Syntrophomonas wolfei is a specialist, evolutionarily adapted for syntrophic growth with methanogens and other hydrogen- and/or formate-using microorganisms. This slow-growing anaerobe has three putative ribosome RNA operons, each of which has 16S rRNA and 23S rRNA genes of different length and multiple 5S rRNA genes. The genome also contains 10 RNA-directed, DNA polymerase genes. Genomic analysis shows that S. wolfei relies solely on the reduction of protons, bicarbonate or unsaturated fatty acids to re-oxidize reduced cofactors. Syntrophomonas wolfei lacks the genes needed for aerobic or anaerobic respiration and has an exceptionally limited ability to create ion gradients. An ATP synthase and a pyrophosphatase were the only systems detected capable of creating an ion gradient. Multiple homologues for β-oxidation genes were present even though S. wolfei uses a limited range of fatty acids from four to eight carbons in length.Syntrophomonas wolfei, other syntrophic metabolizers with completed genomic sequences, and thermophilic anaerobes known to produce high molar ratios of hydrogen from glucose have genes to produce H2 from NADH by an electron bifurcation mechanism. Comparative genomic analysis also suggests that formate production from NADH may involve electron bifurcation. A membrane-bound, iron–sulfur oxidoreductase found in S. wolfei and Syntrophus aciditrophicus may be uniquely involved in reverse electron transport during syntrophic fatty acid metabolism. The genome sequence of S. wolfei reveals several core reactions that may be characteristic of syntrophic fatty acid metabolism and illustrates how biological systems produce hydrogen from thermodynamically difficult reactions.
Syntrophic interactions among microbial species are essential in anaerobic degradation, and syntrophic fatty and aromatic acid metabolism accounts for much of the carbon flux in methanogenic ecosystems (Schink, 1997). Syntrophy is an essential intermediary step in the conversion of natural polymers such as polysaccharides, proteins, nucleic acids and lipids to CO2 and CH4 (McInerney and Bryant, 1981; Schink, 1997). Syntrophy is usually defined as a thermodynamically based interaction where the degradation of a compound such as a fatty acid (Eq. 1) occurs only when degradation products of the compound, usually hydrogen, formate and acetate, are maintained at very low concentrations by a second microorganism, usually a methanogen (McInerney et al., 1979; 1981; Schink, 1997):
For example, when the H2 partial pressure is kept at 1 Pa by the methanogens, the Gibbs free energy change for Eq. 1 becomes −39.2 kJ mol−1, assuming that the concentrations of the butyrate and acetate are each 0.1 mM. Even under optimal conditions, the Gibbs free energy changes for syntrophic metabolism are close to equilibrium (Schöcke and Schink, 1997; Scholten and Conrad, 2000; Jackson and McInerney, 2002) and this energy must be shared among the syntrophic partners (Schink, 1997). For these reasons, growth rates (< 0.005 h−1) and growth yields (2.6 g dry weight mole−1 of propionate)(McInerney and Bryant, 1981; Scholten and Conrad, 2000) are low, which makes biochemical investigations very difficult. How organisms conserve energy and grow when the thermodynamic driving force is very low is an important physiological question. Many syntrophic associations form highly organized, multicellular structures where the partners are in close physical proximity to each other (Sekiguchi et al., 1999; Boetius et al., 2000; Ishii et al., 2006). Little is known about the molecular mechanisms involved in the formation and maintenance of these catalytic units.
Syntrophomonas wolfei and other members of the Syntrophomonadaceae form a coherent phylogenetic group that is distantly related to other members of the phylum Firmicutes (Fig. S1) (Zhao et al., 1990). Members of Syntrophomonadaceae, like S. wolfei, are metabolic specialists that metabolize fatty acids syntrophically in association with hydrogen/formate-using microorganisms and are unable to use other substrates or electron donor and acceptor combinations (McInerney et al., 2008). Some species are able to grow in pure culture with unsaturated fatty acids such as crotonate (Beaty and McInerney, 1987; Amos and McInerney, 1990; McInerney et al., 2008). Members of Syntrophomonadaceae are also unusual in that they stain Gram-negative and have a complex cell wall ultrastructure although they are members of the Firmicutes (McInerney et al., 1981). Syntrophomonas wolfei serves as the model microorganism for the study of syntrophic fatty acid metabolism. It uses only four to eight-carbon fatty acids and several unsaturated fatty acids in co-culture with hydrogen/formate-using microorganisms (McInerney et al., 1979; 1981) and grows in pure culture with unsaturated fatty acids such as crotonate (Beaty and McInerney, 1987; Amos and McInerney, 1990). Syntrophomonas wolfei metabolizes fatty acids by the β-oxidation pathway, forms most of its ATP by substrate-level phosphorylation (SLP), and has menaquinone and a c-type cytochrome (Wofford et al., 1986; McInerney and Wofford, 1992; Wallrabenstein and Schink, 1994).
