 4 weeks after bile duct ligation BDL; Elsing, Hepatology 16:119A, 1992), and during liver regeneration, i.e. 2h following PHX (DNlenbach, d Hepato 16 : $7, 19925. This appeared, at least in part, attributable to an increased H+ sensitivity of the antiport. A possible NHA regulation at the transcriptional level, however, remained unexplored. AIM: To determine whether hepatocellular NHA is regulated at the mRNA level after BDL and/or PHX. METHODS: Poly-A RNA was prepared from rat livers 2, 8 and 20h after 2/3 PHX (or sham operation, SH) and from rat hepatoeytes isolated 4 weeks alter BDL (or SH), respectively. Northern blots were hybridized with a specific probe for the rat NHA (RAI; Krapf, JCI 88: 783, 1991). Autoradiographs were quantitated deneltometdcly using glyceraldehyde-3-phosphate dehydregenase (GAPD) as internal standard. RESULTS: PAl and GAPD hybridized as single 5.8 kb and 1.3 kb bands, respectively, with mRNA extracted from both, BDL (or SH) hepatocytes and PHX (or SH) livers. Normalisation of individual RAI to respective GAPD signals revealed steady state mRNA levels of hepatocetlular NHA to be. increased from 2.6• 0.6 (arbitrary units; n=3) after SH to 5.3• (n=5) after BDL 00<0.025. However, 2, 8 and 20h after PIN, steady state NHA mRNA levels averaging 0.51 • (n=3), 0.66• in=4) and 0.7• (n=5), respectively were not signi~cently different from those after SH (0.4• (n=2), 0.6• (n=3) and 0.8• (n=2), respectvey). CONCLUSION.: Rat hepatocyte NHA is regulated at the mRNA level in BC, i.e. 4 weeks after BDL, but not during liver regeneration after PHX. Cause(s5 and mechanism(s) responsible for the increased steady state NHA mRNA levels in BC remain to be determined.

L'Hostis, C., Geindre, M. and Deshusses, J.; Dept. de Biochimie, CH-1211 Geneva 4

The characteristics of proline transport in the proeyclic form of Trvpanosoma brucei were studied on the cells and membrane vesicles. In the cells, uptake displayed Michaelis-Menten kinetics (Km=19 uM and Vm=17 nmol/min/108 cells). The transport system was found to be highly concentrative (more than 200 fold). In competition studies with analogues only L-alanine, L-cysteine and L-azetidine-2-carboxylate inhibited proline uptake. Experiments on the cells with ionophores showed that proline uptake is not energized via an ion co-transport. Iodoacetate (10 raM) decreased proline uptake by hall KCN (lmM) inhibited proline uptake to a lesser extent and the degree of inhibition was proportional to the intracellular ATP concentration. Proline transport experiments on T. brucei membrane vesicles showed an uptake by the vesicles but only in the presence of intravesicular ATP. These results suggest that this proline carrier system is highly specific, ATPdriven and independent of an ion co-transport.

S UBSTRATE DEPENDENCE OF NA-CA EXCHANGE ' GATING ' CURRENTS E. Niggli & P. Lipp, Department of Physiology, University of Bern, 3012 Bern. Flash photolysis of 'caged' Ca :+ was used to produce intracellular Ca 2+ concentration jumps in sin gl e cardiac myocytes. The rapid increase of intracelhtlar Ca 2+ activates Na-Ca exchange current and a brief 'gating' current. The gating current dc~) has been associated with a conformational change of the Na-Ca ex changer molecules after binding of Ca z § The amount of charge moved with I ~ reflects the number of exchanger molecules with the Ca 2+ binding sites facing the inside of the membrane. Combined application of a rapid extracellular solution switcher and photorelease of intracellular Ca 2+ enabled us to analyze the effects of Na+o and Ca2+o on the steady-state distribution of exchanger states. Compared to substrate-free conditions (Li + or NMG solutions), a short (-2 s) exposure to 10 mM Na + reduced I=f to 92.2% 5: 2.4% (mean .-hS.E, n = 11). Addition of 1 mM Ca 2 § had no significant influence on I~ (99% -6 3.9%; n = 9). These remits do not reflect the inward shift of molecules predicted by a consecutive transport model for the Na-Ca exchanger. We conclude that In, f may arise from a conformational change of the Na-Caexehanger molecules after Ca2+ binding other than a membrane crossing transition. Alternatively, the findings could be explained by assuming a two-step simultaneous mechanism for the Na-Ca exchange. Increases of/f observed after prolonged exposure to substrate-free solutions may result from uncontrolled changes of in frog oocytes has established some principles on the structure of this protein. To make it amenable to biochemical analysis an expression system producing large amounts of the protein is needed. To this purpose we have used the vaccinia virus system, a system that allows great flexibility in the choice of the expressing cell lines and which is faster than other viral systems.

Two different cell lines (HeLa and COS) were tested so far: in both significant expression of the exchanger-activity (Na+-dependent/Ca2+ uptake or release) could be detected. After 35S-Met labeling of the cells a band of a molecular mass of 120 kDa was observed, accounting for 3-5% of the total newly syntesized proteins.

An important step in the cardiac signal transduction coupling electrical excitation to contractile activity (ec-eoupling) is the activation of Ca 2+release from the sarcoplasmic reticulum (SR). Evidence is growing that in addition to the "'physiological" activator (Ca:+-current) the Na*-current may serve as a release trigger. Release most likely results from an activation of the Na+/Ca:'~-exchange to transport Ca2*-ions into the cell. But until now evidence supporting this hypothesis was indirect. We tLge, d.a new approach to measure the intracellular Ca 2+-concentration ratiometrically with a mixture of the two fluorescent Ca2+-indicators Fluo-3 and Fura-Red. Ca2+-current and Na+-current induced Ca2+-trausients were analyzed i~ the absence and presence of 20 p/vl ryanodine. The Ca z+transients obtained under the different conditions were clearly distinguishable in agreement with the concept of Ca2+-induced Ca TM release: In addition we were able to demonstrate small Ca2+-transients obtained under conditions where Ca:+-release was blocked by ryanodine. In the presence of ryanodine kinetic differences between Na+-current and Ca2 § induced Ca2+-transients could be found with the fast Ca TM indicator Fluo-3. These results therefore directly demonstrate the existence of Na*-current induced Ca2 § that under conditions of a functional SR may serve as a trigger for CaZ*-release. In addition these findings support the existence of a restricted space underneath the sarcolemma.

Supported by the SNF.

Horisberger, J.-D., Canessa, C. and Rossier, B. C. Institut de Pharmacologic, Lausanne.

A cystein to phenylalanin (Cl13F) mutation in the outer third of the first transmembrane dcraain of the a sulmmit of Na-K-ATPase has been shown to decrease the binding rate of ouabain (~HO J. 11:1681 (~HO J. 11: , 1992 . We have studied the Is + activation kinetics of this mutant by expressing it in Xenopus oocytes and measuring Na-K-ptr~p current under voltage clamp conditions. In the absence of external Na +, the apparent affinity for K + W~S increased, K 89 at -50 mV 45 + 6 pM in the C113F mutant vs 135 + 24 pM in wild type. The maximal K + ~ current (Imax)-was about 50% smaller in the Cl13F mutant. We studied then the ouabain-sensitive charge translocaticn under conditions of Na+/Na + exchange (El .Na <-->E2 + Na). The rate constant for a -50 to +50 mV voltage step was markedly slower in t~e CI 13F mutant, 53 + 3 s-' than in wild type, 288 _+ 12 s-', while the backwazd rate constant of the same step was not affected. Kinetic analysis demonstrates that the change of this single forward step rate constant can explain the slower ouabain binding rate, the lower Imax and the nigher apparent affinity for external K +.

Beron, J. and Verrey, F., Physiologisches Institut der Universit~it Zfirich, CH-8057 Zi~rich.

Aldosterone induces in A6 cells an early increase in active Na pumps, detected by measuring the initial rate of ouabain binding (Pellanda et al., Am. J. Physiol. 262: C899-C906, 1992). It has been postulated that this effect, which precedes a transcriptionally mediated increase in the rate of Na,K-ATPase subunit synthesis, could involve the translocation of Na pumps from an intracellular location to the cell surface. To test this hypothesis we used A6 cells maintained on filter cultures in serum-free medium. As expected, aldosterone induced an increase of the initial rate of ouabain binding (2.5 fold in 3 hours). This effect was parallel to but independent of the increase in transepithelial Na reabserption. Changes in the pool of cell-surface Na,K-ATPase were evaluated using labeling of cell-surface proteins with sulfoanceinimidobiotin or enzyme-mediated radioiodination. Three hours after aldosterone (106 M) addition there was no increase in cell-surface Na,K-ATPase, while after 20 hours and 5 days there was a 1.7 and 2.5-fold increase, respectively. This increase was parallel to that of the total Na, K-ATPase pool measured by Western blotting and to that of the total pump sites measured by ouabain binding. We conclude that the late response to aldosterone is characterized by an increase in total and cell-surface Na,K-ATPase, while there is no indication that aldosterone promotes the translocation of pumps to the cell-surface. The early effect might be mediated by the activation of previously silent cell-surface pumps.

Francesco Hofmann, Peter James, Thomas Vorherr and Emesto Carafoli Institute of Biochemistry, Swiss Federal Institute of Technology (ETH), 8092 Ziirich, Switzerland The C-terminal portion of the plasma membrane Ca2+ATPase (PMCA) contains different regulatory domains. A recombinant C-terminal fragment of the hPMCAlb isoform (E1079 -P1180) was used to study the role of two acidic amino acid stretches located on either side of the calmodulin-binding domain. The molecular mass of the recombinant C-terminal fragment, as determined by electrospray ionization mass spectrometry, was 39 mass units higher than the calculated value (12055 Da). This difference was due to an "EGTA insensitive" Ca 2+ ion, which was located by chymotryptic proteolysis in a fragment corresponding to the last 37 amino acids of the expressed protein. Fluorescence experiments on the dansylated recombinant C-terminal fragment titrated with increasing amounts of free Ca 2+ revealed two additional Ca 2+ binding sites with affinities corresponding to kDS of about 30 and 300 nM, respectively. "Stains all" spectra of different synthetic peptides, corresponding to subdomains of the expressed protein, indicated that the site with the kD of 30 nM was probably located in the acidic sequence on the N-terminal side of the CaM-binding domain; the site with the 300 nM kD was apparently located on the C-terminal side of the calmodulin binding domain or, alternatevely, formed by the appropriate three dimensional arrangement of sequentially distant residues of the domain.

Beggah, A.T., Jaunin, P., Peitsch, M., and Geering, K., Institut de Pharmacologie et de Toxicologic de l~niversitt, rue du Bugnon 27, CH-1005 Lausanne, Switzerland Na,K-ATPase is a plasma membrane transporter that is responsible for the maintenance of the ionic homeostasis of animal cells. In the active form, Na,K-ATPase is composed of a catalytic ct-subunit (Sa) and a 13-subunit (SI~). Subunit assembly is needed for the structural and functional maturation of the enzyme as well as for its transport from the ER to the plasma membrane. In order to characterize more precisely the structural determinants in SI3 which are implicated in subunit assembly, we have introduced point mutations in the last 11 amino acids of the C-terminal. By following the cellular accumulation of Set, the glycosylation processing of SI3 and the cell surface expression of functional Na,K-ATPase in Xenopus oocytes after injection of cRNAs, we could show that mutations in the hydrophobic domain of the C-terminus abolish subunit assembly. On the other hand, mutations in the hydrophilic charged domain have no significant effect on the assembly process. Our data indicate that the hydrophobic domain in the C-terminal of SI~ either directly participates in the interaction with Sa or contributes to the acquisition of an assembly competent configuration of S13. The carbohydrate structures of blotted glycoproteins can be analysed by probing them with lectins (referred to here as lectin-blot analysis, extension of the widely used Western blotting technique). Lecdas detect gtycosylation by non-covalent binding to specific sugar groupings. We compared the lectin binding patterns of Na, K-ATPase (NKA), an integral membrane protein of intriguing heterogeneity, isolated from kidney and brain of different rat strains (spontaneously hypertensive rots SLIP,, Wistar Kyoto normotensive controls WKY, Milan hypertensive and normotensive strains MHS, MNS.). NKA is composed of an ct catalytic subanit (112 kD) and a heavy glycosylated I~ subunit (55 kD). Yet, little information is available on possible heterogenous sugar groupings in NKA isoforms. Loctin-blot analysis was performed with a detection limit of about 0.1 lag glycoprotein with 5 different lectins, chosen for their well described and rather narrow binding specificities; these are conjugated to the steroid hapten digoxigenin by chemical coupling of digoxigenin-carboxymethyl-N-hydroxysuccinimide ester (Haselbeck A. et al. Anal. Biochem. 191 (1990) . The different NKA preparations, electroblotted to a nitrocellulose sheet after gel electrophoresis ( SDS PAGE), were incubated with the digoxigenin labelled lectins for lh at 20~ The bound lecfins were detected by an alkaline phosphatase labelled antidigoxigenin antibody. The staining panems were significantly different among the tested rat brain and kidney NKA preparations suggesting that sugar groupings confer additional heterogeneity to NKA. The human plasma membrane calcium ATPase (hPMCA) gene family consists of at least four members; isoform variability is further increased by alternative RNA splicing at a number of "hot spots". We have used quantitative polymerase chain reaction (PCR) methodology to assess the distribution of the hmnan PMCA1, 2, 3 and 4 mRNAs in various tissues.

The results show that hPMCA1 and 4 correspond to ubiquitous isoforms whereas the expression of significant amounts of hPMCA2 and 3 is restricted to cerebral cortex and skeletal muscle. In contrast to the relative non-specific pattern of overall hPMCA gene expression, the different alternative spliced sub-forms of each gene transcript show a highly tissuespecific pattern of expression with a predominance of complex splicing options performed in the human cerebral cortex. Novel alternative spliced human PMCA mRNAs were detected by RT-PCR for isoforms 2, 3 and 4. A new nomenclature for the precise description of PMCA isoforms has been developed that takes alternative splicing at all "hot spots" into account. 

Maize (Zea mays L, cv LG 11) root membranes were fractionated by dextran density gradient centdfugation. Marker enzymes were used to study the distribution of the different membranes in the gradients and a filtration technique was developed to measure 45Ca2+ transport in sealed vesicles. Most of the Ca 2+ transport activity was associated with the ER. However, a small part of this activity was associated with the tonoplast (corresponding to the activity of the H+/Ca 2+ antiport) and to the plasma membrane. When the Ca 2+ transport was measured in the presence of exogenous calmodulin (1 ~tM), the calmodulin-stimulated activity was associated with the tonoplast vesicles only. This calmodulin-sfimulated Ca 2 § transport was insensitive to monensin, a proton ionophore, ruling out a direct effect of calmodulin on the H+/Ca 2 § antiport. We propose that a calmodulin-stimulated Ca2+-ATPase is associated with the tonoplast of young maize root cells. Data will be presented On the characterization of this calmodulin-sfimulated Ca2+-transport activity.

Two proton-translocating activities are present on the tonoplast of higher plant cells: an anion-sensitive ATPase and a cation-sensitive pyrophosphatase (PPase). Both pumps generate an electrochemical potential difference -of protons, serving as the driving force for several transport processes (uniport, antiport and symport) across the vacuolar membrane. To enhance our understanding of the molecular properties and transport mechanism of these two pumps, it was necessary to reconstitute them into liposomes and to compare the H+-pumping activity with that of native tonoplast vesicles. The ATP-and pyrophosphatedependent proton pumps from tonoplast-enriched vesicles prepared from Rubus hispidus cell cultures were solubilized in the presence of Triton X-100 and reconstituted into liposomes of soybean phosphilipids, using Bid-Beads SM-2 to remove the detergent. The specific activity of the two pumps was greatly increased by the solubilization-reconstitution procedure. Identical characteristics were found for pyrophosphatedependent proton transport in native and reconstituted vesicles. However, the ATP-dependent proton transport of the reconstituted vesicles was no longer inhibited by KNO s.

Siedler M., Ringli, C., and Dudler R., Institut fiJr Pflanzenbiologle, Universitfit Z~rich, CH-8008 ZiJrich

The products of multiple drug resistance (MDR)-like genes are integral membrane proteins that function as energy-dependant transporters of a diverse array of substrates in many eukaryotes. Best known are the mammalian P-glyeoprotein drug transporters which are responsible for the multiple drug resistance phenomenon of many tumor cells. While some P-glycoproteins have recently been shown to have dual functions as drug transporters and chloride channels, other related proteins are involved in signal sequence-independent transport of peptides across membranes and heavy metal detoxification. To explore the possible function of MDR-like gene products in processes such as herbicide detoxification and signal-molecule transport in plants, we decided to analyze such genes in Arabidopsis thaliana. The genome of this species appears to contain a small number of MDR-like genes, two of which we have cloned and analyzed (atpgpl and atpgp2). We have studied the spacial and temporal expression pattern of atpgpl in transgenie plants using promoter/reporter gene chimeric constructs. To explore the function of this gene we have constructed transgenic plants overexpressing either the sense strand or the anti-sense strand of atpgpl.

The results of these experiments will be presented. The membrane topology of two alkane-inducible cytochrome P450s from the yeast Candida tropicalis, alkl and alkZ, was tested by construction of fusion proteins with part of invertase and histidinol dehydrogenase (invHIS4C) and expression in a 5accharomyces cerevisiae his4 mutant. Depending on the localization of invHIS4C on the endoplasmic retieulum (ER) cytoplasmic or luminal side, the enzyme converts histidinol to histidine and allows the his4 yeast strain to grow on histidinol supplemented medium. The N-terminal segments of alkl and alk;' were fused to invHIS4C at different locations. The combination of this in vivo assay with subcellular immunoprecipitations of the expressed fusion proteins allowed us to establish that both P450s contain only one transmembrane domain with their N-terminus located in the ER lumen. Deletions removing the transmembrane domain of alkl resulted in a less efficient targeting to the ER membrane but did not prevent insertion in these membranes. Furthermore deletion of a negatively charged peptide preceding the alkl transmembrane domain in an invHIS4C protein fused after this domain caused the N-terminal end to have a positive net charge and to be oriented in the cytoplasm, thus translocating the remaining of the fusion protein into the ER lumen. P815 mastocytoma cells, expressing the glycosylphosphatidylinositol-anchored (GPI-anchored) murine T cell marker Thy-i were mutagenized by treatment with ethyl methanesulfonate. Thy-i negative cells were selected by successive negative sorting of anti-Thy-l-stained cells. The synthesis of an intracellular Thy-i protein was demonstrated by immunoprecipitation in one clone selected for analysis. Furthermore, the expression of another GPI-anchored polypeptide was shown to be affected, suggesting a defect in GPI anchoring. Analysis of the putative intermediates in GPI biosynthesis using metabolic labeling with [3H]myo-inositol and TLC confirmed this interpretation. This new glycosylphosphatidylinositol anchoring mutant is therefore a potential host for DNA mediated complementation using cosmids or cDNA expression libraries. Light-harvesting pigment-protein complex of photosystem II (LHCII) transferred into a ternary system (isooctane-phospholipids-water) showed i) efficient chlorophyll ~ to ~ excitation energy transfer and 2) energetic uncoupling of accessory xanthophyll pigments from chlorophylls. The light-induced fluorescence decrease of LHCII was found to be monophasic in the law water system and biphasic in aqueous buffer. During a dark intervall of 30 minutes the fluorescence changes were 50% reversible in buffer but completely reversible in the solvent system, indicating that only the fast component of fluorescence quenching, independent of energy transfer to xanthophylls, is active in the latter system. Non-conventional low water systems may provide unique possibilities to study chlorophyll deexcitation processes in photosynthesis.

C. Beghdadi-Rais, M. Schreyer, C. Bron and N. Fasel, Institut de Biochlmie, Universit6 de Lausanne.

All surface proteins anchored in the plasma membrane via a glycosylated form of phospbatidylinositol (GPI) are originally synthesized with a COOH-terminal hydrophobic region which is cleaved soon after translation and replaced by the preformed anchor. The signal which directs glycolipid addition is located in the COOH-terminal region of the precursor protein and has been shown to require a well-defined cleavage/attachment site positionned 10-12 residues upstream of the hydrophobic C-terminal domain. The GPl-anchored Thy-I glycoprotein was used as a model system to further characterize the amino acid requirements of the spacer region separating the anchor acceptor site and the C-terminal hydrophoblc segment. The spacer domain was replaced by different amino acid sequences and the correctly processed, GPI-anchored, molecules were selected by panning. This approach and the use of random amino acid sequences to modify the spacer element will provide informations on the structural requirements of this important domain.

The highly conserved ORF 31 downstream of psbB in Chlamydomonas reinhardtii, appears to be required for maintaining Photosystem II complex activity under conditions of limited chloroplast protein synthesis.

Caroline Monod, Michel Goldschmidt-Clermont, Jean-David Rochaix. University of Geneva, Dept. of Molecular Biology, Sciences II, CH-1211 Geneva, Switzerland.

We have identified for the first time a potential sUess protein required by C. reinhardtii under conditions of reduced chloroplast protein synthesis. The importance of this potential stress protein is underlined by its conservation through evolution from C. reinhardtii to higher plants, not only with respect to its amino acid sequence but also with respect to its location in the chloroplast genome: ORF 31 is always located downstream of the psbB gene, that encodes one of the larger polypelotides of the Photosystem II complex. Having begun to characterize the function of the ORF 31 product is especially satisfying, since approximately half of the subunits comprising the Photosystem II complex are low molecular weight proteins of which the functions are unknown. Several of these small proteins are conserved from cyanobacteria to higher plants and are therefore believed to play a basic role in the structure and/or the functional capability of the PS II complex. Determining the role of these low molecular weight proteins is necessary to complete our understanding of the function of all the coding sequences that comprise the chloroplast genome in algae and higher plants.

The carbohydrate part of the glycopbesphatidyitnoeito[-arlchore (GPI-anchors) of Saccharomyces cerevisiae has recently been tound lo have the same core structure as other eukaryctio GPI-anchors, namely Protehl-ethanolaralne-PO4-(Man(xl,2-or3Manul,2->)

Man~l,2Manctl, 6Manc~l,4GIcNal,6Myolnosltoll-PO4-11ptd (side chair= tin parenthesis). Although

[3H]myo-inositol becomes integrated into newly made GPi-acohored proteines within minutes after addition to growing yeast cells, completed [3H]rnye-inositol labeled precursor giycolipids ready to be transferred onto proteins have not been Iound. Yet such tree GPIs have been descried in trypanoeomes and mammalian cells.When yeast cells were metabeiioaity labeled with [3Hlmyo-inosito[ in the presence ol manoosarNne (2-amino-2-deoxy-D-mecr~ose) we observed novel lipide some ol which were sensitive to nitrous acid and to mild base treatment and resistent to PI-PLC. Since this combination of treatments is quite diagnostic for tater GPI precursors, these novel lipids might be mannosarninecontaining GPI-intermediatea the further elongation of which is arrested because subsequent transfer of mannose onto the C2 position ol mannosamine Is no more possible or they might accumulate because some mannosyitrarlst erase of the OPI-pathway cannot work with SOP-mannosamine.Further studies address the int raceilular addition of the Man~l,2 or 3Manor1,2-side chain. For this we used metabolic labeling with [3H]rnyo-inositol of secretion mutant sec18. Labeled proteins retained either in the ER or the Goigi were isolated and the carbohydrate part of the anchor was released by hydrofluorio acid and then analyzed by paperchromatography in parallel with the appropriate standards. It turns out that the side chaitl is already present on the proteins retained in the ER and must be added in the ER. The cytoskeletal component vinculin has been demonstrated by hydrophobic photolabelling to insert into bilayers of acidic phospholipids. We now show that the higher molecular weight variant metavinculin as well as ~-actinin share the same property. Interestingly, preincubation of ~-actinin with vinculin or metavinculin prior to the addition of liposomes strongly inhibited labelling of e-actinin under conditions of non-limiting liposome surface, but enhanced labelling of vinculin. In contrast, vinculin and metavinculin did net mutually influence their labelling. Using gel filtration chromatography, we could demonstrate that ~-actinin still bound to the vinculin/liposome complex under conditions similar to those used for hydrophobic photolabelling with nonlimiting lipid surface; Our results suggest formation of a triple complex consisting of vinculin,~-actinin and phospholipids. In this complex, both proteins interact at the bilayer, resulting in an altered conformation of the two proteins and,as a consequence,in modified bilayer interactions. features which may contribute to the host's inability to eliminate the pathogenic spirochete. An electron microscopal (EM) study was performed to reveal uttrastructural alterations in human peripheral blood monocytes and murine bone marrow-derived macrophages caused by the uptake of this organism. Scanning EM showed that the monocytes/macrophages engulfed only parts of the extremely long spirochetes, while the remaining parts were projecting out of the cells.

When two phagocytes started to internalize. the same spirochete, transmission EM revealed that these cells fused at their contact regions. In addition, the routine macrophages developed enlarged cisterns of rough endeplasmic reticulum (rER) which were arranged in parallel rows along the cell periphery. These rER configurations separated from the macrophages creating cell membrane-bound rER bodies. In conclusion, release of catabolic enzymes out of leaky phagolysosomes, formation of syncytial cells of the foreign body type, and rER alterations are observed with monocytes/macrophages following the phagocytosis of B.

burgdorferi. These findings may help to explain the host's inability to overcome the spirochetal infection in chronic disease situations. Liposomes are used to transfer drugs as well as various macromolecules to cells; they are believed to enter cells by endocytosis but it is not known if this is true for lymphocytes. We determined the liposome uptake by lymphocytes isolated from healthy donors: fluorescent liposomes prepared in the presence of 6-carboxyfluorescein (CF) by the detergent dialysis method were incubated with lymphocytes at either 4~ or 37~ daring times ranging from 5 to 30 minutes in RPM/-1640 medium. As a control, lymphocytes were incubated in presence of free CF at a concentration corresponding to the total CF released by detergent lysed liposomes. After incubation, the cells were washed in PBS, observed under a fluorescence microscope and analysed either in a spectrofluorometer or in a FACScan. Fluorescence microscopy showed that the lymphocytes incubated with CF-|iposomes were all diffusely fluorescent; they were more fluorescent when incubated at 37~ than at 4~ However the control cells incubated with free CF also displayed some fluorescence although it was less pronounced. The fluorometry and the FACScan results confirmed and quantified these observations: cells were 4-5 times more fluorescent when incubated with CF-liposomes as with free CF which is probably taken up by pinocylosis. Moreover, the lymphecytes became more fluorescent with increased incubation time in the presence of CF-liposomes. The temperature dependence strongly suggests that the liposomes enter the lymphocytes by an energy-requiring process. SNSF 31-25666.88 Glucosidase II is a resident glycoprotein of the endoplasmie retieulum in hepatocytes. It catalyzes the removal of the two inner al,3-1inked glucose residues of the oligosaccharide precursor chains of N-glycosylated glycoproteins. The aim of this study was to clone the cDNA encoding pig gtucosidase lI in order to elucidate its protein structure. Two different approaches were established. (i) a ;~gtl 1 cDNA library was screened with antibodies raised against the SDS-denatured glucosidase II. Five positive clones were purified to homogeneity and their fusion proteins further analyzed. One of them showed immunoreactivity with antibodies recognizing the native enzyme. The fusion protein of another clone and the purified glucosidase II were digested with different proteases. The fragments were separated by SDS PAGE followed by Western blot analyses. The results indicate, that several fragments are shared by the fusion protein and glucosidase H. (ii) proteolytic fragments of glucosidase II were microsequenced. Oligonucleotide primers were synthesized according to these protein sequences and used for PCR amplification of DNA fragments coding for glueosidase I/. Pig liver mRNA was prepared as a template. The amplified DNA fragments wiU be used for screening the ~,gtl 1 library. Combination of these two approaches should yield a full length nucleotide sequence information. SRP, a cytoplasmic ribonucleoorotein, plays an essential role in sorting proteins to the ER. In addition to its signal recognition and targeting activity, SRP also contains an elongation arrest function which effects a soecific arrest or pause in the biosynthesis of'ER-targeted proteins in vitro. Two ~roteins, SRP 9 and SRP 14, and the Alu sequences of SRP RNA are essential for this function. SRP 9 and SRP 14 specifically bind SRP R?TA anly as a heterodimer. Furthermore, the two oroteins form a dimer in the absence ef SRP RNA. The heterodimer is therefore most likely an intermediate in the assembly of the ~A-protein comolex. The analysis of a series of ~-and C-terminal deletion mutants of SRP 9 and SRP 14 has allowed us to define regions in the proteins that are essential for dimerisation and for SRP RNA binding. These observations suggest that both proteins are involved in RNA binding.

te Heesen, S., Knauer, R. 1, Lehle, L. 1 and Aebi, M., Institut ftir Molekularbiologie I, Universitat ZUrich, Htnggerberg, 8093 Ztirich, Switzerland, and lInstitut f~ir Zellbiologie, Universit/it Regensburg, W-8400 Regensburg, Germany. N-linked glycosylation is an essential protein modification occuring in all eukaryotic ceils. The central step is the co-translational transfer of the core-oligosaccharide assembled on the lipid carrier dolichoi to selected Asn-X-Ser/Thr residues of nascent polypeptide chains. This reaction occurs in the endoplasmic reticulum and is catalysed by the enzyme N-oligosaccharyl transferase. In yeast, WbpIp is an essential component of this enzyme. The SWPI gene was isolated as an allele-specific suppressor of a wbpI mutation. SWP1 encodes an essential 30 kD type I transmembrane protein. The depletion of the protein results in a reduced enzyme activity similar to Wbplp. Chemical crosslinking experiments indicates that Swplp and Wbplp form a protein complex essential for oligosacchary transferase activity.

The sulfation of proteins on tyrosine residues is a posttranslational modification that occurs specifically in the trans-Golgi. Proteins containing tyrosine sulfation sites can be selectively labeled with [ssS]sulfate in the trans-Golgi. In order to study intracellular transport out of this organelle for proteins that are not naturally sulfated we fused a nonapeptide corresponding to the tyrosine sulfation site of rat cholecystokinin precursor to the carboxy terminus of a type II transmembrane protein (subunit H1 of the human asialoglycoprotein receptor) as well as to a secreted protein (human ~-l-proteinase inhibitor, A1Pi). Upon transfection into COS-7 cells these fusion proteins were sulfated in vivo. By performing pulse-chase experiments with [asS]sulfate and measuring the appearance of radiolabeled A1Pi fusion protein in the medium we were able to measure the rate of transport of this protein from the trans-Golgi to the cell surface. A number of natural mutants of A1Pi which are defective in secretion, tagged with the tyrosine sulfation peptide, are currently being analyzed.

TARGETING OF ERGIC-53 TO THE ER-GOLGI iNTERMEDIATE COMPARTMENT Itin, C.,-Schindler, R., Kappeler, F. and Haud, H.-P., Department of Pharmacology, Biocenter of the University, CH-4056 Basel, Switzedand ERGIC-53 is a marker protein for the ER-Golgi intermediate compartment (ERGIC) . Cloning of the cDNA of ERGIC-53 revealed a transmembrane protein with type I topology. $urpdzingly the 12 amino acid long cytoplasmic tail eardes a KKXX ER retention motif although ERGIC-53 localizes to a post-ER compartment. This raised the question whether the KKXX is a functional motif. Therefore we replaced the cytoplasmic tail of a surface membrane protein (human CD4) with that of ERGIC-53. The construct was efficiently retained when transiently expressed in Vero and COS cells showing an ER-Iike pattern by immunofluorascence microscopy cleady different from endogenous ERGIC-53. Replacement of the two lysines by sednes in this construct lead to surface appearance that was indistinguishable from expressed human CD4. This indicates that the KKXX-moUf in the cytoplasmic tail of ERGIC-53 acts as a retention signal without conferring ERG IC localization. In contrast, overexpression of E RG IC-53 in Vero and COB cells lead to surface appearance suggesting saturation of ERGIC-53 retention, whereas low levels of expression in Vero cells revealed correct intracenular localization to the ERGIC. To test if the transmembrane 'domain o! ERGIC-53 affects retention we attached the transmembrane and the cytoplasmic domain of ERGIC-53 to the luminal portion of CD4. This construct was efficiently retained in the ER without missorting to the plasma membrane indicating that the transmembrane together with the cytoplasmic domain is not sufficient for correct targeting to the ERGIC. A mutational analysis of the retention motif in the intact ERGIC-53 is in progress. We tested the role of coat proteins (COPs) of non-clathrin coated vesicles in biosynthetic membrane transport by microinjection of epitope-specific antibodies into fibroblasts. The transmembrane g[ycocprotein (G) of the temparature-sensitive mutant tsO45 of vesicular stomatitis virus (VSV) was used as a model system for transport through the secretory pathway. One of the antibodies raised against synthetic peptides of 6-COP inhibits transport of VSV-G to the cell surface and reduces secretion of endogenous proteins. VSV-G accumulates in the cis-Golgi network (CGN) in injected cells at a site proximal to acquisition of resistance to Endoglycosidase H. Since transport of VSV-G from the trans-Gelgi network to the cell surface is not inhibited under these conditions, we conclude that B-COP is involved in a transport step between the CGN and the stack of Golgi cisternae.

