Correlating Levels of Type III Secretion and Secreted Proteins with Fecal Shedding of Escherichia coli O157:H7 in Cattle

Abstract

The locus of enterocyte effacement (LEE) of Escherichia coli O157:H7 (O157) encodes a type III secretion system (T3SS) for secreting LEE-encoded and non-LEE-encoded virulence proteins that promote the adherence of O157 to intestinal epithelial cells and the persistence of this food-borne human pathogen in bovine intestines. In this study, we compared hha sepB and hha mutants of O157 for LEE transcription, T3SS activity, adherence to HEp-2 cells, persistence in bovine intestines, and the ability to induce changes in the expression of proinflammatory cytokines. LEE transcription was upregulated in the hha sepB and hha mutant strains compared to that in the wild-type strain, but the secretion of virulence proteins in the hha sepB mutant was severely compromised. This reduced secretion resulted in reduced adherence of the hha sepB mutant to Hep-2 cells, correlating with a significantly shorter duration and lower magnitude of fecal shedding in feces of weaned (n = 4 per group) calves inoculated with this mutant strain. The levels of LEE transcription, T3SS activity, and adherence to HEp-2 cells were much lower in the wild-type strain than in the hha mutant, but no significant differences were observed in the duration or the magnitude of fecal shedding in calves inoculated with these strains. Examination of the rectoanal junction (RAJ) tissues from three groups of calves showed no adherent O157 bacteria and similar proinflammatory cytokine gene expression, irrespective of the inoculated strain, with the exception that interleukin-1β was upregulated in calves inoculated with the hha sepB mutant. These results indicate that the T3SS is essential for intestinal colonization and prolonged shedding, but increased secretion of virulence proteins did not enhance the duration and magnitude of fecal shedding of O157 in cattle or have any significant impact on the cytokine gene expression in RAJ tissue compared with that in small intestinal tissue from the same calves.

INTRODUCTION

Enterohemorrhagic Escherichia coli (EHEC) O157:H7 (referred to here as O157) causes a broad spectrum of diarrheal illnesses, including uncomplicated diarrhea, hemorrhagic colitis (HC), and hemolytic-uremic syndrome (HUS) (34). O157 infections are the major cause of acute renal failure in young children in the United States (34). Shiga toxins encoded by the stx₁ and stx₂ genes (52) are the major virulence factors responsible for the development of HC and HUS. Among ruminants, cattle are considered the major reservoir for O157 and animals colonized with these bacteria serve as major direct and indirect sources of this pathogen (28). O157 colonizes the large intestines of cattle, especially the cecum and tissues of the rectoanal junction (RAJ) (7, 37). Colonization requires intimate adherence of O157 bacteria to intestinal epithelial cells, a process that culminates in the formation of characteristic histopathological lesions called attaching and effacing (A/E) lesions. A/E histopathology results from the rearrangements of cytoskeletal elements of mucosal epithelial cells and the formation of small pedestals that cup attached bacterial cells.

The genes responsible for the production of A/E lesions are located on a pathogenicity island termed the locus of enterocyte effacement (LEE) (31). The LEE of O157 is about 43.3 kb in size and contains 41 open reading frames (ORFs) that are organized into five major operons (LEE1 through LEE5), an organization that is identical to that of the LEE of enteropathogenic E. coli (EPEC) (13, 40). The expression of LEE is positively regulated by Ler, which is encoded by the ler gene of LEE1 (12). Several transcriptional regulators affect LEE expression by modulating the expression of Ler (3, 27, 51). In previous studies, we found that the product of the hha gene, which was originally identified as a hemolysin-modulating protein (39), exerts a negative effect on LEE expression by repressing the transcription of ler and that strains of E. coli O157:H7 with hha deletions showed increased LEE transcription and adherence to HEp-2 cells (47, 50).

Several genes contained in the LEE1, LEE2, LEE3, and LEE4 operons encode proteins for the synthesis of a type III secretion system (T3SS), which spans the inner and outer bacterial cell membranes in the form of a needlelike structure or a syringe (16, 23). This syringe is connected to a filament produced by the protein EspA of LEE4. The syringe and the EspA filament constitute the translocon or conduit of the T3SS. At the distal end of this translocon are LEE4-encoded proteins EspB and EspD, which make 3- to 5-nm pores in the host epithelial cell membranes through which various LEE-encoded and non-LEE-encoded proteins are secreted. The energy for the secretion of these proteins by the T3SS is provided by an ATPase encoded by the sepB gene of LEE3 (17, 25). The ATPase is associated with several other LEE proteins at the base of the T3SS on the cytoplasmic membrane. The loss of the sepB-encoded ATPase function because of either a deletion or transpositional inactivation of sepB abrogates the T3SS-mediated secretion of LEE-encoded proteins, resulting in reduced adherence and colonization of epithelial cells (11, 24, 25). The intimate adherence of bacteria to host epithelial cells occurs because of the interactions between a bacterial outer membrane adhesin (intimin) and a translocated intimin receptor (Tir) located in host epithelial cell membranes (15). Both intimin and Tir are the products of ORFs present in LEE5, but only Tir requires the T3SS for its translocation and subsequent insertion into the host epithelial cell membranes by adopting a hairpin loop conformation. Intimin, on the other hand, belongs to an autotransporter family of proteins, which uses its unique structural features to translocate to bacterial outer cell membranes (2).

The importance of LEE-encoded proteins in the colonization of animal intestines has been demonstrated by the inability of O157 mutants with eae, espA, espB, tir, or combinations of these genes disrupted or deleted to colonize the intestines of ruminant and nonruminants (1, 5, 8, 32, 35). However, there are no published reports correlating enhanced LEE expression and secretion of LEE-encoded proteins with increased intestinal colonization, persistence, and fecal shedding of O157 in cattle.

