Background to the invention The principle behind vaccination is to induce an immune response in the recipient, thus providing protection against subsequent challenge with a pathogen. This may be achieved by inoculation with a live attenuated strain of the pathogen, i. e. a strain having reduced virulence such that it does not cause the disease caused by the virulent pathogen while still stimulating a broad immune response.

Using modern genetic techniques, it is now possible to construct site-directed attenuated bacterial strains in which stable attenuating deletions have been created. A number of site-directed mutants of Salmonella have been created using this type of technology (2,7, 9,14, 19, 35, 36, 37). Mutations in a large number of genes have been reported to be attenuating, including the aro genes (e. g. aroA, aroC, aroD and arobe (15,18)), pur, htrA (4), ompR, ompF, ompC (2), galE (14), cya, crp (7), phot (13, 19), rfaY (48), dksA (48), hupA (48), sipC (48) and clpB (48).

One class of bacterium that has been attenuated by such modern genetic techniques is enterotoxigenic Escherichia coli (ETEC), which causes diarrhoea. The virulence of (ETEC) strains depends on their expression of fimbrial colonization factor antigens (CFAs) which allow them to attach to and colonize the mucosal surface of the small intestine of their host species. Human adapted ETEC strains express a number of CFAs, the most frequently occurring of which are CFA/I, CFA/II (comprising CS3 expressed with either CS 1 or CS2) and CFA/IV (comprising CS6 expressed alone or with either CS4 or CS5). Depending on the geographic location, CFA/I, CFA/11 and CFA/IV account for between 50% and 80% of ETEC strains. Many other CFAs have been described, but each of them is found in only a small proportion of ETEC strains (33). Evidence indicates that anti-CFA immune responses are important for protection against ETEC disease (6,24, 28,29, 30).

Colonization of the small intestine is accompanied by the secretion of enterotoxins.

Two types of enterotoxins have been identified in ETEC strains, the heat labile toxin (LT) and the heat stable toxin (ST). LT is highly homologous in structure to the cholera toxin, a multi-subunit protein of the form AB5. The A subunit is the active component of the toxin, which functions to increase the activity of adenylate cyclase.

This is delivered into host cells by the B subunits, which bind to gangliosides on the cell surface. ST is a small (19 amino acid) non-immunogenic polypeptide that has guanylate cyclase stimulating activity. In addition, it has been demonstrated recently that a large proportion of ETEC strains also produce EAST1, a heat-stable toxin similar in size and mode of action to ST but different in sequence, which was originally identified in enteroaggregative E. coli strains (34).

It has been proposed that derivatives of ETEC strains, which have lost the ability to produce toxins, may be effective live vaccines against virulent isolates. A derivative of a wild-type ETEC strain, E1392/75, that has spontaneously lost the ST and LT activities but which continues to express CFA/11 was identified and designated E1392/75-2A (5). In human volunteer studies, oral vaccination with 2x101° cfu E1392/75-2A gave 75% protection against challenge with a toxin-expressing ETEC from a different serotype but which expressed the same CFAs (reviewed by (30)).

However, approximately 15% of vaccinees experienced mild diarrhoea as a side effect of the vaccine. It was concluded that further attenuation of this strain was required before it could be considered for use as a live vaccine against ETEC infections.

Two derivatives of E1392/75-2A were generated by targeted deletion of potential attenuating genes and evaluated in clinical trials (32,38). It was demonstrated that both of the derivatives (PTL002, AaroClAompR, and PTL003, AaroClAompCI AompF) were attenuated when compared to the parent strain and caused no clinical symptoms in volunteers who ingested up to 5x109 cfu of freshly harvested live organisms. All volunteers receiving the maximum dose of these candidate vaccines generated specific immune responses against the CFA/11 antigen expressed by the strains.

Summary of the Invention An effective vaccine against ETEC must immunize against CFA/I, CFA/II and CFA/IV as a minimum, and therefore attenuated strains expressing all of these antigens are required. Thus, it is required that the genes expressing the toxins LT, ST and EAST1 are inactivated or deleted from strains expressing all of these CFAs.

Toxin minus strains have previously been suggested as a starting point for developing a live attenuated multi-strain vaccine against ETEC (Chatfield, 38). There was, however, no explanation in Chatfield as to how such strains might be generated.

We have now found that there are particular difficulties associated with the generation of a strain expressing CFA/I or a strain expressing CS5 and CS6 from which the toxin genes, especially the ST gene, have been deleted. We have devised a novel strategy and suicide vector for overcoming these difficulties and producing toxin minus forms of these strains.

Without wishing to be bound by this theory, we believe that the reason that ST minus forms of strains expressing CFA/I or CS5 and CS6 were difficult to generate was that the CFA/CS genes are closely linked to the ST gene and are on the same plasmid. In a global review of epidemiological studies in which sufficient data had been collected (33) it is reported that of 204 CFA/I expressing strains, all 204 of them expressed ST, either alone (149/204) or in combination with LT (55/204). Further, more recent studies have confirmed this finding, e. g. Qadri et al (22), where all of 87 CFA/I expressing strains isolated in a two year period in Bangladesh expressed ST, either alone or in combination with LT. No strains were identified which expressed CFA/I and LT alone, suggesting an extremely tight genetic linkage between the ST and CFA/I loci. Numerous scientific papers document the close linkage between CFAII and ST genes in ETEC strains, to the extent that whenever an effort has been made to derive a strain which has lost one of these loci, the other has always been lost concurrently. In no instance of which we are aware has it been possible to separate the two loci by deletion or inactivation of the ST gene and produce a strain that still expresses CFA/I (42-46).

The invention provides a bacterial cell which expresses colonization factor antigen CFA/I from a native plasmid but does not express heat stable toxin (ST). The invention also provides a bacterial cell which expresses colonization factor antigen CS5 from a native plasmid and/or expresses colonization factor antigen CS6 from a native plasmid, but does not express heat stable toxin (ST). The LT gene and the EAST1 gene may also be deleted or inactivated in the cells of the invention. The cells generally contain further attenuating mutations, such as mutations in each of the aroC, ompF and ompC genes, in order to make them acceptable for use in vaccines.

The cells of the invention may be genetically engineered to express a heterologous antigen, such as a non-toxic component or form of LT, or a colonization factor antigen (CFA). Such cells induce an immune response against the heterologous antigen as well as the native antigens and hence improve the protection provided by a vaccine.

The invention includes a vaccine against diarrhoea containing the cells of the invention. Preferably, the vaccine includes a blend of different cells which between them carry all the most common CFAs, namely CFA/I, CFA/11 and CFA/IV.

Furthermore, the invention provides a suicide vector and a method which makes possible the reliable and rapid isolation of the cells of the invention and other bacterial cells containing deleted, inactivated or replaced genes. The vector represents an improvement over known suicide vectors in that it allows more specific and more reliable targeting than known vectors. The vector is less than 5 kb in size (e. g. from 2.5 to 5 kb or from 2.5 to 4 kb) and comprises the sacB region which codes for a product that is toxic to bacteria when grown on sucrose, in which region the IS 1 insertion sequence is deleted or inactivated. The small size of the vector and the absence of the IS 1 insertion sequence help to prevent the vector from targeting to the wrong place in the cellular DNA.

Detailed Description of the Invention Bacteria useful in the Invention The bacterial cells of the invention are generally derived from enterotoxigenic E. coli (ETEC) cells by deletion or inactivation of the ST gene and optionally other toxin genes. As mentioned above, ETEC is a class of E. coli that cause diarrhoea. They colonise the small intestine. They can be isolated from human clinical samples, typically stools produced whilst suffering from diarrhoea. A standard ETEC strain is H10407, deposited at the ATCC under catalogue #35401.

Infections of ETEC are the single most frequent cause of travellers diarrhoea, causing 3-9 million cases per year amongst visitors to developing countries. In endemic areas, ETEC infections are an important cause of dehydrating diarrhoea in infants and young children, resulting in up to 400,000 deaths a year, predominantly in this age group. In developing countries, the incidence of ETEC infections leading to clinical disease decreases with age, indicating that immunity to ETEC infection can be acquired. In contrast, naive adults from industrialized countries who visit endemic areas are highly susceptible to ETEC infections. However, with prolonged or repeated visits to endemic areas susceptibility to ETEC infections diminishes, suggesting that a live attenuated approach to ETEC vaccination may prove successful.

A vaccine to protect against ETEC diarrhoea in humans must provide protection against the seven major colonization factors and, as a minimum, the heat labile toxin (LT) to ensure that protection against different strains is obtained. In order to achieve this, the same attenuations could be made in a range of different ETEC strains, each with a different colonization factor. This would involve deleting the toxins from all such strains. The present invention provides a panel of suitable strains from which all toxin genes have been completely deleted which can provide the starting point for the generation of a multi-strain vaccine. Alternatively, it may be

possible to express multiple colonization factors in a smaller number of strains from which the toxins have been similarly deleted.

Toxin-deleted strains of the present invention were derived from wild-type clinical isolates obtained from a long-term epidemiological study carried out in Egypt by scientists at the US Navy NAMRU3 facility in Cairo. A list of the strains provided is given in the following table. Strain Code Phenotype CFA LT ST EAST1 WS-1858B A 071 : H- CFA/I-+ + WS-4437A B 0128 : H12 CFA/I + WS-6117A C 0153 : H45 CFA/1-+ + WS-2560B D 025: H- CS4, CS6 + + + WS-2773E E 039: H12 CS5, CS6 + + + WS-4150D F 06: H16 CS2, CS3 + WS-6170A G 017: H18 CS2, CS3 + WS-3504D H 0141 : H5 CS2, CS3 + + + WS-3517A I 06: H- CS2, CS3-+ + WS-2252A J 015 : H18 CS4, CS6 + + + WS-2511A K 04: H- CS4, CS6-+ + WS-2556A L 06: H1 CS4, CS6-+ + WS-4046A M 039 : H- None identified +-N. D.

It will be clear to those skilled in the art that other strains may be equally suitable as a starting point for the generation of a toxin-deleted, attenuated multi-strain vaccine.

The strains with the codes A, B, E, H & J were attenuated by the specific removal of all known toxin genes and then further manipulated, as described in the accompanying Examples. Resulting toxin-minus strains and strain PTL003 described above were deposited by Acambis Research Limited of Peterhouse Technology

Park, 100 Fulbourn Road, Cambridge, CB 1 9PT, United Kingdom with the European Collection of Cell Cultures (ECACC), CAMR, Salisbury, Wiltshire SP4 OJG, United Kingdom on 3d September 2001 (accession numbers 010903**) or 29th August 2002 (accession numbers 020829**) in accordance with the Budapest Treaty. The characteristics of the strains and the accession numbers of those which were deposited are as follows: Strain Parent Strain Accesesion LPS:flagellin Antibiotic CS proteins Regulator Toxin Genes Example number Resistance E1392/75-2A E1392/75 N/A O6:H16 Strep CS1 CS3 rns None PTL003 E1392/75-2A 01090302 O6:H16 Strep CS1 CS3 rns None (submitted as ACM2005) ACAM2008 PTL003 02082965 O6:H16 None CS1 CS3 rns None WS-2773E (strain E) N/A N/A O39:H12 None CS5 CS6 ?csvR ST EAST LT 3 WS-2773E-Tox WS-2773E 01090305 O39:H12 None CS5 CS6 ? csvR None minus (submitted as ACM2002) ACAM2006 WS-2773E-Tox minus N/A O39:H12 None CS5 CSW6 ? csvR None ACAM2012* ACAM2006 02082968O39:H12 None CS5 CS6 ? csvR None WS-3504D (strain H) N/A N/A O141:H5 Amp CS2 CS3 rns EAST 4 WS-3504D-Tox WS-3504D 01090304 O141:H5 Amp CS2 CS3 rns None minus (submitted as ACM2003) ACAM2007 WS-3504D-Tox minus 02082964 O141:H5 None CS2 CS rns None WS-1858B (Strain A) N/A N/A O71:H- Amp/Tmp/Smz CFA/I rns ST EAST 2 WS-1858B-Tox WS-1858B N/A O71:H- Amp/Tmp/Smz CFA/I rns None minus ACAM2010 WS-1858B-Tox minus 02082967 O71:H- None CFA/I rns None WS-2252A (Strain J) N/A N/A O15:H18 None CS4 CS6 cfaD ST EAST LT 4 WS-222A-Tox WS-2252A 01090306 O15:H18 None CS4 CS6 cfaD None minus (submitted as ACM2004) ACAM2009 WS-2252A-Tox minus 02082966 O15:H18 None CS4 CS6 cfaD None WS-2511A (Strain K) N/A N/A O4:H- None CS4 CS6 cfaD ST EAST x 2 N/A Strain K WS-2511A-Tox minus N/A O4:H- None CS6 cfaD ST EAST x 2 *ACAM2006 contains a lysogenic phage in its chromosome - ACAM202 is a derivative of ACAM2006 from which a large part of the<BR> genome, including several genes critical for phage assembly, have been deleted.

In addition to the strains in the above table, a toxin minus derivative of strain B described in Example 2 below was deposited under accession number 01090303.

The invention includes"descendents"of these deposited cells. A"descendent"is any cell derived from a deposited cell. The descendents of a deposited cell include cells with one or more further attenuating mutations, such as the mutations described below. Descendents also include cells which have been engineered to express heterologous antigens, such as the heterologous antigens described below.

Although the bacteria of the invention are generally E. coli bacteria, other types of bacteria may be used. A plasmid in accordance with the invention may be constructed by deleting or inactivating the ST gene in a plasmid native to enterotoxigenic E. coli, and then transferring the resulting plasmid to another bacterium. In this way, the other bacterium may be made to carry the CFA/I antigen or the CS5 and CS6 antigens.

The bacteria that are used to make the vaccines of the invention are generally those that infect by the oral route. The bacteria may be those that invade and grow within eukaryotic cells and/or colonize mucosal surfaces. The bacteria are generally gram negative but in some embodiments gram positive bacteria may be used. The bacteria are generally pathogens.

The bacteria used may be from the genus Escherichia, Salmonella, Shigella or Vibrio.