Gr0nberg, J., Luginb0hl, U. and Sterehi, E., Institut fLir Biochemie und Molekularbiologie, Universit~t Bern.

Maturation of Lactase-Phlofizin Hydrolase (LPH) (EC 3.2.1.23-62) requires proteolytic processing of a precursor (pro-LPH) to the mature enzyme (m-LPH). Subcellular site and function of this processing are not known. We studied the synthesis, processing and sorting of human LPH permanently expressed in MDCK cellls. LPH was synthesized and proteolytically processed normally by the transfected MDCK cells, the molecular species corresponding to those found in human intestinal epithelial cells. To investigate sorting of LPH, transfected MDCK cells cultured on transwell filters, labelled with [35S]-methionine and LPH was detected by surface immunoprecipitation. Most of the LPH synthesized was inserted into the apical (microvillus) membrane and small amounts were found basolateral. 42% of the LPH immunoprecipitated from the apical membrane was in the mature, i.e. proteoytically processed form. On the basolateral membrane only 20% was in mature form. These data show that sorting of LPH occurs prior to and independantly of proteolytic processing.

Lactase-phlorizin hydrolase (LPH), an intestinal brush border glycoprotein, is synthesized as a single chain precursor (pro-LPH)(Mr = 215 -230 kDa) that undergoes intracellular cleavage prior to insertion in the apical membrane. To assess the role of intracellular cleavage on the transport, function and sorting of pro-LPH/LPH a stable MDCK cell line was generated that expresses LPH. Biosynthetic labeling experiments demonstrated that the transport kinetics and posttranslational processing pattern of LPH in this cell line are virtually similar to those in intestinal cells. Moreover, the enzymatic activity was found to be indistinguishable from that of brush border LPH. The sorting of pro-LPH and LPH was studied by biosynthetic labelings of cells grown on filters.

Here, we could demonstrate that the cleaved LPH molecule was directly sorted to the apical membrane, while uncleaved pro-LPH was found on the apical as well as on the basolateral membrane. The data suggest a correlation between cleavage and direct apical transport of LPH in MDCK cells. A novel secretory pathway has been described in which small vesicles bud from maturing granules, carrying with them an aliquot of soluble granular constituents for post-granular constitutive release. Proinsulin conversion yields equimolar amounts of insulin and C-peptide in granules. Whereas insulin is crystalline in granules, C-peptide remains soluble, and should thus be found enriched relative to insulin in the novel pathway. To study this, transformed B-cells were labelled (20';[3H]leu) then chased. Labelled insulin (INS*) and C-peptide (CP*) were quantified by HPLC. CP*/INS* in lh chase medium was 2.0, in keeping with exaggerated CP* release via the novel pathway. CP*/INS* in cells was 0.61, indicating that 39% of CP* was lost from them. Since only 1.2% CP* had been released, 37.8% was left unaccounted for. There was no CP* degradation in the medium; it was thus intracellular. To determine whether such degradation of CP* occurred in granules, release was stimulated (>40-fold) to discharge granule stores. CP*/INS* in medium after lh stimulation was 1.1, reflecting a normal equimolar ratio in granules. Selective CP* degradation thus arose elsewhere in the cell, most probably in lysosomes. The post-granular constitutive pathway thus exists in transformed B-cells, but it only accounts for a minute loss of CP when compared with that due to intracellular degradation. The same is believed to apply to primary Bcells.

POTENTIALLY INVOLVED IN EXOCYTOSIS Hodel, A., Schaefer, T. and Burger M. M., Friedrich Miescher-Institut, CH-4002 Basel Two membrane proteins potentially involved in the exocytotic process are currently purified and characterized. i) Proteins of light membrane fractions of bovine adrenal medullary tissue are solubilized in the detergent HECAMEG. From the family of proteins binding to fixed chromaffin granules in an affinity chromatography step a 130kD MW protein was selected. This protein can be found in medullary but not in cortical tissue fractions. Tryptie digestion followed by peptide separation on HPLC and microsequencing of selected pepfides revealed 4 novel amino acid sequences. Degenerate oligonucleotides were used to prime PCR amplification reactions on a eDNA template synthesized from adrenal medullary mRNA. The DNA sequence of the 900bp PCR product was determined. No significant homology was found in databanks. The DNA fragment hybridizes to 8kb mRNA from adrenal medulla in Northern analysis. It is currently used to screen an adrenal medulla eDNA library. ii) Syntaxin A, a plasma membrane protein of 35kD MW from rat brain syuaptosomes was postulated to play a role in neurotransminer release in nerve terminals (Bennett et al., Science 257 (1992) ). Using the PCR technique eDNA from adrenal medulla coding for bovine syntaxin was amplified and subcloned into an expression vector. The protein lacking its C-terminal hydrophobic stretch is expressed in E.CoIL The soluble form of syntaxin is used a) to characterize its interaction with purified chromaffin granules in vitro, and b) to test its role in exocytosis upon introduction into permeabilized chmmaffin cells. Proinsulin (PI) is normally converted to insulin in the regulated pathway of pancreatic beta ceils. However, we have reported that some conversion of both human and rat II PI occurs in transfected FAO (hepatoma) cells, which release proteins ofily via the constitutive pathway. We have further studied PI conversion in this pathway as follows: -1) FAO cells were stably transfected with eDNA for rat I PI, which, unlike human or rat II PI, presents putative consensus sequences for cleavage by furin (a candidate constitutive pathway conversion endoprotease) at both its conversion sites. Products released to the medium were analysed by HPLC/radioimmunoassay. Insulin was the major form, with some PI but no conversion intermediates. This contrast with FAO cells expressing human or rat II PI which release mainly intermediates and PI.

-2) COS cells were transiently transfected with human PI eDNA; there was no conversion. Insulin and proinsulin but no intermediates were released when ceils were cotransfected with furin, thereby confn-rning that furin can promote PI conversion. In conclusion, proinsulin can be converted in the constitutive pathway. The interplay between endoproteases (including furin) and proinsulin structural domains determines the conversion products generated. Dept. of Internal Medicine. Univ. Hospital, H LAB 8, CH-8091 Z,,rich, Switzerland Secretion of digestive enzymes involves zymogen granules in the pancreatic acinar cells. This energy utilizing activity may involve nucleoside phosphatases and/or specific attachment of cytosolio molecules to carbohydrate-bearing components of the zymogen granule membrane (ZGM). To characterize the carbohydrate-groups of the few ZGM proteins we used digoxigenin labelled lectins and anti-digoxigenin antibodies labelled with alkaline phosphatase. ZGM proteins were separated by non-denaturing electrophoresis in a 5-15% gradient polyacrylamide gel containing 0.2% CHAPS and 0.1 M TRIS/HCI, pH 9.5. At least four distinct protein bands were detected, all glycoproteins as confirmed with PAS-staining. The proteins were electroblotted to PVDF membranes and incubated with digoxigenin labelled lectins GNA, SNA, MAA, PNA, and DSA. Then anti-cligoxigenin antibody conjugated to alkaline phosphatase was added to probe the bound lectins, 4-Nitroblue tetrazolium chloride was used as substrate for the alkaline phosphatase producing a violet color. The presence of glycosyl residues near the active sites of the nucleoside phosphatases was investigated by preincubafing the gel stripes with substrates AMP, ADP, and ATP and with lectins Con A, WGA, SBA and RCA120. The nucleoside phosphatase activities were detected by visualization of the resulting lead phosphate precipitates in the gel forming in the presence of Pb § The various lectins bind to different glycoproteins of the ZGM and inhibit nucleoside phosphatase activities towards different substrates. The function of these glycoproteins during exocytosis remains to be elucidated. We mutated the first amino acid of the C-peptide of human proinsulin from Gin to Pro to determine the effect of proline immediately C-terminal to one of the two cleavage sites implicated in conversion. C1Pro or native proinsuli n eDNA were stably transfected in AtT20 (pituitary corticotroph) cells. Kinetics of conversion of labelled (10'; [3H]Leu) C1pro proinsulin showed an important cellular accumulation of des-64,65 split proinsulin with only trace amounts of insulin appearing after 120 minutes chase, whereas conversion of native proinsulin progressively generated insulin in cells with a modest accumulation of des-31,32 split proinsulin. HPLC analysis of cellular immunoreactive material in steady state conditions revealed the majority as des-64,65 split C1Pro proinsuliu; for native proinsulin transfectants the major product was fully processed insulin. C1Pro proinsulin is thus not cleaved efficiently between B-chain and Cpeptide. To determine if this was simply due to replacement of the polar (Glu) C1 residue or a unique effect of C1Pro, we deleted the first 4 residues of C-peptide (highly conserved acidic domain), bringing Leu to the new C1 position; conversion was only modestly affected. We conclude that Pro C-terminal to the B-chain/C-peptide cleavage site induces a conformational change and thereby inhibits conversion.

Transfected AtT20 ceils consdtutivelv release des-31.32 snlit nroinsulin.

Jeantet. Centre Mtdical Universitaire, 1211 Gen~ve 4.

Transfected AtT20 cells convert proinsnlin to fully processed insulin and stock it in secretory granules. To see whether these ceils also secrete insulin-like material via the constitutive pathway, we analysed stably transfected clones by pulse-chase experiments: cells were labelled for 10 min with [3H]Leu and then chased for up to 120 min. Radioactivity in cell extracts and medium was analysed by HPLC. During the chase, des-31.32 split proinsulin accounted for approx. 50% of the released material while in cell extracts des-31.32 split prdinsulin never exceeded 15% of the total insulin-like material. To distinguish between regulated and constitutive secretion, cells were first chased for 30 min under basal conditions and then incubated a further 30 min under basal or stimulated (forskolin/IBMX) conditions. Radioactivity in basal medium was 34% insulin, 51% des-31.32 split proinsulin and 15% proinsulin, In stimulated medium there was 46% insulin, 34% des-31.32 split proinsulin and 20% proinsulin. Insulin and proinsulin release was stimulated 2.8 fold while that of des-31.32 split proinsnlin only 1.4 fold. These results suggest that a significant portion of newly synthesised proinsulin in AtT20 cells is released through the constitutive pathway after partial conversion to des-31,32 split proinsulin presumably by furin or a related endoprotease active in the constitutive pathway. The cytoskeleton of fibroblasts cultured in three-dimensional collagen lattices or as monolayers was studied by indirect immunofluorescence. The growth of fibroblasts in three-dimensional collagen gels closely approximates the in vivo environment, where they are surrounded by extracellular matrix. Here the cells were spindle-shaped and showed intense F-actin bundles of different thickness as well as cortical F-actin. The tubulin network was not very elaborated. In contrast, dermal fibroblast cultured as monolayers exhibited F-actin-microfibrils of equal thickness arranged in parallel arrays and intense tubulin networks. The effect of some substances used in wound healing were tested on cytoskeleton organization. In fibroblasts, cultured in three dimensional collagen lattices, Ascorbic Acid and the hemodialysate Solcoseryl resulted in dense F-actin-bundles, whereas TGF-I~ and Retinoic Acid lead actin to be organized in aggregates disposed along the plasma membrane. None of these effects were observed in monolayer cultures.

In additional experiments, we studied both the intracellular microfibrils and the extracellular fibronectin de, position during cell migration in monolayer cultures. Actin and tubulia were organized in thin diffuse microfibrils not arranged in linear arrays. In the absence of any migration stimulating substance, only low amountS of fibronectin were deposited along their migration pathways. However, is presence of TGF-t~ and the hemodialysate SolcoseryI, large amounts of fibronectin microfibrils were deposited oriented along the direction of cell translocation. Synthesis rates and structural organization of cytoskeletal elements may strongly depend on a celrs phenotype, as this is for instance the case for fibroblasts during wound repair. These changes of cell shape and function have been emulated in vitro by cultivating dermal fibroblasts within collagen gel matrices. Here we have investigated the spatial arrangement of microflaments, intermediate flaments, microtubules and fibronecdn with respect to (i) the fibroblasts' location within the matrix and (ii) the progression of in vitro wound contraction by high resolution CLSM of muhi-fluorescently labelled specimens. Serial optical sectioning enabled us to simultaneously observe the cytoskeleton, cellcell contacts and cell-matrix interactions at near-theoretical resolution (-0.3.txm in x, y and z) at various depths within the collagen matrix. Whereas the cytoskeletal organisation of the fibroblasts within the 3-D matrices was comparable to that in vivo, it markedly differed from that of cells growing at the surface of the matrix or conventionally in monolayer cultures. Recent work from this laboratory has established that thrombin causes rapid neurite retraction via activation of cell surface receptors and modulation of intracellular metabolism ($uidan et al. 1992; Neuron, 8:363) . Therefore, we examined what were the cytoskeletal reorganizations that occurred concomitantly with the changes in the neuronal cytoarchitecture. Anti-tubulin antibodies and phalloidin staining were used to label microtubules and actin filaments, respectively and analysis was carried out by confocal microscopy. We observed that after thrombin treatment microtubules and actin filaments appear to colocalize in arrays projecting from what seems to be an MTOC. The protein kinase inhibitor staurosporine which inhibits thrombin-induced neurite retraction prevented this cytoekeletal reorganization. The phosphatase inhibitor phenylarsine oxide mimics the effect of thrombin in causing neurite retraction but this involves changes in cytoskeletal arrangement which were distinct from those caused by thror0bin. By using molecular biological, immunological and immunocytochemical methods we have recently identified two strong, non-variable antigens MARPI (microtubule associated repetitive protein I) and GMS, that could be characterized as highly repetitive proteins located on the cytoskeleton of the parasite. A current study in our lab indicates that MARPI and GM6 represent only two members of a large family of proteins that are high molecular weight components from the cytoskeleton and that at least in part exhibit antigenic properties. Future investigations will show whether these antigens can be used as agents in immunodiagnosis of African trypanosomiasis in human and cattle.

J.E. Rickard and T.E. Kreis, Drpartment de Biologie Cellulaire, Sciences III, Universit6 de Gen~ve, CH-1211 Gen~ve 4 CLIP-170 is a microtubule-binding protein implicated in endosomemicrotubule interactions (Pierre et aI., 1992, Cell, 70:887) . It has also been localized to desmosomal plaques in polarized epithelial cells (Wacker et al., 1992, J. Cell Biol., 117:813) , suggesting that it may function to link diverse cytoplasmic organelles to microtubules. In fibroblast-like cells, CLIP-170 accumulates at the microtubule plus ends (Rickard & Kreis, 1990 , J. Cell Biol., 110:1623 , the site of polymer turnover. The interaction of CLIP-170 with microtubules is regulated by phosphorylation in vitro, and the phosphorylation of CLIP-170 in vivo is influenced by perturbation of the microtubule polymer state (Rickard & Kreis, 1991, J. Biol. Chem., 266:17597) . These data suggest a relationship between the phosphorylation and micTotubule-binding activity of CLIP-170 and microtubule dynamics, and are consistent with a higher affinity of CLIP-170 for the stabilizing 'GTP-cap' conformation of tubulin. A working model will be discussed, and results of experiments in progress to test this model will be presented. Localization of CLIP-t70 to micmtubule plus ends, combined with its ability to interact with other cytoplasmic structures, may provide a cellular mechanism for control of microtubule turnover as well as promoting the microtubule-organelle interactions essential for cytoplasmic organization. cytoskeletons were prepared by Triton-X-100 extraction. Rats were immunized with these cytoskeletal proteins to raise an immuneresponse, the serum was collected and used for western-blots to detect the recognized proteins. One of these proteins, termed Nr. VI, was further characterized with immunofluorescence and immunogold-electronmicroscopy and found to be a MAP (microtubule associated protein).

Ortega Perez, R., Arrighi, J-F., Irminger-Finger*, I., van Tuinen, D., Micheli, A.J., Capelli N., Andrey-T*, I., Edelstein*, S. and Turian, G.

Heat-stable proteins of the crude extract ofN. crassa were purified by affinity chromatography on a calmodulin-sepharose 4B column (Lee, Y. C., and Wolff, J. 1984. J.Biol.Chem.259: 1226-30) and further purified by chromatography on a MonoQ column (Pharmacia). Using monoclonal antibodies against MAP 2, MAP 1, and tau (Sigma) or polyclonal antibodies against the 205K MAP from Drosophila (Goldstein, L.S.B., Laymon, R.A., and Mclntosh, J.R. 1986 . J. Cell Biol. 102: 2076 -2087 we were able to demonstrate for the f'trst time the presence of these cytoskeletal proteins in filamentous fungi. These proteins could be phosphorylated with a calcium-calmodulin dependent protein kinase also isolated from Neurospora (van Tuinen, D., Ortega Perez, R. D., Marmr, D., and Turian, G. 1984. FEBS Lett. 176: 317-320) . The isoelectric point of each of these MAPs was determined. MAP-calmodulin complexes were obtained on a native slab gel using 3H-calmodulin and were analyzed by fluorography of the gel. The MAPs thus isolated from N. crassa were able under appropriate conditions to induce the assembly of porcine brain tubulin into microtubules as visualized by EM using negative staining with uranyl acetate.

Besides the microtubular axoneme, the flagellum of the hemoflagellate Trypanosoma brucei contains another cytoskeletal element, called the paraflagellar rod (PFR). This structure is a highly ordered, three-dimensional fibrous network of yet unknown function, which runs parallel and in close contact to the axoneme. The major structural components of the PFR are two immunologically and biochemically related proteins of 69 and 73kD.

Two different PFR gene loci, each containing two identical tandemly linked genes (A and B, C and D respectively), have been found. It was shown that the gene-A codes for the smaller, 69kD PFR protein, whereas in the case of the gene-B no related full length transcript could be identified yet. Further we know that the bigger, 73kD PFR protein is encoded by the PFR gene-C in the second gene locus. The gene-D is only partially analized yet. Surprisingly gene-C only shows partial (65%) similarity to the gene-A and B, although coding for a very similar protein. Changes in microtubule dynamics and nucleation play a major role in the reorganization of the cytoplasm and in the establishment of polarity in epithelial cells (Bacallao et al., 1989 , J. Cell Biol. 109:2817 . Specific microtubule-binding proteins (MBPs) are presumably involved in these changes. We have identified MBP-115, a protein of 115 kDa in HeLa cells. Sequencing of a full-length cDNA clone coding for indicates that it is a novel protein. MBP-115 is predominantly expressed in cell lines of epithelial origin (Coco-2; MDCK) and it is localized on a subset of microtubules. In Caco-2 cells, the MBP-115 expression level increases upon polarization. Vero fibroblasts, which appear not to contain MBP-115, were used to investigate the role of MBP-115 on microtubule organization by transient expression. In such transfected ceils, microtubules appear to be randomly nucleated and have become more resistant to nocodazole treatment. We are currently studying the role of MBP-115 in the establishment of cell polarity. MAP2c is an embryonic microtubule-associated protein. It is a splice variant of MAP2b, lacking a central region, and is therefore of smaller molecular weight (65-70kD) than MAP2b (280kD). It is present in axons and in glial cells (Garner and Matus, J. Cell Biol. 106: 779, 1988; Tucker etal., J. Comp. Neural. 271: 44, 1988) ). In this study we have investigated the occurrence of MAP2c in aggregated rat telencephalon cultures, consisting either of a mixed population of neuronal and glial cells, or enriched in either neurons or glial cells. These cultures were tested with several monoclonal antibodies against MAP2, and with neurofilament-and glial fibrillary acidic protein-specific antibodies as markers for cell types. Differences in the composition of MAP2 variants were determined by electrophoresis and Western blots. MAP2c was found in two variants, one several kD larger than the other. Both variants existed in a phosphorylated form in mixed cell aggregates. The smaller variant was predominant in neurons, and the larger one in gila[ cells. Supported by SNF grants 31-26624.89, 31-33447.92 and 31-28809.90. 336 

Roger Wepf #, Andreas 13remcr, Ueli Aebl and Heinz Gross # #Institute for Cell Biology, ETH H6nggerberg; *Maurice E. Mfdler-Institute, University of Basel

The function and structural state of F-actin filaments is controlled and modulated by actin-binding proteins (ABP). Visualization and structural analysis of such protein-protein interactions is difficult due to the low signal-to-noise ratio in TEM micrographs. Moreover the low signal-to-noise ratio found in TEM micrographs may not be increased considerably by averaging procedures, since most of the ABP's do not expose the same periodicity as the underlying F-actin helim With the introduction of the "in-lens" Field Emission SEM (FESEM) a powerful tool for macromolecular structure determination has become available. We have found that the signal-to-noise ratio in secondary electron images from a FESEM is much higher than in TEM rnicrographs. This advantage of FESI~M makes imaging of single proteins within a complex possible e.g. the two domain structure of the G-actin molecule within the F-actin filament. FESEM allows us to analysis the structure of single additional proteins (ABP's) or protein-complexes, such as the troponin complex, on the F-actin filament.

Coexpression of cytokeratin filaments and vimentin filaments are described for "mesenehymar' epithelial cells, e.g. lens, thymus, as well as for some epitheliallike cells of mesodermal origin, e.g. Sertoli-eells, mesothelial cells. Endothelial cells of large blood vessels are known to possess vimentin filaments alone correlative with the mesodermal origin of these cells. We are engaged in studies to prove morphologioel heterogeneity of cultured endothelial cells. For this reason, we obtained endothelial cells by scraping the intima of the aorta or the vena eava of the cow. Confluent colonies of separate morphology were obtained from both blood vessels 14 days after cell seeding: "Cobble stone" monolayer for phentoype 1 and polygonal monolayer for phenotype 2. As detected by immunofluorescence localization, phenotype 1 showed a ring of actin filaments, bundled vimentin filaments, a meshwork ot cytokeratin filaments, and single cilia. Phenotype 2 displayed a network of actin filaments and vimentin filaments, yet no cytokeratin filaments and no cilia. Only phenotype 1 expressed cell-celladhesion molecules of the E-cadhedn family on the cell surface. We conclude: Separate endothelial cells can be distinguished in culture at the level of phase contrast microscopy. The occurrence of eytokeratin-pesifive cells questions whether endothelial cells are of mesodermal odgin only.

Towards an atomic structure of F-actin: facts and fiction Andreas Bremer 1 . Roger Wepf 2, Christima Hema 1 , and Uell Aebi 1'3 1Maurice E. Mailer-Institute. University of Basel; 2Institute for Cell Biology, ETH Hgnggerberg; 3Dept. for Cell Biology and Anatomy, Johns Hopkins University School of Medicine, Baltimore, USA Actin is a part of the macromoleeulur machinery that controls intracelinlar transport and the cell shape as well as motility of ceils and tissues. F-actin filaments polymerized in vitro from G-actin monomers are composed of two strands of subunits. An atomic model of the actin filament has been proposed [Holmes etal., Nature 3,17 (1990) 44] and we have investigated whether high-resolution EM data can be used to test this model and help developing it into an atomic structure. We have optimized our strategy for digital image processing of F-actin filaments by including a number of refinement steps. The resulting 3-D reconstructions allow for meaningful comparisons since they are remarkably similar to electron densitiy maps computed from the atomic model. The resolution of our 3-D reconstructions is high enough that we can trace the hydrophobic loop of the actin molecule. However, there are also subtle differences between the EM data and the atomic model: we have found reproducibly less mass in region of the DNase I-binding loop of the actin molecule. In fact, recent refinement of the atomic model [K.-C. Holmes, personal communication] has now yielded a conformation of the actin molecule that is better in line with our reconstructions. Recent modelling [K.-C. Holmes, personal communication] has now also confirmed our previous mapping of the phalloidin molecule in 3-D reconstructions of phalloidinstabilized F-actin filaments [Bremer et at., J. Cell Biol. 115 (1991) 689]. We are currently investigating the molecular basis of the structural changes that accompany the hydrolysis of the bound nucleotide during polymerization to ultimately explain them on the atomic level. 

Real-time fluorescence anisotropy measurements on chemotacticallystimulated neutrophils using trimethylamino diphenylhexatriene (TMA-DPH) as a rotational motion probe revealed sinusoidal oscillations in plasma membrane rigidity. These correspond to similarly-induced oscillations in actin filament assembly and cross-linking, apparent pseudopod size, and membrane-associated respiratory burst activity (Wymann etal., J. Biol Chem. 264, 15829, 1989; ibid. 265, 616, 1990) . Increases in membrane rigidity are correlated with increases in membrane-anchored actin filament levels and burst activity. A more rigid environment would allow longer access to substrate and/or cofactors, and inhibit dissociation of the multisubunit regulatory burst enzyme (NADPH oxidase), suggesting a mechanism for its oscillatory behavior. Support for this idea comes from the fact that inhibiting subunit dissociation by chemical cross-linking stabilizes the activity (Tamura et al., J. Biol. Chem. 267, 7529, 1992) . F-actin has been reported to occur within the nuclei of Physarum polycephalum, Neurospora crassa and of higher eucaryotic ceils. However, evidence for a functional role of nuclear actin in these systems is lacking. By electron microscopy, we have observed nuclear dense granules on thin sections of conidia of N. crassa prevented to germinate, not only after heat-shock (46"C), but also in those heated with the anti-actin drugs cytochalasin A (125 gM) or cytochalasin B (167 I.tM). Dense RNA granules migrate out of the nucleus and disperse into the cytoplasm during conidial recovery of germination ability, after their shift-down to 25~ (1). Our results confirm and extend a preliminary study (2) suggesting that actin microfilaments are someway involved in the export mechanism of presumed preribosomal RNA into the cytoplasm.

(1) Ton That, T.C. and Turian, G. (1984) . Protoplasma 120: 165-171.

(2) Turian, G., Barja, F. and Caesar-Ton That, T.C. (1992) . CellBiol.

Intern. Rep. In Press. The SAP-related protein of Drosophila melanogaster has eight conserved domains that are homologous to the sphingolipid activator proteins, a group of lipid-binding proteins involved'in the degradation of glycolipids. We have obtained a P element insertion (A177.1M3, from H. Bellen, Houston) located within a few basepairs of the putative 5'-end of the SAPrelated transcript, at chromosomal location 100B1,2. Flies carrying this insertion on both chromosomes are viable and produce SAP-related. After mobilization of this P element, we have isolated 14 recessive mutations, which could be divided into two complementatiorr groups. Eleven mutations showed severely reduced viability and number of progeny, 3 mutations showed embryonal lethalif~,. On Southern blots, chromosomal defects could be detected in all i4 mutants. Deletions were found in 11 mutants, 3 mutants carried apparent insertions that resulted from incomplete excisions of the P element. The 3 largest deletions that were not lethal all extended towards the centromere, removing Sapr completely. The 3 lethal mutations all had large deletions that extended at least 8 kb towards the telomere, probably causing a lethal mutation in a gene located distally of Sapr. the PG molecular species of the group 2 increased at the expense of those of the group I. On the other hand, low temperatures favored the formation of 18:2/16:0 and 18:3/16:0 as well as 18:2/16:1(3t) and 18:3/16:1(3t) molecular species. These results will be discussed in terms of the influence of growth temperature on the degree of unsaturation in these different PG molecular species. (Supported by the SNSF 3100-33693.92).

Several authors (Rifldn, 1978 & Hajduk et al. 1989 reported the association of a human serum factor, toxic to T ryp_ anosoma brucei brucei (TrypanoLyric Factor = TLF) with High Density Lipoproteins (I-IDL). Analysis of the distribution of TLF in density gradients and gel filtrations of 6 individual sera indicated variations in the properties which did not parallel the distribution of typical HDL as regarding the size-, density-and apolipoprotein (AI, AII, CIII, E, J) -profiles. TLF activity in all sera could be found in a size and density range of typical HDL (200-400kD, 1.063-1.21g/ml), but in some sera additionally in fractions containing LDL sized molecules (and apolipoprotein B) and/or in a density range exceeding that of I-IDL (> 1,24 g/ml). Sera can contain one or two differently sized populations and one or two density classes of TLF. There was no correlation between any of the observed distributions and the lyric activity of whole serum which also varies considerably. These findings challenge the view of a classical HDL moiety being the lyric factor. Siegentha/er, G,, Hotz, R., Chatel/ard-Gruaz, D., Jaconi, S,, Saurat, J.-H, C/inique de Dermatologie, Hdpital Cantonal Universitaire, CH-1211 GenOve 14 Fatty acid-binding proteins (FABPs) are tissue specific cytosolic proteins involved in fatty acid metabolism. Five types of FABP have been characterized but an epitheliaI-FABP has not been identified so far. Since lipids play an important role in skin, we investigated the presence of FABP in this tissue. Using PAGE-Autoradioblotting, we characterized an epidermal (E)-FABP distinct from liver-, heart-, intestine-and adipose tissue-FABPs. FABP analysis could be performed directly on protein extracts without prior partial purification. E-FABP has a Mr of -15 kDa and binds oleic acid with high affinity but does not bind all-trans-, 13-cisand 9-cis-retinoic acid nor all-transretinol suggesting a high specificiW for lipids. Expression levels of E-FABP were low in epidermis, higher in human cultured keratinocytes and strong in abnormal epidermal differentiation (psoriasis). These findings suggest that epidermal cells might have a distinct fatty acid metabolism compared to other tissues.

Tappy, L., Schneiter, Ph., Riou, J.P., J~quier, E., Institut de Physiologie de l'UniversitG CH-1005 Lausanne et INSERM U 197, Lyon, France. An increased fasting hepatic glucose production has been demonstrated in non insulin dependent diabetes mellitus (NIDDM). In order to determine the nature of this increase in hepatic glucose production, glucose kinetics were assessed in 7 obese NIDDM patients (34+-6% fat mass) and 6 lean controls (18_+4%) from 10 pm to 7 am using 6,6 2H-glucose and U-13C glucose tracer infusions. Total hepatic glucose output was increased in NIDDM (256+_59 mg/min vs 182!38 in leans, p<0.05). Cori cyde activity, calculated as the difference in glucose kinetics obtained with 6,6 2H glucose and U-13C glucose, was increased by 110% in NIDDM, and accounted for the totality of the increase in hepatic glucose production. Cori cycle activity correlated positively with fasting plasma glucose (r=0.685, p<0.001). It is concluded that increased Cori cycle activity is responsible for the increase in hepatic glucose production in NIDDM.

Tschantz, J-Cl., Chiesa, A. and Sunahara, G.I., Nestee SA Research Centre, Vers-chez-les-Blanc, CH-1000 Lausanne 26.

A rapid and routine micro-affinity chromatography method coupled with spectrophotometric analysis was evaluated to determine the concentration of total and free cholesterol, triglycerides and phospholipids in the ~-and &-lipoproteins taken from 50 D1 of serum.This procedure overcomes many of the limita tions presented by other methods such as ultracentrifugation, electrophoresis and precipitation. Using this affinity chromatography method, we report that in 35 male gerbils (50-75 g; 3-4 months old) fed a comraereial diet, the ~-lipops content (in mmol/L) is: 2.85 • 0.81 [SD] and 0.33 • 0.14, total and free cholesterol, respectively; 1.99 • 0.41 phospholipids and 0.64 • 0.42 triglycerides; whereas the S-lipoprotein contains: 0.92 • 0.26 and 0.22 • 0.06, total and free cholesterol, respectively; 0.28 • 0.17 phospholipids; and 0.49 • 0,31 triglycerides.

Glycogen kinetics in rat ~eletal muscle during glucose infusiorL Erik Haesler, Lue Tappy, Jean-Pierre Felber, Eric Jfquier. Institute of Physiology, Faculty of Medicine, University of Lausanne, Lausanne, Switzerland. Muscle glycogen synthesis is a major determinant of insulin sensitivity in man. Alterations of glycogen synthesis and/or mobilization could be of importance in the pathogenesis of insulin resistance. To assess whether simultaneous muscle glycogen synthesis and breakdown take place in conditions promoting net glycogen synthesis, three groups of anesthetized rats were compared: a) infused during 2 hours with 14Clabeled glucose followed by 2 hours of lZC-glucose (4-hours group), b) infused during 2 hours with t4C-glucose (2-hours group), c) not infused (control group). Glycogen concentrations ( gg/mg muscle protein) and glycogen specific activities (SA; Bq/gg glycogen) were measured in the soleus muscles (mean _+ sd):

control 2-hours 4-hours glycogen conc.

