• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Abscisic Acid Stimulates a Calcium-Dependent Protein Kinase in Grape Berry1[W]

Posted on: Wednesday, 1 March 2006, 03:02 CST

By Yu, Xiang-Chun; Li, Mei-Jun; Gao, Gui-Feng; Feng, Hai-Zhong; Et al

It has been demonstrated that calcium plays a central role in mediating abscisic acid (ABA) signaling, but many of the Ca^sup 2+^- binding sensory proteins as the components of the ABA-signaling pathway remain to be elucidated. Here we identified, characterized, and purified a 58-kD ABA-stimulated calcium-dependent protein kinase from the mesocarp of grape berries (Vitis vinifera Vitis labrusca), designated ACPK1 (for ABA-stimulated calcium-dependent protein kinase1). ABA stimulates ACPK1 in a dose-dependent manner, and the ACPK1 expression and enzyme activities alter accordantly with the endogenous ABA concentrations during fruit development. The ABA- induced ACPK1 stimulation appears to be transient with a rapid effect in 15 min but also with a slow and steady state of induction after 60 min. ABA acts on ACPK1 indirectly and dependency on in vivo state of the tissues. Two inactive ABA isomers, (-)-2-cis, 4-trans- ABA and 2-trans, 4-trans-()-ABA, are ineffective for inducing ACPK1 stimulation, revealing that the ABA-induced effect is stereo specific to physiological active (+)-2-cis, 4-trans-ABA. The other phytohormones such as auxin indoleacetic acid, gibberellic acid, synthetic cytokinin N-benzyl-6-aminopurine, and brassinolide are also ineffective in this ACPK1 stimulation. Based on sequencing of the two-dimensional electrophoresis-purified ACPK1, we cloned the ACPK1 gene. The ACPK1 is expressed specifically in grape berry covering a fleshy portion and seeds, and in a developmental stage- dependent manner. We further showed that ACPK1 is localized in both plasma membranes and chloroplasts/plastids and positively regulates plasma membrane H+-ATPase in vitro, suggesting that ACPK1 may be involved in the ABA-signaling pathway.

The phytohormone abscisic acid (ABA) regulates many important events during both vegetative and reproductive growth of plants as well as in plants' adaptation to their environment. During vegetative growth, ABA is a central signal of plant response to various environmental challenges including drought, salt, and cold stresses (for review, see Koornneef et al., 1998; Leung and Giraudat, 1998; Finkelstein and Rock, 2002). In reproductive organ seeds, ABA is responsible for the seed storage reserve synthesis, acquisition of desiccation tolerance and dormancy, and induction of stress tolerance (for review, see Finkelstein et al., 2002). Fleshy fruits are also the essential portions of reproductive organs and economically important harvest organs, as are crop seeds. In fleshy fruits as in seeds, ABA regulates various processes concerning assimilate uptake and metabolism to enhance reserve accumulation in these economic sinks (Yamaki and Asakura, 1991; Rock and Quatrano, 1995; Wayne and John, 1996; Opaskornkul et al., 1999; Peng et al., 2003; Pan et al., 2005). Grape berry (Vitis vinifera and also Vitis labrusca) is one of the most widely cultivated fruit trees in the world and also one of the typical non-climacteric fruits. The ripening of grape berry is considered to be independent of the hormone ethylene but to be triggered essentially by ABA (for review, see Coombe, 1992). However, the ABA-signaling pathway in regulating development of the fleshy fruits remains essentially elusive.

ABA signal transduction has been extensively studied in the past years. Numerous cellular components that modulate ABA responses have been identified, leading to considerable progress in understanding the ABA-signaling pathway (for review, see Finkelstein et al., 2002; Himmelbach et al., 2003; Fan et al, 2004). Reversible protein phosphorylation, catalyzed by protein kinases and phosphatases, has been believed to play central roles in ABA signal transduction. Calciummodulated protein phosphatases (PPs) 2C ABIl and ABI2 are the two most characterized, homologous negative regulators of ABA signaling (Leung et al., 1994, 1997; Meyer et al., 1994; Sheen, 1998; Gosti et al., 1999; Merlot et al., 2001). They interact with multiple cellular targets such as a calcineurin B-like (CBL) protein kinase CIPK15 and a CBL Ca^sup 2+^-binding protein ScaBPS. CIPK15 and one of its homologs CIPK3 and ScaBPS are all involved in ABA signaling as negative regulators (Guo et al., 2002; Kim et al., 2003), possibly providing PP2Cs ABIl and ABI2 with the Ca^sup 2+^ sensor (Pandey et al., 2004) when forming a protein complex for perceiving the upstream signal Ca^sup 2+^ (Alien et al., 1999). A link has been established between CIPK15 and an AP2 transcription factor AtERT7 that negatively regulates ABA response as a kinase substrate of CIPK5 (Song et al., 2005). A nuclear-localized transcription factor homeodomain protein ATHB6 has also been identified as a more downstream component of the ABIl and ABI2 (Himmelbach et al., 2002). In contrast to PP2Cs ABIl and ABI2, a PP 2A, RCNl, has been identified as a positive regulator of ABA response involved in early events of ABA signaling (Kwak et al, 2002).

