Metabolic processes of Methanococcus maripaludis and potential applications

Methanococcus maripaludis is a rapidly growing, fully sequenced, genetically tractable model organism among hydrogenotrophic methanogens. It has the ability to convert CO2 and H2 into a useful cleaner energy fuel (CH4). In fact, this conversion enhances in the presence of free nitrogen as the sole nitrogen source due to prolonged cell growth. Given the global importance of GHG emissions and climate change, diazotrophy can be attractive for carbon capture and utilization applications from appropriately treated flue gases, where surplus hydrogen is available from renewable electricity sources. In addition, M. maripaludis can be engineered to produce other useful products such as terpenoids, hydrogen, methanol, etc. M. maripaludis with its unique abilities has the potential to be a workhorse like Escherichia coli and S. cerevisiae for fundamental and experimental biotechnology studies. More than 100 experimental studies have explored different specific aspects of the biochemistry and genetics of CO2 and N2 fixation by M. maripaludis. Its genome-scale metabolic model (iMM518) also exists to study genetic perturbations and complex biological interactions. However, a comprehensive review describing its cell structure, metabolic processes, and methanogenesis is still lacking in the literature. This review fills this crucial gap. Specifically, it integrates distributed information from the literature to provide a complete and detailed view for metabolic processes such as acetyl-CoA synthesis, pyruvate synthesis, glycolysis/gluconeogenesis, reductive tricarboxylic acid (RTCA) cycle, non-oxidative pentose phosphate pathway (NOPPP), nitrogen metabolism, amino acid metabolism, and nucleotide biosynthesis. It discusses energy production via methanogenesis and its relation to metabolism. Furthermore, it reviews taxonomy, cell structure, culture/storage conditions, molecular biology tools, genome-scale models, and potential industrial and environmental applications. Through the discussion, it develops new insights and hypotheses from experimental and modeling observations, and identifies opportunities for further research and applications. Electronic supplementary material The online version of this article (doi:10.1186/s12934-016-0500-0) contains supplementary material, which is available to authorized users.

simple substrate such as CO 2 as the sole carbon source, and N 2 as the sole nitrogen source [6]. However, it needs H 2 or formate (HCOOH) for energy [1,7,8]. In other words, given a renewable source of H 2 , M. maripaludis has the potential to capture and convert the main source of global climate concerns, namely the CO 2 emissions, into a useful fuel (methane).
Over the years, M. maripaludis S2 has become a wellstudied model organism in the literature [9,10] with welldeveloped genetic tools. However, in spite of more than 100 publications exploring its genetics and biochemistry, a comprehensive review of its metabolic processes is missing in the literature. This article presents a holistic and integrated view of its metabolic processes, and suggests some potential applications for this promising organism. In this article, we divide the entire metabolism of M. maripaludis into eight subsystems. We discuss six of them in detail in the main text, but defer, for the sake of brevity, the remaining two along with other relevant topics such as taxonomy and cultivation to Additional file 1. For each subsystem, we describe the key steps and their salient features, and identify the existing gaps in the metabolism. Then, we discuss molecular biology tools for manipulating the genome of M. maripaludis, and some limited systems biology work. Finally, we highlight potential applications for M. maripaludis. Figure 1 shows a simplistic view of the M. maripaludis cell. It is a weakly-motile coccus of 0.9-1.3 µm diameter [3]. This non-spore forming mesophile grows best between 20 and 45 °C at a pH ranging from 6.5 to 8.0 [11]. As in Fig. 1, its cell wall is a single, electron-dense, proteinaceous S-layer lacking peptidoglycan molecules. The S-layer proteins and flagellins have been discussed in the literature [12,13]. Its cell wall lyses rapidly in low concentrations of detergents [3] and low-osmolarity buffers [14], which makes the isolation of DNA easier. Despite the differences in the presence of some amino acids, the primary structure of S-layer proteins shows a high degree of identity (38-45 %) with other microbes [15]. Ether lipids recovered from M. maripaludis mainly include glycolipids [14.2 mg ether lipid/g dry cell weight (DCW)] and polar lipids (0.4 mg ether lipid/g DCW) [16]. The motility in methanococci is due to the presence of flagella, however strong attachments by M. maripaludis to various surfaces require both flagella and pili [17]. Although the specific roles of the pili in M. mariplaudis are still unknown [18], if archaeal pili are similar to their bacterial counterparts, then they could be involved in functions related to cell-to-cell twitching, motility, attachment, biofilm formation, etc. Figure 2 provides a comprehensive and consolidated picture of the metabolic system of M. maripaludis. As shown, it has eight major subsystems: methanogenesis, reductive tricarboxylic acid (RTCA) cycle, non-oxidative pentose phosphate pathway (NOPPP), glucose/glycogen metabolism, nitrogen metabolism, amino acid metabolism, and nucleotide metabolism. Methanogenesis, or the reduction of CO 2 to methane, being its only pathway for energy generation, forms the foundation for its survival and growth [19,20]. In other words, both methanogenesis and cell growth compete for the carbon source. The remaining seven subsystems provide the essential precursors for cell growth via two key intermediates: acetyl CoA and pyruvate. We now discuss the first six subsystems in detail.

