Structure and properties of the coated vesicle (H+)-ATPase

Clathrin-coated vesicles play an important role in both receptor-mediated endocytosis and intracellular membrane traffic in eukaryotic cells. The coated vesicle (H⁺)-ATPase functions to provide the acidic environment within endosomes and other intracellular compartments necessary for receptor recycling and intracellular membrane traffic. The coated vesicle (H⁺)-ATPase is composed of nine different subunits which are divided into two distinct domains. The peripheral V₁ domain, which has the structure 73₃:58₃:40₁:34₁:33₁, possesses the nucleotide binding sites of the (H⁺)-ATPase. The integral V₀ domain, which has the composition 100₁:38₁:19₁:17₆, contains the pathway for proton conduction across the membrane. Topographical analysis indicates a structure for the coated vesicle (H⁺)-ATPase very similar to that of the F-type ATPases. Reassembly studies have allowed us to probe the function of particular subunits in this complex and the activity properties of the separate domains. These studies have led to insights into possible mechanisms of regulating vacuolar acidification.     

Cysteine 254 of the 73-kDa A subunit is responsible for inhibition of the coated vesicle (H+)-ATPase upon modification by sulfhydryl reagents.

The vacuolar class of (H+)-ATPases are highly sensitive to sulfhydryl reagents, such as N-ethylmaleimide. The cysteine residue which is responsible for inhibition of the coated vesicle (H+)-ATPase upon modification by N-ethylmalemide is located in subunit A and is able to form a disulfide bond with the cysteine moiety of cystine through an exchange reaction. This unique property distinguishes this cysteine residue from the remaining cysteine residues of the (H+)-ATPase. Using this reaction, we selectively labeled the cystine-reactive cysteine residue of subunit A with fluorescein-maleimide. After complete digestion of the labeled subunit A by V8 protease, a single labeled fragment of molecular mass 3.9 kDa was isolated and the amino-terminal sequence was determined. This fragment contains 2 cysteine residues, Cys240 and Cys254. Since Cys254 is conserved among all vacuolar (H+)-ATPases whereas Cys240 is not, it is likely that Cys254 is the residue which is responsible for the sensitivity of the vacuolar (H+)-ATPase to sulfhydryl reagents.     

Immunolocalization of the vacuolar-type (H+)-ATPase from clathrin-coated vesicles.

Proton-translocating ATPases of the vacuolar class (V-ATPases) are found in a variety of animal cell compartments that participate in vesicular membrane transport, including clathrin-coated vesicles, endosomes, the Golgi apparatus, and lysosomes. The exact structural relationship that exists among the V-ATPases of these intracellular compartments is not currently known. In the present study, we have localized the V-ATPase by light and electron microscopy, using monoclonal antibodies that recognize the V-ATPase present in clathrin-coated vesicles. Localization using light microscopy and fluorescently labeled antibodies reveals that the V-ATPase is concentrated in the juxtanuclear region, where extensive colocalization with the Golgi marker wheat germ agglutinin is observed. The V-ATPase is also present in approximately 60% of endosomes and lysosomes fluorescently labeled using alpha 2-macroglobulin as a marker for the receptor-mediated endocytic pathway. Localization using transmission electron microscopy and colloidal gold-labeled antibodies reveals that the V-ATPase is present at and near the plasma membrane, alone or in association with clathrin. These results provide evidence that the V-ATPases of plasma membrane, endosomes, lysosomes, and the Golgi apparatus are immunologically related to the V-ATPase of clathrin-coated vesicles.     

