- Email: [email protected]
A WNK kinase binds and phosphorylates V-ATPase subunit C
FEBS Letters 580 (2006) 932–939
A WNK kinase binds and phosphorylates V-ATPase subunit C Anne Hong-Hermesdorfa, Angela Bru¨xa, Ardina Gru¨berb, Gerhard Gru¨berb, Karin Schumachera,* b
a Universita¨t Tu¨bingen, ZMBP-Plant Physiology, Auf der Morgenstelle 1, 72076 Tu¨bingen, Germany Nanyang Technological University, School of Biological Sciences, 60 Nanyang Drive, Singapore 637551, Singapore
Received 11 November 2005; revised 14 December 2005; accepted 4 January 2006 Available online 18 January 2006 Edited by Michael R. Sussman
Abstract WNK (with no lysine (K)) protein kinases are found in many eukaryotes and share a unique active site. Here, we report that a member of the Arabidopsis WNK family (AtWNK8) interacts with subunit C of the vacuolar H+-ATPase (V-ATPase) via a short C-terminal domain. AtWNK8 is shown to autophosphorylate intermolecularly and to phosphorylate Arabidopsis subunit C (AtVHA-C) at multiple sites as determined by MALDI-TOF MS analysis. Furthermore, we show that AtVHA-C and other V-ATPase subunits are phosphorylated when V1-complexes are used as substrates for AtWNK8. Taken together, our results provide evidence that V-ATPases are potential targets of WNK kinases and their associated signaling pathways. 2006 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Vacuolar H+-ATPase; V1-complex; WNK; Phosphorylation; Arabidopsis
1. Introduction pH is an essential parameter for most biological processes and vacuolar H+-ATPases (V-ATPases) create and maintain the pH-differentials suited to the biochemical functions of the endomembrane compartments of all eukaryotic cells. V-ATPases are multi-subunit enzyme complexes composed of the cytosolic ATP hydrolyzing V1 subcomplex (subunits A–H) and the membrane bound proton translocating V0 subcomplex (subunits a, c, c 0 , c00 , d and e) [1,2]. ATP-dependent proton transport energizes secondary active transport and controls cellular pH homeostasis and it is therefore not surprising that a plethora of processes including receptor-mediated endocytosis, protein processing, virus entry and bone resorption has been shown to depend on V-ATPase activity. Moreover, it has recently been shown that the V0-complex is directly involved in membrane fusion events. Despite the remarkable diversity of functions that V-ATPases serve, knowledge of their regulation and in particular their integration into cellular signaling networks is limited. V-ATPases exist in a dynamic equilibrium between fully assembled complexes and reversibly disassembled V1 and V0 * Corresponding author. Fax: +49 7071 293287. E-mail address: [email protected] (K. Schumacher).
Abbreviations: WNK, with no lysine (K); V-ATPase, vacuolar H+ATPase; VHA-C, vacuolar H+-ATPase subunit C; MALDI-TOF MS, matrix assisted laser desorption-ionisation/time-of-flight mass spectrometry
subcomplexes [3–5]. Depending on the energy status of the cell, this equilibrium can be rapidly shifted [3]. Since neither the V1 nor the V0 subcomplex is fully active on its own [6], reversible disassembly provides a means to save energy in times of starvation. Moreover, the reversibility of the process enables the cell to quickly adapt to energy re-supply. Conventional glucose signal transduction is not involved in this process [7] and the molecular features of this mechanism remain to be elucidated. Subunit C occurs in substoichiometric amounts in free V1 complexes and in a subunit C-deficient yeast strain V0 and V1 subcomplexes are present, but do not assemble and are both inactive [8–10]. Therefore, it seems likely that subunit C plays a key role in reversible assembly. Structural analysis of subunit C revealed an elongated, boot like shape [11,12] with three distinct domains: a globular head domain, an elongated ‘neck’, and a globular foot domain. The head domain is thought to connect to the V1 complex, while the foot and lower neck domains bind to the V0 complex [11,12]. The structure of subunit C fits its proposed function in coupling the activities of V1 and V0 subcomplexes [3,13]. Moreover, it has recently been shown that the subunit C from Manduca sexta interacts with actin filaments [14] and that subunits C from Arabidopsis and yeast undergo major structural changes after binding of adenine nucleotides [15]. Both properties are ideally suited for a protein controlling the assembly state of the V-ATPase complex, but it remains to be determined if either actin or nucleotide binding are indeed necessary for this process. The regulator of the H+-ATPase of vacuolar and endosomal membranes complex (regulator of the H+-ATPase of vacuolar and endosomal membranes), consisting of Skp1p, Rav1p, and Rav2p, was shown to enhance V-ATPase assembly through interaction with the V1 moiety in yeast and it is therefore conceivable that interactions of subunit C with other proteins are necessary for complex formation [16,17]. Protein phosphorylation is ubiquitous and the most widespread and best studied signaling mechanism in eukaryotic cells [18]. Surprisingly, phosphorylation of V-ATPase subunits by protein kinases has not been demonstrated so far, although it has been reported that AP50, a subunit of the AP2-complex of bovine clathrin-coated vesicles, phosphorylates the V-ATPase subunit B in vitro [19]. In plants, it has recently been shown that a calcium-dependent protein kinase affects V-ATPase activity and is able to phosphorylate vacuolar proteins [20], however it remains to be determined if the V-ATPase itself is phosphorylated. WNK (with no lysine (K)) kinases are a subfamily of serine/ threonine protein kinases related to the STE20/PAK-like family [21]. A lysine residue in kinase subdomain II, which is essential for the coordination of ATP in the active center and
0014-5793/$32.00 2006 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2006.01.018
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
conserved among all other kinases, is missing in this subfamily [22,23]. It was shown to be replaced by a lysine residue in subdomain I [22] that is characteristic for members of the WNK family. WNK kinases have been found in numerous eukaryotes, but are not found in yeast and it has been proposed that they are restricted to multicellular organisms [21,24]. Mammalian WNKs have been shown to regulate ion transport processes in epithelial cells by reducing either insertion or retention of transport proteins in the plasma membrane [25]. This can be mediated either by a direct interaction with transport proteins as shown for WNK4 and NCCT1 [26] or by interaction with trafficking proteins in the case of WNK1. Upon phosphorylation by WNK1 membrane binding properties of synaptotagmin 2 change [27] and this could affect trafficking events controlling the membrane presence of transport proteins. Information about the functions of WNK kinases from other organisms is limited. The model plant Arabidopsis has been reported to have nine members of the WNK family [28] but only one of them has so far been characterized. AtWNK1 phosphorylates the putative circadian clock component APRR3 in vitro and might be involved in a signal transduction cascade regulating its biological activity [29]. To gain further insight into the molecular properties of the V-ATPase subunit C of Arabidopsis (AtVHA-C), we identified potential interacting proteins using the yeast two-hybrid system. Here, we show that AtWNK8, a member of the Arabidopsis WNK family, binds to and phosphorylates AtVHA-C.
2. Materials and methods 2.1. Yeast two-hybrid assay The yeast strain AH109 (MATa, trp1-901, leu2-3,112, ura3-52, his3-200, gal4D, gal80D, LYS2: GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3: MEL1UAS-MEL1TATALacZ; Clontech) was used to test interaction between AtVHA-C and the C-terminal sequences of AtWNK1-10. AtVHA-C was PCRamplified and after sequence analysis cloned into pGBKT7 (BamHI/ PstI). AtWNK 1–5, 7–10 C-termini were amplified by PCR using a Arabidopsis Col-0 cDNA library as template, subcloned for sequence analysis and cloned into pGADT7 vector (Cfr9I/ XhoI). Yeast cells were cotransformed according to standard protocols with bait and prey plasmids and transformants were selected on synthetic dropout plates lacking Trp and Leu. To test for interaction, cells were grown on selective synthetic media lacking Trp, Leu, His and Ade. 2.2. Constructs, mutagenesis, proteins pEG-AtWNK8 was used for site-directed mutagenesis at position D157 to alanine (D157A) and the mutation verified by DNA sequencing. AtWNK8 FL (full length), DN1, DN2, DC1, DC2 were amplified by PCR (template pEG-AtWNK8 D157A) and cloned into pGEX4T3 – translationally fused to an N-terminal GST-tag. Restriction sites used for cloning were BamHI and Bsp120I. A C-terminal His-tag was introduced through the reverse primers. Proteins were expressed in and purified from Escherichia coli strain BL21DE3 (B F dcm, ompT, hsdSðr B mB Þ gal k (DE3)) by GSH-affinity chromatography according to standard protocols. Wildtype and mutant (K41M/ D157A) AtWNK8 was expressed using the IMPACTe-CN system (New England Biolabs, Frankfurt, D). The coding sequence (1692 bp) was amplified by PCR using pEG-AtWNK8 or pEG-AtWNK8 D157A as template and cloned into pTYB2 (NdeI/Cfr9I). The mutation K41M was introduced by site-directed mutagenesis, using AtWNK8 in pTYB2 as template. Proteins were expressed in E. coli BL21DE3 cells and purified by chitin affinity chromatography. Purification of subunit C from Arabidopsis (AtVHA-C) and Saccharomyces cerevisiae (Vma5p) was performed as described in [11,15], respectively. The C-subunit of Caenorhabditis elegans (CeVha11) was
933 amplified by PCR and cloned into pET28a (BamHI/HindIII). The plasmid for expression of M. sexta V-ATPase C subunit was kindly provided by Prof. H. Wieczorek (University of Osnabru¨ck). Proteins were expressed and purified from E. coli BL21DE3 cells by Ni2+-affinity chromatography according to standard protocols. The integrity of the proteins was tested by Coomassie staining and their concentration measured by Bradford analysis. 2.3. Far-western analysis Proteins were slot blotted onto a nitrocellulose membrane and incubated in blocking buffer (TBS containing 2% (w/v) protein powder) (1 h), and the putative interaction partner was added for >1 h in the same buffer (2–10 lg/ml). After intensive washing (TBS-T) the interaction was detected by immunostaining of the bound protein using a specific antibody (>1 h at RT). Unspecifically bound antibodies were eliminated through washing steps with TBS-T. Membrane was incubated with HRP conjugated goat-anti-rabbitIgG antibodies (blocking buffer; 1 h at RT). Signals were detected using the SuperSignal West Pico Chemiluminescent (Pierce Perbio, Bonn, D). 2.4. In vitro phosphorylation assay AtWNK8 kinase (0.05 lg) was incubated with 0.3 lg substrate for 30 min at 30 C in optimized kinase assay buffer (25 mM Tris–HCl, pH 7.0, 0.2 mM EDTA, and 5 mM MnCl2). The ATP concentration was 25 lM containing 5 lCi [c-32P]ATP (0.1 lM) per reaction. Reactions were stopped by adding 1 volume 2· SDS sample buffer (4% SDS, 140 mM Tris–HCl, pH 6.8, 20% Glycerol, 0.01% Bromophenol blue, and 10% b-mercaptoethanol) and incubating at 95 C for 5 min. Samples were separated on SDS–PAGE, and after gel drying exposed on an X-ray screen for 1–2 h at RT. 2.5. Tryptic digest and MALDI-TOF analysis Phosphorylated and unphosphorylated AtVHA-C containing bands were cut from the gel and destained overnight with a solution of 50 mM ammonium bicarbonate, 40% ethanol. The protein was digested in gel with trypsin (Promega) according to Roos et al. [30] except that the bands had been washed three times with acetonitrile before drying them in a speed vacuum concentrator. Digested samples were desalted with a C18ZipTip (Millipore) and eluted with CHCA (10 mg/ml a-cyano-4-hydroxycinnamic acid in 50% acetonitrile, 0.1% trifluoroacetic acid) or FA (8 mg/ml 3-methoxy-4hydroxycinnamic acid in 50% acetonitrile, 0.1% trifluoroacetic acid) matrix solution. 1–2 ll of matrix-analyte solution was spotted onto the MALDI plate and allowed to dry. Peptide mass mapping was performed by matrix assisted laser desorption-ionisation/time-offlight mass spectrometry (MALDI-TOF MS) using a Bruker Reflex IIIe MALDI-TOF spectrometer (Bremen, Germany). The peptide map was acquired in reflectron positive-ion mode with delayed extraction at a mass range of 500–6000 Da. The instrument was calibrated using a calibration mixture (Applied Biosystems). For interpretation of the protein fragments the PEPTIDEMASS [31] program available at Expasy web site (www.expasy.ch/tools/peptide-mass. html) was used. 2.6. Accession numbers and data analysis Megalign (Clustal method) was used for sequence alignments; the phylogenetic tree is a result of a full heuristic distance analysis with 1000 bootstraps replicates in Paup4.0b10. For all sequences, GenBank Accession Nos. are supplied; for Arabidopsis sequences additional loci characterization is provided: AtWNK1 – At3g04910, AB084266; AtWNK2 – At3g22420, AB084267; AtWNK3 – At3g48260, AB085616; AtWNK4 – At5g58350, AB084268; AtWNK5 – At3g51630, AB084269; AtWNK6 – At3g18750, AB084270; AtWNK7 – At1g49160, AB085617; AtWNK8 – At5g41990, AB084271; AtWNK9 – At5g28080 NM_122691; AtWNK10 – At1g64630, NM_105138; AtWNK11 – At5g55560, NM_124938; OsWNK1 – XP_478792; OsWNK2 – AAU44135; OsWNK3 – XP_476903; OsWNK4 – BAD27820; OsWNK5 - XP_478910; OsWNK6 – AAX95458; OSWNK7 – AK072172; HsWNK1 – Q9H4A3; HsWNK2 – CAI12344; HsWNK3 – CAI43129; HsWNK4 – AAK91995; DmWNK – AAF51744; EcWNK – CAD26582; PbWNK – ZA6636; PsWNK, GiWNK – XM_766287; AtMKK7 – At1g18350, NM_101693.
