H-NS controls metabolism and stress tolerance in Escherichia coli O157:H7 that influence mouse passage

Background H-NS is a DNA-binding protein with central roles in gene regulation and nucleoid structuring in Escherichia coli. There are over 60 genes that are influenced by H-NS many of which are involved in metabolism. To determine the significance of H-NS-regulated genes in metabolism and stress tolerance, an hns mutant of E. coli O157:H7 was generated (hns::nptI, FRIK47001P) and its growth, metabolism, and gastrointestinal passage compared to the parent strain (43895) and strain FRIK47001P harboring pSC0061 which contains a functional hns and 90-bp upstream of the open-reading frame. Results The hns mutant grew slower and was non-motile in comparison to the parent strain. Carbon and nitrogen metabolism was significantly altered in the hns mutant, which was incapable of utilizing 42 carbon, and 19 nitrogen sources that the parent strain metabolized. Among the non-metabolized substrates were several amino acids, organic acids, and key metabolic intermediates (i.e., pyruvate) that limit carbon acquisition and energy generation. Growth studies determined that the parent strain grew in LB containing 14 to 15% bile or bile salts, while the hns mutant grew in 6.5 and 9% of these compounds, respectively. Conversely, log-phase cells of the hns mutant were significantly (p < 0.05) more acid tolerant than the parent strain and hns mutant complemented with pSC0061. In mouse passage studies, the parent strain was recovered at a higher frequency (p < 0.01) than the hns mutant regardless of whether log- or stationary-phase phase cells were orally administered. Conclusion These results demonstrate that H-NS is a powerful regulator of carbon and nitrogen metabolism as well as tolerance to bile salts. It is likely that the metabolic impairments and/or the reduced bile tolerance of the E. coli O157:H7 hns mutant decreased its ability to survive passage through mice. Collectively, these results expand the influence of H-NS on carbon and nitrogen metabolism and highlight its role in the ability of O157:H7 strains to respond to changing nutrients and conditions encountered in the environment and its hosts.


Background
Escherichia coli O157:H7 is a human pathogen that causes hemorrhagic colitis in humans and in some cases may incite hemolytic uremic syndrome [1]. Cattle are a principal reservoir and contaminated ground beef is most frequently implicated in foodborne outbreaks associated with this human pathogen [2,3]. Adult cattle harboring serotype O157:H7 strains are asymptomatic, and the numbers and duration of shedding are influenced by a number of factors including feed, the microbial composition of the intestinal tract, the time of year, and environmental sources [4]. The complex interplay between the bovine host and E. coli O157:H7 persistence has not been fully elucidated although Naylor et al. [5] recently demonstrated localization and adherence to mucosal epithelium in the recto-anal junction. The changing conditions encountered by E. coli O157:H7 in the bovine intestinal tract and the environment require that it closely control metabolism and stress protection.
H-NS is a cytoplasmic protein that is abundant in the E. coli nucleoid [6]. It plays a role in the condensation of chromosomal DNA by binding to curved DNA and consequently can influence the expression of many genes either directly or indirectly [7,8]. The level of H-NS in cells remains relatively constant throughout log-and stationary-phases although hns transcription is subject to auto repression [9,10]. The production of over 60 proteins in E. coli are either transcriptionally or translationally controlled by H-NS [8,9,11]. Several genes modulated by H-NS are responsive to changes in environmental conditions like pH, osmolarity, and temperature [8,9,[11][12][13][14][15]. The involvement of H-NS in modulating genes responsive to environmental stimuli make it an important regulator of virulence in several pathogens including Shigella and Salmonella [12,13,[16][17][18][19].
Although there has been considerable work on gene regulation by H-NS, there has been comparatively little work on its impact on bacterial metabolism and stress tolerance. Considering the varied and changing environments encountered by E. coli O157:H7, a detailed understanding of the consequences of modulators that are responsive to environmental conditions is particularly important. To this end, an hns mutant was constructed to define the growth and metabolic changes and any subsequent impact on gastrointestinal passage in this important human pathogen.

