(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Both pathogen and host dynamically adapt pH responses along the intestinal tract during enteric bacterial infection [1]

['Sarah E. Woodward', 'Department Of Microbiology', 'Immunology', 'University Of British Columbia', 'Vancouver', 'Michael Smith Laboratories', 'Laurel M. P. Neufeld', 'Jorge Peña-Díaz', 'Wenny Feng', 'Antonio Serapio-Palacios']

Date: 2024-08

Enteric pathogens navigate distinct regional microenvironments within the intestine that cue important adaptive behaviors. We investigated the response of Citrobacter rodentium, a model of human pathogenic Escherichia coli infection in mice, to regional gastrointestinal pH. We found that small intestinal pH (4.4–4.8) triggered virulence gene expression and altered cell morphology, supporting initial intestinal attachment, while higher pH, representative of C. rodentium’s replicative niches further along the murine intestine, supported pathogen growth. Gastric pH, a key barrier to intestinal colonization, caused significant accumulation of intra-bacterial reactive oxygen species (ROS), inhibiting growth of C. rodentium and related human pathogens. Within-host adaptation increased gastric acid survival, which may be due to a robust acid tolerance response (ATR) induced at colonic pH. However, the intestinal environment changes throughout the course of infection. We found that murine gastric pH decreases postinfection, corresponding to increased serum gastrin levels and altered host expression of acid secretion-related genes. Similar responses following Salmonella infection may indicate a protective host response to limit further pathogen ingestion. Together, we highlight interlinked bacterial and host adaptive pH responses as an important component of host–pathogen coevolution.

Funding: This work was supported by grants from the Canadian Institutes of Health Research (CIHR; cihr-irsc.gc.ca) to BBF (FDN-159935). SEW is a CIHR CGS-D Graduate Scholar (GSD-154171) and was supported by a University of British Columbia (UBC; ubc.ca) Four Year Fellowship and Dmitry Apel Memorial Scholarship. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

This study investigates the behavior and virulence of C. rodentium in response to variations in host intestinal pH. We profiled pH across the intestine of one of the most commonly used laboratory mouse strains (C57BL/6) to define a range of physiologically relevant pH encountered by enteric pathogens upon initial host contact. We identified pH-specific regulation of the T3SS and epithelial attachment by C. rodentium, as well as significant changes to stress adaptation. Additionally, within-host adaptation increased C. rodentium’s acid tolerance, stimulated by colonic pH conditions. The pH of the gut also changed during established C. rodentium infection, with a significant drop in gastric pH which correlated with pathogen burden. This effect was consistent across additional models of intestinal damage, suggesting a possible host defense mechanism against further pathogen ingestion. Together, these data emphasize the interplay between pathogens and host intestinal pH in the context of enteric infection.

It is clear that EPEC/EHEC/C. rodentium respond to intestinal pH during host colonization. In particular, the extreme acidic conditions of the gastric environment are a major bottleneck to both C. rodentium and EHEC colonization [ 23 – 26 ]. Furthermore, within the small intestine C. rodentium responds to alkaline bicarbonate ions secreted by the host to neutralize acidic chyme by up-regulating genes involved in initial epithelial attachment [ 27 , 28 ]. Despite this work, much remains unclear about how EPEC/EHEC/C. rodentium respond to regional differences in gastrointestinal pH. Furthermore, while the GI environment is known to change as a consequence of diarrheal disease, little investigation has been done to profile intestinal pH postinfection [ 29 ].

Intestinal pH is often overlooked with regard to enteric infection, despite its potential role in affecting infection dynamics, host–pathogen interactions, and possibly even therapeutic success. By definition, diarrheal disease agents interact with the gastrointestinal (GI) environment to establish infection. Pathogenic Escherichia coli, including enteropathogenic E. coli (EPEC) and enterohemorrhagic E. coli (EHEC), are a diverse group of diarrheal disease agents which result in diarrhea, fever, vomiting, and even death in infected hosts [ 11 , 12 ]. Citrobacter rodentium, a related pathogen that shares similar virulence factors, is used to model murine EPEC/EHEC infection [ 13 – 15 ]. C. rodentium naturally infects mice, causing transmissible murine colonic hyperplasia (a form of colitis) with diarrheal disease symptoms [ 16 ]. All 3 pathogens are transmitted through the fecal-oral route and use a type III secretion system (T3SS) for successful colonization [ 13 , 17 – 21 ], translocating protein effectors directly into host cells for bacterial control over host cell processes [ 22 ].

