key: cord-0005095-l7ojx5tc
authors: Calza, Roger; Huttner, Eric; Vincentz, Michel; Rouzé, Pierre; Galangau, Fabienne; Vaucheret, Hervé; Chérel, Isabelle; Meyer, Christian; Kronenberger, Jocelyne; Caboche, Michel
title: Cloning of DNA fragments complementary to tobacco nitrate reductase mRNA and encoding epitopes common to the nitrate reductases from higher plants
date: 1987
journal: Mol Gen Genet
DOI: 10.1007/bf00331162
sha: 90acfb9f6a0ff2f1a560ddc2ccda19f0f410a32d
doc_id: 5095
cord_uid: l7ojx5tc

Messenger RNAs encoding the nitrate reductase apoenzyme from tobacco can be translated in a cell-free system. Poly(A)(+) mRNA fractions from the 23-32 S area of a sucrose gradient were used to build a cDNA library in the expression vector λgt11 with an efficiency of cloning of approximately 10(4) recombinants/ng mRNA. Recombinant clones were screened with a rabbit polyclonal antibody directed against the corn nitrate reductase, which cross reacts specifically with the nitrate reductases from dicotyledons. Among 240000 recombinant plaques, eight clones were isolated containing inserts of sizes ranging from 1.6 kb to 2.1 kb and sharing sequence homologies. Seven of these clones contained a common internal 1.6 kb EcoRI fragment. The identity of these clones was confirmed as follows. A fusion protein of 170 kDa inducible by IPTG and recognized by the rabbit nitrate reductase antibody was expressed by a lysogen derived from one of the recombinants. The antibodies binding the fused protein were eluted and shown to be inhibitory to the catalytic activity of tobacco nitrate reductase. Two monoclonal antibodies directed against nitrate reductase were also able to bind the hybrid protein. The 1.6 kb EcoRI fragment was sequenced by the method of Sanger. The open reading frame corresponding to a translational fusion with the β-galactosidase coding sequence of the vector shared strong homology at the amino acid level with the heme-binding domain of proteins of the cytochrome b5 superfamily and with human erythrocyte cytochrome b5 reductase. When the 1.6 kb EcoRI fragment was used as a probe for Northern blot experiments a signal corresponding to a 3.5 kb RNA was detected in tobacco and in Nicotiana plumbaginifolia mRNA preparations but no cross-hybridization with corn mRNAs was detected. The probe hybridized with low copy number sequences in genomic blots of tobacco DNA.

Nitrate reductase (NR) is a key enzyme involved in the first step of nitrate assimilation in plants (Kleinhofs et al. 1985) , and is also found in bacteria and fungi. Although in these different organisms it catalyzes a similar reduction of nitrate involving NADH or NADPH as electron donor, the enzyme displays a variety of structures and dissimilarities. In Escherichia coli the enzyme is an heterotrimer bound to the bacterial membrane (Rasul Chaudhry and MacGregor 1953) and in ChloreIla NR is an homotetramer (Howard and Solomonson 1982) found in the pyrenoid. In plants the enzyme is an homodimer possibly interacting with the chloroplast outer membrane. Apart from NADH, the three other cofactors involved in the reduction of nitrate by NR are FAD, cytochrome b557 and the molybdenum cofactor. The NR from several plant species have been purified to homogeneity (Redinbaugh and Campbell 1985; Campbell and Wray 1983) ; and shown to catalyze, apart from nitrate reduction, other reactions, such as the reduction of ferric ion (Campbell and Redinbaugh 1984) , which may be of physiological significance. The subunits of these plant NR have a molecular weight close to 110 kDa.

The regulation of NR activity in plants appears to be rather complex and many studies have been devoted to the description of this regulation (Hewitt 1975) . For instance, the catalytic process of nitrate reduction takes place in the leaves of numerous plant species, but it can also occur exclusively in the roots of other species such as white lupin. Molecular tools have been lacking to study more deeply the features of these regulations and, for instance, NR monoclonal antibodies have been available only recently. Nucleic acid probes would be useful for studying the transcription of the NR structural gene under various physiological conditions. The isolation and characterization of a cDNA encoding more than 50% of the tobacco NR mRNA is presented here.

Extraction and fractionation of poly(A) + mRNA. Total RNA was extracted from leaf tissues according to a procedure derived from Chirgwin et al. (1979) . Tobacco leaves were frozen in liquid N2 and ground in a mortar. The powder was rapidly mixed with 100 ml of 50 mM Tris-HC1 buffer, pH 8.0, containing 10 mM EDTA, 5 M guanidinium thiocyanate, 2% sodium-N-lauroylsarcosine and 5% (v/v) 2-mercaptoethanol. After incubation at 65°C for 20 min and centrifugation at 20 000 x g for 30 rain, the supernatants were adjusted to 0.1 g/ml CsC1 and layered on 12 ml cushions of 5.7 M CsC1, 50 mM EDTA, pH 8.0, in 30 ml polyallomer tubes. The tubes were centrifuged for 20 h at 25000 rpm in a SW27 rotor at 15 ° C. The RNA pellets were dissolved in 10 mM Tris, pH 8.0, 1 mM EDTA, 0.05% SDS and incubated at 55°C for 30 min in the presence of 200 lag/ml proteinase K. After PMSF treatment to inactivate proteinase K, samples were extracted by phenol, phenol/chloroform and chloroform, and ethanol precipitated. The yield was generally 200 300 lag total RNA per gram fresh tissue. Poly(A) + mRNAs were purified by two cycles of selection on an oligo(dT) cellulose column (PL Biochemicals) according to the procedure of Bantle et al. (1978) , including a step of denaturation of the RNA in 80% DMSO at 55 ° C for the second cycle of selection.

