WO2001069259A1 - Method for detecting guanylyl cyclase and use thereof for diagnosing metastatic colorectal tumours - Google Patents

Abstract

The invention concerns the use a conjugate of STa peptide and an active molecule wherein said STa peptide has a conformational structure enabling it to be fixed on the GC-C receptor for preparing a pharmaceutical composition for treating or diagnosing metastatic colorectal tumours. More particularly, the invention concerns a method for diagnosing metastatic colorectal tumours which consists in immunologic detection of GC-C receptors in a patient's biological sample, based on the use of the chimeric protein formed by a ClpG protein and the STa peptide and specific antibodies directed against the Clp-G protein or the chimeric protein.

Description

 METHOD FOR THE DETECTION OF GUANYLYL CYCLASE-C AND ITS USE FOR THE DIAGNOSIS OF METASTATIC COLORECTAL TUMORS.

The present invention relates to the detection of Guanylyl Cyclase-C (GC-C) which is the transmembrane receptor for the thermostable enterotoxin (STa) produced by Escherichia coll. Guanylyl Cyclase-C has been described as a specific marker for metastases originating from colorectal cancers (Carithers, SL et al., 1996, PNAS, USA, 93: 14827-14832; Carithers, SL et al., 1996, Dis. Colon Rectum., 39: 171-181). Indeed, it has been shown that GC-C is expressed in healthy cells of the intestinal mucosa and in colorectal metastasized cancer cells, but not in healthy extra-intestinal tissues or in non-colorectal metastasized cancer cells, this which makes it possible to envisage its use as a specific marker for colorectal metastases (aldman, SA et al., 1998, Cancer epidemiology, biomar e.rs & prevention, 7: 505-514; Waldman, SA et al., 1998, Dis. Colon Rectum., 41: 310-315). The invention therefore also relates to the detection of metastases of colorectal origin for the diagnosis of metastatic colorectal tumors. The occurrence of metastases is the major cause of death for patients with colonic cancer. Also, the search for circulating tumor cells with very sensitive and very specific methods constitutes an interesting means of screening and monitoring of cancer patients (Le Maire, V. et al., 1998, Ann. Phar. Fr., 56: 9 -17). To date, there are no simple technologies that are both sensitive and specific for this type of screening. Conventional immunological methods which use a variety of monoclonal antibodies directed against common markers such as ACE, CEA, CA19-9 17-1-A, CA50, CA195, CA72-4, CA242, TAG 72, TPA, villin, LAP, plasma prolactin and the cytokeratins CK19 and CK20, have been shown to be insufficient due to their lack of specificity and sensitivity (Wyld, DK et al., 1998, Int. J. Cancer (Pred. Oncol.), 79: 288 -293). One of the major problems lies in the accessibility of tumor cells to antibodies.

Today, it is possible to easily detect secondary etastatic foci by cumbersome and expensive methods such as magnetic resonance or bone scintigraphy. On the other hand, hematogenous screening techniques such as flow cytometry or immunohistochemistry have limited sensitivity and specificity. Recently, biology techniques PCR-based molecular tests to detect mRNAs encoding specific antigens of the colonic cell, CGM2, CK19 and CK20 (Le Maire, V. et al., 1998, Ann. Pharm. Fr., 56: 9-17) have allowed to detect the presence of circulating cells in the blood with a sensitivity far superior to conventional immunological methods. However, the high frequency of detection, from 30% to 100%, of these antigens in the blood of healthy subjects, due to an illegitimate transcription of tissue-specific genes, indicates a low specificity of the molecular tool which, for the moment , cannot be used routinely to demonstrate the hematogenous spread of colonic cells.

Choosing the GC-C receptor as a molecular target has many advantages:

- Only 5% of healthy patients have been shown to have GC-C mRNA in the blood (Bustin, S. A. et al. 1999, British J. Cancer, 79: 1813-1820).

- Unlike the antigens mentioned above which are only tissue-specific, and while rare antigens are really specific for tumor cells, the GC-C receptor is a specific marker for colorectal metastases.

The affinity of the STa toxin for its receptor is very high and because of its small size (2 Kd), the STa toxin can easily access tumor cells via its GC-C receptor.

The inventors have envisaged a method for detecting the GC-C receptor which does not have the drawbacks of the techniques based on the detection of GC-C mRNAs by RT-PCR proposed in the prior art.

The research carried out in the context of the present invention has led the inventors to develop a method for detecting the GC-C receptor by its exogenous ligand the toxin STa. However, such an approach presents difficulties because:

- The STa toxin is devoid of immunological properties. As a result, it is difficult to produce specific anti-STa toxin antibodies by immunization with the STa toxin alone.

- The STa toxin is weakly antigenic, which limits the use of all the anti-toxin STa antibodies obtained with STa conjugated to an immunogenic factor to target the STa toxin linked to its receptor.

The inventors have now succeeded in preparing hybrid proteins comprising the peptide constituting the toxin STa and an immunogenic carrier protein in which the conformal structure of STa toxin remains compatible with its binding to the GC-C receptor.

The STa toxin is an extracellular peptide produced by E. enterotoxigenic coli (ETEC) responsible for severe diarrhea and "traveler's disease (tourista)". The toxin secreted by the bacteria in the gastrointestinal tract binds to its GC-C receptor, activates it and enters the enterocytic cells by endocytosis to induce an accumulation of intestinal fluid. The peptide constituting the toxin STa, hereinafter also designated the peptide STa, comprises 18 to 19 amino acids and has three disulfide bridges necessary for its enterotoxic activity. Indeed, the integrity of the tridi entional structure of the STa toxin is essential for its binding with the GC-C receptor.

The invention therefore relates to the use of a conjugate of the peptide STa and of an active molecule in which the said peptide STa has a conformal structure allowing its attachment to the GC-C receptor for the preparation of a pharmaceutical composition intended in the treatment or diagnosis of metastatic colorectal tumors.

Advantageously, the active molecule is a protein and the conjugate is a fusion protein, also hereinafter designated chimeric protein or hybrid protein, produced by fusion of the genes coding for the peptide STa and said active protein.

A pharmaceutical composition according to the invention comprises one or more chimeric proteins in a vehicle compatible with said chimeric protein and acceptable for therapeutic or diagnostic use in humans or animals.

The chimeric proteins according to the invention are very particularly useful for the detection of GC-C receptors and the diagnosis of metastatic colorectal tumors. In this case, the active protein entering the constitution of the chimeric protein according to the invention is advantageously an immunogenic carrier protein capable of conferring on the chimeric protein immunogenic properties without modifying the conformal structure of the STa peptide allowing its attachment to the receptor GC- C and therefore the detection of the fusion protein attached to the GC-C receptor.

The inventors have specific knowledge and know-how in the expression of heterologous proteins and peptides in the major protein subunit ClpG of CS31A fimbriae without affecting the formation of these fimbriae on the surface of E. coli.

Thus, the international patent application published under No. W094 / 14967 reports the work of inventors to define regions of the ClpG protein into which heterologous peptides can be introduced without disturbing the biogenesis of the CS31A protein capsule. The inventors have used their previous work on CS31A fimbriae to use the ClpG subunit as a protein conjugated to the STa peptide. The inventors have succeeded in inducing neutralizing and protective antibodies directed against the STa toxin by immunization with chimeric proteins resulting from

genetic fusions between the STh peptide (Sta of human origin) and the ClpG protein.

Four distinct chimeric proteins comprising a heterologous sequence of 17 to 28 amino acids were prepared as described for example in PCT patent application WO94 / 14967, one having the peptide STh at the N-terminal end of the protein ClpG and the other three having the STh peptide at the C-terminus of the ClpG protein. The characterization of the four hybrids shows that, surprisingly, the N and C-terminal fusions are secreted into the extracellular medium in a free monomeric form and not in a fimbrial polymeric form exposed on the surface of bacteria.

The results of this work are remarkable in that all chimeric proteins ClpG-STh are associated with a strong enterotoxic activity comparable to that of the natural toxin STa. However, as indicated above, it is known that this activity, evaluated in vivo using the newborn mouse test, depends on the native structure of the STa toxin stabilized by the existence of three disulfide bridges, also present in the STh toxin.

These results show for the first time that the STa part of these hybrid proteins retains its natural characteristics allowing it to recognize its GC-C receptor and therefore to keep its toxicity by inducing an accumulation of intestinal fluids characteristic of diarrhea.

The invention therefore relates very particularly to a chimeric protein formed by a ClpG protein and the STa peptide where the STa peptide is introduced at the level of a permissive region of the ClpG protein, and its use for in vitro detection of GC receptors -VS. In fact, the antigenic character of the ClpG part of the chimeric proteins makes it possible to use anti-ClpG antibodies for in vitro or in vivo detection of GC-C receptors and therefore for screening for colorectal metastases. A person skilled in the art is capable of preparing such chimeric proteins from the teaching of the inventors' previous work on the ClpG protein and its permissive regions as well as on the basis of the experimental results reported in the experimental part reported below. .

A simple and rapid form of diagnosis of metastatic colorectal tumors according to the invention consists in an immunological detection of the GC-C receptors in a biological sample of a patient, based on the use of a chimeric protein formed by a ClpG protein and STa peptide and specific antibodies raised against the ClpG protein or against the chimeric protein.

They may be monoclonal or polyclonal antibodies prepared by conventional techniques well known to those skilled in the art of the type described in the experimental part below.

A diagnostic method according to the invention comprises the following steps:

- bringing said sample into contact with a chimeric protein formed by a ClpG protein and the STa peptide under conditions allowing the binding of the chimeric protein to the GC-C receptors, which are strongly expressed by metastatic cells, by means of the STa part of the chimeric protein, then - the detection of the chimeric proteins attached to the GC-C receptors using antibodies directed against the ClpG part of the chimeric protein, or against the chimeric protein itself.

The specific anti-ClpG or anti-chimeric antibody can be advantageously coupled to enzymatic or fluorescent chromophores.

The biological sample of a patient on which the diagnostic method of the invention is practiced is for example colonic cells circulating in the blood or coming from a biopsy. The method of the invention can therefore be carried out using blood (ELISA method) or biopsy (immunohistochemical method).

As an example of implementation of the detection of the GC-C receptor for the thermostable toxin STh in a biological extract according to the invention, we can therefore cite: the ELISA technique (Enzyme-linked Immunosorbent Assay) by competition from blood extracts or tissue materials, or immunohistochemistry techniques from biopsy of extraintestinal tissue.

As an example, in the case of a test based on an ELISA technique by microtiter plate competition, the wells are sensitized with all or part (natural or synthetic) of the semi-purified or purified receptor from the intestinal cell lines. T84 or Caco 2 in culture; if a part of the receptor (extracellular part) is used, it must contain the toxin recognition site. Unoccupied wells can be saturated with blocking agents such as BSA (bovine serum albumin). The immobilized receptors are brought into contact with dilutions of the blood plasma (the plasma could contain the extracellular domain of the receptor) or with dilutions of receptors obtained after mechanical homogenization of tissue materials and solubilization of membranes. The purified ClpG-STh fusion proteins are then added to the wells. Another possibility is that the immobilized receptors are brought into contact with the sample + ClpG-STh protein mixture pre-incubated beforehand at 37 ° C / 30 min. After incubation and washing, specific primary antibodies (serum polyclonal IgGs [purified or semi-purified], or monoclonal of one type or a mixture of several types) obtained from either the purified ClpG protein, and, therefore, directed only against the ClpG part of the hybrid protein, ie of the chimeric protein ClpG-STh, and therefore directed against both the ClpG and STh parts. The binding of the STh part of the hybrids to the immobilized receptor is detected indirectly by the use of secondary antibodies, for example, labeled with peroxidase, and capable of binding to the primary antibodies. Another possibility is that the primary antibodies are directly conjugated to the peroxidase, thus avoiding the use of secondary antibodies. The reaction is then revealed by colorimetry through an enzymatic activity which can be amplified if necessary. The intensity of the color reaction is inversely proportional to the amount of receptor in the sample. Reading optical densities at a certain wavelength of standards (different concentrations of purified receptors) will be used to construct a standard curve for measuring the quantity of receptors in the samples.

If the blood extracts contain GC-C receptors, this will fix the ClpG-STh proteins and the whole will therefore not be able to bind to the immobilized receptors because there will be a phenomenon of competition between immobilized GC-C and GC- C free in the biological extracts, with respect to the ClpG-STh proteins: for a high concentration of receptors (positive results), a very pale colored reaction or an absence of coloration will be observed.

If the extracts do not contain receptors (negative result), the ClpG-STh hybrids bind to the receptor fixed on the solid phase of the microtiter plates. During the colorimetric development, the reaction will be colored.

The invention therefore also relates to a kit for implementing this method, characterized in that it comprises:

- a microtiter plate where each well is sensitized with all or part of the GC-C receptor (natural or synthetic) retaining the toxin binding site, - at least one positive control containing purified GC-C receptor.

- the purified ClpG-STh chimeric protein, primary antibodies concentrated in solution containing serum polyclonal or monoclonal IgG (mouse) anti-ClpG or anti-hybrid ClpG-STh,

- labeled secondary antibodies,

- stamps.

As a detailed example, such a kit includes:

- Breakable microtiter plates: 6 strips of 16 wells on a support. Each well is sensitized with all or part of the GC-C receptor retaining the toxin binding site and contains a stabilizer.

- A positive control: lyophilized standard containing GC-C receptor purified to a given concentration which will be diluted with water before use.

The purified, lyophilized and quantified ClpG-STh chimeric protein.

Primary antibodies concentrated in solution containing serum IgG polyclonal (rabbit, goat or others) or monoclonal (mouse) anti-ClpG or anti-hybrid ClpG-STh which will be diluted with antibody buffer at the last moment. Secondary antibodies concentrated in solution containing polyclonal IgG (animal origin depending on that of primary antibodies) conjugated to peroxidase. - A blocking buffer: Tris HCI buffer with

BSA or normal serum (rabbit, goat, mouse or others depending on the origin of the primary and secondary antibody used in the rest of the test).

- A washing pad. - An antibody buffer.

- A stop buffer: 0.3 M sulfuric acid.

Peroxidase substrate A: o-phenylenedi amine or dehydrated diaminobenzidine.

Peroxidase substrate B: hydrogen peroxide solution.

However, the invention also makes it possible to envisage the use of the ClpG-STh fusion proteins in immunohistochemistry on frozen sections or on paraffin sections of tissues of extraintestinal organs (lymph nodes, liver, kidney, spleen, etc.) in colorectal cancer patients to detect the GC-C receptor inside or on the surface of metastatic tumor cells. Using primary polyclonal or monoclonal anti-ClpG or anti-hybrid ClpG-STh antibodies previously labeled (Peroxidase, fluorochro es, biotin, etc ..) either directly or indirectly by secondary antibodies against primary bodies, it is possible to visualize by electron microscopy the presence of tumor cells, to estimate their proportion, and thus to evaluate the degree of cell proliferation (degree of malignancy and extension). The advantage of this system is based on the STh toxin / GC-C receptor recognition which is more sensitive than the GC-C receptor / anti-receptor antibody recognition usually used in this kind of immunohistochemical approach.

The invention mainly finds its application in the field of diagnostics, but it also has therapeutic applications in the treatment of metastases of colonic origin. Indeed, as indicated previously, the STa toxin penetrates into the enterocytes after being fixed on its GC-C receptor (Urbanski, R. et al., 1995, Biochi. Biophys. Acta, 1245: 29-36), and the STa peptide can therefore be used as a vehicle for internalizing the conjugate of the invention. In this case, the active molecule, more particularly an active protein, is an anti-cancer therapeutic agent advantageously having an activity: anti-cancer specific for tumor cells, ie capable of killing them without affecting healthy cells,

- only on cells undergoing multiplication,

- capable of diffusing remotely from a receptor tumor cell onto other peripheral cancer cells which have not incorporated the therapeutic agent. An example of such a therapeutic agent is thymidine kinase type 1 of the herpes virus. The chimeric protein can be administered systemically so as to selectively target the extra-intestinal rectocolic tumor cells and to selectively eliminate them after internalization of the chimeric protein via the GC-C receptor. It is therefore the toxic character determined by the structure three-dimensional STa peptide that will allow this selective action. The very small size of the STa peptide, its absence of immunogenicity and its high selectivity for its receptor gives the STa peptide a very particular interest for vectorizing active agents in colorectal metastases. The existence of Smad3 strains of transgenic mice capable of specifically developing colorectal cancer provides a complementary tool for the experimental study of applications of the invention (Zhu, Y. et al., 1998, Cell., 94: 703-14).