Gastrointestinal pH is a key component of gut physiology, known to vary across regions of the intestine [ 1 , 2 ]. In humans, intestinal pH is characterized by a progressive increase from the duodenum to the distal ileum (pH 6.0 to 7.5), a drop at the cecum to pH 6.4, followed by increasing pH from the proximal to the distal colon (pH 7 to 7.5; Fig 1A ) [ 1 – 3 ]. In mice, reported intestinal pH is much lower, with the intestinal contents of female BALB/c mice ranging from 2.98 in the stomach to 5.24 in the ileum ( Fig 1A ) [ 4 ]. In both species, maintaining intestinal pH is important to homeostasis and gut function. Disruption to intestinal ion transport is associated with spontaneous onset of colitis symptoms in mice and increased susceptibility to induced colitis [ 5 – 7 ]. pH also influences which microbial species colonize the gut, with mildly acidic pH resembling proximal colon conditions inhibiting the growth of Bacteroides spp. [ 8 – 10 ]. Therefore, maintenance of gut pH is important to preventing overgrowth of colonizing bacteria, whether commensal or pathogenic [ 8 , 9 ].

Statistical analysis was performed in Graphpad Prism ( www.graphpad.com ) and clarified in figure legends. Unless otherwise stated, analysis was performed using a Mann–Whitney test to compare 2 groups and a one-way ANOVA with Šidák’s multiple comparisons test for more than 2 groups. Aggregate results represent the mean +/− SD and statistical significance is represented by *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

To analyze intestinal inflammation of mice colonized with either E. coli Mt1B1, ΔescN C. rodentium, or WT C. rodentium as compared to naïve animals, colon tissue samples were collected at 4 days postinoculation by first removing intestinal contents. Samples were collected into 1 ml of PBS containing a protease inhibitor cocktail (Roche Diagnostics) and homogenized using a FastPrep-24 (MP Biomedicals) at 5.5 m/s for 2 min. Homogenate was then spun down at 16,000 × g for 20 min at 4°C for supernatant collection. Inflammatory cytokines were measured using a mouse inflammation cytometric bead array (CBA) kit (BD Biosciences, Cat. no. 552364) according to manufacturer’s instructions. Measurements were made on an Attune NxT flow cytometer (Thermo Fisher Scientific), normalizing cytokine concentrations to tissue weight.

Serum calprotectin levels were analyzed using the mouse S100A9 ELISA kit (Hycult), as per the manufacturer’s instructions. Serum was collected from non-fasted mice by allowing whole blood to clot for 2 h on ice before spinning to collect serum. Absorbance at 450 nm was measured using a Synergy H1 plate reader (Biotek).

Blood was collected from naïve and day 8-infected mice following a 12-h fast. Whole blood was allowed to clot for 2 h on ice before serum collection. Gastrin levels were analyzed using the mouse gastrin enzyme-linked immunosorbent assay (ELISA) kit (Novus Biologicals, Cat. No. NBP3-08148). Absorbance at 450 nm was measured using a Synergy H1 plate reader (Biotek).

Assays were performed using a protocol adapted from [ 40 , 41 ]. In brief, a roGFP2 C. rodentium reporter strain was grown overnight in LB containing 100 μg/ml of carbenicillin. Overnight cultures were diluted 1:25 into 50 ml of pre-warmed LB without antibiotics and were grown for 5 h at 37°C in a shaking incubator. Bacterial strains were washed and resuspended in saline solution (0.9% NaCl w/w). Cultures were then exposed to the different pH of interest in a black 96-well plate with a volume of 200 μl and an OD 600 of 1. Intra-bacterial redox was assessed by measuring fluorescence with an excitation at 405 and 480 nm and an emission at 510 nm. Background fluorescence was controlled by measuring the fluorescence of WT C. rodentium. Normalized 405/480 nm ratios were calculated by using fully oxidized (100 mM H 2 O 2 ) and reduced (10 mM DTT) control conditions.

To determine whether C. rodentium shed from the mouse gut at early (day 2) and peak infection (day 7) differed in acid tolerance, mice were infected with C. rodentium as described above and fecal pellets were collected and homogenized fresh in 1 ml of PBS. Fecal homogenate was inoculated in LB pH 3.5 as described above and plated for enumeration of viable C. rodentium at 30-min intervals. A subset of fecal pellets was not homogenized to simulate natural encapsulation of C. rodentium during gastric acid exposure. Instead, fecal pellets were cut in half, with one half plated for CFU to determine the inoculating population, and the second half dropped into LB pH 3.5 as is for incubation as before. Pellets were removed from pH 3.5 after 120 min and homogenized for plating and CFU counts.

To test for the acid tolerance response (ATR) following preinduction at intestinal pH, C. rodentium was cultured overnight at either neutral pH 7 or colonic pH 5.7 under anaerobic conditions (90% N 2 , 5% CO 2 , 5% H 2 ) or aerobically at 37°C without agitation. Overnight cultures were diluted 1:20 into unadjusted LB (the baseline “untreated” condition), or LB adjusted to gastric pH 3.5 and plated at 60-min intervals as previously described.

To determine whether the deletion of cpxRA would affect C. rodentium survival at gastric pH, we cultured C. rodentium at pH 3.5 as described [ 23 ]. In brief, WT, ΔcpxRA [ 39 ], or ΔcpxRA::cpxRA C. rodentium strains were inoculated 1:20 from overnight culture into unadjusted LB (the baseline “untreated” condition) or LB adjusted to gastric pH 3.5. At 30-min intervals, subcultures were sampled and diluted in PBS for plating on neutral LB agar to determine viability by quantifying surviving CFU. Viable CFU were compared to the inoculum population to determine percent survival over time.

