Nucleic Acid Probes for Detecting E. Coli O157:H7

The invention relates to genetic engineering and provides nucleic acid probes for detecting pathogenic Escherichia coli O157:H7 in food and fecal samples.

E. coli O157:H7 is a virulent food-borne pathogen that causes acute hemorrhagic colitis (bloody inflammation of the colon), and sometimes a severe hemolytic uremic syndrome (HUS), in children. These conditions can be fatal, or cause permanent kidney damage and shortened life expectancy. Treatment options for this infection are limited; prevention of human infection is therefore of the greatest importance.

Epidemics of E. coli O157:H7 generally have been traced to contaminated water or meats, especially undercooked hamburger. Following the 1993 massive outbreak in Washington State, the strain responsible for the epidemic disappeared soon after the recall of the incriminated vehicle. Hence, ingestion of contaminated beef, and not person to person spread, appears to be the chief source of human infection.

Because of the rapid onset of life-threatening complications, expedient diagnosis is of utmost importance to those infected with this strain of E. coli, especially when the patient is a child. However, because the early symptoms of this disease can mimic other disorders, patients infected with this organism during the initial stages of new epidemics may not be tested for E. coli O157:H7 until other causes have been ruled out. Once E. coli O157:H7 is suspected, diagnostic testing methods currently available require about a day to complete before a definitive diagnosis can be confirmed. Any means for shortening this dangerous delay would reduce the risk to infected patients and family members who might become infected through person to person transmission.

Of utmost importance for controlling outbreaks of E. coli O157:H7 is the rapid detection of the contaminated food source. A common method for detecting E. coli O157:H7 includes incubation on agar containing sorbitol, a substrate that supports growth of most fecal E. coli strains, but which the O157:H7 strain cannot metabolize. However, failure to ferment sorbitol does not provide a definitive diagnosis, as non-pathogenic strains of E. coli exist that also cannot ferment sorbitol.

A latex agglutination test for identification of the serogroup O157 is available from Oxoid (Unipath Limited, Basingstoke, Hampshire, England), which recommends that the test be applied to isolates that already have been determined to lack the ability to ferment sorbitol. This antibody-based test lacks specificity: Some strains of E. hermnaii share the antigen detected by the antiserum; hence this test must be confirmed by further fermentation testing before an isolate can be conclusively identified as E. coli O157:H7. Furthermore, noncytotoxic E. coli may possess the E. coli O157 antigen, leading to false positive reactions in such 0157 antigen detection systems.

Other diagnostic methods for E. coli O157:H7 have been suggested. Immunofluorescent examination of fresh stool with labeled O157-specific antiserum has recently been proposed as a rapid technique for the detection of patients whose stools contain E. coli O157:H7 (Park et al., Am. J. Clin. Path. 101:91-94, 1994). Antibody-coated magnetic beads (Dynal) are available that react with E. coli O157:H7 However, antigen detection tests should be accompanied by a confirmatory stool culture (Tarr, Clinical Infectious Diseases, 20:1-8, 1995).

A test kit for E. coli O157:H7 is available which reportedly can detect this organism in meat (Organon Teknika, Durham, N.C.). This is an antigen-based technology, in which bacterial antigens are used as targets to identify the organism. The test requires a day (i.e., overnight) to produce a presumptive positive test. An additional technique which has been proposed is based on a rapid dipstick immunoassay to detect enterohemorrhagic E. coli O157:H7 in retail ground beef (Kim et al., Applied Environmental Micobiology 58: 1764-1767, 1992).

For diagnostic purposes, conventional microbiological and antibody-based testing techniques may be insufficient for management of epidemic outbursts of HUS. By the time HUS appears, about two-thirds of patients no longer have E. coli O157:H7 in their stools. Moreover, the numbers of organisms present in contaminated food samples often are too low to be readily detected. DNA probes specific for this pathogen would provide an alternative method that is capable of detecting small numbers of organisms. Most efforts in the past to develop such probes have focused on the gene encoding the Shiga-like toxins (SLTs) believed to be responsible for many of strain O157's pathogenic effects. See: Levine et at., The Journal of infectious Diseases 156(1):175-182, 1987; Samadpour et al., Applied and Environmental Microbiology 56(5):1212-1215, 1990; Pollard et al., Journal of Clinical Microbiology 28(3):540-545, 1990; Pollard et al., The Journal of Infectious Diseases 162:1195-1198, 1990. However, the pathogenecity of many strains containing SLT genes is questionable (Tarr, Clinical Infectious Diseases, 20:1-8, 1995).

