Commensal Bacteria Influence Escherichia coli O157:H7 Persistence and Shiga Toxin Production in the Mouse Intestine
http://www.100md.com
感染与免疫杂志 2006年第3期
Department of Molecular Genetics, Biochemistry and Microbiology, University of Cincinnati, Cincinnati, Ohio 45267
Division of Infectious Diseases, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio 45229
ABSTRACT
The presence of commensal flora reduced colonization of Escherichia coli O157:H7 and production of Shiga toxin (Stx) in the murine intestine. Stx production was not detected in mice colonized with E. coli that were resistant to the Shiga toxin phage, but it was detected in mice colonized with phage-susceptible E. coli.
TEXT
Escherichia coli O157:H7 causes approximately 73,000 cases of food-borne illness each year in the United States (12). Diarrheal disease can progress to hemorrhagic colitis and, in about 10% of cases, hemolytic uremic syndrome develops. Shiga toxin (Stx), an AB5 toxin, is a major virulence factor of E. coli O157:H7. The A-subunit has N-glycosidase activity and inhibits protein synthesis by cleaving a specific adenine in the host cell rRNA, ultimately leading to cell death (5). The A-subunit is delivered to target cells by the B-pentamer. Pathogenic strains of E. coli O157:H7 can produce one or both of two antigenically distinct forms of Stx, Stx1 and Stx2 (17).
Both Stx1 and Stx2 are encoded on distinct lambda-like phages lysogenized in the E. coli O157:H7 chromosome (9, 14, 18). The phage late gene promoter controls expression of Stx. As a consequence, the Stx genes are not expressed during lysogeny. However, when the phages enter the lytic cycle, the late gene promoter is activated, and new phage particles and Stx are produced and released from the cell by bacterial lysis (25). Under laboratory growth conditions, a few bacteria will spontaneously enter the lytic cycle, resulting in low levels of phage and Stx production. However, some conditions will induce lytic phage growth in the majority of the bacteria in the population. For example, treatment with antibiotics such as the quinolones can induce the phage lytic cycle in E. coli O157:H7 (11), resulting in high levels of Stx production both in vitro and in a mouse model of disease (10, 27). Furthermore, epidemiological studies suggest that antibiotic treatment can increase the risk of developing hemolytic uremic syndrome (26), likely due to higher levels of Stx production.
Lytic infection of commensal E. coli has been shown to influence Stx production. Incubation of phage-susceptible laboratory strains of E. coli with the Stx-encoding phage in vitro resulted in lytic phage growth and high levels of Stx in the culture supernatant (8). Similar results were seen in vivo. Mice were coinfected with an E. coli strain lysogenized with Stx phage and either a phage-susceptible or phage-resistant laboratory strain of E. coli. Higher levels of fecal Stx were recovered when mice were coinfected with the phage-sensitive strain (8).
These studies suggest that human intestinal flora could also influence Stx production and, consequently, the severity of human disease. In a survey of human fecal isolates, about 10% of the E. coli strains were found to be susceptible to Stx phage (7). Interestingly, one fecal E. coli strain was found to actually neutralize the effects of Stx2 on Vero cells (6, 7). In this study, we modified the streptomycin-treated mouse model of E. coli O157:H7 infection (24) to assess the impact of nonpathogenic commensal E. coli on E. coli O157:H7 survival and, importantly, Stx production.
Streptomycin-treated mouse model of E. coli O157:H7 infection. The streptomycin-treated mouse model used was similar to that described previously (8). ECOR strains were obtained from the Michigan State University STEC Center ECOR collection (http://foodsafe.msu.edu/whittam/ecor/index.html). Spontaneous streptomycin-resistant mutants were selected for all strains (Table 1). Briefly, male 6-week-old CD-1 mice were given streptomycin in their drinking water to reduce the presence of facultative intestinal flora. Commensal E. coli was established by inoculating the mice with 109 phage-resistant, phage-susceptible, or toxin-neutralizing E. coli cells, and colonization was allowed to establish for 1 week. A control group did not receive commensal E. coli. The mice were challenged with 106 streptomycin-resistant E. coli O157:H7 strain 185 cells on day zero. All mice were housed separately. To determine the extent of bacterial survival starting on day zero, all the feces in the cage were collected daily, weighed, and diluted 1:5 (wt/vol) in phosphate-buffered saline. Bacterial load was determined by plating serial dilutions on MacConkey agar plates containing selective antibiotics. Persistent colonization of all E. coli strains had occurred as evidenced by recovery of the organisms for weeks after inoculation (data not shown).
A total of four trials were performed, two in each study. In the first study (trials 1 and 2), mice received E. coli O157:H7 and/or commensal E. coli. In the second study (trials 3 and 4), mice were treated with ciprofloxacin in addition to receiving the same strains as in the first study.
Influence of commensal E. coli on survival of E. coli O157:H7. In trial 1, replication of E. coli O157:H7 strain 185 occurred in the mice lacking commensal E. coli, and on days 1 to 3, about 108 to 109 CFU/g feces was recovered (Fig. 1). For comparison with the other mice in the same trial, these results are plotted on all three panels of Fig. 1 ("185 alone" data). The impact of commensal E. coli on survival of E. coli O157:H7 was examined in mice colonized with phage-resistant (Fig. 1A), phage-susceptible (Fig. 1B), or toxin-neutralizing (Fig. 1C) E. coli. All three strains had become established at similar levels, with about 107 to 108 CFU/g feces (Fig. 1, day 0). Infection with E. coli O157:H7 strain 185 did not affect survival of the phage-resistant strain, 158-R (Fig. 1A, days 1 to 3); however, reduced survival of the phage-susceptible strain, 183-S (Fig. 1B), and the Stx-neutralizing strain, 160-TN (Fig. 1C), was observed on day 3.
The presence of any strain of commensal E. coli reduced survival of E. coli O157:H7 (Fig. 1A to C) compared to the mice lacking commensal E. coli (Fig. 1A to C). For the phage-resistant strain, a statistically significant (P < 0.05) reduction in E. coli O157:H7 survival was seen on days 1 to 3 (Fig. 1A). For the phage-susceptible strain (Fig. 1B) and Stx-neutralizing strain (Fig. 1C), the difference was statistically significant (P < 0.05) only on days 1 and 2, corresponding to the times when the commensal bacteria were present in high numbers.
A second trial was performed, and very similar results were obtained (data not shown). The presence of any strain of commensal E. coli significantly reduced the ability of E. coli O157:H7 to survive in the mouse intestine. Interestingly, in trial 2 the phage-susceptible strain colonized very poorly; on day zero, only about 105 CFU/g of feces was recovered. However, even colonization at this low level reduced survival of E. coli O157:H7 by more than 100-fold.