Two members of the family of SNFl-related protein kinase, PKABAl from wheat (Triticum aestivum) and AAPK from broad bean (Vicia faba), were characterized as the ABA-stimulated protein kinases and positive regulators of ABA response (Anderberg and Walker-Simmons, 1992; Li and Assmann, 1996; Gomez-Cadenas et al., 1999, 2001; Li et al., 2000, 2002; Johnson et al., 2002). PKABAl mediates ABA- suppressed gene expression (Gomez-Cadenas et al., 1999), and phosphorylates and activates the ABA-responsive basic-domain Leu- zipper (bZIP) transcription factor TaABF, a homolog of ABI5 of Arabidopsis (Arabidopsis thaliana; Johnson et al., 2002). AAPK mediates ABA-induced stomatal closure and anion channels (Li et al., 2000) and modulates an RNA-binding protein AKIPl by phosphorylating it and inducing its translocation into subnuclear speckles in guard cells (Li et al., 2002). An AAPK ortholog, OSTl, has also been identified as a regulator of stomatal response to ABA (Mustilli et al, 2002). Recently, three homologs of AAPK and OSTl in rice (Oryza saliva), SAPK8, SAPK9, and SAPKlO, were shown to be activated by ABA, suggesting their potential role in ABA signaling (Kobayashi et al., 2004).

Particular members of another class of important protein kinases, mitogen-activated protein kinase (MAPK), have been reported to be also activated by ABA (Knetsch et al., 1996; Burnett et al., 2000; D'Souza and Johri, 2002; Lu et al, 2002; Xiong and Yang, 2003). An Arabidopsis MAPK, AtMPKS, and a rice MAPK, OsMAPKS, have been identified as ABA-activated MAPKs. The former mediates postgermination arrest of development by ABA (Lu et al, 2002), and the latter is involved in disease resistance and abiotic stress tolerance (Xiong and Yang, 2003).

It has been well accepted that calcium is a central regulator of plant growth and development (Hepler, 2005). Specific calcium signatures are recognized by different calcium sensors to transduce specific calciummediating signal into downstream events (Sanders et al, 1999; Harmon et al., 2000; Rudd and Franklin-Tong, 2001). In plants, there are several known classes of Ca^sup 2+^-binding sensory proteins, including calmodulin (CaM) and CaM-related proteins (Zielinski, 1998; Luan et al., 2002), CBL proteins (Luan et al., 2002), and calcium-dependent protein kinases (CDPKs; Harmon et al., 2001; Cheng et al., 2002). Among them, CDPKs, a novel dass of Ca^sup 2+^ sensors having both kinase and CaM-like domain, are the best characterized and involved in most of calciumstimulated protein kinase activities iden tified in plants (Harmon et al., 2001; Cheng et al., 2002). Encoded by a large multigene family with possible redundancy and/or diversity in their functions (Harmon et al., 2001; Cheng et al., 2002), CDPKs are believed to be important components in plant hormone signaling (Cheng et al, 2002; Ludwig et al., 2004). Using a protoplast transient expression system, two Arabidopsis CDPKs, AtCPKlO and AtCPKSO, have been demonstrated to activate a stress- and ABA-inducible promoter, showing the connection of CDPKs to ABA-signaling pathway (Sheen, 1996). However, a previous report showed that ABA had no effect on the induction of AtCPKlO mRNA (Urao et al., 1994), and it has been unclear whether AtCPKSO is stimulated by ABA. This suggests complexity of CDPK signaling in mediating ABA signal, m addition, calcium-dependent histone-phosphorylating activity was reported to be activated by ABA in rice seedling, but the CDPK gene responsible for this activity was not identified (Li and Komatsu, 2000). In tobacco (Nicotiana tabacum), the transcripts of a CDPK gene, NtCDPKl, were enhanced unspecifically by various phytohormones ABA, indole-3-acetic acid (IAA), GA3, and synthetic cytokinin benzyladenine and other growth substances such as jasmonic acid, but it is unknown whether the NtCDPKl is activated by ABA in its kinase activities (Yoon et al., 1999). Thus, information is currently lacking about ABA-activated CDPKs identified by combined biochemical and molecular approaches as potential ABA signal transducers. In this report, we identified, characterized, and purified a 58-kD ABA specifically stimulated CDPK from grape berry, designated ACPKl. We isolated the cDNA coding for the ACPKl based on sequencing of the purified ACPKl protein. Also, we showed that ACPKl positively regulates plasma membrane (PM) H+-ATPase in vitro, suggesting that AC\PKl may be involved in ABA-signaling pathway in grape berry.

