Escherichia coli O157:H7 epithelial adhesin

Abstract

Polypeptides encoded by a continuous segment of chromosomal DNA from E. coli O157:H7, isolated on plasmid pSC(overlap)(ATCC No. 69648), that encodes an adhesin (SEQ ID NO:5) that mediates bacterial colonization of bovine intestines, vaccines derived therefrom, and antibodies directed against the adhesin are disclosed.

Description

FIELD OF THE INVENTION

The invention relates to genetic engineering and particularly to the demonstration that a contiguous segment of chromosomal DNA from E. coli O157:H7 encodes an adhesin that mediates colonization of the gastrointestinal tracts of bovines, and possibly humans, with E. coli O157:H7 and bacteria using structurally related adherence mechanisms.

BACKGROUND OF THE INVENTION

E. coli O157:H7 is a virulent and common foodborne pathogen. Most outbreaks, and many sporadic cases (38,42; see the appended Citations), have been attributed to food of bovine origin. Most E. coli O157:H7 infections are sporadic, but this organism can cause massive epidemics by contamination of ground beef (19) and water (69). E. coli O157:H7 is transmissible from person to person, but the disappearance of the strain which caused the massive 1993 outbreak in Washington State soon after recall of the incriminated vehicle demonstrates that ingestion of contaminated beef, and not person to person spread, is the chief source of human infection.

E. coli O157:H7 organism elaborates Shiga-like toxins (SLT) I and/or II. SLT I and II inhibit protein synthesis by disrupting a glycosidic bond at a specific adenine (A4324) in 28S rRNA of the 60S ribosomal subunit. SLT-producing E. coli (SLTEC) are ubiquitous in food (62) and animals (47). The vast majority are probably not human pathogens.

Current data suggest that E. coli O157:H7 is the most common and medically significant SLTEC. Only one outbreak of bloody diarrhea caused by SLTEC other than E. coli O157:H7 has ever been reported (11). Additionally, even when sought appropriately, non-O157:H7 SLTEC are rarely found in stools submitted for bacterial culture in North America compared to their frequency in the environment (8,52,59). Moreover, E. coli O157:H7 is the predominant precipitant of the hemolytic uremic syndrome (HUS), the most important complication of enteric infection with E. coli O157:H7. For example, E. coli O157:H7 was found in 96% of HUS patients if stool was obtained within the first six days of diarrhea (72). Even though non-O157:H7 SLTEC have caused some cases of HUS in several foreign series (10,11,12,35,40), these strains have never been reported to cause HUS in the United States. These data suggest that E. coli O157:H7 is the most important cattle-borne human pathogen threatening the food supply of this country today.

Cattle are the only reservoir of E. coli O157:H7 so far identified. Approximately 1 in 200 apparently healthy northwestern United States dairy and beef cattle carry E. coli O157:H7, and 8 to 16% of herds have at least one infected animal (25). Similar carriage rates have been detected nationwide (26). These are probably minimum carriage rates, because the technique used to culture E. coli O157:H7 is relatively insensitive.

A very low inoculum of E. coli O157:H7 can cause human disease. Person to person spread occurs rather easily in outbreaks and among sporadic cases (5,6,60). Microbiologic analysis of the contaminated hamburger from the 1993 Western United States outbreak demonstrated that only approximately 200 E. coli O157:H7 were present in each of the contaminated patties (46). It is probable that the inadequate cooking that was applied reduced this concentration by at least one log suggesting that very few E. coli O157:H7, perhaps in the range of 1-10 bacteria, can cause clinically apparent infection.

Data suggest that the incidence of diseases caused by E. coli O157:H7 has increased in the United States, independent of ascertainment bias by diagnosing physicians (44,70). Additionally, an increasing rate of antibiotic resistance in Washington State human isolates of E. coli O157:H7 might portend an increased prevalence of this pathogen in animals administered antibiotics. For example, before 1988, none of 56 strains of E. coli O157:H7 were resistant to a wide variety of antibiotics tested, whereas after 1988, 7.4% of 176 strains were resistant to the same combination of antimicrobials (streptomycin, sulfamethoxazole, and tetracycline). It is probable that the selective pressure for the acquisition of antibiotic resistance in E. coli O157:H7 occurred in farm animals. This emerging resistance is of considerable concern because such strains might achieve a selective advantage over other coliform bacilli in cattle given antibiotics, thereby increasing the frequency with which food of bovine origin is contaminated with this pathogen.

Because of the ease with which E. coli O157:H7 can cause human disease, it is crucial to reduce this pathogen in, or eliminate it from, its ecological niche, namely the gastrointestinal tracts of healthy cattle.

The molecular mechanisms used by E. coli O157:H7 to adhere to epithelial cells and colonize animals are poorly characterized. However, the adhesive properties of E. coli O157:H7 have been noted by several investigators. Most North American strains of E. coli O157:H7 displayed D-mannose-resistant adherence patterns to HEp-2 or Henle 407 cells (57). Most strains adhere in the form of localized microcolonies, a phenotype strongly linked to diarrhea in epidemiological studies of enteropathogenic E. coli (EPEC) (13,16). A 60 MDa plasmid is present in all stains of E. coli O157:H7, and one group associated the expression of sparse D-mannose-resistant adhesion to Henle 407 cells to the presence of this plasmid (34). Plasmid-cured E. coli O157:H7 expressed no fimbriae and were nonadherent, and a 60 MDa plasmid from E. coli O157:H7 conferred weak adherence to non-adherentE. coli C600. However, other investigators have shown that plasmid-less E. coli O157:H7 were fimbriated, whereas laboratory E. coli strains were not (79). Furthermore, plasmid-cured E. coli O157:H7 adheres to epithelial cells as well or better than its parent (22,33). Only one of five adherent strains of E. coli O157:H7 studied by Sherman et al. (66) was fimbriated, but this fimbriated strain also agglutinated erythrocytes. The agglutination was sensitive to D-mannose, suggesting that this adherence was due to type I fimbriae. Taken together, these data suggest that an identifiable fimbrial structure is not responsible for the adherence of most E. coli O157:H7 to Henle 407 cells.

Outer membranes of E. coli O157:H7 competitively inhibit adherence to HEp-2 cells, an inhibition which is not due to H7 flagellin or O157 lipopolysaccharide (65). Adherence of E. coli O157:H7 to HEp-2 cells was reduced, but not abolished, by antibody to a 94 kDa outer membrane protein (64). Antibodies to enterotoxigenic E. coli colonization factor antigens I and II do not detect surface structures on E. coli O157:H7 (78). E. coli O157:H7 do not have sequences homologous to the EPEC adherence factor plasmid or to the diffuse adherence adhesin (71).

Some investigators have suggested that the epithelial cell adhesin of E. coli O157:H7 is encoded by its eae gene (17). E. coli O157:H7 eae is related to inv, which encodes Yersinia invasin, which also functions as an adhesin, and EPEC eae, which encodes intimin. An eae deletion mutant of E. coli O157:H7 neither adhered to HEp-2 cells nor caused the attaching and effacing (AE) lesion in newborn pigs (17). When deletion mutants were complemented in trans by an intact eae gene, the strain could again cause the AE lesion, but still could not adhere in vitro. However, data from other groups suggest that the eae gene product is not an adhesin for E. coli O157:H7. First, despite sequence homology to inv in its bacterial localization and transmembrane domains, the receptor binding domain of E. coli O157:H7 eae is quite dissimilar (4,82). Second, an eae insertional mutant in E. coli O157:H7 retained the ability to adhere to HEp-2 cells in a quantitative adherence assay (41). Third, an eae gene product does not confer adherence on nonadherent laboratory strains of E. coli. (Jerse, A., et al., Proc. Natl Acad. Sci. USA 87:7839-7843, 1990) Thus, a molecule other than the eae gene. product in E. coli O157:H7 appears to be the primary adhesin of E. coli O157:H7 for bovine epithelial cells, enabling this human pathogen to colonize the bovine gastrointestinal tract.

