* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Contribution of defined amino acid residues to the immunogenicity
Artificial gene synthesis wikipedia , lookup
Signal transduction wikipedia , lookup
Point mutation wikipedia , lookup
G protein–coupled receptor wikipedia , lookup
Genetic code wikipedia , lookup
Ribosomally synthesized and post-translationally modified peptides wikipedia , lookup
Interactome wikipedia , lookup
Amino acid synthesis wikipedia , lookup
Polyclonal B cell response wikipedia , lookup
Magnesium transporter wikipedia , lookup
Biochemistry wikipedia , lookup
Metalloprotein wikipedia , lookup
Protein purification wikipedia , lookup
Protein structure prediction wikipedia , lookup
Protein–protein interaction wikipedia , lookup
Expression vector wikipedia , lookup
Proteolysis wikipedia , lookup
Western blot wikipedia , lookup
Two-hybrid screening wikipedia , lookup
FEMS Microbiology Letters 192 (2000) 223^229 www.fems-microbiology.org Contribution of de¢ned amino acid residues to the immunogenicity of recombinant Escherichia coli heat-stable enterotoxin fusion proteins Isabelle Batisson, Maurice Der Vartanian * Laboratoire de Microbiologie, Institut National de la Recherche Agronomique, Centre de Recherches de Clermont-Ferrand-Theix, 63122 Saint-Gene©s-Champanelle, France Received 28 June 2000; received in revised form 11 September 2000; accepted 19 September 2000 Abstract We investigated whether the toxicity-associated receptor-binding domain of the non-immunogenic Escherichia coli heat-stable enterotoxin (STh) as a fusion with a carrier protein and the inclusion of an appropriate spacer are critical factors for eliciting antibody responses against the native toxin. The immunological properties of three toxic and one non-toxic fusion proteins, consisting of STh N-terminally joined to the C-terminus of the major subunit ClpG of E. coli CS31A fimbriae, were compared. In contrast to the non-toxic hybrid STh with glycine and leucine simultaneously substituted for the receptor-interacting Pro13 and Ala14 amino acids, the toxic chimeras responded by producing high serum levels of anti-STh antibodies in immunized animals. On the other hand, only the toxic ClpG-STh construct with the natural peptide 47 KSGPESM53 of Pro-STh as spacer stimulated STh-neutralizing responses against both native toxin and enterotoxigenic live E. coli cells. Altogether, these findings suggest a close relationship between conformational similarity to the native structure of STh and the ability to elicit specific antibody responses against STh. ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Fusion protein; Neutralizing antibody ; Serum reactivity ; STa enterotoxin ; Heat-stable enterotoxin ; Escherichia coli 1. Introduction Escherichia coli heat-stable enterotoxin (STa), produced by some enterotoxigenic E. coli (ETEC) strains after colonization of the small intestine, is one of the causative agents of childhood diarrhea in developing countries, in addition to traveler's diarrhea [1]. The two types of STa, which vary only slightly, STp and STh, respectively from porcine and human strains of ETEC, are typical extracellular peptides consisting of 18 (STp) and 19 (STh) amino acid residues. STh is synthesized in the cytoplasm as a 72 amino acid precursor consisting of Pre, Pro and mature STh regions [2,3]. After binding to the transmembrane guanylyl cyclase C receptor [4] on the intestinal brushborder membrane, mature STh activates guanylyl cyclase * Corresponding author. Tel. : +33 (73) 624-243; Fax: +33 (73) 624-581; E-mail: [email protected] C and the resulting rise in mucosal cGMP mediates the electrolyte-rich, watery diarrhea characteristic of the illness. The receptor-binding and enterotoxic properties of STh have been mapped to a highly conserved domain including six cysteine residues forming three intramolecular disul¢de bonds that are absolutely necessary for toxicity of the molecule [5]. Because STa is non-immunogenic in its native form, several di¡erent approaches have been explored to obtain non-toxic immunogenic molecules for safe vaccine designs, one of which is genetic coupling of STa to appropriate carrier proteins [6^9]. In general, these constructions either failed to elicit neutralizing antibodies or retained some