Reports

The biosynthetic pathway of coenzyme F430 in methanogenic and methanotrophic archaea

See allHide authors and affiliations

Science  21 Oct 2016:
Vol. 354, Issue 6310, pp. 339-342
DOI: 10.1126/science.aag2947

Enzymes for making (or breaking) methane

The last enzymatic step of microbial methanogenesis, and the first step of microbial methane oxidation, relies on the nickel-containing tetrapyrrole coenzyme F430. The successful metabolic engineering of any organism to enzymatically consume methane thus also needs the appropriate machinery to synthesize this compound. Using comparative genomics, Zheng et al. identified several candidate genes responsible for coenzyme F430 biosynthesis. Cloning and expression of all the subsequent proteins in Escherichia coli confirmed the complete in vitro conversion of sirohydrochlorin into mature F430.

Science, this issue p. 339

Abstract

Methyl-coenzyme M reductase (MCR) is the key enzyme of methanogenesis and anaerobic methane oxidation. The activity of MCR is dependent on the unique nickel-containing tetrapyrrole known as coenzyme F430. We used comparative genomics to identify the coenzyme F430 biosynthesis (cfb) genes and characterized the encoded enzymes from Methanosarcina acetivorans C2A. The pathway involves nickelochelation by a nickel-specific chelatase, followed by amidation to form Ni-sirohydrochlorin a,c-diamide. Next, a primitive homolog of nitrogenase mediates a six-electron reduction and γ-lactamization reaction before a Mur ligase homolog forms the six-membered carbocyclic ring in the final step of the pathway. These data show that coenzyme F430 can be synthesized from sirohydrochlorin using Cfb enzymes produced heterologously in a nonmethanogen host and identify several targets for inhibitors of biological methane formation.

Methanogenic archaea are a major player in the global carbon cycle, producing nearly 1 billion metric tons of methane annually (1, 2). The terminal step of methanogenesis is catalyzed by methyl-coenzyme M reductase (MCR) and involves the conversion of coenzyme B (CoB-SH) and methyl-coenzyme M (MeS-CoM) to the mixed heterodisulfide CoB-S-S-CoM and methane (3) (Fig. 1). MCR uses the unique nickel-containing tetrapyrrole coenzyme F430 to carry out its catalytic function (4) (Fig. 1). Recently, anaerobic methanotrophic archaea (ANME) have been shown to contain a homolog of MCR and catalyze the anaerobic oxidation of methane (AOM) (5). AOM is thought to operate, at least in part, as the reverse of methanogenesis, with MCR catalyzing the critical first step in the pathway, the activation of methane with CoB-S-S-CoM (6).

Fig. 1 The MCR-catalyzed reaction, the structure of coenzyme F430, and the identified coenzyme F430 biosynthesis (cfb) gene cluster from M. acetivorans C2A.

Colored arrows indicate the relative size and orientation of each cfb gene.

There is great interest in strategies to convert methane to liquid fuel or other more easily transported commodity chemicals. The development of a bioconversion process for methane that uses AOM is an attractive solution (7); however, efforts to engineer industrially viable anaerobic methanotrophic strains are hindered by the lack of genetic and biochemical information about the biosynthesis of coenzyme F430 and the formation of holo-MCR.

At present, the genomes of >60 species of methanogenic archaea have been sequenced. We searched these genomes for homologs of known chelatase genes, whose products are responsible for metal ion insertion into tetrapyrrolic cofactors (8). Analysis of the genomic contexts of the chelatase homologs, along with knowledge of the chemistry required for the conversion of known precursors of C2 and C7 methylated tetrapyrroles to coenzyme F430 (9), led to the identification of five genes, conserved in all methanogens, that are potentially involved in coenzyme F430 biosynthesis (Fig. 1 and fig. S1). These genes are also present in the genome of an ANME-2d strain (Candidatus Methanoperedens nitroreducens) (fig. S1) (10).