Other advantages and characteristics of the invention will appear from the following examples concerning the preparation of chimeric proteins formed by a ClpG protein and the peptide STa and the demonstration that the chimeric proteins ClpG-STh are associated with a strong comparable enterotoxic activity. to that of the natural toxin STa.

A - MATERIAL AND METHODS.

1) Bacterial and plasmid strains.

The different bacterial strains used during this work are presented in Table 1 below.

Table 1

Genotype / Characteristics

DH5α F ^" supEé ά, A (argF-lacZYA) U169 (080 lacZΔMlΞ), hsdRl 7 (rk ^~ , mk ⁺ ), recAl, endAl, gyrA96 (Nal ^r ), thil, relAl

B41 ETEC strain of bovine origin (O101: K ^~ : H ^" , K99, F41) producing STa.

JCB570 AraDl39A (araABC-Leu) 7679, galU, galK, A (lac) X74, rpsL, thi, phoR, zih 12:: Tnl 0 JCB571 JCB570 dsbA:: kanl

JCB819 JCB570 (malF-lacZl 02) dsbB:: kanl

JCB818 JCB570 (malF-lacZl 02) dsbA:: kanl, dsbB:: kanl

The B41 strain is a wild reference strain producing STp toxin. The JCB571, JCB818, JCB819 strains derived from JCB570 are mutated in the Dsb system involved in the formation of disulfide bridges. The bacteria are cultivated at 37 ° C. in LB medium (Luria Bertani) liquid or agar supplemented with the appropriate antibiotics: ampicillin (Ap), 50 μg / l; chloramphenicol (Cm), 25 μg / ml; tetracycline (Te), 12.5 μg / ml; streptomycin (S) 50 μg / ml.

The different plasmids used in this work are described in Table 2 below.

Table 2

pHSG575 Low copy number cloning vector; origin of replication pSClOl, Cm ^R pEH52 pHSG575 with the clp operon coding for fimbriae

CS31A

PDSPH52 pEH524 deleted from the clpG pSK + gene Cloning vector pBluescript SK (+); origin of replication pColEl, Ap ^E

PDEV41155 pSK + with clpG between BcoRV and Xbal PHPCOH838 pDEV41155 with a GTT codon just before the last codon of clpG pSTN2 pHPCOH838 with STh in Sphl pSTC22 pHPCOH838 with STh between the Hpal sites of clpG and Xbal of the pProSTTh28 pST sites pSTB22 pST Smal pSTC17 PSTC22 deleted from the Hpal / Smal fragment pEHSTN24 pΞH524 with clpG-STh from pSTN24 in place of clpG pEHSTC22 pEH524 with clpG-STh from pSTC22 in place of clpG pΞHProSTC28 pEH524 with clpG psth28 1 Cloning vector; origin of replication pColEl,

Tc ^R pSXIOS pSelect-1 with clpG in which the Spel and Bglll sites were created pSTVIOS pSXIOS with STh inserted in Spel PSTVD41 pSTVIOS deleted from the BamΑl / BglII fragment PEHSTVD41 pΞH524 with clpG-STh from pSTVDIT prc instead of p99 expression having the inducible promoter

Ptrc; origin of replication pColEl, Ap ^R , Lacl *, rmB pTrcSTN24 pTrc99A with clpG-STh from pSTN24 pTrcProSTC28 pTrc99A with clpG-STh from pProSTC28 pSTCDMl pHPC0838 with STh mutated for 1 'ent rotoxicity pGP clp HG sites with STh between Sphl and Hpal pSTE531 PSTN24 with duplex DNA E531 / E351 (Table 10) in

Pstl pSTDG17 PSTE531 deleted from the SnaBl / Hpal fragment pSTDG126 PHPC0838 with STh between Sphl and Hpal pGMSTC7 pSTC22 with duplex DNA GMCST3 / GMCST5 (Table 10) between Sphl and Hpal pGMSTC71 PGMSTC7 deleted from the HpalC / Hpal7 fragment fragment ppalMM7 ppalMM7 ppal Smal pGMSTC13 pProSTC28 with duplex DNA GMCST3 / GMCST5 (Table 10) between Sphl and Hpal pGMTSTCl3 / 19 PGMSTC13 deleted from the Hpal / Hpal fragment pGMSTC13 / 28 PGMSTC13 deleted from the HcoRI / EcoRI fragment The plasmid pEH524 carrying the operon clp contains all the genes necessary for the biogenesis of CS31A. The plasmid pDSPH524 is a derivative of pEH524 in which the clpG gene has been completely deleted. The plasmids pDEV41155 and pSXIOS are derived respectively from the vectors pBluescript SK (Stratagene) and pSelect 1 (Promega) in which only the clpG gene has been cloned. The plasmid pSXIOS has a unique Spel restriction site created by site-directed mutagenesis. Asparagine 203 from ClpG is thus replaced by a threonine. The plasmid pHPC0838 derives from pDEV41155 by substitution of the unique restriction site A fllll in a unique site B rill, by creation of a unique site Hpal and by addition of a valine residue between the last two residues (tyrosine and asparagine) C- ClpG terminals. In order to be able to control the expression of the different hybrid proteins, the fusion genes were placed downstream of the P _trc promoter inducible by IPTG

(isopropyl-β-D-thiogalactopyranoside) in the expression vector pTrc99A. Induction is carried out with 2 mM IPTG.

2) Purification of the plasmids (Qiagen kit). One hundred milliliters of liquid LB medium containing the appropriate antibiotics are seeded with 500 μl of a bacterial preculture and incubated overnight at 37 ° C with stirring. After centrifugation, the pellet is taken up in 4 ml of Pi buffer. Then 4 ml of lysis buffer P2 are added and the whole is incubated for 5 min at room temperature. Four milliliters of P3 buffer are then added and the whole is left for 5 min in ice, then centrifuged for 30 min at 15,000 g at 4 ° C. The supernatant is deposited on a QIAGEN-100 column previously equilibrated with 3 ml of QBT buffer. The column is then washed with 10 ml of QC buffer and the DNA is eluted with 5 ml of QF buffer. The eluate is added with two volumes of absolute ethanol and incubated for 5 min in ice. After centrifugation (30 min, 15,000 g, 4 ° C), the DNA pellet is dried for 5 min under vacuum and taken up in sterile double-distilled water. The quantity and purity of the DNA obtained are estimated by measuring the OD at 260 nm and at 280 nm. The recombinant DNA thus purified is then analyzed by sequencing.

3) Purification of DNA fragments in agarose gel.

The enzymatic digestions of the DNA are carried out in accordance with the indications given by the suppliers. The samples, in solution in loading buffer, are deposited on a 0.7% to 1.4% agarose gel, containing ethidium bromide at a final concentration of 0.5 μg / ml. After migration in buffer of TBE electrophoresis, the DNA fragments are visualized under ultraviolet (UV). The 1 kb marker scale (Gibco-BRL) makes it possible to identify the size of the DNA fragments. These are then directly cut from the agarose gel and then purified using the QIAquick Gel Extraction kit (QIAGEN) according to the supplier's instructions. The agarose block containing the DNA is dissolved at 50 ° C for 10 min in three volumes of QX1 buffer. The sample is loaded onto a QIAquick column which is centrifuged at 10,000 g for 1 min. Then 750 μl of PE washing buffer are deposited on the column which is incubated for 5 min at room temperature, before being centrifuged at 10,000 g for 1 min. Additional centrifugation (10,000 g, 1 min.) Makes it possible to remove the residual traces of PE buffer. The column is placed in a new Eppendorf tube and the DNA is eluted by adding 50 μl of H ₂ 0 and centrifuging at 10,000 g for 1 min.

4) Duplex of synthetic oligonucleotides. All the synthetic oligonucleotides purified by HPLC (Eurogentec) used to make the different constructions are shown in Table 3 below. Table 3

Oligos Sequences (5'_ 3 ')

Sth72N3 AGGATCTGCΛGGGTAACATCCTGEACAGGCAGGATTACAACAAAGTTCA

CAGCAGTACAGAΓCTGCCGCATG

Sth72N5 CGGCAGATCTGTACTGCTGTGAACTTTGTTGTAATCCTGCCÎΓGΓACAGG ATGTTACC CTGCAGATCCTCATG

Sth73C3 CTAGTTAGTAACATCCTGΓACAGGCAGGATTACAACAAAGTTCACAGCA

GTAGTTCCCGGGGTTATCAGGGTT

Sth69C5 AACCCTGATAACCCCGGGAACTACTGCTGTGAACTTTGTTGTAATCCTG CC TGTACAGGATGTTACTAA

Sth33PC3 GCTAGAATTCATCGATTCAGGACCACTTTTGTT

Sth33PC5 AACAAAAGTGGTCCTGAATCGATGAATTCTAGC

Sth69V3 CTAGGATCCCCGTAACATCCΓGÏΆCAGGCAGGATTACAACAAAGTTCAC AGCAGTAGTTAGAAGAATΓC

Sth69V5 CTAGGAARRC TCTAACTACTGCTGTGAACTTTGTTGTAATCCTGCCRG

ΓACAGGATGTTACGGGGATC

E351 TTATΛCGTΛATTΛΛTTΛΛGCTTGCA

E531 AGC TTAATTAAT TACGTATAATGCA

SPGST3 AACTTAGTAACATCCTGTACAGGCAGGATTACAACAAAGTTCACAGCAG TAGTTAGAAGAATΓGCATG

SPGST5 CGAATΓCTTCTAACTACTGCTGTGAACTTTGTTGTAATCCTGCCO'GTAC

AGGATGTTACTAAGTT

GMCST3 AACTGCΛG ^ TTCATGTTAÂCCGCATG GMCST5 CGGTTAACNTGAATTCTGCAGTT

STNDG2 AACTTAGTAACATCCTGTACAGGCAGGATTACAACAAAGTTCACAGCAG

TACAGATC2GCCGCATG

STNDG1 CGGCAGATCTGTACTGCTGTGAACTTTGTTGTAATCCTGCCTGTACAGG

ATGTTACTAAGTT

STGL673 CTAGAΓCÎTAATAACAACCAGTACAÇAG

STGL635 ACCATTACAACACAGTTCACAACAATAATTACΓCGAGTT

AACTCGAGTAATTATTGTTGTGAACTGTGTTGTAATGGT

CTGTGTACTGGTTGTTATTAAGAT

In Table 3 above, the sequences coding for all or part of the STh toxin are indicated in bold characters. The restriction sites for the enzymes Bgl III (AGATCT in Sth72N3, Sth72N5; STNDG1, STNDG2; GMCST3, GMCST5; STGL673, STGL635), Hpal (GTTAAC in GMCST3, GMCST5), Pstl (CTGCAG in Sth72N3, Sth725)

(CCCGGG in Sth73C3, Sth69C5), Clal (ATCGAT in Sth33PC3, Sth33PC5), EcoRl (GAATTC in Sth33PC3, Sth33PC5; Sth69V3, Sth69V5), BamKl (GCATCC in Sth69V3, Sth69V5), BsrGI

(TGTACA in Sth72N3, Sth72N5; Sth73C3, Sth69C5; SPGST3, SPGST5; STNDG1, STNDG2), Xhol (CTCGAG in STGL673, STGL635) and Pacl (TTAATTAA) and Sna l (TACGTA) in E531 and E1. Codons modified from the original STh toxin sequence are underlined.

The single-stranded complementary oligonucleotides are mixed in TE buffer to a final concentration of 2 μg / μl. They are heated to 80 ° C for 1 min then to 65 ° C for 5 min and cooled in ice until the temperature reaches about 25 ° C. The oligonucleotide duplexes thus formed are stored at -20 ° C.

5) Cloning, transformation and selection of the recombinant clones.

Passenger DNA and vector DNA are mixed in a 1: 1 or 3: 1 molar ratio at a concentration of total DNA between 15 and 50 ng / μl. The fragments are ligated in a final volume of 20 μl of ligase buffer (BRL) containing 1 to 5 U of T4 DNA ligase for 18 h at 16 ° C or overnight at 4 ° C. The recipient bacterial cells, previously treated with CaCl _2r are then added (200 μl) and then incubated for 40 min in ice. After a thermal shock (2 min at 42 ° C then 5 min in ice), 800 μl of sterile liquid LB medium preheated to 37 ° C are added. The whole is incubated for 45 min at 37 ° C. with shaking. Then 100 μl of the transformation mixture are spread on selective agar medium and incubated overnight at 37 ° C. The individual screening of the colonies was carried out by analysis of the restriction fragments and by sequencing of the fusion genes. The restriction fragments are obtained from a mini-preparation of partially purified plasmid DNA. To do this, a bacterial colony is removed and cultured overnight at 37 ° C. with stirring in liquid LB medium supplemented with appropriate antibiotics. Then 1.5 ml of the bacterial suspension are centrifuged (5 min, 15000 g, 4 ° C) and the pellet is taken up with 150 μl of buffer Pi and 200 μl of lysis buffer P2. After incubation for 5 min in ice, the bacterial membranes, proteins and chromosomal DNA are precipitated by the addition of 150 μl of P3 buffer and left for 10 min in ice. After centrifugation (15 min, 15,000 g, 4 ° C), two volumes of absolute ethanol are added to the supernatant. The solution is placed in ice for 15 min, then centrifuged (15 min, 15,000 g, 4 ° C). The DNA pellet is dried under vacuum for 5 min and taken up in 50 μl of sterile double-distilled water. This DNA (5 to 10 μl) is then analyzed after endonuclease digestion and electrophoretic migration.

6) Antibodies.

The anti-ClpG antiserum specific for the major subunit of fimbriae CS31A was obtained in rabbits after three intra-dermal injections carried out at twenty days apart with 250 μg of native CS31A protein in the presence of incomplete Freund's adjuvant.

7) Preparation of protein extracts. The extracts from solid cultures are obtained in the following manner: 1 ml of a preculture carried out in liquid LB medium is poured onto the surface of an LB agar containing the appropriate antibiotics. Petri dishes are placed overnight at 37 ° C

(cover at the top) in a humid atmosphere. The bacterial carpet is collected in PBS buffer, subjected to agitation using a vortex and centrifuged for 10 min at 12,000 g. The supernatant is then immediately analyzed or stored at -20 ° C. Forty Petri dishes per strain were thus prepared in order to recover enough supernatant to ensure all the experiments undertaken during this work. The supernatants of liquid cultures overnight at 37 ° C. are recovered after centrifugation (10 min, 12000 g). They are then concentrated using a NANOSEP microconcentrator (Pall Filtron Laboratories) which retains proteins with molecular mass greater than 10 kDa. The native ClpG subunit is extracted from the strain DH5α (pEH524) cultivated in solid LB medium as described above. The protein concentration of the extract is estimated by the ESL colorimetric method (Exact, Sensitive, Lo interference) using the ESL Protein Assay kit (Boehringer). The ClpG proteins are visualized on a 15% polyacrylamide gel after staining with silver nitrate.

8) Immunodetection of hybrids. a) On colonies.

The colonies obtained after transformation are cultured in an orderly fashion on agar LB dishes and incubated overnight at 37 ° C. The replication is carried out by directly applying a nitrocellulose membrane to the surface of the agar. The membrane is saturated for 1 h at room temperature in blocking buffer, before being incubated for 2 h with the first antibody diluted 1/500 in TBS buffer, 0.1% Tween 20. After three washes of 5 min in TBS buffer, 0.05% Tween 20, the membrane is incubated for 1 h with an appropriate secondary antibody conjugated to peroxidase (Nordic Immunological Laboratories) diluted 1/1000 in TBS buffer, 0.1% Tween 20 After two washes of 5 min with TBS buffer, 0.05% Tween 20 and one washing of 5 min with TBS buffer, the membrane is incubated in a solution containing 1 H ₂ 0 ₂ and 4-chloro-1-naphthol. The colorimetric reaction is stopped by rinsing the membrane in several baths of double-distilled water. b) Electron microscopy.

