Complementation of the Saccharomyces cerevisiaePlasma Membrane H+-ATPase by a Plant H+-ATPase Generates a Highly Abundant Fusicoccin Binding Site*

Accumulating evidence suggests that the H+-ATPase of the plant plasma membrane is activated by a direct, reversible interaction with 14-3-3 proteins involving the displacement of the C-terminal autoinhibitory domain of the enzyme. The fungal phytotoxin fusicoccin (FC) appears to stabilize this H+-ATPase·14-3-3 complex, thus leading to a persistent activation of the H+-ATPase in vivo. In this study we show that functional replacement of the Saccharomyces cerevisiae H+-ATPase genes by a Nicotiana plumbaginifolia H+-ATPase (pma2) results in the generation of a high affinity fusicoccin binding site that is exceptionally abundant. Acquisition of FC binding capacity is accompanied by a significant increase in the amount of plasma membrane-associated yeast 14-3-3 homologs. The existence of a (plant) PMA2·(yeast)14-3-3 complex was demonstrated using two-dimensional gel systems (native/denaturing). After expression of PMA2 lacking most of its C-terminal region, neither H+-ATPase·14-3-3 complex formation nor FC binding activity could be observed. Furthermore, we obtained direct biochemical evidence for a minimal FC binding complex consisting of the C-terminal PMA2 domain and yeast 14-3-3 homologs. Thus we demonstrated unambiguously the relevance of this regulatory ATPase domain for 14-3-3 interaction as well as its requirement for FC binding.

Solute transport across the plasma membrane of plants and fungi is driven by an electrochemical proton gradient, which is generated by the activity of a H ϩ -ATPase. The pivotal role of the plasma membrane H ϩ -ATPase in nutrient uptake and growth makes it a primary target for regulatory mechanisms (1,2).
At the C terminus of plant and yeast H ϩ -ATPases, an autoinhibitory domain is present that is thought to play a role in regulation of enzyme activity. In plants, the fungal phytotoxin fusicoccin (FC) 1 seems to cause the displacement of this inhibitory domain from the catalytic site, and as a consequence, it induces an activation of the H ϩ -ATPase in vivo (3,4).
Accumulating evidence indicates that 14-3-3 proteins are mediators of FC action as well as regulators of plant H ϩ -ATPase activity. 1) One component of the high affinity FCbinding protein of the plant plasma membrane was shown to belong to the family of eukaryotic 14-3-3 proteins (5-7). Members of this family are highly conserved hydrophilic proteins that serve diverse regulatory functions by mediating proteinprotein interactions as well as regulating protein kinase activities (reviewed in Ref. 8). 2) Recent observations suggest that 14-3-3 dimers interact directly with the H ϩ -ATPase, involving the displacement of the C-terminal autoinhibitory domain of the enzyme. The H ϩ -ATPase⅐14-3-3 complex represents an activated state of the enzyme that can be stabilized by FC binding in vivo (9 -11) or in vitro (9). 3) Circumstantial evidence has been presented that indicates a role of the C-terminal ATPase regulator in complex formation with 14-3-3 homologs (9, 10) as well as FC binding capability of the corresponding 14-3-3s (9). However, direct biochemical evidence for the relevance of the C terminus of the H ϩ -ATPase in these processes has been elusive.
We analyzed the regulation of the plant H ϩ -ATPase by 14-3-3 homologs after heterologous expression in the yeast Saccharomyces cerevisiae. In this species, the H ϩ -ATPase is encoded by two genes (PMA1 and PMA2), only one of which (PMA1) is constitutively expressed and essential for growth (12). Yeast transformed with pma2, a H ϩ -ATPase isoform from Nicotiana plumbaginifolia, was still able to grow when the endogenous H ϩ -ATPase genes were deleted (13). On the contrary, the three Arabidopsis thaliana H ϩ -ATPase genes AHA1, AHA2, and AHA3 were not able to functionally replace the yeast PMA1 unless, as shown for AHA2, the enzyme was activated by deletion of its C terminus (14 -16). Apparently, the N. plumbaginifolia PMA2 was active in yeast despite of its Cterminal autoinhibitory domain. The yeast S. cerevisiae possesses two 14-3-3 isoforms, BMH1 and BMH2 (17,18), which are candidate activators of the heterologously expressed plant H ϩ -ATPase.