Results
Generation and confirmation of the hns::nptI mutant A 695-bp fragment that includes the hns structural gene and upstream regulatory region was amplified by PCR from genomic DNA of E. coli O157:H7 ATCC 43895. The nucleotide sequence of the EcoRI-BamHI fragment was determined and was identical to the sequenced gene in E. coli O157:H7 EDL933 [20]. Transconjugants resulting from the conjugation of pSC0063 from E. coli SM10 λpir to E. coli O157:H7 strain ATCC 43895 were kanamycin resistant, ampicillin sensitive, sucrose positive, and O157 positive. Confirmation of a double crossover in which wild-type hns was replaced with hns::nptI was confirmed by PCR (data not shown). PCR analysis of genomic DNA from the parental strain (ATCC 43895) with primers choi9917 and choi9918 generated a 375-bp fragment; whereas, genomic DNA from the hns::nptI mutant (FRIK47001P) resulted in a DNA fragment approximately 1.6-kb in length. The 1.6-kb fragment is in agreement with the projected size of the DNA fragment containing wildtype hns (375 bp) and the nptI gene (1.2 kb).

Growth and serological characteristics
The growth of the hns mutant (FRIK47001P) in LB at 37°C was slower and reached a lower maximum optical density than the parent strain or FRIK47001P containing pSC0061 (data not shown). The doubling times of the parent, complemented, and mutant strains were 27, 34, and 36 minutes, respectively. The maximum OD 600 nm achieved by the parent and complemented strains was 1.3 to 1.4 after 18 h of incubation while the hns mutant reached a maximum OD 600 nm = 1.0. Similar to the results in LB, the growth of the hns mutant in LB supplemented with mucin or in fecal slurries was slower and reached a lower final density than the parent strain. In addition, the colony size of the hns mutant on agar media was smaller than the parent and complemented strains. The presence of kanamycin in the growth media did not influence the growth of the mutant. These results are in agreement with results with hns mutants of non-pathogenic E. coli which had greater doubling times than the parent strains [11,21]. One contributor to the longer doubling time is the increased σ s level in the hns mutant, which is partially suppressed in hns rpoS double mutants [11]. The hns mutant tested negative for the H7 antigen and was nonmotile in motility medium (data not shown); however, the O157 antigen was still detected. The absence of H7 is consistent with previous data showing that H-NS is involved in the control of flagella gene expression and that hns mutants are non-motile [22,23].

Growth inhibition by pH, NaCl, and bile
In addition to slower growth, the hns mutant was more susceptible to inhibition by low pH, NaCl, bile (Table 1) and cefixime (data not shown) compared to the parent and complemented strains. Low temperature further slows the growth of hns mutants [24]. Thus, the growth of the mutant was further restricted when low temperature was combined with another inhibitor. The minimum growth pH of the hns mutant at the different temperatures was 4.5 to 5.0 which was 0.5 to 1.0 pH unit higher than the parent and complemented strains with the greatest difference observed at the lowest incubation temperature (16.5°C). Both the parent and complemented strains grew at pH 4.0 at each incubation temperature. Similar to the results with pH, the hns mutant grew in LB supplemented with a maximum of 4-6% NaCl compared to 6-7% for the parent and complemented strains. The test strains tolerated the highest concentration of NaCl when incubated at 25°C. There was one exception in which the hns mutant complemented with pSC0061 grew in a maximum of 4% NaCl at 16.5°C which was equivalent to the maximum level tolerated by the hns mutant. It is unclear why the complemented strain differed from the parent strain in this case where in all other growth conditions the results were the same. An unexpected finding was the ability of the E. coli O157:H7 to grow in LB containing bile compounds at 14-15%. The hns mutant was significantly more sensitive to bile salts and bovine bile preparations and grew in the presence of a maximum of 6.5% and 9.0% of these compounds, respectively. The growth of the strains in the presence of bile was further studied in LB containing 5% bile salts. Although the hns mutant grew, it had a greatly extended lag period (approximately 40 h) in comparison to the parent and complemented strains (<12 h) (Figure 1).

