key: cord-0005022-slqje8cq authors: Orlik-Eisel, Gabriele; Lutz, Frieder; Henschen, Agnes; Eisel, Ulrich; Struckmeier, Martin; Kräuter, Josef; Niemann, Heiner title: The cytotoxin of Pseudomonas aeruginosa: Cytotoxicity requires proteolytic activation date: 1990 journal: Arch Microbiol DOI: 10.1007/bf00245265 sha: 70b66b198de92af54ee901f6ad83b9e0a61af295 doc_id: 5022 cord_uid: slqje8cq The primary structure of a cytotoxin from Pseudomonas aeruginosa was determined by sequencing of the structural gene. The cytotoxin (31,700 Mr) lacks an N-terminal signal sequence for bacterial secretion but contains a pentapeptide consensus sequence commonly found in prokaryotic proteins which function in a TonB-dependent manner. The cytotoxin gene has a [G+C]-content of 53.8% which is considerably lower than generally observed for genes from Pseudomonas aeruginosa. The cytotoxin gene was exclusively detected in strain 158 but not in three other clinical isolates, as determined by Southern and Northern hybridization. The latter technique revealed that the toxin is translated from monocistronic mRNA. The promoter of the cytotoxin is inactive in Escherichia coli. Upon site-directed modification of the 5′-noncoding region by the polymerase chain reaction the gene was expressed under control of the trcpromoter. The gene product obtained in Escherichia coli was nontoxic. Toxicity was induced by subsequent treatment with trypsin. [(35)S]methionine-labeled cytotoxin with high specific radioactivity was obtained by in vitro transcription/translation. Like [(125)I] labeled material from Pseudomonas aeruginosa this polypeptide bound to membrane preparations from Ehrlich ascites cells, as evidenced by sedimentation through a sucrose gradient at neutral pH. Offprint requests to: H. Nlemann periplasm of the bacterium (Kluftinger et al. 1989 ) and becomes liberated by autolysis rather than by secretion (Scharmann 1976) . Isolated from bacterial autolysates, the cytotoxin has been characterized as a protein of 25,000 to 29,000 Mr which acts primarily on the plasma membranes of mammalian cells (Baltch et al. 1987; Kluftinger et al. 1989; Lutz 1979 ) by binding to a high affinity binding sites (Lutz 1986) . As a consequence, pores of about 2 nm diameter are formed resulting in a breakdown of the cellular gradient for low molecular substances. The role of the cytotoxin in the manifestation of the Pseudomonas aeruginosa infection, however, has not been thoroughly investigated. In this paper, we present the sequence of the cytotoxin. We show that a posttranslational activation step involving proteolytic removal of a 3,000 Mr peptide from the carboxy-terminal end takes place during or after autolysis. Materials. Enzymes were purchased from Boehringer (Mannhelm, FRG). 7[3ZP]ATP, ~[35S]dCTP, and L-[35S]methionine were from Amersham (Braunschweig, FRG). Rabbit reticulocyte lysate was obtained from Amersham or Promega (Heidelberg, FRG), nucleotides and ribonuclease lnhibator were from Pharmacia (Freiburg, FRG), and anti-rabbit IgG from Dako (Copenhagen, Denmark). Diethylpyrocarbonate was from Sigma (Mfinchen, FRG). Bacterial stratus and plasmids. Pseudomonas aerugmosa strain 158 (0 : 6; H : ao, a2, a3, pyocin : 38e) was a clinical isolate from bovine mastitis milk. Pseudomonas aeruginosa strains 032 and 037 were isolated as swab samples from horse vagina or dog ear. The 054 strain was isolated from feces of a septicemic cow. All strains were propagated at 30~ on TSA/TSB (Difco. Detroit). Escherichia coli strains HB101 and JM101 were grown at 37~ in 2YT or M9minimal medium. Plasmids pUC18/19 (Messing and Vieira 1982) , M13mp18/19 (Norrander et al. 1983 ) and pRN653A,B,C (H. Niemann, A. Smid, M. Rosing, and E. Amann, unpublished) were used for establishing DNA-libraries, for DNA-sequenclng, and for combined in vitro transcription/translation, respectively. For tightly regulated expression of the cytotoxin gene in Escherichia coli the IPTG-inducible vector pTrc99a (Amann et al. 1988 ) was used. -500 1 500 1~00 | l l , , j i l i i p i l l l (Lutz 1979) . The N-terminal amino acid sequence of the purified cytotoxin was determined by Edman degradation (Edman and Henschen 1975) . Hybridization conditions with synthetic oligonucleotides. Chromosomal DNA was isolated from logarithmically growing cultures as described (Meade et al. 1982) . The mixture of heptadecamere oligonucleotides [ATGAA(C/T)GA(G/A)AT(C/T/A)-GA(C/T)AC] was synthesized with an