Bacterial adhesins, when used as immunogens, prevent disease or colonization of mucosal surfaces by bacteria in many animals (1,18,21,29,30,36,49,50,55,61, 68,81, which are hereby incorporated by reference). The reduction of E. coli O157:H7 at its bovine source would enhance the microbiologic safety of food derived from cattle, and lessen the environmental biohazard risk posed by the approximately 100,000 cattle detectably infected with E. coli O157:H7 at any one time in the United States. The availability of antibody for passive immunization would greatly mitigate the harm engendered by outbreaks of this infection.

SUMMARY OF THE INVENTION

Transposon-mediated mutations of E. coli O157:H7 have been isolated that do not adhere to HeLa cells and that have lost the ability to colonize bovine intestines. A HeLa cell in vitro system has been established that provides a means of assaying variants of E. coli O157:H7 for their ability to colonize cattle. The gene into which the transposon inserted have been sequenced. In a seperate approach, two overlapping 40 kb segments of chromosomal DNA from E coli O157:H7 have been cloned that confer D-mannose resistant adherence to nonadherent strains of E. coli. The overlapping region has been cloned, and E. coli HB101 expressing this overlapping region on a plasmid (ATCC No. 69648) have acquired the ability to adhere to epithelial cells of both human (HeLa) and bovine (MDBK) origin. These findings demonstrate that a contiguous segment of chromosomal DNA from E. coli O157:H7 encodes an adhesin, and that this same adhesin mediates both bacterial adherence to HeLa cells.

The adhesin-encoding region of pear has been identified as the nucleotide sequence of SEQ ID NO:4. Also described are recombinant expression vectors containing the adhesin-coding region, as are bacterial cells which are transformed with the recombinant expression vector. The recombinant adhesin preferably has the amino acid sequence of SEQ ID NO:4. Also described are immunological binding partners that bind to the recombinant adhesin. Vaccine formulations of the invention contain the recombinant adhesin encoded by a nucleic acid molecule that hybridizes under stringent conditions to the nucleotide sequence of SEQ ID NO:4.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows shedding rates of E. coli O157:H7 from calves inoculated with adherent strain 86-24 and non-adherent strain F4. Bars=SE.

FIG. 2 is a map demonstrating overlap region between cloned regions of pSC(A-G6) and pSC(T-H12), and an adherence-conferring subclone region of the overlap, pear.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Escherichia coli O157:H7 causes severe and potentially fatal infections in humans, especially children. This pathogen is harbored by apparently healthy cattle. The molecular mechanisms underlying the carriage of E. coli O157:H7 by cattle are not understood. The invention provides an isolated adhesin that enables E. coli O157:H7 to adhere to epithelial cells in vitro and to colonize cattle.

Transposon-based mutants of E. coli O157:H7 have been identified that have lost the ability to adhere to, epithelial cells. These same mutants colonize calves poorly compared to parental E. coli O157:H7. This finding validates the use of the HeLa cell adherence assay. Additionally, a region of the E. coli O157:H7 chromosome has been cloned that expresses the adhesin. This chromosomal region has been designated the epithelial adherence region (ear).

Two complementary approaches were used to identify the gene encoding the adhesin that enables E. coli O157:H7 to colonize bovine intestines. The first approach identified a transposon mutant of E. coli O157:H7 that has lost the ability to adhere to HeLa cells and that also colonizes cattle poorly. For the second approach, a segment (pear) of chromosomal DNA from E. coli O157:H7 was cloned which, when expressed in derivatives of nonadherent E. coli K12, provides these strains with the ability to adhere to HeLa cells.

Our sequence analysis demonstrates that the candidate adhesin of E. coli O157:H7 is a homolog of the IrgA protein in Vibrio cholerae (Goldberg, M. B., et al., Proc. Natl. Acad. Sci. USA 88:1125-1129, 1991). This molecule is related to the TonB dependent family of outer membrane proteins in E. coli (Goldberg, M. B., et al., Molecular Microbiology 6(16):2407-2418, 1992). The DNA sequence analysis of the irgA homolog which is encoded by pear, and the homology to genes encoding IrgA and CIR, are described in the following Examples.

EXAMPLE 1

Identification of Mutants That Have Lost the Ability to Adhere to HeLa Cells

A prototype adherent strain, E. coli O157:H7 substrain 86-24, was isolated in 1986 from a patient with hemorragic colitis whose illness was traced to ground beef the patient had consumed in a fast food restaurant (23). The transposon TnphoA was used to mutagenize substrain 86-24 according to the procedures in the appended citations 73 and 74, which are hereby incorporated by reference. TnphoA, a Tn5 derivative, carries the alkaline phosphatase gene (phoA) and inserts in random locations throughout the E. coli genome but only in 1-3 sites per target cell. Hence, an in-frame integration of phoA of TnphoA into regions of genes encoding an extracellular domain of a protein will result in the expression of PhoA.

PhoA-expressing TnphoA mutants of E. coli O157:H7 were tested in the HeLa cell assay to find nonadherent mutants. This assay is described in detail in citation 8, which is hereby incorporated by reference. Bias was avoided by coding each organism, and the examining microscopist was blinded to which samples were controls and which were mutants. EPEC strain B171 was used as the positive control for localized adherence in all experiments (58). Negative controls were nonadherent E. coli HB101and/or NM554 (56). Besides the prototype adherent substrain of E. coli O157:H7, 6 other substrains tested adhered in a localized pattern to HeLa cells in the presence of D-mannose.

As previously reported by others (75), day to day variability in the degree of adherence of E. coli O157:H7 to HeLa cells was typically observed. However, three of 177 PhoA-expressing transconjugants screened for adherence to HeLa cells proved to be consistently nonadherent when tested in the coded assay (strains A5, F4, and N11). Strains A5, F4, and N11 retained all other phenotypic and genotypic characteristics of the parent strain of E. coli O157:H7.

Southern blot analysis determined the locations of the transposon insertions in adherent and non-adherent TnphoA mutants. DNA from strains A5, F4, and N11 and from adherent mutants H8, P11, and P12 were digested with MluI, which does not cleave DNA within TnphoA. Resulting fragments were separated in an agarose gel, transferred to Nytran, and probed with a fragment from the Tn5 central region of TnphoA. Interestingly, the results indicated that there were two TnphoA insertions in the nonadherent mutants A5 and N11, and three in F4, in apparently identical MluI bands of 23 and 16 kb length. Single integrations of TnphoA are demonstrated in each of the three adherent transconjugants. TnphoA integrated in the chromosome of stains A5, F4, and N11, and not in plasmid DNA

EXAMPLE 2

Animal Testing

The nonadherent strain F4 and wild type E. coli O157:H7 were tested for their ability to colonize conventional Holstein calves (<1 week old). After an initial feed of colostrum, calves were placed in individual holding pens in an isolation facility, and reared on whole cow's milk with free-choice access to water, alfalfa hay, and a high protein grain mixture. It was demonstrated at the outset that the calves were not excreting E. coli O157:H7 by culturing their feces on sorbitol-MacConkey agar (SMA). Four animals received either 10⁸ adherent E. coli O157:H7 86-24 NalR or 10⁹ nonadherent mutant strain F4. In dual challenge experiments, each of four calves simultaneously received 10⁸ adherent E. coli O157:H7 86-24 and 10⁹ nonadherent TnphoA mutant F4. E. coli O157:H7 was a spontaneously nalidixic acid resistant mutant selected on agar plates containing nalidixic acid. TnphoA encodes kanamycin resistance.