Autoinducer production. E. coli can communicate in a cell-density-dependent manner through the process of quorum sensing (21, 22), mediated by the signaling molecules autoinducer 2 (AI-2) and AI-3. AI-3 has been shown to upregulate expression of the locus of enterocyte effacement genes for intimate attachment by E. coli O157:H7 (20), and production of AI-3 by commensal E. coli has been proposed to enhance intestinal colonization by E. coli O157:H7. Production of both AI-2 and AI-3 in E. coli requires the LuxS enzyme (20). We examined the ability of the E. coli strains used in this study to produce AI-2 as a marker for LuxS activity and production of AI-3. Bioluminescence by the Vibrio harveyi reporter strain BB170 (sensor 1– sensor 2+) was used to measure AI-2 production as described previously (19, 21). Light production was measured using a Luminoskan Ascent luminometer (Labsystems, Franklin, MA). Culture supernatant from the V. harveyi control strains BB120 (AI-1+ AI-2+) and BB152 (AI-1– AI-2+) induced a greater-than-10-fold increase in bioluminescence in V. harveyi reporter strain BB170, whereas supernatant from the negative control, E. coli DH5, did not cause a substantial induction in bioluminescence (Table 2). The three nonpathogenic E. coli strains and the E. coli O157:H7 strain used in this study all produced AI-2 (Table 2), as evidenced by a nearly 10-fold increase in bioluminescence. These results suggest that the contribution of autoinducer produced by commensal E. coli on enhancing colonization of E. coli O157:H7 is negligible compared to the strong inhibitory effects exerted by commensal E. coli.
Colicin production. The ability of commensal E. coli to reduce survival of E. coli O157:H7 could be due to competition for nutrients or colonization sites. Alternatively, the commensal E. coli could actively kill the E. coli O157:H7 by production of toxins such as colicins. The strains were tested for colicin production. None of the nonpathogenic E. coli produced colicins active against the indicator strains of DH5 or E. coli O157:H7 strain 185. However, both DH5 and E. coli O157:H7 strain 185 were killed by colicin B, which was produced by strain DM1187(pCLB1) (4). These results suggest that competition with the commensal E. coli, not active bactericidal activity, was likely responsible for the reduced survival of E. coli O157:H7.
Influence of commensal E. coli on Stx production. Fecal Stx levels were measured by enzyme-linked immunosorbent assay according to the manufacturer's instructions (Premier EHEC ELISA; Meridian Bioscience, Inc., Cincinnati, OH). In both trials 1 and 2, Stx was frequently detected in the feces of mice lacking commensal E. coli (Fig. 2). Increased production of Stx in the absence of commensal E. coli is likely due to the higher intestinal survival levels of E. coli O157:H7 in the absence of commensal E. coli.
While the presence of any of the strains of commensal E. coli caused a similar reduction in survival of E. coli O157:H7 (Fig. 1), the different strains of commensal E. coli were observed to exert differential effects on Stx production. Fecal Stx was never recovered from mice colonized with the phage-resistant strain, 158-R (Fig. 2A and B). On day 2 of trial 1, significantly less (P < 0.05) toxin was recovered from mice colonized with the phage-resistant strain, 158-R, and the phage-resistant toxin-neutralizing strain, 160-TN (Fig. 2A), compared to mice lacking commensal E. coli (Fig. 2A). In contrast, Stx was recovered from mice colonized with the phage-susceptible strain (183-S) on all 3 days. On day 2, the amount of Stx recovered from the mice colonized with the phage-susceptible strain was not statistically different from that in the mice lacking commensal E. coli. About 50-fold fewer E. coli O157:H7 microorganisms were recovered from mice colonized with the phage-susceptible E. coli compared to the mice lacking commensal E. coli (Fig. 1B), and the increased toxin production by fewer E. coli O157:H7 cells could be due to lytic infection of the phage-susceptible strain by the E. coli O157:H7 Stx-encoding phage.
Fecal toxin recovery from the second trial (Fig. 2B) was similar to that seen in the first trial. Stx was most commonly detected in the feces of mice lacking commensal E. coli. Similarly, Stx was rarely recovered from mice colonized with phage-resistant (158-R) or phage-resistant, toxin-neutralizing (160-TN) E. coli. However, unlike trial 1, for the mice colonized with the phage-susceptible strain (183-S), fecal Stx was detected only for one mouse on day 3. The phage-susceptible strain survived poorly in trial 2, with only about 105 CFU/g, compared to between 107 and 108 CFU/g for trial 1, and the failure to observe amplified Stx production could have been due to low numbers of phage-susceptible bacteria. However, as noted above, even this low colonization level significantly reduced survival of E. coli O157:H7.
Overall, in these two trials, Stx was detected in 17 of 36 fecal samples from mice lacking commensal E. coli. Stx was not detected in fecal samples from mice colonized with the phage-resistant strain, 158-R. Stx was detected in 1 of 36 fecal samples from mice colonized with the phage-resistant, toxin-neutralizing strain, 160-TN. In contrast, Stx was detected in 8 of 36 stool samples for mice colonized with the phage-susceptible strain, 183-S. Thus, the sensitivity or resistance to the Stx phage can influence Stx production in the intestinal environment.
Influence of ciprofloxacin on survival of E. coli O157:H7. Treatment with the antibiotic ciprofloxacin has been shown to induce the phage lytic cycle and result in increased Stx production in vitro and in vivo in mice infected with E. coli O157:H7 (27). We examined the influence of phage-susceptible and phage-resistant commensal E. coli on ciprofloxacin-induced Stx production. All of the strains used in this study were equally susceptible to ciprofloxacin (MIC, 0.015 μg/ml). Preliminary studies established that a 10-μg intraperitoneal dose of ciprofloxacin would allow the examination of the role of ciprofloxacin on phage induction with little effect on bacterial viability in vivo (data not shown). An experimental protocol similar to that described in trial 1 was used to study the effects of ciprofloxacin on E. coli O157:H7 survival and Stx production, except a single dose of 10 μg of ciprofloxacin was administered to mice (intraperitoneally) on day 2 post-E. coli O157:H7 infection.
In the first ciprofloxacin trial (trial 3), in the absence of commensal E. coli, E. coli O157:H7 strain 185 survived at very high levels, between 109 and 1010 CFU/g feces (Fig. 3). As observed in Fig. 1, the presence of any strain of commensal E. coli significantly reduced (P < 0.05) the survival of E. coli O157:H7 strain 185 (Fig. 3A to C). Recovery of all strains of E. coli was slightly reduced by ciprofloxacin treatment (Fig. 3, day 3). Results for the second ciprofloxacin trial (trial 4) were similar to those for trial 3 (data not shown).