The bacterial strains are cultured on solid LB medium overnight at 37 ° C. The bacterial carpet is recovered and then immersed in 4% paraformaldehyde, 0.1% glutaraldehyde, 0.2 M PBS (pH 7.4). After centrifugation, the pellet is rinsed three times for 15 min in phosphate buffer and dehydrated in 50%, 75% then 85% ethanol (3 x 15 min). It is then placed in a 1: 1 mixture of medium LR White resin: 85% ethanol for 1 h and impregnated in pure LR White resin (3 × 1 h) before being included in a gelatin capsule. After polymerization, ultra-thin sections

(70 nm thick) are produced using an ultramicrotome (Ultracut E. Reichert) and placed on nickel grids. The latter are saturated for 20 min with PBS, 1% BSA and then rinsed three times with PBS. They are then placed on a drop of anti-ClpG antibody diluted 1/10 for 1 h at room temperature. After three rinses with PBS, the grids are placed on a drop of anti-rabbit IgG goat IgG marked with colloidal gold (10 nm) (Sigma) and diluted 1/50 in PBS. They are rinsed three times in PBS then in distilled water and contrasted with uranyl acetate for 5 min. The observations are carried out on a Philips EM400 80 KV transmission electron microscope. c) Western immunostaining.

The supernatants containing the ClpG-STh hybrids are heated for 5 min at 100 ° C. in Laemmli buffer, before being deposited on a 15% polyacrylamide gel under denaturing conditions. A scale of colored SDS-PAGE markers (Low Range, Biorad) makes it possible to follow the migration during the electrophoresis. The gel, a 0.2 μm nitrocellulose membrane (Biorad Trans Blot) and two filter papers (Biorad) are incubated for 15 min in transfer buffer in the presence of 15% methanol. The gel is deposited on the membrane and the assembly is sandwiched between the two filter papers. The proteins are then transferred onto the nitrocellulose membrane by semi-dry electrotransfer (SemiPhor TE 70, Hoeffer Scientific Instruments) for 1 h at an intensity of 1 mA / cm ² . The membrane is then saturated for 1 h in blocking buffer and incubated for one night with the first specific antibody diluted to 1/500 in TBS buffer, 0.1% Tween 20. The membrane is stained according to the procedure described for immunodetection in if you are on colonies. In order to amplify the response, an antibody specific secondary coupled with biotin (Sigma) (1/1000 dilution) was used. In this case, after incubation for 1 h, the membrane is washed three times for 5 min in TBS buffer, 0.05% Tween 20 and incubated for 30 min with a 1/100 dilution of streptavidin conjugated to peroxidax (250 μg / ml, Sigma) in TBS buffer, 0.05% Tween 20. After three new washes of 5 min each, the reaction is stopped by rinsing the membrane in several baths of distilled water.

9) ELISA.

96-well microtiter plates (Immulon 2, Dynatech) were covered with 100 μl / well of antibodies or antigens diluted in 50 rtiM carbonate buffer (pH = 9.6) and incubated overnight at 4 ° C. . After aspiration of the content of the wells, 10 μl of saturation buffer are deposited and the plates are again incubated overnight at 4 ° C. After three washes with PBS, 0.05% Tween 20 and two washes with PBS, the plates are used immediately or stored at -20 ° C. Three types of ELISA tests have been performed:

- "Classic" ELISA: the quantification of the hybrid proteins present in the supernatants obtained from bacterial cultures in LB agar medium is determined by depositing different dilutions of the extracts in antibody buffer on plates coated with anti-ClpG rabbit antiserum. In parallel, in order to establish a standard curve, different concentrations of ClpG proteins were also deposited. The plates are then incubated for 2 h at 37 ° C., washed twice with PBS buffer, 0.05% Tween 20 and once with PBS buffer. After the addition of 100 μl per well of an anti-ClpG mouse antiserum and an incubation of 90 min at 37 ° C. followed by several washes, 100 μl of goat anti-mouse IgG IgG coupled to peroxidase and diluted to 1/1000 are then added. After a 2 h incubation at 37 ° C and several washes, 100 μl per well of a staining solution prepared extemporaneously are added. The plates are then placed in the dark at room temperature for 20 min before reading the results at an absorbance of 405 nm on a Dynatech MR 5000 reader. To measure non-specific reactions, the absorbances at 405 nm obtained with the supernatant of the DH5 strain (pDSPH524) are deduced from the values obtained with the ClpG -STh hybrids.

- "Double sandwich" ELISA: the supernatants obtained from bacterial cultures produced in LB agar medium are diluted in antibody buffer and distributed on plates covered with the 11C anti-STa monoclonal antibody (8 μg / well). The plates are incubated for 2 h at 37 ° C, then washed. The anti-ClpG antibody diluted 1/500 in antibody buffer is then deposited at the rate of 100 μl per well. After incubation for 90 min at 37 ° C and several washes, 100 μl of anti-rabbit IgG goat IgG coupled to peroxidase and diluted to 1/1000 are added. After a 2 h incubation at 37 ° C. and after several washes, the plates are stained according to the procedure described above. The use of the supernatant of the strain DH5α (pEH524) makes it possible to measure the non-specific reactions.

- ELISA "competition": the E. coli ST EIA kit

(Oxoid) including a 96-well plate coated with the synthetic toxin STp is used to detect the presence of STa toxins in protein extracts and to quantify the hybrids. This kit is based on a competitive reaction between the toxin fixed on the plate and the exogenous toxin contained in the samples in the presence of a given quantity of anti-STa antibodies. The protein extracts diluted in PBS buffer are distributed at the rate of 100 μl per well. A bacterial supernatant provided by the kit serves as a positive control. Different known amounts of purified STp toxin (Calbiochem) are also deposited to establish a standard curve. An anti-STa monoclonal antibody solution included in the kit (10 μl) is added to each well. After one 90 min incubation at 37 ° C, the plates are then washed twice with PBS buffer, 0.05% Tween 20 and once with PBS buffer. One hundred μl of a substrate solution included in the kit is then added to each well and the plates are placed for 30 min, protected from light. The colorimetric reaction is quantified by reading the absorbance at 490 nm, after the addition of 1.5 N sulfuric acid.

10) In vitro adhesion to cell cultures.

Caco-2 cells are cells derived from human colon carcinoma. They retain the characteristics of enterocytes from the small intestine when they are differentiated. This cell line expresses the ClpG receptor and that of STa (GC-C). Intestine-407 cells (Flow Laboratories Inc.), derived from jejunum and human embryonic ileum, also express the ClpG receptor. The Intestine-407 and Caco-2 cells are cultured under a partial pressure of 5% of CO ₂ at 37 ° C., in flasks (Falcon 3084 boxes, Becton Dickinson Lalware, Oxnard, Calif.) Containing 25 ml of medium respectively. of modified Eagle (MEM) and 25 ml of minimum medium of Eagle modified by Dulbecco (DMEM) (Flow Laboratories Inc.). The media are supplemented with 10% to 20% of fetal calf serum (SVF), 1% of amino acids non-essential, 200 U / l of penicillin and 50 mg / 1 of streptomycin. After confluence, the cells are washed twice with PBS without Ca ²⁺ or Mg ²⁺ and then trypsinized for 5 min at 37 ° C. The cells, thus recovered, are rapidly diluted in 20 ml of culture medium, then centrifuged for 5 min at 1200 g. The pellet is taken up with 4 ml of culture medium. After counting on a Malassez cell, a 24-well Falcon dish (Becton Dickinson Lalware, Oxnard, Calif.), With a coverslip at the bottom of each well, is inoculated with 1 ml of culture medium containing 2.10 ⁵ to 4.10 ⁵ cells per well . The Intestine-407 cells are used at confluence while the Caco-2 cells are used 15 days later, the time necessary for their differentiation. Before the adhesion test, the cells are washed once with PBS

(pH = 7.2) without Ca ²⁺ or Mg ²⁺ . A suspension of 10 ⁸ bacteria / ml of culture medium containing 1%

(weight / volume) of D-mannose is added to the cell culture and the whole is incubated for 3 h at 37 ° C. After three washes with PBS, the cells are fixed for 10 min in methanol and stained for 20 min with a 20% Giemsa solution. The coverslips are then glued to a slide, Caco-2 cells upwards, and Intestine-407 cells between the slide and the coverslip. After drying for 1 h at room temperature, the slides are examined under a photon microscope at the x100 objective, in immersion.

11) Enterotoxicity. The bacterial strains, containing the various recombinant plasmids, are cultured overnight at 37 ° C. on agar LB dishes. After centrifugation (10 min, 1600 g), the supernatant is separated from the bacterial pellet and stored. The pellet is suspended in PBS, centrifuged (10 min, 1600 g) and washed a second time with PBS. The washed pellet is finally taken up in a volume of PBS corresponding to the starting volume. The bacteria count is determined after one night at 37 ° C. by spreading of various dilutions of the bacterial suspension on MacConkey lactose agar (DIFCO) selective medium containing the appropriate antibiotics. Half of the suspension of washed whole bacteria is kept. The other half is sonicated and then centrifuged at 12,000 g for 5 min. The supernatant recovered corresponds to the cell lysate containing the intracellular proteins.

The enterotoxic activity of supernatants, washed whole bacteria and lysates is examined in newborn mice. Three-day Swiss 0F1 mice from different litters are separated from their mother immediately before use, randomly mixed and separated into groups of at least three mice. An aliquot of 0.1 ml of each sample is delivered directly to the stomach of the mice using a flexible plastic tube 3 cm long and 0.3 mm in internal diameter connected to a 1 ml syringe fitted with a 26 G needle. Three hours after injection, mice are suffocated by prolonged exposure to chloroform vapors. The entire intestine is removed from the abdomen, weighed, and the gut weight / carcass weight ratio (I / C) is then calculated for each mouse (Figure 18). An I / C ratio ≥ 0.09, corresponding to a significant accumulation of intestinal fluids, is interpreted as positive enterotoxicity.

12) Measurement of β-galactosidase activity. The β-galactosidase activity present in the extracellular medium was determined by the use of the β-galactosidase Enzyme Assay System kit (Promega) according to the recommendations of the supplier. The bacterial cultures are carried out on agar medium. The bacterial carpet suspended in PBS buffer is adjusted to 5.10 ¹⁰ bacteria / ml. This bacterial suspension is centrifuged (1600 g, 5 min) and 200 μl of supernatant are mixed with 100 μl of lysis buffer (kit). Then 150 μl of the mixture are added to 150 μl of 2X assay buffer (Annex 11) which contains the ONPG substrate (0-ni trophenyl -β-D- galactopyranoside). The reaction mixture is incubated for 30 min at 37 ° C. The reaction is stopped by adding 500 μl of 1M sodium carbonate. After mixing, the absorbance is measured at 420 nm. To carry out the positive control, 200 μl of culture containing 5.10 ¹⁰ bacteria / ml are directly mixed with 100 μl of lysis buffer before centrifugation (1600 g, 5 min). The supernatant is then treated as described above.

13) Quantification of thiol groups. The method for quantifying free thiol groups in the supernatants obtained from bacterial cultures produced in agar medium containing 5 inM β-mercaptoethanol is based on the use of a commercial test (kit "Thiol and Suifide Quantitation", Molecular Probes ™, Netherlands) for quantifying sulfhydryl groups. Thiol groups reduce an inactive disulfide derivative of papain

(papain-SSCH ₃ ) stoichiometrically releasing the active enzyme (papain-SH). The activity of this enzyme is then evaluated in the presence of the chromogenic substrate of papain, N-benzoyl-L-arginine p-nitroanilide (L-BAPNA), after the release of the chromophore p-nitroanilide which is detected by measuring the absorbance at 410 nm. This colorimetric test is 100 times more sensitive than that performed with the Ellman reagent and can detect up to 0.2 μM of thiols. Different concentrations of an L-cysteine solution (0.1 mM) were used to establish a standard curve.

14) Determination of the oxidation-reduction state of the fusion proteins. In order to detect the formation of disulfide bridges in the STh portion of the hybrid, the state of oxidation-reduction of the fusion proteins was determined in the intracellular and extracellular compartments of the recombinant bacteria. The state of redox proteins in the extracellular medium was evaluated as follows: 25 ml of liquid LB medium are inoculated at OD _550nm = 0.05 from an overnight preculture and incubated at 37 °. C with stirring (225-240 rpm). At OD _550nm = 0.5 the culture is centrifuged for 10 min at 1600 g. One milliter of supernatant is removed and stored at -20 ° C to detect possible enterotoxicity in the mice test. The rest of the supernatant is immediately treated with TCA (trichloroacetic acid) at a final concentration of 5% to concentrate the proteins and to avoid any oxidation of the cysteines during sample handling. If necessary, the supernatants are incubated for 15 min at 37 ° C. in the presence of 100 mM of dithiothreitol (DTT) which reduces all the disulphide bridges of the STa before precipitation by TCA.

The determination of the protein redox state in the intracellular medium was carried out by precipitating, with a final 5% TCA, 25 ml of a bacterial culture in the exponential phase. The reduced control of the intracellular medium was obtained by directly cultivating the strain in the presence of 5 mM of DTT. After incubation for 1 h in ice and centrifugation for 10 min at 16,000 g at 4 ° C, the pellets are washed with 1 ml of cold acetone and centrifuged. The protein pellet is then dried in the open air and taken up in 200 .mu.l of solution containing 1 AMS (4-acetamide '-maléimidylstilbène-2, 2' - sulphonic acid disulphide, ^s' els disodium) prepared extemporaneously. AMS chemically binds to sulfhydryl groups released after the reducing action of DTT on oxidized cysteines. The fixation of each AMS molecule on the reduced proteins makes it possible to increase their molecular weight by 490 daltons and to separate them on an SDS-PAGE gel at 12% in the absence of reducing agents and thermal denaturation. The oxidized and reduced forms are visualized by "Western" immunostaining. The supernatants (500 μl), obtained in stationary phase from cultures produced on agar medium, were treated according to the same procedure as that used for the supernatants originating from cultures, in exponential phase, carried out in liquid medium.

16) Insulin reduction test. This test makes it possible to verify the capacities of the reduced STa to exchange its thiol groups with oxidized insulin and is based on the measurement of the appearance of a white precipitate due to the release of the B chain of insulin after reduction of its disulphide bridges.

A stock solution of insulin (Sigma) is prepared at a concentration of 10 mg / ml (1.67 mM) by diluting 50 mg of insulin in 4 ml of 0.05 M TrisHCl buffer pH = 8, then adjusting the pH a first time at 2 or 3 with 1 M HCl and a second time at 8 with 1 M NaOH. The volume is then adjusted to 5 ml with double distilled water. The insulin solution must be perfectly clear before being stored at -20 ° C. 1 mg / ml insulin solution

(0.167 mM) is then prepared extemporaneously in 0.1 M phosphate buffer, 2 mM EDTA from the mother solution of insulin. Then 450 μl of this solution are mixed in a spectrophotometric tank with the sample to be tested in a final volume of 600 μl. The reaction starts when 2 μl of 100 mM DTT is added. The whole is mixed vigorously then the tank is placed in the spectrophotometer. The time when precipitation occurs, defined by the increase in absorbance at 650 nm (At _550nm ) by 0.02 above the baseline, and the precipitation rate at 650 nm, defined as the increase absorbance at 650 nm during the one minute time interval (ΔA _550nm x min ^-1 ) for absorbances between 0 and 1, are then determined.

16) Immunogenicity of hybrids.

The supernatants were prepared as described above. The recombinant live bacteria are prepared in the same way before each series of inoculation: after centrifugation of the bacterial mat (5 min, 1600 g), the bacterial pellet is taken up in PBS. The bacteria are listed on MacConkey lactose agar selective medium before immunization of the animals. a) Antibody production in rabbits.

A New Zealand breed rabbit about

11 weeks is used for each type of antigen. The inoculation protocols are determined according to the nature of this antigen. Thus, each rabbit receives either 500 μl of supernatant in the presence of incomplete Freund's adjuvant (v / v) intra-dermally on days D ₀ , D ₂₁ , D ₂₈ and D ₃₅ , i.e. four injections one week apart of a suspension of 5.10 ⁸ bacteria live (≈ 500 μl) intravenously. The blood is collected by section of the jugular vein ten days after the last booster. b) Antibody production in mice and protection test. - Intra-peritoneal route: 200 μl of a suspension containing 2 to 4.10 ⁸ live bacteria diluted in PBS buffer or 200 μl of supernatant are inoculated into groups of 5 non-consanguineous Swiss OF1 mice (Iffa Credo, France) aged eight at twelve weeks on days D ₀ , D ₇ and D _i4 . The serum of each mouse is collected individually one week after the last booster by intra-cardiac puncture under general anesthesia (Vétranquil + Imalgène, Rhône Mérieux).

