A Comprehensive Evaluation of Colonic Mucosal Isolates of Sutterella wadsworthensis from Inflammatory Bowel Disease

Inflammatory bowel disease (IBD) arises in genetically susceptible individuals as a result of an unidentified environmental trigger, possibly a hitherto unknown bacterial pathogen. Twenty-six clinical isolates of Sutterella wadsworthensis were obtained from 134 adults and 61 pediatric patients undergoing colonoscopy, of whom 69 and 29 respectively had IBD. S. wadsworthensis was initially more frequently isolated from IBD subjects, hence this comprehensive study was undertaken to elucidate its role in IBD. Utilizing these samples, a newly designed PCR was developed, to study the prevalence of this bacterium in adult patients with ulcerative colitis (UC). Sutterella wadsworthensis was detected in 83.8% of adult patients with UC as opposed to 86.1% of control subjects (p = 0.64). Selected strains from IBD cases and controls were studied to elicit morphological, proteomic, genotypic and pathogenic differences. This study reports Scanning Electron Microscopy (SEM) appearances and characteristic MALDI-TOF MS protein profiles of S. wadsworthensis for the very first time. SEM showed that the bacterium is pleomorphic, existing in predominantly two morphological forms, long rods and coccobacilli. No differences were noted in the MALDI-TOF mass spectrometry proteomic analysis. There was no distinct clustering of strains identified from cases and controls on sequence analysis. Cytokine response after monocyte challenge with strains from patients with IBD and controls did not yield any significant differences. Our studies indicate that S. wadsworthensis is unlikely to play a role in the pathogenesis of IBD. Strains from cases of IBD could not be distinguished from those identified from controls.


Introduction
Inflammatory bowel disease (IBD) is an idiopathic inflammatory disorder that is comprised of two major phenotypes, Crohn's disease (CD) and ulcerative colitis (UC). The understanding of its aetiopathogenesis has taken rapid strides in the last decade, with current investigations focusing heavily on aberrations in hostmicrobe interactions at the luminal intestinal surface. Genetic defects in pathogen recognition and primary handling of microbes by the innate immune system compounded with distinct changes in the luminal microbiome or dysbiosis form the current backbone of this pathogenic hypothesis [1,2]. Despite this, researchers in the field have been striving to identify and delineate a solitary microorganism that can explain the initiation and perpetuation of this chronic inflammatory process.
In this regard, anaerobic and microaerophilic bacteria residing in the intestinal lumen have often been the neglected species, primarily on account of the intrinsic difficulty in culturing and isolating these organisms by using traditional microbiological techniques. Molecular studies have demonstrated that a substantial proportion of bacterial species (up to 30-40% of dominant species) in patients with active IBD belong to phylogenetic groups that are unusual when compared to healthy subjects [3,4]. With this premise in mind, our laboratory has focused on enhanced and improved bacteriological conditions for the optimum growth of microaerophilic bacteria from colonic biopsy samples [5,6]. In our pilot studies we noted the unusual preponderance of the rare microaerophilic Gram negative bacterium Sutterella wadsworthensis from cultures of biopsy samples from patients with IBD. This unusual organism has been encountered before by Mangin et al. who used 16S rRNA gene sequencing to create molecular inventories of the dominant fecal bacteria in four CD patients and four controls. They found that bacterial species which were not commonly dominant in healthy individuals were over-represented in CD. One of these species included S. wadsworthensis, which belonged to the dominant microbiota of one of the four CD patients [7].
These bacteria were first reported when performing biochemical characterization and susceptibility testing of Campylobacter gracilis clinical isolates from patients with diverse infections of the GI tract [8]. These organisms could be differentiated from C. gracilis mainly by their bile resistance and cell wall fatty acid patterns. 16S rRNA gene sequencing confirmed that these unusual organisms were not related phylogenetically to any of the Campylobacters, including C. gracilis, with the closest taxa belonging to unrelated aerobes. This was ratified by a subsequent report which demonstrated that the majority of S. wadsworthensis strains were isolated from GI infections, only occasionally being isolated from non-abdominal specimens, and were more likely to be involved in serious infections than C. gracilis [9]. The supposition was that S. wadsworthensis was a putative human pathogen. Three other species, Sutterella stercoricanis, Sutterella morbirenis and Sutterella parvirubra belonging to the same genera have subsequently been identified from canine and human feces [10,11].
The role of this group of bacteria has not been clearly elucidated in the aetiopathogenesis of IBD. This study has for the first time outlined the role of S. wadsworthensis in patients with IBD and performed a comprehensive phenotypic, genotypic, proteomic and pathogenetic characterization of this bacterial species, which will serve as a useful benchmark for future studies.

