Pathways and Bioenergetics of Anaerobic Carbon Monoxide Fermentation

In autotrophic, hydrogen-utilizing methanogens, CO is an intermediate of the anabolic reductive acetyl-CoA pathway, a pathway functionally similar to the Wood–Ljungdahl pathway in bacteria (Berg et al., 2010). The archaeal CODH required for growth is a nickel-dependent enzyme, just as observed in bacteria (Hammel et al., 1984). For CO to be utilized as energy source by methanogens, a CODH has to be present that can function in the CO oxidizing direction, and suitable cofactors for electron transfer to the methanogenic metabolism should be available. Thermophilic Methanothermobacter thermoautotrophicus is capable of growing on CO, but slowly; at CO pressures up to 50 kPa at a rate of 1% compared to its growth rate on H2/CO2 (Daniels et al., 1977). The genome of M. thermoautotrophicus (Smith et al., 1997) codes for a single CODH enzyme with ACS as a neighboring gene, which are used for anabolism during growth on hydrogen (Stupperich et al., 1983). Growing M. thermoautotrophicus on CO as sole electron donor, results in methane formation and small amounts of H2, suggesting hydrogen is an intermediate or side product of the metabolism. In hydrogenotrophic methanogens an EcH is present, which is involved in the reduction of ferredoxin with H2, driven by a proton gradient (Thauer et al., 2010). Generation of reduced ferredoxin by CODH allows for the reverse reaction to take place, generating a proton gradient. This additionally results in the formation of hydrogen, which subsequently can be used for reduction of CO2 to methane (Figure 5). Metabolic activity using CO, with H2 as intermediate, is also observed in Methanosarcina barkeri (Fischer and Thauer, 1990) and Methanosarcina thermophila (Terlesky and Ferry, 1988; Zinder and Anguish, 1992). Experiments with cell extracts and washed cells of M.

thermophila show increased hydrogenogenic activity upon exposure to CO (Zinder and Anguish, 1992). When growing M. thermophila on acetate, hydrogen production was detected besides methane formation, suggesting coupling of CO oxidation to proton reduction (Terlesky and Ferry, 1988). M. barkeri also has the capability of oxidizing CO coupled to formation of hydrogen, resulting in the formation of a proton gradient (Bott and Thauer, 1989). Upon addition of methanogenic inhibitors, cell suspensions of M. barkeri were still capable of utilizing CO, producing hydrogen and additionally resulting in formation of ATP (Bott et al., 1986). If a hydrogenogenic metabolism can sustain growth in M. barkeri is unclear. Without methanogenic inhibitors, the formed hydrogen is further oxidized to form methane, allowing for growth (O’Brien et al., 1984). Judging from the carboxydotrophic generation time of M. barkeri (∼65 h) and M. thermoautotrophicus (∼200 h), the methanogenic metabolism on CO is not very efficient. For both strains, hydrogen accumulation is observed during methanogenic carboxydotrophic growth (Daniels et al., 1977; O’Brien et al., 1984). This suggests hydrogenases, required for the formation of methane from the intermediately formed hydrogen, are inhibited by CO. The involved hydrogenases, such as the heterodisulfide reductase-associated hydrogenase and the F420-reducing hydrogenase, are of the [Ni–Fe] type (Thauer et al., 2010) and can be expected to be relatively resistant, but not insensitive, to CO (De Lacey et al., 2007). Additionally, M. thermoautotrophicus contains an iron-dependent methylenetetrahydromethanopterin dehydrogenase, which is involved in the conversion of methenyltetrahydromethanopterin to methylenetetrahydromethanopterin directly using hydrogen as a donor (Zirngibl et al., 1990).

This hydrogenase is found in several hydrogenotrophic methanogens, and is mainly expressed under nickel-deprived conditions (Afting et al., 1998; Afting et al., 2000). The enzyme is susceptible to CO (Lyon et al., 2004) and is therefore a potential target for inhibition during carboxydotrophic growth. Despite the indication that hydrogenases are a limiting factor for carboxydotrophic methanogenesis, inhibition of other enzymes by CO cannot be ruled out as only limited information is available on the mechanisms of CO toxicity in methanogens.