- Oral use: groups of four female mice, deprived of water for at least 5 h, receive an inoculum of 10 ¹⁰ live bacteria in the presence of 3% NaHC0 ₃ on days D ₀ , D _x , D ₂ before being mated on day D ₁₄ then re-immunized on days D ₂₈ , D ₂₉ , D ₃₀ . After parturition, the blood, the whole intestine and the faeces are removed. The three-day-old newborn mice are then tested with toxin to assess vaccine protection. For this, each litter of mice is divided into two groups. The first group receives 12.5 ng (1.6 U) of purified STp toxin (Calbiochem) and the second group, 2 U of supernatants containing the homologous hybrids produced by the recombinant bacteria which were used during the primary immunization. c) Preparation of the extracts and detection of anti-ClpG and anti-STa antibodies. - Serum: the immune blood is incubated for at least 4 h at 37 ° C and then centrifuged for 10 min at 3000 g. The serum is stored at -20 ° C.

- Intestinal contents: the entire intestine is ligated at one end, then 200 μl of washing buffer are introduced through the other end. After ligation of the latter, and by light pressure along the intestine, the buffer is transferred several times from one end to the other. The intestinal contents are collected, centrifuged (10 min, 16,000 g) and the supernatant is stored at -20 ° C.

- Feces: they are crushed in the washing buffer (200 μl / 100 mg faeces) and then centrifuged for 10 min at 16,000 g. The supernatant is stored at -20 ° C.

Serial dilutions of the different samples to be tested are made in buffer antibodies and distributed at a rate of 100 μl / well on the ELISA plates coated either with the ClpG antigen (Iμg) or with the STa toxin (kit E. coli ST EIA) to detect the anti-ClpG antibodies and the anti antibodies respectively Sta. The antigen-antibody complexes are detected according to the so-called "classic" procedure described in paragraph 9 above, after incubation of the plates with goat anti-rabbit IgG or anti-mouse Ig (A or G) goat immunoglobulins (Nordic Immunological Laboratories) coupled with peroxidase. The antibody titers, expressed in log ₁₀ of the inverse of the dilution, are calculated from the absorbance values read at 405 nm deduced from the values obtained after immunization of the animals either with live DH5α bacteria (pDSPH524), or with the corresponding supernatant.

17) Neutralization test.

Each immune serum is incubated for 18 h at

4 ° C in the presence of either purified STp toxin (Calbiochem), or ClpG-STh hybrids (supernatants), or live E. coli B41 bacteria producing

Please. Then 100 μl of the mixture are administered intragastrically to newborn mice 3 days old. After 3 h of incubation, the intestine weight / carcass weight ratio is then calculated (see Equipment and Methods, § 11). Anti-STa antibodies described in the literature as being neutralizing were used as positive controls. These antibodies are: the polyclonal antibodies G108-8 and G104-6 and the monoclonal antibodies ST1-.3, ST: G8 and 11C.

B - RESULTS.

I - Construction and characterization of ClpG-STh hybrids.

1) Choice of regions and insertion of STh into ClpG.

Recent work has shown that certain regions of the major ClpG subunit can be modified without harming the biogenesis of the polymer CS31A. Thus, epitopes of different types and sizes, fused at the NH ₂ -terminal end or inserted into the variable central region V3 of ClpG (amino acids 186 to 221), are exposed on the surface of the bacteria. These sites were therefore chosen to introduce the STh enterotoxin into ClpG. The C-terminal end of ClpG also appeared interesting to us because numerous works have shown that blocking of the C-end STa terminal could be damaging for both secretion and antigenicity of hybrids.

It has been chosen to insert different "linkers" between ClpG and STh to have the maximum chance of obtaining hybrid proteins which are expressed and secreted on the bacterial surface. Indeed, various studies have shown that the introduction of certain amino acids, in particular proline residues, between STa and the carrier protein increases the flexibility and accessibility of STa in the hybrid molecule.

The plasmids and oligonucleotides used in this work are presented respectively in Tables 2 and 3 and the various constructions described in detail below are shown diagrammatically in FIG. 1. FIG. 1 represents the structure of the fusion proteins ClpG-STh. In Figure 1, the amino acids are numbered from the N-terminus of the mature protein ClpG. The nucleotide and peptide sequences relating to ClpG are in bold characters. The STh portion of the fusion protein is represented by a colored frame: ST15 (YCCELCCNPACTGCY), ST16 (NYCCELCCNPACTGCY) and ST19 (NSSNYCCELCCNPACTGCY) correspond respectively to the last 15, 16 and 19 amino acids of the mature toxin STh. Restriction sites are underlined and those lost after cloning are in parentheses. The asterisks denote the restriction sites created by site-directed mutagenesis. The vertical arrow represents the signal peptidase cleavage site. The nucleotide sequence of clpG is shown diagrammatically by one frame and the cloning vectors by another frame.

To ensure better flexibility of the hybrid at the ClpG-STh join, the STh toxin was inserted at the N-terminal end of ClpG and separated from the carrier protein by a "linker" of 6 amino acids (PADPHA ). Thus, the plasmid pSTN24 was obtained by insertion, into the Sp h 1 site of pHPC0838, of the double stranded oligonucleotide Sth72N5 / Sth72N3. This duplex codes for the last fifteen amino acids of the STh enterotoxin. After cloning, only the 5 'Sphl site is preserved. The initial cleavage site (AHA: WT) of the signal peptide was converted into a site (AHA: AD) more frequently present in E. coli.

For the same reasons as those explained above, a "linker" containing two prolines and a glycine (NPDNPG) was placed between ClpG and STh. The plasmid pSTC22 was then produced by cloning of the double-stranded oligonucleotide Sth69C5 / Sth73C3 coding for the last 16 amino acids of STh between the Hpal and Xbal sites of pHPC0838. The Xbal site was lost during cloning. The STh toxin is thus introduced after the Tyr ²⁵⁶ residue of ClpG. In order to confirm the importance of such a "linker" between the carrier protein and the toxin STa in the secretion and the antigenicity of the hybrid, the plasmid pSTCD17 was produced by deletion of the Hpal / Smal fragment of pSTC22. In this case, ClpG and STh are no longer separated except by a glycine residue.

It was considered that the choice of the peptide comprising amino acids 46 to 56 (NKSGPESMNSS) of natural pro-STh as a junction between ClpG and STh could represent the best joint to promote conformational folding close to that of the free toxin. For this, the plasmid pProSTC28 was obtained by inserting between the Hpal and Smal sites of pSTC22 a double-stranded oligonucleotide (Sth33PC5 / Sth33PC3) coding for amino acids 46 to 53 (NKSGPESM) of the protein Pro-STa ₃ followed by the first three amino acids (NSS) of the mature STh toxin. A glycine residue separates the tripeptide NSS and the rest of the STh sequence.

STh was also introduced into the variable central region V3 of the ClpG protein. For this, the plasmid pSTVIOS was created by insertion of the synthetic double-stranded oligonucleotide Sth69V3 / Sth69V5, coding for the mature whole STh toxin, in the S el site of pSXIOS. Another type of insertion into V3, represented by the plasmid pSTVD41, was obtained by deletion of the BazπHI / BglII fragment of pSTVIOS coding for amino acids 204 to 215 of ClpG (Figure 1).

Six different hybrids comprising a foreign sequence of 17 to 28 amino acids were thus genetically constructed. One chimera has STh at the N-terminus of ClpG, two chimeras have STh in the central region V3, and three others, at the C-terminus.

To overcome a possible instability or toxicity of the genetic constructs due to a too high level of expression linked to the number of multiple copies of the plasmids, it has been tried to reintroduce into the clp operon carried by the plasmid pEH524 the fusion genes clpG- sth in place of the wild-type clpG gene. FIG. 2 represents the subcloning of the different fusion genes of the clp operon. In FIG. 2, the blocks represent the different genes of the clp operon contained in the 8.5 kb HindIII / BcoRI fragment. The blackened part of each block indicates the signal sequence. The horizontal arrow indicates the direction of the transcription. The names of the genes are indicated below each block. The clpG gene has been replaced in the clp operon by the various fusion genes originating from the plasmids pSTN24 (1), pSTC22 (2), pProSTC28 (3) and pSTVD41 (4). Thus, the plasmids pEHSTN24 and pEHSTC22 were obtained by replacing the Swal / Hpal fragment of pEH524 respectively by the fragments __7coRV / _3cI136II of pSTN24 and pSTC22. The plasmids pEHProSTC28 and pEHSTVD41 were constructed by substituting the Muni / Hpal fragment of pEH524 respectively by the Mi.nI/.Ecll3611 fragments of pProSTC28 and pSTVD41. Numerous attempts to place the fusion genes carried by pSTC17 and pSTVIOS in the clp operon have been unsuccessful. Consequently, these two plasmids were the only ones to be transcomplemented with the plasmid PDSPH524 (Table 2).

All the genetic constructs were verified by sequencing after purification of the recombinant DNAs.

2) Detection of hybrids in the extracellular medium.

The presence of the hybrid proteins in the extracellular medium was studied using supernatants derived from cultures in solid or liquid medium of the recombinant DH5α bacteria harboring the plasmids presented in table 2.

FIG. 3 represents the analysis carried out by "Western" immunostaining under denaturing conditions (SDS-PAGE), using the anti-ClpG antiserum (FIG. 3A and B) and different monoclonal antibodies (11C, 20C1) directed against the STa toxin (Figure 3C and D).

The antigenicity of the fusion proteins was analyzed from the supernatants of the recombinant DH5α bacteria cultivated in liquid medium (A) or agar (B, C, D). The liquid culture supernatants are concentrated ten times (A) and the supernatant of the DH5 bacteria (pEH524) obtained on agar medium is diluted five times (B lane 1). Immunodetection is carried out after migration on 15% polyacrylamide gel under denaturing conditions either with the anti-ClpG polyclonal antibody (A, B) or with the anti-STa 11C (C) or 20C1 (D) monoclonal antibodies. In the latter cases (C, D), the colorimetric reaction is amplified using a secondary antibody coupled to biotin which is recognized by streptavidin conjugated to peroxidase. The horizontal arrow indicates the position of the ClpG protein (~ 29 kDa).

The electrophoretic profiles of the proteins obtained from cultures on agar (Figure 3B) or liquid (Figure 3A) medium are identical. However, the comparison of the production of hybrids under different culture conditions shows that, for the same volume of supernatant, the quantity of recombinant proteins collected under liquid culture conditions (FIG. 3A) is lower than under culture conditions on agar medium (Figure 3B). Consequently, the supernatants obtained from cultures on agar medium were therefore preferably used for the rest of this study.

All chimeric proteins having STh inserted at the N-terminal (Figure 3B, lane 2) or C-terminal (Figure 3B, lane 3 to 5) end of the ClpG protein, are secreted into the extracellular medium. The hybrids expressed by pEHSTN24 (lane 2) and pEHSTC22 (lane 3) migrate in the form of a single protein with a molecular mass higher than that of ClpG (approximately 29 kDa, lane 1), probably resulting from the presence of STh. In contrast, the electrophoretic profile of the constructions pEHProSTC28 (lane 4) and [pSTC17 + pDSPH524] (lane 5) shows two bands reacting with the anti-ClpG antiserum (FIG. 3B). The upper band would correspond to the entire hybrid while the lower band, located at the level of the native ClpG subunit, would represent a cleaved derivative of the hybrids (FIG. 3A and B, lanes 4 and 5). No hybrids corresponding to STh insertions were observed in two sites (amino acids 202-204 and 204-215) of the variable central region V3 of ClpG ([pSTVl0S + pDSPH52] and pEHSTVD41) which indicates that these two sites are not compatible with the production and / or secretion of the hybrids (not shown). For this reason, these constructions have not been the subject of other studies during this work.

The supernatants were then tested for their reactivity towards the 11C and 20C1 monoclonal antibodies directed against the STa toxin (FIG. 3C and D). The hybrid expressed by pEHSTN24 (lane 2) does not react with any of these antibodies, which suggests a poor accessibility of STh to these antibodies, probably following a change in conformation of the toxin after fusion with the carrier protein. On the contrary, the protein coded by pEHSTC22 (lane 3) and the uncleaved protein expressed by pEHProSTC28 (lane 4) are recognized by the two anti-STa antibodies. On the other hand, the degradation product of pEHProSTC28 (lane 4) is not detected with the anti-STa antibodies, which suggests that part of the hybrid encoded by this plasmid is cleaved on the C-terminal side of ClpG releasing thus the STh portion. On the other hand, the two proteins expressed by [pSTC17 + pDSPH524] react with the two anti-STa antibodies (lane 5). Cleavage of this hybrid could therefore occur in the N-terminal region of ClpG. The non-cleaved proteins expressed by pEHSTC22, pEHProSTC28 and [pSTC17 + pDSPH524] therefore represent the whole hybrids since they are recognized both by anti-ClpG antibodies and by anti-STa antibodies. FIG. 4 reports the presence of the hybrids in the culture supernatants determined by “double sandwich” ELISA. The hybrids in the culture supernatants produced on agar medium are detected using anti-ClpG and 11C antibodies directed respectively against the ClpG and STh portions of the hybrid. The supernatant DH5α (pEH524) containing ClpG but no STh toxin is used as a negative control and makes it possible to measure the aspecific reactions. Initially, the recombinant proteins are fixed by the STh portion of the hybrid to the anti-HC antibodies of mice fixed on the microtitration plate. The free ClpG part of the hybrid is then recognized by the anti-ClpG rabbit antiserum used as secondary antibody. The presence of the ClpG-STh hybrids in the samples is detected after the addition of a goat anti-rabbit IgG antiserum coupled with peroxidase. The results show the presence of the recombinant proteins in all the supernatants since they are recognized by both anti-STa and anti-ClpG antibodies. These results confirm those observed by "Western" immunostaining (FIG. 3) and prove that the lack of reactivity of the hybrid expressed by pEHSTN24 with respect to the 11C and 20C1 monoclonal antibodies (FIG. 3) is not due to the absence of the STh portion. In fact, these hybrids are detected by ELISA in the diluted samples, which suggests a low affinity of the antibodies for the protein.

The presence of recombinant proteins in the supernatants could result from lysis of the bacteria. To ensure that this was not the case, we measured the β-galactosidase activity in the culture supernatants and the E. coli DH5α lysates hosting both the pSK ⁺ vector expressing β-galactosidase and the 'one of the recombinant plasmids derived from pEH524. Β-galactosidase activity was mainly associated with cell lysates and negligible in the supernatants.

All of these results demonstrate the high permissivity of the N- and C-terminal ends of the ClpG protein for the expression and secretion of STh in the extracellular medium. Conversely, no recombinant protein could be detected in the culture supernatants corresponding to constructs having an ST molecule in the internal region V3 of ClpG. These data suggest that STh fused to ClpG must have one of its free ends to be secreted into the extracellular medium. Consequently, the non-permissive constructions were not used for the rest of the work. These results also show that the coupling of the toxin to the C-terminal of ClpG makes it possible to conserve The antigenicity of STh in its new molecular environment and this independently of the "linker" placed between ClpG and STh. On the other hand, although the fusion of STh at the N-terminal end of ClpG does not affect the secretion of the hybrid, it decreases its antigenic properties.

3) Quantification of hybrids.

First, the quantification of the hybrids was carried out by ELISA on microtiter plates covered with the anti-ClpG antiserum. FIG. 5 reports the quantification by ELISA of the fusion proteins from their ClpG portion:

- 5A: The quantity of hybrids present in each supernatant is estimated by reference to the standard curve representing the absorbance at 405 nm (A405nm) as a function of different concentrations of ClpG. Inlay: concentration and production of ClpG. The bacteria count is determined on MacConkey lactose selective agars containing the appropriate antibiotics.

- 5B: The supernatants come from cultures on agar medium from the strain DH5α hosting the plasmids indicated.

FIG. 6 reports the quantification by ELISA of the fusion proteins from their STh portion. - 6A: The quantity of hybrids present in each supernatant is estimated by reference to the standard curve representing the absorbance at 490 nm (A490nm) as a function of different concentrations of purified STa toxin (Calbiochem). Only A490nm between 0.2 and 1.2 allow quantification of the toxin. Inlay: concentration and production of STa. The bacteria count is determined on MacConkey lactose selective agars containing the appropriate antibiotics. The B41 strain is used as a reference strain producing the natural STp toxin, na, not applicable: indicates a lack of affinity of anti-ST antibodies.