Acid tolerance
The acid tolerance of the parent, hns mutant, and complemented strains was tested using both log-and stationaryphase cells. The survival of log-phase cells of the hns mutant was significantly greater (p < 0.05) than the parent strain when challenged in synthetic gastric fluid (pH 2.0) ( Figure 2A). These findings are consistent with the derepression of σ s (rpoS)-regulated genes in the absence of H-NS. Stationary phase increased the acid tolerance com-pared to log phase in all three strains and the numbers decreased by less than 1 log 10 CFU/ml after 3 h of acid challenge ( Figure 2B). However, the parent and hns-com-Effect of bile salts on the growth of E. coli O157:H7 43895 (parent) and hns mutant (FRIK47001P)   plemented strains exhibited similar acid tolerance that was significantly greater (p < 0.05) than the hns mutant.
In contrast to the findings with acid challenges conducted in synthetic gastric fluid that is acidified with hydrochloric acid, there was no difference in the survival of the parent and hns mutant strains challenged in LB acidified with acetic or lactic acids (data not shown).

Glutamate decarboxylase activity
In an attempt to discern the reduced acid tolerance in stationary-phase cells of the hns mutant, glutamate decarboxylase (GAD) activity was measured (ΔpH/h/10 8 CFU). The hns mutant had significantly less (p < 0.01) GAD activity than the parent strain and strain FRIK47001P complemented with pSC0061 (p < 0.001) when examining stationary-phase cells (data not shown). These findings provide one possible explanation for the reduced acid tolerance in comparison to the parent strain of the hns mutant in stationary phase.

Sole carbon and nitrogen sources
Metabolic characterization using Phenotype Microarrays showed that the hns mutant was incapable of utilizing 42 carbon and 19 nitrogen sources that the parent strain metabolized (Tables 2 and 3). The ability of the hns mutant to uptake and/or use a variety of substrates as sole carbon sources was significantly altered. The parent strain was able to use six amino acids as sole carbon sources that the hns mutant was unable to utilize. The mutant did not use several disaccharides (rhamnose, lactulose, sucrose, maltose) and trehalose, dulcitol, D-ribose, and dextrin. In addition, eight organic acids, including pyruvate, were not used as a sole carbon sources by the hns mutant (Table 3). H-NS is involved in the regulation of genes associated with carbohydrate metabolism, like the maltose operon [25], but there were no carbon sources that were metabolized by the mutant that were not utilized by the parent strain. The parent strain metabolized six amino acids as sole nitrogen sources that the hns mutant was unable to use; however, the mutant was capable of using several dipeptides (Table 3). Nitrite was one nitrogen source that was utilized by the hns mutant but not the parent strain (data not shown). The findings from this metabolic characterization greatly expand the role of H-NS as a regulator of carbon and nitrogen metabolism in E. coli.

Gastrointestinal passage
Testing of fecal samples prior to inoculation determined that mice were E. coli O157:H7 negative and that few bacteria grew on the selective agar that were similar to the inoculation strain (i.e., sorbitol negative). The administration of the parent strain, hns mutant (FRIK47001P), and hns mutant complemented with pSC0061 did not result in signs of illness in ICR mice. Although mice were tested for 3 days post inoculation, all mice that shed the inoculated E. coli O157:H7 strain tested positive one day after inoculation and some mice continued to shed the strain at day 2. No fecal samples tested positive at day 3. The results from three trials are shown in Table 4. The parent Acid tolerance of E. coli O157:H7 43895 (parent) and hns mutant (FRIK47001P)

Discussion
E. coli O157:H7 must respond to changing environmental conditions whether it is present in the gastrointestinal tract or the environment [26]. Metabolic pathways must be carefully controlled and stress-protection systems triggered in a timely manner to effectively compete and survive. H-NS is a major component of the bacterial nucleoid, having pleiotropic effects on gene expression, genome stability and DNA recombination [9,27]. H-NS is