RESULTS

Characterization of a 58-kD Membrane-Associated CDPK in Grape Berry

Grape is a berry fruit, comprising the fleshy portion and seeds. The fleshy portion covers three parts, i.e. pericarp and endocarp that both are several cell layers thick, and mesocarp that makes the large fleshy portion. Using in-gel assays, we detected in the microsomes of the grape mesocarp a calcium-dependent autophosphorylated protein with a molecular mass of 58 kD (Fig. IA). This 58-kD protein phosphorylates histone III-S, one of the best exogenous substrates for assaying CDPKs (Roberts and Harmon, 1992), and this kinase activity depends also on the presence of Ca^sup 2+^ (Fig. 1A). The 58-kD kinase was Lmmunorecognized by an antiserum raised against the CaM-like domain of soybean (Glycine max) CDPKa (Bachmann et al., 1996; Fig. IA). In the soluble fractions or in the absence of free Ca^sup 2+^, neither the autophosphorylation nor histonephosphorylating kinase activity was detected (Fig. IA). These results suggest that the 58-kD membrane-associated kinase is a member of CDPK family.

Figure 1. Biochemical characterization of a 58-kD membrane- associated Ca^sup +^-dependent protein kinase in grape berry. A, A 58-kD membrane-associated Ca^sup 2+^-dependent protein kinase is present in grape berry. Soluble fraction (lanes labeled S, 40 g of protein) and microsomes (lanes labeled M, 20 g of protein) were separated on a 12% SDS-polyacrylamide gel. The separating gel for assaying kinase catalytic activity was polymerized in the presence of histone III-S. Autophosphorylation and kinase activities were assayed in the presence (+Ca^sup 2+^) or absence (-Ca^sup 2+^) of free Ca^sup 2+^ as described in "Materials and Methods." The molecular masses of protein standards are shown at the left of the sections in kD. Autophosphorylating (gels above) and histonephosphorylating activity (gels below) of 58-kD kinase are shown in the left and middle sections, and immunoblotting in the right section. Immunoblotting was done with the antiserum directed against the CaM-1 ike domain of soybean CDPKa as described in "Materials and Methods." B, Ca^sup 2+^-dependent electrophoretic mobility shift of the 58-kD kinase in both the in-gel autophosphorylation (a) and histone-phosphorylating activity (b) assays. Ca^sup 2+^ or EGTA to a final concentration of 2 mM was added to the microsomal proteins dissolved in SDS-PAGE sample buffer. After SDS-PAGE, the in-gel phosphorylation assays were done in the presence of Ca^sup 2+^. -Ca^sup 2+^ and +Ca^sup 2+^ indicate the absence and presence of Ca^sup 2+^ in the SDS-PAGE sample buffer, respectively. C, Ca^sup 2+^ dependence of the in-gel autophosphorylation (indicated by 58-kD with arrow) and in vitro activity (columnar figures) of the kinase in a medium pH at 7.5 and in the presence of Mg^sup 2+^ at 10 mMand EGTA at 0.45 mM. The in vitro and in-gel histone III-S-phosphorylating activities and the in- gel autophosphorylation of the kinase were assayed as described in "Materials and Methods." Values in the columnar figures are means SE (n = 5). D, The same assay as in C but in different concentrations of Mg^sup 2+^ and in the presence of 0.55 mM Ca^sup 2+^. E, The same assay as in C but in different medium pHs and in the presence of 0.55 mM Ca^sup 2+^ and 10 mM Mg^sup 2+^. F, Inhibition of both the in-gel autophosphorylation (indicated by 58- kD autophosph) and histone-phosphorylating activity (indicated by 58- kD histone phosph) of the 58-kD kinase by CaM antagonists or kinase inhibitors. CaM was used at 5 M; TFP or WS at 250 M; W7 at 100 (W7*) or 250 M (W7**), and K252a at 10 M. These reagents were added, respectively, to the phosphorylation reaction medium (buffer B as described in "Materials and Methods") for a preincubation and a subsequent reaction incubation for ^sup 32^P-labeling to the kinase or its substrate histone as described in "Materials and Methods;" - and + indicate the absence and presence of Ca^sup 2+^ in the reaction buffer, respectively.