- 6B: The supernatants are from cultures in agar medium of the strain DH5α harboring the plasmids indicated. The amount of STa toxin present in the sample is inversely proportional to the absorbance at 490 nm.

The absorbance measurements at 405 nm (A _405nm ) of different deposits containing known amounts of ClpG made it possible to establish a standard curve (FIG. 5A). With reference to this curve, the value of A _405rraι corresponding to the undiluted samples (Figure 5B) made it possible to express the amount of hybrids present in the supernatants in μg / ml and in μg / 10 ¹⁰ bacteria (Figure 5A). Measurement recorded with the undiluted supernatant of the strain DH5α

(pDSPH524) (negative control) makes it possible to evaluate the non-specific reactions. This value is deduced from the values obtained for the different samples tested. These data indicate that the amount of ClpG produced by the recombinant bacteria appears much lower (10 to

200 times less) than the amount of ClpG produced by DH5α

(pEH524). This is all the more true since certain values obtained are probably overestimated due to the cleavage of the hybrids (Figure 3). Thus, the fusion of STh to the ClpG subunit decreases the production and / or the secretion of ClpG.

In a second step, the quantification of the hybrids was carried out by competition, for a given quantity of an anti-STa antibody, between the STa toxin fixed on the microtitration plate (kit E. coli ST EIA, Oxoid) and the toxin STa present in the sample to be tested. Consequently, the higher the amount of STa toxin, the more it will bind anti-STa antibodies which, therefore, can no longer interact with the STa fixed on the ELISA plate. The colorimetric reaction linked to the quantity of anti-STa antibodies fixed on the microtiter plate is then weak. Conversely, a high absorbance value at 490 nm (A _490nm ) is representative of a small amount of toxin present in the sample. of the known amounts of purified STp toxin (Calbiochem) were used to establish a standard curve (Figure 5A). Absorbance values at 490 nm greater than or equal to 1.2 indicate either the absence of STh (and therefore ClpG-STh hybrids) or the lack of affinity of the STh part towards antibodies. anti-STa. The values of A _490nm , obtained with different dilutions of supernatants containing the hybrids, are shown in FIG. 6B. Only the A _490rm obtained with the DH5 supernatant (pEHSTN24) invariably remain high regardless of the dilution of the sample. This lack of reactivity reflects a lack of affinity of the anti-STa antibodies rather than an absence of the hybrid in the supernatant since a positive response in “Western” immunostaining (FIG. 3) and in ELISA (FIG. 4) is observed. .

The quantification technique on ELISA plates coated with the STa toxin seems more reliable and more representative of the actual amount of secreted hybrids. Indeed, the STa toxin is a small molecule very resistant to different proteases and therefore not very prone to proteolytic cleavage. Consequently, the number of STa toxin molecules detected is directly proportional to the number of hybrid molecules exported in the extracellular medium, even if a cleavage in the ClpG part of the hybrid occurs after secretion. However, as has been shown previously (FIGS. 3, 4 and 6), the existence of a difference in affinity of the STh portion of the hybrids with respect to the anti-STa monoclonal antibody used cannot be excluded. If we consider the values thus obtained, we can see that only the hybrid expressed by [pSTC17 + pDSPH524] is secreted in a lower quantity than the toxin produced by the wild strain B41 (FIG. 6A). The data obtained by quantification from the ClpG portion (FIG. 4A) show a 1/3 ratio between the quantities of hybrids expressed respectively by pEHSTN24 and pEHSTC2. Since these two hybrids are not cleaved (Figure 3), and if it is assumed that this ratio is retained after the quantification based on the STh portion, the hybrid expressed by pEHSTN24 would then be secreted at the rate of 0.1 μg / 10 ¹⁰ bacteria i.e. 0.8 μg / ml.

4) Production of hybrids on the bacterial surface. The presence of hybrid CS31A fimbriae on the surface of the recombinant bacteria was analyzed in situ by the ability of the bacteria to adhere to Caco-2 and Intestine-407 cells in culture and by immunostaining with colloidal gold. The bacterial strains DH5α (pEH524) and DH5α

(pDSPH524) were used respectively as a control positive and negative control. None of the chimeric strains are capable of adhering to intestinal cells in vitro or of binding polyclonal anti-ClpG antibodies. In view of these data, it is clear that the ClpG-STh fusion proteins are not assembled in the form of polymers on the surface of the recombinant bacteria, but that they are directly secreted in the extracellular medium in the form of free monomers. Furthermore, we have never been able to demonstrate multimeric forms in SDS-PAGE under non-denaturing conditions as in the case of native CS31A fimbriae. Thus, the coupling of STh to the ends of ClpG prevents the formation of hybrid CS31A fimbriae, which explains the inability of the recombinant bacteria to adhere to Caco-2 and Intestine-407 cells. However, this does not rule out the possibility that the modified mono eric forms of ClpG can still recognize their receptors and thereby adhere to these cells in vitro.

5) Enterotoxicity.

One enterotoxic unit (U) is the smallest amount of STa toxin needed to cause an accumulation of intestinal fluids. This minimum effective dose (DEM) is evaluated in three-day-old newborn mice by measuring the weight ratio of intestine / carcass weight (I / C) after inoculation of different dilutions of each test sample. FIG. 7 reports the determination of the minimum effective dose (DEM) of the living cells (A) and of the supernatants (B, C) of culture of E. coli strains cultivated in solid (A, B) or liquid (C) medium. The dotted line represents the toxicity threshold.

The lower toxicity threshold was set by experience at an I / C ratio equal to 0.09 by intragastrically injecting different quantities of purified STa toxin (Calbiochem). Figure 8 reports the determination of the minimum effective dose (DEM) of toxin required to produce a positive response in the newborn mice test. Groups of 3 to 10 mice are inoculated with 100 μl of PBS containing 1 to 100 ng of purified and marketed STp toxin (Calbiochem). The dots represent the value of the gut weight / carcass weight (I / C) ratio for each mouse. The horizontal lines indicate the average I / C ratios for each batch of mice. The values of the I / C ratios> 0.09 (dotted line) are considered to be positive.

The minimum dose of toxin producing an accumulation of intestinal fluids is 8 ng of pure STp toxin. Table 4 below reports the enterotoxic activity. The biological activity of live recombinant bacteria and the corresponding supernatants obtained in liquid medium or agar was studied in vivo in newborn mice, and compared to the activity of the natural STp toxin secreted by the B41 strain.

Table 4

^a I / C, intestine weight / carcass weight ratio. Each value represents the average of the I / C ratios of 3 to 15 young mice (± standard deviation). A response> 0.090 is considered positive. ^b %, percentage determined by the ratio of the number of sick mice / total number of mice tested. ^c U, enterotoxic unit which is defined as the enterotoxic activity corresponding to the minimum effective dose (DEM) giving an I / C ratio equal to 0.090. For the supernatants, the DEM corresponds to the greatest dilution giving an I / C ratio equal to 0.090. The reverse of this dilution represents the titer of the supernatants in enterotoxic units. For whole cells, the DEM is the minimum number of bacteria giving an I / C ratio equal to 0.090. The calculation of the DEMs is explained in Figure 7. ^d positive control of toxicity ^e negative control of toxicity ^f nd, not detectable.

All of the I / C ratios obtained with the supernatants of liquid or solid cultures and with the live bacteria previously washed with PBS shows that all the hybrids are released in vitro and in vivo in the extracellular medium with a strong enterotoxic activity. . The biological activity of whole cells is variable according to the constructions with, perhaps, a slight superiority for DH5α (pEHProSTC28) whose response is similar to that of the strain B41. Supernatants from the recombinant bacteria have an enterotoxic activity comparable to that of the supernatants derived from the B41 strain regardless of the culture conditions (Table 4). However, the hybrids produced in liquid culture appear 25 to 50 times more toxic than the homologous hybrids synthesized in agar medium. As we will see later, it is likely that this difference in toxicity reflects a difference in the oxygenation of bacterial cultures. All of these results designate the fusion proteins coded by pEHSTN24 and pEHProSTC28 as being the most toxic. These data are in agreement with the values of the production rates (μg ST / 10 ¹⁰ bacteria) measured in competitive ELISA from the culture supernatants produced on agar medium (FIG. 5A). This correlation confirms that the enterotoxic activity is proportional to the quantity of ST toxin detected in the samples and that the specific activity of the hybrids (8 ng / U to 35 ng / U) is not very different from that of the toxin. STa produced by the strain B41 (16 ng / U).

In order to assess the thermostability of the biological activity of these hybrids, each supernatant collected on solid medium was adjusted to 2 U enterotoxic and heated between 90 ° C and 110 ° C for 15 min. The results in Table 5 below indicate that the hybrids are at least as heat resistant as the STp toxin.

Table 5

The supernatants were prepared from solid medium and titrated to 2 U as shown in Figure 7. The purified STp toxin (2 U) (Calbiochem) is used as a positive control. ^a Each value represents the average of the I / C ratios of three to eight mice. %, the value expressed as a percentage indicates the ratio of the number of sick mice / total number of animals tested. An I / C ratio> 0.09 and a percentage of sick mice greater than 50% are considered positive. ^b NT, not tested. Data on enterotoxicity of hybrids

ClpG-STh demonstrate that neither the coupling with ClpG nor the gastrointestinal environment alter the biological activity of the STh portion, and that the latter adopts a spatial conformation close to that of the free natural enterotoxin. The accumulation of intestinal fluids in newborn mice, after inoculation of live bacteria previously washed with PBS, suggests the existence of an in vivo secretion of the hybrids since the previous results underlined the absence of ClpG-STh polymers at the surface of the bacteria and the absence of toxicity of the cell lysates obtained from live bacteria washed with PBS.

6) Role of the clp operon in the secretion and enterotoxicity of the hybrids.

The absence of hybrid fimbriae on the bacterial surface suggests that the biogenesis system of

CS31A is not necessary for the secretion of chimeric proteins into the extracellular medium. However, the release of ClpG-STh hybrids, in vi tro and in vivo, under a biologically active form is unexpected in the light of the results described by other work which highlights the very low export of STa coupled to a heterologous carrier protein. Our observations could then be explained by the presence, in the clp operon, of genes other than clpG capable of encoding proteins capable of intervening in secretion and toxicity.

(formation of disulfide bridges) of hybrids. In order to verify this hypothesis, the toxic activity of different cell extracts from bacteria harboring plasmids containing only the clpG-sth fusion genes was measured and is reported in Table 6 below in which: ' ^b ' ^c idem Table 4. The titers were calculated by the same procedure as that explained in FIG. 7. ^d negative control of toxicity. ^e nd, not detectable. ^f nt, not tested.

Table 6

The supernatants of DH5α (pSTN24) and DH5oc

(pProSTC28) from cultures in agar medium show a significant enterotoxicity which is however lower than that of the homologous supernatants obtained with bacteria containing all the genes of the operon clp

(Table 4). Likewise, the whole DH5α bacteria (pProSTC28) previously washed with PBS and the liquid culture supernatant corresponding to this same recombinant are the only ones to exhibit significant enterotoxic activity (Table 6). However, their enterotoxicity is lower than that of the homologous samples having the functional CS31A system. In addition, no significant toxic activity was detected in the lysates of the recombinant bacteria previously washed in PBS. These results indicate that the rest of the clp operon does not seem necessary for the secretion and enterotoxicity of the ClpG-STh hybrids. However, they suggest that the transport process linked to the CS31A system contributes to better production or maturation of the hybrids since in the presence of this system all the cell extracts exhibit enterotoxic activity (Table 4) whereas in its absence, only a few extracts appear to be toxic (Table 6). 7) Formation and role of disulfide bridges. In order to confirm that the enterotoxicity of the hybrids is linked to the presence of the disulphide bridges in the STh portion, the various recombinant strains were cultured on agar medium with or without β-mercaptoethanol (5 mM). This reducing agent prevents the formation of disulfide bridges.

Figure 9 reports the effect of redox conditions on enterotoxic activity. (A) Enterotoxicity of the culture supernatants of recombinant E. coli DH5α cultured on agar medium in the absence (a) or in the presence (b) of β-mercaptoethanol (5 mM). In the latter case, half of the supernatant is tested immediately (b) and the other half is subjected to the action of oxygen in the air overnight at room temperature (c).

(B) Enterotoxicity of living bacteria (2.10 ¹⁰ bacteria) previously washed with PBS after culture on agar medium containing 5 mM β-mercaptoethanol. n = number of mice tested for each sample.

Unlike the supernatants from cultures produced in the absence of β-mercaptoethanol (Figure 9A, a), the supernatants from cultures containing this agent reducing agent (Figure 9A, b) do not reveal any enterotoxic activity, regardless of the recombinant bacteria tested. When the supernatants are exposed to air overnight, they regain their toxic power (Figure 9A, c). This indicates that the abolition of toxicity by β-mercaptoethanol is not due to the direct effect of this reducing agent on the cells but probably to the absence of the disulfide bridges. The presence of these disulfide bridges is therefore not necessary for the passage of the hybrids through the external membrane. In addition, the disulfide bridges, previously broken in the presence of β-mercaptoethanol, can be reformed under the action of oxygen from the air. Indeed, the calculation of the concentration of thiol groups present in these supernatants left in the air and taken at different times reported in Table 7 above shows a decrease in the amount of free thiols and an increase in enterotoxicity depending on the duration of exposure to air.

Table 7

Table 7 above reports the effect of air oxidation on the enterotoxic activity of the supernatants. The supernatants of the recombinant DH5α bacteria are harvested from cultures produced on agar medium containing 5 mM of β-mercaptoethanol and subjected to the action of air for 0, 18 and 42 hours. The quantity of thiol groups is determined relative to the standard curve. Each value represents the average of the I / c ratios of 4 to 12 young mice (± standard deviation). A G / C ratio ≥ 0.09 is considered positive. The percentage of enterotoxic activity is determined by the ratio of the number of diseased mice to the total number of inoculated mice. The percentage of oxidation at time T18 and T42 is calculated relative to the quantity of free thiols at time T0. Thus, there is a direct correlation between the appearance of the enterotoxic activity of the hybrids and the oxidation by air of the free sulfhydryl groups created beforehand under reducing conditions. On the other hand, the recombinant bacteria cultivated in the presence of β-mercaptoethanol and then washed with PBS induce an accumulation of intestinal fluids in newborn mice (Figure 9B). This strongly suggests that de novo synthesis of the hybrid proteins can be carried out in vi vo or that there are in vi vo environmental conditions favorable to the formation of disulfide bridges.

In order to verify more precisely whether the formation of the disulfide bridges did take place at the level of the STh portion of the hybrid, we determined in vi tro the oxidation-reduction state of the hybrid proteins in the culture supernatants collected in stationary phase and exhibiting enterotoxic activity. The migration on polyacrylamide gel under non-reducing conditions of these supernatants treated with different combinations of DTT and AMS makes it possible to distinctly visualize the hybrid proteins in reduced form and in oxidized form after immunostaining with anti-ClpG antibodies.

FIG. 10 represents the oxidation-reduction state of the hybrid proteins. Bacteria recombinant cells were cultured on agar medium. The supernatants are (+) or not (-) treated with 100 mM DTT before precipitation with final 5% TCA. The pellet is taken up in a buffer containing (+) or not (-) 10 mM AMS. After migration on 12% polyacrylamide gel under non-reducing conditions, the proteins are revealed with the anti-ClpG antibody (A and B, a) or with the mixture of anti-ST 11C + 20C1 antibodies (B, b). The colorimetric reaction is revealed after the use of anti-rabbit IgG antibodies or anti-mouse IgG antibodies coupled with biotin and streptavidin conjugated to peroxidase. Red stands for reduced form and ox for oxidized form. The fixation of each AMS molecule on the reduced proteins delays their migration and the oxidized forms, which do not react with AMS, migrate more quickly (cf. Materials and Methods, §14). Analysis of the supernatants treated only with AMS (lane 3) or additionally with DTT (lane 2) clearly shows that the hybrids present in the extracellular medium are oxidized (lane 4). This is confirmed by the electrophoretic profiles of the reduced forms not coupled to AMS (lane 1) which are indistinguishable from the oxidized forms (lanes 3 and 4). Indeed, all the hybrids, treated simultaneously with DTT and AMS, are delayed in their migration while no change is observable in presence of AMS without DTT. Knowing that the ClpG protein is devoid of cysteine residues, these results indicate that AMS is chemically linked to the sulfhydryl groups of STh. In order to verify that the oxidized form of the hybrids has the STh portion, we analyzed the reactivity of a single supernatant, which was chosen as an example, vis-à-vis a mixture of monoclonal anti-STa antibodies (Figure 10 B, b). The results confirm that the oxidized form contains the STh portion and, therefore, that the disulfide bridges are formed at the STh level. On the other hand, the reduced form does not react with anti-STa antibodies, probably due to loss of the antigenic properties of STh caused by a severe change in the conformational structure of the toxin. It therefore appears that the genetic fusion of STh at the N- and C-terminal permissive ends of the ClpG carrier protein has made it possible to obtain four ClpG-STh hybrids which are secreted in vi tro and in vivo in the extracellular medium in the form of monomers biologically active. The STh portion of hybrids is affected in its antigenic properties when it is fused to the N-terminal end of ClpG or when it is subjected to an environment unfavorable to the formation of its disulfide bridges. The latter have proved to be essential for enterotoxic activity, but not for passage of hybrids through the outer membrane of bacteria.