A defining feature of syntrophic metabolism is the need for reverse electron transport (McInerney et al., 2009). Hydrogen (E′ of −261 mV at 1 Pa H2) and formate (E′ of −258 mV at 1 µM formate) production from electrons generated in the oxidation of acyl-CoA intermediates to their respective enoyl-CoA intermediates (E′ of −10 mV) (Sato et al., 1999) is energetically unfavourable (ΔE′ of ∼ −250 mV) and energy input is required by a process called reverse electron transport. Syntrophus aciditrophicus, which syntrophically degrades fatty and aromatic acids, contains genes for a unique Rnf-type ion-translocating, electron transfer complex and a membrane-bound iron–sulfur oxidoreductase that may function in reverse electron transport (McInerney et al., 2007). How other syntrophic metabolizers accomplish reverse electron transport is not clear although an NADH : acceptor oxidoreductase has been partially purified from syntrophically grown S. wolfei cells that may be involved in reverse electron transport (Müller et al., 2009).
The S. wolfei genome sequence delineates a core set of reactions characteristic of syntrophic fatty acid metabolism and illuminates poorly understood metabolic processes critical for H2 and formate production and energy conservation from thermodynamically challenging substrates.
Results and discussion
General features of the genome
Syntrophomonas wolfei has one circular chromosome consisting of 2.94 mega base pairs (MB) (Table 1; Fig. S2). The genome size is comparable to the genomes of syntrophic specialists such as Pelotomaculum thermopropionicum (3.0 MB) and S. aciditrophicus (3.2 MB) and to metabolic specialists in the class Clostridia (Table S1). The chromosome has a well-defined G + C skew with inflection points at the origin of replication and termination site (see Fig. S2). Approximately 76.7% of the genes are transcribed from the leading strand as is typical of members of the phylum Firmicutes (Rocha, 2002). Eighty-five per cent of the chromosome is predicted to be coding sequence with 2677 genes of which 2574 are open reading frames (ORFs) predicted to code for proteins and 103 code for RNA genes (Table 1). About 56.3% of the ORFs have functional assignments, slightly lower than most of the genomes in Table S1 and 37.5% of the ORFs lack functional assignment but have similarity to sequences in databases. Interestingly, only a small percentage of ORF have no similarity with sequences in databases even though S. wolfei is only distantly related to other sequenced organisms with completed genomic sequences. Sequence analysis suggests that 201 ORFs contain information for translocation to the outside of the cell membrane and 301 ORFs contain two or more predicted transmembrane spanning helices. There is one small region of DNA (from 1 505 786 to 1 521 252 bases) with a much lower % G + C content (26–34%) than the average % G + C of ∼45. The low % G + C region contains genes for 12 hypothetical proteins (Swol_1303 through Swol_1315) and an integrase gene (Swol_1316) and is flanked by portions of a tRNA gene, consistent with its acquisition by horizontal gene transfer.
Table 1. General features of the Syntrophomonas wolfei genome.
2 936 195 bp
2 489 888 bp
G + C content
Genes total number
Protein coding genes
Genes with function prediction
Genes without function prediction
There are three putative rRNA operons; each has an unusual organization in that they contain multiple 5S rRNA genes (see Fig. S3; Table 1 and Table S1). Also, each of the 16S rRNA and 23S rRNA genes has a different length (see Fig. S3). Kosaka and colleagues (2008) showed that 23S rRNA genes of P. thermopropionicum and S. wolfei contain two intervening sequences. Fragmentation of the 23S rRNA was confirmed in P. thermopropionicum (Kosaka et al., 2008). The S. wolfei genome has 46 tRNAs, 68 transposase genes and 10 RNA-directed, DNA polymerases (reverse transcriptases) genes. The S. wolfei genome has 70 pseudogenes, most of which are partial copies of integrases. Syntrophomonas wolfei also has six CRISPR regions, which have recently been shown to provide immunity from foreign nucleic acids (Sorek et al., 2008; Horvath et al., 2009). This is a large number of CRISPR regions for a genome of this size.
Comparison with other genomes
When S. wolfei ORFs were compared pair-wise to individual microbial genomes, best reciprocal blast (BRB) hits revealed the closest associations to members of the phylum Firmicutes (see Fig. S4): Desulfotomaculum reducens MI-1 (1213 BRB hits), Desulfitobacterium hafniense Y51 (1179), P. thermopropionicum SI (1150), and Alkaliphilus metalliredigens QYMF (1145). Approximately 1000 genes are well conserved across the Firmicutes species shown in Fig. S4. The remaining genes (c. 1500) represent a novel complement within the S. wolfei genome.