Here we show that the ability to secrete LEE-encoded proteins promotes the colonization and persistence of wild-type O157 and an isogenic hha deletion mutant strain in bovine intestines, in contrast to an isogenic hha sepB double deletion mutant strain lacking the ability to secrete LEE proteins. We also demonstrate that enhanced expression of LEE in the hha mutant strain resulted in increased T3SS activity with concomitant increases in the secretion of virulence proteins and enhanced in vitro adherence to HEp-2 cells but caused no significant enhancements of this mutant's colonization of and persistence in bovine intestines relative to those of the wild-type strain, which is capable of a lower basal level of T3SS activity. The ability of E. coli O157:H7 and its isogenic mutants to colonize and persist in the bovine intestinal tract was monitored by quantifying the shedding of these bacterial strains in animal feces and by examining their adherence to tissues collected from the RAJ at the termination of the experiment. The RAJ tissues were also examined for cytokine gene expression to determine if differential expression of LEE and secreted proteins in O157 would correlate with the induction of proinflammatory immune response indicators in the host.

MATERIALS AND METHODS

Bacterial strains, culture media, and growth conditions.

E. coli O157:H7 strain 86-24 (20) was used as the wild-type parent strain to construct isogenic mutants. For routine gene cloning, we used vector pCR2.1 TOPO or pCR XL and E. coli TOP10 as a bacterial host strain (Invitrogen, Carlsbad, CA). All E. coli strains were cultivated in Luria-Bertani broth (LB) or LB containing 1.5% agar (Sigma-Aldrich, St. Louis, MO). Dulbecco's modified Eagle medium (DMEM; Invitrogen, Carlsbad, CA) was used when bacterial cultures were needed for RNA isolation or trichloroacetic acid (TCA) precipitation of secreted proteins from culture supernatants. Media were also supplemented with antibiotics at the following final concentrations: streptomycin, 100 mg liter⁻¹; kanamycin, 50 mg liter⁻¹; ampicillin, 100 mg liter⁻¹.

Construction of O157 hha and hha sepB deletion mutants.

A previously described allelic replacement procedure (50) was used to construct in-frame hha and sepB deletions in E. coli O157:H7 strain 86-24, a spontaneous streptomycin-resistant mutant of a Shiga toxin 2-producing strain associated with a disease outbreak in Washington State (20, 53). We used plasmids pSM122 and pSM302, carrying temperature-sensitive origins of replication, to delete the hha and sepB genes, respectively, from strain 86-24. The hha sepB double deletion mutant was constructed in the hha mutant strain by using allelic-replacement plasmid pSM302. The recombinant procedures used for the construction of plasmids pSM122 and pSM302 have been described previously (49, 50).

Quantitative reverse transcriptase PCR (QRT-PCR).

An RNeasy Mini kit (Qiagen Inc., Valencia, CA) was used to isolate RNA from bacterial strains grown in DMEM to an A₆₀₀ of 1.0. RNA was treated with TURBO DNase for DNA removal (Applied Biosystems/Ambion, Austin, TX). QRT-PCR was performed by adding 25 ng of RNA, 0.75 μM (each) antisense and sense primers (Table 1), 0.25 μM TaqMan probe (labeled at the 5′ and 3′ ends with 6-carboxyfluorescein reporter and 6-carobxytetramethylrhodamine quencher dyes, respectively) to a QRT-PCR Master Mix (Agilent Technologies, Inc., Santa Clara, CA). The reaction mixtures were incubated in an Mx3005P QPCR system (Agilent Technologies, Inc., Santa Clara, CA) for cDNA synthesis (50°C for 30 min), amplification, and real-time detection of amplified products (95°C for 10 min; 40 cycles of 95°C for 30 s, 55°C for 60 s, and 72°C for 30 s) specific for LEE genes eae, espA, ler, escJ, and escV of O157 (14). A comparative quantification program of the Mx3005 system allowed the calibration of target gene expression in the mutant strains to the expression of the corresponding genes in the parent strain. In addition, the transcriptional level of each of the LEE genes in the QRT-PCR was normalized to the expression of rpoA, a gene whose expression is not affected by the hha or sepB mutation, in order to account for any minor variations in the amount of RNA across samples. The expression data for each target gene in the mutant strains were plotted as n-fold changes by adjusting the expression of the same gene to 1 in the wild-type calibrator strain.

Table 1.