Ca^sup 2+^-binding proteins such as CDPKs migrate in gels at different rates in the Ca^sup 2+^-bound versus Ca^sup 2+^-free state (Roberts and Harmon, 1992). To investigate this phenomenon, we added Ca^sup 2+^ or EGTA to the protein sample just before electrophoresis and then the in-gel phosphorylation was analyzed in the presence of Ca^sup 2+^. The assays of both in-gel autophosphorylation and histone-phosphorylating activity of the 58-kD kinase showed a clear mobility shift when the kinase migrates in the presence of Ca^sup 2+^ (Fig. IB). The in-gel autophosphorylation and in vitro histone- phosphorylating activities of the 58-kD kinase depended strongly on Ca^sup 2+^ (Fig. 1C) and Mg^sup 2+^ (Fig. 1D) concentrations and in a weaker extent on medium pHs (Fig. 1E). The assays of in-gel kinase activity gave the same results as the in vitro assays (data not shown).

We also analyzed the effects of the CaM antagonists trifluoperazine (TFP) and N-(6-aminohexyl)-5-chloro1-naphthalene sulfonamide (W7) and inhibitor of Ser/ Thr protein kinases K252a on the 58-kD kinase. The calcium-dependent in-gel autophosphorylation and histone-phosphorylating activity of the 58-kD kinase could be inhibited by TFP, W7, and K252a, and the dose dependence of W7 for the inhibition was observed (Fig. 1F). By contrast, CaM and N-(6- arninohexyl)1-naphthalene sulfonamide (W5, an inactive analog of W7), had no apparent effect on autophosphorylation or phosphorylating activity of the 58-kD kinase (Fig. 1F). Taken together, all of the results consistently indicate that the 58-kD membrane-associated kinase is a CDPK.

ABA Stimulates the 58-kD CDPK

To investigate the effects of ABA on grape CDPK, we used throughout this study a system to incubate in vivo the berry tissues in the ABA-containing medium essentially according to the technique of Peng et al. (2003), as detailed in "Materials and Methods." The assays of in vitro phosphorylation of the total microsomal proteins extracted from ABA-free incubated tissues allowed the detection of two bands of calciumindependent phosphoproteins (CIPs; Fig. 2A, a, lane 1) and two other bands of calcium-dependent phosphoproteins (CDPs; Fig. 2A, a, lane 2). The molecular masses of the CIPs are 72 (CIP72) and 40 kD (CIP40), and those of CDPs 64 (CDP64) and 58 kD (CDP58), respectively (Fig. 2A, a). ABA appeared to stimulate three of them, CIP72, CDP64, and CDP58, but not CIP40 (Fig. 2Aa). However, only one band of phosphorylation at the point of 58 kD was detected by in-gel phosphorylation (Fig. 2A, a and b), probably because the other in vitro-detected phosphoproteins may be only substrates of protein kinase, or that phosphorylation of diverse proteins may require the complex in vitro system where total membrane proteins are present and protein-protein interactions related to the phosphorylation may be realizable. The CDP58 phosphoprotein detected in the in vitro system should be the identified 58-kD kinase (Fig. 1). ABA strongly stimulated the 58-kD kinase in both its autophosphorylation and histone-phosphorylating kinase activities (Fig. 2A, b and c). In accordance with this, the amount of the 58- kD kinase, estimated by immunodetection with the antiserum to CaM domain of soybean CDPKa, was also enhanced by ABA treatment (Fig. 2A, d). The ABA-induced stimulation of the 58-kD kinase was dependent on the dose of ABA application, and the ABA concentration of 10 M maximized the effects (Fig. 2A, b-d). When ABA at concentrations above 10 M was applied, the ABA-induced effects declined but still remained at steady, higher levels (Fig. 2A, b- d). It is noteworthy that the concentrations of ABA within the treated tissues corresponding to the concentrations of applied ABA below 10 M did not overpass the highest level of the physiological concentrations in grape berries (Fig. 2A; also see Fig. 8).

