https://orcid.org/0000-0001-6750-7867
Figures
Abstract
Previously, we had demonstrated that Escherichia coli O157:H7 (O157) strain 86–24 expresses proteins involved in survival rather than virulence in vitro in rumen fluid from dairy cattle limit fed a maintenance diet. Here, we verified if this observation would be true for different O157 strains grown in vitro in rumen fluid from, and in vivo in the rumen of, animals on contrasting maintenance (high fiber) and lactation (high energy-protein) diets usually limit fed to dairy cattle. For the in vivo studies, an economical, novel, reusable and non-terminal rumen-fistulated animal model permitting simultaneous evaluation of multiple bacterial strains in the bovine rumen was developed. All experiments were conducted in duplicate using different animals to account for host-related variations. The O157 strains included, 86–24, EDL933 and the super shed SS-17. E. coli NalR (#5735), derived from a bovine intestinal commensal E. coli, was included as a control. As expected, diet influenced ruminal pH and volatile fatty acid (VFA) composition. The pH ranged from 6.2–7.0 and total VFA concentrations from 109–141 μM/ml, in animals fed the maintenance diet. In comparison, animals fed the lactation diet had a ruminal pH ranging between 5.18–6.0, and total VFA of 125–219 μM/ml. Strain dependent differences in O157 recovery from the rumen fluid of cattle fed either diet was observed, both in vitro and in vivo, with O157 strains 86–24 and EDL933 demonstrating similar survival patterns. Analysis of the O157 proteomes expressed in the rumen fluid/rumen verified previous observations of adaptive responses. Any difference in the adaptive response was mainly influenced by the animal’s diet and growth conditions (in vitro and in vivo) and not the O157 strain. These new insights into the O157 responses could help formulate modalities to control O157 across strains in cattle at all stages of husbandry.
Citation: Kudva IT, Trachsel J, Biernbaum EN, Casey T (2022) Novel reusable animal model for comparative evaluation of in vivo growth and protein-expression of Escherichia coli O157 strains in the bovine rumen. PLoS ONE 17(5): e0268645. https://doi.org/10.1371/journal.pone.0268645
Editor: Pierre Germon, INRAE Centre Val de Loire: Institut National de Recherche pour l’Agriculture l’Alimentation et l’Environnement Centre Val de Loire, FRANCE
Received: December 7, 2021; Accepted: May 4, 2022; Published: May 26, 2022
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the paper and its Supporting Information files. In addition, the complete R script used to perform statistical test and figure generation is available at https://github.com/USDA-FSEPRU/O157_rumen_proteomics_1/.
Funding: This work was supported by USDA-ARS CRIS project 5030-32000-100-00D and 5030-32000-112-00D. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Importance of this study
This study demonstrates that O157, irrespective of strain, have similar survival and protein expression patterns in rumen fluid with variations being influenced primarily by rumen fluid composition and the in vitro or in vivo environment. Hence, evaluating microbes within the host is critical when seeking to understand colonization dynamics and identifying targets for control strategies. Most importantly, the method presented here for conducting such in vivo studies in the rumen will permit reuse of the animal and allow for multiple evaluations.
Introduction
Shiga toxin-producing Escherichia coli (STEC) is the third leading cause of foodborne illness after Campylobacter and Salmonella, implicated in 265,000 illnesses annually in the US [1,2]. Globally, STEC infections are reported to cause approximately 2.8 million acute illnesses with 230 deaths annually [3]. Escherichia coli O157:H7 (O157) was the first STEC serotype to be associated with diarrhea, bloody diarrhea or hemorrhagic colitis and hemolytic uremic syndrome (HUS) in humans [4–7] and has since caused frequent outbreaks in the US [2,8]. STEC such as O157 can infect humans in low doses (~10 viable bacteria) [5,9–11] and are acquired through the fecal-oral route following ingestion of bacteria-contaminated food or water, or after contact with infected animals and humans [5,9,12,13]. No specific therapies are available for treating STEC infections in humans. Cattle are considered the primary STEC reservoirs as most outbreaks are directly or indirectly associated with cattle [14,15]. Cattle remain asymptomatic [16], although STEC can be isolated from various gastrointestinal tract sites and tend to persist at the rectoanal junction (RAJ) [17–19]. Cattle shed STEC in a seasonal pattern, with increased shedding in warmer months and decreased shedding in winter [15]. Average duration of bovine O157 carriage is 30 days, although colonization of up to 1 year has been reported [20–22]. The typical shedding rate of O157:H7 from the RAJ is 10–100 colony forming units per gram (CFU/g) of feces [23]. Animals shedding greater than 104 CFU/g feces, termed “super-shedders” contribute to herd prevalence of STEC [24,25].
Studies evaluating colonization sites within the bovine gastrointestinal tract (GIT) have invariably recovered some O157 from the rumen even if in relatively low numbers compared to other GIT compartments [19,26]. Considering the challenging environment in the rumen with feed, digesta, diverse microbiota, and various chemicals, it is of interest to understand mechanisms of bacterial survival and chemical sensing in this GIT compartment that enables downstream colonization of other intestinal sites [18,19,27–31]. Chemical sensing studies examining AHL (acyl-homoserine lactones) sensing have been conducted to either establish the role of SdiA, an AHL receptor, in the bacterial-bovine commensal relationship [27,32] or to examine the effect of AHLs on STEC colonization after consumption of various diets [33]. For instance, rumen cannulated heifers fed a grain diet [32] or steers fed a grain or hay (forage) diet [33] were inoculated with STEC and changes in rumen AHLs were analyzed using thin-layer chromatography and correlated with ruminal bacteria recovered. Studies like these could provide insights into reducing STEC shedding by manipulating specific chemical sensing pathways and by feeding cattle a diet less conducive to STEC colonization. Probiotic, microbiome, and proteomic analyses have also been conducted using rumen fluid. Bertin et al. explored the use of Lactobacillus reuteri strains as potential probiotics for finishing beef cattle prior to slaughter to reduce ruminal STEC concentrations [34]. In that study, glycerol-supplemented rumen fluid from cannulated Holstein cattle was co-inoculated with L. reuteri and STEC strains FCH6 and EDL933 under anaerobiosis in vitro, to demonstrate STEC suppression by the putative probiotic bacteria [34]. A similar study by Zhou et al. assessed the impacts of Tasco, an Ascophyllum nodosum extract marketed by Acadian Seaplants Ltd., Dartmouth, NS, Canada, on the ruminal microbiome and STEC in rams to serve as a feed additive to reduce fecal STEC shedding [31]. Tasco decreased E. coli populations in the rumen 10-fold and reduced O45, O103, O111, and O121 in the feces [31]. Additionally, rumen fluid has been used as a means of analyzing bacterial survival. One such study by Free et al. compared in vitro survival of O157 and non-O157 STEC in rumen fluid medium and observed significant decline in viable counts for serotypes O103:K:H8 and O26:H11 post-24 hr incubation [35].
In our previous study, we cultured O157 strain 86–24 in vitro in different preparations of rumen fluid and LB media and observed that O157 primarily expresses proteins involved in bacterial adaptation and not virulence in rumen fluid [28]. However, most studies including ours, evaluating STEC growth in bovine rumen fluid and making inferences were either conducted in vitro, or in terminal in vivo experiments with limited bacterial recovery from the rumen. Hence, we sought to develop a method that would permit non-terminal in vivo studies in rumen cannulated animals with ensured recovery of bacterial inoculum. Once the method was standardized, we evaluated three different O157 strains in the rumen of cows on contrasting diets usually fed to dairy cattle and compared the in vivo bacterial profiles to that observed in vitro in the same rumen fluid to determine differences in O157 survival and adaptive responses.
Materials and methods
Bacterial strains
The four different E. coli strains used in this study included, (i) E. coli NalR (NADC stock #5735; stx1-, stx2-, eae-), a non-pathogenic, nalidixic-resistant derivative of a bovine commensal E. coli, (ii) O157 strain 86–24 (stx2+, eae+), a wild-type, human, O157:H7 isolate from an outbreak in Washington State [36,37], (iii) O157 strain EDL933 (ATCC®43895™; stx1+, stx2+, eae+), a wild-type, human, O157:H7 isolate associated with the 1982 multi-state outbreak involving contaminated hamburgers (6), and (iv) O157 strain SS17 (stx2+, eae+), a bovine, wild-type, super-shed O157 [38].
Rumen fistulated cattle
All animal experiments were done using protocols approved by the Institutional Animal Care and Use Committee at the National Animal Disease Center (NADC, Ames, IA). A total of six rumen-fistulated Holstein heifers/cows between 1 to 4 years of age and routinely used as rumen fluid ‘donors’ at the NADC were utilized for this project. One animal (animal #A) was part of the pilot study standardizing the ruminal-challenge method with E. coli NalR (#5735) and five (animals #A2, B, C, D, E) were used in the test experiments evaluating the three O157 strains with E. coli NalR (#5735) as a control. Feed intake and body temperatures were monitored on a daily basis. Animals were housed in the field barn until ready for ruminal-challenge with O157 strains when they were moved in to the BSL-2 containment barns at the NADC. At the end of each study, following confirmation of their non-O157 shedding status, all animals were returned to the field barn. The animals were limit fed a total mixed ration maintenance diet (M diet) for the first half of the study lasting for 2 weeks following which the diet was gradually switched to the lactation diet (L diet) over 2 weeks; the L diet was continued into the second half of the study lasting for another 2 weeks. Water was provided ad libitum throughout. The maintenance diet is usually formulated such that feed energy intake will result in no net loss or gain of energy in the cow’s body thereby maintaining overall weight [39,40]. On the other hand, a lactation diet is designed to meet the high energy needs of the cow during milk production [39,40]. All dietary compositions were monitored by an animal nutritionist on site for our study to closely match the defined requirements. The M diet consisting of 25% grass hay, 65% corn silage, and 10% Steakmaker 40–20 (product composition label is shown in S1 Fig. Land O’Lakes, Inc., Arden Hills, Mn.) was limit fed at ~19 lbs/head with access to pasture or at ~30 lbs/head without pasture. The L diet was limit fed at ~99–100 lbs/head and included 2.5% grounded corn, 62% corn silage, 2.5% soybean meal, 12% legume hay and 21% lactation premix (product composition label is shown in S1 Fig).
Bacterial culture
(i) Inoculum preparation.
Overnight bacterial cultures grown in Luria-Bertani (LB) broth, with or without nalidixic acid (100 μg/ml; LB-Nal), at 39°C with aeration, were used to set up log-phase sub-cultures in 50 ml LB/LB-Nal broth, under the same growth conditions. Bacteria harvested from the log-phase cultures at an OD600 0.5–0.6, washed and re-suspended in sterile 0.9% saline, were used to charge dialysis tubings or cartridges as described below under “In vitro and in vivo evaluation of bacterial strains”. All O157 cultures were confirmed serologically using latex agglutination kits (E. coli O157 latex, Oxoid Diagnostic Reagents, Oxoid Ltd., Hampshire, United Kingdom) and plated for viable counts as described below.