The respective antibiotic resistances of these strains were exploited to identify E. coli O157:H7 in fecal samples by screening for shed challenge organisms on sorbitol MacConkey agar (SMA) containing nalidixic acid with or without kanamycin. Antibiotic resistant, sorbitol nonfermenting colonies were confirmed to be E. coli O157:H7 by their reactivity in the O157 latex particle agglutination test (Oxoid E. coli 0157 Test; Unipath Limited, Hampshire, England). The nonadherent strain was detectable for fewer days and at lower concentrations as shown in FIG. 1, which summa the results of all challenges. The animals showed no ill effects which could be attributed to the E. coli O157:H7.

The shedding index (cfu/g of stool x number of days shed) was significantly greater for the adherent than for the non-adherent strain when analyzed by non-parametric rank sum analysis (p=0.028). Strain F4 grows as well as strain 86-24 NalR in fresh bovine stool and rumen contents, and in liquid broth, incubated aerobically overnight. These data suggest that the abbreviated excretion of the TnphoA mutant by the challenged calves is not related to decreased viability of the mutant compared to the parent strain, even though it is difficult to simulate in vitro the exact conditions of the calf gastrointestinal tract. By demonstrating that calves retain the adherent stain more effectively than the nonadherent strain, these results validate the use of the HeLa cell in vitro adhesion assay for use in development of other reagents relevant to vaccine preparation.

EXAMPLE 3

Expression of a Recombinant Adhesin Using Chromosomal DNA from E. coli O157:H7

A segment of DNA has been derived from the chromosome of E. coli O157:H7 strain 86-24 NalR that renders nonadherent E. coli NM554 adherent. To clone this segment, approximately 2000 PhoA expressing and nonexpressing TnphoA mutants of E. coli O157:H7 86-24 NalR were screened, and one transconjugant (20D2B) was found that no longer reacted in the O157 latex particle agglutination test. This sorbitol negative mutant produced SLT II, was H7 antigen positive and β-glucuronidase negative, and possessed the same API score as the parental strain. (API score refers to a product produced by Analytab, Plainview, N.Y., which determines multiple bacterial growth characteristics. A score is given for each characteristic; taken in total, the score speciates bacteria. Within a species, there may be multiple scores.) However, 20D2B was highly adherent to HeLa cells. A partial Sau3a digest of genomic DNA of the hyperadherent strain 20D2B was ligated into plasmid Supercos (pSC) (20,77), packaged, and used to transduce nonadherent laboratory strain NM554. This experiment yielded 2200 transductants with an average of 40 kb of DNA inserted into the BamHI site of pSC.

The 2200 cosmid clones were screened for adherence to HeLa cells, and two adherent clones were identified and designated pSC(A-G6) and pSC(T-H12). E. coli NM554 containing pSC(A-G6) and pSC(T-H12)) adhered to HeLa cells in a diffuse rather than localized pattern although nascent clusters were sometimes seen. Southern blotting demonstrated that: (a) the A-G6 and T-H12 determinants overlap by approximately 15 kb; (b) these inserts are derived from E coli O157:H7 chromosomal DNA; (c) the inserts do not encode eae, bfp (which encodes the bundle forming pilus adhesin of EPEC), or SLT II, and (d) the overlap region is conserved in each of 9 E. coli O157:H7 tested, but not in E. coli HB101,DH5α, or EPEC B171.

As shown in FIG. 2, a deletion mutant of pSC(A-G6), designated “pSC(overlap)” (ATCC No. 69648), retains the overlapping segment between pSC(A-G6) and pSC(T-H12). Interestingly, nonadherent E. coli HB101 transformed with pSC(overlap) display diffuse adherence to Madin-Darby bovine kidney cells (MDBK), and, in a preliminary experiment, localized adherence to HeLa cells. An 8 kb subclone of pSC(overlap), designated “pear”, restores adherence to non-adherent strain A5. (pear, and the irgA homologous subclone described below, display diffuse adherence to HeLa cells.)

The data summarized above suggest that: (1) an identifiable adhesin from E. coli O157:H7 expressed in E. coli HB101 (pear) enablesE. coli O157:H7 to adhere to epithelial cells of human (HeLa) and bovine (MDBK) origin in vitro; and (2) this adhesin is the same molecule which permits E. coli O157:H7 to remain in the gastrointestinal tracts of bovines. Further identification and characterization of the subject recombinant E. coli O157:H7 adhesin is described below.

EXAMPLE 4

Identification of the Genes on the Adherence Conferring Plasmid (Pear)

pSC(overlap) itself consists of 4 kb of pSC DNA and approximately 15 kb of E. coli O157:H7 DNA pear consists of 8 kb of chromosomal DNA plus the SK+ vector (Stratagene). To identify the adhesin expressed by pear, the entire fragment was sequenced and open reading frames were determined. The results are described below.

The appended SEQ ID NO:1 shows the 8,041 base pair nucleotide sequence of pear. Almost all of the sequence has been confirmed. Ambiguous DNA (in regions not encoding the candidate adhesin) is noted by N in the appended sequence. The pear insert contained three open reading frames (ORFs) of sufficient length to encode potential virulence or adherence factors. Two of these are homologous to genes necessary for resistance to tellurite (Jobling, M G, et al., Gene 66:245-258, 1988). These terE and terD homologs are shown in SEQ ID NO:2 and SEQ ID NO:3, corresponding respectively to nucleotides 7024-6449 and 7670-7092 of SEQ ID NO:1. The other ORF is homologous to a gene encoding a homolog of IrgA (Goldberg, M B, et al., Molecular Microbiology 6:2407-2418, 1992). This irgA homolog is shown in SEQ ID NO:4, which corresponds to nucleotides 3036-5126 of SEQ ID NO:1. IrgA is an outer membrane protein of V. cholerae, and is believed to be important for colonization of mice in an experimental system (Goldberg, M B, et al, Infection and Immunity, 58:55-60, 1990). The E. coli O157:H7 adhesin (SEQ ID NO:4) is also homologous to the E. coli colicin I receptor (CIR) (Griggs, D. W., et al., J. Bacteriol. 168:5343-5352, 1987). The amino acid homologies of the candidate adhesin to IrgA and to CIR are demonstrated by comparing SEQ ID NO:5 and SEQ ID NO:6, and SEQ ID NO:7 and SEQ ID NO:8, respectively.

EXAMPLE 5

Mutations in the irgA homolog, as cloned into an expression vector, lead to loss of adherence. Transposon (TnphoA) insertions in the irgA homolog of E. coli O157:H7 ablate adherence of laboratory strains of E. coli transformed with a plasmid vector into which an adherence conferring region has been inserted. We cannot state with certainty the exact site of the two TnphoA insertions which ablated adherence, but the regions of the insertions are between nucleotides 3271-3310 and 3801-3840 of SEQ ID NO:1.

EXAMPLE 6

A product of a single gene (i.e., the irgA homolog) confers adherence to nonadherent E. coli. We first performed PCR using as primers the sequences 5′GGGGATCCAATTCTGGCATGCCGAGGCAGTCG3′ (SEQ ID NO:9), corresponding to nucleotides 2895-2914 of SEQ ID NO:1) and 3′GGACCGCCTTGTCACCGTTGCTCTTAGATCTGG5′ (SEQ ID NO:10, corresponding to nucleotides 5176-5196 of SEQ ID NO:1) from which DNA on pear was amplified. These sequences were cloned into the BamHI and XbaI sites of pSK+. We also amplified the same gene using as template DNA from E. coil O157:H7. In this latter case,. the primers used were 5′GGAAGGATCCCCGACACGCCATACGGATAGCTG3′ (SEQ ID NO:11, corresponding to nucleotides 2867-2890 of SEQ ID NO:1) and 3′GCAACGGTGACGTTGAGGACCGCCAGATCTAAAGG5′ (SEQ ID NO:12, corresponding to nucleotides 5159-5183 of SEQ ID NO:1). This latter PCR product was also cloned into pSK+, using the same BamHI and XbaI sites. In both cases, multiple laboratory strains of nonadherent E. coli were rendered adherent to HeLa cells by these cloned single genes.