Other than E. coli, examples of the species of bacteria that can be used in the invention are Salmonella typhimurium-the cause of salmonellosis in several animal species; Salmonella typhi-the cause of human typhoid; Salmonella enteritidis-a cause of food poisoning in humans; Salmonella choleraesuis-a cause of salmonellosis in pigs; and Salmonella dublin-a cause of both a systemic and diarrhoel disease in cattle, especially of new-born calves.

Strains of E. coli and Salmonella are particularly useful in the invention. Salmonella are potent immunogens and are able to stimulate systemic and local cellular and antibody responses. In particular an attenuated strain of Salmonella typhi is preferred for use in the invention. Preferred Salmonella typhi strains for use in the present invention include CVD908-htrA (AaroC lSaroD AhtrA) and CVD908 (AaroC AaroD) (49). Strains of E.coli other than ETEC that may be used include enteropathogenic E.coli (EPEC), enteroinvasive E.coli (EIEC) and enterohemorrhagic E. coli (EHEC).

As used herein, references to a"native"plasmid which expresses CFA/I or CS5/6 mean a plasmid which exists in wild-type cells, for example ETEC cells isolated from a person with diarrhoea. They exclude a plasmid constructed in the laboratory for expression of CFA/I or CS5/6. In a cells of the invention, the ST gene in the native plasmid is deleted or inactivated. However, the CFA/I or CS5/6 gene in the plasmid is functional, i. e. it expresses CFA/I or CS5/6.

Further mutating the bacteria In order for the bacteria to be used in a vaccine, they must generally be further attenuated. The attenuation may, for example, be brought about by deleting or inactivating one or more of the following genes: aroA, aroC, aroD, aroE, pur, htrA, ompC, ompF, ompR, cya, crp, phoP, surA, rfaY, dksA, hupA, sipC and clpB.

Preferred combinations of genes include: - at least one aro gene (e. g. aroA, aroC, aroD or aroE) and at least one omp gene (e. g. ompC, ompF or ompR) ; - at least one aro gene (e. g. aroA, aroC, aroD or aroE) and the htrA gene; -aroC, ompF and ompC.

The further attenuating mutations may be introduced using the suicide vector and methods of the invention or by methods known to those skilled in the art (see ref.

25). Appropriate known methods include cloning the DNA sequence of the wild-type

gene into a vector, e. g. a plasmid, and inserting a selectable marker into the cloned DNA sequence or deleting a part of the DNA sequence, resulting in its inactivation.

A deletion may be introduced by, for example, cutting the DNA sequence using restriction enzymes that cut at two points in or just outside the coding sequence and ligating together the two ends in the remaining sequence. Alternatively, and more usually now, a mutant allele in which the flanking regions of a target gene are amplified separately and linked directly together in a separate overlap PCR reaction, with omission of the intervening target sequence, can be constructed (32). A plasmid carrying the mutated DNA sequence can be transformed into the bacterium by known techniques such as electroporation and conjugation. It is then possible by suitable selection to identify a mutant wherein the inactivated DNA sequence has recombined into the chromosome of the bacterium and the wild-type DNA sequence has been rendered non-functional by homologous recombination.

Furthermore, the antibiotic resistance genes must generally be removed from the bacteria before they are used in a vaccine. Bacteria isolated from the wild often contain antibiotic resistance genes, such as resistance genes against ampicillin, streptomycin, sulphmethoxazole, kanamycin, trimetheprim and tetracyclin. These genes can be removed using the suicide vector and methods of the invention or by methods known to those skilled in the art.

The nature of the mutations The mutations introduced into the bacterial vaccine to prevent expression of enterotoxins or other virulence genes delete or inactivate the gene. They generally knock-out the function of the gene completely. This may be achieved either by abolishing synthesis of any polypeptide at all from the gene or by making a mutation that results in synthesis of non-functional polypeptide. In order to abolish synthesis of polypeptide, either the entire gene or its 5'-end may be deleted. A deletion or insertion within the coding sequence of a gene may be used to create a gene that synthesises only non-functional polypeptide (e. g. polypeptide that contains only the

N-terminal sequence of the wild-type protein). In the case of a toxin gene, the mutation may render the gene product non-toxic.

The mutations are generally non-reverting mutations. These are mutations that show essentially no reversion back to the wild-type when the bacterium is used as a vaccine. Such mutations include insertions and deletions. Insertions and deletions are preferably large, typically at least 10 nucleotides in length up to the length of the entire gene or coding sequence, for example from 10 to 600 nucleotides. Preferably, the whole coding sequence or whole gene is deleted.

The mutations are typically site-directed. They may be specific or selective to the toxin gene or other virulence factor gene. In the case of deleting or inactivating the ST gene in a CFA/I or CS5/CS6 strain, the mutation must specifically target the ST gene without deleting or inactivating the (closely-linked) CFA/I gene, CS5 gene or CS6 gene.

Expression of heterologous antigens An attenuated bacterium of the invention may be genetically engineered to express an antigen that is not expressed by the native bacterium (a"heterologous antigen"), so that the attenuated bacterium acts as a carrier of the heterologous antigen. In the case that the bacterium is an ETEC bacterium, the antigen may be from another species or from another strain of ETEC, so that the vaccine provides protection against the other species or strain. Furthermore, the bacterium may be engineered to express more than one heterologous antigen, in which case the heterologous antigens may be from the same or different species or strains.

The heterologous antigen may be a complete protein, a part of a protein containing an epitope or a fusion protein. The antigen may be from another bacterium, a virus, a yeast or a fungus. More especially, the antigenic sequence may be from a pathogenic

strain of E. coli (e. g. ETEC). Useful antigens include ETEC colonization factor antigens, non-toxic components or non-toxic mutants of E. coli LT (e. g. the B subunit and mutants of the A subunit), LT-ST fusion proteins and cholera toxin B subunit (CT-B) (50-55).

The ETEC CFAs and components thereof are prime candidates for expression as heterologous antigens. To instigate diarrhoeal disease, pathogenic strains of ETEC must be able to colonize the intestine and elaborate enterotoxins. For most strains of ETEC, CFAs that are responsible for adhesion to the intestinal mucosa have been identified. In almost all cases CFAs are expressed as fimbriae on the outer surface of the bacteria. A large number of CFAs have been identified, the most prevalent being CFA/I, CFA/11 (includes CS1, CS2, CS3) and CFA/IV (includes CS4, CS5, CS6).

Additional antigens include CS17, CS7, CS9, CS 14, CS 12, PCFO159, PCFO 166.

The DNA encoding the heterologous antigen may be expressed from a promoter that is active in vivo. Promoters that have been shown to work well are the nirB promoter (10, 39), the htrA promoter (10), the pagC promoter (57) and the ssaXpromoter (58).

For expression of the ETEC colonization factor antigens or derivatives of LT, CT or ST, the wild-type promoters could be used.

A DNA construct comprising the promoter operably linked to DNA encoding the heterologous antigen may be made and transformed into the attenuated bacterium using conventional techniques. Transformants containing the DNA construct may be selected, for example by screening for a selectable marker on the construct. Bacteria containing the construct may be grown in vitro before being formulated for administration to the host for vaccination purposes.

Plasmid stabilisation In order to prevent loss of the plasmid expressing the heterologous antigen or of a native plasmid, an element may be added to the plasmid which enhances its stability.

It is generally the case that the plasmids found in ETEC strains encoding the various colonization factor antigens are low copy number and stable enough to ensure their maintenance over many generations in the absence of specific selection mechanisms.

However, following manipulation of these plasmids in accordance with the invention, for example to delete genes for toxins such as ST or other virulence determinants, these stable properties might be impaired. This problem may be alleviated by employing methods for improvement of plasmid stability.

There are a number of"toxin/antitoxin"plasmid stability determining systems known, for exampleparDE (23) from plasmid RP4 (1), and hok/sok (also known as part from plasmid R1 or pndAB from plasmid R483 (11,12)) which could be used to do this. These systems encode two functions: firstly a toxic entity that would kill cells in which it is expressed, which has a long biological half-life, and secondly an antitoxic entity that prevents this killing but has a short biological half-life. In the event that a plasmid encoding these functions is segregated during division the daughter cell which does not contain the plasmid exhausts its supply of antitoxin and is killed by the more persistent toxin moiety. Thus, only cells that continue to harbour the plasmid are maintained in the growing population.

Another system that may be used to enhance the stability of a plasmid in accordance with the invention is a multimer resolution system. Multimer resolution systems confer stability by resolving plasmid multimers into single plasmid copies and hence decreasing the chance of plasmid free daughter cells being generated by random segregation at cell division. A number of site-specific recombination systems which act to resolve plasmid multimers into monomers have been identified. In accordance with such a system, the plasmid to be stabilised contains a recognition site for a site- specific recombinase and the host cell contains a DNA sequence encoding a site- specific recombinase. The recombinase acts on the recognition site and thereby directs proper segregation of the plasmid during cell division. The recombinase may be encoded on the plasmid to be stabilised or in the chromosome of the host cell.

The recombinase is generally a resolvase. Examples of resolvases which may be used in the invention include the Cre recombinase of plasmid P 1, the E. coli XerC (ArgR) protein, the D protein recombinase of plasmid F, the ParA recombinases of plasmids RP4 and RK2, the site-specific recombinase of plasmid Rl, resolvases encoded by the Tn3-like transposable genetic elements and the Rsd resolvase from the Salmonella dublin virulence plasmid.

The recognition elements which may be used in the present invention include those for the above recombinases. Any recognition element recognised by the site-specific recombinase employed may be used. Suitable recognition elements include those sites recognised by the XerC site-specific recombinase, such as the cer site of plasmid ColEl and the similar ckr site of plasmid ColK (59), the psi site of plasmid pst 101 and the cer like site of plasmid pHS-2 from Shigellaflexneri. Other recognition elements which may be used include the crs site from the Salmonella dublin virulence plasmid, the loxP site of plasmid P1, the rfs site of the F plasmid and the res site of the Tn3-like transposable genetic element In a particularly preferred embodiment of the invention, the recombinase is the Rsd resolvase which acts via the crs recognition element. The Rsdlcrs system is described in detail in our copending application, UK Patent Application No.

0024203.2.

Formulation of vaccines The invention provides a vaccine against diarrhoea comprising an E. coli cell of the invention and a pharmaceutically acceptable carrier or diluent. Generally, the vaccine includes a blend of different toxin minus cells (e. g. 2,3, 4,5 or 6 cells) which between them carry all the most common CFAs. For example, the vaccine may contain five different strains of toxin minus cell as follows: (i) a cell which expresses CFA/I (e. g. ECACC Accession No. 01090303 or 02082967) (ii) a cell which expresses CS5 and CS6 (e. g. ECACC Accession No. 01090305 or 02082968);

(iii) a cell which expresses CS4 and CS6 (e. g. ECACC Accession No. 01090306 or 02082966); (iv) a cell which expresses CS2 and CS3 (e. g. ECACC Accession No. 01090304 or 02082964); and (v) a cell which expresses CS1 and CS3 (e. g. PTL003, ECACC Accession No.

01090302 or 02082965).

As mentioned above, the cell deposited under ECACC Accession No. 01090302 (PTL003) has already been tested in two clinical trials and been shown to be safe and immunogenic.

The vaccine may be formulated using known techniques for formulating attenuated bacterial vaccines. The vaccine is advantageously presented for oral administration, for example as a dried stabilised powder for reconstitution in a suitable buffer prior to administration. Reconstitution is advantageously effected in a buffer at a suitable pH to ensure the viability of the bacteria. In order to protect the attenuated bacteria and the vaccine from gastric acidity, a sodium bicarbonate preparation is advantageously administered with each administration of the vaccine. Alternatively the vaccine is presented in a lyophilised encapsulated form.

The vaccine may be used in the vaccination of a mammalian host, particularly a human host. An infection caused by a microorganism, especially a pathogen, may therefore be prevented by administering an effective dose of a vaccine prepared according to the invention. The dosage employed may ultimately be at the discretion of the physician, but will be dependent on various factors including the size and weight of the host and the type of vaccine formulated. However, a dosage comprising the oral administration of from 107 to 1011, e. g. from 108 to 10 10, bacteria per dose may be convenient for a 70 kg adult human host.

EXAMPLES The Examples described in this section serve to illustrate the invention.

Brief description of the drawings Figure 1: Map of suicide vector plasmid pDM4. u = unknown sequence, unknown length.

Figure 2: Map of improved suicide vector pJCB 12.

Figure 3: Diagram of method used to create specific gene deletion constructs by overlap extension PCR. Step 1 = PCR amplification of two DNA fragments. Step 2 = overlap extension PCR using DNA products from reaction 1 and reaction 2 of step 1 and amplification of the overlap extension PCR product. R and S stand for restriction enzyme sites.

Figure 4: Diagram of method used to demonstrate correct integration of suicide vector in to targeted locus by linkage PCR.

Figure 5: Sequence of the ST gene and flanking regions of the plasmid in strain B showing the open reading frame of the structural gene and the position of all oligonucleotides described in the text and detailed in Table 1.

Figure 6: Sequence of the EAST1 gene and flanking regions from IS1414 (Genbank &num AF143819) showing the open reading frame of the structural gene and the position of all oligonucleotides described in the text and detailed in Table 1.

Figure 7: Sequence of the LT locus showing the position of the oligonucleotides used to create the deletion construct. The underlined ATG codon near the end of the sequence is the start codon of LTA.

Figure 8: Sequence of the 3'flanking region of the ST-1 gene determined from strain E showing the position of oligonucleotides described in the text and detailed in Table 1. The open reading frame of the structural gene and the ATG and TAA start and stop codons are underlined.

Figure 9: Diagram of the individual stages involved in the attenuation of strain E.

Figure 10: Diagram of the individual stages involved in the attenuation of strain H.

Figure 11: Sequence of the aroC gene locus showing the position of the oligonucleotides used to create the deletion construct and of others mentioned in the text and described in Table 1.

Figure 12: Diagram of the individual stages involved in the attenuation of strain J.

Figure 13: Gel showing PCR amplified LT-B coding sequence from strain E. The left hand arrow indicates the LTB PCR product and the right hand arrow indicates a 600bp marker.

