Conserved Asp684 in Transmembrane Segment M6 of the Plant Plasma Membrane P-type Proton Pump AHA2 Is a Molecular Determinant of Proton Translocation*

The mechanism of proton pumping by P-type H+-ATPases is still unclear. In the plant P-type plasma membrane H+-ATPase AHA2, two charged residues, Arg655 and Asp684, are conserved in transmembrane segments M5 and M6, respectively, a region that has been shown be contribute to ion coordination in related P-type ATPases. Substitution of Arg655 with either alanine or aspartate resulted in mutant enzymes exhibiting a significant shift in the P-type ATPase E 1P-E 2P conformational equilibrium. The mutant proteins accumulated in theE 1P conformation, but were capable of conducting proton transport. This points to an important role of Arg655 in theE 1P-E 2P conformational transition. The presence of a carboxylate moiety at position Asp684 proved essential for coupling between initial proton binding and proton pumping. The finding that the carboxylate side chain of Asp684 contributes to the proton-binding site and appears to function as an absolutely essential proton acceptor along the proton transport pathway is discussed in the context of a possible proton pumping mechanism of P-type H+-ATPases.

transfer of ion(s) across the membrane (8). High resolution three-dimensional crystal structures of the E 1 and the E 2 conformations, respectively, have been solved for the animal SR Ca 2ϩ -ATPase (7,9). These structures indicate substantial domain movements of P-type ATPases in both cytoplasmic and transmembrane parts during catalysis. Significant progress has been made to map the transport pathways in representative P-type ATPases such as Ca 2ϩ -and Na ϩ /K ϩ -ATPases (7, 9 -16). These studies have established the presence of cationbinding pockets, each of which are formed by several coordinating groups situated in transmembrane helices M4, M5, M6, and M8. However, the molecular determinants of cation specificity have not been determined for any P-type ATPase and the mechanism of proton transport by P-type ATPases is not well characterized.
Protons in solution (as H 3 O ϩ ) form solvation structures characteristic of cationic solvation. This suggests that the plant H ϩ -ATPase possibly could mediate hydronium ion transport (17). However, in a number of different proton pumps that have been described in molecular detail, H ϩ seems to be the transported species, and H ϩ is in these proton pumps believed to be transported along a proton transduction chain (a proton wire) comprising alternating groups of proton donors and proton acceptors. Whether P-type proton pumps also could mediate H ϩ transport along a proton transduction chain is unknown.
In this study, we investigated whether two charged residues, Arg 655 AHA2 and Asp 684 AHA2 , situated in transmembrane segments M5 and M6, respectively, might play a role in the proton pumping mechanism of plasma membrane H ϩ -ATPase. Transmembrane segment M5 and M6 are connected by a very short stretch of extracellularly located amino acid residues and are therefore expected to form a hairpin loop. During proteolysis of the related Na ϩ /K ϩ -and H ϩ /K ϩ -ATPases, inclusion of potassium ions stabilizes the M5M6 hairpin loop, and it has been suggested that this moiety moves in and out of the membrane during catalysis (18,19). It is evident from the crystal structures of SERCA1 Ca 2ϩ -ATPase in two conformations (7,9) that the M5M6 segment is a mobile structure, somehow coupling ATP hydrolysis to cation transport. The selected residues Arg 655 AHA2 and Asp 684 AHA2 are conserved in all plant plasma membrane H ϩ -ATPases and are close to the predicted protonbinding site in atomic models of plant (20) and fungal (21) P-type H ϩ -ATPases. Asp 684 AHA2 is strictly conserved in H ϩ -, Na ϩ /K ϩ -(Asp 808 ), H ϩ /K ϩ -(Asp 804 ), and Ca 2ϩ -ATPases (Asp 800 ). In the SERCA1 Ca 2ϩ -ATPase it coordinates both of the two Ca 2ϩ ions bound (7). In a homology model of AHA2 (20), Arg 655 AHA2 seems to be equivalent to Glu 771 SERCA , a residue involved in direct coordination of Ca 2ϩ in SERCA1. One of the residues, the negatively charged Asp 684 AHA2 , has previously been implicated in enzyme conformational changes in the plant H ϩ -ATPase (22). The other, the positively charged Arg 655 AHA2 , has not been assigned a role so far.