Influence of ciprofloxacin on Stx production. As observed in trials 1 and 2, commensal E. coli affected Stx production in trials 3 and 4. Significantly more (P < 0.05) Stx was recovered from the feces of mice lacking commensal E. coli than from mice colonized with nonpathogenic E. coli on every day in both trials (Fig. 4). The impact of ciprofloxacin treatment can be seen on day 3. In trial 3 (Fig. 4A), in the mice lacking commensal E. coli a 60-fold increase in Stx was observed, with a mean value of 10,437 ng/g recovered on day 3 compared to a mean value of 175 ng/g on day 2, and this increase was statistically significant (Fig. 4).
Ciprofloxacin treatment had less of an impact on Stx production in mice colonized with commensal E. coli (Fig. 4). Stx was rarely detected in mice colonized with the phage-resistant E. coli strain, 158-R, or the phage-resistant, toxin-neutralizing strain, 160-TN. In contrast, in trial 3 on day 3, fecal Stx was detected in six of eight mice colonized with the phage-susceptible E. coli strain (183-S), and the amount of Stx recovered on day 3 after ciprofloxacin treatment was significantly greater (P < 0.05) than that recovered before ciprofloxacin treatment on day 1 or 2 (Fig. 4). Furthermore, on day 3 the amount of toxin recovered from the mice colonized with the phage-susceptible strain was significantly (P < 0.05) greater than that recovered from mice colonized with either of the phage-resistant strains (Fig. 4). However, there was still significantly more (P < 0.05) Stx in the feces of mice lacking commensal E. coli compared to mice with phage-susceptible E. coli.
In trial 4 (Fig. 4B), the response to ciprofloxacin was not as dramatic as observed in trial 3. Fecal Stx recovery in the mice lacking commensal E. coli in trial 4 was higher after ciprofloxacin treatment than before ciprofloxacin treatment; however, it was not as high as observed in trial 3. Less fecal Stx was recovered from the mice with commensal E. coli compared to the mice lacking commensal E. coli on all days.
The intestinal environment is very complex, and many factors, including the presence of commensal bacteria, can influence the colonization of enteric pathogens such as E. coli O157:H7. The administration of nonpathogenic bacteria, or probiotics, into the intestine has been suggested as a method to prevent future colonization by enteropathogens and protect against disease (reviewed in reference 2), and bacterial species such as Bifidobacteria breve (1), Clostridium butyricum (23), and Lactobacillus rhamnosus (16) inhibit growth of E. coli O157:H7 in the intestine.
Nonpathogenic E. coli can also reduce colonization by E. coli O157:H7. Miranda et al. (13) demonstrated that a human K-12 E. coli isolate could out-compete E. coli O157:H7 strain EDL933 in the murine intestine, and they proposed that competition for glycolytic substrates was responsible. In this study, commensal E. coli did not out-compete E. coli O1457:H7, but it did cause a modest decrease in their numbers. Reduced recovery of the E. coli O157:H7 strain was not due to autoinducer or colicin production. In studies carried out to 18 days, the E. coli O157:H7 strain was still recovered at high levels (105 to 107 CFU/g feces) when commensal E. coli was present. The E. coli O157:H7 strain used in this study was a clinical isolate, and it may be particularly well adapted to survive in the intestine.
Commensal E. coli may not always exert a beneficial effect. While all of the strains of nonpathogenic E. coli characterized in this study had similar abilities to reduce colonization of E. coli O157:H7, their impacts on Stx production were not the same. Production of Stx was more commonly observed in mice colonized with phage-sensitive E. coli than in mice colonized with phage-resistant E. coli, suggesting that lytic infection of the phage-susceptible strain contributed to the Stx production. Phage-susceptible strains of E. coli are not uncommon; about 10% of human fecal E. coli isolates were found to be susceptible to lytic infection and produce Stx when incubated with the phage (7).
We also examined the influence of a toxin-neutralizing strain on Stx production in vivo. In previous studies, an E. coli isolate from a healthy child was shown to neutralize the cytotoxic effects of Stx2 on Vero cells, but it did not neutralize the related toxin, Stx1 (6). Stx1 and Stx2 share 55% amino acid homology, and strains of E. coli O157:H7 can produce Stx1, Stx2, or both (17). However, the ability to produce Stx2 has been associated with progression to severe disease, including hemolytic uremic syndrome (3). Since Stx2 is more toxic than Stx1, the toxin-neutralizing strain could be useful for preventing severe disease in humans. Both 160-TN, the toxin neutralizing strain, and 158-R, the non-toxin-neutralizing strain, are resistant to the phage. A detectable level of Stx was rarely observed in mice colonized with either phage-resistant strain, and it was not possible to assess if toxin neutralization was occurring. The mice were treated with ciprofloxacin to increase Stx production; however, even after ciprofloxacin treatment, very little Stx was recovered from mice colonized with either phage-resistant strain. These studies suggest that phage resistance may be more important than toxin neutralization.
The intestinal environment cannot be replicated in vitro, and animal studies are absolutely essential for understanding human disease. Overall, we observed a tremendous variability with respect to colonization by the nonpathogenic E. coli and pathogenic E. coli O157:H7 and an even greater variability with respect to production of Stx. Fecal Stx levels ranged from below the limit of detection (25 ng/g) to about 1,000 ng/g for untreated mice, and in one case, treatment with ciprofloxacin resulted in over 200,000 ng/g. The composition of the intestinal bacteria exerts a strong influence on production of Stx, and we have shown that even low numbers of commensal E. coli can substantially reduce colonization by E. coli O157:H7. While we verified that streptomycin treatment eliminated endogenous E. coli as evidenced by lack of bacterial growth on MacConkey agar plates, streptomycin-resistant microbes such as bacterial anaerobes or even fungal species could be present and could also influence colonization by E. coli and Stx production.
Epidemiologic studies strongly suggest that the amount of Stx produced in the intestine during an E. coli O157:H7 infection is important in determining the severity of disease (26). The health implications of an extremely high toxin burden are profound and may determine whether infection by E. coli O157:H7 will lead to self-limiting diarrheal disease or life-threatening complications, such as hemolytic uremic syndrome. Understanding the complex factors that influence Stx production could hold the key to treating this important disease.
ACKNOWLEDGMENTS
We thank Joel Mortensen, Cincinnati Children's Hospital Medical Center, for the clinical E. coli O157:H7 isolate and Bonnie Bassler, Princeton University, for the V. harveyi strains.