Utilized by the parent strain but not the hns mutant
involved in the control of several genes in E. coli associated with the metabolism of specific substrates as well as stress protection [8,9,15]; such as Lrp (leucine-responsive regulatory protein), σ s (rpoS), and two-component sys-tems (PhoP, EvgA and YedW) [8]. Thus, H-NS is a master regulator affecting the expression of individual genes and regulons with several of the phenotypes observed in hns mutants occurring as a result of indirect effects on gene expression [8,27].
In our studies of an hns mutant of E. coli O157:H7, the growth was slower than the parent strain. Although it is not completely understood, the increased doubling time in hns mutants is partially related to the increased level of σ s and the possibility of secondary site mutations that affect growth [11]. H-NS negatively influences the level of σ s at the posttranscriptional level [9,28,29], and the level of σ s increases in log-phase cells of hns mutants [11,30]. Another factor that contributes to the level of σ s in hns mutants is the increase in the heat-shock chaperone DnaK which enhances σ s stability [15]. The increase in σ s results in the production of σ s -regulated proteins some of which are involved in stress protection [8,15,30]. Results from this study using phenotypic microarrays showed that H-NS has a significant impact on carbon and nitrogen metabolism that likely impact the growth of the mutant (Tables 3 and 4). Among the substrates that were not metabolized were several carbon and nitrogen sources, including several key metabolic intermediates (Figure 3) that likely influence growth and energy generation. It is probable that the slower growth of the hns mutant is a result of the combined effects of the increased level of σ s [11] and altered substrate metabolism. However, another possibility that cannot be excluded is the generation of a second site mutation that influenced growth [11,40].    The growth of the hns mutant was also more susceptible to inhibition by NaCl and bile salts than the parent strain (Table 1). These differences were accentuated by sub-optimal growth temperature, often encountered when outside of the intestinal tract. A significant finding from these growth comparisons was the high tolerance of the parent E. coli O157:H7 strain to bile and bile salts (15 and 14%, respectively), while the growth of the hns mutant was limited to a maximum of 9.0 and 6.5%, respectively. Bile salts are produced by the liver to breakdown fat and are deleterious to biological membranes. E. coli must tolerate bile salts that can reach concentrations of 20 mM in the duodenum and even higher in the small intestine [31]. The reduced capacity of the mutant to tolerate bile salts might be a consequence of membrane changes in the hns mutant. H-NS is known to influence expression of at least 5% of the genes in E. coli, many that encode for membrane proteins [8]. The possibility of an altered mem-brane is supported by the inhibition of the mutant by the cephalosporin-like antibiotic, cefixime, while the parent strain grew in media containing 0.4 μg cefixime/ml. Future studies with H-NS-deficient strains should be useful in identifying the proteins involved in bile tolerance in E. coli O157:H7 strains.

Amino acids and derivatives Peptides Other
Acid tolerance is one characteristic that influences the persistence of E. coli O157:H7 shedding in cattle [32] and may promote survival during gastric passage in humans [33]. The increased acid tolerance of the hns mutant in comparison to the parent strain in log-phase cells is consistent with the expression of σ s -regulated proteins [9]. In stationary-phase cells, the acid tolerance increased in all three strains (parent, hns mutant, and hns complemented strains) in comparison to log-phase cells. However, the hns mutant was less acid tolerant than the parent strain in stationary phase. The gad system comprised of GadA and Key carbon sources that were not metabolized by the E. coli O157:H7 hns mutant  [55,56]. Glycyl-L-proline was tested (c). α-ketoglutarate was not metabolized as a sole carbon source by neither the parent strain nor the hns mutant (d). GadB, two isoforms of glutamate decarboxylase, and gadC that encodes for a glutamate/γ-aminobutyrate antiporter [34][35][36][37][38] is one contributor to acid tolerance. Results from the phenotype microarrays showed that the hns mutant did not utilize glutamate as a sole carbon source; therefore, we compared glutamate decarboxylase activity in stationary-phase cells of the mutant and parent strains. The specific activity of glutamate decarboxylase was reduced in the hns mutant (p < 0.01) compared to the parent strain. The decreased specific acitivity of glutamate decarboxylase may explain the difference in acid tolerance between the hns mutant and parent strain in stationaryphase cells.