Poly(A) + mRNAs represented generally 1.0%-1.2% of the total leaf RNA and contained less than 10% ribosomal RNA. Poly(A) + RNAs (100 gg/tube) were fractionated on 5%-20% linear 12 ml sucrose gradie~ats made in TE buffer (Tris-HC1 10 mM pH 8.0, EDTA 1 raM), by centrifugation for 16 h at 100000xg in a SW40 rotor as previously described (Comm6re et al. 1986 ).

Cell-free translation of mRNAs and immunoselection of translation products. Total or fractionated poly(A) + mRNAs were translated in a rabbit reticulocyte cell-free translation system (Amersham) using 35S-labelled methionine (900 Ci/mmol, I mCi/ml, Amersham). Approximately 0.05-0.1 lag mRNA and 1.2 laCi 35S-methionine were added per microlitre of lysate. After incubation for 1 h at 30 ° C, aliquots of the translation mixture were analysed by SDS-PAGE. Translation products corresponding to 50 lal of lysate were immunoselected for nitrate reductase polypeptide fragments by passing through a protein A-Sepharose 4B column on which 20 lal of NR-specific antiserum had been bound in TNE buffer (50 mM Tris-HC1, pH 7.5, I mM EDTA, 0.5% (w/v) Tween 20, 150 mlV[ NaC1, 0.2% NAN3). After adsorption of the lysate diluted in TNE buffer, the column was washed with TNE buffer containing 1 mM cold methionine and 1 M NaC1 and then with 50 mM Tris-HC1 buffer (pH 7.5). The protein A-bound material was then eluted and denatured in Laemmli buffer for SDS polyacrylamide gel electrophoresis (PAGE).

Construction of cDNA libraries b7 2GT/1. cDNAs were synthesized according to Gubler and Hoffman (1983) with the following modifications. Two hundred nanograms of total poly(A) + mRNAs were heat-denatured at 70 ° C for 2 min in water, and reverse transcribed at pH 8.3 using 14 U of AMV reverse transcriptase (Genofit) in a final volume of 40 lal in the presence of 0.1 lag oligo(dT) (15 mers, Pharmacia) at 42 ° C for 40 rain. The efficiency of synthesis of the first strand varied from 8%-25%. The second strand was synthesized from the first strand in the presence of 0.8 U of RNAse H (BRL), and 4 U of E. Coli DNA ligase (Biolabs) using 40 U of DNA Pol I holoenzyme (Amersham) in a 90 lal final volume. The incubation was performed at 12°C for 1 h and then at 16°C for an additional hour. The double stranded DNA was healed by incubation with 1 U ofT4 DNA polymerase at 37°C for 15 min. The effi-ciency of synthesis of the second strand was close to 100%. The cDNA obtained (approximately 100 ng) was subsequently methylated with 20 U of EcoRI methylase. To improve the efficiency of cloning (tenfold) a step of S1 nuclease treatment followed by a second healing step was included in the procedure of cDNA synthesis between the healing and methylation steps, cDNAs were treated with 2 U of S1 nuclease (Boehringer) in a volume of 100 lal for 15 min at 30°C in order to eliminate abnormal ends, as suggested by Lapeyre (1985) . This resulted in a 5% decrease in the amount of high molecular weight material in the cDNA preparation. After addition of EcoRI linkers to the cDNA (phosphorylated 12 mers, PL Biochemicals) and two successive digestions with concentrated EcoRI, cDNAs were fractionated on a Sephacryl $300 column and the material corresponding to cDNAs larger than 0.5 kb was pooled. The cDNA was ligated with dephosphorylated 2gtl 1 vector cut with EeoRI under conditions specified by Young and Davis (1983) . The ligation mixture was packaged using commercially available packaging lysates (Gigapack, Clontech labs, CA). Libraries were checked for the proportion of recombinants on strain RY1088 in the presence of XGal and IPTG: they contained 70%-80% white plaques. A significant proportion of these white plaques detected in the first libraries turned to pale blue upon retesting on strain RY1090, and were lacking detectable inserts. This was found to be related to the poor elimination of small cDNA fragments and linkers by chromatography on Sepharose 4B columns and was improved by using Sephacryl $300. After optimization, yields were approximately 50000-100000 white plaques per nanogram cDNA, among which 50% contained detectable inserts.

