Single point mutations distributed in 10 soluble and membrane regions of the Nicotiana plumbaginifolia plasma membrane PMA2 H+-ATPase activate the enzyme and modify the structure of the C-terminal region.

The Nicotiana plumbaginifolia pma2 (plasma membrane H+-ATPase) gene is capable of functionally replacing the H+-ATPase genes of the yeast Saccharomyces cerevisiae, provided that the external pH is kept above 5.0. Single point mutations within the pma2 gene were previously identified that improved H+-ATPase activity and allowed yeast growth at pH 4.0. The aim of the present study was to identify most of the PMA2 positions, the mutation of which would lead to improved growth and to determine whether all these mutations result in similar enzymatic and structural modifications. We selected additional mutants in total 42 distinct point mutations localized in 30 codons. They were distributed in 10 soluble and membrane regions of the enzyme. Most mutant PMA2 H+-ATPases were characterized by a higher specific activity, lower inhibition by ADP, and lower stimulation by lysophosphatidylcholine than wild-type PMA2. The mutants thus seem to be constitutively activated. Partial tryptic digestion and immunodetection showed that the PMA2 mutants had a conformational change making the C-terminal region more accessible. These data therefore support the hypothesis that point mutations in various H+-ATPase parts displace the inhibitory C-terminal region, resulting in enzyme activation. The high density of mutations within the first half of the C-terminal region suggests that this part is involved in the interaction between the inhibitory C-terminal region and the rest of the enzyme.

The Nicotiana plumbaginifolia pma2 (plasma membrane H ؉ -ATPase) gene is capable of functionally replacing the H ؉ -ATPase genes of the yeast Saccharomyces cerevisiae, provided that the external pH is kept above 5.0. Single point mutations within the pma2 gene were previously identified that improved H ؉ -ATPase activity and allowed yeast growth at pH 4.0. The aim of the present study was to identify most of the PMA2 positions, the mutation of which would lead to improved growth and to determine whether all these mutations result in similar enzymatic and structural modifications. We selected additional mutants in total 42 distinct point mutations localized in 30 codons. They were distributed in 10 soluble and membrane regions of the enzyme. Most mutant PMA2 H ؉ -ATPases were characterized by a higher specific activity, lower inhibition by ADP, and lower stimulation by lysophosphatidylcholine than wild-type PMA2. The mutants thus seem to be constitutively activated. Partial tryptic digestion and immunodetection showed that the PMA2 mutants had a conformational change making the C-terminal region more accessible. These data therefore support the hypothesis that point mutations in various H ؉ -ATPase parts displace the inhibitory C-terminal region, resulting in enzyme activation. The high density of mutations within the first half of the C-terminal region suggests that this part is involved in the interaction between the inhibitory C-terminal region and the rest of the enzyme.
The plasma membrane H ϩ -ATPase in plants and fungi is an electrogenic pump that couples ATP hydrolysis to proton transport out of the cell. This enzyme is composed of a single type of subunit of about 100 kDa and belongs to the P-type ATPase family characterized by a phosphorylated intermediate during catalysis (1). In the yeast Saccharomyces cerevisiae, H ϩ -AT-Pases are encoded by two genes, PMA1 and PMA2 (2,3), but only PMA1 is highly expressed and essential. The yeast H ϩ -ATPase has been well characterized at biochemical and genetic levels (for reviews, see . In plants, on the other hand, H ϩ -ATPases are encoded by a multigenic family, the members of which are differentially expressed. An important question to consider is whether all plant H ϩ -ATPase isoforms have identical kinetics and whether they are all subjected to similar metabolic regulation. Gene duplication that led to the two most expressed H ϩ -ATPase subfamilies predated the divergence of current plant families (8) and might have led to function specialization. Unfortunately, the expression of several isoforms in a same organ (9 -11) prevents their individual biochemical characterization and therefore invites the above question.
