Identification and Characterisation of Pseudomonas 16S Ribosomal DNA from Ileal Biopsies of Children with Crohn's Disease

Molecular analysis of bacterial 16S rRNA genes has made a significant contribution to the identification and characterisation of bacterial flora in the human gut. In particular, this methodology has helped characterise bacterial families implicated in the aetiology of inflammatory bowel disease (IBD). In this study we have used a genus specific bacterial 16S PCR to investigate the prevalence and diversity of Pseudomonas species derived from the ileum of children with Crohn's disease (CD), and from control children with non-inflammatory bowel disease (non-IBD) undergoing their initial endoscopic examination. Fifty eight percent of CD patients (18/32) were positive using the Pseudomonas PCR, while significantly fewer children in the non-IBD group, 33% (12/36), were PCR positive for Pseudomonas (p<0.05, Fischer's exact test). Pseudomonas specific 16S PCR products from 13 CD and 12 non-IBD children were cloned and sequenced. Five hundred and eighty one sequences were generated and used for the comparative analysis of Pseudomonas diversity between CD and non-IBD patients. Pseudomonas species were less diverse in CD patients compared with non-IBD patients. In particular P.aeruginosa was only identified in non-IBD patients.


Introduction
Crohn's disease (CD) is a chronic inflammatory condition of the gastrointestinal tract, predominantly affecting the ileo-caecal region. CD is increasing in incidence in numerous countries throughout the world [1,2,3]. One model of CD suggests that it is initiated in genetically susceptible individuals by an infectious agent capable of triggering a breakdown in the regulatory constraints on the mucosal immune system and leads to an immune mediated tissue injury. However, the infectious agent(s) that trigger these responses is unclear. Several putative susceptibility genes have been implicated in CD, including CARD15, IL-23R and TLR-4 [4,5,6]. The role of CARD15 in sensing bacterial peptidoglycan, and the association between mutant forms of this gene and CD, further suggest that micro-organisms may play a key role in disease aetiology.
Mycobacterium avium subspecies paratuberculosis [7] has been the most widely implicated aetiological agent of CD [8][9][10]. Adherent-Invasive E.coli [11], Listeria monocytogenes [12], Pseudomonas [13,14] and Yersinia [15] have also been proposed as triggers of CD. Some authors have also argued that CD results from a loss of tolerance to the normal flora of the gastrointestinal tract [16]. This hypothesis is supported by observations that CD does not occur in gnotobiotic rats, and that some CD patients have significantly higher antibody levels to commensal gut bacteria [17].
It has also been suggested that CD is the result of dysbiosis i.e. an imbalance between 'beneficial' (non-inflammatory) commensals and 'harmful' (pro-inflammatory) commensals [18]. Using bacterial 16S ribosomal RNA (rRNA) gene sequencing, the accepted standard of bacterial identification [19], studies have demonstrated a significant difference in the gut flora of CD patients compared with control populations, providing support for the dysbiosis hypothesis [20][21][22]. Such studies report that members of the phylum Bacteriodetes and Proteobacteria are increased in CD patients. A preliminary study reported by Eckburg and Relman using 16S rDNA PCR also identified significantly more Proteobacteria phylumassociated 16S rDNA sequences (preliminary E. coli and Pseudomonas species) from colonic tissue samples obtained from patients with CD than from non CD patients [23] Bacteria within the genus Pseudomonas have also been associated with the development of CD. Several Pseudomonas species are opportunistic pathogens including P.aeruginosa and P.fluorescens [25,24]. Some, such as P.aeruginosa and P.putida, are also minor members of the normal gastrointestinal microbiota [26][27][28]. Early studies reported that variants of Pseudomonas-like bacteria were detected in the gastrointestinal tract of CD patients and found to be absent in the gastrointestinal tract of patients without inflammatory bowel disease (IBD) [29][30][31]. More recently, a Tcell superantigen (I2), encoded by P.fluorescens, has been associated with lesions in CD patients [13,14]. However, many of these studies were performed on adult patients with long standing disease, who had experienced a variety of treatment regimes, including antibiotics that would have altered intestinal flora.
A series of preliminary subtractive hybridisation experiments conducted in our laboratory comparing DNA extracted from CD and non-IBD gut biopsies detected Pseudomonas specific 16S rDNA in the differentially expressed populations (Kirkwood, unpublished observations). These experiments implied that the role of Pseudomonas species in early onset CD should be investigated further.
Here we present a study investigating the role of Pseudomonas species in CD and non-IBD paediatric patients presented for initial diagnosis prior to medical treatment. Specifically, the diversity of Pseudomonas species in the mucosa of gastrointestinal tracts of children with and without CD was investigated. We have conducted a Pseudomonas specific 16S ribosomal analysis on ileal tissue obtained from CD and non-IBD patients, and subsequently cloned and sequenced PCR products from each patient. The resulting Pseudomonas gene libraries were compared using LIB-SHUFF (LIBrary SHUFFling) analysis [32] and DOTUR (Distance Based Operational Taxonomic Unit and Richness Determination) analysis [33].

