The binding site for regulatory 14-3-3 protein in plant plasma membrane H+-ATPase: involvement of a region promoting phosphorylation-independent interaction in addition to the phosphorylation-dependent C-terminal end.

14-3-3 proteins constitute a family of well conserved proteins interacting with a large number of phosphorylated binding partners in eukaryotic cells. The plant plasma membrane H+-ATPase is an unusual target in that a unique phosphothreonine motif (946YpTV, where pT represents phosphothreonine) in the extreme C-terminal end of the H+-ATPase interacts with the binding cleft of 14-3-3 protein (Wurtele, M., Jelich-Ottmann, C., Wittinghofer, A., and Oecking, C. (2003) EMBO J. 22, 987-994). We report binding of 14-3-3 protein to a nonphosphorylated peptide representing the 34 C-terminal residues of the Arabidopsis plasma membrane H+-ATPase isoform 2 (AHA2). Following site-directed mutagenesis within the 45 C-terminal residues of AHA2, we conclude that, in addition to the 946YpTV motif, a number of residues located further upstream are required for phosphorylation-independent binding of 14-3-3. Among these, Thr-924 is important for interaction with 14-3-3 protein even when Thr-947 is phosphorylated. We suggest that the role of phosphorylation, which is accentuated by fusicoccin, is to stabilize protein-protein interaction between 14-3-3 protein and several residues of the H+-ATPase C-terminal domain.

14-3-3 proteins bind with high affinity to a large number (Ͼ100) of different proteins in eukaryotic cells (for a recent list, see Ref. 1). Since this interaction often results in changed activity of the target protein, 14-3-3 proteins have emerged as important regulators of enzyme function (1)(2)(3)(4)(5). Binding to a given target most often requires a phosphorylated binding motif (1,4,6,7), and signal transduction components such as protein kinases and protein phosphatases therefore become crucial for determining whether interaction between 14-3-3 protein and the target will occur in vivo (4,8). 14-3-3 proteins regulate the activity of several important plant enzymes. Among them are nitrate reductase (9,10), F 0 F 1 -ATP synthase (11), and the plasma membrane H ϩ -ATPase (12,13). Most secondary transporters in plant cells depend upon the activity of the plasma membrane H ϩ -ATPase, which is essential for generating the electrochemical gradient across the plasma membrane (for a review, see Ref. 14), and it has therefore generated considerable interest that 14-3-3 proteins have appeared as positive regulators of this proton pump (15).
No canonical 14-3-3 binding motifs are present in the plasma membrane H ϩ -ATPase, but deletion or substitution of the three C-terminal amino acid residues 946 YTV from the H ϩ -ATPase abolishes interaction with 14-3-3 protein (16 -18). In order for the 14-3-3 protein to bind to the plasma membrane H ϩ -ATPase, phosphorylation of the penultimate residue Thr-947 (to generate 946 YpTV) is a prerequisite (16 -18). Thr-947 is phosphorylated in vivo (19,20). Increased phosphorylation at this position occurs e.g. in stomatal guard cells as a response to blue light exposure concomitant with activation of proton pumping by the plasma membrane H ϩ -ATPase (20 -22).
Fusicoccin (FC) 1 is a fungal toxin that activates the plasma membrane H ϩ -ATPase in vivo (23). FC binding requires the simultaneous presence of 14-3-3 and the C-terminal end of H ϩ -ATPase (15,24,25). A plant 14-3-3 protein was recently co-crystallized with FC and a phosphorylated peptide (QSYpTV) representing the five C-terminal residues of a tobacco plasma membrane H ϩ -ATPase isoform (26). In the crystal structure, FC fills a cavity in the phosphopeptide/14-3-3 interaction surface where FC and the phosphopeptide mutually increase the binding affinity of each other by almost 2 orders of magnitude.
