The coupling ion in the methanoarchaeal ATP synthases: H+ vs. Na+ in the A1Ao ATP synthase from the archaeon Methanosarcina mazei Gö1

Authors

  • Kim Y. Pisa,

    1. Molecular Microbiology & Bioenergetics, Institute of Molecular Biosciences, Johann Wolfgang Goethe Universität Frankfurt, Frankfurt, Germany
    Search for more papers by this author
  • Claudia Weidner,

    1. Molecular Microbiology & Bioenergetics, Institute of Molecular Biosciences, Johann Wolfgang Goethe Universität Frankfurt, Frankfurt, Germany
    Search for more papers by this author
  • Heiko Maischak,

    1. Molecular Microbiology & Bioenergetics, Institute of Molecular Biosciences, Johann Wolfgang Goethe Universität Frankfurt, Frankfurt, Germany
    Search for more papers by this author
  • Holger Kavermann,

    1. Molecular Microbiology & Bioenergetics, Institute of Molecular Biosciences, Johann Wolfgang Goethe Universität Frankfurt, Frankfurt, Germany
    Search for more papers by this author
  • Volker Müller

    1. Molecular Microbiology & Bioenergetics, Institute of Molecular Biosciences, Johann Wolfgang Goethe Universität Frankfurt, Frankfurt, Germany
    Search for more papers by this author

  • Present address: Heiko Maischak, Department of Bioorganic Chemistry, Max Planck Institute for Chemical Ecology, 07745 Jena, Germany.
    Holger Kavermann, Roche Diagnostics GmbH, Penzberg, Germany.

  • Editor: Aharon Oren

Correspondence: Volker Müller, Molecular Microbiology & Bioenergetics, Institute of Molecular Biosciences, Johann Wolfgang Goethe Universität Frankfurt, Max-von-Laue-Str. 9, 60438 Frankfurt, Germany. Tel.: +49 69 79829507; fax: +49 69 79829306; e-mail: vmueller@bio.uni-frankfurt.de

Abstract

To establish a system to analyze ATP synthesis by the archaeal A1Ao ATP synthase and to address the nature of the coupling ion, the operon encoding the A1Ao ATP synthase from the mesophile Methanosarcina mazei Gö1 was cloned in an expression vector and it was expressed in the F1Fo ATP synthase-negative mutant Escherichia coli DK8. Western blot analyses revealed that each of the subunits was produced, and the subunits assembled to a functional, membrane-embedded ATP synthase/ATPase. ATP hydrolysis was inhibited by dicyclohexylcarbodiimide but also by tributyltin, which turned out to be the most efficient inhibitor of the Ao domain of A1Ao ATP synthase known to date. ATP hydrolysis was not dependent on the Na+ concentration of the medium, and inhibition of the enzyme by dicyclohexylcarbodiimide could not be relieved by Na+. The enzyme present in the cytoplasmic membrane of E. coli catalyzed ATP synthesis driven by an artificial ΔpH but not by ΔpNa or ΔμNa+.

Introduction

Methanogenic archaea are a nutritionally uniform group of microorganisms that thrive by the conversion of a limited number of substrates such as methanol, N-methyl compounds, H2+CO2 or acetate to methane (Thauer, 1998). The most common substrate combination is H2+CO2, which is converted to methane according to

image

Energy conservation during methanogenesis is by a chemiosmotic mechanism (Deppenmeier et al., 1999). The ultimate step in the pathway, the heterodisulfide reductase system, is coupled to vectorial proton translocation across the membrane (Deppenmeier, 2002), whereas the penultimate step, the methyltetrahydromethanopterin : coenzyme M methyltransferase, is coupled to vectorial Na+ translocation across the membrane (Gottschalk & Thauer, 2001). Therefore, methanogenic archaea are unique among organisms because they couple the pathway of energy conservation with the primary extrusion of both Na+ and H+ (Deppenmeier et al., 1996; Schäfer et al., 1999). Claims have been made for H+- as well as Na+-driven ATP synthesis in Methanosarcina mazei Gö1 and Methanothermobacter thermautotrophicus, and inhibitor studies suggested the presence of H+- coupled A1Ao and Na+, coupled F1Fo ATPases (Smigan et al., 1992, 1994, 1995; Becher & Müller, 1994). In contrast, genome studies clearly excluded the presence of F1Fo ATPase genes in M. mazei Gö1 (Deppenmeier et al., 2002) and M. thermautotrophicus (Smith et al., 1997) but revealed the A1Ao ATPase genes. Therefore, the question of how the Na+ gradient is converted to ATP synthesis is still open and, clearly, more defined systems are required to address this important question.