In another comparison, the best blast hit (BBH) to any microbial gene was determined (Fig. S5) and showed 225, 221, 214 and 179 closest hits to the genomes of P. thermopropionicum SI, D. reducens MI-1, Moorella thermoacetica ATCC 39073, and Heliobacterium modesticaldum Ice1 respectively. Like the three other syntrophic metabolizers with completed genomic sequences (Syntrophobacter fumaroxidans, P. thermopropionicum and S. aciditrophicus), a small number of archaea-related genes were identified by the BBH approach. These included nine BBHs to Methanosarcina mazei Gö1 and six to Methanococcus maripaludis S2, suggesting the possibility of lateral gene transfer events from these potential syntrophic partners. The genes potentially acquired by horizontal gene transfer include those encoding for ABC-type uptake systems, B12-dependent enzymes and cell surface proteins. Genome to genome comparisons at the protein sequence level revealed a few small regions of synteny between the S. wolfei genome and the genomes of Carboxydothermus hydrogenoformans Z-2901, D. reducens MI-1 and P. thermopropionicum SI, but not with the other genomes listed in Table S1 (data not shown). The regions of synteny included genes involved in replication, translation, transcription and biosynthesis. A large region indicative of a core genome characteristic for syntrophy was not detected.
Fatty acid metabolism
A metabolic reconstruction of the core physiological traits of S. wolfei, which combines the known enzymatic machinery and predictions based on the gene inventory, can be seen in Fig. 1. Syntrophomonas wolfei uses only four- to eight-carbon fatty acids and several unsaturated fatty acids in co-culture (McInerney et al., 1979) and in pure culture growth occurs only with unsaturated fatty acids such as crotonate (Beaty et al., 1987; Amos and McInerney, 1990). Enzymatic studies showed that S. wolfei metabolizes fatty acids by the β-oxidation pathway (Wofford et al., 1986; McInerney and Wofford, 1992) and genes for all of the β-oxidation enzymes were detected (Fig. 1; Table S2). Surprisingly, there are multiple homologues for all of the β-oxidation genes despite S. wolfei's extremely restricted substrate range (see Table S2). The S. wolfei genome contains nine acyl-CoA dehydrogenase genes, five enoyl-CoA hydratase genes, six 3-hydroxyacyl-CoA dehydrogenase genes and five acetyl-CoA acetyltransferase genes. Müller and colleagues (2009) detected acyl-CoA dehydrogenases encoded by Swol_1933 and Swol_2052 in syntrophically grown co-cultures. Genome-scale metabolic modelling suggests that metabolically limited organisms, such as S. wolfei, rely on gene duplication of essential functions rather than the possession of alternate pathways to maintain metabolic robustness in case a deleterious mutation occurs (Mahadevan and Lovley, 2008). Alternatively, S. wolfei may express different homologues under different environmental conditions (Strittmatter et al., 2009).
β-Oxidation of butyrate generates two acetyl-CoAs, one of which is used to make ATP by the action of phosphotransacetylase (Swol_0767 gene product) and acetate kinase (Swol_0768 or Swol_1486 gene products) (Fig. 1; Table S2). The second acetyl-CoA is used for the activation of butyrate (Fig. 1). Seven CoA transferase genes are present that could function in fatty acid activation (see Table S1). Two acyl-CoA synthase (AMP-forming) (ligases) (Swol_1144 and Swol_1180) genes arealso present, but their functions are most likely in coenzyme A biosynthesis (Swol_1180) and poly-β-hydroxyalkanoates (PHA) metabolism (Swol_1144). β-Oxidation of butyrate results in the net synthesis of one ATP per butyrate by SLP. Two-thirds of the ATP made by SLP is predicted to be used to drive reverse electron transport as discussed below (Schink, 1997). The remaining one-third of the ATP made by SLP is available for growth.
H2 and formate production
Multiple formate dehydrogenase and hydrogenase genes are present, consistent with the reoxidation of reducing equivalents (NADH and reduced electron transfer flavoprotein) generated during β-oxidation for the production of hydrogen or formate (McInerney et al., 1981; Wallrabenstein and Schink, 1994). Genomic analysis predicts that there are five formate dehydrogenases (Fdh), two of which are externally oriented, and three hydrogenases, one of which is externally oriented (see Table S3; Fig. 1). All three hydrogenases are [FeFe]-type hydrogenases, normally associated with hydrogen production (Vignais et al., 2001). The [NiFe]-type and [NiFeSe]-type hydrogenases, usually associated with H2 oxidation, were not detected. The presence of cytoplasmic as well as externally oriented hydrogenases and formate dehydrogenases may suggest that a proton motive force could be formed by separation of proton-consuming reactions (H2 or formate production internally) and proton-producing reactions (H2 or formate oxidation externally) across the membrane (Odom and Peck, 1981; Heidelberg et al., 2004). However, without NiFe-hydrogenases, it is unclear whether S. wolfei can oxidize H2.