PCR assays to detect O157:H7 strains using the uidA gene have also been reported (Feng et at., Applied and Environmental Microbiology 57(1):320-323, 1991; Feng, Molecular and Cellular Probes 7:151-154, 1993.).

More recently, a multiplex PCR assay was reported which uses three sets of primers, two of which are directed to conserved regions within genes encoding for SLT-I and SLT-II, and the third set directed to the uidA gene (Cebula et at., Journal of Clinical Microbiology 33(1):248-250, 1995).

The invention provides a nucleic acid probe for detecting the presence of enterohemorrhagic E. coli O157:H7, comprising an isolated nucleic acid molecule that hybridizes under stringent conditions to SEQ ID NO:1 or its complement and to the DNA of E. coli O157:H7 but not to the DNA of the clonally related enteropathogenic E. coli O55:H7.

The invention provides a DNA fragment (SEQ ID NO:1) that is useful for detecting E. coli O157:H7 in food and fecal samples. This DNA fragment, isolated from E. coli O157:H7, does not cross-hybridize with the known genes for Shiga toxins in Shigella dysenteriae, or the uidA gene which encodes for .beta.-glucoronidase in E. coli. Rather, this DNA fragment unexpectedly contains two genes that are related by homology to the putative transmembrane export protein of Yersinia pseudotuberculosis (Kessler et al., Journal of Bacteriology 175(5):1412-1422, 1993), and to the rfbE gene which reportedly encodes perosamine synthetase in Vibrio cholerae (Stroeher et ai,, Proc. Natl. Acad. Sci, U.S.A. 89:2566-2570, 1992).

The subject DNA fragment was serendipitously identified through efforts to determine the role of the O157 lipopolysaccharide in the pathogenesis of E. coli O157:H7. Briefly summarized, TnphoA insertion mutagenesis was used to create mutants of E. coli O157:H7 that did not express this antigen. It was noted that such mutants were hyperadherent, that is, they had a pronounced increase in their ability to adhere to HeLa cells. Two O157-hyperadherent mutants, designated 20D2B and F12, were selected for detailed analysis after screening ca. 3000 TnphoA mutants of E. coil O157:H7 for the loss of the antigen by the latex particle agglutination test. Both mutant strains were approximately eight-fold more adherent to HeLa cells than the parent strain. Southern blots indicated that these mutants contained two and one TnphoA insertions, respectively. Nucleotide sequence analysis of the DNA regions flanking the F12 insertion surprisingly revealed two open reading frames that share significant homology with the putative transmembrane export protein of Yersinia pseudotuberculosis and the rfbE gene of Vibrio cholerae. Using PCR primers (identified below) from these flanking regions, the subject DNA fragment was cloned from E. coil O157:H7. This fragment is referred to herein as "the F12 region" and its nucleotide sequence is depicted in SEQ ID NO:1.

The invention provides nudeic acid probes derived from the F12 region. Preferred probes are selected from the two open reading flames corresponding to nucleotides 1-1026 (SEQ ID NO:2) and nucleotides 1029-2123 (SEQ ID NO:3) shown in SEQ ID NO:1. By way of example, use of a nucleic acid probe (SEQ ID NO:4), corresponding to nucleotides 913-2199 shown in SEQ ID NO:1, is described below. Other nucleic acid probes can be readily selected from among the contiguous nucleotides within the disclosed sequence(s) and screened for both sensitivity (for strain O157:H7) and specificity (lack of sensitivity for other microbial contaminants of food and feces) using an appropriate panel of positive and negative genomic controls.