C. rodentium was cultured to mid-log phase (3.5 h, 37°C with agitation) at gastrointestinal pH before being spun down and resuspended in PBS with the addition of FM 1–43 Dye (N-(3-Triethylammoniumpropyl)-4-(4-(Dibutylamino) Styryl) Pyridinium Dibromide) (Cat. no. T3163, Invitrogen). Images were taken on a Zeiss X10 light microscope and cell area was measured using CellProfiler 4.2.1 [ 37 , 38 ].

To assess pH-induced biofilm formation, C. rodentium was inoculated in LB at physiological pH to an OD 600 of 0.5 in a black, clear bottom 96-well plate. The plate was incubated at 37°C, standing, for 48 h to allow for biofilm formation. After 48 h, culture media were removed and the OD 600 of the resultant biofilm was measured in a Synergy H1 plate reader (Biotek), and 100 μl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) at a concentration of 0.5 mg/ml in LB was added to the biofilms and allowed to incubate at 37°C for 10 min. MTT was removed and replaced with 100 μl of 100% DMSO for resuspension before measuring OD at 570 nm to approximate formazan production as a readout of metabolic activity.

Eukaryotic RNA was isolated from mouse tissue samples (stomach and colon) stored in RNAprotect Tissue Reagent (Qiagen) at −70°C before extraction using a GeneJET RNA Purification Kit (Cat. no. K0731, Thermo Scientific). cDNA synthesis and Quantitative RT-qPCR was performed as described above using various primer pairs ( S3 Table ). β-2-microglobulin (B2M) was used as an endogenous control.

C. rodentium was subcultured for 3.5 h 1:40 in LB base media adjusted to gastrointestinal pH, as before. After subculture samples were preserved in RNAprotect Bacterial Reagent (Qiagen) and stored at −70°C until extraction. Bacterial RNA extractions were done using the RNAse-Free DNase Set (Cat. no. 79254, Qiagen) as per the manufacturer’s instructions. A QuantiTect Reverse Transcription Kit (Cat. no. 205313, Qiagen) was used for cDNA synthesis and Quantitative RT-qPCR was performed using a QuantiNova SYBR Green RT-PCR Kit (Cat. no. 208056) in combination with primers ( S3 Table ) for target gene mRNA. dnaQ was used as an endogenous control. Data was normalized to efficiency of primers by target gene (91.6% to 103.5% for all primers used).

Evaluation of actin pedestal formation was performed as follows. Caco-2/TC7 and CMT-93 cells were cultured in a 6-channel slide (Cat. No. 80606; ibidi) to 80% confluence before infection as before with C. rodentium preconditioned to mid-log phase at pH 4.8 or pH 7.0. Following the 4-h infection, cells were washed with PBS, fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.1% of Triton X-100 for 10 min, and blocked with 3% BSA–PBS for 30 min. Cells were immunostained using a polyclonal rat anti-Tir antibody [ 17 ] overnight, followed by a goat anti-rat IgG (H+L) secondary antibody conjugated to Alexa Fluor 568 (Cat. no. A-11077; Thermo Fisher Scientific) for 45 min. F-actin was stained with Alexa Fluor 488 phalloidin (Cat. no. A12379). Nuclei and bacteria were stained with DAPI. Cells were imaged with a Leica Stellaris 5 confocal microscope using a 63× objective and analyzed using LAS-X software.

Murine rectal carcinoma cell line CMT-93 (ATCC CCL-223) and human adenocarcinoma cell line Caco-2/TC7 were maintained in DMEM supplemented with 10% fetal bovine serum, 1% Glutamax, and 1% nonessential amino acids. Cells were used between passages 4 to 10. Cells were seeded in 12-well plates at 90,000 cells/well and incubated for 48 h prior to infection. Confluent CMT-93 cells were washed once with PBS before infection. The C. rodentium inoculum was prepared by subculturing 1:40 from overnight culture to mid-log phase (3.5 h, 37°C, shaking) in either LB or DMEM media adjusted to intestinal pH. Subcultures were washed and resuspended in cell culture media at multiplicity of infection (MOI) 100. Infected plates were centrifuged (1,000 rpm, 5 min) to synchronize the infection before incubation for 4 h at 37°C and 5% CO 2 . Infected plates were washed 5 times with PBS before detachment using 0.1% Triton X-100 in PBS for dilution and plating on neutral LB agar. Adherent bacterial CFU were quantified and data was normalized to exact inoculum dosage (determined by retrospective plating) to account for differences between pH conditions.

Bacterial growth at gastrointestinal pH was determined by inoculating bacteria at an optical density at 600 nm (OD 600 ) of 0.005 from an overnight culture into 250 μl total volume of the media of interest (LB or DMEM base media adjusted to gastrointestinal pH). Cultures were incubated at 37°C with agitation for 20 h in a Synergy H1 plate reader (Biotek) with OD 600 measurements at 10-min intervals.