So far, the A1Ao ATP synthase from the hyperthermophile Methanocaldococcus jannaschii is the only methanoarchaeal A1Ao ATP synthase that has been purified (Lingl et al., 2003). Although this enzyme has allowed the first insights into the structure of A1Ao ATP synthases (Coskun et al., 2004), it proved to be disadvantageous for studying ATP synthesis in a proteoliposome system due to the instability of liposomes at 80 °C (Weidner, 2003). For that purpose, enzymes from mesophiles or moderate thermophiles are more suitable. Because all attempts to purify A1Ao ATP synthases from membranes of mesophilic methanogens have failed so far (Müller et al., 1999; Schäfer et al., 1999), a procedure was developed to produce subcomplexes of the A1Ao ATP synthase in Escherichia coli. Overproduction of a functional A1 subcomplex (A3B3CDF) allowed the first functional and structural characterization of an A1 subcomplex (Grüber et al., 2001; Lemker et al., 2001, 2003; Coskun et al., 2002). The initial overproduction of an A3B3CDF subcomplex has been followed up and a system has been developed to produce the entire A1Ao ATP synthase in a functional state in E. coli and the ion specificity of the A1Ao ATP synthase has been addressed.

Materials and methods

Organisms and plasmids

Methanosarcina mazei Gö1 (DSM 3647) was obtained from the ‘Deutsche Sammlung für Mikroorganismen und Zellkulturen,’ Braunschweig, Germany, and grown under strictly anaerobic conditions as described (Hippe et al., 1979). Escherichia coli DK8 (1100 Δ[uncB-uncC] ilv::Tn10) (Klionsky et al., 1984) and DH5α (supE44 ΔlacU169 Φ80lacZΔM15 hsdr17 recA1 endA1 gyrA96 thi1 relA1) (Hanahan, 1983) was obtained from the ‘Deutsche Sammlung für Mikroorganismen und Zellkulturen,’ Braunschweig, Germany. The cells were grown in Luria–Bertani (LB) media at 37 °C, and at an OD600 nm of 0.5 gene expression was induced by 0.2% arabinose. At an OD600 nm of 1.5–1.7, the cells were harvested by centrifugation (10 000 g; 20 min; 4 °C) in a Sorvall Superspeed RC2-B. The plasmids used were pRT103 (Lemker et al., 2001), pVSBAD2 (Prof. Dr C. Baron, Hamilton, Canada), pMALc2 (Fa. NEB) and pBAD-TOPO (Fa. Invitrogen) and pRT1.

Construction of the plasmid pRT1

The genes encoding the A1Ao ATP synthase are organized in an operon in the order 5′-ahaH, ahaI, ahaK, ahaE, ahaC, ahaF, ahaA, ahaB, ahaD, ahaG-3′ (Fig. 1). The entire operon was cloned in two steps. By restriction of the plasmid pRT103 (Lemker et al., 2001) with the endonucleases XbaI and SacI, a 6581-bp fragment was obtained that contains the 3′-end of ahaE, the genes ahaC, ahaF, ahaA, ahaB, ahaD and ahaG and 723 bp of the gene hypF (Fig. 1) and cloned into pVSBAD2, giving pRT0. In a second step, a fragment that includes the genes ahaH, ahaI, ahaK and the 5′-end of ahaE was produced using genomic DNA of M. mazei and two primers 5′-ahaH(SalI) [5′-ATATTAAATGTCGACAGACGGAGATTC-3] and 3′-ahaEup(XbaI) [5′-CGGTATCTAGAAGTT-3] for insertion of the required cleaving sites. The fragment was cloned using a pBAD TOPO TA Cloning® Kit, giving pHS3. In a third step, the SalI and XbaI fragment of pHS3 was cloned in pRT0, giving pRT1, which includes the complete aha-operon under control of an araBAD promoter and the gene for the repressor AraC. For further information, see Fig. 1. The identity of the construct was confirmed by restriction mapping and DNA sequence analyses.

Figure 1.

 Strategy for cloning pRT1. The enzymes used for cloning are boxed. Genes encoding for hydrophobic proteins are indicated with asterisks. For further information, see text.