Methanogenesis
Methanogenesis (Fig. 3) is the biological production of methane via the reduction or disproportionation of relatively simpler carbon substrates such as CO 2 , formate, acetate, and methanol as follows.  [3], formate [21], or electricity [22]. In M. maripaludis, methanogenesis occurs via the reduction of CO 2 with H 2 /formate/electricity, or the disproportionation of formate. Lohner et al. [22] demonstrated H 2 -independent electromethanogenesis from CO 2 in both wild-type M. maripaludis strain S2 and hydrogenase mutant strain MM1284. The mutant strain under the same conditions showed a factor of 10 lower methane production rates as compared to the wild-type strain S2 [22]. However, their attempts to prove biomass growth were inconclusive.
The formate-dependent methanogenesis involves an additional endergonic step where formate is oxidized to CO 2 via formate dehydrogenase with a simultaneous reduction of coenzyme F420 [23]. The reduced coenzyme F420 serves as the electron carrier for two intermediary steps in methanogenesis, but H 2 is probably not an intermediate [21,24]. As shown in Fig. 3, the resulting CO 2 feeds into the first step of methanogenesis.
A recent study [8] showed the effects of H 2 and formate limitation/excess on growth yield and regulation of methanogenesis using a continuous culture of M. maripaludis. They concluded that the growth yield (g DCW/mol CH 4 ) decreased remarkably with excess H 2 or formate. While they speculated energy spilling or dissipation to be a possible cause, the exact cause is still unclear.
While M. maripaludis can also assimilate other carbon substrates, such as acetate and pyruvate, they are not physiologically relevant for methane production [25,26]. No methane production from acetate (17), and extremely low methane from pyruvate [26] (only 1-4 % compared to that for H 2 ) have been reported.

Mechanism
The structures and functions of the cofactors and coenzymes involved in methanogenesis are listed in Table 2. The first step in methanogenesis is the reduction of CO 2 . It involves the simultaneous oxidation of low-potential reduced ferredoxins and capture of CO 2 by methanofuran (MFR) to form formyl-MFR (ΔG 0 = 0 kJ/mol) [27]. These extremely low-potential ferredoxins could come from two pools [28]. One is EchA that not only uses one H 2 , but also consumes proton-motive force (PMF) to generate ferredoxins. This accounts for only 4 % of the reduced ferredoxins as shown in a ∆5H 2 ase mutant [24]. The second and the major pool is Vhu/Hdr bifurcation complex that consumes two H 2 and generates one pair of relatively high potential electrons to reduce CoB-S-S-CoM and another pair of extremely low potential electrons to reduce the ferredoxins. The formyl group from formyl-MFR is then transferred to THMPT (ΔG 0 = −5 kJ/mol) to form formyl-THMPT, and the latter is then dehydrated to methenyl-THMPT (ΔG 0 = −5 kJ/mol) [29]. In the next two steps, the reduced F420 gets oxidized by supplying electrons to reduce methenyl-THMPT to methylene-THMPT (ΔG 0 = +6 kJ/mol) and methylene-THMPT to methyl-THMPT (ΔG 0 = −6 kJ/mol) [30]. These reactions are fully reversible, as evidenced by their near-zero free energy changes. The oxidized F420 is then reduced (ΔG 0 = −11 kJ/mol) in the presence of H 2 [27]. Next, the methyl group from methyl-THMPT is transferred to coenzyme M (HS-CoM) in an exergonic step (ΔG 0 = −30 kJ/mol) coupled with 2Na + translocation by a membrane-bound enzyme complex [31]. This reaction builds up an electrochemical Na + gradient, which drives energy production via ATP synthase [27]. The final step of methanogenesis is the reductive demethylation of methyl-S-CoM to methane and CoM-S-S-CoB (ΔG 0 = −30 kJ/mol). Subsequently, this CoM-S-S-CoB gets reduced with the help of H 2 to form HS-CoM and HS-CoB (ΔG 0 = −39 kJ/mol) [32]. This reduction of CoM-S-S-CoB mediates via an electron bifurcation mechanism [27]. This step along with the earlier step involving the Na + translocation supplies the major energy demand of M. maripaludis.

Hydrogenases
The key to the survival of M. maripaludis on CO 2 is its ability to take up external H 2 and generate electrons from H 2 → 2H + + 2e − with the help of seven hydrogenases (Fig. 3). These are Fru, Frc, Vhu, Vhc, Hmd, EchA, and EchB, which can be categorized in different manners. The first five are cytoplasmic and the last two are membranebound [33]. Fru and Frc use cofactor F420 [34]; Vhu and Vhc use ferredoxin and CoM/CoB [35]; Hmd uses direct H 2 [34]; and EchA and EchB use ferredoxins as electron carriers [24]. M. maripaludis needs four pairs of electrons to reduce one mole of CO 2 to methane. Fru/Frc can supply two pairs, Vhu/Vhc can supply two pairs, Hmd can supply one pair, and EchA/EchB can supply one pair each.
Of the above, Fru/Frc and Vhu/Vhc play a major role in H 2 uptake. Fru and Frc reduce two molecules of coenzyme F420 with the help of two H 2 [34]. One F420 (rd) gets oxidized by reducing methenyl-THMPT and the other by reducing methylene-THMPT. Vhu and Vhc facilitate the flow of electrons from H 2 to heterodisulfide reductase (Hdr) complex [35], which in turn catalyzes the reductions of CoM/CoB and ferredoxins via an electron bifurcation mechanism [32]. These reduced ferredoxins are the major electron suppliers during the first step of methanogenesis.
Hmd uses H 2 to reduce methenyl-THMPT to methylene-THMPT [34] without any carrier. As shown in Fig. 3, Mtd also catalyzes the same reaction, but with the help of reduced F420 as an electron carrier. Hendrickson et al. [34] demonstrated that during the growth on H 2 and CO 2 , Hmd is not essential in the presence of active Fru/Frc, but is essential otherwise. In contrast, the ΔFru, ΔFrc, and ΔHmd mutants grew normally during formate-dependent growth, proving that formate acted as the electron donor.
EchA generates a small portion of low-potential reduced ferredoxins required for the first step of methanogenesis. Its role in M. maripaludis is anaplerotic, because it is required only under certain conditions such as (1) to replenish the intermediates of methanogenesis cycle, and (2) imperfect coupling during electron bifurcation [24]. Lie et al. [24] showed this by eliminating all nonessential pathways of H 2 metabolism and using formate as the sole electron donor. In this case, both Hdr complex and EchA independently provided the electrons for growth.
In contrast, EchB supplies electrons to anabolic oxidoreductases for the synthesis of precursors such as pyruvate and acetyl CoA [33,36]. EchB mutants affect the autotrophic growth severely, but it is unclear how they still survive. When conditions limit growth, anabolic CO 2 fixation is unimportant, but methanogenesis continues. Under such a scenario, EchA is essential, but EchB could be detrimental [24].
During formate-dependent growth, the H 2 required for the essential anaplerotic (EchA) and anabolic (EchB) functions is produced from formate. This H 2 production can occur via two pathways as demonstrated by Lupa et al. [23] in M. maripaludis. One involves Fdh1-Fru/Frc, and the other involves Fdh1-Mtd-Hmd. Of these two, the former seems to be predominant (~90 %), as the deletion of either Fdh1 or Fru/Frc reduced H 2 production rates severely [23].