II - Mechanism of secretion of hybrids.

1) Role of the proteins DsbA and DsbB.

The results are presented in table 8 below in which: + or -, presence or absence of the ancillary genes of the clp ^b I / C operon, ratio of the weight of the intestine / weight of the carcass. Each value represents the average of the I / C ratios of 3 to 14 mice (± standard deviation). A response ≥ 0.09 is considered positive. ^c %, percentage determined by the ratio of the number of sick mice / total number of mice tested. ^of negative toxicity control ^e nt, not tested.

^Ê nd, not detectable. Table 8

Strain Solid medium Liquid medium plasmid CS31 / \ * I / C% ^c U / 10 ^iU I / C% U / 10 ^1U bacteria bacteria

JCB570dsbA ⁺ dsbB ⁺ pEH524 ^d + 0.062 ± 0.020 0 nt ^e nt nt nt pEHSTN24 + 0.102 + 0.019 75 0.8 0.065 ± 0.008 0 nd ^f pEHSTC22 + 0.124 + 0.021 100 1.0 0.061 ± 0.008 0 nd pEHProSTC28 + 0.142 + 0.010 100 3.6 0.066 ± 0.030 0 nd

_P STC17 + pDSPH524 + 0.129 + 0.041 75 2.7 0.096 ± 0.033 25 216 pSTN24 - 0.131 + 0.031 100 3.5 0.166 ± 0.033 75 233 pSTC22 - 0.149 + 0.008 100 3.6 0.150 + 0.012 100 1082 pProSTC28 - 0.154 + 0.012 100 8.3 0.153 ± 0.019 100 1082 pSTC17 - 0.161 ± 0.009 100 5.3 0.161 ± 0.019 100 900

JCB571dsbAdsbB ⁺ pEH524 + 0.059 + 0.003 0 nt nt nt nt pEHSTN24 + 0.122 + 0.029 100 1.0 0.074 ± 0.010 0 nd pEHSTC22 + 0.127 ± 0.038 90 2.0 0.066 ± 0.012 0 nd pEHProSTC28 + 0.135 + 0.033 100 7.9 0.081 0.0 25 nd pSTC17 + pDSPH524 + 0.137 ± 0.030 100 5.8 0.156 + 0.049 75 800 pSTN24 - 0.143 ± 0.012 100 5.3 0.150 ± 0.013 100 847 pSTC22 - 0.126 + 0.008 100 18.2 0.148 ± 0.037 100 1165 pProSTC28 - 0.141 ± 0.011 100 13.7 0.136 ± 0.025 100 2041 pSTC17 - 0.133 + 0.020 100 11.9 0.151 ± 0.020 100 1429

JCB819dsbA ⁺ dsbB ^' pEH524 + 0.061 + 0.001 0 nt nt nt nt pEHS! N24 + 0.164 ± 0.009 100 3.7 0.073 ± 0.003 0 nd pEHSTC22 + 0.139 ± 0.036 100 1.1 0.064 ± 0.004 0 nd pEHProSTC28 + 0.134 + 0.020 100 2.4 0.058 ± 0.002 0 nd pSTC17 + pDSPH524 + 0.139 ± 0.037 100 1 0.144 ± 0.004 100 139 pSTN24 - 0.150 + 0.025 100 3.4 0.131 ± 0.028 100 694 pSTC22 - 0.141 ± 0.017 100 9.1 0.170 ± 0.010 100 723 pProSTC28 - 0.134 ± 0.023 100 16.9 0.140 ± 0.029 100 1136 pSTC17 - 0.137 ± 0.024 100 17.0 0.151 ± 0.026 100 1037

J B818dsb dsbB- pEH524 + 0.061 + 0.007 0 nt nt nt nt pEHSTN24 + 0.113 + 0.036 75 2.9 0.065 ± 0.011 0 nd pEHSTC22 + 0.134 + 0.043 75 6.3 0.070 ± 0.027 25 nd pEHProSTC28 + 0.115 + 0.021 100 2, 1 0.075 ± 0.015 12 nd pSTC17 + pDSPH524 + 0.092 ± 0.025 75 nd 0.085 ± 0.021 37 505 pSTN24 - 0.167 ± 0.006 100 10.0 0.157 ± 0.073 100 806 pSTC22 - 0.143 ± 0.026 100 10.1 0.161 ± 0.007 100 1364 pProSTC28 - 0.166 ± 0.014 100 15.6 0.157 ± 0.039 100 2263 pSTC17 "0.160 ± 0.008 100 16.7 0.145 ± 0.025 100 614 Table 8 above reports the effect of the dsbA and dsbB mutations on the enterotoxic activity of the recombinant proteins. The Dsb system is frequently involved in the formation of the disulfide bridges of many bacterial proteins, including the STa toxin. In order to determine its role in the formation of the disulfide bridges of the STh portion present in the hybrids, the various recombinant plasmids were transferred into the strain JCB570

(dsbA ⁺ dsbB ⁺ ) and in the derived strains JCB571 (dsbA ^~ dsbB "), JCB819 (dsbA ⁺ dsbB ^~ ) and JCB 818 (dsbA ^' dsbB ^~ ) (Table

8). The enterotoxic activity of the supernatants obtained from cultures carried out in liquid medium or agar was then determined (Table 8). The results indicate that all the supernatants originating from cultures in agar medium of dsb ^* or dsb ^" strains, with or without the ancillary genes of the clp operon, induce an accumulation of intestinal fluids in newborn mice and exhibit enterotoxic activity. However, except for the extracts corresponding to the construction [pSTCl7 + pDSPH524], no enterotoxic activity is detected in the culture supernatants in liquid medium of the strains having the annexed clp genes. This could be attributable to a differential production and / or secretion of these hybrids, resulting from: characteristics of the vector plasmid carrying the fusion genes - properties of the host bacterial strain hosting the recombinant plasmid

- whether or not the clp annex genes influence the plasmid and the host strain used.

With the dsb ⁺ or dsb ^~ strains derived from the E. coli JCB570 strain cultivated in a liquid medium, the enterotoxic activity (or the secretion) is greater in the absence than in the presence of the annexed clp genes. The results are reversed with the strain DH5α cultivated under the same conditions. It is also clear that all the supernatants originating from cultures in liquid medium of the dsb ⁺ or dsb ^~ strains have an enterotoxic activity 40 to 300 times greater than the homologous supernatants derived from cultures in solid medium. This difference can be explained by better oxygenation of the liquid cultures subjected to continuous agitation. Another interesting fact is that the dsb ^' mutants reveal an enterotoxicity at least as high as the isogenic dsb ⁺ strain.

In summary, all of these results highlight the following conclusions: - the clp genes other than clpG do not seem to intervene in the secretion and enterotoxicity of the ClpG-STh hybrids produced by the dsb ⁺ and dsb ^~ strains.

- the Dsb system responsible for the periplasmic formation of the disulfide bridges of the natural STa toxin is not involved in that of the disulfide bridges of our hybrids.

2) Role of ClpG. The unmodified ClpG monomer can only be exported to the surface of the bacteria in the presence of the annex proteins involved in the biogenesis of CS31A fimbriae. However, ClpG-STh hybrids derived from bacteria lacking a functional CS31A system pass through the external membrane to be released into the extracellular medium in a biologically active form. In order to determine the role of ClpG in this process, various constructs having the oligonucleotide encoding STh fused in phase directly after the signal sequence of ClpG were made (FIG. 11). The latter aim to obtain recombinant plasmids which differ from plasmids pSTN24, pSTC22, pProSTC28 and pSTCl7 (Figure 1) only by the absence of the clpG gene.

FIG. 11 represents the structure of the fusions between the signal sequence of clpG and different portions of sth. The nucleotide sequences from the constructs shown in Figure 1 are in bold. The STh portion of the fusion protein is colored blue: ST15 (YCCELCCNPACTGCY), ST16 (NYCCELCCNPACTGCY) and ST19 (NSSNYCCELCCNPACTGCY) correspond respectively to the 15, 16 and 19 C-terminal amino acids of the mature STh toxin. Enzyme restriction sites are underlined. SSclpG, signal sequence of clpG. The plasmid pSPGST1 was obtained by insertion of the double-stranded oligonucleotide SPGST3 / SPGST5 between the Sphl and Hpal sites of the plasmid pHPC0838 (Figure 1). This plasmid, thus devoid of the clp G gene, carries the oligonucleotide coding for the mature STh toxin placed immediately after the signal sequence of clpG. Another construction called pSTE531 was carried out by inserting the double stranded oligonucleotide E351 / E531 into the Pst1 site of the plasmid pSTN24. The newly introduced HindIII, PacI and SαaBI sites change the amino acids PADPHA present between STh and ClpG in pSTN24 to PASLINYV. The deletion of the S'naBI / ipal fragment of pSTE531 leads to the plasmid pSTDG17 which corresponds to the plasmid pSTN24 without the clpG gene and with the hexapeptide PASLIN instead of PADPHA at the C-terminal of STh. After verification of the cloning by sequencing, it turned out that the restriction enzyme SnaBl had not cut at the TAC / GTA site but at the TA / CGTA site. This modification resulted in a shift in the reading frame, thus generating a premature TAA stop codon. The substitution of the SphI / Hpal fragment of pHPC0838 by the synthetic duplex DNA STNDG2 / STNDG1 leading to the plasmid pSTDG126 makes it possible to have the equivalent of pSTN24 deleted from clpG with the C-terminal end free of the toxin. Plasmids pGMSTC7 and pGMSTC13 are derived from plasmids pSTC22 and pProSTC28, respectively. In these cases, 1 double-stranded oligonucleotide GMCST3 / GMCST5 was introduced between the Sphl and Hpal sites of pSTC22 and pProSTC28. In order to gradually eliminate the amino acids present between the signal peptide of ClpG and the STh toxin, the Hpal / Hpal and Hpal / Smal fragments of pGMSTC7 were deleted to result in the plasmids pGMSTC71 and pGMSTC710 respectively. Finally, the plasmids pGMSTC71 and pGMSTC710 differ from pSTC22 and pSTC17 respectively only by the absence of the clpG gene. Likewise, the constructions pGMSCT13 / 19 and pGMSTCl3 / 28 were obtained respectively after deletion of the Hpal / Hpal or EcoRI / EcoRI fragment from pGMSTC13. Only pGMSTC13 / 19 corresponds to the plasmid pProSTC28 deleted from the clpG gene.

These different constructions were verified by sequencing and introduced into the different mutated strains for the proteins DsbA and / or DsbB, before being tested for their enterotoxic activity and their secretion in the extracellular medium. FIG. 12 represents the secretion and the toxicity of the hybrids corresponding to the different fusions between the signal peptide of ClpG and STh. In FIG. 12, the accumulation of intestinal fluid is observed by the undiluted supernatants originating from cultures in solid medium of the recombinant bacteria. The secretion of the hybrids was determined by reading the absorbance at 490 nm of the ELISA plates coated with the ST toxin (kit E. coli ST EIA, Oxoid) as indicated in FIG. 6. + or -, presence or absence of secretion . I / C, average established from a group of 3 to 9 young mice per sample. An I / C ≥ 0.09 is considered positive.

The results show that the mature STh toxin preceded by the signal sequence of ClpG is sufficient to induce, in the majority of cases, the secretion of the biologically active toxin and this independently of the Dsb system. Only the fusion protein coded by pSTDGl7 is not detected in the extracellular medium, unlike that coded by pSTN24 which differs from the previous one essentially by the presence of ClpG and the hexapeptide PADPHA instead of PASLIN at the C end -terminal of STh. The presence of this hexapeptide (PASLIN) downstream of STh (pSTDGl7) therefore seems to inhibit the secretion of the toxin since its suppression (pSTDG126) restores the function secretory (Figure 12). However, when the ClpG protein

(257 amino acids) and the six amino acids (PADPHA) between

ClpG and STh are present at the C-terminal end of STh

(pSTN24), the hybrid is secreted. This suggests that coupling of ClpG to the C-terminus of STh would promote the passage of STh across the outer membrane. On the other hand, all the hybrid proteins having the C-terminal part of free STh are secreted into the extracellular medium in a biologically active form, whatever the peptide sequence placed between the signal sequence of ClpG and the STh toxin. These results indicate that

- A heterologous signal peptide such as that of ClpG may be sufficient to export the STh toxin alone to the periplasm of E. col i. This suggests that STh, once in this cell space, is capable by itself of translocating through the outer membrane.

- The secretion of the STh toxin alone is influenced by the nature of the heterologous peptide sequence present at its C-terminal end, but not at its N-terminal end.

- ClpG is not necessary for the secretion of STh when the C-terminal end thereof is free.

- ClpG is not involved in the formation of disulfide bridges, neither in dsb ⁺ strains, nor in dsb ^' strains, since the STh toxin alone fused to the signal peptide of ClpG remains biologically active when it is secreted.

3) Redox state of the hybrid proteins.

Figure 13 shows the redox state of a hybrid protein produced by the dsb ⁺ and dsJ ^~ strains in the extracellular (A) and intracellular (B) compartments. Supernatants from liquid cultures (25 ml) in stationary (A) or exponential phase

(B) treated (+) or not (-) with 100 mM DTT before precipitation with 5% final TCA. The pellet is taken up in a buffer containing (+) or not (-) 10 mM AMS. After migration on polyacrylamide gel (12%) under non-reducing conditions, the proteins are revealed by “Western” immunostaining using rabbit anti-ClpG antibody, goat anti anti-rabbit IgG antibodies coupled with biotin and strptavidin conjugated to peroxidase. Red stands for reduced form and ox for oxidized form. Cox is the oxidation control corresponding to the extracellular hybrid produced by a dsb ⁺ strain hosting the plasmid pEHProSTC28.

The oxidation-reduction state of the extracellular ClpG-STh hybrid secreted by the mutants dsbA ^' and dsbB hosting the plasmid pProSTC28 was analyzed using supernatants from liquid cultures in stationary phase. (Figure 13A). Treatment of the supernatants with DTT and AMS (lane 1) makes it possible to position the reduced form of the hybrid on the immunoblots. Analysis of the supernatant treated only with AMS (lane 2) clearly shows that the hybrid present in the extracellular medium (lane 3) is oxidized since it does not bind to AMS. These results are in agreement with those relating to the toxicity of the hybrids (Table 8). The oxidation-reduction state of the same fusion protein in the intracellular compartment of the dsbA * dsbB * and dsbA ^' dsbB ^' strains was also determined after precipitating 25 ml of a fresh culture taken at DO _550rm = 0 with _TCA . 5 (Figure 13B). The reduced control (lane 2) corresponds to a culture carried out in the presence of 5 mM of DTT. The migration on polyacrylamide gel of the unheated samples free of reducing agents indicates that the hybrid protein is present in a reduced form independently of the dsb mutation (lanes 1 and 3). All these observations are against the formation of disulfide bridges in the intracellular compartment of Dsb ⁺ or Dsb ^" ^ cells, but suggest that these are probably created as the hybrids are exported to the extracellular medium. would explain the toxic power of the supernatants originating from the Dsb strains ^" and the non-toxicity of the cell lysates originating from the Dsb ⁺ strains. In order to confirm this hypothesis, the Kpnl / Bglll fragment of pProSTC28 (FIG. 1) was inserted between the JCpjQl / βamHI sites of the plasmid pTrc99A (Table 2). Thus, the fusion gene carried by pProSTC28 was placed under the control of the promoter P _trc inducible by IPTG. The plasmid thus obtained, pTrcProSTC28, was introduced into the strains dsbA * (JCB570) and dsbA ^' (JCB571). These two strains were cultivated in liquid LB medium up to OD ₅₅₀ - _m = 0.5. The culture is then divided in two. One half contains 2 mM IPTG to allow the induction of the P _trc promoter controlling the production of the hybrid. The other half, without IPTG, is used as a negative control for expression and secretion of the hybrid. Culture supernatant (25 ml) is collected at different times. Two milliliters were kept to test their enterotoxicity in the newborn baby mouse test. The rest was precipitated with TCA and added with AMS.