This work was supported by grant R21-AI-02-008 to A.A.W. and grant T32-AI055406 to S.D.G.
REFERENCES
1. Asahara, T., K. Shimizu, K. Nomoto, T. Hamabata, A. Ozawa, and Y. Takeda. 2004. Probiotic bifidobacteria protect mice from lethal infection with Shiga toxin-producing Escherichia coli O157:H7. Infect. Immun. 72:2240-2247.
2. Bengmark, S. 1998. Ecological control of the gastrointestinal tract. The role of probiotic flora. Gut 42:2-7.
3. Boerlin, P., S. A. McEwen, F. Boerlin-Petzold, J. B. Wilson, R. P. Johnson, and C. L. Gyles. 1999. Associations between virulence factors of Shiga toxin-producing Escherichia coli and disease in humans. J. Clin. Microbiol. 37:497-503.
4. Bullock, J. O., S. K. Armstrong, J. L. Shear, D. P. Lies, and M. A. McIntosh. 1990. Formation of ion channels by colicin B in planar lipid bilayers. J. Membr. Biol. 114:79-95.
5. Endo, Y., K. Tsurugi, T. Yutsudo, Y. Takeda, T. Ogasawara, and K. Igarashi. 1988. Site of action of a Vero toxin (VT2) from Escherichia coli O157:H7 and of Shiga toxin on eukaryotic ribosomes. RNA N-glycosidase activity of the toxins. Eur. J. Biochem. 171:45-50.
6. Gamage, S. D., C. M. McGannon, and A. A. Weiss. 2004. Escherichia coli serogroup O107/O117 lipopolysaccharide binds and neutralizes Shiga toxin 2. J. Bacteriol. 186:5506-5512.
7. Gamage, S. D., A. K. Patton, J. F. Hanson, and A. A. Weiss. 2004. Diversity and host range of Shiga toxin-encoding phage. Infect. Immun. 72:7131-7139.
8. Gamage, S. D., J. E. Strasser, C. L. Chalk, and A. A. Weiss. 2003. Nonpathogenic Escherichia coli can contribute to the production of Shiga toxin. Infect. Immun. 71:3107-3115.
9. Huang, A., S. de Grandis, J. Friesen, M. Karmali, M. Petric, R. Congi, and J. L. Brunton. 1986. Cloning and expression of the genes specifying Shiga-like toxin production in Escherichia coli H19. J. Bacteriol. 166:375-379.
10. Kimmitt, P. T., C. R. Harwood, and M. R. Barer. 1999. Induction of type 2 Shiga toxin synthesis in Escherichia coli O157 by 4-quinolones. Lancet 353:1588-1589.
11. Kimmitt, P. T., C. R. Harwood, and M. R. Barer. 2000. Toxin gene expression by shiga toxin-producing Escherichia coli: the role of antibiotics and the bacterial SOS response. Emerg. Infect. Dis. 6:458-465.
12. Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607-625.
13. Miranda, R. L., T. Conway, M. P. Leatham, D. E. Chang, W. E. Norris, J. H. Allen, S. J. Stevenson, D. C. Laux, and P. S. Cohen. 2004. Glycolytic and gluconeogenic growth of Escherichia coli O157:H7 (EDL933) and E. coli K-12 (MG1655) in the mouse intestine. Infect. Immun. 72:1666-1676.
14. Plunkett, G., III, D. J. Rose, T. J. Durfee, and F. R. Blattner. 1999. Sequence of Shiga toxin 2 phage 933W from Escherichia coli O157:H7: Shiga toxin as a phage late-gene product. J. Bacteriol. 181:1767-1778.
15. Selander, R. K., D. A. Caugant, and T. S. Whittam. 1987. Genetic structure and variation in natural populations of Escherichia coli, p. 1625-1648. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 2. American Society for Microbiology, Washington, D.C.
16. Shu, Q., and H. S. Gill. 2002. Immune protection mediated by the probiotic Lactobacillus rhamnosus HN001 (DR20) against Escherichia coli O157:H7 infection in mice. FEMS Immunol. Med. Microbiol. 34:59-64.
17. Slutsker, L., A. A. Ries, K. D. Greene, J. G. Wells, L. Hutwagner, and P. M. Griffin. 1997. Escherichia coli O157:H7 diarrhea in the United States: clinical and epidemiologic features. Ann. Intern. Med. 126:505-513.
18. Smith, H. W., P. Green, and Z. Parsell. 1983. Vero cell toxins in Escherichia coli and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chickens and pigs. J. Gen. Microbiol. 129:3121-3137.
19. Sperandio, V., J. L. Mellies, W. Nguyen, S. Shin, and J. B. Kaper. 1999. Quorum sensing controls expression of the type III secretion gene transcription and protein secretion in enterohemorrhagic and enteropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 96:15196-15201.
20. Sperandio, V., A. G. Torres, B. Jarvis, J. P. Nataro, and J. B. Kaper. 2003. Bacteria-host communication: the language of hormones. Proc. Natl. Acad. Sci. USA 100:8951-8956.
21. Surette, M. G., and B. L. Bassler. 1998. Quorum sensing in Escherichia coli and Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 95:7046-7050.
22. Surette, M. G., M. B. Miller, and B. L. Bassler. 1999. Quorum sensing in Escherichia coli, Salmonella typhimurium, and Vibrio harveyi: a new family of genes responsible for autoinducer production. Proc. Natl. Acad. Sci. USA 96:1639-1644.
23. Takahashi, M., H. Taguchi, H. Yamaguchi, T. Osaki, A. Komatsu, and S. Kamiya. 2004. The effect of probiotic treatment with Clostridium butyricum on enterohemorrhagic Escherichia coli O157:H7 infection in mice. FEMS Immunol. Med. Microbiol. 41:219-226.
24. Wadolkowski, E. A., J. A. Burris, and A. D. O'Brien. 1990. Mouse model for colonization and disease caused by enterohemorrhagic Escherichia coli O157:H7. Infect. Immun. 58:2438-2445.
25. Wagner, P. L., M. N. Neely, X. Zhang, D. W. Acheson, M. K. Waldor, and D. I. Friedman. 2001. Role for a phage promoter in Shiga toxin 2 expression from a pathogenic Escherichia coli strain. J. Bacteriol. 183:2081-2085.
26. Wong, C. S., S. Jelacic, R. L. Habeeb, S. L. Watkins, and P. I. Tarr. 2000. The risk of the hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections. N. Engl. J. Med. 342:1930-1936.