EXAMPLE 7

The Adherence of Δ-Ear Mutants to HeLa Cells is Diminished

Strain F12 of E. coli O157:H7 is a hyperadherent mutant that has been mutated by TnphoA such that the O157 antigen is no longer expressed (Bilge, S. S., et al., Abstract B-7, American Society of Microbiology, 21-25 May 1995). F12 is probably hyperadherent because the lack of expression of the O157 antigen enables the adhesin to be more completely exposed on the bacterial cell surface.

We deleted the entire 8041 base pair KpnI-KpnI region (SEQ ID NO:1) of pear from strain F12 as follows. We cloned pear into a suicide vector, pCVD442 (Donnenberg, M S, et al., Infection and Immunity, 59:4310-7, 1991), which was then mated into strain F12 using E. coli SM10 lambda pir as a donor. Sucrose resistant, ampicillin sensitive strains were analyzed to find mutants with the SEQ ID NO:1 region deleted.

When the ear is deleted from strain F12, the organism was observed to be nonadherent or severely adherence deficient (1 cluster of microcolonies per 1-3 high powered fields) by an observer blinded to the identity of the F12 and the ear deletion mutant. (In comparison, the parent strain F12 displayed much higher levels of adherence to HeLa cells, approximately 0.5-1 cluster per cell.) This striking adherence deficiency could be complemented by the cloned genes of irgA from either the plasmid or the chromosome. Hence, this loss of adherence from the deletion of the ear is not caused by a polar effect of the deletion.

In summary, our data demonstrates that the PCR product of a single allele, an irgA homolog in E. coli O157H7, confers an adherent phenotype when cloned into an appropriate vector and transformed into laboratory strains of E coli. Tested strains include: E. coli NM554 (Raleigh, E A, et al, Nucleic Acids Research, 16:1563-75, 1988); E. coli HB101; and E. coli ORN172 (Woodall, L D, et al., Journal of Bacteriology 175: 2770-8, 1993), which is an E. coli K12 strain from which genes encoding type I pili have been deleted. Our deletion mutation data confirm that the epithelial adherence region (ear) encodes an E. coli O157:H7 adhesin. Sequence data suggest that this adhesin is a homolog of IrgA of V. cholerae.

We have also performed TnphoA mutagenesis of E. coli O157:H7, and identified three nonadherent mutants (strains A5, F4, and N11), each of which sustained a TnphoA insertion in the same allele (SEQ ID NO:9). One of these strains, stain F4, was deficient in its ability to colonize in calves in an oral challenge experiment performed at the Washington State University in Pullman, Wash. Sequence analysis suggests that the TnphoA insertion in the same allele among the three nonadherent mutants may have taken place in the midst of a duster of genes, at least one of which has homology to pro-secretory proteins in Yersinia enterocolitica (YscJ) (Michiels, T., et al., Journal of Bacteriology 173:4994-5009, 1991), Rhizobium fredii (Nolt) (Meinhardt, L. W., et al., Molecular Microbiology 6:2407-2418, 1992), and Xanthomonas compestris (HrpB) (Fenselan, S., et al., Molecular Plant-Microbe Interactions 5:390-396, 1992), and it is possible that the secretion of the E. coli O157:H7 adhesin is controlled by this secretory mechanism.

EXAMPLE 8

Construct Recombinants and Deletion Mutants for Bovine Challenge Experiments

Our data suggest that adherence to HeLa cells by E. coli O157:H7 correlates with optimal colonization of calves with this organism. However, the E. coli O157:H7 adhesin is not closely linked to a locus mutagenized by TnphoA in nonadherent strain F4. To exclude the possibility that separate in vitro and in vivo, adherence mechanisms exist for E. coli O157:H7, we may (1) colonize calves with a laboratory E. coli expressing the recombinant O157:H7 adhesin; (2) immunize animals with a recombinant adhesin and ascertain if these animals are protected from challenge with E. coli O157:H7; or (3) create an isogeneic deletion mutant of E. coli O157:H7 and determine if this strain has lost its ability to adhere to HeLa cells and to colonize calves.

EXAMPLE 9

Study Adhesin Deletion Mutants of E. coli O157:H7 in Calves

We shall determine if adhesin-deletion mutants of E. coli O157:H7 lose the ability to colonize challenged animals in the calf model. We have demonstrated that wild type E. coli O157:H7 colonizes the gastrointestinal tracts of calves longer and at higher concentrations than does a nonadherent TnphoA mutant. Ideally, we would challenge calves with a laboratory strain of E. coli expressing the recombinant adhesin of ear, and follow excretion of this organism However, it is very unlikely that a recombinant host stain will survive in the gastrointestinal tract of animals. For example, plasmids of enterotoxigenic E. coli encoding K88 (F4) or K99 (F5) pilus adhesins did not confer the ability to colonize pig small intestine when transferred to laboratory E. coli strain K12, despite the clearly demonstrated role of these adhesins to colonizations of the small intestines in the wild type strains (48). Therefore, we use specific deletion mutants to authenticate the role of the in vitro adhesin in the calf model.

To validate the role of the ear adhesin in bovine colonization, we created an isogeneic, in-frame deletion mutant of almost all of the irgA homolog in E. coli O157:H7. To do this, we cloned regions (ca. 0.45 kb) flanking the gene encoding the IrgA homolog adhesin, up to and including nucleotide 3110 at the 5′ region, and nucleotide 5051 at the 3′ region of the gene encoding the IrgA homolog adhesin, such that the 1941 nucleotides between nucleotides 3110 and 5051, inclusive, are deleted from this construct, corresponding to 647 amino acids. These truncated sequences of the IrgA homolog adhesin were then inserted into pCVD442, and deletion mutants of strain 86-24 NalR were identified by virtue of their ampicillin sensitivity and sucrose resistance. The deletion was confirmed in two mutants by Southern blotting using as probes the internal (i.e., deleted) portion of irgA homologs and DNA flanking the deleted sequences. The deletion mutant is termed 86-24 (Δ irgA). The parent strain (86-24 NalR) and a ten-fold excess of deletion mutant 86-24 (Δ irgA) will be fed simultaneously to cattle, and fecal excretion will be monitored by culturing stool on SMA plates containing nalidixic acid. A representative number of strains will be collected at multiple time points after challenge, until clearance of the challenge strains, or until 60 days post challenge, whichever comes first. These strains will be individually tested for the presence of the irgA deleted sequences using the polymerase chain reaction, to determine which of the strains colonizes the cattle.

In a representative experiment, animals are challenged with dual inocula (adherent and excess nonadherent deletion mutant of E. coli O157:H7). Based on previous challenges, we can anticipate that at least 6 calves will excrete E. coli O157:H7 for at least 21 days after inoculation.

Study the humoral and secretory immune response of the calves challenged with E. coli O157:H7: The immune response of bovines to enteric colonization with E. coli O157:H7 is unknown. The calf challenge experiments described above provide a model to study this response. In particular, the humoral and secretory immune response to various E. coli O157:H7 components, including but not limited to the recombinant adhesin, can be quantified to determine if the development of a response is correlated to the clearance of this organism from the gastrointestinal tract.

To study the immune response of the challenged bovines to E. coli O157:H7, serum and saliva are obtained from each calf prior to challenge and at each sampling time, and stored at −70° C. Intestinal secretions are saved at the end of the experiment, when animals are necropsied.