degree of STa-associated toxicity, suggesting that the immunogenic properties of STa are in£uenced by conformation and amino acid residues associated with toxicity. To investigate this possibility, we constructed toxic and non-toxic fusion proteins in which the N-terminus of STh was variously fused to the C-terminus of the major protein subunit ClpG of E. coli CS31A ¢mbriae [10], which was used as a carrier and a provider of a signal peptide. We compared the immunological properties of these hybrids after parenteral immunizations of mice 0378-1097 / 00 / $20.00 ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 4 3 9 - 0 FEMSLE 9658 26-10-00 224 I. Batisson, M. Der Vartanian / FEMS Microbiology Letters 192 (2000) 223^229 and rabbit via di¡erent routes with preparations containing either ClpG-STh proteins or whole recombinant E. coli cells. 2. Materials and methods 2.1. Bacterial strains and growth conditions E. coli DH5K (Gibco BRL) was used as the host strain for all plasmids. The bovine enterotoxigenic E. coli B41 isolate was used as the reference strain producing STp enterotoxin [11]. Bacteria were grown routinely at 37³C in Luria^Bertani (LB) broth or on LB agar plates supplemented with appropriate antibiotics. 2.2. Construction of plasmids Plasmids used in this study are shown in Fig. 1. Plasmids pEH838, pEHSTC22 [12,13] and pEHProSTC28 [12,13] derived from pEH524 [14], a vector carrying the clp gene cluster that codes for the ClpG export machinery required for the biogenesis of CS31A ¢mbriae [14] and the secretion of ClpG-STh fusion proteins by E. coli DH5K [13]; the wild-type clpG gene from pEH524 was replaced either by the mutated clpG gene from pHPCO838 [12,13], giving pEH838, or by the clpG: :sth fusion genes from pSTC22 and pProSTC28, leading to pEHSTC22 and pEHProSTC28, respectively. Plasmid pSTCDM1 was made by inserting a double-stranded oligonucleotide coding for the entire mature STh with glycine and leucine simultaneously substituted for Pro13 and Ala14 amino acid residues, between the HpaI and XbaI sites of pHPCO838. Because of many unsuccessful attempts to subclone the fusion genes from pSTC17 [12,13] and pSTCDM1 into pEH524, we trans-complemented pSTC17 and pSTCDM1 with pDSPH524 [15], which is pEH524 with clpG deleted, for permitting secretion of fusion proteins. Gene fusions were checked by sequencing. pended cells on MacConkey lactose agar medium containing the appropriate antibiotic. 2.4. Detection, antigenicity and quanti¢cation of fusion proteins by ELISA Detection, antigenicity and quanti¢cation of STh in the fusions were determined by means of a competitive enzyme-linked immunosorbent assay (ELISA) using the commercially available assay kit for E. coli STa (COLI ST EIA) produced by Denka Seiken Co., Ltd., Tokyo, Japan [16] including microtiter plate-attached STa as solid-phase antigen and peroxidase-conjugated STa-monospeci¢c antibody as secondary antibody [12,13], and by means of a double antibody sandwich ELISA using mouse anti-STa monoclonal antibody as coating agent, rabbit anti-ClpG polyclonal antibody as secondary antibody and goat anti-rabbit peroxidase-conjugated IgG as tertiary antibody [13]. Detection and antigenicity were also determined by Western immunoblotting [13]. 2.5. Enterotoxicity assay The suckling mouse enterotoxicity assay [17] was used to assess any STa-related toxic activity of fusion proteins and live recombinant bacteria, as described previously [12,13]. Samples (0.1 ml) were injected intragastrically into 3 days old Swiss OF1 mice which were killed 3 h after injection and examined for increased gut-to-carcass weight ratio as described [17]. A gut/carcass ratio v 0.090 indicated positive toxicity. One mouse unit (MU) was de¢ned as the enterotoxin activity corresponding to a minimum e¡ective dose that gave a gut/carcass ratio of 0.090. Determination of minimum e¡ective dose and enterotoxin titration were performed as described in detail elsewhere [13]. The minimum e¡ective dose of STa necessary to produce an activity of 1 MU was 8 ng, as determined by using pure toxin STp (Calbiochem). 