AU4023599A - Antigens and their detection - Google Patents

Description

WO99/61458 PCT/AU99/00385 - 1 Antigens and Their Detection TECHNICAL FIELD The invention relates to novel nucleotide sequences located in a gene which encodes a bacterial flagellin 5 antigen, and the use of those nucleotidessequences for the detection of bacteria which express particular flagellin antigens, on the basis of that antigen alone, or in conjunction with the 0 antigen expressed by that strain. 10 BACKGROUND ART The flagellum of many bacteria appears to be made up of a single protein known as flagellin. The serotyping schemes of E. coli and Salmonella enterica are based on highly variable antigenic surface structures which include the 15 lipopolysaccharide which carries the 0 antigen and flagellin which is now known to be the carrier of the classical H antigen. In many strains of S. enterica there are two loci (fliC and fljB) which encode flagellin, and a regulatory system which allows one only to be expressed at 20 any time; and which also provides for expression to rapidly alternate between the two forms first identified as two phases (H1 and H2) for the H antigen of most strains. In E. coli there are 54 forms of H antigen recognised and until recently they were all thought to be 25 encoded at the fliC locus, as has been'shown for E. coli K-12. However in the 1980s Ratiner [Ratiner Y A "Phase variation of the H antigen in Escherichia coli strain Bi327-41, the standard strain for Escherichia coli flagellin antigen H3" FEMS Microbiol. Lett 15 (1982) 33 30 36; Ratiner Y A "Presence of two structural genes determining antigenically different phase-specific flagellins in some Escherichia coli strains" FEMS Microbiol. Lett. 19 (1983) 37-41; Ratiner Y A "Two genetic arrangements determining flagellin antigen specificities 35 in two diphasic Escherichia coli strains" FEMS Microbiol. Lett. 29 ,(1985) 317-323; Ratiner Y A "Different alleles of the flagellin gene hagB in Escherichia coli standard H WO99/61458 PCT/AU99/00385 - 2 test strains" FEMS Microbiol Lett. 48 (1987) 97-104.] showed that in some cases there are two loci and that expression can alternate. The matter was further complicated by a recent paper by Ratiner [Ratiner Y A 5 (1998)"New flagellin-specifying genes in some Escherichia coli strains" J. Bacteriol. 180 979-984] showing three loci (flk, fl and flm) for flagellin in addition to fliC although the fljB locus has not been found in E. coli. However E. coli strains are normally identified by the 10 combination of one 0 antigen and one H antigen [and K antigen when present as a capsule (K) antigen], with no problems reported for the vast majority of cases with alternate phases, while S. enterica strains are normally identified by the combination of O, H1 and H2 antigens. It 15 is still not clear how widespread in E. coli H antigens determined by flagellin genes other than fliC are. Typing is typically carried out using specific antisera. The incidence of pathogenic E. coli in association with human and animal disease supports the 20 need for suitable and rapid typing techniques. DESCRIPTION OF THE INVENTION In a first aspect, the present invention provides a novel nucleic acid molecule encoding all or part of an E. 25 coli flagellin protein. The present invention provides, for the first time, full length sequence for a flagellin gene for the following E. coli type strains: H6 (SEQ ID NO: 8), H9(SEQ ID NO: 11), H10(SEQ ID NO: 12), H14(SEQ ID NO: 15), 30 H18(SEQ ID NO: 18), H23(SEQ ID NO: 22), H51(SEQ ID NO: 50), H45(SEQ ID NO: 43), H49(SEQ ID NO: 48), H19(SEQ ID NO: 19), H30(SEQ ID NO: 29), H32(SEQ ID NO: 31), H26(SEQ ID NO: 25), H41(SEQ ID NO: 39), H15(SEQ ID NO: 16), H20(SEQ ID NO: 20), H28(SEQ ID NO: 27), H46(SEQ ID NO: 35 44), H31(SEQ ID NO: 30), H34(SEQ ID NO: 33), H43(SEQ ID NO: 41) and H52(SEQ ID NO: 51). Corrected full length sequences have been obtained for H7(SEQ ID NO: 9) and WO99/61458 PCT/AU99/00385 - 3 H12(SEQ ID NO: 14) type strains. Partial flagellin gene sequence, including the central variable region, has been obtained for the following E. coli H type strains: H40(SEQ ID NO: 38), 5 H8(SEQ ID NO: 10), H21(SEQ ID NO: 21), H47(SEQ ID NO: 46), Hll(SEQ ID NO: 13), H17(SEQ ID NO: 17), H25(SEQ ID NO: 24), H42(SEQ ID NO: 40), H27(SEQ ID NO: 26), H35(SEQ ID NO: 34), H2(SEQ ID NO: 67), H3(SEQ ID NO: 68), H24(SEQ ID NO: 23), H37(SEQ ID NO: 35), H50(SEQ ID NO: 49), H4(SEQ ID 10 NO: 6), H44(SEQ ID NO: 42), H38(SEQ ID NO: 36), H39(SEQ ID NO: 37), H55(SEQ ID NO: 53), H29(SEQ ID NO: 28), H33(SEQ ID NO: 32), H5(SEQ ID NO: 7), H54(SEQ ID NO: 52) and H56(SEQ ID NO: 54). Comparison of sequences demonstrates that unique 15 flagellin genes have now been sequenced (partially or completely) for the following E. coli H type strains: H1, H2, H3, H5, H6, H7, H9, H11, H12, H14, H15, H18, H19, H20, H21, H23, H24, H25, H26, H27, H28, H29, H30, H31, H32, H33, H34, H35, H37, H38, H39, H41, H42, H43, H45, H46, 20 H48, H49, H51, H52, H54, and H56 and either H8 or H40, H10 or H50 and H4 or H17. By comparison of these sequences, the present inventors were able to identify specific sequences for each of the above H serotypes. 25 The present invention also provides fliC sequences from 10 different H7 strains, in addition to that from the H7 type strain, and two sequences specific to H7 of 0157 and 055 E. coli strains. The present invention encompasses all or part of the 30 flagellin genes sequenced for H2, H3, H5, H6, H9, H11, H14, H18, H19, H20, H21, H23, H24, H25, H26, H27, H28, H29, H30, H31, H32, H33, H34, H35, H37, H38, H39, H41, H42, H43, H44, H45, H46, H47, H48, H49, H51, H52, H54, H55, H56, HS, H40, H15, H10, or H50, H4 and H17 type 35 strains. Of these flagellin genes sequenced, those from the type strains for H8 and H40 are identical, those from type strains HI0 and H50, H1 and H12, H38 and H55, H21 and WO99/61458 PCT/AU99/00385 -4 H47, and H4, H17 and H44 type strains are highly similar. The invention also encompasses newly provided sequence for H7 and H12 as well as novel primers for the specific amplification of H1, H7, H12 and H48 as well as 5 for the other above mentioned newly sequenced flagellin genes. By cloning and expression of these sequenced flagellin genes in a fliC deletion E. coli K-12 strain, and use of anti-H antiserum, we have confirmed the H 10 specificities encoded by 39 falgellin genes. The 39 H specificities are H1, H2, H4, H5, H6, H7, H9, HI0, H11, H12, H14, H15, H16, H18, H19, H20, H21, H23, H24, H26, H27, H28, H29, H30, H31, H32, H33, H34, H38, H39, H41, H42, H43, H45, H46, H49, H51, H52, and H56, encoded by 15 flagellin genes obtained from H type strains for H1, H2, H4, H5, H6, H7, H9, H10, H11, H12, H14, H15, H3, H18, H19, H20, H21, H23, H24, H26, H27, H28, H29, H30, H31, H32, H33, H34, H38, H39, H41, H42, H43, H45, H46, H49, H51, H52, and H56 respectively. 20 The nucleic acid molecules of the invention may be variable in length. In one embodiment they are oligonucleotides of from about 10 to about 20 nucleotides in length. The oligonucleotides of the invention are specific for the flagellin gene from which they are 25 derived and are derived from the central region of the gene. In one embodiment, oligonucleotides in accordance with the present invention, which also include oligonucleotides from the previously sequenced E. coli H1, H7, H12 and H48 genes, are those shown in Table 3. 30 The 45 sequences (see Table 3) provide a panel to which newly sequenced genes can be compared to select specific oligonucleotides for those newly sequenced genes. In a second aspect the invention provides a method of detecting the presence of E. coli of a particular H 35 serotype in a sample, the method comprising the step of specifically hybridising at least one nucleic acid molecule derived from a flagellin gene, wherein the at WO99/61458 PCT/AU99/00385 -5 least one nucleic acid molecule is specific for a particular flagellin gene associated with the H serotype, to any E. coli in the sample which contain the gene, and detecting any specifically hybridised nucleic acid 5 molecules, wherein the presence of specifically hybridised nucleic acid molecules identifies the presence of the H serotype in the sample. In one preferred embodiment the detection method is a Southern blot method. More preferably, the nucleic acid 10 molecule is labelled and hybridisation of the nucleic acid molecule is detected by autoradiography or detection of fluorescence. Preferred nucleic acid molecules for the detection of particular flagellin genes are listed in Table 3. 15 In a third aspect the invention provides a method of detecting the presence of E. coli of a particular H serotype in a sample, the method comprising the step of specifically hybridising at least one pair of nucleic acid molecules to any E. coli in the sample which contains the 20 flagellin gene for the particular H serotype, wherein at least one of the nucleic acid molecules is specific for the particular flagellin gene associated with the H serotype, and detecting any specifically hybridised nucleic acid molecules, wherein the presence of 25 specifically hybridised nucleic acid molecules identifies the presence of the H serotype in the sample. In one preferred embodiment the detection method is a polymerase chain reaction method. More preferably, the nucleic acid molecules are labelled and hybridisation of 30 the nucleic acid molecule is detected by electrophoresis. It is recognised that there may be instances where spurious hybridisation will arise through the initial selection of a sequence found in many different genes but this is typically recognisable by, for instance, 35 comparison of band sizes against controls in PCR gels, and an alternative sequence can be selected.

WO99/61458 PCT/AU99/00385 - 6 In a fourth aspect the invention provides a method for detecting the presence of a particular 0 serotype and H serotype of E. coli in a sample, the method comprising the following steps: 5 (a) specifically hybridising at least one nucleic acid molecule, derived from and specific for a gene encoding a transferase or a gene encoding an enzyme for the transport or processing of a polysaccharide or oligosaccharide unit, the gene being involved in the 10 synthesis of a particular E. coli 0 antigen, to any E. coli in the sample which contain the gene; (b) specifically hybridising at least one nucleic acid molecule derived from and specific for a particular flagellin gene associated with that H serotype, to any E. 15 coli in the sample which contain the gene; and (c) detecting any specifically hybridised nucleic acid molecules. Preferred nucleic acid molecules for the detection of particular flagellin genes are listed in Table 3. 20 In one preferred embodiment, the sequence of the nucleic acid molecule specific for the 0 antigen is specific to the nucleotide sequence encoding the 0111 antigen. More preferably, the sequence is derived from a gene selected from the group consisting of wbdH 25 (nucleotide position 739 to 1932 of Figure 5), wzx (nucleotide position 8646 to 9911 of Figure 5), wzy (nucleotide position 9901 to 10953 of Figure 5), wbdM (nucleotide position 11821 to 12945 of Figure 5) and fragments of those molecules of at least 10-12 nucleotides 30 in length. Particularly preferred nucleic acid molecules are those set out in Tables 8 and 8A, with respect to the above mentioned genes. In another preferred embodiment, the sequence of the nucleic acid molecule specific for the 0 antigen is 35 specific to the nucleotide sequence encoding the 0157 antigen. More preferably, the sequence is derived from a gene selected from the group consisting of wbdNl WO99/61458 PCT/AU99/00385 - 7 (nucleotide position 79 to 861 of Figure 6), wbdO (nucleotide position 2011 to 2757 of Figure 6), wbdP (nucleotide position 5257 to 6471 of Figure 6), wbdR (nucleotide position 13156 to 13821 of Figure 6), wzx 5 (nucleotide position 2744 to 4135 of Figure 6) and wzy (nucleotide position 858 to 2042 of Figure 6) and fragments of those molecules of at least 10-12 nucleotides in length. Particularly preferred nucleic acid molecules are those set out in Tables 9 and 9A, with respect to the 10 above mentioned genes. In one preferred embodiment the detection method is a Southern blot method. More preferably, the nucleic acid molecule is labelled and hybridisation of the nucleic acid molecule is detected by autoradiography or detection of 15 fluorescence. In a fifth aspect the invention provides a method for detecting the presence of a particular 0 serotype and H serotype of E. coli in a sample, the method comprising the following steps: 20 (a) specifically hybridising at least one pair of nucleic acid molecules, at least one of which is derived from and specific for a gene encoding a transferase or a gene encoding an enzyme for the transport or processing of a polysaccharide or oligosaccharide unit, the gene being 25 involved in the synthesis of the particular E. coli 0 antigen, to any E. coli in the sample which contain the gene; (b) specifically hybridising at least one pair of nucleic acid molecules, at least one of which is derived 30 from and specific for a particular flagellin gene associated with the particular H serotype, to any E. coli in the sample which contain the gene; and (c) detecting any specifically hybridised nucleic acid molecules. 35 Preferred nucleic acid molecules for the detection of particular flagellin genes are listed in Table 3.

WO99/61458 PCT/AU99/00385 - 8 In one preferred embodiment, the sequence of the nucleic acid molecule specific for the 0 antigen is specific to the nucleotide sequence encoding the 0111 antigen. More preferably, the sequence is derived from a 5 gene selected from the group consisting of wbdH (nucleotide position 739 to 1932 of Figure 5), wzx (nucleotide position 8646 to 9911 of Figure 5), wzy (nucleotide position 9901 to 10953 of Figure 5), wbdM (nucleotide position 11821 to 12945 of Figure 5) and 10 fragments of those molecules of at least 10-12 nucleotides in length. Particularly preferred nucleic acid molecules are those set out in Tables 8and 8A, with respect to the above mentioned genes. In another preferred embodiment, the sequence of the 15 nucleic acid molecule specific for the 0 antigen is specific to the nucleotide sequence encoding the 0157 antigen. More preferably, the sequence is derived from a gene selected from the group consisting of wbdN(nucleotide position 79 to 861 of Figure 6), wbdO (nucleotide position 20 2011 to 2757 of Figure 6), wbdP (nucleotide position 5257 to 6471 of Figure 6), wbdR (nucleotide position 13156 to 13821 of Figure 6), wzx (nucleotide position 2744 to 4135 of Figure 6) and wzy (nucleotide position 858 to 2042 of Figure 6) and fragments of those molecules of at least 10 25 12 nucleotides in length. Particularly preferred nucleic acid molecules are those set out in Tables 9 and 9A, with respect to the above mentioned genes. In one preferred embodiment the detection method is a polymerase chain reaction method. More preferably, the 30 nucleic acid molecules are labelled and hybridisation of the nucleic acid molecule is detected by electrophoresis. The present inventors believe that based on the teachings of the present invention and available information concerning 0 antigen gene clusters, and 35 through use of experimental analysis, comparison of nucleic acid sequences or predicted protein structures, nucleic acid molecules in accordance with the invention WO99/61458 PCT/AU99/00385 -9 can be readily derived for any particular 0 antigen of interest. Suitable bacterial strains can typically be acquired commercially from depositary institutions. There are currently 166 defined E. coli 0 antigens. 5 Samples of the 166 different E. coli 0 antigen serotypes are available from Statens Serum Institut, Copenhagen, Denmark. The inventors envisage rare circumstances whereby two genetically similar gene clusters encoding serologically 10 different 0 antigens have arisen through recombination of genes or mutation so as to generate polymorphic variants. In these circumstances multiple pairs of oligonucleotides may be selected to provide hybridisation to the specific combination of genes. The invention thus envisages the 15 use of a panel containing multiple nucleic acid molecules for use in the method of testing for 0 antigen in conjunction with H antigen, wherein the nucleic acid molecules are derived from genes encoding transferases and/or enzymes for the transport or processing of a 20 polysaccharide or oligosaccharide unit including wzx or wzy genes, wherein the panel of nucleic acid molecules is specific to a particular 0 antigen. The panel of nucleic acid molecules can include nucleic acid molecules derived from 0 antigen sugar pathway genes where necessary. 25 The inventors also found two mutated flagellin genes from H type strains for H35 and H54 which have insertion sequences inserted into normal flagellar genes identical or near identical to that that of the H11 and H21 type strains respectively. Thus, primers for H11 and H21 30 (listed in Table 3) would also amplify fragments in H35 and H54, which differ in sizes to those in H11 and H21 respectively. The inventors also provide two pairs of primers each for H35 and H54 based on the insertion sequence (see H35 and H54 columns in Table 3). The use of 35 one of them in combination with one of the H11 or H21 primers will generate a PCR band only in H35 or H54 respectively, and this will also differentiate H35 and H54 WO99/61458 PCT/AU99/00385 - 10 from H11 and H21 respectively. The present invention also relates to methods of detecting the presence of particular E. coli H antigens or H antigen and 0 antigen combinations where one or more 5 nucleic acid molecules which generate a particular size fragment indicative of the presence of that H antigen are used or in which the combination of one antigen specific primer for that H antigen with another primer for a related H antigen provides for the detection of the 10 particular H antigen by hybridisation to the relevant gene. Preferably, the H antigen is H11, H21, H35 or H54. The pairs of nucleic acid molecules where the method of the fifth aspect is used may both hybridise to the relevant H or 0 antigen gene or alternatively only one may 15 hybridise to the relevant gene and the other to another site. The inventors recognise in applying the methods of the invention for detecting combinations of 0 and H antigens to samples, that the methods do not indicate 20 whether a positive result for a particular 0 and H antigen combination arises because the 0 and H antigen are present on a single E. coli strain present in the sample or are present on different E. coli strains present in the sample. Because the ability to identify the presence of 25 E. coli strains with particular 0 and H antigen combinations is highly desirable (due to the relationship between particular combinations and pathogenicity) the determination that a particular combination is present in a sample can be followed by isolation of single colonies 30 and checking whether the they contain the relevant combination by using the same method again or using antibody labelled magnetic beads to separate cells expressing the particular 0 or H antigen and then testing the isolated cells for the other serotype. 35 In addition, as mentioned above, the present inventors have established the existence of H7 primers specific to the 0157 and 055 serotypes. Using such WO99/61458 PCT/AU99/00385 - 11 primers it is possible to detect particular 0 and H antigen combinations with the use of H specific nucleic acid molecules. In a sixth aspect the invention provides a method for 5 detecting the presence of a particular 0 serotype and H serotype of E. coli in a sample, the method comprising the following steps: (a) specifically hybridising at least one nucleic acid molecule, derived from and specific for a gene 10 encoding a flagellin associated with a particular E. coli H antigen serotype to any E. coli carrying the gene and present in the sample; and (b) detecting the at least one specifically 15 hybridised nucleic acid molecule, wherein the at least one nucleic acid molecule is specific for the particular combination of 0 and H antigen. Preferably the combination is 055:H7 or 0157:H7. The ability to detect the 0157:H7 combination from a 20 particular H7 primer or pair is of particular use given the association of this combination with pathogenic strains. In a seventh aspect the present invention provides a method for testing a food derived sample for the presence 25 of one or more particular E. coli 0 antigens and H antigens comprising testing the sample by a method of the fourth, fifth or sixth aspect the invention. In an eighth aspect the present invention provides a method for testing a faecal derived sample for the 30 presence of one or more particular E. coli 0 antigens and H antigens comprising testing the sample by a method of the fourth, fifth or sixth aspect the invention. In a ninth aspect the present invention provides a method for testing a patient or animal derived sample for 35 the presence of one or more particular E. coli 0 antigens and H antigens comprising testing the sample by a method of the fourth, fifth or sixth aspect the invention.