EXPERIMENTAL PROCEDURES
Expression in Yeast-The Saccharomyces cerevisiae strain RS-72 (23) was transformed and cultured essentially as described previously (24). In RS-72 (MATa ade1-100 his4 -519 leu2-3, 112), the natural constitutive promoter of the endogenous yeast plasma membrane H ϩ -ATPase PMA1 is replaced by the galactose-dependent promoter of GAL1. Since PMA1 is essential for yeast growth, RS-72 grows only on galactose medium. Using this strain, plasmid-borne plant H ϩ -ATPases brought under control of the constitutive PMA1 promoter can be tested for their ability to rescue pma1 mutants on glucose medium.
pma1 Complementation Test in Yeast-Yeast was grown for 3 days at 30°C in liquid medium containing 2% galactose. Approximately 10 3 cells in 10 l were spotted onto solid synthetic minimal medium containing either 2% galactose at pH 5.5 or 2% glucose at pH values of 6.5, 5.5, and 4.5, respectively. Growth was recorded after incubation for 3 days at 30°C.
Site-directed Mutagenesis-As a starting point for mutagenesis experiments, a modified cDNA of the Arabidopsis thaliana plasma membrane H ϩ -ATPase isoform AHA2 under control of the PMA1 promoter subcloned into the yeast multicopy vector YEp-351 (25) was used. The modified cDNA (pMP-1280; Ref. 26) has eight unique novel restriction enzyme sites created by silent mutagenesis for facilitation of cassette mutagenesis and furthermore encodes the plant H ϩ -ATPase AHA2 with (i) a COOH-terminal deletion of 73 amino acid residues and (ii) a COOH-terminal Met-Arg-Gly-Ser-His 6 (MRGSH 6 ) tag. The COOH-terminal deletion transforms the enzyme into a constitutively activated state, which simplifies functional studies. Addition of the His 6 tag allows purification of the recombinant enzyme by affinity chromatography, whereas the additional residues (Met-Arg-Gly-Ser) provide the H ϩ -ATPase with a convenient antibody epitope.
Site-directed mutagenesis was performed by overlap extension polymerase chain reaction (27) with mutagenic primers designed to replace the codons for Arg 655 and Asp 684 . The codon for Arg 655 (CGT) was replaced with codons encoding Lys (AAA), Ala (GCC), and Asp (GAC), respectively. The codon for Asp 684 (GAC) was replaced with codons for Glu (GAA), Asp (AAC), Ala (GCC), Val (GTC), and Arg (CGC), respectively. Products of the polymerase chain reaction were used to replace the corresponding fragments in pMP-1280. DNA sequencing of the cassettes was used to verify the presence of the desired mutations and the absence of unwanted base changes. Arg 655 and Asp 684 double mutants were constructed by replacing respective fragments in Asp 684 mutant backgrounds with Arg 655 mutations.
Purification of ATPase by Ni 2ϩ Affinity Chromatography-Growth of yeast for protein purification was the same as described in Ref. 22. Microsomal yeast membrane vesicles were isolated as described previously (24). Microsomal membranes were solubilized by the n-dodecyl-␤-D-maltoside at a detergent to protein ratio of 3:1, after which heterologously expressed plant H ϩ -ATPases were purified by Ni 2ϩnitrilotriacetic acid affinity chromatography as described (22), except that the elution buffer contained 0.5 mg/ml of L-phosphatidylcholine. Samples of pure protein were frozen in liquid nitrogen and stored at 80°C. All experiments were repeated with at least three independent membrane preparations.
