Screening of sulfate-reducing bacteria in colonoscopy samples from healthy and colitic human gut mucosa

Authors

  • Vitaly Zinkevich,

    Corresponding author
    1. University of Portsmouth, School of Pharmacy and Biomedical Sciences, St Michael's Building, White Swan Road, Portsmouth PO1 2DT, UK
    Search for more papers by this author
  • Iwona B Beech

    1. University of Portsmouth, School of Pharmacy and Biomedical Sciences, St Michael's Building, White Swan Road, Portsmouth PO1 2DT, UK
    Search for more papers by this author

*Corresponding author. Tel. and Fax: +44 (23) 92 842145 vitaly.zinkevich@port.ac.uk

Abstract

A PCR-based approach combined with microbiological cultivation methods was employed to determine the occurrence of sulfate-reducing bacteria (SRB) in colon biopsy samples from ulcerative colitis patients and from non-colitic controls. The detection of mucosa-associated SRB was carried out by digoxigenin-dUTP-labelled PCR amplification, in liquid Postgate medium B and in a new liquid medium, termed VM medium I. Using Postgate medium B, the growth of SRB was confirmed in 92% of the colitic specimens and in 52% of non-colitic samples. However, PCR analysis and incubation in VM medium I detected SRB in 100% of biopsy material indicating ubiquitous presence of SRB in human colon mucosa.

1Introduction

The inflammatory bowel diseases (IBDs) are characterised as two distinct disorders, Crohn's disease and ulcerative colitis (UC). The latter is an acute and chronic inflammatory disease of the large bowel. Epidemiological data show that UC is most common in the USA, UK, Canada and Scandinavia, where the prevalence is approximately 1 individual in 1000 [1]. UC has also been associated with a rising incidence of colonic cancer [2].

UC and Crohn's are disorders of multifactorial aetiology, i.e. susceptibility genes combine with environmental factors to produce the diseased phenotype [3,4]. Several theories of UC aetiology are currently under consideration, including the reaction to persistent intestinal bacterial infection [5]. A number of studies have suggested that compared with the weak reaction to faecal microflora, an increased systemic and local immune response against the mucosal bacterial flora may play an important role in UC [6].

Several reports suggested the possible involvement of sulfate-reducing bacteria (SRB), a group of phylogenetically diverse anaerobic microorganisms [7,8], in both the initiation and maintenance of UC [9,10]. SRB are ubiquitous in nature, often colonise the animal and human intestine and recently have been implicated in pathological processes [11,12]. SRB comprise many genera and species widely differing in their growth rates and physiological activities, including the ability to reduce sulfate. In contrast to the amount of information on SRB in aquatic and soil habitats, the ecology of intestinal SRB and their metabolic activity remain uncharacterised and only recently efforts were made to apply molecular ecological approaches to screen SRB populations in animal and human specimens [13,14]. Furthermore, compared with the volume of literature on faecal SRB, little is known about the real quantity and composition of mucosa-associated SRB.

Most of the SRB utilise sulfate or other sulfur compounds such as thiosulfate, sulfite and sulfur as terminal electron acceptors. The main product of SRB metabolism, hydrogen sulfide, is toxic to human cells and can damage the epithelial barrier function by impairing the oxidation of n-butyrate, subsequently leading to inflammatory responses characteristic of IBD [15]. It has been documented that the mucus layer is compromised in UC and this in turn could result in a breakdown of the protective layer of the colonic epithelium [16–18]. The persistence of fast-growing, actively metabolising (i.e. H2S-generating) SRB strains associated with mucosal and crypt surfaces could lead to a continuous irritation of the colon. In addition to the effect of sulfide, extracellular polymeric substances produced by SRB are species-specific [19–21] and may contain highly immunogenic components such as O-antigenic portions of lipopolysaccharides [22]. Such antigens may trigger an immune response in genetically predisposed individuals and initiate or contribute to the inflammatory process characteristic of UC [23].