A peptide representing the last 16 amino acids of the C terminus of Arabidopsis thaliana H ϩ -ATPase isoform 2 (AHA2) binds 14-3-3 protein, provided that Thr-947 is phosphorylated, whereas a nonphosphorylated peptide will not bind even in the presence of FC (16). However, in the absence of any detectable phosphorylation, fusion proteins comprising the H ϩ -ATPase C-terminal hydrophilic domain (ϳ100 residues) are able to bind 14-3-3 protein, provided that FC is present (16,27,28). This indicates that specific residues or secondary structures located further upstream in the C terminus of the H ϩ -ATPase must be important for the nonphosphorylated, FC-induced binding of 14-3-3 to the target sequence. In this study, we show that several residues upstream of the 946 YTV sequence of AHA2 stabilize phosphorylation-independent interaction between 14-3-3 protein and AHA2. Among these, Thr-924 is essential for both fusicoccin binding and activation of AHA2 even when Thr-947 is phosphorylated.

EXPERIMENTAL PROCEDURES
Expression of AHA2 in the Yeast Saccharomyces cerevisiae-The yeast strain RS-72 was transformed and cultured essentially as described previously (29). 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, has been replaced by the galactose dependent promoter of GAL1. The cells were harvested, and plasma membranes were isolated as described (29,30). Growth tests on solid medium were performed as described (30).
Expression of Fusion Proteins in Escherichia coli-Expression and purification of MRGSH 6 -tagged Arabidopsis GF14-14 14-3-3 protein (15) and a fusion between 14-3-3 protein and cyan fluorescent protein (16) were as described. Glutathione S-transferase was fused to the AHA2 C-terminal amino acids Phe-860 to Val-948 of AHA2 by first amplifying DNA from the cloned AHA2 cDNA (31) using the forward primer 5Ј-CCG GAA TTC GCT TTC ACG ATG AAG AAA GA-3Ј and the reverse primer 5Ј-GGC GAG CTC TAC TAC ACA GTG TAG TGA CTG-3Ј. The resulting product was digested with EcoRI and XhoI and subcloned into the pGEX-4T-1 vector (Amersham Biosciences), resulting in the desired fusion between glutathione S-transferase and the C-terminal end of AHA2. Deletion mutants from the 5Ј end of AHA2 were made after the same principle but using alternative forward primers. To amplify amino acids Arg-880 to Val-948 of AHA2, the forward primer 5Ј-CCG GAA TTC CAA AGG ACA CTT CAC GGT-3Ј was employed. The segment encoding residues Ala-890 to Val-948 of AHA2 was amplified using the forward primer 5Ј-CCG GAA TTC GAA GCT GTT AAC ATC TTC-3Ј. The forward primers 5Ј-CCG GAA TTC  AGT TAC AGA GAA TTG TCT-3Ј, 5Ј-CCG GAA TTC CAA GCT AAA  AGA AGA GCT GAG-3Ј, 5Ј-CCG GAA TTC GCT GCT GAG ATG GCT-3Ј, 5Ј-CCG GAA TTC CTT AGG GAG CTG CAC AC-3Ј were used to  amplify AHA2 cDNA fragments encoding Tyr-900 to Val-948, Ala-910 to  Val-948, Ala-914 to Val-948, and Leu-919 to Val-948, respectively. SDS-PAGE and 14-3-3 Protein Overlay Assay-SDS-PAGE and Western blotting was performed as described (12). Equal amounts of plasma membrane protein were loaded in each well. The 14-3-3 protein overlay assay employing GF14-14 14-3-3⅐cyan fluorescent protein was as described (16). For detection of phosphothreonine residues, a specific antibody (71-8200; Zymed Laboratories Inc.) was employed. For detection of plasma membrane H ϩ -ATPase, an antibody directed against the N terminus of the AHA3 protein was used (32).
Synthesis of Peptides-Peptides biotinylated at their N termini were synthesized and characterized by Thistle Research (Glasgow, Scotland, UK). The purity of the peptides was Ͼ95%. The sequences and the nomenclature of the peptides were as follows: TP9113, Biotinamidocaproic-EIARELHTLKGHVESVVKLKGDIETPSHYpTV; TP9115, Biotin-amidocaproic-EIARELHTLKGHVESVVKLKGDIETPSHYTV.