TCA cycle
In mouse passage studies, recovery of the hns mutant was significantly less than the parent strain ( Table 2). The mutant was recovered from a single mouse in each of three trials conducted with log-phase and stationaryphase cells. It is possible that the reduced ability of the hns mutant to survive mouse passage is a consequence of a depleted energy supply, resulting from its limited capacity to metabolize available substrates, which is necessary for the optimal functioning of homeostasis or stress protection systems (i.e., ATPase). Gene expression profiles of mucus-grown E. coli have identified genes involved in the catabolism of gluconate, N-acetylglucosamine, sialic acid, glucuronate, mannose, fucose, and ribose are induced and considered preferred carbon sources [39]. Mutants incapable of metabolizing these seven carbohydrates have a decreased capacity to colonize streptomycin-treated mice. Although the mouse passage assay employed in this study did not evaluate colonization, the results from Phenotype Microarrays found that five of the seven carbohydrates identified by Chang et al. [39] as preferred carbon sources were utilized by the hns mutant while the E. coli O157:H7 parent strain used all seven carbon sources (see Tables 2  and 3). In addition, the hns mutant grew in mucin-supplemented LB broth and fecal slurries to a concentration essentially equivalent to that of the parent strain although the lag-phase was longer and the growth rate slower than that of the parent strain which is similar to the results obtained in growth studies conducted in LB. Therefore, a defect in the ability to utilize a mucin carbohydrate is probably not the reason for the reduced passage of the hns mutant.
There are a number of possible antimicrobial factors in the gastrointestinal tract that may be responsible for the diminished capacity of the hns mutant to survive passage through ICR mice. Acetate can reach a concentration of 70 mM in the intestinal tract [41,42]. Although log-phase cells of the mutant were more acid tolerant in challenges conducted with hydrochloric acid, additional challenges were conducted with acetic and lactic acids because results from phenotype microarrays found that the mutant did not utilize acetate or lactate as sole carbon sources. It was reasoned that tolerance to these acids might be impaired, but acid challenges with lactic acid and acetic acid at pH 4.4 found no difference in the survival of the parent and hns mutant strains with log-or stationary-phase cells (data not shown). Another possibility is that bile salts may have affected passage of the hns mutant through mice. In addition to being inhibitory, bile in association with H-NS is known to repress virulence genes [43]. However, additional experiments are needed with bile-sensitive mutants and measurements of bile acid concentrations in the mouse intestinal tract to confirm that bile was responsible for the reduced recoveries of the hns mutant. Considering that the intestinal tract contains at least 500 bacterial species [39], it is also possible that one of the commensal microbes produces an antimicrobial factor that is detrimental to the hns mutant. Studies with gnotoxenic mice would be particularly valuable in addressing the contributions of competing flora on the survival of the hns mutant. Based upon the results from this study, the most plausible explanation for the reduced passage of the hns mutant through mice is the regulatory impairments that influence growth and stress protection.

Conclusion
Results from this study significantly expand the importance of H-NS in carbon and nitrogen metabolism ( Figure  3) as well as stress tolerance that collectively influence the ability of E. coli O157:H7 to respond to changing conditions it encounters in its host and the environment. This is the first report of the high tolerance of E. coli O157:H7 strains to bile and bile salts that may be of significance in localization and persistence within its bovine host and possibly during human infection. Studies are in progress to identify, which proteins that are influenced by H-NS contribute to bile tolerance. The results from this study also establish H-NS as a potential target for control of this important human pathogen.

Bacterial strains and culture conditions
The bacterial strains and plasmids used in this study are listed in Table 5. E. coli were grown in Luria-Bertani (LB) broth at 37°C. Antibiotics (Sigma Chemical Company, St. Louis, MO) were added to media when appropriate: ampicillin, 100 μg/ml; kanamycin, 50 μg/ml; and tetracycline, 10 μg/ml. For log-phase and stationary-phase cultures, 5 μl of an overnight culture was used to inoculate 5 ml of LB and incubated with shaking (150 rpm) at 37°C to an optical density at 600 nm = 0.5 (log phase, ~10 8 CFU/ml) and optical density at 600 nm = 1.1 (stationary phase, ~10 9 CFU/ml), respectively.