(ii) Isolation of bacteria from dialysis tubings/cartridges.
Serial dilutions of bacterial suspensions, aspirated from the dialysis tubings/cartridges post-incubation, were prepared using sterile saline and plated on to LB-Nal and/or sorbitol MacConkey (BD Biosciences) with 4-methylumbelliferyl-β-d-glucuronide (100 mg/liter; Sigma; SMAC-MUG) agar to determine viable colony counts in CFU/ml. Bacteria adhering to the tubing or cartridge walls were scrapped off, using sterile inoculum loops, into the suspension within before aspirating the same. Bacterial suspensions and colonies were evaluated for the O157 serotype using latex agglutination kits (E. coli O157 latex, Oxoid).
(iii) Rumen fluid and fecal sample testing for O157.
To determine the presence of any O157 in the animal, rumen fluid and fecal samples were cultured, over two weeks, both at the beginning and end of study. In an unrelated study, direct inoculation of the rumen with 8 ml of 109 cfu/ml O157 strain 86–24 resulted in the recovery of the bacteria in cattle by selective-enrichment from feces within 24 h of inoculation and from the rumen fluid after a week (data not shown). Hence, a two week period was chosen to monitor any possible contamination of the animals used in this study. Previously standardized non-enrichment and selective-enrichment culture protocols were used with modifications [41–43]. Briefly,10 ml rumen fluid or 10-g fecal sample was added to 50 ml Trypticase soy broth (BD Bioscience, San Jose, Ca.) supplemented with cefixime (50 μg/liter; U.S. Pharmacopeia. Washington D.C), potassium tellurite (2.5 mg/liter; Sigma-Aldrich Corp., St. Louis, Mo.), and vancomycin (40 mg/liter; Alfa Aesar, Haverhill, Ma.) (TSB-CTV) or LB-Nal broth and mixed well. Serial dilutions of each sample were prepared with sterile 0.9% saline both before and after overnight incubation of the TSB-CTV or LB-Nal suspension at 37°C with aeration. The dilutions prepared before incubation (non-enrichment cultures) and after overnight incubation (selective-enrichment cultures) were spread plated onto SMAC-MUG and LB-Nal plates. SMAC-MUG plates were read after overnight incubation at 37°C and colonies that did not ferment sorbitol or utilize 4-methylumbelliferyl-β-d-glucuronide (non-fluorescent under UV light) were further evaluated to be O157 serologically.
Rumen fluid preparation and analysis
Rumen fluid samples collected from rumen-fistulated cows, fed the M diet (MRF) or the L diet (LRF), were used for culture and to set up the in vitro experiments for this study. Up to five liters of either rumen fluid was collected 2–3 hr post-feeding to allow for rumination to occur, at each sampling [44,45]. As previously described, the rumen fluid was strained through cheesecloth to remove large feed particles and poured into collection flasks [28]. Following recording of the pH and freezing of aliquots for volatile fatty acid (VFA) analysis, the remaining strained rumen fluid was aliquoted into flasks for in vitro studies as described below. The pH of the rumen fluid was determined using pH paper (pH range 5.0–8.0; Micro Essential Laboratory Inc., Brooklyn, NY) and VFA concentrations determined by capillary gas chromatography, as described previously [46,47], on an Agilent 6890 N gas chromatograph (Agilent Technologies, Inc., Santa Clara, CA). Three technical replicates were used per rumen fluid sample to determine VFA concentrations. The panel of substrates evaluated to determine total VFA included various by-products of carbohydrate and protein fermentation such as, acetate, propionate, butyrate, isobutyrate, formate, lactate, valerate, isovalerate, caproate, oxalate, phenylacetate, succinate and fumarate [39,40,48]. However, three VFA concentrations, acetate, propionate, and butyrate, were specifically studied as these influence the growth and energy dynamics of the bovine host [39,40,48]. The rumen fluid pH and VFA were recorded before and after introduction of O157 strains at the end of each incubation, in both in vitro and in vivo studies. Intermittent sampling was avoided to prevent any inadvertent influence of repeated exposure of rumen fluid/rumen to external environment.
In vitro and in vivo evaluation of bacterial strains
All experiments were done in duplicate and followed the plan as shown in S2 Fig. Since the overall goal was to determine the influence of diet on bacterial gene expression in the rumen fluid, steps were taken to minimize any other host-related effects. This was done by mixing rumen fluid collected from different animals for the in vitro studies and by alternating the animals used for the in vivo studies between the first and second challenges.
(A) In vitro evaluation in rumen fluid.
Rumen fluid collected from two different animals were mixed to account for any host-related variations in the processing of the feed and distributed in 300 ml aliquots per sterile flasks with rubber stoppers. To create anaerobic culture conditions, the flasks were transferred into the anaerobic Coy Chamber for 72 hr, charged with bacteria containing dialysis tubings or cartridges within the chamber, sealed and then removed for appropriate incubation. Dialysis tubings or cartridges fitted with membranes were used to contain the test bacteria within, allowing for easy recovery, while permitting diffusion of rumen fluid through. (i) With dialysis tubing: Bacteria were grown to log-phase in LB broth (OD600 0.5), washed and resuspended in sterile phosphate buffered saline (PBS, pH7.4). Four ml of each bacterial suspension was added to separate sterile dialysis tubing (Spectra/Por Type F, PVDF: 80,000 kDa cut off; Serva Electrophoresis, Heidelberg, Germany). A total of 2–4 dialysis tubings per bacterial strain were then suspended in separate flasks of anaerobic rumen fluid in the Coy Chamber, prior to sealing the flasks [49]. To correlate with the feed turnover rates in the rumen [28,50], the flasks were incubated for 48 h at 39°C, anaerobically, before harvesting bacteria from the tubings for microbiological and proteomics analysis [51]. An aliquot of the aspirated bacterial suspension was used to determine colony counts. The rest of the aspirated suspension was pelleted by centrifugation at 7,000 rpm, 15 min at 4°C, washed three times with an equal volume of ice-cold sterile PBS, and processed to obtain cell lysate and pellet fractions for iTRAQ proteomics analysis [51]. (ii) With dialysis cartridge: Using the same experimental set up and processing conditions as described above, sterile dialysis cartridges (Ultra-test Dialyzer Chamber, Harvard Apparatus, Hollister, MA), charged with the 4 ml of the bacterial suspension in PBS and sealed at both ends with polycarbonate membranes (0.01um pore size, Harvard Apparatus), were used in place of dialysis tubings (Fig 1). A total of 2–4 dialysis cartridges per bacterial strain were used.
(B) In vivo evaluation in the rumen.
Two animals were used to set up the in vivo studies per dietary condition. The animals were alternated between the first and second challenge to account for possible host-related influences on the bacteria besides the diet itself. To ensure retrievability from the rumen, the dialysis tubings or cartridges, charged with bacteria as described above, were placed in mesh laundry bags tethered to a 3 ft. fishing line attached to the rumen cannula cap (Fig 1). All bags with test bacteria containing dialysis tubings or cartridges were introduced into the animal. Bags were retrieved from the rumen after 48 hr. and transported to the lab on ice where the integrity of the dialysis tubings/cartridges was verified.
Proteomics
(i) Protein sample preparation and labelling for Isobaric Tags for Relative and Absolute Quantification (iTRAQ) analysis.
Following each in vitro and in vivo experiment, harvested bacteria of same strain type and from same growth condition, were pelleted together. The pellets were washed three times with ice-cold sterile phosphate buffered saline (PBS; pH 7.4), frozen and subsequently processed to obtain bacterial cellular and membrane proteins for proteomic analysis as described previously [51]. For iTRAQ, the protein fractions were quantified [52], and dissolved in denaturant buffer (0.1% SDS (w/v) and dissolution buffer (0.5 M triethylammonium bicarbonate, pH 8.5) in the iTRAQ 8-plex kit (AB SCIEX Inc., Foster City, CA). A total of 100 μg of protein per sample were reduced, alkylated, trypsin-digested, and resulting peptides labeled with either of the iTRAQ tags 114, 115, 116,117, 118, 119 or 121, according to the manufacturer’s instructions (AB SCIEX Inc.). Peptides from bacteria grown in MRF-in vitro or -in vivo were labeled separately based on strain-growth condition and then mixed in one tube; the LRF-in vitro and LRF-in vivo peptides were processed likewise. The mixed labeled peptides were dried and held at -80°C until ready for processing as follows. The labeled peptides were desalted with C18-solid phase extraction and dissolved in strong cation exchange (SCX) solvent A (25% (v/v) acetonitrile, 10 mM ammonium formate, and 0.1% (v/v) formic acid, pH 2.8).
(ii) Strong cation exchange fractionation and reverse-phase liquid chromatography with tandem mass spectrometry (LC-MS/MS).
The peptides were fractionated using an Agilent HPLC 1260 with a polysulfoethyl A column (2.1 × 100 mm, 5 μm, 300 Å; PolyLC, Columbia, MD, USA). Peptides were eluted with a linear gradient of 0–20% solvent B (25% (v/v) acetonitrile and 500 mM ammonium formate, pH 6.8) over 50 min followed by ramping up to 100% solvent B in 5 min. The absorbance at 280 nm was monitored and a total of 10 fractions per sample were collected. The fractions were lyophilized and resuspended in LC solvent A (0.1% formic acid in 97% water (v/v), 3% acetonitrile (v/v)). A hybrid quadrupole Orbitrap (Q Exactive Plus) MS system (Thermo Fisher Scientific, Bremen, Germany) was used with high energy collision dissociation (HCD) in each MS and MS/MS cycle. The MS system was interfaced with an automated Easy-nLC 1000 system (Thermo Fisher Scientific). Each sample fraction was loaded onto an Acclaim Pepmap 100 pre-column (20 mm × 75 μm; 3 μm-C18) and separated on a PepMap RSLC analytical column (250 mm × 75 μm; 2 μm-C18) at a flow rate at 350 nl/min during a linear gradient from solvent A (0.1% formic acid (v/v)) to 30% solvent B (0.1% formic acid (v/v) and 80.0% acetonitrile (v/v)) for 95 min, to 98% solvent B for 15 min, and hold 98% solvent B for additional 30 min. Full MS scans were acquired in the Orbitrap mass analyzer over m/z 400–2000 range with resolution 70,000 at 200 m/z. The top ten most intense peaks with charge state ≥ 2 were isolated (with 2 m/z isolation window) and fragmented in the high energy collision cell using a normalized collision energy of 28% selected. The maximum ion injection time for the survey scan and the MS/MS scans were 250 ms, and the ion target values were set to 3e6 and 1e6, respectively. The selected sequenced ions were dynamically excluded for 60 sec.