Methanofuran
Carbon-carrier cofactors involved in the first step of the methanogenic reduction of CO 2 to formyl-methanofuran [124,125] Tetrahydro-methanopterin Carries single carbon fragments between formyl and methyl oxidation levels in methanogens, typically formyl, methenyl, methylene or methyl group [5,125] Coenzyme B Function as electron carrier in metabolism of methanogens [125,126] Coenzyme M Required for methyl-transfer reactions in the metabolism of methanogens., HS-CoB reacts with Methyl-S-CoM to release methane and CoM-S-S-CoB in the final step of methanogenesis [125,127] Coenzyme F420 Major electron transfer currency. Transfer electrons from H 2 to the consecutive intermediates of methane biosynthesis e.g. coenzyme F420 hydrogenase, 5,10-methylene-THMPT reductase and methylene-THMPT dehydrogenase [34,125] Cofactor F430 This cofactor is the prosthetic group of the methyl-CoM reductase which catalyzes the release of methane and CoB-S-S-CoM in the final step of methanogenesis [125,128]

Energy generation/conservation
In most organisms, electron movement along the cell membrane is the key to energy transduction. Substrate oxidation releases electrons that move along the membrane-bound cytochrome carriers and extrudes protons out of the cell to generate a potential gradient. The potential difference drives the protons back into the cell, while at the same time synthesizing ATP from ADP and Pi via ATP synthase [37]. Hydrogenotrophic methanogens such as M. maripaludis lack such an electron transport chain [38]. In place of cytochrome carriers, M. maripaludis uses methyl-THMPT: HS-COM methyltransferase (Mtr), the only membranebound enzyme complex in the core methanogenic pathway, to extrude Na + /H + out of the cell [27,31]. This creates a Na + /H + ion motive force (positive outside), which on their translocations into the cell generate ATP via an A1A0-type ATP synthase [39]. However, a direct experimental evidence specifically for Na + or H + gradient does not exist in the literature. To conserve ATP, M. maripaludis uses reduced ferredoxins as low-potential electron carriers for the highly endergonic reduction of CO 2 to formyl-MFR. As discussed earlier, these ferredoxins are supplied predominantly by the Hdr complex [32] and supplemented by EchA. The genome sequence of M. maripaludis indicates the presence of membrane-bound A1A0-type ATPases (chimeric ATP synthases) instead of the F1F0-type ATPases found in Bacteria and Eukarya [40,41]. The catalytic unit of the A1A0-type ATPase is structurally homologous to the V-type ATPase and functionally homologous to the F1F0-type ATPase [42]. But, the membrane-embedded motors in the A1A0-type ATP synthases are exceptional due to their novel functions and structural features [43].

Acetyl-CoA synthesis
M. maripaludis can synthesize acetyl-CoA from either CO 2 or acetate [7,25]. The CO 2 -based synthesis occurs with the help of carbon monoxide decarbonylase/acetyl-CoA synthase complex (CODH/ACS) [7]. Sequencing studies have confirmed the existence of CODH/ACS in a single cluster (MMP0980-MMP0985) [41]. During the CO 2 -based synthesis, methyl-THMPT, an intermediate of methanogenesis, contributes the methyl carbon of acetyl-CoA, while the CO generated from the reduction of CO 2 in the presence of reduced ferredoxins by CODH, contributes the carboxyl carbon [7]. The acetate-based synthesis is accomplished by AMP-forming acetate CoA ligase (MMP0148, acsA) in M. maripaludis [25]. Shieh et al. [25] showed that M. maripaludis can assimilate up to 60 % of its cellular carbon from exogenous acetate. A sequencing study [41] also showed the presence of ADPforming acetyl-CoA synthetase gene (MMP0253, acd) in M. maripaludis, which catalyzes acetate formation and ATP synthesis from acetyl CoA, ADP and Pi. However, no literature study has experimentally demonstrated the biosynthesis of free acetate by M. maripaludis.