Primers used for PCR and QRT-PCR

Primer	Nucleotide sequence (5′–3′)a	Locationb
E. coli primersc
lerF	GTAAACACCTTTCGATGAGTTCC	4688913–4688935
lerR	GAGTCGATTCAGAAGCAGATTAC	4689001–4689023
lerP	CGAGCGAGTCCATCATCAGGC	4688938–4688958
escJF	TACCTGGTGTTATCGATTGCAG	4679687–4679708
escJR	AGTTAACCTCTGGCGAGCTG	4679788–4679769
escJP	CTAAAACGGCTGCTGAGGAGGGTTG	4679764–4679740
escVF	CCACAATCCTGTTGATTACGAC	4677820–4677799
escVR	AGGAGTAAATAATGTCACCCGC	4677719–4677740
escVP	TGCGCCTGTCGCTCAGTGTTAGTAC	4677793–4677769
espAF	ATCTAAAGCGTCAACCACGG	4662491–4662472
espAR	ACGTCTTGAGGAAGTTTGGC	4662381–4662400
espAP	TTCGCATTCTTATCAGTGCTACTCTGAACATCA	4662402–4662434
eaeF	GGCGGATTAGACTTCGGCTA	4666554–4666535
eaeR	CGTTTTGGCACTATTTGCCC	4666404–4662423
eaeP	AACGCCGATACCATTACTTATACCGCGACG	4666526–4666497
rpoAF	GGTGAGAGTTCAGGGCAAAG	4242997–4242978
rpoAR	GGCTTGACGATTTCGACATC	4242887–4242906
rpoAP	TGAAGTTATTCTTACCTTGAATAAATCTGGCATTG	4242976–4242942
LEE3F	GATCTCTAGAGTTTTCCGATTGATTAATGTTGTTC	4678972–4678948
LEE3R	GCGTCTAGACGCAACATGTGTATATCAATATGGAC	4673159–4673184
Bovine cytokine/chemokine and transcription factor QRT-PCR primer setsd
RPS9	F, CGCCTCGACCAAGAGCTGAAG; R, CCTCCAGACCTCACGTTTGTTCC
IL-1β	F, ATGGGTGTTCTGCATGAG; R, AAGGCCACAGGAATCTTG
IL-8	F, AAGCTGGCTGTTGCTCTC; R, GGCATCAGAAGTTCTGTACTC
IL-10	F, TTACCTGGAGGAGGTGATG; R, GTTCACGTGCTCCTTGATG
IL-13	F, CCTGACGAGCAGCATGTACTGT; R, TTGGATGACACTGCAGTTGGA
IL-17	F, CACAGCATGTGAGGGTCAAC; R, GGTGGAGCGCTTGTGATAAT
IL-22	F, GACTGTGGAGTTTGGCTCCCCTTTTC; R, CAGATTAAAGTCATTGGAGAACTGAAC
TNF-α	F, CGGGGTAATCGGCCCCCAGA; R, GGCAGCCTTGGCCCCTGAAG
RORγT	F, GCAAAGCCCAGGACTCAG; R, GGTGATAACCCCGTAGTGGA
GATA-3	F, AACCGGGCATTACCTGTGTA; R, AGGACGTACCTGCCCTTCTT
T-bet	F, CCTGGACCCAACTGTCAACT; R, GGTAGAAACGGCTGGAGATG

^aThe nucleotide sequences of E. coli primers used in this study were selected from the published genome sequence of E. coli O157:H7 strain EDL933 with accession number AE005174.2. F, forward; R, reverse.

^bPosition of the primer sequence in the genome of strain EDL933.

^cSubscripts F, R, and P indicates forward primer, reverse primer, and probe, respectively.

^dSource or reference: 14, 30, 51.

Construction of plasmids for complementation of hha and sepB deletion mutations.

Previously described plasmid pSM197R, containing hha and its cognate promoter, was used to complement the hha deletion mutant (47). For complementation of the sepB (escN) deletion mutant, a 5.8-kb fragment containing the entire LEE3 operon except for 151 bp in the 3′ region of sepQ was PCR amplified using primers LEE3_F and LEE3_R (Table 1). The amplified LEE3 DNA fragment was resolved on a 1% agarose gel and purified using a gel extraction kit and the manufacturer's recommended procedure (Qiagen, Valencia, CA). The purified LEE3 fragment was cloned into the pCRXL-TOPO cloning system by following the manufacturer's instructions (Invitrogen, Carlsbad, CA). The recombinant pCRXL-LEE3 plasmid (pSM619) was electroporated into the hha sepB deletion mutant strain by a previously described procedure (50). The vector pCRXL was electroporated into the wild-type and hha sepB deletion mutant strains to construct vector control strains.

Detection of secreted proteins in culture supernatants.

Bacterial cultures were grown in DMEM to an A₆₀₀ of 1.0 and centrifuged (6,000 × g for 10 min) to pellet cells. Culture supernatants were filtered (0.22-μm filters), mixed with TCA to a final concentration of 10%, incubated at 4°C overnight, and centrifuged (30 min at 18,500 × g). The pellets were washed once with acetone, chilled to −20°C, and centrifuged, and acetone was decanted. The protein pellet was dried under vacuum and resuspended in TBS (0.1 M Tris-HCl, 0.154 M NaCl, pH 7.4). Ten microliters of resuspended samples was analyzed on a 4 to 15% Tris-buffered SDS-polyacrylamide gel (Bio-Rad Laboratories, Hercules, CA). The gel was stained with GelCode Blue Stain Reagent (Thermo Fisher Scientific, Rockford, IL) and photographed. The molecular masses of proteins in TCA precipitates were approximated from a protein ladder (Bio-Rad Laboratories, Hercules, CA).

HEp-2 cell adherence assays.

Bacterial adherence to HEp-2 cells was assayed by a previously described procedure (46). Briefly, 0.5-ml volumes of freshly propagated HEp-2 cells (5.0 × 10⁴) were added to the wells of a chamber slide and incubated at 37°C overnight in an atmosphere of 5% CO₂ in air. After washing, 0.5-ml volumes of RPMI medium containing 1% fetal bovine serum and 50 μl (5 × 10⁷ CFU) of an overnight bacterial culture were added to the slide chambers. After incubation at 37°C for 3 h, the slides were washed in phosphate-buffered saline (PBS), fixed for 1 min in 95% ethanol, and stained with toluidine blue for 15 s. Slides were washed in distilled H₂O, dipped in 95% ethanol to fix the stain, washed briefly in water, and air dried. A coverslip was mounted on the slide before microscopic examination at ×100 magnification. Ten visual fields containing approximately the same number of confluent HEp-2 cells were randomly selected for enumeration of the number of adherent bacterial cells per HEp-2 cell.

Animal experiments.