Intensity Based Absolute Quantification (iBAQ) analysis of the LC-MS/MS data
Label-free iBAQ values for each protein within each rumen fluid were determined using MaxQuant to analyze the raw LC-MS/MS data [53]. For protein identification, a custom E. coli database was generated by combining protein sequences from 162 O157 genomes available from NCBI and is available in the supplementary file: NCBI_Ecoli_used_for_reference.txt. Redundant proteins were collapsed with CD-HIT [54]. Additionally, we included the Uniprot E. coli panproteome UP000000625 [55] to ensure the most comprehensive reference possible. MaxQuant was run using the default parameters, with the selection of calculating label free quantification iBAQ values; detailed information about the parameters used is available in supplementary file: MQ_O157_parameters.txt. iBAQ values were normalized, using R scripts, by dividing them by the sum of iBAQs in the sample to obtain relative iBAQs (riBAQs) [56]. These riBAQ values were used to compare relative protein abundances between the two different rumen fluids. We calculated log2 fold change (L2FC) values between MRF and LRF using riBAQ values (Log2(riBAQlrf/riBAQmrf)) and considered any protein with an absolute L2FC value of >1 to be enriched in the respective rumen fluid.
Isobaric Tags for Relative and Absolute Quantitation (iTRAQ) analysis of the LC-MS/MS data
(i) Database searching.
Tandem mass spectra were extracted by Proteome Discoverer (Thermo-Fisher) version 2.2.0.388. Charge state deconvolution and deisotoping were not performed. All MS/MS samples analyzed were processed by a thorough database searching considering biological modification and amino acid substitution against a non-redundant Uniprot E. coli strain K12 database (4,448 entries), E. coli O157:H7 strain EDL933, UP000028484 (5,270 entries) and E_coli_O157_20160822 database (75243 entries) with decoy option using MASCOT 2.4 (Matrix Science Inc., Boston, MA, USA) assuming the digestion enzyme trypsin. Mascot was searched with a fragment ion mass tolerance of 0.6–0.02 Da and a parental ion tolerance of 20–10 ppm. Methylthio of cysteine and iTRAQ8plex of lysine and the n-terminus were specified in Mascot as fixed modifications. Gln->pyro-Glu of the n-terminus, deamidated of asparagine and glutamine, methyl of lysine, oxidation of methionine, phospho of serine, threonine and tyrosine, nmethylmaleimide of cysteine and iTRAQ8plex of tyrosine were specified in Mascot as variable modifications.
(ii) Criteria for protein identification.
Scaffold (version Scaffold_4.8.7, Proteome Software Inc., Portland, OR) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95.0% probability by the Peptide Prophet algorithm [57] with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at greater than 95.0% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm [58]. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters.
(iii) Data analysis of peptides/proteins comprising the O157 proteome expressed in rumen fluid from animals fed the maintenance diet.
Scaffold Q+ (version Scaffold_4.8.7, Proteome Software Inc., Portland, OR) was used to quantitate Label Based Quantitation (iTRAQ, TMT, SILAC, etc.) peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95.0% probability by the Peptide Prophet algorithm [57] with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at greater than 95.0% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm [58]. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters. The false discovery rate (FDR) at the peptide level was set with the integrated PSPEP tool in the Protein Pilot Software to be at 1.0%. Channels were corrected by the matrix [0.000,0.000,0.929,0.0689,0.00220]; [0.000,0.00940,0.930,0.0590,0.00160]; [0.000,0.0188,0.931,0.0490,0.001000]; [0.000,0.0282,0.932,0.0390,0.000700]; [0.000600,0.0377,0.933,0.0288,0.000]; [0.000900,0.0471,0.933,0.0188,0.000]; [0.00140,0.0566,0.933,0.00870,0.000]; [0.000,0.000,0.000,0.000,0.000]; [0.00270,0.0744,0.921,0.00180,0.000] in all samples according to the algorithm described in i-Tracker [59]. Normalization was performed iteratively (across samples and spectra) on intensities, as described in Statistical Analysis of Relative Labeled Mass Spectrometry Data from Complex Samples Using ANOVA [60]. Medians were used for averaging. Spectra data were log-transformed, pruned of those matched to multiple proteins and those missing a reference value, and weighted by an adaptive intensity weighting algorithm. Of 9136 spectra in the experiment at the given thresholds, 5895 (65%) were included in quantitation.
(iv) Data analysis of peptides/proteins comprising the O157 proteome expressed in rumen fluid from animals fed the lactation diet.
Scaffold Q+ (version Scaffold_4.8.7, Proteome Software Inc., Portland, OR) was used to quantitate Label Based Quantitation (iTRAQ, TMT, SILAC, etc.) peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95.0% probability by the Peptide Prophet algorithm [57] with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at greater than 95.0% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm [58]. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters. FDR at the peptide level was set with the integrated PSPEP tool in the Protein Pilot Software to be at 1.0%. Normalization was performed iteratively (across samples and spectra) on intensities, as described in Statistical Analysis of Relative Labeled Mass Spectrometry Data from Complex Samples Using ANOVA [60]. Medians were used for averaging. Spectra data were log-transformed, pruned of those matched to multiple proteins and those missing a reference value, and weighted by an adaptive intensity weighting algorithm. Of 4924 spectra in the experiment at the given thresholds, 4739 (96%) were included in quantitation.
(v) iTRAQ Quantitation with (iTRAQ analysis #1) or without (iTRAQ analysis #2) reference comparison.
For quantification, only MS/MS spectra that were unique to a particular protein and where the sum of the signal-to-noise ratios for all the peak pairs at > 9 were used. The ratios with p-values less than 0.05 present in at least two replicates were considered significant. Only the significant ratios from the replicates were used to calculate the average ratio for the protein. For a protein to be determined as differentially expressed, it must have been identified and quantified with at least two unique peptides. The iTRAQ quantitation was done either in comparison to a reference (iTRAQ analysis #1) or in a reference-free manner (iTRAQ analysis #2).
For iTRAQ analysis #1, individual proteomes expressed by the strains, in vitro or in vivo, were compared separately against the proteome expressed by the O157 strain 86–24 under in vitro conditions (reference proteome) for each type of rumen fluid used in this study. Any O157 protein with L2FC, set in Scaffold to a minimum +/- 1 FC, equaling a 2-fold change in comparison to the reference was determined to be differentially expressed in the in vitro or in vivo growth conditions, in either MRF or LRF. The total L2FC value determined for each protein in the 10 fractions per strain sample was used to determine differential expression.
For iTRAQ analysis #2, reporter ion intensities across the strains were compared, for the in vitro or in vivo growth conditions, in each type of rumen fluid. The reporter intensities were averaged for all three strains in each growth condition, then the formula Log2(AVG_vitro / AVG_vivo) used for the final L2FC calculation. Differentially expressed proteins, between in vitro and in vivo conditions, in MRF and LRF were determined by applying Mann-Whitney Test with unadjusted significance level p < 0.05 corrected by Benjamini-Hochberg.
Statistics and Bioinformatics
The 2-way ANOVA test was used to compare VFA concentrations and a value of p<0.05 was considered significant. Several different methods were used to gain biological insights into the proteins enriched under various conditions. First, the predicted cellular location of all proteins was determined using PSORTb v3.0.2 [51,61]. Next, Gene Ontology (GO) terms to all proteins were annotated in the reference databases with InterProScan5 [62]. For sets of differentially expressed proteins enriched GO-terms were calculated with the R package TopGO [63–66]. The R package vegan [67] was used to perform multivariate statistical tests on the composition and dispersion of protein expression patterns between various conditions. The complete R script used to perform statistical test and figure generation is available at https://github.com/USDA-FSEPRU/O157_rumen_proteomics_1/.
Results
Rumen fluid pH and VFA concentrations varied with diet
The two diets had differing influences on the ruminal pH and volatile fatty acid (VFA) profiles (Table 1A and 1B in Fig 2 and Table 2A and 2B in Fig 3). The concentrations of three VFAs, acetate, propionate and butyrate that vary the most with changes in diet and influence bacterial and host growth, were closely analyzed in the context of the total VFA [28,48].
Graphs shown represent survival characteristics of three O157 strains in comparison to E. coli NalR (#5735), in vitro (A) and in vivo (B), in MRF. Bacterial survival characteristics depicted are following anaerobic incubation for 48 h, in vitro in flasks with MRF or in vivo in the rumen of animals fed the maintenance diet. Viable counts in colony forming units [cfu]/ml, with the standard error of means, are shown in both graphs. Key for each bacterial strain tested is also shown. Inserted tables show pH and VFA composition of MRF used in the in vitro (1A) and in vivo (1B) experiments.
Graphs shown represent survival characteristics of three O157 strains in comparison to E. coli NalR (#5735), in vitro (A) and in vivo (B), in LRF. Bacterial survival characteristics depicted are following anaerobic incubation for 48 h, in vitro in flasks with LRF or in vivo in the rumen of animals fed the lactation diet. Viable counts in colony forming units [cfu]/ml, with the standard error of means, are shown in both graphs. Inserted tables show pH and VFA composition of LRF used in the in vitro (2A) and in vivo (2B) experiments.
(i) In MRF. In the in vitro experiments, as shown in Table 1A in Fig 2, the ruminal pH ranged from 6.2 to 6.5 irrespective of the presence of the test bacteria. The total VFA concentrations ranged between 115–181 μM/ml, among animals fed the maintenance diet, with slightly increased VFA post-exposure to test bacteria. In the in vivo experiments, the ruminal pH ranged from 6.2 to 7.0 and the total VFA between 109–141 μM/ml (Table 1B in Fig 2).
(ii) In LRF. In the in vitro experiments, the ruminal pH ranged from 5.2–5.6, without any influence from the test bacteria (Table 2A in Fig 3). The total VFA concentrations ranged from 203–260 μM/ml; higher values were observed in post-exposure to test bacteria (Table 2A in Fig 3). Likewise, in vivo, the ruminal pH ranged from 5.1–6.1 and total VFA concentrations were between 125–212 μM/ml (Table 2B in Fig 3). Percent differences in butyrate concentration were more pronounced in vivo in LRF (Table 1B in Fig 2 and Table 2B in Fig 3; 8–9% in MRF versus 11–14% in LRF) as expected with a high energy diet [39,40]. In addition, total VFA concentrations were significantly different (p<0.0001) between LRF and MRF and may have contributed to the decrease in pH in this rumen fluid compared to MRF.