To analyze the possible involvement of endogenous 14-3-3 proteins in the regulation of the plant PMA2 expressed in the yeast strain deleted of its own H ϩ -ATPase genes, the FC binding activity of this strain was investigated. Analyzing plasma membranes, we observed the presence of a high affinity FC binding site that is exceptionally abundant and the presence of a (plant)PMA2⅐(yeast)14-3-3 complex. Moreover, we obtained direct biochemical evidence of a PMA2 C terminus⅐14-3-3 complex representing the minimal complex with FC binding capacity.

EXPERIMENTAL PROCEDURES
Nomenclature-pma2 and PMA2 designate the N. plumbaginifolia H ϩ -ATPase gene and gene product, respectively. PMA1 and PMA2 and PMA1 and PMA2 designate the S. cerevisiae H ϩ -ATPase genes and gene products, respectively.
Media and Strains-Yeast cells were grown in rich medium containing 2% (w/v) glucose, 2% (w/v) yeast extract, and 20 mM KH 2 PO 4 adjusted with KOH to pH 6.5.
* 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.
Plasma Membrane Preparation and Solubilization of Protein-Plasma membranes were prepared as described (20) with the modifications mentioned in Morsomme et al. (21).
Mild Tryptic Treatment of Plasma Membranes-Mild proteolysis of plasma membranes to remove the C-terminal domain of the H ϩ -ATPase followed the procedure of Palmgren et al. (22). Briefly, plasma membranes (2.5 mg ml Ϫ1 in 25 mM MOPS-KOH, 250 mM sucrose, 4 mM ATP, 5 mM EDTA, 2 mM dithiothreitol, pH 7.5) were mixed with an equal volume of trypsin (100 g ml Ϫ1 ) in 25 mM MOPS-KOH, 250 mM sucrose, 5 mM EDTA, pH 7.5 (protein to trypsin ratio 25:1, w/w). Incubation was at 20°C, and proteolysis was stopped after 5 min by the addition of trypsin-inhibitor (trypsin to inhibitor ratio 1:20, w/w) in the buffer used for trypsin above. Separation of membrane-bound and proteolytically released proteins was achieved by centrifugation at 200,000 ϫ g for 25 min (Beckman TL 100).
Determination of the Activity of the H ϩ -ATPase-The activity of the plasma membrane H ϩ -ATPase was measured as the release of inorganic phosphate from ATP according to the method of Hodges and Leonard (25) as modified by Grä f and Weiler (26).
Gel Electrophoresis and Protein Immunoblotting-Denaturing gel electrophoresis was performed according to Laemmli (27). The discontinuous systems consisted of 4% stacking gels and 9% resolving gels.
Blue native (BN) electrophoresis followed the procedure of Schä gger et al. (28) using gradient gels of 4 to 15% acrylamide. Coomassie Blue (SERVA Blau G) was added from a 5% stock solution in 500 mM aminocaproic acid to solubilized plasma membrane proteins to obtain a detergent to Coomassie Blue ratio of 1:1 (w/w). Electrophoresis was performed at 4 -7°C and started at 100 V until the sample had entered the stacking gel. It was continued with constant voltage of 400 V.
The antiserum against the N terminus of the plant H ϩ -ATPase was kindly provided by Dr. R. Serrano, Valencia, Spain. The antiserum against the 14-3-3 homologs of S. cerevisiae, BMH1 and BMH2, was kindly provided by Dr. G. P. H. van Heusden, Leiden, Netherlands.