Study subjects, specimen collection and processing
Adult patients were recruited from the Department of Gastroenterology at the Aberdeen Royal Infirmary. These subjects were recruited for a previous study looking at the role of enterohepatic Helicobacter in UC [5]. A total of sixty-nine patients with a diagnosis of UC made on the basis of histology of colonoscopic biopsies were recruited and assessed. Sixty-five healthy controls were contacted prior to their index colonoscopy as part of the bowel cancer screening programme and recruited for the study if they had documented absence of both macroscopic and microscopic inflammation. Children were recruited from the Departments of Paediatric Gastroenterology, Hepatology and Nutrition at the Royal Aberdeen Children's Hospital and the Royal Hospital for Sick Children (Yorkhill), Glasgow as part of an ongoing study to investigate the role of microaerophilic colonic microbiota in de-novo paediatric IBD (Bacteria in Inflammatory bowel disease in Scottish Children Undergoing Investigation before Treatment: BISCUIT study). Twenty-nine paediatric patients with newly-presenting, treatment naïve IBD and thirty-two paediatric controls undergoing routine colonoscopy were included in this present study [12]. The extent and severity of disease was assigned using the Montreal criteria [13]. Subjects were excluded if they received antibiotics within three months prior to recruitment. Mucosal colonic biopsies were obtained during the colonoscopy procedures. One to two biopsies were used for culture work and the rest were then transferred to a 280uC freezer for storage pending DNA extraction and analysis.

Ethics
Ethical approval for the study was granted by the North of Scotland Research Ethics Service, UK (reference numbers 04/ S0802/8 and 09/S0802/24). Written informed consent was obtained from all adult subjects and from the parents of all paediatric subjects in the study. Informed assent was also obtained from older children who were deemed capable of understanding the nature of the study.

Bacterial strains
Sutterella wadsworthensis strains. A total of twenty-seven S. wadsworthensis strains were used in this study. The type strain, DSM 14016 ( = ATCC 51579) was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ). Another twenty-six S. wadsworthensis strains obtained during the course of this study were also examined.
Other bacterial strains. Twenty-eight other bacterial strains were used in this study, obtained from international culture collections as well as from clinical and environmental sources:

Sutterella wadsworthensis growth conditions
S. wadsworthensis isolates were obtained by culturing 1-2 mucosal biopsy samples on five selective plates, the details of which are listed in Table 1. Biopsies were first ground in brucella broth before the resultant suspension was added to the plates. 50 ml was added to each plate with the exception of the filtered blood plate where 200 ml was first passed through a 0.45 mm filter. Cultures were incubated in a micro-aerophilic atmosphere, comprising of 5.9% oxygen, 7.2% carbon dioxide, 3.6% hydrogen and 83.3% nitrogen at 37uC. This atmosphere was generated using AnoxomatH Atmosphere Generating System, from MartH Microbiology b.v. (9200 JB Drachten, Netherlands). Plates were reviewed twice weekly for up to one month. Any bacterial isolate deemed Gram-negative and oxygen sensitive (by virtue of failed subculture in room air) was identified by sequencing of the 16S rRNA gene and sequence search on NCBI BLAST (http://blast. ncbi.nlm.nih.gov/Blast.cgi).