Dual activity of the H1-H2 domain of the (Na(+)+K+)-ATPase

(Na⁺+K⁺)-ATPase is a target receptor of digitalis (cardiac glycoside) drugs. It has been demonstrated that the H1-H2 domain of the α-subunit of the (Na⁺+K⁺)-ATPase is one of the digitalis drug interaction sites of the enzyme. Despite the extensive studies of the inhibitory effect of digitalis on the (Na⁺+K⁺)-ATPase, the functional property of the H1-H2 domain of the enzyme and its role in regulating enzyme activity is not completely understood. Here we report a surprise finding: instead of inhibiting the enzyme, binding of a specific monoclonal antibody SSA78 to the H1-H2 domain of the (Na⁺+K⁺)-ATPase elevates the catalytic activity of the enzyme. In the presence of low concentration of ouabain, monoclonal antibody SSA78 significantly protects enzyme function against ouabain-induced inhibition. However, higher concentration of ouabain completely inactivates the (Na⁺+K⁺)-ATPase even in the presence of SSA78. These results suggest that the H1-H2 domain of the (Na⁺+K⁺)-ATPase is capable of regulating enzyme function in two distinct ways for both ouabain-sensitive and -resistant forms of the enzyme: it increases the activity of the (Na⁺+K⁺)-ATPase during its interaction with an activator; it also participates in the mechanism of digitalis or ouabain-induced inhibition of the enzyme. Understanding the dual activity of the H1-H2 domain will help better understand the structure-function relationships of the (Na⁺+K⁺)-ATPase and the biological processes mediated by the enzyme.     

Subunit interactions in the clathrin-coated vesicle vacuolar (H(+))-ATPase complex.

The vacuolar (H(+))-ATPases (or V-ATPases) are structurally related to the F(1)F(0) ATP synthases of mitochondria, chloroplasts and bacteria, being composed of a peripheral (V(1)) and an integral (V(0)) domain. To further investigate the arrangement of subunits in the V-ATPase complex, covalent cross-linking has been carried out on the V-ATPase from clathrin-coated vesicles using three different cross-linking reagents. Cross-linked products were identified by molecular weight and by Western blot analysis using polyclonal antibodies raised against individual V-ATPase subunits. In the intact V(1)V(0) complex, evidence for cross-linking of subunits C and E, D and F, as well as E and G by disuccinimidyl glutarate was obtained, while in the free V(1) domain, cross-linking of subunits H and E was also observed. Subunits C and E as well as D and E could be cross-linked by 1-ethyl-3-(dimethylaminopropyl)carbodiimide, while subunits a and E could be cross-linked by 4-(N-maleimido)benzophenone. It was further demonstrated that it is possible to treat the V-ATPase with potassium iodide and MgATP in such a way that while subunits A, B, and H are nearly quantitatively removed, significant amounts of subunits C, D, E, and F remain attached to the membrane, suggesting that one or more of these latter subunits are in contact with the V(0) domain. In addition, treatment of the V-ATPase with cystine, which modifies Cys-254 of the catalytic A subunit, results in dissociation of subunit H, suggesting communication between the catalytic nucleotide binding site and subunit H. Finally, the stoichiometry of subunits F, G, and H were determined by quantitative amino acid analysis. Based on these and previous observations, a new structural model of the V-ATPase from clathrin-coated vesicles is proposed.     

Characterization of gastric mucosal membranes. X. Immunological studies of gastric (H+ + K+)-ATPase

Gastric mucosal homogenates from hog were fractionated by differential and density gradient centrifugation and free-flow electrophoresis. The two major membrane fractions (FI and FII) thus obtained are distinct both enzymically and in terms of transport reactivity. This heterogenicity extends to their antigenic activity. Purified antibodies which were raised against the K+-ATPase-containing H+ transport fraction FI were of two types: inhibitory and non-inhibitory. Inhibitory antibodies reduced the K+-ATPase activity by approximately 80% and the K+-p-nitro-phenylphosphatase activity by approximately 40% in a concentration-dependent manner, while the small Mg++-dependent component of the enzyme activity was unaffected. Antibodies inhibiting the K+-ATPase also inhibited H+ transport. These antibodies did not cross-react with the other major membrane fraction isolated by free- flow electrophoresis, FII, and gave a single band on rocket immunoelectrophoresis. Antibodies against this FII fraction also did not react with the K+-ATPase and were heterogeneous, giving at least four bands with rocket immunoelectrophoresis and inhibiting both the 5'- nucleotidase and Mg++-ATPase of this fraction. Immunofluorescent staining of tissue sections showed that the FI was derived from the parietal cell of gastric tissue and was localized to the supranuclear area of the cell. Staining of isolated rat gastric cell suspensions by FI antibodies confirmed the selectivity of the antibody and showed a polar, plasma membrane localization. FII antibodies also largely stained the parietal cells in tissue sections. In the 16 hog tissues tested, FI antibodies cross-reacted only with gastric fundus, thyroid and weakly with thymus. Immunoelectronmicroscopy showed that FI antibodies reacted strongly with the secretory membrane at the apical cell surface of the parietal cells and at the secretory canaliculi, weakly with the apical surface of the zymogen cell, and not with the basal-lateral surface of the cells. Thus, the protontranslocating ATPase is localized in the parietal cells and in the region postulated to be the site of acid secretion.     