Surface Plasmon Resonance Spectroscopy-The BIAcore 3000 system from Biacore AB (Uppsala, Sweden) was used. Biotinylated peptides (400 resonance units) were captured on an Biacore SA sensor chip precoated with streptavidin. The interaction was followed in real time at different 14-3-3 and FC concentrations as indicated at 25°C. Premixed solutions of 14-3-3 protein and FC in running buffer (20 mM Hepes, 140 mM NaCl, 10 mM KCl, and 5 mM MgCl 2 , pH 7.0) were injected into the flow cell and passed over the peptide surfaces at a continuous flow rate of 10 l/min to monitor association. Immediately after this injection, buffer without 14-3-3 protein but including FC was injected, followed by injection of running buffer without 14-3-3 as well as FC. Each association-dissociation cycle was followed by two 30-s injections with 10 mM NaOH to remove 14-3-3 protein bound to the immobilized peptides. After regeneration, a 30-min wait period was introduced before the next cycle of binding and regeneration.
All sensorgrams were corrected for bulk refractive indexes by reference-subtracting with sensorgrams from a flow cell in which only streptavidin was immobilized. The data were evaluated using BIAevaluation 3.1 (Biacore AB). To obtain K D and R max values, a 1:1 Langmuir binding model was fitted to the sensorgrams. In experiments evaluating the effect of varied FC concentrations, the observed rate constants (k obs ) were used to estimate association.
ATPase Assays-ATP hydrolytic activity was measured essentially as described (29). Typically, the assay medium (pH 7.0) included 5 M FC and 2.5 g of 14-3-3 protein and the reaction was initiated by the addition of 3-5 g of yeast plasma membranes. The specific ATP hydrolytic activities of the nonactivated AHA2 plasma membrane H ϩ -ATPase expressed in yeast were in the range of 1.2 Ϯ 0.2 mol of P i released/mg of membrane protein/min. The basal activities of mutant enzymes were similar. In peptide experiments, the assay media (pH 6.5) contained 20 M FC and 5 g of 14-3-3 protein, and the peptides were preincubated in the assay buffer 30 min prior to the assay was started. Peptide assays were started by the addition of microsomes isolated from yeast expressing AHA2.
[ 3 H]FC Binding Assay-FC binding was measured to endoplasmic reticulum isolated from transformed yeast cells. Binding assay was performed in 200 l of buffer containing 50 mM Mes (pH 6.5), 5 mM MgCl 2 , and 50 g of GF14-14 14-3-3 protein. The assay was started by the addition of 25 g of plasma membranes. After incubation at 25°C for 3 h, the membranes was precipitated by centrifugation (10 min, 25 ϫ g, 4°C). The pellet was washed three times with buffer and afterward dissolved in 100 l of 7% trichloroacetic acid. Radioactivity was measured in a scintillation counter. All samples ware made as triplicates. Nonspecific [ 3 H]FC binding was monitored in parallel for each sample by the addition of an excess of unlabeled FC (10 Ϫ4 M).

A Nonphosphorylated Peptide Corresponding to the C-terminal 34 Residues of AHA2 Inhibits Activation of AHA2 by 14-3-3
Protein-The binding site for 14-3-3 on plant plasma membrane H ϩ -ATPase involves the phosphorylated tripeptide motif 946 YTV in the C-terminal end ( Fig. 1A) (16 -18, 26), but additional structural features might also be involved in formation of the complex (16,28). Fusion proteins involving fragments of different lengths of the AHA2 C terminus, all including the 946 YTV motif were expressed in E. coli and used for 14-3-3 overlay assays. Thereby, we could identify the minimal sequence necessary for phosphorylation independent interaction with 14-3-3 proteins, namely 35 amino acids as seen in Fig. 2. Additional upstream residues seemed to stabilize the binding further as noted previously (28).