The three cytoplasmic formate dehydrogenases (gene products of Swol_0658, Swol_1028 and Swol_1830) and one of the cytoplasmic hydrogenases (gene products of Swol_1017-1019) appear to be NADH-linked because these genes cluster with those that encode subunits of NADH : quinone oxidoreductase (chains E and F) (see Table S3; Fig. 2). Recently, a trimeric hydrogenase has been purified from Thermotoga maritima that couples the thermodynamically favourable production of H2 from reduced ferredoxin to drive the unfavourable production of H2 from NADH by a process called electron bifurcation (Schut and Adams, 2009). Swol_1017-1019 are homologous to the genes for the bifurcating hydrogenase in T. maritima (Schut and Adams, 2009). Genes homologous to the electron-bifurcating hydrogenase with the same synteny as found in S. wolfei are present in the genomes of the syntrophic metabolizers, P. thermopropionicum and S. fumaroxidans, and anaerobes known to produce high molar ratios of hydrogen (> 3 H2 per glucose) from glucose (Fig. 2). In the syntrophic fatty and aromatic acid degrader, S. aciditrophicus, the gene for the [FeFe]-hydrogenase is adjacent to a gene for NADH : quinone oxidoreductases subunit (Fig. 2). The [FeFe]-hydrogenase and the NADH dehydrogenase chain F in S. wolfei share a high degree of homology at the amino acid level (> 57% and > 46% respectively) with their respective counterparts in other syntrophic metabolizers and anaerobes known to produce high molar ratios of hydrogen from glucose even though these organisms are phylogenetically very distinct. The association of the gene for the catalytic subunit of formate dehydrogenase with genes for NADH : quinone oxidoreductases subunits (Fig. 2) suggests that electron bifurcation may also be involved in formate production from NADH.
Recently, a NADH : acceptor oxidoreductase was partially purified from cell-free extracts of S. wolfei that contained subunits encoded by Swol_0783, Swol_0785 and Swol_0786, predicted to encode for the NADH-linked formate dehydrogenase, and Swol_1017, Swol_1018 and Swol_1019, predicted to encode for the NADH-linked hydrogenase (Müller et al., 2009). This experimental evidence coupled with the comparative genomic analyses (Fig. 2) argues that the bifurcation mechanism (Schut and Adams, 2009) may be a universal approach for reverse electron transport for syntrophic H2 and formate production from NADH. While the bifurcation mechanism seems quite plausible, it is unclear how S. wolfei makes reduced ferredoxin needed to drive the bifurcation mechanism. The S. wolfei genome lacks genes for the ion-translocating Rnf complex, which could use the ion gradient to drive the unfavourable reduction of ferredoxin by NADH and has been implicated in reverse electron transport in S. aciditrophicus (McInerney et al., 2007).
Alternatively, S. wolfei could use the proton motive force to drive the unfavourable production of H2 or formate from reduced menaquinone. Such a mechanism is supported by experimental evidence for the presence of menaquinone and an externally oriented hydrogenase (Wallrabenstein and Schink, 1994). Hydrogen (E′ of −261 mV at 1 Pa H2) and formate (E′ of −258 mV at 1 µM formate) production from reduced menaquinol (E′ of −74 mV) (Thauer et al., 1977) would require an energy expenditure of ∼36 kJ per mole, which could be provided by the hydrolysis of two-thirds of the ATP made by SLP to translocate two protons across the membrane. In support of this hypothesis, S. wolfei contains genes predicted to encode for an externally oriented hydrogenase (Swol_1925-27) and two externally oriented formate dehydrogenases (Swol_0798-800 and Swol_1825-1827). Each gene cluster contains a gene predicted to encode for a membrane-bound b-type cytochrome (Table S3). The gene arrangement suggests that a typical, membrane-bound oxidoreductase complex is formed where the b-type cytochrome accepts electrons from menaquinone, transfers the electrons to an iron–sulfur protein, which then transfers the electrons to the catalytic subunit, either a hydrogenase or a formate dehydrogenase (Jormakka et al., 2002) (Fig. 1).