934
3. Results 3.1. Domain structure of WNK kinases A GAL4-based yeast two-hybrid screen of an Arabidopsis cDNA library with AtVHA-C as the bait protein was performed to identify potential regulators of the V-ATPase. Among the positive clones we identified five independent inserts encoding fragments of the C-terminal domain of AtWNK8. Four unrelated bait proteins were tested and did not show interaction with AtWNK8 thus excluding the possibility of a false positive result (data not shown). In order to evaluate the specificity of this interaction, we searched databases for related proteins of AtWNK8. These searches revealed that the genome of the dicotyledenous model plant Arabidopsis encodes 11 kinases (AtWNK1-11) whereas the monocotyledonous model rice encodes only seven kinases (OsWNK1-7) with the unique lysine placement characteristic of the WNK family. Compared to mammalian WNKs, plant WNKs are much smaller proteins with a predicted molecular weight of about 60–70 kDa. With the exception of AtWNK11 and OsWNK5, which consist only of the highly conserved Nterminal kinase domain, all other plant WNKs have a highly divergent C-terminal domain of about 300 amino acids. The only region of conserved sequence found within this domain overlaps with the autoinhibitory domain identified in mammalian WNK1 [21] (Fig. 1A). Only a subset of plant WNKs (AtWNK3, 6, 7, 8, 10; OsWNK2, 4) contains a predicted short coiled-coil domain at their extreme C-terminal end, which shows weak similarity to the C-terminal coiled-coil domain of mammalian WNKs. Outside this region, similarity
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
can only be detected between family members that arose from recent duplications (AtWNK6 and 7, 1 and 9, 8 and 10). The coiled-coil domain predicted adjacent to the autoinhibitory segment of mammalian WNKs, as well as the S- and P-rich regions and the SH3-domain binding motifs [32] are absent from plant WNKs. Size and domain structure of plant WNKs are similar to PkpA [33], the first member of the WNK family identified in the filamentous fungus Phycomyces blakesleeanus. Interestingly, the genomes of the unicellular eukaryotes Chlamydomonas reinhardtii, Encephalitozoon cuniculi, Phytophtora sojae and Giardia lamblia do encode WNK kinases. Their size is similar to plant WNKs and the consensus sequence of the potential autoinhibitory segment is conserved (Fig. 1B) in some of the WNKs from unicellular organisms. Phylogenetic analysis of kinase domain sequences from all kingdoms (Fig. 1C) indicates that the WNKs represent an ancient clade of highly conserved protein kinases that have evolved considerable variability outside their catalytic domains. 3.2. Specific interaction between AtWNK8 and AtVHA-C To determine if the interaction was specific for AtWNK8, the C-terminal domains of the other Arabidopsis WNK family members were cloned into a GAL4-AD vector and tested for interaction with AtVHA-C (Fig. 2, data not shown for AtWNK6). Neither AtWNK10, the protein most closely related to AtWNK8, nor any other family member was found to interact with AtVHA-C in the yeast two-hybrid system, indicating that the interaction between the two proteins is highly specific.
Fig. 1. WNKs represent an ancient family of protein kinases. (A) Schematic illustration comparing the domain structure of AtWNK8 and mammalian WNK1. Shown are the kinase domain (KD), the autoinhibitory-domain (AID), predicted coiled-coil domains (Coil) as well as proline(P-) and serine- (S-)-rich regions. (B) Sequence alignment of the autoinhibitory domain found in WNKs from diverse eukaryotes. (C) Distance based tree of the WNK kinase domain sequences, a full heuristic bootstrap analysis (1000 replicates) was performed, numbers indicate the percentage of bootstrap support for the respective node. (At: Arabidopsis thaliana, Ce: Caenorhabditis elegans, Cr: Chlamydomonas reinhardtii, Dm: Drosophila melanogaster, Ec: Encephalitozoon cuniculi, Gl: Giardia lamblia, Hs: Homo sapiens, Os: Oryza sativa, Pb: Phycomyces blakesleeanus, Ps: Phytophtora sojae, see Section 2 for GenBank Accession Nos).
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
Fig. 2. Specific interaction of AtWNK8 and AtVHA-C. Yeast twohybrid analysis of AtVHA-C (bait) with AtWNK C-termini (prey). Yeast cells (strain AH109) were cotransformed with prey- and baitconstructs and selected for transformation on SD–Trp–Leu medium. Interaction of bait and prey proteins was shown by growth on SD– Trp–Leu–His–Ade.
3.3. The C-terminal domain of AtWNK8 binds AtVHA-C To substantiate the yeast two-hybrid analysis and to determine the interaction domain of AtWNK8, we analyzed the interaction with AtVHA-C in Far-Western experiments. Full length, as well as N- and C-terminal deletion constructs of AtWNK8 fused to GST, was expressed in E. coli (Fig. 3A). Because wildtype AtWNK8 could not be expressed as a GSTfusion protein, all the constructs used for the interaction studies carried an amino acid exchange (D157A, see below for details). After native purification (Fig. 3B), the proteins were slot blotted on to a membrane, which was then incubated with purified AtVHA-C and subsequently probed with an antibody against AtVHA-C. In comparison to GST, which served as a negative control, AtVHA-C was found to bind to full length AtWNK8 (Fig. 3C). Strong binding was also detected with both DN1 and DN2, which consists of the most C-terminal 110 amino acids and includes the predicted coiled-coil domain. Weaker signals that were not clearly different from the background of GST, were detected for the C-terminal deletions. These results confirmed the interaction detected in the yeast two-hybrid system and indicated that the C-terminal coiled-coil domain is sufficient for binding of AtWNK8 to AtVHA-C. 3.4. AtWNK8 is an active kinase and shows transautophosphorylation To test for kinase activity, AtWNK8 was expressed in E. coli. Whereas GST-tagged wildtype AtWNK8 could not be obtained, expression as an intein-fusion protein allowed purification of the native kinase and the two mutants K41M and D157A. K41 corresponds to the essential lysine residue conserved in all WNKs, whereas D157 represents the essential aspartate found in the active center of all protein kinases and both mutants are therefore presumed to be kinase-inactive. In contrast to the purified mutant proteins, the wildtype protein ran at a higher molecular weight (Fig. 4A). Incubation with [c-32P]ATP resulted in phosphorylation of wildtype AtWNK8,
935
Fig. 3. Binding is mediated by the C-terminal domain of AtWNK8. Farwestern analysis of N- and C-terminal deleted AtWNK8 proteins with AtVHA-C. (A) Schematic representation of AtWNK8 full length (FL) and truncated proteins (DN1, DN2, DC1, DC2). Numbers indicate positions of deletion. (B) Coomassie stained gel. (C) Overlay with AtVHA-C. AtWNK8 full length and deletion proteins (B) were slot blotted onto a nitrocellulose membrane in equal molarities (35 and 7 pmol, respectively) and incubated with AtVHA-C. Binding was detected using an anti-VHA-C antibody.
whereas the two mutants were inactive as expected (Fig. 4A), thus demonstrating that AtWNK8 is an active protein kinase that displays autophosphorylation. The GST-fusion proteins used in the Far-Western analysis carried either the D157A mutation, or a deletion of the entire kinase domain and should thus not have kinase activity. We therefore used them as substrates in kinase assays to determine if autophosphorylation of AtWNK8 can take place intermolecularly (Fig. 4B). Except for DC2, all the kinase-inactive GST-fusions were substrates for phosphorylation by wildtype AtWNK8, indicating that autophosphorylation can take place intermolecularly and is not restricted to the kinase domain of
Fig. 4. AtWNK8 is an active kinase and shows trans-autophosphorylation. (A) Comparison of wildtype AtWNK8 and mutants K41M and D157A. Coomassie stained gel and autoradiograph are shown. (B) Phosphorylation of GST-tagged full length and truncated inactive AtWNK8 by wildtype AtWNK8. The proteins used were the same as shown in Fig. 3.