We further analyzed time course of the ABA-induced effects. ABA rapidly stimulated the 58-kD kinase in its autophosphorylation and kinase activity, and the ABA-induced effects appeared to be transient (Fig. 2B, a and b). An incubation of 15 min could induce the effects, and the effects were maximized with a 60-min incubation (Fig. 2B). A longer incubation over 60 min allowed still a steady, higher level of the 58-kD kinase stimulation, though the effects declined (Fig. 2B, a and b). It is noteworthy, however, that the increase of the apparent amount of the 58-kD kinase began at the point of 45-min incubation, significantly later than did its autophosphorylation and kinase activity, and this enzyme amount did not decrease after 60 min of incubation, which, contrasting with the enzyme activities, showed a typical saturation curve (Fig. 2B, c). These data suggest that the rapid phase (15-30 min) of the 58-kD kinase stimulation by ABA may be due only to a posttranslational modification, but the slow phase (45 min and later) may involve the events of both posttranslational regulation and de novo protein synthesis of the enzyme (Fig. 2B). Whatever mechanism is involved, all above results allow the identification of the 58-kD kinase as an ABA-stimulated CDPK, and so we designate it ACPKl.

Figure 2. ABA stimulates the 58-kD Ca^sup 2+^-dependent protein kinase. A, ABA stimulates the 58-kD Ca^sup 2+^-dependent protein kinase in a dose-dependent manner, a, ABA-stimulated in vitro phosphorylation in total microsomal proteins. The berry tissues were incubated in a medium (pH 5.5) containing O (control) or 10 M()-ABA for 1 h. The microsomes prepared from the tissues were subjected to phosphorylation in vitro and separated by SDS-PAGE, and then phosphoproteins were detected by autoradiography. The detailed procedures of these assays were described in "Materials and Methods." The phosphoproteins from the control tissues are shown after the in vitro phosphorylation in the absence (lane 1) o\r presence (lane 2) of Ca^sup 2+^. The phosphorylation level of some phosphoproteins increases in the microsomes from the ABA-treated tissues after the in vitro phosphorylation in the presence of Ca^sup 2+^ (lane 3). The calculated molecular masses according to the mobility rate of protein standards are shown at the left of the section in kD; - and + indicate the absence and presence of Ca^sup 2+^, respectively, b to d, ABA-induced in-gel-stimulation of the 58- kD Ca^sup 2+^-dependent protein kinase is dose dependent. The microsomal proteins prepared from the ABA-treated tissues were separated by SDS-PACE and then subjected to in-gel autophosphorylation (b) and histonephosphorylating activity (c) assays and immunoblotting analysis with rabbit polyclonal antibodies to the CaM-like domain of soybean CDPKα (d). The procedures of the assays were described in "Materials and Methods." The numbers displayed in line [ABA] indicate the exogenous ABA concentrations applied to the berry tissues, and those in line [ABA]* show the ABA concentrations within the treated tissues determined by radioimmunoassay as described in "Materials and Methods." The ABAstimulated in-gel autophosphorylated protein (b) corresponds to the 58-kD phosphoprotein shown in section a. The molecular mass of the kinase phosphorylating histone III-S (c) or that of the signal immunodetected by anti-soybean CDPKa serum (d) is shown to be 58 kD. B, Time course of the ABA-induced stimulation of the 58-kD Ca^sup 2+^-dependent protein kinase. Berry tissues were incubated for different durations of time from O to 180 min in the medium containing 10 M ABA (ABA) or the same amount of ethanol for solubilizing ABA (Ethanol) as described in "Materials and Methods." The microsomes prepared from the treated tissues were used for analysis of the 58-kD Ca^sup 2+^-dependent protein kinase as described in A. a, In-gel autophosphorylation of the 58-kD kinase. b, In-gel histone-phosphorylating activity of the 58-kD kinase. c, 58-kD immunosignal detected by the rabbit antiserum directed against the calmudulin-like domain of soybean CDPKα.

ABA-Induced Stimulation of ACPKl Is Specific to the Physiologically Active Form of ABA

To determine the stereo specificity of ABA in the ACPKl stimulation, two ABA isomers (-)-2-cis, 4-trans-ABA [(-)-ABA], and 2- trans, 4-trans-()-ABA (trans-ABA), were used in the experiment. The used ()-ABA in the experiments is a mixture of (+)-2-cis, 4-trans- ABA [(+)-ABA] and (-)-ABA with (+)-ABA being physiologically active.