Syntrophomonas wolfei has genes to encode for heterodisulfide reductase subunits A, B, C adjacent to genes for iron–sulfur proteins and an NAD(P)-binding oxidoreductase (Table S3). In methanogens that lack cytochromes, a soluble Hdr is thought to couple the endergonic reduction of ferredoxin with H2 with the exergonic reduction of CoM-S-S-CoB with H2 by a flavin-based electron bifurcation analogous to that for NADH-linked hydrogenases (Schut and Adams, 2009) and acyl-CoA dehydrogenase–electron transfer flavoprotein (Bcd/EtfAB) complex (Herrmann et al., 2008; Li et al., 2008). The genes for the S. wolfei Hdr are not clustered with those for hydrogenase or formate dehydrogenase, suggesting that Hdr in S. wolfei may have a function other than reverse electron transport to H2 or formate. In addition, reactions to produce reduced ferredoxin or a disulfide required as substrates by Hdr are not known.
A homologue for the electron-conducting pili gene in Geobacter sulfurreducens (Reguera et al., 2005) that could allow direct electron transfer between the syntrophic partners was not detected in the S. wolfei genome. Also absent in the S. wolfei genome are genes for the inner and outer membrane cytochromes believed to be needed to transfer electrons to nanowires.
Electron flow from acyl-CoA dehydrogenase
The above discussion suggests that reverse electron transport in S. wolfei occurs via menaquinone interacting directly with hydrogenase or formate dehydrogenase complexes or by a NADH : acceptor oxidoreductase (Müller et al., 2009) interacting with hydrogenases or formate dehydrogenases that make H2 or formate by a bifurcation mechanism (Fig. 1). The remaining question is how S. wolfei makes reduced menaquinone from electrons derived in β-oxidation. The S. wolfei genome has three gene clusters encoding for electron transfer flavoprotein (ETF) (Fig. 3; Tables S2 and S3). ETF accepts electrons from the acyl-CoA dehydrogenase and typically interacts with membrane ETF : quinone oxidoreductases (Beckmann and Frerman, 1985a,b; Husain and Steenkamp, 1985; Zhang et al., 2004). One set of S. wolfei ETF genes is adjacent to a gene predicted to encode for a FeS-oxidoreductase (Table S3; Fig. 3). This gene arrangement is also observed in the phylogenetically distant S. aciditrophicus (McInerney et al., 2007). The amino acid sequences of the S. aciditrophicus (SYN_02638) and S. wolfei (Swol_0698) gene products share 43.45% sequence identity. Both FeS oxidoreductases contain two 4Fe–4S clusters, two conserved cysteine domains and six predicted membrane-spanning helices, suggesting that each protein is a membrane-bound FeS oxidoreductase. Based on a comparison with known quinone-binding motifs in photosynthetic systems (Fisher and Rich, 2000), the FeS oxidoreductase may have a quinone binding site although the sequences for quinone binding are not highly conserved and difficult to discern from sequence information. Many bacteria contain genes homologous to Swol_0698, but very few have adjacent ETF genes. The organisms with completed genomic sequences that have this type of gene arrangement include: S. wolfei, S. aciditrophicus, C. hydrogenoformans, Geobacter lovleyi, Desulfococcus oleovorans, Desulfotalea psychrophila, G. sulfurreducens, D. hafniense (DCB-2 and Y51), D. reducens, H. modesticaldum, Geobacter uraniumreducens and Geobacter metallireducens. Phylogenetic analysis shows that the S. wolfei and S. aciditrophicus gene products are closely related to each other and more distantly related to the C. hydrogenoformans and D. hafniense (DCB-2 and Y51) gene products (Fig. S6). The high degree of homology between S. wolfei and S. aciditrophicus gene products and the clustering of the FeS oxidoreductase gene with ETF genes suggests a role in syntrophic fatty acid metabolism. The C-terminal portion of the Swol_0698 gene product was detected in a partially purified acyl-CoA dehydrogenase activity obtained from syntrophically grown S. wolfei cells (Müller et al., 2009), consistent with the hypothesis that the Swol_0698 gene product funnels electrons from β-oxidation to membrane redox carriers (McInerney et al., 2007).
A second set of ETF-encoding genes is associated with fix genes (Table S3; Fig. 3). This gene cluster contains genes for the two subunits of the electron transfer flavoprotein (fixA and fixB), a ferredoxin (fixX), and a membrane-bound, ETF-oxidoreductase (fixC). In nitrogen-fixing bacteria, Fix catalyses reverse electron transport from ETF to ferredoxin (Earl et al., 1987; Weidenhaupt et al., 1996; Edgren and Nordlund, 2004; Sperotto et al., 2004). The Fix system has also been implicated in anaerobic carnitine reduction in Escherichia coli where the caiTABCDE operon is found adjacent to the fixABCX (Eichler et al., 1995; Walt and Kahn, 2002). Syntrophomonas wolfei does not fix nitrogen or reduce carnitine, and genes required for these activities were not detected in the genome. The location of the fix genes with those for acyl-CoA dehydrogenase, enoyl-CoA hydratase and a CoA transferase implicates their role in electron transfer during β-oxidation. The Fix system could be used to supply reduced ferredoxin needed for the unfavourable synthesis of pyruvate from acetyl-CoA or for hydrogen and formate production by the bifurcation mechanism.