Identification of Pseudomonas in clinical specimens obtained from CD and non-IBD patients
Pseudomonas specific 16S (P16S) ribosomal DNA PCR was performed on DNA extracted from 64 ileal biopsy specimens. Fifty eight percent of CD patients (18/32) were Pseudomonas positive, compared with 33% (12/36) non-IBD patients. Statistical analysis revealed a significant difference between Pseudomonas positive patients in the CD group compared with the non-IBD group (p = ,0.05, Fisher's exact test).

Pseudomonas 16S sequences diversity in CD and non-IBD patients
The diversity of Pseudomonas species present between the two patient groups was analysed by comparing sequences of the 617 bp P16S PCR products from 13 CD and 12 non-IBD patients. A minimum of 20 sequences were analysed for each PCR product, and were used to create a partial P16S gene library for each patient group. Each library of P16S sequences were aligned using ClustalW. A PHYLIP generated DNA distance matrix was calculated using the Jukes-Cantor method and used for LIB-SHUFF. LIBSHUFF analysis showed that the P16S gene libraries of CD and non-IBD control patients were significantly different (p,0.025).
All 581 sequences from the P16S gene library were organised into operational taxonomic units (OTU) using DOTUR. Each OTU consists of a group of P16S sequences with 97% sequence similarity. The P16S sequences were grouped into 6 OTUs of which five were identified in the CD clone library and all six in the non-IBD clone library (supporting table S3). Five OTUs were common to both clone libraries. One OTU was unique to the non-IBD clone library. The majority of sequences (67.2% CD and 43.3% non-IBD) were found in the OTU 1. The second most commonly identified sequences were found in OTU 2 (18.8% CD and 13% non-IBD). OTU 3 sequences were present in 11.3% of the CD group and in 8.4% of the non-IBD group. OTU 4 represented 10% of sequences in the non-IBD group whilst this OTU was only a minor fraction within the CD group (0.6%). OTU 5 sequences were only identified in the non-IBD group (18%). OTU 6 was the least common type, identified in 7.3% of non-IBD and 2.2% of CD sequences ( Figure 1). The number of sequences in each individual OTU was tested against the number of sequences in the remaining OTUs between the two patient groups using Fischer's exact test. A significant difference was identified in OTU 1 and OTUs 4, 5, and 6 between CD and non-IBD patient groups (p,0.01) ( Figure 1 marked with *).

Phylogenetic analysis of Pseudomonas 16S gene library from CD and non-IBD patients
A consensus sequence of each of the 6 OTUs was generated, representing the 581 sequences determined from Pseudomonas 16S PCR products. Figure 2 shows the phylogenetic analysis of each OTU consensus sequences and sequences from 85 Pseudomonas reference strains. Each individual OTU clustered into distinct groups with specific Pseudomonas reference strains. OTU 1 grouped with P.migulae, P.proteolytica, P.brenneri, and P.panacis. OTU 2 grouped with P.mendocina, P.alcaliphila and P.pseudoalcaligenes. OTU 3 grouped P.mosselii, P.plecoglossicida and P.monteilii. OTU 4 grouped with P.fragi, P.lundensis, and P.psychrophila. OTU 5 formed a distinct group with P.nitroreducens and OTU 6 formed a distinct P.aeruginosa group.