Figure 14: Sequence of the LT-B gene cloned from strain E.

Figure 15: Diagram showing construction of plasmid for the expression of LT-B under control of the nirB promoter. The LTB DNA was digested with BglII and NheI. The starting vector, pNCAT4, was digested with the same enzymes and the larger vector fragment was isolated. The LTB DNA and vector fragment were ligated together. The KpnI and BglII sites in the final plasmid, pNLTB, may be used for cloning of the LTB promoter.

Figure 16: Western blot showing the expression of LT-B in attenuated ETEC vaccine strain PTL003, in both the cytoplasm and periplasm.

Figure 17: Gel showing PCR amplified LTAB promoter sequence from strain E.

The left hand arrow indicates a 200 bp maker and the right hand arrow indicates the LT promoter PCR product.

Figure 18: Sequence of LTAB promoter cloned from strain E.

Figure 19: Western blot showing the expression of LT-B in attenuated ETEC vaccine strain PTL003, under control of the native LT promoter. PLLTB-1, -2 and-3 are three independent colonies of PTL003 transformed with pLLTB. The three left had arrows indicate background, the top right hand arrow indicates LTA and the bottom right hand arrow indicates LTB.

Figure 20: Diagram of the"plasmid rescue"technique to obtain the sequence of the 3'flanking region of ST of strain B. The technique is described in detail in Example 2. An ST cassette was made by PCR (primers 4764 and 4765) and cloned into plasmid pJCB 12. The resulting pJCB 12-ST plasmid was introduced into strain B by conjugation. Plasmid DNA from one of the transconjugants was isolated and digested with a restriction enzyme (BglII) which cuts at a site distant from the target gene. The 5'and 3'ends of the resulting cut plasmid were ligated together to close the plasmid. The plasmid was propagated in SY327 pir and then sequenced with a primer (4792) which will only hybridise to the native ST gene (not to the ST gene of pJCB12-ST). The sequence obtained using this primer may be used to design further primers to extend the sequence.

EXAMPLE 1: AN IMPROVED PROCESS FOR THE INTRODUCTION OF MULTIPLE GENTIC MUTATIONS INTO BACTERIAL VACCINE STRAINS This section describes the generation of a novel suicide vector plasmid, pJCB 12, and its use in the generation of an optimised procedure for the introduction of mutations into chromosomal or plasmid encoded gene loci. It also describes many of the standard methods used in this and subsequent examples of the specification, at the end of the section. The sequence and purpose of the oligonucleotides used in PCR for construction or analysis of constructs are given in Table 1 at the end of the Examples.

An improved process for the introduction of multiple genetic mutations into bacterial vaccine strains The generation of an attenuated ETEC strain from a wild-type strain requires mutation of a number of different genetic loci. These mutations are introduced sequentially, each using a process that requires several different steps. When a number of ETEC strains require attenuation, this amounts to a significant number of steps, each potentially with its associated difficulties. It is therefore vital that the process of introducing mutations is optimised fully.

Suicide vector plasmids such as pDM4 (20), pJCB12, pCVD442 (8) and others can be used to introduce defined genetic constructs into specific targets in the bacterial genome. Plasmid pJCB 12 is a new, optimised suicide vector based on the previously constructed suicide vector pDM4. The defined genetic construct to be introduced into the bacterial genome may be a deletion mutation of a specific gene, or a more complex structure such as, for example, an insertion of a gene within another and expressed from a chosen promoter from within the construct. Generally, the

extremities of the constructs will consist of nucleotide sequences derived from the region of the genome to be targeted.

Suicide vectors pDM4 and pJCB12 possess a number of key components (see Figures 1 and 2): An origin of replication which directs replication of the vector in some strains of bacteria but not in others, oriR6K. oriR6K is the origin of replication derived from the naturally occurring plasmid R6K. This origin requires the R6Kpir gene for replication, which is absent from the suicide vectors. Three laboratory E. coli strains are available that carry the pir gene on their chromosome, which are SY327Xpir, SM10Xpir, and DH5aRpir. All three of these strains may be used to propagate pDM4, pJCB12 and their derivatives.

A transfer origin that directs conjugative transfer of the vector from one bacterial strain to another, mobRP4. mobRP4 is the transfer origin from the naturally occurring plasmid RP4. This allows the conjugative transfer of pDM4 and pJCB12 and their derivatives to recipient bacterial strains. In order to function, mobRP4 requires the genes encoding the RP4 transfer functions to be present in the donor bacterial cell. Laboratory E coli strain SM10Xpir carries these genes on its chromosome, and so this strain can be used as a donor strain for pDM4, pJCB12 and their derivatives.

A gene encoding a product that is toxic to bacterial cells when the cells are grown under defined conditions, sacB. sacB codes for levansucrase which produces a product that is toxic to Gram-negative bacteria when grown on sucrose.

A selectable marker, cat. cat codes for chloramphenicol acetyltransferase and confers resistance to the antibacterial chloramphenicol.

A multiple cloning site (MCS), i. e. a site into which defined genetic constructs may be cloned for introduction into a recipient bacterial cell.

It is clear to us from our use that existing suicide vectors such as pDM4 and pCVD442 frequently does not target the correct locus, and screening of a large number of transconjugants or transformants is necessary to identify one that is correctly targeted (if a correctly targeted one can be identified at all). In addition, pCVD442 carries an ampicillin resistance determinant to allow selection of recombinants. However, selection on ampicillin is not as efficient as chloramphenicol and attempts to select for very small numbers of recombinants within a very large mixture of bacteria, such as when performing conjugation experiments, are often thwarted by"background growth"."Background growth"can take the form of smears of bacteria within which identification of transconjugants is not possible, or small colonies, which upon further analysis are not transconjugants.

We believe that the problems of incorrect targeting are partly due to the relatively large size (6-7 kb) of these known suicide vectors which are themselves constructed from components of naturally occurring plasmids. The suicide vectors could therefore target naturally occurring plasmids within a bacterial strain that carry similar nucleotide sequences. In particular, transfer functions among conjugative plasmids are relatively conserved, and the mobRP4 transfer region of pDM4 and pCVD442 is approximately 2. 5kb. In addition, in light of the problem of incorrect targeting we carried out sequencing. Our nucleotide sequence data obtained for the sacB region of pDM4 showed that it includes approximately 600bp of an Il-lié insertion sequence, an insertion sequence which is prevalent in the genomes of many bacteria. We deduced that this could be at least partly responsible for the incorrect targeting.

Suicide vector pJCB 12 is a modified version of pDM4 in which much of the intergenic and non-functional DNA has been removed. Therefore, there is much less opportunity for incorrect targeting using this suicide vector. Whereas pDM4 is approximately 7 kb in size, pJCB12 is only 3 kb but retains all the key components.

In particular, the mobRP4 region of pJCB12 is merely 0.15 kb, and the ISl-like nucleotide sequences have been removed from the sacB region. These modifications are particularly advantageous when manipulating ETEC strains which generally harbour many plasmids that could act as undesirable targets of homologous recombination with components of the suicide vector. In addition, the smaller size of pJCB 12 allows easier in vitro manipulation and construction of derivatives because smaller DNA molecules ligate together and transform into E. coli hosts more efficiently, improving the chances of obtaining derivatives of the correct construction. The smaller size also allows greater efficiency when introducing the constructs into recipient bacteria by transformation rather than by conjugation.

Laboratory E. coli strain SM I Okpir can be used to transfer pJCB12 and its derivatives to recipient bacterial strains by conjugation because it has the tra functions from plasmid RP4 inserted into its chromosome. However, strain SM I Okpir shows relatively low transformation frequencies. For this reason, strain DH5 (xXpir would normally be used for the construction of pJCB12 derivatives, and once derivatives of the correct construction have been identified these would be transferred to SM I Okpir for introduction to recipient strains by conjugation.

Construction of suicide vector pJCB12 Suicide vector pJCB12 was constructed by several rounds of overlap extension PCR (31, Figure 3) using pDM4 plasmid DNA as template. Initially, four fragments were amplified from pDM4 by PCR using the high fidelity DNA polymerase, Pfu TurboTM. These were the oriR6K fragment, amplified using oligonucleotides 4714 and 4715; the mobRP4 fragment amplified using oligonucleotides 4716 and 4717; and the cat gene that was amplified in two parts using oligonucleotides 4718 with

4719 and 4720 with 4721. This was done in order to remove an EcoRI restriction enzyme site within the cat gene. The oriR6K fragment and the mobRP4 were then joined in an overlap extension PCR reaction using oligonucleotides 4714 and 4717.

Likewise, the cat fragments were joined using oligonucleotides 4718 and 4721.

These two resulting fragments were then joined in a final overlap extension PCR reaction using oligonucleotides 4717 and 4718. The resulting PCR product was ligated and transformed into SY327#pir cells and transformants were selected on L- agar supplemented with chloramphenicol at 20 pg/ml. Transformants harbouring plasmids of the correct size were obtained and one of these, called pDM4A7, was chosen for further manipulation.

At this stage, clearly the oriR6K and cat components of the plasmid pDM4A7are functional. However, in order to confirm that the mobRP4 locus was functional plasmid pDM4A7 was transformed into strain SM I Okpir. These transformants were picked onto L-agar supplemented with chloramphenicol at 15µg/ml and naladixic acid at 5yg/ml. This L-agar was cross-streaked with cells of strain SY327Xpir.

While chloramphenicol selects those bacterial cells which harbour pDM4A7, nalidixic acid selects for SY327Xpir. After overnight incubation, many colonies grew where the strains were cross-streaked, but none grew elswhere on the plate, confirming that pDM4A7 is mobilisable from strain SM10Xpir and that the mobRP4 locus is functional.

Plasmid pDM4A7 was then digested with EcoRI, treated with Pfzl TurboTM DNA polymerase and ligated in order to remove the EcoRI restriction enzyme site to generate plasmid pDM4A7AEcoRI. A short HindIII fragment from pDM4 which includes the multiple cloning site was then ligated into pDM4A7AEcoRI digested with HindIII. The ligation reaction was transformed into SY327Xpir and transformants selected on L-agar supplemented with 20, ug/ml chloramphenicol.

Oligonucleotide R6K-01 hybridises within the short HindIII fragment from pDM4 which includes the multiple cloning site. Therefore, transformants were screened by PCR using oligonucleotides R6K-01 and 4720 in order to identify those harbouring the desired plasmid construct. A number of such transformants were identified, and one of these, called pDM4A7AE, was chosen for further manipulation.

Plasmid pDM4A7AE carries three EcoRI sites very close together on the short HindIII fragment from pDM4 which includes the multiple cloning site. The two very short EcoRI fragments of pDM4A7AE were therefore removed by digestion with EcoRI followed by ligation. This resulted in a pDM4A7AE derivative that possess only one EcoRI site which was called pJCB 10. The region of pJCB 10 that includes oriR6K and the MCS was amplified using oligonucleotides 4715 and 4917 and nucleotide sequence determinations for part of this fragment were performed using oligonucleotide 4917. This presented us with the nucleotide sequence across the MCS which was previously unknown.

The sacB gene was then amplified using Pfu DNA polymerase and oligonucleotides 4722 and 4723. The 1.6 kb product was ligated with the plasmid vector pPCR- ScriptrM (Stratagene) and transformed into E. coli XL10 GoldTM cells (Stratagene).

Transformants were obtained and the functionality of the sacB gene was confirmed by plating the clones onto L-agar and 5% sucrose agar. One construct gave good growth on L-agar, and none on 5% sucrose agar, and so was chosen as the source of the sacB gene. The sacB gene was then digested from this clone using the restriction enzyme PstI, sites for which were incorporated into oligonucleotides 4722 and 4723 for this purpose, and ligated with pJCB 10 also digested with PstI. Colonies were checked by PCR using oligonucleotides 4716 and 4766, yielding a product of the expected size (-1700bp). Again the functionality of the gene was confirmed by plating the clones onto L-agar and 5% sucrose agar. One construct grew on L-agar, but not on 5% sucrose agar. Sequencing of this construct using oligonucleotides 4716 and 4766 respectively indicated the orientation of the sacB gene. This construct was called pJCB 12.

Principle of use of pJCB12 Once a defined genetic construct has been ligated into pJCB 12 to give a pJCB 12- derivative, the plasmid is transferred into a recipient strain such as an ETEC strain.

This may be done according to methods well known in the art, either by conjugation from the pJCB 12 host strain SM 1 ORpir, or by transformation of the purified pJCB 12- derivative directly into the recipient strain.

Transconjugants or transformants are selected on bacteriological growth medium supplemented with the antibiotic chloramphenicol. Since the suicide vector pJCB12 is unable to replicate in the absence of the pir gene, any transconjugants or transformants that grow will generally have resulted from fusion of the pJCB 12- derivative with another replicon by homologous recombination. Using pDM4 and other larger suicide vectors will often result in incorrect targeting, and consequently mutants cannot be isolated in subsequent time-consuming steps.

Although targeting by pJCB 12 is much improved over pDM4, incorrect targeting can still occur. Therefore, in order to optimise fully the defined mutation process, a novel approach was taken to screen transformants or transconjugants using PCR to identify those in which the pJCB 12-derivative has targeted the desired region of the genome. For this, one oligonucleotide is designed which hybridises within the pJCB12 nucleotide sequences adjacent to the MCS where the defined genetic construct has been inserted. The other oligonucleotide is designed to hybridise to the region of the genome to be targeted, adjacent to but outside of the defined genetic construct. Transformants or transconjugants that are positive using this PCR will have the pJCB 12-derivative targeted to the correct region of the genome (see Figure 4).

Once the correct recombinants have been identified, derivatives need to be isolated in which the pJCB12 vector has been lost. Such derivatives may be selected by supplementing the bacteriological growth medium with 5% sucrose. This sucrose

selection may be made more efficient using a modified L-medium in which the NaCI ingredient is absent and supplemented with 5% sucrose. Under these conditions the sacs gene of pJCB 12 is toxic, and only derivatives where the sacB gene has been lost will grow. This event again occurs by homologous recombination and has a number of outcomes. Firstly, a reversion event will result in the targeted region remaining as it was. Secondly, homologous recombination may result in the defined genetic construct being swapped with the targeted region resulting in the defined construct being incorporated at the target region. In addition, if the targeted region is part of a plasmid, such as many of the toxin genes of ETEC strains, then two additional events may occur. These are, thirdly, an undefined spontaneous deletion event, resulting in the loss of a part of the targeted region which may extend beyond the boundaries of the defined genetic construct, and, fourthly, the loss of the whole plasmid, an event which may be termed"specific plasmid curing".