The weaned male calves (8 to 10 weeks of age) were housed in climate-controlled biosafety level 2 barns in accordance with the guidelines of the American Association for Laboratory Animal Care. The calves were fed twice daily with pelleted feed and alfalfa hay cubes in amounts equal to 1% of their body weight, and water was offered ad libitum. All animal protocols were approved by the National Animal Disease Center Animal Care and Use Committee. Calves (n = 4 per strain) were inoculated with 50 ml of PBS containing 10¹⁰ CFU of the wild-type or hha or hha sepB mutant strain. Feces were collected every day of the 1st week, every other day of the 2nd week, and every 3rd day of the last 2 weeks by rectal palpation. One-hundred-microliter samples of 10-fold serially diluted feces were plated on sorbitol-MacConkey agar containing 100 μg/ml streptomycin. After the plates were incubated at 37°C for 24 h, sorbitol-negative colonies that were off-white in appearance were counted and confirmed as O157 by agglutination with an anti-O157 antiserum (Oxoid, Unipath Ltd., Ogdensburg, NY). The limit of bacterial detection by direct plating was 10² bacterial CFU/g of feces. Crude DNA preparations of 48 randomly selected colonies (12 colonies from feces of each of the four calves inoculated with the wild-type strain) identified as E. coli O157:H7 by their sorbitol-negative phenotype and agglutination with anti-O157 antiserum were tested by PCR to confirm the presence of genes encoding Shiga toxin 2 (stx₂) and intimin (eae_O157:H7). The methods for crude DNA preparation and PCR detection of stx₂ and eae_O157:H7 have been described in a previous study (48).

Histological studies.

Intestinal tissues of animals were collected at necropsy and fixed in 10% formalin before being transported to the laboratory. Tissues were embedded in paraffin, sectioned, dehydrated, stained with hematoxylin and eosin, and examined microscopically to determine the presence of any gross abnormalities and A/E lesions (7, 37).

Detection of tissue cytokine expression by QRT-PCR.

Total DNA-free RNA was isolated from RAJ and small intestine tissues by using an RNeasy Mini kit and treating the column-bound RNA with RNase-free DNase I according to the manufacturer's instructions (Qiagen, Valencia, CA). One microgram of total RNA was reverse transcribed using oligo(dT)_12-18 primers (Invitrogen, Carlsbad, CA) to cDNA, which was then amplified by a Sybr green-based real-time PCR using the 7300 Real-Time PCR system (Applied Biosystems, Carlsbad, CA) and primers (Table 1) to quantify the expression of genes encoding interleukin-1β (IL-1β), IL-8, IL-10, IL-17, IL-22, tumor necrosis factor alpha (TNF-α), T-bet, GATA-binding protein 3 (GATA3), and retinoid-related orphan receptor γt (RORγT). Final relative quantification was done by the 2^−ΔΔCT method, where the amount of target gene was normalized to an endogenous control (bovine ribosomal protein S9) and expressed relative to small intestinal tissues of the same calf since lymphoid tissue of the colorectal junction in the large intestine is considered the primary colonization site for E. coli O157 (36).

Statistical analyses.

One-way analysis of variance (ANOVA) with Bonferroni's multiple-comparison test (GraphPad Software, San Diego, CA) was used to analyze the QRT-PCR gene expression data for LEE and cytokine genes and the data from in vitro adherence assays. Changes in gene expression or adherence to HEp-2 cells were considered significant at P < 0.05. The significance (P < 0.05) of the difference in the duration or magnitude of fecal shedding between the mutant and wild-type strains was assessed by one-way ANOVA with Tukey's multiple-comparison test. Comparisons of weighted simple linear regression (SLR) models (TableCurve 2D v.5.00; AISN Software) of fecal shedding as a function of days postinoculation between the mutant and wild-type strains were done by obtaining a full and reduced model F-test statistic (38). Standard weights of 1/variance were used in the weighted regressions to compensate for replication of CFU on a single sample day. To use the SLR model and account for large numbers of zero data points, the amount of shedding was converted from log₁₀ CFU to ln (CFU + 0.01) and plotted as a function of days. Three bacterial strains used for animal inoculations were considered three conditions in the models. Slopes and intercepts of the equations for the three conditions were compared by overlap of the 95% confidence intervals from the SLR equations. Further comparisons of the 95% confidence intervals from the weighted SLR equations showed days at which differences in CFU counts among the 3 conditions occurred.

RESULTS

Deletion of the sepB gene has no effect on the increased LEE transcription in the hha mutant strain.

Since mutations affecting hha result in increased LEE transcription (50) and mutations abolishing sepB function render the T3SS unable to secrete LEE proteins (24, 25), we questioned whether a deletion in sepB would lower the elevated transcriptional levels of LEE presumably due to a feedback inhibition caused by the intracellular accumulation of LEE proteins in the hha sepB mutant strain. QRT-PCR analysis showed that the transcriptional levels of ler (representing the LEE1 operon) and escJ, escV, espA, and eae (representing operons LEE2 to LEE5) were higher by 4-fold and ≥5-fold, respectively, in the hha mutant than in the wild-type strain (P < 0.05) (Fig. 1). The transcriptional levels were increased by 3-fold for ler and 5-fold or more for the other four LEE genes in the hha sepB mutant strain (P < 0.05), and these increased transcriptional levels were not significantly different from those in the hha mutant strain (Fig. 1). The greatest increases (about 15-fold) were observed for escJ transcriptional levels in both the hha and hha sepB mutant strains (Fig. 1).

Fig 1.

Determination of the effects of sepB deletion on LEE transcription. DNA-free RNA was subjected to one-step QRT-PCR in the presence of sense and antisense primer sets and fluorogenic TaqMan probes. The relative changes in the expression of LEE genes ler, escJ, escV, espA, and eae representing operons LEE1 to LEE5 in hha and hha sepB mutant strains were determined by adjusting the expression values of the same genes in the wild-type strain to 1. The expression values were computed from three independent experiments and by testing each RNA sample in duplicate. Values are shown as means ± the standard errors of the means. One-way ANOVA with Bonferroni's posttest was used to assess the significance of the difference between the expression values of each gene in the mutant and wild-type strains.

The T3SS-mediated secretion of LEE proteins is reduced by deleting sepB in the hha mutant strain.