27. Zhang, X., A. D. McDaniel, L. E. Wolf, G. T. Keusch, M. K. Waldor, and D. W. Acheson. 2000. Quinolone antibiotics induce Shiga toxin-encoding bacteriophages, toxin production, and death in mice. J. Infect. Dis. 181:664-670.(Shantini D. Gamage, Angel)
Division of Infectious Diseases, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio 45229
ABSTRACT
The presence of commensal flora reduced colonization of Escherichia coli O157:H7 and production of Shiga toxin (Stx) in the murine intestine. Stx production was not detected in mice colonized with E. coli that were resistant to the Shiga toxin phage, but it was detected in mice colonized with phage-susceptible E. coli.
TEXT
Escherichia coli O157:H7 causes approximately 73,000 cases of food-borne illness each year in the United States (12). Diarrheal disease can progress to hemorrhagic colitis and, in about 10% of cases, hemolytic uremic syndrome develops. Shiga toxin (Stx), an AB5 toxin, is a major virulence factor of E. coli O157:H7. The A-subunit has N-glycosidase activity and inhibits protein synthesis by cleaving a specific adenine in the host cell rRNA, ultimately leading to cell death (5). The A-subunit is delivered to target cells by the B-pentamer. Pathogenic strains of E. coli O157:H7 can produce one or both of two antigenically distinct forms of Stx, Stx1 and Stx2 (17).
Both Stx1 and Stx2 are encoded on distinct lambda-like phages lysogenized in the E. coli O157:H7 chromosome (9, 14, 18). The phage late gene promoter controls expression of Stx. As a consequence, the Stx genes are not expressed during lysogeny. However, when the phages enter the lytic cycle, the late gene promoter is activated, and new phage particles and Stx are produced and released from the cell by bacterial lysis (25). Under laboratory growth conditions, a few bacteria will spontaneously enter the lytic cycle, resulting in low levels of phage and Stx production. However, some conditions will induce lytic phage growth in the majority of the bacteria in the population. For example, treatment with antibiotics such as the quinolones can induce the phage lytic cycle in E. coli O157:H7 (11), resulting in high levels of Stx production both in vitro and in a mouse model of disease (10, 27). Furthermore, epidemiological studies suggest that antibiotic treatment can increase the risk of developing hemolytic uremic syndrome (26), likely due to higher levels of Stx production.
Lytic infection of commensal E. coli has been shown to influence Stx production. Incubation of phage-susceptible laboratory strains of E. coli with the Stx-encoding phage in vitro resulted in lytic phage growth and high levels of Stx in the culture supernatant (8). Similar results were seen in vivo. Mice were coinfected with an E. coli strain lysogenized with Stx phage and either a phage-susceptible or phage-resistant laboratory strain of E. coli. Higher levels of fecal Stx were recovered when mice were coinfected with the phage-sensitive strain (8).
These studies suggest that human intestinal flora could also influence Stx production and, consequently, the severity of human disease. In a survey of human fecal isolates, about 10% of the E. coli strains were found to be susceptible to Stx phage (7). Interestingly, one fecal E. coli strain was found to actually neutralize the effects of Stx2 on Vero cells (6, 7). In this study, we modified the streptomycin-treated mouse model of E. coli O157:H7 infection (24) to assess the impact of nonpathogenic commensal E. coli on E. coli O157:H7 survival and, importantly, Stx production.
Streptomycin-treated mouse model of E. coli O157:H7 infection. The streptomycin-treated mouse model used was similar to that described previously (8). ECOR strains were obtained from the Michigan State University STEC Center ECOR collection (http://foodsafe.msu.edu/whittam/ecor/index.html). Spontaneous streptomycin-resistant mutants were selected for all strains (Table 1). Briefly, male 6-week-old CD-1 mice were given streptomycin in their drinking water to reduce the presence of facultative intestinal flora. Commensal E. coli was established by inoculating the mice with 109 phage-resistant, phage-susceptible, or toxin-neutralizing E. coli cells, and colonization was allowed to establish for 1 week. A control group did not receive commensal E. coli. The mice were challenged with 106 streptomycin-resistant E. coli O157:H7 strain 185 cells on day zero. All mice were housed separately. To determine the extent of bacterial survival starting on day zero, all the feces in the cage were collected daily, weighed, and diluted 1:5 (wt/vol) in phosphate-buffered saline. Bacterial load was determined by plating serial dilutions on MacConkey agar plates containing selective antibiotics. Persistent colonization of all E. coli strains had occurred as evidenced by recovery of the organisms for weeks after inoculation (data not shown).
A total of four trials were performed, two in each study. In the first study (trials 1 and 2), mice received E. coli O157:H7 and/or commensal E. coli. In the second study (trials 3 and 4), mice were treated with ciprofloxacin in addition to receiving the same strains as in the first study.
Influence of commensal E. coli on survival of E. coli O157:H7. In trial 1, replication of E. coli O157:H7 strain 185 occurred in the mice lacking commensal E. coli, and on days 1 to 3, about 108 to 109 CFU/g feces was recovered (Fig. 1). For comparison with the other mice in the same trial, these results are plotted on all three panels of Fig. 1 ("185 alone" data). The impact of commensal E. coli on survival of E. coli O157:H7 was examined in mice colonized with phage-resistant (Fig. 1A), phage-susceptible (Fig. 1B), or toxin-neutralizing (Fig. 1C) E. coli. All three strains had become established at similar levels, with about 107 to 108 CFU/g feces (Fig. 1, day 0). Infection with E. coli O157:H7 strain 185 did not affect survival of the phage-resistant strain, 158-R (Fig. 1A, days 1 to 3); however, reduced survival of the phage-susceptible strain, 183-S (Fig. 1B), and the Stx-neutralizing strain, 160-TN (Fig. 1C), was observed on day 3.
The presence of any strain of commensal E. coli reduced survival of E. coli O157:H7 (Fig. 1A to C) compared to the mice lacking commensal E. coli (Fig. 1A to C). For the phage-resistant strain, a statistically significant (P < 0.05) reduction in E. coli O157:H7 survival was seen on days 1 to 3 (Fig. 1A). For the phage-susceptible strain (Fig. 1B) and Stx-neutralizing strain (Fig. 1C), the difference was statistically significant (P < 0.05) only on days 1 and 2, corresponding to the times when the commensal bacteria were present in high numbers.
A second trial was performed, and very similar results were obtained (data not shown). The presence of any strain of commensal E. coli significantly reduced the ability of E. coli O157:H7 to survive in the mouse intestine. Interestingly, in trial 2 the phage-susceptible strain colonized very poorly; on day zero, only about 105 CFU/g of feces was recovered. However, even colonization at this low level reduced survival of E. coli O157:H7 by more than 100-fold.