Generation of malE-fusion clones, expression conditions and purification of the fusion proteins and generation of antibodies

The generation of antibodies against subunits A, B, C, D, E, F, H (Lemker et al., 2001) and c has been described (Lemker et al., 2001). The hydrophilic part of ahaI encoding subunit a was amplified by PCR by introducing restriction sites at the 5′end and the 3′end with the primers ‘ahaI.5′(BamHI)’ [5′-GGGCGGTCCGGATCCAGTAGACCTAAAGAG-3] and ‘ahaI.cD3′(PstI)’ [5′-GCCTGGACTGCAGATCCATAA-3]. The 3′ primer was created with a stop codon to terminate translation. The fragment was cloned in frame in front of malE in pMal-c2 (pHK21), and transformed in E. coli DH5α. Expression of the fusion genes was induced at OD600 nm=1 with 0.5 mM isopropyl-β-d-thiogalactopyranoside. After 2 h of growth, cells were harvested, washed and disrupted in a French press. The cell extract was centrifuged and applied to an amylose resin to purify the MalE-fusion. Proteins were separated and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Schägger & von Jagow, 1987) and stained with Coomassie Brilliant Blue G-250. Because there is no MalE in M. mazei Gö1 and because a MalE antibody does not cross-react with a cell-free extract of M. mazei Gö1, the entire fusion protein was used to immunize rabbits. Western blotting with SDS polyacrylamide gels was performed as described (Towbin et al., 1979; Lemker et al., 2001).

Preparation of membranes

Five to 7 g E. coli cells were washed in TMGD-buffer [50 mM Tris, 5 mM MgCl2, 1 mM DTE, 0.1 mM phenylmethanesulfonyl fluoride, and 10% (v/v) glycerol (pH 7.5)] and then resuspended in 12 mL TMGD-buffer g−1 cell material. Cells were disrupted by three passages through a French press (1200 psig). After cell debris was removed by centrifugation (27 000 g; 30 min; 4 °C), membranes were pelleted by ultracentrifugation (145 000 g; 1.5 h; 4 °C). The membranes were washed twice in TMGD buffer. The protein concentration was determined as described (Bradford, 1976).

Measurement of ATP concentration in cells

ATP was determined as described (Becher & Miller, 1994).

ATPase activity

ATPase activity was measured in an assay mixture containing 100 mM morpholinoethane sulfonic acid (MES), 40 mM Na-acetate, 40 mM Na2S2O5, 10 mM MgSO4, 10% (v/v) glycerol, pH 5.2, and enzyme solution as described previously (Pisa et al., 2007).

Generation of artificial driving forces for ATP synthesis in cell suspensions of E. coli DK8(pRT1)

Different artificial driving forces were applied to cell suspensions to induce ATP synthesis. Therefore, 10 mL cell suspensions (1 mg mL−1) were incubated in 58 mL glass bottles at 37 °C on a rotary shaker (70 r.p.m.). For ΔpH-induced ATP synthesis, HCl was added from stock solution to cell suspensions in 20 mM imidazole/HCl, pH 7.0, 20 mM MgSO4, and 50 mM NaCl. A potassium ion diffusion potential (ΔΨ) was applied by addition of 20 μM valinomycin to cells incubated in 20 mM imidazole, pH 7, 20 mM MgSO4, and 50 mM KCl. A proton diffusion potential was created by addition of KOH (final pH 10.0) to cell suspensions incubated in 10 mM imidazole, 10 mM Tricine, 5 mM potassium citrate, 20 mM MgSO4, and 50 mM NaCl, pH 5.0. To mediate proton flux, the buffer contained 25 μM of the protonophores tetrachlorosalicylanilide or SF6847.

In case of inhibitor studies, cell suspensions were preincubated with the indicated inhibitor for 20 min at 37 °C before the experiment. Ionophores and inhibitors were added as ethanolic solutions; controls received the solvent only.

Results

Production and cellular localization of the A1Ao ATP synthase of M. mazei Gö1 in E. coli DK8(pRT1)

Plasmid pRT1 that contains the A1Ao ATP synthase genes in the order 5′-ahaH, ahaI, ahaK, ahaE, ahaC, ahaF, ahaA, ahaB, ahaD, ahaG-3′ (Fig. 1; see ‘Materials and methods’) was transformed into the F1Fo ATP synthase-negative mutant E. coli DK8 (Klionsky et al., 1984). Transformants were grown on LB, gene expression was induced by 0.2% arabinose, cells were harvested, cell-free extract was prepared and separated into cytoplasm and membrane fraction. Western blot analyses revealed the production of every single subunit of the A1Ao ATP synthase, including the membrane-bound subunits. Subunit c was found exclusively in the cytoplasmic membrane; the others were found in both fractions, indicating that part of the A1Ao ATP synthase was stripped of the membranes during preparation of the membranes. Accordingly, ATPase activity was detected in the cytoplasm (54 U mg−1 protein), as well as in the membrane (67 U mg−1 protein) whereas there was no activity detectable in E. coli DK8 under these conditions either at membranes or in the cytoplasm. This is clear evidence for the functional production of the A1Ao ATP synthase in the bacterial host and is targeted at the cytoplasmic membrane.