Testing of sucrose resistant derivatives by PCR can identify the desired recombinants. For this, oligonucleotides that hybridise at each end of the targeted region and outside of the defined genetic construct are used. If the PCR product is the same size as prior to introduction of the pJCB 12-derivative construct, then a reversion event has occurred. If, for example the genetically defined construct is a deletion mutation, then the PCR product should be smaller than previously and of a predictable size. Specific plasmid curing and undefined spontaneous deletion will normally result in no PCR product or non-specific products of unexpected size in this type of PCR reaction.

In summary, vector pJCB 12 (or another similar vector of the invention) may be used in a method for producing a bacterial cell in which a target gene (e. g. a toxin gene such as ST, LT or EAST1 or a chromosomal gene such as an omp or aro gene) is deleted, inactivated or replaced, which method comprises transferring the vector into a bacterial cell containing the target gene and selecting for a cell in which the target gene has been deleted, inactivated or replaced. The selection may be carried out using a multi-stage procedure along the following lines:

- Selecting for a colony of cells which contains the selectable marker. If the cell into which the vector is transferred is one that does not support replication of the vector from the origin of replication in the vector, selecting for such a colony of cells identifies cells in which the vector has become incorporated into a cellular replicon; - Carrying out PCR to select for a cell in which the vector has correctly targeted to the target gene, wherein one of the primers used in the PCR hybridizes to vector sequence adjacent to the cloning site and the other hybridizes to a site in the cellular DNA adjacent to the target gene. A positive PCR indicates that the vector has targeted to the target gene.

- Selecting for a cell from which vector sequence has been lost by growing the cell under conditions which make effective the gene encoding a product that is toxic to the cells when grown under defined conditions. Survival of a cell indicates that vector sequence has been lost. Where the gene encoding the toxic product is sacB, the cell may be grown in medium supplemented with sucrose and from which NaCl is absent; the product of sacB is toxic when the cells are grown in this medium.

- Finally, PCR may be carried out using primers which hybridize at positions outside, and adjacent to each end of, the target gene, wherein a PCR product smaller than the product obtained from a wild-type cell indicates a deletion mutation.

Materials and Methods Bacterial strains used. ETEC strains as described elsewhere in the specification and lab strains of E. coli: Strain Reference or Source SY327kpir Miller and Mekalanos (56) DH5aRpir P. Barrow, Institute for Animal Health, Compton SMIOw Simon et al. , 1983 (47)

Bacteriological growth media. ETEC strains were routinely grown in L-broth and on L-agar and incubated at 37°C overnight. L-broth consists of lOg/1 peptone, 5g/l yeast extract and 5g/l of NaCl dissolved in 1 1 of deionsed water. L-agar is L-broth supplemented with 15g/l agar. Growth medium containing 5% sucrose was as described above, but without the 5g/l of NaCl. To optimise expression of CFAs ETEC strains were harvested from CFA-agar (1% casamino acids, 2% agar, 0.15% yeast extract, 0.005% MgS04, 0.0005% MnCl2). Chloramphenicol was used at a concentration of 10 llg/ml, tetracycline at 15 Ug/ml and streptomycin at 20 u. g/ml.

Bacterial Conjugations were performed by mixing donor and recipient ETEC strains on L-agar and incubating at 37°C for 3 to 18 h. Bacterial growth was scraped off into L-broth and plated onto L-agar plates supplemented with chloramphenicol and another appropriate antibiotic to select ETEC strains (streptomycin for strain B, tetracycline for other ETEC strains) that had incorporated the pJCB 12-derivative.

Identification of correctly targeted recombinants. Transconjugants or transformants obtained by growth on L-agar supplemented with chloramphenicol following introduction of pJCB 12-derivative constructs were tested by PCR in order to identify those in which the desired genetic locus had been targeted. For this, one of the oligonucleotides hybridised within the pJCB12 nucleotide sequences adjacent to the multiple cloning site (MCS) where the defined genetic construct had been inserted. The other oligonucleotide hybridised to the genome, adjacent to but outside of the defined genetic construct. In such a PCR, the generation of a fragment indicated that the binding sites for the respective oligonucleotides had become linked, which could occur only if the pJCB 12-derivative had targeted the correct region of the genome.

Excision of pJCB12 from transconjugants by growth in the presence of 5% sucrose. Transconjugants or transformants having the pJCB12-derivative targeted to the correct region of the genome were then streaked onto fresh L-agar supplemented with chloramphenicol and another appropriate antibiotic to select ETEC strains (see above), and incubated at 37°C to allow colonies to grow. L-broth cultures inoculated from these fresh plates were then grown. Cells from these cultures were harvested, resuspened in 5% sucrose broth, and incubated overnight prior to plating serial

dilutions on 5% sucrose agar in order to select recombinants in which the pJCB12- derivative had excised. The inoculated sucrose agar plates were then incubated overnight and the resulting colonies tested by PCR using relevant oligonucleotides in order to identify mutants.

DNA manipulations were performed using standard procedures (25). Plasmid DNA was prepared using plasmid purification kits from QIAGEN (Crawley, UK) and DNA fragments were isolated from agarose gels using the QIAquickTM gel extraction kit from QIAGEN.

PCR reactions were performed routinely using Taq DNA polymerase (Life Technologies); when high fidelity PCR was required, such as in the construction of pJCB12, Pfu"Turbo"DNA polymerase (Stratagene) was used. Both DNA polymerases were used in accordance with the suppliers instructions. Routinely, the following PCR cycle was used: Step 1 ; 95°C, 50 sec Step 2 ; 95°C, 10 sec Step 3; 55°C, 1 min Step 4; 72°C, 1 min Step 5; repeat steps 2 to 4 twenty-five times Step 6; 72°C, 1 min For construction of pJCB 12, a pDM4 plasmid DNA preparation was used as the template in PCR reactions. For routine PCR screening a pick from a bacterial colony was used as the template. For construction of toxin deletion mutations, plasmid DNA preparations extracted from suitable ETEC strains were used. Plasmid pJCB 12 and DNA fragments incorporating deletion mutations were generated using a modification of overlap extension PCR (31). In the modified version, there is no overlap in the fragments amplified. Instead, the fragments flank the region to be

deleted, and the complementary sequences which allow joining of the two fragments are included in the 5'-ends of the relevant oligonucleotides; see Figure 3.

SDS-PAGE This was performed essentially as described by Laemmli (16) and was used for visualisation of expressed CFAs and for checking LPS profiles. For LPS profiles, cells were grown overnight in L-broth, harvested and resuspended in water so as to give a cell suspension with an A600 of 20/cm. The cell suspension was mixed with an equal volume of 50mM TrisHCl pH6.8, 2% (w/v) SDS, 10% (v/v) glycerol, 0.25% (w/v) bromophenol blue, 2% (v/v) 2-mercaptoethanol and boiled for 5 minutes. Proteinase K was then added to a final concentration of 0.2 mg/ml and the samples incubated at 60°C for 1 h before loading 5 to 10 pl onto SDS-PAGE gels.

For CFAs, bacteria were harvested from CFA-agar and resuspended in water so as to give a cell suspension with an A600 of 20/cm. The cell suspension was heated to 65°G for 10 mins, the samples were then centrifuged to remove whole cells. An equal volume of 50mM TrisHCl pH6.8, 2% (w/v) SDS, 10% (v/v) glycerol, 0.25% (w/v) bromophenol blue, 2% (v/v) 2-mercaptoethanol was then added. Volumes of 3 to 10 ul were then loaded onto 12% Tris-glycine SDS-PAGE gels (Invitrogen).

CFAs were visualised by Coomassie blue staining (Sambrook et al 1989, ref 25) while LPS was visualised using a SilverXpressTM silver stain kit (Invitrogen).

EXAMPLE 2: REMOVAL OF TOXIN GENES AND INTRODUCTION OF ATTENUATING MUTATIONS INTO STRAINS EXPRESSING CFA/I Manipulation of Strain A to produce strain ACAM 2010 Strain A (WS-1858B) expresses the colonisation factor antigen CFA/I and the heat- stable toxin ST. In order to generate an ST-negative derivative that continues to express the CFA/I colonisation factors the strain was first transformed with a plasmid conferring tetracycline resistance. This was to allow subsequent selection of the strain from conjugation mixtures. The plasmid conferring tetracycline resistance is a derivative of plasmid pACYC184 (40) in which the chloramphenicol resistance

determinant had been inactivated by deletion of a BsmBI restriction enzyme fragment using standard DNA manipulation procedures. This plasmid was termed"pACYC- Tc"and was introduced into strain A by electro-transformation.

The TetR derivative of Strain A was conjugated with SM10Xpir harbouring plasmid pJCB 12-STI. This pJCB 12 derivative contains a fragment of the STI gene amplified from strain B using oligonucleotides 4764 and 4765 cloned into the MCS. The sequence of this construct is given in Figure 5. Note that the STI fragment in this derivative is not a deletion; the aim is to target the locus with the markers encoded by pJCB 12 in order to allow selection for deletion events resulting in the correct outcome.

Transconjugants were selected by plating the mixture on L-agar supplemented with tetracycline and chloramphenicol and the colonies obtained were confirmed as transconjugants by PCR testing using oligonucleotides 4720 and 4721 which amplifies the chloramphenicol resistance determinant. Three transconjugants were identified and called A13, A14 and A18 respectively. Each transconjugant was grown on 5% sucrose medium and colonies obtained were screened for ST using oligonucleotides 4764 and 4765. One of the ST-negative derivatives identified was called A18-34 and was tested by PCR for the presence of the cfaB gene using oligonucleotides BgIIIFOR and BglmodREV, for the presence of the cfaC gene using oligonucleotides 4727 and 4728, and for the presence of the cfaR gene using oligonucleotides 4785 and 4786. All three of these PCR reactions were positive, confirming the presence of the relevant genes. Derivative strain A18-34 was then grown on CFA-agar and processed for SDS-PAGE in order to visualise CFA/I expression. This confirmed that Strain A18-34 expressed CFA/I and, therefore, that an ST-negative, CFA/I expressing strain had been isolated.

In the next step an ampicillin sensitive derivative of strain A18-34 was identified.

For this, strain A18-34 was grown in LB-broth through three passages. The resulting culture was then diluted and plated onto LB-agar. When small colonies were visible they were replica-plated onto LB-agar supplemented with ampicillin at 20011g/ml and then incubated to allow colonies to grow. One colony was identified that was absent

from the ampicillin supplemented replica plate and subsequently confirmed as being ampicillin sensitive. This strain was called A18-34 Aps.

In the next step, a trimethoprim sensitive derivative of strain A18-34 Aps was isolated. For this strain A18-34 Ap was replica plated as described above but onto LB-agar and M9 minimal salts agar supplemented with 0.4% glucose and trimethoprim at 25 pg/ml. Two colonies were identified that had not grown on the trimethoprim supplemented replica plates, and one of these which was called A18-34 Aps Tps was chosen for further manipulations.

The EAST1 toxin gene was then deleted from the strain A18-34 Aps Tps. For this, conjugations were performed with SM 1 Okpir harbouring a pJCB12 plasmid derivative carrying a defined EAST1 deletion as the donor strain and strain A18-34 Aps Tps as the recipient strain. Transconjugants were selected on L-agar supplemented with tetracycline at 15g/ml and chloramphenicol at 10 ug/ml. One transconjugant was identified by PCR using oligonucleotides 4917 and 4778 in which the pJCB 12-derivative was inserted at the correct location. This tetracycline and chloramphenicol resistant transconjugant was then grown in 5% sucrose medium in order to select recombinants in which the pJCB 12-derivative had excised (as described in Materials and Methods). Colonies that grew on L-agar supplemented with 5% sucrose were then tested by PCR using oligonucleotides 4749 and 4752.

Three isolates were identified that were negative in this PCR reaction indicating that the EAST1 locus had been lost.

Defined deletion mutations were then introduced into the aroC, ompC and ompF genes in order to further attenuate the A18-34 Aps Tps XEAST1 derivative as described in other Examples. Initial attempts to introduce the AaroC mutation met with little success as was the case when constructing a strain J-derivative aroC deletion mutation (see below). Therefore, attempts were made to introduce the SaroC'mutation (see below for details of the mutation). Following transfer of pJCB12-aroCJ into the A18-34 Aps Tps dEASTI derivative a transconjugant was identified by PCR using oligonucleotides 4917 and 4742. This transconjugant was then grown in 5% sucrose medium in order to select recombinants in which the

pJCB 12-derivative had excised (as described in Materials and Methods). Colonies that grew on L-agar supplemented with 5% sucrose were then tested by PCR using oligonucleotides 47116 and 47117, and a derivative that carried the #aroCJ defined deletion mutation was identified. Aromatic amino acid auxotrophy was demonstrated in this A18-34 ApSTpS #AEAST1 #aroCJ strain by streaking on minimal agar and minimal agar supplemented with"aro mix. "The strain grew only on the aro mix supplemented agar, whereas the parental strain A 18-34 Aps Tps zIEAST1 grew both in the presence and absence of"aro mix. " Next the Amp defined deletion mutation was introduced into the A18-34 Aps Tps AEASTI #aroCJ-derivative as described for strain H (Example 4) and a SompC deletion mutant was identified. Expression of CFA/I and LPS by this A18-34 Aps Tps #EASTI#aroCJ #ompC strain was confirmed by SDS-PAGE.

Next the AompF defined deletion mutation was introduced into the A18-34 Aps Tps EAST #aroCJ#ompc-derivative as described for strain H (Example 4) and a AompF deletion mutant was identified. Expression of CFA/I and LPS by this A18- 34 Aps Tps SEAST1 aaron #ompC zlompF strain was confirmed by SDS-PAGE.