The two ABA isomers (-)-ABA and trans-ABA are structurally similar to physiologically active (+)-ABA, but functionally inactive (Balsevich et al., 1994; Hill et al., 1995; Walker-Simmons et al., 1997). We previously showed also that the two ABA isomers were unable to be bound to ABA-specific binding proteins that possess potential receptor nature (Zhang et al., 1999, 2001, 2002). The assays of both the in-gel autophosphorylation and histone- phosphorylating activities showed that the two ABA analogs were unable to stimulate the ACPKl kinase (Fig. 3A), revealing that the ACPKl kinase stimulation induced by ABA is stereo specific, only physiologically active (+)-ABA being effective.

We further analyzed the effects of some other phytohormones on activities of the ACPKl kinase. In the assayed substances auxin IAA, GA^sub 3^, synthetic cytokinin N-benzyl-6-aminopurine, and brassinolide, none of them had any effect on ACPKl in either its autophosphorylation or histone-phosphorylating kinase activity (Fig. 3B). This result indicates that the ACPKl kinase appears to be stimulated specifically and exclusively by ABA.

ABA Indirectly Stimulates ACPKl, and Influx of Ca^sup ++^ May Be Involved in This Stimulation

To investigate whether ABA directly acts on the ACPKl kinase, ABA was added directly to the reaction medium of the in-gel phosphorylation instead of incubating the tissues in vivo in the ABA- containing medium. The results showed that ABA had no direct effect on the ACPKl kinase in either its autophosphorylation or kinase activity (Fig. 3C), indicating that ABA indirectly stimulates the ACPKl kinase.

We further investigate possible roles of Ca^sup 2+^ influx in mediating the ABA-induced ACPKl stimulation. Two Ca^sup 2+^ chelators, 1,2-bis(2-aminophenoxy)-ethane-N,N,N,N-tetraacetic acid (BAPTA) and EGTA, were used to incubate the tissues in vivo together with ABA. This treatment of eliminating apoplasmic Ca^sup 2+^ significantly reduced the ACPKl stimulation by ABA, though BAPTA was more effective compared to EGTA (Fig. 3D). A further experiment was done with A23187, a Ca^sup 2+^ ionophore that allows diffusion of Ca^sup 2+^ through plasma membranes in an electrically neutral manner, thus creating an artificial influx of Ca^sup 2+^ from apoplasm into cytoplasm. However, the treatment with A23187 had no significant effect on the ABA-induced ACPKl stimulation (Fig. 3D, lanes 5 and 6). The treatment with A23187 alone did not stimulate ACPKl (Fig. 3D, lane 7). These data suggest that influx of Ca^sup 2+^ from apoplasm into cytoplasm may be involved in the ABAinduced events, which requires a mechanism such as calcium channels to mediate influx of Ca^sup 2+^, but ABA signaling in these events may be more complex than dependent on influx of Ca^sup 2+^.

Purification and Molecular Cloning of ACPK1

We adopted a technique to purify ACPKl kinase from SDS- polyacrylamide gels after the total membrane proteins were separated by SDS-PAGE, because the approaches based on columnar chromatography that we used for this purification were unsuccessful. The exact position of the 58-kD kinase on gels (Fig. 4A) was determined by both autophosphorylation and immunodetection with antiserum against the CaM-like domain of soybean CDPKa, and thus the portions of the gels containing the 58-kD kinase were excised and eluted. The lyophilized eluate was analyzed for assessing the purification. A total amount of 400 g purified proteins was obtained from 30 mg crude microsomal proteins, i.e. 50 gels 10 lanes with 60 g crude microsomal proteins on each lane.

The purified proteins migrated as a single protein band at the point of 58-kD on SDS-polyacrylamide gel (Fig. 4B). The purified 58- kD proteins could autophosphorylate (Fig. 4C) and phosphorylate histone III-s (Fig. 4D) in the presence of Ca^sup 2+^, but both of their autophosphorylation and kinase activities were lost in the absence of Ca^sup 2+^ (data not shown). It is particularly noteworthy that SDS-PAGE of the apparently purified protein in the presence of calcium displayed two bands (Fig. 4E), of which one possessing Ca^sup 2+^-dependent electrophoretic mobility shift should be the 58-kD kinase, and another may be a protein other than this kinase. This nature of Ca^sup 2+^-dependent electrophoretic mobility shift of the 58-kD kinase was further shown in the assays of autophosphorylation (Fig. 4F), histonephosphorylating activity (Fig. 4G), and immunoblotting with antiserum against the CaM-like domain of soybean CDPKα (Fig. 4H).