WO99/61458 PCT/AU99/00385 - 12 Preferably, the method of the seventh, eighth or ninth aspect of the invention is a polymerase chain reaction method. More preferably the oligonucleotide molecules for use in the method are labelled. Even more 5 preferably the hybridised nucleic acid molecules are detected by electrophoresis. In the above described methods it will be understood that where pairs of nucleic acid molecules are used one of the nucleic acid molecules may hybridise to a sequence 10 that is not from the 0 antigen transferase, wzx or wzy gene or the flagellin gene. Further where both hybridise to these genes the 0 antigen molecules may hybridise to the same or a different one of these genes. In a tenth aspect the present invention provides a kit 15 for identifying the H serotype of E. coli, the kit comprising: at least one nucleic acid molecule derived from and specific for an E. coli flagellin gene. In an eleventh aspect the present invention provides 20 a kit for identifying the H and O serotype of E. coli, the kit comprising: (a) at least one nucleic acid molecule derived from and specific for an E. coli flagellin gene; and (b) at least one nucleic acid molecule derived from 25 and specific for a gene encoding a transferase or a gene encoding an enzyme for the transport or processing of a polysaccharide or oligosaccharide unit, the gene being involved in the synthesis of a particular E. coli 0 antigen. 30 The nucleic acid molecules may be provided in the same or different vials. The kit may also provide in the same or separate vials a second set of specific nucleic acid molecules. Particularly preferred nucleic acid molecules for 35 inclusion in the kits are those specified in Tables 3, 8, 8A, 9 and 9A as described above.

WO99/61458 PCT/AU99/00385 - 13 DEFINITIONS In this specification, we have used term "flagellin gene" in many cases where previously one would have used "fliC", 5 to allow for the uncertainty as to locus introduced by recent observations. However, uncertainty as to the locus does not alter the fact that most E. coli strains express a single H antigen and that a single flagellin gene sequence per strain is required to give the genetic basis 10 for H antigen variation . Any use of the name fliC in this specification where a different locus is later shown to be involved would not affect the validity of conclusions drawn regarding application of information based on the sequence, where the conclusions do not relate to the map 15 position. Thus it is generally the nucleic acid molecule itself which is of importance rather than the name attributed to the gene. When it is known or suspected that the gene encoding the H antigen is not in the fliC locus, we use the term flagellin rather than fliC. 20 The phrase, "a nucleic acid molecule derived from a gene" means that the nucleic acid molecule has a nucleotide sequence which is either identical or substantially similar to all or part of the identified gene. Thus a nucleic acid molecule derived from a gene 25 can be a molecule which is isolated from the identified gene by physical separation from that gene, or a molecule which is artificially synthesised and has a nucleotide sequence which is either identical to or substantially similar to all or part of the identified gene. While some 30 workers consider only the DNA strand with the same sequence as the mRNA transcribed from the gene, here either strand is intended. Transferase genes are regions of nucleic acid which have a nucleotide sequence which encodes gene products 35 that transfer monomeric sugar units. Flippase or wzx genes are regions of nucleic acid which have a nucleotide sequence which encodes a gene WO99/61458 PCT/AU99/00385 - 14 product that flips oligosaccharide repeat units generally composed of three to six monomeric sugar units to the external surface of the membrane. Polymerase or wzy genes are regions of nucleic acid 5 which have a nucleotide sequence which encodes gene products that polymerise repeating oligosaccharide units generally composed of 3-6 monomeric sugar units. The nucleotide sequences provided in this specification are described as anti-sense sequences. This 10 term is used in the same manner as it is used in Glossary of Biochemistry and Molecular Biology Revised Edition, David M. Glick, 1997 Portland Press Ltd., London on page 11 where the term is described as referring to one of the two strands of double-stranded DNA usually that which has 15 the same sequence as the mRNA. We use it to describe this strand which has the same sequence as the mRNA. NOMENCLATURE Synonyms for E. coli 0111 rfb Current names Our names Bastin et al. 1991 20 wbdH orfl gmd orf2 wbdI orf3 orf3.4* manC orf4 rfbM* manB orf5 rfbK* 25 wbdJ orf6 orf6.7* wbdK orf7 orf7.7* wzx orf8 orf8.9 and rfbX* wzy orf9 wbdL orf10 30 wbdM orf11 * Nomenclature according to Bastin D.A., et al. 1991 "Molecular cloning and expression in Escherichia coli K-12 of the rfb gene cluster determining the 0 antigen of an E. coli 0111 strain". Mol. Microbiol. 5:9 2223-2231. 35 Other Synonyms wzy rfc wzx rfbX rmlA rfbA 40 rmlB rfbB rmlC rfbC rmlD rfbD glf orf6* wbbl orf3#, orf8* of E. coli K-12 WO99/61458 PCT/AU99/00385 - 15 wbbJ orf2#, orf9* of E. coli K-12 wbbK orfl#, orf10* of E. coli K-12 wbbL orf5#, orf 11* of E. coli K-12 # Nomenclature according to Yao, Z. And M. A. Valvano 1994. 5 "Genetic analysis of the O-specific lipopolysaccharide biosynthesis region (rfb) of Eschericia coli K-12 W3110: identification of genes the confer groups-specificty to Shigella flexineri serotypes Y and 4a". J. Bacteriol. 176: 4133-4143. * Nomenclature according to Stevenson et al. 1994. "Structure of 10 the O-antigen of E. coli K-12 and the sequence of its rfb gene cluster". J. Bacteriol 176: 4144-4156. * The O antigen genes of many species were given rfb names (rfbA etc) and the 0 antigen gene cluster was often referred to as the rfb cluster. There are now new names for the rfb genes as shown 15 in the table. Both terminologies have been used herein, depending on the source of the information. In the claims that follow and in the summary of the invention, except where the context requires otherwise due to express language or necessary implication, the word 20 "comprising" is used in the sense of "including", i.e. the features specified may be associated with further features in various embodiments of the invention. BRIEF DESCRIPTION OF THE DRAWINGS 25 Figure 1 shows Eco R1 restriction maps of cosmid clones pPR1054, pPR1055, pPR1056, pPR1058, pPR1287 which are subclones of E. coli 0111 0 antigen gene cluster. The thickened line is the region common to all clones. Broken lines show segments that are non-contiguous on the 30 chromosome. The deduced restriction map for E. coli strain M92 is shown above. Figure 2 shows a restriction mapping analysis of E. coli 0111 0 antigen gene cluster within the cosmid clone pPR1058. Restriction enzymes are: (B: BamHl; Bg: BglII, 35 E: EcoRl; H: HindIII; K: KpnI; P: PstI; S: Sall and X: Xhol. Plasmids pPR1230, pPR1231, and pPR1288 are deletion derivatives of pPR1058. Plasmids pPR 1237, pPR1238, pPR1239 and pPR1240 are in pUC19. Plasmids pPR1243, pPR1244, pPR1245, pPR1246 and pPR1248 are in 40 pUC18, and pPR1292 is in pUC19. Plasmid pPR1270 is in WO99/61458 PCT/AU99/00385 - 16 pT7T319U. Probes 1, 2 and 3 were isolated as internal fragments of pPR1246, pPR1243 and pPR1237 respectively. Dotted lines indicate that subclone DNA extends to the left of the map into attached vector. 5 Figure 3 shows the structure of E. coli 0111 0 antigen gene cluster. Figure 4 shows the structure of E. coli 0157 O antigen gene cluster. Figure 5 shows the nucleotide sequence (SEQ ID NO: 10 45) of the E. coli 0111 0 antigen gene cluster. Note: (1) The first and last three bases of a gene are underlined and of italic respectively.; (2) The region which was previously sequenced by Bastin and Reeves 1995 "Sequence and anlysis of the 0 antigen gene (rfb) cluster of 15 Escherichia coli 0111" Gene 164: 17-23 is marked. Figure 6 shows the nucleotide sequence (SEQ ID NO: 56) of the E. coli 0157 0 antigen gene cluster. Note: (1) The first and last three bases of a gene (region) are underlined and of italic respectively (2) The region 20 previously sequenced by Bilge et al. 1996 "Role of the Escherichia coli 0157-H7 0 side chain in adherence and analysis of an rfb locus". Inf. and Immun 64:4795-4801 is marked. Figures 7 to 9 show the nucleotide sequences (SEQ ID 25 NOS: 66 to 68 respectively) obtained for flagellin genes from E. coli type strains for H1 to H3 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 10 to 18 show the nucleotide sequences (SEQ 30 ID NOS: 6 to 14 respectively) obtained for flagellin genes from E. coli type strains for H4 to H12 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 19 and 20 show the nucleotide sequences (SEQ 35 ID NOS: 15 to 16 respectively) obtained for flagellin genes from E. coli type strains for H14 and H15 respectively. The primer positions listed in Table 3 are WO99/61458 PCT/AU99/00385 - 17 based on treating the first nucleotide of each of these sequences as No. 1. Figures 22 and 26 show the nucleotide sequences (SEQ ID NOS: 17 to 21 respectively) obtained for flagellin 5 genes from E. coli type strains for H17 and H21 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 27 to 39 show the nucleotide sequences (SEQ 10 ID NOS: 22 to 34) obtained for flagellin genes from E. coli type strains for H23 to H35 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 40 to 49 show the nucleotide sequences (SEQ 15 ID NOS: 35 to 44) obtained for flagellin genes from E. coli type strains for H37 to H46 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 50 to 55 show the nucleotide sequences (SEQ 20 ID NOS: 46 to 51) obtained for flagellin genes from E. coli type strains for H47 to H52 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 56 to 58 show the nucleotide sequences (SEQ 25 ID NOS: 52 to 54) obtained for flagellin genes from E. coli type strains for H54 to H56 respectively. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figure 59 shows the nucleotide sequence (SEQ ID NO: 30 55) obtained for the flagellin gene from E. coli H7 strain M1179. The primer positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figures 60 to 68 show the nucleotide sequences (SEQ 35 ID NOS: 57 to 65 respectively) obtained for flagellin genes from E. coli strains M1004, M1211, M1200, M1686, M1328, M917, M527, M973, and M918 respectively. The primer WO99/61458 PCT/AU99/00385 - 18 positions listed in Table 3 are based on treating the first nucleotide of each of these sequences as No. 1. Figure 69 shows the nucleotide sequence (SEQ ID NO: 1) of the fliC gene and DNA flanking the fliC gene from 5 the H25 type strain. Figure 70A shows the nucleotide sequence (SEQ ID NO: 2) obtained from the 5' end of the insert of plasmid pPR1989. The insert of plasmid pPR1989 encodes the second flagellin gene of the H55 type strain. 10 Figure 70B shows the nucleotide sequence (SEQ ID NO: 3) obtained from the 3' end of the insert of plasmid pPR1989. The insert of plasmid pPR1989 encodes the second flagellin gene of the H55 type strain. Figure 71 shows the nucleotide sequence (SEQ ID NO:4) 15 obtained from the 5' end of the insert of plasmid pPR1993. The insert of plasmid pPR1993 encodes the second flagellin gene of the H36 strain. Figure 72 shows the nucleotide sequence (SEQ ID NO:5) obtained from the 3' end of the insert of plasmid pPR1993. 20 The insert of plasmid pPR1993 encodes the second flagellin gene of the H36 type strain. Figure 73 A shows the sequence of polylinker and the SD sequence of plasmid pTrc99A. Figure 73B shows the sequence of the junction region 25 between the SD sequence and the start of flagellin gene in the plasmids used for the expression of flagellin genes. BEST METHOD OF CARRYING OUT THE INVENTION In carrying out the methods of the invention with 30 respect to the testing of particular sample types including samples from food, patients, animals and faeces the samples are prepared by routine techniques routinely used in the preparation of such samples for DNA based testing. The steps for testing the samples using 35 particular nucleic acid molecules in assay formats such as Southern blots and PCR are performed under routinely determined conditions appropriate to the sample and the WO99/61458 PCT/AU99/00385 - 19 nucleic acid molecules. H antigen Materials and Methods 5 1. Bacterial strains and plasmid: There are 54 H types in E. coli [Ewing, W.H.: Edwards and Ewing's identification of the Enterobacteriaceae., Elsevier Science Publishers, Amsterdam, The Netherlands, 1986]: note H antigens from 1 to 57 were listed and that 10 13, 22 and 57 are not valid. All the standard H type strains except H16 were obtained from the Institute of Medical and Veterinary Science, Adelaide, Australia. The primary stocks are hold at the Statens Serum Institut, Copenhagen, Denmark. 15 The additional H7 strains used are listed in Table 1. We do not have the type strain for H16. It is known that the H3 type strain is biphasic and can also express the H16 flagellin gene [Ratiner, Y. A. (1985) "Two genetic arrangements determining flagellar antigen specificities 20 in two diphasic E. coli strains. FEMS Microbiol Lett 19: 317-323]. We have sequenced and cloned the H16 flagellin gene from the H3 type strain (see below). E. coli K-12 strain C600 hsm hsr fliC::TnlO [Kuwajiwa, G. (1988) "Flagellin domain that affects H 25 antigenicity of E. coli K-12" J. Bacteriol. 170; 485-488] (laboratory stock no. M2126) was obtained from Dr Benita Westerlund-Wikstrom of the Department of Biosciences, University of Helsinkin, Finland. E. coli K-12 strain EJ2282 (laboratory no. P5560) is a fliC deletion strain, 30 and was obtained from Dr Masatoshi Enomoto of the Department of Biology, Okayama University, Japan [Tominaga, A. M. A.-H. Mahmound, T. Mokaihara and M. Enomoto (1994) "Molecular characterization of intact but cryptic, flagellin genes in the genus Shigella.: Mol. 35 Microbiol. 12: 277-285]. Plasmid pTrc99A was purchased from Pharmacia LKB (Melbourne, VIC, Australia).

WO99/61458 PCT/AU99/00385 - 20 2. Antisera Antisera against H1, H3, H8, H14, H15, H17, H23, H24, H25, H26, H29, H30, H31, H32, H33, H35, H36, H37, H38, 5 H39, H43, H44, H46, H47, H48, H49, H52, H53, H54, H55, and H56 were obtained from the Institute of Medical and veterinary Science, Adelaide, Australia. Antisera against H2, H4, H5, H6, H7, H9, H10, H11, H12, H16, H18, H19, H20, H21, H27, H28, H34, H40, H41, H42, H45, and H51 were 10 obtained from Denka Seiken Co., Ltd, Tokyo, Japan. Antisera to type H50 was not available from any known source. The antisera available were checked against the appropriate type strains to confirm the specificities of 15 both flagellin H antigen and H antisera: 52 sera (all those except anti-H16 serum listed above) gave a positive reaction with the corresponding type strains for that serum. 20 3. Agglutination test: Bacteria from 1 ml of an overnight culture grown in Luria broth (Difco Tryptone, 10g/l; Difico yeast extract, 5g/l; NaCl, 0.5 g/l; pH 7.2) at 30oC was centrifuged ( 4000 rpm/10 min) and the bacteria pellet resuspended in 25 100 ml of saline. The agglutination test was carried out by mixing equal volumes (5 ml) of both the cells and antiserum on a slide. The slide was rocked for 1 minute and then observed for agglutination. For all agglutination tests, saline containing no antiserum was mixed with cells 30 to be used as a negative control. For testing the H specificities of strain M2126 or strain P5560 carrying plasmid containing cloned flagellin genes, cells of M2126 or P5560 were used as an additional negative control. 35 All agglutination tests were first carried out using undiluted antisera (note that the antisera we used have been diluted before reaching our hands), except for anti- WO99/61458 PCT/AU99/00385 - 21 H11I, anti-H34, anti-H52 and anti-H26 serum for which we used 1:10 dilutions to avoid background agglutination. In cases for which cross-reactions have been reported, we carried out agglutination tests using serial dilutions of 5 sera (see section 10.1) 4. Motility test: The motility of strain M2126 or strain P5560 carrying cloned flagellin genes was examined microscopically. 1 ml 10 of overnight culture grown in Luria broth (Difco Tryptone, 10g/l; Difico yeast extract, 5g/l; NaCl, 0.5 g/l; pH 7.2) at 30oC was inoculated into 10 ml of Luria broth, and the culture was shaken at 100 rpm at 30oC to early log phase (OD 625 = 0.2). A loopful of culture was placed on a slide 15 and examined under a microscope. Motility of individual cells was easily distinguished from Brownian movement and streaming, and presence or absence of motility recorded. 5. Isolation of chromosomal DNA: 20 Chromosomal DNA from all the 53 H type strains and the strains listed in Table 1 was isolated using the Promega Genomic isolation kit (Madison WI USA). Each chromosomal DNA sample was checked by gel electrophoresis of the DNA and by PCR amplification of the mdh gene using 25 oligonucleotides based on the E. coli K-12 mdh gene [Boyd, E.F., Nelson, K., Wang, F.-S., Whittam, T.S. and Selander, R.K.: Molecular genetic basis of allelic polymorphism in malate dehydrogenase (mdh) in natural populations of Escherichia coli and Salmonella enterica. Proc. Natl. 30 Acad. Sci. USA 91 (1994) 1280-1284]. 6. PCR amplification of flagellin gene: Flagellin genes from different strains were first PCR amplified using one of the following four pairs of 35 oligonucleotides: #1285 (5'-atggcacaagtcattaatac) and #1286 (5'-ttaaccctgcagtagagaca); WO99/61458 PCT/AU99/00385 - 22 #1417 (5'-ctgatcactcaaaataatatcaac) and #1418 (5'-ctgcggtacctggttggc); #1431 (5'-atggcacaagtcattaatacccaac) and #1432 (5'-ctaaccctgcagcagagaca): 5 #1575 (5'-gggtggaaacccaatacg) and #1576(5'-gcgcatcaggcaatttgg) PCR reactions were carried out under the following conditions: denaturing, 94 0 C/30'; annealing, temperature varies (refer to Table 2)/30'; extension, 72 0 C/1'; 30 10 cycles. The PCR product was purified using the Promega Wizard PCR purification kit (Madison WI USA) before being sequenced. The H36 and H53 type strains gave two PCR bands using primer pairs #1431/#1432 and #1417/#1418 respectively, and 15 were not sequenced. 7. Enzymes and buffers: Restriction endonucleases and DNA T4 ligase were purchased from Boehringer Mannheim (Castle Hill, NSW, 20 Australia). Restriction enzymes were used in the recommended commercial buffer. 8. Sequencing of the flagellin genes: Each PCR product was first sequenced using the 25 oligonucleotide primers used for the PCR amplification. Primers based on the obtained sequence were then used to sequence further, and this procedure was repeated until the entire PCR product was sequenced. The sequencing reactions were performed using the 30 DyeDeoxy Terminator Cycle Sequencing method (Applied Biosystems, CA, USA), and reaction products were analysed using fluorescent dye and an ABI377 automated sequencer (CA, USA). Sequence data were processed and analysed using 35 Staden programs [Sacchi CT, Zanella R C, Caugant D A, Frasch C E, Hidalgo N T, Milagres L G, Pessoa L L, Ramos S R, Camargo M C C and Melles C E A "Emergence of a new WO99/61458 PCT/AU99/00385 - 23 clone of serogroup C Neisseria meningitidis in Sao Paulo, Brazil" J. Clin. Microbiol. 30 (1992) 1282-1286; Staden, R.: Automation of the computer handling of gel reading data produced by the shotgun method of DNA 5 sequencing. Nucl. Acids Res. 10 (1982a) 4731-4751; Staden, R.: An interactive graphics program for comparing and aligning nucleic acid and amino acid sequences. Nucl. Acids Res. 10 (1982b) 2951-2961; Staden, R.: Computer methods to locate signals in nucleic 10 acid sequences. Nucl. Acids Res. 12 (1984a) 505-519; Staden, R.: Graphic methods to determine the function of nucleic acid sequences. A summary of ANALYSEQ options. Nucl. Acids Res. 12 (1984b) 521-538; Staden, R.: The current status and portability of our 15 sequence handling software. Nucl. Acids Res. 14 (1986) 217-231]. We were able to PCR amplify flagellin genes from H type strains for H7, 23, 12, 51, 45, 49, 19, 9, 30, 32, 26, 41, 15, 20, 28, 46, 31, 14, 18, 6, 34, 48, 43, 10, 52, 20 and also from H7 strains m1004, m527, m1686, m1211, m1328, m973, m1179, m1200, m917, and m918 using primers #1575 and #1576 which are based on sequences 51-34 bp upstream and 37-54 bp downstream of start and end of the E. coli K-12 flick gene respectively. Thus, the full sequence of the 25 flagellin gene from these strains was obtained and the use of flanking sequence for primers makes it highly likely that they are at the fliC locus. For other strains, we were only able to amplify the flagellin gene using one or more of the other three pairs 30 of primers, which are based on sequence within the fliC gene, and thus only partial sequence was obtained. These amplicons may be of the fliC gene or one of the alternative flagellin genes. The flagellin gene sequences from H type strains for H40, 8, 21, 47, 11, 27, 35, 2, 3, 35 24, 37, 50, 4, 44, 38, 55, 29, 33, 5, and 56 obtained are lacking 18 and 14 codons at 5' and 3' ends respectively. The flagellin gene sequence of H39 obtained using primers WO99/61458 PCT/AU99/00385 - 24 #1285/#1286 lacks 18 and 19 codons at 5' and 3' ends respectively. The flagellin gene sequence of H type strains of H17, 25 and 42 lack 23 and 21 codons at 5' and 3' ends respectively. The flagellin gene sequence of the 5 H type strain for H54 lacks 23 and 12 codons at the 5' and 3' ends respectively. There is very little variation in the sequence at the two ends of flagellin genes and antigenic variation is due to variation in the central region of the gene. The absence of sequence for the ends 10 of some of the flagellin genes is not important for the purpose of the present invention relating to the detection of antigenic variation by DNA sequence based means. The fliC genes from H type strains of H1, H7 and H12 have been sequenced previously [Schoenhals, G. and 15 Whitfield, C.: Comparative analysis of flagellin sequences from Escherichia coli strains possessing serologically distinct flagellar filaments with a shared complex surface pattern. J. Bacteriol. 175 (1993) 5395-5402] and we did not sequence the gene from the HI strain. 20 We have sequenced fliC genes from a set of H7 strains with different 0 antigens, including that of fliC from the H7 type strain as one of the set: we have found four differences from the published H7 sequence (GenBank accession number L07388) which we believe are due to 25 errors in the published sequence. We have also re-sequenced the fliC gene from the H12 type strain, and have found one difference from the published H12 sequence (GenBank accession number L07389) which we believe is due to an error in the published 30 sequence. The flagellin genes from type strains H35 and H54 were also amplified using primers #1431/#1432, which are based on sequence within the fliC gene. Sequence data revealed that these two genes would be non-functional due 35 to insertion sequence inserted in the middle of them. We have sequenced them to facilitate selection of primers for the functional flagellin genes.