Genotypic Characterization
DNA Extraction. Genomic DNA was extracted from the colonic mucosal biopsies using the QIAamp DNA Mini Kit (Qiagen, Crawley, UK) according to an established modification of the manufacturer's instructions, optimised in-house for colonic biopsy tissue as described previously [5].
S. wadsworthensis specific primer design. For the design of a new S. wadsworthensis -specific primer, 16S rRNA gene sequences of all the S. wadsworthensis strains and other related bacterial strains were obtained from the NCBI database (http:// www.ncbi.nlm.nih.gov) for multiple alignment. Nearly full length 16S rRNA sequences of S. wadsworthensis strains isolated from clinical specimens in this study were also used for the alignment analysis. Several primer sequences were initially designed by using the Primer 3 software [14] and then the most suitable one was identified manually using the BioEdit software package (version 7.0.5.3) (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). The downloaded S. wadsworthensis specific sequences were aligned with the newly designed primer sequences and the ones that matched to the highly conserved 16S rRNA gene sequences of the target species but were variable among the other bacterial species, were selected. The 20 base pair S. wadsworthensis specific primer sequences designed (SWF and SWR) are outlined in Table 2. The primer sequences were then subjected to BLAST analysis against all the other sequences within the BLAST database to ensure that at least one of the primers did not share any identical sequence with any microorganisms other than S. wadsworthensis.
Optimization of S. wadsworthensis PCR assay. The specificity of the newly developed S. wadsworthensis primer pair was tested against a wide panel of bacterial species as listed earlier in the methods section. The PCR assay was tested using 40 ng of bacterial DNA in a 50 ml reaction mixture consisting of 10 pmol of each primer (SWF and SWR [Sigma-Aldrich, UK]), 1X PCR buffer (Roche, UK), 250 nM of each deoxy-nucleotidetriphosphate (Bioline, UK), 2 mM MgCl 2 (Roche, UK) and 1 U of Taq polymerase (Roche,UK). The sensitivity of the PCR assay was determined by serially diluting a known quantity of target DNA (50 ng/ml to 0.005 pg/ ml) to detect the minimum concentration that would yield a visible amplicon after gel electrophoresis.
Restriction fragment length polymorphism (RFLP) analysis was also done using the 555 bp S. wadsworthensis-specific PCR products using EcoR I and Hha I enymes to confirm that the PCR products were from a single bacterial species.
Sequence Analysis. Sequencing was done on an Applied Biosystems model 3730 automated capillary DNA sequencer using the S. wadsworthensis -specific primers SWF and SWR or 16S rRNA specific universal bacterial primers for the whole gene amplification of the S. wadsworthensis strains. The sequences obtained were compared to those of the National Center for Biotechnology Information GenBank database using the basic local alignment search tool (BLAST) search program. Multiple alignments and phylogenetic analysis were performed using Bioedit (http://www. mbio.ncsu.edu/BioEdit/bioedit.html) (version 7.0.5.3) and a dendogram was constructed using MEGA version 4 software [15]. The evolutionary distances were calculated according to Kimura's two-parameter model. A phylogenetic tree was inferred using the neighbor-joining algorithm, and the tree topology was statistically evaluated by 1,000 bootstrap resamplings.
GenBank Sequence Submission. All 16S rRNA gene sequences derived from the sequencing of S. wadsworthensisspecific PCR products were submitted to GenBank with the accession numbers from JN664092-JN664117.

Phenotypic Characterization
Basic biochemical tests including Gram staining, catalase, urease and oxidase tests were conducted using conventional manual methods to characterize the isolates. The strains were further characterized using the API Campy commercial biochemical kit according to the manufacturer's instructions (BioMerieux, La Balme Les Grottes, France). The purpose of using the API kits in this study was not to identify or name the bacteria, but to use the biochemical tests to further characterize the S. wadsworthensis strains. By comparing the resultant biochemical profiles, any potential differences between the strains might be elicited. Of all the API kits available it was assumed that API Campy would yield the best results, as S. wadsworthensis is phenotypically similar to Campylobacter species.