To learn more about the phosphorylation-independent interaction between AHA2 and 14-3-3 protein, we synthesized two longer peptides derived from the C terminus of AHA2 (Glu-915 to Val-948; 34 amino acids) and introduced threonine or phos- A, the 54 C-terminal residues of AHA2. The 946 YTV motif interacting with the binding cleft of 14-3-3 protein is marked by a blue box. The autoinhibitory region II is marked by a red box. Residues in this region that when substituted produce an activated pump (Ref. 30 and this work) are marked with yellow (class B) and red (class C) squares above the sequence. Thr-924, shown to be essential for the interaction between 14-3-3 protein and H ϩ -ATPase is marked in boldface letters. B, minimal C-terminal sequence of AHA2 (35 amino acid residues) that, when fused to glutathione S-transferase, interacts with 14-3-3 protein. C, sequences of peptides used in this work. Phosphorylated Thr-947 is marked by a circled P below this residue. phothreonine residues, respectively, at the penultimate positions (Fig. 1C). We subsequently tested the effect of the AHA2derived peptides for their ability to inhibit activation of AHA H ϩ -ATPase by 14-3-3 proteins. This was done in the presence of FC to promote the interaction between the binding partners.
The addition of 14-3-3 protein increased ATP hydrolytic activity of full-length AHA2 about 3-fold (pH 7.0). As seen in Fig.  3, the phosphorylated peptides (C t 16-P and C t 34-P) inhibited the 14-3-3 protein induced activation with comparable potency and at low concentrations (K i Ϸ 0.1 M) without affecting basal ATPase activity. The long nonphosphorylated peptide (C t 34) similarly inhibited the 14-3-3 protein-induced increase in activity, but significantly higher concentrations of peptide were required to observe this effect (K i ϭ 3 M). The shorter nonphosphorylated peptide (C t 16) had no effect even at the highest concentration tested (30 M). This indicated that, in contrast to the shorter peptide (C t 16), the long nonphosphorylated peptide (C t 34) interacts directly with 14-3-3 protein in the presence of FC. This further reinforces the notion that residues upstream of Val-933 in AHA2 are able to promote a phosphorylationindependent interaction with 14-3-3 protein.
Interaction between 14-3-3 Protein and Peptides Corresponding to the C-terminal End of AHA2 Detected by Surface Plasmon Resonance Spectroscopy-We next employed the synthetic peptides for protein-protein interaction studies using surface plasmon resonance spectroscopy (Table I). By employing this technique, it has previously been shown that an nonphosphorylated peptide representing the 16 C-terminal amino acid residues of AHA2 does not bind 14-3-3 protein irrespective of the presence of FC (16) ( Table I). The two longer peptides of 34 residues (C t 34 and C t 34-P) were immobilized to biosensor chips, and 14-3-3 protein (100 nM) and FC (varying concentrations) were injected into the flow system (Fig. 4). This allowed for detection of association between peptides and 14-3-3 protein. Next, 14-3-3 protein was omitted from the flow buffer, whereas FC was kept constant to follow dissociation of already bound 14-3-3 protein. Finally, only buffer was injected into the flow system to measure dissociation of 14-3-3 protein in the absence of FC. Under these conditions, the immobilized peptides did not bind bovine serum albumin control protein, and an unrelated immobilized control peptide representing the N terminus of pig kidney ␣ 1 Na,K-ATPase did not bind 14-3-3 protein (data not shown).
Indeed, both the nonphosphorylated and the phosphorylated peptides bound 14-3-3 protein in this system, although the phosphorylated peptide C t 34-P exhibited the highest affinity for 14-3-3 protein (Table I). However, it was also evident that 14-3-3 protein had an approximately 10 times higher maximum binding response (R max ) for phosphorylated than for nonphosphorylated peptide (Table I). Equal amounts of peptide had been immobilized on the biosensor chips, and mass spectrometry analysis verified that the compositions of the peptides were as expected (data not shown). Thus, it remains a possibility that the nonphosphorylated peptide exists in more than one conformation and that only one of these conformations is capable of interacting with 14-3-3 protein.