A cDNA library enriched for high molecular weight mRNA sequences was made as follows. A sample of 100 lag of mRNAs was fractionated on a sucrose gradient from which the 23-32 S area was recovered (10 lag). This fraction was further fractionated in a 10 mM methylmercuryhydroxide 1.5% agarose gel. The area corresponding to the 2-5 kb molecular sizes was sliced out, mRNAs were eluted, phenolextracted, and used (100 ng) to build a cDNA library according to the above procedure.

Amplifieation and immunological screening of libraries. Libraries were amplified on strain RYI088 at a density of 3 x 104-1 x 105 pfu per 90-mm dish. Plaques were allowed to grow at 42 ° C for 4 h, resulting in a 1000-fold amplification. Amplified libraries were plated on strain RYI090 at a density of 1.0-1.5 x 104 pfu per 90-ram dish and allowed to grow at 42 ° C for 34 h. Nitrocellulose filters (Millipore HATF) were soaked in 10 mM IPTG and dried. Lawns overlaid with these filters were incubated overnight at 37 ° C and then cooled at 4 ° C. The corresponding filters were blocked with 3% BSA (bovine serum albumin) in TBS (50 mM Tris-HC1, pH 8.0, 150 mM NaC1) for 30 rain, incubated with antiserum (1/200 diluted in TBS) for 1 2 h, and washed three times in TBS containing 0.05% Tween 20. Filters were incubated with 0.5 laCi 125I-protein A in TBS containing 50 g/1 skimmed milk and 0.05% Tween 20 for 1-2 h, washed three times in TBS, milk, Tween 20 and twice in TBS. All steps were performed at room temperature. Filters were autoradiographed on Kodak XAR5 film with Dupont Cronex amplifying screens for 18-24 h. An average of 1-3 spots corresponding to putative recombinants were detected per filter. Plugs corresponding to these spots were taken with the wide end of Pasteur pipetes and incubated overnight at 4°C in i ml of lambda diluent (Tris-HCll0mM pH8.0, MgSO¢ 10. Approximately 3000 plaques per plug were rescreened for confirmation. Two further rounds of screening at low density (100 200 plaques/dish) were required for the purification to homogeneity of recombinants.

Immunological reagents for the screening of libraries. The purification of nitrate reductase from corn and the obtainment of the rabbit polyclonal antibodies (eMNR) against this enzyme (Comm~re et al. 1986 ) have been described previously. To neutralize anti E. coli antibodies from the c~MNR immunoglobulins, the serum was diluted fivefold in TBS-BSA and incubated with two volumes of an E. coli crude lysate (20 mg/ml protein). After incubation for 2 h at 4 ° C with swirling, the mixture was centrifuged in Eppendorf tubes and the procedure of lysate addition and clarification was repeated twice.

Lysogens were obtained by infection of the Hfl strain RY1089 by recombinant phages. Bacterial extracts were obtained from lysogens induced or non-induced by IPTG as described previously (Riva 1985) . Extracts were analysed by SDS-PAGE, Western blotting of the fractionated proteins on nitrocellulose by electrotransfer, and immunodetection with eMNR antiserum and a second antiserum from sheep directed against rabbit IgG coupled with peroxidase (Ch6rel et al. 1985) .

Immunoselection of antibodies recognizing the hybrid proteins expressed by recombinants. Nitrocellulose transfers of lawns of strain RY1090 infected by recombinant phages were saturated for I h with 30 mg/ml BSA and after washes in TBS they were incubated in 5 ml of 1/100 diluted non-immune serum (~So) or ~MNR antiserum. The antibodies adsorbed on the filters were eluted by a 30 s incubation in 5 ml of 0.2 M glycine-HC1 buffer, pit 2.2. After removal of the filter the elution buffer was rapidly neutralized with Tris base and non-immune antiserum was added at a 1/200-fold dilution in a final 10 ml volume. Eluted proteins were concentrated and dialysed against PBS against PBS (Na-phosphate 10 mM pH 7.5, NaC1 150 mM) using an Amicon cell equipped with a PM30 membrane. Eluted antibodies were added to tobacco NR partially purified by affinity chromatography and kindly provided by T. Moureaux. After 5 min of incubation, NR activity was assayed as previously described (Ch6rel et al. 1986 ).

Recognition of the fusion protein in an ELISA test using monoclonal antibodies. The isolation and characterization of the monoclonal antibody 96(9)25, which inhibits the activity of various plant nitrate reductases, has been described previously (Ch6rel et al. 1985) . Several monoclonal antibodies recognizing the tobacco enzyme have been obtained using a similar procedure (C. Meyer, unpublished results). The monoclonal antibodies 7(113)NP15 and 24(48)NP19 recognize native NR from tobacco but have no effect on the activity of the enzyme. Recognition of the fusion protein using monoclonal antibodies was performed using a two sites ELISA test according to a previously described procedure (Ch6rel et al. 1986 ). Reactants were added to the microtitration plates in the following order: monoclonal anti- Phage DNA preparation, subcloning of inserts and DNA sequencing. Bacterial and phage procedures were essentially as in Miller (1972) . The strains used are described by Huynh et al. (1985) . Phage production, purification and DNA extraction were performed according to the procedure of Huynh et al. (1985) described for ,~gtl0.

