Among microbes, assimilatory sulfate reduction is the most widespread and essential biochemical process linked to sulfur and provides a mechanism by which microorganisms can reduce sulfate to elemental sulfur in order to satisfy physiological requirements (Peck, 1961). This process is mediated by phosphoadenosine-5′-phosphosulfate (PAPS) reductases (Figure 1), which are widely distributed among microbes and other living organisms except animals (for example 2924 hits for annotated nucleotide sequences and 10,402 for protein sequences were found in the NCBI database). Assimilatory sulfate reduction can also be performed by other enzymatic pathways (Vaneldere et al., 1991; Seitz and Leadbetter, 1995) and likely represents a crucial step in sulfur cycling as it allows for significant disposal of sulfate and conversion to necessary organic sulfated compounds.

While the process of assimilatory sulfate reduction is widespread among microbes, only restricted microbial groups are capable of dissimilatory sulfate reduction (Figure 2). The sulfate-reducing bacteria (SRB) are notable as the only microbes in intestinal ecosystems that rely on inorganic sulfate for conservation of energy. This sulfate respiration pathway has been conserved in five bacterial and two archaeal phyla (Wagner et al., 1998). Most of the SRB belong to the Deltaproteobacteria with more than 25 genera, followed by the Gram-positive SRB within the Clostridia (Desulfotomaculum, Desulfosporosinus, and Desulfosporomusa genera). The SRB use sulfate as a terminal electron acceptor for respiration, with the concomitant production of H₂S (Peck, 1961). In the colon, sulfate reduction is generally associated with dihydrogen oxidation (Carbonero et al., 2012), but the electrons may also be provided from the oxidation of organic compounds, such as lactate (Flint et al., 2009). SRB are ubiquitously present in the human intestinal mucosa (Nava et al., 2012) and have been enumerated by cultivation-dependent methods from human stool in numbers ranging from 10³ to 10¹¹/g (Leclerc et al., 1980; Gibson et al., 1988). In a culture-based study by Gibson et al., the principal SRB were lactate- and hydrogen-utilizing Desulfovibrio spp. (64–81%), acetate-utilizing Desulfobacter spp. (9–16%), propionate- and hydrogen-utilizing Desulfobulbus spp. (5–8%), lactate-utilizing Desulfomonas spp (reclassified within the genus Desulfovibrio) (3–10%), and acetate- and butyrate-utilizing Desulfotomaculum spp. (2%). However, these observations are based on cultivation methods, which underestimate true bacterial diversity. More recently molecular-based techniques have been applied successfully to describing SRB diversity in various environments. SRB have thus far been detected and quantified from stool and colonic mucosa samples (Zinkevich and Beech, 2000; Fite et al., 2004; Nava et al., 2012). The genus Desulfovibrio has generally been found in higher numbers (Fite et al., 2004; Marchesi et al., 2009), with lower abundances of Desulfobacter, Desulfobulbus, and Desulfotomaculum (Nava et al., 2012). The biochemistry of dissimilatory sulfate reduction has been investigated most extensively within species of Desulfovibrio. In the environment, SRB are able to utilize a wide range of substrates as electron donors and acceptors and can even adopt non-sulfidogenic lifestyles (Plugge et al., 2011). For example Desulfobulbus spp. can ferment organic matter under certain conditions (Kuever et al., 2005). It is not known if colonic SRB possess the ability to switch from sulfidogenic to non-sulfidogenic lifestyles in the colon, but an ability to adapt to varying sulfate levels would confer a competitive advantage.