Finally, the pACYC-Tc plasmid that was introduced into strain A to confer tetracycline resistance was specifically cured from the A18-34 Aps Tps aEAST1 daroC'dompC dompF strain. For this the plasmid pJCB12-pACYCori described in Example 4 was introduced by electrotransformation. Transformants harbouring pJCB 12-pACYCori were selected by growth on L-agar supplemented with chloramphenicol. One transformant was then grown in 5% sucrose medium in order to select derivatives from which the pJCB12-pACYCori plasmid had been cured (as described in Materials and Methods). Colonies that grew on L-agar supplemented with 5% sucrose were then picked onto L-agar medium supplemented with chloramphenicol and onto L-agar medium supplemented with tetracyclin. This identified a number of derivatives which were sensitive to both these antibiotics, confirming that the pJCB 12-pACYCori plasmid had indeed specifically cured the

strain of the pACYC-Tc marker plasmid and then itself been cured by growth in the presence of sucrose.

One of these antibiotic sensitive derivatives was chosen and tested by PCR for the presence ofthe cfaA, cfåC, and cfaD genes using oligonucleotide pairs 47104 with 47105,4729 with 4730, and 4785 with 4786 respectively. In addition, the presence of the SaroCJ, SompC and SompF deletion mutations was confirmed using the oligonucleotide pairs 47116 with 47117,4732 with 4743, and 4733 with TT1.

Expression of CFA/I and LPS was confirmed in this strain using SDS-PAGE and the nucleotide sequence across the aroCJ, AompC and zlompF mutations was confirmed.

Manipulation of Strain B to produce strain ACAM 2011 Strain B (WS-4437A) expresses the colonisation factor antigen CFA/I and the heat- stable toxin ST. Initially, isolation of a spontaneous ST-deletion mutation was attempted by targeting ST using plasmid pJCB 12-STI. This plasmid was introduced into Strain B by conjugation from SM10Xpir and transconjugants were obtained by growth on L-agar supplemented with chloramphenicol. After several attempts, a number of ST-negative mutants were identified, but these were all negative for CFA/I.

It was therefore decided to construct a defined STI deletion mutation specific for Strain B. The first requirement was to determine nucleotide sequence flanking the 3' end of the ST gene. The process used to do this is illustrated in Figure 20. Plasmid DNA was isolated from a transconjugant strain in which pJCB 12-STI was targeted to the ST gene. Purified DNA was subjected to restriction endonuclease digestion with BglII which cuts the ETEC plasmid but not the ST gene or the pJCB 12 construct..

Digested DNA was ligated and electrotransformed into SY327Xpir and transformants were selected on L-agar supplemented with chloramphenicol. This process is termed"plasmid rescue"and results in the re-isolation of the pJCB 12 replicon incorporating a large fragment of DNA that includes the whole of the STI region and a large amount of flanking DNA. The sequence of the flanking DNA was obtained. Nucleotide sequence data obtained from this then allowed further

oligonucleotides (4797 and 4798) to be designed and used to determine additional nucleotide sequence further downstream of the STI gene in Strain B. Figure 5 shows the determined sequence and the location of all oligonucleotide binding sites.

Using the determined nucleotide sequence, an ST-deletion mutation was constructed.

This was done by amplifying two fragments from the STI locus using oligonucleotides 47101 with 47114, and 47115 with 47100. The two resulting fragments were then joined by overlap extension PCR using oligonucleotides 47100 and 47101 and ligated into the MCS of pJCB12 using standard techniques.

The construct was introduced into strain B by conjugation from SM 1 Okpir and chloramphenicol resistant transconjugants in which ST had been correctly targeted were identified by PCR using oligonucleotides 4917 and 4799. These transconjugants were grown in the presence of sucrose and colonies obtained were then screened using oligonucleotides 47100 and 47101 in order to identify ST deletion mutants. One derivative was found which was negative using these oligonucleotides, suggesting that a spontaneous deletion had resulted in loss of the entire ST locus. This derivative was positive by PCR using oligonucleotides 4727 with 4728 which amplify part of the cfaC gene. It was also positive using oligonucleotides BglIIFOR and BglmodREV which amplify the cfaA gene, and using oligonucleotides 4785 and 4786 which amplify the CFA/I regulator gene, cfaD.

Functional assay for ST confirmed that this derivative was ST-negative, while SDS- PAGE confirmed that it continued to express CFA/I.

Defined deletions were then introduced into the aroC, ompC and ompF genes in order to attenuate the Strain B ST-negative derivative. Initial attempts to introduce the AaroC deletion were unsuccessful. Therefore the jaroCJ deletion was used (see below for details of the mutation). Following transfer of the pJCB 12aarocJ deletion into Strain B ST-negative derivative, transconjugants were identified by PCR using oligonucleotides 4917 with 4742. A transconjugant was identified and then grown in 5% sucrose medium in order to select recombinants in which the pJCB12 derivative had been excised (as described in materials and methods). Colonies that grew on L- agar supplemented with 5% sucrose were then tested by PCR using oligonucleotides

47116 with 47117 and a derivative that carried the #aroCJ deletion was identified.

The CFA/I status of the exconjugant was checked by PCR using oligonucleotides 4727 with 4728 for cfaC, 4785 with 4786 for cfaD and BglIIFOR with BglIImodREV for cfaA.

The Amp defined deletion mutation was then introduced into the Strain B ST- negative zaroCJ derivative as described for strain H (Example 4) and a dompC deletion mutant identified. The CFA/I status of the exconjugant was checked by PCR using oligonucleotides 4727 with 4728 for cfaC, 4785 with 4786 for cfaD and BglIIFOR with BglIImodREV for cfaA.

The AompF defined deletion mutation was introduced into the Strain B ST-negative sarod zompF Amp derivative as described for strain H (Example 4) and an ompF deletion mutant identified. The CFA/I status of the exconjugant was checked by PCR using oligonucleotides 4727 with 4728 for cfaC, 4785 with 4786 for cfaD and BglIIFOR with BglIImodREV for cfaA.

Finally the pStrep plasmid, that confers streptomycin resistance, was specifically cured from the Strain B ST-negative aaron AompF aompC derivative. For this a plasmid derivative of pJCB12 was constructed which incorporated the pStrep replication origin. This was done by shotgun cloning pStrep fragments generated by restriction endonuclease digestion with SphI into the Sphl site of pJCB 12, and transforming the DNA into XL-10 Gold ultracompetent E. coli cells (Stratagene).

Transformants were plated on L-agar supplemented with chloramphenicol. Because pJCB12 is a suicide vector requiring special host strains that carry the pir gene, transformants that grew in the presence of chloramphenicol were assumed to be derivatives of pJCB 12 which carried the pStrep replication origin. Plasmid DNA from four such transformants showed that in each case an Sphl DNA fragment of approximately 2kb had been cloned into pJCB12. This pJCB 12-derivative was called pJCB12-pStrep-ORI The pJCB 12-pStrep-ORI plasmid was introduced into the Strain B ST-negative #aroCJ #ompF #ompC derivative by conjugation from strain SMlORpir, and the

resulting transconjugants selected by growth on agar medium supplemented with chloramphenicol. Colonies which grew were used to inoculate broth supplemented with chloramphenicol and, following incubation to allow growth, dilutions of this culture were plated on agar supplemented with chloramphenicol. Colonies which grew on this medium were picked onto agar supplemented with streptomycin in order to identify those from which the pStrep plasmid had been lost. One of these streptomycin sensitive colonies was then grown in L-broth supplemented with 5% sucrose in order to select for a derivative from which the pJCB 12-pStrep-ORI plasmid had been lost. Dilutions from this broth culture were plated on L-agar supplemented with 5% sucrose and after incubation some of the resulting colonies were picked onto agar supplemented with chloramphenicol in order to identify those from which the pJCB12-pStrep-ORI plasmid had been lost.

The CFA/I status of the exconjugant was again checked by PCR using oligonucleotides 4727 with 4728 for cfaC, 4785 with 4786 for cfaD and BglIIFOR with BglIImodREV for cfaA. CFA/I protein expression was then checked by SDS- PAGE and Western blotting. The LPS profile was also checked. In addition the aaroCJ, åompC and AompF mutations were confirmed by sequencing, and the aromatic amino acid dependence of the strain was also confirmed.

EXAMPLE 3: A DERIVATIVE OF A VIRULENT WILD-TYPE ETEC STRAIN WHICH EXPRESSES CSS, CS6, LT, ST AND EAST1 WHEREIN THE LT, ST AND EAST1 GENES HAVE BEEN DELETED Strain E expresses CS5, CS6 and the toxins EAST1, ST and LT. In order to facilitate genetic manipulation plasmid pACYC-Tc (described in Example 2 above) was introduced into strain E by electro-transformation in order to confer tetracycline resistance.

The EAST1 toxin gene was deleted from Strain E first. This required the construction of a pJCB 12 derivative which carries a defined EAST1 deletion mutation. Such a deletion mutation was generated by amplifying EAST1 fragments

from the ETEC strain H10407, using oligonucleotides 4749 with 4750, and 4751 with 4752, to generate two DNA fragments flanking the EAST1 gene. Figure 6 shows the sequence of the EAST1 locus derived from IS 1414 (Genbank accession number AF 143 819) and all oligonucleotides used. These were then fused by an additional overlap extension PCR reaction using primers 4749 and 4752 and the resulting fragment was cloned into pJCB 12 using the SalI and SphI restriction sites.

Thus, a pJCB 12 derivative had been constructed which incorporated this deletion mutation.

Strain SM10Xpir harbouring pJCB12-AEASTI was conjugated with Strain E and transconjugants selected. Colonies that grew on L-agar supplemented with chloramphenicol and tetracycline were screened using the oligonucleotides 4917 and 4753 (equivalent to oligos 4 and 1 in Figure 4). Two transconjugants were obtained which were positive in both this PCR reaction indicating that the pJCB12-AEASTl plasmid had correctly targeted the EAST1 gene.

These chloramphenicol resistant transconjugants were then grown in 5% sucrose medium in order to select recombinants in which the pJCB 12-derivative had excised.

Four exconjugants were obtained on the sucrose agar, and these were negative when tested for the presence of the EAST1 gene by PCR using oligonucleotides 4749 and 4752, indicating that the whole of the EAST1 region had been lost in this derivative.

The strain E EAST1-negative derivative was then tested by PCR using oligonucleotides 4738 with 4780 which are specific for the CS5 operon, and oligonucleotides 4740 with 4781 which are specific for the CS6 operon. These PCR reactions were positive for all four exconjugants confirming that they continued to harbour the CS5 and CS6 genes.

Next, an LT-deletion mutant of the Strain E EAST 1-negative derivative was constructed. For this a pJCB 12 derivative plasmid was constructed that carried a defined deletion of the LT-A gene in the LT locus. See Figure 7.

A defined LT-A deletion was constructed by PCR amplification using oligonucleotides 4772 with 4773, and 4774 with 4746, to generate two DNA fragments flanking the LT-A gene. These were then fused by an additional overlap extension PCR reaction and the resulting fragment was cloned into pJCB12 using the SalI and SphI restriction sites.

The pJCB12-ALT-A construct was introduced into the Strain E zlE5T7 derivative by conjugation from the pJCB12 host strain SM I Okpir and transconjugants were selected on L-agar supplemented with chloramphenicol and tetracycline. Resulting colonies were screened by PCR using oligonucleotides 4762 and R6K-01 which identified two transconjugants in which the LT locus had been correctly targeted.

These chloramphenicol resistant transconjugants were then grown in 5% sucrose medium in order to select recombinants in which the pJCB 12-derivative had been excised. Colonies that grew on the sucrose agar were tested by PCR using oligonucleotides 4762 and 4746 in order to identify LT-A deletion mutants. Three derivatives were identified which were negative in this PCR reaction, suggesting that they had lost the entire LT locus. PCR reactions performed on these three LT- negative derivatives using oligonucleotides 4738 with 4780, and oligonucleotides 4740 with 4781 confirmed the presence of the CS5 and CS6 genes respectively.

Next, the ST gene was deleted from this Strain E EAST 1-negative LT-negative derivative. For this, plasmid pJCB 12-STI (as described in Example 2) was transferred into Strain E from SM I Okpir and transconjugant selected on L-agar supplemented with chloramphenicol. A transconjugant was identified by PCR using oligonucleotides 4917 and 4794, and this transconjugant was then grown in 5% sucrose medium to select for derivatives from which the pJCB 12-STI had been lost.

Of 133 colonies screened, only 3 had lost STI, the others being revertants, and all of these 3 STI-negative derivatives were negative also for CS5 and CS6. These results suggested that the STI locus was present on the same plasmid as the CS5 and CS6 genes (as did Southern hybridization data) and that the STI locus was relatively stable, so that revertants or derivatives were obtained in which only specific plasmid curing had occurred.

It was therefore decided that a defined STI-deletion mutation specific for strain E was needed. In order to make this construct additional nucleotide sequence data was required for the region downstream of the STI gene in Strain E. Therefore, one of the pJCB 12-ST transconjugants was used for sequence determinations at the STI locus using the plasmid rescue procedure described in Example 2 and illustrated in Figure 20. A pJCB12-derivative was obtained that incorporates a large fragment of DNA that includes the ST-I gene and a large amount of flanking DNA from Strain E.

This plasmid preparation was used as template in nucleotide sequence determination reactions using oligonucleotide 4764 to determine sequence through the STI gene, and oligonucleotide 4792 to determine sequence downstream of the STI gene. The new sequence data allowed an additional oligonucleotide, 47106, to be designed which was used in further nucleotide sequence determinations. This additional new sequence data enabled oligonucleotides 47112,47120 and 47121 to be designed for construction of the deletion cassette. The sequence of the ST-1 gene and flanking regions showing the binding sites of all oligonucleotides used is given in Figure 8.

As with other deletion mutations, two DNA fragments from the ST region were amplified by PCR using oligonucleotides 4764 with 47120, and 47121 with 47112, and the two fragments generated were fused by an additional overlap extension PCR reaction using oligonucleotides 4764 and 47112. The resulting fragment was cloned into pJCB 12 using the SacI and SalI restriction endonuclease sites of pJCB 12 and those incorporated into oligonucleotides 4764 and 47112 respectively.