The third set of ETF genes (Swol_0266-268) is adjacent to a gene for an acyl-CoA dehydrogenase (bcd/etfAB) (Table S2; Fig. 3). None of the gene products are predicted to be membrane-bound. It is possible that the three gene products form a soluble reverse electron transport complex analogous to that in Clostridium kluyveri (Herrmann et al., 2008; Li et al., 2008). In C. kluyveri, a soluble Bcd/EtfAB complex couples the energetically favourable reduction of crotonyl-CoA to butyryl-CoA by NADH with the unfavourable reduction of ferredoxin (Fd) by NADH through a bifurcation mechanism (Eq. 1).
In S. wolfei, this system would operate in the opposite direction as that shown in Eq. 2 to reduce NAD+ with electrons derived from butyryl-CoA and reduced ferredoxin (Fig. 3). However, the acyl-CoA dehydrogenase partially purified from cell-free extracts of syntrophically grown S. wolfei was not a Bcd/EtfAB complex (Müller et al., 2009). The genes (Swol_1933 and Swol_2052) that encode the acyl-CoA dehydrogenases detected in S. wolfei cell-free extracts (Müller et al., 2009) are not linked to ETF genes.
Respiratory and other primary ion-translocating systems
Genomic analysis indicates that S. wolfei has few options to generate a proton motive force needed for reverse electron transport, nutrient uptake and motility. The key proton translocating genes found in the genome are an ATP synthase and a pyrophosphatase (Table S3). Syntrophomonas wolfei like S. aciditrophicus (McInerney et al., 2007) lacks the genes for aerobic and anaerobic respiration. The cytochrome-encoding genes detected were b-type cytochromes associated with hydrogenases and formate dehydrogenases, and a c-type cytochrome associated with genes predicted to encode for a periplasmic, nitrite reductase (Swol_1521 and Swol_1522). The role of the periplasmic, nitrite reductase may be detoxification (Greene et al., 2003) as S. wolfei does not grow with nitrite (McInerney et al., 1979; 1981). There is a cluster of hypothetical genes, some of which are predicted to encode for membrane proteins and/or redox proteins (Swol_1956 to Swol_1968) (Table S3). The function of these genes in S. wolfei is unknown. Most likely, S. wolfei uses a proton-transporting ATP synthase to hydrolyse ATP made by SLP to generate a proton motive force (Table S3). A gene for a proton-translocating pyrophosphatase (Swol_1064) is present, but a continual source of pyrophosphate is not available because S. wolfei uses a CoA transferase rather than a ligase reaction to activate fatty acids.
Syntrophomonas wolfei grows in a defined medium with crotonate as the energy source and thiamine, lipoic acid, cyanocobalamin and para-aminobenzoic acid as required growth factors (Beaty and McInerney, 1990). Syntrophomonas wolfei has genes for pyruvate synthase (pyruvate : ferredoxin oxidoreductase) and pyruvate carboxylase that convert acetyl-CoA to pyruvate and pyruvate to oxaloacetate respectively (Table S2; Fig. 1). Also present are the genes for pyruvate : phosphate dikinase to convert pyruvate to phosphoenolpyruvate and those to form glucose from phosphoenolpyruvate (Table S2; Fig. 1). Syntrophomonas wolfei uses the oxidative arm of the tricarboxylic acid cycle to synthesize 2-oxoglutarate and succinyl-CoA from acetyl-CoA (Table S2; Fig. 1). This is an unusual approach for an anaerobe as most strict anaerobes lack the genes for 2-oxoglutarate dehydrogenase. The Swol_0375 gene product shares 25% identity with an E-value of 3E-26 to the CKL_0973 gene product, which has been shown to encode for a re-citrate synthase in C. kluyveri (Li et al., 2007). The gene cluster, Swol_0375-378, contains genes for a homocitrate synthase, an aconitase and isocitrate dehydrogenase, similar to arrangement found for the re-citrate synthase gene cluster in C. kluyveri (Li et al., 2007). Genes to encode for si-type citrate synthase, succinate dehydrogenase, malate dehydrogenase, fumarate reductase, the glyoxylate shunt and a complete gamma-aminobutyric acid (GABA) shunt were not detected in the S. wolfei genome.
Assignment of the 1507 functionally described genes into pathways revealed that most pathways needed for biosynthetic self-sufficiency (e.g. biosynthesis of amino acids, purines, pyrimidines, membrane lipids, polysaccharides and some cofactors) are present, but many pathways have missing steps. By manual curation, we were able to detect most of the missing genes involved in amino acid biosynthesis with the exceptions of the genes encoding for threonine-ammonia lyase, methionine synthase, and the arginosuccinate synthase and lyase.