Cysteine degradation to H₂S mediated by L-cysteine desulfhydrase (Figure 3) was described in the 1950s in the Escherichia coli model, a well-known inhabitant of the human gut (Metaxas and Delwiche, 1955). Since that time, L- or D-cysteine desulfhydrase have been described in diverse taxa (Kumagai et al., 1975) including a few intestinal pathogens such as Salmonella thyphimurium (Kredich et al., 1972), Mycobacterium tuberculosis (Wheeler et al., 2005), H. pylori (Kim et al., 2006), and the protozoan Leishmania major (Nowicki et al., 2010). L- or D-cysteine desulfhydrase have also been studied in various oral microorganisms as a source of malodor and abscess formation (Igarashi et al., 2009; Yoshida et al., 2010, 2011). Those genera found in the oral cavity, namely Streptococcus, Prevotella and Fusobacterium, are also common inhabitants of the human colon. A search in the GenBank and EBI databases revealed that cysteine desulfhydrase-encoding genes have been annotated in several common colonic genera such as Clostridium, Enterobacter, Klebsiella, Streptococcus, and, surprisingly, the SRB genus Desulfovibrio. Although desulfhydrase activity has not been quantified in intestinal ecosystems and the prevalence of the gene among intestinal taxa remains undefined, it appears likely that desulfhydrase-harboring microbes are important producers of H₂S in the human colon.

Other microbial enzymes possess cysteine desulfhydrase activity, and, more generally, the ability to convert cysteine or cysteine analogs is widespread. Cysteine desulfidases, that carry out reactions similar to other desulfhydrases, have been described in environmental archaea (Tchong et al., 2005). Beta C-S lyases or cystathionases mediates the cleavage of cystathionine to homocysteine in E. coli (Zdych et al., 1995) and Lactobacillus spp. (De Angelis et al., 2002; Irmler et al., 2008). Another example is the SRB Desulfovibrio desulfuricans, which is able to obtain sulfate from cysteine degradation (Forsberg, 1980). Conversely, the prevalent O-acetyl-L-serine sulfhydrylase (OASS) and homocysteine synthase mediate the microbial synthesis of cysteine from O-acetyl-L-serine and sulfide in bacteria and archaea (Borup and Ferry, 2000; Tai and Cook, 2001).

A recent paper provided evidence that a high proportion of bacterial genomes harbor orthologs of mammalian cystathionine β-synthase (CBS), cystathionine γ-lyase (CSE), and 3-mercaptopyruvate sulfurtransferase (3MST), three enzymes involved in H₂S production (Shatalin et al., 2011). The authors confirmed that pathogenic strains, including E. coli and Staphylococcus aureus, converted cysteine to H₂S through activities of one of those enzymes. Thus, it may be hypothesized that many gut microbes are capable of cysteine conversion to H₂S. Further, the authors demonstrated that sulfide production reduces oxidative stress, thereby increasing antibiotic resistance potential (Shatalin et al., 2011). Indeed, those microbial metabolic pathways are very poorly documented in gut commensals, and characterization of these enzymes and associated genes will be crucial to understanding if microbial-generated sulfide plays a role in shaping the colonic ecosystem.

Only a few genera are able to gain energy through sulfite respiration rather than direct sulfate respiration (Pukall et al., 1999; Soulimane et al., 2011). Among those, the only intestinal resident characterized to date is Bilophila wadsworthia, a close relative of Desulfovibrio spp. (Baron et al., 1989). This unique bacterial species has attracted much attention because of its links with various disease symptoms (Finegold et al., 1992; Arzese et al., 1997; Aucher et al., 1998; Nakayama et al., 2000; Devkota et al., 2012). It reduces sulfite after an initial degradation of taurine and harbors a different dissimilatory sulfite reductase than other SRB (Laue et al., 2001) (Figure 4). A sulfite oxidation metabolic pathway has also been demonstrated in the food-borne human pathogen Campylobacter jejuni (Kelly and Myers, 2005); however, the existence of similar pathways in intestinal commensal microbes has not been described.