WO99/61458 PCT/AU99/00385 - 25 9. Cloning of flagellin genes DNA was digested for 2 hr at 37 0 C with appropriate restriction enzyme(s). The reaction product was then 5 extracted once with phenol, and twice with ether. DNA was precipitated with 2 vols of ethanol and resuspended in water before the ligation reaction was carried out. Ligation was carried out O/N at 4 0 C and the ligated DNA was electroporated into one of the E. coli fliC mutant 10 strains. 9.1. Cloning of flagellin genes from type strains for HI, H2, H3, H5, H6, H7, H9, H10, H11, H12, H14, H15, HI8, H19, H20, H21, H24, H26, H27, H28, H29, H31, H34, H38, H39, 15 H41, H42, H43, H45, H46, H49, H51, H52, and H56: The full flagellin gene was PCR amplified using primers #1868 and #1870 (Table 3A). Both these primers are based on the sequences of the H7 flagellin gene of the H7 type strain. #1868 is the 5' primer: there is an NcoI 20 site incorporated into the primer (Table 3B) and the flagellin gene starts at base 3 of the NcoI site. The 3' primer #1870 has a BamHI site incorporated downstream of the stop codon of the flagellin gene (Table 3B). PCR reactions were carried out under the following conditions: 25 denaturing, 94oC/30'; annealing, temperature varies (refer to Table 3A)/30'; extension, 72oC/1'; 30 cycles. The PCR product was purified using the Promega Wizard PCR purification kit (Madison WI USA) before being digested by restriction enzymes NcoI and BamHI and cloned into the 30 NcoI/BamHI sites of plasmid pTrc99A. Plasmid pTrc99A has a strong trc promoter upstream of the polylinker. Downstream of the promoter, it contains the ribosome binding site (SD sequence, see Fig 73) which is located 8bp upstream of the ATG site within the NcoI 35 site. The polylinker and the SD sequence of pTrc99A are shown in Fig 73. The plasmids generated were given pPR numbers, and WO99/61458 PCT/AU99/00385 - 26 they are listed in Table 3A. In these plasmids, the expression module consists of the trc promoter, the SD sequence, and the full flagellin gene. The sequence of the junction region between the SD sequence and the start of 5 flagellin gene is shown in Fig 73. For flagellin genes from type strains for H6, H7, H9, HI0, H12, H14, H18, H19, H20, H26, H28, H31, H41, H43, H45, H46, H49, H51, and H52, we have the full sequence for each gene and the primer sequences (#1868 and #1870) are 10 conserved among them. The cloned genes therefore have the same sequence as those from the type strains. For flagellin genes from type strains for H1, H15 and H34, we also have the full sequence. The previously published sequence of the flagellin gene from the H1 type 15 strain was extracted from GenBank (accession number L07387) and used. Primer #1868 is conserved in all three. But, primer #1870 has the third base of the fifth last codon in the HI sequence changed from A to G, and the third base of the second last codon changed from C to T in 20 the H15 and H34 sequences: these changes did not change the amino acid coded, so the cloned genes encode the same gene products as those of the type strains. For flagellin genes from type strains for H2, H3, H5, H11, H21, H24, H27, H29, H38, H39, H42, and H56, we do not 25 have the full sequences. In the plasmids carrying genes from these type strains, the expression module consists of the trc promoter, the SD sequence, and the full flagellin gene with the first and the last 21 base pairs being determined by the primer sequences which are based on the 30 H7 flagellin gene of the H7 type strain. The sequence of the junction region between the SD sequence and the start of flagellin gene is shown in Fig 73. 9.2. Cloning of the flagellin gene from type strain of 35 H23: The full flagellin gene was PCR amplified using primers #1868 and #1869 (Table 3A). #1868 is the 5' WO99/61458 PCT/AU99/00385 - 27 primer: there is an NcoI site incorporated into the primer (Table 3B) and the flagellin gene starts at base 3 of the NcoI site. The 3' primer #1869 has a SalI site incorporated downstream of the stop codon of the flagellin 5 gene (Table 3B). PCR reactions were carried out under the following conditions: denaturing, 94oC/30'; annealing, 55oC/30'; extension, 72oC/1'; 30 cycles. The PCR product was purified using the Promega Wizard PCR purification kit (Madison WI USA) before being digested by restriction 10 enzymes NcoI and SalI and cloned into the NcoI/SalI sites of plasmid pTrc99A to give plasmid pPR1942. Plasmid pTrc99A has a strong trc promoter upstream of the polylinker. Downstream of the promoter, it contains the ribosome binding site (SD sequence, see Fig 73) which 15 is located 8bp upstream of the ATG site within the NcoI site. The polylinker and the SD sequence of pTrc99A are shown in Fig 73. The expression module of pPR1942 consists of the trc promoter, the SD sequence, and the full flagellin gene. 20 The sequence of the junction region between the SD sequence and the start of flagellin gene is shown in Fig 73. 9.3. Cloning of flagellin genes from type strains of H30, 25 H32 and H33: The full flagellin gene was PCR amplified using primers #1868 and #1871 (Table 3A). #1868 is the 5' primer: there is an NcoI site incorporated into the primer (Table 3B) and the flagellin gene starts at base 3 of the 30 NcoI site. The 3' primer #1871 has a PstI site incorporated downstream of the stop codon of the flagellin gene (Table 3B). PCR reactions were carried out under the following conditions: denaturing, 94oC/30'; annealing, temperature varies (refer to Table 3A)/30'; extension, 35 72oC/i'; 30 cycles. The PCR product was purified using the Promega Wizard PCR purification kit (Madison WI USA) before being digested by restriction enzymes NcoI and PstI WO99/61458 PCT/AU99/00385 - 28 and cloned into the NcoI/PstI sites of plasmid pTrc99A. Plasmid pTrc99A has a strong trc promoter upstream of the polylinker. Downstream of the promoter, it contains the ribosome binding site (SD sequence, see Fig 73) which 5 is located 8bp upstream of the ATG site within the NcoI site. The polylinker and the SD sequence of pTrc99A are shown in Fig 73. For flagellin genes from type strains for H30 and H32, we have the full sequence. Primer #1868 sequence is 10 conserved in both of them. But, primer #1871 has the third base of the fourth last codon in both sequences changed from G to A to remove a PstI site (see Table 3B): this change did not change the amino acid coded. The expression module consists of the trc promoter, the SD sequence, and 15 the full flagellin gene coding for a gene product which is same as that of the type strain. The sequence of the junction region between the SD sequence and the start of flagellin gene is shown in Fig 73. We do not have the full sequence for the flagellin 20 gene from the H33 type strain. In the plasmid containing the H33 type strain gene, the expression module consists of the trc promoter, the SD sequence, and the full flagellin gene with the first and the last 21 base pairs been determined by the primer sequences which were used 25 for the cloning of H30 and H32. The sequence of the junction region between the SD and the start of flagellin gene is shown in Fig 73. 9.4. Flagellin genes from type strains for H4 and H17: 30 For the flagellin genes of H4 and H17 type strains the full sequence was not obtained, and the sequenced parts were PCR amplified and cloned into plasmid pPR1951 to give in each case a gene in which the first 26 and the last 31 codons are based on the sequence of the H7 35 flagellin gene of the H7 type strain. 9.4.1 Construction of expression plasmid vector WO99/61458 PCT/AU99/00385 - 29 pPR1951: The first 26 codons of the H7 flagellin gene was first PCR amplified using primers #1868 and #1872 (Table 3B). #1868 is the 5' primer: there is an NcoI site 5 incorporated into the primer (Table 3B) and the flagellin gene starts at base 3 of the NcoI site. Primer #1872 was made to have the last two codons (codons 25 and 26) changed from CTG TCG (Leucine and Serine) to GGA TCC (Glycine and Serine) to generate a BamHI site. This PCR 10 fragment was digested with NcoI and BamHI before being cloned into the NcoI/BamHI sites of pTrc99A to make plasmid pPR1949. The last 31 codons (including the stop codon) of the H7 flagellin gene was PCR amplified using primers #1884 15 and #1871 (Table 3A). The 5' primer, #1884, has the first two of the 31 codons changed from TCG AAA (Serine and Lysine) to TCT AGA (Serine and Arginine) to generate a XbaI site (Table 3B). The 3' primer #1871 has a PstI site incorporated downstream of the stop codon (Table 3B). This 20 PCR fragment was digested with XbaI and PstI, and then cloned into the XbaI/PstI sites of pPR1949 to make plasmid pPR1951. 9.4.2 Cloning of flagellin genes from the H4 and 25 HI7 type strains: The central regions of flagellin genes from type strains H4 and H17 were PCR amplified using primers #1878 and #1885 (Table 3B), which have a BamHI and a XbaI incorporated at their ends respectively. PCR reactions 30 were carried out under the following conditions: denaturing, 94oC/30'; annealing, 65oC/30'; extension, 720oC/1'; 30 cycles. The PCR product was purified using the Promega Wizard PCR purification kit (Madison WI USA) before being digested by restriction enzymes BamHI and 35 XbaI and cloned into the XbaI/BamHI sites of plasmid pPR1951 to make plasmids pPR1955 (H4) and pPR1957 (H17). The expression module of plasmids pPR1955 and pPR1957 WO99/61458 PCT/AU99/00385 - 30 consists of the trc promoter, the SD sequence, the first 24 codons of the H7 flagellin gene (of the H7 type strain), 2 codons encoding Glycine and Serine, 292 or 293 codons of the central region based on the flagellin gene 5 obtained from the H4 or H17 type strain respectively, 2 codons encoding Serine and Arginine, and then the last 29 codons of the H7 flagellin gene (of the H7 type strain). 10. Expression of flagellin gene plasmids in E. coli 10 strains lacking the fliC gene, and identification of the H antigens encoded by these plasmids: Plasmids carrying flagellin genes as described in section 9 (see Table 3A for a list) were electroporated into strains M2126 or P5560. Strains M2126 and P5560 do 15 not have functional fliC genes, and are not motile when examined under a microscope. Transformants carrying any of the plasmids listed in Table 3A are motile when examined under a microscope. Thus, the flagellin genes in all of the plasmids are expressed. 20 The antigenic specificity of the flagellin of each transformant was then determined by slide agglutination. 10.1 Flagellin genes from type strains for H2, H5, H6, H7, H9, H11, H14, H15, H18, H19, H20, H21, H23, 25 H24, H26, H27, H28, H29, H30, H31, H32, H33, H34, H39, H41, H42, H43, H45, H46, H49, H51, H52, and H56: As shown in Table 3A, strains with plasmids carrying these flagellin genes expressed the same H antigen as their respective flagellin parent strain. 30 For flagellin specificities H2, H5, H6, H7, H9, H14, H15, H18, H19, H20, H23, H24, H26, H27, H28, H29, H31, H33, H39, H51, H52, and H56, there was no cross reaction reported between these flagellins and flagellin antisera for other H antigens [Ewing, W. H.: Edwards and Ewing's 35 identification of the Enterobacteriaceae., Elsevier Science Publishers, Amsterdam, The Netherlands, 1986], and we conclude that we have in each case sequenced the gene WO99/61458 PCT/AU99/00385 - 31 encoding the flagellin of the expected specificity from the respective type strain. It has been observed that cross reactions exist between some type strains and certain antisera at 5 different levels of dilution (of the antisera), being H11 with anti-H21 and anti-H40, H21 with anti-Hll, H30 with anti-H32, H32 with anti-H30, H34 with anti-H24 and anti H31, H41 with anti-H37 and anti-H39, H42 with anti-H6, H43 with anti-H37, H45 with anti-H20, H46 with anti-H17, and 10 H49 with anti-H39 [Ewing, W. H.: Edwards and Ewing's identification of the Enterobacteriaceae., Elsevier Science Publishers, Amsterdam, The Netherlands, 1986]. We have tested strain M2126 or strain P5560 carrying plasmids containing flagellin genes obtained from each of these 15 type strains (H11, H21, H30, H32, H34, H41, H42, H43, H45, H46, and H49) with the appropriate cross-reacting antisera. For strain M2126 or strain P5560 carrying plasmids containing flagellin genes obtained from type strains H11, 20 H34, H41, H42, H43, H45, H46, and H49, no cross reaction was found. We conclude that we have in each case sequenced the gene coding the flagellin of the expected specificity from the respective type strain. Cross reaction was observed for strain P5560 carrying 25 plasmid pPR1948 (containing the flagellin gene obtained from the H30 type strain) with anti-H32 serum, strain P5560 carrying pPR1940 (containing the flagellin gene obtained from the H32 type strain) with anti-H30 serum , and strain M2126 carrying plasmid pPR1995 (containing the 30 flagellin gene obtained from the H21 type strain) with anti-H11 serum. We note that the reported cross reactions between the H30 type strain and anti-H32, the H32 type strain and anti-H30, and the H21 type strain and anti-H11 happened at 35 a higher level of dilution (of antisera) than for all other type strains with the antisera mentioned above [Ewing, W. H.: Edwards and Ewing's identification of the WO99/61458 PCT/AU99/00385 - 32 Enterobacteriaceae., Elsevier Science Publishers, Amsterdam, The Netherlands, 1986]. We conclude that except for these three cases, the antiserum used were supplied at a dilution which did not exhibit cross reactions. For the 5 three strains carrying flagellin genes cloned form type strains for H30, H32, and H21, it was necessary to further dilute the antiserum. Strain P5560 carrying plasmid pPR1948 (containing the flagellin gene obtained from the H30 type strain) 10 agglutinates with anti-H30 serum when the antiserum is diluted to 1:60, but agglutinates with anti-H32 serum only at a dilution of 1:10 and not at a 1:20 dilution (note that the antisera we used have been diluted before reaching our hands). In contrast, strain P5560 carrying 15 plasmid pPR1940 (containing flagellin gene obtained from the H32 type strain) agglutinates with anti-H32 serum when the antiserum is diluted at 1:100, but agglutinates with anti-H30 serum only at a 1:10 dilution and not at a 1:10 dilution. Thus, we conclude that the flagellin genes we 20 sequenced from type strains for H30 and H32 encode flagellins of H30 and H32 specificities respectively. Strain M2126 carrying plasmid pPR1995 (containing the flagellin gene obtained from the H21 type strain) agglutinates with anti-H21 serum when the antiserum is 25 diluted to 1:40, but agglutinates only with undiluted anti-H11 serum and not at a 1:10 dilution (note that the antisera we used have been diluted before reaching our hands). In contrast, strain M2126 carrying plasmid pPR1981 (containing flagellin gene obtained from the H11 type 30 strain) did not agglutinate with anti-H21 serum. Thus, we conclude that the flagellin genes we sequenced from type strains for H21 encodes flagellin of H21 specificity. 10.2 Flagellin genes from type strains of HI and 35 H12: These two genes are very similar in sequence, with 8 a.a difference between the gene products. It has been WO99/61458 PCT/AU99/00385 - 33 known that some cross-reaction exists between anti-H12 serum and the HI type strain and between anti-H1 serum and the H12 type strain [Ewing, W. H.: Edwards and Ewing's identification of the Enterobacteriaceae., Elsevier 5 Science Publishers, Amsterdam, The Netherlands, 19861. Strain M2126 carrying pPR1920 (carrying a flagellin gene from the HI type strain, Table 3A) agglutinates with anti H1 serum when the antiserum is diluted to 1:100, but agglutinates only with undiluted anti-H12 serum and not at 10 a i:10 dilution (please note that the antisera we used have been diluted before reaching our hands). In contrast, strain M2126 carrying plasmid pPR1990 (carrying a flagellin gene from the H12 type strain, Table 3A) agglutinates with anti-H12 serum when the antiserum is 15 diluted at i:100, but agglutinates only with undiluted anti-Hi serum and not at a i:10 dilution. Thus, we conclude that the flagellin genes we sequenced from type strains for H1 and H12 encode flagellins of HI and H12 specificities respectively. 20 10.3. Flagellin gene coding for H16: Strain P5560 carrying plasmid pPR1969 agglutinated with anti-H16 serum. pPR1969 carries a flagellin gene amplified from the H3 type strain. It has been shown that 25 this H3 type strain is a biphasic strain which can express H3 and H16 specificities [Ratiner, Y. A. (1985) "Two genetic arrangements determining flagellar antigen specificities in two diphasic E. coli strains. FEMS Microbiol Lett 19: 317-323]. Thus, the H3 type strain has 30 two flagellin genes coding for H3 and H16 specificities. We conclude that we have cloned and sequenced the H16 flagellin gene from this H3 type strain. 10.4 Flagellin gene coding for H4: 35 The flagellin genes obtained from type strains for H4 and H17 are nearly identical, with 4 a.a. difference in the gene products. Plasmid pPR1955 carries a flagellin WO99/61458 PCT/AU99/00385 - 34 gene from the H4 type strain, and plasmid pPR1957 carries a flagellin gene from the H17 type strain. Strain P5560 carrying plasmid pPR1955 or plasmid pPR1957 agglutinated with anti-H4 serum, but not with anti-H17 serum. It has 5 been shown that the type strain for H17 is a biphasic strain which can express H17 and H4 [Ratiner, Y. A. (1985) "Two genetic arrangements determining flagellar antigen specificities in two diphasic E. coli strains. FEMS Microbiol Lett 19: 317-323]. The flagellin gene obtained 10 from type strain for H44 is also highly similar to that obtained from the H4 type strain, with 2 a.a. difference in the gene products. It has been shown that the H44 type strain has two complete flagellin genes, being H4 and H44 [Ratiner, Y. A. (1998) "New flagellin specifying genes in 15 some E. coli strains" J. Bacteriol 180: 979-984]. Thus, we conclude that all the three flagellin genes (obtained from type strains for H4, H17 and H44, and sequenced) encode the H4 flagellin, and that the flagellin genes for H17 and H44 specificities have not yet been cloned. 20 10.5 Flagellin gene coding for H10: The flagellin genes obtained from type strains for HI0 and H50 are nearly identical, with 3 a.a. difference in the gene products. Strain P5560 carrying plasmid 25 pPR1923 (which carries a flagellin gene from the H10 type strain) agglutinated with anti-H10 serum. We conclude that the sequence obtained from the H10 type strain encodes the H10 flagellin. It is not clear if the sequence obtained from the H50 type strain encodes HI0 or H50 (see below 30 section for H50). 10.6 Flagellin gene coding for H38: The flagellin genes obtained from type strains for H38 and H55 are nearly identical, with only 1 a.a. 35 difference in the gene products. Strain M2126 carrying plasmid pPR1984 (carrying the flagellin gene from the type strain H38) agglutinated with anti-H38 serum, but not with WO99/61458 PCT/AU99/00385 - 35 anti-H55 serum. It also has been shown that the type strain for H55 has two complete flagellin genes coding for H55 and H38 specificities [Ratiner, Y. A. (1998) "New flagellin specifying genes in some E. coli strains" J. 5 Bacteriol 180: 979-984]. Thus, we conclude that both cloned genes encode the H38 flagellin. 10.7 Summary: Flagellin genes coding for 39 H antigens have been 10 identified, being those for specificities H1, H2, H4, H5, H6, H7, H9, H10, H11, H12, H14, H15, H16, H18, H19, H20, H21, H23, H24, H26, H27, H28, H29, H30, H31, H32, H33, H34, H38, H39, H41, H42, H43, H45, H46, H49, H51, H52, and H56. 15 11. Comparison and alignment of the flagellin genes: Programs Pileup [Devereux, J., Haeberli, P. and Smithies, 0.: A comprehensive set of sequence analysis programs for the VAX. Nucl. Acids Res. 12 (1984) 387 20 395]and Multicomp [Reeves, P.R., Farnell, L. and Lan, R.: MULTICOMP: a program for preparing sequence data for phylogenetic analysis. CABIOS 10 (1994) 281-284] were used. The previously published sequence of H1 (GenBank 25 accession number L07387) was extracted from GenBank and used. Because we did not sequence H36 and H53 flagellin genes and we did not have the H16 type strain, we only compared 51 flagellin genes of H type strains and the fliC genes from the additional 10 H7 strains. 30 Among the H7 fliC genes, the percentage of DNA difference ranged from 0.0 to 2.39%. The flagellin genes from type strains for H40 and H8 are identical. Some others are nearly identical: H21 and H47 (1.5% difference), H12 and HI (2.6% difference), H10 and H50 35 (0.3% difference), H38 and H55 (0.1% difference), H4, H44 and H17 are very similar, the pairwise difference ranging from 0.33% to 0.87%.