Included among these genes are homologs of the genes for sirohydrochlorin cobaltochelatase (cbiXS) and cobyrinic acid a,c-diamide synthetase (cbiA2), which are involved in the biosynthesis of cobalamin (11). Also present are homologs of the nitrogenase genes nifD and nifH, which (together with nifK) encode subunits of the two-component metalloenzyme responsible for the adenosine triphosphate (ATP)–dependent reduction of dinitrogen to ammonia (nitrogen fixation) (12). Nitrogenase is structurally and functionally related to the dark-operative protochlorophyllide oxidoreductase (DPOR), which is involved in chlorophyll and bacteriochlorophyll biosynthesis (13). However, methanogens are not photosynthetic microorganisms, and not all methanogens are diazotrophic (i.e., fix nitrogen). The presence of the nifD (methanogenesis marker 13) and nifH homologs in all methanogens was noted previously, and these genes were found to be constitutively expressed and the encoded proteins shown to associate with one another (14, 15). Methanogenesis markers are found in prokaryotic genomes if, and only if, the species is an archaeal methanogen. The fifth gene is homologous to murD, a gene that encodes an ATP-dependent Mur ligase (uridine diphosphate N-acetylmuramoyl-l-alanine:d-glutamate ligase) involved in bacterial cell wall biosynthesis (16). Each of these genes (except for cbiA2) was targeted in a genome-wide transposon mutagenesis experiment in the methanogen Methanococcus maripaludis and were all found to be essential (17).

We cloned the identified genes from Methanosarcina acetivorans C2A and ligated them into expression vectors for heterologous production of the encoded enzymes in Escherichia coli. The enzymes were then purified as N-terminal His6-tagged fusion proteins and systematically tested for activity (supplementary materials and fig. S2).

We tested the “small” sirohydrochlorin cobaltochelatase (CbiXS) homolog, which we designate as CfbA, for nickelochelatase activity with enzymatically prepared sirohydrochlorin and dihydrosirohydrochlorin (precorrin 2) (18). Precorrin 2 is the immediate biosynthetic precursor of sirohydrochlorin and is two-electrons more reduced than the latter (19). Because coenzyme F430 is a highly reduced tetrapyrrole, we reasoned that precorrin 2 might be the substrate of CfbA. However, no nickel chelation activity was observed with precorrin 2 under any of the assay conditions tested. We therefore turned our attention to sirohydrochlorin. We observed that, unlike other divalent transition metal ions (e.g., Fe2+, Co2+, and Zn2+), there is no evidence of rapid, nonenzymatic insertion of Ni2+ into sirohydrochlorin under our assay conditions. However, in the presence of both Ni2+ and CfbA (and only if the His6-tag of CfbA was first removed by thrombin cleavage), the reaction mixture changed from the bright magenta color characteristic of sirohydrochlorin to a deep purple (Fig. 2 and fig. S3). Analysis of the reaction mixtures by reversed-phase high-performance liquid chromatography (HPLC) showed the near-complete conversion of sirohydrochlorin (which has a retention time of 16.1 min) to a new compound that elutes at 20.2 min (Fig. 2). The ultraviolet (UV)–visible absorption properties [wavelengths of maximum light absorption (λmax) = 386 and 590 nm] and mass spectrum of this compound are consistent with those of Ni-sirohydrochlorin [calculated mass-to-charge ratio (m/z) of the protonated molecule ([M + H]+calc) = 919.22 m/z] (20). Thus, CfbA is a sirohydrochlorin nickelochelatase.

Fig. 2 In vitro activity assays of the coenzyme F430 biosynthesis enzymes.

Reversed-phase HPLC traces, liquid chromatography–mass spectrometry (LC-MS) data, and UV-visible spectra for each of the biosynthetic reactions are shown. (A) Sirohydrochlorin prepared from porphobilinogen by using HemC, HemD, SirA, SirC, S-adenosyl-l-methionine, and nicotinamide adenine dinucleotide (phosphate). (B) Ni-sirohydrochlorin prepared by adding CfbA and NiCl2 to the sirohydrochlorin reaction (along with CfbB to alleviate product inhibition). (C) Ni-sirohydrochlorin a,c-diamide prepared by adding CfbB, glutamine, ATP, and an ATP regeneration system (PEP and PK) to the Ni-sirohydrochlorin reaction (along with CfbCD to alleviate product inhibition). (D) 15,173-seco-F430-173-acid prepared by adding CfbCD, sodium dithionite, ATP, and an ATP regeneration system (PEP and PK) to the Ni-sirohydrochlorin a,c-diamide reaction. (E) Coenzyme F430 prepared by adding CfbE, ATP, and an ATP regeneration system (PEP and PK) to the 15,173-seco-F430-173-acid reaction (along with McrD to alleviate product inhibition). Further experimental details can be found in the supplementary materials. ADP, adenosine diphosphate; Pi, inorganic phosphate.