Discussion
In this study we have conducted a molecular analysis using 16S ribosomal DNA to compare the composition and distribution of Pseudomonas species identified in ileal specimens obtained from CD  and non-IBD paediatric patients. This study extends on previous studies which reported significant differences in bacterial composition between IBD and non-IBD patients at higher bacterial hierarchy levels [20,21,38,39]. In the present study, Pseudomonas 16S (P16S) rDNA was detected in a proportion of ileal biopsy samples analysed from both patient groups. This is consistent with previous findings that describe Pseudomonas as normal, low level members of the gut microbiota [26,28]. Pseudomonas species were found to be significantly more prevalent in biopsy specimens obtained from CD patients (58%) when compared to specimens from non-IBD patients (33%) (p,0.05, Fisher's exact test). This finding suggests that there is an altered mucosal environment in CD patients that favours the growth of Pseudomonas species, which correlates with previous studies that reported Pseudomonas like species in CD patients and not in non-IBD patients [29,30].
The P16S sequence libraries between the two patient groups differed significantly from each other (p,0.025, LIBSHUFF analysis). To analyse the diversity of Pseudomonas species within CD and non-IBD libraries, all 581 P16S sequences were grouped into operational taxonomic units (OTUs). The non-IBD gene library was more diverse than the CD library with 6 OTUs, compared to 5 OTUs in the CD library suggesting that the decline in diversity may occur early in onset disease. All patients used in this study were children, with an average age of 11.7 years. Thus, environmental factors (such as smoking and oral contraceptive use) which typically affect studies on adults with CD, are unlikely to have significantly affected the data [40,41]. These results were not confounded by prolonged treatment with anti-inflammatory drugs or antibiotics since all patients were recruited at initial clinical examination.
Differences in the distribution of Pseudomonas species were identified in the CD and non-IBD P16S libraries. The proportion of sequences found in 4 of 6 OTUs were significantly different between the groups (p,0.01, Fischer's exact test). This shifting in numbers of different Pseudomonas species between patient groups might contribute to an imbalance between beneficial and harmful microbiota leading to the inflammatory changes seen in CD.
It is widely recognised that Pseudomonas is a very diverse group of bacteria, with many species capable of causing human disease. Pathogenic species are contained within each OTU were identified in this study. For example, P.migulae, reported to be present in joints in patients with chronic reactive arthritis [42] was within OTU 1. While P.pseudoalcaligenes, an organism isolated from respiratory tract culture of cystic fibrosis patients [43] was within OTU2. P.mosselii and P.monteilii species were closely related to sequences within OTU 3. These novel species were isolated from clinical specimens such as urine, sputum, placenta, skin, and bronchial samples [44]. The pathogenic strain P.fluorescens indentified in OTU 4 has been postulated to be implicated in CD [13,14]. However, our study does not support a role for P.fluorescens in CD since there was a significant reduction in species within the OTU 4 group in CD compared with non-IBD patients.
The most striking finding of the P16S sequence analysis was the complete lack of Pseudomonas species within OTU 5 in the CD patients, while OTU 5 represented 18% of species in the non-IBD patients. Sequences in OTU 5 were restricted to Pseudomonas aeruginosa. The loss of commensal P.aeruginosa in CD patients may result in dysbiosis, and could play a role in CD. P.aeruginosa is known to inactivate several pro-inflammatory cytokines [45]. In other settings, P.aeruginosa has been reported to be implicated in causing human disease, including chronic granulomatous disease in paediatric patients [46] and as a trigger in cystic fibrosis [47].
The worldwide increase in the prevalence of CD has been associated with an increased use of refrigeration [48]. There are no studies on whether changes in psychotropic bacteria (able to grow at the temperatures maintained inside a refrigerator, 21 to +10 degrees) occur in CD populations. In this study, OTU 1 that included the nonpathogenic species P.panacis and P.brennerii (reported to be present in refrigerated beef [49]) was more prevalent in CD patients than in non-IBD patients. The presence of these species in CD in relation to disease manifestation requires further investigation.
The imbalance in Pseudomonas species seen in our study might be a secondary consequence of defective mucosal barrier function and microbial killing present in IBD patients [50]. This could have led to an aggressive T-cell response and changed the composition of Pseudomonas species early in disease progression. Numerous studies have demonstrated changes in bacterial composition between beneficial and harmful bacterial species in IBD and non-IBD patients. This dysbiosis hypothesis is accompanied by a decrease in microbial diversity in active IBD [20,21,38,39,51,52], similar what was observed with Pseudomonas in this study.
Pseudomonas is one of the most complex genera of Gram-negative bacteria with a wide range of natural habitats which might partly explain their presence in human gut. By using Pseudomonas 16S ribosomal analysis no single Pseudomonas species was identified as a causative agent of CD. However, Pseudomonas species were detected more frequently in CD patients and the Pseudomonas population in these patients was significantly less diverse when compared to that of non-IBD patients. Of particular interest was the lack of P.aeruginosa detected in the gut of CD patients. The dysbiosis between Pseudomonas species and other gut microbiota could be directly involved in CD manifestation or be a secondary consequence after establishment of disease in these children.
Finally, much of the gastrointestinal microbiota remains uncharacterised [53]. It is likely that many of the sequences detected in this study represent novel Pseudomonas species/strains. Full Pseudomonas ribosomal gene analysis (16S-23S) could provide more insights into specific Pseudomonas species associated with CD. Whilst the results of this study accurately represent a part of the bacteriology of active CD, this study did not address the important role that the immune system and immunological tolerance may play in CD. Specifically, it remains possible that an inappropriate immune response to commensal Pseudomonas species is involved in disease aetiology.