Autoinducer production. E. coli can communicate in a cell-density-dependent manner through the process of quorum sensing (21, 22), mediated by the signaling molecules autoinducer 2 (AI-2) and AI-3. AI-3 has been shown to upregulate expression of the locus of enterocyte effacement genes for intimate attachment by E. coli O157:H7 (20), and production of AI-3 by commensal E. coli has been proposed to enhance intestinal colonization by E. coli O157:H7. Production of both AI-2 and AI-3 in E. coli requires the LuxS enzyme (20). We examined the ability of the E. coli strains used in this study to produce AI-2 as a marker for LuxS activity and production of AI-3. Bioluminescence by the Vibrio harveyi reporter strain BB170 (sensor 1– sensor 2+) was used to measure AI-2 production as described previously (19, 21). Light production was measured using a Luminoskan Ascent luminometer (Labsystems, Franklin, MA). Culture supernatant from the V. harveyi control strains BB120 (AI-1+ AI-2+) and BB152 (AI-1– AI-2+) induced a greater-than-10-fold increase in bioluminescence in V. harveyi reporter strain BB170, whereas supernatant from the negative control, E. coli DH5, did not cause a substantial induction in bioluminescence (Table 2). The three nonpathogenic E. coli strains and the E. coli O157:H7 strain used in this study all produced AI-2 (Table 2), as evidenced by a nearly 10-fold increase in bioluminescence. These results suggest that the contribution of autoinducer produced by commensal E. coli on enhancing colonization of E. coli O157:H7 is negligible compared to the strong inhibitory effects exerted by commensal E. coli.
Colicin production. The ability of commensal E. coli to reduce survival of E. coli O157:H7 could be due to competition for nutrients or colonization sites. Alternatively, the commensal E. coli could actively kill the E. coli O157:H7 by production of toxins such as colicins. The strains were tested for colicin production. None of the nonpathogenic E. coli produced colicins active against the indicator strains of DH5 or E. coli O157:H7 strain 185. However, both DH5 and E. coli O157:H7 strain 185 were killed by colicin B, which was produced by strain DM1187(pCLB1) (4). These results suggest that competition with the commensal E. coli, not active bactericidal activity, was likely responsible for the reduced survival of E. coli O157:H7.
Influence of commensal E. coli on Stx production. Fecal Stx levels were measured by enzyme-linked immunosorbent assay according to the manufacturer's instructions (Premier EHEC ELISA; Meridian Bioscience, Inc., Cincinnati, OH). In both trials 1 and 2, Stx was frequently detected in the feces of mice lacking commensal E. coli (Fig. 2). Increased production of Stx in the absence of commensal E. coli is likely due to the higher intestinal survival levels of E. coli O157:H7 in the absence of commensal E. coli.
While the presence of any of the strains of commensal E. coli caused a similar reduction in survival of E. coli O157:H7 (Fig. 1), the different strains of commensal E. coli were observed to exert differential effects on Stx production. Fecal Stx was never recovered from mice colonized with the phage-resistant strain, 158-R (Fig. 2A and B). On day 2 of trial 1, significantly less (P < 0.05) toxin was recovered from mice colonized with the phage-resistant strain, 158-R, and the phage-resistant toxin-neutralizing strain, 160-TN (Fig. 2A), compared to mice lacking commensal E. coli (Fig. 2A). In contrast, Stx was recovered from mice colonized with the phage-susceptible strain (183-S) on all 3 days. On day 2, the amount of Stx recovered from the mice colonized with the phage-susceptible strain was not statistically different from that in the mice lacking commensal E. coli. About 50-fold fewer E. coli O157:H7 microorganisms were recovered from mice colonized with the phage-susceptible E. coli compared to the mice lacking commensal E. coli (Fig. 1B), and the increased toxin production by fewer E. coli O157:H7 cells could be due to lytic infection of the phage-susceptible strain by the E. coli O157:H7 Stx-encoding phage.
Fecal toxin recovery from the second trial (Fig. 2B) was similar to that seen in the first trial. Stx was most commonly detected in the feces of mice lacking commensal E. coli. Similarly, Stx was rarely recovered from mice colonized with phage-resistant (158-R) or phage-resistant, toxin-neutralizing (160-TN) E. coli. However, unlike trial 1, for the mice colonized with the phage-susceptible strain (183-S), fecal Stx was detected only for one mouse on day 3. The phage-susceptible strain survived poorly in trial 2, with only about 105 CFU/g, compared to between 107 and 108 CFU/g for trial 1, and the failure to observe amplified Stx production could have been due to low numbers of phage-susceptible bacteria. However, as noted above, even this low colonization level significantly reduced survival of E. coli O157:H7.
Overall, in these two trials, Stx was detected in 17 of 36 fecal samples from mice lacking commensal E. coli. Stx was not detected in fecal samples from mice colonized with the phage-resistant strain, 158-R. Stx was detected in 1 of 36 fecal samples from mice colonized with the phage-resistant, toxin-neutralizing strain, 160-TN. In contrast, Stx was detected in 8 of 36 stool samples for mice colonized with the phage-susceptible strain, 183-S. Thus, the sensitivity or resistance to the Stx phage can influence Stx production in the intestinal environment.
Influence of ciprofloxacin on survival of E. coli O157:H7. Treatment with the antibiotic ciprofloxacin has been shown to induce the phage lytic cycle and result in increased Stx production in vitro and in vivo in mice infected with E. coli O157:H7 (27). We examined the influence of phage-susceptible and phage-resistant commensal E. coli on ciprofloxacin-induced Stx production. All of the strains used in this study were equally susceptible to ciprofloxacin (MIC, 0.015 μg/ml). Preliminary studies established that a 10-μg intraperitoneal dose of ciprofloxacin would allow the examination of the role of ciprofloxacin on phage induction with little effect on bacterial viability in vivo (data not shown). An experimental protocol similar to that described in trial 1 was used to study the effects of ciprofloxacin on E. coli O157:H7 survival and Stx production, except a single dose of 10 μg of ciprofloxacin was administered to mice (intraperitoneally) on day 2 post-E. coli O157:H7 infection.
In the first ciprofloxacin trial (trial 3), in the absence of commensal E. coli, E. coli O157:H7 strain 185 survived at very high levels, between 109 and 1010 CFU/g feces (Fig. 3). As observed in Fig. 1, the presence of any strain of commensal E. coli significantly reduced (P < 0.05) the survival of E. coli O157:H7 strain 185 (Fig. 3A to C). Recovery of all strains of E. coli was slightly reduced by ciprofloxacin treatment (Fig. 3, day 3). Results for the second ciprofloxacin trial (trial 4) were similar to those for trial 3 (data not shown).