The addition of the cobyrinic acid a,c-diamide synthetase homolog (CfbB) to the reaction mixture enhanced the yield of Ni-sirohydrochlorin (fig. S4). Intermediates in tetrapyrrole biosynthesis often remain tightly bound to their cognate enzyme and are thought to be transferred to the next enzyme in the pathway by substrate channeling (21). Therefore, tetrapyrrole biosynthetic enzymes often exhibit substantial product inhibition, which can be relieved by the addition of the subsequent pathway enzyme in vitro. This suggests that CfbB acts next in the pathway and will accept Ni-sirohydrochlorin as a substrate.

CbiA is a glutamine amidotransferase that catalyzes the ATP-dependent amidation of the a- and c-carboxylic acid moieties of cobyrinic acid in the cobalamin biosynthetic pathway (22). Coenzyme F430 also has amide functional groups at these positions, and we reasoned that CfbB was a Ni-sirohydrochlorin a,c-diamide synthetase. Indeed, addition of both ATP and glutamine to the reaction mixture led to the formation of a new intermediate with a nearly identical UV-visible spectrum to that of Ni-sirohydrochlorin (λmax = 386 and 590 nm)—although its HPLC retention time was shorter by 3 min (17.0 min) and its observed mass was lighter by 1.97 atomic mass units, and with a m/z identical to the [M + H]+calc for Ni-sirohydrochlorin a,c-diamide (Fig. 2 and fig. S5). The yield of Ni-sirohydrochlorin a,c-diamide could again be enhanced by the addition of the subsequent enzyme in the pathway (CfbCD, a complex of CfbC and CfbD), without the reductant required for its activity (vide infra), to alleviate product inhibition (fig. S6). The inclusion of an ATP regeneration system [phosphoenolpyruvate (PEP) and pyruvate kinase (PK)] also helped to drive the CfbB reaction forward (fig. S6).

Two distinctive structural features of coenzyme F430 are the presence of the γ-lactam E ring and the carbocyclic F ring, which form from the c-acetamide and g-propionate side chains of Ni-sirohydrochlorin a,c-diamide, respectively. The high degree of similarity between the UV-visible spectra of Ni-sirohydrochlorin and Ni-sirohydrochlorin a,c-diamide indicates that the product of the CfbB reaction lacks the γ-lactam ring and contains the free amide (Fig. 2).

When the purified CfbCD complex was included in the CfbB reaction, along with the reductant sodium dithionite, ATP, and an ATP regeneration system (all of which were required for activity), the solution changed from the deep purple color characteristic of Ni-sirohydrochlorin a,c-diamide to a pale yellow. Analysis of the reaction mixture by HPLC showed the disappearance of the 17.0-min peak and the formation of a new peak with a retention time of 8.1 min (Fig. 2 and fig. S7). The UV-visible and mass spectra of this new intermediate are indistinguishable from those of the only previously identified intermediate that is unique to the coenzyme F430 biosynthetic pathway, 15,173-seco-F430-173-acid ([M]+calc = 923.30 m/z) (23) (Fig. 2). This intermediate was identical in structure to coenzyme F430, except for the presence of the g-propionate side chain instead of the F ring. Thus, CfbCD effects both the six-electron reduction of the tetrahydroporphyrin ring system of Ni-sirohydrochlorin a,c-diamide and the γ-lactamization of its c-acetamide side chain to form the E ring.

As noted above, CfbC is homologous to the Fe protein (NifH) and CfbD to the NifD subunit of the MoFe protein (NifDK) of nitrogenase, which catalyzes an eight-electron reduction of N2 to NH3 and two protons to H2 (12). The nifD and nifK genes are proposed to have arisen from the paralogous gene duplication and divergence of an ancient shared precursor (24). The nitrogenase homolog DPOR is an analogous two-component system (BchH and BchNB) with similar structural topology and catalyzes a two-electron reduction of the C17=C18 double bond of protochlorophyllide to form chlorophyllide a in the chlorophyll biosynthetic pathway (13). Unlike nitrogenase and DPOR, the Ni-sirohydrochlorin a,c-diamide reductive cyclase contains a homomeric MoFe protein homolog (CfbD) and is thus representative of an early-branching lineage of this enzyme family. A study of the molecular phylogeny of nitrogenase homologs placed the ancestral cfbC and cfbD genes in the last common ancestor of modern organisms and positioned them basal to the emergence of the groups involved in nitrogen fixation and the biosynthesis of photosynthetic pigments (14).

