TEXT

Top Abstract Text References

The importance of Shiga toxin (Stx)-producing Escherichia coli (STEC) has increased since a food-borne infection caused by enterohemorrhagic E. coli (EHEC) was first described (25). Cattle have been implicated as a principle reservoir of STEC (3, 5, 31). The STEC strains are classified into a large number of O serogroups (4, 22). However, the majority of outbreaks and/or sporadic cases of hemorrhagic colitis and hemolytic uremic syndrome have been caused by members of only a few serogroups, such as O26, O111, and O157 (14, 26, 28). The Stx are known to be essential virulence factors of EHEC (6, 21), but the Stx alone may not be sufficient to cause disease (2). The virulence of STEC is known to be mediated through Stxs and attachment factors. The genes coding for these virulence-associated properties are located on phages, a pathogenicity island called the locus of enterocyte effacement (LEE), and plasmids (1, 9, 15). In particular, genes of the LEE are almost always detected in EHEC and are always detected in enteropathogenic E. coli (EPEC) and attaching and effacing E. coli (AEEC), and they express the attaching and effacing (AE) lesion for enteroepithelial cells (9). The LEE consists of genes for adhering intimin (eae), translocated intimin receptor (tir), a type III secretion system, and other secreted proteins (9). These recently discovered mediators of AE lesions are often examined to characterize STEC strains isolated from human clinical samples (29).

The aims of the present study were to describe the prevalence of STEC in healthy cattle in the central part of Japan and to genetically characterize various virulence genes of the STEC strains isolated.

E. coli O157:H7 strains ATCC 43889, ATCC 43890, ATCC 43894, and ATCC 35150, used as positive control strains for stx genes, were obtained from the American Type Culture Collection. E. coli 166, VR299-2, and EPEC108, used as positive control strains for intimin , intimin , the bundle-forming pilus (bfp), and the "EPEC adherence factor" (EAF) plasmid (1, 19, 24), were derived from the stock culture collection of the National Veterinary and Food Research Institute. E. coli Fuku-07, Oki-04, Sai-11, and Shizu-19, previously isolated from cattle in Japan and identified as O157:H7 strains, were used as reference strains.

A total of 358 rectal stool grab samples were collected from healthy dairy cattle (87 calves less than 5 months old, 88 heifers 5 to 12 months old, and 183 cows) on 78 farms located in a region of central Japan called the Kanto-Koshinetsu region between May and October 1998. All rectal stools were sampled by veterinarians from governmental animal hygiene centers. The samples were placed in cool room (4 to 8°C) and taken to the laboratory for immediate processing (usually within 24 h). The stool samples were streaked on sorbitol-MacConkey agar (SMAC) (Eiken, Tokyo, Japan), and 1 g of each stool sample was enriched in 19 ml of modified E. coli broth (mEC) (Eiken). Both SMAC plates and mEC were incubated at 37°C for 16 to 18 h.

A 100-µl aliquot of an enriched mEC culture (described above) was used for stx-PCR (34). It was centrifuged at 13,500 × g for 1 min, and the pellet was washed with 1 ml of phosphate-buffered saline. The bacterial sediment was resuspended in 100 µl of distilled water and heated at 100°C for 5 min. Five microliters was used for stx-PCR performed with a pair of stx common PCR primers (Table 1) (34). (The positions of the stx conserved sequence primers for stx genes are as follows: sense primer VT com-u, positions 280 to 300; and antisense primer VT com-d, positions 778 to 797.) When a sample was negative in the stx-PCR, 1 µl of the PCR mixture was subjected to a nested stx-PCR (Table 1). (The positions of the stx conserved sequence primers for stx genes are as follows: sense primer VT com-nesF, positions 395 to 413; and antisense primer VT com-nesR, positions 547 to 565.)