13.64 +-5.58 15.91 -+2.18 17.89 +3.34 glycogen SA 0.31 +-0.14 0.25 -+ 0.14 The difference in muscle glycogen z4C specific activity observed between both perfused groups was consistent with the dilution of prelabeled glycogen with newly accreted unlabeled glycogen. Therefore, it is concluded that no important muscle glycogen breakdown occurs during net glycogen synthesis. The conversion of acetate to mevatonate during lipogenesis is controlled by 3-hydroxy-3-methyl-glularyl coenzyme A (HMG CoA), We have studied the effect of 2 HMG CoA inhibitors, compactin (C) and 25-hydroxy cholesterol (HC), on the lipogenesis of human reconstructed skin (acetate incorporation and lipid profile). A dose dependent inhibition of the lipogenesis was obtained with HC (IC50=200 nM), whereas a biphasic effect was observed with C (stimulation between 10 and 1000 nM, inhibition at 10 000 nM). Lipid profile analysis indicated that HC increased the proportions of mono-and di-gtycerides as welt as of wax esters. No effect on lipid profile was observed with C. These results suggest that all HMG CoA inhibitors are not acting in the same way on the keratinocytes iipogenesis. Measuremems of the rate of 02 consumption (QO 2) of photoreceptors in the vertebrate retina have always been hampered by the highly heterogeneous structure of the tissue. This difficulty has been circumvented in various ways, but never to full satisfaction. We report here measurements of the time course of stimulus-induced changes in the QO 2 of single rod photoreceptors isolated from the salamander retina. The method consists in pulling a rod photoreeeptor into a close-fining suction pipene so that the cell lies 10-20 pm recessed; the local partial pressure of O 2 immediately adjacent to the cell is then measured continuously, using an O2-sensitive microeleetrode, while the transcellular eta'rent is monitored via the suction pipette. In these conditions, a train of 90 light flashes delivered at 2 s intervals elicits a biphasic change of QO2: the QO 2 increases during the first minute of stimulation, then decreases below its resting value, and finally returns to baseline a few minutes after illumination is stopped. We are now investigating how this change of QO 2 correlates with the intracenular levels of ATP, GTP and creatine phosphate.

Raddatz, E., Rochat, A-C., Kucera, P. and de Ribaupierre, Y. Institut de Physiologie de lVUniversit4, CH-1005 Lausanne.

Spontaneously beating hearts were carefully excised from 2to 4-day-old chick embryos and mounted in an airtight chamber in which the partial pressure of oxygen (PO2) at the tissue level could he strictly controlled. The hearts rapidly (seconds) and reproducibly reacted to changes in PO 2 within the physiological range. The alterations of cardiac activity in response to variations of PO 2 were investigated using a noninvasive computerized microphotometrlc technique. The data acquisition and processing systems allowed us to determine instantaneou~ heart rate, myocardium shortening, velocities of contraction and relaxation. Hypoxia or anoxia transiently increased the beating rate and resulted in an incomplete relaxation (contracture), a decline in amplitude of shortening a~d velocities of contraction and relaxation.

Reoxygenation slowed down or even stopped the cardiac activity. This effect was dependent on ~he severity and duration of the oxygen deprivation in all developmental stages studied. Such a reoxygenation-induced dysfunction appears to be the embryonic analog of the 'oxygen paradox" in whole adult hearts. Moreover, the presence of glucose protected the myocardial function during anoxia and upon reoxygenation.

Kobza, R., Biihrer, A. and Koller, E.A., Department of Physiology, University of Zurich, CH-8057 Zurich

The ECG changes occurring during acute exposure to high altitude are mainly due to the stimulation of the sympatho-adrenal system (FJAP 53:35-42, 1984) . These changes were compared in ten acclimatized mountaineers returning from a Himalaya-expedition and in ten nonacclimatized volunteers (control). Both groups underwent in supine position a standardized stepwise ascent to 6000m (PB 345 mmHg) in a low pressure chamber; the duration of the entire altitude exposure, apart from the adaptation time before ascent, was 2 h. Results: At 6000m the acclimatized subjects show significantly lower increase in heart rate, no significant P-Q shortening, significantly lower Q-T shortening and S-T depression (V4, %'6) than the non-acclimatized control subjects. Conclusions: The hypoxia-induced ECG changes in the acclimatized subjects are similar to those found in betareceptorblocked subjects at altitude and point to the economizing effect of altitude acclimatization which we have described for the respiratory and cardiovascular system (EJAP 62: 67-72, 1991) .

Bernasconi, .P., Bemasc9ni~ A., Biihrer, A., BORE) Pv and Kohl, J., Physiologiscnes mstitut tier Universit~t, CH-8057 Ziiricn

The purpose of the present study was to ..analyse .the effect of training and trairiing type on me incidence of coordination betw~n running-and breathinl r-h~thms under aerobic and anaerobic connitions. Furthermore, ~e effect of coordination on exercise efficiency was analysed. Three groups, of volunteers either sexes l)articipatea m the study: 10. highly train.ca triatmetes, 6 highly tr~.'nea sprin.ters and I0 untfa)n." ed m~di~a/students. Al1 subjects exercised on a treadmill at work loans of 50%, 80% and 110% of their anaerobic threshold. Each work load was aNplied twice, once with spontaneous and once with paced breathing. Respiratory variables, heart rate and leg movements were continuously. recorded 6y an automatic respiratory anatysis system. The degree of spontaneous coornination curing 50%A.T was me sa.me in ~1 3 groups. l-iowever, at 110% AT students coordinated less man at lower work loads and less than trained subjects. At 50% AT sprinters were less successful in increasing coordination by paced breathing than the other groups. The effect of coordination on exercise efficiency differed between groups and between work loads. In conclusion: (a) Training has no effect on incidence of Coordination at low level aerobic exercise. Co) Endurance training increases incidence of coordination espec_ially at anaerobic work load. (c) Higher degree of coordination is mostly associated with lower oxygen uphake for a given work load, however, other factors as e.g. me level or sympathetic tone inttuence this relationship. The reduction of maximal lactic capacity with altitude exposure may depend on a decreased substrate flow through the glycolytic pathway. If this is the case, provided unchanged blood lactate (Lab) kinetics at the end of exercise, the maximal rate of La b accumulation (A[Lab]/At) should be decreased at altitude. 6 males (32+_4 years, x+SD) performed cycloergometric exercises of increasing duration up tO exhaustion (30-45 s) at 200% of VO2max, at sea level (SL1), after 7 (ALTI) and 27 (ALT2) days at 5050 m (Ev-K2-CNR Pyramid), and 7 days after return to sea level (SL2), as well as at a work load Corresponding to VO2max at ALT2 plus VO2naax at SLI (ALT3). La b accumulation (A[Lab]) was determined for each exercise bout. After exhaustion, the kinetics of La b were followed during recovery. A [Lab] was linearly related to exercise duration. The slopes of the individual regression lines, equal to A[Lab]/At , were lower at ALT1 (0.09+0.02 (P<0.01) and at ALT2 (0.17+_0.05) (P<0.05) than at SLI (0.25+_0.05) and SL2 (0.23_+006). There was no difference between ALT2 and ALT3 (0.17--+0.05). The kinetics of appearance and disappearance of [Lab] in the recovery were similar in all conditions. A[Lab]/A is thus reduced by altitude exposure. Such reduction, which is less marked after 27 days of acclimatization, is compatible with the hypothesis of a reduced substrate flow through the glycolytic Pathway at altitude.

B~'trer, A., Otihwiler, M., Schuster, K., Koller, E. A., Physiologisehes Institut, Universit~t Ziirich-Irchel, CH-8057 Ztirich For our low pressure chamber we developed a measurement system for the breath-by-breath analysis and the subject monitoring. The system ear* easily be adapted for the acquisition of up to 24 instrumentation channels (e,g. for motion sensors). For the accurate registration of the measurement parameters under changing ambient conditions (e.g. ambient pressure) it is necessary to take into consideration that scaling factors can alter. For the optimal compensation of such systematic measuring errors we have configured our measuring system as follows: All peripheral devices (mass spectrometer, pneumotachograph, pressure sensor, pulse oxymeter, etc.) send their signals to a 24-channel multi processor system called 'Data Preprocessing Unit' (DPU). This device corrects systematic measurement errors (e.g. time delays introduced by the various instruments) and transfers temporally correlated blocks of data to two personal computers. These handle the data storage, the on-line data processing, the numeric and graphic presentation of the results, the subject monitoring and the control of the whole measurement system. Exhaustive dynamic exercise with large muscle groups in conditions of chronic hypobaric hypoxia may be limited by central rather than peripheral (metabolic) fatigue. Six subjects (32+4 SD yr) performed forearm (at maximum aerobic power) and leg (at VO2max) exercises, at sea level (SL) and after a one month stay at 5050 m (ALT). Before and after exercise lactate (la) and gases were measured in arterialised blood. Electromyographic activity (EMG) of the forearm flexors and of the vastus lateralis was recorded during exercise and integrated (IEMG), and mean (MPF) and centroid (CPF) f~tequencies of the power spectrum were calcutated. In forearm exercise, at the same load exhaustion time (rex) rex was similar in SI and ALT (5.55+2.39 vs 5.63+2.74 min). IEMG (+193%), MPF (-21%) and CPF (-14%) changed similarly during exercise in SL and ALT. By contrast, leg exercise in ALT could be performed at a 21% lower load than at SL for similar tex (SL 6.51+2.84; ALT 5.09+1.87 min). Whereas in SL, exhaustion was accompanied by an increase in IEMG (+66%) and decreases in MPF (-9%) and CPF (-8%), in ALT no EMG changes were found. In ALT La increased less (SL 12.86+1.45 raM; ALT 6.85+1.40) and changes in blood gases were significantly smaller then in SL. It is concluded that in conditions of chronic hypoxia the central nervous system may play a role in limiting exhaustive exercise of large muscle groups before the appearance of peripheral fatigue.

In rats resection of lung tissue is known to trigger a compensatory growth leading to a more or less complete restoration of the lung structure and function. As yet the molecular mechanisms leading to such regeneration have been poorly investigated. Conceptually, the process of lung regeneration after pneumonectomy is analogous to that of liver regeneration after partial hepatectomy, a system which is far better characterised. One could expect that many genes up-regulated in this system would also be up-regulated after pneumonectomy. One example for such a gene is the hepatocyte growth factor (HGF), also called "scatter factor". Its expression is not liver specific. HGF has already teen shown to be induced in the lung, after injury of the organism at distant site~ Moreover, genes involved in remodelling of tissue, such as metallopmteinases and their inhibitors are likely to be found up-regulated in the process of regeneration of the lung. We have prepared a Lambda gtll (Lambda Zap) cDNA library using mRNA isolated from the lungs of pneumonectomized rats 3 days after surgery. Genes potentially involved in lung regeneration, as those described above are being cloned from this library, and their expression at different dine after pneumonectomy is being analysed using Nol~laem blots. Neuropeptide Y (NPY) is an important component of the sympathetic nervous system and acts as a potent effector and modulator of vascular tone. In vivo and in in vitro vascular preparations NPY not only has a vasoconstrictor effect per se but also potentiates the effects of other vasoconstrictors such as norepinephrine and angiotensin II (Angll). The human erythroleukemia (HEL) cell line has proven a valuable tool to study NPY-mediated intracellular effects. These cells express a well defined Y1 subtype NPY receptor coupled to calcium signalling and to inhibition of adenylate cyclase. HEL cells express a whole series of functional membrane receptors, among them an a2-adrenoreceptor. We demonstrate here that they also contain AngII receptors of the AT1 subtype. The receptor density is extremely low (<500 binding sites/cell) but affinity for AnglI is very high (Kd in the sub-nanomolar range). We' also demonstrate that these binding sites are coupled to intracellular calcium signalling. We have further studied potential calcium signalling cross-talk between NPY receptors and a2-adrenergic or AnglI receptors. While there exists no detectable synergism between stimulation by NPY and norepinephrine, the calcium response to AngII is potentiated by over 50% when HEL cells are first stimulated with NPY. The exact mechanism of this effect is under investigation. We wanted to compare respiratory and circulatory responses to electrically induced dynamic leg exercise (EE) with responses to voluntary exercise (VE) or passive movement. For this purpose, torque and/or angular velocity of the lower leg must be the same for these different modes. Unfortunately, force produced by electrical muscle stimulation is not controlled as well as with voluntary contractions. So, for the same force production with EE and VE, voluntary work must be a copy of electrically induced work. For this reason, EE data must not only be recorded by a dynamometer but also shown to the subjects when performing voluntary exercise. Since available dynamometers do not include this possibility, we developed a new one with the following special features: (1) means of the torque-velocity-profiles, i.e. recorded in EE trials, are calculated and shown to the subjects on a feedback system, (2) torque-velocity-profiles are reproduced as velocityprofiles by the dynamometer for passive leg movements and (3) torquevelocity-profiles are presented to the subjects as torque-time-profiles for static reproduction of the dynamic work done before. These new features now allow comparisons of respiratory and circolatory responses to dynamic work, static work, and passive movement.

To estimate the fraction of capillaries that are pcrfasod under physiological conditions and thus available for gas exchange or substrate delivery, a morphometric analysis combined with a perfusion marker is required. To this end we developed a new method to demonstrate in vivo capillary perfusion by light and electron microscopy. To mark the blood plasma, 8 nm colloidal gold particles were prepared, coated with rabbit seram albumin, concentrated by centrifugation and dialysed against Ringer solution. In anaesthctised rabbits 4-5 ml of the final concenUate was injected via a catheter into the right atrium. Blood samples showed a slow decrease of gold particle concentration in plasma with a dose dependent half-time. After variable post-infusion times the circulation was interrupted by a snare around the heart, and the organs were fixed by immersion in Kamovsky fixative. Electron microscopic analyses of capillaries in heart, skeletal muscle, lung, liver, spleen, intestine and brain revealed colloidal gold particles homogeneously dispersed in the blood plasma. Morphometric analysis of heart, skeletal muscle and lung showed that the entire capillary bed was marked by the tracer even within two minutes following infusion. Silver enhancement of light microscopic sections showed areas with different staining intensities suggesting heterogeneous peffusion. We conclude that this plasma tracer allows quantitation of labelled capillaries using electron microscopy and, after silver enhancement, qualitative studies using light microscopy. ( Mammalian cardiomyocytes synthesize and secrete atrial natriuretic peptide (ANP), a vaserelaxant hormone exhibiting natriuretic and diuretic properties. We studied the roles of protein kinase C (PKC) and endogenous prostaglandin production in angiotensin II (Ang II)-induced ANP release in cultured, spontaneously-beating rat cardiomyocytes. Stimulation of cells with 0.1 `aM Ang II lead to increases in particulate-bound PKC activity, cellular prostacyclin (PGI2) and prostaglandin E 2 (PGE2) production, and in ANP secretion. A role for PKC in Ang II-induced ANP release was apparent insofar as the PKC inhibilor CGP 41521 (1 `aM) strongly suppressed Ang II-induced ANP release, as did PKC down-regulation. Immunoblotting experiments indicated that Ang II induced the activation of both Ca+*-senaitive PKC ct and Ca § PKC 8 in these cells. Furthermore, Ang II-induced prostaglandin production and ANP secretion were strongly correlated, suggesting a role for prostaglandins in the response. This was confirmed by finding that the respective phospholipase A 2 and cyclooxygenase inhibitors quinacrine (10 I.tM) and indomethacin (10 aM) fully abolished Ang II-induced ANP secretion, while exogenous PGI2 (1 `aM) and PGE 2 (0.1 `aM) induced significant increases in ANP release in this tissue. Our results suggest that Ang II induces ANP secretion in cardiomyocytes via a PKCdependent pathway involving a myocardial prostanoid receptor. Intestinal motility depends on interactions between myogenic, endocrine and neural factors. The influence of these factors on the contractile activity was studied using a 5-7 cm segment of terminal ileum, perfused arterially with oxygenated fluorocarbon solution and luminally with normal saline. Oral and aboral luminal pressures, luminal output and video images of the segment were recorded simultaneously for 2 hours.

Microscopy showed intact mucosa and enteric neurons indicating good loop viability. 2 motility patterns were observed: 1) High amplitude low frequency contractions (HALFc), with frequency of 0.27/rain and oral and aboral peak pressures of 17 hPa and 15hPa, respectively. They originated mostly in the fkst oral cm of the loop and propagated for 3.5 cm. They were associated with fluid ejection (0.2 ml/contraction). 2) Low amplitude high frequency contractions (LAHEc), with frequency of 27/rain and amplitude of 0.3 hPa. They propagated in both directions. Luminal distension increased frequency and amplitude of HALFc but did not affect LAHFc. Tetrodotoxin abolished HALEc and increased amplitudes and tonus of the LAHF. Butanedione monoxime allowed to describe the active and passive mechanical properties of the loop.

We conclude that the ileal segment separated from the CNS shows coordinated activity (HALFc are controlled by the ENS, LAHFc are of myogenic origin) and allows to study the role of the ENS in intestinal motility.

(Supported by SNRF Grant No.32/26369.89)

Baroffio, A.*, Bochaton-Piallat, M.-L.@, Gabbiani, G.~, Aubry, J.P.+, Kaelin, A. # and Bader, C.R.*, d~partements de physiologie* et de pathologie@, CMU, GENEVA ; +Glaxo Institute for Molecular Biology, Plan-les-Ouates, GENEVA; #d~partement de p~diatrie, HCUG, GENEVA.

The expression of muscle-specific cytoskeletal proteins, such as desmin, ~striated (o~sr) and co-smooth muscle (o~sm) actin isoforms, was studied in quiescent, proliferating and differentiating human muscle satellite cells (HMSC), using immunocytochemical procedures. Quiescent HMSC, studied directly after enzymatic dissociation from muscle fibers, expressed desmin, but neither ~.sr nor o~sm actin. Protiferating mononucleated HMSC cultivated as clones, contained desmin and both actin isoforms, ~sr and ecsm. Inside each clone, at least two subpopulations of cells were found, one containing both desmin and c~sr actin, and another containing only ccsm actin. When clones of HMSC were cultivated in a differentiation medium, part of the cells fused to form myotubes and strongly expressed ~sr actin and desmin, whereas other cells remained mononucleated and contained either desmin or asm actin. These results demonstrate that fl only desmin is expressed in quiescent HMSC, 2) two different muscle actin isoforms are synthetized in HMSC when they proliferate, and 3) single HMSC give rise to an heterogeneous progeny, as concerns the expression of muscle-specific cytoskeletal proteins.

K. Jostarndt; A. Puntschart, H. Hoppeler, R. Billeter Institute for Anatomy, Biildstr.26, CH-3000 Berne 9, Switzerland and Reasearch Institute of the Swigs Sports School Magglthgen, CH-2532 Magglingen, Switzerland.

Human muscle adapts to changes in its panem of use with alterations in the concentrations of key enzymes and structural proteins. Data from animal models indicate that such modulations are accompanied by changes in the steady state concentrations of their respective RNA's. Our focus is the adaptation of human muscle to exercise training. Muscle tissue is obtained from needle biopsies (Bergstrrm technique). RT-PCR allows quantitative assessment of specific RNA's in as little as one cryostat section. A requirement of quantitative PCR is the control of the efficiency of the process during the logarithmic phase which in our hand is followed by the incorporation of radioactive label in its product. Thus we are able to distinguish less than 2 fold differences in RNA concenlration. A further approach involves in situ hybridization in order to localize particular RNAs in the different fiber types. We use 33p labelled RNA probes in these experiments. This isotope has an energy spectrum comparable to 35S, but probes labeled with 33p nncleofides are less prone to oxidative modifications than 35S probes, which eliminates some of the background problems. In situ hybridization with 33p lableled myosin light chain If mRNA on chicken muscle gave signals that were at least as distinct as the ones with ass labeled probes. In human muscles hybridizations with probes for carbonic anhydrase III-and S100o~-mRNA showed preferential localization in type I fibers., which corresponds to the localization of the respective proteins.

Jean-Claude Perriard Institute for Cell Biology, ETH, 8093 Z~irich We present the sorting and the interference with the sarcomeric organization of 2 isoprotein families: the Myosin Light Chains (MLC) and the Actins. MLC form an isoprotein family with members belonging to muscle and to non-muscle tissues; the Actin protein family is composed of six isoforms, which can be subclassed into cytoplasmic, smooth muscle, and striated muscle isoforms. The isoform distribution is studied by microinjection of eDNA constructs, containing an VSV-epitope, into the nucleus of regenerating Adult Rat Cardiomyocytes (ARC). Immunological staining of the epitope showed that muscle isoforms integrate to the sarcomeric organization, which is not the case for cytoplasmic isoforms. The expression of cytoplasmic 7-actin has a striking effect on cell morphology, and affects the myofibrillar organization. We show evidence that the thick filament organization can be independently maintained while the thin filament organization disappears, and that the third filament structure (titin) remains sarcomerically organized. An involvement of the smooth muscle of the corpus cavernosum in 1he pathogenesls of impotence is still hardly debated. UItrastructural changes have been described but il is not known if they are directly correlated to a modification of Ihe muscle/connective tissue ratio.

To try to answer this question a computer based method has been developed allowing ihe quantification of the muscle tissue in Ihe corpus cavernosum, Tissue samples from lhe corpus eavernosum of patients suffering from impotence were fixed in formalin and processed according zo the usual technique for observation by light microscopy. The fight microscop~c images or the Goldner stained sections were dlgilallzed with a CCD camera. TO obtain maximal contrast between the muscle and the surrounding structures the slides were illuminated with a monochromatle light source.

The quantification was performed according to Ihe gray level and the texture differences using an IBM PC compuler.

We present here an image processing technique with allows more reliable measurements by removing the inhomogenlties in illumination, by normalizing each pixeI sensitivily and segmeming the image using a fuzzy logic algorithm.

We believe that a fuzzy logic approach can give more precise results in tlssur quantification, since il gives Ihe possibility of considering several parameters, such as the gray level and Ihe texture.

The preliminary results of these quantiflcatJons do not show significant differences between impotent and control patients. Myomesin is a 185 kD myofibrillar protein which is located in the middle of the sarcomeres, and is one out of the three known M-band proteins: M-protein, myomesin and the muscle specific isoform of creatine kinase. It has been speculated that myomesin might bridge the gap between the titin filament system and the thick-filament system at the M-band, and therefore, it seems to be essential for the fundamental architecture of myofibrils. The aim of this project is to identify eDNA clones of myomesin which will lead to the protein sequence and enable us to examine functional aspects. A chicken heart muscle ~.-gt 11 eDNA expression library was screened by DNA hybridization and by a monoelonal antibody against myomesin. The used probe was a fragment of a cDNA which was found in earlier screenings of a leg muscle library by polyclonal antibodies against M-band protein. This eDNA is thought to contain skeletal muscle specific myomesin. We have identified 40 positive clones with the eDNA probe. Five of them were also detectable in the immunescreening with the monoclonal antibody B4. The resulting derived protein sequences are compared to other myofibrillar components. The expression of these sequences was shown to be specific for sarcomeric muscle.

In order to localize the complementary polymerization sites of fibrin, limited proteolytic digestion of fibrinogen with plasmin resulted in the production of subdomains and smaller peptides. Following the original plasminolytic digestion of fragment D1 in the presence of EDTA we found that EDTA alone was eluted from a Superose 12 column at the same elution volume as the carboxy terminal y-chain peptides. Pure EDTA tetrasodium salt in water showed a pH-dependent change of optical density detected in the ultraviolet region. The most important change was found between pH 4 and 8, corresponding to the deprotonation of the third ionizable group having a pK of 6.2 This observation is interpreted as formation of nanoaggregates by a mechanism which resembles the formation of lipid micelles. The absorbance/size change was reversible and was depended on the addition/removal of only 1 proton equivalent. The existence of nanoaggregates was confirmed by size-exclusion chromatography and dialysis experiments, as well by cryo-electron microscopy.

P.Werner*, U.Hiibscher*, S.Amold ~ and C. Schelling 1" *Institut flit Veterin~biochemie und ~ for Geburtshilfe, Jungtier und Euterkrankheiten, Universit~t Ziirich and %Departement of Veterinary Pathobiology, Texas A&M,College Station, USA Recently idiopathic glomerulonephritis has been diagnosed in Bernese mountain dog (BMD) more than in other breeds. According to the known inbreeding rate a breed predisposition to develop this kidney disease has been demonstrated.Other hereditary glomerulopathies have been recognized in some breeds, e.g. an X-linked recessive glomerulopathie in Samoyed is a model of hereditary nephritis in man, e.g, Alport Syndrom caused by a type IV collagen defect.We collected blood from 150 BMD's of the Swiss breeding population and generated one-dimensional fingerprints by using minisatellites (Jeffrey's probes 33.15/33.6). The therby detected polymorphism between dogs within this breed was very small to directly use fingerprint-pattern as genetic markers in linkage analysis. We therefore started searching restriction fragment length polymorphisms (RFLP) using various single locus probes of human type IV collagen. The results so far indicated that polymorphisms exist in the collagen genes of the BMD. Studies are in progress to further characterize the degree of polymorphism in this breed and any disease association with molecular probes.

Malina H.Z. and Martin X.D., The University of Zurich Department of Ophthalmology, 8091 Zurich Kynurenamine and 3-hydroxykynurenamine, 2 metabolites of tryptophan, are present in the lenses of humans and animals. However, the first enzyme of the tryptophan degradation pathway, indoleamine 2,3-dioxygenase (IDO), has not been previously found in the lens. Here, we measured IDO activity in an extract of bovine lenses. The metabolites formed were detected and measured by HPLC coupled with electrochemical and fluorometric detection. IDO activity was 1.48 nmol/g of lenses/hour. During the reaction, kynurenamine and 3-hydroxyantranilic acid are formed, indicating the presence not only of IDO but also of other enzymes of the tryptophan degradation pathway. IDO works by scavenging oxygen radicals and it probably has great importance in the protection of the lens against oxidative damage, which is considered to be a cause of cataract formation. Supported by Swiss National Fund grant 32-9527.88. To investigate the role played by two natriuretic factors, atrial natriuretic peptide (ANP) and ouabain-displacing factor (ODF), in the regulation of the extracelinlar volume (ECV) under physiological conditions, changes in ECV have been induced in 5 volonteers male subjects. They were overnight-fasted and fluid-deprived before being submitted in random order to the following experiments at a week-time interval: ECV expansion (infusion of 21 NaC1 0.9%), plasma expansion (infusion of 500 ml albumin solution) and ECV reduction (diuretic administa'ation). Changes in ECV were monitored by electrical body impedance (Z) using a new-developed quadripolar, 1 ld-h, battery-operated impedance meter. The correlation between changes in Z and measured changes in ECV from the fluid balance sheet gives a r=-0.98, p<0.001. The bivariant regression between Z and the inulin dilution space (IDS) was significant (r=0.91,p<0.001). Plasma ANP (corrected for passive ,dilution) is linearly related to changes in Z/IDS from -15 to +15 %, without changing the nattiuresis. No significant change independent from passive dilution was seen in plasma ODF. This indicates that ANP but not ODF is responsive to physiological changes in ECV size but that ANP is not nattiurelie in hydropenic conditions. In previous studies we found that the human ciliary body is able to synthetize indoleamines and that the enzyme, indoleamine 2,3-dioxygenase (IDO) which cleaves the pyrrole ring of the indoleamines such as tryptophan and melatonin to produce kynurenines or kynurenamines respectively, is present in aqueous humor, ciliary body, and retina. Kynureninase catalyzes the hydrolysis of 3-hydroxykynurenine (3-HK) (yielding 3-hydroxyantranilate). KA has previously been found in many tissues but not in the eye. In this study we measured KA in iris/ciliary body. The measurement of 3-hydroxyantranilate formed from 3-HK was made by HPLC with electrochemical detechion. The KA was 6 nmol/mg protein/hour. 3-hydroxyantranilic acid is an intermediate in the pathway to nicotinamides nucleotides and its accumulation in the lens could be responsible for senile cataract. l lB-HSD la is the enzyme that converts active glucocorticnids (cortisol, corticosterone) into inactive compounds (cortisone, I 1dehydroeorticosterone) and vice versa. A truncated rat Ilfl-I'/SD l b has been recently cloned. Its coding sequence is identical to that of the previously cloned rat liver llB-HSD but the first 26 amino acids are missing. It has been proposed that ll8-HSD lb could insure mineralocorticoid specificity by inactivating glueocortieoids and preventing their binding to the mineraloeorticoid receptor. To test this hypothesis we have injected 1 lfi-HSD lb cRNA into ooeytes and stably transfected 11B-HSD lb eDNA into a mineraloeorticoid responsive cell line . Oocytes injected with the truncated eRNA were unable to induce dehydrogenase activity whereas oocytes injected with 11g-HSD la cRNA inactivated eorticosterone. Although transfected TBM cells highly expressed the truncated transcript, no enzymatic activity was detected. In contrast, TBM cells transfected with I Ig-HSD la eDNA rapidly converted 11-dehydrocorticosterone into cortleosterone. These data suggest that IlB-HSD lb is inactive and therefore not involved in mineralocortieoid receptor protection. The molecular cloning of toad IIB-HSD isoforms from liver and bladder tissue is in progress. This should help to better understand which isoform plays the critical role in mineraloeortieoid receptor protection. At the end of gestation most amino acids have a higher concentration in fetal man in maternal blood. In conterast, the level of glutamic acid (Ghi) is higher in the maternal than in the fetal circuit. There is no net transfer for Glu across the placenta in either direction, Ha~enatl transport and.metabolism of Ghi was studied using an in vitro I~usiun methoa of neman ptacenta ,mill separate closed cg~mts for maternal and fetal s~de (n=~. Glu was added at 250 Itmol/l together with 1-*'~-Glu (5-10 ttCi) to both an'am'is, t:old ~ radioactive amino acids were analyzed by HPLC ~ by enzymatic reacuon. In addiuun, placental tissues before and after perfusion were hornesenized and the activity of glutamine synthatase was determined by the highly sensitive y-giutamylWansfexase assay according to Levintow (1). Both cold and labelled Glu showed an uptake from both sides with a maternal to fetal gradient. The final Gin concentranon alter 4 hours of perfusiun in the mmernal and fetal circuit were 217_+77 and 120+...53 tlrnol/l, respecnvely. Although no glummine (Gin) was added to the perfusate, fmui cunconlratious of Gin in the matornul and fetal circuit were 85+--22 and 131_+21 pmo..~l, respeelavely. The release of Gin in both circuits w asl,signifieantly corre.mte;a wlm me respeelave up.tare of Glu. HPLC analysis of the "~-label has cleany snown a cunversmn of Glu into Gin with a higher fi'action released into the fetal (20-41%) than into the maternal (11-29%) circuit. The overall ~uction of Gin was i;2-1=5 pd'nol/l~.tissu.e. Tbe tissne hom.ogenates showed a specific transferase activity m ~o-/t~ mu/g assne (n=to, msigmncamt cnange between before and after perfusion), which in analogy with other tissues (1) corresponds to an enzymatic Gin synthesis capacity of about 1-3 pmol/k/g tissue. This range is in good agreement with the actual preducuon of Gin during .the perinsion experiments. The results stt~gest that pla~n.tal GIn. . syntheeeeeeeeeLas may .be liimted .by the respacgve enzyme aetlVlty under pe~. usapn conmuons. I ne acuvity o* gmtamme synthetese Is highly dependend on ATP, indicating that the human placenta during in vitro perfusion is maintaining its energy metabolism. (1) Schutz, Y., Qualm, Y., Cavelaars, A., Schneiter, Ph., Institute of Physiology, Faculty of Medicine, University of Lausanne, Lausanne, Switzerland.

The aim was to develop a new non invasive method to estimate total body water (TBW) based on the utilization of a single dose of urea labelled with 15N (40 mg, 99% enriched, non radioactive isotope) administered per os diluted in water. Fourteen young subjects (63 kg, 22% body fat) were studied. The rate of elimination of 15N urea in urine was determined over a 8-hour period using short urinary timed samples. The extrapolation of the enrichment decay at zero time (t O ) was used to obtain the urea space assumed to be equivalent to TBW. The TBW estimated from ISN urea was found to be significantly lower (31.9+9.1 L, p<.01) than that assessed independently by a combination of bioelectrical impedance and anthropometric measurements (36.7+7.6 L), as well as from ethanol dilution space (36.3+8.8 L). Various factors which may explain the systematic underestimation of TBW based on 15N urea will be outlined.

Kikinis, Z., Cochary, E.F., and Paulson, K.E.. Tufts University, Boston, MA 02111, USA.