Functional Replacement of the Yeast H ϩ -ATPase by the Plant PMA2 Generates a Highly Abundant FC Binding
Site-Based on sequence information obtained from peptides, one component of the FC-binding protein (FCBP) purified from plant plasma membranes has been identified as a 14-3-3 protein (5-7). It has been proposed that the functional FCBP is represented by a reversible complex between a 14-3-3 dimer and the C-terminal domain of the H ϩ -ATPase (9). Formation of the H ϩ -ATPase⅐14-3-3 complex results in activation of the H ϩ -ATPase (9 -11). The reason for this activation seems to be a suppression of the autoinhibitory action of the C-terminal domain of the enzyme (3,4).
Several plant H ϩ -ATPases have been expressed individually in S. cerevisiae and exhibited different capabilities to sustain yeast growth in the absence of the endogenous H ϩ -ATPase PMA1 (13)(14)(15)(16). The three A. thaliana genes AHA1, AHA2, and AHA3 were not able to complement the yeast PMA1 (14,15). Interestingly, increasing the activity of AHA2 by deletion of its C-terminal autoinhibitor allowed the truncated enzyme to functionally replace the yeast H ϩ -ATPase (16). On the other hand, the N. plumbaginifolia H ϩ -ATPase PMA2 sustained yeast growth in the absence of endogenous H ϩ -ATPase genes without deletion of its C terminus (Ref. 13, strain YAKpma2), indicating that PMA2 exists in an apparently active state in transformed yeast. This raised the question as to whether yeast 14-3-3 proteins are involved in regulation of the heterologously expressed plant PMA2. Because participation of 14-3-3 proteins may result in FC binding capacity, we analyzed different yeast strains with respect to FC binding sites.
In agreement with earlier studies showing that wild-type yeast is devoid of any FC binding sites (32), no FC binding activity was found in the yeast strain YPS14-4 expressing the yeast PMA1 (Fig. 1, lower panel). In contrast, FC binding activity could clearly be detected in enriched plasma membranes obtained from YAKpma2 (expressing only the N. plumbaginifolia PMA2; Fig. 1, lower panel; Fig. 2).
Because the FCBP was postulated to be a complex between 14-3-3 homologs and the H ϩ -ATPase (9), we expected the plant PMA2 to recruit yeast 14-3-3 isoforms to the plasma membrane. As shown by immunodetection (Fig. 1), the amount of plasma membrane-associated 14-3-3 polypeptides is significantly increased in the yeast strain YAKpma2 compared with YPS14-4.
Interestingly, expression of PMA2 lacking most of its Cterminal autoinhibitory region (strain YAKpma2-882stop), which represents a per se activated state of the enzyme failed to create FC binding sites. In this case accumulation of 14-3-3 proteins at the plasma membrane did not occur (Fig. 1).
Saturation analysis of [ 3 H]FCol binding to plasma membranes derived from YAKpma2 revealed a high affinity site for FC (k D : 2.42 nM; Fig. 2), resembling plant FCBPs (reviewed in Ref. 33). No low affinity second site could be detected as de-  (34) and Commelina communis (35, Fig. 2) by using the same radioligand.
PMA2 Forms a Complex with Yeast 14-3-3 Homologs-To analyze a possible direct interaction of plant PMA2 and yeast 14-3-3 homologs, we subjected solubilized plasma membrane proteins to a two-dimensional gel system, which combines a native separation of protein complexes according to size with a second dimension under denaturing conditions. As shown in Fig. 3, a substantial part of the yeast 14-3-3 proteins comigrated with the plant PMA2 during native PAGE, establishing a physical association of these polypeptides in YAKpma2. Comigration occurred over a wide range of native molecular mass (300 -900 kDa). FC treatment in vivo (i.e. incubation of yeast before homogenization, Fig. 3, middle panel) or in vitro (i.e. incubation of isolated plasma membranes, Fig. 3, right panel) was necessary to demonstrate the 14-3-3⅐(plant) H ϩ -ATPase complex in the gel system. In the absence of FC, this complex could not be detected (Fig. 3, left panel), probably because of its instability during electrophoresis.