The resulting recombinant plasmid pJBC12-ASTIE was introduced into the Strain E EASTI-negative LT-negative derivative from strain SM10Xpir. The resulting chloramphenicol and tetracycline resistant transconjugants were screened by PCR using oligonucleotides 4917 with 4794, and R6K-01 with 47113. Two transconjugants were identified that were positive, indicating that the STI gene had been correctly targeted.

These chloramphenicol resistant transconjugants were then grown in medium supplemented with sucrose, followed by plating on 5% sucrose agar to obtain

excision of the pJCB 12 derivative. The colonies that grew on sucrose agar were screened using oligonucleotides 4764 with 47112 and a number were identified that were negative, and three colonies were identified that showed the presence of the STI deletion mutation. PCR reactions performed on the STI deletion mutants using oligonucleotides 4738 with 4739 to detect CS5, 4740 and 4741 to detect CS6 and 4783 with 4784 to detect the CS5 regulator gene were all positive, indicating that these genes had been retained in these defined STI-deletion mutants. Additional attenuating mutations in aroC, ompC and ompF were introduced as described (ref 32 and Example 4). See Figure 9.

EXAMPLE 4: REMOVAL OF TOXIN GENES AND INTRODUCTION OF ATTENUATING MUTATIONS INTO STRAINS EXPRESSING CS2/CS3 (STRAIN H) AND CS4/CS6 (STRAIN J) Strain H Manipulation Strain H expresses CS2 and CS3 and the toxins EAST and ST. It also has the genes for LT, but LT protein has not been detected by in vitro assays. In order to facilitate genetic manipulation the plasmid pACYC-Tc (described in Example 2) was introduced into Strain H by electro-transformation in order to confer tetracycline resistance.

The EAST1 toxin gene was the first to be deleted from Strain H. Strain SM10Xpir containing the pJCB12-AEASTl plasmid described in Example 3 was conjugated with Strain H and transconjugants selected on L-agar containing chloramphenicol and tetracycline. Transconjugant colonies in which the EAST1 gene had been correctly targeted were identified by PCR using the oligonucleotides 4917 and 4753.

Transconjugants were then processed in order to select derivatives in which the pJCB12-derivative had excised (described above). Colonies that grew on 5% sucrose agar were tested by PCR using the oligonucleotides 4775 and 4777 and a number were identified that were negative by PCR indicating that the whole of the EAST1 region had been lost. These Strain H EAST 1-negative derivatives were then tested for the presence of a transcriptional activator for colonization factors, rns, by PCR using the oligonucleotides RNS-03 and RNS-04. Two of the Strain H EAST1-

negative mutants were positive for rns. Further testing by PCR for CS2 using oligonucleotides 4712 and 4779, and CS3 using oligonucleotides CS3-02 and CS3- 03 indicated that the mutants were both CS2 and CS3 positive. Testing by PCR for the LT locus using oligonucleotides LT-04 and LT-05, and for the STI locus using oligonucleotides EST-01 and 4765 showed that the mutants were negative for these toxins, indicating that the ST and LT loci had been lost concomitantly with EAST1.

The expression of CS2 and CS3 in the Strain H toxin-negative mutants was confirmed by SDS-PAGE.

Defined deletion mutations were then introduced into the aroC, ompC, and ompF genes in order to further attenuate the Strain H toxin-negative derivative using a method similar to that described previously (32). However, this earlier description for introducing these mutations used the suicide vector pCVD442 described in Example 1. This suicide vector codes for ampicillin resistance, which is not optimal for selecting infrequent transconjugant from a large mixed bacterial population. In addition, at this stage, the Strain H toxin-negative derivative continues to express it's own ampicillin resistance, making pCVD442 useless with this strain. Therefore, the AaroC, AompC and SompF deletion mutations previously cloned into pCVD442 (32) were sub-cloned into pJCB 12. For this, the zl aroC mutation was sub-cloned using the restriction endonuclease sites XbaI and SacI, AompC used SacI and Sall sites, and SompF used SacI and SphI sites. The pJCB 12-derivatives were then transformed into strain SMlO, pir.

The defined AaroC mutation was the first to be introduced into the Strain H toxin- derivative from strain SMIOr Transconjugant colonies that grew on L-agar supplemented with chloramphenicol and tetracycline were screened by PCR using the oligonucleotides 4917 and 4742 in order to identify those in which the aroC locus had been correctly targeted. These transconjugant colonies were then streaked onto fresh L-agar supplemented with chloramphenicol and tetracycline. Following incubation, the colonies that grew were tested for the presence rns by PCR using the oligonucleotide primers RNS-03 and RNS-04 and for CS2 using primers 4712 and 4779. The reactions confirmed that the transconjugants were positive for both these loci. Transconjugants were then processed in order to select derivatives in which the

pJCB 12-derivative had excised (as described in previous Examples). Colonies that grew on 5% sucrose agar were tested by PCR using the oligonucleotide primers 4731 and TT20 in order to identify derivatives that harboured the defined AaroC deletion mutation. One defined AaroC deletion mutant was identified and again checked by PCR to confirm the presence of rns and CS2, as described above. Expression of CS2 and CS3 by this defined AaroC deletion mutant was confirmed using SDS-PAGE, and its LPS was checked using SDS-PAGE.

Next the ompC defined deletion mutation was introduced into this Strain H toxin- negative AaroC derivative by conjugation from strain SM10Xpir harbouring pJCB12- AompC. Transconjugant colonies that grew on L-agar supplemented with chloramphenicol and tetracycline were screened by PCR using the oligonucleotides 4917 and 4743 in order to identify those in which the ompC locus had been correctly targeted. Transconjugants were then processed in order to select derivatives in which the pJCB12-derivative had excised (as described above). Colonies that grew on 5% sucrose agar were tested by PCR using the oligonucleotide primers 4732 and 4743 in order to identify derivatives that harboured the defined DompC deletion mutation.

The defined ompF mutation was then introduced into this Strain H toxin-negative AaroC AompC derivative in a similar fashion to that described for the ompC mutation above, except that transconjugant colonies were screened by PCR using the oligonucleotides R6K-01 and 4733 in order to identify those in which the ompF locus had been correctly targeted. Colonies that grew on 5% sucrose agar were tested by PCR using the oligonucleotides 4733 and TT1 in order to identify derivatives that harboured the defined AompF deletion mutation.

All of the defined deletion mutations of the Strain H toxin-negative AaroC DompC AompF derivative were then checked by PCR for comparison with the wild-type Strain H. For this the oligonucleotides used for aroC were 4731 and TT20, for ompC were 4732 and 4743, and for ompF were 4733 and TT1. The PCR products generated by these reactions were used for nucleotide sequence determinations across the deletion mutations. The oligonucleotides used for the nucleotide sequence determination reactions were TT35, TT38 and TT33 for aroC, ompC and ompF

respectively. All the deletion mutations had the expected nucleotide sequence. The mutant was also checked by PCR for the presence of rns and CS2, and for the absence of the toxin loci, as described above. The presence of the CS3 locus was also confirmed by PCR using the oligonucleotides CS3-03 and CS3-06. Expression of CS2 and CS3 was confirmed using SDS-PAGE, and the LPS was checked using SDS-PAGE.

The ampicillin resistance determinant was then removed from the Strain H toxin- negative AaroC, AompC, AompF derivative. For this the derivative strain was grown through three passages in L-broth and then dilutions plated on L-agar in order to obtain well-separated colonies. The colonies were then replica plated onto L-agar supplemented with ampicillin at 100 llg/ml. After incubation to allow the colonies to re-grow, the L-agar plates were examined in order to identify any colonies present on the L-agar that did not grow on the ampicillin supplemented L-agar. One such colony was identified and used in further experiments.

Finally, the pACYC-Tc plasmid that was introduced into Strain H to confer chloramphenicol resistance was specifically cured from the Strain H toxin-negative AaroC, AompC, AompF derivative. This required the construction of the plasmid pJCB 12-pACYCori. For this the pACYC-Tc origin of replication was amplified by PCR using the oligonucleotides 4760 and 4761 and cloned into pJCB12 using the restriction endonuclease sites SacI and SalI. Strain SM 1 Okpir containing pJCB12- pACYCori was conjugated with the Strain H toxin-negative-AaroC, DompC, AompF derivative, and transconjugants were selected on L-agar containing chloramphenicol.

It was hoped that by introducing a second plasmid carrying the same pACYC 184 replication origin and selecting for its maintenance, the first pACYC-Tc plasmid would be rendered unstable and in the absence of any selection for its maintenance would be more frequently lost spontaneously. Chloramphenicol resistant transconjugant colonies harbouring the pJCB 12-pACYCori plasmid derivative were picked onto L-agar supplemented with tetracycline and one of these colonies was found to be tetracycline sensitive, indicating that the pACYC-Tc plasmid had been lost. This tetracycline sensitive, chloramphenicol resistant colony was then grown on 5% sucrose medium in order to select for derivatives from which the pJCB 12-

pACYCori had been lost spontaneously. Colonies that grew on 5% sucrose agar were picked onto L-agar supplemented with chloramphenicol to confirm that the pJCB 12-pACYCori plasmid had indeed been lost.

These manipulations are shown schematically in Figure 10.

Strain J Manipulation Strain J expresses CS4 and CS6 and the toxins EAST and ST. It also has the genes for LT, but LT protein has not been detected by in vitro assays. In order to facilitate genetic manipulation the plasmid pACYC-Tc (described in Example 2) was introduced into strain J by electro-transformation thus conferring. tetracycline resistance.

The EAST1 toxin gene was deleted from strain J first. Strain SMIOp containing pJCB12-AEASTl was conjugated with Strain J and transconjugants selected on L- agar containing chloramphenicol and tetracycline. Colonies that grew were screened by PCR using the oligonucleotides 4917 and 4753 in order to identify those in which the EAST1 gene had been correctly targeted. These transconjugant colonies were then streaked onto fresh L-agar supplemented with chloramphenicol and tetracycline and incubated to allow colonies to grow. These colonies were tested by PCR to confirm the continued presence of the CS4 operon using the oligonuncleotides 4768 and 4769, and for the CS6 operon using oligonuncleotide 4740 and 4781.

One transconjugant which was positive for both these PCR reactions was processed in order to select derivatives in which the pJCB 12-derivative plasmid had excised (described above). Colonies that grew on 5% sucrose agar were tested by PCR using the oligonucleotides 4749 and 4752 to identify EAST1 mutants. All colonies tested were negative by this PCR reaction indicating that the whole of the EAST1 region had been lost in these derivatives. Further PCR reactions were performed to test for the continued presence of the CS4 and CS6 genes as described above, and expression of CS6 was confirmed by SDS-PAGE.

PCR reactions using oligonucleotides LT-04 and LT-05 amplified a product of the expected size for the LT locus in the Strain J EAST1-negative derivative. Nucleotide sequence determination reactions using these same oligonucleotides showed that this fragment generated by PCR was indeed of the LT locus. Therefore, this locus was targeted using the pJCB12-ALT-A construct in the same way as described in Example 3. LT-negative derivatives were identified in which the whole of the LT locus had been deleted. Further PCR reactions to test for the presence of CS4 using oligonucleotides 4768 and 4769, and CS6 using the oligonucleotides 4740 and 4781 confirmed the continued presence of these loci in the LT-negative derivatives. The expression of CS6 by the Strain J EAST1-negative LT-negative derivatives was confirmed by SDS-PAGE.

PCR reactions performed using the oligonucleotides ST-01 and ST-02 indicated that the ST present in Strain J was that encoded by the transposon Tnl681 (41). The Strain J EAST1-negative LT-negative derivatives were tested by PCR using the oligonucleotides ST-01 and ST-02, and were shown to be ST-negative.

The aroC locus of the Strain J toxin-negative derivative was targeted using pJCB 12- AaroC as described above for Strain H. Correctly targeted transconjugants were identified. Unusually, however, following growth on 5% sucrose medium all derivatives generated in a large number of experiments were identified as revertants.

Therefore a new pJCB 12-AaroC construct was made that incorporated a smaller deletion in the aroC locus. This was constructed by PCR amplification using oligonucleotides 47116 with 47118, and 47119 with 47117, to generate two DNA fragments flanking that region of the aroC gene to be deleted. These were then fused by an additional overlap extension PCR reaction using oligonucleotides 47116 with 47117, and the resulting fragment was cloned into pJCB 12 using the XbaI and SacI restriction endonuclease sites. This construct was called pJCB12-aroCJ and was electrotransformed into SMl OXpir. The sequence of the aroC gene and the binding sites of the oligonucleotides used to construct this novel deletion construct are shown in Figure 11.

While pJCB12-haroCJ was undergoing construction, the ompC defined deletion mutation was introduced into the Strain J toxin-negative derivative using the pJCB 12-AompC construct and procedure described for Strain H. One ompC defined deletion mutant was identified by PCR using oligonucleotides 4732 and 4743.

The defined aroC deletion mutation was then introduced into this Strain J toxin- negative DompC mutant using the new pJCB12-DaroCJ construct by conjugation from strain SMIOXpir. Colonies that grew on L-agar supplemented with chloramphenicol and tetracycline were screened by PCR using the oligonucleotides 4917 and 4742 in order to identify those in which the aroC gene had been correctly targeted. Transconjugants were then processed in order to select derivatives in which the pJCB 12-derivative plasmid had excised (described above). Colonies that grew on 5% sucrose agar were tested by PCR using the oligonucleotides 4731 and TT20 in order to identify those with the incorporated defined arocJ deletion mutation.

The defined ompF deletion mutation was then incorporated into an identified Strain J toxin-negative AompC AaroC derivative. This was done exactly as described above for strain H, and one ompF deletion mutant was identified.

The tetracycline resistant plasmid pACYC-Tc was then specifically cured from the toxin-negative AompC AaroC AompF derivative of Strain J using pJCB 12- pACYCori and then this plasmid itself was removed, again as described above for Strain H.