WO99/61458 PCT/AU99/00385 - 36 For the flagellin genes obtained from type strains for H4, H17 and H44, we have shown that all the three genes encode flagellin with the H4 specificity (see above). For the flagellin genes obtained from type 5 strains fro H21 and H47, and H38 and H55, we have confirmed the specificities for one for each pair and have good reason to conclude that both genes of each pair encode the same H specificity (see above section), being that for H21 and H38 specificities respectively. 10 For the flagellin genes obtained from type strains for H10 and H50, we have confirmed that the one from the H10 type strain encodes H10 specificity. As these two genes are highly similar, we have presumed that they both encode H10 specificity. 15 In the cases where the flagellin gene from two type strains is near identical, we conclude that both genes code for flagellin of the same H specificity and that one or other strain has an additional locus which carries the functional gene, although the flagellin genes sequenced 20 do not appear to be mutated. We have shown by cloning and expression that the flagellin genes obtained from the H1 and H12 type strains encode H1 and H12 specificities respectively (see above section). The neucleotide difference between these two 25 genes is higher at 2.6% (see above), but still within the normal range for variation within a gene in E. coli. The two antigens cross react, and this cross reaction must be due to the high level similarity of the flagellins encoded by these two genes. 30 As discussed above, genes encoding some H antigens have been shown to be located at loci other than fliC. H3, H36, H47, H53 have been shown to be at a locus called flkA, H44 and H55 at fllA, and H54 at flmA [Ratiner Y A (1998)"New flagellin-specifying genes in some Escherichia 35 coli strains" J. Bacteriol. 180 979-984]. However, these strains may carry a fliC in addition to flkA, fllA or flmA [Ratiner Y A (1998)"New flagellin-specifying genes in some WO99/61458 PCT/AU99/00385 - 37 Escherichia coli strains" J. Bacteriol. 180 979-984]. The flagellin gene encoding H48 was previously sequenced from E. coli strain K-12 [Kuwajima G, Asaka J, Fujiwara T, Node K and Kondo E "Nucleotide sequence of the 5 hag gene encoding flagellin of Escherichia coli" J Bacteriol. 168 (1986) 1479-1483]. We have sequenced the fliC gene from the H48 type strain, and found that it is identical to that from K-12. The H54 gene is known to be at flmA [Ratiner Y A 10 (1998)"New flagellin-specifying genes in some Escherichia coli strains" J. Bacteriol. 180 979-984] and the finding of a non-functional presumptive fliC locus in the H54 strain shows that it is present but not expressed. However, we have not amplified and sequenced 15 the functional flmA gene of this strain. Using the 43 unique sequences (being the 39 identified genes with confirmed specificities and the flagellin genes obtained from the H8 (or H40), H25, H37, and H48 type strains) and the sequences from the two non 20 functional flagellin genes (from H type strains H35 and H54) (see Table 3) we have been able to determine antigen specific primers for each of the H antigen specificities and thereby show that it is practicable to detect E.coli strains carrying specific H antigens without false 25 positives from strains of other H types. There is no reason to expect that the addition of 11 sequences to the 43 unique sequences obtained will affect the general conclusion, as unlike previous reports, our study covers flagellin sequences for a substantial majority of known E. 30 coli H antigen specificities. Our study of 11 H7 genes from strains of eight different 0 antigens shows limited variation which was such that the variation within genes for H antigens does not affect the ability to select antigen specific primers. 35 O:H combinations in general define a strain and as some of the strains thus defined were quite distant from each other in a study by Whittam [Whittam T S, wolfe M L, WO99/61458 PCT/AU99/00385 - 38 Wachsmuth I K, Orskov I and Wilson R A "Clonal relationships among Escherichia coli strains that cause hemorrhagic colitis and infantile diarrhea" Infect. Immun. 61 (1993) 1619-1629] the variation we observe is thought 5 to represent that generally present in H7 genes. We also obtained more than one sequences for flagellin genes for H specificities H4, HI0, and H38, and again the level of variation within a given specifities is very low. However, there is a low possibility that primers chosen 10 without knowledge of the variation within genes of each H specificity could fail to give positive results with some isolates due to chance choice of primers which cover a base or bases which contribute to this low level variation. The variation within the H7 genes is in the 15 normal range for variation within a gene in E. coli and if this possibility did occur it would be easy to use an alternate primer pair. For example, if a first primer in a primer pair is unable to hybridise to a target region because of low level variation in that region, a positive 20 result may be achieved by using a second primer in that pair together with a third primer, whether or not the third primer is specific for the flagellin gene. Where the third primer is not specific for the flagellin gene, the specificity of the primer pair derives from the 25 specificity of the second primer. The observation that the overall level of variation within gene for a given H specificity is very low making it extremely unlikely that the regions covered by the two primers specific for H specificity would both have undergone change in the same 30 strain. There are 54 known H antigens for E. coli and of these there are 11 H antigen specificities for which we do not as yet have sequence. It will be easy to determine these sequences and determine primer pairs specific for 35 these H antigens by comparing these sequences with the 45 obtained sequences (see Table 3), and also modify the primers selected for any H antigen for which we already WO99/61458 PCT/AU99/00385 - 39 know the sequence in the unlikely event that there is a possibility of false positives with the primers selected. The sequences for the remaining H antigens can be obtained in one of the following ways: 5 1. where we have two bands by PCR (H36 and H53 type strains), we purify each and sequence, and also clone each into a strain mutated in its fliC gene and determine the H antigen expressed by use of specific sera. In this way a 10 specific sequence can be related to an H antigen specificity. The other band which represents an H antigen gene for a different specificity is expected to include a mutant gene or a gene similar to one of those for a known H specificity, but if not may represent a new specificity 15 for which primer pairs could be selected. It may be difficult to obtain expression of flagellin genes when cloned from E. coli due to cloning together with regulatory sequences which prevent expression. This is easily avoided by cloning the major segment of the gene 20 into a functioning fliC gene to replace the equivalent segment of that gene, using standard site directed mutagenesis to give suitable restriction sites within the cloned gene and incorporating those restriction sites into primers used to amplify the major segment of the gene to 25 be studied to facilitate the cloning. We have cloned and sequenced the PCR bands from the H36 and the H55 type strains using this method (see section 16). 2. Where two or more strains have the same flagellin 30 gene sequence, the genes are cloned as above and the H antigen specificity represented by this sequence is determined. This identifies the strain in which the expected gene is expressed and also those strains for which we have sequenced a gene which is not being 35 expressed. We then clone the gene for the antigen expressed in these strains by making a bank of plasmid clones using chromosomal DNA and select for a clone which WO99/61458 PCT/AU99/00385 - 40 is expressing an H antigen different from the one represented by the known sequence. This can be done by taking advantage of the fact that the H antigen is on flagellin, the protein of the bacterial flagellum used for 5 movement of the bacteria. In the presence of antibodies specific to that flagellum the bacteria cannot swim. For selection the clones are placed in a situation in which motile cells can swim away from the others and be collected. There are many versions of these techniques 10 and any could be used. One version is to place the bacteria on a nutrient agar plate with reduced agar content such that bacteria can swim away from the site of inoculation. This is easily seen as growth on the plate and a sample of the bacteria which are motile can be 15 recovered and cultivated. In this way bacteria carrying cloned H antigen genes can be selected. If the medium in the plate has antibody added to it only bacteria which express an H antigen different to that recognised by the antiserum will be able to swim. Specifically if the 20 antiserum used is specific for the H antigen expressed by the gene for which we have sequence, only clones which express a different H antigen, such as those expressing the H antigen expressed by the H type strains used to make the plasmid, will be selected. Once the clone is 25 obtained, the H antigen gene can be sequenced. Our work has shown that there are at least 7 cases where the H antigen type strains carry two H antigen genes which appear to be complete and have the potential to 30 function. However, while E. coli does not (in general) have a capacity to express more than one flagellin gene, it is striking that there are several loci for flagellin genes [Ratiner Y A (1998)"New flagellin-specifying genes in some Escherichia coli strains" J. Bacteriol. 180 979 35 984]. Several of the pairs of H type strains with identical or near identical sequence do not include any of the H antigen types shown by Ratiner [Ratiner Y A WO99/61458 PCT/AU99/00385 - 41 (1998)"New flagellin-specifying genes in some Escherichia coli strains" J. Bacteriol. 180 979-984] to map other than at fliC although these predominate. This suggests that there are additional cases where the expressed gene is not 5 the only flagellin gene present. However the fact that many of the cases where we obtained flagellin genes of identical or near identical sequence and/or two flagellin genes from one strain involve type strains found by Ratiner [Ratiner Y A (1998)"New flagellin-specifying genes 10 in some Escherichia coli strains" J. Bacteriol. 180 979 984] to map away from fliC are among those near identical to others, indicates that the phenomenon is of limited extent. Nonetheless it remains possible even where only one gene has been obtained by PCR, that it is one of a 15 pair of flagellin genes, the other not being amplified by the primers used, and further that it is the one not amplified which is expressing the H antigen of the strain. It will therefore be necessary to clone as described above each of the flagellin genes we have sequenced and 20 confirm that it expresses the expected antigen to ensure that the invention give results corresponding to those of the traditional serotyping scheme. In the event that it does not, the gene for the type antigen can be cloned and sequenced by the means described above. 25 The 11 H7 fliC sequences fell into three groups, one comprising the genes from the 0157:H7 and 055:H7 strains, which were identical, as expected given the proposed relationship between the clones. It has been shown that E. coli 0157:H7 and 055:H7 clones are closely related 30 [Whittam T S, wolfe M L, Wachsmuth I K, Orskov I and Wilson R A "Clonal relationships among Escherichia coli strains that cause hemorrhagic colitis and infantile diarrhea" Infect. Immun. 61 (1993) 1619-1629] thus it was expected that the H7 fliC genes from 0157 and 055 would be 35 identical. Among the H7 fliC sequences, we can identify primers specific to the H7 fliC gene for each of the three H7 groups. Two of these primers in combination with an H7 WO99/61458 PCT/AU99/00385 - 42 specific primer gave two primer pairs specific for the H7 gene of from the 0157:H7 and 055:H7 clones. 13. Specific oligonucleotide primers for each of the 43 5 flagellin genes Two oligonucleotide primers were chosen based on each of the 43 sequences. None of them had more than 85% identity with any other of 61 flagellin gene sequences. Thus, these primers are specific for each H type. These 10 primers are listed in Table 3. The flagellin gene of the H54 type strain is a mutated gene. It has an insertion sequence (IS1222) inserted into a normal flagellin gene of H21. Thus, primers for H21 would amplify a fragment of different size 15 in H54. We also provide 2 primers based on the insertion sequence (see H54 row in Table 3), and the use of one of them in combination with one of the H21 primers will generate a PCR band only in H54, which will also differentiate those strain carrying the mutated H21 gene 20 from those expressing the H21 flagellin gene. The flic gene of H35 type strain is also a mutated gene. It has an insertion sequence (IS1) inserted into a normal flagellin gene of H11. Thus, primers for H11 would amplify a fragment of different size in H35. We also 25 provide 2 primers based on the insertion sequence (see H35 row in Table 3), and the use of one of them in combination with one of the H11 primers will generate a PCR band only in H35, which will also differentiate those strain carrying the mutated H11 gene from those expressing the 30 H11 flagellin gene. 14. Testing of the H7 specific oligonucleotide primers Primer pair #1806/#1809 (see Table 3) was used to carry out PCR on chromosomal DNA samples of all the 54 H 35 type strains and the H7 strains listed in Table 1. PCR reactions were carried out under the following conditions: denaturing, 94 0 C/30'; annealing, 580C/30'; extension, WO99/61458 PCT/AU99/00385 - 43 720C/1'; 30 cycles. PCR reaction was carried out in an volume of 50ul for each of the chromosomal sample. After the PCR reaction, 5~l PCR product from each sample was run on an agarose gel to check for amplified DNA. 5 Primer pairs #1806/#1809 produced a band of predicted size with all the 11 strains expressing H7, but gave no band with other H type strains. Thus, these primers are H7 specific. 10 15. Testing of oligonucleotide primers specific to H7 of 0157 and 055: Based on a comparison of the fliC sequences of 11 different H7 strains, we have identified two oligonucleotides [#1696 (5'-GGCCTGACTCAGGCGGCC) at 15 positions 178 to 195 in M527 and #1697 (5' GAGTTACCGGCCTGCTGA) positions 1700-1683 in M5271 which are unique to H7 of 0157 and 055. Although not identical to any parts of the fliC sequences of any other H7 strains, these two primers are identical or have high level 20 similarity to fliC genes of some other H types. However a combination of one of these primers with one of the H7 specific primers can give specificity for H7 of 0157:H7 and 055:H7 E. coli. Primer pairs #1696/#1809 and #1697/#1806 were used to 25 carry out PCR on chromosomal DNA samples of all the H type strains and the H7 strains listed in Table 1. PCR reactions were carried out under the following conditions: denaturing, 940C/30'; annealing, 610C/30 ' (for #1696/#1809) or 600C/30'(for#1697/#1806); extension, 30 720C/1'; 30 cycles. PCR reaction was carried out in an volume of 50gi for each of the chromosomal samples. After the PCR reaction, 5gl PCR product from each sample was run on an agarose gel to check for amplified DNA. Both primer pairs produced a band of predicted size 35 with both of the 0157:H7 strains (strains M1004 and M527, see Table 1), and the 055:H7 strain (strain M1686, see Table 1), but gave no band with other strains. Thus, these WO99/61458 PCT/AU99/00385 - 44 two pairs of primers are specific to H7 genes of 0157:H7 and 055:H7 E. coli strains. 16. Identification of flagellin genes for the remaining 15 5 H specificities. 16.1. Sequencing the potential flkA gene coding for the H36 flagellin: Using primers #1431 (5'- atg gca caa gtc att aat acc 10 caa c) and #1432 (5'- cta acc ctg cag cag aga ca), we have amplified two bands from the H36 type strain. PCR reaction was carried out under the following conditions: denaturing, 94oC/30'; annealing, 57oC/30'; extension, 72oC/1'; 30 cycles. These two PCR fragments were then 15 cloned into the pGEM-T vector using the Promega pGEM-T cloning kit (Madison WI USA) to make plasmids pPR1992 and pPR1993. Inserts from both plasmids were first sequenced using the M13 universal primers (which bind to the pGEM-T DNA flanking the insertion site). For pPR1992, primers 20 based on the sequence obtained were then used to sequence further, and this procedure was repeated until the insert was fully sequenced. The sequence of the insert of pPR1992 is identical to that of the H12 flagellin gene sequence except perhaps for 25 the first 8 and last 7 codons which are encoded by the PCR primers in plasmid pPR1992. We have only sequenced the two ends of the insert of plasmid pPR1993 (Figures 71 and 72), and the sequences of the two ends of the insert of pPR1993 are very similar to ends of other sequenced 30 flagellin genes. We conclude that the insert of plasmid pPR1993 encodes a flagellin gene. The full sequence of the insert of plasmid pPR1993 can be obtained using the same method as for the sequencing of the insert of plasmid pPR1992. It is known that flkA gene encodes the H36 35 flagellin [Ratiner, Y. A. (1998) "New flagellin specifying genes in some E. coli strains" J. Bacteriol 180: 979-984], and it is highly likely that plasmid pPR1993 contains the WO99/61458 PCT/AU99/00385 - 45 flkA gene of the H36 type strain. H specificities can be confirmed by slide agglutination. The currently uncharacterised sequence of both ends and of DNA flanking these two sequenced genes can be 5 obtained by PCR walking and sequencing. Methods for PCR walking from a known sequence to an unknown region in chromosomal DNA are available (see [Siebert, P. D. , A. Chenchi, D. E. Kellogg, A. Lukyanov and S. A. Lukyanov (1995) "An improved PCR method for walking in uncloned 10 genomic DNA." Nuc. Acids Res. 23: 1087-1088]). The sequenced genes then can be PCR amplified and cloned using the method(s) described in section 9. Flagellins expressed by strain M2126 carrying these plasmids then can be determined by use of specific sera. 15 The sequences flanking the flkA gene can then be used to PCR amplify other flkA genes (see below). 16.2 The flkA genes coding for H3, H47 and H53: It has been shown that flagellins H3, H47 and H53 are 20 encoded by flkA genes in the type strains [Ratiner, Y. A. (1998) "New flagellin specifying genes in some E. coli strains" J. Bacteriol 180: 979-984]. These genes can be PCR amplified using primers based on the sequences flanking the flkA gene in the H36 type strain. These PCR 25 fragments can then be sequenced, and the genes expressed in strain M2126 for the identification of these genes. 16.3 The fllA genes coding for H44 and H55: It is known that flagellins H44 and H55 are coded by 30 fllA genes. 16.3.1 The H55 flagellin gene: Using primers #1868 and #1870 (Table 3B), we have amplified two bands from the H55 type strain. PCR reaction 35 was carried out under the following conditions: denaturing, 94oC/30'; annealing, 500C/30'; extension, 72oC/1'; 30 cycles. These two PCR fragments were then WO99/61458 PCT/AU99/00385 - 46 cloned into the pGEM-T vector using the Promega pGEM-T cloning kit (Madison WI USA) to make plasmids pPR1994 and pPR1989. Inserts from both plasmids were first sequenced using the M13 universal primers (which bind to the pGEM-T 5 DNA flanking the insertion site). Primers based on the sequence obtained were then used to sequence further, and this procedure was repeated until both inserts were fully or partly sequenced. The sequence of the insert of pPR1994 is highly 10 similar to that of the flagellin gene of the H38 type strain, with 1 amino acid difference in the gene products. We have only sequenced the two ends of the insert of plasmid pPR1989 (figures 70A and 70B), and the sequences of the two ends of the insert of pPR1989 are very similar 15 to ends of other sequenced flagellin genes. We conclude that the insert of plasmid pPR1989 encodes a flagellin gene. The full sequence of the insert of plasmid pPR1989 can be obtained using the same method as for the sequencing of the insert of plasmid pPR1994. It is known 20 that the H55 type strain carries flagellin genes for both H38 and H55, and that the H55 flagellin gene is at the fllA locus [Ratiner, Y. A. (1998) "New flagellin specifying genes in some E. coli strains" J. Bacteriol 180: 979-984]. Thus, it is highly likely that plasmid 25 pPR1989 contains the fllA gene of the H55 type strain. The currently uncharacterised sequence of both ends and of DNA flanking these two sequenced genes can be obtained by PCR walking and sequencing. Methods for PCR walking from a known sequence to an unknown region in 30 chromosomal DNA are available (see [Siebert, P. D. , A. Chenchi, D. E. Kellogg, A. Lukyanov and S. A. Lukyanov (1995) "An improved PCR method for walking in uncloned genomic DNA." Nuc. Acids Res. 23: 1087-1088]). The sequenced genes then can be PCR amplified and 35 cloned using the method(s) described in section 9. Flagellins expressed by strain M2126 carrying these plasmids then can be determined by use of specific sera.