The 58-kD proteins were further identified and purified by isoelectric focusing electrophoresis (IEF)/ SDS-PAGE two- dimensional electrophoresis. Two protein bands were observed on SDS- polyacrylamide gel with the same molecular mass of 58 kD but different pis (Fig. 41, lanes b and c), which is consistent with the above results of SDS-PAGE. One of them (Fig. 41, lane c) was identified as the 58-kD ACPKl kinase by its features of Ca^sup 2+^- dependent autophosphorylation (Fig. 4J), histone-phosphorylating activity (Fig. 4K), and immunorecognition by anti-ACPL1-N^sup 40^ serum (Fig. 4L), a specific antibody to ACPKl (see below).

Figure 3. Some features of ABA-induced ACPK1 kinase stimulation. A, Specificity of the stimulation of the ACPK1 kinase to the physiologically active form of ABA. Berry tissues were incubated in the medium containing various concentrations (indicated by the numbers under lines as M) of ()-ABA or ABA isomers for 1 h as described in "Materials and Methods." In-gel phosphorylation (a) and histone-phosphorylating activity (b) of the 58-kD kinase in the microsomes prepared from the tissues treated by ()-ABA (as a control) or its isomers (-)-ABA or trans-ABA were analyzed. B, ACPK1 appears to be stimulated uniquely by ABA. In-gel phosphorylation (a) and histone-phosphorylating activity (b) of the 58-kD kinase were assayed in the microsomes prepared from the tissues treated as described in A by ()-ABA (as a control) or other plant hormones auxin IAA, CA^sub 3^, synthetic cytokinin N-benzyl-6-aminopurine (6- BA), and brassinolide (BL). The numbers under lines indicate the applied concentrations as M. C, ABA stimulates ACPK1 kinase indirectly. Microsomes were prepared from nontreated tissues. In- gel phosphorylation (a) and histone-phosphorylating activity (b) of the 58-kD kinase were assayed in the microsomes in the presence or absence of 10 M ABA or Ca^sup 2+^ or both; - and + indicate the absence and presence, respectively. D, Effects of apoplasmic Ca^sup 2+^ and influx of Ca^sup 2+^ on ABA-induced stimulation of ACPK1 kinase. Berry tissues were preincubated in a medium (the equilibration buffer as detailed in "Materials and Methods") containing 5 mM BAPTA (a Ca^sup 2+^ chelator, lane3) or 5 M A23187(aCa^sup 2+^ ionophore, lanes 5-7) in the presence of Ca^sup 2+^ (lanes 3, 5, and 7), or 1 mM ECTA in the absence of Ca^sup 2+^ (lanes 4 and 6), and then ABA (10 M except for lane 7 where ABA was absent) was added to the medium for a further incubation of 1 h.The incubation in the presence of Ca^sup 2+^ but absence of ABA (lane 1) was taken as a control, and that in the presence of both Ca^sup 2+^ and ABA (lane 2) as another control. The microsomes were prepared from the treated tissues. In-gel autophosphorylation (a) and histone- phosphorylating activity (b) of the 58-kD kinase in the microsomes were analyzed in th\e presence of Ca^sup 2+^; - and + indicate the absence and presence, respectively.

The IEF/SDS-PAGE-purified 58-kD kinase was sequenced by tandem mass spectrometry (TSM). The data from the first sequencing indicated that the sequence of the purified 58-kD kinase matches that of some plant CDPKs registered in GenBank in their conserved kinase, junction, and CaM-like domains. The sequences of matched peptides were: REIQIMHH, GGELFDRI, HRDLKPENFL, DVWSAGVI, KQFSAMNK, DTDNSGTITFDE, and DQDNDGQIDYGEF (Fig. 5A, indicated by Peptide 1- Peptide 7), which formed the basis for the cloning of ACPKl by PCR. Using the designed degenerate primers that correspond to two of the matched sequences of the 58-kD kinase and also to the conversed domain of most CDPKs (-EIQIMHHL-and -KQFSAMNK-), a cDNA clone putatively coding for ACPKl (Fig. 5A) was obtained and registered in GenBank. A second sequencing of the purified 58-kD kinase showed that the sequence of 58-kD kinase matched the predicted amino acid sequence from the cDNA clone in 10 pieces particularly covering all the N-terminal variable, conserved kinase, junction and CaM-like domains, and variable C-terminal tail domain (Fig. 5A; see also Supplemental Fig. 1; Supplemental Table I), demonstrating that the isolated cDNA clone encodes the 58-kD ACPKl kinase. Based on the cDNA cloning, an ACPKl-specific antiserum against the N-terminal 40 amino acids covering the full variable domain and an adjacent portion of kinase domain (11 amino acids) of ACPKl was prepared and recognized the 58-kD IEF/SDS-PAGE-purified kinase as mentioned above, confirming that the ACPKl is encoded by the isolated cDNA clone.