All recombinant DNA manipulations were performed using standard methods (Maniatis et al. 1982) . EcoRI fragments excised from the vector were subcloned in the dephosphorylated EcoRI site of pUC9 according to standard procedures. The bacterial host strains for pEMBL vectors (Dente et al. 1985) were TG2 or JM109 (Yanish-Perron et al. 1985) . The cDNA insert of clone 13-29 contains a unique KpnI site. It was subcloned as a 2.5 kb KpnI/KpnI (in 2gtl 1) fragment in both orientations into the KpnI site of pEMBL18 and as a 1.2 kb KpnI/SstI (in lacZ of 2gtll) fragment in the KpnI/SstI sites of pEMBLI8. The 120 bp of cDNA sequences contained in the KpnI/SstI fragment were turned round by cloning into the EcoRI/XbaI sites of pEMBL19. Ordered deletions of the KpnI fragment were generated by the ExoIII, $1 nuclease method (Henikoff 1984) . The single-stranded plasmid templates were prepared for sequencing mainly as described previously (Dente et al. 1985; Messing 1983) , except that a single bacterial colony containing a recombinant pEMBL plasmid was first grown overnight in M9 minimal medium (Messing 1983) containing 100 gg/ml ampicillin. D N A sequencing by chain termination (Sanger et al. 1977) and labelling with 35S-nucleotides was performed using standard methods (Messing 1983) . A synthetic oligonucleotide and the 2 reverse primer (Biolabs) were used to complete the sequence. Ninety-eight percent of the sequence was determined from both strands.

Computer programs. Sequence data treatments were performed with the BISANCE package, using the CITI2 facilities on a DPS8 computer, and the programs of R. Staden (MRC, Cambridge, UK) and M. Kanehisa (NIH, Bethesda, USA).

were fractionated by electrophoresis Jin 10 mM methylmercuryhydroxide 1.5% agarose gels for the accurate estimation of molecular weights, or in 3.0% formaldehyde 1.5% agarose gels for routine experiments. After fractionation, mRNAs were transferred to Hybond C (Amersham) membranes as described previously (Thomas 1980) . Total D N A was extracted from young tobacco leaves by a previously published procedure (Dellaporta et al. 1983) , purified by CsC1 gradient equilibrium centrifugation and used for Southern blot analysis as described previously (Deshayes et al. 1985) .

Northern and Southern blots were probed with the nicktranslated 1.6 kb EcoRI fragment from phage 13-29 purified by sucrose gradient centrifugation, or with nick-translated pTA71 carrying a 9 kb wheat rDNA insert (Gerlach and Bedbrook 1979) . Hybridization was performed in 50% formamide at 45 ° C and washed under stringent conditions (3 washes at room temperature in 2 x SSC, 0.1% SDS, followed by two 30-rain washes at 65°C in 0.2 x SSC, 0.1% SDS; Maniatis et al. 1982) . SSC buffer is 0.1 M NaC1 and 0.1 M sodium citrate, pH 7.0.

Immunological screening of a cDNA expression library requires that the new translation product can be recognized by specific antibodies raised against polypeptide chains matured in eucaryotic cells. A polyclonal antibody against corn NR (c~MNR) could be used to immunoselect a tobacco NR polypeptide chain among the cell-free translation products of total poly(A) + RNAs extracted from leaves (Fig. 1) . The signals obtained were very weak, NR representing 2-5/10000 of total tobacco soluble leaf proteins. Cell-free translation products migrated close to the 110 kDa enzyme subunits. Similar results were obtained previously with corn NR m R N A (Comm6re et al. 1986 ).

The cloning vector 2gtll (Young and Davies 1983) was chosen for the construction of cDNA libraries since it was expected that the frequencies of NR cDNAs would be very low, and large libraries required to be screened, cDNA libraries were first built from unfractionated mRNA preparations. The initial efficiencies of cloning and the proportions of true recombinants were low (< 10000 clones/ng cDNA) when the Gubler and Hoffmann (1983) procedure was followed for the synthesis of cDNAs. As described previously, the efficiency of cloning was increased by including a mild $1 nuclease treatment between the healing and methylation steps of cDNA synthesis, and allowed the routine obtainment of cloning efficiencies higher than l0 s recombinants/ ng cDNA.

Since the mRNA encoding for NR was expected to be of a size larger than 3 kb, m R N A preparations were fractionated on a sucrose gradient, and molecules with a sedimentation constant in the 23 32 S range were collected. These mRNAs were further fractionated on a methylmercuryhydroxide gel and the mRNAs migrating in the 3-4 kb area were used to build a cDNA library (4 x l0 s clones, 70% recombinants). These two steps of purification represented a five-to tenfold enrichment of NR mRNA.