Reconstitution of ATPase into Artificial Liposomes and Measurement of Proton Transport-Protein was reconstituted at a lipid to protein ratio of 200:1 as described previously (28). Reconstitution was performed in 10 mM Mes-KOH, pH 6.5, 50 mM K 2 SO 4 , and 20% (v/v) glycerol. Proton transport was measured as described (29) by monitoring fluorescence quenching of 9-amino-6-chloro-2-methoxyacridine (ACMA), a dye that upon protonation accumulates in an impermeant form inside vesicles. The reaction medium contained 10 mM Mes-KOH, pH 6.5, 1 mM ATP, 1 M ACMA, and 50 mM K 2 SO 4 . Proton pumping reactions were started by addition of MgSO 4 to a final concentration of 2 mM.
Gel Electrophoresis and Western Blotting-Membrane fractions or purified proteins were separated by SDS-PAGE on 10% acrylamide using the system of Fling and Gregerson (30). After electrotransfer of the proteins to an Immobilon-P membrane (Millipore), protein blots were probed with an antibody against the MRGSH 6 epitope present at the COOH terminus (Qiagen, Chatsworth, CA).
Phosphorylation and Dephosphorylation of the Phosphorylated Intermediate-Phosphorylation by [␥-32 P]ATP and dephosphorylation initiated with 10 mM EDTA or 10 mM EDTA supplemented 2 mM ADP were performed as described previously (22).
Protein Determination-Protein concentrations were determined by the method of Bradford (31) with the Bio-Rad protein assay reagent and employing bovine serum albumin as a standard.

RESULTS
Selection of Residues to Be Studied-Charged residues in the hairpin loop formed by transmembrane segments M5 and M6 have been shown to contribute to ion coordination in a number of related P 2 -type ATPases (7,32). Fig. 1 illustrates this region of the plant plasma membrane H ϩ -ATPase AHA2 aligned with the corresponding segment of SERCA1 Ca 2ϩ -ATPase, the only P-type ATPase for which high resolution structural data are available (7,9). In transmembrane segments M5 and M6 of this Ca 2ϩ -ATPase, the charged residues Glu 771 and Asp 800 , together with polar side chains, contribute to coordination of bound Ca 2ϩ . To explore the functional role of the charged residues Arg 655 and Asp 684 situated in the corresponding region of the AHA2 plant H ϩ -ATPase, site-directed mutants of these two positions were constructed and expressed in yeast.
Replacement of PMA1 in Yeast with Mutant Plant H ϩ -ATPase-In the yeast S. cerevisiae, PMA1 is an essential gene, encoding a P-type plasma membrane H ϩ -ATPase. To test whether any of the plant plasma membrane H ϩ -ATPase AHA2 substitutions were able to complement a pma1 mutation, each mutant was expressed in the yeast strain RS-72, in which the constitutive promoter of the PMA1 gene has been replaced by the galactose-dependent GAL1 promoter at the chromosomal level (23). Thus, this strain is only able to grow on galactose medium and fails to grow on other carbon sources unless the pma1 deficiency has been complemented. The plant plasma membrane H ϩ -ATPase employed (aha2⌬73; termed wild type (WT) in this study for clarity) carries a truncation of 219 base pairs in the 3Ј end of the cDNA. As nucleotides encoding the COOH-terminal auto-inhibitory domain of AHA2 in this way are removed, this construct readily complements pma1 (22).
Among the Arg 655 substitutions tested, only the R655K substitution was able to fully support yeast growth, whereas the R655A and R655D mutants supported yeast growth at a significantly reduced rate (Fig. 2). Of the Asp 684 -substituted proteins, only D684E was able to support yeast growth to some extent. All other substitutions of Asp 684 (D684N, D684A, D684V, and D684R) failed to replace yeast Pma1p. Altogether, the complementation test pointed to the importance of both Arg 655 and Asp 684 for the proper functioning of plant plasma membrane H ϩ -ATPase. Expression and Purification of Mutant H ϩ -ATPases-To test whether mutant proteins were successfully produced in the heterologous host, isolated yeast microsomal membrane vesicles were examined by Coomassie Blue staining following separation of membrane polypeptides by SDS-polyacrylamide gel electrophoresis (Fig. 3A). In all cases a new band corresponding to a polypeptide with an apparent molecular weight of 92 kDa was observed. The identity of the polypeptides as plant H ϩ -ATPases was confirmed following immunoblotting using an antibody directed against the MRGSH 6 epitope fused to the COOH-terminal end of recombinant plant H ϩ -ATPase (Fig. 3B). Interestingly, mutant ATPase protein typically reached levels above those of the wild-type enzyme (Fig. 3B). Following heterologous expression in transgenic yeast, all of the plant H ϩ -ATPase mutants described in this work could be purified by Ni 2ϩ -nitrilotriacetic acid affinity chromatography to near homogeneity (Fig. 3C). This allowed for a detailed kinetic characterization of the enzymatic properties of the mutant enzymes.