Until recently, the majority of data demonstrating the presence of SRB in UC cases and in healthy individuals were obtained employing traditional microbiological techniques of SRB detection and enumeration, i.e. incubation of faecal samples in nutrient media [10,17]. It is now unequivocally accepted that culture-dependent methods used for the recovery of bacteria from natural habitats are biased, as most microorganisms thriving in natural environments (including the human large intestine) can be cultivated only after their physiological niche is perceived and duplicated experimentally. Molecular biology techniques, which do not require cultivation of bacterial strains, have therefore become increasingly popular, and are now routinely used to detect and characterise bacteria, including SRB, in natural and man-made environments [24–26]. In particular, PCR technology offers a new approach to bacterial detection and quantification [27–29].

Despite the vast amount of literature on detection and enumeration of bacteria in faeces, little is known about the real structure of bacterial communities in the human colon, because many of these bacteria are non-cultivable. Furthermore, relatively few studies investigated microbial diversity in human intestinal mucosa despite the increasing evidence of the importance of both aerobic and anaerobic mucosa-associated microflora, and more specifically their ecology, taxonomy and metabolism, in physiological and pathophysiological processes in the human gastrointestinal tract [6,10,26,30].

To gain a better understanding of the potential role of SRB in the pathogenesis of ulcerative colitis, one of the primary challenges is, undoubtedly, the ability to detect and characterise all SRB present in the human colon, in particular those thriving in mucosa. We have therefore decided to apply a PCR approach to the detection of mucosa-associated SRB. To ensure a broad coverage of SRB genera/strains in the screening process, we have chosen the conserved region of subunit A of the adenosine-5-phosphosulfate (APS) reductase gene as the PCR target. The properties of the APS reductase enzyme characteristic for SRB genera have already been explored in the field of immunoassay-based detection of SRB in environmental samples [31] and recently in molecular ecological analysis of the succession and diversity of SRB in the animal intestine [14]. Also a new liquid medium, referred to as VM medium I, was used for the detection of SRB in the biopsy samples. This medium was developed in our laboratory for the purpose of isolation and quantification of SRB present in biofilms formed on inanimate surfaces in marine and freshwater habitats [32]. It has been found that VM medium I supplemented with lactate and sulfate offered a higher sensitivity of SRB detection (one to two orders of magnitude) in biofilm samples than Postgate medium B (V. Zinkevich, unpublished results).

Most studies of human intestinal SRB that have relied on cultivation-based microbiological analysis of faecal samples reported both a higher frequency and elevated numbers of SRB in specimens from UC cases compared with non-UC patients [9,10,33,34]. The present investigation has used PCR technology and sample cultivation in two different types of liquid nutrient media, to determine whether there was a difference in frequency of mucosa-associated SRB between UC patients and non-UC individuals. No such investigation has previously been reported.

2Materials and methods

2.1Analysis of colonoscopy samples

Specimens of mucosa were taken during colonoscopy from the proximal colon of 13 patients and 61 healthy subjects from the central part of Russia. The biopsy procedure is described elsewhere [35]. Selected individuals were divided into two distinct groups. The first group consisted of eight men and five women ranging in age from 21 to 68 years (mean age was 49 years), suffering from ulcerative colitis. In the second group there were 29 men and 32 women ranging in age from 17 to 78 years (mean age was 26 years) with non-inflammatory conditions. In both groups, none of the subjects had taken antibiotics for at least 6 weeks prior to colonoscopy. Chronic UC was diagnosed in all of the 13 patients from the first group. In six cases, this was clinically, endoscopically and histologically inactive. Seven patients had an active disease with mucosal inflammation.

Following sampling, specimens of mucosa were immediately transferred with a sterile needle from the forceps into an Eppendorf tube containing 0.2 ml of sterile physiological saline solution. The mucosal samples were maintained under an atmosphere of oxygen-free nitrogen gas from the time of the initial collection until use, and then vigorously aspirated into and out of a Pasteur pipette to ensure complete dissociation of bacterial cells from mucosa. The aliquots of mucosal samples were used for inoculating 10-ml vials with liquid Postgate medium B containing lactate and sulfate [7], for PCR reactions and for inoculating 10-ml vials with lactate- and sulfate-supplemented liquid VM medium I.