Dialysis cartridges enabled the development of a novel, non-terminal method for in vivo evaluation of bacteria in the bovine rumen
Dialysis tubings were compared against dialysis cartridges with membranes for the ability to retain test bacterial cultures within while maintaining structural integrity to prevent inadvertent spillage into the rumen fluid in the flask or the cattle rumen in the pilot study. The animals were fed the maintenance diet for this entire experiment. Saline suspensions (4 ml/tubing or cartridge) of log-phase E. coli NalR (#5735), a non-pathogenic bacteria, were used to charge both the tubings and cartridges as described above under Materials and Methods.
(i) In vitro pilot study.
The rumen fluid obtained from animal #A was at pH 6.63 with 54% acetate, 13% propionate, and 8% butyrate. Comparable results, both with bacterial counts and volume of inoculum recovered in 48 h, were observed while using dialysis tubings and cartridges to introduce E. coli NalR (#5735) into the rumen fluid in flasks (S1 Table). As shown in supplementary S1 Table, average of counts obtained from two tubings or cartridges indicated minimal change in viable counts post-anaerobic incubation for 48 h. No leakage of E. coli NalR (#5735), from the intact tubings and cartridges charged with the inoculum, into the surrounding rumen fluid was observed following culture of the same.
(ii) In vivo pilot study.
The rumen fluid in animal #A was at pH 7.06 with 68% acetate, 15% propionate and 9% butyrate at the time of this study. The culture results of this study are shown in the S1 Table and represent average of counts obtained from two tubings or cartridges. Using the dialysis tubing an average of 2-log reduction in the E. coli NalR (#5735) viable count was observed compared to a 1-log reduction with the dialysis cartridges after being in vivo in the bovine rumen for 48h (S1 Table). Of the 2 dialysis tubings introduced into the rumen, only one remained intact and the initial volume of 4 ml was reduced to 200μl; most of the inoculated bacteria was adherent on the tubing material itself. The base of one dialysis tubing was shred in the rumen, and we recovered E. coli NalR (#5735) only from the bovine feces, post-enrichment in LB-Nal broth, at the 48h sampling. In contrast to the tubings, both the dialysis cartridges remained intact, and 3 ml of the 4 ml initial volume was recovered from each of the cartridges. E. coli NalR (#5735) was found adhering to the walls of the cartridges as with the dialysis tubing. All adherent bacteria were scraped off the tubing or cartridge walls, using sterile inoculum loops, into the suspension within before aspirating the same. Overall, the dialysis cartridges performed better than the PVDF tubings in the in vivo experiment as the cartridges were better at resisting the shear forces of ruminal contractions. No leakage of E. coli NalR (#5735), from the cartridges charged with the inoculum, into the rumen of the animal was observed based on the intact assembly of the cartridge and negative rumen fluid- and fecal- cultures over two weeks post-experiment.
O157 strains demonstrated variable survival in MRF and LRF both in vitro and in vivo
Based on the results of the pilot study, dialysis cartridges were used for all the experiments testing bacterial strains simultaneously. For in vitro assays, rumen fluid collected from two animals was mixed prior to distribution into flasks to account for possible host-based variations. In case of in vivo assays, each of the duplicate experiments was conducted in separate animals for similar reasons. Strain dependent differences in O157 survival were observed in the rumen fluid of cattle fed either diet, in both the in vitro and in vivo assays, with O157 strains 86–24 and EDL933 demonstrating similar recovery patterns. In comparison, O157 strain SS-17 and the non-STEC E. coli NalR (#5735) had variable survival in either assay condition.
(i) In vitro survival characteristics in rumen fluid.
Following anaerobic incubation in MRF, in flasks at 39°C for 48 h, O157 strains 86–24 and EDL933 demonstrated 2-log reduction and O157 strain SS-17 a 1-log reduction in viable counts compared to the 3-log reduction of E. coli NalR (#5735) viable counts (Fig 2; S2 Table). Rumen fluid from one animal was used for the pilot-in vitro experiment while rumen fluid from two different animals was mixed for the test-in vitro experiment. This may have presented more comprehensive growth inhibitory factors resulting in the discrepancy in the E. coli NalR (#5735) viable counts recovered at 48h between the pilot and test experiments (S1 and S2 Tables). In LRF, after similar incubation, O157 strains 86–24 and EDL933 had a 2-log reduction and E. coli NalR (#5735) a 1-log reduction in viable counts while O157 strain SS-17 had a 5-log decrease in viable counts (Fig 3; S3 Table). Based on these results, it appeared that O157 strain SS-17 was more sensitive to the low pH, higher VFA conditions, in vitro, in LRF than the other O157 strains. None of the O157 strains tested were recovered from the rumen fluid in the flasks, prior to and post exposure to the bacteria in dialysis cartridges.
(ii) In vivo survival characteristics in the rumen.
Following 48 h incubation within the rumen of fistulated cows fed the maintenance diet, the viable counts of O157 strains 86–24 and EDL933, and E. coli NalR (#5735) reduced by 1-log compared to the 2-log reduction in counts of O157 strain SS-17 (Fig 2; S2 Table). E. coli NalR (#5735) counts recovered at 48h were consistent between the pilot- and test- in vivo experiments verifying the consistency of an in vivo system, and a more dietary than host- related influence on bacterial gene expression since different animals were used for these experiments (S1 and S2 Tables). Interestingly, in vivo in the rumen of animals fed the lactation diet, all O157 strains shared growth patterns that grouped them apart from E. coli NalR (#5735); viable counts decreased by 2-logs for all O157 strains and by 3-logs for E. coli NalR (#5735) (Fig 3; S3 Table). These results differed from the in vitro assays which further verifies the importance of conducting in vivo experiments as some cues present in a live host can be missed in in vitro systems. None of the O157 strains tested were recovered from the rumen fluid and fecal samples of the animals used in the study prior to and post exposure to the bacteria in dialysis cartridges.
Comprehensive analysis of proteomics data allowed identification of O157 proteins uniquely or differentially expressed under varying growth conditions in MRF and LRF
All analysis was done in the context of the diets influencing growth of O157 strains in the rumen fluid in vitro, or in vivo in the bovine rumen.
(i) iBAQ analysis identified O157 proteins uniquely expressed, in MRF and LRF.
iBAQ analysis was used to determine overall differences in O157 proteins expressed, differentially or uniquely, in MRF versus LRF without focusing on individual strains or the in vitro/in vivo growth conditions. Since the iTRAQ experiments were designed to give the highest resolution of differences between growth conditions (in vitro versus in vivo) within each rumen fluid (all iTRAQ labeled MRF samples were mixed and analyzed in one MS run, and all iTRAQ labeled LRF samples in another), the iTRAQ reporter intensity could not be used effectively to compare protein expression between the rumen fluids. However, because all samples pooled in each MS run belonged to the same rumen fluid type, iBAQ values of the same, which excludes labels and merges all data for a particular rumen fluid type, could be used to estimate the differences in protein expression between MRF and LRF. These iBAQ values represented pooled protein expression for all three O157 strains in both the in-vitro and in-vivo culture conditions for each rumen fluid. A much lower total iBAQ value was detected for the O157 proteome in LRF, which may be reflective of the decrease in viable counts of the strains grown under these conditions (S3A Fig). All the same, to avoid any false interpretations due to differences in intensities, iBAQ values were converted to riBAQ values to look at relative protein abundance that is, proportion of each protein within the total protein pool, for each rumen fluid type [68] (S3B Fig; S4A Table). A total of 756 O157 proteins were identified by riBAQ of which 566 proteins were detected in both rumen fluids, 179 were unique to MRF and 11 unique to LRF (Fig 4; S4B Table). The L2FC values were calculated between the two rumen fluids using riBAQ values and any protein with an absolute L2FC value of 1 or greater was considered to be enriched in the respective rumen fluid (Fig 4). Using these criteria 347 O157 proteins were found to be enriched in MRF (S4C Table) and 101 enriched in LRF (S4D Table). These differentially represented O157 proteins were used in a GO term enrichment analysis to identify GO terms that were significantly enriched in each rumen fluid (S5A and S5B Table).
iBAQ: A. Venn diagram showing the number of O157 proteins differentially expressed (L2FC >1) between MRF and LRF. B. Venn diagram showing the number of O157 proteins uniquely expressed in MRF and LRF. iTRAQ analysis #1 (reference-based): A. Venn diagram showing number of differentially expressed proteins in the in vitro and in vivo growth conditions, relative to the 86–24 in-vitro proteome, in MRF. B. Venn diagram showing number of differentially expressed proteins in the in vitro and in vivo growth conditions, relative to the 86–24 in-vitro proteome, in LRF.
(ii) Reference-based iTRAQ analysis #1 identified O157 proteins differentially expressed under in vitro and in vivo growth conditions relative to the 86–24 in vitro proteome, in MRF or LRF.
Post-application of the filtration (protein probability 95%, minimum number of peptides 2, peptide probability 95%) criteria as described in the Materials and Methods section above, a total of 756 and 316 O157 proteins were identified as being expressed, cumulatively, by the three O157 strains in MRF and LRF, respectively. Application of the quantitation criteria (minimum +/- 1 FC equaling a 2-fold change, in comparison to the previously reported [28] reference O157 strain 86–24 proteome expressed in vitro in MRF or LRF) to the total number of O157 proteins enabled the identification of proteins differentially expressed in the in vitro or in vivo growth conditions, with either MRF or LRF. After excluding duplicate hits and proteins with a value of absolute zero (no change) a cumulative list of differentially expressed proteins across the O157 strains was generated. Comparatively, slightly more proteins were differentially expressed in RF from animals fed the lactation diet (119 out of a total of 316 proteins; 38%) than in RF from animals fed the maintenance diet (106 out of a total of 756 proteins; 14%) (S6A and S6B Table). These differences in protein expression were not always O157 strain-specific as majority of the proteins were either up- or down-regulated across all three O157 strains tested with a few exceptions.
Proteins with a FC ≥ +1, in either of the three O157 proteomes relative to the 86–24 in vitro proteome, were designated as up-regulated for the specific growth condition. Taking these up-regulated proteins into account, of the 106 O157 proteins differentially expressed in MRF, 12 were found to be highly expressed in the in vitro condition, 54 were highly expressed in the in vivo condition while the remaining 40 were either down-regulated or equally regulated under both growth condition (Fig 4, Table 1A and 1B). Likewise, of the 119 O157 proteins differentially expressed in LRF, 19 were highly expressed in vitro, 52 were highly expressed in vivo and 48 were down-regulated or equally regulated under both growth condition (Fig 4, Table 1C and 1D).
(Reference:86–24 in vitro proteome).