The humoral and secretory immune responses of the challenged bovines to E. coli O157:H7 proteins are studied by protein immunoblot. Outer membranes, as well as whole cell preparations, of E. coli O157:H7 86-24, E. coli NM554 pSC(A-G6), E. coli NM554 pSC(T-H12), and E. coli HB101 (pear or a derivative) are electrophoresed in adjacent lanes in 10% SDS-PAGE gels and transferred electrophoretically to nitrocellulose membranes (32). Outer membranes from nonadherent E. coli are included as controls. The membranes are probed with calf serum and intestinal secretions, and adherent primary antibodies are detected by murine monoclonal antibodies to bovine IgG1, IgG2, IgM, and IgA, and goat antimouse IgG coupled to horseradish peroxidase. In control blots, the monoclonal antibodies are replaced with isotype-matched monoclonal antibodies to irrelevant antigens.

Protein immunoblot studies are extended to the recombinant adhesin produced above using pGEX (67) as well as to other strains of E. coli O157:H7. Antibodies from calves challenged in this section are also used in an attempt to inhibit in vitro adherence of E. coli O157:H7 strain 86-24 and the recombinant adherent strains to epithelial cells. Finally, the recombinant adhesin is an appropriate target molecule for an enzyme-linked immunosorbent assay to quantify the magnitude of the antibody response to the adhesin in cattle challenged withE. coli O157:H7.

EXAMPLE 10

Immunological Characterization of a Recombinant E. coli O157:H7 Adhesin

Standard immunochemical techniques are used to determine if the cloned adhesin (SEQ ID NO:4) is the same as the adhesin used by E. coli O157:H7 to adhere to HeLa cells. To achieve this objective, outer membrane proteins are prepared from laboratory strain(s) expressing the recombinant adhesin (SEQ ID NO:4). These proteins are analyzed on SDS-polyacrylamide gels, and used to immunize rabbits at three one-month intervals and Holstein cows at 30 and 60 days prepartum. Rabbits are also be immunized with killed whole bacterial cell preparations. A recombinant E. coli NM554, which does not adhere to HeLa cells, is used as a negative control immunogen.

Decomplemented serum from immunized rabbits and cows, and whey from the milk of immunized cows, is first absorbed against E. coli NM554 to deplete antibodies to outer membrane proteins which are irrelevant to adherence. The absorbed sera and milk are tested against E. coli O157:H7 strain 86-24, as well as additional strains, using immunofluorescence and Western blotting to determine if antibody to the recombinant antigen has specific affinity for extracellular antigens or outer membrane proteins, and to determine if this antigen is conserved. Negative control antigens consist of other E. coli, including diarrheagenic strains, as well as other Shiga like toxin-producing E. coli which do not belong to serotype O157:H7. Additional negative antibody controls include preimmune serum, and serum and milk obtained from animals immunized with nonadherent recombinant E. coli NM554.

The function of anti-adhesin antibodies is assessed using the in vitro HeLa cell assay, adapted to quantify the numbers of bacteria adherent to the target cells. E coli O157:H7, as well as adherent and nonadherent recombinant E. coli NM554, are incubated with immune or control sera or milk antibodies before addition to the HeLa cell culture. Antibodies remain in the adherence assay medium. Additionally used in these assays are antibodies in serum and saliva from animals challenged with oral E. coli O157:H7. After the appropriate incubation period, the number of bacteria adherent per cell is enumerated in multiple fields consisting of several hundred eukaryotic cells. The microscopist is blinded to the identity of the strains and antibodies.

The anti-adhesin antibodies raised and selected as described above are also useful for the diagnostic identification of E. coli O157:H7. So too are E. coli O157:H7 nucleotide sequences within or flanking the irgA homolog having the requisite specificity and sensitivity for diagnosing the presence of strain O157:H7 in feed animals, food, and humans, as determined by screening a panel of closely related bacterial strains for specificity, and a panel of E. coli O157:H7 for sensitivity.

EXAMPLE 11

Immunoprophylactic Vaccines

Bacterial adhesins have been used as immunogens to prevent colonization of mucosal surfaces and/or disease in multiple food and laboratory animals. An immunoprophylactic approach to the problem of E. coli O157:H7 carriage using a purified recombinant adhesin (ear or subclone thereof) as a vaccine is considered to be an efficient method to improve the microbiologic safety of food of bovine origin. Preparation and administration protocols for such vaccines are known in the art. See, e.g., U.S. Pat. No. 5,286,484; No. 5,208,024; No. 5,137,721; No. 5,079,165; No. 5,066,596; No. 4,795,803; No. 4,736,017; No. 4,702,911; No. 4,725,435; No. 4,454,116; and No. 3,975,517, which are incorporated by reference herein.

It is well known in the art that vaccines administered to cattle can confer immunity to bacteria that colonize the gut. For example, U.S. Pat. No. 3,975,517, describes operable methods for vaccinating cattle. Other published methods suitable for immunizing animals include: Acres et al., 25 Infection and Immunity 121, 1979; Linggood et al., U.S. Pat. No. 4,971,794, 1990; Sadowski and Prairie, U.S. Pat. No. 4,652,448, 1987; Suchaux et al., 69 Infection and Immunity 2828, 1992; Francis and Willgohs, 52 Am. J. Vet. Res 1051, 1991; Ikemori et al., 53 Am. J. Vet. Res 2005, 1992; Isaacson et al., 29 Infection and Immunity 824, 1980; Morgan et al., 22 Infection and Immunity 771, 1978; Morris et al., 13 J. Med. Microbiol. 265, 1980; Runnels et al., 55 Infection and Immunity 555, 1987; Sojka et al., 11 J. Med. Microbiol. 493, 1978; and Yokoyama et al., 60 Infection and Immunity 998, 1992. Vaccines can be prepared and administered by a variety of routes, many of which are found in Harlow and Lane, Antibodies, 1988, which is hereby incorporated by reference.

Immunization with the ear-encoded IrgA homolog adhesin to result in decreased excretion of E. coli O157H7 requires generation of immune responses with effectors in the gastrointestinal tract of adult and sub-adult cattle, the age groups which are presented for slaughter. In this age group, the presence of a functional rumen presents a major barrier to oral vaccination in a form designed to survive passage through the abomasum (the ‘true’ stomach), as the very large and metabolically active rumenal compartment is thought to block passage of most intact antigens to the abomasum.

Several strategies can result in immune responses in the intestine in adult ruminants. These include: the intramammary route mentioned below; parenteral (intramuscular) immunization, because the resulting antibodies are cleared from circulation into the intestine; intranasal immunization (as antigen presentation by the pharyngeal tonsils is quite effective); and the use of highly stimulating adjuvants such as ISCOMs (immune stimulating complexes such as liposome-antigen combinations) or cholera toxin b-subunit conjugates.

Types Of Vaccines: Purified antigen (prepared using standard recombinant DNA methods) or whole-cell vaccines can be used to stimulate an immune response in pregnant cows, resulting in the presence of protective antibodies in the colustrum and milk produced after parturition. The milk or colustrum can be stored for late administration to newborn calves. Alternatively, protein produced from the irgA homolog can be used to raise monoclonal antibodies which then can be used directly to confer passive immunity to newborn calves. Methods for preparing monoclonal antibodies are well-known in the art, and can be found, for example, in Harlow and Lane, Antibodies, Cold Spring Harbor Laboratory, 1988.

Vaccines Containing Purified Antigen: Purified IrgA homolog adhesin is formulated in sesame or peanut oil, sterile aqueous solution, or other isotonic solution suitable for injection. Stabilizers such as sorbitol or gelatin may be added. The immunogenicity can be enhanced by including various immunoadjuvants in the pharmaceutical preparation. Immunoadjuvants useful for this purpose include the water-in-oil emulsion of Freund, light mineral oil Bayol), the commercially available emulsifiers. Arlacel A and Aracel C, or monophosphoryl lipid A.

Cows are inoculated subcutaneously above the left shoulder twice, approximately 3 and 6 weeks before the onset of calving.

