The F420H2 dehydrogenase from Methanosarcina mazei is a Redox-driven proton pump closely related to NADH dehydrogenases.

The F(420)H(2) dehydrogenase is part of the energy conserving electron transport system of the methanogenic archaeon Methanosarcina mazei Gö1. Here it is shown that cofactor F(420)H(2)-dependent reduction of 2-hydroxyphenazine as catalyzed by the membrane-bound enzyme is coupled to proton translocation across the cytoplasmic membrane, exhibiting a stoichiometry of 0.9 H(+) translocated per two electrons transferred. The electrochemical proton gradient thereby generated was shown to drive ATP synthesis from ADP + P(i). The gene cluster encoding the F(420)H(2) dehydrogenase of M. mazei Gö1 comprises 12 genes that are referred to as fpoA, B, C, D, H, I, J, K, L, M, N, and O. Analysis of the deduced amino acid sequences revealed that the enzyme is closely related to proton translocating NADH dehydrogenases of respiratory chains from bacteria (NDH-1) and eukarya (complex I). Like the NADH-dependent enzymes, the F(420)H(2) dehydrogenase is composed of three subcomplexes. The gene products FpoA, H, J, K, L, M, and N are highly hydrophobic and are homologous to subunits that form the membrane integral module of NDH-1. FpoB, C, D, and I have their counterparts in the amphipathic membrane-associated module of NDH-1. Homologues to the hydrophilic NADH-oxidizing input module are not present in M. mazei Gö1. Instead, the gene product FpoF may be responsible for F(420)H(2) oxidation and may function as the electron input part. Thus, the F(420)H(2) dehydrogenase from M. mazei Gö1 resembles eukaryotic and bacterial proton translocating NADH dehydrogenases in many ways. The enzyme from the methanogenic archaeon functions as a NDH-1/complex I homologue and is equipped with an alternative electron input unit for the oxidation of reduced cofactor F(420) and a modified output module adopted to the reduction of methanophenazine.

Methanosarcina mazei strain Gö1 is a strictly anaerobic methanogenic archaeon that converts a limited number of simple substrates (H 2 ϩ CO 2 , methanol, methylamines, and acetate) to methane. 2-methylthioethanesulfonate is the central intermediate in all methanogenic pathways and is reductively demethylated to methane catalyzed by the 2-methylthioethanesulfonate reductase. The two electrons required for the reduction are derived from 7-mercaptoheptanoylthreonine phosphate, resulting in the formation of a heterodisulfide (CoB-S-S-CoM) 1 of 2-mercaptoethanesulfonate (HS-CoM) and 7-mercaptoheptanoylthreonine phosphate (HS-CoB) (1). An energyconserving step in the metabolism of methanogens is the reduction of CoB-S-S-CoM with either molecular hydrogen or reduced coenzyme F 420 . In recent years, the membrane-bound electron transfer of M. mazei Gö1 has been analyzed in detail, resulting in the discovery of two proton translocating systems referred to as H 2 :heterodisulfide oxidoreductase and F 420 H 2 : heterodisulfide oxidoreductase, respectively (2).
During growth on methylated substrates, part of the methyl groups of the substrates is oxidized to CO 2 , and reducing equivalents are transferred to F 420 . The reduced cofactor (F 420 H 2 ) is reoxidized by the above-mentioned membrane-bound electron transport system consisting of an F 420 H 2 dehydrogenase and a heterodisulfide reductase. The transfer of electrons between the enzymes is most likely mediated by methanophenazine, a hydrophobic cofactor that has been isolated from the cytoplasmic membrane of M. mazei Gö1. The overall process has been shown to be competent in driving proton translocation across the cytoplasmic membrane (3). The resulting electrochemical proton gradient is the driving force for ATP synthesis from ADP ϩ P i catalyzed by an A 1 A 0 -type ATP synthase (2,4).