Applied Biosystems model 380 A DNA synthesizer, 5'-labeled using ~[32p]ATP and T4-polynucleotide kinase and used for hybridization according to Wallace et al. (1981) . Cloning procedures and DNA modifications. DNA modifications were performed according to standard protocols (Maniatis et al. 1982) . The coding sequence for the N-terminus was identified with the 5'-labeled oligonucleotides on a 4 kb KpnI-and a 1.5 kb PstIfragment. In addition, a signal was obtained with a 500 bp Sau3AI/ PstI-fragment. The Sau3A/PstI-fragment was isolated from the gel and cloned under L3-B1 biosafety containment facilities into BamHI/PstI-digested pUC19. The insert was isolated by digestion with PstI and SmaI, nick-translated and used to screen PstI/HincIIand PstI/EcoRV-libraries. 12 overlapping M13 clones, together spanning the entire toxin gene (Fig 1) , were sequenced on both strands to establish the complete structure employing the chain termination method (Sanger et al. 1977) . Computer assisted analyses were performed with the PC-Gene program purchased from Genofit (Geneva, Switzerland). annealing of the oligonucleotides was performed at 45 ~ C for 2 min, the polymerization reaction was at 72~ for 3 min. Through this procedure a singular FokI-site was introduced Into the 5'-noncoding region. Subsequent cleavage with FokI generated 5'-CATG protruding ends that allowed cloning of the amplified gene into the NcoI-site of pTrc99a (Amann et al. 1988 ) to yield pSN3 as detailed in Fig. 2 . Expression of the cytotoxin in Escherichia coli. 10 ml precultures of Escherichia coli strain HBI01 harboring pSN3 were prepared in LBmedium and used to inoculate 90 ml of TSB and 100 mg ampicillin per 1. Cells were grown up to an optical density of OD660 nm of 1.0 and synthesis of the cytotoxin was induced by the addition of 1 ml of 0.5 mM IPTG. The incubation was continued for another 2 h at 37~ when cells were harvested by centrifugation at 16,000 xg, washed once with phosphate buffered saline (PBS), pH 7.4, and resuspended in 1 ml of PBS. The cells were incubated for 12 h at 37 ~ C (autolysis step). Insoluble material was removed by centrifugation (30 rain at 13,000xg) and the supernatant was stored at -20~ Toxicity assays were performed according to Gladstone and van Heyningen (1957) . Proteolytic activation of the cytotoxin was achieved by the addition of 2.7% (w/w protein) TPCK-trypsin (40 U/mg, Boehringer) in lysates containing 3 mM (final concentration) of CaCiz. After 2 h at 37~ reactions were stopped by the addition of 5-fold molar excess of trypsin inhibitor from soybean in 10 mM EDTA. In vitro transcription/translatlons. The coding region of the cytotoxin gene was cloned from the EcoRV site ( Fig. 3) on the 3'-PstI site into SmaI-PstI digested pRN653C to yield pOE65EP33. The sequence of the Y-recombination site was verified by direct sequencing using the SP6 sequencing primer. Plasmid DNA was purified by two consecutive centrifugations on CsCl-density gradients. Transcriptions with SP6-polymerase and translations in rabbit reticulacyte lysate were performed as described previously (Mayer et al. 1988 ). was incubated for 2 h at 4~ and subsequently for 30 min at 30~ with plasma membrane preparations from Ehrlich mouse aseites cells (Kflberg and Christensen 1979) . To assess membrane association, the incubation mixture was then placed on a sucrose cushion prepared in buffer that was either at pH 7.3 or at pH 11.0, as described in detail previously (Mayer et al. 1988 ). The pellet and the supernatant fraction were analyzed by SDS-PAGE. Gel electrophorests and #nmunoprecipitation. SDS-PAGE and processing with DMSO-PPO for autoradiography were performed as described (Niemann and Klenk 1980) . Rabbit antibodies against cytotoxin were purified by binding the antibodies to nitocellulose carrying the purified toxin according to Burke et al. (1982) . Protein A bearing Staphylococcus aureus, strain cowan I, was used to bind immune complexes. To establish the p r i m a r y sequence of the cytotoxin from Pseudomonas aeruginosa, we have cloned a n d sequenced arrow indicates the translation start codon. Note that a mutation of the AAC-codon would lead to a mutation of Asn 2 and thus to a posttranslational removal of the N-terminal Met residue. Taq-Polymerase from Boehrlnger was used to introduce a singular FokIsite m 45 cycles. The products were digested with Xmnl and FokI, purified by agarose gel electrophoresis and cloned into