However, progress has been made through the heterologous expression of plant H ϩ -ATPase genes in the yeast S. cerevisiae. Three H ϩ -ATPase isoforms of Arabidopsis thaliana belonging to the same gene sub-family (aha1-3) were expressed in S. cerevisiae and found to have different kinetics (12). A gene (pma2) from Nicotiana plumbaginifolia belonging to another H ϩ -ATPase sub-family was also expressed in yeast (YAKpma2 strain). Unlike the A. thaliana genes, the N. plumbaginifolia pma2 was able to complement the removal of the two yeast H ϩ -ATPase genes, provided that the external pH was kept above 5.0 (13). More recently, we selected 21 single point mutants of YAKpma2 that were able to grow at external pH 4.0. The mutations conferred better ATPase and proton-pumping activities on the plant H ϩ -ATPase. Most of them were found within the C-terminal region, thus supporting its regulatory role. However, other mutations were located in other regions of the enzyme, thus indicating new residues that are probably involved in regulatory mechanisms (14). These observations invited several questions, such as how many residues are there and which mutation improves the enzyme? Do these mutations have identical effects at structural and functional levels of the H ϩ -ATPase?
In this study, we have identified the majority of point mutations able to activate the plant PMA2 H ϩ -ATPase expressed in yeast, identifying 10 regions implicated in the regulation of the enzyme. Sixty-three percent of the mutations were localized in two domains within the first part of the C-terminal region, but the others revealed new regions, such as the fourth transmembrane span. Progressive tryptic digestion revealed that the C-terminal region of mutants was more accessible than that of the wild-type, suggesting that the mutations result in a conformational change that displaces the inhibitory C-terminal region.

MATERIALS AND METHODS
Media-Yeast cells were grown either in rich medium, containing 2% (w/v) glucose, 2% (w/v) yeast extract (KAT), and 20 mM KH 2 PO 4 , or in minimal medium, containing 0.7% (w/v) yeast nitrogen base without amino acids (Difco), 0.115% (w/v) drop mix (15), supplemented by all * This work was supported by grants from the Interuniversity Poles of Attraction Program-Belgian State, Prime Minister's Office for Scientific, Technical and Cultural Affairs, the European Community's BIOTECH program, and the Belgian Fund for Scientific Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Holder of a Fonds National de la Recherche Scientifique fellowship. § Holders of a Fonds pour la Formation a la Recherche dans l'Industrie et dans l'Agriculture fellowship. amino acids and nucleotides required for growth. Solid media contained 2% agar. The 5-fluoroorotic acid medium was prepared as described in Ref. 15. The medium pH was adjusted with KOH to 6.5 or another pH when indicated.
Isolation and Sequencing of Mutants-Independent cultures, each inoculated with a single colony of YAKpma2, were grown at pH 6.5. At stationary phase, 2-3.10 8 cells from each culture were streaked on plates containing a rich medium at pH 4.0 (HCl). Spontaneous mutants growing under these nonpermissive conditions appeared after 3 days at 30°C. To ensure that each isolated mutant was independent, only one mutant per plate was selected. The 2p(PMA1)pma2 plasmid was retrieved from the YAKpma2 or mutant strains and transferred to Escherichia coli. The plant pma2 gene was sequenced using 11 synthetic primers dispersed throughout the gene.
Plasma Membrane Preparations-Plasma membranes were prepared according to Ref. 18, with the following modifications. After overnight growth in a 1.25-liter culture, the cells were harvested at a density of 80.10 6 cells/ml (rich medium, pH 6.5) and washed three times with ice-cold water. After centrifugation, the pellet was resuspended (15 ml per 10 g of fresh weight) in 250 mM sorbitol, 1 mM MgCl 2 , 50 mM imidazole-NaOH, pH 7.5, 1 mM phenylmethylsulfonyl fluoride, 5 mM dithiothreitol, and the protease inhibitors leupeptin, aprotinin, antipain, pepstatin, and chymostatin at 2 g/ml. Subsequent steps (cell disruption and subcellular fractionation) were unchanged. In the final step, the proteins were resuspended in 10 mM imidazole, pH 7.5 (NaOH) and 1 mM MgCl 2 , then aliquoted, frozen in liquid nitrogen, and stored at Ϫ80°C. The protein concentration was determined by the method of Lowry et al. (19) using bovine serum albumin as the standard.