ATP Hydrolysis and Kinetic Properties-A way to screen for qualitative changes in mutant H ϩ -ATPases is to measure several standard kinetic properties including V max , the K m for ATP, IC 50 for orthovanadate, and pH dependence. Notably, the apparent ATP affinity of all Arg 655 and Asp 684 substitutions remained essentially unaltered (Table I). This is evidence that the nucleotide-binding pocket was unaffected by the substitutions, strongly suggesting that the enzymes were correctly folded. However, when other kinetic parameters were determined for the eight mutants listed above, major differences were observed from the wild-type control.
Among the Arg 655 substitutions, all led to enzymes with significantly reduced ATP hydrolytic activities (Table I). Furthermore, the non-conservative substitutions R655A and R655D led to mutant proteins that were almost insensitive toward the inhibitor vanadate, which specifically binds to the E 2 conformation of P-type ATPases. Thus, importantly, removal of the positive charge at the position of Arg 655 seems to cause an appreciable shift in the equilibrium between the E 1 and E 2 conformational states of the enzyme, since such a shift would be expected to result in a change of K i for vanadate (2). The conservative substitution R655K did not change the vanadate sensitivity of the enzyme.
The pH dependence of P-type H ϩ -ATPases has often been taken as an indicative for their proton affinity, since the activating effect of protons on ATP hydrolytic activity resembles that of the transported cations in other pumps and with comparable affinities. Notably, as was observed in all cases (Fig.  4A), pH profiles of the Arg 655 substitutions were comparable with that of wild-type H ϩ -ATPase.
Substitutions at position Asp 684 similarly led to reduced ATP hydrolytic activities (Table I). In addition, most substitutions had dramatic effects on other kinetic properties of the enzyme. The Asp 684 substitutions were all more or less insensitive toward vanadate. In the D684E substitution, K i for vanadate was radically shifted, whereas the D684N, D684A, D684V, and D684R substitutions were completely vanadate-insensitive. This indicates that albeit all the tested substitutions are able to hydrolyze ATP, they all have difficulties proceeding from the E 1 to the E 2 conformation.
The conservative substitutions D684N and D684E resulted in a minor shift in the pH dependence profile toward more acidic values (Fig. 4B). The slight shift in the pH dependence of the H ϩ -ATPase (Fig. 4B) was more pronounced with non-conservative substitutions such as D684A and D684V. Further- more, a D684R substitution, which replaces a negative charge with a positive charge, produced a dramatic shift in pH dependence (Fig. 4B).
Proton Pumping-To ask whether any of the charged residues are essential for proton pumping, the eight enzymatically active mutants were further assayed for their ability to pump protons in an ATP-dependent manner (Fig. 5). Indeed, all the tested Arg 655 substitutions, including R655D, were able to carry out ATP dependent proton transport, albeit at significantly reduced rates (Fig. 5A). For all Arg 655 substitutions tested, proton transport seemed efficiently coupled to ATP hydrolysis (Table I).
However, among the Asp 684 substitutions, only D684E pumped protons at reduced rates (Fig. 5B). All other attempts to substitute Asp 684 resulted in mutant enzymes that were completely deficient in proton pumping (Fig. 5B). It is notable that mutants having non-conservative substitutions at Asp 684 possessed basal ATP hydrolytic activities (Table I). Thus, in all these mutants ATP hydrolysis appeared to be completely uncoupled from proton translocation. The proton pumping activity of the D684E mutant seemed efficiently coupled to ATP hydrolysis (Table I).   Table I.