the Sinai/ NcoI-dlgested pTrc99a (Amann et al. 1988) domains that could be involved in catalyzing membrane integration. In addition, no signal sequence for secretion and no a-helical transmembrane domains were detected using the programs of Roa and Argos (1986) or Klein et al. (1985) , respectively. A Shine-Dalgarno consensus sequence (AGGA) was found 12 nucleotides upstream from the translation initiator ATG-codon. The [G + C]-content of the coding region (53.8%) is significantly lower than that reported for chromosomally integrated genes of Pseudomonas aeruginosa (West and Iglewski 1988) , indicating that the gene could originally stem from a different organism. This hypothesis is further supported by our finding that the cytotoxin gene is absent in three other clinical isolates of Pseudomonas aeruginosa as evidenced by Southern analyses (Fig. 4A, B) . Even after 45 PCR-cycles (using oligonucleotides binding immediately upstream and downstream from the coding region and 20 ng of chromosomal DNA) these other strains failed to produce a signal in Southern blotting. In addition, Northern blot analyses of R N A from the individual strains also indicated the absence of cytotoxin-specific transcripts (data not shown). The open reading frame was followed by two inverted repeat structures indicated by divergent arrows in Fig. 3 . The free energy values (Tinoco et al. 1973) of these stem-loop structures, -9 2 . 0 5 K J/tool and -7 9 . 5 KJ/mol, suggest that they could function as transcription-termination signals. Northern blot analyses of R N A from strain 158 revealed that the cytotoxin-specific m R N A had a size of about 1100 nucleotides (data not shown). Taken together, these data support the conclusion that the cytotoxin gene is transcribed into monocistronic mRNA. To test the properties of the cloned gene product, we expressed the gene in vitro or in Eseherichia coli and compared the properties of the products with those of unlabeled or [125I] labeled cytotoxin, as isolated from Pseudomonas aeruginosa. For combined in vitro transcription translations, we cloned the coding region from the EcoRV site (cleavage after position 9 of the coding region) to the PstI-site into a modified pSP65-vector system providing an ATG codon for initiation of translation. The resulting construct, pOE65EP33, yielded m R N A encoding a polypeptide of 30,000 Mr which lacked the three N-terminal amino acids of the toxin and contained Translation of the RNA in rabbit reticulocyte lysate produced a major polypeptide of 30,000 Mr (Fig. 5A , lane 1). This molecular species had an electrophoretic mobility that was indistinguishable from material isolated from intact Pseudomonas aeruginosa cells (compare lanes 1 and 2). As demonstrated by Western blotting, the 30,000 intracellular form of the cytotoxin (lane 3) migrated clearly slower in SDS-PAGE than the 28,000 material that was isolated from autolysates (lane 4). Puls-chase experiments of [3 S S]methionine labeled sister cultures of Pseudomonas aeruginosa did not reveal a conversion of the 30,000 species into the 28,000 species (data not shown) indicating that the putative processing step had to occur during autolysis of the bacteria. Expression of the cytotoxin gene in Escherichia coli was inducible with IPTG (compare lanes 1 and 2 in Fig. 5 B) , again yielding material that migrated like the non-processed form of the cytotoxin in SDS-PAGE (compare with lane 4). This material was clearly nontoxic in the granulocyte lysis assay (Fig. 6A) . As shown in lanes 2 and 3 of Fig. 5B , treatment of Escherichia coli lysates with tryt~sin converted the 30,000 species into two smaller species of 28,000 and 26,000 (lane 3). Concomitantly a rapid increase in toxicity was observed (Fig. 6 B) , indicating that the removal of the To see whether this proteolytic processing step altered the binding properties of the cytotoxin to cellular receptors, we performed binding studies of the in vitro synthesized cytotoxin derivative and compared it with iodinated cytotoxin as derived from Pseudomonas aeruginosa autolysates. Binding of the cytotoxin to membrane preparations was assessed by co-sedimentation of the radiolabeled cytotoxin with the membranes through a sucrose cushion of neutral pH. The results are summarized in Fig. 7 . No difference was detected in the binding properties of the in vitro synthesized full-size cytotoxin and the processed cytotoxin. In both instances binding was reversible