2.2Bacterial growth conditions

All SRB strains and marine Pseudomonas NCIMB 2021 were grown in VM medium I with the exception of psychrophilic strains, courtesy of Dr Christian Konblauch of the Max Planck Institute for Marine Microbiology in Bremen, Germany, which were cultivated in Widdel medium at +4°C [36,37].

Desulfovibrio indonensis (NCIMB 13468) and Desulfovibrio alaskensis (NCIMB 13491) isolated from an aquatic biofilm and purified in our laboratory [19,38,39], Desulfovibrio vulgaris, Desulfovibrio vietnamensis, Desulfovibrio gigas, Desulfovibrio desulfuricans ATCC 27774, Desulfomicrobium baculatus (New Jersey) [40] and Desulfobulbus propionicus (NCIMB 12907) were cultivated at 37°C. Desulfococcus multivorans (NCIMB 12965), Desulfosporosinus orientis (NCIMB 8445) and Pseudomonas species (NCIMB 2021) were cultivated at 30°C. Desulfacinum infernum (NCIMB 13416) and Desulfotomaculum nigrificans (NCIMB 8351) were cultivated at 55°C. Escherichia coli K12 and Enterococcus faecalis, both grown in selective medium (information can be obtained from the website http://www.dsmz.de), were cultivated at 37°C.

The VM medium I was composed of (g l−1 distilled water): KH2PO4, 0.5; NH4Cl, 1.0; Na2SO4, 4.5; NaCl, 9.0 (for SRB recovery from the human colon) and 25.0 for all other SRB strains used in this study; CaCl2·2H2O, 0.04; MgSO4·7H2O, 0.06; sodium lactate, 6.0; sodium citrate, 0.3; casamino acids, 2.0; tryptone, 2.0; thioglycolic acid, 0.1; FeSO4·7H2O, 0.5; modified Wolfe's mineral elixir [38] and vitamin solution. The pH of the medium was adjusted to 7.5 prior to adding vitamin solution. The combination of vitamins was as follows (mg l−1 distilled water): ascorbic acid, 100; nicotinic acid, 0.5; vitamin B1, 0.6; vitamin B12, 0.05; vitamin B2, 0.2; vitamin B5, 0.6; vitamin B6, 0.6; vitamin H, 0.01. The vitamin solution and the modified Wolfe's mineral elixir were filter sterilised. Culture tubes (10 ml) were filled with the medium (without vitamins), purged with a N2 flux and autoclaved at 121°C for 30 min. Due to the presence of Fe ions, Postgate medium B and VM medium I act as indicator media for SRB. Upon the production of hydrogen sulfide by active SRB cells the development of a black colour, resulting from the presence of anaerobically formed iron sulfide compounds, indicates bacterial growth. The cultures were inspected for the change of colour at regular time intervals over a period of 28 days. Production of hydrogen sulfide was monitored as stated below.

2.3Production of hydrogen sulfide

A 14-day-old SRB consortium isolated from patients and controls, grown in either Postgate medium B or VM medium I at 37°C, served as a source of inoculum (1 ml) for freshly prepared media (10 ml). The transferred cultures were incubated at 37°C for 10 days and the H2S concentration was measured using a H2S meter (Cititox, City Technology Ltd, UK), with a lower detection limit of 1 ppm.

2.4Purification of chromosomal DNA

Chromosomal DNAs of the following SRB strains were used as APS reductase-positive controls: Desulfovibrio indonensis, Desulfovibrio alaskensis, Desulfovibrio vulgaris, Desulfovibrio vietnamensis, Desulfovibrio gigas, Desulfovibrio desulfuricans ATCC 27774, Desulfomicrobium baculatus (New Jersey), Desulfobulbus propionicus, Desulfococcus multivorans, Desulfosporosinus orientis, Desulfacinum infernum, Desulfotomaculum nigrificans, Desulfofrigus fragile, Desulfofrigus oceanense, Desulfotalea psychrophila, Desulfotalea arctica, Desulfofada gelida. As APS reductase-negative controls DNAs of E. coli K12, E. faecalis and marine Pseudomonas were used.

Chromosomal DNAs from all bacterial strains were obtained using the guanidine isothiocyanate method as described elsewhere [41]. Samples of chromosomal DNA from SRB communities isolated from UC patients and non-UC individuals were prepared using the Genomic-tip 20/G (Qiagen) method according to the manufacturer's instructions.