Studies on (K+ + H+)-ATPase V. Chemical composition and molecular weight of the catalytic subunit

(1) A (K+ + H+)-ATPase preparation from porcine gastric mucosa is solubilized in sodium dodecyl sulfate, and is subjected to gel filtration. (2) A main subunit fraction is obtained, which is a protein carbohydrate lipid complex, containing 88% protein, 7% carbohydrate and 5% phospholipid. The Detailed composition of the protein and carbohydrate moieties are reported. (3) Sedimentation analysis of the subunit preparation, after detergent removal, reveals no heterogeneity, but the subunits readily undergo aggregation. (4) Acylation of the subunit preparation with citraconic anhydride causes a clear shift of the band obtained after SDS gel electrophoresis, but the absence of broadening and splitting of the band pleads against subunit heterogeneity. (5) Treatment of the subunit preparation with dansyl chloride indicates that the NH2 terminus is blocked, which favors the assumption of homogeneity of the protein. (6) Binding studies with concanavalin A indicate that at least 86% of the subunit preparation is composed of glycoprotein. (7) These findings, taken together, strongly suggest that there is a single subunit which is a glycoprotein and which represents the catalytic subunit of the enzyme. From sedimentation equilibrium analysis a molecular mass value of 119 kDa (S.E. 3, n = 6) is calculated for protein + carbohydrate and of 110 kDa (S.E. 3, N = 6) for protein only. (8) In combination with the molecular mass of 444 kDa (S. E. 10, n = 4) obtained for the intact enzyme by radiation inactivation we conclude that the enzyme appears to be composed of a homo-tetramer of catalytic subunits.     

Characterization of a P-type Ca(2+)-ATPase from Flavobacterium odoratum.

In most bacterial cell types studied, low intracellular free calcium is maintained by a variety of secondary exchangers which utilize transmembrane ion gradients. Prokaryotic calcium ATPases appear to be extremely uncommon, and none have been reported in Gram-negative organisms. We demonstrate ATP-dependent calcium uptake in everted membrane vesicles of Flavobacterium odoratum, a common Gram-negative soil and water bacterium. Calcium is transported with an apparent initial rate of 10 nmol/min mg of protein. It is inhibited by 20 microM orthovanadate, a specific P-type ATPase inhibitor, but significantly, it is unaffected by the addition of N-ethylmaleimide, N,N-dicyclohexylcarbodiimide, valinomycin, or nigericin. Because the Ca(2+)-ATPase makes up a high proportion of the total ATPase activity it is easily detected by a soluble ATP hydrolysis assay, with an initial rate for calcium-dependent ATPase activity in vesicles of 25-40 nmol/min.mg at pH 7.8 and 25 degrees C. The calcium-dependent activity is preferentially solubilized by the detergent C12E8 and can be precipitated at 55-80% ammonium sulfate in a fraction free of other contaminating ATPase activities. This partially purified fraction is enriched 15-fold and demonstrates an apparent Km for calcium of 2 microM, and for ATP of 130 microM. The IC50 for vanadate is 1.6 microM. These values are similar to those obtained for the eukaryotic sarcoplasmic reticulum calcium ATPase. The enzyme is rapidly phosphorylated by [gamma-32P]ATP in a calcium-dependent, vanadate-inhibitable manner. The phosphorylated species migrates with an apparent molecular mass of 60 kDa by NaDodSO4-polyacrylamide gel electrophoresis, and the phosphoryl group is sensitive to alkaline conditions, a characteristic of the acylphosphate linkage found in ATPases. These data demonstrate that the majority of calcium transport in F. odoratum is facilitated by a P-type ATPase.     