The E 1 P-phosphorylated Intermediate Accumulates in R655A, R655D, and D684N Mutant H ϩ -ATPases-The amount of phosphorylated intermediate was measured under steadystate conditions for the R655K, R655A, R655D, D684E, D684N, D684A, D684V, and D684R mutants (Fig. 6A). Both the R655A and R655D H ϩ -ATPases accumulated in the form of the phosphorylated intermediate (EP) at levels significantly higher (approximately four times higher) than the WT ATPase (Fig. 6A). In contrast to the E 2 P conformational state of P-type ATPases, the E 1 P conformational state is able to transfer the covalently bound phosphate back to ADP. By initiating dephosphorylation of R655A and R655D proteins with or without ADP, the EP of R655A and R655D was demonstrated to be sensitive to the application of ADP (Fig. 6B). Thus, these proteins accumulate in the E 1 P conformational state, and the low level of proton transport by these proteins can be ascribed to a reduced rate of the E 1 P-E 2 P conformational transition.
Whereas the conservative D684E substitution resulted in an enzyme with EP half of those of the WT enzyme (Fig. 6A), the level of EP was found to be approximately five times higher in the D684N mutant as compared with the unmodified enzyme (Fig. 6A). The phosphorylated intermediate was sensitive toward application of ADP (Fig. 6B). This corresponds to previous observations that the D684N substitution is blocked in the E 1 P-E 2 P transition (22). If protons, like cations in other P-type ATPases, are occluded following phosphorylation and ADP dissociation, these results might imply that the E 1 P (H ϩ ) occluded state is greatly stabilized in the D684N mutant. However, whether protons get occluded within the pump molecule during catalysis is unknown.
With the non-conservative substitutions, accumulation of EP was much less evident. Thus, D684A and D684V failed to accumulate EP to the same degree as the D684N mutant and, with the D684R substitution, only negligible amounts of EP intermediate could be detected (Fig. 6A). These results indicate that the catalytic machinery in the D684A, D684V, and D684R mutants not only is blocked in the E 1 P-E 2 P conformational transition accompanying proton transport, but also is disturbed in a step before phosphorylation, namely the E 1 -E 1 P transition.
Genetic Analysis of Arg 655 and Asp 684 Double Mutants-In a three-dimensional structural model of the AHA2 H ϩ -ATPase (20), the side chains of Arg 655 and Asp 684 are located in close vicinity to each other. From an energetic point of view, arginine residues within the membrane spanning segments are expected to be either strongly hydrogen-bonded or members of ions pairs (33). Therefore, it seems possible that an electrostatic interaction between Arg 655 and Asp 684 could take place. In the yeast plasma membrane H ϩ -ATPase Pma1p, Arg 695 Pma1p and Asp 730 Pma1p in M5 and M6, respectively, were found to be involved in the formation of a structural important salt-bridge (34 Pma1p and Asp 730 Pma1p and insertions of alanines at these positions in the yeast H ϩ -ATPase resulted in active ATPases capable of ATP hydrolysis and proton transport at near normal rates (34).
To address the effect of complementary charge swapping and charge elimination in the plant H ϩ -ATPase, all the single point substituted enzymes analyzed in this work (R655K, R655A, R655D) and (D684E, D684N, D684A, D684V, D684R) were combined. Among the 15 double mutants, R655K,D684E could replace Pma1p to some extent (Fig. 7). All other double mutants failed to support growth of the pma1 mutant yeast strain (Fig. 7), indicating that neither simple charge neither swapping nor charge elimination do result in a functional proton pump.
Biochemical Characterization of Asp 684 and Arg 655 Double Mutants-Two mutant proteins, namely R655D,D684R (charge swapping) and R655A,D684A (charge elimination) were affin- ity purified to apparent homogeneity (data not shown). When reconstituted in lipid vesicles, no proton transport could be demonstrated for any of the double mutated species (Fig. 8B), although both exhibited ATP hydrolytic activity (R655D, D684R: 0.1 Ϯ 0.02 mol of P i produced per min/mg of protein and R655A,D684A: 1.9 Ϯ 0.1 mol of P i produced per min/mg of protein). Because the apparent affinities for ATP were comparable with that of the wild-type enzyme (R655D,D684R: 152 M Ϯ 32 M and R655A,D684A: 103 Ϯ 18 M), the nucleotide-binding site of both substitutions appeared to be intact. Since both double mutants seem to be folded correctly, the two residues in the plant protein are not likely to be engaged in an essential structural salt bridge.