In the yeast strain YPS14-4 (expressing only the yeast PMA1), a complex consisting of the endogenous H ϩ -ATPase PMA1 and 14-3-3 homologs was not detectable, neither after in vivo (Fig. 4) nor after in vitro FC treatment (data not shown). In addition, complex formation between 14-3-3 isoforms and PMA2 lacking part of its C-terminal region (strain YAKpma2-882stop) could not be observed (Fig. 4). These observations are in agreement with the lack of FC binding activity and the low amount of plasma membrane-associated 14-3-3 homologs in these strains (Fig. 1).
The Minimal Complex with FC Binding Activity Consists of the C-terminal PMA2 Domain and a 14-3-3 Dimer-We have established a direct interaction of plasma membrane-associated yeast 14-3-3 isoforms with the plant PMA2, which enables the YAKpma2 strain to bind FC (Figs. 1-3). This complex most likely is characterized by a direct association of the C-terminal domain of PMA2 with 14-3-3 proteins. By analysis of yeast cells expressing PMA2 lacking its C-terminal autoinhibitory region (strain YAKpma2-882stop), neither FC binding activity (Fig. 1) nor formation of a complex between 14-3-3 homologs and the C-terminally truncated PMA2 could be observed (Fig. 4).
Mild tryptic treatment of plasma membranes is known to result in removal of the C terminus of the H ϩ -ATPase (22). To investigate the effect of limited proteolysis on YAKpma2 plasma membranes, we analyzed proteolytically released proteins by native PAGE according to size. Mild tryptic treatment of FC-stabilized yeast plasma membranes liberated a complex consisting of the C-terminal fragment of PMA2 and a 14-3-3 dimer (Fig. 5, lane ϩ). Resolution of this native complex under denaturing conditions showed its individual components, namely 14-3-3 proteins and two peptides of approximately 5-8 kDa (Fig. 6, lane ϩ, A, triangles). These peptides were identified as C-terminal H ϩ -ATPase fragments by use of antibodies raised against the C-terminal region of PMA2 (Fig. 6, lane ϩ,  B). FC treatment was necessary to demonstrate the complex (Fig. 5, lane Ϫ, Fig. 6, lane Ϫ). In the absence of FC, the individual components of the complex were released separately from the plasma membrane. As a consequence, solely 14-3-3 dimers could be detected in the native gel system, whereas the C-terminal PMA2 peptide passes the front of the gel because of its small size.
14-3-3 dimers themselves that have been released by mild tryptic treatment of YAKpma2 plasma membranes in the absence of FC were unable to bind FC (Fig. 5, lower panel). On the other hand, FC binding activity could clearly be demonstrated in the fraction containing the complex between a 14-3-3 dimer and the C terminus of PMA2 (Fig. 5, lower panel). Thus, the C terminus⅐14-3-3 complex represents the minimal unit exhibiting FC binding capability.
Effect of FC treatment-It has been established that an in vivo treatment of intact plant tissues with FC leads to an increase in the amount of plasma membrane-associated 14-3-3 homologs (5, 7). Under these conditions, a direct interaction of 14-3-3 homologs with the H ϩ -ATPase was demonstrated, indicating that FC stabilizes this complex during plant plasma membrane preparation, resulting in its increased recovery (9 -11). In contrast, treatment of YAKpma2 with FC in vivo did not increase the amount of yeast 14-3-3 proteins associated with the plasma membrane ( Fig. 1), indicating that the PMA2⅐14-3-3 complex is unusually stable and persists during yeast plasma membrane preparation, even in the absence of FC. Nevertheless, FC treatment is necessary to demonstrate this complex after native electrophoresis (Fig. 3). Taking into account that FC binding depends on the presence of the H ϩ -ATPase⅐14-3-3 complex (Fig. 5), the data strongly suggest that FC does not promote the interaction between 14-3-3 homologs and the plant PMA2; rather, it stabilizes a preformed complex.
