Halobacterial A-ATP synthase in relation to V-ATPase

The ATP synthase in an extremely halophilic archaebacterium, Halobacterium halobium (salinarium), is an A-ATPase (Mukohata and Yoshida, 1987b; Mukohata and Ihara, 1990); it differs from the F-ATPase that had been considered to be ubiquitous throughout aerobic organisms and to be the central enzyme of ATP synthesis. The difference between the A-ATPase and F-ATPase was first shown enzymologically by the azide insensitivity of halobacterial ATP synthase (Mukohata and Yoshida, 1987a,b) and its A₁-ATPase (Nanba and Mukohata, 1987), then immunochemically, by the relatively small cross reaction of F-ATPase with an anti-halobacterial Ai-ATPase antibody (Mukohata et al. 1987a). (The halobacterial ATPase discussed here is the catalytic head-piece of the ATP synthase and thus corresponds to F₁-ATPase with respect to F₁F_o-ATP synthase. The ATPase is denoted as A₁-ATPase and the ATP synthase as A₁A_o-ATPase, when needed.) To our surprise, the antibody cross reacted with V-ATPase as much as it did with the A-ATPase of another kind of archaebacterium (Mukohata et al. 1987a). This close relationship between A-ATPases and V-ATPases and the remote relationship between F-ATPases and A/V-ATPases were confirmed by the identity (homology) values of the amino acid sequences deduced from the genes coding the ATPases (Denda et al. 1988a,b;,Inatomi et al. 1989; Ihara and Mukohata, 1991).

The relationship was then discussed with reference to the evolution of the proton-translocating ATPase family (Gogarten et al. 1989; Iwabe et al. 1989; Mukohata et al. 1990). The diversion of the A-ATPase from the V-ATPase took place long after the F-ATPase had diverged from the A/V-ATPase.

V-ATPases have been modified on the various endomembranes: vacuole, chromaffin granule, synaptosome, Golgi apparatus, and so on (Forgac, 1989). The A-ATPase has also been modified, possibly in different ways, in the extreme environments of archaebacterial habitats.

Halobacterial ATP synthase (A₁A_o-ATPase) is not inhibited by nitrate, whereas its 0:3/33 head piece (A₁-ATPase) is inhibited (Mukohata and Ihara, 1990). The V-ATPase in situ, which is thought to be equivalent to the ATP synthase (F₁F₀ and A₁ A_o), is inhibited by nitrate but its head piece does not hydrolyse ATP.

In this report, we describe some effects of nitrate and other chaotropic anions that differentiate A-ATPases and V-ATPases. We also describe some features which even differentiate the A-ATPase of one member of the Halobacteriaceae from that of other members.

Halobacterium halobium (salinarium) R₁mR was separated as a spontaneous mutant from Halobacterium halobium R₁ (from D. J. Kushner, University of Ottawa). Halobacterium halobium DSM670, Halobacterium saccharovorum DSM1137, Halobacterium sodomense ATCC33755, Haloferax volcanii ATCC29605, Haloarcula vallismortis ATCC29715, Halococcus morrhuae DSM1307 and Natronobacterium pharaonis DSM2160 were purchased from the Institute for Fermentation (Osaka). Halobacterium and Halococcus were cultivated in media as reported previously (Matsuno-Yagi and Mukohata, 1977). Haloferax and Haloarcula were cultivated in 15.6% NaCl, 2% MgSO₄-7H₂O, 0.4% KC1, 1.3% MgC1₂·6H₂O, 0.1% CaCl₂·2H₂O, 0.02% NaHCO₃, 0.05% KBr, 0.1% glucose and 0.5% yeast extract, pH 7.0. Natronobacterium was cultivated in 25% NaCl, 0.1% KH₂PO₄, 0.1% NH₄Cl, 0.1% sodium glutamate, 0.017% CaSO₄-2H₂O, 0.024% MgSO₄-7H₂O, 0.1% KC1, 0.5% casamino acid and 0.5% yeast extract at pH 9.0. All cultures were vigorously shaken for 1 week at 37–40°C. Cells were collected by centrifugation. Membrane vesicles were prepared by sonication as described by Mukohata et al. (1986). Protein concentration was determined by the Lowry method using bovine serum albumin as a standard.