Scanning Electron Microscopy (SEM)
S. wadsworthensis strains were harvested for SEM. Bacterial cells were suspended in 2% Glutaraldehyde in 0.1 M sodium Cacodylate buffer (pH 7.2) for 3 hours to be fixed. The fixed cell suspensions were then applied to glass coverslips coated with poly-L-lysine and allowed to adhere for five minutes. Coverslips were then rinsed with water and post-fixed with 1% Osmium tetroxide (OsO4) for 30 minutes. They were next rinsed with water and dehydrated through a graded ethanol series (60%, 80%, and 90%) for 10 minutes each. The sample was then left for 30 minutes in absolute ethanol with ethanol changes every 10 minutes. The coverslips were kept in Hexamethyldisilazane (HMDS) SigmaH for 10 minutes, and left to dry overnight. Next day, the specimen was coated with gold using automated sputter coater EMITECH K550 (Emitech limited, Ashford, Kent) and samples were transferred to view S. wadsworthensis using SEM.

Proteomic Characterization
To complement the phenotypic and genotypic methods, matrixassociated laser desorption/ionisation -time of flight mass spectrometry (MALDI-TOF MS) was performed to further characterize selected S. wadsworthensis strains. Sample preparation. A total of 11 bacterial strains were analyzed by MALDI-TOF MS: eight S. wadsworthensis isolates (4 strains obtained from healthy controls and 4 strains from IBD patients), the S. wadsworthensis type strain DSM 14016, C. jejuni and C. concisus strains. Approximately 5-10 mg (half a plastic loop) of bacterial cells were harvested from fresh blood agar plates and proteins were extracted as described by Alispahic et al. (2010) [16]. The protein extract (supernatant) was diluted, with 30% acetonitrile in 0.1% trifluoracetic acid (Applied Biosystems, UK) in water. Subsequently, 1 ml of diluted supernatant was mixed with 1 ml of matrix solution (10 mg HCCA (a-cyano-4hydroxycinnamic acid) matrix (Bruker Daltonics, Germany) to 1 ml of a solution comprising: 500 ml acetonitrile, 475 ml UHQ water and 25 ml trifluoroacetic acid). 1 ml of the sample-matrix mixture was pipetted onto a MALDI ground steel target plate (Bruker Daltonics) and left to dry on air.
MALDI-TOF MS. Mass spectra were acquired using an ultrafleXtreme TM series MALDI-TOF MS (Bruker Daltonics), equipped with a 1 kHz smartbeam-II TM nitrogen laser (l = 337 nm). Spectra were recorded using the linear positive mode for masses in the range of 2 kDa to 20 kDa. The parameter settings were as follows: ion source 1, 25 kV; ion source 2, 23.85 kV; lens, 6.5 kV; pulse ion extraction time, 160 ns. A lowmass gate of 800 Da was used. Insulin, ubiquitin, cytochrome C and myoglobin were used as external calibrants, enabling a mass accuracy of 60.5 parts per 1000. Laser power and the number of laser shots per sample were varied by the operator for optimal performance.
Data analysis. Mass spectra were processed prior to visual inspection using FlexAnalysis 3.3 software (Bruker Daltonics). This included smoothing, baseline subtraction and intensity normalisation. MALDI BioTyper software 2.0 (Bruker Daltonics) was used to generate pseudo gel view representations of each spectrum. The mass spectra and corresponding pseudo gel images for each bacterial strain were then visually inspected and compared.

Preparation of monocytes and S. wadsworthensis (whole cell) stimulation studies
Whole venous blood was collected from six adult healthy individuals who were not on any medications. The blood was diluted 1:1 in RPMI 1640 (Sigma-Aldrich), layered onto Histopaque-1077 (Sigma-Aldrich), and centrifuged at 8006g for 30 min. Peripheral blood mononuclear cells were removed from the interface and monocytes were prepared as published earlier [17]. The effect of S. wadsworthensis whole cell preparations was assessed by stimulating the monocytes for up to 24 hours. Seven S. wadsworthensis strains (4 strains obtained from healthy controls and 3 strains from IBD patients) were used in this study along with an E. coli (S1041) as a positive control. Unstimulated monocytes were used as a negative control.