Patient sample collection
Patients undergoing an initial endoscopy at the Royal Children's Hospital, Melbourne, Australia for a clinical diagnosis of inflammatory bowel disease (IBD) were enrolled in this study. All of the patients had exhibited symptoms consistent with IBD. None of the patients had received antibiotics, immunosuppressives drugs or glucocorticosteroids prior to endoscopy. A total of 68 patients were included in this study, 32 children were diagnosed with Crohn's disease (CD) and 36 children were diagnosed with no pathological abnormalities (non-IBD). Mucosal biopsy specimens obtained from the ileum were used for all analyses. The presence or absence of granulomas was recorded for each CD patient. A detailed description of patient characteristics is listed in supporting tables S1 and S2, and includes presence of granulomas, inflammation, sex and age. Both patient groups had a mean age of 12 years. All tissue specimens were collected and placed immediately into a sterile screw-cap cryotube containing 0.5 ml of RNAlater [Ambion]. The samples were stored at 270uC until DNA extraction. Samples were examined in the order in which they were collected, so that CD and non-IBD cases were examined in a random manner. The clinical diagnosis of CD was established using standard clinical and endoscopic criteria according to the Vienna classification [34] and confirmed by histological examination of gut tissue. Ethics approval for the study was obtained from the Human Ethics Committee of the Royal Children's Hospital (EHRC no. 23003). Written informed consent was obtained from each individual, parent or guardian prior to enrolment in the study.

DNA extraction from biopsy specimens
Tissue samples (3-6 mm 3 ) were homogenised using a sterile plastic pestle in 200 ml of lysis buffer (20 mM Tris-HCl, pH 8.0 and 2 mM EDTA). The homogenate was subjected to three cycles of freeze/ thaw. Lysozyme (20 mg/ml, Sigma) was added to each sample and incubated at 37uC for 30 minutes. A volume of 600 ml buffer AL (QIAamp DNA mini kit, Qiagen) and glass beads (,20%, diameter 150-212 microns, Sigma) were added to each sample and subjected to mechanical disruption in a FastPrep (setting of 6.5 ms 22 for 45 seconds). Glass beads were removed and supernatant was treated with proteinase K (24 mAU) incubated at 56uC for 2 hours with vortexing 2-3 times per hour. DNA was extracted using the QIAamp DNA mini kit according to the manufacturer's instructions [QIAgen]. In an alternative method DNA was extracted as above except the freeze/thaw cycles and the FastPrep cycle were omitted. Briefly, tissue sample was homogenised as above in 200 ml buffer RLT [QIAgen]. Then two 5 mm sterile glass beads and an additional 400 ml buffer RLT were added and homogenate was incubated for one hour with 15 seconds vortexing every 10 minutes. Glass beads were removed and the DNA was extracted using the QIAamp DNA mini kit according to the manufacturer's instructions [QIAgen]. DNA was eluted into 50 ml nuclease free water. Two DNA extraction methods were employed for 16s rDNA amplification, since studies conducted in our lab has shown that the use of different extraction methods can influence the diversity of 16s rDNA sequences obtained (unpublished observations).
All DNA extractions were conducted in biological safety cabinet class II. A mock homogenisation and extraction without sample material was included in each extraction run to exclude contamination with bacteria during the procedure.