H+/ATP stoichiometry for the gastric (K++H+)-ATPase

Summary The initial rate of ATP-dependent proton uptake by hog gastric vesicles was measured at pH's between 6.1 and 6.9 by measuring the loss of protons from the external space with a glass electrode. The apparent rates of proton loss were corrected for scalar proton production due to ATP hydrolysis. For vesicles in 150mm KCl and pH 6.1, corrected rates of proton uptake and ATP hydrolysis were 639±84 and 619±65 nmol/min×mg protein, respectively, giving an H⁺/ATP ratio of 1.03±0.7. Furthermore, at all pH's tested the ratio of the rate of proton uptake to the rate of ATP hydrolysis was not significantly different than 1.0. No proton uptake (<10 nmol/min×mg protein) was exhibited by vesicles in 150mm NaCl at pH 6.1 despite ATP hydrolysis of 187±46 nmol/min×mg (nonproductive hydrolysis). Comparison of the rates of proton transport and ATP hydrolysis in various mixture of KCl and NaCl showed that the H⁺/ATP stoichiometries were not significantly different than 1.0 at all concentrations of K⁺ greater than 10mm. This fact suggests that the nonproductive rate is vanishingly small at these concentrations, implying that the measured H⁺/ATP stoichiometry is equal to the enzymatic stoichiometry. This result shows that the isolated gastric (K⁺+H⁺)-ATPase is thermodynamically capable of forming the observed proton gradient of the stomach.     

Papain fragmentation of the gastric (H+ + K+)-ATPase

Membrane-bound (H+ + K+)-ATPase purified from hog gastric mucosa was exposed to limited papain digestion. Such treatment resulted in a rapid inhibition of the K+-stimulated adenosine triphosphatase and p-nitrophenyl phosphatase activities, with about 90% of these activities lost after 3 min incubation at 37 degrees C with 0.1 units of papain per mg of enzyme protein. Parallel to the inhibition of the enzyme activities, there was a production of a 77 kDa membrane-bound fragment containing the aspartyl phosphate residue of the phospho-intermediate. This fragment accounted for about 45% of the total enzyme protein after the 3 min papain treatment. The digestion barely affected the steady-state level of phosphorylation, allowed the aspartyl phosphate of the 77 kDa fragment to undergo the transition to the E2P form, and did not significantly alter the fraction of ADP-sensitive phosphoenzyme. The presence of KCl, however, depressed the steady-state level of phosphoenzyme formed from [gamma-32P]ATP considerably less than that of the control enzyme. With further exposure to papain the 77 kDa peptide became fragmented into a 28 kDa soluble peptide that retained the phosphorylating site. Binding of fluorescein 5'-isothiocyanate (FITC) to the native enzyme did not affect the sites of papain hydrolysis because the same peptide fragments were obtained. The FITC reaction site was also in the 28 kDa soluble peptide fragment.     

Characterization of the ATP-binding domain of the sarco(endo)plasmic reticulum Ca(2+)-ATPase: probing nucleotide binding by multidimensional NMR.

The skeletal muscle sarco(endo)plasmic reticulum Ca(2+)-ATPase (SERCA1a) mediates muscle relaxation by pumping Ca(2+) from the cytosol to the ER/SR lumen. In efforts aimed at understanding the structural basis for the conformational changes accompanying the reaction cycle catalyzed by SERCA1a, we have studied the ATP-binding domain of SERCA1a in both nucleotide-bound and -free forms by NMR. Limited proteolysis analyses guided us to express a 28 kDa stably folded fragment containing the nucleotide-binding domain of SERCA1a spanning residues Thr357-Leu600. ATP binding activity was demonstrated for this fragment by a FITC competition assay. A nearly complete backbone resonance assignment of this 28 kDa ATP-binding fragment, in both the AMP-PNP-bound and -free forms, was obtained by means of heteronuclear multidimensional NMR techniques. NMR titration experiments with AMP-PNP revealed a confined nucleotide-binding site which coincides with a cytoplasmic pocket region identified in the crystal structure of apo-SERCA1a. These results are consistent with previous site-directed mutagenesis studies of SERCA1a.     

The action of trypsin on the gastric (H+ + K+)-ATPase.