R655D,D684R and R655A,D684A were completely vanadateinsensitive (data not shown). Interestingly, the pH dependence profiles of the double mutants could be super imposed upon the corresponding pH profiles of the respective Asp 684 single point substituted ATPases (D684R and D684A; Figs. 4 and 8A). In fact, for all of the kinetic parameters tested, the double mutants behaved similarly to the respective Asp 684 single point mutants. This would suggest that the rate-limiting step in the catalytic cycle of the Arg 655 and Asp 684 double mutants is determined by the Asp 684 substitution. DISCUSSION The goal of the present study was to determine the role in proton transport of conserved residues Arg 655 and Asp 684 situated in the middle of transmembrane segments M5 and M6, respectively, of the plant plasma membrane H ϩ -ATPase. Our results provide new insight into the mechanism by which Ptype proton pumps transport protons across the membrane.
The Positive Charge at Arg 655 Is Involved in the E 1 P-E 2 P Transition-All tested Arg 655 substitutions (R655K, R655A, and R655D) were capable of catalyzing ATP-coupled proton transport albeit with different efficiencies. Thus, whereas the R655K substituted enzyme was almost as efficient as the wildtype plant H ϩ -ATPase in carrying out ATP hydrolysis and proton transport, the two non-conservative substitutions (R655A and R655D) resulted in enzymes that were severely inhibited in both ATP hydrolysis and proton transport.
R655A and R655D mutant ATPases were almost insensitive toward vanadate, an inhibitor of the E 2 conformation, but, as judged from their slight ability to carry out ATP dependent proton transport, both mutants were able to carry out a complete catalytic cycle. The phosphorylated intermediate accumulated in the R655A and the R655D proteins to levels corresponding to 3-4 times the value in the unmodified protein. It was demonstrated that the high level of EP in these proteins could be ascribed the E 1 P phosphoform. Based on the unaltered pH dependence and this high level of E 1 P accumulating in the R655A and R655D H ϩ -ATPases, this study do not suggest a role for Arg 655 in the initial proton-binding event allowing E 1 P formation to proceed. Rather, in R655A and R655D, the E 1 P-E 2 P conformational transition occurs at a significantly reduced rate. Arginine is the most basic amino acid (pK a ϳ12.5), suggesting that Arg 655 is likely to be found in the protonated form. We conclude that the positive charge of Arg 655 must be ascribed an important, although not essential, role in the E 1 P-E 2 P conformational transition. We did not find evidence for a structural salt bridge between Arg 655 and Asp 684 .
The Amino Acid Side Chain at Position 684 Influences Proton Affinity and the E 1 -E 1 P Conformational Transition-In catalytic models of P-type ATPases, the species to be transported must be bound before any ATP hydrolysis can take place and before the E 1 -E 1 P conformational change. All Asp 684 substitutions of the plant H ϩ -ATPase tested showed capabilities of both formation of the E 1 P phosphorylated intermediate and ATP hydrolysis. Thus, the biochemical characteristics of the amino acid side chain at position 684 appear not to be essential for the specific protonbinding event in E 1 that allows the formation of E 1 P. Alternatively, removal of the negative charge at position Asp 684 mimics the protonated state of this residue.
The Carboxylate of Asp 684 as a Molecular Determinant for Proton Transport-By insertion of a single CH 2 group into the aspartate at position 684 (D684E) in the plant H ϩ -ATPase, ATP hydrolytic activity was hampered (ϳ15% of wild-type). Reconstitution experiments demonstrated that ATP hydrolysis was coupled to proton transport. Therefore, a precise location of the carboxylate at position 684 must be important for efficient ATP hydrolysis and proton transport. However, although the precise location of this carboxylate is not essential for ATP dependent proton transport, it seems that the presence of a carboxylate at position 684 is absolutely necessary for any coupling between ATP hydrolysis and proton transport to take place. Thus, all Asp 684 substitutions in which the proton acceptor capabilities of Asp 684 have been removed (D684N, D684A, D684V, and D684R) showed a complete lack of coupling between ATPase activity and proton translocation. Still the mutant proteins show relatively high levels of ATP hydrolysis and steady-state phosphoenzyme.