Colonization with both a-virulent C. rodentium ΔescN [ 35 ] and commensal E. coli Mt1B1 [ 36 ] were carried out by oral gavage with 2.5–4 × 10 8 CFU from overnight culture, as determined by retrospective plating. Control WT-infected mice were inoculated at the same dosage. S. Typhimurium (ATCC 14028) infections were carried out by oral gavage with 1 × 10 6 CFU of S. Typhimurium from overnight culture. Fecal samples were processed for CFU enumeration as described above.

C. rodentium infections were carried out by gavaging mice orally with 10 8 colony-forming units (CFUs) of C. rodentium DBS100 from overnight culture. Mice were euthanized at experimental endpoint by isoflurane anesthesia followed by carbon dioxide inhalation. Bacterial burden was determined by collecting colon samples into 1 ml of reduced phosphate-buffered saline (PBS) and homogenized in a FastPrep-24 (MP Biomedicals) at 5.5 m/s for 2 min. Sample homogenate was diluted for plating on MacConkey agar (Difco) and incubated for 18 to 20 h at 37°C before counting bacterial colonies.

pH measurements were performed on fresh, undiluted intestinal content, using an Elite pH Spear Pocket Tester (Thermo Scientific). Measured pH represents the average of 2 to 3 readings per sample. All bacterial colonization experiments were done in 7-week-old female C57BL/6 mice fed the Picolab Rodent Diet 20 chow (Cat. no. 5053; LabDiet). Mice were monitored daily throughout the 4 to 8 day infections for weight loss and clinical symptoms.

As mentioned above, animal experiments were performed in accordance with the guidelines of the Canadian Council on Animal Care and the University of British Columbia (UBC) Animal Care Committee according to Animal Care Protocols A20-0187 and A17-0228. C57BL/6 and BALB/c mice were ordered from Jackson Laboratory (Bar Harbor, Maine, United States of America) and maintained in a specific pathogen-free facility at UBC on a 12-h light-dark cycle. Mice were allowed to acclimatize to the facility for 1 week following arrival. All pH measurements and infection models were run after the acclimatization period.

The sacB gene-based allelic exchange method and the suicide vector pRE112 [ 33 ] were used to generate a chromosomal in-frame deletion mutant of ldcC in C. rodentium strain DBS100. The immediate upstream and downstream flanking regions of ldcC were amplified by PCR using primers listed in S2 Table . Purified DNA fragments were ligated into pRE112 by Gibson assembly. The suicide vector was transformed into E. coli strain MFDpir [ 34 ] by electroporation, which was used as the donor strain to transfer the plasmid by conjugation into C. rodentium. After sucrose selection, colonies resistant to sucrose and sensitive to chloramphenicol were screened, and then further verified, for ldcC delection by PCR using primers Check_ldcC_F and Check_ldcC_R.

Bacterial strains are listed in S1 Table . Unless otherwise noted, C. rodentium strain DBS100 was used in all assays [ 30 ]. A fresh bacterial stock was used for all assays as follows: 2 days in advance bacteria were streaked onto a fresh LB agar plate from frozen (−70°C in glycerol) and grown overnight at 37°C. A single colony was re-inoculated into LB (lysogeny broth, Sigma) broth and grown overnight at 37°C, shaking. To test the response of bacteria to gastrointestinal pH, LB broth or DMEM (Dulbecco’s Modified Eagle Medium, Hyclone) was adjusted to gastrointestinal pH by the addition of 37% HCl or 40% NaOH. Due to the high buffering capabilities of DMEM, media were pH-adjusted fresh for use within 12 h. pHs tested were: 3.5, 4.0, 4.4, 4.8, 5.7, 6.0, 6.4, 7.0, and 7.5. To test the effect of pH on human pathogenic E. coli, EHEC strain EDL933 [ 31 ] and EPEC strain E2348/69 [ 32 ] were used, as well as Salmonella enterica serovar Typhimurium strain CDC 6516–60 (ATCC 14028).

Animal experiments were performed in accordance with the guidelines of the Canadian Council on Animal Care and the University of British Columbia (UBC) Animal Care Committee. Approval for the work carried out in this study was granted by the UBC Animal Care Committee according to Animal Care Protocols A20-0187 and A17-0228.