2.3. Preparation of cellular and extracellular fractions 2.6. Immunization of animals and antibody response monitoring LB broth preculture (1 or 2 ml) containing exponentially growing cells was poured onto LB plates which were incubated overnight at 37³C in a humid atmosphere with the agar surface facing up. Bacteria were carefully harvested by being scraped from the agar surface, and the ¢nal suspension volume was made up to 2 ml with phosphate-bu¡ered saline (PBS). After centrifugation at 12 000Ug for 10 min, the resulting supernatants were stored frozen until used as a source of secreted fusion proteins. Cell pellets were suspended and washed in PBS, resuspended in PBS in a ¢nal volume of 2 ml and immediately used as a whole bacterial cell fraction. The bacterial enumeration expressed in CFU (colony-forming units) ml31 was done by spreading out dilutions of PBS-sus- One to four New Zealand white rabbits were immunized intradermally at multiple sites with 15 Wg of fusion protein and emulsi¢ed in incomplete Freund's adjuvant on days 1, 21, 28, and 35 or intravenously with 5U108 live recombinant bacteria suspended in 0.5 ml PBS on days 1, 8, 15, and 22. Rabbit antisera were collected 10 days after the last booster injection, pooled and stored at 320³C until used. Five Swiss OF1 mice were intraperitoneally immunized thrice at 1 week intervals with about 6 Wg of fusion protein or with 3U108 live recombinant bacteria in PBS. Mouse antisera were collected by cardiac puncture under general anesthesia one week after the ¢nal boost, pooled and stored at 320³C until assayed. FEMSLE 9658 26-10-00 I. Batisson, M. Der Vartanian / FEMS Microbiology Letters 192 (2000) 223^229 Anti-STa and anti-ClpG antibody titers in animal antisera were determined using ELISA with microtiter plates precoated with 1 Wg of puri¢ed synthetic STa [16] or with 1 Wg of puri¢ed ClpG protein [12], respectively. Sera from immunized rabbits and mice were serially diluted in PBS (pH 7.2) containing 2% skim milk and 0.5% fetal calf serum added to the wells and incubated for 2 h at 37³C. After washing thrice with PBS containing 0.05% Tween 20 then twice with PBS, bound antibody was detected by adding goat anti-rabbit IgG or goat anti-mouse IgG conjugated to peroxidase (Nordic Immunological Laboratories). After a 20 min enzyme^substrate reaction, the absorbance at 405 nm was read and antibody titers were expressed as the log10 of the reciprocal dilution. To monitor non-speci¢c reactions, absorbances measured with sera from animals immunized with the whole bacterial cell and supernatant fractions from E. coli DH5K (pDSPH524) strain were subtracted from absorbances obtained with test samples. 2.7. Seroneutralization of toxin activity For toxin neutralization, the concentration of pure na- 225 tive STa was adjusted with PBS to 25 MU (200 ng) ml31 , that of ClpG-STh fusion protein to 40 MU (4 to 18 Wg) ml31 and that of native STp from E. coli B41 to 40 MU (320 ng) ml31 , and mixed with an equal volume of sera pooled from groups of immunized or control animals. For neutralization of the B41 cell-bound toxicity, density of bacteria in PBS was adjusted to 1010 CFU ml31 , 2U1010 CFU ml31 , 5U1010 CFU ml31 and 1011 CFU ml31 and mixed with an equal volume of antiserum. The ¢nal dilution of the sera was 1/2 in all cases. After incubation at 4³C for 18 h, samples (0.1 ml) were administered intragastrically to groups of suckling mice as described above. 3. Results 3.1. Toxicity and antigenicity of fusion proteins Our recent studies [13] showed that ClpG-STh fusion proteins from pEHSTC22, pSTC17+pSPH524 and pEHProSTC28 were secreted in the extracellular milieu in a heat-stable toxic form capable of reacting with antibodies to either ClpG or STa (Fig. 1). In addition, they desig- Fig. 1. Structure of fusion proteins. (A) The STh enterotoxin structure. The numbers in boldface above the boxes are the positions of the amino acid residues relative to the Pre-Pro-STh precursor. Only residues 46^53 of the Pro-STh region are shown. The numbers in boldface below boxes are the positions of the residues relative to the mature STh. The toxic domain in mature STh consists of the sequence from Cys6 to Cys18 which includes the receptor-binding region Cys11 -Asn-Pro-Ala-Cys15 . (B) ClpG-STh fusion proteins. The numbers in normal face above boxes are the positions of the amino acid residues relative to the signal peptide cleavage site 31/+1 of the ClpG precursor [10]. The additional valine