Whole Cell Vaccines: A bacterial stain suitable for use as a whole-cell vaccine is a nonpathogenic (to animals and humans) bacterium expressing the homolog adhesin (SEQ ID NO:4). This strain can be administered orally as a live vaccine, or can be used as a killed vaccine. Oral vaccination is an efficient means for stimulating production of secretory IgA in mucosal tissues such as the internal lining of the intestinal tract. Using a killed vaccine is the safest choice, as it obviates the risk that the non-pathogenic vaccine strain will acquire virulence genes by genetic transfer from other bacteria that may be present in the gut.

Where the goal is to acquire materials suitable for providing passive immunization, another advantageous route of administration is to inject the vaccine preparation direly into the teat canal of pregnant or nursing cows. Whether administration is oral or via injection, the vaccinated animal will produce protective antibodies in her milk that will passively immunize recipient calves.

Preparation Of Formalin-Treated Vaccine: One means of vaccine preparation incorporates some of the advantages of both live and killed whole-cell vaccines. For this procedure, a bacterial in expressing the antigenic protein is subjected to a controlled formalin treatment. A broth culture of the vaccine strain is grown to a concentration of about 10⁸ to 10⁹, then incubated under aerobic conditions for 10 to 15 hours in the presence of about 0.04% (vol/vol) formalin (0.016% wt/vol formaldehyde). Alternatively, the vaccine strain is grown on solid media, and the cells scraped off and suspended in broth medium for the formalin treatment. Formalin-treated bacteria remain viable, but optimally the proportion of colony-forming units is reduced by about 1000-fold compared with bacteria not exposed to formalin. For each batch of vaccine, plate counts are performed to ensure that the requisite proportion of bacteria have survived the formalin treatment. If necessary, the time of exposure to formalin and the percentage of formalin added to the broth are adjusted so that the plate counts of post-treatment cultures are 1000-fold reduced compared with controls. The vaccine contains the entire broth culture constituents including metabolic waste products and extracellular proteins.

Virulence of vaccine strains are tested by oral inoculation of 3 to 4 day old suckling mice with either virulent E. coli O157:H7 or the formalin-treated vaccine stain. Each mouse is administered about 10⁶ to 10⁸ organisms in about 0.15 ml. Mice are orally inoculated with serial dilutions of the bacteria, and survival is determined at 40 hours.

Administration Of Vaccine: For oral administration, the formalin-treated vaccine strain is packaged in enteric coated capsules that dissolve only after passing through the stomach. To achieve prophylaxis with monoclonal antibodies, the antibody can be mixed with colostrum, or any pharmaceutical carrier suitable for oral administration.

Injection: Four to eight ml doses of the vaccine are administered intramuscularly or directly into the lactiferous ducts via the teat canals of two year old cows. For lactiferous duct administration, the cows are given a series of 4 to 6 ml doses at 42, 32, 22 and 12 days prior to parturition. For intramuscular vaccination, 4 ml of the formalinized culture is injected in the side of the neck 28 days before parturition, followed by 6 ml 14 days later, and 8 ml at 5 days before the expected calving day. About 1 liter of milk is collected from vaccinated and control cows seven days after parturition. Whey is prepared by centrifuging milk for 44,000 g for two hours, then collecting the supernatant. Whey is stored either in 2 ml aliquots at 4° C., stored frozen at −30° C., or stored lyophilized.

To monitor antibody production in response to the vaccine, 1 l of milk is taken from vaccinated pregnant cows on 2,3,4,7 and 14 days after calving. Whey is prepared from the milk by centrifugation and stored as described above.

Challenging Passively-Immunized Calves: To demonstrate the protective effects of whey from vaccinated teats, newborn calves separated from their mothers are fed one aliquot of post-immune whey every 8 hours for 72 hours. Control calves are fed whey from non-vaccinated cows. Six hours after being fed the first dose of whey, the calves are infected orally with broth culture containing about 10⁹-10¹⁰ colony forming units virulent E. coli O157:H7 per ml. Each day after being exposed to the O157:H7, stool samples from infected calves are cultured to detect the presence or absence of E. coli O157:H7. Three calves from each group are sacrificed after 10 days, and their intestines examined for the presence or absence of E. coli O157:H7.

Mucosal Immunization: An additional route of vaccination exploits the mucosal immune system, which provides a defense against colonization of surfaces lining the gastrointestinal and respiratory systems, and relies on the induction of IgA antibodies secreted from these organs into the lumens where organisms colonize, and viable bacterial vectors which are in themselves not harmful to animas or humans, but which could be genetically engineered to express the E. coli O157:H7 adhesin. For a review of mucosal immunization, see: McGhee, J. R., et al., Seminars in Hematology 30 (4 suppl 4):3-12, discussion 13-15, 1993. For example, an attenuated or harmless commensal organism, including but not limited to a member of the Enterobacteriaceae, is transformed with the E. coli O157:H7 recombinant adhesin. This construct is administered to animals orally, or via intranasal inoculation. The efficacy of this approach is measured by assaying for specific IgA to the antigen, or by challenging vaccinated and control animals with E. coli O157:H7, and following excretion of this pathogen.

Immunization With Recombinant Bacterial Antigen Produced in Transgenic Plants: An alternative approach to the prevention of the carriage of E. coli O157:H7 by food animals, and possibly by humans, is to introduce the E. coli O157:H7 adhesin expressed in transgenic plants. For example, Haq et al. introduced sequences encoding E. coli heat-labile enterotoxin (LT-B) on an expression vector into Agrobacterium tumifaciens. This host bacterial strain was used to transfer the sequences encoding LT-B into tobacco (Nicotiana tabacum cv. Samsun) and potato (Solanum tubrosum variety “Frito-Lay 1607”) plants, which then expressed variable amounts of these toxin antigens, as determined by immunoassay. These bacterial proteins were fed to mice, either in the form of extract of tobacco plants which expressed LT-B, or as potatoes expressing LT-B (i.e., no extract). In both cases, serum IgG and mucosal IgA responses to LT-B were elicited.

Thus, as an analogous approach in the immunization of cattle to prevent carriage of E. coli O157:H7, transgenic plants expressing the E. coli O157:H7 adhesin are fed to food animals. The response to the antigen is measured by assaying for circulating IgG and mucosal IgA antibodies specific for the E. coli O157:H7 adhesin. These animals could then be challenged with E. coli O157:H7, and excretion followed, with appropriate control animals also challenged, to determine if an immunoprotective response has been elicited. The determination of immunoprotection would be analogous to the challenge with the irgA homolog deletion mutants, but in this case, animals immunized with transgenic plants expressing the IrgA homolog adhesin, or a control (i.e., irrelevant) antigen will be challenged with nalidixic acid resistant E. coli O157:H7 The excretion of challenge organisms will be followed by plating stool on SMA with nalidixic acid, and the duration and level of shedding of E. coli O157:H7 in the two groups will be compared.

It is also contemplated that immunizing transgenic plants may be used to protect humans from infection with E. coli O157:H7 and related pathogens expressing the E. coli O157:H7 adhesin, or homologs of this adhesin. In the case of humans, IgA and IgG can be assayed to detect a humoral response to this protein, but challenge experiments cannot be performed because of the exceptional pathogenicity ofE. coli O157:H7.

Use of a Recombinant Adhesin as a Competitive Inhibitor of E. coli O157:H7 in the Gastrointesinal Tracts of Food Animals and of Humans: Large quantities of recombinant adhesin expressed by plants and consumed in the diets of animals and humans may compete with wild type E. coli O157:H7 and related pathogenic bacteria for binding sites on the enterocyte surface. For example, animals are put on such feed in the several days or weeks prior to shipment for slaughter, provided controlled challenge experiments show that the recombinant adhesin, expressed in plants which are then fed to the animals, promotes clearance of E. coli O157:H7 from the gastrointestinal tract of such animals, thereby reducing the load of this pathogen that enters the production line in abattoirs. In a similar approach, children with the early stage of gastrointestinal infection with E. coli O157:H7 are administered these recombinant competitive inhibitors, including but not limited to the recombinant adhesin expressed on plants, to promote clearance of the organism, thereby ameliorating infection, or preventing the development of hemolytic uremic syndrome (HUS). Also, contacts of children with E coli O157:H7 infection, who have yet to display symptoms, are fed recombinant adhesin, with the intent that colonization in such potential secondary cases is prevented.