ATPase Assays-ATPase assays were performed in a 96-hole microplate at 30°C in a medium containing 2 mM MgATP, 1 mM free Mg 2ϩ (MgCl 2 ), 50 mM Mes-Mops-Tris (pH 7.0), and 10 mM sodium azide (mitochondrial ATPase inhibitor), as described by Goffeau and Dufour (18), with the following modifications. The reaction was started by adding 40 l of enzyme (20 g) to 160 l of reaction medium. After 3, 6, and 9 min, 50 l aliquots were taken and mixed with 60 l 5% trichloroacetic acid to stop the reaction. The optimal pH was determined under the same conditions, with the buffer pH adjusted with HCl or NaOH (from pH 5.0 to 8.0). Inhibition curves were performed in the presence of increasing concentrations of vanadate (from 0 to 200 M) or ADP (from 0 to 10 mM). Stimulation by lysophosphatidylcholine (LPC) 1 was performed in the presence of increasing concentrations of LPC (from 0 to 600 g/ml). The specific activity was calculated by linear regression from the slope of the amount of P i released versus time. The total amounts of ATP and Mg 2ϩ to be added to the reaction mixture at the indicated pH were calculated in order to obtain the desired concentrations of MgATP and Mg 2ϩ (20).
Limited Tryptic Digestion-Limited tryptic digestion was performed at 37°C on 2 g of purified plasma membrane fractions with a trypsin/ protein ratio of 1/8. The reaction was performed in 50 mM Tris-HCl (pH 8.5) in a final volume of 500 l. After 0, 0.25, 0.5, 1, 2, 5 and 10 min, 50 l were taken and added to 10 l 10% SDS to stop the reaction. After addition of 20 l of 4-fold concentrated Laemmli's buffer, 10 l of each fraction was resolved on standard polyacrylamide gel and transferred to a nitrocellulose membrane. PMA2 fragments were detected by Western blot using polyclonal antibodies directed against either the 110-C-terminal residues, the E118-Q241 fragment (defining the small cytoplasmic loop) of PMA2, or the S305-S620 fragment (defining the large cytoplasmic loop) of PMA2 expressed in E. coli as a fusion product with the glutathione-S-transferase (pGEX, Amersham Pharmacia Biotech).

Identification of 23 New Mutants of the Plant PMA2-A
yeast strain, without its own two plasma membrane H ϩ -ATPase genes, PMA1 and PMA2, but expressing the N. plumbaginifolia plasma membrane ATPase pma2 gene, was shown to be able to grow at pH 6.0 but not below pH 5.0 (13). Spontaneous mutants able to grow at more acidic pH had been previously isolated and characterized (14). Twenty-one intragenic mutations mainly localized in the C-terminal region of the plant PMA2 H ϩ -ATPase were thus identified. This large number of distinct mutated positions led us to wonder whether new regions were still to be discovered. To address this question, we isolated and characterized 40 new intragenic mutants. Among these, we identified 23 new missense mutations ( Fig. 1), which revealed 14 new positions in the plant PMA2 not yet affected by previously described mutations. These mutants revealed new regions of PMA2 possibly implicated in the activation of the proton pump, such as the first part of the small cytoplasmic loop, the fourth transmembrane span, and the C-terminal end of the large cytoplasmic loop (Fig. 1). Additionally, the mutations in the C-terminal region presented two hot segments, the first one between Ala 852 and Lys 871 and the 1 The abbreviations used are: LPC, lysophosphatidylcholine; Tricine, second one between Arg 879 and Val 895 . Five other mutations were distributed as follows: one nonsense mutation, one deletion, and three duplications of a small sequence, all within the C-terminal region leading either to a frameshift and a shortening of the C-terminal domain or to a longer but modified C-terminal sequence. Twelve mutations corresponded to positions previously discovered. In total, 10 regions of PMA2 were thus affected by point mutations: the N-terminal region, the first transmembrane span, two regions in the small cytoplasmic loop, the fourth transmembrane span, two regions in the last part of the large cytoplasmic loop, and three regions in the C-terminal region.