at the C-terminus of ClpG in pEH838 results of the creation of a unique HpaI site in pEH524 allowing in frame insertion of STh. In pEHProSTC28, the inclusion of a glycine residue between the Ser56 and Asn57 residues of the STh precursor minimizes natural cleavage between Pro-STh and mature STh [13]. Data about the secretion, the enterotoxicity of both ClpG-STh proteins and recombinant whole cells (values represent the speci¢c toxin activity expressed in MU per 1010 CFU) and the antigenicity of ClpG and STh in the fusion summarize results recently published [13], except for the non-toxic mutant speci¢ed by pSTCDM1 for which data were from this study. Asterisks designate mature STh with glycine and leucine simultaneously substituted for Pro13 and Ala14 residues, respectively. FEMSLE 9658 26-10-00 226 I. Batisson, M. Der Vartanian / FEMS Microbiology Letters 192 (2000) 223^229 mice with ClpG-STh hybrids (data not shown). Both rabbits and mice immunized with recombinant bacteria developed serum antibodies with an ELISA anti-ClpG titer close to the serum control in all cases (Fig. 3A), re£ecting the highly immunogenic nature of ClpG within the fusions. Only the non-toxic strain failed to induce any anti-STh antibody formation (Fig. 3B), although eliciting the highest anti-ClpG titer in these animals (Fig. 3A). Responses to toxic bacteria against STh were identical in mouse and variable in rabbit, in which maximum titer was obtained with bacteria carrying pSTC17+pDSPH524 (Fig. 3B). Anti-ClpG titers were always greater than anti-STh Fig. 2. Immunoblot analysis of the non-toxic fusion protein. Supernatants from agar plate-grown cultures containing 5U109 CFU ml31 of DH5K (pEH524) (lane 1) and 5U1010 CFU ml31 of DH5K (pSTCDM1+pDSPH524) (lanes 2 and 3) were mixed with an equal volume of 2ULaemmli bu¡er containing L-mercaptoethanol, boiled and loaded onto 0.1% SDS^15% polyacrylamide gels. After electrophoresis, proteins were electrotransferred to nitrocellulose membranes and incubated either with ClpG-speci¢c antiserum (lanes 1 and 2) or with the STa-speci¢c monoclonal antibody 11C (lane 3) as previously described [13]. Horseradish peroxidase-labeled goat anti-rabbit IgG (lanes 1 and 2) or horseradish peroxidase-labeled goat anti-mouse IgG were used as secondary antibodies (lane 3). Membranes were then developed with H2 O2 -K-chloronapthol as substrate. The arrow at left indicates the native ClpG protein. nated the pEHProSTC28-speci¢ed hybrid as the highest toxic chimera (Fig. 1). In contrast, the secreted fusion protein from pSTCDM1+pDSPH524 with glycine and leucine simultaneously substituted for Pro13 and Ala14 residues of STh revealed no toxin activity (Fig. 1). As assessed by immunoblotting analysis (Fig. 2), this non-toxic chimera reacted with anti-ClpG (lane 2) but not with antiSTa antibodies (lane 3), indicating the lack of STh immunogenicity. In addition, it was detected as a single protein band migrating more slowly than native ClpG (lane 1), consistent with the presence of the additional STh sequence. Most likely, this single band represents the fulllength hybrid molecule. Taken together, these results suggest that Pro13 and Ala14 amino acid residues are required not only for toxicity but also for proper anti-STa antibody binding and that the lack of toxicity and STh immunogenicity for the fusion protein expressed from pSTCDM1 cannot be attributed to cleavage and loss of the enterotoxin sequence. 3.2. Immunogenicity of fusion proteins To evaluate the immunogenic potential of the fusion proteins we compared the anti-ClpG and anti-STh antibody responses following intravenous immunization of rabbits and intraperitoneal immunization of mice with live recombinant bacteria (Fig. 3) or after intradermal immunization of rabbits and intraperitoneal immunization of Fig. 3. ELISA of sera from animals immunized with live bacteria harboring the indicated plasmids for the ability to recognize both ClpG (A) and STh (B). Rabbits and mice were immunized intravenously (IV) and intraperitoneally (IP), respectively. E. coli DH5K (pEH524) was used as negative control and sera from animals immunized with this strain were used as positive and negative controls for reactivity with ClpG and STh, respectively. See Section 2 for details. Error bars represent