Among the substitutions of Asp 684 tested, it appeared that D684N, D684A, D684V, and D684R exhibited a complete block in normal P-type enzymatic E 1 P-E 2 P transition. Thus, in addition to the fact that no proton pumping accompanied ATP hydrolysis, these mutant proteins were completely insensitive toward vanadate. In particular, following replacement of the Asp 684 carboxylate by carbamide (D684N), the mutated H ϩ -ATPase accumulated in the E 1 P conformational state, and the rate-limiting step regarding ATP hydrolysis in this mutant is most likely determined by the rate of spontaneous E 1 P dephosphorylation.
Taken together, the absence of a carboxylate at position 684 in the H ϩ -ATPase completely abolishes any enzymatic transition from E 1 P to E 2 P. Our results therefore demonstrates that the carboxylate group of Asp 684 is absolutely essential for the transfer of protons during the E 1 P-E 2 P conformational transition, which is hypothesized to accompany transport of the bound species across the membrane. This strongly suggests that the carboxylate of Asp 684 function as a molecular determinant for proton translocation across the membrane in plant P-type H ϩ -ATPases.
Mechanism of Proton Transport by P-type Proton Pumps-This study provides some clues into the biochemistry of the initial proton-binding event allowing E 1 P formation (the E 1 proton-binding site). The nature of the amino acid side chain at position 684 influences binding of the proton to this E 1 protonbinding site. Several groups including Asp 684 , none of which need to be essential, may contribute to initial coordination of the transported species but further conformational transitions involve Asp 684 as an essential residue. The E 1 proton-binding site could include a proton acceptor group or a group of atoms coordinating a hydrated proton (H 3 O ϩ ). A H 3 O ϩ coordination center could include the Asp 684 carbonyl side chain oxygen as one of the coordinating groups. In such a scenario, the proton could arrive to an already bound water molecule.
The function of Asp 684 AHA2 could resemble the functions of Asp 85 and Asp 61 in the non-related proton pumps bacteriorhodopsin (BR; 35, 36) and F 1 F 0 -ATPase (37,38), respectively. In an essential protonation event, the carboxylates of these residues receives the proton during catalysis. Among the amino acid residues involved in the proton transfer reaction of BR, Asp 85 is the only indispensable. A substitution to other than glutamate of Asp 85 results in mutant proteins completely devoid of proton transporting activity (39,40). As has been suggested for the Neurospora PMA1 H ϩ -ATPase (21), the presence of a conserved arginine (Arg 655 AHA2 /Arg 82 BR ) close to a conserved aspartate working as the essential proton acceptor group during catalysis (Asp 684 AHA2 /Asp 85 BR ) could resemble the proton-binding site in bacteriorhodopsin. The pK a of Asp 85 BR is influenced by the orientation of Arg 82 BR , and by modulating pK a values of other groups also, Arg 82 BR modulates the proton transfer trough the protein (40,41). Our finding that non-conservative substitutions of Arg 655 in the plant H ϩ -ATPase significantly slows down the E 1 P-E 2 P transition would be in accordance with an equivalent role of Arg 655 in the plant H ϩ -ATPase.
The major finding of this communication is that the carboxyl group of Asp 684 contributes to the proton-binding site of the AHA2 P-type H ϩ -ATPase. A negative charged oxygen of the Asp 684 carboxyl side chain group could function as an essential proton acceptor/donor during transport of protons across the membrane. This suggests that the mechanism of proton transport by P-type H ϩ -ATPases basically could be similar to the mechanism utilized by proton pumps such as BR and F 1 F 0 ATP synthase, in which proton transfer utilizes protonable side chain groups as proton donors and proton acceptors in a timely fashion.