During aceticlastic methanogenesis, CO is an intermediate originating from the splitting of acetyl-CoA by the CODH/ACS complex (Grahame, 2003). Subsequently, CODH is used to further oxidize CO and provide electrons for reduction of the methyl-group to methane. Methanogenic bioreactors fed with acetate accumulated CO to levels up to 0.25 Pa, likely resulting from its role as intermediate in aceticlastic methanogenesis (Hickey and Switzenbaum, 1990). Obligate aceticlastic methanogens such as Methanothrix sp. strain CALS-1, were observed to accumulate CO to low partial pressures while metabolizing acetate (Zinder and Anguish, 1992). Upon addition of low amounts of CO, it was consumed till equilibrium levels of 0.16 Pa. Addition of CO to levels of 2 kPa caused inhibition of growth on acetate. Obligate aceticlastic methanogens contain a low level of hydrogenases and are in general assumed not to utilize these enzymes for their energy metabolism (Deppenmeier et al., 1996; Smith and Ingram-Smith, 2007). This makes it unlikely that inhibition of hydrogenases results in CO toxicity for these methanogens. However, in contrast to EcH containing methanogens, the electrons released from CO oxidation cannot be coupled to proton reduction, requiring other pathways to re-oxidize formed reduction equivalents. The methyl-branch of the reductive acetyl-CoA pathway could theoretically fulfill this role, as the genes are present in the genome of aceticlastic Methanosaeta/Methanothrix species (Zhu et al., 2012). All of these organisms, however, lack the ability to grow on H2/CO2 or formate (Thauer et al., 2008), suggesting no activity of this pathway in the CO2-reducing direction.

The genes of the methyl-branch coded for in these species are phylogenetically similar to genes present in methylotrophic methanogens, which use this branch in the oxidative direction, in order to generate reduction equivalents for biosynthesis (Zhu et al., 2012). This is further supported by C13 labeling studies in Methanosaeta harundinacea which confirms activity in the oxidizing but not in the reducing direction (Zhu et al., 2012). This suggests the methyl-branch of the reductive acetyl-CoA pathway in aceticlastic methanogens is not optimal for re-oxidation of reduced cofactors, as is suggested for hydrogenotrophic methanogens (Figure 5). When exposed to elevated levels of CO this could result in an over-reduced state of the cell, making it difficult for obligate aceticlastic methanogens to utilize it as a substrate.

The only methanogen which appears to deal quite well with CO is Methanosarcina acetivorans, which was initially isolated from marine sediments (Sowers et al., 1984). In addition to methane, acetate and formate were observed to be the main end products from CO (Rother and Metcalf, 2004). Additionally, M. acetivorans was found to produce methylated-thiols from CO (Oelgeschläger and Rother, 2009). The M. acetivorans genome codes for two isoforms of CODH/ACS, Cdh1, and Cdh2, which are both considered to be functional in acetyl-CoA anabolism and catabolism (Matschiavelli et al., 2012). The expression levels of the two isoforms are theorized to be regulated on transcriptional and posttranscriptional level by a CdhA subunit (CdhA3), which is suggested to act as a CO sensor (Matschiavelli et al., 2012). In addition to the CODH/ACS complex, two monofunctional CODH, CooS1F, and CooS2, were found to assist in removal of the CO at high CO partial pressures (Rother et al., 2007). For M. acetivorans no hydrogen formation is observed during growth on CO, which is supported by the fact that it is devoid of any significant hydrogen metabolism (Sowers et al., 1984). Despite the inability of M. acetivorans to grow on H2/CO2, its genome codes for homologs of the methyl-branch of the reductive acetyl-CoA pathway (Galagan et al., 2002). Proteomic analysis shows that these genes are more abundantly expressed during CO-dependent growth when compared to growth on acetate or methanol (Lessner et al., 2006). This suggests that, in contrast to what is proposed for Methanosaeta/Methanotrix species, M. acetivorans uses the methyl-branch of the reductive acetyl-CoA pathway to regenerate its reduction equivalents. However, due to its lack of hydrogenases it is unclear how the organism couples oxidation of CO to reduction of F420, which is required to operate this pathway. M.