The F 420 H 2 dehydrogenase with a molecular mass of 115 kDa has been purified from M. mazei Gö1 and contains iron-sulfur clusters and FAD (5). The isolated enzyme is very similar to the corresponding protein from Methanolobus tindarius (6) and is composed of five different subunits with molecular masses of 40, 37, 22, 20, and 17 kDa. A F 420 H 2 dehydrogenase has also been purified form the sulfate-reducing archaeon Archaeoglobus fulgidus (7).
In this report the gene locus encoding the F 420 H 2 dehydrogenase on the M. mazei genome is described. Furthermore, it is shown that the corresponding enzyme is a novel proton pump * This work was supported by Grant De 488/4-2 of the Deutsche Forschungsgemeinschaft (Bonn-Bad Godesberg), by the Deutsche Forschungsgemeinschaft priority program "Structure of functional modules from energy-transducing complexes in prokaryotes" Grant De 488/6-1, and by a grant from the Ministry of Science and Culture of the state Lower Saxony. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF228525 and AF228526.
contributing to the generation of the electrochemical proton gradient in the methanogenic organism.

EXPERIMENTAL PROCEDURES
Assay Conditions-Washed inverted vesicles of M. mazei Gö1 (DSM 3647) were prepared according to Ide et al. (8). Proton translocation, electron transport, and ATP synthesis were monitored as described previously (8). The isolation and reduction of F 420 as well as the synthesis of 2-OH-phenazine was performed according to Abken et al. (9).
Determination of N-terminal Amino Acid Sequences-The F 420 H 2 dehydrogenase was purified as described previously (5). The subunits were separated on SDS-polyacrylamide gel electrophoresis and electroblotted onto polyvinylidene difluoride membranes (Pall GmbH, Dreieich, Germany). N-terminal sequences were determined by Dr. B. Schmidt (Zentrum Biochemie, University of Göttingen) on an Applied Biosystems Procise Sequencer.
Cloning and Sequencing of the fpo Gene-The complete genomic sequence of M. mazei Gö1 was determined by a whole genome shotgun approach. More than 18,000 clones carrying inserts of approximately 2.5 kilobases in length from small insert libraries representative of the whole genome were sequenced from both ends usind LICOR IL 4200 and ABI PRISM 377 DNA sequencers. The generated sequence readings were assembled into contigs with the Prap software implemented in the STADEN software package.
Computer Analysis-Protein sequence analysis was performed with the following internet servers: PredictProtein server; SignalP V1.1 World Wide Web Prediction Server, Center for Biological Sequence Analysis; and PSORT Prediction server.

RESULTS
Proton Translocation Activity of the F 420 H 2 Dehydrogenase-It was shown that electron transport from F 420 H 2 to CoM-S-CoB as catalyzed by washed inverted vesicles from M. mazei Gö1 is coupled to proton translocation across the cytoplasmic membrane (3). With the identification of methanophenazine as an electron carrier in the membrane the redox-driven proton translocation could be analyzed in more detail. It became evident that the key enzymes of the membrane-bound electron transfer systems were able to interact with 2-OH-phenazine, which is a water-soluble homologue of methanophenazine (9). Reducing equivalents from F 420 H 2 were transferred to 2-OHphenazine by the membrane-bound F 420 H 2 dehydrogenase. Furthermore, the heterodisulfide reductase present in the cytoplasmic membrane was able to use reduced 2-OH-phenazine as electron donor for the reduction of CoB-S-S-CoM (10).