936
AtWNK8. In contrast to DC1, DC2 includes the conserved domain that has been shown to be autoinhibitory in mammalian WNKs. The lack of transphosphorylation of DC2 along with the reduced autophosphorylation signal of full length AtWNK8 indicates that the autoinhibitory function of this domain is conserved in plants. 3.5. AtWNK8 phosphorylates AtVHA-C Having shown that AtWNK8 and AtVHA-C physically interact and that AtWNK8 is an active protein kinase, it remained to be determined if and where AtVHA-C is phosphorylated. Recombinant full length, as well as N- and C-terminally truncated AtVHA-C (134-267), was used as substrate in a kinase assay and was shown to be phosphorylated by AtWNK8 (Fig. 5A). Tryptic digestion of phosphorylated AtVHA-C followed by MALDI-TOF analysis allowed the identification of six phosphorylated peptides (Table 1, Fig. 5B). All peptides contain two or more serine or threonine residues so that individual phosphorylated residues could not be determined. We therefore used a model of AtVHA-C, based on the recently published crystal structure of its yeast homolog VMA5 [12], to assess the likelihood of individual residues being phosphorylated by their predicted surface positions. Combined with in silico phosphorylation site predictions [34], our results suggest that residues S116, T204, S211,
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
S212, S253, and S331 (Fig. 5B) are the most likely sites for phosphorylation of AtVHA-C by AtWNK8. Among the V-ATPase subunits, subunit C is one of the least conserved showing only about 30% sequence identity between yeast and Arabidopsis. Of the potential AtWNK8 target sites, only S211 is conserved in all eukaryotes, whereas all others are only found in a subset of sequences. It was therefore of interest to see if the subunits C from other species are also recognized by AtWNK8. Surprisingly, the AtVHA-C homologs from S. cerevisiae (Vma5p), M. sexta and C. elegans were all phosphorylated by AtWNK8 (Fig. 6A), however binding of AtWNK8 DN2 was only detected for AtVHA-C and weakly for subunit C from C. elegans but not for S. cerevisiae and M. sexta in Far-Western experiments (Fig. 6B). 3.6. AtWNK8 phosphorylation of hybrid V1-complexes We tested whether AtVHA-C could also be phosphorylated when it is integrated into the V1-complex. As it is difficult to isolate sufficient amounts of intact complexes from Arabidopsis, we made use of the fact that V1-complexes lacking subunit C (V1(-C)), can be isolated from the midgut of Manduca larvae [11]. After incubation with AtVHA-C overnight, the protein complex eluted earlier from the size exclusion column compared with the V1(-C) ATPase – indicating that AtVHA-C was indeed reconstituted into the Manduca V1-complex (data not shown). Aliquots of the gel filtration fractions corresponding to V1(-C), V1(+AtVHA-C) and unbound C were used as substrates in phosphorylation assays with and without AtWNK8 and were subsequently separated by SDS–PAGE (Fig. 7A). Autoradiography revealed that not only AtVHAC, but also the Manduca subunits A, G and either B or H, which were not separated after SDS–PAGE due to their similar molecular weights, were phosphorylated (Fig. 7B). The control reactions, in which no kinase was added, showed clearly that the detected phosphorylation signals are not due to autophosphorylation or ATP-binding but are dependent on the presence of AtWNK8.
4. Discussion
Fig. 5. AtWNK8 phosphorylates AtVHA-C. (A) Full length AtVHAC and AtVHA-C 134-267 were used as substrates for phosphorylation by AtWNK8. (B) MALDI-TOF analysis revealed four peptides being phosphorylated by AtWNK8, which are highlighted in the model (compare Table 1). Serine and threonine residues with a high likelihood of phosphorylation (NetPhosServer 2.0) are enlarged in the peptide sequence and appear as spheres in the model. Numbers correspond to their sequence position.
We show here that a member of the Arabidopsis WNK family binds to and phosphorylates the V-ATPase subunit C. WNK kinases have received considerable attention since human WNKs were genetically linked to the regulation of blood pressure [23]. Our attempt to address the specificity of the interaction of AtWNK8 with AtVHA-C has led us to re-examine the available WNK sequences. Arabidopsis has been reported to have the largest WNK family [28]; our analysis confirms this notion and extends the Arabidopsis family to 11 members, whereas the larger rice genome encodes only seven members. Phylogenetic analysis shows that much of the diversification of plant WNKs has taken place before the divergence of mono- and dicotyledenous plants. Based on the fact that the yeast genome does not encode WNKs, it has been proposed that they are restricted to multicellular organisms [22]. However, database searches showed that WNKs are not only found in multicellular but also in unicellular eukaryotes including very distantly related groups like oomycetes and diplomonads. WNKs therefore represent an ancient group of protein kinases and it seems possible that the characteristic placement of the catalytic lysine residue in subdomain I found in all WNKs is
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
937
Table 1 MALDI mass spectrometry analysis of unphosphorylated and phosphorylated peptides from AtVHA-C Start residue
End residue
Measured massa
Measured massb
Sequence
17 112 195 214 242 329
4 131 213 231 261 347
2314.863 2407.307 2383.057 2284.949 2491.273 2379.260
2394.832 2487.281 2463.027 2364.919 2571.243 2459.227
GSHMASMTGGQQMGRGSMTSR YPTMSPLKEVVDNIQSQVAK DWLACYETLTDYVVPRSSK KLFEDNEYALYTVTLFTR EKGFQVRDFEQSVEAQETRK KVRSILERLCDSTNSLYWK
a
Mass of unphosphorylated peptide. Mass of phosphorylated peptide.
b
Fig. 6. AtWNK8 phosphorylates but does not bind subunit C from other eukaryotes. (A) The C subunits of Saccharomyces cerevisiae (Sc), M. sexta (Ms), Caenorhabditis elegans (Ce), and Arabidopsis thaliana (At) were used as substrates for phopshorylation by AtWNK8. (B) The proteins used as phosphorylation substrates in (A) were immobilized on a nitrocellulose membrane (each 30 pmol) and tested for interaction with AtWNK8 DN2. Interaction was detected using an anti-GSTantibody.
Fig. 7. AtWNK8 phosphorylation of hybrid V1-complexes. Pooled gel filtration fractions representing of V1-C (lanes 1, 2), V1 (AtVHA-C) (lanes3, 4), and unbound VHA-C (lanes 5, 6) were used as substrates in phosphorylation assays with AtWNK8 (2, 4, 6) and in controls without kinase (1, 3, 5). (A) Coomassie stained gel, (B) autoradiograph. * marks autophosphorylation of AtWNK8.
actually older than the placement in subdomain II found in all other protein kinases. Outside the highly conserved kinase domain, WNKs differ considerably. The only additional region with identifiable sequence similarity found in all plant and in some unicellular WNKs overlaps with the autoinhibitory segment described in mammalian WNK1 [21]. We have shown that AtWNK8 is an active kinase and its activity is reduced in the presence of DC2, a kinase-inactive deletion that contains the putative autoinhibitory domain. Although this domain is also present in DN2, this deletion protein does not inhibit AtWNK8 autophosphorylation activity. This is reminiscent of results obtained with mammalian WNK1, where the autoinhibitory activity was found to be reduced when the C-terminal flanking coiled-coil domain was present [21]. Our results indicate that regulation via an autoinhibitory sequence is a conserved feature of plant and mammalian WNKs and is possibly also found in some of the WNKs from unicellular organisms. Interestingly, both rice and Arabidopsis contain one WNK gene that encodes only the kinase domain and it will be interesting to determine the activities and potential regulators of these unusual plant WNKs. Although mammalian WNKs have been shown to exist as tetramers, intermolecular phosphorylation has not been detected [21]. Our results show clearly that autophosphorylation of AtWNK8 takes place in an intermolecular fashion, indicating that plant WNKs may also be able to multimerize. As the coiled-coil domain, which has been proposed to be responsible for the multimerization of WNK1 [35] is not found in the plant proteins, it remains to be determined which motifs are necessary for the interaction of AtWNK8 with itself. The presence of multiple binding sites for SH3-domains suggests that the C-terminal domain of mammalian WNKs is able to interact with multiple partners [36], however so far no interactions between this domain and other proteins have been described. A small C-terminal fragment of AtWNK8 including a potential short coiled-coil domain is sufficient for binding of AtVHA-C, demonstrating that the C-terminal domain does indeed mediate protein–protein interactions. The predicted coiled-coil domain in AtWNK8 is comparably short and it remains to be determined if it really adopts a coil structure. The C-terminal end of AtWNK10, the closest relative of AtWNK8, shows about 50% amino acid identity and is also predicted to form a coiled-coil domain and yet it does not interact with AtVHA-C in the yeast two-hybrid system. The specificity of this interaction is accentuated by the fact that the binding domain of AtWNK8 is able to discriminate between subunits C from different eukaryotes, although they are functionally interchangeable [37–39] and therefore most likely have a very similar structure.
938
Protein phosphorylation is a ubiquitous mechanism that has been shown to regulate the activity of a large number of transport proteins including ion channels and proton pumps. Although pharmacological evidence suggests that V-ATPases are targets for protein kinase-mediated signaling [40,41], evidence for a direct interaction and phosphorylation of V-ATPase subunits has so far been missing. We have shown that AtWNK8 phosphorylates AtVHA-C and have identified four phosphorylated peptides each of which includes two or more serine or threonine residues, indicating that phosphorylation by AtWNK8 is not restricted to a particular domain of subunit C. Phosphorylation target sides of WNKs are not well characterized but it has been suggested that the P + 1 residue is not a strong specificity determinant [32]. Although we could not determine individual phosphorylated residues, the likely sites found in the different peptides indicate that the P 1 and P + 1 positions are not conserved. The phosphorylated peptides are either derived from the globular head domain, that has been proposed to bind to the catalytic domain of V1, or from the foot region assumed to be oriented to the membrane [11,12]. Site-directed mutagenesis revealed that the C-terminal foot region of subunit C might be important for stable assembly of V1 and V0 [13]. Furthermore, the nucleotide-binding property of subunit C has been localized to the C-terminal region of subunit C and the structural changes induced by nucleotide-binding might be involved in the regulation of the reversible V1V0 disassembly [15]. In vivo, phosphorylation could either be targeted to free subunit C molecules or intact V-ATPase complexes. Addition of recombinant yeast subunit C to the V1(-C) complex of Manduca significantly increased ATPase activity [11], supporting its functional role in regulation of ATPase activity [42,43]. We found that AtVHA-C can also be reconstituted into the V1-complex from Manduca and were able to show that it can be phosphorylated by AtWNK8 when present in V1. Surprisingly, several other V1-subunits are also substrates for AtWNK8 and it will be of interest to determine if they can interact with AtWNK8 and are phosphorylated in vivo. Similar to V-ATPase subunit C, subunit H is not involved in catalysis but was implicated in regulating the V-ATPase [44,45]. Phosphorylation of VHA-H might therefore be important for this regulatory function. The V-ATPase subunit B seems to be a likely target for a protein kinase, since it was described before as phosphorylated protein [19]. If the phosphorylation of other V1-subunits by AtWNK8 requires tight binding to AtWNK8 and takes place in vivo needs to be determined. Subunit C-homologues from diverse species are all phosphorylated by AtWNK8, although efficient binding is restricted to the Arabidopsis subunit. Taken together, these results indicate that the C-terminal interaction domain of AtWNK8 may represent a direct docking module that provides specificity and might facilitate efficient phosphorylation of subunit C under physiological conditions. Our results show for the first time, that the V-ATPase can be phosphorylated by a protein kinase. Although both AtVHA-C and AtWNK8 are ubiquitously expressed (data not shown), it remains to be determined if the two proteins colocalize and if phosphorylation of the V-ATPase takes place in vivo. Our results provide the framework necessary to address this question and to integrate V-ATPase regulation into known signaling networks.