Influence of ciprofloxacin on Stx production. As observed in trials 1 and 2, commensal E. coli affected Stx production in trials 3 and 4. Significantly more (P < 0.05) Stx was recovered from the feces of mice lacking commensal E. coli than from mice colonized with nonpathogenic E. coli on every day in both trials (Fig. 4). The impact of ciprofloxacin treatment can be seen on day 3. In trial 3 (Fig. 4A), in the mice lacking commensal E. coli a 60-fold increase in Stx was observed, with a mean value of 10,437 ng/g recovered on day 3 compared to a mean value of 175 ng/g on day 2, and this increase was statistically significant (Fig. 4).
Ciprofloxacin treatment had less of an impact on Stx production in mice colonized with commensal E. coli (Fig. 4). Stx was rarely detected in mice colonized with the phage-resistant E. coli strain, 158-R, or the phage-resistant, toxin-neutralizing strain, 160-TN. In contrast, in trial 3 on day 3, fecal Stx was detected in six of eight mice colonized with the phage-susceptible E. coli strain (183-S), and the amount of Stx recovered on day 3 after ciprofloxacin treatment was significantly greater (P < 0.05) than that recovered before ciprofloxacin treatment on day 1 or 2 (Fig. 4). Furthermore, on day 3 the amount of toxin recovered from the mice colonized with the phage-susceptible strain was significantly (P < 0.05) greater than that recovered from mice colonized with either of the phage-resistant strains (Fig. 4). However, there was still significantly more (P < 0.05) Stx in the feces of mice lacking commensal E. coli compared to mice with phage-susceptible E. coli.
In trial 4 (Fig. 4B), the response to ciprofloxacin was not as dramatic as observed in trial 3. Fecal Stx recovery in the mice lacking commensal E. coli in trial 4 was higher after ciprofloxacin treatment than before ciprofloxacin treatment; however, it was not as high as observed in trial 3. Less fecal Stx was recovered from the mice with commensal E. coli compared to the mice lacking commensal E. coli on all days.
The intestinal environment is very complex, and many factors, including the presence of commensal bacteria, can influence the colonization of enteric pathogens such as E. coli O157:H7. The administration of nonpathogenic bacteria, or probiotics, into the intestine has been suggested as a method to prevent future colonization by enteropathogens and protect against disease (reviewed in reference 2), and bacterial species such as Bifidobacteria breve (1), Clostridium butyricum (23), and Lactobacillus rhamnosus (16) inhibit growth of E. coli O157:H7 in the intestine.
Nonpathogenic E. coli can also reduce colonization by E. coli O157:H7. Miranda et al. (13) demonstrated that a human K-12 E. coli isolate could out-compete E. coli O157:H7 strain EDL933 in the murine intestine, and they proposed that competition for glycolytic substrates was responsible. In this study, commensal E. coli did not out-compete E. coli O1457:H7, but it did cause a modest decrease in their numbers. Reduced recovery of the E. coli O157:H7 strain was not due to autoinducer or colicin production. In studies carried out to 18 days, the E. coli O157:H7 strain was still recovered at high levels (105 to 107 CFU/g feces) when commensal E. coli was present. The E. coli O157:H7 strain used in this study was a clinical isolate, and it may be particularly well adapted to survive in the intestine.
Commensal E. coli may not always exert a beneficial effect. While all of the strains of nonpathogenic E. coli characterized in this study had similar abilities to reduce colonization of E. coli O157:H7, their impacts on Stx production were not the same. Production of Stx was more commonly observed in mice colonized with phage-sensitive E. coli than in mice colonized with phage-resistant E. coli, suggesting that lytic infection of the phage-susceptible strain contributed to the Stx production. Phage-susceptible strains of E. coli are not uncommon; about 10% of human fecal E. coli isolates were found to be susceptible to lytic infection and produce Stx when incubated with the phage (7).
We also examined the influence of a toxin-neutralizing strain on Stx production in vivo. In previous studies, an E. coli isolate from a healthy child was shown to neutralize the cytotoxic effects of Stx2 on Vero cells, but it did not neutralize the related toxin, Stx1 (6). Stx1 and Stx2 share 55% amino acid homology, and strains of E. coli O157:H7 can produce Stx1, Stx2, or both (17). However, the ability to produce Stx2 has been associated with progression to severe disease, including hemolytic uremic syndrome (3). Since Stx2 is more toxic than Stx1, the toxin-neutralizing strain could be useful for preventing severe disease in humans. Both 160-TN, the toxin neutralizing strain, and 158-R, the non-toxin-neutralizing strain, are resistant to the phage. A detectable level of Stx was rarely observed in mice colonized with either phage-resistant strain, and it was not possible to assess if toxin neutralization was occurring. The mice were treated with ciprofloxacin to increase Stx production; however, even after ciprofloxacin treatment, very little Stx was recovered from mice colonized with either phage-resistant strain. These studies suggest that phage resistance may be more important than toxin neutralization.
The intestinal environment cannot be replicated in vitro, and animal studies are absolutely essential for understanding human disease. Overall, we observed a tremendous variability with respect to colonization by the nonpathogenic E. coli and pathogenic E. coli O157:H7 and an even greater variability with respect to production of Stx. Fecal Stx levels ranged from below the limit of detection (25 ng/g) to about 1,000 ng/g for untreated mice, and in one case, treatment with ciprofloxacin resulted in over 200,000 ng/g. The composition of the intestinal bacteria exerts a strong influence on production of Stx, and we have shown that even low numbers of commensal E. coli can substantially reduce colonization by E. coli O157:H7. While we verified that streptomycin treatment eliminated endogenous E. coli as evidenced by lack of bacterial growth on MacConkey agar plates, streptomycin-resistant microbes such as bacterial anaerobes or even fungal species could be present and could also influence colonization by E. coli and Stx production.
Epidemiologic studies strongly suggest that the amount of Stx produced in the intestine during an E. coli O157:H7 infection is important in determining the severity of disease (26). The health implications of an extremely high toxin burden are profound and may determine whether infection by E. coli O157:H7 will lead to self-limiting diarrheal disease or life-threatening complications, such as hemolytic uremic syndrome. Understanding the complex factors that influence Stx production could hold the key to treating this important disease.
ACKNOWLEDGMENTS
We thank Joel Mortensen, Cincinnati Children's Hospital Medical Center, for the clinical E. coli O157:H7 isolate and Bonnie Bassler, Princeton University, for the V. harveyi strains.
This work was supported by grant R21-AI-02-008 to A.A.W. and grant T32-AI055406 to S.D.G.