(iii) Reference-free iTRAQ analysis #2 verified minimal association of differentially expressed proteins to specific O157 strains tested. By comparing iTRAQ reporter ion intensities across the strains, without comparing to a reference strain proteome, a total of 667 and 473 proteins were identified as being differentially expressed, cumulatively, by the three O157 strains in the in vitro and in vivo growth conditions of MRF and LRF, respectively. This reference-free iTRAQ analysis was done to remove possible biases introduced by iTRAQ #1 analysis and verify if similar observations would be made. As with iTRAQ#1 analysis, iTRAQ #2 analysis determined more O157 proteins to be up-regulated in the rumen fluid from animals on the lactation diet (154 out of a total of 473 proteins; 33%) compared to those on the maintenance diet (81 out of a total of 667 proteins; 12%). In MRF, 25 proteins were more highly expressed in the in vitro condition and 56 were more highly expressed in the in vivo condition (S7 Table). In LRF, 43 proteins were more highly expressed in the in vitro condition and 111 were more highly expressed in the in vivo condition (S7B Table). Using these differentially expressed proteins, a GO term enrichment analysis was preformed to identify GO terms that were significantly enriched in each of the culture conditions within each rumen fluid (S8A–S8D Table).
The normalized reporter intensities from this iTRAQ analysis #2 were then used to construct a Bray-Curtis dissimilarity matrix (multidimensional scaling or MDS) describing the similarities of the global protein expression between all samples within each rumen fluid (Fig 5). In MRF, the difference in global protein expression between the culture conditions was less evident (PERMANOVA p = 0.08, F = 3.7, R2 = 0.51) and no significant strain effect was detectable PERMANOVA p = 0.67) (Fig 5A, S9A Table). Within LRF, a significant difference in global protein expression patterns was detected between the two culture conditions, (PERMANOVA p = 0.05, F = 5.6, R2 = 0.572) but no significant difference was detected between the different strains (PERMANOVA p = 0.65) (Fig 5B, S9B Table). In both rumen fluids, protein expression patterns in the in vivo culture conditions appeared to have more variable protein expression patterns; this inference was derived from the dispersion distance plots as shown in Supplementary Fig 3. However, statistical testing failed to uncover sufficient evidence for this (PERMDISP2 p = 0.30 for both MRF and LRF).
The plots depicting the similarity of global protein expression between O157 strains tested in (A) MRF or (B) LRF, in vitro and in vivo. This is based on Bray-Curtis dissimilarities generated from Scaffold normalized iTRAQ reporter intensities generated in reference-free iTRAQ analysis #2.
Taken together these analyses indicated that differences in protein expression were mainly attributable to the growth conditions (in vitro and in vivo) and not the O157 strains. The proteome of O157 strain SS-17 appeared to have distinctive features compared to proteomes of 86–24 and EDL933, especially in vitro in LRF, although no statistical significance was observed. This distinction captured by both the MDS, and dispersion plots (Fig 5 and S4 Fig) appeared to reflect the survival patterns observed for SS-17 which varied from 86–24 and EDL933 (Figs 2 and 3).
iBAQ and iTRAQ analyses identified similar O157 proteins associated with adaptive responses
Based on the iBAQ analysis, overall, 179 O157 proteins were determined to be unique to MRF and 11 unique to the LRF (Fig 4; S4B Table). The 179 O157 proteins uniquely expressed in MRF could be functionally assigned to lipid and carbohydrate metabolism, protein and nucleotide metabolism, energy metabolism, regulatory, transcription and translation, transport, cell integrity, cell division, environmental adaption, and respiration (S4B Table). Similarly, the 11 O157 proteins uniquely expressed in LRF could be assigned to lipid and carbohydrate metabolism (FolE, KdsA), protein metabolism (DmlA), energy metabolism (AceF, MaeA, SucC), regulatory (EngB), transcription and translation (FusA1, PrfC, RpsL), and environmental adaptation (OsmC) (S4B Table). No protein directly associated with virulence was identified in the O157 proteomes expressed in both MRF and LRF.
Proteins with similar functions were also identified by iTRAQ analysis #1. In vitro, the O157 strains highly expressed more proteins in LRF (19 proteins) than in MRF (12 proteins), relative to the reference 86–24 in vitro proteome, especially proteins associated with protein, nucleotide and energy metabolism, transcription, and translation (Table 1). This increased expression of proteins in vitro in LRF was associated mainly with the O157 strain SS-17 for 12/19 proteins (AccC, Adk, DeoB, HslU, Lon, DkgA, GltA, PfkA, Pgk, Upp, SucD, RpoB), even though SS-17 had the most suppressed growth (5-log reduction) in this growth condition. This was in contrast to in vitro in MRF where only 5/12 proteins were exclusively high in SS-17 (PckA, HupA, Hfq, RplU, RpsD) (Table 1, S6A and S6B Table, Fig 3). In vivo, a greater number of proteins were up-regulated in both LRF (52 proteins) and MRF (54 proteins), relative to the reference proteome. For the most part the proteins with increased expression in vivo, relative to the reference 86–24 in vitro proteome, could be associated with all three or at least two of the test strains except for 8/52 (GlnA, GlyA, YhcB, GlyS, FhuA, OmpC, OmpF, GadA/B), and 13/54 (AccC, RpiA, AcnB, CydA, GltA, SucC, SucD, Fur, ArgS, GlnS, RpoA, ValS, NarZ) proteins that were primarily high in SS-17 in LRF and MRF, respectively (Table 1, S6A and S6B Table). As observed with LRF in vitro, SS-17 demonstrated suppressed growth in MRF in vivo by 2-logs compared to other test bacteria (Fig 3). The correlation between growth suppression and increased expression of proteins of the lipid and carbohydrate metabolism, energy metabolism and transcription and translation is unclear but may indicate attempt at survival by stressed bacteria [69]. In the reference-free iTRAQ dataset, 17 and 22 proteins overlapped with proteins identified as having increased expression by iTRAQ analysis #1, in MRF and LRF, respectively. These are shown in bold in Table 1 under each of the MRF and LRF growth conditions. The increased expression of proteins in SS-17 may have also contributed to the distinctive profile in MDS and dispersion plots (Fig 5 and S3 Fig). Irrespective of the level of expression, no virulence protein was identified in the O157 proteomes, following in vitro or in vivo growth in either rumen fluids.
Interestingly GadA/B, associated with the glutamate-dependent acid resistance system, was the only acid-resistance related enzyme expressed in LRF, with increased expression in vivo (Table 1, S9A and S9B Table). With MRF, GadA/B was very minimally expressed under in vitro conditions only while expression of SpeB and PotD, associated with the arginine-dependent acid resistance system, was enhanced in all 3 O157 strains in vivo (Table 1, S6A and S6B Table).
Discussion
In this study, we standardized a novel, non-terminal method that allows reuse of animals following in vivo studies with bacterial pathogens in the rumen of canulated cattle. The use of dialysis cartridges fitted with appropriate membranes ensured, (i) containment of bacteria being tested within the cartridges while exposing these bacteria to the rumen environment, (ii) simultaneous testing of multiple bacteria, and (iii) retrievability of these bacteria without euthanizing the animal.
Utilizing this method, survival, and protein expression of three O157 strains was evaluated along with a non-STEC E. coli NalR (#5735) as a comparative control, in vitro and in vivo, in ruminal fluids of contrasting compositions. Feeding dairy cows, a high fiber maintenance diet or a high energy-protein lactation diet resulted in rumen fluids with differing pH and VFA concentrations, termed MRF and LRF respectively. The pH ranged from 6.2–7.0 and total VFA concentrations from 109–141 μM/ml in MRF while LRF had a ruminal pH ranging between 5.2–6.1, and total VFA of 125–219 μM/ml (Table 1A and 1B in Fig 2 and Table 2A and 2B in Fig 3). An increase in the total VFA concentrations was observed post-incubation with test bacteria under in vitro growth conditions, in MRF (VFA of 167–181 μM/ml) and LRF (VFA of 215–260 μM/ml) (Table 1A in Fig 2 and Table 2A in Fig 3). As previously reported [28] this increase in in vitro VFA may have been due to breakdown of the limited nutrients by all microbes in the rumen fluid in the absence of a dynamic turn-over seen in vivo in the rumen of a consistently fed animal demonstrating the limitations in reproducing in vivo rumen conditions in vitro. Key differences were observed in the survival of the bacteria tested in either rumen fluids, under both growth conditions. The O157 strains 8624 and EDL933 had similar survival patterns, in vitro and in vivo, in both rumen fluids (Figs 1 and 2). In contrast, the O157 strain SS17 and E. coli NalR (#5735) had the most varied survival patterns. E. coli NalR (#5735) and O157 strain SS17 demonstrated the highest decrease in viable counts in vitro in MRF (3-log decrease) and LRF (5-log decrease), respectively; the opposite was observed in vivo with a 3-log decrease for E. coli NalR (#5735) in LRF and a 2-log decrease in MRF for SS17 (Figs 1 and 2, S2 and S3 Tables). These observations clearly indicate the need for reliable methods for studying bacteria in vivo in the animal as the environmental cues seen within cannot be fully reproduced in vitro and bacterial strains can respond differently to such cues. Lack of ease in conducting rumen site-specific studies has resulted in limited in vivo in animals compared to in vitro studies [27,31–35] which can be circumvented using our novel method of analysis.
Diet, growth conditions and challenge with O157 strains has been shown to influence ruminal pH and VFA concentrations. For instance, an average ruminal pH of 5.98 over 29 days post-O157 strain 86–24 challenge was observed in vivo in cattle fed a grain diet comprising of 90% corn and 10% alfalfa [32], while an in vitro study displayed a decrease of 1.4 pH units from 7.25 prior to incubation to 5.85 twenty-four hours post-O157 strain EDL933 inoculation in rumen fluid from cattle fed a mixed diet of 80% hay and 20% concentrate [70]. Alongside pH, ruminal VFA concentration was analyzed in the latter study. Prior to incubation, VFA concentration was measured at 115.78 mM, decreasing at hour 6 to 90.8 mM, and increasing to 129.53 mM after 24 hours, with a high acetate concentration of 84.51 mM [71]. Our previous in vitro study corroborates this result, as acetate accounted for the majority of VFAs measured in unfiltered, filtered, and depleted rumen fluid, prepared from rumen fluid from animals fed the high fiber maintenance diet, at 62–72% of VFAs measured, with propionate and butyrate concentrations ranging from 13–18% and 6–13%, respectively [28]. In this study, irrespective of the growth condition, VFA concentrations were higher in LRF compared to MRF, derived from animals fed a high protein versus high fiber diet, respectively. This high VFA concentration may have stemmed from increased fermentation in LRF which in turn contributed to its acidic pH; percent concentration of acetate was slightly higher in MRF while butyrate was higher in LRF (Table 1A and 1B in Fig 2 and Table 2A and 2B in Fig 3).