Finally, PCR reactions were performed to confirm the presence of CS4 using oligonucleotides 4768 with 4769, and CS6 using the oligonucleotides 4740 with 4781. Similarly the presence of the CFA/IV regulatory gene was confirmed using oligonucleotides 4785 and 4786. The nucleotide sequences of the defined aroC, ompC and ompF deletion mutations were determined as described for Strain H, and were as expected. The absence of EAST1, LT and ST loci were again confirmed by PCR as described above, expression of CS6 was confirmed by SDS-PAGE, and the LPS also was checked using SDS-PAGE.

A diagram showing all of the steps involved in these manipulations is shown in Figure 12.

EXAMPLE 5: INCREASING THE STABILITY OF MANIPULATED CFA PLASMIDS This Example describes as an illustration one system for stabilising a CFA plasmid.

The parDE locus was used as the stabilizing moiety.

In order to obtain a derivative of an ETEC strain that does not code for ST but continues to express its CFA genes, a pJCB12-ASTI derivative is constructed where the parDE locus is flanked by regions homologous to the regions flanking the ST gene in the plasmid to be targeted. Introduction of this plasmid into a recipient strain and selection on chloramphenicol will now result in transconjugants in which both loci are present in the CFA plasmid. The functional parDE locus will now ensure that this plasmid is maintained in these cells. Growth on sucrose to identify derivatives from which the pJCB 12 component has been lost will now strongly select for cells containing a recombinant plasmid in which the desired cross over (ie the replacement of the ST structural gene with the parDE locus) has taken place. Cells in which a reversion event has occurred or from which the entire plasmid has been lost will be killed by the action of the toxin. Incorporation of the parDE locus will also function to stabilise the inheritance of the CFA plasmid during subsequent rounds of mutation, for example the introduction of additional attenuating mutations such as in the aroC, ompC and ompF genes.

Oligonucleotides 4789 and 4790 are used to amplify theparDE locus from plasmid RP4. The purified PCR product is then cloned into pJCB12-ASTI using the SoI restriction enzyme sites to give plasmid pJCB 12-ASTI : . parDE. This plasmid is now introduced into a recipient strain by conjugation from SM10kpir or electro- transformation with purified plasmid DNA and the required intermediates and derivatives isolated by methods described in detail in the previous Examples.

EXAMPLE 6: EXPRESSION OF LT-B IN AN ATTENUATED ETEC STRAIN, PTL003, TO INDUCE A PROTECTIVE IMMUNE RESPONSE AGAINST LT Aim The aim of this work was to express the B-subunit of Escherichia coli heat labile toxin in a vaccine strain of ETEC. The LTB gene was derived from strain WS2773-E (NAMRU3, Cairo, Egypt) but it could be derived from any LTB-encoding strain.

Similarly the approach could be extended to include mutants of LT-A, LT-B fused to other proteins (eg ST) or to expression of CT-B or derivatives thereof. Initial plasmid constructs were designed to demonstrate that LT-B could be expressed in an ETEC vaccine strain in the absence of LT-A and could be correctly exported and assembled. Subsequent constructs used the native LT promoter to drive expression.

Ultimately, a similar construct could be inserted into the chromosome of ETEC to create a stable strain without the need for antibiotic selection.

Methods Section 1-PCR amplification of the LTB CDS (protein coding sequence) Primers Genbank sequences were used to design appropriate PCR primers (Table 1). The forward primer (Bfor) was designed to amplify from the start codon of the LTB gene and was based on Genbank sequence M17874 (17). To facilitate cloning into expression vectors, a BglII restriction site was included starting 8 bases 5'prime of the ATG start codon. The reverse primer (Brev) was designed to amplify from 200 bases downstream of the stop codon, such that any transcription terminators would be included in the PCR product, and was based on Genbank sequence AF 190919 (26). An NheI restriction site was included in the primer to facilitate cloning into expression vectors.

Template Plasmid DNA was isolated from Strain WS2773-E (NAMRU3, Cairo, Egypt) for use as a template.

Reaction Reaction mixture 2 pI plasmid DNA from WS2773-E (approx 50 ng/gl) 10 µl 10X Pfu DNA Polymerase buffer (Stratagene) 0.8 pI dNTPs (25 mM of each dNTP) 0.5 Ill primer BFOR (258 ng/µl) 0.5 µl primer BREV (227 ng/lul) 84 pI Hz0 2 pI PfuTurbo DNA Polymerase (2. 5U/µl) Program 1 cycle 94°C x 1 min 30 cycles 94°C x 1 min 56°C x 1 min 72°C x I min 1 cycle 72°C x 10 min Hold 4°C Results A 600 bp PCR product was synthesized and isolated from a 1 % agarose gel using a QIAquickTM gel extraction kit (Qiagen) according to the manufacturer's instructions (Figure 13).

Section 2-Cloning of the LTB PCR product Cloning into pPCR-Script Amp SK+ Gel isolated PCR product was ligated into pPCR-Script Amp SK+TM (Stratagene) according to the instructions in the manufacturer's instruction manual (#211188). 2 u. l ofligation mix was used to transform E. coli XL 1 O-GoldTM Supercompetent cells (Stratagene &num 230350) and correct constructs were identified by digestion of purified plasmid DNA with PvuII. A correct construct was designated pPCRLTB. The LTB gene was fully sequenced (Fig 14) and was compared to Genbank sequence M 17874 (17). Four base changes were found which resulted in two amino acid changes. The PCR was repeated and a second clone was sequenced and gave identical results, showing that the LT-B sequence in strain WS2773-E differs from the database sequence.

Cloning into an inducible expression vector The LTB gene was transferred into expression vector pNCAT4 (Fig 15). pNCAT4 is an expression vector designed for use in Salmonellas typhi and typhimurium. The vector was originally derived from pTETnirlS (3) but had been modified by replacement of the TetC gene by the H. pylori catalase gene and replacement of the bla ampicillin resistance gene by a kanamycin resistance gene. In this plasmid the catalase gene is under the control of the nirB promoter which is upregulated in anaerobic conditions. For the purpose of the present invention many alternative expression plasmids which are well known in the art could be substituted for pNCAT4 or pTETnir 15.

Plasmid pPCRLTB was digested with restriction enzymes BglII and NheI and a 600 bp fragment containing the LTB gene was isolated from a 1 % agarose gel using a QIAquickTM gel extraction kit (Qiagen) according to the manufacturer's instructions.

Plasmid pNCAT4 was digested with restriction enzymes BglII and NheI and a 2.6 kb

vector fragment was isolated from a 1 % agarose gel using a QIAquickTM gel extraction kit (Qiagen) according to the manufacturer's instructions.

The vector was ligated to the LTB gene (25) and the ligation mixture was used to transform E. coli XL10-GoldTM supercompetent cells (Stratagene #230350). Correct clones were identified by restriction enzyme analysis and one (designated pNLTB) was used to transform electrocompetent ETEC-PTL003.

Section 3-Expression and localization of LTB in ETEC-PTL003 Sample preparation Cultures of PTL003 and PTL003-pNLTB were grown overnight in LB medium with appropriate antibiotics. Samples (1 ml) were diluted in 100 ml medium and grown with shaking for-3 h. Flasks were transferred to a static incubator and growth was continued for a further 1.5 h.

Bacteria were fractionated into periplasmic and spheroplast components by the following method: Cells were harvested by centrifugation for 5 min at 10 400 x g.

The cell pellet was resuspended in 5 ml of 200 mM Tris HCI, pH8. 0 and was transferred to a 50 ml glass beaker containing a magnetic flea. To this was added 5 ml of TES buffer (200 mM Tris HCI, pH 8.0, 1 mM Na2EDTA, 1 M sucrose).

Stirring was set in motion and a timer was started. At t=90 s, 1 mg of lysozyme (10 mg/ml solution) was added, and at t=135s, 10 ml of water was added; the mixture was incubated at room temperature for a further 30 min. Spheroplast formation was checked with a light microscope and, if complete, the preparation was centrifuged for 20 min at 47 500 x g at 20°C. The supernatant containing the periplasmic proteins was harvested and concentrated-two-fold in a Centricon MacrosepTM spin- concentrator (3 000 NMWL, Millipore). The pellet, containing spheroplast proteins (cytoplasmic and membrane-bound), was resuspended in PBS. Samples of the cell fractions were mixed with equal quantities of Tris-Glycine SDS-PAGE sample

buffer (Invitrogen LC267) containing 0.1 M dithiothreitol. A portion of the periplasmic sample was heated at 95°C for 5 min, the remainder was kept at room temperature.

Sample analysis Samples were separated by electrophoresis on 18% Tris-Glycine Gels (Invitrogen EC6505) and transferred onto nitrocellulose membranes according to the instructions supplied with the Xcell II Blot Module (Invitrogen EI9051). Blots were incubated as follows: Block 1 hour in PBS, 0.05% Tween 20TM containing 5% dried milk (MarvelTM) with gentle rocking.

Primary antibody 1 hour with rabbit anti-cholera toxin antibody (Sigma C3062) diluted 1: 10 000 in PBS, 0.05% Tween 20TM containing 1% Marvel and a 1 : 2 000 dilution of an E. coli extract (Promega &num S3761). The antibody/extract mixture was pre-incubated for 1 hour before use to reduce non-specific binding to E. coli proteins.

Wash Three 5 min washes in PBS, Tween 20TM, 1% MarvelTM.

Secondary antibody 1 hour with horseradish peroxidase conjugated anti-rabbit antibody (Sigma A4914) diluted 1: 10 000 in PBS, 0.05% Tween 20TM, 1 % MarvelTM.

Wash Three 5 min washes in PBS, Tween 20TM, 1% MarvelTM.

Two 5 min washes in PBS, Tween 20M Two 5 min washes in PBS Develop Blots were developed using a SuperSignal West Pico Chemiluminescent Substrate kit (Pierce).

The results of the Western blots are presented in Figure 16. LTB monomers were detected in both spheroplast and periplasmic boiled fractions. In the periplasmic samples kept at room temperature LTB pentamers were detected indicating that the LTB could be transported across the cell membrane and assembled in the normal way.

Section 4-PCR amplification of the LTB promoter Primers PCR primers were designed based on preliminary sequence data of the upstream region of the LTA gene of WS2773-E (Figure 7). The forward primer (Pfor) annealed-200 bp upstream of the LTA gene. A KpnI restriction site was included in the primer to facilitate cloning into expression vectors. The reverse primer (Prev) annealed just upstream of the start codon of the LTA gene and was designed to introduce a BglII restriction site to allow correct positioning of the promoter fragment with respect to the LTB gene.

Template Plasmid DNA was isolated from Strain WS2773-E (NAMRU3, Cairo, Egypt) for use as a template.

Reaction Reaction conditions were as described for the LTB gene in Section 1.

Results A 200 bp PCR product was synthesized and isolated from a 1 % agarose gel using a QIAquickTM gel extraction kit (Qiagen) according to the manufacturer's instructions.

(Figure 17) Section 5-Cloning of the LT promoter Cloning into pPCR-Script Amp SK+ Gel isolated PCR product was ligated into pPCR-Script Amp SKTM+ (Stratagene) according to the instructions in the manufacturer's instruction manual (#211188). 2 VLI of ligation mix was used to transform E. coli XL 10-GoldTM Supercompetent cells (Stratagene #230350) and correct constructs were identified by digestion of purified plasmid DNA with PvuII. A correct construct was designated pPCRProm. The sequence of this construct is described in Figure 18.

Cloning into an LTB expression vector Plasmid pPCRProm was digested with KpnI and BglII and the 200 bp promoter fragment was isolated from a 1 % agarose gel using a QIAquickTM gel extraction kit (Qiagen) according to the manufacturer's instructions. The nirB promoter of pNLTB was excised by digested with KpnI and BglII (Figure 15) and the remaining 3.7kb vector fragment was isolated from a 1% agarose gel using a QIAquickTM gel extraction kit (Qiagen) according to the manufacturer's instructions. The vector was ligated to the LT promoter fragment (Sambrook et al., 1989, ref 13) and the ligation mixture was used to transform E. coli XL10-GoIdTM supercompetent cells (Stratagene &num 230350). Correct clones were identified by restriction enzyme analysis of purified plasmid DNA and one (designated pLLTB) was used to transform electrocompetent ETEC-PTL003.

Section 6-Expression of LTB under the control of the LT promoter Sample preparation Cultures of PTL003 and PTL003-pLLTB were grown overnight in LB medium with appropriate antibiotics. Absorbance at 600nm was used to determine the concentration of cells in the cultures. Aliquots containing 5 x 107 cells were concentrated by centrifugation and the pellets were resuspended in 10 sul Tris- Glycine SDS-PAGE sample buffer (Invitrogen LC267) containing 0.1 M dithiothreitol.

Sample analysis Samples were analysed by SDS-PAGE (16) and Western blotting exactly as described in Section 3.

Results are presented in Figure 19. LTB was expressed in PTL003 under the control of the LTAB promoter.