The increase in the amount of the H ϩ -ATPase⅐plant 14-3-3 complex obtained after an in vivo FC treatment of plant tissue is accompanied by activation of the H ϩ -ATPase (5, 7). In contrast, FC treatment of YAKpma2 in vivo did not activate PMA2 (Table I), which is in agreement with the unaltered amount of the complex (Fig. 1). Additionally, an in vitro FC treatment had no effect on PMA2 activity (Table I). Taken together, the data suggest that binding of FC to the preformed complex does not activate PMA2; rather, the complex represents an already activated state of the H ϩ -ATPase. This idea is supported by the phosphohydrolytic activity of PMA2 in YAKpma2 (Table I); at pH 7.3, its activity is only slightly reduced (20%) compared with the C-terminal-truncated PMA2 (strain YAKpma2-882stop), representing a hyperactive state of PMA2. Hence, the interaction of the C terminus of PMA2 with yeast 14-3-3 homologs seems to resemble, at least partly, the effect of a deletion of the C-terminal autoinhibitor.

FIG. 2. Saturation analysis for [ 3 H]FCol binding to plasma membranes of the yeast strain YAKpma2 (expressing only the plant PMA2
). Plasma membrane proteins (5 g) were incubated for 90 min at 25°C with increasing concentrations of the radioligand. Nonspecific binding was determined by including a 500-fold excess of unlabeled FC, respectively. The inset shows the analysis of the data according to Scatchard (37). The right panel shows maximal [ 3 H]FCol binding activity of plasma membranes derived from YAKpma2 as well as leaves of C. communis. B, protein-bound radio ligand; B/F, ratio of proteinbound to free radioligand.

Interaction of the C-terminal Domain of the Plant H ϩ -ATPase with a 14-3-3 Dimer Creates a FC-binding Protein-Although
the fusicoccin-binding protein initially has been identified as a 14-3-3 protein (5-7), several observations indicate that plant 14-3-3 homologs themselves are not able to bind fusicoccin. FC binding activity of partially purified plasma membrane-associated 14-3-3 proteins could only be detected when coelution with the H ϩ -ATPase was observed (9). In addition, the expression of plant 14-3-3 isoforms in S. cerevisiae did not result in the generation of a FC binding site. 2 In this study we showed that functional expression of the plant H ϩ -ATPase PMA2 in yeast deleted of its own H ϩ -ATPases (strain YAKpma2) generated a highly abundant FC binding site in the yeast plasma membrane (Figs. 1 and 2). Acquisition of FC binding activity is due to direct interaction of PMA2 with endogenous 14-3-3 homologs (Fig. 3). The C-terminal-truncated PMA2 failed to interact with 14-3-3 proteins (Fig. 4) as well as to create FC binding sites (Fig. 1). Thus, this H ϩ -ATPase domain appears to be essential for both complex formation with 14-3-3 proteins and generation of a functional FCBP. Additionally, in plants, several hints have been obtained suggesting a participation of the H ϩ -ATPase C terminus in complex formation with 14-3-3 protein (9, 10). Nevertheless, the suggestions altogether were based on indirect evidence. With respect to YAKpma2, we obtained direct biochemical evidence for a complex consisting of the C terminus of PMA2 and a yeast 14-3-3 dimer (Fig. 5, 6). Notably, this complex represents the minimal complex with FC binding capacity. In contrast, 14-3-3 dimers themselves were unable to bind FC (Fig.  5).
Interestingly, wild-type yeast expresses 14-3-3 isoforms (17, 18) as well as a plasma membrane H ϩ -ATPase but is devoid of any FC binding activity (32). The C-terminal regulatory domain of the yeast H ϩ -ATPase (PMA1) is not homologous to the corresponding region of the plant enzyme. This indicates that the generation of a FC binding site mainly depends on the ability of the plant H ϩ -ATPase C-terminal domain to interact with a 14-3-3 dimer.
Mechanism of FC Action-FC binding is a feature of the plant H ϩ ATPase⅐14-3-3 complex, especially of a complex consisting of the C-terminal enzyme domain and a 14-3-3 dimer, and does not occur on the separated components (Fig. 5). Thus, we can rule out that FC promotes formation of the complex. In YAKpma2, this complex is exceptionally highly abundant. Binding of FC did not activate PMA2 (Table I). On the other hand, FC treatment was required for demonstration of the complex after native electrophoresis (Figs. 3 and 5). Thus, FC exerts its effect by stabilizing the complex between the plant H ϩ -ATPase and 14-3-3 proteins.