Cytometric bead array
In order to assess whether the whole cell preparations induced an inflammatory response, levels of TNF-a in culture supernatant were measured at the different time points using a human inflammatory cytokine cytometric bead array (CBA;BD Biosciences). Samples were analyzed on a multifluorescence BD FACSCalibur TM flow cytometer using BD CellQuest TM software and BD TM cytometric bead array software. Standard curves were generated using FCAP Array Software TM . Sample results were then calculated by comparing sample CBA results to the respective standard curve. All CBA work was performed in duplicate and cytokine levels induced by stimulation were calculated by subtracting unstimulated cytokine levels from each stimulated value and expressing TNF-a levels as a fold-change relative to TNF-a levels derived from E. coli stimulations.

Statistical Analysis
Statistical analysis was performed using the Pearson Chi Squared, 2-tailed test or the Fisher's exact test wherever appropriate, utilizing Graph Pad software (San Diego, CA).

Assessment of S. wadsworthensis prevalence in mucosal colonic biopsies
Clinical information of the study population. The mean age of the sixty-nine patients with UC was 45.6627.8 years and 46.4% were male as opposed to the sixty-five control patients whose mean age was 64.265.6 years and 59.3% of these subjects were male. There was a statistically significant difference in age between the UC cohort and the control group (p,0.0001). At the time of colonoscopy the adult patients were classified as extensive (24.6%), left sided (63.8%) and proctitis (11.6%) according to the Montreal criteria. During recruitment 10.1% of patients were in clinical remission (S0), 23.2% had mild UC (S1), 46.4% had moderate UC (S2) and the remaining 20.3% had severe UC (S3). A total of 115 biopsy sites were analyzed from UC patients. Twenty-four (34.8%) subjects had a single site analyzed and fortyfive (65.2%) had more than one biopsy site assessed. A single biopsy site was analyzed from each control subject. The adult patients served as clinical sources for the isolation of S. wadsworthensis. The newly developed PCR was utilized to detect the presence of S. wadsworthensis from mucosal biopsies from these patients.
The paediatric IBD patients (n = 29, male 62.1%) had a median age of 12.2 years (interquartile range, IQR, 9.8, 14.1 years) at the time of recruitment out of which fifteen had CD, nine UC and five IBD-unclassified. The paediatric control patients (n = 32, male 78.1%) had a mean age of 10.8 years (IQR, 8.3, 12.7 years) at the time of index colonoscopy. The paediatric cohort was primarily utilized as a source for isolation of S. wadsworthensis, which was then utilized to develop and standardize the detection PCR.
A total of twenty-six S. wadsworthensis strains were isolated from the mucosal colonic biopsy samples of the above mentioned study populations. The majority of the clinical strains were obtained from paediatric patients, ten from control and seven from IBD patients. Six strains were obtained from adult IBD patients and three from adult control patients.
Evaluation of the sensitivity and specificity of the newly developed S. wadsworthensis -specific PCR assay. Twentyseven strains of S. wadsworthensis and twenty-eight other bacterial strains representing a broad spectrum of bacterial species were screened to test for the specificity of the newly designed speciesspecific primers. The primer pair produced an intense band of <555 bp from all the S. wadsworthensis strains (Figure 1). No amplicon was observed from any of the other bacterial strains tested. For detection of S. wadsworthensis DNA from mucosal biopsy samples a nested PCR approach was taken where the first round PCR was done using universal bacterial primers (27F and 1492R) [18] followed by a second round of PCR using the S. wadsworthensis-specific SWF and SWR primers. The optimal thermal cycling conditions required to obtain specificity for the S. wadsworthensis-specific PCR were: 94uC for 5 minutes, 30 cycles of 94uC for 30 seconds, 58uC for 30 seconds, and 72uC for 2 minutes, followed by 72uC for 15 minutes. The optimized PCR was able to detect 0.5 pg/ml of S. wadsworthensis DNA (Figure 2).
The RFLP analysis of all the S. wadsworthensis-specific PCR products also showed the expected banding pattern of 327 bp and 228 bp for EcoR 1 digestion and 420 bp and 135 bp for Hha I digestion respectively.
Prevalence of S. wadsworthensis in adults with UC using species-specific PCR. All the adult mucosal biopsy samples were screened using the newly developed and optimized S. wadsworthensis-specific PCR. S. wadsworthensis was detected in fiftyseven of the sixty-nine subjects with ulcerative colitis and fifty-six of the sixty-five controls. The prevalence of S. wadsworthensis in the UC population was 83.8% which was similar to that in the controls, 86.1% (p = 0.64). There was no statistically-significant difference in the prevalence of S. wadsworthensis in male subjects as opposed to female patients. Furthermore, no age related differences could be demonstrated. Looking specifically at patients with UC, no statistically-significant differences were noted in the prevalence of S. wadsworthensis with respect to extent or severity of UC according to the Montreal classification.
Phylogenetic Analysis. A phylogenetic tree was constructed based on nearly full-length (,1400 bp) 16S rRNA gene sequences of the S. wadsworthensis strains isolated from this study and sequences of S. wadsworthensis strains and other Sutterella species available in GenBank ( Figure 3). All the S. wadsworthensis sequences derived from this study clustered closely with the other S. wadsworthensis sequences from GenBank and presented the highest (99-100%) sequence identity to the S. wadsworthensis type strain ATCC 51579. The S. wadsworthensis strains from different geographical locations (strain ATCC 51579, strain WAL 7877 and strain WAL 9054 originating from USA and the strains from this study being from the UK) did not cluster into different lineages. Also, S. wadsworthensis sequences analyzed from the UC and the healthy controls did not cluster into separate groups in the dendrogram. The GenBank derived sequences from the S. parvirubra, S. stercoricanis and S. morbirenis strains each grouped into different clusters in the dendrogram confirming different species.