Generation and confirmation of an hns::nptI mutant
A DNA fragment containing the hns structural gene and flanking regions was amplified from genomic DNA of E. coli O157:H7 ATCC 43895 by PCR using oligonucleotide primers (choi0005, 5'-TTTGGATCCAAAGCCTGGCTT-GAAGAAGAGATG-3' and choi0006, 5'-TTTGAATTCTTTT-GAATTCCTTACATTCCTGGC-3') containing an EcoRI (choi0005) and BamHI (choi0006) site on the respective 5'-ends. After PCR amplification, the PCR fragment was digested with EcoRI and BamH1 and ligated into pBR322, previously digested with the same enzymes, to produce pSC0061. The nucleotide sequence of the cloned fragment was determined (University of Wisconsin-Biotechnology Center). The suicide vector, pCVD442, was used to generate hns mutant in E. coli O157:H7 ATCC 43895 by homologous recombination. The hns gene in pSC0061 was inactivated by insertion of nptI encoding for aminoglycoside 3'-phosphotransferase (confers resistance to kanamycin) digested from pUC4K [44] using PstI and gap filled with Klenow fragment. nptI was ligated with a partial PstI digest of pSC0061 carrying hns which has a PstI site present within its open reading frame (ORF). The construct (pSC0062) was digested with EcoR1 and BamH1 to liberate the 1.9-kb hns::nptI fragment. The fragment was blunt-ended with Klenow and ligated with SmaI-digested pCVD442 [45] to produce pSC0063. The constructed vector was conjugally mobilized from E. coli SM10 λ pir [46] carrying the tra gene to recipient strain E. coli O157:H7 43895. Conjugation and mutant selection were conducted as described previously [47]. Transconjugants that were kanamycin resistant, ampicillin sensitive and tested positive for the O157 antigen by latex agglutination (Oxoid, Basingstoke, England) were selected for further study. The primers choi9917 (5'-ATGAGCGAAGCACT-TAAAATTCTGAAC-3') and choi9918 (5'-TTCTTCCATT-GCTTTTTTGATTACAGC-3') were used to amplify a portion of the hns ORF and confirm the presence of nptI in hns (1.6-kb fragment) (data not shown). FRIK 47001 (hns::nptI) was complemented by transforming FRIK 47001 with pSC0061.

Phenotype MicroArrays
The analysis of sole carbon and nitrogen source utilization was conducted with Phenotype MicroArrays™ (Biolog, Hayward, Calif.). This method is used to identify the function and global interactions of a gene on cellular metabolism [48,49]. The parent E. coli O157:H7 strain ATCC 43895 and hns::nptI mutant (FRIK47001) were evaluated for their ability to use carbon and nitrogen sources using PM1, PM2, and a PM3 plates, respectively. Strains were grown on R2A agar (Biolog) overnight at 37°C and sufficient cells were transferred to inoculating fluid to equal the 85%T turbidity standard (Biolog). FRIK47001 was resuspended in inoculating fluid containing glucose rather than succinic acid when inoculating the PM3 plate because it was unable to use succinic acid as a carbon source. Each well was inoculated with 100 μl of the standardized cell suspension, incubated at 37°C for 48 h, and the results recorded.

Growth studies
Growth studies were conducted in LB adjusted to the appropriate pH or supplemented with NaCl, bile, bile salts, or cefixime using the parent, hns mutant, and hns mutant complemented with pSC0061. Cells were used to inoculate the adjusted LB medium at a final concentration of 10 5 CFU/ml and incubated at 16.5, 25, and/or 37°C without shaking. The minimum pH for growth was determined in the pH range of 3.5 to 7.0 at 0.5 increments. The following were added to LB at the following ranges: NaCl, 0 to 10% (0.5% increments); bile salts no.3 (Difco), 0 to 16%; bovine bile (Sigma), 0 to 16%; porcine stomach mucin (Sigma), 2.5 mg/ml; bovine submaxillary gland mucin (Sigma), 2.5 mg/ml; and cefixime (Lederle), 0.2 to 0.8 μg/ml. Inoculated tubes were incubated at the designated temperature and observed daily for turbidity for up to 21 days. The growth of the strains in the presence of bile salts and mucin were studied in more detail. Growth in LB supplemented with 5% bile salts or 2.5 mg mucin/ml was monitored spectrophotometrically at 600 nm using a Bioscreen Analyzer at 37°C for 48 h (Labsystems, Helsinki, Finland). Plates were shaken for 30s prior to each optical density reading. The growth of strains was also monitored in fecal slurries (mouse feces diluted 1:10 in sterile deionized water). The fecal slurries were inoculated at a final concentration of approximately 10 5 CFU/ml from overnight cultures grown in LB, and then incubated at 37°C. Samples were removed and the number of CFU/ml determined by plating on MacConkey sorbitol agar (MSA, Difco) supplemented with 2.5 μg potassium tellurite/ml (Fisher) and 0.05 μg cefixime/ml (Lederle Laboratories, Pearl River, NY) [50] and enumerating the number of sorbitol-negative colonies.