Pyruvate synthesis
Pyruvate is the entry point into glycolysis, citric acid cycle, and amino acid metabolism. Acetyl-CoA is converted to pyruvate through pyruvate:ferredoxin oxidoreductases (PORs) [34,44,45] as follows: This is reversible in that PORs also catalyze pyruvate oxidation to acetyl-CoA in the absence of H 2 [26]. However, pyruvate oxidizes very slowly in M. maripaludis and PORs appear to function mainly in the anabolic directions during growth [44].
The PORs containing five polypeptides in M. maripaludis are encoded by one gene cluster (porABCDEF). Of these, porEF is unique to M. maripaludis, because the N-terminal sequences of the first four polypeptides (porABCD) are similar to those in other Archaea [45]. The importance of porEF in M. maripaludis was highlighted by Lin et al. [46]. They showed that porEF mutants of M. maripaludis JJ grew extremely slowly and pyruvate-dependent methanogenesis was completely inhibited. Interestingly, porF mutant failed to restore growth, but restored methanogenesis to wild-type levels. In contrast, porE mutant restored growth partially, but did not restore methanogenesis. This indicates that porF serves as an electron donor to PORs.
Pyruvate is also a precursor for alanine biosynthesis via alanine dehydrogenase (MMP1513, ald) [47]. The same enzyme catalyzes the reverse reaction also, i.e. alanine to ammonia and pyruvate, in M. maripaludis. In addition, alanine transaminase that catalyzes the conversion of alanine to pyruvate in several organisms including Pyrococcus furiosus, Escherichia coli, Mus musculus, and Homo sapiens, may also exist in M. maripaludis. Our inference is based on a BLASTp search with the protein sequences of M. maripaludis. Our search located proteins with high similarity (e-value = 7e−63) to the alanine transaminase from P. furiosus. M. maripaludis uptakes alanine with the help of alanine permease and alanine racemase. While the former transports both l-alanine and d-alanine into the cell, the latter is essential for converting d-alanine to l-alanine [47][48][49], because alanine dehydrogenase is specific for l-alanine only.

Glycolysis/gluconeogenesis and glycogenolysis/ glyconeogenesis
M. maripaludis does not assimilate carbohydrates such as pentoses and hexoses, as it lacks the required transporters [41]. However, it has all the enzymes and cofactors required for glycolysis/gluconeogenesis and glycogenolysis/glyconeogenesis with some unique features. In fact, studies have shown that methanococci such as M. maripaludis synthesize and store glycogen as a reserve metabolite, and use it for methane generation in the absence of exogenous substrates [50]. The bifunctional activity of ADP-dependent phosphofructokinase (PFK)/glucokinase (GK) has been demonstrated experimentally in both M. jannaschii [51] and M. maripaludis [52]. Castro-Fernandez et al. [52] measured the activities of glucose phosphorylation versus dephosphorylation. They unexpectedly observed that the latter was two-folds more efficient than the former. Based on these observations, they indicated that M. maripaludis can catalyze d-glucose formation, and suggested a possibility of methane production from glycogen or d-glucose during starvation in M. maripaludis.
Given the key roles of glycolysis/gluconeogenesis in M. maripaludis, it is critical to understand their regulation. In general, a pathway can be regulated by (1) substrate availability, (2) up-or down-regulating enzyme activities for rate-limiting steps, (3) allosteric regulation of enzymes, and (4) covalent modifications such as phosphorylations of substrates. Essentially, the enzymes catalyzing the irreversible steps are most suited for regulation [57]. In most Archaea, nonphosphorylating NADP + -dependent glyceraldehyde-3-phosphate (G3P) dehydrogenase (GAPN), phosphorylating glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and glyceraldehyde-3-phosphate ferredoxin oxidoreductase (GAPOR) act as the regulatory points in glycolysis [58][59][60]. The genome sequence of M. maripaludis codes for all three genes, namely GAPN (MMP1487), GAPDH (MMP0325), and GAPOR (MMP0945) [41]. GAPOR catalyzes ferredoxin-dependent G3P oxidation, GAPN catalyzes NADP-dependent G3P oxidation, and GAPDH catalyzes G3P synthesis. Based on the activity, transcriptomic, and flux balance analyses in M. maripaludis, Park et al. [61] showed that GAPOR is a post-transcriptionally regulated enzyme that is completely inhibited by the presence of 1 µM ATP, and (unlike GAPN) is most likely involved only under non-optimal growth conditions. Yu et al. [50] mentioned pH-dependent PFK (optimum pH = 6.0) as an important regulatory enzyme in M. maripaludis. The activation and inhibition of PFK was found to be dependent on the presence/absence of various substrates such as ADP, AMP, Pi, cAMP, and citrate. Yu et al. [50] also reported that full activity of pyruvate kinase, another key enzyme in glycolysis, depended on Mn 2+ . In contrast to Mn 2+ , Fe 2+ showed 70 % activity, and Mg 2+ showed 20 % activity of pyruvate kinase, while Zn 2+ , Cu 2+ , Co 2+ , and Ni 2+ showed zero activity. The activity of phosphoglycerate mutase was unaffected by Mg 2+ and AMP, and depended on the presence of reduced dithiothreitol, cysteine hydrochloride, and glutathione.