Taking advantage of 2-OH-phenazine as electron acceptor, it became evident that the F 420 H 2 dehydrogenase is directly involved in the generation of an electrochemical proton gradient ( Fig. 1). Concentrated vesicles were diluted with a sucrose/ thiocyanate solution containing 200 nmol of F 420 H 2 and were pulsed with 2-OH-phenazine under an atmosphere of molecular nitrogen. In the course of electron transport from F 420 H 2 to 2-OH-phenazine a rapid alkalinization of the medium occurred that was due to proton movement into the lumen of the inverted vesicles. Thiocyanate was used as a permeant chargecompensating cation required to exchange for the ejected protons, thus maintaining the electroneutrality across the membrane. After consumption of F 420 H 2 , the energy conserving electron transport stopped leading to a decay of the generated ⌬ H ϩ by passive diffusion of protons from the lumen of the inverted vesicles to the medium. This is indicated in the second reaction phase where a reacidification took place until the base line was reached again. After calibration of the instrument response with NaOH as an internal standard the extent of reversible alkalinization was used to calculate the H ϩ /2e Ϫ ratio. As evident from Fig. 1 (inset), the extent of H ϩ ejection was dependent on the amount of 2-OH-phenazine added. From the linear part of the reaction curve (0 -7 nmol of 2-OH-phenazine/ assay) a stoichiometry of 0.9 Ϯ 0.2 H ϩ /2e Ϫ was determined. Proton translocation was not observed when 2-OH-phenazine was replaced by ethanol or when F 420 instead of F 420 H 2 was added, indicating that proton transfer was specifically coupled to the F 420 H 2 -dependent 2-OH-phenazine reduction. In the presence of the protonophore SF 6847 the generation of a pH gradient was abolished, and a reversible alkalinization was prevented ( Fig. 1).
Coupling of Electron Transport and ATP Synthesis-When the electron transfer from F 420 H 2 to 2-OH-phenazine was analyzed the initial velocity of substrate conversion was 0.42 Ϯ 0.03 mol ϫ min Ϫ1 ϫ mg protein Ϫ1 (Table I). After 1 min, the rate slowly decreased because of the depletion of reduced F 420 . The reaction was coupled to the phosphorylation of ADP as indicated by a rapid increase of the ATP content upon start of the reaction (Fig. 2). The initial rate of ATP synthesis was 0.11 Ϯ 0.03 mol ϫ min Ϫ1 ϫ mg protein Ϫ1 resulting in an ATP/2e Ϫ stoichiometry of about 0.26 within the first minute of the reaction. In the absence of ADP, 2-OH-phenazine reduction slowed down to 0.3 Ϯ 0.02 unit/mg protein (Table I), and ATP synthesis was not possible (Fig. 2). The ATP synthase inhibitor DCCD led to a strong inhibition of phosphorylation of ADP and resulted in a slight decrease of the electron transport rate (0.32 Ϯ 0.04 unit/mg protein). SF 6847 abolished ATP synthesis ( Fig. 2) because of the decay of ⌬ H ϩ . Inhibition of electron transfer in the presence of DCCD or in the absence of ADP was relieved by the addition of SF 6847 (Table I). The results of the experiments outlined are clearly in accordance with the chemiosmotic coupling mechanism of ATP synthesis. Furthermore, the effect of SF 6847, DCCD, and ADP on the rate of electron transport resembles the phenomenon of respiratory control.
FIG. 1. Redox-driven proton uptake by washed inverted vesicles from M. mazei Gö 1. Proton translocation was monitored in a glass vessel connected to a sensitive pH electrode as described previously (8). After gassing with N 2 , the vessel was filled with 3 ml of 40 mM KSCN solution containing 0.5 M sucrose, 1 mg/ml resazurin, 10 mM dithioerythritol, and 200 nmol F 420 H 2 , followed by the addition of 50 -80 l of washed vesicles (1-1.4 mg protein/assay). The assay was continuously stirred, and the pH was adjusted to 6.8 -6.9. The reaction was started by the addition of 20 nmol of 2-OH-phenazine (20 mM stock solution in ethanol) After finishing the experiments the pH changes were calibrated with a NaOH standard solution. SF 6847 was added to a final concentration of 15 nmol/mg protein where indicated. The low control ratio is explained by the fact that about 50% of the vesicles were uncoupled and catalyzed electron transfer without generating an electrochemical proton gradient (2).