The Mutations Conferred New Physiological and Enzymatic Properties-A common feature of mutants is their ability to grow at acidic pH. This physiological property was previously linked to a better H ϩ -pumping activity (14), and this was confirmed for all of the new mutants tested (data not shown). Nevertheless, some mutations could modify in particular some kinetic parameters. To gain an insight into the kinetic parameters, we further analyzed 12 mutants representing each region of the enzyme in which a mutation had been identified by our screening. The ATPase activity of the mutants was either unchanged or increased by up to 2-fold compared with the wild-type plant H ϩ -ATPase, except for the P294Q mutant, showing a reduced activity. However, these differences might reflect a higher ATPase content in the plasma membrane or a higher molecular activity. We therefore quantified the PMA2 H ϩ -ATPase content after SDS-PAGE. The major band around 100 kDa corresponds to PMA2 as confirmed by immunodetection of full-length (wild-type) and shortened (deletion mutant 882ocre) PMA2 analysis (data not shown). This band represents about 10% of the total plasma membrane proteins and was quantified by image analysis. We found that the amount of mutated PMA2 varied from 61 to 126% of the level observed for the wild-type PMA2, except again for the P294Q mutant, showing a 158% increase of H ϩ -ATPase in the plasma membrane compared with the wild-type PMA2 (Table I and Fig. 2). The molecular ATPase activity obtained from the specific ATPase activity corrected with the relative amount of ATPase in the plasma membrane was increased for 11 of the 12 mutants analyzed here and for another 3 mutants analyzed previously (14). We therefore conclude that the larger specific ATPase activity of the mutants is due to a better intrinsic ATPase activity. The only exception is mutant P294Q, for which the ATPase activity did not increase despite a larger ATPase amount in the plasma membrane.
One mutant displayed an important alkaline shift of the optimal pH (7.6) compared with the wild-type (6.6). The other mutants had a more discrete alkaline shift (optimal pH between 6.8 and 7.2) and differences among these mutants were too small to be considered as significant at this stage (Table I).
Other kinetic characteristics revealed particular features when compared with the wild-type PMA2. Two mutants (P72A and P294Q) were 4-or 7-fold, respectively, less sensitive to vanadate, a transition state analog of inorganic phosphate that binds to the E2 form of the enzyme (21), suggesting that these two mutations result in a modification of the E2/E1 ratio in the E1-vanadate insensitive form. Another parameter analyzed was the competitive inhibition of ATP hydrolysis by ADP, which was released at the E1.H-ATP 3 E1.HϳP transition step. All of the mutants tested were 2-4-fold less sensitive to ADP as compared with the wild-type PMA2 (Table I).
Plant H ϩ -ATPases are stimulated by LPC, a detergent-like phospholipid, both in native plant membranes and when expressed in yeast (12,14). Activation of wild-type PMA2 was 5-6-fold, whereas the PMA2 mutants were all less activated, although responses varied (Table I). Some mutants were acti-FIG. 2. PMA2 content in plasma membranes from different mutants. Plasma membrane fractions (10 g of proteins) were resolved by electrophoresis on a 10% Tris-Tricine polyacrylamide gel and stained with Coomassie Blue. The band corresponding to the plant H ϩ -ATPase is indicated by an arrow. MW corresponds to size markers. The protein bands were quantified using ImageMaster software (Amersham Pharmacia Biotech). The intensity of the band corresponding to the H ϩ -ATPase was reported to the total intensity of protein per lane to give the relative H ϩ -ATPase amount in the plasma membrane. This value was given as an arbitrary value of 100% for YAKpma2, and the value for the mutants was expressed as the ratio compared with the wild-type PMA2. The data are given in Table I. vated very slightly or not at all, such as E14D, W75C, W858C, and W883L, whereas others were stimulated 2-3-fold, such as H221N, A917V, N510K, P294Q, E626G, and S298L. Finally, two mutants (P72A and P154R) displayed a particular response because they were inhibited by 300 g/ml LPC (Table I), although they were weakly stimulated by a lower amount of LPC (data not shown).

The Mutations Induced a Different Conformation in the Cterminal
Region-LPC activation has been suggested as occurring through the displacement of the inhibitory C-terminal region (22). The reduced LPC stimulation of mutants might indicate that they have already undergone a conformational change that displaced the C-terminal region. To test this hypothesis, we performed limited tryptic digestion on plasma membranes purified from wild-type and mutant strains and we analyzed the H ϩ -ATPase integrity by means of antibodies directed against either the C-terminal region, the small cytoplasmic loop or the large cytoplasmic loop. The 12 selected mutants representing each affected region of PMA2 and the wild-type PMA2 expressing strain were analyzed. The detection with antibodies directed against the C-terminal region of the enzyme (Fig. 3, left column) revealed that in all cases, the 100-kDa band corresponding to the mutated PMA2 was more rapidly degraded by trypsin compared with the wild-type enzyme. The 100-kDa fragment, corresponding to the full-length en-zyme, disappeared after 5 min for the wild-type PMA2, whereas the degradation was completed between 15 s and 1 min for the mutated PMA2, suggesting that the C-terminal region of the PMA2 mutants contained a more accessible cleavage site for trypsin. In addition to the 100-kDa signal, antibodies against the small cytoplasmic loop detected a ϳ90-kDa fragment resulting from the tryptic digestion of PMA2, confirming that the first region to be cleaved in the wild-type as well as in the mutants was the C terminus (Fig. 3, right  column) and that the C terminus was cleaved more rapidly in the mutated PMA2 as compared with the wild-type PMA2. The degradation of the 100-kDa band associated with the apparition of the 90-kDa degradation product was also observed with antibodies against the large cytoplasmic loop (not shown).