Malic enzyme, the donor of much of the NADPH required for de novo fatty acid synthesis, is tightly controlled by nutritional conditions. We examined the effects of fasting followed by refeeding a high carbohydrate/no fat diet on expression and localization of malic enzyme (ME) and fatty acid synthase (FAS) in mouse liver. In situ hybridization showed that both lipogenic enzymes were expressed at low basal levels in all bepatocytes in livers of mice fed a control diet. The temporal response of dietary-induced expression of both genes was studied at 0, 6, 12, 24, 36, 48 and 72 hours, and 8 days. In situ hybridization showed that ME and FAS mRNA was induced initially in periportal cells within 6 hours. By 24 hours, induced expression in periportal hepatocytes was maximal. This was coincident with expression of ME and FAS mRNAs which appeared to be maximal between 24 and 36 hours as measured by slot blot analysis. Both mRNA levels and portal expression of ME and FAS then decline: Transcription rates were measured by nuclear run-on assay and demonstrated that maximum transcription rates precede maximum mRNA levels by peaking at 12 hours. Furthermore, run-on assays showed that the periportal induction by carbohydrates is mostly a transcriptional response. These results suggest that ME and FAS appear to be regulated at the level of transcription. The honeybee retina is a nervous tissue with a crystal-like structure and an extreme degree of metabolic compartmentation. Quantitative evidence shows that glial cells in situ phosphorylate glucose to glucose-6P and supply with a metabolic substrate the photoreceptors. Alaniqq is likely to be this substrate ,since more than 50% of ~'*C(U)-glucose is transformed to "'*C(U)-alanine by transamination of pyruvate by glutamate. It is released into the extracellular space and then transferred to the photoreceptors where it enters the Krebs cycle. Light stimulation of the photoreceptors increases their O 2 consumption by 400%. This in turn send a signal to the glial cells inducing a rise in the rate of phosphorylation of glucose and of glycogen turnover. In parallel the concentration of alanine rises and the concentration of glutamate and of protine, both decrease by about 50%. Proline is the prefered precursor of glutamate. We conclude that glial ceils, which contain most of the hexokinase activity, oversatisfy the respiratory requirements of photoactivated photoreceptors for metabolic energy substrate. VIP and noradrenaline (NA) have been previously shown to promote glycogenolysis in primary cultures of mouse cerebral cortical astrocytes. Recently, we have observed a second, temporally-delayed, action of VIP or NA : following glycogenolysis, an induction of glycogen re.synthesis is observed, resulting, within 9 hours, in glycogen levels that are 6 to I0 times higher than those measured before the application of either neurotransmitter. VIP pulses as short as one minute are sufficient to double, 9 hours later, glycogen levels. The induction of glycogen resynthesis triggered by VIP or NA is abolished by inhibition of transcription and translation. The effect of VIP and NA is mimicked by dibutyryl cAMP, suggesting the involvement of a cAMP Responsive Element in the long term effect of both neurotransmitters.

Studies performed on the same cultures have shown that astrocytes produce and release great quantities of lactate : in 30 minutes, as much as 1 I.tmol of lactate per mg of protein is measured in the extracellular medium; the release of lactate is linear at least during 6 hours; iodoacetate, a blocker of the glycolytic pathway, almost completely abolishes lactate release, while sodium azide, a blocker of oxidative phosphorylation, induces a 3-fold increase over basal levels. Furthermore, NA but not VIP enhances lactate output from astrocytes. The electrophysiological properties of Schwann cells (SC) were studied on rabbit vagus nerves after complete Wallerian degeneration (DN) of axons. We found that the resting membrane potential of the remaining SC depended on [K+]0 in a similar manner as did normal nerve (NN Mechanisms regulating K + changes in the axonal microenvironment were studied in the rabbit vagus nerve. This preparation is well suited for this type of investigation because it contains a great number of small nonmyelinated axons with high membrane sur[ace/voinme ratio. By combining the techniques of ion selective microelectrodes with the sucrose-gap method, we measured simultaneously the changes in the r potassium concentration ([K + ]o) and the variations of the membrane potential. We have observed that the kinetics of the changes of [K+]o strongly depended on the presence or absence of the perinettral sheath (PS). In its presence, the changes in [K + ]o, induced by electrical activity or by application o[ ouabain, GABA, caffeine and Ba + +, were much larger than in preparations where the PS was absent. We showed that at resting state, the IK*lo in the extracelinlar apace is subtly controlled by the (Na+-K+)ATPase and Ba + +-sensitive K + channels. During and after electrical activity the changes in membrane potential did not correspond only to the variations of [K+Jo -We hypothesize that this deviation of the membrane potential trom the Nerstian behavior comes [rom a contribution of the Schwann cell K + cttrrents, invoived in the [K + In reguiatioa. . Only a few proteins seem to be selectively expressed by central nervous system (CNS) oligodendrocytes as compared to peripheral nervous system (PNS) schwann cells. Because of the importance of oligodendrocytes in myelination, development and regeneration of neurons, we were mainly interested in oligodendrocyte specific c-DNAs. The isolated COP (Central nervous system _oligodendrocyte 12rotein] c-DNA corresponds to a m-RNA of 4.9 kb. The available c-DNA ( 3.4 kb) including the poly A tail does not show homology to any known oligodendrocyte or CNS c-DNA. Further sequence information is provided by isolating the COP gene from the rat chromosome, In northern blots the expression is restricted to the postnatal rat CNS and is higher at P16-20 than in the adult spinal cord. COP expression was detected by in situ hybridisation in developing oligodendrocytes in the rat CNS and in cultured, galC positive, oligodendrocytes. Oligodendrocytes are a major glial cell type in the central nervous system (CNS). Beside their main function of myelinating axons, they have been shown to express, on their surface, proteins that are important for CNS development and regeneration. We differentially screened a rat postnatal day 16 (P16) spinal cord eDNA library with single-stranded eDNA of P16 spinal cord (plus probe) and with sscDNA of X-irradiated, oligodendroeytc free spinal cord (minus probe). Liver mRNA was used to remove "housekeeping genes" from both sets of probes by subtraetive hybridization. Northern analysis showed that most of the clones obtained are either restricted to or more highly expressed in the nervous system. In situ hybridization shewed that many of them are selective for oligodendrocytes. Besides already known genes (e.g. MOG) several clones represent novel genes. Four of these are exclusively expressed by oligodendroeytes, three are found in addition in sciatic nerve, most propably in Schwann cells. Some of the mRNAs have a perinuclear distribution in the cell bodies, but some are present also in the cell processes (as previously described for MBP mRNA). For some genes we have the full lenghl eDNA, for others we are currently isolating and sequencing them. Functional studies and the production of antibodies are under way. Previous work has shown that the regional expression of a marker protein for fiber growth and plasticity, GAP-43, is mostly inverse to the degree of myelination of CNS gray matter areas in normal adult rats. This suggests a possible role of neurite growth inhibitory proteins from oligodendrocytes in suppressing sprouting and synapse rearrangement in the normal mature CNS. in order to study the relation between myelination and GAP-43 expression more directly, we have examined GAP-43 expression in the lumbar spinal cord of rats where myelination was suppressed by neonatal X-irradiation. In these spinal cords the typical regional downregulation of GAP-43 expression (as seen with GAP-43 immunohistochemistry) fails to occur and GAP-43 expression remains high in most parts of the unmyelinated spinal cord. Non-irradiated, myelinated parts of the spinal cord of the same rats show the normal developmental suppression of GAP-43. These results support the hypothesis that myelin and neurite growth inhibitors suppress sprouting in the normal rat CNS.

Oiigodendrocytss play an important role in development and function of the central nervous system (CNS). They provide the basis for fast saltatory conduction by elaborating the myelin sheath around the axons. Moreover, they produce proteins which inhibit neurite outgrowth; these proteins might be used to establish and maintain the wiring in the CNS. In order to understand gene expression in oligodendrocytes in more detail, we cloned putative transcription factors from this particular glial cell type.

A cDNA Ubrary constructed from differentiated oligodendrcytes was screened for members of the KrQoDel family of zinc finger proteins. Eleven different cDNAs were obtained. Complete sequencing of four clones revealed that they code for new members of the zinc finger gene family. All proteins deduced from the cDNAs contain multiple Zinc finger repeats. The corresponding mRNAs range in size from 2 to 5 kb. As expected, all four genes are expressed in spinal cord and brain, the expression of three clones being higher in the CNS than in all peripheral tissues tested (with the exception of testis). In optic nerves, in which the oiigodendrocytes have been eliminated by X-irradiation, the expression of two of the zinc finger clones was strongly repressed, similar to the typical myelin gene MBP. So far, direct cellular localization by kn si t..uu hybridization could be obtained for one of the clones; the corresponding mRNA is present in oligodendrocytes in all CNS areas, as well as in neuronal subpopulations.

Cell proliferation during remyelination in agqreqatinq brain cell cultures.

v. Comte-Miserez, P. Honegger and J.-M Matthieu. Laboratoire de Neurochimie, P6diatrie, CHUV, 1011 Lausanne, Demyelination was induced in aggregating Drain cell cultures using a monoclonal antibody against myelin/oligodendrocyte glycoprotein (MOG) in the presence of complement. Seven days after having removed anti-MOG and complement and restored normal medium conditions, the cultures fully remyelinated. De and remyelination were assessed by measuring myelin basic protein (MBP). 2',3'-Cyclic nucleotide 3'-phosphodiesterase (CNP) was used as an oligodendrocyte marker and glutamine synthetase (GS) as an astrocyte marker During this period, cell proliferation, determined by 14C-thymidine incorporation was similar in remyetinating and control cultures. Platelet derived growth factor (PDGF AA) given to demyelinated cultures induced celt proliferation and inhibited remyelination PDGF AA removal allowed cells to differentiate into oligodendrocytes and myelinate During remyelination if PDGF AA was added to AraC, proliferating cells were killed, but a partial remyelination could be observed, This indicates that mature and non dividing oligodendrocytes are able to remyelinate. Under those experimentaL conditions, GS levels remained stable indicating that PDGF AA affected only oligodendrocytes prol i feration.

Tribollet E., Arsenijevie Y., Marguerat A., Dreifuss J.J., Bertrand D. and K. Maury, Department of Physiology, CMU, CH-1211 Geneva 4. Specific binding sites for vasopressin (AVP) are present in high numbers in the rat superior cervical ganglion (SCG). Binding of AVP to these sites induces a massive increase in the turnover of phosphatidylinositol, which indicates that they are functional V 1 type receptors. Results of a previous study suggested that the major effect of AVP was not exerted on the principal noradrenergic neurones but rather on other cells (Kiraly et al., PNAS, 83, 5335-5339, 1986) . In order to assess the cellular localization of AVP receptors in the SCG, we have labelled these sites with an iodinated V 1 antagonist in primary cell cultures obtained from SCGs of 5-6 days old rats. Binding was performed on living cells, 12 hours after enzymatic dissociation. Large noradrenergie neurones were not labelled, nor were small dopaminergie neurones. Silver grains were found overlying non-neuronal cells which on the basis of morphological criteria may be satellite ceils rather than Schwann cells. We are presently further characterizing the cell type(s) bearing AVP receptors by using antibodies directed against various glial cell markers in combination with autoradiography. 1 (PN-1) , is a sedrle protease inhibitor, that /n v/b'o peomotes neurite outgrowth. In vivo, it is upregulated following lesion of the rat sciatic nerve and following transient global ischemia in the gerbil hippocampus. Therefore, GDN could influence both degenerative events and/or axonal regeneration /n v~o. Spatial and temporal expression of GDN was examined by /n s~u hybddisation and immunoh~ochemisUy.

In the mouse embryo, ~ major sites of GDN expresskxl are cartilage, bone, lung and heart, whereas it is detected in a few restricted areas of the developing newous system. Postnally, GDN expression gradually increases in the brain and it is widely expressed in the adult ne~ous system. In addition, we generated several strains of transgenic mice uanying a chimeric Thy-I-GDN gene. This transgerm directs expression of high levels of r~ GDN in the brairL Transgenic mice of three lines were extensively analysed at RNA and protein level and went through a whole series of behavior tests. Transgenic GDN was shown to be biologically active by its capability to form SDS-stable complexes with thrornbin. Furthermore, the role of GDN was evaluated in CNS regeneration processes. AS part of the central nervous system, the optic nerve (ON) offers excellent possibilities to study regenerative capacities of white matter. Lower vertebrates are capable of regenerating ON and fully regain visual function. Soon after mechanical crush of Xenopus tadpole ON, glial cells phagocytize cell debris, migrate, and rearrange into a loose core. The acquisition of these novel features is accompanied by an alteration in the intermediate filament (IF) complement.

Vimentin synthesis is selectively increased, whereas in the normal ON cytokeratin expression predominates. In order to study whether this shift in IF expression is realized in all glial cells or brought about by certain cell types only, freshly dissociated cells from ON were immunofluorescently stained for IF proteins and glial markers. The glia-derived nexin (GDN) or protease-nexin I (PN I) is an inhibitor of serine proteases which was shown to be constitutively expressed in the olfactory system and re-expressed in neuroglia following ischemic and excitotoxic insults to the brain. So far GDN/PN-I was visualised in the tissues fixed by formaldehyde-type fixatives.

Recently we observed that its immunocyto/ histochemical detection with a monoclonal antibody was markedly enhanced by the addition of picric acid to the fixative.

Since in situ hybridization has shown widespread expression of GDN/PN-I mRNA in the developing rat and mouse brain, we re-investigated the localization of GDN/PN-I protein in the neonatal and adult rat brain fixed with Bouin.

We report here that -in addition to the expected localization in the olfactory bulb -we detected GDN/PN I immunoreactivity in many distinct neuronal and non-neuronal populations of the central nervous system of the rat. Angiotensin II (Ang II), the final member of the cascade resulting from the activation of angiotensinogen by renin, has the same effect.

Ang II receptors, as well as mRNA for renin and angiotensinogen, have recently been found in various tissues, including the brain. Do%~-regulation of GDN is detected at both the mRNA and protein levels. On the other hand, GDN synthesis can be increased 10fold if the cells are treated with synthetic ligands that specifically interfere with the binding of Ang II to its receptors.

Northern blot analysis clearly showed the presence of the type I or "glial-specific" Ang II receptors, but in extremely low amounts. Ang II receptors also regulate GDN synthesis in Schwannoma and glioblastoma cell lines, but not in fibroblasts or neuroblastoma cells. These results indicate that some neuropeptides can regulate extracellular proteolytic activity. The chemoaffinity theory postulates the existence of cell-specific molecular signals that uniquely identify individual developing neurons. Such molecules are thought to promote both accurate axon outgrowth and the formation of correct synaptic connections. Here we report on a candidate molecular label of this type that is expressed exclusively by two pairs of sibling interneurons in the developing central nervous system of the grasshopper. Our experiments indicate that during axogenesis, this molecule becomes enriched at the growth cones of these ceils; while during subsequent synaptogenesis, it becomes concentrated at the cells' developing terminal arbors. In both cases immunoelectron microscopical studies show that the molecule is secreted by the labeled structures and becomes incorporated into the surrounding extracellular matrix. This molecule, which we call TERM-I, is a glycoprotein with a molecular weight of approximately 48 kD. The highly restricted spatiotemporal expression pattern of TERM-1 implies that individual developing neurons can acquire and retain unique molecular labels which may be important for neuron-specific outgrowth and target recognition. (Supported by the Swiss NSF).

Holm, J., Appel, F. and Schachner, M., Department of Neurobiology. Swiss Federal Institute of Technology, HOnggerberg, CH-8093 Ztirich Cell-to-cell interactions have been developed to a high degree of precision in the nervous system. Various cell surface and extracellular matrix molecules have been identified to be involved in neuron-neuron and neuron-gila recognition. In search for their particular functions, their structure was determined with the aim of relating specific functional traits to distinct structural features. Several neural cell adhesion molecules were shown to belong to theimrounoglobulin supeffarnily and to contain structural mctifs characteristic of extracellular matrix molecules, the fil~onectin type III homologou s repeats. The neural ceil adhesion molecule L 1, and its homologs in other species, such as human L 1CAM and rat NILE, is one of these molecules. To analyze structure-function relationships of the neural ceil adhesion molecule L 1, we have expressed different fragments of the extracellular part of this glycoprotein in CliO ceils 1 The cDNAs were inserted into the pEE.14 enioryotic expression vector which carries the haman cytomegalovirns promoter and a glutamine synthetase gune as selection marker. For correct secretion o fintemal L 1 protein fragments a signal sequence and a stop cedon must be added to the cDNA. Therefore we constructed a set of vectors using the signal sequence of L 1 followed by a Bgl II linker and stop codons in all three reading frames. The results showed a secretion of the internal L 1 protein fragments carrying a signal peptide. Scully, J.L. and Otten, U., Dept. of Physiology, University of Basel, CH-4051 Basel Glucocorticoid hormones (GCs) are secreted in response to stress. The hippocampus is a particular target for GCs, and prolonged exposure exacerbates hippocampal damage due to ageing or neurological insults in both rodents and primates. Conversely, there is evidence that adrenalectomy destroys hippocampal granule cells in vivo. GCs have been found to decrease the expression of nerve growth factor (NGF) in rodent fibroblasts in vitro and sciatic non-neuronal cells in vivo. Since NGF and other members of the neurotrophin (NT) family are strongly implicated in the development and maintenance of the hippocampus, it is possible that GC effects are mediated via NTs. We were interested to see whether GCs could regulate NT expression in immortalized hippocampal cell lines which show characteristic features of hippocampal neurons. Using reverse transcription and amplification of DNA by the polymerase chain reaction (PCR), we have found that the expression of NTs in hippocampal ceils is differentially modulated by GCs, and that the neuronal response may be developmentally regulated.

Masuda-Nakagawa, L., Brodbeck, D., Muller, K.J. and Nicholls, J.G., Dept. of Pharmacology, Biocenter, University of Basel, CH-4056 Basel

The aim of our experiments is to investigate how protein molecules in extracellu]ar matrix of leech nervous system induce regeneration of identified nerve cells. In culture the processes of leech neurons on a laminin-like molecule are slender, straight and unbranched; on the plant lectin Con A they are thick, curved and branched, on enriched tenascin fractions the patlern of outgrowth resembles that on Con A. In the animal itself the laminin-like molecule appears during regeneration at the site of injury close to growing fibres. Metabolic labeling of the entire CNS shows that laminin synthesis is increased in the regenerating CNS. When the connective glial cells is killed by intracellular injection of protease, there is new sprouting and laminin accumulates. It therefore seems unlikely that glial cells produce the laminin that is induced in regeneration. Small phagocytes of the brain, microglial cells, also accumulate at the lesion site. Microglial cells isolated from the CNS are immunoreactive to laminin. We are currently investigating whether antibodies against laminin can inhibit regeneration ot axons. Our results suggest that microglial cells may make laminin and bring it to sites of injury. Axonin-1 of the chick and TAG-1 of the rat are homologous members of the immunoglobulin superfamily expressed either as membrane-associated GPIlinked or as soluble forms, and they are believed to play an important role in the axon guidance mechanisms in developing nerve tissues. Here we report the cloning of the human homologue from a fetal brain cDNA library. At the amino acid level, human TAG-1/axonin-1 has a identity of 91% with rat TAG-1 and 75% with chicken axonin-1. The encoded protein was transiently expressed in monkey COS 1 cells, and a stable mouse myeloma cell line was established expressing human TAG-1/axonin-1. The transfected COS 1 and myeloma cells showed immunoreactivity on the cell surface with polyclonal anti-chicken axonin-1 antibodies. Considerable amounts of the expressed protein were also detected in the cell culture medium, lmmunostaining of cryostat sections of embryonic retinas with polyclonal anti-axonin-1 antibodies showed similar expression patterns in chicken and human samples at corresponding developmental stages. Leech neurons in culture growing on leech extracellular matrix (ECM) respond with neurite retraction to electrical stimulation or to depolarization with high extracellular K +. This response is substrate dependent and requires 'influx of calcium (Gmmbacher-Reinert and Nicholls, 1992, J.Exp.Biol. 167:1-14; Neely, J. Neurosci., in press). Time lapse studies reveal that the first morphological changes of neurons exposed to high extracellular K + include loss of fllopodia and rounding-up of growth cones. Phalloidin, an agent that stabilizes microfilaments, prevents the neurites from retracting after depolarization. By contrast, taxol, a chemical that stabilizes ndcrotubules, has no effect. Disruption of microfilaments with cytochalasin induces changes similar to the ones observed after depolarization: loss of fllopodia, rounding-up of growth cones and neurite retraction. Depolymerization of microtubules has no effect on the neurites. Immunocytochemical studies reveal a loss of microfilaments, but not microtuhules in the growth cones of leech neurons on ECM after depolarization. Neurons growing on Concanavalin A do not retract their neurites and show no changes in the cytoskeleton after depolarization. We are now exploring the mechanism by which depolarization causes disruption of the microfllaments in growth cones on ECM. This work was supported by a grant from the Swiss Nationalfond No. 3127814.89 to J.G.N.

Rader, C., Ziegler, U., Stoeckli, E.T., Osterwalder, T., and Sonderegger, P., Biochemisches Institut der Universit~it Ziirich, CH-8057 Ztirich

The axonai surface glycoprotein axonin-1 is thought to be involved in axon guidance mechanisms in the developing nervous system. The neurite outgrowth-promoting activity of axonin-1 has recently been demonstrated to result from a heterophilic interaction with Ng-CAM. Here we present evidence for homophilic binding among axonin-1 molecules. Heterologous expression by myeloma cells generated functionally competent soluble and membrane-bound axonin-1. Cell lines exposing membrane-bound axonin-1 at their surface formed large multicellular aggregates. Pairwise coincubations of transfected and parental myeloma cells revealed homophilic axonin-1 interactions across the intermembrane space as the molecular mechanism promoting stable cell-cell contacts. Interactions of axonin-1 with both Ng-CAM and other axonin-i molecules might contribute to the formation of macromolecular networks at contact sites of growth cones and axons comprising molecules of both membranes, and thus, represent a mechanism for regulating neurite outgrowth and pathfinding.

Genomic clonong of the axonally secreted cell adhesion molecule axonin-1. R.J. Giger, R.A. Zuellig, and P. Sonderegger. Institute of Biochemistry, University of Zurich, CH-8057 Zurich, Switzerland. Axonin-1 is a 135-kd glycoprotein of the chick. It is secreted from the neurites of cultured neurons and occurs also as a membrane-bound form. Axonin-1 is presumably the species homologue of TAG-1 of the rat. The cDNA of axonin-1 has recently been cloned (Zuellig et al., Eur. J. Biochem. 204, 453-463, 1992) . The axonin-1 gene extends over more than 30 kilobases (kb) and consists of 23 exons givirig rise to two mRNAs of 4000 and 7790 nucleotides (excluding the poly(A) tail). The deduced amino acid sequence contains a signal peptide of 23 amino acids, six immunoglobulin (Ig)-like domains, four fibronectin-type-III (FNllI)-like domains and 35 amino acids at the C-terminus. The first exon is located several kb upstream of the rest of the axonin-1 gene and encodes the first part of the leader sequence. The signal peptide is encoded by exon 2 and exon 3. Each Ig and FNIII domain is encoded by two exons. The last exon codes for the C-terminal 35 amino acids which are probably responsible for GPI-anchorage. Ribonuclease protection studies indicated two major transcription initiation sites at -153 and -178 bp upstream of the ATG start site. The 5' leader sequences are G+C rich (67%) and encoded by exon 1 and 2. The interrupting intron is of undetermined size Alteration in the motoneurone contents of heat shock protein (hsp 70) and ubiquitin were studied in rats which had been subject to loose ligation of one common sciatic nerve. This results in a peripheral neuropathy which peaks at 14 days following ligation and is characterized by transient degeneration of both myelineted and mnnyelineted nerve fibers, abnormal motor behaviors, allodynia and hyperalgesia of the hindpaw. Hsp 70 and ubiquitin are involved in protein metabolism and their expression is regulated during cellular stress. The unlesioned side was used as control. Motoueurone stainings for iasp 70 and ubiquitin were differentially altered at the peak of the neuropathy. Axon damage resulted in a decrease in hsp 70 labeling while ubiqultin staining increased. At the same time motoneurones undergoing axon damage overestained for immediate early gene encoded protein c-JUN and for nerve groth factor (rNGF). In contrast, no clear alteration was seen in the intensity of labeling calcitonin gene related peptide (CGRP). This study demonsrates that peripheral neuropathy resulting from loose ligation of the sciatic nerve not only produces sensory alterations as previously but also leads to pronounced alterations in motoneurone functionnlng that could partly explain the abnormal motor behaviours. Recently, a low molecular weight copper and zinc containing protein has been characterised from human brain tissue which is capable of inhibiting neuronal growth and survival in vitro and which was found to be down-regulated in the brains of Alzheimer's disease victims. This protein was shown to have around 70% sequence homology with mammalian metallothioneins (MT-I, MT-2), but with 68 rather than 60-61 amino acids, leading to its classification as MT-3. In order to investigate the structural differences between this protein and other MTs, a readily available source is required. To this end, an analogous protein has been isolated from bovine neural cortex based on its high metal content and similar elution on anion exchange chromatography. After incubation of the crude brain homogenate with excess cadmium salt, the purified protein contains cadmium and copper in the ratio 5:1. The apo-protein is cysteine rich and devoid of aromatic residues with an analytical molecular weight of 6975 obtained by ion-spray mass spectrometry. This compares with a calculated molecular weight of 6968 for the human protein. Difference absorption and circular dichroism measurements suggest that cadmium coordination is similar to that observed in other cadmiumcontaining MTs. Luminescence properties of the protein are consistent with the presence of a Cu(1) chromophore.

Rotzler, S., Herczeg, A. and Brenner, H.R., Dept. of Physiology, University of Basel, CH-4051 Basel.

During neuromuscular synapse formation, muscle fibres express the 'adult' subtype of acetylcholine receptors (AChR) as a consequence of the neural induction of AChR e-subunit at the synapse. To examine whether the induction of e-subunit and of adult AChR induction was dependent on the stage of motor neuron and/or muscle development, their expression was examined at ectopic endplates induced in extrasynaptic segments of adult rat soleus muscle by surgical redirection of the fibular nerve. In-situ hybridization and electrophysiological identification of AChR subtypes showed local induction of e-mRNA and of adult AChRs at ectopic endplate sites, respectively. Similar observations were made when the foreign nerve was cut at the earliest stages of synapse formation and the muscles were kept active by exogenous stimulation. However, stimulation after denervation caused a redistribution of myonuclei expressing e-mRNA. As the neural signal maintaining e-mRNA expression after denervation appears to be associated with the basal lamina (BL) of the muscle fibre, we have now cultured myotubes on BL from muscle and found accumulations of AChRs where myotubes contacted synaptic BL. Experiments are in progress to determine the capacity of isolated synaptic BL to induce e-mRNA in cultured myotubes.

Varga, Z., Blackshaw, S., Muller, K.J. and Nicholls, J.G.; Dept. of Pharmacology, Biocenter, University of Basel, CH-4056 Basel The central nervous system (CNS) of newborn opossums Ilas been used to study mechanisms of repair. The entire CNS is dissected and maintained in culture. Physiological experiments have shown that through-conduction becomes re-established 3-5 days after a lesion to the spinal cord. The principal aim of our experiments was to analyze by fluorescence and electron microscopy how neurites traverse the lesion and extend beyond it to form synaptie connections with targets. Neurons in dorsal root ganglia were stained with the carbocyanine dye (Dil) and their growth followed by video microscopy in living preparations. After photoconversion of Dil it became possible to observe the DRG neurites in electron micrographs and to search directly for synaptie connections. A further problem concerns the stage at which fibers lose the ability to grow through lesions. It is known that adult Opossum CNS cannot regenerate after a lesion. Experiments in progress sh0w that numerous processes traverse lesions made to the spinal cord of 3-6 day opossums; by contrast less repair was found in 11-14 day old opossums although they still survived well in vitro. It seems of interest that neurites grow less readily across lesions at the very lime that oligedendrocytes start to appear. Supported by grants from the Swiss Nationaffonds (31-30047.90) and from the International Foundation for Paraplegia. Nicotinic acetylcholine receptors (nAChRs) play a key role for synaptic transmission in the gangfia of the vertebrate autonomic nervous system. Still little is known about subunit composition, function and regulation of these neuronal nAChRs. We have used the in situ-hybridization technique to determine the expression pattern of neuronal nAChR subunit genes in cryostat sections of rat autonomic ganglia and adrenal medulla. Radiolabelled cRNA probes were prepared in vitro from cDNA clones encoding a2, ct3, a4-1, ~2 and 134 rat neuronal nAChR subunits. Preliminary results suggest that the main nAChR subtype found in both sympathetic and parasympathetic ganglia as well as in the adrenal medulla may consist of ~3 and 134 subunits. Whereas t~2 mRNA expression was somewhat weaker than or3 and 154, the expression of a4-1 and 132 mRNA was very low. These results reveal an interesting contrast to the brain in which a4 and 1~2 seem to be the most common nAChR subunits. We are currently investigating the expression pattern of several other nAChR subunit mRNAs in the rat autonomic nervous system. Supported by Swiss NF grant 31.31018.91 to ABC.

in the rat brain.

Braissant, 0., Lemberger, T., W. Wahli, Institut de Bioloqie Animale, CH-1015 Lausanne

The peroxisome proliferators activated receptors (PPARs) belong to the nuclear hormone receptor superfamilly.

Recently, our laboratory and others have shown that one possible group of putative physiological inducers of PPARs might be several polyunsaturated fatty acids (PFAs), among which figures also arachidonic acid. The central nervous system is exceptionally rich in PFAs, and arachidonic acid was already proposed to be involved both in inter-and intracellular signalling mechanisms.Therefore, we investigated the expression of PPAR in different regions of the rat brain at several developmental stages by using in situ hybridization and Northern blotting experiments. As probe served a PCR-cloned DNA comprising a part of the ligand-binding or E domain of the rat PPAR-alpha. The multiplicity of mRNA molecules hybridizing with our PPAR probe raises the question whether they correspond to different splicing variants or to various isoforms of the receptor. ct-Melanocyte stimulating hormone (c~-MSH) and related peptides are supposed to influence brain development. In view of a more detailed analysis of such actions, we studied the ontogeny of [125I]Nle4,D-Phe 7u-MSH binding sites in fetuses and offspring of time-pregnant Long Evans rats by quantitative in vitro automdiography. Binding sites are detected in lower brainstem, midbrain and thalamus by gestational day (GD) 15/GD 16. In hypothalamus, paraventrieular nucleus and prcoptie area are labelled by GD 18, while other typical locations (arcuate and ventromedial nucleus) acquire significant numbers of receptors only after birth. A striking feature is the high density of receptors in developing striatnm, which increases rapidly from GD 18 and peaks between late fetal and first postnatal days. Cerebral cortex (virtually devoid of binding sites in adulthood) exhibits periods of transient receptor expression, e.g., in piriform cortex by GD 20 and 22, and in parietal cortex at the end of the first postnatal week. The presence of relatively high regional receptor densities at early stages and their transient occurrence suggest a role in developmental processes.

Barakat-Walter, I., and Duc, C., Institut d'histologie et d'embryologie, Facult(~ de m~decine, rue du Bugnon 9, CH-1005 Lausanne Since the action of thyroid hormones on responsive cells is mediated through nuclear 13 receptors (NT3R) we detected the expression of NT3R in the cell population of dorsal root ganglia (DRG) and sciatic nerve in rat. The expression of NT3R was visualized by immunocytochemistry with the specific 2B3 monoelonal antibody. In the DRG of rat embryo, NT3R immunoreactivity was first discretely revealed in a few neurons at E14, then strongly expressed by all the neurons at E17 and in the newborn rat. In adult rat, the DRG neurons continued to possess a clear NT3R-immunostaining which slightly faded with age. In the developing sciatic nerve, Schwann ceils exhibited a transient NT3 R immunoreactivity restricted 1o a short period ranging from E17 up to 10 postnatal days. Afterwards, Schwann cells became rapidly free of any detectable NT3R immunostaining. However, transection of adult sciatic nerve caused a rapid reexpression of NT3R by Schwann cells during axonal degeneration forty-five days after transection, NT3R expression disappeared again in regenerated nerve. These results suggest that thyroid hormones could play a role in sciatic nerve development and regeneration. (SNF N ~ 31-33671.92).

Regional development of t~-and k-opiold receptors in prenatally benzodiazepine-exposed rats S. Inderbitzin, M. Schlumpf, W. Lichtensteiger, Institute of Pharmacology, University of Ziirich, CH-8006 Ziirich

The possible influence of prenatal administration of diazepam (2,5 mg/kg/d, s.c., gestation day 14-20) on/x-opioid receptor development was studied in striatum of female and male rat brains at two different postnatal ages (PN 14, PN 28), using the selective tx-opioid receptor agonist (3H)-DAGO, whereas the influence of prenatal diazepam on kopioid receptor development was investigated in striatum, midbrain and nucleus accumbens/olfactory tubercle at PN 14, PN 28 and 8 weeks in both, male and females, using the selective k-opioid agonist (3H)-U69,593. Preliminary data from Scatchard analysis indicate a probable decrease in the total number of k-opioid binding sites in the nucleus accumbens/olfactory tubercle region of prenatally diazepam-treated PN 14 male rats.

Markus E. Lauber and Waiter Lichtensteiger, Pharmakologisches institut der UniversiNt Ztirich, Gloriastr. 32, CH-8006 Zfirich A central step in the sexual brain differentiation is the intraneuronal conversion of testosterone to estrogen. This conversion is catalyzed by an enzyme complex comprised of cytochrome P450 aromatase (CYP19) and the NADPH-dependent cytochrome P450 reductase. By means of activity assays and immmunocytochemistry, the former enzyme has been localized in specific brain regions, including the preoptic area, the amygdala, the anterior and posterior hypothalamic areas and others. We have designed oligonucleotides specific for rat cytochrome P450 aromatase and we are in the process of examining the expression of this gene in the developing rat brain by means of in situ hybridization and Northern blot analysis. The two methods have already been validated in the ovary of pregnant rats, where high levels of aromatase-specific mRNA have been detected in the corpus luteum. We will present data on the relative expression and regional distribution of mRNA encoding cytochrome P450 aromatase in the rat brain during normal, prenatal and postnatal development.