An oligonucleotide probe specific for both stx1 and stx2 (Table 1) (34) was synthesized and labeled with alkaline phosphatase at the 5' end by using an AP-Oligonucleotide labeling kit (Boehringer, Mannheim, Germany). The DNA probe was used to detect STEC by colony hybridization. For each stool sample 20 sorbitol-positive colonies typical of E. coli were picked from the SMAC plate. When sorbitol-negative coliform colonies appeared on the agar, we also picked up to 20 such colonies. The colonies were examined for the presence of any stx genes by colony hybridization techniques. All stx-positive isolates were confirmed to be E. coli strains by conventional biochemical tests, and if needed, API 20E and/or Micro ID techniques were employed. Then, up to three stx-positive colonies per sample were randomly chosen and subjected to classification as stx type stx1 or stx2 by PCR (Table 1) (18). One isolate per animal was selected for further study if all stx-positive isolates showed the same stx profile. The STEC isolates selected were characterized by PCR to determine the presence of bfp, the EAF plasmid, EHEC hlyA, and the LEE genes eae, espA, espB, espD, and tir (Table 1). Intimin subtyping (, , , , or ) was carried out for eae-positive isolates. All PCRs included 25 or 30 cycles of amplification with the conditions described in Table 1.

Determination of the O-antigen group was carried out by the method described in the Denka Seiken protocol (Denka Seiken, Tokyo, Japan). A strain identified as an STEC strain was incubated on Trypticase soy agar (Difco) at 37°C for 18 h. Approximately 0.3 g of organisms was harvested and suspended in 3 ml of saline solution. After heating at 100°C for 30 min, the suspension was centrifuged at 900 × g for 20 min. The sediment was resuspended in 0.5 ml of saline and used as the O-antigen solution. Sera against each O-antigen group tested were obtained from Denka Seiken and the E. coli Reference Laboratory, Universidad de Santiago de Compostela, Lugo, Spain, or were prepared at the National Institute of Animal Health, Tsukuba, Japan.

Antimicrobial susceptibility was tested by using the following 14 antimicrobial agents, most of which are commonly used for therapeutic treatment in production animals in Japan (the exceptions are ciprofloxacin and nalidixic acid): aminobenzylpenicillin, ceftiofur, dihydrostreptomycin, kanamycin, thiamphenicol, chloramphenicol, oxytetracycline, chlortetracycline, erythromycin, josamycin, colistin, nalidixic acid, ciprofloxacin, and enrofloxacin. Aminobenzyl penicillin, dihydrostreptomycin, oxytetracycline, and chlortetracycline were purchased from Wako Pure Chemical Industries, Ltd., Osaka, Japan, and the other drugs were supplied by the National Veterinary Assay Laboratory, Tokyo, Japan. In vitro susceptibility tests were carried out by the agar dilution method by using the recommendations of the National Committee for Clinical Laboratory Standards (17). Briefly, all drugs were serially twofold diluted, and 1 ml of each dilution was mixed with 9 ml of Mueller-Hinton agar (Difco). After inoculation of 10³ to 10⁴ CFU of organisms, the plates were incubated at 37°C for 24 h. The MIC was the lowest concentration of antimicrobial agent that suppressed visible bacterial growth. Two reference strains, E. coli NIH J2 and Staphylococcus aureus 209P, were included as internal controls in the susceptibility study. Since breakpoints for drugs used in animals have not been established, a strain was defined as resistant when its growth could not be suppressed with four times the MIC at which 90% of the isolates were inhibited or with 16 times the MIC at which 50% of the isolates were inhibited.