assayto-assay variations. FEMSLE 9658 26-10-00 I. Batisson, M. Der Vartanian / FEMS Microbiology Letters 192 (2000) 223^229 227 between immunogenicity and toxicity and that expression of immunogenicity by the toxic fusion proteins is independent on the nature of the spacer joining ClpG to STh. 3.3. Serum neutralization of STa toxicity Sera from rabbit and mice variously immunized with fusion proteins were mixed with either native STa or toxic ClpG-STh chimeras or native STp from E. coli B41 and the enterotoxicity of the mixture was determined by the suckling mouse assay (Fig. 4). Only the antiserum of rabbit intradermally boosted with ClpG-STh protein from pEHProSTC28 in the presence of incomplete Freund's adjuvant exhibited toxin-neutralizing activity against native STa, native STp and ClpG-STh proteins (Fig. 4A), and also against up to 2.5U109 whole E. coli B41 cells (Fig. 4B), explaining why only neutralization results relative to this antiserum, named SL28, were shown. It is probable that SL28 recognizes conformational epitopes associated with toxicity which are common to STa, STp and ClpGSTh from pEHProSTC28. Altogether, these ¢ndings indicate that the route of immunization, the choice of animal, the origin of antigen (free or cell-associated), the toxicity state of immunogen, and the inclusion of an appropriate spacer between the carrier protein and the toxin are important considerations for the development of neutralizing antibodies. 4. Discussion Fig. 4. Neutralization of toxin activity in the suckling mouse assay by the serum SL28 from rabbits immunized intradermally with ClpG-STh from pEHSTC28 in the presence of incomplete Freund's adjuvant. (A) A selected dose of native STa toxin, toxic ClpG-STh fusion proteins and native STp toxin from E. coli B41 was mixed (or not) with SL28. The ¢nal dose administered per infant mouse was 1.25 MU for STa and 2 MU for the two other toxic samples. Serum from a rabbit immunized with the culture supernatant of DH5K (pEH524) was used as a negative control. (B) Various quantities of live E. coli B41 cells were mixed (or not) with SL28. The suckling mouse assay for STa enterotoxicity and the signi¢cance for mouse units (MU) are described in Section 2. Gut/ carcass weight ratios of v0.09 (dotted line) are considered positive for toxicity and, therefore, negative for neutralization. Each bar represents the mean gut/carcass ratio and the standard errors of a group of four to 11 mice. *, P 6 0.05; **, P 6 0.01; ***, P 6 0.005. titers and responses against both ClpG and STh were generally higher in rabbit than in mouse (Fig. 3). Immunizations with fusion proteins led to comparable results, except that the detoxi¢ed antigen elicited low (in rabbit sera) or not detectable (in mouse sera) anti-ClpG titers and that the anti-STh response to toxic samples was, overall, lower (data not shown). All toxic proteins triggered similar antiSTh responses in mouse, but not in rabbit, in which response to the chimeric protein from pEHProSTC28 was higher. These results indicate that there is a correlation Studies on the immunological response to the E. coli heat-stable enterotoxin STa have been hampered by its small size, toxicity, poor antigenicity and lack of immunogenicity. Sanchez et al. [8] have reported the construction of a secretable detoxi¢ed STa fusion protein capable of eliciting the production of antibodies which could recognize native STa in an ELISA. However, detoxi¢cation resulted from a single substitution of a disul¢de-linked cysteine, which drastically disrupted the native three-dimensional structure of the toxin. In addition, whether these antibodies neutralized the biological activity of STa was not assessed. This is an important consideration, since removal of amino acid residues associated with toxicity, may, in such a small molecule, render the molecule incapable of inducing neutralizing antibodies. On the other hand, Cardenas and Clements [18] have described a nontoxic, antigenic, and immunogenic recombinant STa capable of inducing antibodies with neutralizing activity when delivered by a bacterial vector. However, neutralization occurred in the absence of detectable ELISA reactivity against STa, which is di¤cult to explain. In addition, this STa chimera was rendered non-toxic merely