Results

Regional variability in gastrointestinal pH may be influenced by genetic background To assess the response of C. rodentium to intestinal pH, we first defined regional pH ranges within its natural host, the mouse. We compared reported gastrointestinal pH (pH 2 to 6) to pH measured in healthy, female, 7-week-old C57BL/6 mice across gut regions (Fig 1A and 1B). We observed a similar pattern to reported murine and human values, which see pH increase at significant increments towards the ileum-large intestinal regions (Fig 1B). At the cecum, we found a nonsignificant drop in pH which is characteristic of human intestinal pH, but not previously observed in mice [1]. Across all regions, we observed variability between individuals, particularly in the stomach and ileum, demonstrating that gastrointestinal pH likely fluctuates over time and with varied consumption of food and water. Overall however, the pH across both the small intestine (duodenum, jejunum, and ileum) and large intestine (cecum and colon) was more basic than reported murine values in BALB/c mice [4]. Mice in the aforementioned study were fed a different diet, the Teklad Global 18% Protein Rodent Diet (Teklad), with a lower protein content (18%) [42]. Dietary protein is thought to increase intestinal pH by increasing ammonia production as a by-product of bacterial metabolism [2,9,43]. We therefore placed C57BL/6 mice on either the Teklad diet or Picolab Rodent Diet 20 5053 (Picolab; 21% protein) for 2 weeks before sacrifice. We found no significant differences between diets, except in the distal colon where pH was significantly higher on the lower protein Teklad diet (S1A and S1B Fig). This indicates that the more alkaline pH observed across gut regions may instead be due to differences in mouse genetic background (BALB/c versus C57BL/6). Indeed, while the only comprehensive profiling of murine intestinal pH has been done in BALB/c mice, ileal, cecal, and colonic pH has been reported to be more alkaline in FVB/N and ICR mouse backgrounds, and more closely representative of values found in the human gut (Fig 1) [44–46]. To verify whether mouse genetic background can indeed result in altered intestinal pH, we compared the intestinal pH of both BALB/c and C57BL/6 mice fed the Picolab diet. We found a significant difference in intestinal pH in the ileum and colon regions (S1C Fig). As reported previously, we did not observe a drop in intestinal pH of BALB/c mice from the ileum to cecum regions, rather pH increased along the length of the GI tract. This indicates that genetic differences between mouse strains may result in altered intestinal pH, with pH of C57BL/6 mice more closely reflecting the pattern found in humans. Altogether, these data establish important physiological variation within the naïve murine gastrointestinal tract, which could be used by enteric pathogens to signal their biogeographical location.