In the screening assay, E. coli O157:H7 serves as the positive genomic control. Negative genomic controls are selected from among other microbial strains that typically contaminate the type of food, agricultural, or clinical sample to be tested. Such negative genomic controls are preferably selected from among E. coli strains that are genetically and antigenically related and most preferably clonally related to strain O157:H7. The clonal relationships among E. coli strains that cause hemorrhagic colitis and infantile diarrhea are described in Whittam et al., Infection and Immunity 61:1619-1629, 1993. FIG. 4 therein presents a phylogenetic tree of genomically dosely related strains. In the same lineage as strain O157:H7 are the following clonally related strains which can serve as suitable negative controls, notably: the enteropathogenic strains O55:H7 (i.e., the dosest phylogenetically related strain), O111:H12, and O55:H6. Other E. coli strains suitable as negative controls include noncytotoxic and cytotoxic E. coli of serotypes O157:H43, O111:H8, O26:H11, O128:H7, O128:H21, O111:H21, O128:H2, and O111:H2.

Bacterial strains for use in the screening panel can be obtained from the reference collections of: the Centers for Disease Control (CDC), Atlanta, Ga.; the E. coli Reference Center, University Park, Pa.; the American Type Culture Collection (ATCC), Rockville, Md.; the Center for Vaccine Development, Baltimore, Md.; the Division of Microbiology, Food and Drug Administration (FDA), Washington, D.C.; the Center for Disease Control, Ottawa, Canada; the Universitats-Krankenhaus, Hamburg, Germany; and the Statens Seruminstitut, Copenhagen, Denmark.

Such a reference panel of bacteria can be employed to identify nucleic acid probes from the F12 region shown in SEQ ID NO:1 that provide the requisite sensitivity and selectivity for distinguishing E. coli O157:H7 from other bacteria, including related strains of E. coli. In a representative screening system, bacteria in the test panel are plated on trypticase soy agar, as described in Samadpour et at., 1990, which is hereby incorporated by reference. Candidate oligonucleotide probes randomly chosen from contiguous nucleotides within the F12 region are synthesized using a DNA synthesizer, following the procedures recommended by the manufacturer. A probe corresponding to SEQ ID NO:4, which has been shown to possess the requisite specificity, may be used as a control probe. All probes are labeled using conventional techniques. Aliquots of chromosomal DNA extracted from each strain of bacteria in the panel are hybridized under stringent conditions with the labeled candidate and control probes. Wild-type E. coli O157:H7 serves as the positive control. Candidate probes that hybridize under stringent conditions with O157:H7 but only slightly or not at all with the negative control strains are deemed to have the requisite specificity for detecting E. coli O157:H7 in samples of ground meat, bovine feces, and samples of clinical interest.

The F12 fragment (SEQ ID NO:1) can also be employed to identify nucleic acid probes from the E. coli O157:H7 genome flanking the F12 region that provide the requisite sensitivity and selectivity. To identify these contiguous sequences using the F12 region, a library of random DNA fragments is generated by partial endonuclease digestion, cloned into an appropriate vector, and screened with the F12 probe. Cloned DNA is mapped to determine if regions contiguous to the F12 region are present on the inserted DNA identified by homology to the F12 region. The contiguous DNA is sequenced, and tested for specificity and sensitivity as a probe for the detection of E. coli O157:H7.

The nucleotide length of the probe is chosen using customary criterion. Ten nucleotides is generally the lower limit for such nucleotide segments, because sequences smaller than that sometimes fail to form stable hybrid duplexes, or may hybridize nonspecifically with unrelated genes that happen to share short stretches of complementary sequence with the desired target sequence. Accordingly, in preferred embodiments, the probes are at least 10 nucleotides in length, with the oligonucleotide being capable of forming a detectable stable duplex with the F12 region sequence(s). More preferred are probes of at least 15 nucleotides, as such probes are more likely to form stable duplexes with their complementary sequences. Oligonucleotides at least 15 bases long are very likely to be specific, as sequences of this length will occur by chance no more than once in a mammalian genome, hence are extremely unlikely to occur by chance in a bacterial genome. Preferred probes will range from 15 to 50 bases in length, as probes in this range can be prepared at a reasonable cost using commercially available DNA synthesizers or services that synthesize oligonucleotides on request. The invention also contemplates that fragments of DNA even longer than 50 nucleotides may be obtained from the F12 region, by using restriction endonucleases, or by using PCR primers and subsequently cloning these longer fragments.