Characterization of S. wadsworthensis isolates to detect strain difference
As the PCR results revealed almost universal prevalence of S. wadsworthensis in both IBD cases and controls, we hypothesised that there may be differences between the strains isolated from the two groups and so went on to study their phenotypic, morphological, proteomic and pathogenic profiles.
Phenotypic Characterization. S. wadsworthensis was found to be a Gram-negative pleomorphic rod shaped organism with   different morphological structures seen on Gram stain, even though the S. wadsworthensis cultures were pure. All S. wadsworthensis strains tested negative for catalase, urease and oxidase tests. The API Campy kit results showed that all the S. wadsworthensis strains tested (9 clinical isolates and the type strain) were positive for the EST, ArgA and AspA tests, indicating they possess the enzymes esterase, L-arginine arylamidase and L-aspartate arylamidase respectively. The only discrepant result between the S. wadsworthensis strains was for the HIP test; four strains, including the type strain, tested positive and six strains were found to be negative. This tested the ability of the bacteria to hydrolyse sodium hippurate. A positive result indicated the presence of the enzyme hippurate hydrolase. All the assimilation tests on the second half of the strips were negative for all S. wadsworthensis strains.
SEM. Under conventional light microscopy, there were clearly two different morphologies for S. wadsworthensis. For this reason, an intensive literature search for S. wadsworthensis electron micrographs was done. However, there were no published electron micrograph images available to answer the morphology question. Due to this, two S. wadsworthensis strains (SW9 from control and SW13 from IBD) were selected for SEM. The SEM images clearly indicated lattice formation ( Figure 4A) and two distinct morphologies of this bacterium: a long rod (nearly 4mm) and a short cocco-bacillus (nearly 1mm) ( Figure 4B). Interestingly, there were also cells with filamentous shape ( Figure 4C) and a helical form was also observed ( Figure 4D).
MALDI-TOF MS. Matrix-associated laser desorption/ ionization -time of flight mass spectrometry (MALDI-TOF MS) is a well-established technique in proteomics for the analysis of proteins and peptides. Profiling of proteins by MALDI-TOF MS has been developed as a rapid and sensitive technique for microorganism identification [16]. In the MALDI Biotyper system developed by Bruker Daltonics, MALDI-TOF mass spectra of small proteins (range 2-20 kDa) acquired directly from smears of cells or their extracts are compared with reference spectra in the MALDI Biotyper database. The most recent MALDI Biotyper 2.0 database available (version 3.1.2, dated 28 January 2011, containing reference spectra for 3740 microorganisms) did not contain any information on S. wadsworthensis. The purpose of MALDI-TOF MS in this study, as with the API testing kit, was not to identify the bacterial strains but to further characterize them and elicit any differences between them.
Out of the eleven strains tested, triplicate samples of the first six bacterial strains analyzed by MALDI-TOF MS were analyzed to assess how reproducible the method was. All three replicates for these six strains produced almost identical results, so it was decided that obtaining a single mass spectrum for each of the remaining five strains would be sufficient. The remaining five strains were analyzed and the mass spectra for all 11 strains were obtained. It was immediately apparent that there was a characteristic S. wadsworthensis pattern that all of the strains conformed to ( Figure 5A). Notably there was a dominant peak at approximately 9400Da which was common to all S. wadsworthensis strains ( Figure 5B). The Campylobacter strains on the other hand, produced different patterns of peaks, with C. jejuni producing a dominant double peak at approximately 9500Da and 10300Da (data not shown).
Monocyte challenge. A human monocyte challenge system was used to assess the immunogenicity of different S. wadsworthensis strains. A survival test was performed to examine the ability of S. wadsworthensis strains to survive when exposed to the oxygen present in the monocyte stimulation experimental setup. This showed that .50% of S. wadsworthensis cells survived for four hours. Seven S. wadsworthensis strains and an E. coli strain were inoculated onto monocytes that were obtained from six healthy volunteers. Levels of TNF-a were measured at 4 hours of stimulation. All S. wadsworthensis strains induced TNF-a cytokine production from the human monocytes. Compared to the E. coli strain used S. wadsworthensis strains induced higher levels of TNF-a production over the first 4 hours although subsequent measurements at 24 hours indicated that TNF-a levels remained detectable in E. coli stimulated monocytes but that TNF-a was not detectable in S. wadsworthensis stimulated wells, most likely because of the death of S. wadsworthensis organisms (Table 3).