C. rodentium alters growth dynamics, epithelial attachment, and virulence gene expression in response to gastrointestinal pH Having established the range of physiologically relevant pH, we next investigated whether C. rodentium responds to variations in pH by adjusting the pH of LB to mimic the ranges of pH found in both mouse and human GI tracts. Growth of C. rodentium was highest at the range of pH 6.4 to 7.5, with impaired growth at acidic pHs 4.4 and 4.8, resembling the gastric and duodenal regions (Fig 2A). As previously reported [23], no growth was observed at gastric pH of 3.5 to 4. Human enteropathogens EHEC, EPEC, and S. Typhimurium were similarly unable to grow at gastric pH 3.5 and displayed altered growth dynamics in response to pH below 5 (S2A–S2C Fig). This suggests that while acid tolerance and viability rates differ between pathogens [26,47], they are all affected by fluctuations in gastrointestinal pH. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. C. rodentium alters its behavior and virulence in response to regional gastrointestinal pH. (A) Growth in pH-adjusted LB across the range of gastrointestinal pH, measured as OD 600 nm at 10-min intervals over 20 h. (B) Attachment of C. rodentium to CMT-93 murine colonic epithelial cells after preinduction at gastrointestinal pH. Data represent biological replicates (N = 3–4; each the average of 3 technical replicates). Statistical analysis represents an ordinary one-way ANOVA test with Holm–Šidák’s multiple comparisons test. (C) Gene expression analysis by RT-qPCR of key virulence- and early infection-related genes across the range of gastrointestinal pH. Statistical analysis represents t tests with Bonferroni correction for multiple comparisons. Outlier tests were run using the ROUT method (Q = 1%), identifying no outliers. Error represents mean +/− SD. (D) Representative images of actin pedestal formation on Caco-2-/TC7 cells following infection with C. rodentium pre-induced at pH 4.8 and pH 7.0 (N = 3 biological replicates). Scale bar represents 30 μm. Arrows indicate representative Tir-rich pedestals. Statistics represents an unpaired t test. Summary data displayed in Fig 2 can be found in S1 Data. LB, lysogeny broth; OD, optical density. https://doi.org/10.1371/journal.pbio.3002761.g002 We further investigated the relative metabolic activity of C. rodentium subjected to intestinal pH by measuring the reduction of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) to formazan by NADPH-dependent oxidoreductase enzymes in metabolically active bacteria. As expected, bacteria incubated at pH 3.5 produced significantly less formazan compared to neutral pH despite an increased biomass, suggesting decreased metabolic activity (S2D and S2E Fig). These results support both the absence of growth at pH 3.5 and the low recoverability of cells after removal from acidic pH [23]. The observed increase in biomass at pH 3.5 may further suggest the active encapsulation of community members within biofilms. Biofilm formation is an important stress-induced protective behavior of bacteria which allows for the survival of community members shielded within. Aside from direct death of bacteria under low pH, bacteria within biofilms are also known to decrease their metabolic activity under stress, which promotes survival and persistence, especially as many stressors affect the bacterial membrane (such as antibiotics) and have the greatest impacts on actively growing bacteria [48]. Taken together, extreme low pH of the gastric environment has considerable consequences to bacterial fitness. To determine if the altered fitness dynamics observed across gastrointestinal pH also reflect differences in pathogen virulence, we assessed epithelial cell attachment, an essential characteristic of infection. This was done by “pre-inducing” C. rodentium via subculture in pH-adjusted LB before infection of CMT-93 cells, a murine rectal cell line chosen because the rectum is a known large intestinal niche of C. rodentium during host colonization. We expected attachment to increase in response to pH representing key replicative niches of C. rodentium, such as the cecum and colon. However, C. rodentium attachment was significantly increased after preinduction at pH 3.5 and 4.4, representing gastric-duodenal pH (Fig 2B). Attachment was lowest at pH 7.5 which represents the upper range of ileal pH, as well as pH found in the extra-intestinal environment. Assays were repeated using subculture in pH-adjusted DMEM media, which is known to activate expression of the T3SS. Despite high buffering of DMEM media making accuracy difficult, the same trends were noted in both growth and attachment (S2F and S2G Fig). Therefore, LB media was used in all subsequent assays. To further explore pH-mediated virulence regulation, we characterized the gene expression profile of known virulence- and early infection-associated genes following C. rodentium subculture in pH-adjusted LB prior to RNA extraction. Despite the fact that LB media is not known to activate the T3SS, type III secretion-associated virulence genes ler, tir, eae, and effector gene nleB, were highly expressed at pH 4.8 which represents duodenal pH (a 4- to 5-fold increase from baseline). At gastric pH 3.5, we observed down-regulation of virulence genes associated with type III secretion, including both structural and effector-associated genes (Fig 2C), as well as virulence regulators ler and regA [49]. While a known fimbrial gene fimA was also down-regulated, we found an up-regulation of kfcC, a component of the C. rodentium type IV pilus, which may explain the increase in epithelial cell attachment observed at gastric pH [49]. To further investigate the extent to which the T3SS contributes to pH-mediated attachment, we evaluated the ability of a ΔescN C. rodentium strain lacking the T3SS to attach to CMT-93 cells following preinduction as before. We found that infection with the ΔescN strain did not result in decreased epithelial attachment following preinduction in LB at low pH 3.5 and 4.4 (S2H Fig), further supporting that low pH induced epithelial attachment is mediated by non-T3SS factors. To determine whether the induction of T3SS-related genes at pH 4.8 results in increased actin polymerization and pedestal formation—a hallmark of infection by attaching and effacing pathogens such as C. rodentium, EHEC, and EPEC–CMT-93 cells were infected as before with either C. rodentium pre-induced at pH 4.8 or neutral pH 7.0. As CMT-93 cells do not demonstrate robust visual actin polymerization, we have further performed these assays using the human intestinal adenocarcinoma cell line Caco-2/TC7. Following infection, cells were stained for translocated intimin receptor (Tir), a core component of the T3SS used to stain for actin-rich T3SS-induced pedestals on epithelial cells [50]. We found that average number of pedestals per Caco-2/TC7 epithelial cell was significantly higher following 4-h infection with pH 4.8-induced C. rodentium (Fig 2D). As expected, C. rodentium was less able to induce pedestal formation in CMT-93 cells. However, we still observed a trend towards increased pedestal formation following infection with bacteria pre-induced at pH 4.8 (S2I Fig). Together, these data demonstrate that C. rodentium regulates virulence gene expression in a pH-dependent manner, leading to increased attachment-related behavior upon initial entry to the intestinal environment.