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939 Acknowledgments: We are grateful to Joanne Chory for generous support in the initial phase of this work. We thank Claudia Oecking for critical reading of the manuscript and acknowledge the excellent technical assistance provided by Zhao-Xin Wang and Andrea Armbru¨ster. This work was supported by the Deutsche Forschungsgemeinschaft (SFB446 TP A20 to K.S.) and the School of Biological Sciences, NTU, Singapore.
References [1] Nelson, N. (2003) A journey from mammals to yeast with vacuolar H+-ATPase (V-ATPase). J. Bioenerg. Biomembr. 35, 281–289. [2] Nishi, T. and Forgac, M. (2002) The vacuolar (H+)-ATPases – nature’s most versatile proton pumps. Nat. Rev. Mol. Cell Biol. 3, 94–103. [3] Kane, P.M. and Smardon, A.M. (2003) Assembly and regulation of the yeast vacuolar H+-ATPase. J. Bioenerg. Biomembr. 35, 313–321. [4] Kane, P.M. (1995) Disassembly and reassembly of the yeast vacuolar H(+)-ATPase in vivo. J. Biol. Chem. 270, 17025–17032. [5] Sumner, J.P., Dow, J.A., Earley, F.G., Klein, U., Ja¨ger, D. and Wieczorek, H. (1995) Regulation of plasma membrane V-ATPase activity by dissociation of peripheral subunits. J. Biol. Chem. 270, 5649–5653. [6] Kane, P.M. and Parra, K.J. (2000) Assembly and regulation of the yeast vacuolar H(+)-ATPase. J. Exp. Biol. 203, 81–87. [7] Parra, K.J. and Kane, P.M. (1998) Reversible association between the V1 and V0 domains of yeast vacuolar H+-ATPase is an unconventional glucose-induced effect. Mol. Cell Biol. 18, 7064– 7074. [8] Ho, M.N., Hirata, R., Umemoto, N., Ohya, Y., Takatsuki, A., Stevens, T.H. and Anraku, Y. (1993) VMA13 encodes a 54-kDa vacuolar H(+)-ATPase subunit required for activity but not assembly of the enzyme complex in Saccharomyces cerevisiae. J. Biol. Chem. 268, 18286–18292. [9] Doherty, R.D. and Kane, P.M. (1993) Partial assembly of the yeast vacuolar H(+)-ATPase in mutants lacking one subunit of the enzyme. J. Biol. Chem. 268, 16845–16851. [10] Ho, M.N., Hill, K.J., Lindorfer, M.A. and Stevens, T.H. (1993) Isolation of vacuolar membrane H(+)-ATPase-deficient yeast mutants; the VMA5 and VMA4 genes are essential for assembly and activity of the vacuolar H(+)-ATPase. J. Biol. Chem. 268, 221–227. [11] Armbru¨ster, A., Svergun, D.I., Coskun, U., Juliano, S., Bailer, S.M. and Gru¨ber, G. (2004) Structural analysis of the stalk subunit Vma5p of the yeast V-ATPase in solution. FEBS Lett. 570, 119–125. [12] Drory, O., Frolow, F. and Nelson, N. (2004) Crystal structure of yeast V-ATPase subunit C reveals its stator function. EMBO Rep. 5, 1148–1152. [13] Curtis, K.K., Francis, S.A., Oluwatosin, Y. and Kane, P.M. (2002) Mutational analysis of the subunit C (Vma5p) of the yeast vacuolar H+-ATPase. J. Biol. Chem. 277, 8979–8988. [14] Vitavska, O., Merzendorfer, H. and Wieczorek, H. (2005) The VATPase subunit C binds to polymeric F-actin as well as to monomeric G-actin and induces cross-linking of actin filaments. J. Biol. Chem. 280, 1070–1076. [15] Armbru¨ster, A., Hohn, C., Hermesdorf, A., Schumacher, K., Borsch, M. and Gru¨ber, G. (2005) Evidence for major structural changes in subunit C of the vacuolar ATPase due to nucleotide binding. FEBS Lett. 579, 1961–1967. [16] Seol, J.H., Shevchenko, A. and Deshaies, R.J. (2001) Skp1 forms multiple protein complexes, including RAVE, a regulator of VATPase assembly. Nat. Cell Biol. 3, 384–391. [17] Smardon, A.M., Tarsio, M. and Kane, P.M. (2002) The RAVE complex is essential for stable assembly of the yeast V-ATPase. J. Biol. Chem. 277, 13831–13839. [18] Johnson, S.A. and Hunter, T. (2005) Kinomics: methods for deciphering the kinome. Nat. Methods 2, 17–25. [19] Myers, M. and Forgac, M. (1993) The coated vesicle vacuolar (H+)-ATPase associates with and is phosphorylated by the 50kDa polypeptide of the clathrin assembly protein AP-2. J. Biol. Chem. 268, 9184–9186.
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939 [20] McCubbin, A.G., Ritchie, S.M., Swanson, S.J. and Gilroy, S. (2004) The calcium-dependent protein kinase HvCDPK1 mediates the gibberellic acid response of the barley aleurone through regulation of vacuolar function. Plant J. 39, 206–218. [21] Xu, B.E., Min, X., Stippec, S., Lee, B.H., Goldsmith, E.J. and Cobb, M.H. (2002) Regulation of WNK1 by an autoinhibitory domain and autophosphorylation. J. Biol. Chem. 277, 48456– 48462. [22] Xu, B., English, J.M., Wilsbacher, J.L., Stippec, S., Goldsmith, E.J. and Cobb, M.H. (2000) WNK1, a novel mammalian serine/ threonine protein kinase lacking the catalytic lysine in subdomain II. J. Biol. Chem. 275, 16795–16801. [23] Wilson, F.H. et al. (2001) Human hypertension caused by mutations in WNK kinases. Science 293, 1107–1112. [24] Verissimo, F. and Jordan, P. (2001) WNK kinases, a novel protein kinase subfamily in multi-cellular organisms. Oncogene 20, 5562–5569. [25] Gamba, G. (2005) Role of WNK kinases in regulating tubular salt and potassium transport and in the development of hypertension. Am. J. Physiol. Renal Physiol. 288, F245–F252. [26] Wilson, F.H. et al. (2003) Molecular pathogenesis of inherited hypertension with hyperkalemia: the Na–Cl cotransporter is inhibited by wild-type but not mutant WNK4. Proc. Natl. Acad. Sci. USA 100, 680–684. [27] Lee, B.H. et al. (2004) WNK1 phosphorylates synaptotagmin 2 and modulates its membrane binding. Mol. Cell 15, 741–751. [28] Nakamichi, N., Murakami-Kojima, M., Sato, E., Kishi, Y., Yamashino, T. and Mizuno, T. (2002) Compilation and characterization of a novel WNK family of protein kinases in Arabiodpsis thaliana with reference to circadian rhythms. Biosci. Biotechnol. Biochem. 66, 2429–2436. [29] Murakami-Kojima, M., Nakamichi, N., Yamashino, T. and Mizuno, T. (2002) The APRR3 component of the clock-associated APRR1/TOC1 quintet is phosphorylated by a novel protein kinase belonging to the WNK family, the gene for which is also transcribed rhythmically in Arabidopsis thaliana. Plant Cell Physiol. 43, 675–683. [30] Roos, M., Soskic, V., Poznanovic, S. and Godovac-Zimmermann, J. (1998) Post-translational modifications of endothelin receptor B from bovine lungs analyzed by mass spectrometry. J. Biol. Chem. 273, 924–931. [31] Wilkins, M.R., Lindskog, I., Gasteiger, E., Bairoch, A., Sanchez, J.C., Hochstrasser, D.F. and Appel, R.D. (1997) Detailed peptide characterization using PEPTIDEMASS – a World-Wide-Webaccessible tool. Electrophoresis 18, 403–408. [32] Min, X., Lee, B.H., Cobb, M.H. and Goldsmith, E.J. (2004) Crystal structure of the kinase domain of WNK1, a kinase that causes a hereditary form of hypertension. Structure (Camb.) 12, 1303–1311.
939 [33] Ruiz-Perez, V.L., Murillo, F.J. and Torres-Martinez, S. (1995) PkpA, a novel Phycomyces blakesleeanus serine/threonine protein kinase. Curr. Genet. 28, 309–316. [34] Blom, N., Sicheritz-Ponten, T., Gupta, R., Gammeltoft, S. and Brunak, S. (2004) Prediction of post-translational glycosylation and phosphorylation of proteins from the amino acid sequence. Proteomics 4, 1633–1649. [35] Lenertz, L.Y., Lee, B.H., Min, X., Xu, B.E., Wedin, K., Earnest, S., Goldsmith, E.J. and Cobb, M.H. (2005) Properties of WNK1 and implications for other family members. J. Biol. Chem. 280, 26653–26658. [36] Xu, B.E., Lee, B.H., Min, X., Lenertz, L., Heise, C.J., Stippec, S., Goldsmith, E.J. and Cobb, M.H. (2005) WNK1: analysis of protein kinase structure, downstream targets, and potential roles in hypertension. Cell Res. 15, 6–10. [37] Beltran, C., Kopecky, J., Pan, Y.C., Nelson, H. and Nelson, N. (1992) Cloning and mutational analysis of the gene encoding subunit C of yeast vacuolar H(+)-ATPase. J. Biol. Chem. 267, 774–779. [38] Oka, T., Yamamoto, R. and Futai, M. (1997) Three vha genes encode proteolipids of Caenorhabditis elegans vacuolar-type ATPase. Gene structures and preferential expression in an Hshaped excretory cell and rectal cells. J. Biol. Chem. 272, 24387– 24392. [39] Schumacher, K., Vafeados, D., McCarthy, M., Sze, H., Wilkins, T. and Chory, J. (1999) The Arabidopsis det3 mutant reveals a central role for the vacuolar H(+)-ATPase in plant growth and development. Genes Dev. 13, 3259–3270. [40] Carini, R., Grazia De Cesaris, M., Splendore, R. and Albano, E. (2001) Stimulation of p38 MAP kinase reduces acidosis and Na(+) overload in preconditioned hepatocytes. FEBS Lett. 491, 180–183. [41] Nordstrom, T., Grinstein, S., Brisseau, G.F., Manolson, M.F. and Rotstein, O.D. (1994) Protein kinase C activation accelerates proton extrusion by vacuolar-type H(+)-ATPases in murine peritoneal macrophages. FEBS Lett. 350, 82–86. [42] Sun-Wada, G.H., Murata, Y., Namba, M., Yamamoto, A., Wada, Y. and Futai, M. (2003) Mouse proton pump ATPase C subunit isoforms (C2-a and C2-b) specifically expressed in kidney and lung. J. Biol. Chem. 278, 44843–44851. [43] Peng, S.B., Zhang, Y., Tsai, S.J., Xie, X.S. and Stone, D.K. (1994) Reconstitution of recombinant 33-kDa subunit of the clathrincoated vesicle H(+)-ATPase. J. Biol. Chem. 269, 11356–11360. [44] Domgall, I., Venzke, D., Luttge, U., Ratajczak, R. and Bottcher, B. (2002) Three-dimensional map of a plant V-ATPase based on electron microscopy. J. Biol. Chem. 277, 13115–13121. [45] Keenan Curtis, K. and Kane, P.M. (2002) Novel vacuolar H+ATPase complexes resulting from overproduction of Vma5p and Vma13p. J. Biol. Chem. 277, 2716–2724.