For use in the described selection system or for analyzing food or fecal samples, probes must be labeled in order to be detectable in hybrids formed during the assays. Synthetic probes are conventionally labeled with radioactive phosphorus using polynucleofide kinase. Alternatively the probes can be tagged with an affinity reagent such as biotin or streptavidin, thus enabling the probe to bind to an enzyme capable of cleaving a chromogenie substrate (for details, see Sambrook et al.).

Various conventional protocols can be employed for the hybridization step. Generally, the labeled probe is present in a suitable hybridization solution, while the target nucleic acids are fixed to an insoluble substratum, such as derivatized nylon. In the preferred embodiment, target DNA sequences and labeled probes are incubated under stringent hybridization conditions. Such conditions are generally understood to be those that permit only perfectly matched or nearly perfectly matched hybrids to form, and their use will assure that probes hybridize specifically with sequences of E. coli O157:H7. The destabilizing effects of mismatched bases increases as probe length decreases, such that for probes shorter than 20 nucleotides, stringent hybridization conditions will rarely tolerate any mismatched bases. For longer probes, it is possible that a few mismatches may be present in hybrids that form even under stringent conditions. In some situations, practitioners could be misled by results involving mismatched hybrids, but in the present instance, the screening methods provided ensure identification of probes having the appropriate specificity regardless of whether mismatches may be present. That is, the screening procedure identifies probes that react with the targeted O157:H7 strain (and, optimally, with the positive control F 12 region) but not with the negative control bacteria. Hence it is immaterial whether any given probe forms a perfect or only a near-perfect hybrid, as the ability to distinguish these two groups of bacteria is the only pertinent quality in the context of the present test assays.

A variety of operable hybridization conditions are available to one skilled in the art, e.g., as provided in Sambrook et al. For probes longer than about 200 nucleotides in length, hybridization usually is carried out at about 25.degree. C. below the melting temperature of the hybrid, as this gives the optimal rate of duplex formation. For shorter probes, hybridization is usually carried out closer to the calculated melting temperature of a perfect hybrid, e.g., 10.degree.-15.degree. C. below the melting temperature. For probes of various lengths, melting temperatures for perfectly matched hybrids can be determined empirically, or can be calculated using one of the formulas found in Sambrook et at. An exemplary hybridization solution for oligonucleotides contains 6 X SSC (ca. 1M [Na+]), and 0.5% sodium dodecyl sulfate.

The subject sequences can also be used for detection of E. coli 0157:H7 in polymicrobial samples using polymerase chain reaction (PCR) methodology. For example, nucleotide primers which have a high likelihood of being specific for E. coli O157:H7 are determined, using techniques described above, and used in PCR assays to detect genomic sequences specific for E. coli O157:H7 in enrichment cultures likely to contain the target organism. For example, a sample suspected of having E. coli O157:H7 in its natural condition (e.g., ground meat or bovine feces) is incubated in a broth enrichment culture, DNA is extracted using standard techniques, the DNA is annealed to the appropriate primers, in the PCR reaction, and the resulting bands are examined on agarose gels for products of this reaction. A positive band suggests the presence of E. coli O157:H7. It is estimated that such a technique could be performed within several hours of the start of incubation, that is, in a time period considerably less than present techniques which rely on amplification of the organism to an extent that antigens can be detected.

In another embodiment, the expression products (SEQ IDS NO:5 and NO:6, respectively) of the nucleotide sequences shown in SEQ ID NOS: 2 and 3 are considered candidate immunogens for preparing antibody reagents for detection of the O157:H7 strain.

The invention is further illustrated by the following example, which describes detection of E. coli O157:H7 using a nucleic acid probe (SEQ ID NO:4) encompassing the disclosed rfbE homologous sequence (SEQ ID NO:3).