Structure of the fpo Operon Coding for the F 420 H 2 Dehydrogenase from M. mazei Gö1-The entire genome of M. mazei Gö1 is currently sequenced in the Genomic Laboratory Göttingen. In the course of the sequencing process the data were checked for the presence of regions coding for the known N-terminal amino acid sequences obtained from five subunits of the purified enzyme. Finally, one DNA fragment was identified that codes for the 40-, 22-, 20-, and 16-kDa subunits of the purified F 420 H 2 dehydrogenase (Fig. 3). The proposed name for the gene locus is fpo for F 420 H 2 :phenazine oxidoreductase. The fragment comprises 12 genes that were designated fpoA, B, C, D, H, I, J, K, L, M, N, and O. Each gene is preceded by at least one putative ribosome binding site starting with the initiation codon ATG or GTG (fpoL) and is terminated by the stop codons TAA or TGA. Other putative open reading frames could not be identified in the direct neighborhood of the flanking genes fpoA and fpoO, respectively. Upstream of fpoA is an AT-rich region, which contains potential archaeal consensus promoter sequences (Fig. 4A). At the opposite boundary a stem loop was found downstream of fpoO (Fig. 4B), followed by a T-rich region. The RNA duplex stability is Ϫ51 kJ/mol (11). Furthermore, two different repeats of a 9-bp sequence and a 10-bp sequence (TAAAGTGGCT and CTTTATTTT) were identified in this region (Fig. 4B). These structures are similar to transcriptional termination sites of polypeptide-encoding genes from other archaea (12). All 12 structural genes are organized so compactly in the cluster that there is almost no intergenic space for promoter or terminator-like sequences. Therefore, the genes fpoA-O may represent one operon. Evidence for this assumption came from Northern blot analysis. A single 11kilobase signal was obtained when RNA from methanol-grown cells was hybridized with a specific probe (not shown). The size of the transcript is in full accordance with the length of the predicted fpo genes.
The deduced N-terminal amino acid sequences from fpoB, C, D, and I were identical to the N termini of four subunits of the purified F 420 H 2 dehydrogenase. The fifth subunit is encoded by the fpoF gene, which will be described below. Because the gene products of the remaining fpo genes were not found in the homogeneous protein preparation, it is most likely that only a subcomplex of the F 420 H 2 dehydrogenase was purified and that the other subunits were lost during purification. However, the core enzyme composed of FpoB, C, D, F, and I showed catalytic activity with F 420 H 2 as electron donor and several artificial dyes as electron acceptors (10).
The deduced primary sequences of all predicted F 420 H 2 dehydrogenase subunits from M. mazei Gö1 were compared with those of other organisms. Eleven polypeptides showed significant homologies to NADH:plastoquinone oxidoreductases from cyanobacteria or chloroplasts and to NADH:UQ oxidoreductases from mitochondria and bacteria (complex I, NDH-1). Alignments of the fpo gene products A-N indicated similarities of 42-71% and identities of 37-60% to the corresponding subunits of the above-mentioned enzyme complexes (Table II). With the exception of nqo6 from Thermus thermophilus (13) and of nuoI from Pyrococcus abyssi (14), highest scores were obtained for gene products of higher plants and algae. On the other hand the fpo genes are arranged in the same order as bacterial NDH-1 genes. This fact prompted us to number the genes according to the nomenclature of the nuo operon from Escherichia coli (Fig. 3).
Hydropathy plots revealed that the deduced polypeptides from fpoA, H, J, K, L, M, and N are membrane-integral components. The largest subunits are predicted to contain 14 (FpoM, N) to 16 (FpoL) transmembrane helices, and the smaller peptides FpoJ and FpoH are predicted to contain 2 and 8 membrane-spanning helices. Computer programs (PSORT Prediction/SignalP-Server) revealed that FpoA and FpoK may contain N-terminal signal peptides with cleavage sites at amino acid positions 38 and 23, respectively. If this prediction is correct, each of the processed polypeptides would comprise two transmembrane helices. In summary, the membrane integral part of the F 420 H 2 dehydrogenase complex showed high similarities to the corresponding module of bacterial NDH-1 with respect to composition and homology of the amino acid sequences. In the bacterial nuo/nqo operons known so far the genes encoding hydrophobic subunits are clustered at the 3Ј end of the operon with the exception of nuoA/nqo7, which is located at the very beginning of the operon. The same organization is given in the fpo gene cluster from M. mazei Gö1. It is important to note that the hydrophobic subunits of bacterial NDH-1 and of the F 420 H 2 dehydrogenase have their counterparts in mitochondrially encoded complex I subunits from Eukarya (15).