Although the sepB deletion did not affect increased transcriptional levels of LEE caused by the hha deletion, the T3SS-mediated secretion of LEE-encoded proteins was reduced in the hha sepB deletion mutant. The hha mutant secreted larger amounts and greater numbers of proteins into the growth medium than the wild-type and hha sepB mutant strains (Fig. 2). The increased secretion of LEE-encoded proteins in the hha mutant was reduced to the levels secreted by the wild-type strain upon complementation of the hha mutant with plasmid pSM197R. Similarly, transcomplementation of the hha sepB mutant with plasmid pSM619 containing the entire LEE3 operon except for the 3′ truncated sepQ gene restored the T3SS-mediated secretion of LEE-encoded proteins. Among the proteins secreted by the hha and transcomplemented hha sepB mutants, the molecular masses of four proteins were estimated to be 90, 60, 35, and 20 kDa based on the known molecular masses of constituent proteins of a standard protein ladder (lane MW). The secreted proteins intimin, Tir, EspB, and EspA have been predicted to have molecular masses of 97, 58, 33, and 20.5 kDa (29, 33). None of these four proteins were detectable in the wild-type and hha sepB strains in stained SDS-polyacrylamide gels. However, in a previous study, we have demonstrated that these four proteins secreted into the culture supernatants by the wild-type and hha mutant strains show specific reactivity to anti-intimin, anti-Tir, anti-EspB, and anti-EspA antibodies in Western blot assays (49).

Fig 2.

Determination of the effects of hha and sepB deletions on the secretion of LEE proteins by the T3SS. (A) Molecular masses of marker proteins are indicated in kilodaltons on the left (lane MW), and the numbered lanes contain the TCA-precipitated culture supernatants of wild-type O157/vector pCR2.1 (lane 1), hha mutant/vector pCR2.1 (lane 2), and hha/pSM197R (lane 3). (B) Molecular masses of marker proteins are indicated in kilodaltons on the left (lane MW), and the numbered lanes contain the TCA-precipitated culture supernatants of wild-type O157/vector pCRXL (lane 1), hha sepB mutant/vector pCRXL (lane 2), and hha sepB mutant/pSM619 (lane 3). Approximate molecular masses of the secreted proteins intimin, Tir, EspB, and EspA are indicated on the right side of each panel.

Increased adherence of the hha mutant to HEp-2 cells was negatively impacted by the sepB deletion.

Microscopic scanning of HEp-2 cells incubated with the wild type or the isogenic hha or hha sepB mutant strain revealed higher numbers of bacterial cells (indicated by a red arrow) of the hha mutant adhering to HEp-2 cells compared to fewer adherent bacterial cells of the wild-type or hha sepB mutant strain (Fig. 3A). Enumeration of adherent bacterial cells to HEp-2 cells showed that the adherence of the hha mutant was 18-fold greater than that of the wild-type strain (P < 0.05) (Fig. 3B). The deletion of sepB in the hha mutant strain lowered the adherence of the hha sepB mutant by 35-fold relative to that of the hha mutant strain (Fig. 3B) (P < 0.05). Although the mean numbers of adherent hha sepB mutant bacterial cells were 50% lower than those of the wild-type strain in three independent experiments, this difference was not statistically significant although the adherent cells of the double mutant strain were less frequently encountered on HEp-2 cell monolayers than those of the wild-type strain (Fig. 3B).

Fig 3.

(A) Toluidine blue-stained images of HEp-2 cells showing adherent bacterial cells. HEp-2 cells were incubated with the wild-type or hha or hha sepB mutant strain for 3 h and visualized at ×100 magnification. Red arrows point to adherent bacteria, which are seen as small, dark blue, rod-shaped cells. (B) Bar graph generated by enumerating bacteria adherent to HEp-2 cells as shown in panel A. Data shown are the means ± the standard errors of the means of the combined results of 10 randomly selected fields from each of the three independent assays. Differences were considered significant at P < 0.05.

Deletion of sepB in the hha mutant strain reduces the duration of fecal shedding of the hha sepB mutant strain in experimentally inoculated calves.

Since the deletion of sepB in the hha mutant strain caused significant reductions in the secretion of LEE proteins and adherence of the hha sepB mutant to HEp-2 cells, we wanted to determine if calves orally inoculated with the hha sepB mutant strain would show a reduced duration (defined as the day beyond which calves showed no detectable levels of fecal shedding of inoculated bacterial strains) and magnitude of fecal shedding compared to that of the wild-type or hha mutant strain in calves inoculated with these two strains. Fecal shedding data were collected by enumerating sorbitol-negative off-white colonies and confirming them as O157 by agglutination with anti-O157 antiserum. The accuracy of fecal counts determined by the above two criteria was validated by PCR detection of the stx₂ and eae_O157:H7 genes in 100% of the colonies determined to be O157 based on a sorbitol-negative phenotype and a positive reaction with anti-O157 antiserum. Analysis of fecal shedding data showed that the calves (n = 4) inoculated with the hha sepB deletion mutant had the shortest mean duration (12 days) of fecal shedding, in contrast to that of the group of calves inoculated with the hha mutant (25 days) or wild-type (22 days) strain (P < 0.05) (Fig. 4). The duration of fecal shedding in the calves inoculated with the hha sepB mutant ranged from 10 to 14 days, compared to the duration of 16 to 28 days for the calves inoculated with the wild type and 22 to 28 days for the calves inoculated with the hha mutant strain. The mean duration of fecal shedding of the hha sepB mutant-inoculated calves was significantly lower than that of the calves inoculated with the wild-type or hha mutant strain (P < 0.05). Although the differences in the mean duration of shedding between the groups of calves inoculated with the wild-type and hha mutant strains was statistically insignificant (Fig. 4), fecal samples from two of the four calves inoculated with the hha mutant were positive for 22 days, compared to two calves whose fecal samples were positive for 16 days in the group of four calves inoculated with the wild-type strain. Two other calves in each group were positive for 28 days.

Fig 4.

Duration of fecal shedding of the wild-type and hha and hha sepB mutant strains in calves. The mean duration of fecal shedding in calves was estimated by averaging the total number of days in the 28-day fecal sampling period (following oral inoculation with the wild-type or hha or hha sepB mutant strain) that the calves shed detectable levels (≥10² CFU/g feces) of the inoculum strain. The mean duration of shedding of each group of calves is shown as a horizontal line intersected by a vertical line representing the standard error of the mean. The duration of shedding of each calf is indicated by the four shapes surrounding the mean line. The differences were considered significant at P < 0.05.