O157 strains can express specific genes/proteins, to prolong survival in certain environmental conditions as influenced by VFA concentrations and pH in the rumen fluid. In our study, we used riBAQ analysis of the proteomics data to determine differences in proteins expressed between the contrasting rumen fluids. This was done since the proteomics data was generated in two different LC-MS/MS runs with each type of rumen fluid represented only on one run. The riBAQ analysis of the proteomics data, based on relative abundance, was useful in identifying proteins uniquely expressed in LRF (11) versus MRF (179), although caution is required when comparing protein relative abundances between different rumen fluids as compositional effects could led to some erroneous conclusions (more proteins under the limit of detection in LRF could have led to an artificial inflation of the proportions of the detected proteins). In silico analysis of the iTRAQ dataset using reporter intensities attributed differential expression of proteins mainly to culture conditions rather than individual bacterial strains being tested, as evidenced in the MDS plot (Fig 5), irrespective of the rumen fluid. The reference-based analysis of the iTRAQ dataset further identified proteins with increased expression in vitro and in vivo relative to the 86–24 in-vitro proteome, in both LRF and MRF. Under in vitro growth condition expression of more proteins was increased in LRF (19 proteins) than MRF (12 proteins), compared to in vivo growth conditions where greater number of proteins were up-regulated in both LRF (52 proteins) and MRF (54 proteins). Interestingly, SS17 seemed to contribute most towards some of the differentially expressed O157 proteins in vitro in LRF, and in vivo in MRF. While our multivariate statistical tests did not detect robust strain differences, SS17 exhibited some indications of strain level differences with altered growth characteristics and indications of altered protein expression. This distinctive feature of the SS-17 proteome was identified by the MDS, and dispersion plots based on reference-free analysis of the iTRAQ data as well. SS17 demonstrated greater growth suppression in vitro in LRF, and in vivo in MRF, and hence the increased protein expression may correlate with a possible stress response [69].
Transcriptome analyses of O157 during growth in the rumen have revealed several pathways, genes, and proteins that are differentially expressed. Hughes et al. found that SdiA-AHL (acyl-homoserine lactones) sensing alters 49 EHEC genes while in the rumen, including those involved in acid resistance (gad) and the locus of enterocyte effacement (LEE) [32], with similar results observed by Nguyen et al. when using AHL synthase gene yenI [72]. Rumen AHLs sensed by SdiA directly repress LEE gene transcription, but indirectly induces gad acid resistance genes to increase EHEC’s chance of survival in acidic environments like the rumen [32]. A 2018 transcriptome study conducted by Segura et al. found that O157 strain EDL933 differentially expresses 617 genes in the bovine rumen in comparison to minimal media [70]. Upregulation of genes within carbohydrate transport (xylA/F, fucA/O, rbsACD, cmtB), amino acid transport (dctA, tnaA, tdcD/E, asnA), energy production and conversion (appBC, dmsABC, napAH, narGHIJ, nrfEFG), and aerobic respiration (fdoHI, appBC glpD) were observed. Ethanolamine is an important nitrogen source and growth aid for bacteria in the digesta and may also act as a signal for gene regulation while in the rumen, as observed by induction of assimilation and transport genes (eut) in the rumen content [70]. Our previous work also showed expression of several O157 proteins involved in cell structure (OmpA/W, BamA/D, LppC, and AsmA), metabolism (LysU, HyC, TnaA, and LamB), chemotaxis (GgbP, Trg), adherence (AidA-1, EspP, and CsgB/G), and stress responses (KatG, HchA, MoaB, and the Arg acid resistance system) but none associated with O157 virulence in vitro, in the bovine rumen fluid [28]. This led to the conclusion that rumen-specific protein expression allows O157 to adapt and survive within the rumen in order to move throughout and colonize the bovine gastrointestinal tract.
Similarly, in our present study, we observed numerous O157 proteins belonging to nine general function categories were upregulated in the rumen fluid from cows fed the maintenance (MRF) or lactation (LRF) diets. General protein functions included: metabolism (lipid and carbohydrate, amino acid and protein, energy), regulation, translation and transcription, transport, cell membrane synthesis, environmental adaptation, and respiration (Table 1). Proteins expressed in vivo irrespective of diet type with a role in metabolism (YhcB, OmpT), protein translation (RpsF), transport (FhuA, ModA), or environmental adaptation (FliC, SodA) were also identified in other proteomic and transcriptomic rumen fluid studies, described hereafter. An in vitro model by Segura et al. identified transport proteins ModA and FeoA [70] expressed by O157 strain EDL933, in rumen fluid extracted at slaughter from bulls fed a mixed diet of 80% hay and 20% concentrate. In addition to the ModA and FeoA transport proteins, proteins involved in metabolism (MalK, GltA, PckA, SucC/D), or respiration (SdhB) were also upregulated, supporting our findings. Lastly, SpeB, RpsF, and FliC were identified in our previous proteomics work as well, using cattle on a low protein maintenance diet consisting of 25% fiber and 10% protein [28]. Other proteins from the current study were also identified in our earlier work, spanning each general function category- except for regulation- including DegQ, FbaB, OmpC, FtnA, RpoA, Tig, RbsB, BamA/D, TerD, and FdoH, in different preparations of the rumen fluid [28].
Three O157 acid-resistance systems have been identified thus far: glucose-repressed or oxidative, glutamate-dependent, and arginine-dependent [16,73–75]. The glucose-repressed system is activated during stationary phase via RpoS and cAMP receptor protein and repressed with the addition of glucose. The glutamate and arginine systems utilize either GadA/B or AdiA decarboxylases to convert glutamate or arginine to gamma-aminobutyric acid (GABA) or agmatine, respectively, consuming excess protons in the process. Agmatine is either exported via the arginine-agmatine antiporter (aidC) or converted (speB) to putresceine as part of the polyamine biosynthetic pathway and putresceine transported out of the cell via the ABC transporter encoded by potD [73,74]. GABA is transported out of the cell via the antiporter encoded by gadC. The glutamate system seems to provide the most protection against acidic environments (below pH 6) [16,73,74]. In fact, in cattle fed a high grain diet of 90% corn and 10% alfalfa and challenged with O157 strain 86–24, an upregulation of gad transcripts was observed in rumen fluid post-challenge, activated by acyl-homoserine lactones (AHLs) and SdiA [27,32]. In a similar study, gad gene expression was induced in vitro by O157 strain 86–24 engineered to express AHL synthase, yenI, to mimic constant AHL exposure [72]. Other acid resistance systems are activated between pH 5.5–6.9 and used more often than the glutamate system to counter the deleterious effects of protons [73,74]. In our study, GadA/B was the only acid-resistance related enzyme expressed in the acidic LRF with increased expression in vivo (Table 1). In MRF on the other hand, as observed in our previous study, minimal expression of GadA/B was observed in vitro only while SpeB and PotD expression was enhanced in all 3 O157 strains tested in vivo (Table 1).
None of the proteins expressed in LRF and MRF were associated with O157 virulence, irrespective of the in vitro and in vivo growth conditions or the O157 strain tested. Using the reference-based iTRAQ analysis #1, 66 and 71 proteins were determined to have increased expression, in in vitro and in vivo conditions combined, in MRF and LRF, respectively (Table 1, Fig 4). Of these, 17/66 and 22/71 proteins were identified as being up-regulated in either in vitro or in vivo conditions by reference free iTRAQ analysis #2 as well (Table 1 shown in bold) and are of interest along with the acid resistance mechanisms for further evaluation in therapeutic modalities. In conclusion, this study highlights a novel method for introducing and retrieving bacteria from the rumen of animals for in vivo evaluations and the application of different proteomics data analysis methods to derive useful protein expression data. The reusable animal model is very cost-effective as multiple studies with access to the internal rumen environment that cannot be fully reproduced in vitro can be conducted, with minimal number of animals. The observations made reinforced that the maintenance and lactation diets have differing influences on the rumen environment which in turn affects O157 survival and protein expression. The O157 proteins uniquely expressed in vivo in the rumen vary from those expressed in vitro in the rumen fluid but taken together further support our previous conclusions that O157 induces expression of proteins from a broad range of functions to adapt and survive the diverse physiological and microbiome conditions within the bovine rumen. Given that the expression of these proteins is mainly influenced by the animals’ diet and not O157 strain type, preharvest control modalities could be developed with these proteins to target O157 across strains in cattle at all stages of husbandry. Additional experiments are being planned to study the influence of the rumen environment, under similar dietary conditions, on O157, non-O157 and commensal E. coli gene expression at various time-points post-exposure, to obtain a dynamic and comparative profile of the adaptive responses of these bacteria.
Supporting information
S1 Fig. Product labels for ‘Steakmaker’ and ‘Lactation Premix’ used in the animal diets.
https://doi.org/10.1371/journal.pone.0268645.s001
(TIF)
S1 Table. Recovery of E. coli NalR (#5735) in dialysis PVDF tubings versus cartridges exposed to MRF in in vitro and in vivo pilot experiments.
https://doi.org/10.1371/journal.pone.0268645.s005
(DOCX)
S2 Table. Recovery of bacteria from cartridges exposed to MRF in in vitro and in vivo.
https://doi.org/10.1371/journal.pone.0268645.s006
(DOCX)
S3 Table. Recovery of bacteria from cartridges exposed to LRF in in vitro and in vivo.
https://doi.org/10.1371/journal.pone.0268645.s007
(DOCX)
S4 Table.
A: iBAQ and riBAQ values of proteins expressed by O157 strains in MRF and LRF. B: Proteins uniquely expressed by O157 strains in MRF and LRF, irrespective of the in vitro/in vivo growth conditions, identified using iBAQ. C: Proteins more highly expressed by O157 strains in MRF; based on riBAQ values of L2FC < -1 only. D: Proteins more highly expressed by O157 strains in LRF; based on riBAQ values of L2FC > 1 only.
https://doi.org/10.1371/journal.pone.0268645.s008
(ZIP)
S5 Table.
A: GO terms of O157 proteins enriched in MRF; based on riBAQ values of L2FC < -1 and Fisher p values < 0.1. B: GO terms of O157 proteins enriched in LRF; based on riBAQ values of L2FC >1 and Fisher p values < 0.1.
https://doi.org/10.1371/journal.pone.0268645.s009
(ZIP)
S6 Table.
A: iTRAQ quantitative samples view report of proteins differentially expressed by O157 strains in MRF, in comparison to the 86–24 in vitro proteome (reference-based analysis). B: iTRAQ quantitative samples view report of proteins differentially expressed by O157 strains in LRF, in comparison to the 86–24 in vitro proteome (reference-based analysis).
https://doi.org/10.1371/journal.pone.0268645.s010
(ZIP)
S7 Table.