Secondary structure prediction classifies the gene products FpoB, FpoC, FpoD, and FpoI as nonmembrane proteins. This is in agreement with the cellular localization of the homologous polypeptides from bacterial NDH-1 of Nuo BCDI from E. coli (16) and eukaryotic complex I of PSST, 30k, 49k, TYKY from bovine heart (17). These subunits comprise a module that connects the membrane-integral subcomplex to the NADH-oxidizing device (Fig. 3). It is most likely that FpoB, C, D, and I have a similar function in the F 420 H 2 dehydrogenase. In subunit FpoB and FpoI, binding motifs for up to three tetranuclear iron-sulfur centers are present that are invariably conserved in the bacterial and eukaryotic equivalents (Table III). It has been suggested that these prosthetic groups mediate electron transport between the subcomplexes and play an important role in energy conversion of NDH-1 and complex I (18).
The amino acid sequence derived from the open reading  (Table III). The corresponding polypeptide has no counterpart in NDH-1 or complex I and shows no homologies to any known protein. This hydrophilic polypeptide was not copurified with the core enzyme, and its function is unknown.
NADH is oxidized by the hydrophilic NADH dehydrogenase fragment of NDH-1 or complex I, which is composed of NuoF, G, and H (E. coli) or the 24-, 51-, and 75-kDa subunits (bovine heart) (16,17). Genes encoding these subunits are not present in the fpo gene cluster or anywhere else on the M. mazei chromosome, indicating that the input module is different and adjusted to the oxidation of F 420 H 2 as electron donor of the F 420 H 2 dehydrogenase. A suitable candidate for this function is the 37-kDa subunit of the purified core enzyme. The N-terminal sequence of this polypeptide was not found in the deduced amino acid sequences of the fpo gene cluster. However, a gene coding for the fifth subunit (37 kDa) of the purified enzyme was identified at a different location on the chromosome and was referred to as fpoF (F for F 420 ). The deduced amino acid sequences shows motifs for two [Fe4-S4] clusters. It is homologous to the ␤ subunit of F 420 -reducing hydrogenases and to subunits of the F 420 H 2 dehydrogenase from M. tindarius (FfdB; Ref. 19) and A. fulgidus (AF1833; Ref. 20). In the latter organism the fpoF homologue gene AF1833 is part of the operon encoding the F 420 H 2 dehydrogenase. Recently, the gene was overexpressed in E. coli, and the corresponding polypeptide was purified to homogeneity. The subunit contained nonheme iron, acid-labile sulfur, and FAD and was able to oxidize F 420 H 2 when the artificial electron acceptor methylviologen was added. 2 Because AF1833 and FpoF are structurally equivalent and FpoF is part of the purified F 420 H 2 dehydrogenase from M. mazei Gö1, it may function as electron input device of the enzyme.