Unless otherwise noted, ATP hydrolytic activity was determined at 40°C in a reaction mixture of 1.5 mol l⁻¹ Na₂SO₄, 10 mmoll⁻¹ MnSO₄, 4 mmol l⁻¹ ATP and 40mmoll⁻¹ Mes at pH 5.8 (Nanba and Mukohata, 1987). The reaction was started by adding the membrane vesicles, which had been solubilized with 0.7% Triton X-165 at a final protein concentration between 0.2 and 0.6 mg ml⁻¹. Liberated inorganic phosphate (P₁) was determined by the method of Taussky and Shorr (1953) or by the the Malachite Green method of Hess and Derr (1975). ATP synthetic activity was determined by the method of Mukohata et al. (1986), in which membrane vesicles of Halobacterium halobium were loaded with the reaction medium (1 moll⁻¹ NaCl or I moll⁻¹ sodium isethionate, 0.1 moll⁻¹ MgCl₂, 5 mmol l⁻¹ ADP, 20 mmol l⁻¹ P₁ and 50 mmol l⁻¹ Pipes at pH 6.8) by sonication, and washed with substrate-free medium. When needed, 10 mmol l⁻¹ chaotropic reagents were added to the medium. ATP was synthesized by base-acid transition (ΔpH=2.8) (Mukohata et al. 1986) and determined by the luciferin-luciferase method.

Western blotting was performed as described by Towbin et al. (1979). Proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) as described by Laemmli (1970), transferred to a polivinylidene difluoride membrane (Millipore, Bedford) and hybridized with an antibody raised in rabbits against each of the subunits (α and β) of Halobacterium halobium A₁-ATPase. After the second hybridization with alkaline-phosphatase-conjugated goat anti-rabbit IgG, the hybridized proteins were visualized by mixing them with 5-bromo-4-chloro-3-indolylphosphate and tetrazolium.

All other reagents were purchased from Wako Chemical Co. (Osaka) and Nacalai Chemicals (Kyoto).

A₁-ATPase showed complex kinetics for ATP hydrolysis in a 1.5 mol l⁻¹ Na2SO4 assay solution, as was found for the Halobacterium saccharovorum ATPase in 4 mol l⁻¹ NaCl (Schobert and Lanyi, 1989) (Fig. 1). Apparently, the rate of P_i release consists of two components: the rate V₁ found within the initial 1 min, which is not affected by nitrate concentration, and the rate V₁ after 3 min, which is affected by nitrate concentration. Here, we discuss only the nitrate-sensitive V₂. Lineweaver-Burk plots show that the nitrate inhibition is noncompetitive with a fixed K_m for ATP of2.5 mmol l⁻¹ and an apparent K_i for nitrate of about 3 mmol l⁻¹ (data not shown).

A₁-ATPase activity was examined for the effects of several anions, most of which appear in the Hofmeister series (Table 1). Nitrate inhibits the ATPase whereas other reagents, even more chaotropic ones, do not. Although chloride and acetate seem to activate ATPase activity to some extent, only chloride increased the I50 value of nitrate by sixfold (Table 2). When the potassium salts of these anions were added, similar results were obtained.

The effects of these anions on ATP synthesis in the membrane vesicles of Halobacterium halobium were also examined (Table 3). Just as it had little effect in solutions containing 1 moll⁻¹ NaCl as supporting salt (the salt enzymes to maintain their functions), nitrate did not inhibit ATP synthesis in solutions containing 1 mol l⁻¹ sodium isethionate, which does not protect the ATPase from nitrate inhibition (Table 2). Chaotropic reagents much stronger than nitrate, such as thiocyanate and trichloroacetate, did inhibit ATP synthesis.

The subunit sizes and salt-dependency of the ATPases of Halobacteriaceae were examined. Antibodies raised against each of the subunits (αand β) of Halobacterium halobium A₁-ATPase were tested for their cross reactions to the whole-cell proteins of various halobacterial strains. The sizes of the two major subunits of all the strains were similar to the individual subunits of Halobacterium halobium ATPase: 86×10³M_r(a subunit) and 64×10³M_r(J3 subunit) on SDS-PAGE (Fig. 2).

Membrane vesicles prepared from various halobacteria showed little ATPase activity without the addition of detergent. After the vesicles had been solubilized with Triton X-165, ATPase activity became detectable (Table 4). As has been demonstrated with Halobacterium halobium (Nanba and Mukohata, 1987), this result is found because the vesicles were in the right-side-out configuration with catalytic A₁ sectors inside. ATP was hydrolysed only after the membranes had been solubilized and the A₁-ATPase had been released. The ATPases of Halobacterium halobium R₁rnR and DSM670 were much more active in 1.5 mol l⁻¹ Na₂SO₄ at pH 5.8 than in 4 mol l⁻¹ NaCl at pH 7.0. In contrast, the ATPases of Halobacterium saccharovorum, Halobacterium sodomense and Haloarcula vallismortis were more active in 4 mol l⁻¹ NaCl than in 1.5 mol l⁻¹ Na₂SO₄. The ATPase of Haloferax volcanii exhibited only low activity in both media.