The disclosed adhesin may also be used to prevent or ameliorate human infection with E. coli O157:H7, as well as with bacteria which use a homolog of this adhesin, with shared active sites or epitopes. As discussed above, the E. coli O157:H7 adhesin demonstrates striking homology to IrgA. There are domains of high degree of homology (conserved regions) interspersed with relatively nonconserved regions (variable regions). The possibility exists, therefore, that the IrgA homolog in E. coli O157:H7 can be used as an immunogen (vaccine) against V. cholerae infections. An additional possibility is that the IrgA homolog in E. coli O157:H7 may be a useful vaccine against the carriage of other pathogenic Enterobacteriaceae by food animals, such as diarrheagenic E. coli that do not belong to serogroup O157:H7.

The IrgA homologue might also be useful in the induction of protective immunity, or as a competitive inhibitor of colonization, in several additional hosts and in several additional infections. These include (1) a vaccine to prevent infections in humans caused by E. coli O157:7H other diarrheagenic E. coli, and V. cholerae; (2) a vaccine to prevent carriage by cattle, and other animals destined for human food, of diarrheagenic E. coli; (3) a competitive inhibitor of colonizing diarrheagenic E. coli in food animals. To test the efficacy of these approaches, multiple experiments are contemplated. These include (1) the administration of immune globulin (passive immunity) from donors immunized with the IrgA homolog adhesin to high risk patients (i.e., children in contact with a primary case of E. coli O157:H7 infection), in a placebo-controlled fashion, to determine the differences in attack rates between the two groups; (2) a vaccine formulation could be administered to human volunteers subsequently challenged with V. cholerae, or to patients residing in areas of high risk for infection; (3) challenge of immunized food animals with related diarrheagenic E. coli (i.e., not E. coli O157:H7) and determine if colonization can be established. The possibility also exists that animals infected with pathogenic E. coli sharing an IrgA-mediated adherence mechanism cause illness in the animal industry. The IrgA homolog adhesin may, in these cases, be used to diagnose, prevent, or treat the infection. In fact, we have identified pear sequences in RDEC-1 and some E. coli strains from calves with diarrhea.

Glossary

Glossary of abbreviations, strains, plasmids, and genes relevant to this disclosure:

EPEC   enteropathogenic E. coli
MDBK   Madin-Darby bovine kidney cells
PhoA   alkaline phosphatase
PBS    phosphate-buffered saline
pSC    plasmid Supercos
PCR    polymerase chain reaction
SLT    Shiga-like toxins
SLT I  Shiga-like toxin I
SLT II Shiga-like toxin II
SMA    sorbitol-MacConkey agar
XP     5-bromo-4-chloro-3-indolylphosphate

E. coli O157:H7 86-24: An SLT-II producing isolate which caused a large restaurant-associated outbreak of hemorrhagic colitis which included two deaths (23,71).

E. coli O157:H7 NalR: Nalidixic acid resistant adherent mutant of E. coli O157:H7 86-24 used in TnphoA mutagenesis. This strain adheres in a localized pattern to HeLa cells.

E. coli O157:H7 Strains A5, F4, N11: TnphoA mutants of E. coli O157:H7 86-24 NalR which express PhoA, and are non-adherent, kanamycin resistant, and ampicillin sensitive.

E. coli O157:H7 Strains H8, P11, P12: TnphoA mutants of E. coli O157:H7 86-24 NalR which are adherent, kanamycin resistant, and ampicillin sensitive.

E. coli 20D2B: a TnphoA mutant of E. coli O157:H7 NalR which is hyperadherent. This strain does not express the O157 antigen, but retains all other characteristics of the parent strain (H7 antigen positive, sorbitol nonfermenting, SLT II positive, same API score).

E. coli NM554: A laboratory strain of E coli used for cosmid cloning.

Plasmids used:

pSC(A-G6) and pSC(T-H12): overlapping cosmid clones containing 30 kb of E. coli O157:H7 chromosomal DNA expressed in E. coli NM554 which confers D-mannose-resistant adherence on this E. coli in a mostly diffuse pattern.

pSC(overlap): deletion mutant of pSC(A-G6) which retains 15 kb overlap region, and confers adherence to E. coli HB101. pSC(overlap) was deposited on Jun. 24, 1994, under accession number 69648 at the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Md. 20852, U.S.A.

pGEX: A vector which allows the expression of the IrgA homolog adhesin and a molecule which can act as a ligand for an affinity purification step (67).

pear. adherence-conferring 7 kb subclone of pSC(overlap).

irgA: Iron regulated gene A, from V. cholerae. This has regions of homology to the adhesin of E coli O157:H7 that was cloned and is described in this application.

CIR: E. coli colicin I receptor. E coli outer membrane protein which also has regions of considerable homology to the described E. coli O157:H7 adhesin.

Strain F12: another O157 TnphoA mutant which has lost the ability to express the O157 antigen by virtue of having sustained a TnphoA insertion in the rfb locus. This strain is hyperadherent. ear deletions from F12 are considerably less adherent than the parent strain.

Strain 86-24 NalR (Δ irgA): a deletion mutant of E. coli O157:H7 stain 86-24 NalR, which is used in calf challenge experiments.

ear (epithelial adherence region): subclone of 15 kb pSC(overlap) region between pSC(A-G6) and pSC(T-H12), representing approximately 7 kb of cloned E. coli O157:H7 chromosomal DNA which confers the adherence phenotype to nonadherent laboratory E. coli.