2.5PCR primers

A remarkable degree of homology in the conserved region of the APS reductase subunit A gene of Desulfovibrio vulgaris (mesophilic eubacterium) (GenBank accession number Z69372) and Archaeoglobus fulgidus (thermophilic archaeon) [42] allowed the design of oligonucleotide primers for different SRB strains [43]. The primers dCCAGGGCCTGTCCGCCATCAATAC (forward) and dCCGGGCCGTAACCGTCCTTGAA (reverse), which correspond to nucleotide positions 969–992 and 1603–1624, respectively, of the apr locus of D. vulgaris, were used for the amplification of a 658-bp DNA fragment of a conserved region of subunit A of the APS reductase gene from five SRB genera as listed below.

2.6PCR amplification

Bacterial cells from the biopsy samples were pelleted by centrifugation at 20 300×g for 15 min at +4°C, suspended in 50 μl of deionised H2O and disrupted by five freezing and thawing cycles (in liquid nitrogen and +37°C respectively). Cell debris was removed by centrifugation as described above and an aliquot of each supernatant was used for PCR reactions.

The PCR-digoxigenin (DIG)-dUTP Labelling Mix (Boehringer Mannheim) was used to produce DIG-labelled PCR products. A hot start for 4 min at 95°C was used before adding Taq polymerase (Boehringer Mannheim) for all amplifications to minimise non-specific priming. PCR reactions were carried out in a RoboCycler Gradient 96 Temperature Cycler with Hot Top Assembly (Stratagene). Reactions were performed in 50-μl reaction mixtures, containing 2.5 units of Taq polymerase, 5 μl of 10×concentrated Taq PCR buffer, 0.5 μM of each primer and PCR DIG labeling mix, containing 200 μM each of dATP, dCTP, dGTP, 190 μM of dTTP and 10 μM of DIG-dUTP (Boehringer Mannheim). Initial denaturation of the template DNA was completed at 95°C for 5 min and 35 amplification cycles were performed according to the following scheme: denaturation at 95°C for 1 min; annealing at 62°C for 1 min; elongation at 72°C for 1 min. The final PCR step was at 72°C for 10 min. Ten per cent of the reaction volume was analysed by 1% agarose gel electrophoresis (Sigma, Type II, medium EEO) in TAE buffer (National Diagnostics) [44]. The DIG-labeled DNA marker VIII (Boehringer Mannheim) was used as a standard.

PCR products from an agarose gel were transferred to a positively charged nylon membrane (Bio-Rad) by capillary transfer [44] and crosslinked to the membrane on a UV transilluminator for 3 min. The DIG-labelled PCR products on the membrane were located with an alkaline phosphatase-conjugated anti-digoxigenin antibody. The antibody–antigen complexes were visualised using the chemiluminescent substrate CSPD (Boehringer Mannheim). The blot was covered in plastic wrap and exposed to X-ray film (Kodak BioMax Light-2, Sigma) at room temperature for a few minutes.

Chromosomal DNA of different bacterial strains was PCR amplified with the same primers and the same experimental conditions as described above but in standard PCR reactions, i.e. 200 μM of each dNTP (Boehringer Mannheim) was added to PCR mixtures instead of PCR DIG labelling mix. Appropriate PCR products were purified using the QIAquick PCR purification kit (Qiagen) and cloned according to standard methods [39] in pGEM-T Easy Vector (technical manual, Promega). Recombinant plasmid DNA was purified using the Qiagen plasmid mini kit (Qiagen). Plasmid DNA was sequenced by Cambridge Bioscience (UK). Restriction enzymes and T4 DNA ligase used were obtained from New England Biolabs. E. coli JM 109 (Promega) [45] was used as a host strain for molecular cloning. The E. coli JM 109 strain was grown in LB medium [44] and SOC medium (Promega, technical manual 042) at 37°C. The solid LB medium was supplemented with 100 μg ampicillin ml−1, 100 μg X-Gal ml−1 and 0.5 mM IPTG.