The effect of trypsin on gastric (H+ + K+)-ATPase and K+-phosphatase was studied. Loss of both enzymic activities was biphasic, consisting of a fast and slow phase. Several peptides were produced from the original 105,000-dalton region of the sodium dodecyl sulfate electrophoretic separation, but only two, 87,000 and 47,000 daltons, were labeled following incubation with [gamma-33P]ATP. After a 30-min hydrolysis, 35% of the original peptide remained unaltered and appeared to be a glycoprotein. ATP and ADP abolished the second phase of tryptic inactivation of both activities and only two peptides, of 78,000 and 30,000 daltons, were found on the acrylamide gel in addition to the original 105,000-dalton region, neither of which was labeled by [gamma-33P]ATP. The protection was specific for these nucleotides, AMP, beta, gamma-methylene ATP, TTP, and pNPP being ineffective. Na+ and K+ at high concentrations reduced the rate of loss of activity but no change in the peptides produced was found. The level of phosphoenzyme was increased 2-fold by trypsin treatment, whereas the quantity of K+-sensitive phosphoenzyme remained relatively constant. Thus, the 105,000-dalton region is heterogeneous, consisting of a catalytic subunit (the active site is on a 47,000-dalton fragment), a glycoprotein, and another 105,000-dalton peptide. The action of trypsin is initially to prevent interconversion of a K+-insensitive to a K+-sensitive form of the phosphoenzyme, thus inhibiting hydrolysis.     

Characterization of ATPases of plain synaptic vesicle and coated vesicle fractions isolated from rat brains.

The plain synaptic vesicle and the ocated vesicle fractions were isolated from rat brains, and the ATPase [EC 3.6.1.3] activities were characterized in terms of ionic effects, drug effects, and protein components. Coated vesicle fraction contained three times as much actomysin-like proteins as plain vesicle fraction, although both fractions had an identical ratio of actin-like protein to myosin-like protein. The ATPases of these two fractions were activated by both Mg2+ and Ca2+, and, in the presence of either of the cations, were inhibited by KCl. Reserpine activated plain vesicle ATPase only in the presence of Cl-. Colchicine and vinblastine inhibited coated vesicle ATPase only. The results are consistent with the view that actomyosin-like proteins are involved in the synaptic retrieval process.     

Lack of immunological cross reactivity between the transport enzymes (Na+ + K+)-ATPase and (K+ + H+)-ATPase

Goat antisera against (Na⁺ + K⁺)-ATPase and its isolated subunits and against (K⁺ + H⁺)-ATPase have been prepared in order to test for immune cross-reactivity between the two enzymes, whose catalytic subunits show great chemical similarity. None of the (Na⁺ + K⁺)-ATPase antisera cross-reacted with (K⁺ + H⁺)-ATPase or inhibited its enzyme activity. The same was true for the (K⁺ + H⁺)-ATPase antiserum with regard to (Na⁺ + K⁺)-ATPase and its subunits and its enzyme activity. So not withstanding the chemical similarity of their subunits, there is no immunological cross-reactivity between these two plasma membrane ATPases.Number LIII in the series Studies on (Na⁺ + K⁺)-Activated ATPase.     

H+ transport by reconstituted gastric (H+ + K+)-ATPase

Gastric (H+ + K+)-ATPase was reconstituted into artificial phosphatidylcholine/cholesterol liposomes by means of a freeze-thaw-sonication technique. Upon addition of MgATP, active H+ transport was observed, with a maximal rate of 2.1 mumol X mg-1 X min-1, requiring the presence of 100 mM K+ at the intravesicular site. However, in the absence of ATP an H+-K+ exchange with a maximal rate of 0.12 mumol X mg-1 X min-1 was measured, which could be inhibited by the well-known ATPase inhibitors vanadate and omeprazole, giving the first evidence of a passive K+-H+ exchange function of gastric (H+ + K+)-ATPase. An Na+-H+ exchange activity was also measured, which was fully inhibited by 1 mM amiloride. Simultaneous reconstitution of Na+/H+ antiport and (H+ + K+)-ATPase could explain why reconstituted ATPase appeared less cation-specific than the native enzyme (Rabon, E.C., Gunther, R.B., Soumarmon, A., Bassilian, B., Lewin, M.J.M. and Sachs, G. (1985) J. Biol. Chem. 260, 10200-10212).     

Transport ratios of reconstituted (H+ + K+)-ATPase

Gastric (H+ + K+)-ATPase was reconstituted into artificial phosphatidylcholine/cholesterol vesicles by means of a freeze-thaw-sonication procedure. The passive and active transport mediated by these vesicles were measured (Skrabanja, A.T.P., Asty, P., Soumarmon, A., De Pont, J.J.H.H.M. and Lewin, M.J.M. (1986) Biochim. Biophys. Acta 860, 131-136). To determine real initial velocities, the proteoliposomes were separated from non-incorporated enzyme, by means of centrifugation on a sucrose gradient. The purified proteoliposomes were used to measure active H+ and Rb+ transport, giving at room-temperature velocities of 46.3 and 42.5 mumol per mg per h, respectively. A transport ratio of two cations per ATP hydrolyzed was also measured. These figures indicate that the enzyme catalyzes an electroneutral H+-Rb+ exchange.     