by the addition of a 100-fold excess of unlabeled cytotoxin (data not shown). However, binding apparently involved only attachment to peripheral binding sites, since a significant amount of labeled material Fig. 6 . Toxic]ty assays of Escherlchta colt derived cytotoxin on granulocytes before (a) and after (b) treatment with trypsin eluted from the membranes, when the pellet fraction was resuspended and re-sedimented through the sucrose cushion (compare lanes 9 and 10 of Fig. 7) . Furthermore, no co-sedimentation of the labeled material was observed when the sucrose was made up in buffer of pH 11.0, again indicating that the bound material was not converted into an intrinsic membrane protein. It is clear, however, that such observations have to be confirmed by experiments involving binding to intact cells. We have established the sequence of a pore-forming cytotoxin from Pseudomonas aeruginosa by determining the amino-terminal amino acid sequence of the purified protein and sequencing of the structural gene, as identified by a pool of synthetic oligonucleotides. The cytotoxin sequence did not reveal a significant sequence similarity with any other known protein. It is important to note that no proteolytic processing of the N-terminus of the protein occurs during or after bacterial autolysis. As generally observed with procaryotic polypeptide carrying an asparagine residue in position 2, the methionine residue is retained in the mature toxin molecule (Ben-Bassat and Bauer 1987) . In agreement with a previous report 11) were incubated with plasma membrane preparations and sedimented together with the membranes through a sucrose cushion at the pH indicated (Mayer et al. 1988 ). The pellet fractions (lanes 2, 4, 7, 9, and 12) and TCA-precipitable material from the supernatant fractions (lanes 3.5, 8, 10, and 13) were analysed by SDS-PAGE. Samples in lanes 4 and 5, and 9 and 10 are derived from a second centrifugation step (Scharmann 1976) indicating that the cytotoxin was released from the bacteria only after several days of growth, the molecule lacks a secretory signal. Within the N-terminal domain, a remarkable homology to a pentapeptide consensus sequence (TonB-box), commonly found in outer membrane receptor proteins of the Escherichia coli iron transport system, was detected. As yet, the TonB-box was found exclusively in all proteins that function in a TonB-dependent manner . Interestingly, this group of proteins contains also some colicins known to kill closely related bacteria by pore formation. Uptake of such colicins by the target cell occurs in a receptor mediated and TonB-dependent process . Recent modifications of the TonB-box from the FhuA receptor by site-directed mutagenesis (Sch6ffier and Braun 1989) have shown that a replacement of the Va111 residue by aspartic acid only weakened the colicin M sensitivity of the Escherichia coli strain indicating that the interaction between the FhuA receptor and the TonB protein was not completely abolished. At present, we do not know whether the cytotoxin serves a colicin-like function for Pseudomonas aeruginosa, The molecular weight of the cytotoxin purified from bacterial autolysates was 28,000 as determined by SDS-PAGE. This material migrated clearly faster than the 30,000 Mr species obtained by in vitro transcription/ translation or by expression in Escherichia coli. Although the in vitro synthesized material bound specifically to membrane preparations from Ehrlich ascites cells, exhibiting properties indistinguishable from the mature [125I] labeled cytotoxim this non-processed form was nontoxic in the granulocyte lysis assay. Cytotoxicity clearly required proteolytical processing which in autolysates was mediated by endogenous proteases. Trypsintreatment of Escherichia coli lysates also restored cytotoxicity. Such processing could involve only C-terminal sequences, since identical N-termini were determined by Edman degradation and by DNA-sequencing. The mechanism by which pore formation through the cytotoxin is induced is far from being understood at the molecular level. We show here that binding to peripheral acceptor sites does not require proteolytic processing and Coulton et al. 1986 Sauer et al. 1987 Heller and Kadner/985 Pressler et al. 1988 Lundrigan and Kadner 1986 Griggs et al. 1987 Krone et al. 1985 K6ck et al. 1987 Schramm et al. 1987 Mankovich et al. 1984 also does not involve the N -t e r m i n a l sequences, since the in vitro