Maier A., Schlumpf M., Beer H.F., Schubiger P.A. and Lichtensteiger W., Inst. of Pharmacology, Univ. of Zurich, Zurich and Paul Scherrer Inst., Radiopharmacy Div., Villigen PSI

The ontogeny of expression of Dopamine D~ receptor mRNA in the rat brain and binding to the receptor were assessed at several time points from gestationsl day (GD) 14 to postnatal day (PN) 60 by receptor autoradiography and in situ hybridization. Long Evans rat pups and fetuses of different ages were frozen and sectioned eoronally and sagittally at 10 pro. D t receptor binding, determined with the selective antagonist Iz~I-SCH 23982 was first noted on GD 18 in the developing striatum, basal ganglia and the olfactory tubercle, reaching adult levels at around PiN 14. The expression of DL receptor mRNA was studied by using a mixture of 3 specific 3sS-labelled oligonucleotides. On GD 14 messages were noted in the developing striatum, olfactory tubercle and retina. Tiffs study demonstrates that specific binding to D~ receptors is present in midfetal brain.

D.M. Vogt, M.R. Cello, Institut for Histologie und allgemeine Embryologie, Universit~t Perolles, CH-1700 Fribourg

The CajaI-Retzius neurons have been considered as unusual neurons on account of their morphology and their fate during corticogenesis. They are situated in the marginal layer of the cortex and are the first cortical neurons to be generated. The function of CajaI-Retzius cells and their destiny in adulthood is still controversial. Calretinin (CR), an EF-hand Ca2+-binding protein, is a neuronal marker in adult mammals. During the development of rat brain we observed strong CR-like immunoreac-tMty in the cells of the marginal layer. These CR positive cells appear as early as E12. They seem to be more numerous in younger than in older specimens. They exhibit typical morphological features of the Cajal-Retzius cells like horizontal orientation with usually one nearly rectilinear dendrite, a spindle-shaped cell body, round nucleus and, ultrastructurally, many roundish mitochondria. Furthermore strongly labelled CR positive cells were found in the lower part of the cortical plate. Again the cells were strikingly rich in mitochondria. In conclusion CR is expressed in early corticogenesis and seems to be a marker for Cajal-Retzius cells. It can therefore serve as a tool to resolve the long-standing problem of the fate of this conspicuous cell type during development. We inve~gated the effect of neonatal eye removal on the tan~ntlal extent of the barre$~ in mice. Areas were measured in drawings made from ta.r~enflaUy cut Nisd~ained sections of somatosensoty cortex. We compared areas of 29 barrels-shown in figure -corresponding to 29 mystada] vibrissae, between 60 days old mice enudeated at birth (n=13) Trisomy 13 (Ts13) of the mouse i.e. triplication of one of its chromosomes, can be generated in the progeny of mice. The ontogenesis of calbindin D-28K (CaBP) was studied in dorsal root ganglia (DRG) and spinal cord of normal and Ts13 fetal mice. Cryostat sections of lumbosacral region were prepared from normal and Ts13 fetal mice (E12, E14, E16 and E18) and they were immunostained with polyclonal antiserum against CaBP. In DRG, the CaBPimmunoreactivity was detected at E12 in only about 3% of neurons in both normal and Ts13 mice. In later stages the percentage of CaRP-positive neurons dramatically increased in normal DRG to reach 22% at E18. In contrast, the percentage of CaRP-positive neurons in Ts13 DRG displayed a slight increase and it reached only 10% at E18. On the other hand, CaBP-positive neurons in normal as well as Ts13 spinal cord first appeared in ventral horn at E12, in intermediate gray matter at E14 and in dorsal horn at E16. CaBP-positive fibers were distributed in anterior tuniculus from E14 onward. In spinal cord, the expression of CaBP-immunoreactivity did not show any difference between normal and Ts13 mice. In conclusion, CaBP-immunoreactivity was significantly reduced in lumbosacral DRG of Ts13 as compared with normal mice. No difference was seen in the spinal cord. This result suggests that an imbalance involving chromosome 13 in mouse embryo produces a reduction Of CaBP synthesis during neurogenesis of DRG.

Embryologie der Universit~t, CH 1700 -Freiburg. The growing retinal ganglion cell axon reveals three morphologically different parts: the growth cone, a distal and a proximal segment. Electron microscopic studies of the topographic distribution of these segments within the chiasm of embryos throughout the period of development allows conclusions with respect to the integration of optic axons as a function of their arrival time. In 4 to 19 days old embryos we found, firstly, that growth cones are predominantly located within the ventral sub-plat zone of the chiasm. Secondly, the maturity of axonal profiles shows a gradient along a ventro-dorsal axis with the oldest axons being located dorsally. With increasing age the gradient of maturity becomes steeper than during the early phase of development. We conclude that newly arriving axons are in general added at the ventral periphery of the developing chiasm which eventually results in a chronotopic organization of fibers. Calbindin D-28k (CaBP) and Calretinin (Calr) are two calciumbinding proteins which share a high homology. In the chicken dorsal root ganglia, CaBP is expressed by subpopulations of large A-and small R-neurons while Calr-immunoreactMty is restricted only to a subset of large A-neurons. The peripheral projections of these neurons to the tissues of hindlimbs were studied by detecting the CaRP-immunoreactive nerve fibers with a rabbit antiserum to CaBP (a gift from M. Cello, Fribourg) and Calr-immunoreactive axons with a rabbit antiserum to Calr (SWant, code no. 7686 ). In the skin, the CaBP-and Calr-positive thick myelinated axons were innervating specifically Herbst corpuscules while the CaBP-immunoreactive thin axons innervated Merkel corpuscules and the collar region of hair follicles. In muscles, both CaBP-and Calr-positive thick myelinated nerve fibers were forming selectively the equatorial innervation of muscle spindles. However neither CaBP-nor Calr-positive axons could be observed in tendon organs. Most of the peripheral targets innervated by CaBP-or Calr-immunoreactive neurons are indeed known as rapidly adapting mechanoreceptors. These results suggest that these two calcium-binding proteins could be involved in rapid adaptation.

( The expression and role of nerve growth factor (NGF) in degenerative brain disease are poody understood, Many in vitro and a few in vivo studies showed enhanced NGF synthesis by reactive astrocytes. Thus, reactive astrogliosis induced by neuronal necrosis might be expected to enhance NGF content of affected brain tissues. We measured NGF in rat cerebral neocortex from 2 days to 12 weeks after permanent unilateral middle cerebral artery occlusion (MCAO), NGF was measured by an ELISA using anti-NGF antibody clone 27/21 (Boehdnger), MCAO performed in spontaneously hypertensive rats induced about 65% infarction of the ipsilateral cerebral neocortex. Reactive astrogliosts occurred around but not within the infarction most extensively from day 2 to day 14 after MCAO as shown by GFAP immunohistochemistry. Reactive microgliosis of the border zones was marked 2 days after MCAO as shown by OX-42 immunohistochemistry while the infarction was densely infiltrated with immunoreactive macrophages at day 5. The NGF content of infarcted cortex was unchanged 2 days after MCAO but reduced to :~ 50% from 5 days to 12 weeks due to neuronal necrosis. In adjacent cortex comprising the gliotic border zone the NGF content was unchanged except for a modest increase by 30% at 2 weeks. Thus, in contrast to the working hypothesis, reactive gliosis did not lead to a marked and long-term increase in NGF content of the infarction border zone. The resulting pronounced net loss of NGF in the infarcted hemisphere could contribute to secondary post-infarction events, e.g. degeneration of cholinergic basal forebrain neurons.

Global Ischemia H. Paesold, M.T. Miss, R. Kettler, P. Schoch and H.P. Lorez Pharma D/Vision, Preclin. Research, F. Hoffmann-La Roche, Basel, Switzerland Neuronal degeneration induces reactive gliosis. Based mainly on in vitro studies it is assumed that reactive astrocytes produce nerve growth factor (NGF). We used a rat model of transient forebrain ischemia to study temporal changes in the expression of brain NGF content induced by reactive gliosis. Adult rats underwent permanent bilateral common carotid artery ligation. Twenty-four hours later the animals were exposed to 7% 02/93% N2 for 17 minutes. This induced focal neuronal necrosis most markedly in cerebral neocortex and neostdatum and, to a lesser extent, in hippocampus. Two days to 12 weeks after hypoxia brain NGF was determined by a two-site ELISA using the anti-NGF antibody clone 27/21 (Boehringer). Northern blots were hybridized with a digoxigenin-labeled antisense NGF cRNA probe (Scott et al, Nature 302, 538, 1983 ) 2 weeks after hypoxia. Whereas the NGF content remained almost unchanged in hippocampus, a modest increase (by ~; 50 pg/100 mg w.w.) occurred in cerebral cortex and neostriatum at 2 -4 and 12 weeks, respectively. A very marked gliosis was present in cortex and s/datum as shown by the increase in 3H-PK 11195 binding to the mitochondrial benzodiezepine receptor, in monoamine oxidase B activity and in the number of 0)<-42-and GFAP-immunoreactive glial cells. The increase in NGF content at day 14 was not reflected by enhanced NGF gene transcription in either neostdaturn or cortex. Considering the focal loss of neurons we conclude that reactive gliosis following forebrain ischemia leads to a moderately transient expression of NGF. We have investigated how neurons in the optic tecta of chick embryos depend for survival on their afferents from the retina. Activity-mediated effects were tested by blocking retinal action potentials (with tetrodotoxin, TTX); and other, "trophic", ones by blocking intraocular axoplasmic transport (with colchicine).

TIX (150 ng) injected into one eye at embryonic day 16 (El6) induced pyknosis (cell death) of neurons in the deep layer (SGC) of the contralateral tectum as early as 5 h after the injection; the number of pyknotic cells reached a peak at 13 h, and by 24 h the level of pyknosis had fallen to control levels. Neurons in the superficial layers (SGFS) of the tectum contralateral to the injection were not affected by the TTX.

Intraocular injection of colchicine (225 ng) at El6 caused pyknosis in both the superficial and the deep layers of the contralateral tectum. In the SGFS, pyknosis occurred as early as 9 h post-injection, reaching a plateau by 13 h and remaining essentially constant from then until the end of the period tested. In the SGC, significant pyknosis occurred 4 h later than in the superficial layers; the number of pyknotic cells did not seem to reach a plateau during the period tested.

This suggests that the survival of developing tectal neurons depends on an ongoing supply of substances released from the retino-tectal axon terminals, the release being activitv-denendent in the case of the deep neurons but indeoendent of activity in the-case of the superficial ones. It has been generally recognized that ischemia and subsequent reoxygenation produce reactive oxygen radical species, superoxide radicals in particular, in various tissues including brain. Although most of these studies have suggested that superoxide radicals are generated during reoxygenation period, some studies suggest that formation of radicals have occured during cerebral ischemia. Mongolian Gerbils were anaesthetized with fluothane 4% and clips were placed on the common carotid arteries for 6 rain. Body temperature was maintened at 36-37~ At the appropriate survival times, the animal were reanaesthetized and brains quikly removed. After isolation, hippocampi were frozen in isopentane. They were taken before ischemia, after ischemia and after 10, 30 and 60 min postischemia. SOD activity was measured in hippocampus by the modified method, originally developped by Kostyuk and Potapovich (1989) . Results show that the SOD activity increases during ischemic period to reach a maximum at 30 min. At 60 rain the activity is comparable to preischemic values. These data are consistent with the hypothesis that cerebral ischemia is accompanied by hippocampal cell activation with an associated increase in SOD activity providing transient deleterious effects, which may be implicated in the delayed neuronal death process, It is standard to multiply raw counts of cells (or nuclei etc.) by a "correction factor" to compensate for sectioned cells being counted more than once. However, fashionable stereologists currently condemn the available correction factors, preferring newer methods such as the "Disector". We here derive a correction factor (C) that, unlike previous ones, is theoretically unbiased, even for cells of variable size and shape: m

where n c is the corrected count, nr the raw count, T section thickness, m the sample size (i0 to 20 cells), and ai the length of a randomly selected whole cell or nucleus measured perpendicular to the section (by focusing up and down using an oil-immersion objective).

This approach requires T > ai for all cells.

Lukas, W. and Jones, KA., Glaxo Inst. MoL Biol., CH1228 Geneva Calretinin (CR) is a member of a family of calcium-binding proteins that are localized in discrete populations of neurons in the brain. Its distribution in some areas of the hippocampus that receive dense glutamatergic inputs suggests that CR is aiding calcium homeostasis. To test for the possibility that CR can be neuroprotective, we induced neuronal cell death in 2 week-old cortical cultures from rat by treatment with the calcium ionophore A23187, or with brief applications of glutamate, NMDA or KA, and then stained with and antibody against CR. In control cultures 13% of all neurons were immunopositive while in all treated cultures the percentage of CR + neurons was significantly higher. This effect was dose dependent and most striking for cultures treated with the calcium ionophore where, after treatment for 3 hr at 3 I~M, the percentage rose to 35%. We quantified the neuronal loss independently for both CR + and CR-populations. For example, at a concentration of A23187 that induced a 45% loss of all neurons, there was a 50% loss of CR" neurons and only a 15% loss of CR + neurons. This effect was not observed when calcium was removed from the medium. Selective sparing of CR + neurons continued when A23187 was added with glutamate receptor antagonists, and thus the selective cell killing by A23187 can not be explained by differences in intercellular glutamate receptor densities. These data suggest that CR can significantly attenuate short-term neuredegeneration induced by calcium ovedoad. AIDS-patients often develop impairment of cognitive functions which is ascribed to a failure of nerve cells. Since the HIV-virus does not attack neurons, it is postulated that the normal activity of nerve cells is disturbed by infected glial cells. We have studied the fate of "perineuronal nets" (PN's) in the frontal and parietal cortex of six AIDS-patients, by lectin-and immunohistochemistry. PN's are a specialized extracellular matrix produced by glial cells around cortical parvalbumin-neurons. PN's cement the relationship between nerve and glial cells. In all six AIDS-cases we found a disappearance of PN's around parvalbuminimmunoreactive neurons of the cerebral cortex. We conclude that infection with the HIV-virus decreases the capacity of glial cells to produce the molecules of PN's. The absence of this extracellular matrix around parvalbumin-neurons may engender subtle physiological abnormalities in this population of inhibitory cortical interneurons, which could lead to mental deterioration.

Btieler, H., Aguzzi, A.*, Autenried, P.#, Greiner, R. #, Sailer, A., Aguet, M., and Weissmann, C. Institut ftir Molekularbiologie I, Institut ftir Neuropathologie * und Biologisches Zentrallabor #, Universit~t Ziirich.

Prusiner proposed that the transmissible agent (prion) causing spongiform encephalopathies such as scrapie in sheep or Creutzfeld-Jakob disease in man is identical with prpSc, a modified form of the normal cellulurA0rotein PrP C, and that prion propagation occurs by conversion of PrP ~--or its precursor into PrP Sc, elicited by prpSe. Alternatively, the agent may consist of an as yet unidentified nucleic acid genome coated by host-derived PrP. We generated mice with deleted PrP genes (PrP o/o) by homologous recombination in ES cells. Wild type (wt) and prpo/o mice were infected intracerebrally with mouse-derived prions. While wt mice developed disease at 158 +/-11 days and died shortly thereafter, all PrP ~176 mice are healthy, at > 280 days post infection. Moreover, brain extracts from infected wt mace transmitted the disease to wt mice, while brain extracts of infected PrP o/o, to date, did not. This shows that prpC is essential for scrapie pathology and likely also for prion propagation. Introduction of the hamster PrP gene into PrP ~ mice renders them susceptible to hamster-derived but not to mousederived prions, demonstrating a specific relationship between incoming prion and resident PrP. These results are compatible with Prusiner's proposal, but do not, by themselves, exclude other possibilities.

K.F. St&tktdd, N. Gendre, V. Keller and R.F, Stocker

We have isolated and tested several monoelonal antibodies (mab) in order to identify antennal sl~eific proteins.

Mab ca51/2 (kindly provided by A. Hot'bauer, Institute of Genetics, Wtirzburg, Germany) recognizes sensory neurons in the third antennal segment (fnniculus), which can be correlated with basiconic sensflla (BS) . No staining in the funiculus is observed in the mutant lozenge lacking BS. In Western blots of homogenized antennae, the mab recognizes a protein with a molecular weight of 43,3kdal. In a similar preparation made from lozenge antennae, this protein band is missing.

Therefore it seems that the 43,3kdal protein is associated with BS.

Mab ha21/2 (provided by A. Hofbauer) stains sheath cells correlated with coeloconic sensilla (CS) and BS. The latter show an additional labeling in their shaft. However, in lozenge, mab lacks labeling of BS associated sheath cells. On Western blots of a homogenate of wild type antennae the mab recognizes two protein bands of 46,Tkdai and 42,2kdal. The larger band is missing in preparations from lozenge antennae. The mab is likely to cross-react with two colocaiized antigens, the larger protein being correlated with BS.

Idab I24B5 and VG2 generated in our laboratory stain structures in the third antennal segment. VG2 recognizes structures correlated with BS and CS, whereas the antibody I24B5 stains specifically CS.

The molecular mechanisms causing scrapie and other prion diseases are still unknown. Interactions of the PrP isoforms with cellular proteins of normal or infected animals have previously been analyzed using either tigand blots or binding of PrP to native proteins in unfixed sections. A 45 kDa ligand (Pli 45) was shown to be glial fibrillary acidic protein (GFAP) which is produced by astrocytes. We have purified Plis of higher molecular weight using ion exchange and reverse phase chromatography on an HPLC system. A 60 kDa Pli has been purified to homogeneity. Sequence analysis should allow us to clone a corresponding eDNA. Attempts to clone Plis by expression screening of a eDNA library in E coil will also be reported. Further analysis of cellular proteins interacting with PrP may give us insight into the normal function of PrP as well as the cellular processes underlaying the degenerative changes observed in price diseases. Institut fiir Himforschong, Universit~t Ziirieh, CH-8029 Z(lrich Glutamate receptors form part of a major excitatory neurotransmitter system. We have cloned several potential non-NMDA glutamate receptor subunits (GlaR's) in pigeon brain. The primary structures of three glutamate receptor subunits have been analyzed. Based on amino acid sequence similarities the obtained clones revealed that they encode for AMPA-sensitive glutamate receptor homologues to rodent GIuR-B. GIuR-C, and GIuR-D. Furthermore, within the brain regions analyzed by in-situ hybridization histochemistry (i.e. telencephalon, cerebellar cortex, optic tectum and subtectal nuclei) all three subunits showed distinct and regionally specific expression patterns. Most of the differences were seen in cell types of the cerebellar cortex. Functional analysis of these proteins in transfection assays is currently being performed. Polyclonal antibodies to a C-terminal peptide of GIuR-B/C confirm the presence of an immunoreactive band around 100 kDa in Western blots of pigeon brain extracts, In the cerebellar cortex of pigeon and rat these antibodies produced similar immunocytochemical staining patterns. These results indicate a high degree of evolutionary conservation for GIuR subunits.

Raeber, A.J., Komberg, T.B., and S.B. Prusiner Departments of Neurology and Biochemistry and Biophysics, University of California, San Francisco, CA 94143, USA The scrap ie agent belongs to a novel class of transmissible pathogens termed prions that cause spongiform encephalopathies in humans and animals. An abnormal isoform of the host encoded pdon protein (PrP) was found to he a major and essential-component of the infectious scrapie prion.

In order to explore the fruit fly Drosophila melanogaster as a new model system to study neurodegenerative diseases caused by pdons, transgenic flies were generated by introducing, wild-type and mutant Syrian hamster PrP genes into the Drosophila melanogaster germ line by P-elementmediated transformation. Induction of the transgenes placed under the control of the Drosophila heat shock protein 70 gene promoter resulted in the synthesis of full-length hamster PrP. The recombinant protein is shown to be posttranslationally modified similar to authentic hamster PrP, which includes two asparagine-linked oligosaccharidas and a glycolipid anchor. We were unable to detect a phenotype associated with the expression of the PrP transgenes. An immunocytochemical study has been carried out at the level of light microscopy to determine the localization of AMPA-selective glutamate receptor (GluR) subunits in cerebellar slice cultures. Cultures prepared with standard techniques from newborn rats were fixed after 2-3 weeks 'in vitro' and processed for immunocytochemistry using antibodies to the C-termini of GluRt, GluR2/3 and GluR4 (Wenthold et el., J. Biol. Chem. 1992, 267: 501-507) in combination with an avidin-biotin-peroxidase procedure. Antibodies to GluR1 stained astrocytes in the cultured slice and in the outgrowth zone as well as some Purkinje cells. Cell bodies and dendritic trees of Purkinje cells were strongly labeled by antibodies to GluR2/3, while the staining of granule cells was weaker. Immunoreactive Purkinje cell axons could occasionally be followed to their terminals around unstained neurons in deep cerebellar nuclei. Antibodies to GluR4 labeled certain small multipolar cells in the vicinity of Purkinje cells. Whether these cells represent Golgi epithelial cells or interneurons has to be determined. The differential labeling patterns obtained with these antibodies to GluR subunits show major similarities but also interesting differences to those described in cerebella 'in situ'.

Anita L0thi, Beat H. G~thwiler and Urs Gerber, Brain Research Institute, University of Zfirich, CH-8029 Z~u'ich Glutamate, the most prevalent excitatory neumtransmitter in the mammalian brain, activates both ionotropic and metabotropic receptors. We have investigated the interaction between these receptor types by studying the effects of NMDA receptor activation and amchidonic acid application on metabotropie responses. The singleelectrode voltage-clamp method was used to record metabotropic responses in CA3 pyramidal ceils of rat hippecampal slice cultures. Metabotrepic responses were evoked by bath application of 1-amino-cyclopentyl-trans-lS,3R-dicarboxylate (ACPD), a selective agonist. ACPD application (10 p.M, 30see) resulted in an inward current associated with a decrease in potassium conductance. Following bath application of NMDA (10 BM, 30see), the ACPD-induced inward current was potentiated by 40-2: 20~ In contrast, bath application of arachidonic acid (AA, 40 paM) resulted in a decrease of this current by 40+10% The effects of AA were mimicked by stimulation of the endogenous pliospholipase A 2 (PLA2) through melittin from bee venom (0.5 BM) or by direct PLA 2 application (1 p.g/ml). Mepacrine, a PLA 2 inhibitor, partially inhibited the melittin effects. Thus, AA reduces postsynaptic metabotropie responses, however, does not appear to be involved in the potentiation of the ACPD-mediated responses by NMDA receptor activation. Intraeellular recordings were obtained from bioeytin-labelled neurons in cortical layers II-VI in a submerged slice (300 I.tM thick) preparation. Electrical stimulations of the white matter via bipolar electrodes elicited four types of synaptic potentials in all cells tested (n = 56) : the non-NMDA and the NMDA receptor-mediated early and late EPSPs, blocked by CNQX and AP5 respectively, and the GABAA and GABAs receptor-mediated early and late IPSPs, suppressed by bicuculline (BIC) and 2-hydroxysaclofen, respectively.

Following prolonged bath application (I0 rain.) of BIC 20 B.M, cingulate cortex neurons fired in response to electrical stimuli, paroxysmal discharges of i-2 see. duration, and spontaneously emitted epileptiform bursts at a rate of about 1 per rain. Both the stimulus evoked and the spontaneous events presented an early and a late component, consisting of an initial depolarizing shift followed by a long duration depolarization with superimposed action potentials. The late component was suppressed at hyperpolarized membrane potentials and was abolished by AP5 50 p,M, whereas CNQX 20 IxM suppressed both the early and the late components, as well as the spontaneaous activity.

These results suggest that in the absence of GABAergic inhibition synaptic potentials play a dominant role in epileptogenesis in that NMDA receptors would amplify the paroxysmal depolarization initiated by the activation of non-NMDA receptors. Figurov. A. and Muller, D. Department of Pharmacology, CMU, 1211 Geneva 4.

Several recent reports indieate that mechanisms of synaptie transmission and plasticity in hippoeampus involve activation of the enzyme calcium/calmodulin protein kinase II and enhancement of phosphorylation processes. Another way to enhance the activity of this kinase, as shown by phosphorylation assays, is to treat slices with antagonists of phosphatases such as calyculin A. Here we tested the effects of this compound on synaptic responses in hippocampal slices. At the concentration of 1 uM, ealyeulin A reprodueibly increased the size of synaptie responses elicited by stimulation of Schaffer collaterals. This enhancement averaged 29% in 15 experiments. The effect was observed only with AMPA-dependent synaptie responses, but not NMDA-mediated potentials. Other properties of plasticity such as paired-pulse facilitation or long-term potentiation (LTP) were unaffected by calyculin A. LTP could still be induced in the presence of the drug and the enhancement produced by ealyculin A was comparable on control and potentiated responses. These results suggest that the efficacy of AMPA-medlated synaptic transmission in hippocampus can be controlled by a phosphorylation mechanism which is probably not involved in LTP expression (Work supported by FNRS 31-29338.90 and 31.30980.91).

Pralong, E. and Jones, R.S.G., University Department of Pharmacology, Oxford OX1 3QT, UK Recent hypotheses propose an imbalance between dopamine (DA) and glutamate transmission in schizophrenia. We have investigated DAglutamate interactions using intracellular recording of rat enthorinal cortical (EC) cells in vitro. DA (0.1-1 mM) hyperpolarized most layer ]I cells (22/38) but depolarized most layer V cells (15/21). The hyperpolarization was mediated by D2 receptors whereas depolarizing responses were neither D 1 nor D2 mediated. Synaptic responses of layer II cells were complex and exhibited fast and slow excitatory postsynaptic potentials (EPSPs) mediated by AMPA and NMDA receptors, and fast and slow inhibitory postsynaptic potentials (IPSPs) mediated by GABA A and GABA B receptors, respectively. DA (50-500BM) reduced reversibly (in a dose-dependant manner) the amplitude of the untreated synaptic potentials, e.g. 100 BM DA reduced the fEPSP by 29+/--4% (mean +/__ sem, n=23). Pharmacologically isolated components of the synaptic potentials were also studied. These results indicate that both D1 and D2 receptors modulate the NMDA and the AMPA mediated glutamate transmission although D1 receptors play a more prominent role. In addition, the flPSP did not appear to be directly modulated by DA. Supported by the Royal Society and the Swiss National Foundation. We have used organotypic hippocampal slice cultures as a model system with which to investigate the long-term consequences of epileptic activity. Cultures were prepared with standard techniques from 5 day old rat pups, and allowed to mature for 14-18 days. They were then treated with 500 ~M picrotoxin for 2-3 days in serum-free medium. The expression of glutamate receptor (GhiR) mRNAs was examined after 48 hrs by in situ hybridization using 35S-labelled subunit specific oligonucleotides as probes. The most striking change in mRNA expression was observed for two AMPA-sensitive GluR subunits (GhiR-A & GhiR-B). Compared to control cultures, mRNA levels for these subunits in picrotoxin-treated cultures decreased to less than half the levels observed in untreated sister cultures, as judged by the labelling intensity on autoradiographie films. Levels of GABA receptor subunit mRNAs were apparently unaffected. Nissl staining of these cultures following in situ hybridization showed virtually no nuclear or perikaryal damage to the neurons, indicating that cell death did not account for the decrease in mRNA expression. In contrast, with 72 hrs exposure to convulsants, large numbers of degenerating cells in all hippocampal regions were apparent in Nissl stained material. Such experiments will allow us to examine the mechanisms underlying excitotoxic injury and the control of receptor expression under normal and pathological conditions. The rat pineal gland responds to norcpinephrine (NE) treatment with a 100fold increase in cGMP formation. The function of cGMP in the pineal gland, as well as the mechanism of its formation is not known precisely yet.

Since nitric oxide (NO) is a potent stimulator of the soluble form of guanylate cyclase and sodium nitroprusslate elevates cGMP levels in pinealocytes, we have characterized NOS in the pineal gland with our modification of the method published by Bredt et al. (PNAS, .

Pineal NOS activity was about 50% of the activity measured in the cerebellum. The activity is strongly dependent on calcium and is inhibited by CaM inhibitors. These results are similar to those reported for brain NOS, and suggest the presence of a constitutive, Ca++/CaM -dependent form of NOS in the rat pineal gland. Developmental studies indicate that NOS is present before the cGMP response to NE treatment is first detected and circadian studies indicate that it does not change significantly on a 24hour basis in the adult. In male spontaneously hypertensive rats (SHR/Kyo NIbm; systolic tail blood pressure about 230 mm Hg) permanent middle cerebral artery occlusion (MCAO) consistently induced large cortical end occasionally small neostdatal infarctions. We tested the applicability of this model for the evaluation of neuroprotective drugs. MK-801 (DizocilpineR), a non-competitive NMDA receptor antagonist, was administered pre-and/or post-occlusion. The cortical infarction size was measured 2 days after MCAO in serial, frontal paraffin sections stained with toluidine blue, using a computerized analysis system. MK-80t did not unequivocally reduce the infarction volume in SHR in contrast to male normotensive Fischer 344 rats, in which we observed a reduction in infarction size of about 50% after treatment with MK-801. Similar differences in neuroprotective efficacies between SHR and normotensive strains were frequently ascribed to a different size of the penumbra. In support, Fischer 344 rats showed an extended border zone in the caudal cortex exhibiting scattered neuronal necrosis, which was much smaller in our SHR. However, besides different vasculadzation, other mechanisms can not be excluded. Amino acid levels were measured in peffusates from L-shaped rat cerebellar Slices installed in e Krebs-filled three-compartment system following electrical white matter stimulation. The lateral compartments housed white matter and a cortical section containing parallel fibres respectively, whereas the central compartment housed cortical structures at the point of bending. This arrangement allows electrical stimulation while the peffusion medium passing the central chamber can be collected for amino acid analysis with HPLC. Following a 2-minutes stimulation period of either 2 Hz (n=6) or 5 Hz (n=5) Arg levels were significantly rised above levels already present in the Krebs-solution (6.62 + -1.75 pmol/min respectively 7.31-+2.66 pmol/ min), Arg is the precursor of NO, a neuronal messenger in the brain, which is synthesized by NO synthase. Since Arg and the NO synthase are located in different cell types it can be suggested that Arg passes throu,gh the extracellular space in order to replenish the precursor pool for NO synthesis. Cultured neurons from the CNS provide a useful tool for the study of neurotoxic agents and for substances which can prevent neuronal cell death. We have used 14-day old dissociated cultures from the ventral part of the embryonic rat spinal cord (El4). Glutamate (500 ttm, 15-30 rain) was added to the cultures and 24 hrs later, a morphological degeneration of the neurons was observed. Two biochemical assays were used to quantitate the neurotoxicity : lactate dehydrogenase (LDH) was measured in the supernatant and choline acetyltransferase (CHAT) in the cells. The latter proved to be a more sensitive and rapid measure for neurodegeneration. MK801 (20 ttM) and ketamine (100 pM) acted as neuroprotectors as demonstrated by a decrease in LDH and an increase in ChAT activity. Results using methylprednisolone, lazaroids and neurotrophic factors as neuroprotectors will be presented. In chick spinal cord an active synthesis of prostaglandin D 2 (PGD2) was first demonstrated by bioconversion of [14C] arachidonate and further by PGD2 enzyme immunoassay. In an attempt to determine the cell types which synthesize PGD2 in spinal cord, antisera raised to two PGD synthase isozymes (the rat brain or spleen enzymes) were used for immunocytochemical staining. Under specified conditions of processing, the antibody directed to the brain enzyme gave raise to a strong specific immunostaining restricted to : 1) neurons of lamina II and III;

2) small elements apparently arranged around motoneuron perikaryons and dendrites. These elements in the electron microscope correspond to small Nissl bodies located just below the plasmalemma. In contrast, no immunoreaction was observed when spleen PGD synthase antibody was used.