Figure 14 shows the maturation of the hybrid proteins during their secretion by dsb * and dsb ^' strains. The bacteria dsbA ^' (A) or dsbA * (B) containing the plasmid pTrcProSTC28 are cultivated in a liquid medium with (+) or without (-) 2 mM of IPTG. An aliquot (2 ml) of supernatant is removed at different times to analyze its enterotoxic activity (A, a and B, a). Each histogram represents the average of the I / C ratios of 3 to 5 mice. A response> 0.09 is considered positive. At each time and for each condition, 25 ml of supernatant are treated (+) or not (-) with 100 mM of DTT before precipitation with TCA (5% final). The pellet is taken up in a buffer containing (+) or not (-) 10 mM AMS. After migration on 12% polyacrylamide gel under non-reducing conditions, the proteins are revealed by “Western” immunostaining with the anti-ClpG antibody (A, b; B, b ete) or with a mixture of anti-monoclonal antibodies. Sta 11C + 20C1 (A, c). Figure B, b not allowing a clear view of the different protein profiles, a second series of similar experiments was carried out only for the growth times between 90 and 115 minutes (B, c). Immunoblots are visualized by the use of secondary antibodies coupled to biotin and streptavidin conjugated to peroxidase. The stars (~ k) and ("À") indicate respectively the reduced form and the oxidized form of the hybrid. The circle (•) probably indicates another oxidized form of the hybrid. Red stands for reduced form and ox for oxidized form. Cox is the oxidation control corresponding to the extracellular hybrid produced by a dsb * strain hosting the plasmid pEHProSTC18. Culture means "overnight".

In strain JCB571, a band located at the level of the reduced form and another located at the level of the oxidized form of the hybrid protein are observed with the anti-ClpG antiserum from 15 min of culture in the presence of IPTG (FIG. 14A, b, lane 1). No band is detected without induction by IPTG, even after 115 min of culture (FIG. 14A, b, lane 7). Only the oxidized form is revealed after treatment with AMS of an overnight culture in the presence of IPTG (FIG. 14A, b, lane 9) and only these forms are detected by the anti-ST 11C and 20C1 monoclonal antibodies. (Figure 14A, c) as has already been shown previously (Figure 10B). The hybrid detected after 15 to 20 min of culture in the presence of IPTG is toxic (FIG. 14A, a). No toxic activity was detected in the supernatants of the non-induced cultures.

In strain JCB570 (FIG. 14B), the reduced and oxidized forms of the hybrid appear respectively 15 min and 90 min after the addition of the IPTG (FIG. 14B, b, lane 4 and c, lane 1). The appearance of the oxidized form, which translates the presence of the disulfide bridges, is correlated with the acquisition of the toxic power of the supernatant (FIG. 14B, a). The intermediate band (figure

14B, b and c) could correspond to a second oxidized form of the protein as suggested by the protein doublets reacting with anti-STa antibodies

(Figure 14A, c, lane 4) and the absence of this band in the presence of DTT and AMS (Figure 14B, c, lane 3). All of these results demonstrate that the hybrid passes through the outer membrane of the bacterium Dsb ⁺ or Dsb ^" in reduced form. The disulfide bridges then form gradually as the secretions of the reduced hybrids are formed in the medium. extracellular, probably by the action of oxygen in the air, as previously demonstrated (Figure 9), so the periplasmic Dsb system is not involved in the formation of disulfide bridges in the STh portion of hybrids.

4) Reduction of insulin by the toxin STa. The previously reported results point to oxygen in the air as one of the possible candidates responsible for the oxidation of ClpG-STh hybrids in vi tro, and therefore the formation of disulfide bridges in the extracellular medium. However, the oxygen content in the intestinal lumen is low. It is therefore conceivable that a more efficient system could intervene in vivo to oxidize the thiol groups of the STh portion of the hybrids outside the bacterial cell. The STh peptide sequence contains twice the Cys-XX-Cys motif known to be an active site characteristic of proteins having oxidoreductase activity involved in thiol-disulfide bridge exchanges. The STh portion of hybrids may therefore be able to exchange disulfide bridges with other proteins. This hypothesis was verified by testing the capacity of free STh or coupled to ClpG to catalyze in vi tro the reduction of the disulfide bridges of insulin by DTT. Figure 15 shows the reduction of insulin catalyzed by different proteins in the presence of DTT.

The supernatants are obtained from cultures on solid medium in the stationary phase. The JCB818 strain (pTrcSTN24) was cultured in the presence (+) or in the absence (-) of 2 mM IPTG.

DTT reduces the STh toxin which in turn reduces insulin through a disulfide bridge exchange reaction. The release of the B chain from insulin then induces the formation of a white precipitate. The thioredoxin from Spirulina sp (Sigma) was used as a positive control. The insulin (1 mg / ml) incubated in the 0.1M potassium phosphate buffer, 2 mM EDTA containing only 0.33 mM DTT corresponds to the negative control. The purified DsbA protein (Calbiochem) served as a positive control as an oxidoreductase of the Dsb system. The timing of precipitation, and the rate of precipitation of the insulin B chain are the quantitative parameters considered to measure the reduction in insulin. The results are listed in Table 9.

Table 9

Sample Beginning of precipitation Precipitation rate

(minutes) (ΔA _S50nι - X min ^"1 )

Thioredoxin 12 0.078

Negative control 91 0.012

Toxin STp 44 0.030

DsbA 24 protein 0.042

Supernatant B41 27 0.028

supernatant

JCB818 (pTrcSTN24) (+ IPTG) 34 0.037

supernatant

JCB818 (pTrcSTN24) (- -IPTG) 59 0.011

^T 9 above table relates the reduction of insulin by the free STa toxin or coupled to ClpG. The precipitation appearance time is determined by the increase in absorbance to 650 n ( _A650rι 0.02 above the baseline. The precipitation rate represents the maximum increase in absorbance to 650 nm during the one minute time interval for values of A _650rm between 0 and 1. The almost similar values obtained with pure STp toxin, the natural STp toxin from E. coli B41 and DsbA, are lower than those obtained with thioredoxin but higher than those observed with the negative control. In order to verify whether the toxin retains the same properties in its conjugated form, the fusion gene carried by pSTN24 was placed under the control of the P _trc promoter by insertion of the Kpnl / Xbal fragment of pSTN24 between the Kpnl / Xbal sites of the plasmid pTrc99A before being transferred to the JCB818 dsbA ^' dsbB ^' strain. After induction by IPTG, the culture supernatant from the strain JCB818 (pTrcSTN24) thus obtained proves capable of precipitating insulin at a speed and at a rate similar to that of the protein DsbA and the toxin STp. In the absence of induction, precipitation appears later than in the presence of IPTG and the insulin precipitation rate is identical to that of the negative control (Figure 15 and Table 9).

The STp toxin and the ClpG-STh hybrids therefore have properties analogous to the DsbA protein which allow them to catalyze the reduction of insulin via DTT. Thus, the STa toxin, whether or not coupled to ClpG, is active in vitro in the exchange of disulfide bridges. A maturation model of the free or coupled STp toxin with ClpG: (1) the STa toxin alone or associated with ClpG is synthesized in reduced form in the intracellular medium of bacteria Dsb ⁺ or Dsb ^" , (2) it then crosses the external membrane of the bacterium in this form and, (3) once in the extracellular medium, the disulfide bridges intramoleculars would be created gradually by the action of oxygen and / or through an exchange reaction with oxidized proteins other than DsbA and DsbB.

III - Immunogenicity of hybrid proteins.

1) Production of antibodies directed against the recombinant proteins.

Despite their toxicity, the DH5 bacteria harboring different recombinant plasmids (Table 2) or the corresponding culture supernatants were injected into animals (mice, rabbits) by different routes of inoculation (intra-dermal, intravenous, intraperitoneal) , oral) in order to evaluate their capacity to produce antibodies directed against the carrier protein ClpG and especially against the STh part of the hybrid. According to the results obtained, their neutralizing or even protective potentials have also been estimated with a view to the use of these antibodies in immunotherapy. In this study, we included a hybrid "detoxified" by genetic modification of the STh sequence for active vaccination in humans or animals. This non-toxic hybrid was constructed by inserting between the Hpal and Xbal sites of pHPC0838 (Figure 1) the synthetic oligonucleotide STGL673 / STGL635 coding for the STh toxin including the amino acids Pro ¹³ and Ala ¹⁴ , involved in recognition at the receptor GC-C, have been replaced by glycine and leucine respectively.

Figure 14 shows the structure of the "detoxified" ClpG-STh fusion. Amino acids are numbered from the N-terminus of the mature protein ClpG. The sequences relating to ClpG are in bold. The amino acids corresponding to the synthetic oligonucleotide coding for the modified STh toxin are outlined in blue. Restriction sites are underlined. The vertical arrow represents the site of cleavage by signal peptidase. The asterisks indicate the modified amino acids: the glycine and leucine residues replace the proline and alanine residues respectively in the normal STh sequence. pSK +, pBluescript (S +).

The plasmid pSTCDMl thus obtained (FIG. 16), expresses a hybrid protein ClpG-STh extracellular which is recognized in "Western" immunolabelling by polyclonal anti-ClpG antibodies, but weakly by the monoclonal anti-STa 11C and 20C1 antibodies. After quantification by ELISA competition, the ^λ detoxified hybrid "contained in the vaccine preparation is estimated at 10 ng / 10 ¹⁰ bacteria or 80 ng / ml. The use of toxic and non-toxic hybrids as vaccine antigens also aims to identify a possible relationship between enterotoxicity and immunogenicity. For each type of antigen, the live recombinant bacteria were inoculated into a group of five mice and a single rabbit, respectively intraperitoneally and intravenously (FIG. 17). Modalities described in the Material and Methods chapter Figure 17 represents the immunogenicity of the recombinant bacteria: anti-ClpG (A) IgG titers and anti-ST (B) IgG titers after inoculation of the DH5 bacteria harboring the indicated plasmids. were immunized with 2 to 4.10 ^e bacteria and rabbits with 5.10 ⁸ bacteria. sera from each group of mice were mixed advan t the ELISA assay The antibody titers expressed in log of the inverse of the dilution of the sera are calculated after deducing the signal obtained with the serum of animals immunized with the strain DH5α (pDSPH524) used as a negative control. IV, intravenous route; PI, intraperitoneal route.

After recovery of the sera, the anti-ClpG and anti-STh IgG titers were measured by ELISA on microtiter plates covered respectively with the ClpG protein (1 μg) and with the STp toxin (kit E. coli ST EIA, Oxoid). The sera from each group of mice were mixed before being analyzed. Anti-ClpG serum IgG responses were obtained in all cases and the titers, between log ₁₀ = 4 and log ₁₀ = 6, are similar to that of the positive control (pEH524) (FIG. 17A). The anti-STh serum IgG responses are lower since the titers vary from log ₁₀ = 0 ("detoxified" hybrid) to log ₁₀ = 4.3 (pSTC17 + pDSPH524 hybrids) (Figure 17B). There could therefore be a certain relationship between the toxicity and immunogenicity of STh coupled to ClpG. This relationship suggests that the native three-dimensional conformation of the toxin within the hybrid structure is important for the production of anti-STh antibodies. These titles appear more important in rabbits than in mice. The bacteria harboring the plasmid pEHSTC22 or [pSTC17 + pDSPH524] made it possible to obtain the best anti-STh titers in rabbits. All of the inoculations demonstrate that all bacteria are highly immunogenic because they are capable, in most cases, of inducing both the production of anti-ClpG and anti-STh antibodies. The response directed against the ClpG protein is high in all cases and is fairly homogeneous from one bacterium to another within the same species animal. On the other hand, the anti-STh responses are much more variable depending on the animal species and the recombinant bacteria.

For each type of hybrid, culture supernatants were also inoculated into groups of five mice and one rabbit respectively by intraperitoneal and intra-dermal routes, as indicated in the Materials and Methods chapter.

Figure 18 shows the immunogenicity of the supernatants. The animals received an inoculum of 200 μl (mouse) and 500 μl (rabbits) of supernatants from the recombinant DH5α strains cultured on agar medium. The anti-ClpG (A) and anti-STh (B) antibody titres expressed in log of the inverse of the dilution of the sera are calculated after deducing the signal obtained with the serum of animals immunized with the strain DH5α (pDSPH524) used. as a negative control. The approximate concentration in the supernatants is 0.08 μg eq.STh / ml for the non-toxic hybrid and varies between 0.25 to 6.3 μg eq.STh / ml for the toxic hybrids (Figure 6). PI, intraperitoneal route; ID, intra-dermal route; AIF, incomplete adjuvant of Freund.

The anti-ClpG serum IgG titer is log ₁₀ = 1.6 in rabbits immunized with the "detoxified" supernatant. In the case of toxic supernatants, it varies between log ₁₀ = 3.8 and log ₁₀ = 5.6 depending on the animal species and the hybrid used (Figure 18A) and the titer of serum anti-STa IgG between log ₁₀ = 1.7 and log ₁₀ = 2 , 2 in mice, and between log ₁₀ = 1.8 and log ₁₀ = 3.5 in rabbits (Figure 18B). The hybrids encoded by the plasmids pEHProSTC28 and pEHSTC22 made it possible to obtain the highest anti-STa titers in rabbits. As before, the "detoxified" hybrid does not induce an anti-ST response. However, the fact that the corresponding anti-ClpG serum IgG titer is very low (log ₁₀ = 1.6) is probably due to the small quantity of inoculated hybrids (40 ng eq.STh) and could explain the absence antibodies to the toxin.

2) Neutralization of enterotoxic activity.

The first step in this experiment was to establish a neutralization protocol. For this, 500 μl of different polyclonal and monoclonal anti-STa antibodies described in the literature as being neutralizing, were incubated in the presence of 80 ng (10 U) of pure STp toxin, under the neutralization conditions recommended by the various authors, and injected into newborn mice. Only one of these antibodies, the G104-6 antibody, has been shown to neutralize the STp toxin. Consequently, in the rest of our study, we have chosen the neutralization conditions corresponding to this antibody which was therefore chosen as a positive control. We incubated for 18 hours at 4 ° C our various sera in the presence of toxin to test their neutralizing power. Provided _A s, the G104-6 serum neutralized supernatants containing the different ClpG STh-hybrid and one corresponding to the wild strain B41. The results obtained with our sera from rabbits and mice after inoculation of the supernatants or living recombinant bacteria, showed that only the serum of the rabbit immunized intra-dermally with the supernatant corresponding to the strain DH5α (pEHProSTC28) neutralized the effect. pure STp toxin.

Table 10 below shows the seroneutralization of enterotoxic activity.

^a SL28, rabbit serum immunized intradermally with the supernatant corresponding to the hybrid expressed by pEHProSTC28 in the presence of incomplete Freund's adjuvant. (+) and (-), in the presence or absence of the SL28 serum. I / C, average established from a group of 4 to 11 young mice per sample. An I / C> 0.090 is considered positive. The percentage (%) is determined by the ratio of the number of sick mice / total number of mice tested. ^b The inoculates are incubated for 18 hours at + 4 ° C in the presence of different antisera. ^c Live bacteria B41 washed with PBS are placed in the presence or absence of SL28 and administered immediately to young mice. The bacteria are numbered on MacConkey lactose selective boxes.