Equilibrium dissociation constants (K D ) were determined from steady-state binding values predicted from fits to the data in Fig. 4. Comparison of K D values of 14-3-3 in complex with phosphorylated and nonphosphorylated peptides demonstrated that phosphorylation increased the affinity for 14-3-3 protein severalfold (Table I).
The addition of FC increased the maximal binding of 14-3-3 to both peptides significantly (Fig. 4). This implies that a preformed complex between 14-3-3 protein and AHA2 forms the basis for FC-dependent high affinity complexes irrespective of whether they are phosphorylated or not. For both peptides, values for K D and R max were also determined in the presence of  a Sequences of peptides are given in Fig. 1. Numbers designate amino acid residues counted from the C terminus. P symbolizes phosphorylation at position Thr-947, the penultimate residue.
b Data from Ref. 16. c NA, no association could be detected.
Phosphorylation-independent Binding of 14-3-3 Protein 5 M FC. FC at this concentration decreased K D ϳ4-fold for the phosphorylated peptide and 3-4-fold for the nonphosphorylated peptide, whereas R max was not changed significantly by FC (Table I).
Taken together, these data support the notion that there is a region of interaction between AHA2 and 14-3-3 protein in the stretch of amino acids of the longer peptides (C t 34) not represented in the shorter peptides (C t 16), namely between Glu-915 and Val-933.

Kinetics of Interaction between 14-3-3 Protein and Peptides
Corresponding to the C-terminal End of AHA2-The observed rate constants (k obs ) measured in the Biacore system relate to the association rate constant (k a ) as follows: k obs ϭ k a j C ϩ k d . By rearrangement and approximation, this gives k a Ϸ k obs /C, when k d Ͻ Ͻ k obs . Since k d values in the experiments reported here were at least 1 order of magnitude lower than the corresponding k obs values, it is reasonable to take k obs as an approximate measure for k a .
It is notable that for the phosphorylated peptide (Fig. 4C), the observed rate constants (k obs ) are decreasing with increasing FC concentrations. This implies that the build-up of the 14-3-3⅐ATPase complex is slowed by FC. The effect of FC on the dissociation rate constant (k d ) is, however, much more pro-nounced than its effect on the association process (Fig. 4D). For the phosphorylated peptide C t 34-P, the concentration of FC needed to decrease k d to 50% was found to be 120 nM, about 1 order of magnitude lower than the concentration required to reduce the apparent association rate to 50% of maximal reduction. Thus, the overall effect is an efficient stabilization of the complex. When changing the flow to buffer only, dissociation increased, indicating that FC is reversibly bound in the 14-3-3⅐phosphopeptide complex (Fig. 4, A and B).
For the nonphosphorylated peptide (C t 34), FC likewise reduced the apparent dissociation rate (Fig. 5B), where the effect on the dissociation rate constant was much less pronounced (Fig. 5C). However, it is clear from Fig. 5A that the dissociation of 14-3-3 protein from immobilized peptide is increased when FC is omitted from the injection buffer. Thus, FC acts similarly to stabilize the complex between 14-3-3 protein and nonphosphorylated peptide.
Growth of Yeast Expressing AHA2 Mutants-Alanine-scanning mutagenesis was next employed to identify residues in the C-terminal end of the Arabidopsis plasma membrane H ϩ -ATPase AHA2 that affect the strength of the interaction between 14-3-3 protein and H ϩ -ATPase. A convenient screening system was employed in which growth of transformed yeast cells is dependent on the activity status of heterologously expressed plant H ϩ -ATPase. In the yeast S. cerevisiae, PMA1 is an essential gene, encoding a P-type plasma membrane H ϩ -ATPase. Following expression in yeast, full-length AHA2 H ϩ -ATPase promotes slow growth in the absence of the endogenous yeast plasma membrane H ϩ -ATPase Pma1p (Fig. 6). This is due to association of endogenous yeast 14-3-3 protein with a subset of AHA2 H ϩ -ATPases that have been phosphorylated at Thr-947 by endogenous yeast kinase(s) (16, 33). The AHA2-dependent growth is further stimulated by FC (Fig. 4). The role of FC is most likely to promote an interaction between yeast Phosphorylation-independent Binding of 14-3-3 Protein 14-3-3 protein and AHA2, independent of the phosphorylation status of the H ϩ -ATPase (15,33).