A-ATPases and V-ATPases, like F-ATPases, are H⁺-translocating enzymes composed of a catalytic head piece, a stem and a membrane-embedded channel (Pedersen and Carafoli, 1987a; Penefsky and Cross, 1991; Mukohata et al. 1992). A-ATPases and F-ATPases usually synthesize ATP under the proton-motive force, whereas V-ATPases only hydrolyse ATP to pump protons across membrane into endosomal vesicles. The halobacterial ATP synthase of present interest is a type of A-ATPase (A₁A₀) that is similar to V-ATPases (V₁V₀). The halobacterial ATPase under discussion is the head piece separated from the ATP synthase and thus designated the A₁-ATPase. The halobacterial A₁-ATPase appears to be composed of α₃ β₃ subunits (Mukohata et al. 1991); no other subunit is needed for ATP hydrolysis. The F₁-ATPase is composed of α₃ β₃ γor α₃ β₃ δas the functional complex (Yoshida et al. 1977), although an α₃ β₃ complex of the thermophilic bacterium PS3 has been shown to hydrolyse ATP (Miwa and Yoshida, 1989).

Halobacterium halobium A1-ATPase is inhibited by nitrate with an I50 value of about 3mmol l⁻¹ in 1.5moll⁻¹ Na₂SO₄ (Mukohata and Ihara, 1990). Since reagents that are even more chaotropic than nitrate, such as thiocyanate and chlorate, were less effective inhibitors (Table 1), one can conclude that nitrate inhibited the A₁-ATPase specifically. The addition of chloride or acetate apparently activated the ATPase; this effect became very obvious as the concentration of chloride or acetate increased. However, the I50 value increased as the chloride concentration was increased, but not as the acetate, isethionate or sulfate concentration was increased. Therefore, chloride seems to protect the ATPase from nitrate inhibition to some extent. This protective effect of chloride is not due merely to the increase in concentration of salts that support the integrity of the enzyme and/or activate it.

100mmoll⁻¹ nitrate completely inhibits the halobacterial A₁-ATPase. However, removal of nitrate by dialysis restores the initial activity of the sample to the level observed before the addition of nitrate (data not shown). Reversibility of nitrate inhibition was also reported for Halobacterium saccharovorum ATPase (Schobert and Lanyi, 1989).

The activity of Halobacterium halobium ATPase is very low in solutions of NaCl or KC1 as the supporting salt, even at 4 mol l⁻¹, and is much higher in solutions of Na₂SO₃ (1.5moll⁻¹) and sodium citrate (0.8moll⁻¹) (Nanba and Mukohata, 1987). Halobacterium saccharovorum ATPase (Hochstein et al. 1987; Stan-Lotter and Hochstein, 1989; Schobert and Lanyi, 1989) and some halobacterial ATPases show higher activity at pH 7.0 in 4 mol l⁻¹ NaCl than in 1.5 mol l⁻¹ Na₂SO₄ (Table 4). Therefore, the requirement for a high (supporting) salt concentration of halobacterial ATPases in general is largely due to a salting-out type of action, which makes the subunit complex tight enough to hydrolyse ATP. The tightness needed would probably differ among ATPases even in the halobacterial family, depending on the structures of the subunits (see below). The preference for the supporting salt could be explained in a similar way. Activating effects of SO₄²⁻ and HSO₃⁻ are reported in the A₁-ATPase of Sulfolobus acidocaldarius (Konishi et al. 1987; Lübben et al. 1987) and Methanosarcina barkeri (Inatomi, 1986). However, the concentrations of the supporting salts in these reaction media were very low (20–30 mmol l⁻¹). The effect of sulfate may not be same on all A-ATPases.