Consequently, only this serum, named SL28, was tested for its capacity to neutralize the homologous or heterologous supernatants (Table 10). SL28 inhibits the enterotoxic activity of the hybrids expressed by pEHSTC22, pEHProSTC28 and pSTC17 but not that of the hybrid expressed by pEHSTN24. In the latter case, the result is not surprising since the ELISA tests (Figures 4 and 6B) and the "Western" i muno-imprints (Figure 3) have always showed a lack of affinity of STa specific monoclonal antibodies for this hybrid. SL28 is also effective against the natural STp toxin secreted by the B41 strain. In order to get as close as possible to the natural conditions of infection, different quantities of live B41 bacteria washed with PBS were administered, without prior incubation period with the SL28 serum, to newborn mice (Table 10). While B41 bacteria inoculated alone at a concentration of 5.10 ⁹ to 5.10 ¹⁰ bacteria / ml induce an accumulation of intestinal fluids, bacteria administered in the presence of serum SL28 retain their enterotoxicity at a concentration of 5.10 ¹⁰ bacteria / ml but not between 5.10 ⁹ bacteria / ml and 2, 5.10 ^αo bacteria / ml. To be certain that these data resulted from a neutralizing effect and not from a bactericidal effect of the serum, identical amounts of bacteria B41, diluted in 90% of PBS buffer (negative control) or in 90% of SL28 serum, were incubated overnight at room temperature and then counted. No difference was observed in the presence or absence of serum, which eliminates a bactericidal effect of the serum. 3) Protection.

The oral inoculation of 10 ¹⁰ live recombinant acteriae in adult mice was carried out for three consecutive days. Regardless of the E. coli DH5α strain, approximately 10 ⁵ to 10 ⁶ bacteria were still present in the small intestine three hours after the last inoculation, and one day later in the cecum and feces. We have also previously shown (Tables 4 and 10) that these living bacteria injected intragastrically into newborn mice are capable of triggering the accumulation of intestinal fluids 3 h after their inoculation. There is therefore nothing to prevent the use of E bacteria. coli DH5 live recombinants to stimulate mucosal immunity since they can remain in the intestine and secrete the ClpG-STh hybrids there.

The different recombinant DH5α bacteria were then administered to groups of four pregnant pregnant mice by the oral route, as described in the Material and Methods chapter. Each litter of mice from these mothers was separated into two lots at the age of 3 days. The first batch was tested with 1.6 U of purified STp toxin, and the other with 2 U of hybrids used to immunize pregnant mice. The group of mice from mothers immunized with the DH5α strain (pEH524) was used as a negative control group for protection. In most cases, mice are not protected from a 1.6 U dose of pure STp toxin. Only groups of mice from three mothers immunized with recombinant DH5 bacteria

(pEHSTC22), a group of mice from a mother

(mouse N ° 3) immunized with the DH5 bacteria

[pSTC17 + pDSPH524] and a group of mice from a mother (mouse N ° 2) immunized with the "detoxified" bacterium DH5 [pSTCDMl + pDSPH524] seem partially protected. In all of these cases, 40% to 50% of young mice resist 1.6 U of toxin. Some mice from mothers immunized with DH5 (pEHProSTC28) or DH5α bacteria [pSTC17 + pDSPH524] also appear to be protected against 2 U of homologous hybrids since only 20% to 50% of mice develop an accumulation of intestinal fluids. However, the protections obtained with homologous hybrids may be due to antibodies directed only against the ClpG part of the hybrid. On the whole, these results are rather disappointing and do not allow us to conclude on the existence of protection for mice from mothers vaccinated orally.

The presence of anti-ClpG and anti-STh antibodies in the serum, the intestinal contents and the faeces of each mouse was measured three days after parturition. The Anti-ClpG IgG and IgA from the different samples were quantified on microtiter plates coated with 1 μg of ClpG antigen / well. The search for anti-STa antibodies was carried out on an ELISA plate sensitized with the STa toxin (kit E. coli ST EIA, Oxoid). Anti-ClpG IgGs could thus be demonstrated in the serum of a single mouse immunized with the positive control DH5oc (pEH524) (log titer ₁₀ = 1.7) and in the serum of two mice immunized with bacteria " detoxified "(log headings ₁₀ = 1.2 and 2.6). A high background with all the extracts corresponding to the intestinal contents did not make it possible to detect the presence of antibodies directed against the protein ClpG and the toxin ST. Analysis from the faeces was therefore the only way to assess the production of antibodies in the intestinal mucosa. Anti-ClpG IgG antibodies (log titers ₁₀ = 1.5 and 1.2) and anti-ClpG IgA antibodies (log titers ₁₀ = 2.0 and 1.3) were detected in the faeces of two mice immunized with bacteria DH5α [pSTCDMl + pDSPH524]. No anti-ST IgG or IgA response was detected in the different samples.

In summary, immunization of animals parenterally with different vaccine preparations

(supernatants and living bacteria) has, in the majority of cases, led to high antibody titers specific serum directed against the carrier protein ClpG and the STh toxin. In view of this work, it is likely that there is a direct relationship between the immunogenicity and the spatial structure of the STh molecule of the hybrids since only those exhibiting enterotoxic activity induce the production of anti-STa serum IgG antibodies. Only the serum (SL28) of a rabbit immunized intradermally with the culture supernatant of DH5 (pEHProSTC28) in the presence of incomplete Freund's adjuvant is capable of neutralizing the enterotoxic activity of the natural toxin STa. However, these encouraging results require to be reproduced on a larger number of rabbits. The results on the protection of young mice are not convincing. The low level of anti-ClpG and anti-STa IgG or IgA antibodies obtained in the serum and faeces of the mothers 3 days after parturition suggests that the same is true for the milk intended to provide protection in the gut of mice.

IV- Construction and characterization of

GFP-STh hybrids

Green fluorescent protein (GFP) is a

protein extracted from the jellyfish Aequorea Victoria. This

protein of 238 amino acids, which has a molecular mass

from 27 to 30 kDa in SDS-PAGE, has the particularity of emitting fluorescence under UV without the addition of substrate or co-

factors. It has a major absorption peak at 395 nm

and an emission peak at 509 nm. By its ease and its

rapid detection, GFP is a candidate protein

ideal as a cell marker. Indeed, it is

detectable in living cells, which allows

measure a fluorescence emitted by cells or by a

fabric, repeatedly, at different times and in different

experimentally controlled conditions. He is not

necessary to fix or prepare the cell for the

revelation. In addition, GFP is very stable and resistant to

many denaturing physico-chemical conditions

(heat up to 65 ° C, pH 5.5 to 12.6M HCl-guanidium,

8M urea, 1% SDS, glutaraldehyde, paraformaldehyde). The

the gfp gene (714 bp) of GFP is today widely

used as a cell marker to identify then

sort a transformed cell population as a gene

reporter to measure gene expression in

living cells, or as a molecular marker

to visualize, in a living cell, the dynamics

of a protein fused to GFP. Proteins chimeric GFP with cytoplasmic proteins or

nuclear have been described previously.

1) Genetic mergers between GFP and STh.

In our study, we are merging in phase the

gfp gene coding for GFP to the sth gene coding for

thermostable toxin STh in order to visualize by fluorescence

its attachment to the Guanylyl Cyclase-C receptor (GC-C), in the

perspective of the diagnosis of colorectal tumors, and its

cellular internalization in living cells

normal or tumor intestinal, from the perspective of

therapeutic treatment of these tumors.

The different constructions described in

detail below are shown schematically in Figure 19.

- 19A: The peptide sequence of the STh toxin

native is shown. It contains 19 amino acids which are

represented by letters corresponding to a letter code

international . Amino acids are numbered from

the N-terminus of the mature STh protein and the

numbers appear in bold under the sequence

protein. - 19B: The peptide sequence of the green

fluorescent pro tein GFP is shown. It has 238

amino acids which are represented by letters

corresponding to an international letter code. Acids

amines are numbered from the N-terminus

of native protein GFP and the numbers appear in

bold left and right of the sequence

protein. This sequence corresponds to the GFP protein

encoded by the gfp gene carried by the vector plasmid pGFPuv

marketed by CLONTECH (1020 East Meado Circle Palo

Alto, CA 94303-4230, USA) and whose accession number to

GenBank is U62636.

- 19C: The structure of fusion proteins

GFP-STh is shown diagrammatically. The vector pGFPuv marketed

by CLONTECH expresses the gfp gene which codes for the protein

native 238 amino acid GFP which has been used in

our study as a source of the gfp gene for construction

recombinant vector plasmids puvSTC3, puvSTC7,

puvSTCll, which express the different proteins

GFP-STh chimeras studied during this work. These three

last vectors derive from pGFPuv which is a vector to

large number of copies, which has an origin of replication pUC and which confers resistance to ampicillin. The gene

native gfp expressed by the vector pGFPuv and the genes

gfp 'recombinants::' sth expressed by the puvSTC3 vectors,

puvSTC7 and puvSTCll are under transcriptional control

of promoter l a c (nucleotides 95-178) included in the

promoter sequence which includes the binding site of the

CAP protein (nucleotides 111-124), regions - 35

(nucleotides 143-148) and - 10 (nucleotides 167-172), the

transcription initiation site (nucleotide 179) and

the lacO operator sequence (nucleotide 179-199). The

arrows determine the direction of the transcription, "P ac"

the lactose promoter, "atg", the initiation codon for

translation, and "taa" the translation stop codon.

Numbers in small bold type and those in large

bold type placed below the frames represent

amino acids numbered from the N- terminus

term relating to the GFP protein and

STh toxin.

The recombinant protein GFP-STh encoded by the

gfp 'gene::' s th expressed by the vector puvSTC3 is

consisting of an N-terminal region comprising the 235

first amino acids of the GFP protein and of a region C-terminal including amino acids in position 4 to 19

STh toxin. These two regions are separated by the

"Linker 1" of 12 amino acids (NKSGPESMNSSG) for

provide flexibility between the two areas

GFP and STh. The plasmid puvSTC3 was obtained by insertion,

in the j_7cI136II (Sacl) and Eagl sites of pGFPuv, from

Hpal-Eagl DNA fragment of the plasmid pProSTC28 containing the

sth gene.

The recombinant protein GFP-STh encoded by the

gfp 'gene::' s th expressed by the vector puvSTC7 is

consisting of an N-terminal region comprising the 235

first amino acids of the GFP protein and of a region

C-terminal including amino acids in position 4 to 19

STh toxin. These two regions are separated by the

"Linker 2" of 6 amino acids (NPDNPG) constituting a

separator sequence between the two domains GFP and STh. The

plasmid puvSTC7 was obtained by insertion, in the sites

Ec 113611 (Sacl) and Eagl of pGFPuv, of the DNA fragment Hpal-

Eagl of the plasmid pSTC22 containing the sth gene.

The recombinant protein GFP-STh encoded by the

gfp 'gene::' s th expressed by the vector puvSTCll is

consisting of an N-terminal region comprising the 235 first amino acids of the GFP protein and of a region

C-terminal including amino acids in position 4 to 19

STh toxin. These two regions are separated by the

"Linker 3" composed only of the amino acid glycine

(G). The plasmid puvSTCll was obtained by insertion, into

the Sel13611 (Sac1) and Eag1 sites of pGFPuv, of the fragment

Smal-Eagl DNA of the plasmid pSTC22 containing the sth gene.

2) Fluorescence of the GFP-STh chimeric proteins

The fluorescence of the GFP-STh proteins was evaluated from cultures in agar medium (LB agar) or liquid medium (LB broth) of the recombinant DH5 _α bacteria harboring the vector plasmids presented in FIG. 19C. In solid medium, the fluorescence emission was analyzed in a Petri dish from bacterial colonies irradiated with UV rays while in liquid medium it was examined under fluorescence microscopy with a FITC filter from bacterial cells. suspension. As symbolized in Figure 19C (+ or

-) all the fusion proteins emit fluorescence under UV in a manner comparable to the native GFP protein. This indicates that the loss of the last 3 amino acids (LYK) of GFP and the presence in C-terminal of GFP of an additional sequence of 17 to 28 amino acids, including amino acids 4-19 of STh, do not modify not here GFP fluorescence and therefore do not appear to affect its conformational structure.

3) Enterotoxicity of the GFP-STh chimeric proteins

The enterotoxic activity of culture supernatants in LB broth or in LB agar medium of recombinant DH5 _α bacteria harboring the vector plasmids presented in FIG. 19C was examined in vivo in S iss OF1 3-day old mice. An aliquot of 0.1 ml of each sample is delivered directly to the stomach and, 3 h after the injection, the animals are sacrificed. The whole intestine is removed, weighed, and the ratio (I / C) weight of the intestine (I) / weight of the carcass (C) is then calculated for each mouse. An I / C ratio> 0.09 corresponding to a significant accumulation of intestinal fluids, is interpreted as a positive enterotoxicity indicative of the interaction of the toxin with the GC-C receptor. As symbolized in Figure 19C (+ or

-) all the supernatants coming from the recombinant bacteria have a strong enterotoxic activity.

The data on the enterotoxicity of the GFP-STh hybrids demonstrate that these can be secreted into the external environment due to their over-expression. In addition, these results indicate that neither the coupling with GFP nor the gastrointestinal environment alter the biological activity of the STh portion of the hybrids, and therefore that it adopts a spatial conformation. close to that of the native enterotoxin. Thus, despite its high molecular weight (27-30 kDa) GFP does not seem to modify the primary function of the STh toxin which is to bind to its intestinal receptor, guanylyl cyclase-C, to induce an accumulation of intestinal fluids .

4) Conclusion

Other fusion proteins were constructed according to the same strategy used for obtaining the ClpG-STh chimeras. They differ from the previous ones, especially with regard to the “reporter” protein.

In these new fusions, the ClpG protein has been replaced by the GFP encoded by the gfp gene originating from the prokaryotic expression vector pGFPuv (Clontech). The three gfp:: sth fusion genes obtained (insertion in phase of the sth gene at the 3 ′ end of the gfp gene) express the GFP protein in E. coli and the three GFP-STh fusion proteins thus produced are secreted in the external environment, are fluorescent and are enterotoxic.

5) Outlook

In the perspective of the diagnosis of colonic tumors, the recombinant toxins GFP-STh could be used to target the guanylyl cyclase-C receptor (GC-C) expressed by the colonic tumor cells which will thus be detectable simply by fluorescence under UV rays (λ _excitatioI1 = 395 nm; λ _eraissiorι = 509 nm) by techniques classical cyto (histo) logies after examination under an epifluorescence microscope.

In the perspective of the treatment of colonic tumors, the use of chimeric toxins GFP-STh will also make it possible to evaluate the potentiality of the toxin STh as vectorizing agent for anticancer agents by allowing the demonstration of the cellular internalization of STh after visualization direct by confocal or epifluorescence microscopy.

Claims

1) Use of a conjugate of the peptide STa and of an active molecule in which the said peptide STa has a structural conformation allowing its fixation on the receptor GC-C for the preparation of a pharmaceutical composition intended for the treatment or in the diagnosis of metastatic colorectal tumors.

2) Use according to claim 1, characterized in that the active molecule is an active protein and the conjugate is a chimeric protein.

3) Use according to claim 1 for the preparation of a pharmaceutical composition intended for the diagnosis of metastatic colorectal tumors, characterized in that the active protein is an immunogenic carrier protein capable of conferring on the chimeric protein immunogenic properties without modifying the conformal structure STa peptide.

4) Use according to claim 3, characterized in that the immunogenic carrier protein is a ClpG protein. 5) Use according to claim 4, characterized in that the chimeric protein comprises the peptide STa at the N or C-terminal end of a ClpG protein.

6) Method for diagnosing metastatic colorectal tumors in a biological sample from a patient, characterized in that it comprises the following steps:

- bringing said sample into contact with a chimeric protein formed by a ClpG protein and the STa peptide under conditions allowing the binding of the chimeric protein to the GC-C receptors by means of the STa part of the chimeric protein, then

- the detection of chimeric proteins attached to the GC-C receptors using antibodies directed against the ClpG protein or against the chimeric protein.

7) diagnostic method according to claim

6, characterized in that it consists of an ELISA technique by competition from blood extracts or tissue materials. 8) A diagnostic kit for implementing the method according to claim 7, characterized in that it comprises:

- a microtiter plate where each well is sensitized with all or part of the GC-C receptor retaining the toxin binding site,

- at least one positive control containing purified GC-C receptor.

- the purified ClpG-STh chimeric protein, - primary antibodies concentrated in solution containing polyclonal or monoclonal (mouse) serum IgG anti-ClpG or anti-hybrid ClpG-STh,

- labeled secondary antibodies,

- stamps.

9) A diagnostic method according to claim 6, characterized in that it consists of an immunohistochemical technique from biopsy of extra-intestinal tissues.