Tri-carboxylic acid (TCA) cycle
TCA cycle plays an important role in generating electron carriers such as NADH & FAD for energy production [62]. Most aerobes have an oxidative TCA cycle to oxidize complex carbon molecules, such as sugars, to CO 2 and H 2 O to generate energy [62]. However, most anaerobes  have RTCA cycles to reduce CO 2 and H 2 O to synthesize carbon compounds. Methanogens being anaerobes also have RTCA cycles. Furthermore, their TCA cycles are incomplete, as they lack several steps and enzymes [63]. M. maripaludis in particular lacks phosphoenolpyruvate carboxykinase, citrate synthase, aconitate, and isocitrate dehydrogenase [25,64]. The missing steps in M. maripaludis are shown as dashed lines in Fig. 4. As shown in Fig. 4 Fig. 5. Hendrickson et al. [41] reported the complete genome sequence of M. maripaludis and noted that 2-oxoglutarate oxidoreductase, the last enzyme in the TCA cycle, has four subunits (MMP0003, MMP1315, MMP1316, and MMP1687) that are not contiguous. This is in contrast to PORs that are also oxidoreductases, but have contiguous subunits (MMP1502-MMP1507).
Regulation of TCA cycle in M. maripaludis in particular, and Archaea in general, is poorly understood. However, 2-oxoglutarate plays an important role in nitrogen regulation [65]. In M. maripaludis, NrpR protein represses nitrogen fixation in ammonia-rich conditions by binding to the nif promoters [66]. In the absence of ammonia, 2-oxoglutarate is unable to synthesize glutamate, hence its level increases. High levels of 2-oxoglutarate act as the inducer and prevent binding of NrpR to nif promoters, resulting in the activation of nitrogen fixation and glutamine synthetase to bring down 2-oxoglutarate levels.
The TCA regulation in Methanobacterium thermoautotrophicum, another methanogen with an incomplete reductive cycle, can shed some light on the regulation in M. maripaludis. As reported by Eyzaguirre et al. [30] for M. thermoautotrophicum, M. maripaludis may also exhibit unidirectional synthesis of phosphenolpyruvate via phosphenolpyruvate synthetase (ppsA). The activity of this enzyme may be inhibited by AMP, ADP, and 2-oxoglutarate. Similarly, PYC, the ATP-dependent enzyme responsible for pyruvate carboxylation in M. maripaludis, may exhibit anabolic function as reported by Mukhopadhyay et al. [30] in M. thermoautotrophicum. Furthermore, its activity may depend on biotin, ATP, Mg 2+ (or Mn 2+ , Co 2+ ), pyruvate, and bicarbonates; and it may be inhibited by ADP and 2-oxoglutarate.

Pentose phosphate pathway (PPP)
PPP is essential for the syntheses of nucleotides and nucleic acids in M. maripaludis. Glyceraldehyde-3-phosphate and fructose-6-phosphate synthesized during glycolysis/gluconeogenesis form the feeds to PPP and produce xylulose-5-phosphate and erythrose-4-phosphate (E4P) via transketolase (TKL) in the first step. Yu et al. [50] proposed a NOPPP in M. maripaludis (Fig. 6). They suggested the presence of this pathway based on the zero activities of oxidative enzymes [glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase] and high activities of non-oxidative enzymes [transketolase (MMP1113, MMP1115), transaldolase (MMP1308), ribose-5-phosphate 3-epimerase (MMP1114), and ribulose-5-phosphate isomerase (MMP1189)] [41,50] . Tumbula et al. [67] supported this observation by ruling out oxidative PPP based on the labelling patterns of riboses after supplementing the medium with [2-13 C] acetate. They argued that E4P cannot be the precursor for aromatic amino acids (AroAAs), if NOPPP is its only route. Therefore, they conjectured an alternative route (carboxylation of a triose such as dihydroxyacetone phosphate) for E4P. Porat et al. [68] on the other hand showed that E4P is not a precursor for AroAAs in M. maripaludis. They proposed two alternative routes for the syntheses of AroAAs based on the presence of dehydroquinate dehydratase. The details of these routes are provided in the Additional file 1. NOPPP is mainly regulated by substrate availability [69,70]. However, no such regulation has been shown yet in M. maripaludis.

Nitrogen metabolism
M. maripaludis can utilize three nitrogen sources: ammonia, alanine, and dinitrogen (free N 2 ) with ammonia being the most preferred source for growth [47,71] .  Ammonia assimilation occurs in M. maripaludis via glutamine synthetase (encoded by glnA) which makes glutamine from glutamate and ammonia [72]. Glutamine then serves as the precursor for protein synthesis. Cohen-Kupiec et al. [72] observed that glnA mutants are unable to grow even in the presence of exogenous glutamine and alanine indicating the essentiality of glnA and absence of glutamine transporters.
As discussed before, alanine uptake occurs in M. maripaludis via alanine racemase and alanine permease [47]. Moore et al. [47] confirmed that alanine converts to pyruvate and ammonia by alanine dehydrogenase. Free N 2 fixation or diazotrophy in methanogens including M. maripaludis is well established and extensively reviewed in the literature [73][74][75]. A comparison of four nitrogen-fixing hydrogenotrophic methanococci (M. formicicus, M. maripaludis, M. aeolicus, and M. thermolithotrophicus) is given in Table 3 [71,[76][77][78]. Blank et al. [79] showed diazotrophy in wild-type M. maripaludis and its four mutants using transposon insertion mutagenesis. Kessler et al. [73] characterized the nif gene cluster in M. maripaludis based on a sequence analysis. Six nif genes (nifH, D, K, E, N, and X) and two homologues of bacterial nitrogen sensor-regulator glnB (i & ii) exist between nifH and nifD in a single operon in M. maripaludis (Fig. 7a). Although the conserved order of nif genes resembles that in Bacteria, the presence of  (2.2.1.1, tkl)  single operons and two homologues between nif genes is unique to Archaea. As shown in Fig. 7b, diazotrophy is effected by multiprotein nitrogenase complex comprising an Fe protein and a MoFe protein [80]. In the presence of N 2 , the oxidized Fe protein reduces by taking electrons from reduced ferredoxins. The reduced Fe protein then oxidizes in the presence of ATP and reduces the MoFe protein. The MoFe protein donates the electrons to N 2 and reduces it to ammonia in three successive steps: nitrogen to diamine to hydrazine to two ammonia molecules and one H 2 . These reductive steps require electrons from reduced ferredoxins, and their relatively high-energy demand makes N 2 fixation unfavorable in M. maripaludis. Therefore, the cells are less likely to activate this fixation, when ammonia or alanine is available [48,81].