Protection by vanadate or ADP did not modify sensitivity to trypsin proteolysis (data not shown), indicating that the differences in conformation do not affect the E2 to E1 transition step of the catalytic cycle. However, LPC modified the pattern of degradation for the mutated and the wild-type PMA2 as observed with the anti-C terminus antibodies (Fig. 4). Shorter products were observed in the presence of LPC, suggesting either that LPC protects tryptic sites within the C-terminal region (direct effect) or that LPC changes the H ϩ -ATPase structure so that tryptic sites within the C-terminal are less accessible (indirect effect). In addition, the profiles of tryptic digests of the mutant and wild-type forms of the plant PMA2 H ϩ -ATPase in the presence of LPC were very similar even if the degradation was slightly more rapid for the mutant (Fig. 4). Another detergent, Triton X-100, had no effect on wild-type and mutated PMA2 tryptic digestion (data not shown). These results suggest that the differences in conformation between the native and mutated PMA2, observed in the absence of LPC, were almost abolished by this phospholipid. These observations made for E14D (Fig. 4) were reproduced for six other mutants analyzed.

DISCUSSION
Distribution of the PMA2 Mutations-We have previously identified 16 PMA2 residues, the mutation of which led to a more effective enzyme (14). In this study, we have identified 14 new positions and 8 new mutations of previously identified positions. Most amino acids involved in the activation of PMA2 and detectable by our screening should now be identified because several mutations were independently discovered on several occasions and, chronologically, the last 10 mutants confirmed previously identified positions. Completing the cartography of mutations would require a tremendous amount of work, and the results would not be much more informative than the analysis of the currently available 42 distinct mutations. Altogether, we have detected 10 regions of the plant H ϩ -ATPase PMA2 that are possibly implicated in enzyme regulation. They are distributed throughout the protein, but some regions, such as most of the large cytoplasmic loop and the second, third, and last six transmembrane spans, are not affected by any mutation. At least two hypotheses could explain this observation. The first is that these regions are not implicated in the regulation of the protein and that a structural change within them would not lead to a conformational modification capable of activating the H ϩ -ATPase. The second and not exclusive hypothesis is that some of these regions are involved in the catalytic cycle of the H ϩ -ATPase. The large loop, for example, is required for the binding of MgATP and the formation of the aspartyl-phosphate intermediate. Mutations in most of this sequence would probably affect the turnover and reduce activity. These kinds of mutants were therefore not selected by our screening.

Mutants Reveal New Regions Possibly Involved in PMA2
FIG. 3. Limited tryptic digestion of wild-type and mutated PMA2. Purified plasma membranes of the strains expressing the 12 mutated PMA2s, representing each affected region of the H ϩ -ATPase, and the wild-type PMA2 were digested by trypsin as described under "Materials and Methods" for the indicated times. PMA2 was immunodetected by antibodies directed against the C-terminal domain (left column) or against the small loop (right column). With the anti-Cterminal antibodies, we can follow the degradation from the C-terminal region of the full-length enzyme (100 kDa) and with the anti-small loop antibodies, the degradation of the full-length enzyme (100 kDa) associated with the apparition of a 90-kDa degradation product.