These experiments point to neurons as source of PGD2 via a rat brain-like PGD synthase. SNF N ~ 3.397. Prostaglandin (PG) E2 is one of the major prostanoids formed in chick spinal cord homogenates. In the present study variations of PGE2 synthesis are investigated. Results show that the basal level of PGE2 is low. In contrast: 1) addition of arachidonate (AA) greatly enhanced PGE2 formation in a dose and time dependent manner. Others unsaturated fatty acids do not markedly modify basal accumulation of PGE2, but interestingly lower the AA response. 2) addition of Crotalus phospholipase A2 (to induce endogenous AA release) promotes a smaller but noticable PGE 2 formation, as to compare to exogenous AA. This limited synthesis of PGE2 may be due to concomittant release of other unsaturated fatty acids. From these experiments it is concluded that PGE2 synthesis is modulated by unsaturated fatty acids. Since these fatty acids are released in large amounts during injury, they could contribute to limit PG formation. SNF N ~ 3.397-0.86. The effect of vasopressin on hypoglossal neurones was studied in brainstem slices of newborn rats, using CsCl-filled micropipettes and the single-electrode voltage clamp technique. Vasopressin, applied at 0.1 to 1 #M, generated a non-inactivating inward current in almost all neurones tested. Antidromic invasion following electrical stimulation of the hypoglossal nerve and morphological characterization of biocytininjected neurones indicate that part of the vasopressin sensitive cells were motoneurones. Hypoglossal neurones formed two classes. A) Cells in which the effect of vasopressin persisted in the presence of TTX and/or in a low calcium/high magnesium perfusion solution. Currentvoltage relations suggest that the vasopressin current was voltagedependent and reversed at potentials ranging from -20 to 0 inV. B) Neurones in which the action of vasopressin, although postsynaptic, could be suppressed or attenuated by TTX and/or by increasing the concentration of extracellular Ca or Mg. Our results suggest that in class A neurones vasopressin affects membrane channels located at or near the cell soma, whereas in class B neurones the primary effect would be mainly dendritic and transmitted toward the soma by a mechanism which is TTX-sensitive and modulated by divalent cations. The distribution of vasopressin binding sites in the brain of the jerboa was studied using a newly developped fl-radio imager. Brain sections on microscope slides were incubated for 1 hour with [3H]-vasopressin. The slides were then introduced in the radio Unager. Each 73-parttcle emitted from brain sections generated a light spot of I mm in diameter, which was red by a CCD camera. The number of events and the coordinate of the center of gravity were recorded. After 1-20 hours, an image representing the distribution of [3H]-vasopressin bound in brain sections was generated. Specific [3H]-vasopressin binding was detected in brain regions such as cortex, insula Calleja, pailidum, the arnygdala and the hippocampus. The binding was quantified directly in cpm from the images obtained. The linearity of the method allowed a relevant measurement of non-specific binding and thus Its substraction from images representing the total binding. This method provides a high resolution map of histological sections within a few hours, allows direct quantification of emitted J3-partleles and can detect all isotopes emitting .B-particles. Vasopressin increases the excitability of facial motoneurones in newborn rats by generating an inward current which is sodlum-dependent, TTXinsensitive and voltage-dependent (Raggenbass et el., J. Neurosci. 11, 1609 -1616 1991) . When facial neurones were voltage-clamped at their resting membrane potential, reducing the extracellular calcium concentration from 2 to 0.01 raM, while keeping the magnesium ~oncentration at 1 raM, caused a 1.3-to 2.2-fold increase in the amplitude of the vasopressin current, the average relative increase being 1.7 _+ 0.1 (mean _+ SEM, n=13). This effect was fully reversible. It was specific, since lowering extracellular calcium did not affect the inward current elicited by activation of AMPA receptors also present on facial motoneurones (n=3). Lowering extracellular magnesium also increased the vasoprassin response, although only by a factor of 1.3 + 0.1 (n=6). These data suggest that in the physiological solution (2 mM Ca, 1 mM Mg) the response of facial motoneurones to vasopressin is partially blocked. Increasing the extraceUular calcium from 2 to 5 mM did not further reduce the amplitude of this response (n=7). I/V curves suggest that the potentiation of vasopressin current in low calcium may be due to the disappearance of a region of negative slope conductance. A reduced arginine vasopressin (AVP) innervation is seen with aging in some parts of the rat brain and this can be restored by administering testosterone to senescent rats (Goudsmit et al., Brain Res. 473, 306-313, 1988) . Previous studies in which we mapped AVP and oxytocin (OT) binding sites in the rat brain revealed marked changes during development and in relation to gonadal functions (TriboUet et al., J. Neurosci. 9, 1764 -1773 , 1989 ). In the present work, we labelled AVP and OT receptors throughout the brain of young (3 months old, n=4) and old (20 months old, n=4) male rats using two 125I-labelled ligands, one specific for V l AVP receptors, the other for OT receptors. Analysis of autoradiograms by an image analyzer revealed a reduction of OT binding in various areas, but not of AVP binding. OT receptors were markedly decreased in the hypothalamic ventromedial nucleus, the central amygdaloid nucleus, the islands of Calleja, i.e. in regions where binding of OT is reduced by castration in rats of either sex (Tribollet et al., Brain Res. 511, 129-1,10, 1990) . Binding was also reduced in the ventral hippocampus. Experiments are in progress to assess whether these changes can be reversed by administration of gonadal steroids. Somatostatin is an important regulator of endocrine and brain functions. In the central nervous system, somatostatin exerts its functions as a neurotransmittes and as a neuromodulator by binding to specific receptors on the plasma membrane. Based on their pharmacological characteristics several somatostatin receptor subtypes had been postulated. During the past year at least four different SRIF receptors have been cloned by us and by other groups. The deduced amino acid sequence of these receptors display sequence and structural homology to the family of G-protein coupled receptors. The distribution of the receptor mRNAs in mouse and rat tissues was studied by northern blot analysis and in situ hybridization. At least three of the genes are expressed in cortex and hippocampus, although the expression levels vary. In substantia nigra moderate expression of SSTR2 RNA and low expression of SSTR1 RNA was found. SSTRI was also expressed in spinal cord. SSTR2 is the most widely expressed SRIF receptor with relatively high expression levels in cortex, hippocampus, atriatum, spinal cord, substantia nigra, lung and spleen. The highest levels of SSTR3 RNA were found in the cerebellum and the olfactory bulb. SRIF receptors were stably expressed in HEK 293 human embryonic kidney cells. These cell lines are now used as tools to analyse receptor ligands and to characterize the receptor second messenger coupling. cord. Amplitude histograms of mepscs were unimodal and skewed towards larger events. The mean of the modes obtained from i0 experiments was -18 pA with a maximal amplitude range of -4 pA to -160 pA for individual mepscs. Current transients to a short voltage pulse were used to estimate the passive cable parameters of the motoneurons.

The membrane time constant and electrotonic length were 20 • 2 ms and 0.96 • 0.13 (n=lS), respectively. The mepscs were subsequently corrected for imperfect space and voltage clamp. The resulting amplitude histograms could be fitted by the sum of two gaussian curves using a maximum likelihood estimator, revealing a mean quantal size of -48 pA with a coefficient of variation of 0.28. The conductance underlying the elementary event was 750 • 139 pS with probably not more than 54 channnels involved. Our data suggest that quantal size and its variance are masked by the cable properties of neurons, that quantal variance is close to its counterpart at the neuromuscular junction and that simultaneous release of elementary quanta occurs occasionally. ( Activation of protein kinase C (PKC) results in down-modulation of the GABAn receptor. In order to identify phosphorylation sites of PKC, the recombinant subunit combination alB2"r was expressed in Xenopus oocytes. The resulting receptor channel could be modulated either by oleoylacetylglycerol (2 [.tM), or stereospecifically by 413-phorbol 12myristate 13-acetate (13-PMA; 10 nM). 14 Set or Thr residues of consensus phosphorylation sites for PKC in the large intracellular domains of td, 132 and -r were altered individually by site-directed mutagenesis. Mutant subunits were co-expressed with wild type subunits to give an a113272S combination. The mutations S410A in 132 and $327A in 72S led to a 40-50 % reduction of the short-term effect of 13-PMA on the GABA current amplitude. Our results identify two single serine residues as phosphorylation sites of a PKC endogenous to Xenopus oocytes. Upon co-expression, the effect of the two mutant subunits was not additive. This indicates, that phosphorylation of both sites is required for a full, PKC mediated down-regulation of GABA currents.

Dityatev, A.E. and Clamann, H.P. Institute of Physiology, University of Bern Quantal analysis of fluctuations of PSPs in neuromuscular junctions has shown that the coefficient of variation (CVo) of quamal size (Q) obtained by analysis of evoked and by spontaneous miniature potentials is about 30%. Estimates of CVq from neuron-neuron connection are 0 or 5%. Quantal size of miniature currents show a variation of 30% or more. To resolve this paradox we carried out an analysis of the accuracy of the deconvolution method used for estimation of CVq. Histograms of 500 hippocampal EPSPs were simulated with different values of ratio of Q to the standard deviation of noise o (Q/c = 1, 1.5, 2, 2.5, 3) and different values of CVQ (0, 15, 30%). A bootstrap method was used for setting the weights of components used for simulation. Q and CVq were estimated as parameters of a unimodal quautal model because it shows a low bias of estimates. Significant underestimates of CVq (using maximum likelihood method) were obtained for values of Q/a < 2.5. Correct estimates of CVq were obtained for CVq=15% when Q/~ was > 2.5 and in part for CVQ=30% when Q/~ was equal to 3. CVQ could sometimes be estimated correctly by using dependence of the mean estimate of Q on CVq because the estimated Q was maximal for the correct value of CV o, values of CVQ in different preparations can be explained by underestimation of this parameter by the maximum likelihood estimator. Supported by Grant No. 31-27973.89 from the Swiss National Fund.

K. Maury and D. Bertrand Dpt of Physiology, CMU, 1211 Geneva 4 Ionic currents where recorded from neurons isolated from the inferior colliculus of newborn or young rats. Newborn neurons display large voltage activated outward currents that, by analogy, can be classified as IA and IK currents. Recordings from more than 100 ceils never showed significant inward currents that could trigger a regenerative electrical activity. In contrast, neurons dissociated from young rats (30 days) display large and fast inward currents of about 2 nA. Thus, we concluded that newborn neurons are immature in their electrical properties and have not yet expressed channels responsible for regenerative activity. Attempts to identify the activity of ligand-gated receptors in newborn neurons have revealed that, although immature, these cells are already expressing ligand-gated channels. This indicates that the development profile of voltage dependent and ligand-gated channels undergo different time courses. In view of our findings it appears that the inferior colliculus, a critical structure for processing of auditory information, pursue its maturation during the weeks following birth.

Falk-Vairant, J., Cavalll, A. and Dunant, Y.. D6partement de pharmacologie, Centre M6dteaJ Universitaire, CH-1211 Gen~ce 4

The quantal, or pulsatile, nature of transmitter release is an essential feature of rapidly transmitting synapses. We recently succeeded in reconstituting the process of ACh release in Xenopus oocytes injected with mRNA from cholinergic neurons. Antisens polynucleotides complementary to eDNA of the proteolipid mediatophore 15 kDa subunit were shown to parallely inhibit expression of the mediatophore and of transmitter release in mRNA primed oocvtes, whereas other neuronal functions were not affected (with C. Leroy, E. Meunier, N. Morel. M. Israel, CNRS, Gif-sur-Yvette, France).

We are presently investigating whether the release of other neurotransmitter can be express in reconstituted systems as well. Also, a method is being developped to test whether transmitter release in such systems is pulsatile like in the natural synapses. A brief stimulation was applied to the nerve-eleetxoplaque synapse of Torpedo eleetrte organ. During and after this period of activity, synaptte transmission and the metabolism of eulcium were monitored : 4SCalctum accumulated in stimulated tissue as a function of the number of stimuli. Hlstochemical localization showed that the number of calcium deposits in synaptic vesicles transiently increased at the end of activity and declined later. Rapid freezing of the tissue followed by freeze fracture revealed membrane openings (pitsl in the presynaptic membrane. While the number of pits did not change during the tetainc stimulation, it increased soon after, reaching a maximum value at 1 rain after the tetanus. It Is concluded that synaptic vesicles sequester calcium ions in synaptie terminals during activity and expel them afterwards by exocytasls. An other approach consists in by-passing the normal calcium channels by using a Ca-ionophore. The miniature potential obtained in this way have a different mnplitude and time course than those obtained under normal conditions. From this, information can he obtained upon the relationship between calcium entry and pulsatilc tranamitter release.

Dept of Pharmacology C.M.U. 1211 Geneva 4. Ehrlich (1886) showed that methylene blue possesses an exceptionnaly high affinity for both perikarya and fine varicose nerve fibres. When methylene blue (0.0025%, final concentration) is added in the medium of organotypie cultures kept on porous and transparent Millicell membrane (see Stoppini et el. J. Neurnsci. Methods 37 (1991) 173-182), fibers and then soma begin to stain after 15 rain of incubation in a 5% CO2 incubator. At that point only a few cells are stained, giving a Golgi-like picture of the tissue. A longer incubation time results in a progressively greater number of cells that take up methylene blue. This rapid and simple method can be used as an interesting test for the viability of the tissue (dead cells do not stain), and allows the visualization of the fibre network and the cellular organization in a still living tissue. If desired, the tissue can then be fixed with ammonium heptamolybdate which precipitates intracellular methylene blue and processed for either light or electron microscopy. The preservation of the fixed tissue is excellent, thereby allowing to analyze at different levels the connectivity established between different groups of cells. This approach can be useful for the study of the morphological characteristics of identified neurones. ( Neurotrophins stimulate trk-type protein kinase receptors, however, further intracellular signalling steps in brain neurons are poorly understood. The present study aimed at investigating the role of phospholipase-CTl (PLC) and phosphatidylinositol (PI) hydrolysis in the signal transduction mechanisms for BDNF and NT-3. We found that both factors stimulate PI hydrolysis in primary cell cultures of fetal brain neurons in a dose and time dependent manner. The absence of extracelinlar calcium resulted in a decrease of the PI response. The effects on P1 hydrolysis were preceded by a rapid stimulation of phosphorylation of PLC-~ I mediated by activated irk-type kinase receptors. Stimulation of trk-phosphorylation, PLC-ff 1 phosphorylation and PI hydrolysis oecurred in cultures from all major brain areas. While pure neuronal cultures were responsive no effects were observed in non-neuronal cultures. K-252b which selectively blocks neurotrophin actions by inhibiting trk-type receptor proteins prevented the PI hydrolysis and phosphorylation of PLC-T1 mediated by BDNF and NT-3. Chronic treatment of cultures with BDNF or NT-3 downregnlates these responses. The findings provide evidence that irkmediated selective phosphorylation of PLCw1 followed by PI hydrolysis is involved in the signal transduction of BDNF and NT-3 in the brain and that brain neurons functionally respond to these factors during early development. To study the morphological correlate of synaptic transmission and plasticity in hippocampal organotypie cultures, we have developed a method that allows to identify calcium at the ultrastruetural level, and thus to recognize activated from silent synapses, based on the accumulation of calcium triggered by electrical stimulation. For the visualization of calcium, we developed a new technique which consists in precipitating calcium after fixation in a selective and high electrondense reaction product using the oxidative properties of osmium.

Calcium deposits are clearly recognized as fine particles present in various subeellular compartments such as synaptic vesicles and mitochondrin. Accumulation of calcium was also observed in smooth reticulum-like structures localized in presynaptie terminals, dendrites and dendritic spines. Another interesting feature of this technique is the possibility to observe calcium in synaptic clefts and more specifically in active zones. After brief high frequency stimulation trains, changes in the distribution of calcium took place, specifically in postsynaptic spines. A significant increase in the number of calcium-containing spines was observed (about 15% of spines with calcium deposits versus 5% in control conditions), and the calcium deposits were generally accumulated in retieulum-like structures (Work supported by De Reuter Fondation and FNRS 31.29338.90 and 31.30980.91 ). Brain cell aggregate cultures offer a unique model to study brain development and plasticity. In order to adjust the culture conditions for long-term experiments (e.g., to study demyelination and remyelination in vitro), aggregate cultures were prepared either from 15-day or from 16-day rat fetuses, and grown for 50 days in chemically defined medium. The following parameters were measured at different time points to monitor maturation and maintenance: protein, DNA and myelin basic protein content; neuron-and gliaspecific enzyme activities. It was found that in cultures prepared from 15-day fetuses, the presence of a defined mixture of albumin-bound lipids (Albumax, Gibco-BRL) enhanced and prolonged the maturation of neurons and glial cells, and stabilized the cells in longterm cultures. Cultures prepared from brains of 16-day fetuses expressed the highest levels of all parameters measured, and supplementation with Albumax was more effective for neuronal than for glial parameters. Cell density and developmental stage of the cells at culture initiation appeared to be the most critical parameters for optimal conditions in long-term cultures. Serotonergic neurons located at the base of the mammalian brain innervate practically every region of the brain and the spinal cord. These neurons exhibit spontaneous electrical discharges in a rhythmical way. The basic activity of the target neurons is regulated by the rhythmical activation of multiple 5-HT receptors. The firing frequency of the serotonergic neurons is modulated by a feedback mechanism involving serotouln autoreceptors regulating intracallular cAMP levels, while the sensitivity of the target neurons is controlled by the differantial expression of 5-HT receptors. We have investigated the effects of cAMP on the development and the functional properties of serotonergie neurons in culture, and the influence of serotonergic ligands on the expression of their receptors on the target neurons. To analyze gene expression in primary cultures, a quantitative RT-PCR approach using internal standards was developed. Cultures of rat brain serotonergic neurons were continuously treated with cAMP analogues. Increased cAMP levels had three effects: First, the neuronal morphology was changed towards that typical for matu~ scmtonergic neurons. Second, the expression of tryptophan hydroxylase, the rate-limiting enzyme in serotonin production, was increased in dibutyryl-cAMP treated cultures. Third, the expression of the inhibitory autoreceptor (5-HTIA) was down-regulated. These results suggest the existence of a mechanism by which the neurons react to synaptic input regulating intraceilulas cAMP levels. Neuronal cultures from rat cortex were treated with specific 5-HT receptor agoulsts and antagonists and the consequences for receptor expression were analyzed. Ligand binding in rive has been shown to affect serotonin binding. The in v/tro approach allows the analysis of the underlying molecular mechanisms.

SCHMID E. and HORNUNG J.-P., Institute of Anatomy, University of Lausanne, Buguon 9,1005 Lausanne.

The serotonergic neurons are grouped in a column of raphe nuclei extending from the mesencephalon to the pyramidal decussation, divided in an anterior group projecting rostrally and a posterior one caudally. The dorsal and median raphe nuclei located in the mesencephalon innervate the cerebral cortex, in organotypic co-cultures of brainstem and forebrain slices, serotonin (5-HT)-containlng neurons survive and reinnervate the cerebral cortex. This system allow to test whether serotonergic neurons of the anterior or the posterior group grow in a specific cortical target. To this end, a section of the brainstem of the met-myelencephalic border or of the mesencephalon was cocultured with a section of the fronto-parietal cortex from 2-day-old rats. By 2 days in culture (DIV2), numerous growth cones appeared at the tip of the 5-HT-immunoreactive axons. The 5-HT-containing axons colonized the cerebral cortex between DIV4 and D1VT, branched through all layers and started to form a plexus from DIV10 to DIV15. A second process of maturation took place between DIV20 and DIV30: a dense axonal network developed at the bottom of the slice and a second one in the overlying neuropil with regularly spaced varicusities. Although the cerebral cortex is not a repulsive substrate, this second phase of maturation was never observed in cultures containing rhombencephalic serotonergic neurons. IN THE CAT AUDITORY CORTEX de Ribaupierre, F., Simm, G., de Ribaupierre, Y., Vall61ian J.-F. and Clarke, S. Inst. de Physiologie de l'universit6, CH 1005 Lausanne Functional significance of interhemispheric connections for the auditory system was approached by recording 8 spike trains simultaneously from 7 microelectrodes placed in the primary (AI), anterior (AAF) and posterior (PAF) auditory cortices on both sides of a nitrous oxide and nembutal anesthetized cat. Neuronal interactions were studied by cross-correlation techniques for 256 single unit pairs during spontaneous and acoustically driven activity. For 116 pairs with both members within the same hemisphere, interactions observed were typical for the presence of a shared common input (CI), in 17 % of the pairs. These CI occurred between pair members located within the same (AI-AI) or across different areas (AI-AAF), within the same cortical depth or between superficial and deep layers, and for quite a range of best frequency (BF) separations. Pairs with members in opposite hemispheres presented CI for 9 % of them. These CI were also observed for all combination of areas, layers and BFs separations, with a preference for short BFs separations and different associations of binaural properties. It is quite significant that these interhemispheric neuronal interactions were only about half less frequent than those within one hemisphere. They could be due to the presence of local collaterals in the hemisphere of origin, of the callosal axons. 

The mammalian cerebral cortex is innervated by two parallel serotonergic pathways originating from the dorsal and median raphe nuclei. The median raphe projection predominates in the upper layers of the neocortex, where, in certain areas, it forms baskets which surround the soma and proximal dendrites of GABA-containing neurons. Among the heterogenous population of GABAergic cortical interneurons, those contacted by the serotonergic median raphe baskets belong to the subtype containing the calcium-bintlin~ protein calbindin, in rat, marmoset, and human. In the cat cortex, presence of calcium-binding proteins in the target neurons of the serotonergic baskets was investigated using double-labelling immunocytochemistty. No calciumbinding proteins (calbindin, parvalbumin, calretinin) were detected in those neurons. As a second marker in GABAergic neurons, several neuropeptides (neuropeptide "I, somatostatin, substance P, cholecystokinin (CCK)) were analyzed. Of those, only CCK-immunoreactivity was observed in the target neurons, using DAB-reacted sections, and fluorescent labelled preparations viewed with confocal microscopy (BioRad 6000, Cancer Research Institute, Lausanne). Variation in the neurochemical composition of the cortical neurons innervated by serotonergic baskets reveal species-specific modes of regulation of these circuits. Single unit properties of the Dorsal Nucleus of the Lateral Lemniscus (DLL) were studied in six Nistar albino female rats. Spike trains of 71 units were recorded along 9 electrode penetrations in the DLL. Activity was recorded during spontaneous activity and binaural white noise bursts stimulation. In spontaneous activity, the distribution of firing rate showed three distinct unit populations with peaks at 1.5, 3.5 and 8.5 spikes/s. Peri-$timulus Time Histograms {PSTH) were computed on several hundreds of responses to 200 ms noise bursts, delivered at 20 dE above the threshold. The mean latency to the first peak of the response was 21 ms and the signal-to-noise ratio was 125.8. The most common response pattern to noise bursts was an excitatory onset response followed in some cases by an inhibition. The majority of DLL neurons responded to binaural noise bursts stimulation (80%). Most units presented on excitatory response to the contralateral stimulation with an inhibitory response to the ipsilateral stimulation (EI, 18%) or without it (E0, 56% Simultaneous recordings of single unit spike trains were carried out in the auditory sector of the thalamic reticular nucleus (RE) and medial geniculate body (MGB) of nitrous oxide anesthetized cats. Coherence and partial coherence analyses of 171 pairs of units, one in RE and one in MGB, were used to investigate the degree of linear time invariant association between the ceils as a function of frequency. For 79 pairs of units the comparison of coherences and partial coherences during spontaneous activity and during stimulation by white noise bursts, indicate that the functional connectivity is modified by stimulus delivery. Overall the stimulus tended to decrease association of firing at very low frequencies (below 5 Hz) and increased association at frequencies higher than 80 Hz. Average null levels were estimated in order to group pairs of units corresponding to the same thalamic subdivisions. Removal of linear effects by partial coherence indicates that the medial division of MGB and RE are preferentially associated at discharge rates at 15 spikes/sec. The techniques presented here provide a valuable improvement in detecting modulations of neural information processing induced by a stimulus without being necessarily time-locked to its time course.

Kretz, R., Nay, P., and Rager, G., Anatomischcs Institut dcr Universit~it, CH-1700 Fribourg In ten tree shrews (Tupaia belangeri) wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) was injected into electrophysiologically defined areas of the dorsal lateral geniculate nucleus. We investigated the geniculostriate pathway by using the highly selective TMB-AHM method (Olucha et al. 1985, J. Neurosci. Moth. 13,131) in order to detect histochemically the anterogradely transported tracer substance WGA-HRP. The results obtained so far not only confirm the general laminar organization of the geniculostriate projection in the tree shrew, but they also reveal a clear bandlike pattern (patches) in layers I and IIIb. This periodicity -distances between maxima of 250 to 450 ~mcan be seen in coronal and in tangential sections and can easily be distinguished from the more or less regular staining pattern of layer IV and the most basal part of sublaycr IIIc. I have now studied intrahemispheric corticocortical connections of these visual areas. Connections were traced anterogradely with the Nauta-method for degenerating axons from a small lesion in upper area 19. They were distributed discontinuously; dense afferents were found around the lesion and in visual areas V3, V3A, as well as in upper V1 and V2 and lateral V4. Weak afferents were found in VP, V5, lower V2, central V1 and medial V4. Lower V1 was devoid of afferents. The tangential distribution of intrisic connections, studied in the cortex medial to the lesion, varied in density in a columnar fashion. Thus, there is evidence for modular organization of human extrastriate visual areas, namely i) selective connectivity between some human visual areas or parts of them; and ii) columnar organization of intrinsic connections.

Schultz W, Ljungberg T and Apieella P, Iustitut de Physiologie, Univ. de Fribourg, CH-1700 Fribearg Lesion studies showed that midbrain dopamine neurons are involved in movements, motivation and cognition. In order to assess their role in learning and cognition, we investigated their eleclrophysiological activity while monkeys learned to perform a spatial delayed response task via two intermediate tasks. The behavioral apparatus contained two response levers, and the animal had to teach, after a delay of 2-3 s, the correct lever previously indicated by an inslroction cue. While learning each task, 25% of 76 dopamine neurons responded to liquid reward, whereas only 9% of 163 neurons responded to this stimulus when task performance was established. Reward responses were significantly more numerous and pronounced during but not after learning in area A10, as compared to areas A8 and A9. Dopamine neurons responded also to the two conditioned stimuli that were important for task performance, the instruction cue as the first stimulus in each trial indicating the target of the upcoming arm movement (58% of 76 neurons during and 44% of 163 neurons after learning), and the trigger stimulus as conditioned incentive stimulus eliciting an ocular reaction and an arm reaching movement by predicting reward (38% during and 40% after learning). None of the dopamine neurons showed dalay-related sustained activity typical for striatum and frontal cortex. These data demonstrate that dopamine neurons respond to the most significant attentional and motivating stimuli during learning and performance of a cognitive task. Their activity does not reflect the specific cognitive components of the tasks, such as preparation of movement, expectation of external stimuli, expectation of reward, or working memory. Rather, dopamiue neurons appear to participate in attentional and motivational processes necessary for learning of behavioral ~ks. To confirm that close appositions of corticospinal (CS) axon terminals with spinal motoneuronal dendrites or somata, as seen in light microscopy, correspond to direct corticomotoneuronal (CM) connections in the rat, possible CM synapses were investigated with electron microscopy (EM). Injections of wheat germ agglutinin-conjugated horseradish peroxidase ONGA-HRP) into motor cortex, and cholera toxin subunit B (CB) into forelimb muscles, labeled CS axons and spinal motoneurons, respectively. Cervical cord sections were processed for visualizing them and flatembedded in Epon. Semi-thin and ultta-thin sections were then prepared. WGA-HRP labeled CS axon terminals were distributed in Rexed's laminae I-X and, to a lesser extent, in the lateral funiculus of the white matter. CB labeled motoneuronal dendrites extended profusely in the gray matter and in the lateral funiculus of the white matter, widely overlapping the termination zone of CS axons. Synapses containing round vesicles were identified at EM level between the WGA-HRP labeled CS axons and the CB labeled motoneuronal dendrites. In conclusion, direct CM connections, believed to be a prerogative of primates, are also present in the rat. There is also evidence that this connection is of functional importance. 

The t erisoollr P.* IL~d Bchulz t p., Division de Psychopharmacologle, I.U.P.G., CH-1225 GENEVE and * Institut f~r Verhaltenswissenschart, ETH Zentrum, ca-so92 ZURICH Oestregens seem to play a modulatory role in the regulation of emotional bshavlour. In the male, neuroactive oestradlol-17~ (E2) is forlned by aromatizatIsn sf testosterone in various hypothalamic and limbic areas. Aromatase activity (AA) was measured using an in vitro radiometric assay in microdissected samples obtained from RHA and RLA rats, which differ in emotional reactity. Tissue concentration of dopamine IDA) and its metabolite DOPAC were also measured by HpLC/ED in frontoparletal cortical (FPC) samples to assess the activity of the mesocortical dopamlnergic system, which is involved in emotlonal hehavlour. Under baseline conditions, there was a significant difference (l><O.oo5) in AA between RHA and RLA rats at the level of the bed n. of the stria termlnalis (BST), which contains oestrogen-sensitIve neurons projecting to the ventral te~edatal area (VTA). DA and DOPAC in the FPC did not differ. ~n a separate experiment, rats subjected to a mild stress (open field and elevated plus-maze for 10 minl showed a marked decrease in AA in all areas and the BST difference was abolished. However, both DA and DOPAC were markedly increased in the FPC of RLA as compared to RHA rats (p<0.01), indicating differential activation of the mesocorticul dopuminergic system in RLA vs RHA rats. These preliminary results suggent that AA may play a role in the control of emotional behaviour in part through testosterone-derived E2 acting on oestrogen-sensltive neurons in the BST projecting to the VTA, which could regulate the sensitivity of the (1). After 32 days of adaptation to alcohol drinking (alcohol in increasing concentrations was substituted in place of water every other day), alcohol (9% v/v) consumption was measured in a free-choice, two bottle situation during 6 days in single-housed D-and S-rats. Dominant rats consumed significantly (p_<0.05) more alcohol (3.56 g/kg/day) than subordinate rats (2.14 g/kg/day). This finding is in agreement with neurochemical data indicating that dominant behavior is inversely related to the activity of brain 5-HTsystems. On the other hand, low 5-HT activity is related to increased chronic alcohol drinking. Our result suggests that social rank is a nonpharmacological factor which influences alcohol consumption. Based on the use of molecular biological, immunological and immunecytochemical methods we could identify several non-variable antigens from African trypanosomes that are located on the cytoskeleton. A recent study revealed that two of these antigens, proteins I/6 and MARP1, immunoreact with sera from various uninfected hosts (mouse, bovine, human) and that at least one, I/6, is immunologically related to a cytoskeletal component from mouse 3T6 fihroblasts. In the case of VSG (variant surface glycoprotein) analogous results were obtained. In sera from different host organisms native antibodies could be detected which immunoreact both, with VSG specific epitopes and with epitopes from different 3T6 proteins. In addition, evidence could be provided that, in mice, these antibodies are produced by a CD5+Bcell compartment of the immune system. These data indicate that self-reactive specificities in uninfected hosts crossreact with strong trypanosomal antigens and thus create an immunological environment which is pre-primed to trypanosomal infections. We have analysed the effect of antibodies directed against MHC class II antibodies on the proliferation of T-cells. We have shown that anti-class II antibodies exert a direct effect on T-cells causing an acute block in their proliferation. The inhibition does not involve apoptosis. It is also not due to interference with the state of activation of the T-cells, since the trancriptional activator NF-kappaB, which regulates interleukin 2 and interleukin 2 receptor gene expression, remains activated in arrested cells. Interleukin 2, interleukin 2 receptor and c-myc gene expression, which are characteristic for T-cell activation, are also unaffected. In adddition, the rapid arrest of DNA synthesis caused by anti-class II antibodies is entirely reversible. By analysing the cell-cycle phase distribution of inhibited cells, it could be shown that cells in all phases of the cell-cycle are inhibited. The signal transduction pathway that results in inhibition was shown to be independent of protein ldnase C and extracellular Calcium Tyrosine kinase inhibitors, however, partly reduced the level of inhibition and, conversely, phosphatase inhibitors enhanced it. Antibodies directed against MHC class II molecules have been demonstrated in a number of autoimmune conditions and also in patients with AIDS. Anti-MHC class II-mediated inhibition of T-cells may therefore play an important role in a number of autoimmune diseases and contribute to the pathogenesis of HIV-infection.

Diagnosis of trypanosomiasis using phage presentation repertoire libraries P. Weber and T. Seebeck Institut for allg. Mikrobiologie; CH-3012 BERN It has been demonstrated by our group that mammals, infected by African Trypanosomes, raise an early antibody response to a small number of trypanosomal proteins. We have characterized the molecular structure of several of these early antigens since we intend to use them for diagnostic purposes. As a technique for producing these proteins we are currently establishing the expression and presentation of the antigens at the surface of M13 bacteriophage. One of our goals is to use such recombinant phages as polypeptide carriers for immunodiagnosis. The extensive polymorphism displayed by the HLA antigens of the major histocompatibilit;r complex is of crucial importance to understand the regulation of immune responses in normal and pathological situations. Because of their high degree of diversity, HLA genes also represent useful markers for human population genetics. The development of the PCR technology and of the oligonucleotide typing methods now allows large scale analysis of the polymorphic HLA loci. By using a combination of 21 SSO probes, DPB1 allelic polymorphism was analysed in a sample of 135 healthy individuals from a local panel of blood donors. 17 of 22 possible alleles were detected with 4 alleles accounting for >72% of all DP alleles. Significant linkage disequilibrium was found for the following haplotypes: A1-B8-DRB1*0301-DPB1*0101, B44-DRBl*07-DPBI*ll01, DRBI*1302-DPBI*0301, and DRBl*04-DPBl*0201. In our panel, heterozygosity levels are: H=0.87 for HLA-A 0.92 for HLA-B, 0.90 for HLA-DRB1 and 0.82 for HLA-DPB1. These values are strikingly similar despite different frequency distributions.