DISCUSSION
Energetics of the F 420 H 2 Dehydrogenase-F 420 is the central cytoplasmic electron carrier in the methanogen M. mazei Gö1. The cofactor is involved in the process of methanogenesis from H 2 ϩ CO 2 and from methylated compounds such as methanol and methylamines. In the methylotrophic pathway of methane formation (21) three out of four methyl-groups are reduced to methane. The remaining methyl-moiety is oxidatized to CO 2 , and the resulting reducing equivalents are transferred to F 420 . The membrane-bound F 420 H 2 dehydrogenase catalyzes the reoxidation of F 420 H 2 and most likely transfers the electrons to the novel membrane-integral cofactor methanophenazine, which is a linear sesterterpenic ether of 2-OH-phenazine (2). Previous studies indicate that the protein is part of the membrane-bound F 420 H 2 :heterodisulfide oxidoreductase, which is one of the major energy conserving systems of M. mazei Gö1 (3). It was found that the electron transport from F 420 H 2 to the heterodisulfide (CoM-S-S-CoB) is coupled to the transfer of 3-4 protons across the cytoplasmic membrane. Using 2-OH-phenazine as a water-soluble precursor of methanophenazine, it became evident that the overall electron transfer can be divided into two partial reactions catalyzed by the F 420 H 2 dehydrogenase and the heterodisulfide reductase, respectively. Very recently, it was shown that the dihydrophenazine-dependent heterodisulfide reduction (Reaction 2) is coupled to proton translocation exhibiting a stoichiometry of 2H ϩ /2e Ϫ (8). The experiments in this publication show that in the first partial reaction catalyzed by the F 420 H 2 dehydrogenase, a maximum of 0.9 Ϯ 0.2 protons/two electrons were translocated. Taking into account that about 50% of the vesicles are uncoupled (2), a H ϩ /2e Ϫ ratio of about 2.0 could be considered. Then the H ϩ /2e Ϫ stoichiometries of the partial reactions would add up to 4 and support the value of 3-4 H ϩ /2e Ϫ translocated in the overall electron transport from F 420 H 2 to heterodisulfide (3). In summary, the data clearly indicate that the F 420 H 2 dehydro-2 H. Brü ggemann and U. Deppenmeier, unpublished results. genase is a redox-driven proton pump showing a maximal energetic efficiency of about 2 H ϩ translocated per 2e Ϫ transported.
Comparison of F 420 H 2 Dehydrogenase and Proton Translocating NADH Dehydrogenases-The F 420 H 2 dehydrogenase from M. mazei Gö1 resembles eukaryotic complex I and bacte-rial NDH-1 in many ways: 1) The membrane-bound, flavin and iron-sulfur containing enzymes are characterized by a complex subunit composition. 2) The electron donors F 420 H 2 and NADH are both reversible hydride donors with comparable mid-point potentials. 3) Both enzymes take advantage of small hydrophobic nonproteinous electron acceptors namely quinones in case of the NADH dehydrogenase and methanophenazine in case of F 420 H 2 dehydrogenase. The electron acceptors are highly hydrophobic and can diffuse within the cytoplasmic membrane. Several experimental data indicate that semiquinone radicals are produced in the course of the NADH dehydrogenase reaction (22). The same could be true for methanophenazine because the direct precursor 2-OH-phenazine is also reduced in two 1e Ϫ transfer reactions forming a semiphenazine radical as an intermediate in solution (23). 4) It was shown that mitochondrial and bacterial NADH:ubiquinone oxidoreductases as well as the F 420 H 2 dehydrogenase are inhibited by diphenyleneiodonium chloride (24,25). 5) The redox-reaction catalyzed by the proteins is coupled to proton translocation.
It is shown in this publication that the physiological characteristics of the F 420 H 2 dehydrogenase are supported by genetic data. The deduced primary sequences of the M. mazei Gö1 F 420 H 2 dehydrogenase subunits were compared with those of other organisms and with those of some phylogenetically related enzymes. For 11 proteins encoded by fpoA to fpoN related counterparts exist in bacterial NDH-1 and mitochondrial complex I. The organization of the genes resembles operons encoding the proton-translocating NADH:ubiquinone oxidoreductases (Ndh/Nqo) from several bacteria such as E. coli, T. thermophilus, and Rhodobacter capsulatus (13,26,27). The bacterial enzyme is composed of 13 or 14 different subunits (28) that form the following modules (nomenclature according to E. coli) 1) The hydrophilic NADH dehydrogenase fragment composed of NuoE, F, and G catalyzes the oxidation of NADH. 2) NuoA, H, J, K, L, M, and N form the membrane-integral module and are involved in quinone reduction and proton-translocation. 3) NuoB, C, D, and I connect the above-mentioned subunits and catalyze electron transfer from module 1 to module 3 (29).