Deletion of sepB in the hha mutant strain reduces the magnitude of fecal shedding of the hha sepB mutant strain in calves.

The magnitude of fecal shedding, expressed as log₁₀ CFU counts per gram of feces, continued to decline for the first 13 days postinoculation in the animals inoculated with the wild-type strain or the hha or hha sepB mutant strain (Fig. 5). The magnitude of fecal shedding on days 1 to 13 ranged from 5.39 to 1.69 log₁₀ CFU in the calves inoculated with wild-type strain, from 5.51 to 3.04 log₁₀ CFU in the calves inoculated with hha mutant strain, and from 5.49 to 1.57 log₁₀ CFU in the calves inoculated with the hha sepB mutant strain. In the calves inoculated with the hha sepB mutant strain, the magnitude of shedding continued to decline from 5.49 log₁₀ CFU at day 1 postinoculation to 0 at day 19 through day 28. In the group of calves inoculated with the hha mutant strain, fecal shedding increased from 3.04 log₁₀ CFU on day 13 to 4.65 log₁₀ CFU on day 19 and then declined to 1.58 and 1.39 log₁₀ CFU by days 25 and 28, respectively. In the group of calves inoculated with the wild-type strain, fecal shedding increased from 1.69 log₁₀ CFU to 2.72 log₁₀ CFU between days 13 and19, declined to 1.69 log₁₀ CFU on day 22, and increased again from 3.46 to 3.98 log₁₀ CFU between days 25 and 28.

Fig 5.

Determination of the magnitude of fecal shedding profiles of calves inoculated with the wild-type (●) or hha (■) or hha sepB (▲) mutant strain. Bacterial counts of inoculated strains recovered in feces were log₁₀ transformed and plotted against fecal sampling days. The calves that were culture negative by direct plating were assigned a random CFU value of 1 so that the log₁₀ transformation of 1 CFU equals 0. The log₁₀ CFU count per gram of feces (y axis) plotted for the corresponding sampling day (x axis) represents the mean of the four calves in each group. The standard error of the mean shows the range of shedding on each sampling day for the four calves in each group.

When data on the magnitude of fecal shedding were analyzed by weighted SLR, fecal shedding [expressed as ln (log CFU + 0.1)] was significantly lower in the calves inoculated with the hha sepB mutant strain starting at day 11 postinoculation and become undetectable by direct fecal plating at days 19 through 28 (P < 0.05) (Table 2). On the other hand, the magnitudes of fecal shedding were not considered significantly different between the wild-type and hha mutant strains by the SLR analysis despite the recovery of somewhat higher numbers of bacteria from feces of calves inoculated with the hha mutant strain than from those inoculated with the wild-type strain (Table 2). Histological analysis provided no evidence of adherent bacteria on the tissue sections prepared from the colorectal junctions of calves inoculated with the wild-type strain or either of the two mutant strains (data not shown).

Table 2.

Weighted SLR analysis of magnitude of fecal shedding in calves

Day and strain	Predicted no. of CFU	95CILa	95CIUb	Significancec
1
Wild type	11.3	10.27	12.32	A
hha	11.42	10.35	12.48	A
hha sepB	11.04	9.367	12.71	A
3
Wild type	10.32	9.46	11.19	A
hha	10.63	9.74	11.52	A
hha sepB	9.686	8.178	11.2	A
5
Wild type	9.35	8.592	10.11	A
hha	9.841	9.029	10.65	A
hha sepB	8.331	6.971	9.69	A
7
Wild type	8.376	7.646	9.106	A
hha	9.054	8.195	9.912	A
hha sepB	6.976	5.748	8.204	A
9
Wild type	7.402	6.613	8.191	AB
hha	8.266	7.275	9.277	A
hha sepB	5.621	4.498	6.746	B
11
Wild type	6.428	5.51	7.347	A
hha	7.478	6.247	8.709	A
hha sepB	4.266	3.215	5.318	B
13
Wild type	5.454	4.361	6.548	A
hha	6.691	5.202	8.179	A
hha sepB	2.911	1.891	3.932	B
16
Wild type	3.993	2.59	5.397	A
hha	5.509	3.596	7.422	A
hha sepB	0.8792	−0.1787	1.937	B
19
Wild type	2.532	0.7887	4.276	A
hha	4.328	1.968	6.687	A
hha sepB	−1.153	−2.337	0.03044	B
22
Wild type	1.071	−1.028	3.17	A
hha	3.146	0.3273	5.965	A
hha sepB	−3.185	4.56	−1.811	B
25
Wild type	−0.3899	−2.853	2.073	A
hha	1.964	−1.32	5.249	A
hha sepB	−5.218	−6.825	−3.61	B
28
Wild type	−1.851	−4.683	0.9812	A
hha	0.7829	−2.972	4.538	A
hha sepB	−7.25	−9.117	−5.383	B

^a95CIL, lower limit of the 95% confidence interval.

^b95CIU, upper limit of the 95% confidence interval.

^cPredicted means of ln(CFU + 0.01)-transformed values from SLR equations for three conditions (the wild-type and two mutant strains) within a day followed by the same letter are not significantly different based on overlap of their respective 95% confidence intervals.

Deletion of sepB in the hha mutant strain enhanced IL-1β gene expression in the RAJ.