The last enzyme encoded by the cfb cluster, CfbE, is homologous to an ATP-dependent Mur ligase. Mur ligases use ATP to activate a carboxylic acid group as an acyl-phosphate for nonribosomal peptide bond formation during the biosynthesis of peptidoglycan (25). We reasoned that CfbE could use similar chemistry to activate the g-propionate side chain for intramolecular C–C bond formation to produce the carbocyclic F ring and thus function as a coenzyme F430 synthetase. As expected, addition of CfbE to reaction mixtures containing 15,173-seco-F430-173-acid resulted in the production of a new compound in low yield, which had an identical HPLC retention time (7.6 min), UV-visible spectrum, and isotopic mass distribution to authentic coenzyme F430 (fig. S8).

We hypothesized that the low yield of coenzyme F430 was due to product inhibition of the CfbE reaction. The mcr gene cluster encoding the α, β, and γ subunits of MCR, which has been identified previously and is distinct from the cfb cluster, contains two genes, mcrC and mcrD, of unknown function (26). Recently, McrC was identified as a component of a large reductase complex capable of reducing coenzyme F430 to the Ni1+ form, and thus it may play a role in MCR activation (27). McrD has been shown to physically interact with MCR through coprecipitation experiments, though it is not required for in vitro MCR activity (28). We postulated that McrD may function as a chaperone protein that could bind coenzyme F430 and deliver it to apo-MCR. We therefore cloned mcrD from M. acetivorans C2A and expressed and purified the encoded protein to determine whether it was capable of accepting coenzyme F430 from CfbE and alleviating the observed inhibition. Consistent with this expectation, nearly full conversion of 15,173-seco-F430-173-acid to coenzyme F430 was observed when McrD was included in the reaction mixtures (Fig. 2 and figs. S9 and S10).

Each of the identified coenzyme F430 biosynthetic enzymes represents a new target for inhibitors of methanogenesis. The data show that these enzymes are sufficient for the synthesis of coenzyme F430 from the common tetrapyrrolic intermediate sirohydrochlorin and can be produced in an active form in E. coli. Furthermore, if McrD is confirmed as a coenzyme F430–binding protein that chaperones the coenzyme to MCR, this protein will also be required for the heterologous production of holo-MCR. Taken together, these findings set the stage for metabolic engineering efforts using MCR for anaerobic methane conversion.

Supplementary Materials

www.sciencemag.org/content/354/6310/339/suppl/DC1

Materials and Methods

Figs. S1 to S10

Table S1

References and Notes

Acknowledgments: This work was supported by the U.S. Department of Energy Advanced Research Projects Agency–Energy (ARPA-E) (grants DE-AR0000428 and DE-AR0000433). S.O.M., P.D.N., and K.Z. are inventors on patent application 62332658 submitted by Auburn University, which covers the synthesis of coenzyme F430 using the cfb gene products for the production of holo MCR. The data reported in this paper are available in the supplementary materials. S.O.M. provided the scientific direction and overall experimental design for the studies. K.Z. cloned the sirC gene from M. acetivorans, expressed and purified the coenzyme F430 biosynthetic enzymes, and performed the in vitro enzyme assays. P.D.N. cloned the mcrD gene from M. acetivorans, expressed and purified the encoded enzyme, and aided in the initial HPLC and LC-MS experiments. V.L.O. cloned the hemC and hemD genes from E. coli, expressed and purified the encoded enzymes, and aided in the initial synthesis of sirohydrochlorin. X.-p.Y. cloned the sirA and cfbA, -B, -C, -D, and -E genes; developed the initial purification procedures for the encoded enzymes; and confirmed the activity of SirA. S.O.M., K.Z., P.D.N., and V.L.O. wrote and edited the manuscript.
View Abstract

Navigate This Article