The detailed reaction mechanism of the enzyme is unknown. Several intrinsic redox components were detected that are involved in NADH-dependent quinone reduction (18). The NuoE, F, and G module contains most of the redox-active prosthetic groups (16). One noncovalently bound FMN and at least five EPR detectable iron-sulfur clusters form a long electron transfer chain guiding electron transfer from NADH to the [Fe4-S4] cluster N2 that is part of the amphipathic connecting fragment   NuoB, C, D, and I. It is still a matter of debate whether this cluster is bound to NuoB or NuoI. However, it is well established that reduced N2 successively injects single electrons into the membraneous subcomplex, thereby activating a serial array of quinones that are directly involved in H ϩ translocation (30).
The native structure of the F 420 H 2 dehydrogenase is still unknown. However, the primary sequence informations, operon structure, and the homology to bacterial NDH-1 allow composition of a tentative model (Fig. 5). The gene product FpoF forms the input module that oxidizes F 420 H 2 by hydride transfer. FAD present in this subunit catalyzes a two-electron/ one-electron switch to reduce the [Fe4-S4] clusters. It is still an open question whether the gene product of fpoO is also part of the input module. This polypeptide is predicted to be hydrophilic and probably contains [Fe2-S2]-clusters. On the other hand this polypeptide was not copurified with the core enzyme, indicating that it is not essential for catalytic activity. From the F 420 H 2 -oxidizing device the electrons are then channeled to the amphipathic connecting fragment composed of FpoB, C, D, and I, which is highly homologous to the corresponding module of NDH-1. Because all iron-sulfur signatures are conserved in FpoB and FpoI, it is reasonable to assume that a FeS-cluster comparable with N2 is present in one of these subunits. In analogy to NDH-1 and complex I cluster N2 should transfer electrons to the membrane integral module composed of FpoA, H, J, K, L, M, and N. In spite of the fact that the composition of the membraneous part of the F 420 H 2 dehydrogenase and NDH-1 is identical, the further electron transport pathway of the F 420 H 2 -dependent enzyme is difficult to predict because methanogenic archaea do not contain quinones. Therefore, the reaction mechanism of the F 420 H 2 dehydrogenase must be different at this point and must involve the electron carrier methanophenazine. The mid-point potential of 2-OH -phenazine, which is a potential precursor of methanophenazine, was determined to be Ϫ255 mV (23). With the assumption that the redox potential of methanophenazine is similar, the change of free energy (⌬G o Ј) coupled to the F 420 H 2 -dependent methanophenazine reduction is only Ϫ20.2 kJ/mol compared with a ⌬G o Ј of Ϫ80.9 kJ/mol for the NADH-dependent reduction of ubiquinone. These thermodynamic facts are reflected by the coupling efficiencies of the enzymes because the maximal H ϩ / 2e Ϫ ratio of the F 420 H 2 dehydrogenase is 1.8, in contrast to NDH-1/complex I, which translocates four or even more protons across the membrane per reaction cycle (30).
Despite the aforementioned differences, the F 420 H 2 dehydrogenase represents a NDH-1 homologue in the methanogenic archaeon M. mazei Gö1, which is equipped with an alternative input device and a modified proton translocating machinery. Further analysis of the enzyme may contribute to the understanding also of the reaction mechanism of NADH dehydrogenases.
Interestingly, fpo-like gene clusters were not detected in the methanogenic archaea Methanococcus jannaschii (31) and Methanobacterium thermoautotrophicum (32), indicating that a F 420 H 2 dehydrogenase is absent in these organisms. This fact is in accordance with the finding that the electron transport chains from obligate hydrogenotrophic methanogens of the or-ders Methanobacteriales and Methanococcoles are different from those of methylotrophic methanogens belonging to the order Methanosarcinales. (2)