Acid challenge
Acid tolerance was assessed in synthetic gastric fluid adjusted to pH 1.5 and LB acidified with acetate or lactate (pH 4.4) as previously described [51]. Log (OD 600 = 0.5) and stationary (OD 600 = 1.1) phase cells were used to inoculate flasks containing 50 ml of synthetic gastric fluid or acidified LB to achieve a final concentration of ca. 10 5 CFU/ml. Flasks were incubated at 37°C with shaking (150 rpm). Samples were removed periodically and then plated on tryptic soy agar (TSA, Difco) using a Model D Spiral Plater (Spiral Systems Inc., Cincinnati, OH). The number of colony forming units (CFU) was determined after 24 h of incubation at 37°C. The survival curves shown are the mean values from at least three trials.

GAD activity
The cells were examined for GAD activity as previously described [52,53] with the following modifications. Strains were cultured overnight in LB at 37°C for 21 h. The cells from 1.0 ml were pelleted by centrifugation and washed with 0.9% NaCl. The cells were resuspended in 3.0 ml of GAD reagent (1 g of L-glutamic acid, 0.05 g of bromocresol green, 90 g of NaCl, and 3 ml of Triton X-100 per liter) and incubated for 3 h at 35°C. The color and pH change of the reaction mixture was recorded every 30 minutes. GAD activity was reported as the ΔpH/h/10 8 cells.
The mean values were calculated from at least three trials.

Mouse inoculation studies
The passage studies were conducted as previously described [32]. In brief, strains were grown in LB at 37°C to log and stationary phase. After harvesting and washing with PBS, the cells were resuspended in 10% sucrose or phosphate buffer (0.01 M, pH 7.4) to a concentration of 10 4 , 10 3 , or 10 2 CFU/ml and administrated to individually housed ICR mice (ca. 20 g, Harlan Sprague Dawley Inc., Madison, Wis). Three trials were conducted, and in each trial three mice were inoculated with each concentration of the test strain. Fecal samples were collected each day for 3 days after inoculation and tested for the presence or absence of E. coli O157:H7. Fecal samples were diluted (1:10) in modified EC broth supplemented with novobiocin (20 μg/ml) and incubated at 37°C with shaking (150 rpm) for 18-24 h. Following enrichment, 20 μl of Dynabeads ® anti-E. coli O157 (Dynal Biotech ASA, Oslo, Norway) was added to 0.5 ml of the enriched fecal culture and wash buffer (0.5 ml), incubated, and washed as described by the manufacturer. The resuspended beads were streaked on MacConkey sorbitol agar supplemented with potassium tellurite/ml and cefixime/ml [50]. The plates were incubated at 42°C for 24 h and examined for sorbitol-negative E. coli O157:H7 colonies. Suspect colonies were confirmed as O157 by agglutination (E. coli O157:H7 test kit, Oxoid). The minimum detection level of this procedure was approximately 100 CFU/g.

Statistical analyses
The data presented are the mean values from at least three trials. The acid tolerance of strains was compared using the inverse of slopes from linear regression plots of survivor curves. The data from acid challenges were analyzed for statistical differences by one-way analysis of variance. A z-test for was used to evaluate significant differences in the proportions of mice testing positive between inoculation strains [54]. All calculations were conducted using SigmaStat software (Jandel Scientific, San Rafael, Calif.).