Characterization of membrane-bound ATPase activity

ATPase activity was maximal at pH 5.2. Methanol in concentrations up to 1.2% (v/v) had no effect on ATPase activity. However, as observed before with the A1 domain (Lemker et al., 2002), ATPase activity was stimulated by sodium acetate (2.2-fold by 40 mM) or sulfite (1.9-fold by 40 mM). Furthermore, the enzyme was inhibited by diethylstilbestrol with an I50 of 200 nmol diethylstilbestrol mg−1 protein (Fig. 2a).

Figure 2.

 Inhibition of ATPase activity by diethylstilbestrol (a) or dicyclohexylcarbodiimide (b). The activity was measured at membranes of Escherichia coli DK8(pRT1). The protein concentration was 1 mg mL−1. The enzyme was preincubated with diethylstilbestrol or dicyclohexylcarbodiimide for 30 min at 37°C, before the reaction was started by the addition of 4 mM ATP. Diethylstilbestrol and dicyclohexylcarbodiimide were added as ethanolic solutions; controls received the solvent only. One hundred per cent activity corresponds to 64 mU mg−1.

A prerequisite for addressing the nature of the coupling ion is a coupled enzyme. Diethylstilbestrol acts on the A1 domain and does not allow to estimate the amount of coupled enzymes in the preparation. Dicyclohexylcarbodiimide, a known inhibitor of the A1Ao ATP synthase, blocks ion flow through the AO domain and thereby inhibits ATPase activity. Therefore, dicyclohexylcarbodiimide inhibition of ATP hydrolysis is indicative of a coupled enzyme. As can be seen from Fig. 2b, dicyclohexylcarbodiimide inhibited ATP hydrolysis. Dicyclohexylcarbodiimide inhibited the enzyme with an I50 of 500 μmol dicyclohexylcarbodiimide mg−1 protein. This experiment indicates a functional coupling of A1 and AO in the preparation. To strengthen this conclusion, it was attempted to solubilize an intact, coupled A1Ao ATP synthase from the membranes of E. coli DK8(pRT1). Dodecylmaltoside solubilized 60% of the ATPase activity; 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate or Triton X-100 were less efficient (Table 1). Because dicyclohexylcarbodiimide did not completely block membrane-bound ATPase activity, either in E. coli DK8(pRT1) or in M. mazei Gö1 (Wilms et al., 1996), the effect of the inhibitor tributyltin that was described to interact with the Fo domain of F1Fo ATPases was tested (Pedersen & Carafoli, 1987; von Ballmoos et al., 2004). As can be seen from Fig. 3, tributyltin inhibited the A1Ao ATPase activity very effectively. Complete inhibition was observed at 250 μM, and half-maximal activity at 60 μM. In contrast, tributyltin did not inhibit ATP hydrolysis catalyzed by the A1 domain. These experiments not only revealed tributyltin to be an efficient inhibitor of the Ao domain of the A1Ao ATPase but also underlined the conclusion that the A1 and Ao domains are functionally coupled in the preparation.

Table 1.   Solubilization of the Methanosarcina mazei A1AO ATP synthase from Escherichia coli DK8(pRT1) membranes with different detergents
Detergent*g g−1ATPase activity
Membrane (U)SupernatantPelletTotal activity (U)Recovery (%)
(U)(%)(U)(%)
  • *

    Detergents were used at a concentration of 1 or 2 g g−1 membrane protein; total protein concentration was 29 mg mL−1, volume was 1 mL.

Dodecylmaltoside
 RT12011555251680
22012605251785
 37°C12013654201785
22012602101470
Triton X-100
 RT120630147020100
2201050136523115
 37°C12042013651785
22042014701890
CHAPS
 RT1203159451260
2204206301050
 37°C1203158401155
220420420840
Figure 3.