A WNK kinase binds and phosphorylates V-ATPase subunit C Anne Hong-Hermesdorfa, Angela Bru¨xa, Ardina Gru¨berb, Gerhard Gru¨berb, Karin Schumachera,* b
a Universita¨t Tu¨bingen, ZMBP-Plant Physiology, Auf der Morgenstelle 1, 72076 Tu¨bingen, Germany Nanyang Technological University, School of Biological Sciences, 60 Nanyang Drive, Singapore 637551, Singapore
Received 11 November 2005; revised 14 December 2005; accepted 4 January 2006 Available online 18 January 2006 Edited by Michael R. Sussman
Abstract WNK (with no lysine (K)) protein kinases are found in many eukaryotes and share a unique active site. Here, we report that a member of the Arabidopsis WNK family (AtWNK8) interacts with subunit C of the vacuolar H+-ATPase (V-ATPase) via a short C-terminal domain. AtWNK8 is shown to autophosphorylate intermolecularly and to phosphorylate Arabidopsis subunit C (AtVHA-C) at multiple sites as determined by MALDI-TOF MS analysis. Furthermore, we show that AtVHA-C and other V-ATPase subunits are phosphorylated when V1-complexes are used as substrates for AtWNK8. Taken together, our results provide evidence that V-ATPases are potential targets of WNK kinases and their associated signaling pathways. 2006 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Vacuolar H+-ATPase; V1-complex; WNK; Phosphorylation; Arabidopsis
1. Introduction pH is an essential parameter for most biological processes and vacuolar H+-ATPases (V-ATPases) create and maintain the pH-differentials suited to the biochemical functions of the endomembrane compartments of all eukaryotic cells. V-ATPases are multi-subunit enzyme complexes composed of the cytosolic ATP hydrolyzing V1 subcomplex (subunits A–H) and the membrane bound proton translocating V0 subcomplex (subunits a, c, c 0 , c00 , d and e) [1,2]. ATP-dependent proton transport energizes secondary active transport and controls cellular pH homeostasis and it is therefore not surprising that a plethora of processes including receptor-mediated endocytosis, protein processing, virus entry and bone resorption has been shown to depend on V-ATPase activity. Moreover, it has recently been shown that the V0-complex is directly involved in membrane fusion events. Despite the remarkable diversity of functions that V-ATPases serve, knowledge of their regulation and in particular their integration into cellular signaling networks is limited. V-ATPases exist in a dynamic equilibrium between fully assembled complexes and reversibly disassembled V1 and V0 * Corresponding author. Fax: +49 7071 293287. E-mail address: [email protected] (K. Schumacher).
Abbreviations: WNK, with no lysine (K); V-ATPase, vacuolar H+ATPase; VHA-C, vacuolar H+-ATPase subunit C; MALDI-TOF MS, matrix assisted laser desorption-ionisation/time-of-flight mass spectrometry
subcomplexes [3–5]. Depending on the energy status of the cell, this equilibrium can be rapidly shifted [3]. Since neither the V1 nor the V0 subcomplex is fully active on its own [6], reversible disassembly provides a means to save energy in times of starvation. Moreover, the reversibility of the process enables the cell to quickly adapt to energy re-supply. Conventional glucose signal transduction is not involved in this process [7] and the molecular features of this mechanism remain to be elucidated. Subunit C occurs in substoichiometric amounts in free V1 complexes and in a subunit C-deficient yeast strain V0 and V1 subcomplexes are present, but do not assemble and are both inactive [8–10]. Therefore, it seems likely that subunit C plays a key role in reversible assembly. Structural analysis of subunit C revealed an elongated, boot like shape [11,12] with three distinct domains: a globular head domain, an elongated ‘neck’, and a globular foot domain. The head domain is thought to connect to the V1 complex, while the foot and lower neck domains bind to the V0 complex [11,12]. The structure of subunit C fits its proposed function in coupling the activities of V1 and V0 subcomplexes [3,13]. Moreover, it has recently been shown that the subunit C from Manduca sexta interacts with actin filaments [14] and that subunits C from Arabidopsis and yeast undergo major structural changes after binding of adenine nucleotides [15]. Both properties are ideally suited for a protein controlling the assembly state of the V-ATPase complex, but it remains to be determined if either actin or nucleotide binding are indeed necessary for this process. The regulator of the H+-ATPase of vacuolar and endosomal membranes complex (regulator of the H+-ATPase of vacuolar and endosomal membranes), consisting of Skp1p, Rav1p, and Rav2p, was shown to enhance V-ATPase assembly through interaction with the V1 moiety in yeast and it is therefore conceivable that interactions of subunit C with other proteins are necessary for complex formation [16,17]. Protein phosphorylation is ubiquitous and the most widespread and best studied signaling mechanism in eukaryotic cells [18]. Surprisingly, phosphorylation of V-ATPase subunits by protein kinases has not been demonstrated so far, although it has been reported that AP50, a subunit of the AP2-complex of bovine clathrin-coated vesicles, phosphorylates the V-ATPase subunit B in vitro [19]. In plants, it has recently been shown that a calcium-dependent protein kinase affects V-ATPase activity and is able to phosphorylate vacuolar proteins [20], however it remains to be determined if the V-ATPase itself is phosphorylated. WNK (with no lysine (K)) kinases are a subfamily of serine/ threonine protein kinases related to the STE20/PAK-like family [21]. A lysine residue in kinase subdomain II, which is essential for the coordination of ATP in the active center and
0014-5793/$32.00 2006 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2006.01.018
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
conserved among all other kinases, is missing in this subfamily [22,23]. It was shown to be replaced by a lysine residue in subdomain I [22] that is characteristic for members of the WNK family. WNK kinases have been found in numerous eukaryotes, but are not found in yeast and it has been proposed that they are restricted to multicellular organisms [21,24]. Mammalian WNKs have been shown to regulate ion transport processes in epithelial cells by reducing either insertion or retention of transport proteins in the plasma membrane [25]. This can be mediated either by a direct interaction with transport proteins as shown for WNK4 and NCCT1 [26] or by interaction with trafficking proteins in the case of WNK1. Upon phosphorylation by WNK1 membrane binding properties of synaptotagmin 2 change [27] and this could affect trafficking events controlling the membrane presence of transport proteins. Information about the functions of WNK kinases from other organisms is limited. The model plant Arabidopsis has been reported to have nine members of the WNK family [28] but only one of them has so far been characterized. AtWNK1 phosphorylates the putative circadian clock component APRR3 in vitro and might be involved in a signal transduction cascade regulating its biological activity [29]. To gain further insight into the molecular properties of the V-ATPase subunit C of Arabidopsis (AtVHA-C), we identified potential interacting proteins using the yeast two-hybrid system. Here, we show that AtWNK8, a member of the Arabidopsis WNK family, binds to and phosphorylates AtVHA-C.
2. Materials and methods 2.1. Yeast two-hybrid assay The yeast strain AH109 (MATa, trp1-901, leu2-3,112, ura3-52, his3-200, gal4D, gal80D, LYS2: GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3: MEL1UAS-MEL1TATALacZ; Clontech) was used to test interaction between AtVHA-C and the C-terminal sequences of AtWNK1-10. AtVHA-C was PCRamplified and after sequence analysis cloned into pGBKT7 (BamHI/ PstI). AtWNK 1–5, 7–10 C-termini were amplified by PCR using a Arabidopsis Col-0 cDNA library as template, subcloned for sequence analysis and cloned into pGADT7 vector (Cfr9I/ XhoI). Yeast cells were cotransformed according to standard protocols with bait and prey plasmids and transformants were selected on synthetic dropout plates lacking Trp and Leu. To test for interaction, cells were grown on selective synthetic media lacking Trp, Leu, His and Ade. 2.2. Constructs, mutagenesis, proteins pEG-AtWNK8 was used for site-directed mutagenesis at position D157 to alanine (D157A) and the mutation verified by DNA sequencing. AtWNK8 FL (full length), DN1, DN2, DC1, DC2 were amplified by PCR (template pEG-AtWNK8 D157A) and cloned into pGEX4T3 – translationally fused to an N-terminal GST-tag. Restriction sites used for cloning were BamHI and Bsp120I. A C-terminal His-tag was introduced through the reverse primers. Proteins were expressed in and purified from Escherichia coli strain BL21DE3 (B F dcm, ompT, hsdSðr B mB Þ gal k (DE3)) by GSH-affinity chromatography according to standard protocols. Wildtype and mutant (K41M/ D157A) AtWNK8 was expressed using the IMPACTe-CN system (New England Biolabs, Frankfurt, D). The coding sequence (1692 bp) was amplified by PCR using pEG-AtWNK8 or pEG-AtWNK8 D157A as template and cloned into pTYB2 (NdeI/Cfr9I). The mutation K41M was introduced by site-directed mutagenesis, using AtWNK8 in pTYB2 as template. Proteins were expressed in E. coli BL21DE3 cells and purified by chitin affinity chromatography. Purification of subunit C from Arabidopsis (AtVHA-C) and Saccharomyces cerevisiae (Vma5p) was performed as described in [11,15], respectively. The C-subunit of Caenorhabditis elegans (CeVha11) was
933 amplified by PCR and cloned into pET28a (BamHI/HindIII). The plasmid for expression of M. sexta V-ATPase C subunit was kindly provided by Prof. H. Wieczorek (University of Osnabru¨ck). Proteins were expressed and purified from E. coli BL21DE3 cells by Ni2+-affinity chromatography according to standard protocols. The integrity of the proteins was tested by Coomassie staining and their concentration measured by Bradford analysis. 2.3. Far-western analysis Proteins were slot blotted onto a nitrocellulose membrane and incubated in blocking buffer (TBS containing 2% (w/v) protein powder) (1 h), and the putative interaction partner was added for >1 h in the same buffer (2–10 lg/ml). After intensive washing (TBS-T) the interaction was detected by immunostaining of the bound protein using a specific antibody (>1 h at RT). Unspecifically bound antibodies were eliminated through washing steps with TBS-T. Membrane was incubated with HRP conjugated goat-anti-rabbitIgG antibodies (blocking buffer; 1 h at RT). Signals were detected using the SuperSignal West Pico Chemiluminescent (Pierce Perbio, Bonn, D). 2.4. In vitro phosphorylation assay AtWNK8 kinase (0.05 lg) was incubated with 0.3 lg substrate for 30 min at 30 C in optimized kinase assay buffer (25 mM Tris–HCl, pH 7.0, 0.2 mM EDTA, and 5 mM MnCl2). The ATP concentration was 25 lM containing 5 lCi [c-32P]ATP (0.1 lM) per reaction. Reactions were stopped by adding 1 volume 2· SDS sample buffer (4% SDS, 140 mM Tris–HCl, pH 6.8, 20% Glycerol, 0.01% Bromophenol blue, and 10% b-mercaptoethanol) and incubating at 95 C for 5 min. Samples were separated on SDS–PAGE, and after gel drying exposed on an X-ray screen for 1–2 h at RT. 2.5. Tryptic digest and MALDI-TOF analysis Phosphorylated and unphosphorylated AtVHA-C containing bands were cut from the gel and destained overnight with a solution of 50 mM ammonium bicarbonate, 40% ethanol. The protein was digested in gel with trypsin (Promega) according to Roos et al. [30] except that the bands had been washed three times with acetonitrile before drying them in a speed vacuum concentrator. Digested samples were desalted with a C18ZipTip (Millipore) and eluted with CHCA (10 mg/ml a-cyano-4-hydroxycinnamic acid in 50% acetonitrile, 0.1% trifluoroacetic acid) or FA (8 mg/ml 3-methoxy-4hydroxycinnamic acid in 50% acetonitrile, 0.1% trifluoroacetic acid) matrix solution. 1–2 ll of matrix-analyte solution was spotted onto the MALDI plate and allowed to dry. Peptide mass mapping was performed by matrix assisted laser desorption-ionisation/time-offlight mass spectrometry (MALDI-TOF MS) using a Bruker Reflex IIIe MALDI-TOF spectrometer (Bremen, Germany). The peptide map was acquired in reflectron positive-ion mode with delayed extraction at a mass range of 500–6000 Da. The instrument was calibrated using a calibration mixture (Applied Biosystems). For interpretation of the protein fragments the PEPTIDEMASS [31] program available at Expasy web site (www.expasy.ch/tools/peptide-mass. html) was used. 2.6. Accession numbers and data analysis Megalign (Clustal method) was used for sequence alignments; the phylogenetic tree is a result of a full heuristic distance analysis with 1000 bootstraps replicates in Paup4.0b10. For all sequences, GenBank Accession Nos. are supplied; for Arabidopsis sequences additional loci characterization is provided: AtWNK1 – At3g04910, AB084266; AtWNK2 – At3g22420, AB084267; AtWNK3 – At3g48260, AB085616; AtWNK4 – At5g58350, AB084268; AtWNK5 – At3g51630, AB084269; AtWNK6 – At3g18750, AB084270; AtWNK7 – At1g49160, AB085617; AtWNK8 – At5g41990, AB084271; AtWNK9 – At5g28080 NM_122691; AtWNK10 – At1g64630, NM_105138; AtWNK11 – At5g55560, NM_124938; OsWNK1 – XP_478792; OsWNK2 – AAU44135; OsWNK3 – XP_476903; OsWNK4 – BAD27820; OsWNK5 - XP_478910; OsWNK6 – AAX95458; OSWNK7 – AK072172; HsWNK1 – Q9H4A3; HsWNK2 – CAI12344; HsWNK3 – CAI43129; HsWNK4 – AAK91995; DmWNK – AAF51744; EcWNK – CAD26582; PbWNK – ZA6636; PsWNK, GiWNK – XM_766287; AtMKK7 – At1g18350, NM_101693.