3Results and discussion

3.1Detection of SRB in Postgate medium B

Incubation of colonoscopy samples in the indicator liquid Postgate B medium revealed the presence of SRB in 12 out of the 13 samples from patients with UC (92%) and in 32 out of the 61 samples from healthy volunteers (52%). These results are similar to findings reported elsewhere for faecal specimens [9,33,34]. The ubiquitous presence of SRB in colonoscopy samples from patients with UC as well as high rates of hydrogen sulfide production in faecal samples obtained from some diseased individuals indicate that SRB may play a role in the UC aetiology [9,17,46,47]. Alternatively, the fact that SRB were not detected in all of the mucosal samples recovered from non-UC patients may reflect quantitative differences and the inability to cultivate small numbers of SRB in the chosen indicator medium.

As stated earlier, the abundance and diversity of SRB in environmental samples has usually been analysed applying cultivation techniques [9] and through hybridisation with rRNA-targeted oligonucleotide probes for Gram-negative SRB [48]. A recent study by Hristova et al. [13], based on the use of genus- and subgenus-specific 16S rRNA hybridisation probes, demonstrated that Gram-positive SRB of the genus Desulfotomaculum are readily detected in human and animal intestines. Furthermore, members of this genus were shown to be a predominant SRB group in the neonatal mouse intestine [14]. These results further emphasise that care should be exercised when interpreting data on microbial recovery and species diversity relying solely on the cultivation-based microbiological analysis.

3.2PCR screening and SRB growth in VM medium I

PCR primers designed for the conservative region of the APS reductase subunit A gene were used for PCR with chromosomal DNA from different SRB strains and strains purified from human biopsy samples. PCR products of the expected size (658 bp) were obtained (Figs. 1 and 2). PCR products of novel Desulfovibrio indonensis and D. alaskensis strains were cloned and sequenced (EMBL, accession numbers are AJ271651 and AJ271652 respectively). Comparison of their sequences showed that they exhibited high levels of homology with the sequence of Desulfovibrio vulgaris. Sequences of pairs of D. vulgarisD. indonensis and D. vulgaris–D. alaskensis revealed 89.1% and 88.2% similarity, respectively. The corresponding identities of the amino acid sequences were 93.1% and 92.6%. The main difference between the DNA sequences was in the third letters of the codons. This is a strong indication that the region should be conserved in APS reductases of all SRB strains and that the described primers and PCR conditions can be used as a diagnostic tool for SRB detection.

Figure 1.

Analysis of PCR products. Template DNAs from positive control SRB strains (lanes 1, 2) and from SRB communities isolated from patients with ulcerative colitis (lanes 3–5) and control individuals (lanes 6–8) were PCR amplified. The position of PCR products is indicated by the arrow on the left of the figure. The positions of 1018-bp and 517-bp molecular markers are indicated by the arrows on the right side. Lane 1, Desulfovibrio indonensis; lane 2, Desulfovibrio alaskensis; lanes 3–5, PCR products from UC specimens; lanes 6–8, PCR products from control individuals; lane 9, 1-kb DNA ladder (Gibco BRL).

Figure 2.

Analysis of PCR products. The position of PCR products is indicated by the arrow on the left of the figure. The positions of 1018-bp and 517-bp molecular markers are indicated by the arrows on the right side. PCR amplification of the 658-bp APS reductase fragment from positive control SRB strains: lane 1, Desulfotomaculum nigrificans; lane 2, Desulfofaba gelida; lane 3, Desulfosporosinus orientis; lane 4, Desulfomicrobium baculatus; lane 5, Desulfococcus multivorans; lane 6, Desulfocinum infernum; lane 7, Desulfofrigus fragile; lane 8, Desulfofrigus oceanense; lane 9, Desulfotalea psychrophila; lane 10, Desulfobulbus propionicus. PCR amplification of DNA from negative control bacterial strains did not yield APS reductase amplicons: lane 11, E. coli K12; lane 12, E. faecalis; lane 13, Pseudomonas NCIMB 2021. Lane 14, 1-kb DNA ladder (Gibco BRL) as size marker. Lane 15, Desulfovibrio desulfuricans ATCC 27774.