Although it was specific for tobacco NR in tobacco leaf extracts, the antibody eMNR recognized several E. coli proteins in Western blot experiments (see Fig. 3 ). This resulted in high background levels when libraries were immunologically screened. The antibody c~MNR was therefore incubated with large amounts of E. coli crude protein extracts, and confirmed to be still able to detect tobacco NR at the nanogram level. Among a total of 40 clones initially selected by screening 240000 recombinant plaques, 11 clones expressing reproducibly a signal upon rescreening were further purified to homogeneity. This eventually required, for the highly expressing clone 13-29, the plating, picking and amplification of isolated plaques in the absence of IPTG induction to prevent the appearance of non-expressing clones during plaque growth (results not shown).

Purified clones were tested for the presence of DNA inserts by EcoRI digestion. Seven clones (13-18, 13-20, 13-27, 13-29, 13-33, 13-34, 13-36) contained EcoRI inserts of similar size (1.6 kb). They gave consistently signals of differing intensities upon immunological detection. A strong signal was detected for clone 13-29 and to a lesser extent for clone 13-28, whereas other clones gave low signal intensities. The 1.6 kb EcoRI fragment from clone 13-36 was subcloned in pUC9 (Vieira and Messing 1982) and used as a probe for the study of insert homologies among the ten other recombinants. Seven clones cross-hybridized with this probe. Six of them (13 18, 13 20, 13-27, 13-29, 13-33, 13-34) had been classified as containing an insert of similar size, suggesting that these inserts were possibly identical. Clone 13 28 contained a 2 kb EcoRI insert. The structure and orientation of cDNAs in the cloning site of the vector were studied using the restriction enzymes Kpnl and Sstl.

It was found that the size of inserts in some of the recombinants was larger than 1.6 kb and that the 1.6 kb EcoRI "l---s, 

Lysogens of phages 13-29 and 13-36 carrying inserts in opposite orientations were constructed in strain RY1089. These lysogens were tested for the expression of a hybrid protein by Western blot analysis (Fig. 3) . Using the polyclonal antibody ~MNR, a protein of an approximate molecular weight of 175 kDa was found in the extracts of a lysogen for phage 13-29 induced with IPTG, but not in the non-induced controls. This hybrid protein was also detected after Coomassie blue staining of the corresponding SDS-PAGE fractionated extracts (Fig. 3) . No hybrid protein was detected in a lysogen for phage 13-36. These results are in agreement with the insert of clone 13-29 being integrated in frame with the coding sequence of/~-galactosidase and encoding a polypeptide chain of an approximate size of 60 kDa, which is in reasonable agreement with the size of the corresponding cDNA insert. The isolation by immunological screening of clone 13-36 carrying an insert oriented opposite to the/~-galactosidase gene, and the study of the corresponding lysogen, confirm the weak expression of inserted sequences under the control of another phage promoter in 2gtl 1, as previously discussed (Lapeyre 1985) . The absence of detection of hybrid proteins in Western blots of extracts from a lysogen of the recombinant clone 13-36 may be attributed to the low level of expression of this hybrid protein. On the basis of these results, phage lambda 13-29 was chosen for further studies and the corresponding insert was subcloned.

Although the antiserum c~MNR used for the screening procedure was reasonably specific towards tobacco NR, this Fig. 4 . Inhibition of the catalytic activity of tobacco nitrate reductase (NR) by immunoglobulins immunoselected by adsorption on plaques formed by phages 2gtll and 13 29 grown in the presence of IPTG. Immunoglobulins were added in the enzyme incubation mixture at the indicated dilution, based on the initial amount of antiserum used for adsorption on phage plaques. Immune antiserum c~MNR adsorbed on 2gtll plaques (e) or on recombinant clone 13-29 plaques (D). Non-immune antiserum ~So adsorbed on plaques from clone 13-29 (o). A small but significant stimulation of NR activity was induced by adding non-immune antiserum (~So) in the preparation of nitrate reductase. This is attributed to a stabilizing effect of serum proteins on the enzyme. The specific activity of the extract was 0.4 pM NOz/pg protein per minute strain RY 1090 lawns were infected by phages 2 g t l l and 13-29. After induction with I P T G the synthesized proteins were transferred on nitrocellulose filters. Control (eSo) and ~M N R polyclonal antibodies were incubated with the filters and the corresponding immunoselected immunoglobulins were eluted from the filters at low pH. These selected antibodies were then assayed for their ability to inhibit the nitrate reducing activity of tobacco NR. As shown in Fig. 4 , a 40% inhibition of tobacco N R activity was obtained with a 1/200 dilution of the selected immunoglobulins, in reasonable agreement with the inhibitory activity of the c~MNR serum (70% inhibition with a 1/200 dilution of this serum). Proteins expressed by clone 13-29 could bind specifically antibodies inhibitory to the tobacco N R in the c~MNR serum. This suggests also that the amount of hybrid protein available for imnmnoselection was in excess of the amount of immunoglobulins to be adsorbed in the serum. Indeed we found that c~MNR antiserum used for the confirmation of immunologically positive clones could not be used for further screening of new recombinants, probably due to a depletion of immunoglobulins directed against NR.