Table 1: Oligonucleoides used<BR> Name Nucleotide sequence Target locus; use<BR> cotA; detection of CS2 locus<BR> oriR6K; construction of pJCB2<BR> oriR6K; construction of pJCB2<BR> moRP4; construction of pJCB2<BR> mobRP4; construction of pJCB2<BR> cat; construction of pJCB2<BR> cat; construction of pJCB2<BR> cat; construction of pJCB2<BR> cat; construction of pJCB2<BR> sacB; construction of pJCB2<BR> sacB; construction of pJCB2<BR> cfaC; detection of CFA/I<BR> cfaC; detection of CFA/I<BR> aroC; confirmation of linkage with pJCB12 and mutation<BR> ompC<BR> ompF<BR> detection of CS5<BR> detection of CS5<BR> cssA; detection of CS6<BR> cssA; detection of CS6<BR> aroC; confirmation of mutation<BR> ompC<BR> eltB; construction of #LT-A<BR> EAST1; confirmation of linkage with pJCB12<BR> EAST1; construction of EAST1 deletion mutation<BR> EAST1; construction of EAST1 deletion mutation<BR> EAST1; construction of EAST1 deletion mutation<BR> EAST1; construction of EAST1 deletion mutation<BR> EAST1; to check for linkage with pJCB12 ori pACYC184; construction of pJCB12-pACYCori<BR> ori pACYC184; construction of pJCB12-pACYCori<BR> LT, construction of #LT-A<BR> ST strain A; construction of of ST deletion mutation<BR> ST strain A; construction of of ST deletion mutation<BR> cat; nucleotide sequence deterinations into sacB<BR> csfA; detection of CS4<BR> csfA; detection of CS4<BR> LT, construction of #LT-A<BR> LT, construction of #LT-A<BR> LT, construction of #LT-A<BR> detection of EAST1<BR> detection of EAST1<BR> cotC; detection of CS2<BR> detection of CS5<BR> cssB; detection of CS6<BR> csvR; detection of CS5 regulator gene<BR> csvR; detection of CS5 regulator gene<BR> cfaD; detection of CFA/I regulator gene.<BR> cfaD; detection of CFA/I regulator gene<BR> parDE; construction of pJCB12-#STB::parDE<BR> parDE; construction of pJCB12-#STB::parDE<BR> STIB, nucleotide sequence downstream of STI in StrainB<BR> STI, sequencing upstream of STI and confirmation of linkage<BR> STIB, nucleotide sequence downstream of STI in StrainB<BR> STIB, nucleotide sequence downstream of STI in StrainB<BR> STIB, construction of of pJCB12-#STIB<BR> STIB, construction of of pJCB12-#STIB<BR> STIB, construction of of pJCB12-#STIB<BR> STIE, nucleotide sequence downstream of STI in StrainB<BR> STIE, construction of pJCB12-#STIE<BR> STIE, confirmation of linkage of STI in Strain E<BR> STIB, construction of pJCB12-#STIB STIB, construction of of pJCB12-#STIB<BR> aroC; construction of pJCB12-#aroCJ<BR> aroC; construction of pJCB12-#aroCJ<BR> aroC; construction of pJCB12-#aroCJ<BR> aroC; construction of pJCB12-#aroCJ<BR> STIE, construction of pJCB12-#STIE<BR> STIE, construction of pJCB12-#STIE<BR> cat of pJCB12; confirmation of linkage.<BR> <P>LT-B;<BR> cfaB; detection of CFA/I<BR> cfaB; detection of CFA/I<BR> LT-B;<BR> detection of CS3 genes<BR> detection of CS3 genes<BR> detection of CS3 genes<BR> detection of ST genes<BR> eltA; detection of LT genes<BR> eltB; detection of LT genes<BR> LT promoter<BR> LT promoter<BR> oriR6K; confirmation of linkage<BR> rns; detection of CFA/II regulatory gene<BR> rns; detection of CFA/II regulatory gene<BR> Detection of STTn1681<BR> Detection of STTn1681<BR> ompF<BR> aroC<BR> ompF; nucleotide sequence determinations across #ompF<BR> aroC; nucleotide sequence determinations across #aroC<BR> ompC; nucleotide sequence determinations across #ompC

Table 2 GenBank Accession numbers for sequence data EAST1 (astA) AF143819 ST (estA) M18346 LT-A (pelta) V00275 LT-B (eltB) M17874 CFA/I operon M55661 CS2 operon Z47800 CS3 operon X16944 CS4 operon AF296132 CS5 operon AJ224079 CS6 operon U04844 cfaD M55609 csvR X60106 rns J04166 parDE RK2 L05507 sacB X02730 oriR6K M65025 mobRP4 X54459 cat V00622

References 1. Burkardt, H. J. , G. Riess, and A. Puhler, Relationship ofgroup PI plasmids revealed by heteroduplex experiments. RP1, RP4, R68 and RK2 are identical. J Gen Microbiol, 1979.114 (2): p. 341-8.

2. Chatfield, S. N. , et al., Role of ompR-dependent genes in Salmonella <BR> <BR> <BR> <BR> typhimurium virulence : mutants deficient in both ompC and ompF are attenuated in vivo. Infect Immun, 1991.59 (1) : p. 449-52.

3. Chatfield, S. N. , et al., Use of the nirB promoter to direct the stable expression of heterologous antigens in Salmonella oral vaccine strains development of a single- dose oral tetanus vaccine. Biotechnology (N Y), 1992.10 (8): p. 888-92.

4. Chatfield, S. N. , et al., Evaluation of Salmonella typhimurium strains <BR> <BR> <BR> <BR> harbouring defined mutations in htrA and aroA in the murine salmonellosis model.

Microb Pathog, 1992.12 (2): p. 145-51.

5. Cravioto, A. 1980, PhD Thesis, University of London, London, United Kingdom.

6. Cravioto, A. , et al., Risk of diarrhea during the first year of life associated with initial and subsequent colonization by specific enteropathogens. Am J Epidemiol, 1990.131 (5): p. 886-904.

7. Curtiss, R. , 3rd and S. M. Kelly, Salmonella typhimurium deletion mutants lacking adenylate cyclase and cyclic AMP receptor protein are avirulent and immunogenic. Infect Immun, 1987.55 (12): p. 3035-43.

8. Donnenberg, M. S. and J. B. Kaper, Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive-selection suicide vector. Infect Immun, 1991.59 (12): p. 4310-7.

9.. Dougan, G. , et al., Construction and characterization of vaccine strains of Salmonella harboring mutations in two different aro genes. J Infect Dis, 1988.

158 (6): p. 1329-35.

10. Everest, P. , et al., Expression of LacZfrom the htrA, nirB and groE promoters in a Salmonella vaccine strain influence of growth in mammalian cells.

FEMS Microbiol Lett, 1995.126 (1) : p. 97-101.

11. Gerdes, K. , P. B. Rasmussen, and S. Molin, Unique type of plasmid maintenance function. postsegregational killing ofplasmid-free cells. Proc Natl Acad Sci U S A, 1986.83 (10) : p. 3116-20.

12. Gerdes, K. , et al., The hok killer gene family in gram-negative bacteria. New Biol, 1990.2 (11) : p. 946-56. r 13. Hohmann, E. L. , et al., phoP/phoQ-deleted Salmonella typhi (Ty800) is a safe and immunogenic single-dose typhoidfever vaccine in volunteers. J Infect Dis, 1996.

173 (6): p. 1408-14.

14. Hone, D. , et al., Construction of defined galE mutants of Salmonella for use as vaccines. J Infect Dis, 1987.156 (1) : p. 167-74.

15. Jones, P. W. , et al., Oral vaccination of calves against experimental salmonellosis using a double aro mutant of Salmonella typhimurium. Vaccine, 1991.

9 (1) : p. 29-34.

16. Laemmli, U. K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 1970.227 (259): p. 680-5.

17. Leong, J. , A. C. Vinal, and W. S. Dallas, Nucleotide sequence comparison between heat-labile toxin B-subunit cistrons from Escherichia coli of human and porcine origin. Infect Immun, 1985.48 (1) : p. 73-7.

18. Levine, M. M. , et al., Attenuated Salmonella as live oral vaccines against typhoidfever and as live vectors. J Biotechnol, 1996.44 (1-3) : p. 193-6.

19. Miller, S. I. , A. M. Kukral, and J. J. Mekalanos, A two-component regulatory system (phoP phoQ) controls Salmonella typhimurium virulence. Proc Natl Acad Sci U S A, 1989.86 (13) : p. 5054-8.

20. Milton, D. L. , et al., Flagellin A is essentialfor the virulence of Vibrio anguillarum. J Bacteriol, 1996.178 (5): p. 1310-9.

21. Picard, D. , et al., Characterization of defined ompR mutants of Salmonella typhi : ompR is involved in the regulation of Vi polysaccharide expression. Infect Immun, 1994.62 (9): p. 3984-93.

22. Qadri, F. , et al., Prevalence of toxin types and colonization factors in enterotoxigenic Escherichia coli isolated during a 2-year periodfrom diarrheal patients in Bangladesh. J Clin Microbiol, 2000.38 (1) : p. 27-31.

23. Roberts, R. C. , A. R. Strom, and D. R. Helinski, The parDE operon of the broad-host-range plasmid RK2 specifies growth inhibition associated with plasmid loss. J Mol Biol, 1994.237 (1) : p. 35-51.

24. Rudin, A. , L. Olbe, and A. M. Svennerholm, Monoclonal antibodies against <BR> <BR> <BR> <BR> fimbrial subunits of colonization factor antigen I (CFAlI) inhibit binding to human enterocytes andprotect against enterotoxigenic Escherichia coli expressing heterologotis colonization factors. Microb Pathog, 1996.21 (1) : p. 35-45.

25. Sambrook, J. , E. F. Fritsch, and T. Maniatis, Molecular cloning : a loboratory manual. 2nd ed. 1989: Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y.

26. Schlor, S. , et al., Genetic rearrangements of the regions adjacent to genes encoding heat-låbile enterotoxins (eltAB) of enterotoxigenic Escherichia coli strains.

Appl Environ Microbiol, 2000.66 (1) : p. 352-8.

27. Strugnell, R. , et al., Characterization of a Salmonella typhimurium aro vaccine strain expressing the P. 69 antigen of Bordetella pertussis. Infect Immun, 1992.60 (10): p. 3994-4002.

28. Svennerholm, A. M. , et al., Roles of different coli surface antigens of <BR> <BR> <BR> <BR> colonization fåctor antigen II in colonization by and protective immunogenicity of enterotoxigenic Escherichia coli in rabbits. Infect Immun, 1990.58 (2): p. 341-6.

29. Tacket, C. O. , et al., Enteral immunization and challenge of volunteers given enterotoxigenic E coli CFA/II encapsulated in biodegradable microspheres.

Vaccine, 1994.12 (14): p. 1270-4.

30. Tacket, C. O. and M. M. Levine, Vaccines against enterotoxigenic Escherichia coli infections, in New Generation Vaccines, M. M. Levine, et al. , Editors. 1997, Marcel Dekker, Inc. , New York, N. Y.

31. Tao, B. Y. and K. C. P. Lee, Mutagenesis by PCR, in PCR Technology current innovations, H. G. Griffin and A. M. Griffin, Editors. 1994, CRC Press, Inc.: Boca Raton, Florida. p. 69-83.

32. Turner, A. K. , et al., Construction and characterization of genetically defined aro omp mutants of enterotoxigenic Escherichia coli andpreliminary studies of safety and immunogenicity in humans. Infect Immun, 2001.69 (8): p. 4969-79.

33. Wolf, M. K., Occurrence, distribution, and associations of O and H serogroups, colonization factor antigens, and toxins of enterotoxigenic Escherichia coli. Clin Microbiol Rev, 1997.10 (4): p. 569-84.

34. Yamamoto, T. and P. Echeverria, Detection of the enteroaggregative Escherichia coli heat-stable enterotoxin 1 gene sequences in enterotoxigenic E. coli strains pathogenic for humans. Infect Immun, 1996.64 (4): p. 1441-5.

35. Dougan EP-B 0322237.

36. Dougan EP-B 0400958.

37. Dougan EP-B 0524205.

38. Chatfield WO 99/49026.

39. Charles WO 92/15689.

40. Chang, A. C. and S. N. Cohen, Construction and characterization of amplifiable multicopy DNA cloning vehicles derivedfrom the P15A cryptic miniplasmid. J Bacteriol, 1978.134 (3): p. 1141-56.

41. So, M. and B. J. McCarthy, Nucleotide sequence of the bacterial transposon Tnl 681 encoding a heat-stable (ST) toxin and its identification in enterotoxigenic Escherichia coli strains. Proc Natl Acad Sci U S A, 1980.77 (7): p. 4011-5.

42. Echeverria, P. , et al. , Plasmids coding for colonization factor antigens I and II, heat-labile enterotoxin, and heat-stable enterotoxin A2 in Escherichia coli. Infect Immun, 1986.51 (2): p. 626-30.

43. Levine, M. M. , et al., Colonization factor antigens I and II and type 1 somatic pili in enterotoxigenic Escherichia coli: relation to enterotoxin type. Infect Immun, 1983.39 (2): p. 889-97.

44. McConnell, M. M. , et al. , Plasmids coding for colonization factor antigen I and heat-stable enterotoxin production isolated from enterotoxigenic Escherichia coli : comparison of their properties. Infect Immun, 1981.32 (2): p. 927-36.

45. Murray, B. E. , et al., CFA/I-ST plasmids: comparison of enterotoxigenic Escherichia coli (ETEC) of serogroups 025,063, 078, and 0128 and mobilization from an R factor-containing epidemic ETEC isolate. J Bacteriol, 1983.153 (1) : p.

566-70.

46. Reis, M. H. , et al., Transfer of a CFA/I-ST plasmid promoted by a conjugative plasmid in a strain of Escherichia coli of serotype 0128ac : H12. Infect Immun, 1980.29 (1) : p. 140-3.

47. Simon, R. , U. Priefer, and A. Puhler, A broad host range mobilisation system for in vivo genetic engineering: transposon mutagenesis in Gram negative bacteria.

Bio/Technology, 1983. 1 : p. 784-791.

48. Turner et al (1998) Infection and Immunity 66 (5) 2099-2106.

49. Tacket et al, 1997, Infection & Immunity, 65 (2), 452 to 456.

50. Cieplak et al (1995) Journal of Biol. Chem. 270 (51), 30545-30550.

51. Chong et al (1998) Vaccine 16, 732-740.

52. Rappuoli et al (1999) Immunology Today 20 (11), 493-500.

53. Clements (1990) Infect. & Immun. 58 (5), 1159-1166.

54. Aitken and Hirst (1993) Vaccine 11 (2), 227-233.

55. Sanchez et al (1990) Res. Microbiol. 141,971-979.

56. Miller, V. L. and J. J. Mekalanos, Synthesis of cholera toxin is positively regulated at the transcriptional level by toxR. Proc Natl Acad Sci U S A, 1984. 81 (11): p. 3471-5.

57. Dunstan, S. , Simmons, C. and Strugnell, R. Use of in-vivo regulated promoters to to deliver antigens from attenuated Salmonella typhimurium.

Infection and Immunity (1999) 67, 5133-5141.

58. Valdivia, R. and Falkow, S. Fluorescence-based isolation of bacterial genes expressed within host cells. Science (1997), 277,2007-2011.

59. Summers et al, Mol. Genet. Genes. , 201 (2): 334-338.