934
3. Results 3.1. Domain structure of WNK kinases A GAL4-based yeast two-hybrid screen of an Arabidopsis cDNA library with AtVHA-C as the bait protein was performed to identify potential regulators of the V-ATPase. Among the positive clones we identified five independent inserts encoding fragments of the C-terminal domain of AtWNK8. Four unrelated bait proteins were tested and did not show interaction with AtWNK8 thus excluding the possibility of a false positive result (data not shown). In order to evaluate the specificity of this interaction, we searched databases for related proteins of AtWNK8. These searches revealed that the genome of the dicotyledenous model plant Arabidopsis encodes 11 kinases (AtWNK1-11) whereas the monocotyledonous model rice encodes only seven kinases (OsWNK1-7) with the unique lysine placement characteristic of the WNK family. Compared to mammalian WNKs, plant WNKs are much smaller proteins with a predicted molecular weight of about 60–70 kDa. With the exception of AtWNK11 and OsWNK5, which consist only of the highly conserved Nterminal kinase domain, all other plant WNKs have a highly divergent C-terminal domain of about 300 amino acids. The only region of conserved sequence found within this domain overlaps with the autoinhibitory domain identified in mammalian WNK1 [21] (Fig. 1A). Only a subset of plant WNKs (AtWNK3, 6, 7, 8, 10; OsWNK2, 4) contains a predicted short coiled-coil domain at their extreme C-terminal end, which shows weak similarity to the C-terminal coiled-coil domain of mammalian WNKs. Outside this region, similarity
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
can only be detected between family members that arose from recent duplications (AtWNK6 and 7, 1 and 9, 8 and 10). The coiled-coil domain predicted adjacent to the autoinhibitory segment of mammalian WNKs, as well as the S- and P-rich regions and the SH3-domain binding motifs [32] are absent from plant WNKs. Size and domain structure of plant WNKs are similar to PkpA [33], the first member of the WNK family identified in the filamentous fungus Phycomyces blakesleeanus. Interestingly, the genomes of the unicellular eukaryotes Chlamydomonas reinhardtii, Encephalitozoon cuniculi, Phytophtora sojae and Giardia lamblia do encode WNK kinases. Their size is similar to plant WNKs and the consensus sequence of the potential autoinhibitory segment is conserved (Fig. 1B) in some of the WNKs from unicellular organisms. Phylogenetic analysis of kinase domain sequences from all kingdoms (Fig. 1C) indicates that the WNKs represent an ancient clade of highly conserved protein kinases that have evolved considerable variability outside their catalytic domains. 3.2. Specific interaction between AtWNK8 and AtVHA-C To determine if the interaction was specific for AtWNK8, the C-terminal domains of the other Arabidopsis WNK family members were cloned into a GAL4-AD vector and tested for interaction with AtVHA-C (Fig. 2, data not shown for AtWNK6). Neither AtWNK10, the protein most closely related to AtWNK8, nor any other family member was found to interact with AtVHA-C in the yeast two-hybrid system, indicating that the interaction between the two proteins is highly specific.
Fig. 1. WNKs represent an ancient family of protein kinases. (A) Schematic illustration comparing the domain structure of AtWNK8 and mammalian WNK1. Shown are the kinase domain (KD), the autoinhibitory-domain (AID), predicted coiled-coil domains (Coil) as well as proline(P-) and serine- (S-)-rich regions. (B) Sequence alignment of the autoinhibitory domain found in WNKs from diverse eukaryotes. (C) Distance based tree of the WNK kinase domain sequences, a full heuristic bootstrap analysis (1000 replicates) was performed, numbers indicate the percentage of bootstrap support for the respective node. (At: Arabidopsis thaliana, Ce: Caenorhabditis elegans, Cr: Chlamydomonas reinhardtii, Dm: Drosophila melanogaster, Ec: Encephalitozoon cuniculi, Gl: Giardia lamblia, Hs: Homo sapiens, Os: Oryza sativa, Pb: Phycomyces blakesleeanus, Ps: Phytophtora sojae, see Section 2 for GenBank Accession Nos).
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
Fig. 2. Specific interaction of AtWNK8 and AtVHA-C. Yeast twohybrid analysis of AtVHA-C (bait) with AtWNK C-termini (prey). Yeast cells (strain AH109) were cotransformed with prey- and baitconstructs and selected for transformation on SD–Trp–Leu medium. Interaction of bait and prey proteins was shown by growth on SD– Trp–Leu–His–Ade.
3.3. The C-terminal domain of AtWNK8 binds AtVHA-C To substantiate the yeast two-hybrid analysis and to determine the interaction domain of AtWNK8, we analyzed the interaction with AtVHA-C in Far-Western experiments. Full length, as well as N- and C-terminal deletion constructs of AtWNK8 fused to GST, was expressed in E. coli (Fig. 3A). Because wildtype AtWNK8 could not be expressed as a GSTfusion protein, all the constructs used for the interaction studies carried an amino acid exchange (D157A, see below for details). After native purification (Fig. 3B), the proteins were slot blotted on to a membrane, which was then incubated with purified AtVHA-C and subsequently probed with an antibody against AtVHA-C. In comparison to GST, which served as a negative control, AtVHA-C was found to bind to full length AtWNK8 (Fig. 3C). Strong binding was also detected with both DN1 and DN2, which consists of the most C-terminal 110 amino acids and includes the predicted coiled-coil domain. Weaker signals that were not clearly different from the background of GST, were detected for the C-terminal deletions. These results confirmed the interaction detected in the yeast two-hybrid system and indicated that the C-terminal coiled-coil domain is sufficient for binding of AtWNK8 to AtVHA-C. 3.4. AtWNK8 is an active kinase and shows transautophosphorylation To test for kinase activity, AtWNK8 was expressed in E. coli. Whereas GST-tagged wildtype AtWNK8 could not be obtained, expression as an intein-fusion protein allowed purification of the native kinase and the two mutants K41M and D157A. K41 corresponds to the essential lysine residue conserved in all WNKs, whereas D157 represents the essential aspartate found in the active center of all protein kinases and both mutants are therefore presumed to be kinase-inactive. In contrast to the purified mutant proteins, the wildtype protein ran at a higher molecular weight (Fig. 4A). Incubation with [c-32P]ATP resulted in phosphorylation of wildtype AtWNK8,
935
Fig. 3. Binding is mediated by the C-terminal domain of AtWNK8. Farwestern analysis of N- and C-terminal deleted AtWNK8 proteins with AtVHA-C. (A) Schematic representation of AtWNK8 full length (FL) and truncated proteins (DN1, DN2, DC1, DC2). Numbers indicate positions of deletion. (B) Coomassie stained gel. (C) Overlay with AtVHA-C. AtWNK8 full length and deletion proteins (B) were slot blotted onto a nitrocellulose membrane in equal molarities (35 and 7 pmol, respectively) and incubated with AtVHA-C. Binding was detected using an anti-VHA-C antibody.
whereas the two mutants were inactive as expected (Fig. 4A), thus demonstrating that AtWNK8 is an active protein kinase that displays autophosphorylation. The GST-fusion proteins used in the Far-Western analysis carried either the D157A mutation, or a deletion of the entire kinase domain and should thus not have kinase activity. We therefore used them as substrates in kinase assays to determine if autophosphorylation of AtWNK8 can take place intermolecularly (Fig. 4B). Except for DC2, all the kinase-inactive GST-fusions were substrates for phosphorylation by wildtype AtWNK8, indicating that autophosphorylation can take place intermolecularly and is not restricted to the kinase domain of
Fig. 4. AtWNK8 is an active kinase and shows trans-autophosphorylation. (A) Comparison of wildtype AtWNK8 and mutants K41M and D157A. Coomassie stained gel and autoradiograph are shown. (B) Phosphorylation of GST-tagged full length and truncated inactive AtWNK8 by wildtype AtWNK8. The proteins used were the same as shown in Fig. 3.