In this study, all stx-positive organisms isolated from cattle were identified as E. coli on the basis of biochemical properties. Therefore, an stx-PCR-positive sample was defined as STEC positive in this report. The STEC prevalence values for cows and heifers were 127 of 183 animals (69%) and 58 of 88 animals (66%), respectively, and thus were higher than the value for calves (40 of 87 animals [46%]). However, nested stx-PCR could detect STEC in all the samples examined. As determined by stx colony hybridization for 20 colonies per sample, we recovered 962 (12.4%) STEC colonies from a total of 7,745 coliforms. Sorbitol-negative STEC were detected only rarely among the sorbitol-negative colonies (20 of 585 colonies [3.4%]). STEC were recovered from 17 calves (20%), 27 heifers (31%), and 35 cows (19%). Of 962 STEC isolates, 92 from 74 animals on 54 farms were selected and examined by serotyping, virulence factor analysis, and antimicrobial susceptibility testing. STEC were isolated by colony hybridization from 76 of 79 (96%) stx-PCR-positive samples but only from 3 of 123 stx-PCR-negative samples. The latter three samples might have contained PCR inhibitors, because there were enough STEC organisms in the samples to be detected by colony hybridization.

The 92 STEC strains selected belonged to 25 different O serogroups, and 18 strains were nontypeable. Only one isolate was a member of serogroup O157, and 37 of the 74 typeable isolates could be classified into five serogroups (O8, O26, O84, O113, and O116) (Table 2). A connection was found between the serogroups and the types of stx genes; all of the serogroup O26 strains contained stx1, whereas the serogroup O84, O113, and O116 strains contained stx2. Another connection was found between serogroup and the presence of eae. The majority of the strains belonging to serogroups O26 (eight of nine strains), O84 (all six strains), O103 (all three strains), and O111 (both strains) contained eae and the LEE-related genes espA, espB, espD, and tir (Table 2). The serogroup O26, O111, and O157 strains, which are most frequently detected as causal agents of hemorrhagic colitis and hemolytic uremic syndrome, together accounted for more than 10% of the STEC isolates examined (12 of 92 strains). These STEC strains originated from nine farms. Thus, of 78 (11.5%) of the cattle farms analyzed were reservoirs of potentially EHEC. All 92 STEC strains were negative for either bfp or the EAF plasmid. The EHEC hlyA gene was prevalent in STEC strains; 66 of 92 (72%) of the strains were EHEC hlyA positive as determined by PCR. Of the 22 eae-positive STEC strains, 16 could be intimin typed and placed in established subtypes. The eight eae-positive isolates belonging to O serogroup O26 were intimin type , and all three O serogroup O103 strains were intimin type ; the other eae-positive isolates were intimin type  or nontypeable (Table 2). The only O157 isolate and the reference O157 strains were classified as intimin type ; no intimin type  STEC strains were detected among the non-O157 serogroup strains (Table 2).

Only eight isolates were resistant to some of the antimicrobial drugs tested (Table 2). Six of these isolates were resistant to aminoglycosides (kanamycin and dihydrostreptomycin) and/or tetracyclines (chlortetracycline and oxytetracycline), and the other two showed resistance to colistin (Table 2). Resistance to these drugs, which are conventionally used in veterinary medicine, is common in E. coli strains from cattle. The LEE genes were found in all five drug-resistant strains which were members of O serogroups other than serogroup O8 but not in serogroup O8 strains.

In early reports workers described STEC infection rates in cattle of 8% for cows and 19% for calves in the United States (28), 17% for cows and 9% for bulls in Germany (16), and 9% for cows and 25% for calves in Canada (32). According to these reports, remarkable differences were found in the rates of recovery of STEC isolates for adult and young animals; however, these reports were published in the first half of the 1990s. In this study we recovered STEC by colony hybridization at slightly higher frequencies (20% for calves, 31% for heifers, and 19% for cows). Blanco et al. have reported that the STEC prevalence in Spain is 35% for cows and 37% for calves (5) and that there is not a significant difference between calves and cows. The use of different detection methods and the rapid development of the methods make it difficult to compare STEC prevalence rates reported for various geographic areas. For the same reason it is impossible to judge whether the prevalence rate in cattle has really increased as reflected in the reports from the last 10 years. However, the results of STEC screening by PCR indicate that STEC are widespread in healthy cattle in central Japan. Using nested stx-PCR, we detected stx genes from each cow. We suggest that cattle always discharge various amounts of STEC, but when the number is low, the STEC cannot be detected by stx-PCR. We found that many calves were STEC positive and STEC negative as determined by stx-PCR when several samples collected at 2- to 8-week intervals were examined (data not shown).