by fusion with the carrier but not by modi¢cation of some amino acids, making the in£uence of speci¢c residues on the immunogenic properties of STa misunderstood. FEMSLE 9658 26-10-00 228 I. Batisson, M. Der Vartanian / FEMS Microbiology Letters 192 (2000) 223^229 We have previously shown that the three toxic ClpGSTh fusions used in this study were secreted in the extracellular milieu in an antigenic disul¢de-bonded form [13], indicating that spatial conformation of STh in the fusion was close to that of the native toxin, which must be very £exible since the STh moiety in the di¡erent hybrids is certainly not quite the same. Here we showed, in addition, that they were highly immunogenic. Indeed, sera from animals immunized via di¡erent routes with either chimeric molecules or whole recombinant E. coli cells contained antibodies with high titers that were able to recognize both the carrier protein ClpG and the native toxin in vitro. In disagreement with other investigators [6,18], we found that inclusion of an appropriate linker between the carrier and STa moieties was not absolutely required for expression of STa antigenicity and production of antibodies capable of interacting with native toxin. By contrast, maximum toxicity and ability of the fusion protein to elicit serum-neutralizing antibodies against both native toxin and enterotoxigenic live E. coli cells were obtained only after inclusion of the peptide 47 KSGPESM53 of Pro-STh as a spacer between the ClpG and STh domains. Given the involvement of the Pro region in the maturation pathway of STa [19], it is conceivable that this natural linker favors folding of the STh moiety in a more native way. Therefore, it can be assumed that the structure which best resembles the conformation of the native toxin within a fusion protein would be the most e¡ective for eliciting neutralizing antibodies. To investigate whether alteration of some amino acid residues that interact directly with the toxin receptor guanylyl cyclase C [20,21] in£uenced the immunogenicity of the chimeric STh, we substituted simultaneously glycine and leucine for Pro13 and Ala14 residues in STh. Because single substitution of Pro13 by glycine or Ala14 by leucine resulted in marked reduction in their ability to bind to guanylyl cyclase C and in the toxic activity without, however, completely abolishing them [20^22], double mutation was performed. Unlike the usual disul¢de-linked cysteine residues [8], replacement of Pro13 and Ala14 by glycine and leucine residues does not cause structural changes [20], and thus does not drastically disrupt conformational epitopes which are common to native toxin. In contrast to toxic ClpG-STh fusions, and in spite of its ability to be secreted normally and to trigger high antibody responses against ClpG, the non-toxic mutant was incapable of eliciting serum antibodies that could either recognize STh antigenically or neutralize the biological activity of STh. Thus, its lack of toxicity corresponded to its inability to develop antibodies against the STh moiety, suggesting either that substitution of Pro13 and Ala14 induced some changes in toxicity-associated STh conformation, thus a¡ecting immunogenicity or that these amino acid residues belong to a STh epitope required for the development of antibodies directed against STh. Acknowledgements This study was supported by grants from the Conseil Rëgional Auvergne (France) and the Institut National de la Recherche Agronomique (France). T. Takeda is gratefully acknowledged for donation of the mouse anti-STa monoclonal antibody 11C. We thank M. Chavarot, C. De Martrin, A. Garrivier, B. Ja¡eux and G. Vert for technical assistance and S. Dutilloy and F. Magne for secretarial assistance. References [1] Black, R.E. (1993) Epidemiology of diarrhoeal disease: implications for control by vaccines. Vaccine 11, 100^106. [2] Okamoto, K. and Takahara, M. (1990) Synthesis of Escherichia coli heat-stable enterotoxin STp as a pre-pro form and role of the pro sequence in secretion. J. Bacteriol. 172, 5260^5265. [3] Rasheed, J.K., Guzman-Verduzco, L.-M. and Kupersztoch, Y.M. (1990) Two precursors of the heat-stable enterotoxin of Escherichia coli : evidence of extracellular processing. Mol. Microbiol. 