936
AtWNK8. In contrast to DC1, DC2 includes the conserved domain that has been shown to be autoinhibitory in mammalian WNKs. The lack of transphosphorylation of DC2 along with the reduced autophosphorylation signal of full length AtWNK8 indicates that the autoinhibitory function of this domain is conserved in plants. 3.5. AtWNK8 phosphorylates AtVHA-C Having shown that AtWNK8 and AtVHA-C physically interact and that AtWNK8 is an active protein kinase, it remained to be determined if and where AtVHA-C is phosphorylated. Recombinant full length, as well as N- and C-terminally truncated AtVHA-C (134-267), was used as substrate in a kinase assay and was shown to be phosphorylated by AtWNK8 (Fig. 5A). Tryptic digestion of phosphorylated AtVHA-C followed by MALDI-TOF analysis allowed the identification of six phosphorylated peptides (Table 1, Fig. 5B). All peptides contain two or more serine or threonine residues so that individual phosphorylated residues could not be determined. We therefore used a model of AtVHA-C, based on the recently published crystal structure of its yeast homolog VMA5 [12], to assess the likelihood of individual residues being phosphorylated by their predicted surface positions. Combined with in silico phosphorylation site predictions [34], our results suggest that residues S116, T204, S211,
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
S212, S253, and S331 (Fig. 5B) are the most likely sites for phosphorylation of AtVHA-C by AtWNK8. Among the V-ATPase subunits, subunit C is one of the least conserved showing only about 30% sequence identity between yeast and Arabidopsis. Of the potential AtWNK8 target sites, only S211 is conserved in all eukaryotes, whereas all others are only found in a subset of sequences. It was therefore of interest to see if the subunits C from other species are also recognized by AtWNK8. Surprisingly, the AtVHA-C homologs from S. cerevisiae (Vma5p), M. sexta and C. elegans were all phosphorylated by AtWNK8 (Fig. 6A), however binding of AtWNK8 DN2 was only detected for AtVHA-C and weakly for subunit C from C. elegans but not for S. cerevisiae and M. sexta in Far-Western experiments (Fig. 6B). 3.6. AtWNK8 phosphorylation of hybrid V1-complexes We tested whether AtVHA-C could also be phosphorylated when it is integrated into the V1-complex. As it is difficult to isolate sufficient amounts of intact complexes from Arabidopsis, we made use of the fact that V1-complexes lacking subunit C (V1(-C)), can be isolated from the midgut of Manduca larvae [11]. After incubation with AtVHA-C overnight, the protein complex eluted earlier from the size exclusion column compared with the V1(-C) ATPase – indicating that AtVHA-C was indeed reconstituted into the Manduca V1-complex (data not shown). Aliquots of the gel filtration fractions corresponding to V1(-C), V1(+AtVHA-C) and unbound C were used as substrates in phosphorylation assays with and without AtWNK8 and were subsequently separated by SDS–PAGE (Fig. 7A). Autoradiography revealed that not only AtVHAC, but also the Manduca subunits A, G and either B or H, which were not separated after SDS–PAGE due to their similar molecular weights, were phosphorylated (Fig. 7B). The control reactions, in which no kinase was added, showed clearly that the detected phosphorylation signals are not due to autophosphorylation or ATP-binding but are dependent on the presence of AtWNK8.
4. Discussion
Fig. 5. AtWNK8 phosphorylates AtVHA-C. (A) Full length AtVHAC and AtVHA-C 134-267 were used as substrates for phosphorylation by AtWNK8. (B) MALDI-TOF analysis revealed four peptides being phosphorylated by AtWNK8, which are highlighted in the model (compare Table 1). Serine and threonine residues with a high likelihood of phosphorylation (NetPhosServer 2.0) are enlarged in the peptide sequence and appear as spheres in the model. Numbers correspond to their sequence position.
We show here that a member of the Arabidopsis WNK family binds to and phosphorylates the V-ATPase subunit C. WNK kinases have received considerable attention since human WNKs were genetically linked to the regulation of blood pressure [23]. Our attempt to address the specificity of the interaction of AtWNK8 with AtVHA-C has led us to re-examine the available WNK sequences. Arabidopsis has been reported to have the largest WNK family [28]; our analysis confirms this notion and extends the Arabidopsis family to 11 members, whereas the larger rice genome encodes only seven members. Phylogenetic analysis shows that much of the diversification of plant WNKs has taken place before the divergence of mono- and dicotyledenous plants. Based on the fact that the yeast genome does not encode WNKs, it has been proposed that they are restricted to multicellular organisms [22]. However, database searches showed that WNKs are not only found in multicellular but also in unicellular eukaryotes including very distantly related groups like oomycetes and diplomonads. WNKs therefore represent an ancient group of protein kinases and it seems possible that the characteristic placement of the catalytic lysine residue in subdomain I found in all WNKs is
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939
937
Table 1 MALDI mass spectrometry analysis of unphosphorylated and phosphorylated peptides from AtVHA-C Start residue
End residue
Measured massa
Measured massb
Sequence
17 112 195 214 242 329
4 131 213 231 261 347
2314.863 2407.307 2383.057 2284.949 2491.273 2379.260
2394.832 2487.281 2463.027 2364.919 2571.243 2459.227
GSHMASMTGGQQMGRGSMTSR YPTMSPLKEVVDNIQSQVAK DWLACYETLTDYVVPRSSK KLFEDNEYALYTVTLFTR EKGFQVRDFEQSVEAQETRK KVRSILERLCDSTNSLYWK
a
Mass of unphosphorylated peptide. Mass of phosphorylated peptide.
b
Fig. 6. AtWNK8 phosphorylates but does not bind subunit C from other eukaryotes. (A) The C subunits of Saccharomyces cerevisiae (Sc), M. sexta (Ms), Caenorhabditis elegans (Ce), and Arabidopsis thaliana (At) were used as substrates for phopshorylation by AtWNK8. (B) The proteins used as phosphorylation substrates in (A) were immobilized on a nitrocellulose membrane (each 30 pmol) and tested for interaction with AtWNK8 DN2. Interaction was detected using an anti-GSTantibody.
Fig. 7. AtWNK8 phosphorylation of hybrid V1-complexes. Pooled gel filtration fractions representing of V1-C (lanes 1, 2), V1 (AtVHA-C) (lanes3, 4), and unbound VHA-C (lanes 5, 6) were used as substrates in phosphorylation assays with AtWNK8 (2, 4, 6) and in controls without kinase (1, 3, 5). (A) Coomassie stained gel, (B) autoradiograph. * marks autophosphorylation of AtWNK8.
actually older than the placement in subdomain II found in all other protein kinases. Outside the highly conserved kinase domain, WNKs differ considerably. The only additional region with identifiable sequence similarity found in all plant and in some unicellular WNKs overlaps with the autoinhibitory segment described in mammalian WNK1 [21]. We have shown that AtWNK8 is an active kinase and its activity is reduced in the presence of DC2, a kinase-inactive deletion that contains the putative autoinhibitory domain. Although this domain is also present in DN2, this deletion protein does not inhibit AtWNK8 autophosphorylation activity. This is reminiscent of results obtained with mammalian WNK1, where the autoinhibitory activity was found to be reduced when the C-terminal flanking coiled-coil domain was present [21]. Our results indicate that regulation via an autoinhibitory sequence is a conserved feature of plant and mammalian WNKs and is possibly also found in some of the WNKs from unicellular organisms. Interestingly, both rice and Arabidopsis contain one WNK gene that encodes only the kinase domain and it will be interesting to determine the activities and potential regulators of these unusual plant WNKs. Although mammalian WNKs have been shown to exist as tetramers, intermolecular phosphorylation has not been detected [21]. Our results show clearly that autophosphorylation of AtWNK8 takes place in an intermolecular fashion, indicating that plant WNKs may also be able to multimerize. As the coiled-coil domain, which has been proposed to be responsible for the multimerization of WNK1 [35] is not found in the plant proteins, it remains to be determined which motifs are necessary for the interaction of AtWNK8 with itself. The presence of multiple binding sites for SH3-domains suggests that the C-terminal domain of mammalian WNKs is able to interact with multiple partners [36], however so far no interactions between this domain and other proteins have been described. A small C-terminal fragment of AtWNK8 including a potential short coiled-coil domain is sufficient for binding of AtVHA-C, demonstrating that the C-terminal domain does indeed mediate protein–protein interactions. The predicted coiled-coil domain in AtWNK8 is comparably short and it remains to be determined if it really adopts a coil structure. The C-terminal end of AtWNK10, the closest relative of AtWNK8, shows about 50% amino acid identity and is also predicted to form a coiled-coil domain and yet it does not interact with AtVHA-C in the yeast two-hybrid system. The specificity of this interaction is accentuated by the fact that the binding domain of AtWNK8 is able to discriminate between subunits C from different eukaryotes, although they are functionally interchangeable [37–39] and therefore most likely have a very similar structure.
938
Protein phosphorylation is a ubiquitous mechanism that has been shown to regulate the activity of a large number of transport proteins including ion channels and proton pumps. Although pharmacological evidence suggests that V-ATPases are targets for protein kinase-mediated signaling [40,41], evidence for a direct interaction and phosphorylation of V-ATPase subunits has so far been missing. We have shown that AtWNK8 phosphorylates AtVHA-C and have identified four phosphorylated peptides each of which includes two or more serine or threonine residues, indicating that phosphorylation by AtWNK8 is not restricted to a particular domain of subunit C. Phosphorylation target sides of WNKs are not well characterized but it has been suggested that the P + 1 residue is not a strong specificity determinant [32]. Although we could not determine individual phosphorylated residues, the likely sites found in the different peptides indicate that the P 1 and P + 1 positions are not conserved. The phosphorylated peptides are either derived from the globular head domain, that has been proposed to bind to the catalytic domain of V1, or from the foot region assumed to be oriented to the membrane [11,12]. Site-directed mutagenesis revealed that the C-terminal foot region of subunit C might be important for stable assembly of V1 and V0 [13]. Furthermore, the nucleotide-binding property of subunit C has been localized to the C-terminal region of subunit C and the structural changes induced by nucleotide-binding might be involved in the regulation of the reversible V1V0 disassembly [15]. In vivo, phosphorylation could either be targeted to free subunit C molecules or intact V-ATPase complexes. Addition of recombinant yeast subunit C to the V1(-C) complex of Manduca significantly increased ATPase activity [11], supporting its functional role in regulation of ATPase activity [42,43]. We found that AtVHA-C can also be reconstituted into the V1-complex from Manduca and were able to show that it can be phosphorylated by AtWNK8 when present in V1. Surprisingly, several other V1-subunits are also substrates for AtWNK8 and it will be of interest to determine if they can interact with AtWNK8 and are phosphorylated in vivo. Similar to V-ATPase subunit C, subunit H is not involved in catalysis but was implicated in regulating the V-ATPase [44,45]. Phosphorylation of VHA-H might therefore be important for this regulatory function. The V-ATPase subunit B seems to be a likely target for a protein kinase, since it was described before as phosphorylated protein [19]. If the phosphorylation of other V1-subunits by AtWNK8 requires tight binding to AtWNK8 and takes place in vivo needs to be determined. Subunit C-homologues from diverse species are all phosphorylated by AtWNK8, although efficient binding is restricted to the Arabidopsis subunit. Taken together, these results indicate that the C-terminal interaction domain of AtWNK8 may represent a direct docking module that provides specificity and might facilitate efficient phosphorylation of subunit C under physiological conditions. Our results show for the first time, that the V-ATPase can be phosphorylated by a protein kinase. Although both AtVHA-C and AtWNK8 are ubiquitously expressed (data not shown), it remains to be determined if the two proteins colocalize and if phosphorylation of the V-ATPase takes place in vivo. Our results provide the framework necessary to address this question and to integrate V-ATPase regulation into known signaling networks.
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939 Acknowledgments: We are grateful to Joanne Chory for generous support in the initial phase of this work. We thank Claudia Oecking for critical reading of the manuscript and acknowledge the excellent technical assistance provided by Zhao-Xin Wang and Andrea Armbru¨ster. This work was supported by the Deutsche Forschungsgemeinschaft (SFB446 TP A20 to K.S.) and the School of Biological Sciences, NTU, Singapore.