The stx genes have been detected in members of approximately 90 to 100 E. coli O serogroups (4, 33). Most STEC strains of bovine origin seem to belong to 20 O serogroups: O4, O8, O22, O25, O32, O45, O82, O84, O103, O111, O113, O116, O121, O136, O146, O153, O157, O171, O172, and OX3 (O174) (4, 22). The 92 STEC strains used in our study were classified into 25 O serogroups, but only 10 serogroups (O8, O22, O84, O103, O111, O113, O116, O136, O153, and O157) were the same as the major bovine STEC O serogroups. Less than one-half of the Japanese isolates (44 of 92 isolates) belonged to these major O serogroups. Moreover, a comparison of the serogroup data for Spanish (5) and Japanese cattle showed that members of 13 of the 25 Japanese O serogroups (O1, O15, O28ac, O38, O55, O73, O88, O104, O111, O119, O123, O125, and O158) were found only in Japanese cattle. Similarly, of the 27 Spanish O serogroups, 18 (O4, O9, O20, O41, O74, O77, O78, O82, O90, O91, O92, O105, O132, O146, O150, O162, O165, and O171) were represented only by Spanish STEC strains. The distributions of O serogroups in cattle were thus markedly different in Spain and Japan. A comparison of cattle STEC strains from France (23) and Japan produced similar results. Based on STEC O serogroups, it seems that the cattle in Japan are isolated from the cattle in European countries in spite of the internationalization of the livestock market.

To assess the hazard of bovine STEC for human health, it is necessary to analyze a broad range of virulence-associated genes from isolates obtained from diseased humans and from cattle. Human EHEC typically possess the LEE genes together with stx (12) and sometimes the per (plasmid-encoded regulation) gene group on the EAF plasmid, which activate the LEE cassette (11, 27). The LEE is a gene group that is essential for microorganisms to cause AE lesions in intestinal cells (12). As most STEC strains isolated from diseased humans carry the LEE cassette (29), a detailed analysis of these genes is justified. The LEE was present in 22 of 92 STEC strains from cattle, whereas bfp or the EAF plasmid could not be found in any of these strains. Organisms with the EAF plasmid seem to be isolated very rarely from cattle, as no reports have described E. coli isolates with the EAF plasmid from cattle samples. Moreover, no reports have described AE lesion expression of LEE-possessing E. coli in weaned calves or adult cattle.

As the size of the LEE cassette is 35 kbp and the LEE cassette contains 41 predicted open reading frames (7), it is impossible to determine by molecular techniques alone whether a strain is an AEEC (including EHEC and EPEC). Therefore, observation of AE lesion expression in cell cultures is a common biological assay used to confirm the presence of AEEC (13). Although the biological assay is very important for confirming bacterial pathogenicity of a clinical sample, using this assay would be a very difficult task for epidemiological investigations because of the number of isolates to be examined. As all AE lesion-positive organisms are included among strains that possess the LEE, detecting LEE genes by a simple method such as PCR should be a very useful means of screening for epidemiological studies. In order to avoid false-negative results for LEE, we developed several PCRs to detect part of espA, espB, espD, and tir, genes which are essential parts of the type III secretion system of the LEE. In this study, all 22 STEC strains that possessed LEE could be detected by the eae-PCR alone. This result was confirmed by other PCRs for the LEE genes. One strain, however, was negative for espA, and another was negative for tir. Thus, the strains could be subtyped on the basis of their LEE genes by using the new PCR assays. We expect that these new PCRs will be useful tools for analyzing the LEE cassette.

Finally, as a method for direct detection of EHEC O157 in human samples by using multiplex PCR has been developed (22), a combination PCR for stx and intimin type  may be an effective prescreening method for detection of EHEC O157 in cattle fecal samples because pathogenic O157 strains always contain a type  eae gene in the LEE.