Taurine, a sulfated compound secreted by the eukaryotic host that is crucial for chelating bile acids in addition to contributing to osmoregulation, cardiac function, and retinal development is degraded by B. wadsworthia using a taurine:pyruvate aminotransferase to obtain sulfite (Laue et al., 1997; Laue and Cook, 2000) (Figure 4). An intriguing link between dietary fat and the utilization of taurochloric acid as a source of taurine for B. wadsworthia is suggested by two recent studies (Swann et al., 2011; Devkota et al., 2012). Compared to mice fed a low fat or high polyunsaturated fat diet, mice fed a high milk fat diet demonstrated an overrepresentation of B. wadsworthia, which was determined to be mediated by taurine-conjugated bile acids (Devkota et al., 2012). Furthermore, milk fat was essential for colonization by B. wadsworthia in the germ-free mouse colon. Swann et al. (2011) demonstrated that the lack of an established microbiota results in a dominance of taurine-conjugated bile acids over unconjugated or glycine-conjugated bile acids, suggesting that bile acids, whose composition is modulated by the gut microbiota, may serve as signaling molecules in liver and other tissues (Swann et al., 2011). Although taurine has also been demonstrated to be a substrate for SRB in marine microbial mats (Visscher et al., 1999), it is not known if colonic SRB are capable of taurine utilization.

Bile acids can be sulfated by the action of host sulfotransferases, which allow their subsequent detoxification and disposal (Alnouti, 2009). It was demonstrated in a rat model that the opposite conversion, bile acid desulfation, can be mediated by microbial enzymes with arylsulfatase activity (Robben et al., 1988). Arylsulfatases were purified from Clostridium perfringens (Huijghebaert and Eyssen, 1982; Huijghebaert et al., 1982; Robben et al., 1986) inhabiting the rat intestine and are also widespread in several abundant colon genera. Little arylsulfatase activity was reported in human stool (McBain and Macfarlane, 1998), which does not preclude a more extensive activity of microbial arylsulfatase in the colonic ecosystem where significant concentrations of sulfated bile acids are found (Palmer, 1967; Hofmann et al., 2008).

Some colonic microbes have evolved to degrade colonic mucins (Roberton and Stanley, 1982; Derrien et al., 2004, 2008). In particular, microbial glycosulfatases catalyze the release of sulfate from sulfomucins (Corfield et al., 1992; Roberton and Wright, 1997). Glycosulfatase enzymes have been described primarily in the intestinal resident Prevotella strain RS2 and Bacteroides fragilis (Roberton et al., 2000), but also in the stomach resident Helicobacter pylori (Slomiany et al., 1992). A suite of glycosulfatase-like enzymes were also described in the intestinal Peptococcus niger, using less abundant substrates such as estrogen-3-sulfates and phenylsulfates (Vaneldere et al., 1991). Relative abundance and activity of glycosulfatase-harboring bacteria may greatly influence colonic sulfur metabolism; however, little is known about prevalence, abundance, and activity of bacterial glycosulfatases. A search in the GenBank and EBI databases confirmed that glycosulfatases encoding genes have been annotated in important colonic genera such as Prevotella, Bacteroides (Arumugam et al., 2011), and Akkermansia (Belzer and De Vos, 2012).

The two major forms of inflammatory bowel disease (IBD), Crohn's disease (CD), and ulcerative colitis (UC), afflict 0.1–0.5% of individuals in Western countries (Podolsky, 2002; Hanauer, 2006). It is commonly accepted that IBD results from multifactorial interactions among genetic and environmental factors that lead to a dysregulation of the innate immune response to the intestinal microbiota in genetically predisposed individuals (Podolsky, 2002).

Substantial evidence exists for a potential pathogenic role of H₂S in IBD, particularly in UC (Pitcher and Cummings, 1996). Healthy colonic epithelial cells depend on the availability of short-chain fatty acids such as butyrate for nutrition. Butyrate is produced during colonic fermentation and oxidized by colonocytes via the enzyme acyl-CoA dehydrogenase. Because this enzyme is inhibited by H₂S, oxidation of butyrate is impaired by H₂S (Babidge et al., 1998). Patients with UC were shown to have reduced breath excretion of CO₂ after administration of butyrate intrarectally, reflecting reduced colonic butyrate oxidation (Den Hond et al., 1998). UC patients with reduced colonic butyrate oxidation also exhibit increased intestinal permeability (Den Hond et al., 1998). Indeed, as ion absorption, mucin synthesis, membrane lipid synthesis, and detoxification processes in colonocytes depend on the oxidation of butyrate (Roediger et al., 1997), decreases of butyrate oxidation would be expected to compromise epithelial barrier function (Ramakrishna et al., 1991).