Regulation-Most residues affected by mutations are well conserved among plant H ϩ -ATPases but not in fungal H ϩ -ATPases (42), suggesting that the mutations involve a regulation mechanism that is specific to plants. As an exception, three mutated residues are conserved in plant and yeast H ϩ -ATPases: Pro 72 , Pro 154 , and Pro 294 . The last two are also conserved in the rabbit sarcoplasmic Ca 2ϩ -ATPase SERCA1. As a matter of fact, both P72A and P294Q clearly displayed a lower sensitivity to vanadate, whereas P72A and P154R displayed a particular response to LPC (inhibition at high concentration). This suggests that these positions, although involved in regulatory aspects of the plant H ϩ -ATPase, might also play a role in the basic catalytic activity. In this respect, the structural and functional properties of these mutants might turn out to be different from those of the other mutants, even though they were all selected for their better growth at low pH.
Some PMA2 mutants identified in this study could be compared with mutants of the S. cerevisiae plasma membrane H ϩ -ATPase (PMA1) and the rabbit sarcoplasmic Ca 2ϩ -ATPase (SERCA1), obtained by directed mutagenesis. Up to now, the N-terminal domain had not been indicated as playing an important role in the yeast PMA1 or mammalian SERCA1 activity, suggesting that the N-terminal region of PMA2, which includes E14D, might be involved in a regulatory mechanism specific to the plant enzyme. Glu 14 is also the first residue conserved in all plant H ϩ -ATPases known so far. In contrast, the first transmembrane span of the yeast PMA1 seems to be a very sensitive domain at the structural level. Proline 123 of the yeast PMA1 (scP123) corresponding to the proline 72 of the plant PMA2 (npP72) even though not essential, was suggested as playing a structural role in acting on conformational changes or optimizing the interaction between helices of the yeast PMA1 (24,25). In the small cytoplasmic loop, the Pro 154 residue of the plant PMA2 is conserved in the yeast PMA1 (scP198) and the mammalian SERCA1 (sercaP147). The scP198A mutation of the yeast PMA1 showed a dominant lethal phenotype (26). The region encompassing the three other mutations within the PMA2 small loop (between Val 220 and His 229 ) is well conserved among plant and yeast H ϩ -ATPases. The yeast R271T mutant is dominant lethal (23), and the L275S mutant of PMA1 was found as a second-site revertant of the G158D mutation located in the second transmembrane span (27). The fourth span of both PMA1 and SERCA1 was systematically analyzed by site-directed mutagenesis (28,29) and contained residues essential for ion transport. The proline conserved in SERCA1 (sercaP236) or PMA1 (scP339) corre-sponding to Pro 294 in the plant PMA2 largely affected Ca 2ϩ (30) and H ϩ (28) transport. From this survey, we can conclude that many PMA2 mutations from the first to the fourth transmembrane spans affect residues that are within or close to regions involved in catalysis and where mutations in the yeast PMA1 and/or the mammalian SERCA1 were deleterious or even lethal. In two cases, the mutations touched a PMA2 residue conserved in the yeast H ϩ -and mammalian Ca 2ϩ -ATPase. However, the mutation (P154R or P294Q) improved the plant PMA2 but was deleterious in the other two cases. As the mutations were not the same, this discrepancy could be further investigated by a site-directed mutagenesis interchanging the mutation in each enzyme. The observation that the majority of the PMA1 and SERCA1 mutations had a negative or no significant effect on enzyme activity, whereas the mutations found in the plant PMA2 had a positive effect, emphasizes the significance of the positive selection (yeast growth at low pH) used in this study.