A. O157 proteins significantly different between the in vitro and in vivo culture conditions in MRF determined by using reference-free analysis of iTRAQ data. B. O157 proteins significantly different between the in vitro and in vivo culture conditions in LRF determined by using reference-free analysis of iTRAQ data.
https://doi.org/10.1371/journal.pone.0268645.s011
(ZIP)
S8 Table.
A. GO terms of O157 proteins enriched in the in vitro growth condition, in MRF; based on reference-free, Scaffold normalized iTRAQ intensities resulting in proteins with L2FC values < -0.5 and Fisher p values < 0.1 only. B. GO terms of O157 proteins enriched in the in vivo growth condition, in MRF; based on reference-free, Scaffold normalized iTRAQ intensities resulting in proteins with L2FC values > 0.5 and Fisher p values < 0.1. C. GO terms of O157 proteins enriched in the in vitro growth condition, in LRF; based on reference-free, Scaffold normalized iTRAQ intensities resulting in proteins with L2FC values < -0.5 and Fisher p values < 0.1 only. D. GO terms of O157 proteins enriched in the in vivo growth condition, in LRF; based on reference-free, Scaffold normalized iTRAQ intensities resulting in proteins with L2FC values > 0.5 and Fisher p values < 0.1.
https://doi.org/10.1371/journal.pone.0268645.s012
(ZIP)
S9 Table.
A. PERMANOVA results detailing strain and condition effects within MRF. Based on Bray-Curtis dissimilarities generated from reference-free, Scaffold normalized iTRAQ intensities. B. PERMANOVA results detailing strain and condition effects within LRF. Based on Bray-Curtis dissimilarities generated from reference-free, Scaffold normalized iTRAQ intensities.
https://doi.org/10.1371/journal.pone.0268645.s013
(RAR)
Acknowledgments
Expert rumen fistulation surgeries conducted by NADC veterinarians, Dr. Rebecca Cox and Dr. Jean Hall, is gratefully acknowledged. Skillful assistance with general husbandry, post-surgery monitoring and rumen sampling was provided by the NADC animal caretakers. Bryan Wheeler and Matt Inbody provided excellent technical assistance with the collection and filtration of rumen fluid, and placement/harvesting of ruminal cartridges. Assistance provided by technicians Sam Humphrey, Stephanie Jones and interns, Aparna Ajjarapu, Axel M. A. Morales, with the volatile fatty acid analysis at various stages of the study is acknowledged. iTRAQ proteomics was done at the Proteomics Division, ICBR, University of Florida, Gainesville, FL. We gratefully thank Dr. John D. Lippolis at the Ruminant Disease and Immunology Research Unit, NADC for providing expert guidance in the use of iBAQ for proteomics data analysis and the insightful review of our manuscript.
This research was supported by an appointment to the Agricultural Research Service (ARS) Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the U.S. Department of Energy.
(DOE) and the U.S. Department of Agriculture (USDA). ORISE is managed by ORAU under DOE contract number DE- SC0014664. All opinions expressed in this paper are the authors’ and do not necessarily reflect the policies and views of USDA, ARS, DOE, or ORAU/ORISE.
Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.
References
- 1. Tack DM RL, Griffin PM, et al. Preliminary Incidence and Trends of Infections with Pathogens Transmitted Commonly Through Food—Foodborne Diseases Active Surveillance Network, 10 U.S. Sites, 2016–2019. MMWR Morb Mortal Wkly Rep 2020;69:509–14. pmid:32352955
- 2. Scallan E, Griffin PM, Angulo FJ, Tauxe RV, Hoekstra RM. Foodborne Illness Acquired in the United States—Unspecified Agents. Emerging Infectious Diseases. 2011;17(1):16–22. pmid:21192849
- 3. Majowicz SE, Scallan E, Jones-Bitton A, Sargeant JM, Stapleton J, Angulo FJ, et al. Global incidence of human Shiga toxin-producing Escherichia coli infections and deaths: a systematic review and knowledge synthesis. Foodborne Pathog Dis. 2014;11(6):447–55. pmid:24750096
- 4.
Kaper JB, O’Brien AD. Escherichia coli O157:H7 and Other Shiga Toxin-Producing E. coli Strains. Washington, DC: ASM Press; 1998.
- 5. Paton JC, Paton AW. Pathogenesis and Diagnosis of Shiga Toxin-Producing Escherichia coli Infections. Clin Microbiol Rev. 1998;11:450–79. pmid:9665978
- 6. Riley LW, Memis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, et al. Hemorrhagic colitis associated with a rare Escherichia coli serotype. New Eng J Med. 1983;308(12):681–5. pmid:6338386
- 7. Mead PS, Griffin PM. Escherichia coli O157:H7. Lancet. 1998;352:1207–12. pmid:9777854
- 8. Batz MB, Hoffmann S, Morris JG Jr. Ranking the disease burden of 14 pathogens in food sources in the United States using attribution data from outbreak investigations and expert elicitation. J Food Prot. 2012;75(7):1278–91. pmid:22980012
- 9. Bell BP, Goldoft M, Griffin PM, Davis MA, Gordon DC, Tarr PI, et al. A multistate outbreak of Escherichia coli O157:H7-associated bloody diarrhea and hemolytic uremic syndrome from hamburgers. JAMA. 1994;272(17):1349–53. pmid:7933395
- 10. Mayer CL, Leibowitz CS, Kurosawa S, Stearns-Kurosawa DJ. Shiga toxins and the pathophysiology of hemolytic uremic syndrome in humans and animals. Toxins (Basel). 2012;4(11):1261–87. pmid:23202315
- 11. Rivas M, Chinen I, Miliwebsky E, Masana M. Risk Factors for Shiga Toxin-Producing Escherichia coli-Associated Human Diseases. Microbiol Spectr. 2014;2(5). pmid:26104362
- 12. Kaper JB, Nataro JP, Mobley HL. Pathogenic Escherichia coli. Nat Rev Microbiol. 2004;2(2):123–40. pmid:15040260
- 13. Ray PE, Liu X. Pathogenesis of Shiga toxin-induced hemolytic uremic syndrome. Pediatr Nephrol. 2001;16:823–39. pmid:11605791
- 14. Persad AK, LeJeune JT. Animal Reservoirs of Shiga Toxin-Producing Escherichia coli. Microbiol Spectr. 2014;2(4):EHEC-0027-2014. pmid:26104194
- 15. Kudva IT, Hovde CJ. Escherichia coli O157:H7 in Ruminants-A Review. Pandalai SG, editor. Trivandrum India: Research Signpost; 1998. 665–92 p.
- 16. Nguyen Y, Sperandio V. Enterohemorrhagic E. coli (EHEC) pathogenesis. Front Cell Infect Microbiol. 2012;2:90. pmid:22919681
- 17. Naylor SW, Low JC, Besser TE, Mahajan A, Gunn GJ, Pearce MC, et al. Lymphoid follicle-dense mucosa at the terminal rectum is the principal site of colonization of enterohemorrhagic Escherichia coli O157:H7 in the bovine host. Infect Immun. 2003;71(3):1505–12. pmid:12595469
- 18. Keen JE, Laegreid WW, Chitko-McKown CG, Durso LM, Bono JL. Distribution of Shiga-Toxigenic Escherichia coli O157 in the Gastrointestinal Tract of Naturally O157-Shedding Cattle at Necropsy. Appl Environ Microbiol. 2010;76(15):5278–81. pmid:20543036
- 19. Wells JE, Berry ED, Kim M, Bono JL, Oliver WT, Kalchayanand N, et al. Determination of gastrointestinal tract colonization sites from feedlot cattle transiently shedding or super-shedding Escherichia coli O157:H7 at harvest. J Appl Microbiol. 2020;129(5):1419–26. pmid:32350973
- 20. Fegan N, Vanderlinde P, Higgs G, Desmarchelier P. The prevalence and concentration of Escherichia coli O157 in faeces of cattle from different production systems at slaughter. J Appl Microbiol. 2004;97(2):362–70. pmid:15239703
- 21. Widiasih DA, Ido N, Omoe K, Sugii S, Shinagawa K. Duration and magnitude of faecal shedding of Shiga toxin-producing Escherichia coli from naturally infected cattle. Epidemiol Infect. 2004;132(1):67–75. pmid:14979592
- 22. Lim JY, Li J, Sheng H, Besser TE, Potter K, Hovde CJ. Escherichia coli O157:H7 colonization at the rectoanal junction of long-duration culture-positive cattle. Appl Environ Microbiol. 2007;73(4):1380–2. pmid:17189448
- 23. Cobbold RN, Hancock DD, Rice DH, Berg J, Stilborn R, Hovde CJ, et al. Rectoanal junction colonization of feedlot cattle by Escherichia coli O157:H7 and its association with supershedders and excretion dynamics. Appl Environ Microbiol. 2007;73(5):1563–8. pmid:17220263
- 24. Omisakin F, MacRae M, Ogden ID, Strachan NJ. Concentration and prevalence of Escherichia coli O157 in cattle feces at slaughter. Appl Environ Microbiol. 2003;69(5):2444–7. pmid:12732509
- 25. Matthews L, Low JC, Gally DL, Pearce MC, Mellor DJ, Heesterbeek JA, et al. Heterogeneous shedding of Escherichia coli O157 in cattle and its implications for control. Proc Natl Acad Sci U S A. 2006;103(3):547–52. pmid:16407143
- 26. Keen JE, Laegreid WW, Chitko-McKown CG, Durso LM, Bono JL. Distribution of Shiga-toxigenic Escherichia coli O157 in the gastrointestinal tract of naturally O157-shedding cattle at necropsy. Appl Environ Microbiol. 2010;76(15):5278–81. pmid:20543036
- 27. Sperandio V. SdiA sensing of acyl-homoserine lactones by enterohemorrhagic E. coli (EHEC) serotype O157:H7 in the bovine rumen. Gut Microbes. 2010;1(6):432–5. pmid:21468228
- 28. Kudva IT, Stanton TB, Lippolis JD. The Escherichia coli O157:H7 bovine rumen fluid proteome reflects adaptive bacterial responses. BMC Microbiology. 2014;14(48). pmid:24559513
- 29. Lima FS, Oikonomou G, Lima SF, Bicalho ML, Ganda EK, Filho JC, et al. Prepartum and postpartum rumen fluid microbiomes: characterization and correlation with production traits in dairy cows. Appl Environ Microbiol. 2015;81(4):1327–37. pmid:25501481
- 30. Auffret MD, Dewhurst RJ, Duthie CA, Rooke JA, John Wallace R, Freeman TC, et al. The rumen microbiome as a reservoir of antimicrobial resistance and pathogenicity genes is directly affected by diet in beef cattle. Microbiome. 2017;5(1):159. pmid:29228991
- 31. Zhou M, Hünerberg M, Chen Y, Reuter T, McAllister TA, Evans F, et al. Air-Dried Brown Seaweed, Ascophyllum nodosum, Alters the Rumen Microbiome in a Manner That Changes Rumen Fermentation Profiles and Lowers the Prevalence of Foodborne Pathogens. mSphere 2018;3(1): e00017–18; pmid:29404417
- 32. Hughes DT, Terekhova DA, Liou L, Hovde CJ, Sahl JW, Patankar AV, et al. Chemical sensing in mammalian host-bacterial commensal associations. Proc Natl Acad Sci U S A. 2010;107(21):9831–6. pmid:20457895
- 33. Sheng H, Nguyen YN, Hovde CJ, Sperandio V. SdiA aids enterohemorrhagic Escherichia coli carriage by cattle fed a forage or grain diet. Infect Immun. 2013;81(9):3472–8. pmid:23836826
- 34. Bertin Y, Habouzit C, Duniere L, Laurier M, Durand A, Duchez D, et al. Lactobacillus reuteri suppresses E. coli O157:H7 in bovine ruminal fluid: Toward a pre-slaughter strategy to improve food safety? PLoS One. 2017;12(11):e0187229. pmid:29091926
- 35. Free AL, Duoss HA, Bergeron LV, Shields-Menard SA, Ward E, Callaway TR, et al. Survival of O157:H7 and non-O157 serogroups of Escherichia coli in bovine rumen fluid and bile salts. Foodborne Pathog Dis. 2012;9(11):1010–4. pmid:22957973
- 36. Tarr PI, Neill MA, Clausen CR, Newland JW, Neill RJ, Moseley SL. Genotypic variation in pathogenic Escherichia coli O157:H7 isolated from patients in Washington, 1984–1987. J Infect Dis. 1989;159(2):344–7. pmid:2644374
- 37. O’Brien AD, Melton AR, Schmitt CK, McKee ML, Batts ML, Griffin DE. Profile of Escherichia coli O157:H7 pathogen responsible for hamburger-borne outbreak of hemorrhagic colitis and hemolytic uremic syndrome in Washington. J Clin Microbiol. 1993;31(10):2799–801. pmid:8253989
- 38. Cote R, Katani R, Moreau MR, Kudva IT, Arthur TM, DebRoy C, et al. Comparative analysis of super-shedder strains of Escherichia coli O157:H7 reveals distinctive genomic features and a strongly aggregative adherent phenotype on bovine rectoanal junction squamous epithelial cells. PLoS One. 2015;10(2):e0116743. pmid:25664460
- 39.