Three N R monoclonal antibodies raised against N. plumbaginifolia and corn N R and a control monoclonal antibody directed against a swine coronavirus (Table 1) were used to characterize further the hybrid proteins. Among the four monoclonal antibodies tested, clones 96(9)25 and 24(48)NP19 specifically recognized antigenic determinants from an extract of a lysogen for clone 13-29, made in the bacterial strain RY1089. No significant recognition was detected for clone 7(I13)NP15 and GET. The monoclonal antibody 96(9)25 has a higher affinity for tobacco N R than the monoclonal antibody 24(48)NP19. The recognition of the hybrid protein by these two monoclonal antibodies suggests a similar situation. At least two different epitopes can be detected on the hybrid protein that are also recognized on tobacco NR. The absence of the epitope recognized by the monoclonal antibody 7(113)NP15 shows also, as expected, that tobacco N R and the hybrid protein are antigenically different, due to a lack of part of the N R coding sequence in the cloned cDNA.

The nucleotide sequence of clone 13 29 is shown in Fig. 5 .

This partial c D N A is 1662 bp long, including the EcoRl sites, and does not contain polyadenylation sequences. Although the c D N A was cloned in the vector after methylation and linker addition, no linker sequences are found at the ends of the cDNA. This strongly suggests that the methylation step had not worked properly and that an internal Fig. 6 46), As shown in Figs. 5 and 6, two stretches of high homology were found. From position 360 to 435 the tobacco nitrate reductase showed sequence homology ranging from 35%-50% with proteins from the cytochrome b5 superfamily (Guiard and Lederer 1979a) . L~ and Lederer (1983) have isolated, by chymotrypsin proteolysis, the heine-binding domain from Neurospora crassa nitrate reductase and shown its homology with the cytochrome b5 family. Homology between the tobacco nitrate reductase amino acid sequence and the fungal nitrate reductase heine-binding domain is also found, as expected.

From residues 476 to 552 the tobacco nitrate reductase aminoacid sequence showed 57% identity with the N-terminal sequence of the human cytochrome b5 reductase (Yubisui et al. 1984 ). This FAD flavoenzyme is known to form a binary complex with cytochrome b5, performing electron transport from NADH for various redox physiological processes. No definite homologies were found with other polypeptides in the sequence data banks or with selected sequences, such as the partial coding sequence for the alpha subunit of E. coli respiratory nitrate reductase.

Messenger RNAs were extracted from tobacco, Nicotiana plumbaginifolia and corn leaves of plants grown on a nitrate-containing medium. Northern blots of mRNAs fractionated on formaldehyde-agarose gels were probed with Fig. 6 . Alignment of the predicted aminoacid sequence of tobacco nitrate reductase (NR) in the heme-binding domain (B1 to B120) and FAD/NADH domain (R1 to R90) with proteins of the cytochrome b5 superfamily and with human erythrocyte eytochrome b5 reductase (RDHUB5). Identity between any sequence and tobacco NR is shown by boxed residues. Residues of the heme-binding domain have been numbered as conventional for bovine cytochrome b5 (CBBO5, Ozols and Strittmatter 1969) with the two heme-liganding histidines numbered His-39 and His-63 stars. The FAD/NADH domain has been numbered from the N-terminal residue of cytochrome b5 reductase. CBRT5M, rat mitochondrial cytochrome b5 (Lederer et al. 1983 ); CBRT5, rat microsomal cytochrome b5 (Ozols and Heinemann 1982) ; CBHO5, horse microsomal eytochrome b5 (Ozols et al. 1976 ); CBRB5, rabbit microsomal cytochrome b5 (Tsugita et al. 1970 ); CBHU5, human microsomal cytoehrome b5 (Nobrega and Ozols 1971) ; CBCH5, chicken microsomal cytochrome b5 (Nobrega and Ozols 197J) ; NRTOB, tobacco nitrate reductase (this paper); NRNEU, Neurospora crassa nitrate reductase (L~ and Lederer 1983, and unpublished data) ; FLAVB2, yeast flavocytochrome b2 ~Lederer et al. 1985); SULFOX0 chicken sulfite oxidase (Guiard and Lederer 1979 b) ; CONSEN, consensus sequence the 1.6 kb c D N A insert (Fig. 7 ). An hybridization signal was detected after 3 days of exposure in tobacco and N. plumbaginifolia m R N A preparations. No signal was detected even after prolonged exposure in the case of corn m R N A s . Hybridation of the same blots with a r D N A wheat repeat unit showed that m R N A s homologous to the 1.6 kb insert were migrating slightly faster than the 28 S ribosomal R N A . A size of 3.5-3.6 kb was assigned to the m R N A when migrations were performed in methylmercuryhydroxide gels under fully denatured conditions in the presence of single-stranded D N A molecular weight markers. Fig. 7 . Northern blot analysis of mRNA sequences hybridizing with the 1.6 kb EcoRI insert from phage 2 13-29. Total RNA was extracted from the leaves of tobacco (7), N. plumbaginifolia (P) and corn (M) plantlets grown on a nitrate-containing medium. Poly(A) + mRNAs were purified from these samples. Lanes 1 and 10, total corn RNA (5 gg) and total tobacco RNA (2 gg) stained with ethidium bromide; lanes 2 and 9, the same samples probed on Northern blots with the wheat rDNA fragment from plasmid pAT7]; lanes 3-8, Northern blots of poly(A) ÷ mRNAs probed with the 1.6 kb EcoRI fragment from phage 13-36 subcloned in pUC9 (sp. act. 109 cpm/l~g, 1 week autoradiography). Lanes 3 and 4, 1 and 2 pg corn mRNA; lanes 5 and 6, 1 and 2 pg tobacco mRNA; lanes 7 and 8, 2 and 5 gg N. plumbaginifolia mRNA