4, 265^273. [4] Schulz, S., Green, C.K., Yuen, P.S.T. and Garbers, D.L. (1990) Guanylyl cyclase is a heat-stable enterotoxin receptor. Cell 63, 941^948. [5] Gariëpy, J., Judd, A.K. and Schoolnick, G.K. (1987) Importance of disul¢de bonds in the structure and activity of Escherichia coli enterotoxins ST1b. Proc. Natl. Acad. Sci. USA 84, 8907^8911. [6] Clements, J.D. (1990) Construction of a nontoxic fusion peptide for immunization against Escherichia coli strains that produce heat-labile and heat-stable enterotoxins. Infect. Immun. 58, 1159^1166. [7] Lo«wenadler, B., Lake, M., Elmblad, A., Holmgren, E., Holmgren, J., Karlstro«m, A. and Svennerholm, A.-M. (1991) A recombinant Escherichia coli heat-stable enterotoxin (STa) fusion protein eliciting antiSTa neutralizing antibodies. FEMS Microbiol. Lett. 82, 271^278. [8] Sanchez, J., Svennerholm, A.-M. and Holmgren, J. (1988) Genetic fusion of a heat-stable enterotoxin-related decapeptide antigen to cholera toxin B-subunit. FEBS Lett. 241, 110^114. [9] Aitken, R. and Hirst, T.R. (1993) Recombinant enterotoxins as vaccines against Escherichia coli-mediated diarrhoea. Vaccine 11, 227^ 233. [10] Girardeau, J.-P., Bertin, Y., Martin, C., Der Vartanian, M. and BEuf, C. (1991) Sequence analysis of the clpG gene, which codes for surface antigen CS31A subunit: evidence of an evolutionary relationship between CS31A, K88 and F41 subunit genes. J. Bacteriol. 173, 7673^7683. [11] So, M., He¡ron, F. and McCarthy, B.J. (1979) The E. coli gene encoding heat stable toxin is a bacterial transposon £anked by inverted repeats of IS1. Nature 277, 453^456. [12] Batisson, I. and Der Vartanian, M. (2000) Extracellular DsbA-insensitive folding of Escherichia coli heat-stable enterotoxin STa in vitro. J. Biol. Chem. 275, 10582^10589. [13] Batisson, I., Der Vartanian, M., Gaillard-Martinie, B. and Contrepois, M. (2000) Full capacity of recombinant Escherichia coli heatstable enterotoxin fusion proteins for extracellular secretion, antigenicity, disul¢de bond formation, and activity. Infect. Immun. 68, 4064^4074. [14] Martin, C., BEuf, C. and Bousquet, F. (1991) Escherichia coli CS31A ¢mbriae: molecular cloning, expression and homology with the K88 determinant. Microb. Pathog. 10, 429^442. [15] Der Vartanian, M., Mëchin, M.-C., Ja¡eux, B., Bertin, Y., Fëlix, I. and Gaillard-Martinie, B. (1994) Permissible peptide insertions surrounding the signal peptide-mature protein junction of the ClpG prepilin: CS31A ¢mbriae of Escherichia coli as carriers of foreign sequences. Gene 148, 23^32. FEMSLE 9658 26-10-00 I. Batisson, M. Der Vartanian / FEMS Microbiology Letters 192 (2000) 223^229 [16] Scotland, S.M., Willshaw, G.A., Said, B., Smith, H.R. and Rowe, B. (1989) Identi¢cation of Escherichia coli that produces heat-stable enterotoxin STA by a commercially available enzyme-linked immunoassay and comparison of the assay with infant mouse and DNA probe tests. J. Clin. Microbiol. 27, 1697^1699. [17] Giannella, R.A. (1976) Suckling mouse model for detection of heatstable Escherichia coli enterotoxin : characteristics of the model. Infect. Immun. 14, 95^99. [18] Cardenas, L. and Clements, J.D. (1993) Development of mucosal protection against the heat-stable enterotoxin (ST) of Escherichia coli by oral immunization with a genetic fusion delivered by a bacterial vector. Infect. Immun. 61, 4629^4636. [19] Yamanaka, H. and Okamoto, K. (1996) Amino acid residues in the Pro region of Escherichia coli heat-stable enterotoxin I that a¡ect e¤ciency of translocation across the inner membrane. Infect. Immun. 64, 2700^2708. 229 [20] Sato, T., Ozaki, H., Hata, Y., Kitagawa, Y., Katsube, Y. and Shimonishi, Y. (1994) Structural characteristics for biological activity of heat-stable enterotoxin produced by enterotoxigenic Escherichia coli: X-ray crystallography of weakly toxic and nontoxic analogs. Biochemistry 33, 8641^8650. [21] Ozaki, H., Sato, T., Kubota, H., Hata, Y., Katsube, Y. and Shimonishi, Y. (1991) Molecular structure of the domain of heat-stable enterotoxin produced by a pathogenic strain of Escherichia coli. J. Biol. Chem. 266, 5934^5941. [22] Waldman, S.A. and O'hanley, P. (1989) In£uence of a glycine or proline substitution on the functional properties of a 14-amino-acid analog of Escherichia coli heat-stable enterotoxin. Infect. Immun. 57, 2420^2424. FEMSLE 9658 26-10-00