Intestinal pH activates transcription of diverse bacterial stress responses Bacterial stress responses and virulence are closely co-regulated [26,51]. Therefore, we next investigated the relationship between intestinal pH and the transcriptional expression of stress-related genes. A key strategy of food-borne pathogens under acid stress is to increase intracellular pH, which can be done via production of amino acid decarboxylases which allow for the consumption of cytoplasmic protons [52]. Interestingly, we did not find homology within the C. rodentium genome to either the transcriptional regulators GadW and GadX of the glutamate decarboxylase acid-resistance system, nor to AdiC of the arginine acid-resistance system found in some E. coli strains, including EHEC [53,54]. We did however identify 70% protein sequence homology between the E. coli CadA, a lysine decarboxylase, and C. rodentium LdcC (S3A Fig). As expected, we found that transcript levels of ldcC increased at low pH from 4.4 to 5.7 (Fig 3A), indicating a potential role for LdcC in survival of acid stress. To further investigate this, we challenged wild-type C. rodentium and a C. rodentium strain lacking ldcC to gastric pH 3.5 over a range of 0 to 120 min. We found that viability of the ΔldcC strain was significantly impaired at 120 min as compared to the wild-type strain (S3B Fig). These data indicate that C. rodentium uses amino acid decarboxylation to combat acid-induced stress despite lacking homology to 2 major systems in related E. coli. However, the effect of ldcC deletion was subtle, with no significant difference observed upon initial exposure to low pH, indicating that C. rodentium likely employs several strategies of acid adaptation. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Stress gene regulation is affected by gastrointestinal pH. (A) Expression of acid stress-associated genes across the range of gastrointestinal pH by RT-qPCR. Data represent biological replicates (average of 3 technical replicates; N = 4). (B) Expression of general stress-associated genes across the range of gastrointestinal pH by RT-qPCR. Data represent biological replicates (average of 3 technical replicates; N = 3–4). (C) Intracellular ROS accumulation at 3 h postinoculation into LB at gastrointestinal pH, as determined using a roGFP2 reporter C. rodentium strain. Data represent biological replicates (average of 3 technical replicates; N = 3). Statistical analysis represents t tests with Bonferroni. Outlier tests were run using the ROUT method (Q = 1%), identifying no outliers. Summary data displayed in Fig 3 can be found in S1 Data. LB, lysogeny broth; ROS, reactive oxygen species. https://doi.org/10.1371/journal.pbio.3002761.g003 Another strategy of food-borne pathogens in dealing with acid stress is proton efflux [52]. The F 1 F 0 -ATPase proton efflux membrane ATPase is conserved across many bacterial pathogens, including C. rodentium, in which it is encoded by the atpD gene. We observed a 7-fold increase in expression of atpD at pH 4.4 compared to neutral media (Fig 3A). Additionally, upon searching the C. rodentium genome for genes annotated as involved in acid resistance, we chose to further investigate the expression of yodD, a stress-induced peroxide/acid resistance protein, and asr, the acid resistance repetitive basic protein, both of which are shared with EHEC. Both genes were up-regulated at gastric pH 3.5, and asr in particular demonstrated a 300-fold increase in expression at pH 4.4, suggesting that both of these genes are important to Citrobacter’s acid resistance strategy within the upper GI tract. Aside from direct mechanisms of acid resistance, we further profiled the effect of intestinal pH on nonspecific stress responses which have evolved as general mechanisms to overcome environmental stress. We investigated the expression of rpoS, a gene encoding the stationary phase sigma factor present in gram-negative bacteria which controls general stress responses to stressors such as oxidative or osmotic stress. RpoS is important for C. rodentium survival under heat and H 2 O 2 conditions, as well as full pathogen virulence [55,56]. Mutation of rpoS in EHEC has also been found to decrease resistance to acid stress [57]. While rpoS expression was therefore expected to increase with decreased pH, it was unaltered across the pH range tested (Fig 3B). This was surprising, given that when we exposed a roGFP2 reporter strain of C. rodentium, which can be used to measure intra-bacterial redox dynamics, to different pHs we found that gastric pH 3.5 to 4 significantly increased oxidative stress (Fig 3C) [41]. We further investigated expression of fur, encoding a ferric iron transcription regulator which may regulate transcription of genes that protect against ROS damage [52]. While we observed a trend towards increased fur expression at pH 4.4, there was no significant up- or down-regulation of the system induced by altered pH. Despite the lack of ROS-specific gene expression, we did observe evidence that bacteria at low pH respond to general cell damage. For example, we found that expression of recA which encodes a recombinase protein used to repair DNA double-strand breaks increased with increased acidity of the culture media, suggesting cell stress and DNA damage at pH 4.8 and lower (Fig 3B) [58].