REFERENCES
1. Asahara, T., K. Shimizu, K. Nomoto, T. Hamabata, A. Ozawa, and Y. Takeda. 2004. Probiotic bifidobacteria protect mice from lethal infection with Shiga toxin-producing Escherichia coli O157:H7. Infect. Immun. 72:2240-2247.
2. Bengmark, S. 1998. Ecological control of the gastrointestinal tract. The role of probiotic flora. Gut 42:2-7.
3. Boerlin, P., S. A. McEwen, F. Boerlin-Petzold, J. B. Wilson, R. P. Johnson, and C. L. Gyles. 1999. Associations between virulence factors of Shiga toxin-producing Escherichia coli and disease in humans. J. Clin. Microbiol. 37:497-503.
4. Bullock, J. O., S. K. Armstrong, J. L. Shear, D. P. Lies, and M. A. McIntosh. 1990. Formation of ion channels by colicin B in planar lipid bilayers. J. Membr. Biol. 114:79-95.
5. Endo, Y., K. Tsurugi, T. Yutsudo, Y. Takeda, T. Ogasawara, and K. Igarashi. 1988. Site of action of a Vero toxin (VT2) from Escherichia coli O157:H7 and of Shiga toxin on eukaryotic ribosomes. RNA N-glycosidase activity of the toxins. Eur. J. Biochem. 171:45-50.
6. Gamage, S. D., C. M. McGannon, and A. A. Weiss. 2004. Escherichia coli serogroup O107/O117 lipopolysaccharide binds and neutralizes Shiga toxin 2. J. Bacteriol. 186:5506-5512.
7. Gamage, S. D., A. K. Patton, J. F. Hanson, and A. A. Weiss. 2004. Diversity and host range of Shiga toxin-encoding phage. Infect. Immun. 72:7131-7139.
8. Gamage, S. D., J. E. Strasser, C. L. Chalk, and A. A. Weiss. 2003. Nonpathogenic Escherichia coli can contribute to the production of Shiga toxin. Infect. Immun. 71:3107-3115.
9. Huang, A., S. de Grandis, J. Friesen, M. Karmali, M. Petric, R. Congi, and J. L. Brunton. 1986. Cloning and expression of the genes specifying Shiga-like toxin production in Escherichia coli H19. J. Bacteriol. 166:375-379.
10. Kimmitt, P. T., C. R. Harwood, and M. R. Barer. 1999. Induction of type 2 Shiga toxin synthesis in Escherichia coli O157 by 4-quinolones. Lancet 353:1588-1589.
11. Kimmitt, P. T., C. R. Harwood, and M. R. Barer. 2000. Toxin gene expression by shiga toxin-producing Escherichia coli: the role of antibiotics and the bacterial SOS response. Emerg. Infect. Dis. 6:458-465.
12. Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607-625.
13. Miranda, R. L., T. Conway, M. P. Leatham, D. E. Chang, W. E. Norris, J. H. Allen, S. J. Stevenson, D. C. Laux, and P. S. Cohen. 2004. Glycolytic and gluconeogenic growth of Escherichia coli O157:H7 (EDL933) and E. coli K-12 (MG1655) in the mouse intestine. Infect. Immun. 72:1666-1676.
14. Plunkett, G., III, D. J. Rose, T. J. Durfee, and F. R. Blattner. 1999. Sequence of Shiga toxin 2 phage 933W from Escherichia coli O157:H7: Shiga toxin as a phage late-gene product. J. Bacteriol. 181:1767-1778.
15. Selander, R. K., D. A. Caugant, and T. S. Whittam. 1987. Genetic structure and variation in natural populations of Escherichia coli, p. 1625-1648. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 2. American Society for Microbiology, Washington, D.C.
16. Shu, Q., and H. S. Gill. 2002. Immune protection mediated by the probiotic Lactobacillus rhamnosus HN001 (DR20) against Escherichia coli O157:H7 infection in mice. FEMS Immunol. Med. Microbiol. 34:59-64.
17. Slutsker, L., A. A. Ries, K. D. Greene, J. G. Wells, L. Hutwagner, and P. M. Griffin. 1997. Escherichia coli O157:H7 diarrhea in the United States: clinical and epidemiologic features. Ann. Intern. Med. 126:505-513.
18. Smith, H. W., P. Green, and Z. Parsell. 1983. Vero cell toxins in Escherichia coli and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chickens and pigs. J. Gen. Microbiol. 129:3121-3137.
19. Sperandio, V., J. L. Mellies, W. Nguyen, S. Shin, and J. B. Kaper. 1999. Quorum sensing controls expression of the type III secretion gene transcription and protein secretion in enterohemorrhagic and enteropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 96:15196-15201.
20. Sperandio, V., A. G. Torres, B. Jarvis, J. P. Nataro, and J. B. Kaper. 2003. Bacteria-host communication: the language of hormones. Proc. Natl. Acad. Sci. USA 100:8951-8956.
21. Surette, M. G., and B. L. Bassler. 1998. Quorum sensing in Escherichia coli and Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 95:7046-7050.
22. Surette, M. G., M. B. Miller, and B. L. Bassler. 1999. Quorum sensing in Escherichia coli, Salmonella typhimurium, and Vibrio harveyi: a new family of genes responsible for autoinducer production. Proc. Natl. Acad. Sci. USA 96:1639-1644.
23. Takahashi, M., H. Taguchi, H. Yamaguchi, T. Osaki, A. Komatsu, and S. Kamiya. 2004. The effect of probiotic treatment with Clostridium butyricum on enterohemorrhagic Escherichia coli O157:H7 infection in mice. FEMS Immunol. Med. Microbiol. 41:219-226.
24. Wadolkowski, E. A., J. A. Burris, and A. D. O'Brien. 1990. Mouse model for colonization and disease caused by enterohemorrhagic Escherichia coli O157:H7. Infect. Immun. 58:2438-2445.
25. Wagner, P. L., M. N. Neely, X. Zhang, D. W. Acheson, M. K. Waldor, and D. I. Friedman. 2001. Role for a phage promoter in Shiga toxin 2 expression from a pathogenic Escherichia coli strain. J. Bacteriol. 183:2081-2085.
26. Wong, C. S., S. Jelacic, R. L. Habeeb, S. L. Watkins, and P. I. Tarr. 2000. The risk of the hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections. N. Engl. J. Med. 342:1930-1936.
27. Zhang, X., A. D. McDaniel, L. E. Wolf, G. T. Keusch, M. K. Waldor, and D. W. Acheson. 2000. Quinolone antibiotics induce Shiga toxin-encoding bacteriophages, toxin production, and death in mice. J. Infect. Dis. 181:664-670.(Shantini D. Gamage, Angel)