Discussion
The role of S. wadsworthensis in human gastrointestinal diseases has been documented in the past but its specific involvement in inflammatory bowel disease has never been firmly established [8,9]. The fortuitous isolation of this unusual bacterial strain from mucosal biopsies of patients with IBD prompted a series of experiments that have been outlined in this paper and has culminated in detailed characterization of this putative pathogen. Despite this detailed microbial analysis no specific differences have been identified between cases and controls. However, this negative result does not preclude the importance of successful isolation of this fastidious organism from colonic biopsy samples by adhering to a strict microaerophilic culture environment and stringent growth conditions. Our study provides an important proof of concept that will enable future studies to satisfy Koch's postulates to tie in other luminal microaerophilic bacteria as causative factors in IBD. This is the first report of the appearances of this bacterial species under scanning electron microscopy (SEM). As opposed to previous descriptions, the bacteria are pleomorphic, with strains existing in predominantly two morphological forms, long rods and coccobacilli. Filamentous and helical forms were also documented. The bacteria appeared in clusters on Gram stain and this was further corroborated with the appearance of a lattice type network on SEM. The functional role of this lattice formation is unclear, but given the high prevalence of this organism that we have demonstrated in the human colonic mucosa, further investigation of the interaction of S. wadsworthensis in the complex luminal microbiota may be of interest. This lattice structural network may paradoxically harbour either commensal bacteria or pathogenic bacterial strains in close juxtaposition to the epithelial surface with diametrically opposite consequences.
All S. wadsworthensis strains were negative for urease, catalase and oxidase, consistent with the initial description by Wexler et al [8]. The authors had described S. wadsworthensis as asaccharolytic, and certainly none of the sugar assimilation tests from the API kit were found to be positive [8]. As S. wadsworthensis is phenotypically closely related to the Campylobacter genus, the API Campy kit, specifically for the identification of Campylobacter species was trialed. This yielded more promising results. All of the S. wadsworthensis strains tested were positive for the EST, ArgA and AspA tests, meaning they possessed the enzymes esterase, Larginine arylamidase and L-aspartate arylamidase respectively. Esterases are often required to breakdown dietary residue and are mostly derived from the intestinal microbiota, with E. coli, Bifidobacterium and Lactobacillus species all known to be sources [19]. Colonic esterases have been shown to release powerful antioxidants, present in cereal bran, from complex ester-linked structures which could otherwise not be absorbed [20]. It is possible that the esterases produced by S. wadsworthensis have a similar role.
This study utilized a newly developed PCR to detect the presence of S. wadsworthensis in a subset of adult patients with UC and adult control patients. The rates of detection was similar in both these groups (83.8% vs. 86.1%, p = 0.64), suggesting that this bacterial strain is probably a commensal. The only other study that has looked at the prevalence of this pathogen in patients with gastrointestinal disease was by Engberg et al. who used PCR and DNA sequencing of the 16S rRNA gene to characterize bacterial populations in fecal samples [21]. They detected S. wadsworthensis in only 7 out of 1483 (0.47%) patients with GI disorders, and only 1 out of 107 (0.93%) healthy individuals. This wide disparity in prevalence between this study and our cohort of patients and controls suggests that S. wadsworthensis is closely adherent to the mucosal lining and is more likely to be identified from biopsy samples as opposed to feces. Additionally, this newly designed PCR for this bacterial strain was validated using the type strain of S. wadsworthensis as well as several related bacterial strains and will form a useful tool in future studies.