Most Mutations Affect the First Half of the C-terminal Region-The concentration of 63% of the mutations into two segments in the first half of the C-terminal region indicates the critical role of this region in the regulation of plant H ϩ -AT-Pases. The C-terminal region is involved in glucose-induced regulation of yeast H ϩ -ATPase (31,32), in calmodulin binding in mammalian Ca 2ϩ -ATPases (33), and in regulation of plant H ϩ -ATPases (22). However, the C-terminal sequence is very different in each of these three ATPases, reflecting distinct mechanisms of regulation. Activation of plant H ϩ -ATPases had previously been induced by in vitro LPC or in vivo fusicoccin treatment (34), or by tryptic digestion of the C terminus (22). Deletions by site-directed mutagenesis within the C-terminal region of A. thaliana H ϩ -ATPases had the same effect (35). All of these modes of activation led to a slight alkaline shift in the optimum pH and an increase in ATP hydrolysis and protonpumping activity. These results were generally interpreted as the removal of the inhibitory C-terminal region aside the rest of the enzyme. The mutations identified in this study affecting the C terminus were generally point mutations but also deletions or insertions in the C-terminal region modifying the sequence and possibly the structure of the end of the protein. The great majority of the mutations affected the first half of the C-terminal region, possibly resulting in a reduction of or a disruption to the interaction between this region and the rest of the enzyme. This hypothesis is supported by the observation that trypsin had better access to this region in the mutants. Except for A917V, none of the mutations were found in the FIG. 4. Effect of LPC on limited tryptic digestion of wild-type and mutated PMA2. Purified plasma membranes from the wild-type PMA2 and E14D mutant expressing strains were digested by trypsin in the presence or absence of LPC as described under "Materials and Methods." PMA2 fragments were immunodetected by antibodies directed against the C-terminal domain of the plant PMA2. second half. This suggests that the first half of the C-terminal region interacts with the rest of the H ϩ -ATPase and possibly reduces its activity and that the second half does not participate directly in enzyme autoinhibition. The mutations localized in the other regions of PMA2 might represent areas possibly interacting with the first half of the C-terminal region. The fact that some of these mutations concern residues located within or near predicted transmembrane spans is not surprising if we consider that the mutations also modify H ϩ -pumping capacity of the enzyme (14).
Significance of the Mutations-All mutations seem to result in a similar structural modification of the PMA2 as shown by the tryptic treatment, thus reflecting a shift between a latent or activated state of PMA2. We therefore suggest that the mutations mimic structural modifications of PMA2 that occur in the plant following a regulatory signal. This signal might well be LPC, a phospholipid known to activate plant H ϩ -ATPases (22,36). In the presence of LPC, trypsin accessibility to PMA2 was modified, and there were no longer any differences between wild-type and mutated PMA2. We therefore suggest that mutations transform PMA2 into its activated form, as does LPC. Moreover, LPC also restricts access to the tryptic sites located at the beginning of the C-terminal region, suggesting either direct interaction of this part with LPC or that LPC binding occurs somewhere else but results in the structural modification of the first residues of the C-terminal region. Gomes et al. (37) suggested that LPC binds to the H ϩ -ATPase in sites that are not located on the C-terminal region. However, the deleted mutant used during their study (aha2⌬92) still contained a small part of the C-terminal region that encompasses, for instance, the Ala 852 , Gly 855 , Ala 857 , Trp 858 , and Leu 862 residues, which are mutated in PMA2. Therefore, we cannot exclude the possibility that LPC effectively binds to the first part of the C-terminal region.
Regulatory mechanisms other than LPC might also be involved. The most obvious is (de)phosphorylation by a kinase/ phosphatase system. Although phosphorylation of plant H ϩ -ATPases occurs (38,39,43), little is known concerning the relationship between this type of modification and enzyme structure and activity.
Finally, the recent identification of a complex formed by the plant H ϩ -ATPase and the regulatory 14 -3-3 proteins (40,41,44) introduces a third regulatory mechanism that might be mimicked by the PMA2 mutations. It has been shown that 14-3-3 proteins bind to the C-terminal region (40,41,44). This binding can be stabilized by the phytotoxin fusicoccin and this results in an increased H ϩ -ATPase activity, a property shared by the PMA2 mutants. We suggested above that the first half of the C-terminal region contributes to the binding to and partial inhibition of the rest of the H ϩ -ATPase. This would leave the second half of the C-terminal region for binding to the regulatory 14-3-3 proteins. We might therefore suggest that 14-3-3 binding modifies the interaction of the first half of the Cterminal region with the rest of the enzyme. According to this model, mutations obtained in the first half would bypass the positive regulation by a 14-3-3 protein.
In conclusion, this study has shown that activation of the plant H ϩ -ATPase PMA2 expressed in yeast could be reached by single point mutations localized in 10 regions of the enzyme, with a large concentration in two small areas within the first part of the C-terminal region. Activation of PMA2 was correlated to a change in conformation characterized by the greater accessibility of the tryptic site localized in the first part of the C-terminal region, thus mimicking regulatory modifications of the plant H ϩ -ATPase. The availability of a large amount of homogenous latent (wild-type) and activated (mutant) forms of PMA2 should lead to a detailed comparison of their structural and functional properties, thus sidestepping the problem that in plant material, several isoforms coexist, each of which is possibly represented by a mixture of latent and activated forms.