The RAJ has been implicated as the major site of E. coli O157 colonization, although this organism can be found throughout the gastrointestinal tracts of cattle. We utilized real-time PCR to determine whether there was differential cytokine/chemokine expression in tissue isolated from this site at 28 days postinoculation with the three isogenic strains. In addition, selected transcription factors known to regulate cytokine/chemokine gene expression were examined. As shown in Table 3, IL-1β mRNA levels were significantly higher in calves inoculated with the hha sepB mutant than those in calves inoculated with the other two isogenic strains. We did not observe differences between the wild-type and mutant strains in the induction of IL-8 gene expression but rather found a severalfold elevation in IL-8 in RAJ compared to the small intestine. Interestingly, we observed increases in transcription factors GATA3 and RORγt in the RAJs of calves inoculated with the three strains but not in the expression of cytokine IL-13, IL-17, or IL-22, which are known to be regulated by these transcription factors.

Table 3.

Quantification of cytokine expression in RAJ tissues at day 28 by QRT-PCR

Cytokine/chemokine or transcription factor and E. coli strain	Fold changea	P valueb
IL-1β
Wild type	0.692
hha	0.596	NSc
hha sepB	2.340	<0.01
IL-8
Wild type	8.505
hha	6.870	NS
hha sepB	9.636	NS
IL-10
Wild type	0.927	NS
hha	0.542	NS
hha sepB	0.307	NS
IL-13
Wild type	0.055
hha	0.035	NS
hha sepB	0.883	NS
IL-17
Wild type	0.625
hha	0.294	NS
hha sepB	0.144	NS
IL-22
Wild type	1.088
hha	0.960	NS
hha sepB	0.195	NS
TNF-α
Wild type	0.821
hha	0.32	NS
hha sepB	0.287	NS
T-bet
Wild type	0.615
hha	0.355	NS
hha sepB	0.191	NS
GATA3
Wild type	6.06
hha	10.016	NS
hha sepB	4.668	NS
RORγT
Wild type	4.629
hha	3.555	NS
hha sepB	2.447	NS

^aFold change (calculated as 2^−ΔΔCT) in RAJ tissue cytokine gene expression relative to small intestinal tissue using the bovine ribosomal protein S9 coding gene as an endogenous control. The observed variability in the ratios of target genes to RS9 among the three treatment groups was approximately 5%.

^bThe level of significance of the fold change in cytokine/chemokine or transcription factor gene expression for E. coli mutants compared to the wild type was determined by ANOVA using Bonferroni's post hoc test.

^cNS, not significant.

DISCUSSION

In this report, we describe the impact of the T3SS and the levels of secreted LEE proteins on the abilities of O157, an important food-borne human pathogen, to adhere to epithelial cells and to colonize bovine intestines. Of the three isogenic strains evaluated, the greatest secretion of LEE proteins was observed in the hha mutant strain, which correlated with increased in vitro adherence of this mutant strain to HEp-2 cells. In previous studies, we demonstrated that increased adherence of the hha mutant to HEp-2 cells is caused by increased transcriptional levels of LEE due to derepression of the ler promoter compared to the wild-type parent strain, in which LEE transcription is highly repressed in vitro despite growth in DMEM (47, 50). The highly reduced HEp-2 cell adherence of the hha sepB mutant, which expressed LEE at levels comparable to those in the hha mutant strain, was attributed to its diminished capacity to secrete LEE proteins. The reduced secretion of LEE proteins could presumably be attributed to the inability of the hha sepB mutant to synthesize the sepB-encoded ATPase that is responsible for generating energy for the operation of the T3SS (18, 25). These results corroborated findings from earlier reports demonstrating that sepB mutants of O157 adhered poorly to HeLa cells (11). Similarly, reduced adherence to HEp-2 cells has also been demonstrated in the sepB mutants of EPEC strains harboring a LEE that is highly homologous in its structure and function to the LEE of EHEC O157 (24). Both hha and sepB deletion mutations were complemented when the hha and sepB genes cloned on multicopy plasmids were introduced into these mutant strains, respectively. These results indicated that in-frame chromosomal deletions of these two genes had no effect on the expression of downstream genes. For complementation of sepB deletion, however, the entire LEE3 operon was required since sepB is the third gene in this operon and the translation of the very first ORF of LEE3 is important for the translation of downstream genes in LEE3 (54).

Although the importance of various LEE-encoded proteins, such as Tir, intimin, EspA, and various effector proteins, in the colonization of animal intestines has been affirmed using mice, rabbits, sheep, gnotobiotic piglets, and cattle (6, 8, 10, 26, 41), the requirement of the T3SS for the colonization of animal intestines has been demonstrated only in infant rabbits (42) or 10- to 14-day-old calves fed milk replacer diets (5, 55). In this study, we used 8- to 10-week-old calves fed an adult bovine diet for oral inoculations with wild-type and hha and hha sepB mutant strains to demonstrate that the T3SS is required for increased intestinal colonization of these animals, as only the strains with a wild-type copy of the sepB gene were detected at greater magnitudes and increased duration in the animal feces. In addition, the hha sepB deletion mutant was cleared from calves more rapidly than the hha mutant or wild-type strain. These data indicated that increases in the expression of the T3SS and secretion of LEE proteins via the T3SS caused by the hha deletion needs to be coupled with a functional T3SS for increased cellular attachment and potentially enhanced persistence of O157 in cattle.