Despite similar rates of detection of S. wadsworthensis from patients with IBD and controls, our clinical isolates gave us an excellent opportunity to further characterize this organism. Several bacterial species have been noted to have distinctly different phenotypic and genotypic profiles when isolated from patients with clinical disease as opposed to controls. This has been recently documented in Campylobacter concisus and mucosa associated E. coli wherein distinct genomospecies have been documented with varying pathogenic potential [22,23]. Sequence analysis of strains identified during the study from cases and controls did not cluster in distinct groups confirming that this phenomenon does not extend to S. wadsworthensis.
Proteomic analysis of S. wadsworthensis strains utilizing MALDI-TOF MS is being reported for the first time in this paper. MALDI-TOF MS has recently emerged as an alternative to phenotypic and genotypic methods for the fast and reliable identification of Figure 5. Mass spectra obtained for S. wadsworthensis strains. (A) Spectra obtained for 6 S. wadsworthensis strains: S. wadsworthensis type strain, SW1, SW4, SW5, SW6 and SW7. All of the S. wadsworthensis strains conform to a common pattern. a.u. arbitary unit; m/z, mass-to-charge ratio; Da, Daltons; (B) Overlaid spectra of five S. wadsworthensis strains overlaid. This is a close-up of the mass spectra from 8,700 Da to 10,500 Da. It shows how closely each S. wadsworthensis mass spectrum, aside from intensity, matches the others. a.u., arbitary unit; m/z, mass-to-charge ratio; Da, daltons. doi:10.1371/journal.pone.0027076.g005 Table 3. TNF-a levels (depicted as percentage) derived from human monocytes (N = 6) exposed to whole cell preparations of various S. wadsworthensis strains.  [16,24,25,26]. It can be used to classify closely related bacteria, which may be indistinguishable by conventional methods [24]. It was therefore employed to investigate the diversity of the S. wadsworthensis protein profiles. A characteristic pattern was clearly demonstrated on the analysis of S. wadsworthensis strains, with a dominant peak at approximately 9400Da. Although no major differences were detected between the strains, MALDI-TOF MS proved to be a highly sensitive and reproducible method for the characterization of the bacterial mass spectra. Very small mass of sample were required, which was ideal for analyzing the slow and modest growing S. wadsworthensis. This finding will provide a useful reference for future studies and an alternative to genotypic methods like 16S rRNA gene sequencing. The in-vitro cytokine analysis after monocyte challenge failed to show any distinct differences between strains isolated from patients with IBD and controls. This last finding essentially closes the loop in the series of studies looking at the possible role of S. wadsworthensis in IBD. Our experiments have conclusively shown that the prevalence of this bacterium is similar in IBD patients and controls and that there is no phenotypic, genotypic, proteomic or pathogenic characteristic to distinguish bacteria isolated from these two groups of patients. It is quite likely that S. wadsworthensis is a commensal and a harmless bystander amidst the inflammatory cascade that is typical of inflammatory bowel disease.