WO99/61458 PCT/AU99/00385 - 47 16.3.2 The H44 flagellin gene: The sequence information for DNA flanking the fllA gene in the H55 type strain can then be used to PCR, 5 sequence and identify the fllA gene in the H44 type strain. 16.4 The flmA gene coding for H54: This gene can be cloned by making a bank of plasmid 10 clones in strain M2126 using chromosomal DNA of the H54 type strain and selecting for a transformant which is motile on an agar plate. This is done by taking advantage of the fact that the H antigen is on flagellin, the protein of the bacterial flagellum used for movement of 15 the bacteria. Strain M2126 lacks flagellin. Once the clone(s) is obtained and identified by use of anti-H54 serum, the flagellin gene can be sequenced. It is possible that clones expressing different flagellin specificities can be obtained, and each of them can be identified by 20 using different sera. 16.5 The flagellin genes obtained from the H37 and H48 type strains: We have used primers #1868 and #1869 (both were based 25 on the sequence obtained from the H48 type strain, also see section 9) and primers #1868 and #1870 (both were based on the sequences of the H7 flagellin gene of the H7 type strain, also see section 9) to PCR amplify and clone the sequenced flagellin genes from the H48 and H37 type 30 strains respectively. Strain P5560 carrying the plasmid containing either the cloned gene was not motile and did not react with the appropriate antisera. It is highly likely that mutaions have occured due to PCR errors. This can be resolved by re-amplification and re-cloning of the 35 genes. 16.6 The flagellin gene obtained from the H25 type WO99/61458 PCT/AU99/00385 - 48 strain: The flagellin gene sequence we first obtained from the H25 type strain lacks 23 and 21 codons at 5' and 3' ends respectively. We could not amplify the full gene from 5 the H25 type strain using primers based on the H7 flagellin gene of the H7 type strain, and it was necessary to get the full sequence of this flagellin gene by other means. We have used primers (#2650: 5' - cag cga tga aat act 10 tgc cat and #2648: 5' - caa tgc ttc gtg acg cac) based on the genes (fliD and fliA respectively) flanking fliC gene in E. coli K-12 [Blattner, F. R., G. I. Plunkett, C. A. Bloch, N. T. Perna, V. Burland, M. Riley and et al. (1997) "The complete genome sequence of E. Coli Kil2" Science 15 277: 1453-1474] and primers (#2658: 5' - gcc tga gtc aga cct ttg and # 2653 5' - aac ctg tct gaa gcg cag) based on the flagellin sequence obtained from the H25 type strain to PCR amplify both ends of the flagellin gene. The PCR product was then sequenced, and we have now obtained the 20 full flagellin gene sequence and sequence for the DNA flanking the flagellin gene from type strain H25 (Figure 69). Now, it is straightforward to PCR amplify, clone and express, and identify this gene using the methods described in sections 9 and 10. 25 16.7 The flagellin genes obtained from the H8 and H40 type strains: The flagellin gene sequences obtained from both the H8 and H40 type strains lack 18 and 15 codons at 5' and 3' 30 ends respectively. We have used primers based on the H7 flagellin gene of the H7 type strain to PCR amplify and clone the full genes from these two strains. Strain M2126 carrying plasmid made this way was not motile under microscope and did not react with the appropriate 35 antisera. This could be due to PCR errors as mentioned in section 16.5 or perhaps the first and last few amino acids encoded by the primers (based on H7 flagellin gene) are WO99/61458 PCT/AU99/00385 - 49 uncompatible in this case. The full sequence of the full gene can be obtained using method described in section 16.6. The flagellin gene can then be PCR amplified, cloned and expressed, and 5 identified using the methods described in sections 9 and 10. The gene products of the flagellin genes obtained from the H8 and H40 type strains are identical. Thus, one of these two H specificities must be encoded by a unknown 10 gene, and it can be cloned and identified using the method described in the section 16.8. 16.8 Flagellin genes coding for H17, H35, and H50: 15 As mentioned above, the sequenced flagellin genes from the H17 and H50 type strains encode H4 and H10 specificities respectively. The flagellin gene sequence obtained from the H35 strain has a insertion and encodes a non-functional gene (see section 8). Thus, genes coding 20 for these flagellins have not been identified, and their location is unknown. One can use primers based on DNA flanking fliC, fllA, flkA, and flmA to do PCR on the type strain for each of the flagellin antigen. PCR products can then be sequenced, and possible genes can be cloned, 25 expressed and identified then. If the target gene is not PCR amplified using primers based on sequence of these loci or sequence flanking these loci, it can be cloned by making a bank of plasmid clones in strain M2126 using chromosomal DNA of the type strain 30 and selecting for a transformant which is motile on an agar plate. This is done by taking advantage of the fact that the H antigen is on flagellin, the protein of the bacterial flagellum used for movement of the bacteria. Strain M2126 lacks flagellin. Once the clone(s) is 35 obtained and identified by use of antisera, the flagellin gene can be sequenced. It is possible that clones expressing different flagellin antigens can be obtained, WO99/61458 PCT/AU99/00385 - 50 and each of them can be identified by using different antisera. Antiserum for H50 can be prepared using standard methods [Ewing, W.H.:Edwards and Ewing's identification of the Enterobacteriaceae., Elsevier 5 Science Publishers, Amsterdam, The Netherlands, 1986]. 0 antigen Materials and Methods-part 1 The experimental procedures for the isolation and 10 characterisation of the E. coli 0111 0 antigen gene cluster (position 3,021-9,981) are according to Bastin D.A., et al. 1991 "Molecular cloning and expression in Escherichia coli K-12 of the rfb gene cluster determining the 0 antigen of an E. coli 0111 strain". Mol. Microbiol. 15 5:9 2223-2231 and Bastin D.A. and Reeves, P.R. 1995 "Sequence and analysis of the 0 antigen gene(rfb)cluster of Escherichia coli 0111". Gene 164: 17-23. A. Bacterial strains and growth media Bacteria were grown in Luria broth supplemented as 20 required. B. Cosmids and phage Cosmids in the host strain x2819 were repackaged in vivo. Cells were grown in 250mL flasks containing 30mL of culture, with moderate shaking at 30 0 C to an optical 25 density of 0.3 at 580 nm. The defective lambda prophage was induced by heating in a water bath at 45 0 C for 15min followed by an incubation at 37 0 C with vigorous shaking for 2hr. Cells were then lysed by the addition of 0.3mL chloroform and shaking for a further 10min. Cell debris 30 were removed from imL of lysate by a 5min spin in a microcentrifuge, and the supernatant removed to a fresh microfuge tube. One drop of chloroform was added then shaken vigorously through the tube contents. C. DNA preparation 35 Chromosomal DNA was prepared from bacteria grown overnight at 37 0 C in a volume of 30mL of Luria broth. After harvesting by centrifugation, cells were washed and WO99/61458 PCT/AU99/00385 - 51 resuspended in 10mL of 50mMTris-HCl pH 8.0. EDTA was added and the mixture incubated for 20min. Then lysozyme was added and incubation continued for a further 10min. Proteinase K, SDS, and ribonuclease were then added and 5 the mixture incubated for up to 2hr for lysis to occur. All incubations were at 370C. The mixture was then heated to 65 0 C and extracted once with 8mL of phenol at the same temperature. The mixture was extracted once with 5mL of phenol/chloroform/iso-amyl alcohol at 4 0 C. 10 Residual phenol was removed by two ether extractions. DNA was precipitated with 2 vols. of ethanol at 4 0 C, spooled and washed in 70% ethanol, resuspended in l-2mL of TE and dialysed. Plasmid and cosmid DNA was prepared by a modification of the Birnboim and Doly method 15 [Birnboim, H. C. and Doly, J. (1979) "A rapid alkaline extraction procedure for screening recombinant plasmid DNA" Nucl. Acid Res. 7:1513-1523]. The volume of culture was 10mL and the lysate was extracted with phenol/chloroform/iso-amyl alcohol before precipitation 20 with isopropanol. Plasmid DNA to be used as vector was isolated on a continuous caesium chloride gradient following alkaline lysis of cells grown in IL of culture. D. Enzymes and buffers. Restriction endonucleases and DNA T4 ligase were 25 purchased from Boehringer Mannheim (Castle Hill, NSW, Australia)or Pharmacia LKB (Melbourne, VIC Australia). Restriction enzymes were used in the recommended commercial buffer. E. Construction of a gene bank. 30 Individual aliquots of M92 chromosomal DNA (strain Stoke W, from Statens Serum Institut, 5 Artillerivej, 2300 Copenhagen S, Denmark) were partially digested with 0.2U Sau3Al for 1-15mins. Aliquots giving the greatest proportion of fragments in the size range of 35 approximately 40-50kb were selected and ligated to vector pPR691 previously digested with BamHl and PvuII. Ligation mixtures were packaged in vitro with packaging extract.

WO99/61458 PCT/AU99/00385 - 52 The host strain for transduction was x2819 and recombinants were selected with kanamycin. F. Serological procedures. Colonies were screened for the presence of the 0111 5 antigen by immunoblotting. Colonies were grown overnight, up to 100 per plate then transferred to nitrocellulose discs and lysed with 0.5N HC1. Tween 20 was added to TBS at 0.05% final concentration for blocking, incubating and washing steps. Primary antibody 10 was E. coli O group 111 antiserum, diluted 1:800. The secondary antibody was goat anti-rabbit IgG labelled with horseradish peroxidase diluted 1:5000. The staining substrate was 4-chloro-l-napthol. Slide agglutination was performed according to the standard procedure. 15 G. Recombinant DNA methods. Restriction mapping was based on a combination of standard methods including single and double digests and sub-cloning. Deletion derivatives of entire cosmids were produced as follows: aliquots of 1.8mg of cosmid DNA were 20 digested in a volume of 20ml with 0.25U of restriction enzyme for 5-80min. One half of each aliquot was used to check the degree of digestion on an agarose gel. The sample which appeared to give a representative range of fragments was ligated at 40C overnight and transformed by 25 the CaC1 2 method into JM109. Selected plasmids were transformed into sf174 by the same method. P4657 was transformed with pPR1244 by electroporation. H. DNA hybridisation Probe DNA was extracted from agarose gels by 30 electroelution and was nick-translated using [a-32P] dCTP. Chromosomal or plasmid DNA was electrophoresed in 0.8% agarose and transferred to a nitrocellulose membrane. The hybridisation and pre-hybridisation buffers contained either 30% or 50% formamide for low 35 and high stringency probing respectively. Incubation temperatures were 420C and 37 0 C for pre-hybridisation and hybridisation respectively. Low stringency washing of WO99/61458 PCT/AU99/00385 - 53 filters consisted of 3 x 20min washes in 2 x SSC and 0.1% SDS. High-stringency washing consisted of 3 x 5min washes in 2 x SSC and 0.1% SDS at room temperature, a lhr wash in 1 x SSC and 0.1% SDS at 58 0 C and 15min wash in 5 0.1 x SSC and 0.1% SDS at 58 0 C. I. Nucleotide sequencing of E. coli 0111 0 antigen gene cluster (position 3,021-9,981) Nucleotide sequencing was performed using an ABI 373 automated sequencer (CA, USA). The region between map 10 positions 3.30 and 7.90 was sequenced using uni-directional exonuclease III digestion of deletion families made in PT7T3190 from clones pPR1270 and pPR1272. Gaps were filled largely by cloning of selected fragments into Ml3mpl8 or Ml3mpl9. The region from map 15 positions 7.90-10.2 was sequenced from restriction fragments in Ml3mpl8 or Ml3mpl9. Remaining gaps in both the regions were filled by priming from synthetic oligonucleotides complementary to determined positions along the sequence, using a single stranded DNA template 20 in M13 or phagemid. The oligonucleotides were designed after analysing the adjacent sequence. All sequencing was performed by the chain termination method. Sequences were aligned using SAP [Staden, R., 1982 "Automation of the computer handling of gel reading data produced by the 25 shotgun method of DNA sequencing". Nuc. Acid Res. 10: 4731-4751; Staden, R., 1986 "The current status and portability of our sequence handling software". Nuc. Acid Res. 14: 217-231]. The program NIP [Staden, R. 1982 "An interactive graphics program for comparing and aligning 30 nucleic acid and amino acid sequence". Nuc. Acid Res. 10: 2951-2961] was used to find open reading frames and translate them into proteins. J. Isolation of clones carrying E. coli 0111 0 antigen gene cluster 35 The E. coli 0 antigen gene cluster was isolated according to the method of Bastin D.A., et al. [1991 "Molecular cloning and expression in Escherichia coli K- WO99/61458 PCT/AU99/00385 - 54 12 of the rfb gene cluster determining the O antigen of an E. coli 0111 strain". Mol. Microbiol. 5(9), 2223 2231]. Cosmid gene banks of M92 chromosomal DNA were established in the in vivo packaging strain x2819. From 5 the genomic bank, 3.3 x 10 3 colonies were screened with E.coli 0111 antiserum using an immuno-blotting procedure: 5 colonies (pPR1054, pPR1055, pPR1056, pPR1058 and pPR1287) were positive. The cosmids from these strains were packaged in vivo into lambda particles and 10 transduced into the E. coli deletion mutant Sf174 which lacks all O antigen genes. In this host strain, all plasmids gave positive agglutination with 0111 antiserum. An Eco R1 restriction map of the 5 independent cosmids showed that they have a region of approximately 11.5 kb 15 in common (Figure 1). Cosmid pPR1058 included sufficient flanking DNA to identify several chromosomal markers linked to 0 antigen gene cluster and was selected for analysis of the 0 antigen gene cluster region. K. Restriction mapping of cosmid pPR1058 20 Cosmid pPR1058 was mapped in two stages. A preliminary map was constructed first, and then the region between map positions 0.00 and 23.10 was mapped in detail, since it was shown to be sufficient for 0111 antigen expression. Restriction sites for both stages 25 are shown in Figure 2. The region common to the five cosmid clones was between map positions 1.35 and 12.95 of pPR1058. To locate the 0 antigen gene cluster within pPR1058, pPR1058 cosmid was probed with DNA probes covering 0 30 antigen gene cluster flanking regions from S. enterica LT2 and E.coli K-12. Capsular polysaccharide (cps) genes lie upstream of 0 antigen gene cluster while the gluconate dehydrogenase (gnd) gene and the histidine (his) operon are downstream, the latter being further 35 from the 0 antigen gene cluster. The probes used were pPR472 (3.35kb), carrying the gnd gene of LT2, pPR685 (5.3kb) carrying two genes of the cps cluster, cpsB and WO99/61458 PCT/AU99/00385 - 55 cpsG of LT2, and K350 (16.5kb) carrying all of the his operon of K-12. Probes hybridised as follows: pPR472 hybridised to 1.55kb and 3.5 kb (including 2.7 kb of vector) fragments of Pstl and HindIII double digests of 5 pPR1246 (a HindIII/EcoRl subclone derived from pPR1058, Figure 2), which could be located at map positions 12.95 15.1; pPR685 hybridised to a 4.4 kb EcoRl fragment of pPR1058 (including 1.3 kb of vector) located at map position 0.00-3.05; and K350 hybridised with a 32kb EcoRl 10 fragment of pPR1058 (including 4.0kb of vector), located at map position 17.30-45.90. Subclones containing the presumed gnd region complemented a gnd-edd strain GB23152. On gluconate bromothymol blue plates, pPR1244' and pPR1292 in this host strain gave the green colonies 15 expected of a gnd'edd - genotype. The his' phenotype was restored by plasmid pPR1058 in the his deletion strain Sf174 on minimal medium plates, showing that the plasmid carries the entire his operon. It is likely that the 0 antigen gene cluster region 20 lies between gnd and cps, as in other E. coli and S. enterica strains, and hence between the approximate map positions 3.05 and 12.95. To confirm this, deletion derivatives of pPR1058 were made as follows: first, pPR1058 was partially digested with HindIII and self 25 ligated. Transformants were selected for kanamycin resistance and screened for expression of 0111 antigen. Two colonies gave a positive reaction. EcoRl digestion showed that the two colonies hosted identical plasmids, one of which was designated pPR1230, with an insert which 30 extended from map positions 0.00 to 23.10. Second pPR1058 was digested with Sall and partially digested with Xhol and the compatible ends were re-ligated. Transformants were selected with kanamycin and screened for 0111 antigen expression. Plasmid DNA of 8 35 positively reacting clones was checked using EcoRl and Xhol digestion and appeared to be identical. The cosmid of one was designated pPR1231. The insert of pPR1231 WO99/61458 PCT/AU99/00385 - 56 contained the DNA region between map positions 0.00 and 15.10. Third, pPR1231 was partially digested with Xhol, self-ligated, and transformants selected on spectinomycin/ streptomycin plates. Clones were screened 5 for kanamycin sensitivity and of 10 selected, all had the DNA region from the Xhol site in the vector to the Xhol site at position 4.00 deleted. These clones did not express the 0111 antigen, showing that the Xhol site at position 4.00 is within the 0 antigen gene cluster. One 10 clone was selected and named pPR1288. Plasmids pPR1230, pPR1231, and pPR1288 are shown in Figure 2. L. Analysis of the E. coli 0111 0 antigen gene cluster (position 3,021-9,981) nucleotide sequence data Bastin and Reeves [1995 "Sequence and analysis of 15 the 0 antigen gene(rfb)cluster of Escherichia coli 0111". Gene 164: 17-23] partially characterised the E.coli 0111 O antigen gene cluster by sequencing a fragment from map position 3,021-9,981. Figure 3 shows the gene organisation of position 3,021-9,981 of E. coli 0111 0 20 antigen gene cluster. orf3 and orf6 have high level amino acid identity with wcaH and wcaG (46.3% and 37.2% respectively), and are likely to be similar in function to sugar biosynthetic pathway genes in the E. coli K-12 colanic gene cluster. orf4 and orf5 show high levels of 25 amino acid homology to manC and manB genes respectively. orf7 shows high level homology with rfbH which is an abequose pathway gene. orf8 encodes a protein with 12 transmembrane segments and has similarity in secondary structure to other wzx genes and is likely therefore to 30 be the 0 antigen flippase gene. Materials and Methods-part 2 A. Nucleotide sequencing of 1 to 3,020 and 9,982 to 14,516 of the E. coli 0111 0 antigen gene cluster 35 The sub clones which contained novel nucleotide sequences, pPR1231 (map position 0 and 1,510), pPR1237 (map position -300 to 2,744), pPR1239 (map position 2,744 WO99/61458 PCT/AU99/00385 - 57 to 4,168), pPR1245 (map position 9,736 to 12,007) and pPR1246 (map position 12,007 to 15,300) (Figure 2), were characterised as follows: the distal ends of the inserts of pPR1237, pPR1239 and pPR1245 were sequenced using the 5 M13 forward and reverse primers located in the vector. PCR walking was carried out to sequence further into each insert using primers based on the sequence data and the primers were tagged with M13 forward or reverse primer sequences for sequencing. This PCR walking procedure was 10 repeated until the entire insert was sequenced. pPR1246 was characterised from position 12,007 to 14,516. The DNA of these sub clones was sequenced in both directions. The sequencing reactions were performed using the dideoxy termination method and thermocycling and reaction 15 products were analysed using fluorescent dye and an ABI automated sequencer (CA, USA). B. Analysis of the E. coli 0111 O antigen gene cluster (positions 1 to 3,020 and 9,982 to 14,516 of Figure 5) nucleotide sequence data 20 The gene organisation of regions of E. coli 0111 0 antigen gene cluster which were not characterised by Bastin and Reeves [1995 "Sequence and analysis of the 0 antigen gene(rfb)cluster of Escherichia coli 0111." Gene 164: 17-23],(positions 1 to 3,020 and 9,982 to 14,516) is 25 shown in Figure 3. There are two open reading frames in region 1. Four open reading frames are predicted in region 2. The position of each gene is listed in Table 9. The deduced amino acid sequence of orfl (wbdH) 30 shares about 64% similarity with that of the rfp gene of Shigella dysenteriae. Rfp and WbdH have very similar hydrophobicity plots and both have a very convincing predicted transmembrane segment in a corresponding position. rfp is a galactosyl transferase involved in 35 the synthesis of LPS core, thus wbdH is likely to be a galactosyl transferase gene. orf2 has 85.7% identity at amino acid level to the gmd gene identified in the E.