A stretch of 23 consecutive residues in the C-terminal domain of full-length AHA2 H ϩ -ATPase, from Glu-908 through Glu-930, were substituted individually with Ala or Ser. As described previously (30), substitution of a number of residues in the autoinhibitory region II situated between Ser-904 to Leu-919 results in H ϩ -ATPase mutants that can substitute for Pma1p (marked by yellow and red squares in Fig. 1A). To test whether any of the plant plasma membrane H ϩ -ATPase AHA2 substitutions were unable to respond to FC, each mutant was expressed in the yeast strain RS-72, in which the constitutive promoter of the PMA1 gene has been replaced by the galactosedependent GAL1 promoter at the chromosomal level (34). 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. Growth on glucose was tested both in the presence and absence of FC (Fig. 6). Strikingly, only one substitution, T924A, failed to complement Pma1p regardless the presence of FC in the growth medium. However, other substitutions, such as E915A, H923A, L925A, and V929A, showed a reduced response to FC.
Identification of Single Residues Essential for Phosphorylation-independent Binding of 14-3-3 to the C Terminus of AHA2-Binding of 14-3-3 protein to AHA2 can be studied in vitro in an overlay assay following separation of yeast plasma membrane polypeptides by SDS-PAGE and transfer to a nitrocellulose membrane. AHA2 substitutions in the stretch between Glu-921 and Ser-931 were tested for their ability to bind 14-3-3 protein in the absence of FC (Fig. 7C). 14-3-3 protein interacted strongly with a 97-kDa band in the preparation expressing wild-type AHA2. No polypeptides in the same molecular weight region were detected in a control yeast expressing Pma1p but not AHA2. Interestingly, H923A and T924A substitutions of AHA2 showed a strongly reduced ability to interact with 14-3-3 protein. Other substitutions, such as L925A, K926A, G927A, V929A, and E929A, also showed a somewhat reduced ability to interact with 14-3-3 protein.
Reduced ability of some AHA2 substitutions to bind 14-3-3 in vitro and to replace Pma1p in vivo could simply be due to a failure of yeast to produce mutant enzymes in sufficient quantities. However, the amount of mutated AHA2 in yeast plasma membranes, as detected in Western blots by a plant H ϩ -ATPase-specific antibody (Fig. 7A), were not reduced compared with wild type AHA2, indicating that there was no correlation between expression levels of the plant enzyme and the ability to either substitute for Pma1p or to bind 14-3-3 protein.
Phosphorylation Status of AHA2 Substitutions Expressed in Yeast-The possibility remained that the impaired ability of some AHA2 mutants to support yeast growth and to bind 14-3-3 protein was correlated with a decreased phosphorylation level of Thr-947. Reduced phosphorylation at Thr-947 is expected to lead to reduced interaction with 14-3-3 and lower H ϩ -ATPase activities in vivo. Fig. 7B demonstrates that a 97-kDa band in plasma membranes of yeast cells expressing plant AHA2, but absent in yeast expressing Pma1p, was recognized by an anti-phosphothreonine antibody. Apparently, wild type AHA2 as well as mutant proteins were all phosphorylated at Thr residues at comparable levels.
It was therefore of interest to investigate whether this ability was preserved in full-length AHA2 proteins carrying substitutions in the region around Thr-924. Mutant proteins carrying substitutions in this region had specific ATP hydrolytic activi- ties comparable with those of wild type AHA2 protein (data not shown), suggesting that they do not contribute to C-terminal autoinhibition of the enzyme. However, in contrast to wild type AHA2, the H923A and T924A substitutions did respond with only a minor increase in activity following the addition of 14-3-3 protein and FC to the assay medium (Fig. 8). This observation supports the notion that His-923 and Thr-924, and possibly also other nearby residues, are important for the activation mechanism of 14-3-3 proteins.