Amplification of Pseudomonas 16S ribosomal DNA
A Pseudomonas specific 16S ribosomal (P16S) PCR (617 bp) was conducted using Pseudomonas specific primers described by Spilker et al., [35]. Five ml of purified DNA was used in a 50 ml PCR reaction. PCR reactions were performed using a final concentration of 16 PCR Buffer, 2 mM MgCl, 2.5 U Taq polymerase [Applied Biosystems], 250 mM of each deoxynucleoside triphosphate [Invitrogen] and 500 nM of the Pseudomonas genus specific primers, PA-GS-F (59-GACGGGTGAGTAATGCCTA3-9), and PA-GS-R (59-CACTGGTGTTCCTTCCTATA-39) [35]. PCR conditions were: 95uC for 5 minutes, 10 cycles of 94uC for 15 seconds, 53uC for 30 seconds and 72uC for 45 seconds, this was repeated for another 25 cycles with the exception that the 72uC elongation step was increased by 5 seconds every cycle. A final extension phase of 72uC for 10 minutes was used. A nested PCR assay was conducted using all DNA samples which were negative or weakly positive (not sufficient for gel extraction and cloning). Universal bacterial 16s primers described by Wang et al., [36] were used in the first round PCR. PCR reactions were performed in a total volume of 50 ml using a final concentration of 16  . PCR conditions were as above except an annealing temperature of 57uC was used. E.coli DNA was used as a positive control for the first round and as negative control for the second PCR. P.aeruginosa and P.putida DNA was used as a positive control for both rounds of PCR. In all PCR assays a mock extraction and a negative PCR reaction were included. PCR products were separated by electrophoresis using a 1% TBE agarose gel and visualized under UV light using ethidium bromide. Amplicons of the correct size were extracted from the agarose gel and purified using a QIAquick Gel Extraction kit [QIAGEN], according to the manufacturer's instructions.

Cloning and sequencing of Pseudomonas 16S rDNA fragments
To evaluate the sequence diversity within the Pseudomonas species, the P16S rDNA PCR products CD patients and non-IBD patients were cloned and sequenced. DNA was extracted using the QIAquick gel extraction kit [QIAgen] and cloned into pCR2.1 vectors [TA Cloning Kit, Invitrogen] and transformed into One Shot Top 10 chemically competent E.coli [TA Cloning Kit, Invitrogen], as described by the manufacturer's instruction. Blue/ white screening was performed and white colonies were replated for insert confirmation. A minimum of 24 white colonies from each PCR product were selected and grown overnight at 37uC in 96 well agar plates. Plates were then sent to the Australian Genome Research Facility [AGRF, Brisbane, Australia] for culture, plasmid isolation and bi-directional sequencing using the vector-targeted M13 forward and reverse primers. Sequencing reactions were performed with Big Dye version 3.1 on an AB3730x1 sequencing platform. All sequences were analysed using the Sequencher program, (version 4.7) to remove low quality data and vector sequences, and used to prepare consensus sequences of each clone.

Bioinformatics analysis
Consensus sequences were aligned using ClustalW, available from GenomeNet Computation service from Kyoto University Bioinformatics Center (http://align.genome.jp). A PHYLIP generated DNA distance matrix (dnadist) with Jukes-Cantor option was calculated using the Mobyle Pasteur program suite (http:// mobyle.pasteur.fr). The PHYLIP dnadist matrix was used for LIBSHUFF gene library comparison and for the generation of operational taxonomic units using DOTUR. The LIBSHUFF analysis compares two gene libraries and provides a statistical test for the null hypothesis to determine if two libraries are significantly different from one another [32]. In this study webLIBSHUFF version 0.96 was used (Henrikson, JR. 2004 http://libshuff.mib. uga.edu). The DOTUR program (http://www.plantpath.wisc. edu/fac/joh/dotur.html) [33] was used to assess the richness of Pseudomonas collection associated with biopsy specimens obtained from CD and non-IBD patients. In our study we compared the Pseudomonas membership (the list of OTUs in a community) and the membership structure (the abundance of each OTU) at a 97% sequence similarity threshold. The number of sequences in each unique OTU and the proportion of sequences from each patient group in each unique OTU was calculated and analysed.

Phylogenetic analysis
Phylogenetic analyses was conducted using MEGA version 3.1 (Kumar, Tamura, Nei 2004) [37]. Consensus sequences from each individual OTU group were generated using Sequencher program (version 4.7) and sequences for 85 Pseudomonas reference strains for use in the phylogenetic tree were obtained from the Ribosomal Database Project II (release 9.59 at http://rdp.cme.msu.edu). All sequences were adjusted to 617 bp and OTUs consensus sequences were aligned with Pseudomonas type strains using ClustalW from MEGA then a phylogenetic tree was constructed. Trees were produced by the neighbor-joining program with the Kimura correction and Boot strapping 500 times.

Statistical analysis
Fisher's exact test was used to analyse the association between two binary variables, such as the presence or absence of Pseudomonas between CD and non-IBD patient groups. Tests were either one or two tailed and a p value of ,0.05 was considered significant.