WO99/61458 PCT/AU99/00385 - 58 coli K-12 colanic acid gene cluster and is likely to be a gmd gene. orf9 encodes a protein with 10 predicted transmembrane segments and a large cytoplasmic loop. This inner membrane topology is a characteristic feature 5 of all known 0 antigen polymerases thus it is likely that orf9 encodes an 0 antigen polymerase gene, wzy. orflO (wbdL) has a deduced amino acid sequence with low homology with Lsi2 of Neisseria gonorrhoeae. Lsi2 is responsible for adding GlcNAc to galactose in the 10 synthesis of lipooligosaccharide. Thus it is likely that wbdL is either a colitose or glucose transferase gene. orfll (wbdM) shares high level nucleotide and amino acid similarity with TrsE of Yersinia enterocolitica. TrsE is a putative sugar transferase thus it is likely that wbdM 15 encodes the colitose or glucose transferase. In summary three putative transferase genes and an 0 antigen polymerase gene were identified at map position 1 to 3,020 and 9,982 to 14,516 of E. coli 0111 0 antigen gene cluster. A search of GenBank has shown that there 20 are no genes with significant similarity at the nucleotide sequence level for two of the three putative transferase genes or the polymerase gene. Figure 5 provides the nucleotide sequence of the 0111 antigen gene cluster. 25 Materials and Methods-part 3 A. PCR amplification of 0157 antigen gene cluster from an E. coli 0157:H7 strain (Strain C664-1992, from Statens Serum Institut, 5 Artillerivej, 2300, Copenhagen S, 30 Denmark) E. coli 0157 O antigen gene cluster was amplified by using long PCR [Cheng et al. 1994, "Effective amplification of long targets from cloned inserts and human and genomic DNA" P.N.A.S. USA 91: 5695-569] with 35 one primer (primer #412: att ggt agc tgt aag cca agg gcg gta gcg t) based on the JumpStart sequence usually found in the promoter region of O antigen gene clusters [Hobbs, WO99/61458 PCT/AU99/00385 - 59 et al. 1994 "The JumpStart sequence: a 39 bp element common to several polysaccharide gene clusters" Mol. Microbiol. 12: 855-856], and another primer #482 (cac tgc cat acc gac gac gcc gat ctg ttg ctt gg) based on the gnd 5 gene usually found downstream of the 0 antigen gene cluster. Long PCR was carried out using the Expand Long Template PCR System from Boehringer Mannheim (Castle Hill NSW Australia), and products, 14 kb in length, from several reactions were combined and purified using the 10 Promega Wizard PCR preps DNA purification System (Madison WI USA). The PCR product was then extracted with phenol and twice with ether, precipitated with 70% ethanol, and resuspended in 40mL of water. B. Construction of a random DNase I bank: 15 Two aliquots containing about 150ng of DNA each were subjected to DNase I digestion using the Novagen DNase I Shotgun Cleavage (Madison WI USA) with a modified protocol as described. Each aliquot was diluted into 45ml of 0.05M Tris -HC1 (pH7.5), 0.05mg/mL BSA and 10mM 20 MnC1 2 . 5mL of 1:3000 or 1:4500 dilution of DNaseI (Novagen) (Madison WI USA) in the same buffer was added into each tube respectively and 10ml of stop buffer (100mM EDTA), 30% glycerol, 0.5% Orange G, 0.075% xylene and cyanol (Novagen) (Madison WI USA) was added after 25 incubation at 15 0 C for 5 min. The DNA from the two DNaseI reaction tubes were then combined and fractionated on a 0.8% LMT agarose gel, and the gel segment with DNA of about 1kb in size (about 1.5mL agarose) was excised. DNA was extracted from agarose using Promega Wizard PCR Preps 30 DNA Purification (Madison WI USA) and resuspended in 200 mL water, before being extracted with phenol and twice with ether, and precipitated. The DNA was then resuspended in 17.25 mL water and subjected to T4 DNA polymerase repair and single dA tailing using the Novagen 35 Single dA Tailing Kit (Madison WI USA). The reaction product (85ml containing about 8ng DNA) was then extracted with chloroform:isoamyl alcohol (24:1) once and WO99/61458 PCT/AU99/00385 - 60 ligated to 3x 10-3 pmol pGEM-T (Promega) (Madison WI USA) in a total volume of 100mL. Ligation was carried out overnight at 4 0 C and the ligated DNA was precipitated and resuspended in 20mL water before being electroporated 5 into E. coli strain JM109 and plated out on BCIG-IPTG plates to give a bank. C. Sequencing DNA templates from clones of the bank were prepared for sequencing using the 96-well format plasmid DNA 10 miniprep kit from Advanced Genetic Technologies Corp (Gaithersburg MD USA) The inserts of these clones were sequenced from one or both ends using the standard M13 sequencing primer sites located in the pGEM-T vector. Sequencing was carried out on an ABI377 automated 15 sequencer (CA USA) as described above, after carrying out the sequencing reaction on an ABI Catalyst (CA USA). Sequence gaps and areas of inadequate coverage were PCR amplified directly from 0157 chromosomal DNA using primers based on the already obtained sequencing data and 20 sequenced using the standard M13 sequencing primer sites attached to the PCR primers. D. Analysis of the E. coli 0157 0 antigen gene cluster nucleotide sequence data Sequence data were processed and analysed using the 25 Staden programs [Staden, R., 1982 "Automation of the computer handling of gel reading data produced by the shotgun method of DNA sequencing." Nuc. Acid Res. 10: 4731-4751; Staden, R., 1986 "The current status and portability of our sequence handling software". Nuc. Acid 30 Res. 14: 217-231; Staden, R. 1982 "An interactive graphics program for comparing and aligning nucleic acid and amino acid sequence". Nuc. Acid Res. 10: 2951-2961]. Figure 4 shows the structure of E. coli 0157 0 antigen gene cluster. Twelve open reading frames were predicted 35 from the sequence data, and the nucleotide and amino acid sequences of all these genes were then used to search the GenBank database for indication of possible function and WO99/61458 PCT/AU99/00385 - 61 specificity of these genes. The position of each gene is listed in Table 9. The nucleotide sequence is presented in Figure 6. orfs 10 and 11 showed high level identity to manC 5 and manB and were named manC and manB respectively. orf7 showed 89% identity (at amino acid level) to the gmd gene of the E. coli colanic acid capsule gene cluster (Stevenson G., K. et al. 1996 "Organisation of the Escherichia coli K-12 gene cluster responsible for 10 production of the extracellular polysaccharide colanic acid".J. Bacteriol. 178:4885-4893) and was named gmd. orf8 showed 79% and 69% identity (at amino acid level) respectively to wcaG of the E. coli colanic acid capsule gene cluster and to wbcJ (orfl4.8) gene of the Yersinia 15 enterocolitica 08 0 antigen gene cluster (Zhang, L. et al. 1997 "Molecular and chemical characterization of the lipopolysaccharide O-antigen and its role in the virulence of Y. enterocolitica serotype 08".Mol. Microbiol. 23:63-76). Colanic acid and the Yersinia 08 0 20 antigen both contain fucose as does the 0157 0 antigen. There are two enzymatic steps required for GDP-L-fucose synthesis from GDP-4-keto-6-deoxy-D-mannose, the product of the gmd gene product. However, it has been shown recently (Tonetti, M et al. 1996 Synthesis of GDP-L 25 fucose by the human FX protein J. Biol. Chem. 271:27274 27279) that the human FX protein has "significant homology" with the wcaG gene (referred to as Yefb in that paper), and that the FX protein carries out both reactions to convert GDP-4-keto-6-deoxy-D-mannose to GDP 30 L-fucose. We believe that this makes a very strong case for orf8 carrying out these two steps and propose to name the gene fcl. In support of the one enzyme carrying out both functions is the observation that there are no genes other than manB, manC, gmd and fcl with similar levels of 35 similarity between the three bacterial gene clusters for fucose containing structures. orf5 is very similar to wbeE (rfbE) of Vibrio WO99/61458 PCT/AU99/00385 - 62 cholerae 01, which is thought to be the perosamine synthetase, which converts GDP-4-keto-6-deoxy-D-mannose to GDP-perosamine (Stroeher, U.H et al. 1995 "A putative pathway for perosamine biosynthesis is the first function 5 encoded within the rfb region of Vibrio cholerae" 01. Gene 166: 33-42). V. cholerae 01 and E. coli 0157 0 antigens contain perosamine and N-acetyl-perosamine respectively. The V. cholerae 01 manA, manB, gmd and wbeE genes are the only genes of the V. cholerae 01 gene 10 cluster with significant similarity to genes of the E. coli 0157 gene cluster and we believe that our observations both confirm the prediction made for the function of wbe of V. cholerae, and show that orf5 of the 0157 gene cluster encodes GDP-perosamine synthetase. 15 orf5 is therefore named per. orf5 plus about 100bp of the upstream region (postion 4022-5308)was previously sequenced by Bilge, S.S. et al. [1996 "Role of the Escherichia coli 0157-H7 0 side chain in adherence and analysis of an rfb locus".Infect. Immun. 64:4795-4801]. 20 orfl2 shows high level similarity to the conserved region of about 50 amino acids of various members of an acetyltransferase family (Lin, W., et al. 1994 "Sequence analysis and molecular characterisation of genes required for the biosynthesis of type 1 capsular polysaccharide in 25 Staphylococcus aureus". J. Bateriol. 176: 7005-7016) and we believe it is the N-acetyltransferase to convert GDP perosamine to GDP-perNAc. orfl2 has been named wbdR. The genes manB, manC, gmd, fcl, per and wbdR account for all of the expected biosynthetic pathway genes of the 30 0157 gene cluster. The remaining biosynthetic step(s) required are for synthesis of UDP-GalNAc from UDP-Glc. It has been proposed (Zhang, L., et al. 1997 "Molecular and chemical characterisation of the lipopolysaccharide O-antigen and 35 its role in the virulence of Yersinia enterocolitica serotype 08".Mol. Microbiol. 23:63-76) that in Yersinia enterocolitica UDP-GalNAc is synthesised from UDP-GlcNAc WO99/61458 PCT/AU99/00385 - 63 by a homologue of galactose epimerase (GalE), for which there is a galE like gene in the Yersinia enterocolitica 08 gene cluster. In the case of 0157 there is no galE homologue in the gene cluster and it is not clear how 5 UDP-GalNAc is synthesised. It is possible that the galactose epimerase encoded by the galE gene in the gal operon, can carry out conversion of UDP-GlcNAc to UDP GalNAc in addition to conversion of UDP-Glc to UDP-Gal. There do not appear to be any gene(s) responsible for 10 UDP-GalNAc synthesis in the 0157 gene cluster. orf4 shows similarity to many wzx genes and is named wzx and orf2 which shows similarity of secondary structure in the predicted protein to other wzy genes and is for that reason named wzy. 15 The orfl, orf3 and orf6 gene products all have characteristics of transferases, and have been named wbdN, wbdO and wbdP respectively. The 0157 0 antigen has 4 sugars and 4 transferases are expected. The first transferase to act would put a sugar phosphate onto 20 undecaprenol phosphate. The two transferases known to perform this function, WbaP (RfbP) and WecA (Rfe) transfer galactose phosphate and N-acetyl-glucosamine phosphate respectively to undecaprenol phosphate. Neither of these sugars is present in the 0157 structure. 25 Further, none of the presumptive transferases in the 0157 gene cluster has the transmembrane segments found in WecA and WbaP which transfer a sugar phosphate to undecaprenol phosphate and expected for any protein which transferred a sugar to undecaprenol phosphate which is 30 embedded within the membrane. The WecA gene which transfers GlcNAc-P to undecaprenol phosphate is located in the Enterobactereal Common Antigen (ECA) gene cluster and it functions in ECA synthesis in most and perhaps all E. coli strains, and 35 also in 0 antigen synthesis for those strains which have GlcNAc as the first sugar in the O unit. It appears that WecA acts as the transferase for WO99/61458 PCT/AU99/00385 - 64 addition of GalNAc-1-P to undecaprenol phosphate for the Yersinia enterocolitica 08 0 antigen [Zhang et al.1997 "Molecular and chemical characterisation of the lipopolysaccharide 0 antigen and its role in the 5 virulence of Yersinia enterocolitica serotype 08" Mol. Microbiol. 23: 63-76.] and perhaps does so here as the 0157 structure includes GalNAc. WecA has also been reported to add Glucose-1-P phosphate to undecaprenol phosphate in E. coli 08 and 09 strains, and an 10 alternative possibility for transfer of the first sugar to undecaprenol phosphate is WecA mediated transfer of glucose, as there is a glucose residue in the 0157 0 antigen. In either case the requisite number of transferase genes are present if GalNAc or Glc is 15 transferred by WecA and the side chain Glc is transferred by a transferase outside of the 0 antigen gene cluster. orf9 shows high level similarity (44% identity at amino acid level, same length) with wcaH gene of the E. coli colanic acid capsule gene cluster. The function of 20 this gene is unknown, and we give orf9 the name wbdQ. The DNA between manB and wdbR has strong sequence similarity to one of the H-repeat units of E. coli K12. Both of the inverted repeat sequences flanking this region are still recognisable, each with two of the 11 25 bases being changed. The H-repeat associated protein encoding gene located within this region has a 267 base deletion and mutations in various positions. It seems that the H-repeat unit has been associated with this gene cluster for a long period of time since it translocated 30 to the gene cluster, perhaps playing a role in assembly of the gene cluster as has been proposed in other cases. Materials and Methods - part 4 To test our hypothesis that 0 antigen genes for 35 transferases and the wzx, wzy genes were more specific than pathway genes for diagnostic PCR, we first carried out PCR using primers for all the E. coli 016 0 antigen WO99/61458 PCT/AU99/00385 - 65 genes (Table 7). The PCR was then carried out using PCR primers for E.coli 0111 transferase, wzx and wzy genes (Table 8, 8A). PCR was also carried out using PCR primers for the E. coli 0157 transferase, wzx and wzy 5 genes (Table 9, 9A). Chromosomal DNA from the 166 serotypes of E. coli available from Statens Serum Institut, 5 Artillerivej, 2300 Copenhagen Denmark was isolated using the Promega Genomic (Madison WI USA) isolation kit. Note that 164 of 10 the serogroups are described by Ewing W. H.: Edwards and Ewings "Identification of the Enterobacteriacea" Elsevier, Amsterdam 1986 and that they are numbered 1-171 with numbers 31, 47, 67, 72, 93, 94 and 122 no longer valid. Of the two serogroup 19 strains we used 19ab 15 strain F8188-41. Lior H. 1994 ["Classification of Escherichia coli In Escherichia coli in domestic animals and humans pp 31-72. Edited by C.L. Gyles CAB international] adds two more numbered 172 and 173 to give the 166 serogroups used. Pools containing 5 to 8 samples 20 of DNA per pool were made. Pool numbers 1 to 19 (Table 4) were used in the E. coli 0111 and 0157 assay. Pool numbers 20 to 28 were also used in the 0111 assay, and pool numbers 22 to 24 contained E. coli 0111 DNA and were used as positive controls (Table 5). Pool numbers 29 to 25 42 were also used in the 0157 assay, and pool numbers 31 to 36 contained E. coli 0157 DNA, and were used as positive controls (Table 6). Pool numbers 2 to 20, 30, 43 and 44 were used in the E. coli 016 assay (Tables 4 to 6). Pool number 44 contained DNA of E. coli K-12 strains 30 C600 and WG1 and was used as a positive control as between them they have all of the E. coli K-12 016 0 antigen genes. PCR reactions were carried out under the following conditions: denaturing 94 0 C/30"; annealing, temperature 35 varies (refer to Tables)/30"; extension, 72 0 C/1'; 30 cycles. PCR reaction was carried out in an volume of 25mL for each pool. After the PCR reaction, 10mL PCR WO99/61458 PCT/AU99/00385 - 66 product from each pool was run on an agarose gel to check for amplified DNA. Each E. coli chromosomal DNA sample was checked by gel electrophoresis for the presence of chromosomal DNA 5 and by PCR amplification of the E. coli mdh gene using oligonucleotides based on E. coli K-12 [Boyd et al. (1994) "Molecular genetic basis of allelic polymorphism in malate degydrogenase (mdh) in natural populations of Escherichia coli and Salmonella enterica" Proc. Nat. 10 Acad. Sci. USA. 91:1280-1284.] Chromosomal DNA samples from other bacteria were only checked by gel electrophoresis of chromosomal DNA. A. Primers based on E. coli 016 0 antigen gene cluster 15 sequence. The O antigen gene cluster of E. coli 016 was the only typical E. coli 0 antigen gene cluster that had been fully sequenced prior to that of 0111, and we chose it for testing our hypothesis. One pair of primers for each 20 gene was tested against pools 2 to 20, 30 and 43 of E. coli chromosomal DNA. The primers, annealing temperatures and functional information for each gene are listed in Table 8. For the five pathway genes, there were 17/21, 13/21, 25 0/21, 0/21, 0/21 positive pools for rmlB, rmlD, rmlA, rmlC and glf respectively (Table 7). For the wzx, wzy and three transferase genes there were no positives amongst the 21 pools of E. coli chromosomal DNA tested (Table 7). In each case the #44 pool gave a positive 30 result. B. Primers based on the E. coli 0111 0 antigen gene cluster sequence. One to four pairs of primers for each of the 35 transferase, wzx and wzy genes of 0111 were tested against the pools 1 to 21 of E. coli chromosomal DNA (Table 8). For wbdH, four pairs of primers, which bind WO99/61458 PCT/AU99/00385 - 67 to various regions of this gene, were tested and found to be specific for 0111 as there was no amplified DNA of the correct size in any of those 21 pools of E. coli chromosomal DNA tested. Three pairs of primers for wbdM 5 were tested, and they are all specific although primers #985/#986 produced a band of the wrong size from one pool. Three pairs of primers for wzx were tested and they all were specific. Two pairs of primers were tested for wzy, both are specific although #980/#983 gave a band 10 of the wrong size in all pools. One pair of primers for wbdL was tested and found unspecific and therefore no further test was carried out. Thus, wzx, wzy and two of the three transferase genes are highly specific to 0111. Bands of the wrong size found in amplified DNA are 15 assumed to be due to chance hybridisation of genes widely present in E. coli. The primers, annealing temperatures and positions for each gene are in Table 8. The 0111 assay was also performed using pools including DNA from O antigen expressing Yersinia 20 pseudotuberculosis, Shigella boydii and Salmonella enterica strains (Table 8A). None of the oligonucleotides derived from wbdH, wzx, wzy or wbdM gave amplified DNA of the correct size with these pools. Notably, pool number 25 includes S. enterica Adelaide 25 which has the same 0 antigen as E. coli 0111: this pool did not give a positive PCR result for any primers tested indicating that these genes are highly specific for E. coli 0111. Each of the 12 pairs binding to wbdH, wzx, wzy and 30 wbdM produces a band of predicted size with the pools containing 0111 DNA (pools number 22 to 24). As pools 22 to 24 included DNA from all strains present in pool 21 plus 0111 strain DNA (Table 5), we conclude that the 12 pairs of primers all give a positive PCR test with each 35 of three unrelated 0111 strains but not with any other strains tested. Thus these genes are highly specific for E. coli 0111.