Moran J. How the Rumen Works. Tropical dairy farming: feeding management for small holder dairy farmers in the humid tropics: Landlinks Press; 2005. p. 312.
- 40. Erickson PS, Kalscheur KF. Nutrition and feeding of dairy cattle. Animal Agriculture2020. p. 157–80.
- 41. Kudva IT, Hatfield PG, Hovde CJ. Effect of diet on the shedding of Escherichia coli O157:H7 in a sheep model. Appl Environ Microbiol. 1995;61(4):1363–70. pmid:7747956
- 42. Kudva IT, Hatfield PG, Hovde CJ. Characterization of Escherichia coli O157:H7 and other Shiga toxin-producing E. coli serotypes isolated from sheep. J Clin Microbiol. 1997;35(4):892–9. pmid:9157149
- 43. Kudva IT, Hunt CW, Williams CJ, Nance UM, Hovde CJ. Evaluation of dietary influences on Escherichia coli O157:H7 shedding by sheep. Appl Environ Microbiol. 1997;63(10):3878–86. pmid:9327551
- 44. Van Gylswyk NO, Wejdemar K, Kulander K. Comparative Growth Rates of Various Rumen Bacteria in Clarified Rumen Fluid from Cows and Sheep Fed Different Diets. App Environ Microbiol. 1992;58:99–105.
- 45. Nikkhah A. Bioscience of ruminant intake evolution: feeding time models. Advances in Bioscience and Biotechnology. 2011;02(04):271–4.
- 46. Salanitro JP, Muirhead PA. Quantitative Method for the Gas Chromatographic Analysis of Short-Chain Monocarboxylic and Dicarboxylic Acids in Fermentation Media. App Microbiol. 1975;29:374–81. pmid:1167776
- 47. Lambert MA, Moss CW. Preparation and analysis of the butyl esters of short-chain volatile and non-volatile fatty acids. J Chromatograph 1972;74:335–8.
- 48.
C.M.B. M. Anatomy and Physiology of the Rumen. In: Millen D., De Beni Arrigoni M., LP R., editors. Rumenology: Springer, Cham.; 2016.
- 49. Anderson KL, Whitlock JE, Harwood VJ. Persistence and differential survival of fecal indicator bacteria in subtropical waters and sediments. Appl Environ Microbiol. 2005;71(6):3041–8. pmid:15933000
- 50. Welch JG. Rumination, particle size and passage from the rumen. J Animal Sci. 1982;54:885–94.
- 51. Kudva IT, Krastins B, Sheng H, Griffin RW, Sarracino DA, Tarr PI, et al. Proteomics-based expression library screening (PELS): a novel method for rapidly defining microbial immunoproteomes. Mol Cell Proteomics. 2006;5(8):1514–9. pmid:16737953
- 52. Lippolis JD, Bayles DO, Reinhardt TA. Proteomic changes in Escherichia coli when grown in fresh milk versus laboratory media. J Prot Res 2009;8:149–58. pmid:19053534
- 53. Tyanova S, Temu T, Cox J. The MaxQuant computational platform for mass spectrometry-based shotgun proteomics,. Nat Protocols. 2016;11:2301–19. pmid:27809316
- 54. Fu L, Niu B, Zhu Z, Wu S, Li W. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics. 2012;28(23):3150–2. pmid:23060610
- 55. UniProt C. UniProt: the universal protein knowledgebase in 2021. Nucleic Acids Res. 2021;49(D1):D480–D9. pmid:33237286
- 56. Shin JB, Krey JF, Hassan A, Metlagel Z, Tauscher AN, Pagana JM, et al. Molecular architecture of the chick vestibular hair bundle. Nat Neurosci. 2013;16(3):365–74. pmid:23334578
- 57. Keller A, Nesvizhskii AI, Kolker E, Aebersold R. Empirical Statistical Model To Estimate the Accuracy of Peptide Identifications Made by MS/MS and Database Search. Anal Chem 2002;74:5383–92. pmid:12403597
- 58. Nesvizhskii AI, Keller A, Kolker E, Aebersold R. A Statistical Model for Identifying Proteins by Tandem Mass Spectrometry. Anal Chem 2003;75:4646–58. pmid:14632076
- 59. Shadforth IP, Dunkley TP, Lilley KS, Bessant C. i-Tracker: for quantitative proteomics using iTRAQ. BMC Genomics. 2005;6:145. pmid:16242023
- 60. Oberg AL, Mahoney DW, Eckel-Passow JE, Malone CJ, Wolfinger RD, Hill EG, et al. Statistical analysis of relative labeled mass spectrometry data from complex samples using ANOVA. J Proteome Res. 2008;7(1):225–33. pmid:18173221
- 61. Yu NY, Wagner JR, Laird MR, Melli G, Rey S, Lo R, et al. PSORTb 3.0: improved protein subcellular localization prediction with refined localization subcategories and predictive capabilities for all prokaryotes. Bioinformatics. 2010;26(13):1608–15. pmid:20472543
- 62. Jones P, Binns D, Chang HY, Fraser M, Li W, McAnulla C, et al. InterProScan 5: genome-scale protein function classification. Bioinformatics. 2014;30(9):1236–40. pmid:24451626
- 63.
R, Core, Team. R: A language and environment for statistical computing. R Foundation for statistical Computing, Vienna, Austria URL https://wwwR-projectorg/. 2020.
- 64. Alexa A, Rahnenfuhrer J. topGO: Enrichment Analysis for Gene Ontology. R package version 2.42.0. 2020.
- 65. Ashburner M, Ball CA, Blake JA, Botstein D, al. e. Gene Ontology: tool for the unification of biology. Nat Genet. 2000;25:25–9. pmid:10802651
- 66. Consortium. TGO. The Gene Ontology resource: enriching a GOld mine. Nucleic Acids Res. 2021;49:D325–D34. pmid:33290552
- 67. Oksanen J, Guillaume Blanchet F, Friendly M, Kindt R, Legendre P, McGlinn D, et al. Vegan: Community Ecology Package. R package version 25–7 https://CRANR-projectorg/package=vegan. 2020.
- 68. Krey JF, Wilmarth PA, Shin JB, Klimek J, Sherman NE, Jeffery ED, et al. Accurate label-free protein quantitation with high- and low-resolution mass spectrometers. J Proteome Res. 2014;13(2):1034–44. pmid:24295401
- 69. Bartholomaus A, Fedyunin I, Feist P, Sin C, Zhang G, Valleriani A, et al. Bacteria differently regulate mRNA abundance to specifically respond to various stresses. Philos Trans A Math Phys Eng Sci. 2016;374(2063). pmid:26857681
- 70. Segura A, Bertoni M, Auffret P, Klopp C, Bouchez O, Genthon C, et al. Transcriptomic analysis reveals specific metabolic pathways of enterohemorrhagic Escherichia coli O157:H7 in bovine digestive contents. BMC Genomics. 2018;19(1):766. pmid:30352567
- 71. Bertin Y, Segura A, Jubelin G, Duniere L, Durand A, Forano E. Aspartate metabolism is involved in the maintenance of enterohaemorrhagic Escherichia coli O157:H7 in bovine intestinal content. Environ Microbiol. 2018;20(12):4473–85. pmid:30109758
- 72. Nguyen YN, Sheng H, Dakarapu R, Falck JR, Hovde CJ, Sperandio V. The acyl-homoserine lactone synthase YenI from Yersinia enterocolitica modulates virulence gene expression in enterohemorrhagic Escherichia coli O157:H7. Infect Immun. 2013;81(11):4192–9. pmid:23980115
- 73. Tsai MF, McCarthy P, Miller C. Substrate selectivity in glutamate-dependent acid resistance in enteric bacteria. Proc Natl Acad Sci U S A. 2013;110(15):5898–902. pmid:23530225
- 74. Tsai MF, Miller C. Substrate selectivity in arginine-dependent acid resistance in enteric bacteria. Proc Natl Acad Sci U S A. 2013;110(15):5893–7. pmid:23530220
- 75. Foster JW. Escherichia coli acid resistance: tales of an amateur acidophile. Nat Rev Microbiol. 2004;2(11):898–907. pmid:15494746