Urs Chdstan, Paul Jan6*, and Josef Gut; Departments of Pharmacology and *Biochemistry, Biocanter of the University, CH-4056 Basel A monoform anti-TFA antibody, directed against tritlucroacetyl-pratein adducts (TFA-protein adducts) that are elicited in tissues of experimental animals and humans upon exposure to the anesthetic agent halothane, recognizes crossreactive proteins of 64 kDa and 52 kDa in several tissues of rats and the liver of humans not previously exposed to the drug. These cross-reactive proteins were shown to confer molecular mimicry of TFA-protein adducts. Hera, by the use of the anti-TFA antibody as an immunoaffinity matrix, the protein of 64 kDa was purified from rat head. The amino acid sequence of six internal tryptic peptides exhibited 100% identity with the corresponding deduced amino acid sequence of the dihydrolipoamide acetyltransferase component (E2 subunit) of the rat liver pyruvate dehydrogensse (PDH) complex, as encoded by the clone pRMIT [Gershwin, M.E. at al, (1987) J. Immunol, 138, 3525-3531] , Similar to the haptan-dedvative N-e-trifluoroacetyI-L-lysine, free lipoic acid, the prosthetic group of the E2 subunit, inhibited on immunoblots the recognition by anti-TFA antibody of the E2 subunit, of the not yet identified protein of 52 kDa, and of TFA-pratein adducts. Additionally, both lipoic acid and N-e-tritluoroacetyI-Llysine did abolish the precipitation of the native E2 subunit by anti-TFA antibody from solubilized rat heart mitochondda. These data suggest that lipoic acid is actively involved in the molecular mimicry of epitopes present on TFA-protein addusts by the E2 subunit of the PDH complex. The stimulus-secretlon model of CTL-mediated lysis proposes the release of the lytic pore-forming protein perforin into the intercellular space between killer and target. CTL lines, in contrast to tumor target cells, are resistant to cell-mediated lysis and they also withstand lysis by the isolated perforin.

These observations suggest a selfprotection mechanism of CTLs against their own lytic protein.

We radioactively labeled proteins of cytotoxic granules and studied the binding of perforin to different cell lines. Studies performed at 4~ revealed that equal amounts of perforin bound to susceptible target cells as well as to resistant CTLs in a Ca2+-dependent manner. Cell-bound perforin was treated with trypsin. Perforin on the CTL line B6.1 was more accessible to the protease than perforin bound to the tumor target cell K562. This indicates a difference in conformation or accessibility. We propose a mechanism whereby CTLs have an inhibitory molecule, which prevents membrane insertion and subsequent polymerisation of perforin.

Calreticulin (Ro/SS-A auroantigen) is a granule-associated molecule of cytolytic T lymphocytes and is released during target cell lysls Dupuis, M., Sch~rer, E., Krause, K.-H., and Tschcpp, J..

The delivery of lytic components harbored in granules of CTL3 by exocytosis is an effective mechanism for target cell destruction. Purified granules contain perforin I granzymes and proteoglycans as major components. We report the purification of an additional major 60-kD granule-assoclated protein from human LAK cells and mouse cytolytic T cells. The Na2-terminal amino acid sequence of the polypeptide was found to be identical to calreticulin. Calreticulln is a calcium storage protein. In CTLs, calreticulln colocalizes with the lytic perforin to the lysesome-like secretory granules, as confirmed by double-label immunofluorescence confocal microscopy. Moreover, when the release of granuleassociated proteins was triggered by stimulation of the T cell receptor complex, calreticulin was released along with granzymes A and D. Since perforin is activated and becomes lyric in the presence of calcium, we propose that the role of calreticulin is to prevent organelle autolysis due to the proteinTs calcium chelator capacity.

Heiko M~hl and Josef Pfeilschifter, Department of Pharmacology,Biocenter, University of Basel, CH-4056 Basel Treatment of mesangial cells with recombinant human interleukin 1 (IL-l~) dose-dependently increased nitdte formation due to the induction of a mcarophage-type of nitric oxide synthase. Addition of the immunosuppressive drug cyclosporin A or its derivatives, cyclosporin G and cyclosporin H, dose-dependently inhibited IL-1 [3induced nitritie generation. Half-maximal inhibition was observed at concentrations of 0.9 ~M, 2.0 ~M and 3.8 ~M of cyclosporin A, cyclosporin G and cyclosporin H, respectively. Two other immunosuppressive drugs, FK506 and rapamycin (10 p.M each) did not affect IL-1 ~-induced nitrite production, suggesting that inhibition of nitric oxide synthase activity is not related to the immunosuppressive activities of the investigated drugs. In summary, cyclosporins may protect the kidney from damage resulting from cytokine-stimulated mediator release and subsequent inflammatory reactions. Production of low levels of nitric oxide (NO) by human neutrophils has been demonstrated recently. The possible retationship between this pathway and the superoxide (02-)-generating system was evaluated. Incubation of neutrophils with L-arginine (L-arg) (1 mM,lh at 37~ did not increase 02 production except with PMA as a stimulus (18.7+/-6.8 vs 14.3+/6.1nmol/5min.10~cetls,n=9,p<0.01). In contrast, no effect of L-arg was found on NO production measured as nitrite (NO2-) with the Griess reagent. Maximum NO2 production was obtained during adherence of neutrophils to plastic surfaces in comparison to resting cells in suspension (2.49+/-0.23 vs 0.74+/-0.19 nmol/h.10~cells). NO2-could be further stimulated by agonists only in cells in suspension. Upon adhesion, 02-was also produced at a higher level compared to cells in suspension but could be further stimulated with PMA. L-canavanine inhibited NO2-and 02-production at different maximal concentrations of 10 and 5 mM respectively. These data suggest at the same time a relationship and a dissociation between N02 and 02 release. The specificity of CsA binding to the major intracellular receptor protein, cyclophilin A and B as well as the interaction with the phosphatase calcineurin was investigated. Binding of PL-CS, a photo~tnlty probe of CsA, to recombinant human eyelophflin A and B is saturable and specific. Although PL-CS binds calcineurin in the absence of eyelophflln and calmodulin, this binding is non-specific, since it does not discriminate between active and inactive CS derivatives. In the presence of eyclophflin cyclosporin-calcineurin binding becomes specific. , Ternary complexes containing an equimomr rauo oi cyclophflln A or B, PL-CS and calcineurin are resolved using chemical-crosslinldng technique. The formation of these complexes is specific, calcium, but not ealmodulin dependent and is inhibited only by cyelosporines, which bind cyclophflin. The drug-immunophilin complex binds only to the calcineurin A subunlt; The proteolytie 40 kD product of calcineurin A retains the binding properties, suggesting that the N-terminal domains are not necessary for the complex formation. Similarly, a trtmeric complex of FKBP, FK506 and calcineurin can be demonstrated. As expected these complexes are only competed by homologous derivatives. In rive labelling of Jurkat T-cells revealed, that similar immunophflindrug-calcineurin complexes Occur in rive. (15)) and mudne (L929) cell lines. TNF-~ bioactivity was detected in the supematant of the transfected cells using a standard L929 bioassay or a newly developed PK(15) bioassay. This bioassay is 100 to 1000 times more sensitive than the L929 bioassay suggesting a partial species-specificity of the porcine TNF receptor(s). We previously found that prenatal exposure of rats to diazepam resulted in a marked reduction of lymphooyte proliferation and of LPSstimulated TNF-a release from rat splenocytes during postnatal development.

Both of these defects recovered in adulthood. We further examined the mitogen-stimulated release of IL-6, a product of activated macrophages and lymphocytes. IL-6 was significantly lowered at 2, 4, and 8 weeks of age in splenocytes after prenatal diazepam exposure. In contrast to TNFa, this decreased release was still observed in adulthood. In search for mechanisms underlying the disturbed cytokine release, we are currently investigating the liberation of IFN-7 and PGE 2.

Car, B.D.. Ozanen, L., Garotta, G, Steckholzer, U. and Ryffel, B. Institut mr Toxikologte der ETH und der Universitat Zfirich Schorenstr. 16, 8603 Schwerzenbach, Switzerland.

Production of Interferon y (IFN-T) by activated T-lymphocytes and natural killer cells results In a pleotropic array or responses, including antiviral activity, stimulation of macrophages and NK-cella and the Induction of MHC class II molecules In a variety of cell types. That these responses are transduced through a single cellular receptor prompted our investigation Into the topographic distribution of this receptor in mice, Affinity purified rabbit-anti mouse antibodies directed against the murine IFN-7 receptor were utilized In morphologtc and fluorocytometric investigations. Striking patterns of highly positive cells were observed in the liver, spleen and kidney. Hepatic staining was restricted to Kupffer cells which were rnarkediy positive, while coexpresslng the macrophage marker F4/80 and MHC class II molecules. This coexpression by macmphagedendritic cells of the splenic red pulp was confirmed by Immunofluorescence and FACS analyses. Renal positivity was restricted to glomeruli which contained cells expressing much IFN-T receptor antigen In the mesangtum. By FACS analysis activated T-and B-lymphseytes and well as NK-cells were determined to express the IFN-y receptor. The close proxlmity of IFN-T receptor bearing cells to ceils capable of producing IFN-T Implicate these cells in IFN-y-mediated pathologic events in the mouse. Hepatic graft-versus-host inflammatory responses and certain routine glomerulonephridities are proposed as disease processes Involving inappropriate stImulation of the IFN-7-receptor. .the ability to proliferate in an uncontrolled manner. High levels of activated NF-kappaB can be found in their nucleus and the IL2R and IL2 genes are constitutively expressed. Expression of these genes is critically dependent on the presence of the parasite in the host cell cytoplasm. To establish whether IL2R/Tac gene expression is essential for continuous proliferation, anti-sense IL2R/Tac RNA experiments were carded out. Theileria-infected cells were permanently transfected with plasmid constructs designed to express either sense or anti-sense IL2R/Tac RNA under control of a cadmium-inducible metallothionein promotor. The neomycin resistance gene was co-transfected to provide a selection marker. Several clones of ceils transfected with either sense or anti-sense constructs were generated. It was found that clones transfected with anti-sense IL2R/Tac constructs grew more slowly than cells containing sense constructs, also in the absence of cadmium. This was due to the continuous basic activity of the promotor. Reduced amounts of IL2R/Tac mRNA were also found in cells transfected with anti-sense constructs and, in proliferation assays, significant inhibition of ceils containing anti-sense constructs could be induced by the addition of cadmium to the culture medium. These data show that proliferation of Theileria-infected T-cells used in this study is at least in part dependent on the expression of a fully functional IL2R. Melanoma cells can secrete several cytokines and express various cell surface molecules, such as the intereeUular cell adhesion molecule ICAM-1, class II histocompatibility antigens, and the CALLA antigen, typically found in cells of the immune system. We have investigated the possible expression of interleukin-2 (IL-2) receptors in melanomas using monselonal antibodies specific for the p55/a chain (TAC antigen) and the p57~ subanit. Flow cytometric analysis of cultured melanoma cells showed the presence at low level of the TAC antigen and of the [~ chain on the cell surface of several lines. Similar results were obtained in vivo by immunohistochemistry on cryoseetions prepared from cutaneous and ocular melanoma explants. Positive staining of various intensity was readily observed for the a chain oftbe IL-2 receptor in a high percentage of tumor cells. The 13 chain could also be detected, although with a weak staining and in a limited number of specimens. Analysis of RNA from melanoma cell lines by Nonhero bolt showed the presence of typical 4.0 Kb transcrilxs of the p75 subunit, while mRNA for the p55 chain could be detected using combined reverse transcription/PCR analysis. Together these results suggest that melanoma cells may express high affinity receptor for 1I.-2. Interferons were discovered by virtue of their antiviral activity but they exert additional activities including inhibition of cell growth, antitumor action, effects on cellular differentiation, and immunoregulatory functions. The ~x and fi (type I) IFNs are thought to act through a common receptor; the gene encoding one of its chains was cloned, however at least one more chain is required for activity (Uz6 et at.). To define the physiological role of IFN-~J~, we generated mice devoid of the known receptor chain by homologous recombination in ES cells. Homozygous IFN-~0R 010 mice were viable, anatomically normal and fertile. The type I receptor was functionally inactive as shown by the lack of antiviral response of primary IFN-~J3R 0/0 embryonic fibroblasts (PMEFs) to natural murine type I IFN-m'0; the response to IFN-? was unaltered. Northern blot analysis of PMEF-RNA with probes derived from several IFN-ieducible genes showed that mutant cells did not respond towards natural mudne type I IFNs, rlFN-I3 and human recombinant IFN-~x2/al (to which normal mouse cells are responsive), indicating that the disrupted receptor chain is essential for the response to the IFNs tested. Hematological and immunological characterization of the targeted animals is underway.

Interleukin-I dependent stimulation of interleukin-6 production in culti-t vated human dermal fibroblasts is linked to the presence of fetal calf serum or platelet-derived growth factor Interleukin-6 receptor density is higher in cultivated human follicular outer root sheath cells than in interfollicular epidermal keratinocytes and decreases with confluency E.R. Waelti +, A. Limat*, S.P. Inaebnit ~ T. Hunziker*, U.N. Wiesmann ~ Dept. of Pathology + , Dermatology* and Pediatrics ~ , University of Berne, Switzerland.

In skin, effects of interleukin-6 (IL-6) on cell motility, proliferation and differentiation of keratinocytes are discussed. We therefore studied IL-6 receptor densities on follicular outer root sheath cells (ORS) and interfollicular epidermal keratinocytes (IEK) in preconfluent and confluent cell cultures. Binding studies and Scatchard analysis were performed using recombinant 125I-IL-6 added in low to saturating doses. The number of binding sites for IL-6 on ORS cells grown to preconfluency ranged from 1380 to 7350 per cell. ORS cells grown to confluency showed a much lower number of binding sites, i. e. 240 to 430 per cell. In contrast, the number of binding sites on IEK was 680 per cell when grown to preconfluency and 80 to 130 per cell in confluent cultures. The Scatchard plot disclosed one type of high affinity receptor with a Kd-value of 43 pM and 7300 receptors per cell on ORS cells grown to preconfluency. IEK grown to preconfluency exhibited a Kd-value of 51 pM and 580 receptors per cell. Our results support the concept that IL-6 is involved in enhanced cell motility and proliferation of ORS ceils, which occurs e.g. in re-epithelialization of superficial wounds. In some pathologies, proteolytic enzymes released by pulmonary cells may affect the biological functions of hormones. To adress this question, we measured the activities of various peptidases and the level of cytokines (IL6 and TNF-~) by bioassay and immunoassay in bronchoalveolar lavage fluid of patients with various pathologies. We found that neutral aminopeptidase (APN/CDI3) activity was correlated to the number of neutrophils and dipeptidylaminepeptidase IV (DPPIV/CD26) activity to that of lymphocytes. TNF-o~ levels measured by immunoassay and bioassay were well correlated (n=68, r=0.87), but we demonstrated the presence of a substance interfering with the bioassay of IL6.

Interfering capacity was more marked in samples with low IL6 biological and immunological levels. It is hypothetized that, depending of its N-terminal sequence, IL6 might be a substrate for DPPIV/CD26, but not for APN/CDI3. However it is premature to link proteolytic activities and loss of IL6 biological activities.

Interleukin-1 and cell-to-cell contacts stimulate interleukin-6 in fibroblasts in fibroblast/keratinocytes co-cultures W. Liu ~ Y. Cao ~ S,P. Inaebnit ~, E.R. Waelti +, A. Limat*, T. Hunziker* and U.N. Wiesmann a, Dept. of Pediatrics ~ Pathology + and Dermatology*, University of Berne, Switzerland.

Interleukin-6 0L-6) production of cultured dermal fibroblasts can be stimulated by adding interleuldn-1 (IL-1) to the culture medium. Mixed co-cultures of fibroblasts with keratinocytes produce large amounts of IL-6 (Waelti et al., J. Invest. Dermatol. 1992: 18, 805-808) . The production of IL-6 in fibroblasts is stimulated to a certain extent by keratinocyte conditioned medium. In a two-chamber co-culture system keratinocytes secreted 1L-1 and flbroblasts subsequently produced considerably increased amounts of IL-6, but less than in co-cultures of the two cell type on either side of a porous membrane. Cell contacts through the pores greatly stimulated IL-6 production. The extent of IL-6 production was dependent on the pore size (3 Ixm pores > 0.45 Ixm pores). The I].,-6 synthesis could partially be suppressed by 10 pal hydrocortisone. Our results support the evidence that in fibroblast/keratinocyte cultures the production of IL-6 is stimulated by IL-1 but also by factors depending on cell-to-cell contacts. Interleukin 113 (IL-I~) and tumor necrosis factor c~ (TNFc~) stimulate PLA 2 and PGE 2 synthesis and secretion by mesangial cells. Coincubation with a neutralizing monoclonal antibody to PLA 2 attenuates IL-I~ and TNFc~-induced PGE 2 production by 45% and 52%, respectively. Addition of purified PLA 2 to unstimulated mesangial cells causes a marked release of arachidonic acid and an increased synthesis of PGE 2. Moreover, addition of purified PLA 2 to glomerular epithelial and glomerular endothelial cells augmented both, arachidonic acid release and PGE 2 synthesis, with the endothelial cells being especially sensitive. These data suggest that expression and secretion of PLA 2 triggered by proinftammatory cytokines crucially participates in glomerular inflammatory processes. Clinical evidence suggests that acute withdrawal of chronic glucocorticoid therapy with suppression of endogenous cortisol production in RTX patients is associated with an increased susceptibility to rejection. Furthermore, experimental and clinical observations revealed an enhanced immunological and inflammatory response in glucocorticoid deficiency. A key enzyme of the inflammatory response is the PLA2, which releases amchidonic acid. In order to elucidate whether glucocorticoid deficiency causes PLA2 up-regulation, the expression of PLA2 mRNA and protein, as well as the enzyme activity, were quantified in rats with and without ADX. The mRNA of PLA2 was quantified by poiymerase chain reaction using a constant amount of a modified PLA2 cDNA transcript as an internal standard. PLA2 message and activity varied from organ to organ with the lowest values in kidneys. The PLA2 mRNA was increased by 116 +_ 24 % in various tissues of ADX rats. This up-regulation of PLA2 mRNA was not due to a nonspecific effect of ADX since the mRNA levels of other proteins (cfos, c-myc, c-erb8, methallothionein-II) remained unchanged. The increase in PLA2. mRNA in ADX rats was reflected by a corresponding increase in tissue (kidney, lung, spleen, liver) PLA2 enzyme activity and protein content as assessed by quantitative Western blot analysis. Thus, ADX increases PLA2 message, protein and activity, an observation in accordance with the increased susceptibility to inflammatory reactions in glucocorticoid deficiency. The gag gene of Rous sarcoma virus (RSV) encodes a protein, pl0 of unknown function, located between MA and CA in Pr76 ga~. This protein has a high content of prcline and glycine residues and is very hydrophobie. We have inserted Mlu-I linkers (Thr-Arg) in its sequence at five different positions: 934, 973, 1003, 1048 and 1075. Upon transfection in CEFs cells at 36~ these mutants produced as many virus particles as the wild type. The mutant at 973 showed a very low infectivity, while the others were fully infectious. However, at 41~ the mutations at positions 973 and 1075 inhibited virus particle release by a factor of 20 and abolished infectivity. Interestingly, both linker insertions were near proline rich regions of the protein. Intracellular viral Drotein analysis showed that the mutants synthesized Pr76 g'ag at a rate similar to that of the wild type. Together, the data support a role for pl0 in budding processes as well as in an early step on infection. We are now constructing cell lines that would constitutively shed these defective thermosensitive retrovirus particles. Further characterization of these pl0 thermosensitive mutants is under way, and results will be presented. Human MxA and murine Mxl are closely related interferon-induced proteins with intrinsic antiviral activities. Both proteins inhibit influenza virus multiplication, although at different steps of the replication cycle. Mx proteins are members of a new family of GTPbinding proteins and possess an intrinsic GTPase activity. GTP-binding or GTPase activity of Mx proteins appears to be necessary fur their antiviral function. To further elucidate the molecular mechanism of influenza virus inhibition by Mx proteins, we attempt to test the function of purified Mx proteins in in vitro transcription and translation systems of influenza virus. First, we expressed recombinant Mxl and MxA proteins, with a His-tag at their N-terminus, in E. coil or in Swiss 3T3 cells and purified them under nondenaturing conditions by Nichelate affinity chromatography. Mx proteins carrying mutations in the GTP-binding site were used as controls. Mxl and MxA but not the mutant proteins hydrolysed GTP to GDP. The Km values for the GTPase activity of recombinant Mxl and MxA were in the range of 0.1 to 1.0. The hydrolysing activity of Mx proteins was specific for GTP. We currently test whether the recombinant Mx proteins are able to inhibit distinct influenza virus replication steps using in vitro-systems. The interferon-induced murine Mxl protein possesses an intrinsic antiviral activity with selectivity for influenza viruses. Mxl accumulates in the nucleus and inhibits the replication of influenza virus at the level of primary transcription. To examine, whether viral proteins interact with Mxl, we stably transfected a routine cell line derived from a A2G mouse with expression-vectors carrying cDNAs of the virus polymerase subunits PB1 and PB2 or of the nucleoprotein (NP) of the viral transcriptioncomplex. A2G cells (Mx+) are resistant to influenza virus infection when pretreated with interferon-a/I]. Interferon treatment of A2G cells expressing PB1, PB2 and NP induced Mxl synthesis to high levels. Stable expression of PB1 and NP in A2G cells had no influence on the antiviral activity of Mxl. However, the PB2 expressing cells remained partially sensitive to influenza virus infection. Neutralization of the Mxl protein by expression of PB2 protein, suggest that Mxl interferes with the normal function of the influenza virus RNA polymerase either by removing the PB2 subunit from the functional complex or by competing with PB2 for the relevant partners in the polymerase complex. The nature of this interaction is now under investigation.

Michel, M.R., Studer, E., Favre, D. and Arrigo, A.-P.*, Institute of Medical Microbiology, University of Berne, CH-3010 Berne, Switzerland; Molecular and Cellular Genetics, CNRS UMR-106, Claude Bernard University, Lyon, France.

It is well documented that transiently, thermotolerant cells will survive otherwise lethal heat shock challenges. Biologically active C protein of SFV acts in the target cell as a complex pleiotropic regulator of host cellular protein synthesis. Here we show that low amounts of C pro-I 4 rein (10 -10 molecules per cell) transferred by electroporation conferred heat resistance to the target cell. In contrast, high amounts (105-106 molecules per cell) resulted in a significant thermosensitivity. Moreover, in a cell-free translation system, low concentrations of added C protein led to an increased translational efficiency at 30~ and a translational tolerance at the supra-optimal temperature of 43~

In contrast, high concentrations were responsible for the shut-off of protein synthesis at 30~ but did not abolish translational tolerance at 43~ Bovine herpesvims 1 (BHV-1) is an important pathogen of cattle. Like human herpesvirus HSV-1, BHV-1 establishes a latent state in the peripheral nervous system. For HSV-1 the viral DNA genome was shown to be linear in virions, a rolling circle mechanismn was postulated during lytic infection and some evidence for joined genome ends during latency is available.

Here I will present some data about the genome configuration of BHV-1 analyzed by pulsed field gelelectrophoresis (PFG) and Southern blot technique. To examine the latent DNA in the bovine trigeminal ganglia (a tissue with high percentage of connective and nervous fibers) a novel tissue preparation technique was developed without any shearing of DIx/A for characterizatior~ by PFG.

As expected BHV-t DNA in virions was shown to be linear with size of about 140 kb. BHV-1 DNA obtained from tytically infected bovine tissue culture cells exhibited mainly a linear unit size of 140 kb, in addition a smear above and below this band was observed. BHV-1 DNA was detested in one of three ganglia from seropositve cattle. In this ganglia of a latent infected animal, most of the DNA was found to be circular and a small amount linear unit size. Most interestingly the circular DNA appeared smaller than unit size. Southern blot analysis using genomic subfragments of BHV-1 revealed some evidence for deletion of certain BHV-1 DNA sequences in the circular genomes during latency. The coexistence of a high amount of circular genomes with deletions and a low amount of linear unit size genomes gives a new perspective for regulation of neurotropic herpesvirus latency! 5O8 COMPARISON OF CORONAV]RAL CONSERVED SEQUENCES Bridgen, A., Tobler, K. and Ackermann, M. Institut for Virologie, Winterthurerstrasse 266a, 8057 Z0rich.

Our work with the coronavirus porcine epidemic diarrhoea virus (PEDV) has led us to look more closely at RNA and protein sequences and sequence motifs which are conserved between different coronaviral genomes. Many of these are specific to coronaviruses, while others are shared with other positive stranded RNA viruses, for example Berne virus and equine arteritis virus. Functions can be assigned to some of these conserved regions, for example the spike and membrane protein trans-membrane regions or the intergenic "leader RNA" binding site. Other well conserved domains, including regions within the membrane and nucleocapsid proteins, have as yet no defined role. The study of such conserved sequences has two purposes 1) the identification of functionally important regions of the viral RNA and proteins and 2) application to genome amplification by polymerase chain reaction as an aid to cloning. In this poster we compare some of these conserved regions. We also describe the use of the latter technique for the cloning of 4 kbp of the PEDV genome, including the membrane and nudeocapsid protein genes, at a time when nothing was known of the PEDV sequence or of its relationship to other coronaviruses. This technique is therefore applicable to the cloning of other coronaviruses since it is fast, efficient and avoids extensive initial clone screening.

SEQUENCE ANALYSIS OF THE PORCINE EPIDEMIC DIARRHOEA VIRUS MEMBRANE AND NUCLEOCAPSID GENES Tobler, K., Bridgen, A. and Ackermann, M. Institut for Virologie, Winterthurerslrasse 266a, 8057 Z0rich.

We cloned 4 kbp of the porcine epidemic diarrhoea virus (PEDV) genome following polymerase chain amplification of the cDNA using degenerate primers based on conserved coronaviral sequences. Sequence analysis of this cloned DNA revealed two major open reading frames. The translated products showed good homology to other coronavirus membrane and nuoleooapsid proteins, particularly to those of human coronavirus 229E (HCV 229E) and porcine transmissible enteritis virus (TGEV). The PEDV membrane protein showed greater homology to those of other coronaviruses than did the nucleocapsid protein, which was longer than the corresponding HCV 229E or TGEV proteins. A second, smaller, open reading frame was present within the PEDV N gone, similar to the f gene of bovine coronavirus. The PEDV genome also contained RNA motifs typical of coronaviruses including a 7 base intergenie sequence and an 11 base sequence near the poly A tail. These results confirm the previous, provisional, classification of PEDV as a coronavirus. We recently reported the detection of rearranged genomic and viral DNA by indirect in-situ PCR in cell preparations (Diagn Mol Pathol 1: 85-97, 1992) using single primer pairs for in-situ DNA amplification and oligonucleotide probes for subsequent detection of PCR-products by in-situ hybridization (ISH) . In this study we tested the feasability of different in-situ PCR techniques in detection of CMV and HBV DNA in tissue sections. We compared I. indirect in-situ PCR with single primer pairs and oligo-nucleotide probes, 2. indirect in-situ PCR with multiple primer pairs and genomic probes, and 3. direct in-situ PCR incorporating labeled nucleotides into PCR-products and immtmohistochemical detection without ISH.

In-situ DNA amplification using multiple primer pairs followed by ISH with genomic probes provided clear evidence of amplification on tissue sections. Insitu PCR with single primer pairs and an oligonucleotide probe yielded negative results. Direct in-situ PCR resulted in non-specific nuclear signals, which were thermal cycling and Taq-polymerase dependent and were also observed in experiments with omission of primers in the PCR reaction mixture, excluding mis-prlming as a possible cause. These non-specific nuclear signals may be related to internal priming and incorporation of labeled nucleotides through DNA repair. Indirect in-situ PCR provides more specific results and can be applied to tissue sections when modified to include multiple primer pairs and genomic probes. Direct in-situ PCR is not suitable for tissue sections with presently used protocols.

Hcrtig, C.; Stalder IIP. and Peterhans, E., Institute of Veterinary Virology, University of Bern, L~inggass-Strasse 122, CH-3012 Bern, Switzerland

We have amplified and sequenced parts of the genomes of 12 laboratory strains of bovine viral diarrhoea/mucosal disease vires. The virus strains originated from North America and Europe.The cumulative sequence heterogeneity of the amplified fragments located in the region of the nonstmctural protein pg0 was 24% as compared to 47% in the gp53 region. The nucleotide changes did not alter the deduced amino acid sequence in pS0. In contrast, the nucleotidc changes located in gp53 resulted in mutations of 42% of deduced amino acids. These results indicate that the viral surface glycoprotein gp53 is under considerable immunological pressure while conservation in pg0 indicates that this protein is subject to constraints related to its function in the replication of the virus. Analysis of the genome of virus from persistently infected animals immunotolerant to their "own" virus revealed that both the pS0 and gp53 fl'agments remained identical over a period of 11 months. Although restricted to selected regions of the viral gcnome, our analysis suggests that immunological tolerance may limit or even "freeze" the evolution of BVD/MD viruses despite the high mutation rate typical of RNA viruses. Parasltc wasps of the genus Chelonus inject their eggs into eggs of various lepidopterous species. The parasitoid larvae develop within the developing host larvae; parasitisation causes hosts to enter metamorphosis one instar earlier than nonparasitised larvae. The wasp females inject, together with the parasitoid egg, venom and calyx fluid which contains polydnavirus. This virus appears to be essential for successful parasitoid development. We observed that the viral genome consists of at least 10 different classes of circular doublestranded DNA molecules with a length of 7 -31 kb. Various circles were cloned and characterized by restriction analysis. Southern blot experiments revealed that viral DNA exists in the wasp in two forms, either integrated in the genome or extrachromosomally. Both forms exist even in tissues from males which are not known to produce polydnavirus particles. Replication of extrachromosomal viral DNA was seen to begin in females at a specific stage of pupal development. We are currently analysing the structure of integration sites of individual viral DNAs in the wasp genome. Our earlier work on the recognition of bacteriophage QI3 plus strand RNA by QI~ replicase bad shown that specific binding occurred at two internal RNA sites, the 8-and the M-site, but that binding at the 3'-end (the site of initiation of synthesis) was not detectable. We now found that doublelooped structures resulting from simultaneous binding at the 3'-end and at the two internal sites could be demonstrated by electron microscopy after initiation of RNA synthesis in presence of host factor, GTP and ATP. Unexpectedly, similar structures were also found for complexes consisting of plus strand RNA and host factor without replicase. This suggests that the role of the host factor on the plus strand template is to bring the 3'-end into the proximity of the S-site/M-slte domain, where replicase can initiate on it. Kinetic studies of initiation by replicase showed that the plus strand, even in presence of host factor, was a far slower and less efficient template than the minus strand; for the latter, host factor had a substantial inhibitory activity. Our findings are easily explained in the context of a template specificity model based on steric exclusion developed earlier.

Antigenic variation in Borrelia burgdorferi after passage through Ixodes ricinus and Ixodes hexagonus ticks.

Hu C.M., Gern L., Toutoungi L.A., Aeschlimann A. Institut de Zoologie de rUniversit6, 2000 Neuch~teL In Europe, Borrelia burgdorferi (Bb), the causative agent of Lyme borreliosis is transmitted to humans and/or animals by Ixedes ricinus (Ir) and L hexagonus (Ih).This means that Bb spend part of its life cycle within the arthropod, an environment completely different from that found within the vertebrate hosts. In this study, 9 in vitro grown Bb isolates were artificially reintroduced into Ir and Ih and afterwards recultivated to determine wether the passage through ticks influences the protein and antigen profiles of Bb. All isolates and reisolates were examined by SDS-Poge and Western Blot using poly-and monoclonal antibodies. Comparing initial isolates with the reisolates, 7/9 presented modifications of the protein and antigen profiles with respect to outer surface proteins A, B and the 22kD protein after passage through Ir, The same modifications were observed after passage through Ir and Ih using a cloned isolate, The fact that the phenotype of Bb changed during residence in the tick suggests that ticks provide the environment which leads to the variation of one or the other epitope of Bb antigens.

Z~ckert, W.R., Filipuzzi-Jernly, E., and Meyer, J. zahn~rztliches Institut, Abt. PZMOM, CH-4051 Basel Like other Borrelia species, B. burgdorferi disDlays a quite unique feature among prokaryotes: Its genome consists of a linear chromosome compl~entedby circular and linear plasmida. We cloned a 3.8 kb EcoRI fragment from the 29 kb circular plasmid of the type strain B31, which hybridized to several additional restriction fragments of the 29 kb circular (1.4, 6.4, 7.0 kb EcoRI) and the 50 kb linear plasmid (4.5 HindIII) of this strain. These were also cloned and their crosshybridizing segments identified on subclones. Partial DNA sequence analysis revealed > 90 % homology over several hundred bp. Whether a second region of homology occurs on some of these fragments is still under investigation. Signals with linear and circular plasmids of other Swiss and American B. burgdorferi isolates were also detected. The repeated DNA seems to be speciesspecific and therefore may be useful in DNA diagnostics of Lyme disease. We have recently reported the existence of four distinct chromosomaUy located sites femAB, femC and femD, by Tn551mutagenesis, lowering the level of resistance in methicillin resistant Staphylococcus aureus strain BB270 (mec). We now report the cloning and sequencing of a 2.6 kb HindlIl fragment containing the 1.0 kb right end of Tn551 and an adjacent 1.6 kb chromosomal DNA junction fragment of the femC region from strain BB589 (BB270, f~2005 femC::Tn551). Tn551 has integrated upstream of an open reading frame that has 76.5% identity at the amino acid level with the Bacillus cereus glutamine synthetase (GS). By location and analogy with the B. cereus GS operon, we suspect the insertion has occurred in a possible S. aureus GS repressor. Northern blots probed with the putative GS coding region revealed a 1.9 kb transcript in BB270 that was replaced by a weak, approximately 7.0 kb transcript in BB589. f~2005 (femC::Tn551) thus alters the size and relative amount of the putative GStranscript, suggesting that the lowering of metbicillin resistance observed in the femC mutant is due to this alteration.