There are 40 ATP-binding cassette (ABC) transporters predicted to be involved in nutrient uptake and antibiotic resistance (see Table S4; Fig. 1) (Ren et al., 2007). There are multiple genes for arsenite resistance, calcium antiporters, multi-drug pumps and uptake systems for potassium, zinc, ferrous iron, magnesium, phosphate and sugars (Table S4).
Syntrophomonas wolfei stains Gram-negatively and has a complex cell wall that suggested the presence of an outer membrane (McInerney et al., 1981). However, phylogenetic analysis placed S. wolfei in the phylum Firmicutes (Zhao et al., 1990), whose members lack lipopolysaccharides. The S. wolfei genome contains genes for peptidoglycan synthesis but lacks those for lipopolysaccharide biosynthesis, consistent with the placement of S. wolfei in the phylum Firmicutes.
Regulation and signal transduction
The S. wolfei genome contains a prototypical bacterial RNA core polymerase (rpoA, rpoB, rpoC) that, along with 17 sigma factors, confers promoter specificity (Table S5). The sigma factors include a general housekeeping 70 factor (rpoD), a heat shock 32 factor (rpoH), a flagella biogenesis 1 factor (fliA) and a 54 factor (rpoN) similar to that used for general nitrogen control in E. coli. The genome contains three sigma 54-interacting transcriptional regulators.
Genes for numerous two-component regulatory systems (19 histidine kinase-type sensor transmitters, 19 response regulatory proteins and 9 receiver-only domain proteins) are present in the genome. Roughly two thirds are genetically linked to a cognate two-component protein member, which suggests an interacting partner in sensory transduction. Compared with other Gram-positive microbes, S. wolfei has a relatively small number of primary transcription factors containing a helix–turn–helix motif (c. 58 genes). Instead, S. wolfei appears to have adopted a minimalist regulatory strategy reliant on a core set of signalling pathways, consistent with its highly specialized metabolism.
An important question is whether S. wolfei communicates with its methanogenic partners to initiate syntrophy. The genetic systems necessary for quorum sensing such as acyl-homoserine lactone production are absent. The genome contains a non-ribosomal peptide synthetase gene (Swol_1084) that could be involved in synthesizing short peptides, which can act as signalling molecules. Alternatively a flagellum recognition system, similar to the system in P. thermopropionicum (Shimoyama et al., 2009), is possible. Syntrophomonas wolfei does produce flagella, but it would be difficult at this point to infer from the genome if the methanogen can recognize the flagellum as a signal to initiate a cooperative relationship.
Syntrophomonas wolfei has evolved multiple systems for adapting to stress. This includes motility, PHA formation and potentially sporulation. Oxidative stress can be dealt with by genes encoding rubrerythrin (Swol_0670) and superoxide dismutase (Swol_1711), although the catalase gene is absent.
Motility and taxis
Motility and the presence of flagella have been observed with laboratory cultures (McInerney et al., 1981). Syntrophomonas wolfei has a set of flagellar structural proteins (M-ring, motor, hook and rod) along with the associated flagellar biogenesis and regulatory genes including a σ 28 and an anti-28 factor (flgM) and E. coli-type master switch proteins, flhCD (see Table S6). Nine methyl-accepting chemotaxis sensory proteins (three soluble and six membrane associated) act to detect as yet unknown attractants and/or repellants for signal transduction. The genome contains three genes for each of the cheB, cheC, cheR and cheW plus four cheA and one cheD gene, suggesting potentially parallel signal transduction pathways with differing outputs. No cheZ or cheV genes were detected.
Synthesis of PHA
Members of the genus Syntrophomonas are among the few microorganisms within the Firmicutes that make PHA (McInerney et al., 1992). Syntrophomonas wolfei is unique in that most microorganisms produce PHA when nutrients other than carbon, usually nitrogen or oxygen, are limiting for growth (Anderson et al., 1990; McInerney et al., 1992), while S. wolfei synthesizes PHA during the exponential stage of growth in pure culture or co-culture apparently in the absence of any nutrient limitation (Amos and McInerney, 1989). The production of PHA may provide a mechanism for S. wolfei to obtain energy when thermodynamic conditions make the degradation of fatty acids unfavourable (Amos and McInerney, 1989; Beaty and McInerney, 1989; McInerney et al., 1992). The genome contains all of the genes necessary to synthesize and metabolize PHA. Interestingly, these genes do not appear to be clustered into functional units on the chromosome. There are two genes for 3-hydroxybutyryl-CoA dehydratases (Swol_0650 and Swol_2566) that are distinguishable from the enoyl-CoA hydratases involved in β-oxidation, three genes for poly-(3)-hydroxyalkanoate synthetases (Swol_1099, Swol_1145 and Swol_1241) and two genes for acetoacetyl-CoA reductases (Swol_0651 and Swol_1910). The presence of two sets of genes for poly-(3)-hydroxyalkanoate synthesis is consistent with previous work that showed that S. wolfei uses two routes for PHA synthesis (Amos and McInerney, 1993), the condensation of two acetyl-CoA molecules to acetoacetyl-CoA and its subsequent reduction to d(-)-3-hydroxybutyryl-CoA, and the conversion of a β-oxidation intermediate to d(-)-3-hydroxybutyryl-CoA without carbon–carbon bond cleavage. Having two routes for PHA production shows that energy storage is a very important process in S. wolfei's lifestyle.