PCR analysis with DNA samples of Desulfovibrio indonensis, Desulfovibrio alaskensis, Desulfovibrio vulgaris, Desulfovibrio vietnamensis, Desulfovibrio gigas, Desulfovibrio desulfuricans ATCC 27774, Desulfomicrobium baculatus New Jersey, Desulfococcus multivorans, Desulfobulbus propionicus (NCIMB 12907), Desulfofrigus fragile, Desulfofrigus oceanense, Desulfotalea psychrophila, Desulfotalea arctica, Desulfofada gelida, Desulfocinum infernum (NCIMB 13416), Desulfotomaculum nigrificans (NCIMB 8351) and Desulfosporosinus orientis (NCIMB 8445), used as positive controls, yielded APS reductase PCR amplicons of the correct size (Figs. 1 and 2). Data are not presented for DNA from strains Desulfovibrio vulgaris, Desulfovibrio vietnamensis, Desulfovibrio gigas and Desulfotalea arctica. In contrast, for the negative control DNA samples (E. coli K12, E. faecalis and Pseudomonas species NCIMB 2021) no PCR products were detected, thus confirming primer specificity (Fig. 2). The results obtained using positive and negative controls justified the strategy of PCR-based SRB detection using these primers.

Indeed, the results of PCR DIG analysis of colonoscopy material from UC patients and non-UC individuals demonstrated the presence of SRB in all samples, thus contradicting the results based on the detection of SRB in faeces using cultivation methods [9,33,34]. Examples of analysed PCR DIG products are shown in Fig. 3. Products of the expected size were obtained with all samples.

Figure 3.

Analysis of DIG-labeled PCR products. DNA amplified from biopsy samples obtained from patients with ulcerative colitis (A) and controls (B). The position of molecular marker VIII (Boehringer Mannheim) is indicated by arrows, and size is given in bp. A: Lanes 1–8, PCR products from UC samples; lane 9, molecular marker VIII; lanes 10–14, PCR products from UC samples 9R–13R, respectively; lane 15, PCR product from Desulfovibrio indonensis. B: Lanes 1–16, PCR products from healthy control samples; lane 17, molecular marker VIII; lane 18, PCR product from Desulfovibrio indonensis.

Molecular strategies that target conserved regions of genes encoding hydrogenases [49] or dissimilatory sulfite reductase [50] have been applied in environmental studies of SRB populations. However, these approaches have scarcely been used to investigate intestinal SRB communities. The validity of our approach in detecting SRB based on targeting of the APS reductase gene has been recently confirmed in an independent investigation. In the latter study, a different segment (396 bp) of this gene was utilised for PCR-based metabolic molecular ecology studies of the succession of intestinal SRB in mice [14].

The PCR DIG method is highly sensitive and is therefore prone to contamination and false-positive results. To ensure that the PCR DIG results revealing the presence of SRB in all biopsy samples were not due to artefacts, the aliquots of these samples were inoculated into VM medium I. The latter has been shown to be superior to Postgate medium B for culturing SRB [32]. The growth of SRB and the production of H2S were observed in 100% of cultures, thus confirming the results of PCR analyses. Clearly, the use of VM medium I allowed a better recovery of SRB from colonoscopy samples than Postgate medium B. Results of SRB detection based on cultivation methods and PCR are summarised in Table 1. Due to the high number of control samples (64 individuals) only the results from the first 13 control patients are presented in the table as they were representative of the whole group.

Table 1.  Detection of SRB in colonoscopy samples obtained from UC patients (UC) and healthy individuals (H) based on cultivation using Postgate B and VM I media and PCR amplification with primers specific for the aprA gene
 Cultivation in Postgate medium BCultivation in VM medium IPCR DIGH2S (ppm)
UC1+++21
UC2+++52
UC3+++60
UC4+++120
UC5+++89
UC6+++91
UC7++109
UC8+++49
UC9+++202
UC10+++82
UC11+++39
UC12+++67
UC13+++196
H1+++64
H2+++40
H3++72
H4+++81
H5++169
H6++90
H7+++61
H8+++143
H9++56
H10+++23
H11++201
H12+++36
H13++54
H2S levels were measured in SRB consortia recovered from colonoscopy samples using VM I medium. All values are means of three determinations. The standard error of the method was within 5%.