It has been shown in previous studies that to detect A/E lesions and tissue damage in the large intestines of calves experimentally infected with O157:H7, the levels of shedding must be ≥10⁶ CFU/g of feces (7). In concurrence with this report, the tissues obtained on day 28 postinoculation from the RAJs of calves inoculated with the wild-type or hha or hha sepB mutant strain appeared normal and the mean shedding per gram of feces in these calves ranged from 10² CFU for the wild type to 10⁴ CFU for the hha mutant to undetectable levels for the hha sepB mutant strain on day 28 postinoculation. Consistent with these results were the observations indicating no significant differences in the levels of proinflammatory cytokines, including TNF-α, in the RAJ tissues of the animals inoculated with the wild-type or hha or hha sepB mutant strain. However, IL-1β expression was significantly elevated in RAJ tissues of calves inoculated with the hha sepB mutant strain. These findings agree with in vitro studies demonstrating that injection of effector proteins by the T3SS is essential for the subversion of signaling cascades and inhibition of proinflammatory cytokine expression in epithelial cells (22, 56). The notable exception was that the expression of IL-8 (CXCL8), a neutrophil chemokine, was elevated in RAJ tissues of animals irrespective of the inoculated bacterial strain and without any obvious signs of inflammation at the RAJ. Thus, it is conceivable that non-LEE-encoded factors that are expressed equally by these three strains and do not require the T3SS for their secretion might be responsible for IL-8 expression or this cytokine is naturally expressed at higher levels at this intestinal site. For example, in vitro EPEC infections of epithelial cell lines of small and large intestinal origin have been shown to cause IL-8 production by modification of multiple kinase signaling pathways and independent of LEE-encoded effectors (44). In addition, we also observed severalfold increased expression of the transcriptional factors GATA3 and RORγt but without parallel increases in the expression of IL-13 or IL-17 and IL-22, cytokines which are known to be regulated by these transcriptional factors. It could be that other, as yet unidentified, cytokines are induced by these transcription factors during O157 infection or that cytokines such as IL-13, IL-17, and IL-22 are elevated at time points other than those examined. Our interpretation of the cytokine expression data is based on the use of small intestinal tissues as controls to determine relative increases in cytokine gene expression in the tissues of the colorectal junction of each animal. The use of control tissues derived from the small intestine is valid in accordance with published studies identifying the colorectal junction as a primary site of O157 colonization in the large intestines of cattle (7, 36). In addition, studies have reported the use of small or large intestinal tissues of healthy individuals as controls to determine relative increases in the expression of inflammatory cytokines in human patients with inflammatory bowel disease (4).

Fecal shedding of O157 in cattle is reported to be highly intermittent, with variable number of animals in a given herd shedding detectable levels of O157 for several weeks (21, 45). These reports support our findings that the fecal shedding profile of O157 in calves inoculated with the wild-type or hha mutant strain continued to show a negative trend for 2 weeks before establishing a low-level intermittent shedding response over the course of the following 2 weeks. Interestingly, the SLR analysis of fecal shedding over a 4-week period did not show any significant difference in the magnitude of fecal shedding between the hha mutant and wild-type strains, suggesting that increased secretion of LEE proteins might not be sufficient by itself to enhance the magnitude of fecal shedding. Thus, it is reasonable to speculate that additional factors might be required in conjunction with increased T3SS activity and secretion of LEE proteins to enhance the duration and magnitude of fecal shedding of the hha mutant in cattle.

In a previous study, we demonstrated that hha deletion mutants of O157 had reduced motility due to reduced expression of the flagellar gene fliC and increased expression of curli fimbriae with concomitant increases in its abilities to produce biofilms and adhere to HEp-2 cells (46). Thus, it is possible that reduced expression of flagellar genes might have negated the ability of the hha mutant strain to establish increased colonization of bovine intestines, as no significantly greater duration or magnitude of fecal shedding than that of the wild-type strain was observed. For example, it has recently been shown that fliC mutants of O157 had significantly reduced adherence to cultured primary rectal epithelial cells, which are considered the principal sites of colonization by O157, compared to that of the fliC-complemented mutant strain (30). In addition, these studies also demonstrated that flagellar expression is temporally regulated, as O157 bacterial cells show an abundance of flagella on their cell surfaces during early stages of adherence to epithelial cells but flagellar expression was reduced during the formation of A/E lesions on these cells. Many EPEC serotypes have also been shown to require flagella for adherence and formation of microcolonies on HeLa or HEp-2 cells (19), but unlike H7 flagella, purified flagella of EPEC serotypes failed to adhere to rectal epithelial cells (30), implying a specificity of H7 flagella for rectal epithelial cells.

Contrary to the studies citing the importance of flagella in the initial adherence of O157 to rectal epithelial cells, a recent report has shown that a fliC mutant strain of O157 that is unable to produce flagella colonized the RAJ as effectively as the wild-type flagellated strain did (9). It was argued in this report that O157 flagella may be a liability for bacterial survival in the bovine gastrointestinal tract, as flagellar biosynthesis consumes a large proportion of cellular energy, flagella invoke a host immune response compromising the persistence of O157 in the host, and potential interference of flagellar structures with the T3SS compromises intimate adherence of O157 to epithelial cells. However, data from our studies indicate that reduced flagellation (indicated by the reduced expression of fliC and decreased motility relative to those of the wild-type strain [46]) in the hha mutant strain, despite enhanced secretion by the T3SS of LEE-encoded adherence and effector proteins, did not confer a selective advantage over the wild-type strain with respect to the persistence and magnitude of fecal shedding. One possible explanation is that flagella are important in the initial adherence of O157 to epithelial cells and when flagella are not expressed or expressed poorly, alternative adherence factors might contribute to the initial adherence of the O157 hha mutant to epithelial cells. Since the expression of curli fibers is significantly enhanced in the hha mutant strain (46), it is possible that these fibers might serve as adhesive structures during the initial adherence of this mutant to rectal epithelial cells. A recent study has demonstrated that sorbitol-fermenting O157:NM strains, which are frequently associated with human infections in Germany and have an increased propensity to produce HUS in infected individuals, adhered at significantly higher levels to human colonic epithelial cells and expressed increased amounts of curli (43).

In conclusion, we have demonstrated that the T3SS and secretion of LEE-encoded proteins are essential requirements for O157 adherence to epithelial cells and colonization of bovine intestines. Deletion of the sepB gene in the hha sepB mutant rendered the T3SS inactive and reduced both the duration and the magnitude of fecal shedding compared to the hha mutant, even though the level of LEE gene expression was similarly higher in the two mutant strains than in the wild-type strain. However, increased expression of LEE and increased secretion of LEE-encoded proteins did not enhance the colonization and persistence of the hha mutant relative to those of the wild-type strain.