The genomic complexity of the sequences homologous to the 1.6 kb EcoRI fragment was studied using DNA extracted from tobacco. The results of a Southern blot performed under high stringency conditions are presented in Fig. 8 . Signals corresponding to low copy numbers of the corresponding sequences were detected in the digests, as indicated by a comparison with hybridization signals obtained by probing with a rDNA wheat repeat unit. When genomic DNA was cut by EcoRI a signal was detected in the positions of a 3.4 kb and a 5.0 kb fragment and a weak hybridization was also observed corresponding to a 2.9 kb fragment. Two functional structural genes are assumed to be present in the tobacco genome (Mtiller 1983) . Taking into account that the 1.6 kb EcoRI fragment is internal to the cDNA, this suggests that intervening sequences must be present in the corresponding area of at least one of the two structural genes.

The present report provides evidence for the cloning of cDNAs encoding a part of a protein antigenically related to tobacco nitrate reductase. A family of clones was isolated by immunological screening of a cDNA library built in Fig. 8 . Southern blot analysis of tobacco genomic sequences homologous to the 1.6 kb EeoRI insert from phage 2 13-29. Lane 1, molecular weight markers; lane 2, 10 ttg of tobacco genomic DNA digested with EeoRI, probed with the 1.6 kb EcoRI fragment (sp. acty. 10 9 cpm/pg, 3 days autoradiography); lane 3, 1 gg of the same DNA cut by EcoRI probed with the wheat rDNA sequences from plasmid pTA71 (sp. act. 108 cpm/gg, I day autoradiography) the expression vector 2gtll. These recombinant clones encode polypeptide fragments sharing homologies with the active site of the nitrate reductase from tobacco. The cloned fragments hybridize with a 3.5 kb mRNA species and to low copy number genomic sequences. These cloned sequences have been confirmed to correspond to fragments of the nitrate reductase mRNA by sequence data analysis. Plant and fungal nitrate reductases contain a cytochrome b557 domain (Redinbaugh and Campbell 1985) . As shown by L~ and Lederer (1983) by amino acid sequencing, the heme-binding domain of Neurospora crassa nitrate reductase is a member of the cytochrome b5 superfamily. Our results show that the tobacco enzyme also shows a significant homology with different members of the superfamily. The cytochrome b5 homology entirely covered the berne-binding domain of the various proteins: we thus propose that this sequence defines the heine-binding domain of tobacco nitrate reductase. All the 13 conserved residues of the superfamily (Lederer et al. 1983 ) are present in our tobacco NR sequence, which appears to be closer to chicken cytochrome b5 than to any other member of the superfamily. The relative dissimilarity between the plant and fungal nitrate reductases indicates a divergent evolution for that enzyme. Since X-ray determinations of the spatial structure of beef microsomal cytochrome b5 (Mathews et al. 197J ) and flavocytochrome b2 (Xia et al. 1987 ) have been performed, it is hoped from the observed homology that the three dimensional conformations will be predictable.

A significant homology of the coding sequence of the cloned cDNA with human flavoprotein cytochrome b5 re-ductase was also detected. This protein is known to catalyse the reduction of cytochrome b5, using NADH as an electron donor and FAD as a redox intermediate. This may be functionally compared to the catalytic reaction of nitrate reduction involving NADH, FAD, and cytochrome b5 as cofactors. The suggestion is that this second homologous sequence is the N-terminal moiety of' the NADH/FAD domain of tobacco nitrate reductase. Nitrate reductase would therefore appear to be a protein in which the heme-binding site and the reducing flavoprotein domain lie side by side in the polypeptide chain. To the best of our knowledge no sequences are available for the rnolybdopterin-binding domain from redox enzymes, which would be helpful to further analyse the functional structure of assimilatory nitrate reductases.

Nitrate reductase represents a small proportion (2-5/10000) of the soluble proteins. Cell-free translation data and Northern experiments suggest that the corresponding mRNA accounts for approximately 1/10000 of total poly(A) + mRNA. The frequency of immunologically positive clones isolated from the cDNA library is 1/20000. This value seems low since this library was made from purified mRNAs and should contain one NR cDNA among 1000-2000 recombinants. When the library was screened by hybridization with the 1.6 kb EcoRI fragment it was found that indeed the proportion of inserts homologous to the probe was larger (1/5000). We conclude that most of these cDNAs were too small to express an immunogenic protein. The library was screened with a polyclonal antibody raised against the corn NR. The preferential isolation of relatively similar cDNAs may reflect the limited number of epitopes that can be recognized on tobacco NR by an antiserum raised against a heterologous NR. Previous work has shown that among six different epitopes of the corn NR identified by a family of monoclonal antibodies, only one epitope, a conformational epitope involved in the catalytic activity of the enzyme, was found on the tobacco enzyme (Ch6rel et al. 1985) . This epitope is apparently also present on the hybrid protein expressed by recombinant clone 13-29. It will be of interest to see whether the hybrid protein is able to catalyze some of the oxidoreductive steps involved in nitrate reduction by nitrate reductase.