There were notable differences in the growth dynamics of SRB populations. The time required for individual cultures to give a positive reaction in the indicator medium varied from 3 to 28 days, regardless of the origin of mucosal sample. Hence, similar to results recorded for faecal specimens, reported by Pitcher et al. [10], there was no correlation between the presence of rapidly growing SRB populations and the source of biopsy specimens, i.e. samples obtained from either UC or control patients (data not shown). These authors also demonstrated the presence of SRB in all colitic and healthy samples of human faeces using Postgate medium E.

Monitoring of H2S production in VM medium I in mixed SRB cultures recovered from control and UC biopsies revealed that the levels of H2S varied from approximately 20 to 200 ppm within each group (Table 1). These results demonstrated the lack of correlation between the source of SRB population (i.e. non-UC or colitic samples) and H2S levels in culture media. Although our findings are in agreement with reports of other authors [51], who showed no overall difference between faecal sulfide levels in non-drug-treated UC patients and controls, they are at variance with the study of Pitcher et al. [10]. The latter authors suggested that there is a greater net production of H2S from gut bacteria in UC cases than from microorganisms in healthy individuals. As results of in vitro studies on sulfide generation in human faeces cannot be directly compared with data from culturing of mucosa-associated SRB, the role of biotically driven H2S production by the latter type of bacteria in the pathology of UC has yet to be determined.

Our results confirm the difficulties encountered when trying to recover SRB from the human colon using standard microbiological techniques, emphasise the advantage of a molecular approach in detecting SRB populations in the human mucosa and demonstrate for the first time that in a healthy human colon mucosa-associated SRB are ubiquitous.

4Conclusions

The use of a PCR-based approach and a culturing method employing a new liquid VM medium I developed in our laboratory demonstrated the presence of mucosa-associated SRB in all human colonoscopy samples, regardless of whether the specimens originated from UC patients or non-UC individuals. In contrast, Postgate medium B did not facilitate the recovery of SRB from all colonoscopy samples. It also became apparent that the mere presence of SRB in the mucosal samples of UC patients does not indicate that these bacteria are likely to play a role in the pathogenesis of this disease. The ubiquitous occurrence of SRB in colon biopsies from both colitic and control individuals suggests that if SRB are involved in UC aetiology, this is likely due to qualitative and perhaps quantitative differences in structure and/or physiology between bacterial communities found in mucosa and possibly in faeces from healthy and diseased patients. The cultivation-based microbiological analysis of SRB in faeces of UC individuals demonstrated that there was a lack of significant difference in bacterial counts between diseased and control specimens. However, total viable SRB counts were significantly related to the clinical severity grade [10]. Whether a quantitative difference in mucosa-associated SRB exists between healthy and diseased colon and whether this is of relevance for UC aetiology remains to be determined.

Apart from differences in bacterial numbers, significant physiological and phylogenetic differences among strains of SRB recovered from human faeces have also been reported [10]. Studies aimed at elucidating possible dissimilarities between mucosa-associated SRB populations isolated from control and colitic colon biopsy specimens are currently in progress in our laboratory. Recent results (unpublished), based on the analysis of SRB consortia from colonoscopy samples using antibodies raised against extracellular polymers produced by a specific strain of Desulfovibrio, indicate that there are highly significant differences between the type of mucosa-associated SRB detected in colitic and control colons. Antibodies raised against extracellular polymeric substances from another SRB species of the same genus did not show any correlation.

These pilot data indicate that the key to understanding the role of SRB in UC may indeed lie in the in-depth study of differences in the SRB community structure between colitic and non-UC individuals. The use of molecular ecological approaches that target specific metabolic genes such as APS reductase or sulfide reductase genes coupled with the application of SRB-specific 16S rDNA probes would prove invaluable in such an investigation.

Acknowledgements

The authors would like to thank the University of Portsmouth for the financial support to V.Z., Dr O. Plotnikov for his assistance in obtaining samples from the UC patients and non-UC individuals and Ms M. Feio and Dr D. Gorecki for helpful discussions.

Ancillary