Syntrophomonas wolfei has not been shown to form spores, yet its genome contains a full complement of genes needed for the developmental programs of spore formation and germination (Table S7). Sporulation has been shown to occur for other members of the genus Syntrophomonas (Wu et al., 2006a,b; 2007a; Sousa et al., 2007), particularly under bicarbonate-limited conditions (Wu et al., 2007b). The genome contains 59 genes involved in sporulation, which rivals C. hydrogenoformans for lowest number of genes involved in sporulation, if S. wolfei in fact forms spores.
Genomic analysis shows that S. wolfei specializes in syntrophic fatty acid metabolism in association with methanogens and other hydrogen- and formate-using microorganisms. The high degree of specialization and acquisition of genes from its methanogenic partner supports the idea that syntrophic metabolizers evolved as syntrophic specialists by interacting with niche-associated methanogens (Kosaka et al., 2008). Its very restricted metabolic potential probably reflects the outcome of evolutionary selection for energy conservation systems that are highly efficient at using small amounts of free energy. The inability to respire and the dearth of primary ion pumps makes the generation of ion gradients difficult, but may be advantageous to syntrophic specialist like S. wolfei because there are few routes for proton leakage. This may allow for more effective control of the production and use of the proton motive force.
Genomic comparisons did not reveal a large, contiguous, core set of genes that could be identified as being involved in syntrophy, but did reveal that genes for electron bifurcation for hydrogen (Schut and Adams, 2009) and possibly formate production are present in genomes of syntrophic metabolizers and anaerobes that yield high amounts of H2 from glucose (Fig. 2). A second potential set of core syntrophic reactions may be the reactions involved in electron transfer from acyl-CoA dehydrogenases to membrane complexes. By comparative genomic analyses, we identified a novel, membrane-bound, iron–sulfur oxidoreductase that may funnel electrons from reduced ETF to membrane redox carriers (McInerney et al., 2007). While we have made great progress in delineating the genetic systems involved in syntrophic fatty and aromatic acid metabolism (McInerney et al., 2007), we have very little understanding of what other gene systems may be needed for syntrophy. It is unclear whether cell–cell communication systems, specialized recognition factors and/or attachment systems are needed. However, genomic information provides us with the tools to test these hypotheses experimentally.
Media and growth conditions
Syntrophomonas wolfei ssp. wolfei strain Göttingen DSM 2245B was grown in basal medium (Beaty and McInerney, 1987) without rumen fluid with 20 mM sodium crotonate. Multiple 1 L cultures were used to obtain cells for DNA isolation, which was performed using the CTAB containing method (Wilson, 2001), which can be found at http://www.jgi.doe.gov.
Syntrophomonas wolfei genomic DNA, isolated by standard methods (above), was sequenced at the Joint Genome Institute (JGI) using a combination of 3 kb, 8 kb and 40 kb DNA libraries. All general aspects of library construction and sequencing performed at the JGI can be found at http://www.jgi.doe.gov. The Phred/Phrap/Consed software package (http://www.phrap.com) was used to assemble all three libraries and to assess quality (Ewing and Green, 1998; Ewing et al., 1998; Gordon et al., 1998). Possible mis-assemblies were corrected, and gaps between contigs were closed by editing in Consed, custom primer walks or PCR amplification (Roche Applied Science, Indianapolis, IN). The error rate of completed genome sequence of S. wolfei is less than 1 in 50 000. Pair-wise graphical alignments of whole genome assemblies (e.g. synteny plots) were generated by using the MUMmer system (Delcher et al., 1999a,b). The sequence of S. wolfei can be accessed using the GenBank Accession No. NC_008346.
This work was supported by Department of Energy contracts DE-FG03-96-ER-20212 (to M.J.M.) and DE-FG02-08ER64689 (to R.G. and M.J.M.), and the National Science Foundation Grant NSF EF-0333294 (to R.P.G. and M.J.M.).