 Inhibition of ATP hydrolysis by tributyltin. Solubilized A1Ao ATP synthase (□) from Escherichia coli DK8(pRT1) with a protein concentration of 11.5 mg mL−1 or the A1 subcomplex (▴) from E. coli DK8(pTL2) with a protein concentration of 1.75 mg mL−1 were incubated in buffer with the inhibitor for 30 min at 37°C before the reaction was started by addition of 2.5 mM Na2-ATP. One hundred per cent corresponds to 0.091 U mg−1 for the A1Ao ATP synthase and to 0.196 U mg−1 for the A1 ATPase.

ATP synthesis by an artificial ΔpH

The above-mentioned experiments unequivocally demonstrated a coupled ATPase activity by the A1Ao ATP synthase, a prerequisite for ATP synthesis. To verify whether the enzyme is capable of ATP synthesis in E. coli, artificial driving forces were applied to cell suspensions of E. coli DK8(pRT1) and the cellular ATP content was monitored. When an artificial ΔpH was created across the membrane, ATP was synthesized almost instantaneously (Fig. 4a), and the increase in cellular ATP content was dependent on the magnitude of the ΔpH generated (Fig. 4b). Cells carrying only the vector did not synthesize ATP. ΔpH-driven ATP synthesis was inhibited by dicyclohexylcarbodiimide or diethylstilbestrol (Fig. 5). These experiments clearly demonstrate that the A1Ao ATP synthase of M. mazei Gö1 produced in E. coli is assembled in a functional state that allows for ATP synthesis, a prerequisite for further functional studies. Furthermore, the data demonstrate that H+ is used as a coupling ion.

Figure 4.

 An artificial ΔpH drives ATP synthesis in whole cells of Escherichia coli DK8(pRT1). Cells were incubated in 20 mM imidazole, 20 mM MgSO4, 20 mM K2HPO4 and 50 mM NaCl, pH 7, at a protein concentration of 1 mg mL−1. At the time point indicated by the arrow, HCl was added to lower the pH to 1 (◆), 2 (□), 3 (▴), 4 (•), 5 (⋄), 6 (▪) and 7 (▵). The test buffer was adjusted to a pH of 7. Samples were taken, and the intracellular ATP content was monitored. (b) Depicts the intracellular ATP content as a function of ΔpH.

Figure 5.

 Inhibition of the ΔpH-driven ATP synthesis with diethylstilbestrol, dicyclohexylcarbodiimide or hexestrole. Cell suspensions of Escherichia coli DK8(pRT1) with a protein concentration of 1 mg mL−1 were incubated in buffer in the presence of 1% ethanol (♦), 1 mM dicyclohexylcarbodiimide (▴), 200 μM diethylstilbestrol (▪) or 200 μM hexestrole (•). The cell suspensions were preincubated with the inhibitors for 20 min at 37°C. At the time point indicated by the arrow, the pH was decreased to 1 by addition of HCl.

Na+ as a coupling ion in the A1Ao ATP synthase from M. mazei?

The F1Fo ATP synthases from Acetobacterium woodii, Propionigenium modestum and Ilyobacter tartaricus are known to use Na+ as a coupling ion (Laubinger & Dimroth, 1987; Reidlinger & Müller, 1994; Neumann et al., 1998). The Na+-binding site is located in the rotor subunit c and shown to be made up of residues Gln32, Glu65, Ser66 and Val63 in I. tartaricus (Meier et al., 2005). This binding motif is conserved in subunit c from the known Na+ F1Fo ATP synthases, but also in subunit c from methanoarchaeal A1Ao ATP synthases (Fig. 6), indicating that the A1Ao ATP synthase from M. mazei may use Na+ as a coupling ion. However, a stimulation of ATP hydrolysis by Na+ (tested as low as 0.072 mM) was not observed, and there was no protection from dicyclohexylcarbodiimide inhibition of ATP hydrolysis by Na+. ATP synthesis in whole cells of A. woodii could be driven by a proton diffusion potential in a Na+-dependent manner (Heise et al., 1991), an elegant way to demonstrate ΔμNa+-dependent ATP synthesis. However, neither a potassium ion diffusion potential nor a proton diffusion potential could drive the synthesis of ATP via the A1Ao ATP synthase. The absence of features characteristic for Na+ F1Fo ATP synthases are not in favor of a Na+-coupled A1Ao ATP synthase in M. mazei.

Figure 6.