Molecular cloning of the rat stomach (H+ + K+)-ATPase

We have isolated cDNA clones for the rat stomach (H+ + K+)-ATPase by employing a novel procedure involving the use of oligonucleotides corresponding to conserved amino acid sequences of related cation transport ATPase and a cross-hybridization with the sheep kidney (Na+ + K+)-ATPase alpha-subunit cDNA. The complete nucleotide sequence of the cDNA has been determined and the amino acid sequence of the protein deduced. The ATPase consists of 1,033 amino acids and has an Mr of 114,012. Amino acid homology and hydropathy plot comparisons between the gastric ATPase and the (Na+ + K+)-ATPase catalytic subunit demonstrate a striking similarity which suggests that their higher order structure and mechanism of action are virtually identical. The greatest homology occurs in the phosphorylation site region and in domains which may be involved in nucleotide binding and energy transduction. The most substantial differences occur in the N-terminal region and in the transmembrane domains. In addition, we report the presence of an open reading frame 5' to the translation initiation site of the gastric ATPase, which raises the possibility that the mRNA is polycistronic.     

Interaction of fluorescein isothiocyanate with the (H+ + K+)-ATPase

Fluorescein isothiocyanate was used to covalently label the gastric (H+ + K+)-ATPase. FITC treatment of the enzyme inhibited the ATPase activity while largely sparing partial reactions such as the associated p-nitrophenylphosphatase activity. ATP protected against inhibition suggesting the ligand binds at or near an ATP binding site. At 100% inhibition the stoichiometry of binding was 1.5 nmol FITC per mg Lowry protein a value corresponding to maximal phosphoenzyme formation. Binding occurred largely to a peptide of 6.2 isoelectric point, although minor labelling of a peptide of pI 5.6 was also noted. Fluorescence was quenched by K+, Rb+ and Tl+ in a dose-dependent manner, and the K0.5 values of 0.28, 0.83 and 0.025 mM correspond rather well to the values required for dephosphorylation at a luminal site. Vanadate, a known inhibitor of the gastric ATPase produced a slow Mg2+-dependent fluorescent quench. Ca2+ reversed the K+-dependent loss of fluorescence and inhibited it when added prior to K+. This may relate to the slow phosphorylation in the presence of ATP found when Ca2+ was substituted for Mg2+ and the absence of K+-dependent dephosphorylation. The results with FITC-modified gastric ATPase provide evidence for a conformational change with K+ binding to the enzyme.     

Studies on (K+ + H+)-ATPase III. Binding of adenylyl imidodiphosphate

1. Adenylyl imidodiphosphate (AMPPNP) binds to (K+ + H+)-ATPase from pig gastric mucosa with a dissociation constant (Kd) of 50 microM for the AMPPNP-enzyme complex. 2. Monovalent cations reduce the amount of AMPPNP bound in the following order of effectiveness Tl+ greater than K+ greater than Rb+ greater than Cs+ greater than Na+, Li+, choline+. 3. AMPPNP binding to the enzyme has a pH optimum at pH 7.0--7.5 in the absence of added ions, which is shifted to pH 8 upon addition of MgCl2. 4. Cyclodiaminotetraacetic acid (CDTA, Tris salt) inhibits binding of AMPPNP. This inhibition is not due to chelation of Mg2+. It may be due to direct binding of CDTA to the enzyme or to removal of stabilizing cations other than Mg2+. 5. Binding curves determined in the presence of various concentrations of Mg2+ show that at low Mg2+ concentrations (less than 0.5 mM), the apparent number of binding sites is reduced, while at higher Mg2+ concentrations (greater than or equal to 0.5 mM), the binding of AMPPNP is inhibited in a competitive way. 6. From these observations it is concluded that the enzyme has two binding sites for AMPPNP and only one for Mg-AMPPNP (or two with strong anti-cooperativity), and that Mg2+ inhibits binding of Mg-AMPPNP. This finding is interpreted in terms of a model involving a dimeric form of the enzyme.