synthesized c y t o t o x i n had similar binding p r o p e m e s like the m a t u r e molecule. W i t h the c y t o t o x i n gene at h a n d and the d e v e l o p m e n t o f various deletion m u t a n t s t h e r e o f further studies on the pore f o r m a t i o n process can n o w be u n d e r t a k e n . Appendix. While this manuscript was in preparation. Hayashi et al. (1989) published their data on the nucleotide sequence and the expression of the cytotoxin gene. The authors also came to the conclusion that cytotoxlcity was posttranslationally generated by proteolytic cleavage. Tightly regulated tac promoter vectors for the expression of unfused and fused proteins in Escherichia coll Production of cytotoxin by clinical strains of Pseudomonas aeruginosa Amino-terminal processing of proteins ed) Molecular basis of viral and microbial pathogenesis A monoclonal antibody against a 135-K Golgi membrane glycoprotein Protein fusions of fi-galactosidase to the ferrichromeiron receptor of Escherichm coli K-12 Sequence determinatmn". In. Needleman SB (ed) Protein sequence determination Heyningen WE van 0957) Staphylococcal leucocidins Cloning and promoter identification of the iron-regulated cir gene of Escherichla coll Pseudomonas aerugmosa cytotoxin: The nucleotide sequence of the gene and the mechanism of activation of the protein Nucleotide sequence of the gene for the vitamin B12 receptor protein in the outer membrane of Escherichia coll Electron-transferring enzymes in the plasma membrane of Ehrlich ascites tumor cell The detection and classification of membrane-spanning proteins Pseudomonas aeruginosa cytotoxin: Periplasmic localization and inhibition of macrophages Primary structure of colicin M, an inhibitor of murein biosynthesis Characterization of the pCol V-K30 encoded cloacin DFI 3 aerobactin outer membrane receptor protein ofEscherichia coli: isolation and purification of the protein and analysis of its nucleotide sequence and primary structure A simple method for displaying the hydropathlc character of a protein Nucleotide sequence of the gene for the ferrienterochelin receptor FepA in Escherlchia coll Purification ofa cytotoxic protein IYom Pseudomouas aeruginosa Interaction of Pseudomonas aerugblosa cytotoxin with plasma membranes from Ehrlich ascites tumor cells Cytotoxic protein from Pseudomonas aeruginosa: Formation of hydrophilic pores in Ehrlieh ascites tumor cells and effect on cell viability Molecular cloning: a laboratory manual Organization of the colicin I b gene Membrane integration and intracellular transport of the coronavirus glycoprotein El, a class III membrane glycoprotean Ausubel FM (1982) Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis A new pair of M13 vectors for selecting either DNA strand of double-digest restriction fl'agments Ecology, clinical significance, and antimicrobial susceptibility of Pseudomonas aerugO~osa Coronavirus glycoprotein El, a new type of viral glycoprotein Construction of improved M13 vectors using oligodeoxynucleotide-directed mutagenesis Gen-etJcs of the Iron dicitrate transport system of Escherichia colt A conformational preference parameter to predict helices in integral inembrane proteins DNA sequencing with chain-terminating inhibitors Ferric-coprogen receptor FhuE of Escherichia coll. Processing and sequence common to all tonB-dependent outer membrane receptor proteins Formation and isolation of leucocidin from Pseudomonas aeruginosa Transport across the outer membrane of Escherwhia coti wa the FhuA receptor is regulated by the TonB protein of the cytoplasmic mernbrane Nucleotide sequence of the cohcin B activity gene cba: Consensus peptapeptide among TonB-dependent colicins and receptors Improved estimation of secondary structure in ribonucleic acids A set of synthetic oligodeoxyribonucleotide primers for DNA sequencing in the plasmid vector pBR322 Codon usage in Pseudomonas aeruginosa Acknowledgements. We thank V. Braun for valuable discussion and for drawing our attention to the presence of a TonB-box within the cytotoxin sequence. We are grateful to G. Kielweln for supply, A. Bauernfeind and R. Onsorg. for typing of the Pseudomonas aeruginosa strain 158, and R. Jungblut for Ehrlich ascites plasma membranes. This work was supported by the Wilhelm-Sander-Stiftung, Neustadt/Donau (FRG), to F. L., and by grants from the Deutsche Forschungsgemeinschaft (SFB 249 and Nie 175-5/2). This work was part of the PhD-thesis of G. O.-E.