FC-binding Properties of Mutated AHA2 Protein-The binding site for FC is complex, since this compound interacts with both 14-3-3 protein and the C-terminal region of plant plasma membrane H ϩ -ATPase (15,26). To investigate the role of Thr-924 and nearby residues in FC binding to nonphosphorylated H ϩ -ATPase, full-length AHA2 proteins carrying substitutions in the stretch from Glu-921 to Leu-925 were expressed in yeast, and membranes of yeast endoplasmic reticulum enriched in recombinant plant enzyme were isolated. It has previously been demonstrated that heterologous AHA2 accumulating in yeast endoplasmic reticulum does not get phosphorylated at Thr-947 (33). The membranes were mixed with recombinant 14-3-3 protein produced in E. coli and incubated with tritiated [ 3 H]FC. It appeared that the T924A substitution had completely lost its ability to bind FC (Fig. 9). Other substitutions in the region around Thr-924, such as Glu-921, Leu-922, His-923, and Leu-925, had a lowered ability to bind FC, indicating that these residues also are important for the interaction of H ϩ -ATPase with FC. Since these residues are separated in space from the FC binding site observed in the crystal structure (26), the effect is most likely indirect and reflects insufficient binding of H ϩ -ATPase to 14-3-3 protein, resulting in reduced formation of the FC binding site. DISCUSSION The data reported here throw light on a number of important questions related to the unusual interaction between plant plasma membrane H ϩ -ATPase and 14-3-3. (i) How is phosphorylation-independent complex formation possible? (ii) How does the C-terminal end of plant H ϩ -ATPase interact with 14-3-3? (iii) How does FC act to stabilize the complex?
Structural Basis for Phosphorylation-independent Binding of H ϩ -ATPase to 14-3-3 Protein-We show that a peptide comprising the 34 C-terminal residues (Glu-915 to Val-948) of AHA2 is sufficient for forming a phosphorylation-independent interaction with 14-3-3 even in the absence of FC. In contrast, a peptide comprising the 16 C-terminal residues (Val-933 to Val-948) has to be phosphorylated at Thr-947 before interaction can take place. This points to residues situated between Glu-915 and Val-933 as being important for establishing the phosphorylation-independent contact. By systematically analyzing Cterminal residues by site-directed mutagenesis, we identified Thr-924 in this region as being essential for the formation of the 14-3-3⅐AHA2 complex. In addition, a number of other residues in the same region are important for the interaction between H ϩ -ATPase and 14-3-3 protein. These data support the notion that residues in the C-terminal end of AHA2 apart from the 946 YTV motif are interacting with the 14-3-3 protein.
Thr-924 is conserved in most plant plasma membrane H ϩ -ATPases and might allow for specific bonding with 14-3-3 residues outside the binding cleft involved in interaction with the YTV sequence.
It has previously been observed that interaction between 14-3-3 protein and a binding partner can occur outside the conserved 14-3-3 binding cleft. Thus, the structural analysis of the 14-3-3⅐serotonin N-acyltransferase complex has revealed that there are extensive interactions between the two binding partners beyond those involving the phosphothreonine motif of serotonin N-acyltransferase, which docks into the 14-3-3 binding groove (35). In fact, most of the interaction surface, involving a total of 37 14-3-3 residues, is distant from the epitope binding groove. Whereas the contact between 14-3-3 and the phosphothreonine region is expected to mostly involve polar contacts, less specific hydrophobic contacts are used for the rest of the interface. However, interaction between the AHA2 residue Thr-924 and the surface of 14-3-3 protein may be more specific.