Total bacterial DNA (chromosome and plasmid) was prepared from a variety of bacteria, including: E. coli O157:H7 strain 86-24, a strain which caused an outbreak of infection in a fast food restaurant in Walla Walla, Wash., in 1986 (Griffin et al., Annals of Internal Medicine, 109:705-712, 1988) as the positive control (the F12 region was derived from this strain); E. coli NM544 (a laboratory strain) as the negative control (Raleigh et al., Nucleic Acids Research, 16:1563-1575, 1988); five test strains from diarrheal E. coli group 5 (all E. coli O55:H7); a clone of E. coli which is closely related to E. coli O157:H7 (Whittam et al., Infection and Immunity, 61:1619-1629, 1993); and five test strains or E. coli O157:H43. The DNA from each of these organisms was digested with EcoRI, separated electrophoretically in 0.7% agarose, and transferred to a nylon support membrane. The immobilized DNA was probed with a fragment of F12 DNA (SEQ ID NO:4), and homologous DNA sequences were detected in: E. coli O157:H7 strain 86-24; and four of the five E. coli O157:H43 strains probed, though at lesser intensity. This probe (SEQ ID NO:4) identified an EcoRI fragment of ca. 6 kb in the strains which were positive. The probe did not identify homologous DNA in the negative control, the E. coli O55:H7 strains, and one of the E. coli O157:H43 strains. It is important to note, however, that the nonreactive E. coli O157:H43 strain was only very weakly positive in the Oxoid latex particle agglutination test (unlike the other E. coli O157:H43 strains), and its identity as a member of the O157 serogroup is not certain.

Plasmid pF12 containing SEQ ID NO:7 (i.e., nucleotides 32-2199 of SEQ ID NO:1) was deposited on Apr. 14, 1995, at the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Md. 20852, U.S.A. pF12 is a PCR product of the cloned homologs of the V. cholerae rfbE gene and the gene encoding the putative transmembrane export protein of Y. pseudotuberculosis from E. coli O157:H7 strain 86-24, as well as some flanking sequence. It was cloned by using as primers the sequences 5'CTT CTG GCA TGA TTG ATT GGC3' (SEQ ID NO:8) and 3'CGA GTG GGG CGG TGG AAT TG5' (SEQ ID NO:9) from the deduced nucleotide sequences of the cloned DNA segments surrounding the TripboA insertion in strain F12. The primers were modified to contain BamHI and EcoRI sites, respectively, which were used to insert the PCK product into pSK+ (Stratagene) at these sites. Confirmatory sequencing from the PCR products and of the cloned regions flanking the TnphoA insertion demonstrate two errors in these cloned segments. (SEQ ID NO:1 is the correct sequence.) These include a T at position 964 (which is an erroneous C in the cloned insert deposited at ATCC) and a C at position 1510 (which is an erroneous T in the cloned insert deposited at ATCC).

The practice of the present invention will employ, unless otherwise indicated, a number of conventional techniques of molecular and cellular biology, recombinant DNA, microbiology, and immunology, which are within the skill of the art. Such techniques are explained fully in the scientific literature and detailed experimental protocols are available in a number of technical manuals; see for example: F. M. Ausubel et at. (eds.), "Current Protocols in Molecular Biology," (1987 and 1993); Kriegler, M. (ed.), "Gene Transfer and Expression, a Laboratory Manual," (1990), W. H. Freeman Publishers; Sambrook, Fritsch, and Maniatis, "Molecular Cloning: A Laboratory Manual," Second Edition (1989); Erlich, H. A. (ed.) "PCR Technology: Principles and Applications for DNA Amplification," (1989), Stockton Press; Langone, J. J. and H. Van Vunakis (eds.), "Immunological Techniques, Part I: Hybridoma technology and monoclonal antibodies," Methods in Enzymology 121:1-947 (1986); Hurrell, J. G. R. (ed.), "Monoclonal Hybridoma Antibodies: Techniques and Applications," (1982), CRC Press, Boca Raton, Fla.; Coligan, J. E. et al. (eds), "Current Protocols in Immunology," (1991), Greene Publishing and Wiley-Interscience, N.Y.; Weir, D. M., "Handbook of Experimental Immunology," (1986), Blackwell Scientific; Kurstack, E., "Enzyme Immunodiagnostics," (1986), Academic Press, San Diego; Polak, J. M. and S. Van Noorden (eds.), "Immunocytochemistry: Practical Applications in Pathology and Biology," (1983), John Wright PSG, Littleton, Mass.; Stemberger, L. A., "Immunocytochemistry," (1986), Wiley, N.Y.; and primary references cited therein. All publications, patents, and patent publications mentioned herein are hereby incorporated herein by reference in their entirety.

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.