 Alignment of c subunits from Na+ F1Fo ATP synthases and methanoarchaeal A1Ao ATP synthases. The sequences of the c subunits from F1Fo ATP synthases from Ilyobacter tartaricus (I.tar), Propiogenium modestum (P.mo) and Acetobacterium woodii (A.wo) and A1Ao ATP synthases from Methanosarcina mazei Gö1 (M.ma), and Methanocaldococcus jannaschii (Mj2, second hairpin), Methanococcus maripaludis (M.mari2, second hairpin), Methanopyrus kandleri (M.kand), Methanothermobacter thermautotrophicus (M.the1, first hairpin), Methanothermobacter thermautotrophicus (M.the2, second hairpin), Methanosarcina barkeri (M.bar), Methanocaldococcus jannaschii (M.j3, third hairpin), Methanococcus maripaludis (M.mari3, third hairpin), Methanosarcina acetivorans C2A (M.ac). It should be noted that some methanoarchaeal c subunits arose from gene multiplications but the sodium ion-binding motif is not conserved in every copy. Therefore, only the domains that have retained the potential sodium ion-binding site are shown.

Discussion

Expression of archaeal genes in bacteria can be driven from different artificial promoters. Here, the pBAD promoter has been used, which is tightly repressed by AraC and derepression is achieved by addition of arabinose (Guzman et al., 1995). This tight regulation prevents unwanted expression from leaky promoters and the production of amounts of ATP synthase that might be toxic to E. coli. However, this expression system made it impossible to check a functional complementation of the ATPase-negative phenotype of E. coli. This is usually done by demanding growth of the transformant on a nonfermentable carbon source such as succinate that requires the presence of electron transport-driven phosphorylation. Unfortunately, the inducer arabinose is also a carbon and energy source for E. coli DK8. However, reproducibly higher final optical densities (twofold) in cultures of E. coli DK8(pRT1) were observed compared with E. coli DK8(pVSBAD2) grown on LB plus 0.2% arabinose. This might result from a functional complementation and may argue for ATP synthesis in E. coli catalyzed by the A1Ao ATP synthase.

Most importantly, it could be demonstrated here for the first time that an A1Ao ATP synthase is indeed able to synthesize ATP. Moreover, the data clearly demonstrated ΔpH as a driving force and thus H+ as a coupling ion. The rotor subunit c has the conserved Na+-binding motif, and very recently, a predicted Na+ dependence of ATP hydrolysis by another archaeal ATPase from Pyrococcus furiosus could be experimentally verified. The enzyme was predicted for the same reason to be a Na+ A1Ao ATPase (Pisa et al., 2007). Here, a Na+ dependence could not be observed even using different approaches. This could mean that the KM for Na+ is well below the lowest concentration of Na+ that could be obtained in the assay. However, this concentration was suitable to detect the Na+ dependence of ATP hydrolysis in A. woodii (Reidlinger & Müller, 1994) and P. furiosus (Pisa et al., 2007). On the other hand, the possible argument that the Na+ concentration in the assay buffer was above the KM for Na+ cannot hold true for the other approach used. If there is a ΔμNa+-driven ATP synthesis, then the proton diffusion potential should have led to the synthesis of ATP, even in the absence of added Na+ (considering contaminating Na+ concentrations above the KM). However, even after addition of high amounts of Na+, ATP synthesis was not observed. Furthermore, Na+ is thought to bridge the rotor subunits in the rotor (c ring) of Na+ F1Fo ATP synthases, thus leading to the extreme stability of the rotor. Again, this was not observed for the A1Ao ATP synthase from M. mazei. The apparent discrepancy between prediction and experimental evidence may result from a yet to be identified additional Na+-binding site, for example in the other motor subunit, subunit a, which may be absent in methanoarchaeal ATPases.

In summary, the production of an archaeal ATP synthase containing nine subunits in a bacterial host has been demonstrated for the first time. The enzyme is targeted to the membrane and integrated into the membranes in a functional, ATP-synthesizing state. The enzyme uses H+ as a coupling ion and evidence for Na+ as a coupling ion was not obtained. Therefore, the Na+ gradient established by the methyltetrahydromethanopterin : coenzyme M methyltransferase may be converted by a Na+/H+ antiporter to a proton gradient that then drives the synthesis of ATP.

Acknowledgements

This work was supported by grants from the Deutsche Forschungsgemeinschaft (SPP 1070 and SFB 472).

Authors' contribution

K.Y.P., C.W. and H.M. contributed equally to this work.

Ancillary