Regulation
Cohen-Kupiec et al. [72] hypothesized that both nif and glnA are regulated by the same nucleotide binding sequences (Fig. 7c) residing in the nif promoter region of M. maripaludis. Lie et al. [82] confirmed this by isolating NrpR protein. NrpR represses both N 2 fixation and glnA expression by binding to the aforementioned nitrogen operator sequences (nifOR). While NrpR represses nif transcription fully in the presence of ammonia, it represses it partially in the presence of alanine, and represses fully in the presence of free N 2 . As discussed before, NrpR in turn is regulated by 2-oxoglutarate levels in the TCA cycle [65]. As shown in Fig. 7c, the binding of NrpR to nifOR weakens during nitrogen deficiency, allowing 2-oxoglutarate to induce nif transcription.
Experiments have also demonstrated that cell growth with alanine was only marginally lower than with ammonia, while significantly reduced with free nitrogen [48].
glnB proteins (i & ii), encoded within the nif gene cluster as shown in Fig. 7a, play a key role in nitrogen sensing and regulation in M. maripaludis [81]. glnB + strain lost nitrogenase activity within an hour of ammonia addition [83], while glnB mutant did not. This indicates that the glnB proteins switch off the nitrogenase activity in the presence of ammonia. Kessler et al. [71] reported the importance of molybdenum-dependent nitrogenases during nitrogen fixation in M. maripaludis.
The above six subsections along with amino acid and nucleotide metabolisms described in the Additional file 1 complete our picture for the metabolic processes of M. maripaludis. In spite of the extensive literature on this organism, we could identify at least a few gaps in our current understanding, which offer opportunities for further research. While a comprehensive picture for the metabolic network and its individual components is useful for the research community, we need to note that all these subsystems reside in a single cell, interact with each other in a complex manner, and result in various cellular phenotypes. Furthermore, the picture so far has been largely qualitative and experimental. As recent research [42,[84][85][86] has demonstrated the importance of synergizing the experimental with the computational, the qualitative with the quantitative, and the elemental with the systemic, we now complement our picture with a quantitative and systems biology perspective. This is essential for fruitful applications of M. maripaludis. However, very little work exists on the systems biology of M. maripaludis.

Selectable markers
It is difficult to identify antibiotic resistant markers in methanogens due to different ribosome structures and the absence of peptidoglycans in their cell walls. Puromycin resistance in M. maripaludis was reported by transforming it with pKAS100 and pKAS102 plasmids [89]. To aid vector transformation, an optimized polyethylene glycol (PEG) method was proposed [91]. This method increased transformation frequency by an order of four to five (2 × 10 5 transformants/µg of insertion vector) as compared to the natural transformation method. Subsequent methylation of plasmid with PstI methylase increased transformations at least four-folds, thus approaching those obtained for E. coli. Addition of divalent cations inhibit transformation [91]. Neomycin is the second selectable marker reported for M. maripaludis [88], for which aminoglycoside phosphotransferase genes APH3'I and APH3'II were cloned under the control of Methanococcus voltae methyl reductase promoter. 500-1000 µg/ml of neomycin delayed the growth of M. maripaludis, and 1000 µg/ml inhibited it completely. Kanamycin and geneticin are non-inhibitory for M. maripaludis [92,93].

Shuttle vectors
Tumbula et al. [9] constructed a shuttle vector pDLT44 for M. maripaludis JJ using plasmid pURB500 (from M. maripaludis C5) and pMEB.2 (E. coli vector containing a methanococcal puromycin resistance marker). This shuttle vector was found to be stable in E. coli under ampicillin selection. This was the first report of a plasmid replicated independently in a methanogen, which can be manipulated in E. coli. Although pURB500 was originated from a methanococcus, it did not replicate in M. voltae. Another study [94] reported expression shuttle and integrative vectors for M. maripaludis using histone promoter (P hmvA ) and multiple cloning sites from M. voltae for overexpressing ilvBN and ppsA. These expression vectors may be useful for studying the physiology and biochemistry of M. maripaludis. However, their transformation efficiencies vary from one strain to another [10]. For instance, M. maripaludis S2 showed much lower efficiency than M. maripaludis JJ, but it can be improved by manipulating a shuttle vector. Walters et al. [10] showed that a significantly smaller shuttle vector pAW42 was sufficient to maintain in M. maripaludis S2 and provided 7000-fold increase in transformation efficiency for pURB500-based vectors.