Physiological Significance of the PMA2⅐14-3-3 Complex in YAKpma2-In plasma membranes obtained from YAKpma2, we observed a high amount of a complex consisting of the plant PMA2 and yeast 14-3-3 homologs and able to bind FC (Figs. [1][2][3]. In contrast to the situation in plants, in vivo FC treatment of YAKpma2 did not increase the amount of the H ϩ -ATPase⅐14-3-3 complex (Fig. 1), indicating that it is extremely stable, even in the absence of FC. The exceptionally high abundance of FC binding sites in YAKpma2 plasma membranes compared with plant plasma membranes (Fig. 2) may thus be explained either by the higher amount of the complex in vivo and/or by the higher stability of the complex during yeast membrane isolation.
Is there any physiological significance for the formation of a plant PMA2⅐yeast 14-3-3 complex in YAKpma2? PMA2 is capable of functionally replacing the yeast H ϩ -ATPase, suggesting an active state of PMA2 despite of the presence of its 2 M. Piotrowski and C. Oecking, unpublished observation. complete C-terminal domain. Obviously, this autoinhibitory region is involved in complex formation with 14-3-3 proteins (Fig. 5). In plants the H ϩ -ATPase⅐14-3-3 complex represents an activated state of the H ϩ -ATPase in which the autoinhibitory action of its C-terminal region is suppressed (9 -11). Formation of a stable PMA2⅐14-3-3 complex in yeast might well be the reason why this plant isoform is capable of sustaining yeast growth in the absence of endogenous H ϩ -ATPase. We therefore propose that binding of yeast 14-3-3 homologs results in the formation of an activated state of PMA2 in YAKpma2. The lack of activation of PMA2 by FC in YAKpma2 can be explained by supposing that, because of the uncommon stability of the complex, PMA2 is saturated with yeast 14-3-3 homologs and consequently maximal activated. Therefore, in vivo FC treatment could not lead to a higher complex amount and consequently could not result in activation of the PMA2 pool. However, the 14-3-3⅐PMA2 complex does not represent the most active form of PMA2 since point mutations in various domains of PMA2 were shown to result in increased H ϩ pump activity (21), suggesting that in this case the H ϩ -ATPase undergoes a conformational change that results in full enzyme activity.
While this manuscript was in preparation, it was shown that yeast expressing the H ϩ -ATPase isoform AHA2 from A. thaliana was able to grow in the absence of yeast PMA1 but only on the condition that FC was added to the growth medium (36). FC binding sites in the yeast expressing AHA2 amounted to 3 pmol mg Ϫ1 plasma membrane protein (36), a value significantly lower than that determined for plant plasma membranes by using the same radioligand (Ref. 5, [ 3 H]dihydrofusicoccin: 25-30 pmol mg Ϫ1 ). The exogenous addition of yeast 14-3-3 isoforms in the presence of FC resulted in 5-to 10-fold higher FC binding activity and 2-fold higher ATPase activity of AHA2 (36). In contrast, in the work reported here, a much higher FC binding activity was demonstrated in yeast plasma membranes expressing the N. plumbaginifolia PMA2 (YAK-pma2) in the absence of any exogenous 14-3-3 ( Fig. 2: 113 pmol mg Ϫ1 versus 6 pmol mg Ϫ1 determined for plant plasma membranes by using [ 3 H]FCol as radioligand). Moreover, FC had no effect on ATPase activity (Table I). These data suggest that the N. plumbaginifolia PMA2 has a structure allowing a more stable binding of yeast 14-3-3 proteins than the A. thaliana AHA2. This might be an explanation for the capacity of PMA2  Conclusions-14-3-3 homologs bind directly to the C-terminal autoinhibitory domain of the plant plasma membrane H ϩ -ATPase, resulting in an activation of the enzyme. The complex consisting of a 14-3-3 dimer and the C terminus of the H ϩ -ATPase represents the binding site for the fungal phytotoxin fusicoccin and is stabilized by FC treatment.