The NR structural gene is one of the few examples of plant genes for which it is possible to select for or against its expression. Furthermore the impairment in nitrate assimilatory functions still allows the regeneration of fertile plants. Numerous NR mutants have been isolated in our laboratory and are currently being clharacterized. A probe corresponding to part of the NR apoenzyme coding sequence will be useful to analyse these mutants which belong to seven different complementation groups (Mfiller AJ and Gabard J, unpublished results; Gabard et al. 1987) , and to study the NR structural gene. Genomic clones homologous to this probe have been isolated and are under study.

Specificity of oligo (dT)-cellulose chromatography in the isolation of polyadenylated RNA

Ferric-citrate reductase activity of nitrate reductase and its role in iron assimilation by plants

Purification of barley nitrate reductase and demonstration of nicked subunits

Monoclonal antibodies identify multiple epitopes on maize leaf nitrate reductase

Immunological comparisons of nitrate reductases of different plant species using monoclonal antibodies

Isolation of biologically active RNA from sources enriched in ribonucleases

In vitro translation of nitrate reductase messenger RNA from maize and tobacco and detection with an antibody directed against the enzyme of maize

A plant DNA miniprep: version II

The pEMBL family of single-stranded vectors

Liposomemediated transformation of tobacco mesophyll protoplasts by an Escherichia coli plasmid

Isolation and characterization of Nicotiana plumbaginifolia nitrate-reductase deficient mutants: genetic and biochemical analysis of the nia complementation group

Cloning and characterization of ribosomal RNA genes from wheat and barley

A simple and very efficient method for generating cDNA libraries

The "cytochrome b5 fold": structure of a novel protein superfamily

Amino acid sequence of the b5-1ike heine-binding domain from chicken sulfite oxidase

Unidirectional digestion with exonuclease III creates targeted breakpoints for DNA sequencing

Assimilatory nitrate and nitrate reduction

Quaternary structure of assimilatory NADH : nitrate reductase from Chlorella

Construction and screening cDNA libraries in lambda gtl 0 and lambda gtl 1. In: Glover DM (ed) DNA cloning: A practical approach

Current progress towards an understanding of the genetics and molecular biology of nitrate reductase in higher plants

On the presence of a heme-binding domain homologous to cytochrome b5 in Neurospora crassa assimilatory nitrate reductase

Two homologous cytochrome b5 in a single cell

Complete amino acid sequence of flavocytochrome b2 from baker's yeast

The structure of calf liver cytochrome b5 at 2.8 A resolution

Experiments in Molecular Genetics

Genetic analysis of nitrate reductase-deficient tobacco plants regenerated from mutant cells. Evidence for duplicate structural genes

Amino acid sequences of tryptic peptides of cytochromes b5 from microsomes of human, monkey, porcine and chicken liver

Chemical structure of rat liver cytochrome b5. Isolation of peptides by high-pressure liquid chromatography

Correction of the amino acid sequence of calf liver microsomal cytochrome b5

Proteolytic cleavage of horse liver cytochrome b5. Primary structure of the heine-containing moiety

Escheriehia coli nitrate reductase subunit A: Its role on the catalytic site and evidence for its modification

Quaternary structure and composition of squash NADH nitrate reductase

Deux approches nouvelles ~i l'6tude des ARN poly-m4rases de Saccharomyces cerevisae: la formation d'enzymes hybrides et l'isolement des gbnes de structure

DNA sequencing with chain termination inhibitors

Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose

Comparative study of the primary structures of cytochrome b5 from four species

the pUC plasmid, an M13 mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers

Three-dimensional structure of flavocytochrome b2 from baker's yeast at 3.0-A resolution

Amino acid sequence of NADH-cytochrome b5 reductase of human erythrocytes

Improved M13 phage cloning vectors and host strains: Nucleotide sequences of the M13 rap18 and pUC19 vectors

Efficient isolation of genes by using antibody probes

No striking homologies were found between the partial protein sequence of nitrate reductase and this sequence

Acknowledgements. We would like to thank Nicole Guiso and Roxane Predeleanu for providing us with iodinated protein A and Christiane Henry for typing the manuscript. We would like also to thank Bruno Lapeyre, Michel Riva, Jean de Gunzburg, and Charles Auffray for their encouragement and helpful discussions during the development of this work. We acknowledge Frangoise Vedele and Claude Mugnier for their patient efforts and continuous help in the sequence and data analysis, and Kim H6 Diap L~ and Florence Lederer for communication unpublished data on N. crassa nitrate reductase. Computer facilities and grants were ob-tained from the french "Minist6re de la Recherche et de la Technologic" (programme mobilisateur Essor des Biotechnologies).