Integrative plasmid and gene replacement
Stathopoulos et al. [95] constructed an integration vector pIJA03-cysS for M. maripaludis to determine the essentiality of cysS gene coding for cysteinyl-tRNA synthetase. They successfully replaced cysS by constructing a pBD1 vector by using another plasmid pPH21310. Several other mutants of M. maripaludis have also been constructed using the techniques of integrative plasmids and gene replacement. For example, acetate auxotrophs were isolated by random insertional mutagenesis in the wild type M. maripaludis with the help of pWDK104 [96]. Using transposon insertion mutagenesis, mutations were made in and around nifH gene to study nitrogen fixing abilities of four transformants. In another study of transposon insertion mutagenesis, an 8-kb region corresponding to the nif gene cluster was confirmed for nitrogen fixation [73].

Markerless mutagenesis
Moore [47] demonstrated markerless mutagenesis in M. maripaludis to show the roles of genes with an unusual ability to use d-alanine or l-alanine. They used a negative selection based system with hpt and upt genes encoding for hypoxanthine and uracil phosphoribosyl transferases present in M. maripaludis. The Hpt system was used to produce markerless in-frame deletion mutations in three genes (ald, alr, and agcS) coding for alanine dehydrogenase, alanine racemase, and alanine permease. hpt was used together with upt to restore the function of wild type ald.

Systems biology
In recent years, systems biology models have been developed for many microbes [97], plants [98], and animals [99] to understand, analyze, and quantify the extent and impact of intracellular interactions and genetic perturbations [100]. However, no such model existed for M. maripaludis, until Goyal et al. [101] reported the first constraint-based genome-scale metabolic model (iMM518). The model comprised of 570 reactions, 556 metabolites, and 518 genes (30 % ORF coverage) across 52 pathways. The details are available in [101]. The only other model [102] available in the literature is for a coculture of M. maripaludis and D. vulgaris, developed for studying methane production from lactate, and understanding the syntrophic association between the two. Unlike iMM518, this model is limited to the reactions in central metabolism only.
Goyal et al. [101] studied the essentialities of genes and reactions in M. maripaludis. Of the 518 genes in their model, 278 proved essential and 240 non-essential. 282 of the 570 reactions proved essential for cell growth. In a previous study, Sarmiento et al. [87] had experimentally analyzed the gene functions in M. maripaludis using whole-genome libraries of Tn5 transposon mutant. 34 of the metabolic genes deemed essential for growth by Sarmiento et al. [87] were also in the list of essential genes by Goyal et al. using iMM518.
Goyal et al. [101] compared the effectiveness of various carbon, hydrogen, and nitrogen sources on cell growth and methane production. Both their model and our recent experiments (unpublished work) suggest that M. maripaludis cannot grow on formate alone (in the absence of CO 2 ). iMM518 also shows that it cannot grow on acetate alone, but no experiments have proven this. Lupa et al. [23] have reported growth on formate under N 2 /CO 2 and Wood et al. [103] have reported growth on formate under H 2 /CO 2 . iMM518 predicts both. This suggests that both H 2 and formate can act as electron donors, while CO 2 is the sole carbon substrate. Between H 2 and formate, H 2 is a better hydrogen source (or electron donor) as growth on CO 2 /H 2 is higher than CO 2 / formate.
Based on iMM518, ammonia is better than N 2 as a nitrogen source for growth, but worse for methane production. Thus, free N 2 is more advantageous for methane production than ammonia with the added benefit of reduced biomass. The reason seems to be the additional ATPs required by nitrogenases for fixing N 2 to ammonia first before cellular assimilation. Since methanogenesis is the only energy producing pathway in M. maripaludis, the additional energy is supplied by enhanced methanogenesis.
iMM518 enabled Goyal et al. [101] to identify the best gene combinations whose deletions would maximize methane production in M. maripaludis. In addition to the above, iMM518 also successfully predicted several experimentally observed phenotypes such as the roles of leuA that encodes the first enzyme for leucine biosynthesis, porE and/or porF whose deletions affect the growth and oxidation of pyruvate, and nif and glnA expressions during nitrogen availability. With its quantitative power and useful insights into the metabolic processes, such a validated model can help tremendously in reengineering M. maripaludis for desired ends.
A recent experimental study [104] measured extracellular fluxes of CO 2 , H 2 , and CH 4 in M. maripaludis during growth on CO 2 as the sole carbon substrate. Using these fluxes and a systems biology model (in this case, iMM518), the study also proposed a procedure for estimating maintenance energy parameters (growth associated maintenance, GAM and non-growth associated maintenance, NGAM).

Potential applications
Methanogens play a key role in the global carbon cycle by reducing atmospheric CO 2 [27]. Their unique characteristics in general, and those of M. maripaludis in particular, offer potential for applications in wastewater treatment, carbon capture and utilization, value-added chemicals production, GTL (Gas to Liquid) applications, and methane production from renewable hydrogen or via electromethanogenesis. Although M. maripaludis has not been used in an industrial setup so far, its attractive features offer much potential for these applications. We now briefly discuss the existing work on each potential application, and assess the promise of M. maripaludis.