Nitrate only inhibits the halobacterial ATP synthesis slightly in 1 mol l⁻¹ NaCl, even at 100mmoll⁻¹, a level that almost completely inhibits ATP hydrolysis in 1.5 mol l⁻¹ Na₂SO₄ (Mukohata and Ihara, 1990). As in the case of ATP hydrolysis (Table 2), 1 mol l⁻¹NaCl loaded into the vesicles may protect ATP synthesis from nitrate inhibition. However, ATP is synthesized in 1 mol l⁻¹ sodium isethionate in amounts similar to or even greater than those in 1 mol l⁻¹ NaCl. Nitrate does not inhibit the ATP synthesis either in sodium isethionate or in NaCl (Table 3). Since isethionate does not protect ATP hydrolysis from nitrate inhibition (Table 2), these results indicate that ATP synthesis is not affected by nitrate. The ATP synthase of Halobacterium halobium is the membrane-anchored holoenzyme (A₁A_o), whereas the isolated Ai-ATPase is the α₃ β₃ head piece of the synthase. There is probably a nitrate binding site(s), which is characterized by non-competitive inhibition. When the head piece is detached as the A,-ATPase from the ATP synthase, the nitrate binding site should be exposed or its affinity for nitrate should be increased, which eventually inhibits ATP hydrolysis.

Slight inhibition of ATP synthesis by nitrate may be due to the chaotropic nature of nitrate. As shown in Table 3, the degree of inhibition of ATP synthesis roughly follows the Hofmeister series: SCN⁻, CCl₃COO⁻ >1⁻ > ClO₄⁻ > NO₃⁻ Additions of chaotropic anions to the external medium also depressed ATP synthesis. Moreover, these anions caused a faster decay of the light-induced pH shift of the membrane vesicle suspension (data not shown). Therefore, the reagents seem to make the membrane leaky to protons, which should diminish the proton-motive force and decrease ATP synthesis.

By contrast, chaotropic anions destroy V-ATPases by releasing the head piece, which carries no ATP hydrolysing activity (Rea et al. 1987; Bowman et al. 1989; Moriyama and Nelson, 1989; Arai et al. 1989). Halobacterial A₁-ATPase was not released from the membrane vesicles to any marked extent by various chaotropic anions (data not shown). Thus, the mode of action of chaotropic anions differs between A-ATPases and V-ATPases.

By analogy with F-ATPase, the N,N′-dicyclohexylcarbodiimide (DCCD)-binding protein is considered to be part of the membrane-embedded component of H⁺-translocating A-ATPase/synthase. [¹⁴C]DCCD labeled two polypeptides of 78 and l2×10³M_r (by SDS-PAGE) in parallel with the inhibition of ATP synthesis in Halobacterium halobium (Mukohata et al. 1987a,b). The A-cyclohexyl N′-[4-(dimethylamide)-a-naphthyl] carbodiimide (NCD-4)-binding protein of 10×10³M_r (by gel filtration in the presence of SDS) was isolated (K.-I. Sugimura, S. Watanabe, K. Ihara and Y. Mukohata, unpublished results). The DCCD-binding protein of 10 362 Da was also isolated from Sulfolobus acidocaldarius (Denda et al. 1989). The size of this DCCD-binding protein is almost in the range of the c subunit of F₁F₀-ATPase but not in the range of the corresponding subunit of V-ATPase, which has a duplicated mass of l6 ×lO³M_r (Mandel et al. 1988; Nelson and Nelson, 1989). This is another difference between A-ATPases and V-ATPases.

The relative molecular masses of the two major subunits of Halobacterium halobium ATPase are similar to those of other A-ATPases and V-ATPases (Mukohata et al. 1991, 1992). The apparent relative molecular masses of the Halobacterium halobium ATPase subunits, estimated by SDS-PAGE, are 86×10³ for the α subunit and 64×10³ for the β subunit (Nanba and Mukohata, 1987). These M_r values are much larger than those estimated from amino acid sequences deduced from the encoding genes: 64 104 Da for the α1; subunit and 51956 Da for the β subunit (Ihara and Mukohata, 1991). This discrepancy is probably due to the excess content of acidic amino acids (approximately 20%) that is characteristic of halophilic enzymes. Western blot analysis using the antibody to Halobacterium halobium ATPase revealed that various halobacterial strains carry A-ATPase subunits of similar sizes (Fig. 2). The size diversity is a little wider among the β subunits than among the α subunits, which carry the ATP-binding region. Since the sizes as well as the amino acid sequences of the subunits are well conserved, even among the A-ATPases of the three different groups (Denda et al. 1988a,b; Inatomi et al. 1989; Ihara and Mukohata, 1991), the actual sizes of these ATPase subunits of Halobacteriaceae should be similar. The apparently larger subunit sizes on SDS-PAGE suggest that, like Halobacterium halobium ATPase, all of these subunits contain large amounts of acidic amino acids (Ihara and Mukohata, 1991). Such high contents (and the distribution) of acidic amino acids probably result not only in the apparent diversity of the subunit sizes on SDS-PAGE but also in the different salt preferences.