Theoretically, two separate and distinct sites for 14-3-3 binding could exist in the C-terminal domain of the H ϩ -ATPase, one involving the 946 YpTV motif, required for phosphorylation-de-FIG. 8. Ability of different AHA2 C-terminal substitutions to become activated by 14-3-3 protein in the presence of FC. In order to test the ability to become activated by 14-3-3 protein and FC, ATP hydrolytic activity was measured in purified plasma membranes from yeast (2.5 g) expressing the different AHA2 mutants. Control values (100%) correspond to the activity in the absence of 14-3-3 protein and FC. The presented data are the average of three independent experiments. pendent interaction, and the other involving upstream residues allowing for phosphorylation-independent binding. However, this model can be discarded, since, following deletion of the C-terminal amino acids 946 YTV from the full-length AHA2 H ϩ -ATPase, binding of 14-3-3 protein is completely abolished even in the presence of FC (16). Therefore, phosphorylationindependent binding not only includes upstream residues but also involves the 946 YTV motif. It is possible that the residues act in concert to bind 14-3-3 protein and thus compensate for the lack of phosphorylation at Thr-947.
A number of cases have been described in the literature where 14-3-3 proteins interact with a target in a phosphorylation-independent manner. Screening of a random peptide phage library revealed several nonphosphorylated peptides that bind 14-3-3 protein (36). In the most frequently occurring epitope, R18, it appears that a cluster of negative charges (WLDLE) compensate for the lack of the negatively charged phosphate group when interacting with 14-3-3 protein (36). Co-crystallization demonstrated that R18 binds in the conserved amphipathic groove of 14-3-3 (37). In this context, it is noteworthy that in AHA2 there are no negative charges in the immediate vicinity of Thr-947.
Exoenzyme S (ExoS) from Pseudomonas aeruginosa likewise binds 14-3-3 in a nonphosphorylated form (38,39). The site of interaction between ExoS and 14-3-3 protein has been localized to four residues in the C-terminal region of ExoS (40,41). The sequence of this region ( 424 DADL) includes negatively charged residues. Mutational analysis of 14-3-3 proteins has demonstrated that ExoS selectively employs residues in the amphipathic binding groove in order to bind (37,39) and uses the same charged residues in 14-3-3 as pRaf-1 as contact sites (39). Recently, an additional nonphosphorylated 14-3-3 binding epitope, the guanine nucleotide exchange factor p190RhoGEF, was identified, and the binding site was narrowed down to 1370 IQAIQNL (40). The structural basis for this binding is not known. Does 14-3-3 Protein Interact with Autoinhibitory Regions of Plant H ϩ -ATPase?-Within the C-terminal regulatory domain of plant H ϩ -ATPase, there are two autoinhibitory regions, regions I and II. In AHA2, these regions comprise residues Lys-863 to Leu-885 and Ser-904 to Leu-919, respectively. Based on deletion studies and site-directed mutagenesis, it has been suggested that 14-3-3 may interact with region II of NpPMA1, a tobacco plasma membrane H ϩ -ATPase (28). We have confirmed that this is also true for AHA2, an H ϩ -ATPase belonging to another phylogenetic subgroup than tobacco Nicotiana plumbaginifolia H ϩ -ATPase isoform 2 (14). However, by alanine-scanning through all residues in region II, we observed no substitutions in this segment that completely abolished binding of 14-3-3 protein, although interaction was significantly reduced following substitution of residues in a consecutive stretch between Glu-908 to Glu-915 (data not shown). This would suggest that, in addition to residues further downstream of region II, interaction between 14-3-3 and region II indeed takes place but involves several residues, none of which are essential for complex formation. Since the constraint on enzyme activity by the C-terminal regulatory domain of AHA2 is released following 14-3-3 interaction, complex formation could possibly stabilize a conformation of region II that does not interact with the rest of the AHA2 polypeptide.
Stabilization of the H ϩ -ATPase⅐14-3-3 Protein Complex by FC-An unexpected observation in this study was that FC slows down the build-up of the complex between the H ϩ -ATPase peptide and 14-3-3 protein. Possible explanations could be that FC, after binding to the active site in 14-3-3 proteins (26), constitutes a steric hindrance for binding of peptide and/or reduces the mobility of peptide during the association process. However, since FC strongly inhibits dissociation of the peptide from 14-3-3 protein once it is bound, the overall effect of this fungal toxin is to efficiently stabilize the protein-protein complex.