Wastewater treatment
Tabatabaei et al. [105] has summarized the characteristics of methanogenic populations used in wastewater treatment. The production of biogas from the anaerobic degradation of waste relies on a symbiotic relationship between syntrophic bacteria (Syntrophomas, Synthrophospora, and Syntrophobacter) and methanogens [106]. The former convert acid-phase products into acetates and H 2 for use by the latter. New evidence [107][108][109] also suggests the possibility of direct electron transfer through nanowires/electrically conductive pili between the two. Without the methanogens removing the acetates and H 2 , acetogenesis cannot proceed [110].
Of all the methanococci, M. maripaludis has many advantages such as autotrophic growth, short doubling time, N 2 -fixation, and stimulation of growth by acetate/ amino acids. Compared to mesophilic methanococci (e.g. M. vannielli, M. voltae, and M. aeolicus), M. maripaludis seems a better choice for a pure culture. M. vannielli has a long doubling time of 8 h, M. voltae does not show autotrophy and requires both acetate and amino acids for its growth, while M. aeolicus requires a higher temperature of ~46 °C for optimum growth. The genera such as methanocaldococcus, methanotorris, and methanothermococcus are thermophilic and require even higher temperatures (65-85 °C) for optimum growth. Thus, M. maripaludis has an advantage over other methanococci for maintaining a low partial pressure of H 2 during wastewater treatment at low temperatures.

Carbon capture and conversion
The flue gas exhausts from power plants typically have 3-15 % CO 2 in majority N 2 . Global CO 2 emissions are the major cause for global warming and climate change [111]. While the world's leading nations have committed to reduce CO 2 emissions in the future, the possibility of sequestration has been losing favor due to various geological and societal reasons. In such a scenario, the interest in converting CO 2 to useful products or fuels is increasing. Given the availability of renewable H 2 from wind, solar, nuclear, etc., M. maripaludis with its ability to uptake CO 2 in the presence of N 2 offers a potential route to capture and convert CO 2 simultaneously from CO 2 emissions to a useful fuel such as methane [83]. M. maripaludis in a microbe consortium with other methanogens such as M. aeolicus [77], M. thermolithotrophicus [76], M. formicicus [78] can be in principle used as to capture and convert CO 2 from power and chemical plant emissions. Given that all methanogens are anaerobic, such a possibility does demand some pretreatment of the flue gas emissions that typically contain some residual oxygen. Further studies are clearly needed to test the feasibility and economics of such an application at large scale.

Methane from renewable energy
The main challenge in using methanogens for large-scale biomethane production is the need for H 2 . The only viable source for this H 2 is a renewable energy source such as solar, tidal, nuclear, or wind. Formate and H 2 are being studied [112] as potential options for energy storage to temporally balance the availability of energy with the use of electricity. M. maripaludis along with other methanogens offers the potential for converting formate and H 2 along with CO 2 into useful fuels such as methane and methanol.
The other possibility is to convert surplus renewable electricity directly into methane via electrochemical methanogenesis. This is already established for a mixedculture of methanogenic microbes comprising Methanobacterium sp. (>93 %) and Methanobrevibacter (~5 %) [113]. Recently, Lohner et al. [22] also demonstrated the uptake of electrons by a hydrogenase-mutant of M. maripaludis, although methane production relative to the wild type was only 1/10 as discussed before. However, these studies do point to the potential of methanogens such as M. maripaludis as biocatalysts for the electrochemical conversion of CO 2 .
Hydrogen production H 2 production by various aerobic and anaerobic microbes has been reported in the literature [114,115]. While genetically engineered E. coli strains can produce 1.7 µmol/mg DCW min to 4.2 µmol/mg DCW min [116] of H 2 from formate during growth on glucose, the wildtype M. maripaludis S2 can readily produce 1.4 µmol/mg DCW min from formate with only CO 2 [23]. Thus, scope exists to enhance this production rate further by reengineering M. maripaludis.

Other applications
Several applications remain unexplored for M. maripaludis in spite of its unique advantages. For instance, it can be studied for the production of high value-added pharmaceuticals, vitamins, amino acids, corrinoids, and terpenoids. Our recent studies (unpublished work) with iMM518 suggests that geraniol, a useful flavoring agent [117], can be produced by up-regulating hmgA and/or idi1 in M. maripaludis. The biotransformation of 2,4,5-trinitrotoluene, a priority pollutant, and metabolic conversion of 5-methylfurfurals and 2-methylfufurals (formed during the concentration of aqueous wastes in the paper and pulp industries) to furfurals have been studied with Methanococcus spp. (strain B) [118,119]. Given the availability of ready genetic tools, such biotransformations can also be explored with M. maripaludis. Other possible applications include the production of liquid biofuels such as methanol, butanol, etc., as being explored [120] by some research groups around the world.

Conclusions
M. maripaludis is a model methanogen with some unique metabolic features (e.g. methanogenesis from simple carbon substrates such as CO 2 , glycolysis via modified EMP pathway, conversion of carbon to a fuel rather than accumulation into biomass, whole cell biocatalysis, diazotrophy, synthesis of all 22 amino acids, and RTCA) that other common workhorses such as E. coli and yeasts do not. This paper presented an integrated and comprehensive review of its metabolic processes, which is missing in the literature in spite of extensive fundamental research on specific aspects of its biochemistry and genetics. We classified and described the salient features of its eight major subsystems and integrated them into a holistic schematic. Our review suggests that further efforts are required towards understanding regulation, acetate biosynthesis, amino acid synthesis, electron transport chain, glycolysis, value-added chemicals/ fuels production, metabolism under stress, carbon capture and utilization applications, and systems biology models. Recent tools for next generation sequencing and -omics, and their integration with systems biology models can deepen our understanding of M. maripaludis at the molecular level and promote further research into this interesting microbe. We hope that this review will benefit both modelers and experimentalists to familiarize themselves with M. maripaludis as an opportunistic methanogen and pursue further studies on our identified gaps within its biochemical pathways and applications. Authors' contributions NG wrote the manuscript, IAK and ZZ edited the manuscript. All authors read and approved the final manuscript.