Patients with UC have been shown to consume more protein and, thereby, more sulfur amino acids than control subjects (Tragnone et al., 1995). Removing foods such as milk, cheese, and eggs improved symptoms in UC patients (Truelove, 1961). Further, the numbers of SRB and rate of sulfidogenesis were reported greater in UC patients than control cases (Gibson et al., 1991; Pitcher et al., 2000). In a study comparing patients with IBD to healthy individuals or patients with other gastrointestinal symptoms, the prevalence of Desulfovibrio piger detected via PCR was significantly higher in the IBD patients (Loubinoux et al., 2002). Levine et al. (1998) found that production of H₂S from feces of patients with UC was 3–4 times greater than from feces of healthy controls. However, this difference in H₂S production may not have been due to colonization by a greater number of SRB, as qPCR did not show that patients with active UC harbored more SRB in stool or rectal mucosa than healthy controls (Fite et al., 2004). A common treatment regimen for patients with UC may contribute to conflicting results regarding the density of SRB populations in UC patients. Specifically, 5-aminosalicylic acid (5-ASA), an anti-inflammatory medication commonly prescribed for UC, also inhibits SRB growth and production of H₂S (Pitcher et al., 2000; Edmond et al., 2003). For example, stool sulfide concentrations between patients with UC and non-colitic controls did not differ when the use of salicylates in colitic patients was not controlled (Moore et al., 1998); in those patients with UC who were not administered 5-ASA, fecal sulfide concentrations were significantly greater (Pitcher et al., 2000).

Additional evidence for the role of H₂S in UC is the observation that SRB were found in surgically constructed ileo-anal pouches of UC patients but not in pouches of patients with familial adenomatous polyposis (FAP). Furthermore, H₂S production in UC pouches was 10 times greater than that in FAP pouches (Duffy et al., 2002). Finally, in a later study, the severity of pouchitis was positively correlated with fecal concentrations of H₂S (Ohge et al., 2005), possibly reflecting a pathogenic role for this gas. Coffey et al. (2009) proposed that the surgical creation of ileo-anal pouches in UC patients might lead to colonic metaplasia, which, in turn, could result in increased production of sulfomucin and, thus, higher colonization by SRB. The adverse consequences of such colonization would be greater exposure to potentially proinflammatory concentrations of H₂S (Coffey et al., 2009).

Increased activity of mucin sulfatase, an enzyme belonging to the glycosulfatase group, was observed in patients with active UC but not CD (Tsai et al., 1995). In most patients, fluctuations in fecal sulfatase activity corresponded with severity of clinical disease, suggesting that the increased fecal sulfatase activity contributed to perpetuation of the disease. In this scenario, individuals genetically predisposed to a high SRB carriage rate who then experience increased sulfatase activity might be at increased risk of UC due to the increased availability of endogenous sulfate for SRB sulfide production. Similarly, diets high in exogenous sources of sulfate would represent the greatest risk for those genetically predisposed to a higher abundance of SRB.

Recently, a culture-dependent study demonstrated that mucosal biopsies from IBD patients were more often colonized by Fusobacterium spp. than those from matched healthy controls (Strauss et al., 2011). It was further demonstrated that Fusobacterium isolates from IBD patients trigger inflammatory pathways in colonic cells (Dharmani et al., 2011). Fusobacterium spp. produce H₂S through cysteine desulfhydrase activity. It also appears possible that direct consumption of cysteine can reduce in some instances sulfate availability for SRB, thereby possibly explaining the conflicting reports of SRB association with IBD. The recent demonstration of selection for B. wadsworthia in the gut microbiota of mice fed a high-saturated fat diet that was associated with higher occurrence and severity of colitis compared to those fed high-unsaturated fat (Devkota et al., 2012) provides a potential mechanistic model for the development of IBD in susceptible individuals. Notably, this effect was mediated by an increase of specific bile acids, particularly taurocholic acid, a potential source of taurine for B. wadsworthia.