WO99/61458 PCT/AU99/00385 - 68 C. Primers based on the E. coli 0157 0 antigen gene cluster sequence. Two or three primer pairs for each of the 5 transferase, wzx and wzy genes of 0157 were tested against E. coli chromosomal DNA of pools 1 to 19, 29 and 30 (Table 9). For wbdN, three pairs of primers, which bind to various regions of this gene, were tested and found to be specific for 0157 as there was no amplified 10 DNA in any of those 21 pools of E. coli chromosomal DNA tested. Three pairs of primers for wbdO were tested, and they are all specific although primers # 1211/#1212 produced two or three bands of the wrong size from all pools. Three pairs of primers were tested for wbdP and 15 they all were specific. Two pairs of primers were tested for wbdR and they were all specific. For wzy, three pairs of primers were tested and all were specific although primer pair #1203/#1204 produced one or three bands of the wrong size in each pool. For wzx, two pairs 20 of primers were tested and both were specific although primer pair #1217/#1218 produced 2 bands of wrong size in 2 pools, and 1 band of wrong size in 7 pools. Bands of the wrong size found in amplified DNA are assumed to be due to chance hybridisation of genes widely present in E. 25 coli. The primers, annealing temperatures and function information for each gene are in Table 9. The 0157 assay was also performed using pools 37 to 42, including DNA from 0 antigen expressing Yersinia pseudotuberculosis, Shigella boydii, Yersinia 30 enterocolitica 09, Brucella abortus and Salmonella enterica strains (Table 9A). None of the oligonucleotides derived from wbdN, wzy, wbdO, wzx, wbdP or wbdR reacted specifically with these pools, except that primer pair #1203/#1204 produced two bands with Y. 35 enterocolitica 09 and one of the bands is of the same size with that from the positive control. Primer pair #1203/#1204 binds to wzy. The predicted secondary WO99/61458 PCT/AU99/00385 - 69 structures of Wzy proteins are generally similar, although there is very low similarity at amino acid or DNA level among the sequenced wzy genes. Thus, it is possible that Y. enterocolitica 09 has a wzy gene closely 5 related to that of E. coli 0157. It is also possible that this band is due to chance hybridization of another gene, as the other two wzy primer pairs (#1205/#1206 and #1207/#1208) did not produce any band with Y. enterocolitica 09. Notably, pool number 37 includes S. 10 enterica Landau which has the same 0 antigen as E. coli 0157, and pool 38 and 39 contain DNA of B. abortus and Y. enterocolitica 09 which cross react serologically with E. coli 0157. This result indicates that these genes are highly 0157 specific, although one primer pair may have 15 cross reacted with Y. enterocolitica 09. Each of the 16 pairs binding to wbdV, wzx, wzy, wbdO, wbdP and wbdR produces a band of predicted size with the pools containing 0157 DNA (pools number 31 to 36). As pool 29 included DNA from all strains present in 20 pools 31 to 36 other than 0157 strain DNA (Table 6), we conclude that the 16 pairs of primers all give a positive PCR test with each of the five unrelated 0157 strains. Thus PCR using primers based on genes wbdN, wzy, wbdO, wzx, wbdP and wbdR is highly specific for E. coli 25 0157, giving positive results with each of six unrelated 0157 strains while only one primer pair gave a band of the expected size with one of three strains with 0 antigens known to cross-react serologically with E. coli 0157.

WO99/61458 PCT/AU99/00385 - 70 TABLE 1 H7 strains used in this work in addition to the H 5 antigens type strains Name used Serotype Original Source* in this name study M527 0157:H7 C664-1992 a M917 Ol8ac:H7 A57 IMVS M918 Ol8ac:H7 A62 IMVS M973 02:H7 A1107 CDC M1004 0157:H7 EH7 b M1179 Ol8ac:H7 D-M3291/54 IMVS M1200 07:H7 A64 c M1211 Ol9ab:H7 F8188-41 IMVS M1328 053:H7 14097 IMVS M1686 055:H7 TB156 d * a. Statens Serum Institut, Copenhagen, Denmark. 10 b. Dr R. Brown of Royal Children's Hospital, Melbourne, Australia. c. Max-Planck Institut fur molekulare Genetik, Berlin, 15 Germany. d. Dr P. Tarr of Children's Hospital and Medical Center, University of Washington, USA. 20 IMVS, Institute of Medical and veterinary Science, Adelaide, Australia. CDC, Centers for Disease Control and prevention, Atlanta,

USA.

WO 99/61458 PCT/AU99/00385 - 71 Table 2 Oligonucleotides used to PCR amplify fliC genes from different H type strains for sequencing H Type Strains Annealing Temperature (oC) Primers Used 1 55 #1575/#1576 2 55 #1285/#1286 3 55 #1285/#1286 4 50 $1431/#1432 5 60 #1285/#1286 6 55 #1575/#1576 7 55 #1575/#1576 8 55 #1431/#1432 9 60 #1575/#1576 10 55 #1575/#1576 11 55 #1285/#1286 12 60 #1575/#1576 14 60 #1575/#1576 15 60 #1575/#1576 16 60 #1575/#1576 17 60 #1417/#1418 18 60 #1575/#1576 19 60 #1575/#1576 20 60 #1575/#1576 21 55 #1285/#1286 23 60 #1575/#1576 24 60 #1285/#1286 25 60 #1417/#1418 26 60 #1575/#1576 27 50 #1431/#1432 28 60 #1575/#1576 29 60 #1285/#1286 30 60 #1575/#1576 31 60 #1575/#1576 32 60 #1575/#1576 33 60 #1285/1286 34 55 #1575/#1576 35 50 #1431/#1432 37 60 #1285/#1286 38 60 #1285/#1286 39 55 #1285/#1286 40 55 #1285/#1286 41 60 #1575/#1576 42 60 #1285/#1286 43 60 #1575/#1576 44 60 #1285/#1286 45 60 #1575/#1576 46 60 #1575/#1576 47 55 #1285/#1286 48 60 #1575/#1576 49 60 #1575/#1576 50 60 #1285/#1286 51 60 #1575/#1576 52 60 #1575/#1576 54 50 #1431/#1432 55 60 #1285/#1286 56 60 #1285/#1286 WO 99/61458 PCT/AU99/00385 - 72 Table 3 Summary of the flagellin sequences obtained and specific H type oligonucleotide primers H type strain(s) the H specificity H type strain from Positions of Positions of sequenced gene(s) coded by the which the flagellin gene primer 1 primer 2 obtained from gene(s) sequence was used for primer choice 1 1 1 892-909 1172-1189 2 2 2 568-587 1039-1056 4,17,44 4 4 466-483 628-648 5 5 5 697-714 877-897 6 6 6 565-585 799-816 7 7 7 553-570 1483-1500 (primer #1806) (primer #1809) 9 9 9 616-633 838-855 10(50)* 10 10 559-579 697-717 11 11 11 586-606* 791-810* 12 12 12 892-909 1172-1189 14 14 14 586-606 793-813 15 15 15 640-660 817-834 3 16 3 649-666 925-942 18 18 18 589-606 802-819 19 19 19 607-624 538-855 20 20 20 574-591 760-780 21,47 21 21 676-693** 862-879** 23 23 23 637-654 1336-1353 24 24 24 496-516 772-792 26 26 26 553-570 772-789 27 27 27 685-702 799-819 28 28 28 592-609 778-798 29 29 29 538-555 757-774 30 30 30 814-831 943-962 31 31 31 571-588 790-807 32 32 32 514-831 1057-1074 33 33 33 553-570 718-735 34 34 34 568-585 796-816 38,55 38 38 553-573 709-729 39 39 39 556-573 718-735 41 41 41 598-615 784-801 42 42 42 547-567 715-735 43 43 43 580-597 844-861 45 45 45 640-657 943-963 46 46 46 565-582 781-801 49 49 49 589-609 754-771 51 51 51 565-582 1042-1059 52 52 52 598-615 829-846 56 56 56 697-714 877-897 8 and 40 8 562-579 1045-1062 25 25 529-549 703-723 35 non-functional H 11 gene 769-789* 1045-1065* 37 37 520-537 715-735 48 48 568-585 835-852 54 non-functional H21 gene 988-1008" 1344-1364** * See section 13 for choice of primers for the flagellin gene of H11 ** See section 13 for choice of primers for the flagellin gene of H21 ** See text WO 99/61458 PCT/AU99/00385 - 73 Table 3A Cloning, expression and identification of flagellin genes H type strain Primers used Annealing Plasmid Host strain Anti-serum H antigen from which for PCR temperature carrying the H used for which reacts encoded by the H antigen amplification of (oC) used for antigen gene expression with an E Coli the cloned gene was the H antigen PCR fliC deletion gene amplified gene amplification strain carrying the plasmid Hi #1868 & #1870 55 pPR1920 M2126 H1 H1 H2 #1868 & #1870 55 pPR1977 P5560 H2 H2 H3 #1868 & #1870 55 pPR1969 P5560 H16 H16 H4 #1878 & #1885 65 pPR1955 P5560 H4 H4 H5 #1868 & #1870 60 pPR1967 M2126 H5 H5 H6 #1868 & #1870 55 pPR1921 P5560 H6 H6 H7 #1868 & #1870 55 pPRi919 P5560 H7 H7 H9 #1868 & #1870 55 pPR1922 P5560 H9 H9 H10 #1868 & #1870 55 pPR1923 P5560 H10 H10 H11 #1868 & #1870 55 pPR1981 M2126 H11 H11 H12 #1868 & #1870 60 pPR1990 M2126 H12 H12 H14 #1868 & #1870 55 pPR1924 P5560 H14 H14 Hi5 #1868 & #1870 55 pPR1925 P5560 H15 H15 H17 #1878 & #1885 65 pPR1957 P5560 H4 H4 H18 #1868 & #1870 55 pPR1986 M2126 H18 H18 H19 #1868 & #1870 55 pPR1927 P5560 H19 H19 H20 #1868 & #1870 55 pPR1963 M2126 H20 H20 H21 #1868 & #1870 55 pPR1995 M2126 H21 H21 H23 #1868 & #1869 55 pPR1942 P5560 H23 H23 H24 #1868 & #1870 55 pPR1971 M2126 H24 H24 H26 #1868 & #1870 65 pPR1928 P5560 H26 H26 H27 #1868 & #1870 55 pPR1970 M2126 H27 H27 H28 #1868 & #1870 60 pPRi944 P5560 H28 H28 H29 #1868 & #1870 55 pPR1972 M2126 H29 H29 H30 #1868 & #1871 55 pPR1948 P5560 H30 H30 H31 #1868 & #1870 65 pPRi965 M2126 H31 H31 H32 #1868 & #1871 55 pPR1940 P5560 H32 H32 H33 #1868 & #1871 55 pPRi976 M2126 H33 H33 H34 #1868 & #1870 65 pPR1930 P5560 H34 H34 H38 #1868 & #1870 48 pPR1984 M2126 H38 H38 H39 #1868 & #1870 48 pPR1982 M2126 H39 H39 H41 #1868 & #1870 65 pPR1931 P5560 H41 H41 H42 #1868 & #1870 50 pPR1979 M2126 H42 H42 H43 #1868 & #1870 65 pPR1968 M2126 H43 H43 H45 #1868 & #1870 60 pPR1943 P5560 H45 H45 H46 #1868 & #1870 60 pPR1966 M2126 H46 H46 H49 #1868 & #1870 60 pPR1985 M2126 H49 H49 H51 #1868 & #1870 65 pPR1941 P5560 H51 H51 H52 #1868 & #1870 65 pPR1935 P5560 H52 H52 H56 #1868&#1870 50 pPR1978 M2126 H56 H56 WO 99/61458 PCT/AU99/00385 - 74 Table 3 B Oligonucleotide primers used for PCR amplification and cloning of H antigen genes #1868 5'- cat gce atg gca caa gtc att aat acc -3' NcoI #1869 5'- ata tgt cga ctt aac cct gca gca gag aca g -3' SalI #1870 5'- atg gat cct taa ccc tgc age aga gac ag -3' BamHI #1871 5' - aac tgc agt taa ccc tgt age aga gac ag -3' PstI #1872 5' - cgg at ccc gca gac tgg ttc ttg ttg at - 3 ' BamHI #1878 5' - cgg gat eca ct cta tcg age gcc tctct - 3' BamHI #1884 5'- gct ctaga cgc aga tca ttc age agg cc -3' XbaI #1885 5' - gcLgagac atg ttg gac act tog gtc gc - 3' XbaI WO 99/61458 PCT/AU99/00385 - 75 TABLE 4 Pool Strains of which chromosonal DNA included in the pool Source* No. 1 E. coli type strains for O serotypes 1, 2, 3, 4, 10, 16, 18 and 39 IMVSa 2 E. coli type strains for O serotypes 40, 41, 48, 49, 71, 73, 88 and 100 IMVS 3 E. coli type strains for O serotypes 102, 109, 119, 120, 121, 125, 126 and IMVS 137 4 E. coli type strains for O serotypes 138, 139, 149, 7, 5, 6, 11 and 12 IMVS 5 E. coli type strains for O serotypes 13, 14, 15, 17, 19ab, 20, 21 and 22 IMVS 6 E. coli type strains for O serotypes 23, 24, 25, 26, 27, 28, 29 and 30 IMVS 7 E. coli type strains for O serotypes 32, 33, 34, 35, 36, 37, 38 and 42 IMVS 8 E. coli type strains for O serotypes 43, 44, 45, 46, 50, 51, 52 and 53 IMVS 9 E. coli type strains for O serotypes 54, 55, 56, 57, 58, 59, 60 and 61 IMVS 10 E. coli type strains for O serotypes 62, 63, 64, 65, 66, 68, 69 and 70 IMVS 11 E. coli type strains for O serotypes 74, 75, 76, 77, 78, 79, 80 and 81 IMVS 12 E. coli type strains for O serotypes 82, 83, 84, 85, 86, 87, 89 and 90 IMVS 13 E. coli type strains for O serotypes 91, 92, 95, 96, 97, 98, 99 and 101 IMVS 14 E. coli type strains for O serotypes 103, 104, 105, 106, 107, 108 and 110 IMVS 15 E. coli type strains for O serotypes 112, 162, 113, 114, 115, 116, 117 and IMVS 118 16 E. coli type strains for O serotypes 123, 165, 166, 167, 168, 169, 170 and See b 171 17 E. coli type strains for O serotypes 172, 173, 127, 128, 129, 130, 131 and See c 132 18 E. coli type strains for O serotypes 133, 134, 135, 136, 140, 141, 142 and IMVS 143 19 E. coli type strains for O serotypes 144, 145, 146, 147, 148, 150, 151 and IMVS 152 * a. Institute of Medical and Veterinary Science, Adelaide, Australia b. 123 from IMVS; the rest from Statens Serum Institut, Copenhagen, Denmark c. 172 and 173 from Statens Serum Institut, Copenhagen, Denmark, the rest from

IMVS

WO 99/61458 PCT/AU99/00385 - 76 TABLE 5 Pool Strains of which chromosonal DNA included in the pool Source* No. 20 E. coli type strains for O serotypes 153, 154, 155, 156, 157, 158, 159 and IMVS 160 21 E. coli type strains for O serotypes 161, 163, 164, 8, 9 and 124 IMVS 22 As pool #21, plus E. coli 0111 type strain Stoke W. IMVS 23 As pool #21, plus E. coli 0111:H2 strain C1250-1991 See d 24 As pool #21, plus E. coli 0111:H12 strain C156-1989 See e 25 As pool #21, plus S. enterica serovar Adelaide See f 26 Y. pseudotuberculosis strains of O groups IA, IIA, IIB, IIC, III, IVA, IVB, See g VA, VB, VI and VII 27 S. boydii strains of serogroups 1, 3, 4, 5, 6, 8, 9, 10, 11, 12, 14 and 15 See h 28 S. enterica strains of serovars (each representing a different O group) IMVS Typhi, Montevideo, Ferruch, Jangwani, Raus, Hvittingfoss, Waycross, Dan, Dugbe, Basel, 65,:i:e,n,z,15 and 52:d:e,n,x,z15 d. C1250-1991 from Statens Serum Institut, Copenhagen, Denmark d. C1250-1991 from Statens Serum Institut, Copenhagen, Denmark e. C156-1989 from Statens Serum Institut, Copenhagen, Denmark f. S. enterica serovar Adelaide from IMVS g. Dr S Aleksic of Institute of Hygiene, Germany h. Dr J Lefebvre of Bacterial Identification Section, Laboratoroie de Sant& Publique du Qubbec, Canada WO 99/61458 PCT/AU99/00385 - 77 TABLE 6 Pool Strains of which chromosonal DNA included in the pool Source* No. 29 E. coli type strains for O serotypes 153, 154, 155, 156, 158, 159 and 160 IMVS 30 E. coli type strains for O serotypes 161, 163, 164, 8, 9, 111 and 124 IMVS 31 As pool #29, plus E. coli 0157 type strain A2 (0157:H19) IMVS 32 As pool #29, plus E. coli 0157:H16 strain C475-89 See d 33 As pool #29, plus E. coli 0157:H45 strain C727-89 See d 34 As pool #29, plus E. coli 0157:H2 strain C252-94 See d 35 As pool #29, plus E. coli 0157:H39 strain C258-94 See d 36 As pool #29, plus E. coli 0157:H26 See e 37 As pool #29, plus S. enterica serovar Landau See f 38 As pool #29, plus Brucella abortus See g See h 39 As pool #29, plus Y. enterocolitica 09 40 Y. pseudotuberculosis strains of O groups IA, IIA, IIB, IIC, III, IVA, IVB, VA, See i VB, VI and VII 41 S. boydii strains of serogroups 1, 3, 4, 5, 6, 8, 9, 10, 11, 12, 14 and 15 See j 42 S. enterica strains of serovars (each representing a different O group) Typhi, IMVS Montevideo, Ferruch, Jangwani, Raus, Hvittingfoss, Waycross, Dan, Dugbe, Basel, 65:i:e,n,z15 and 52:d:e,n,x,z15 43 E. coli type strains for O serotypes 1,2,3,4,10,18 and 29 IMVS 44 As pool #43, plus E. coli K-12 strains C600 and WG1 IVMS See k d. 0157 strains from Statens Serum Institut, Copenhagen, Denmark e. 0157:H26 from Dr R Brown of Royal Children's Hospital, Melbourne, Victoria f. S. enterica serovar Landau from Dr M Poppoff of Institut Pasteur, Paris, France g. B. Abortus from the culture collection of The University of Sydney, Sydney, Australia h. Y. enterocolitica 09 from Dr. K. Bettelheim of Victorian Infectious Diseases Reference Laboratory Victoria, Australia. i. Dr S Aleksic of Institute of Hygiene, Germany J. Dr J Lefebvre of Bacterial Identification Section, Laboratoroie de Sant6 Publique du Quebec, Canada k. Strains C600 and WG1 from Dr. B.J. Backmann of Department of Biology, Yale University, USA.