pH-induced stress alters membrane regulation and cellular morphology An important outcome of stress-induced gene regulation is the repair and maintenance of the bacterial cell membrane, which can even induce adaptive changes to cellular morphology. To visualize potential morphological changes induced by pH, C. rodentium was pre-induced at GI pH before staining with FM 1–43 Dye. We observed bacterial cell lengthening at pH 4.0 to 4.8, an indicator of cell stress, and a trend towards small size and abnormal cell shape at pH 3.5 (Fig 4A and 4B). We also observed a decrease in membrane intensity in a subpopulation of cells at pH 3.5, possibly reflecting a loss of membrane integrity [23]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Membrane regulation in response to intestinal pH. (A) Representative cellular morphology after 3.5 h at relevant gastrointestinal pHs by light microscopy and cell membrane staining. Scale bar represents 5 μm. (B) Cell area measurements from light microscopy across gastrointestinal pH (N = 3 biological replicates, each the sum of 2 technical replicates). (C) Expression of genes by RT-qPCR associated with the Cpx stress response following culture at gastrointestinal pH. Statistical analysis represents t tests with Bonferroni correction for multiple comparisons. Outlier tests were run using the ROUT method (Q = 1%), identifying no outliers. Data represent biological replicates (average of 3 technical replicates; N = 3). (D) Survival of wild-type, ΔcpxRA, ΔcpxRA::cpxRA C. rodentium in gastric pH 3.5 over the range of average gastric emptying time. Each point represents the average of 3–5 biological replicates (3 technical replicates per biological replicate). Statistical analysis represents a mixed-effects model with Geisser–Greenhouse correction and Šidák’s multiple comparisons test. Color of stars indicates significance of Δcpx (pink) or complement (green) strains compared to the WT strain. Summary data displayed in Fig 4 can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3002761.g004 The Cpx envelope stress response is an important mechanism to overcome conditions which result in misfolding of envelope proteins, such as alkaline pH, and has been shown to be required for C. rodentium virulence [51,59]. We found that cpxA, part of the CpxRA two-component system which mediates the Cpx response, was most highly expressed at low pH 4.4 [51]. CpxRA is known to regulate both the expression of periplasmic proteases such as DegP, as well as T3SS-related virulence. We found that expression of degP, which is induced by stress [60], was significantly up-regulated in response to subculture at pH 3.5, 4.4, 5.7, and 6.0 compared to neutral base media (Fig 4C). degP was not up-regulated at pH 4.8 which resulted in a strong activation of T3SS-related genes, consistent with reported inverse regulation of these systems (Fig 2C), and supporting Cpx stress response regulation under low upper intestinal pH conditions. The above data suggest that membrane protein synthesis and degradation must be tightly regulated to allow for successful pathogen adaptation to variable intestinal pH. If this is the case, interfering with low pH-induced membrane regulation will impact C. rodentium survival under acidic conditions. We therefore tested survival at gastric pH 3.5 of wild-type C. rodentium or a strain deficient in the CpxRA two-component system (ΔcpxRA), and which is unable to properly regulate its membrane stress response [51]. While there was no difference upon initial exposure to pH 3.5 (0 min) between wild-type, ΔcpxRA, and ΔcpxRA::cpxRA strains, the ΔcpxRA strain exhibited significantly reduced survival after 30 to 120 min, with up to an average of 3.4% viability in the absence of cpxRA after 120 min, which represents the upper limit of gastric emptying time, as compared to 53.5% in the complement strain (Fig 4D) [61]. The Cpx response is therefore not only necessary for C. rodentium virulence, but is also critical to gastric passage and overall host–host transmission. Overall, these data suggest that C. rodentium alters its membrane and overall shape under gastrointestinal pH conditions to adapt to its surrounding environment.

Host-adaptation increases C. rodentium acid tolerance which may support host transmission During an established infection, C. rodentium shed from the gut at peak infection is known to display a hypervirulent state characterized by altered T3SS expression and tissue colonization dynamics, and the ability to colonize new hosts at a reduced infectious dose, allowing for more effective transmission. This hypervirulent state is quickly lost following ex vivo passage, indicating that it results from within-host adaptation rather than evolutionary change [62,63]. We hypothesized that a crucial part of this host-adapted (HA) state would include adaptation to better tolerate the gastric environment. While the conditions that trigger host-adaptation remain unknown, since host-adaptation occurs following passage through the mouse gut, we investigated whether in vitro culture at colonic pH representing the key in vivo niche of C. rodentium, ahead of gastric pH exposure, could induce an ATR and mimic the protection observed after gastrointestinal transit. Indeed, we observed a striking ATR following stationary phase growth at colonic pH 5.7 under anaerobic conditions (Fig 5A). While culture at neutral pH resulted in only 20.6% survival after 2 h at gastric pH, subculture at colonic pH 5.7 resulted in an average survival rate of 83.6%. Anaerobic conditions were chosen to mimic passage through the low oxygen large intestinal environment. We did not observe an ATR under aerobic conditions nor in response to short-term exposure to low pH (preinduction to mid-log phase) (S4A Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. HA C. rodentium is better able to survive acid-induced stress. (A) C. rodentium survival at gastric pH after initial anaerobic preinduction at either colonic pH 5.7 or neutral pH. Each point represents the average of 3 biological replicates (3 technical replicates per biological replicate). Statistics represent a two-way ANOVA with Šidák’s multiple comparisons test. (B) Survival of C. rodentium in homogenized feces collected on days 2 and 7 postinfection, representing non-adapted and host-adapted states (N = 6–8). ND, not detected. Summary data displayed in Fig 5 can be found in S1 Data. HA, host-adapted; LB, lysogeny broth. https://doi.org/10.1371/journal.pbio.3002761.g005 Having demonstrated a robust ATR under colonic pH conditions, we next sought to determine whether HA-C. rodentium conditioned within the host gut, demonstrates increased acid tolerance. We evaluated the survival of C. rodentium shed in the feces of infected mice at day 2, representing an early, non-HA state, and C. rodentium shed at day 7 postinfection, representing the HA state (Fig 5B). Though we found similar percent survival at 0 t 30 min postexposure to pH 3.5, none of the day 2-shed C. rodentium was able to survive 120 min at gastric pH. As these fecal samples were homogenized before exposure to pH 3.5 rather than naturally encapsulated in the fecal pellet, we repeated this assay on day 2 feces without homogenization. Despite an increase to 8% survival, we still observed significant population loss (S4B Fig). This indicates that while natural encapsulation, such as within contaminated food, contributes to “naïve” non-HA pathogen passage of the gastric barrier it must be combined with acid adaptation to facilitate high efficiency transmission. Altogether, these data suggest that C. rodentium adapts to the acidic conditions of the colon during colonization, possibly conferring a competitive advantage to initial gut seeding.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002761

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.