Colorectal cancer is the third most frequent cancer worldwide and the fourth most common cause of cancer death, with responsibility for 4,92,000 deaths annually (Weitz et al., 2005; Ferlay et al., 2010). Genetic and environmental factors play a significant role in the development of colorectal cancer (Kinzler and Vogelstein, 1996; Rhodes and Campbell, 2002; de la Chapelle, 2004). Although etiologically divided into sporadic (90% of the cases), hereditary (5–10%), and IBD-associated (2%), all colorectal cancers show multistep development with several mutations (Kinzler and Vogelstein, 1996; Rhodes and Campbell, 2002; de la Chapelle, 2004). Doll and Peto (1981) estimated that over 90% of gastrointestinal cancers are determined by environmental factors such as diet. It has been suggested that environmental cancer risk is determined by the interaction between diet and colonic microbial metabolism (O'Keefe et al., 2007). Particularly, there is strong epidemiologic and experimental evidence that diets with high animal fat and protein (meat) are associated with increased risk of colorectal cancer (Willett et al., 1990; Sandhu et al., 2001; Norat et al., 2002). As discussed earlier, meat contains relatively high dietary sulfur, which can promote microbial sulfate reduction in the colon.

Kanazawa and colleagues (1996) demonstrated that H₂S concentrations were significantly greater in patients who had previously undergone surgery for sigmoid colon cancer and who later developed new epithelial neoplasia of the colon, compared to individuals of similar age with a healthy colon. The ability of the colon to detoxify H₂S is also reduced in patients with colon cancer (Ramasamy et al., 2006). The association of H₂S with colon cancer is further supported by the finding that H₂S induces colonic mucosal hyperproliferation, with this effect reversed by butyrate (Christl et al., 1996). The mucosal hyperproliferation effect of H₂S may be mediated by mitogen-activated protein kinase (MAPK)-mediated proliferation (Deplancke and Gaskins, 2003). Hydrogen sulfide is also a potent genotoxin that induces direct radical-associated DNA damage (Attene-Ramos et al., 2006, 2007). Colon cancer in UC and, perhaps, sporadic colon cancer in general may reflect genomic instability resulting from exposure to H₂S (Attene-Ramos et al., 2007). As the number of SRB was reported to be not significantly different (Balamurugan et al., 2008) or reduced in colorectal cancer patients when compared to healthy controls (Scanlan et al., 2009), impaired detoxification of H₂S may be critical to the role of this compound in colon cancer.

Increased detection of Fusobacterium spp. in CRC tumors has been reported in two independent studies (Kostic et al., 2011; Castellarin et al., 2012). As described previously, Fusobacterium spp. produce H₂S through cysteine desulfhydrase activity. As H₂S has been demonstrated to be genotoxic, it can be hypothesized that it may be one causative metabolite in the etiology of CRC. It also appears possible that direct bacterial consumption of cysteine would limit sulfate availability for SRB, thereby possibly explaining the conflicting reports of SRB association with CRC.

Irritable bowel syndrome (IBS) is a functional bowel disorder characterized by chronic abdominal pain, bloating, and abnormal bowel habits (Mayer, 2008). Diarrhea or constipation may predominate or these symptoms may alternate. Microbial dysbioses have been described, and as microbial gases excreted in breath are higher in IBS patients, abnormal microbial pathways are a possible cause (King et al., 1998).

Consistent with this possibility, it was demonstrated recently that exogenous H₂S (NaHS) inhibits in vitro motor patterns in the human, rat and mouse colon and jejunum mainly through an action on multiple potassium channels (Gallego et al., 2008). Exogenous H₂S (NaHS) inhibited nociception induced by colorectal distension in both healthy and post-colitic rats (Distrutti et al., 2006) but, conversely, triggered visceral nociceptive behavior when administered intracolonically in the mouse (Matsunami et al., 2009). Recently, a functional dysbiosis was described in constipation-predominant IBS patients with, in particular, a significant increase of SRB and fecal sulfide, indicating a potential role for sulfide in the development of IBS symptoms (Chassard et al., 2012).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.