S. BAS and T.L. VISCHER, Division de Rhumatologie, H6pital cantonal universitaire, Geneva, Switzerland.

Reactive arthritis may follow urethritis or cervicitis caused by CT but the process leading to joint inflammation from a distant infection is not known, An important question concern the presence or not of viable CT in the joint. To resolve whether whole bacteria or merely antigenic fragments were present in reactive arthrilis synovial fluid, we tried to detect bacterial rRNA or DNA.

The presence of CT rRNA was tested on SFC of 9 suspect patients and 55 controls patients. This was performed by hybridization with a chemilumineseent labeled, singlestranded DNA probe complementary to the rRNA of CT (GEN-PROBE PACE 2 SYSTEM, San Diego, California) in a liquid reaction mixture [1] . Without cells, the sensitivity of the probe was of 1000 elementary bodies. In presence of 28.5 to 60.106 SFC, the sensitivity was of 10 000 elementary bodies. The RLU (relative light units) values were comparable for both groups. The presence of CT DNA was searched in SFC of 14 suspects patients and 13 controls patients. The polymerese chain reaction (PCR) was used to detect a sequence of CT plasmid DNA considered to be essential for chlamydial growth. After amplification by PCR in presence of specific oligonucleotide primers amplifying a sequence of 201 bp [2] and hybridization with an internal digoxigenin labelled DNA probe of 114 bp, the sensitivity of the method in presence of 5 Ilg of SFC DNA was of 10 elementary bodies. In these conditions we found 7 positive results from the suspected patients and none from the controls. The working conditions have been established with this system and we are trying now to confirm these results with two other sequences (one for plasmid DNA and one for genomic DNA). For studies of human specific activation of promutagenic compounds, a cDNA coding for the human cytochrome P450 enzyme CYP1A1 was functionally expressed in Saccharomyces cerevisiae. With the idea of further increasing the heterologous enzyme activity, a human NADPHcytochrome 17450 reduetase (hOR) eDNA was coexpressed in the strain expressing human CYP1A1. Furthermore, to characterize the bOR enzyme in more detail the hOR expression plasmid was introduced into a strain where the yeast homolog CPR1 had previously been disrupted. From these experiments it was concluded that: (1) The CPR1 gene is not essential for growth.

(2).The cprl ::URA3 disrupted strain exhibits no cytochrome-crednctase activity. (3) The cprl::URA3 disruption confers sensitivity towards ketoconazole. (4) Heterologous hOR expression complements the cprh:URA3 disruption. (5) The hOR enzyme exhibits a four times higher affinity than the yeast enzyme for the substrate NADPH. (6) A., Sengstag, Ch. and Wiirgler, F.E., Inst. fiir Toxikologie der ETH und Univ. Ziirich, Oncogene activation and tumor suppressor gene inactivation can be attributed to genetic changes such as point mutations, chromosome losses, mitotic recombinations and gene conversions, which take place within the future tumor cell and eventually lead to the development of a tumor. The ability of a chemical to induce such genetic changes can be easily tested in the S.cerevisiae test strain YHE2, a derivative of strain D7.

In order to detect test substance induced gene conversions, we work with two mutant alleles of the TRP5 locus. One of these, the allele trp5a, carries a stable mutation (reversion rate <10-7). Therefore we were interested to characterize it in order to gain further insight into the observed gene conversion events.

Our analysis revealed four mutations in the trp5a allele. In contradiction to the Kozak rule, the transition A to G at position -3 before ATG has no influence on gene expression. The transition C to T at position 168 after ATG results in an amber stop codon which abolishes gene expression. Two further mutations in the 3' untranslated region of the trp5a gene, transitions G to A lying 182 and 413 bp downstream of the open reading frame, also contribute to the phenotype of yeast strain YHEI (trp5a), since its tryptophane prototrophy is not amber-suppressable.The extent of these contributions remains to be determined. Increased BZRLA has been related to the mechanism of HE. Acute liver failure was induced in adult male Sprague-Dawley rats by intraperitoneal injections of thioacetamide (TAA, 600mg/kg/day for 3 days). All 17 TAA-rats developed HE stage Ill-IV (behavior, activity monitor, ASAT (1898+_1359 U/I); bilirubin (36+_27 limol/I) and centrolobular necrosis in the liver). Respective values for the 14 control rats were: ASAT 45+5 U/I, blilirubin 1.5_+0.5 i~mol/I. Whole brain BZRLA activity measured with a 3H-Flumazenil binding competition assay was similar in both TAA (52.7+34.1 ng/g brain weight) and control rats (44.3+18.2 ng/g, n.s.). In a control experiment with the same model diazepam (lmg/kg/day p.o.) was given for 3 days. BZRLA activity in TAA-rats was 223_+65ng/g, (n=5), in controls 103+23ng/g, n=5, p=0.002. We conclude that the increased BZRLA activity in brain of HE rats is neither necessary nor sufficient to explain the pathogenesis of HE. The pharmacological activity of the racemic drug CIT, a selective serotonergic reuptake inhibiting antidepressant resides probably rather in the S-(+)-CIT than in the R-(-)enantiomer (nyttel et al., J N Transm 88 (1992) 157 ). An HPLC-method has been developed for the quantitative analysis of the enantiomers of CIT, its demethylated metahelites and of the propionic acid derivative (PROP) in the plasma of depressive patients treated with CIT. The basic and acidic compounds were isolated from 2 ml plasm~ samples using liquid-liquid or solid phase extraction and derivatised with HFBA or methyl iodide, respectively. The extracts were submitted to HPLC (Chiracel OD; fluorescence detection). Identification of the enantiomers was performed with an optical rotation detector. Preliminary results about the stereoselectivity in the metabolism and steady-state kinetics of GIT in patients will be presented. Inst., Univ. of Zfirich, Z0rich, Switzerland; Max-Planck-lnst. for Psychiatry, Munich, FRG Pathophysiological theories of schizophrenia have emphasized dysfunctions in dopaminergic (DA) systems, although a role of EAA systems has also been proposed. Lumbar CSF of 19 patients with schizophrenic disorders (according to DSM-III-R; all patients were off drugs for at least I year) and of 16 age-and sex-matched controls were analyzed by optimized precolumn ophtalaldehyde derivatization HPLC. Performing an analysis of covariance with age and sex as eovariates, taurine and tyrosine were found to be significantly decreased in the patients. The remaining amino acids evaluated (e,g., aspartate, glutamate and sulfur containing EAA) as well as the catecholamines metabolites did not differ between patients and controls. However, the level of a compound (P15.5) which we recently identified as~'-glutamyl-glutamine (as it coeluted with the dipeptide at 3 different pH) was significantly decreased by 13-18% in the patients. The results rather do support a possible role of~-Glu dipeptides in the pathogenesis of schizophrenic disorders than is the case for DA or EAA systems. is preferentially bound to human serum albumin [HSA]: its total binding constant is 9 nearly 3-fold higher than that of (+)-(S)-C. Question arose whether this stereoselective difference is dependent on the temperature (range: 4 -480C), pH-value (range: 5 -10.2) and serum albumin species (bovine, rabbit and rat). In all experiments equilibrium dialysis was performed in order to determine % binding of the enantiomers. Using HSA it is shown that the stereoselectively caused binding difference depends not on the pH-value, yet decreases markedly with increasing temperature. Species specificity of the stereoselective different binding becomes obvious when other serum albumins are used: e.g. bovine serum albumin binds -at the opposite to HSA -(+)-(S)-C to a much higher extent than (-)-(R)-C.

E. Martinelli, S. Mfihlebach, Dept. of Pharmacy, Kantonsspital Aarau, Switzerland

More than 40 hospital-admitted patients with acute cerebral disorders (aneurisms, polytrauma, cerebral surgery, epilepsy and others) running a lfigh risk for posttraumatic seizures, were treated prophylactically with phenytdin. Drug dosage and serum level monitoring was prospectively followed up over 17 + 20 days (mean + SD). The drug regimen consisted in a rapid i.v. loading dose (15 mg/kg infixsed over 4 hours), followed by 175 mg (-< 70 kg) or 200 mg (> 70 kg) every 12 hours, for 5 days. Thereafter, individual dose adapatation (i.v. or p.o.) was based on calculation of Michaelis-Menten kinetics Using Bayesian forecasting three serum level determinations (one prior to the second dose, the others prior to the morning doses of day 4 and 5) allowed a first dose recommendation. Dose adaptation was given after each additional serum level determination (once to twice a week). 16 hours after initiation of therapy more than 85% of the patients showed a plasma trough level of 9-20 rag/L; before starting individualized dosing more than 60% of patients remained within this range, The serum level prediction showed an over all accuracy of 4.6 mg/L _+ 0.35 (mean of absolute error + SEM) and a bias of -0.21 mg/L 5:0.6 (mean + SEM). Dose range was betwecu 100 and 650 rag/day (i.v.). About 85% of the first rnaintainance doses had to be adapted. In the patients studied neither posttraumatic seizures nor major intoxications occurred. Therefore, independent of comedication (drug interaction potential), the results of this study indicate a clear patient benefit when using the phenytoin dosage regimen presented. Bayesian forecasting allows rational dose individualization of compounds showing more complex pharmacokincties and small therapeutic indices. Thereby safety and effectiveness of drug therapy may be greatly enhanced. In the search for an orally active form of Desferal| for the treatment of iron overload, we have substituted the 3 hydroxamates and the amine group of desferrioxamine B. A considerable number of modified desferrioxamine B derivatives (prodrugs) have been synthesized and their in vivo activity evaluated in animal models. The substitution of the charged and polar functions by hydrophobic groups increases the lipophilicity of the molecules and should improve the absorption of drugs when given orally. Furthermore the modifications of hydroxamates which are the Fe-binding sites of the molecule prevent possible chelation of Fe from food in the GI tract. Many of the tetrasubsfituted prodrags induced significant Fe excretion in both rats and monkeys, indicating that once absorbed, prodrugs are hydrolysed to the active parent compound desferrioxamine B. Several of them were also orally active. Most of the prodrugs, when given subcutaneously have shown long duration of action. This observation has led us to consider these prodrugs as long acting "depot" forms of Desferal. Desfcrrioxamine B derivatives which were monosubstituted at the amino group have shown similar deferralizing properties and in addition improved chemical stability.

Beatrice Baldinger, Markus Hasenfratz and Karl B~.ttig, Behavioral Biology Laboratory, Swiss Federal Institute of Technology, ETH-Zentrum, 8092 Zurich, Switzerland

As found in an earlier study with cigarettes of varying nicotine and tar yields, tar yield seemed to be more important for regulating smoke intake than nicotine yield, since the smokers compensated only while smoking low tar/low nicotine cigarettes (ultra-light) but not medium tar/low nicotine cigarettes (test cigarette). Hypothesizing compensatory intensification of puffing and inhaling, the present study examined puffing behavior. Twelve female smokers participated in the experiment smoking cigarettes with tar and nicotine yields of 11.4 and 0.9 (habitual); 2.0 and 0.21 (ultra-light); and 9.3 and 0.08 mg/cigarette (test), respectively, resulting in tar/nicotine ratios of 12.& 9.5, and 116.25. Physiological parameters and subjective ratings confirmed the results of a former switching study. Whereas the test cigarettes produced only weak physiological effects and low taste scores, the effects of the habitual cigarettes were in the expected range and those of the ultra-light cigarettes about hallway in between. The present puffing behavior measurements revealed a greater number of puffs and greater total puff volumes for ultra-light than for habitual and test cigarettes while puff interval was longest for the habitual cigarettes, shorter and similar for the test and the ultra-light cigarettes. These results support the eadier conclusion that ultra4ight cigarettes were smoked eompensatorily whereas the test cigarettes with a very high tar/nicotine ratio did not produce a comparable behavior.

Hintermann E., and Meyer U. A., Department o! Pharmacology, Biocenter of the University, CH-4056 Basel

Amine acetylation plays important roles in the regulation of physiological processes as well as in the metabolism of drugs and environmental chemicals. In vertebrates, for example, an N-Acetyltransferase (NAT) which acetylates serotonin, is involved the in seasonal regulation of reproduction and photo-pedodism, In insects, dopamine is acetylated, forming an intermediate necessary for sclerotization of the insect cuticle. In addition, NATs may be involved in neurotransmitter inactivation. It is so far unknown whether a single or several different NATs are responsible for all these activities. Using tryptamine acetylation as activity assay, we have fractionated D. melanogaster homogenate by conventional chromatographic procedures. Two activities could be separated on a MonoQ column by e/ution with a pH gradient. The electrophoretical/y almost homogenous enzymes (800x increase of specific activity) have a rel. mol. weight of approximately 30kD. The purified proteins will be used to raise antibodies and to develop probes for molecular cloning. With these tools, we would like to investigate their functional roles and possible relationships to vertebrate NATs.

CIGARETrE SMOKING RELATED VARIATIONS OF PULSE AND ACTMTY UNDER FIELD CONDITIONS Albert Jacober, Markus Hasenfratz & Karl B~itti~ Behavioural Biology Laboratory, Swiss Federal Institute of Technology, ETH-Zentrum, CH-8092 Zurich, Switzerland Small portable devices were used to record 30-see averages of heart rate and motor activity over 2 consecutive days and nights. On smoking days, subjects had to press a marker button whenever they lit a cigarette. On abstinence days, when smoking was not allowed, the moments of strong desire to smoke had to be marked. Smoking related pulse (SRP) and activity (SRA) were computed by averaging the 30-see data across all cigarettes for the 1O minutes preceding and the 10 minutes following the lighting of the cigarettes. The grand means of SRP and SPA started to increase 5 minutes before and reached a maximum with lighting the cigarettes, dropped immediately after lighting to an absolute minimum and while SPA remained at prelightlng levels, SRP recovered continuously from 5 minutes after lighting onwards to a maximum. Qualitatively equal curves were obtained for both sexes, for workdays and days off and for subjects with different professions, suggesting an ubiquitous presence of the phenomenon. In contrast to the SRP and SRA of the first 5 cigarettes of the day, which showed the same pattern with even more pronounced increases of SRP 5 minutes after lighting, the SRP of the last 5 cigarettes of the day remained at prelighting levels after cigarette lighting. These findings support the concept of nicotine induced heart rate increases which disappear during the day, and the SRP and SRA of the markings of smoking desire on abstinence days confirmed tkls view since no SRP increases after these markings occurred. The pronounced increases of SRP and SPA before the lighting of the cigarettes might be interpreted as conditioned unrest, preceding the lighting of a cigarette.

METABOLIZING P450 ISOZYMES BY CARBON MONOXIDE Bonnabry, P., Leemann, T., and Dayer, P., Division de Pharmacologic clinique, H6pital Cantonal Universitaire, CH-121 ] Gen~ve 14 Carbon monoxide (CO) binding to protoporphyrin IX is a typical feature of all cytoehrome P450 enzymes and inhibition of monooxygenation reactions by CO is often used to identify the catalyst as a P450, The selectivity of CO binding to specific P450 isozymes was investigated by studying the inhibition by CO of prototype reactions for 3 major drug metabolizing P450s in microsomes from 6 different human livers: dexlxomethorphan (DM) O-demethylation (P450Dm, CYP2D6), diclofenac (DC) 4'-hydroxylation (P450TB, CYP2C subfamily) and midazolam (MZ) l'-hydroxylati0n (P450NF, CYP3A subfamily). Substrates were incubated under identical conditions (10 min, 20 I.tg microsomal protein in 200 txl). Reactions were monitored (metabolite production -HPLC) in 2 different conditions: CO saturation 1) only before starting the reaction (8ml/min bubbling for 2 rain), and 2) continuously (before and throughout incubation). Inhibition was virtually complete (> 94 %) for all 3 reactions under continuous saturation. When saturation was not maintained throughout CO displayed different affinity for P450DBI, P450TB and P4501,ll=: 82 _+ sd5, 61 + 7 and 46 + 5 % inhibition, respectively (p < 0.01, ANOVA). Neither substrate concentration, Kin, Vmax nor CLint (Vm,x/Km) appeared to have a major influence on CO inhibition, Selective inhibition by carbon monoxide may be a useful marker of specific human cytochrome P450 isozymes functions. Cytochrome P450 genes (CYP450) of insects are involved in the biosynthesis of steroids and the metabolism of other compounds which makes them necessary in development and physiology. In the field of toxicology they are crucial for chemotoxicity and genotoxicity. First they are associated with insecticide resistance (e.g. DDT). Second there is evidence that they are metabolizing genotoxic substances which can be detected in the somatic mutation and recombination test (SMART).

In order to characterize CYP450 genes of Drosophila melanogaster we started a cloning project. A first approach, where we screened a third instar larval cDNA library with various human CYP450 cDNAs, failed. This can be explained as follows: either the over all sequence homology is not conserved enough between human and Drosophila, or homologous genes are not expressed at this developmental stage. Therefore, we have been using oligonucleotides as hybridization probes, coding for the conserved heme-binding domain, for screening a genomic library. Independent clones obtained were further characterized by sequencing. Prelimanary results will be shown. The waste water of a hospital is monitored for genotoxie activity by the umuC test, a bacterial test which measures the activation of the $05 repair system. Salmonella thyphfmurium TAr555 carrying the plasmid pSK]O02 with a umuC-lacZ fusion gene is used. The induction of the fusion gene is assayed colorimetrical]y. Since a change in the piasmid copy number would influence the number of inducible genes and by this the amount of gene product measured, we performed minipreparation of the plasmid after incubating bacteria with different waste waters and known drugs. With chloramphenicol an amplification of the plasmid could be shown after 16 hours, but not yet after 2 hours, the standard incubation time for the umuC test. With mitomycin C, a known genotoxic compound (Albertini at el, Env.Mal.Mut I2: 353-363 ,1988) , no influence on the copy number was found. Interestingly, waste water samples, which produced a significant induction in the umuC test, showed a reduced yield o1 the p}asm~& ]-his effect could be due either to the isolation procedure influenced by waste water components or by a particular cytotoxic mechanism.

In addition selected waste water samples were analysed for their mutegenfc potential in the Ames test and for their potentia] to induce the formation of micronuclei in e eukaryotic ceil }ine. Cytomegalovirus retinitis Is the most frequent infectious ocular complication among AIDS patients. Ganciclovir and foscarnet are widely used vlrostatic drugs, which have been proven to be effective in stopping the progression of the retinitis. A new way of application these compounds was tested using transscleral Iontophoresis in pigmented rabbits. The analytical work consisted in the quantification of these drugs in different ocular tissues and fluids. The procedure for the quantification of gancyclovir involves the use of acyclovir, an antiviral drug structurally related to ganciolovir as the internal standard. After a two steps sample preparation, the extract was injected on a HPLC-UV system fllchrosorb RP 18,5 Ix, 200 x 4.6 ram) using an isocratic program wlth a phosphate buffer:acetonltrfle mobile phase 11 ml/mn). The effluent was monitored at the UV maximum (254 nm) and the method was specific and sufficiently sensitive to allow quantification in a range of concentration from 0.1 Ixg to 30 l~g/ml. For GC-MS confirmation, after derivatization with MSTFA, TMS derivatives were obtained allowing a very specific detection of ganclclovlr and acyclovir.

The same strategy is now used for the quantification and the confirmation of foscarnet.

Portmann, R. and ~olzer, R., Laboratory Spiez, CH-3700 Spiez Toxicity of a substance is only a relative and not an absolute value. In many circumstances it is therefore satisfactory to be able to evaluate toxicity with a reproducible method. Since Toxicity may also be dependent on metabolic turnover, easily available only in intact organism as e.g. Artemia salina, being no warmblooded animal, no vertebrate, easy to handle etc and allowing a low cost bioassay for toxicity. But there are almost no comparable data available. We therefore decided to standardize the assay condition and use mierotiterplates for the test. To speed up the evaluation of different test substances a computer guided program was developped controlling and calculating the needed dilutions and pipetting. An image analyzing unit is recording at preseleeted time intervals automatically the number of total organism per cavity as well as the swimming velocity. This allows us to evaluate easily the lethal concentration and the effective concentration (Lc 50 / Ec 50) with the special designed analysis program. Data so far obtained, show that the sensibility towards pesticides is comparable to results obtained with rats or mice. A description and photographs of the apparatus as well as a table of over 50 substances tested so far will be presented.

Oxidatively modified proteins in iron overloaded mice Wieland, P., Sehranz, Chr. and Lauterburg, B.H., Institute of Clinical Pharmacology, University of Berne, Switzerland Iron overload in humans is associated with hepatic injury and a high incidence of hepatocellular carcinoma. The toxicity of iron has been attributed to iron dependent Fenton reactions generating highly reactive oxygen species. Yet, the evidence for oxidative damage in chronic iron overload is scarce. We therefore looked for oxidative alterations in the liver of a hemochromatosis model. Mice were fed regular chow containing 1% carbonyl iron for five months. Liver homogenates and isolated mitochondria were incubated with 2,4-dinitrophenyl hydrazine (DNPH). The proteins were washed extensively and analyzed spectrophotometrically after solubilization in 6M guanidine. Carbonyl functions of oxidatively altered amino acids react with DNPH to form a hydrazone absorbing at 370 nm, Sections of liver from iron fed mice showed grade 2-3 stainable iron in maerophages, Kupffer cells and hepatocytes (preferential in pert-portal zone) but no structural damage by light microscopy. In iron fed mice the concentration of protein carbonyls was 5.4 _+ 0.6 compared to 2.4 -+ 0.2 nmol/mg protein in the controls. In mitoehondria the concentration of protein earbonyts was 1.4 + 0.2 and 0.8 _+ 0.1 nmol/mg protein in iron-loaded and control mice, respectively. The present data indicate that iron overload results in a significant increase in the steady state concentration of oxidatively modified proteins in the liver. Oxidatively altered proteins may eventually result in functional and structural cell injury.

Singh, A., MacLellan, K., Chu*, I. and Villeneuve*, D.C. Atlantic Veterinary College, Charlottetown, PE, and "Environmental Health Directorate, Ottawa, Canada.

Groups of 20 weanling Sprague-Dawley rats, divided equally between sexes were given PCB congener #126 (3,3',4,4',5-pentachlorobiphenyl) in their diets at 0, 0.1, 1.0, 10.0 or 100 ppb for 13 weeks. Many liver parenchymal ceils from animals fed the congener contained proliferated smooth reticulum, increased number of lipid droplets, and initochondria with abnormalities. In addition, lipofuscin particles were numerous in the cells from animals of 100 ppb group. Sex differences in lesions were not discernible except in the highest dose group females where lipofuscin particles were relatively higher in number in the hepatocytes.

Ottr results indicate that the congener causes liver alterations in rats at even 0.1 ppb level. Pseudomonas aeru~inosa as a hospitalism agent is a common contaminant in laboratory animal facilities and is transmitted to mice through drinking water. It can cause septicemia and death in experimer.taly i~m~/nosupressed mice and is therefore undesirable in SPF animals. Different regimes of treating drinking water are currently in use to prevent infection of mice with P. aeru@incea. This study evaluated the effect of different chlorine concentrations (0, 2, 4, 8, and i0 ppm) in drinking water on existing infections or Constant reinfections with P. aeru@inosa in mice. Animals that were initially infected and recieved water containing 6 or 8 ppm chlorinated water cleared the infection within one week and remained free for the duration of the stud~. All other animals stayed infected. Mice that recieved 6 ppm chlorinated drinking water and were weekly reinfected with P. aeruginosa withstood the infection for three weeks. AI--I other animals were infected within one or two weeks.

Interleuldn-1 ~ (IL-1) is a potent inducer of various go-inflammatory cytokines in vivo and in cultured cells. Factors such as platelet-derived growth factor (PDGF) or FGF are assumed to amplify induction by co-stimulation of cytokines and/or upregulating IL-1 receptor numbers on cells. Fetal calf sentm (FCS) contains various polypeptide factors, including PDGF, and stimulates interleukin-6 (IL-6) production. Therefore cultured human dermal fibroblasts were incubated in MEM supplemented with 0 to 10% FCS for 24h. Without IL-1 stimulation the concentration of IL-6 in the supematant rose in a dose dependent manner (0 to 7.5 ng IL-6/ml)

The role of PDGF-BB in the production of IL-6 after IL-1 stimulation was investigated by incubating fibroblasts in medium containing 0.5% FCS with ILl (15 U/ml), PDGF-BB (20 ng/ml), and both IL-1/PDGF-BB, respectively, for 24h: IL-1 moderately stimulated IL-6 production

ng/ml), whereas PDGF-BB alone had only little effect (1 to 15 ng IL-6/ml)

IL-1 together with PDGF-BB enhanced IL-6 production to very high levels of 400 to 450 ng/ml. Our findings suggest that IL-6 production in fibroblasts stimulated by IL-1 is dependent on the presence of serum. PDGF is one major though possibly not the only factor in serum acting synergistically with IL-1

REGULATION OF IL-8, GRO~ AND ENA-78 IN LUNG TYPE-II EPITHEUAL CELLS, FIBROBLASTS AND MONOCYI'ES

Switzedand Intedeukin-8 (IL-8), GROo. and ENA-78 are potent neutrophil-stimulating peptides inducing chemotaxis, respiratory burst and the release of granule enzymes. All three peptides are structurally related and belong to the CXCfamily of the chemokines. We have studied the kinetic of induction of IL-8, GRQ~ and ENA-78 by northern blot analysis, neutrophil-stimulating activity and ELISA measurements in human adult lung fibroblasts, lung type-II epithelial cell line A549 and blood monocytes

In ltpepolysaccharide stimulated monocytes, however, ENA-78 mRNA was transiently induced at 18 to 24 hours, whereas IL-8 and GRO~ mRNA reached maXimal levels at 4 to 8 hrs after induction. This difference in regulation has been confirmed with specific ELISA measurements, demonstrating that ENA-78 peptide is released from monocytes when 11--8 activity is downregulated

The modification of proteins by covalent attachment of fatty acids, such as palmitic acid, is a common but mandatory event in the maturation of enveloped viruses. We are studying the acylation of envelope proteins of Semliki Forest virus (SFV), an RNA virus containing the spike proteins El, E2 and B3. B1 and E2 become acylated in vertebrate (Vero) and invertebrate cells (A. albopictus C6/36). We found considerable differences between these cells. In C6/36 cells, mainly El acted as primary acceptor of palmitic acid whereas p62, the precursor of E2 and E3, was weakly labeled. The palmitoylation of E2 (or p62) seems to be delayed as compared to El and might be coupled with the cleavage of p62. Further data suggest that the small amount of acylated p62 contains less bound palmitic acid residues within its cytoplasmic domain than its cleavage product E2. This domain is responsible for the spikenucleocapsid interaction during virus budding. The role of acylation in this interaction will be discussed. In previous work we demonstrated the usefulness of chemically induced PMV for studying virus binding per se omitting endocytosis. Here we present data characterizing SF viral binding to the cell surface using this model and, for comparison, living cells.Thereby the need of divalent cations for virus attachment could clearly be demonstrated.Our results further suggest an involvement of sialic acids in SFV binding sites. Both, removal of sialic acid residues from PMV by neuraminidase treatment and free sialic acid present in the incubation mixture led to a reduction of PMV binding ability. These results could be confirmed by experiments using living Vero cells. Similar treatments strongly reduced the susceptibility of the cells to SFV infection. Studying fusion between viral and plasma membranes represents another application of PMV.

Measles virus (MV) rarely induces lethal diseases of the human central nervous system characterized by reduced expression of the viral envelope proteins, and by lack of viral budding. The MV envelope contains two integral membrane proteins, fusion (b0 and hemagglutinin (H), and a membrane-associated matrix (M) protein. Previously, analysis of MV genes from autopsy material indicated that the M protein and the F protein intracellular domain are often drastically altered by mutations.Here we present evidence that truncation of the F protein intracellular domain does not impair fusion function and we suggest that this alteration interferes with viral budding. Unexpectedly, certain combinations of functional F and H proteins were unable to induce syncytia formation, an observation suggesting that specific F-H protein interactions are required for cell fusion. We also found that three of four H proteins of persistent MVs are defective in intracellular transport, oligosaccharide modification, dimerization, and fusion-helper function. Thus, MVs replicating in brains at the terminal stage of infection are typically defective in M protein and in the two integral membrane proteins. Whereas the M protein appears dispensable altogether, partial preservation of F and H protein function seems to be required, presumably to allow local cell fusion. Certain subtle alterations of the F and H proteins might be instrumental for disease development.

CAN INCOMING CAPSID PROTEINS CONTROL VIRAL DNA REPLICATION?

The T-lymphocyte-derived cell line EL4 is a host for the immunosuppressive swain of minute virus of mice MVMi. The prototype sWain MVMp, in contrast, grows specifically in A9 fibroblasts. For lyric growth of MVM in lymphocytes, two regions of the i genome are required: iE mapping from nt 1084 to nt 2070 and iL from nt 3523 to nt 4339. For growth in fibroblasts, pL (the region from MVMp analogous to ilL) is sufficient.Here we focus on how iL and pL function in EL4 cells. Coinfection experiments using MVMi and a recombinant MVMi virus modified to contain pL in place of iL (MVMp3) showed that MVMi cannot help the replication of the hybrid p3 viral DNA in trans. Since pL lies in the coding region for the capsid proteins, this result led to a hypothesis that incoming capsid proteins remain associated with viral DNA and inhibit viral replication in cis. To test this hypothesis we are transfecting purified MVMp3 RF DNA into MVMi-infected EIA ceils. If the incoming capsids of MVMp were responsible for the absence of MVMp3 DNA replication, the transfected p3 DNA should be able to replicate.

Sammy Frey', Mark Marsh* Toon Stegmann' 'Biocenter, Basel and *MRC, London Human immunodeficiency virus (HIV), the causative agent of AIDS, is an enveloped virus. The virus has one integral membrane protein, a non-covalently linked complex of gp120 and gp41, which are derived from the same precursor, gp160, a product of the env. gene. While gp120 recognizes the viral receptor, CD4, gp41 is thought to be responsible for the ability of the viral membrane to fuse with the host cell membrane.To study the fusion mechanism, large quantities of concentrated gp 120/gp41 are needed. To avoid having to deal with live virus, part of the HIV genome containing the env gene and the viral regulatory element REV was cloned into a vector for glutamine synthetase selection and CHO cells were transfected with it. After subcloning, stable expression of the protein on the cell surface was obtained, as demonstrated by indirect immunofluorescence with antibodies specific for gp41. Here, it is shown, that fusion of these gp 120/ gp41 expressing cells with CD4 expressing cells, but not with CD4 negative cells, was obtained. We intend to functionally reconstitute the protein from these cells to study the fusion mechanism. Bovine herpesvirus-1 (BHV-1) contains four immediate-early (IE) genes involved in regulation of the productive cycle of replication. A spliced IE RNA (2.9 kb) and an unspliced early RNA (2.6 kb) are 3' coterminal and exhibit a common open reading frame with a coding potential of 676 amino acids. The putative protein contains a cysteine-rich zinc finger domain and is a strong transactivator in transient expression assays. Because of the structural and functional homology to ICP0 of herpes simplex virus, we named it blCPO. In order to identify the viral protein, we produced polyclonal antisera against synthetic peptides in rabbits. The peptides were synthesized according to the predicted amino acid sequence of blCPO. Each peptide represented a specific epitope of the predicted protein, including the C-and N-termini. Sera against the Cterminus and the N-terminus stained nuclei of BHV-1 strain Jura or K22 infected cells in an in situ immunoassay (black plaque). The blCPO protein was expected to enter the cell nucteus because of its gene regulating activity. Western Blots revealed a viral protein of about 100 kD, which was synthesized both under IE conditions and at later times of the BHV-1 infection. Expression of phol encoded acid phosphatase of S.pombe has been reported to be regulated by phosphate. Here we show that in addition it is regulated by adenine.

Starving adenine auxotrophs for adenine, we fmd a drastic increase ofphol encoded acid phosphatase activity. This derepression ofphol occurs at the mRNA level and is dependent on non limiting phosphate concentrations in the media. In order to find genes, that are involved in this adenine-dependent regulation, we screened for mutants exhibiting high aPase activity in adenine excess. The mutants define four complementations groups (anrl, anr2, anr3, anrS) and they also display defective phosphate regulation of phoI. Adenine as well as phosphate are necessary to repress phol expression.We speculate, that the regulating signal is a compound of the adenine nucleotide pool To learn more about genes regulated by the adenine nucleotide pool, we screened a cDNA gene library by differential hybridization and found a gene, which is also repressed by adenine as well as dependent on the anr genes. A characterization of this gene will be given.