References [1] Nelson, N. (2003) A journey from mammals to yeast with vacuolar H+-ATPase (V-ATPase). J. Bioenerg. Biomembr. 35, 281–289. [2] Nishi, T. and Forgac, M. (2002) The vacuolar (H+)-ATPases – nature’s most versatile proton pumps. Nat. Rev. Mol. Cell Biol. 3, 94–103. [3] Kane, P.M. and Smardon, A.M. (2003) Assembly and regulation of the yeast vacuolar H+-ATPase. J. Bioenerg. Biomembr. 35, 313–321. [4] Kane, P.M. (1995) Disassembly and reassembly of the yeast vacuolar H(+)-ATPase in vivo. J. Biol. Chem. 270, 17025–17032. [5] Sumner, J.P., Dow, J.A., Earley, F.G., Klein, U., Ja¨ger, D. and Wieczorek, H. (1995) Regulation of plasma membrane V-ATPase activity by dissociation of peripheral subunits. J. Biol. Chem. 270, 5649–5653. [6] Kane, P.M. and Parra, K.J. (2000) Assembly and regulation of the yeast vacuolar H(+)-ATPase. J. Exp. Biol. 203, 81–87. [7] Parra, K.J. and Kane, P.M. (1998) Reversible association between the V1 and V0 domains of yeast vacuolar H+-ATPase is an unconventional glucose-induced effect. Mol. Cell Biol. 18, 7064– 7074. [8] Ho, M.N., Hirata, R., Umemoto, N., Ohya, Y., Takatsuki, A., Stevens, T.H. and Anraku, Y. (1993) VMA13 encodes a 54-kDa vacuolar H(+)-ATPase subunit required for activity but not assembly of the enzyme complex in Saccharomyces cerevisiae. J. Biol. Chem. 268, 18286–18292. [9] Doherty, R.D. and Kane, P.M. (1993) Partial assembly of the yeast vacuolar H(+)-ATPase in mutants lacking one subunit of the enzyme. J. Biol. Chem. 268, 16845–16851. [10] Ho, M.N., Hill, K.J., Lindorfer, M.A. and Stevens, T.H. (1993) Isolation of vacuolar membrane H(+)-ATPase-deficient yeast mutants; the VMA5 and VMA4 genes are essential for assembly and activity of the vacuolar H(+)-ATPase. J. Biol. Chem. 268, 221–227. [11] Armbru¨ster, A., Svergun, D.I., Coskun, U., Juliano, S., Bailer, S.M. and Gru¨ber, G. (2004) Structural analysis of the stalk subunit Vma5p of the yeast V-ATPase in solution. FEBS Lett. 570, 119–125. [12] Drory, O., Frolow, F. and Nelson, N. (2004) Crystal structure of yeast V-ATPase subunit C reveals its stator function. EMBO Rep. 5, 1148–1152. [13] Curtis, K.K., Francis, S.A., Oluwatosin, Y. and Kane, P.M. (2002) Mutational analysis of the subunit C (Vma5p) of the yeast vacuolar H+-ATPase. J. Biol. Chem. 277, 8979–8988. [14] Vitavska, O., Merzendorfer, H. and Wieczorek, H. (2005) The VATPase subunit C binds to polymeric F-actin as well as to monomeric G-actin and induces cross-linking of actin filaments. J. Biol. Chem. 280, 1070–1076. [15] Armbru¨ster, A., Hohn, C., Hermesdorf, A., Schumacher, K., Borsch, M. and Gru¨ber, G. (2005) Evidence for major structural changes in subunit C of the vacuolar ATPase due to nucleotide binding. FEBS Lett. 579, 1961–1967. [16] Seol, J.H., Shevchenko, A. and Deshaies, R.J. (2001) Skp1 forms multiple protein complexes, including RAVE, a regulator of VATPase assembly. Nat. Cell Biol. 3, 384–391. [17] Smardon, A.M., Tarsio, M. and Kane, P.M. (2002) The RAVE complex is essential for stable assembly of the yeast V-ATPase. J. Biol. Chem. 277, 13831–13839. [18] Johnson, S.A. and Hunter, T. (2005) Kinomics: methods for deciphering the kinome. Nat. Methods 2, 17–25. [19] Myers, M. and Forgac, M. (1993) The coated vesicle vacuolar (H+)-ATPase associates with and is phosphorylated by the 50kDa polypeptide of the clathrin assembly protein AP-2. J. Biol. Chem. 268, 9184–9186.
A. Hong-Hermesdorf et al. / FEBS Letters 580 (2006) 932–939 [20] McCubbin, A.G., Ritchie, S.M., Swanson, S.J. and Gilroy, S. (2004) The calcium-dependent protein kinase HvCDPK1 mediates the gibberellic acid response of the barley aleurone through regulation of vacuolar function. Plant J. 39, 206–218. [21] Xu, B.E., Min, X., Stippec, S., Lee, B.H., Goldsmith, E.J. and Cobb, M.H. (2002) Regulation of WNK1 by an autoinhibitory domain and autophosphorylation. J. Biol. Chem. 277, 48456– 48462. [22] Xu, B., English, J.M., Wilsbacher, J.L., Stippec, S., Goldsmith, E.J. and Cobb, M.H. (2000) WNK1, a novel mammalian serine/ threonine protein kinase lacking the catalytic lysine in subdomain II. J. Biol. Chem. 275, 16795–16801. [23] Wilson, F.H. et al. (2001) Human hypertension caused by mutations in WNK kinases. Science 293, 1107–1112. [24] Verissimo, F. and Jordan, P. (2001) WNK kinases, a novel protein kinase subfamily in multi-cellular organisms. Oncogene 20, 5562–5569. [25] Gamba, G. (2005) Role of WNK kinases in regulating tubular salt and potassium transport and in the development of hypertension. Am. J. Physiol. Renal Physiol. 288, F245–F252. [26] Wilson, F.H. et al. (2003) Molecular pathogenesis of inherited hypertension with hyperkalemia: the Na–Cl cotransporter is inhibited by wild-type but not mutant WNK4. Proc. Natl. Acad. Sci. USA 100, 680–684. [27] Lee, B.H. et al. (2004) WNK1 phosphorylates synaptotagmin 2 and modulates its membrane binding. Mol. Cell 15, 741–751. [28] Nakamichi, N., Murakami-Kojima, M., Sato, E., Kishi, Y., Yamashino, T. and Mizuno, T. (2002) Compilation and characterization of a novel WNK family of protein kinases in Arabiodpsis thaliana with reference to circadian rhythms. Biosci. Biotechnol. Biochem. 66, 2429–2436. [29] Murakami-Kojima, M., Nakamichi, N., Yamashino, T. and Mizuno, T. (2002) The APRR3 component of the clock-associated APRR1/TOC1 quintet is phosphorylated by a novel protein kinase belonging to the WNK family, the gene for which is also transcribed rhythmically in Arabidopsis thaliana. Plant Cell Physiol. 43, 675–683. [30] Roos, M., Soskic, V., Poznanovic, S. and Godovac-Zimmermann, J. (1998) Post-translational modifications of endothelin receptor B from bovine lungs analyzed by mass spectrometry. J. Biol. Chem. 273, 924–931. [31] Wilkins, M.R., Lindskog, I., Gasteiger, E., Bairoch, A., Sanchez, J.C., Hochstrasser, D.F. and Appel, R.D. (1997) Detailed peptide characterization using PEPTIDEMASS – a World-Wide-Webaccessible tool. Electrophoresis 18, 403–408. [32] Min, X., Lee, B.H., Cobb, M.H. and Goldsmith, E.J. (2004) Crystal structure of the kinase domain of WNK1, a kinase that causes a hereditary form of hypertension. Structure (Camb.) 12, 1303–1311.
939 [33] Ruiz-Perez, V.L., Murillo, F.J. and Torres-Martinez, S. (1995) PkpA, a novel Phycomyces blakesleeanus serine/threonine protein kinase. Curr. Genet. 28, 309–316. [34] Blom, N., Sicheritz-Ponten, T., Gupta, R., Gammeltoft, S. and Brunak, S. (2004) Prediction of post-translational glycosylation and phosphorylation of proteins from the amino acid sequence. Proteomics 4, 1633–1649. [35] Lenertz, L.Y., Lee, B.H., Min, X., Xu, B.E., Wedin, K., Earnest, S., Goldsmith, E.J. and Cobb, M.H. (2005) Properties of WNK1 and implications for other family members. J. Biol. Chem. 280, 26653–26658. [36] Xu, B.E., Lee, B.H., Min, X., Lenertz, L., Heise, C.J., Stippec, S., Goldsmith, E.J. and Cobb, M.H. (2005) WNK1: analysis of protein kinase structure, downstream targets, and potential roles in hypertension. Cell Res. 15, 6–10. [37] Beltran, C., Kopecky, J., Pan, Y.C., Nelson, H. and Nelson, N. (1992) Cloning and mutational analysis of the gene encoding subunit C of yeast vacuolar H(+)-ATPase. J. Biol. Chem. 267, 774–779. [38] Oka, T., Yamamoto, R. and Futai, M. (1997) Three vha genes encode proteolipids of Caenorhabditis elegans vacuolar-type ATPase. Gene structures and preferential expression in an Hshaped excretory cell and rectal cells. J. Biol. Chem. 272, 24387– 24392. [39] Schumacher, K., Vafeados, D., McCarthy, M., Sze, H., Wilkins, T. and Chory, J. (1999) The Arabidopsis det3 mutant reveals a central role for the vacuolar H(+)-ATPase in plant growth and development. Genes Dev. 13, 3259–3270. [40] Carini, R., Grazia De Cesaris, M., Splendore, R. and Albano, E. (2001) Stimulation of p38 MAP kinase reduces acidosis and Na(+) overload in preconditioned hepatocytes. FEBS Lett. 491, 180–183. [41] Nordstrom, T., Grinstein, S., Brisseau, G.F., Manolson, M.F. and Rotstein, O.D. (1994) Protein kinase C activation accelerates proton extrusion by vacuolar-type H(+)-ATPases in murine peritoneal macrophages. FEBS Lett. 350, 82–86. [42] Sun-Wada, G.H., Murata, Y., Namba, M., Yamamoto, A., Wada, Y. and Futai, M. (2003) Mouse proton pump ATPase C subunit isoforms (C2-a and C2-b) specifically expressed in kidney and lung. J. Biol. Chem. 278, 44843–44851. [43] Peng, S.B., Zhang, Y., Tsai, S.J., Xie, X.S. and Stone, D.K. (1994) Reconstitution of recombinant 33-kDa subunit of the clathrincoated vesicle H(+)-ATPase. J. Biol. Chem. 269, 11356–11360. [44] Domgall, I., Venzke, D., Luttge, U., Ratajczak, R. and Bottcher, B. (2002) Three-dimensional map of a plant V-ATPase based on electron microscopy. J. Biol. Chem. 277, 13115–13121. [45] Keenan Curtis, K. and Kane, P.M. (2002) Novel vacuolar H+ATPase complexes resulting from overproduction of Vma5p and Vma13p. J. Biol. Chem. 277, 2716–2724.