WO 99/61458 PCT/AU99/00385 - 78 M U u U U U U U U U U U 0 0 0 0 0 0 0 0 0 0 0 IO'D0 'D \0 IO In) \0 iO '.o In) Lr) Q.) 0. %.C J 0 t CD4 C.- C: C ) ) CD C wJ S4q~ C z s 0.4 -4 m 3\ -) 0 C: ) 00 LO -i N0 ON 0 CU300 00 LO '' ON 00 OI -r) Lf l r : ON in CY) N ON G\ In) M* eN C> D \0 0 0~r -- 4\ 0 N- 00 C\ --4 00) 00 u' O U 0 C CD LO In 1 4 00 0 0 '0 ,4 mn CC) Lr 00 00 \()'. 0 O Ini ad ON L- i) N ON - CO I >.C m \0 \D 0 Nt N N N N" 00 00 00 w w m 00 m~ CO\ C) 00 L ON N CD ~ 0\ 1 N .. ONN M4 Yl N4 000t 0 "a 4 6 00 U) ON CD N ON mO -1 \0 N Nl e ~ N 00 00 '0 -- - I I4 Iq ON~~~0 N N '0 -~~C .D!i) N Z ~ . w C C) CO a\ en ir 00 CD \0-0 00 00 \0 Q) 0 Z l 0 M N~ C\ 4 -4 0 In LO ON CD 00 0 Cl mO mO ILO N 0 00 m ~ eu W~ ,4 I I I .9 ~ (<u ~ In Cl O Cl N I ~4 N ~ ~ .'~u Nz CO ml '.0 Cl ON ' ON~c 0- - 0 n O -I-Ia 400) '0 m 5 m ~ m ~ -z -z 0

C

0 ~~~~~- 0u ONw ~ j I w CCO ~ C -~ ~ - - ~ 0 ) 0 CU 4- 4- 4- 4- 0u -cz lN C - -0 1. -Z. ~ u . - -~- WO 99/61458 PCT/AU99/00385 - 79 0 0 0 0 0 04 0 0 0 0 0 0 0 I' 'D \. .D LO 110 to '0 '0 . . 'f). C'4- 0 )C fn co 0 n MS. m-4 1 L) \ WC) C N . \D D .0 CD LO N ON C ) - o 14. -4 00 'j1 to) C4 LO 0 00 mO 00 **~ r-i r-q h .4 b , C ~ .0 N l O It- 0lO N N 0-t N \.D A ON N CO N 00 C CN ~ -4 NlC. C -4 N- 00 00 1 N '., vLO N 0 ,,, Q COC O 00 CA - O fUa NN CO N 3 O\ 00 Ij4 00 ONC\'. 'r-4 '- - ON '- C Z C) r4 Cl V)C w N 00 '. 0 N N N N '0C -4 ON N \'0 00 ON ON\ a% C C C O0 0N\N0 \0 00 0 N 00 \ 0c14 0 C) N It It'. 0 CY ON C C N r4 o ' ' 00 ON ON4 1- 1- 04 C C SN ON ON CC. ON 1 4 ;4 Cl Cl cy4 -4 C ON 0 0 C:) 00 ed ON ON) N -. 0 C-4 No ~ a\ CD CD00 N \'0 00 00 Q 01 o wI to0 0l z Nl ON a\ ON a,\ COC) C:) all 00 uco CDr '-I--0 ~~A _ * * WO 99/61458 PCT/AU99/00385 -80 0 0 0 0 0 0 0 0 0 0 0 0 0 110 [C) \.D tO \ \ f 0 \) 0 (UU 0 -0 Cf Cf) to CD) [CO CAJ c'J c %41 C C14 I'D \ '0 C tUo t a,\C cl 00 Id44 ICO C4 ) 110 00 co) 00 a( -- -- -\ Nt-C (ONa N mC N 00 a' CD~ -1 C4 Cl 0 CD .- " 0 c -t . cc CL > 00 ON a' ca - 4 C)m'4 ON t" \ 00 a\ C C)N~ - 0-4 CC),- \' 0 C4 0 a' .4q ON 'I- a'* 00 C) C4 -,t -t~' I0* C) CD 1- Cr LO '.O a') -1 CA 0 00 . 1u: 0 'f) '01 m C0a ~ 4 IH 01NO "0 00 a\' t 00 C, aN '.0 C0 0 0 G\ C) [C (U r4 -4 m~ 00 N \0 C) -C - 00 000C C4 LO 00 ON I- .. ' m- a',4.

WO 99/61458 PCT/AU99/00385 - 81 -- 4 u u u U U U U U U U U u U u u U 16 0~ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C A LO LO) LO LO) \0 %O LO UO \D LO) L mL A 1 0100000 0 0 00000 O'i o o aXI N ~0 = * C) * ) co CD C) C) CD * Co co r CO0 W 0 E* 00 * 00000 u z 0O 0. ot t"~~C LNUl 4 m a U). 10 'D m0 g '00 N 00 m ". "D -:I, LOwo , L N\ O1 \O HLm D m mN m M~ C l0 \o C) \N M M -4 1110 ,t _10 N NO O r- CD c N \ 1 m-4\DO N C - 1 --\0 -- N- -4 0 H a\ -- N C o 0 Co 0 0 a. Co Co N C .CO 10 N o N 0 L\ Ot\0 a\0 N aoc ,~~ ~ N -4 1-4 m. CO ml~ . O Cl ~ '0 ' w~ ~ ~ ~ 000 :)N ON CD C.J 0C0 Nc w r-4 " 4 " l C l C C CD CD Cl N l C Co Co 1 4$ 4 -4 -4 v -4 -1 - 4 ~-4 -4 -1 C r- - -6 0 00

-

I . I - - - S m m N C C C?.4 N 0 N ' Cl C'1 Cl '.01 N 10 M, 01~ U3 11.0 CD Nm m l 0 r- Co m O' C l N C oo S*0 0 ml Co CO t"4 l C) U-) It 10 \0 0 0-4 .

000' 0 C Co) cO m- t 10 \ cq Cl N N m U)7\ ON - 5 Z:! -- " l l C Cl !:Cl ! 011 Co Co M N all C')' Co -4 -4I ocj4 m , -4 m Z t '4 ' l C 0C CD ~ ' C4 Nl a\0' -4 C b 1-0 0' 0' 1 Co N(t 0C4 N 10 4 Cl r Cl44'0 00 C4 1 'h ~ j- 0 N10 0 01 0o 0 cl C l 10) Co 0 w 0 L0 V.)V. V. N N Nq wV (IV Q) Q) V.)0 0. 0. V) 4-) m. CIAI 4- z0 0 WO 99/61458 PCT/AU99/00385 -82 ~ u u u u u u u u u u u u u u u u 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 LOO LO -i Cco CC) C) 0 - ) C: Y O (D L ) C '4- Co I' *O\o \ ' D L)\D L 0 L O \ 00 U, ~C: CD CN N :N CO C N C) C) ICD C C IfC) o .0 O\ LC) Cl ',D 0, Itl IC) NO- 'f) \O ) a', 0 C' ) L Cl 06 Om 00 FN '0 00,- I) Ic) F4 '.0 0r ) b "M Iq) '.o o) NZ m4 M. -\.o C' LO CL CO ce -- CO LC N Cl ,- ~-r-lCl C 4 C . I) '0 -4 4 _ 14 -4 cq C ON CO N CO C LO q t Co .- q \O N '. q C > 10 0 I' 1-4 \D) all CO CO J N CO C l CO C 00 C N O \ C O N A.J '.0 ,-'. o 4 ON -4 '.N N of - - n C CO0 M ~ 4 t.1! m. C'4 C 4) 0) ON0 0 C) \0 (al \D r- \0 11 0\ C4 \0 N' C1 Cl C 0 - -4 r-4 .- ,-- N It Cl) 10 O \0 C le' CN (CI 4 C 1-4 85Cl C Cl Cl4 Clr ClD Cl -" ClCt >, C1 4 -C 4 C -4 r-4 ' -4 -4 - -4 N4 N N Cl l0 0) W -1 C 1 4 t C4 C4 " C4 " C4 C- C4 M4 0 I-I) -d -4 r -) N c M-- 00 C4 N C 4" L O L - -4 G\ 1-. C) 00 &f 00) -4 C)& 'R~ NO :s' 0 t - 2 N~ 00-4 N mn N N Cl C Cl t.0C N C) Cl) Cl 00 ~ 4- C iz !:! m.00 C D O Cl 0 -4 CO) NON Cl 4 N NO cO' (n Cl C20r4 4 C l C Cl IC) LC) ICO C-4 N 4 CO C4 C) C) 1 r- rCl N 04 r4 " INr ' 0 -1 N C CO Cl 0 4 C 4 NON Cl 4 N C l C O M ON ON4 C4 0' Cl~.4 4 4, - ' l C o1 4 ~ ~* N 1f) C4 ., 0 C1 -0 Ut 1 -Ni- C l ) ~ ~ 0 Zl CO N CO -~ C CO 0 C) '.0 cq0 a jt~ N CC) Nr Cl -4 ~~~~ COCCC !)C En (n V.) N N ) 1-4 V) V) o~~E o~ c)'-4.-4 ' C) 0 0 U)tzcuQ 0 ON\ r ( 00 1C) C4 U * * *

X.)

Claims (33)

1. A nucleic acid molecule encoding all or part of an E. coli flagellin protein, provided that the nucleic acid 5 molecule does not encode a protein expressed by the E. coli H1, H7, H12 or H48 type strains.

2. A nucleic acid molecule according to claim 1 wherein the molecule is derived from a fliC gene. 10

3. A nucleic acid molecule including all or part of a sequence according to any one of SEQ ID NOs:1 to 68.

4. A nucleic acid molecule consisting of all or part of 15 a sequence according to any one of SEQ ID NOs: 1 to 68.

5. A nucleic acid molecule according to any one of claims 1-4 wherein the molecule is from about 10 to 20 nucleotides in length. 20

6. A nucleic acid molecule according to claim 5 wherein the molecule is capable of hybridising to the central region of a flagellin gene from which the molecule is derived. 25

7. A nucleic acid molecule selected from the group of nucleic acid molecules shown in Table 3.

8. A method of detecting the presence of E. coli of a 30 particular H serotype in a sample, the method comprising the step of specifically hybridising at least one nucleic acid molecule derived from a flagellin gene, wherein the at least one nucleic acid molecule is specific for a particular flagellin gene associated with the H serotype, 35 to any E. coli in the sample which contain the gene, and detecting any specifically hybridised nucleic acid molecules, wherein the presence of specifically SUBSTITUTE SHEET (Rule 26) (RO/AU) WO99/61458 PCT/AU99/00385 -84 hybridised nucleic acid molecules identifies the presence of the H serotype in the sample.

9. A method according to claim 8 wherein the at least 5 one nucleic acid molecule is according to any one of claims 1 to 7.

10. A method according to claim 8 wherein the specifically hybridised nucleic acid molecules are 10 detected by Southern Blot analysis.

11. A method of detecting the presence of E. coli of a particular H serotype in a sample, the method comprising the step of specifically hybridising at least one pair of 15 nucleic acid molecules to any E. coli in the sample which contains the flagellin gene for the particular H serotype, wherein at least one of the nucleic acid molecules is specific for the particular flagellin gene associated with the H serotype, and detecting any 20 specifically hybridised nucleic acid molecules, wherein the presence of specifically hybridised nucleic acid molecules identifies the presence of the H serotype in the sample. 25

12. A method according to claim 11 wherein the at least one pair of nucleic acid molecules is according to any one of claims 1 to 7.

13. A method according to claim 11 wherein the 30 specifically hybridised nucleic acid molecules are detected by the polymerase chain reaction.

14. A method for detecting the presence of a particular 0 serotype and H serotype of E. coli in a sample, the 35 method comprising the following steps: (a) specifically hybridising at least one nucleic acid molecule derived from and specific for a gene SUBSTITUTE SHEET (Rule 26) (RO/AU) WO99/61458 PCT/AU99/00385 -85 encoding a transferase or a gene encoding an enzyme for the transport or processing of a polysaccharide or oligosaccharide unit, the gene being involved in the synthesis of a particular E. coli 0 antigen, to any E. 5 coli in the sample which contain the gene; (b) specifically hybridising at least one nucleic acid molecule derived from and specific for a particular flagellin gene associated with that H serotype, to any E. coli in the sample which contain the gene; and 10 (c) detecting any specifically hybridised nucleic acid molecules, wherein the presence of specifically hybridised nucleic acid molecules identifies the presence of the particular H serotype and 0 serotype of E. coli in the sample. 15

15. A method according to claim 14 wherein the at least one nucleic acid molecule of step (a) is selected from the group consisting of: wbdH (nucleotide position 739 to 1932 of Figure 5), 20 wzx (nucleotide position 8646 to 9911 of Figure 5), wzy (nucleotide position 9901 to 10953 of Figure 5), wbdM (nucleotide position 11821 to 12945 of Figure 5), wbdN (nucleotide position 79 to 861 of Figure 6), wbdO (nucleotide position 2011 to 2757 of Figure 6), 25 wbdP (nucleotide position 5257 to 6471 of Figure 6), wbdR (nucleotide position 13156 to 13821 of Figure 6), wzx (nucleotide position 2744 to 4135 of Figure 6) and wzy (nucleotide position 858 to 2042 of Figure 6). 30

16. A method according to claim 14 wherein the at least one nucleic acid molecule of step (a) is selected from the group of nucleic acid molecules shown in Tables 8, 8A, 9 and 9A. 35

17. A method according to claim 14 wherein the at least one nucleic acid molecule of step (b) is according to any one of claims 1 to 7. SUBSTITUTE SHEET (Rule 26) (RO/AU) WO99/61458 PCT/AU99/00385 -86

18. A method according to claim 14 wherein the specifically hybridised nucleic acid molecules are detected by Southern Blot analysis. 5

19. A method for detecting the presence of a particular 0 serotype and H serotype of E. coli in a sample, the method comprising the following steps: (a) specifically hybridising at least one pair of 10 nucleic acid molecules derived from and specific for a gene encoding a transferase or a gene encoding an enzyme for the transport or processing of a polysaccharide or oligosaccharide unit, the gene being involved in the synthesis of a particular E. coli O antigen, to any E. 15 coli in the sample which contain the gene; (b) specifically hybridising at least one pair of nucleic acid molecules derived from and specific for a particular flagellin gene associated with that H serotype, to any E. coli in the sample which contain the 20 gene; and (c) detecting any specifically hybridised nucleic acid molecules, wherein the presence of specifically hybridised nucleic acid molecules identifies the presence of the particular H serotype and 0 serotype of E. coli in 25 the sample.

20. A method according to claim 19 wherein the at least one pair of nucleic acid molecules of step (a) is selected from the group consisting of: 30 wbdH (nucleotide position 739 to 1932 of Figure 5), wzx (nucleotide position 8646 to 9911 of Figure 5), wzy (nucleotide position 9901 to 10953 of Figure 5), wbdM (nucleotide position 11821 to 12945 of Figure 5), wbdN (nucleotide position 79 to 861 of Figure 6), 35 wbdO (nucleotide position 2011 to 2757 of Figure 6), wbdP (nucleotide position 5257 to 6471 of Figure 6), wbdR (nucleotide position 13156 to 13821 of Figure 6), SUBSTITUTE SHEET (Rule 26) (RO/AU) WO99/61458 PCT/AU99/00385 -87 wzx (nucleotide position 2744 to 4135 of Figure 6) and wzy (nucleotide position 858 to 2042 of Figure 6).

21. A method according to claim 19 wherein the at least 5 one pair of nucleic acid molecules of step (a) is selected from the group of nucleic acid molecules shown in Tables 8, 8A, 9 and 9A.

22. A method according to claim 19 wherein the at least 10 one nucleic acid molecule of step (b) is according to any one of claims 1 to 7.

23. A method according to claim 19 wherein the specifically hybridised nucleic acid molecules are 15 detected by the polymerase chain reaction.

24. A method for detecting the presence of a particular O serotype and H serotype of E. coli in a sample, the method comprising the following steps: 20 (a) specifically hybridising at least one nucleic acid molecule derived from and specific for a gene encoding a flagellin associated with a particular E. coli H antigen serotype, to any E. coli carrying the gene and present in the sample; 25 and (b) detecting the at least one specifically hybridised nucleic acid molecule, wherein the at least one nucleic acid molecule is specific for the particular combination of 0 and H antigen. 30

25. A method according to claim 24 wherein the at least nucleic acid molecule is according to any one of SEQ ID NOS: 9, 55, 57 to 65. 35

26. A method for testing a food derived sample for the presence of one or more particular E. coli 0 antigens and H antigens, wherein the particular E. coli 0 and H SUBSTITUTE SHEET (Rule 26) (RO/AU) WO99/61458 PCT/AU99/00385 -88 antigens in the food derived sample are detected using the method of any one of claims 8, 11, 14 or 19.

27. A method for testing a faecal derived sample for the 5 presence of one or more particular E. coli 0 antigens and H antigens wherein the particular E. coli 0 and H antigens in the faecal derived sample are detected using the method of any one of claims 8, 11, 14 or 19. 10

28. A method for testing a patient or animal derived sample for the presence of one or more particular E. coli 0 antigens and H antigens wherein the particular E. coli 0 and H antigens in the patient or animal derived sample are detected using the method of any one of claims 8, 11, 14 15 or 19.

29. A kit for identifying the H serotype of E. coli, the kit comprising at least one nucleic acid molecule according to any one of claims 1 to 7. 20

30. A kit for identifying the H and 0 serotype of E. coli, the kit comprising: (a) at least one nucleic acid molecule derived from and specific for an E. coli flagellin gene; and 25 (b) at least one nucleic acid molecule derived from and specific for a gene encoding a transferase or a gene encoding an enzyme for the transport or processing of a polysaccharide or oligosaccharide unit, the gene being involved in the synthesis of a particular E. coli 0 30 antigen.

31. A kit according to claim 30 wherein the at least one nucleic acid molecule of (a) is selected from the group consisting of: 35 wbdH (nucleotide position 739 to 1932 of Figure 5), wzx (nucleotide position 8646 to 9911 of Figure 5), wzy (nucleotide position 9901 to 10953 of Figure 5), SUBSTITUTE SHEET (Rule 26) (RO/AU) WO99/61458 PCT/AU99/00385 -89 wbdM (nucleotide position 11821 to 12945 of Figure 5), wbdN (nucleotide position 79 to 861 of Figure 6), wbdO (nucleotide position 2011 to 2757 of Figure 6), wbdP (nucleotide position 5257 to 6471 of Figure 6), 5 wbdR (nucleotide position 13156 to 13821 of Figure 6), wzx (nucleotide position 2744 to 4135 of Figure 6) and wzy (nucleotide position 858 to 2042 of Figure 6).

32. A kit according to claim 30 wherein the at least one 10 nucleic acid molecule of (a) is selected from the group of nucleic acid molecules shown in Tables 8, 8A, 9 and 9A.

33. A kit according to claim 30 wherein the at least one 15 nucleic acid molecule of (b) is according to any one of claims 1 to 7. SUBSTITUTE SHEET (Rule 26) (RO/AU)