acetivorans was shown to express a sodium dependent “RnF-like” complex when metabolizing acetate (Li et al., 2007; Schlegel et al., 2012). It is speculated that this RnF complex couples ferredoxin oxidation to reduction of methanophenazine, subsequently passing on the electrons to the heterodisulfide reductase (Hdr) complex, involved in HS-CoM/HS-CoB regeneration (Li et al., 2007; Figure 6). Proteomic data of M. acetivorans show that in cells grown on CO the F420-oxidizing:Fpo complex is relatively more abundant, which suggests a role in CO metabolism (Lessner et al., 2006). This protein complex possibly operates in combination with the RnF complex, to catalyze ferredoxin-dependent F420 reduction (Lessner et al., 2006; Figure 6). An alternative for coupling ferredoxin oxidation to F420 reduction is via a subunit from the Fpo complex which partly resides in the cytoplasm: FpoF. This subunit was found to catalyze ferredoxin:F420 reduction in EcH knockout mutants of Methanosarcina mazei, whereas knockout mutants of the FpoF subunit did not show this activity (Welte and Deppenmeier, 2011). Despite the lack of genes coding for formate dehydrogenases in the genome of M. activorans (Galagan et al., 2002), formate is produced in addition to acetate and methane during carboxydotrophic growth. Formate is suggested to originate from activity of formyl-methanofuran dehydrogenase and is theorized to act as a redox exhaust of the cell during CO driven growth (Matschiavelli and Rother, 2015).

Methanosarcina acetivorans is not the only archaeon producing acetate and formate from CO, as also the hyperthermophilic, sulfate-reducing Archaeoglobus fulgidus generates these products during carboxydotrophic growth in absence of sulfate. In A. fulgidus, however, both the RnF and EcH complexes are not coded for in the genome (Hocking et al., 2014). A FpoF homolog, FqoF, is encoded for in the genome of A. fulgidus (Brüggemann et al., 2000), which could take part in the coupling of ferredoxin oxidation to F420 reduction. The absence of an EcH complex poses also a question for the redox balance of the organism, as hydrogen cannot be used as a redox exhaust. It is possible that formate is formed in a similar way as theorized for M. acetivorans, but it can also be formed via F420-dependent formate dehydrogenases, allowing favorable redox balance during growth.

Comparing the three main variants of fermentative CO metabolism, it is obvious that similar sets of enzymes and pathways are used. Hydrogenogenic microorganisms seem to be the most efficient CO-utilizers due to their relatively simple and “redox-closed” energy metabolism. This allows the water–gas shift mechanism to operate at a rate independent of the rest of the metabolism, minimizing metabolic stress. To maximize the energetic yield on CO, a methanogenic or acetogenic microorganism would first have to perform hydrogenogenesis, conserving energy via an EcH complex, subsequently gaining energy from the use of hydrogen for methanogenesis or acetogenesis. Hydrogenotrophic methanogens and acetogens like M. thermoacetica seem to make use of this strategy, maximizing the ATP yield per CO utilized. However, despite these optimized energy yields, these organisms have no outstanding growth performance on CO. This is likely related to the CO sensitivity of hydrogenases, essential for this type of metabolism. Acetogens which employ RnF complexes, seem less prone to inhibition by CO, but need to couple CO oxidation stoichiometrically to the Wood–Ljungdahl pathway in order to prevent an over-reduced state of the cell. In relatively fast growing acetogenic bacteria, this is prevented by utilizing alternative pathways to re-oxidize cofactors faster. Loss of “energy-rich” reduction equivalents in these re-oxidizing reactions lowers the overall energy yield of the cell. However, this is expected to allow for better maintenance of redox balance, resulting in reduced stress for the overall metabolism. Still net energy is assumed to be conserved by CO-utilizing acetogens, such as C. ljungdahlii, during acetyl-CoA formation.

This suggests a large and diverse range of products can be formed from carboxydotrophic metabolism, making it interesting for future research and bio-based applications.

MD, Drafted and wrote the manuscript; AS, reviewed and revised the manuscript; DS, reviewed and revised the manuscript. All authors gave approval for publication of the manuscript.

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.

The authors involved were financially supported by an ERC grant (project 323009) and the Gravitation grant (project 024.002.002) of the Netherlands Ministry of Education, Culture and Science and the Netherlands Science Foundation (NWO).