Citations

• 1. Acres, S. D., et al., Infect Immun 1979; 25:121-126.
• 2. Baga, M., et al., Escherichia coli. EMBO J 1985; 4:3887-3893.
• 3. Bakker, D., et al., Escherichia coli. Mol Microbiol 1991; 5:875-886.
• 4. Beebakhee, G., et al., FEMS Microbiol Lett 1992; 91:63-68.
• 5. Bell, B, personal communication.
• 6. Belongia, E. A., et al., JAMA 1993; 269:883-888.
• 7. Bilge, S. S., et al., J Bacteriol 1989; 171:4281-4289.
• 8. Bokete, T. N., et al., Gastroenterology 1993; 105:1724-1731.
• 9. Borczyk, A. A., et al., Lancet 1987; 1:98.
• 10. Caprioli, A, et al., J Infect Dis 1992; 166:154-158.
• 11. Caprioli, A, et al., J Infect Dis 1994; 169:208-211.
• 12. Cordovez A., et al., J Clin Microbiol 1992; 30:2153-2157.
• 13. Cravioto, A, et al., Lancet 1991; 337:262-264.
• 14. Donnenberg, M. S., et al., Infect Immun 1990; 58:1565-1571.
• 15. Donnenberg, M. S., et al., Infect Immun 1991; 59:4310-4317.
• 16. Donnenberg, M. S., et al., Infect Immun 1992; 60:3953-3961.
• 17. Donnenberg, M. S., et al., J Clin Invest 1993; 92:1418-1424.
• 18. Duchet-Suchaux, M., et al., Infect Immun 1992; 60:2828-2834.
• 19. Enteric Diseases Branch, CDC, Morbid Mortal Wkly Rep 1993; 42:85-86.
• 20. Evans, G. A., K. et al., Gene 1989; 79:9-20.
• 21. Francis, D. H., et al., Am J Vet Res 1991; 52:1051-1055.
• 22. Fratamico, P M, et al., J Med Microbiol 39:371-381, 1993.
• 23. Griffin, P. M., et al., Ann Intern Med 1988; 109:705-712.
• 24. Griffin P. M., et al., Epidemiol Rev 1991; 13:60-98.
• 25. Hancock, D. D., et al., Epidemiology and Infection 113:119-207, 1994.
• 26. Hancock, D. D., et al., National Prevalence Study for Escherichia coli O157:H7 in United States Dairy Calves. Submitted.
• 27. Hancock, R. E. W., et al., J Bacteriol 1978; 136:381-90.
• 28. Henikoff S., Gene 1984; 28:351-359.
• 29. Ikemori Y., et al., Am J Vet Res 1992; 53:2005-2008.
• 30. Isaacson, R E., et al., Infect Immun 1980; 29:824-826.
• 31. Jacobs, A. A. C., et al., J Bacteriol 1987; 169:735-741.
• 32. Johnstone, A., et al., Immunochemistry in Practice, 2nd Ed. Blackwell Scientific Publications, Oxford, 1987, pp 190-196.
• 33. Junkins, A., et al. Curr Microbiol 1989; 19:21-27.
• 34. Karch, H., et al., Infect Immun 1987; 55:455-461.
• 35. Karmali M. A., et al., J Infect Dis 1985;151:775-782.
• 36. Kimura A., et al., Infect Immun 1990; 58:7-16.
• 37. Krogfelt, K. A., Rev Infect Dis 1991; 13:721-735.
• 38. LeSaux, N., et al., J Infect Dis 1993; 176:500-502.
• 39. Lindberg, F., et al., Nature 1987; 325:84-87.
• 40. Lopez E. L., et al., J Infect Dis 1989;160:469-475.
• 41. Louie, M., et al., Infect Immun 61:4085-4092, 1993.
• 42. MacDonald K. L., et al., JAMA 1988;259:3567-3570.
• 43. Marshall, B., et al., Proc Natl Acad Sci USA 1990; 87:6009-6613.
• 44. Martin, D. L., et al., N Engl J Med 1990; 323:1161-1167.
• 45. Martin M. L., et al., Lancet 1986; ii:1043.
• 46. McNamara, A. M., personal communication.
• 47. Montenegro, M. A., et al., J Clin Microbiol 1990; 28:1417-1421.
• 48. Moon, H. W., et al., Am J Clin Nutrition 1979; 32:119-127.
• 49. Morgan, R L., et al., Infect Immun 1978; 22:771-777.
• 50. Morris, J. A., et al., J Med Microbiol 1980; 13:265-271.
• 51. Oudega, B., et al., Antonie van Leeuwenhoek 1988; 54:285-299.
• 52. Pai C. H., et al., J Infect Dis 1988;157:1054-1057.
• 53. Pai C. H., et al., Infect Immun 1986; 51:16-23.
• 54. Pararuchuri, D. K., et al., Proc Natl Acad Sci USA 1990; 87:333-337.
• 55. Pecha, B., et al., J Clin Invest 1989; 83:2102-2108.
• 56. Raleigh, E. A., et al., Nucl Acid Res 1988; 16:1563-75.
• 57. Ratnam S., et al., J Clin Microbiol 1988; 26:2006-2012.
• 58. Riley, L. W., et al., Infect Immun 1987; 55:2052-2056.
• 59. Ritchie M., et al., J Clin Microbiol 1992:30;461-464.
• 60. Rowe, P. C., et al., Epidemiol Infect 1993; 110:9-16.
• 61. Runnels, P. L., et al., Infect Immun 55:555-558, 1987.
• 62. Samadpour M., et al., Appl Environ Microbiol 1994; in press.
• 63. Sancar, A., et al., J Bacteriol 1979; 137:692-693.
• 64. Sherman, P., et al., Infect Immun 1991; 59:890-899.
• 65. Sherman, P. M. et al., J Med Microbiol 1988; 26:11-17.
• 66. Sherman, P., et al., Infect Immun 1988; 56:756-761.
• 67. Smith, D. B., et al., Gene 1988; 67:31-40.
• 68. Sojka, W. J., et al., J Med Microbiol 1978. 11:493-499.
• 69. Swerdlow, D. L., et al., Ann Intern Med 1992; 117:812-819.
• 70. Tarr, P. I., et al., Am J Epidemiol 1989; 129:582-586.
• 71. Tarr, P. I., et al., J Infect Dis 1989; 159:344-347.
• 72. Tarr, P. I., et al., J Infect Dis 1990; 162:553-556.
• 73. Taylor, R. K., et al., J Bacteriol 1989; 171:1870-1878.
• 74. Taylor, R. K., et al., Proc Natl Acad Sci USA 1987; 84:2833-2837.
• 75. Toth, I., et al., Infect Immun 1990; 58:1223-1231.
• 76. Wadolkowsi, E. A., et al., Infect Immun 1990; 58:2438-2445.
• 77. Wahl, G. M., et al., Proc Natl Acad Sci USA 1987; 84:2160-2164.
• 78. Wells, J. G., et al., J Clin Micro 1983; 18:512-520.
• 79. Wells, J. G., et al., J Clin Microbiol 1991; 29:985-989.
• 80. Wessels, M. R. et al., Proc Natl Acad Sci USA 1991; 88:8317-8321.
• 81. Yokoyama, H., et al., Infect Immun 1992; 60:998-1007.
• 82. Yu, J., et al.,Mol Microbiol 1992; 6:411-7.

While the preferred embodiment of the invention has been illustrated and described, it will be appreciated that various changes can be made therein without departing from the spirit and scope of the invention.

Claims

1-6. (canceled)

7. A vaccine formulation for preventing and/or treating infection by a pathogenic member of the family Enterobacteriaceae in a mammalian subject comprising a peptide encoded by a nucleic acid molecule that hybridizes under stringent conditions to a nucleic acid molecule selected from the group of the nucleotide sequence of SEQ ID NO:4 and the nucleotide sequence complementary to SEQ ID NO:4, and a pharmaceutically acceptable carrier.

8. The vaccine formulation of claim 7 wherein the pharmaceutically acceptable carrier is suitable for administration by a route selected from the group consisting of an oral route, an intranasal route and a parenteral route.

9. The vaccine formulation of claim 7 wherein the pharmaceutically acceptable carrier is selected from the group consisting of sesame oil, peanut oil, sterile aqueous solution, and isotonic solutions suitable for injection.

10. The vaccine formulation of claim 7 wherein said pharmaceutically acceptable carrier comprises liposomes.

11. The vaccine formulation of claim 8 wherein said vaccine formulation is packaged in enteric coated capsules.

12. The vaccine formulation of claim 7 which further comprises a stabilizer.

13. The vaccine formulation of claim 12 wherein the stabilizer is selected from the group consisting of sorbitol and gelatin.

14. The vaccine formulation of claim 7 which further comprises an immunoadjuvant.

15. The vaccine formulation of claim 14 wherein the immunoadjuvant is selected from the group consisting of: the water-in-oil emulsion of Freund, light mineral oil, Arlacel A, Arlacel C, liposome-antigen combinations, cholera toxin B subunit conjugates and monophosphoryl lipid A.

16. The vaccine formulation of claim 7 which comprises a peptide that is encoded by a nucleic acid molecule that hybridizes under stringent conditions to the nucleotide sequence of SEQ ID NO:4 or its complement.

17. The vaccine formulation of claim 16 wherein said peptide is derived from a recombinant organism.

18. The vaccine formulation of claim 16 wherein said peptide is produced in a transgenic plant.

19. The vaccine formulation of claim 16 which comprises an amount of the peptide effective to promote the clearance of the pathogenic enterobacteriaceae from the gastrointestinal tract of a mammalian subject infected by a pathogenic member of the family Enterobacteriaceae when administered thereto.