Figure 4. Purification of the ACPK1 kinase. The portions at the 58-kD point (arrow) of SDS-polyacrylamide gels (A, Coomassie Brilliant Blue R-252-stained gels) were excised, crushed, and eluted for purification of 58-kD proteins as described in "Materials and Methods." The collected 58-kD proteins were subjected to SDS-PACE (B, Coomassie Brilliant Blue R-252-stained gels), immunorecognized by anti-soybean CDPKα serum (C), and detected by in-gel autophosphorylation assay (D). The purified 58-kD proteins were further analyzed for their Ca^sup 2+^-dependent electrophoretic mobility shift with the assays of SDS-PAGE (E, Coomassie Brilliant Blue R-252-stained gels), in-gel phosphorylation (F), histone- phosphorylating activity (G), and immunoblotting by anti-soybean CDPKα serum (H). IEF/SDS-PAGE two-dimensional electrophoresis of the purified 58-kD proteins displays two polypeptides at the 58- kD point but with different pl (I, Coomassie Brilliant Blue R-252- stained gels). One of the two polypeptides (lane c in sections I-L) is shown to be the 58-kD ABA-stimulated Ca^sup 2+^-dependent protein kinase with the assays of in-gel autophosphorylation (J), histone- phosphorylating activity (K), and immunoblotting by anti-ACPK1- N^sup 40^ serum (see "Materials and Methods"), an antibody specific to the 58-kD ACPKI kinase (L). The molecular masses of protein standards are shown at the left of the sections in kD. Lane a in section I shows the protein standards on gel; -Ca^sup 2+^ and +Ca^sup 2+^ in sections E to H indicate the absence and presence of Ca^sup 2+^, respectively.

Sequence Features of ACPKl

ACPKl is a full-length clone with an open reading frame (ORF) having 1,491 nucleotides and encoding a protein of 496 amino acids with a predicted molecular mass of 56 kD and a pi at 5.3. The predicted amino acid sequence exhibits the modular structure typical for CDPKs (Harmon et al., 2001) comprising an N-terminal variable region, a kinase domain, an autoinhibitory junction domain, and a CaM-like domain with four conserved EF-hand motifs implicated in Ca^sup 2+^ binding. Five potential myristoylation sites are present in the full-length amino acid sequence of ACPKl, and two of them are localized in the N-terminal region, suggesting possible membrane- associated nature of ACPKl via N-terminal myristoylation (Martin and Busconi, 2000; Lu and Hrabak, 2002; Rutschmann et al., 2002; Dammann et al., 2003). This modification by acylation may be partly responsible for the higher molecular mass of the biochemically identified 58-kD kinase than that of the predicted 56 kD. Comparison of the deduced ACPKl amino acid sequence with a soybean CDPK from seeds (GmS-CDPKa) shows a strong homology that extends all over the conserved domains (Fig. 5A). The highest identity is shared with potato (Solanum tuberosum) StCDPK (83%), broad bean VfCDPK (81%), and two soybean CDPKs (81%), GmCDPKα (soybean CDPKα) and GmS-CDPKa (Fig. 5B). The two soybean CDPKs should be orthologs with amino acid sequence identity of 99% between them (Fig. 5B). The other closer homologs of ACPKl are Arabidopsis AtCPK4, AtCPK11, and AtCPK12; soybean CDPKβ (GmCDPKβ); soybean seed CDPKb (GmS- CDPKβ); rice OsCDPK; and maize (Zea mays) ZmCDPK (Fig. 5B).

Figure 5. Alignment and phylogenetic relationships of ACPK1 and its related members of the CDPK family. We added the following prefixes to the names of some cited CDPKs in this paper for distinguishing them easily if their original names do not contain the prefixes: At for Arabidopsis, Gm and GmS (S indicates seed) for soybean, Os for rice, St for potato, Vf for broad bean, and Zm for maize. A, Alignment of deduced amino acid sequences of ACPKI and one of its closest homologs, soybean seed CDPKa (GmS-CDPKa), shows the presence of conserved features in Ser/Thr (S/T) protein kinase domain, junction domain, and CaM-like domain and also variable features in N-terminal domain and C-terminal tail. Numbers on the right column indicate numbers of amino acid residues in the predicted sequences. Gaps, indicated by dashes (-), were introduced to maximize alignment. Identical amino acid residues are indicated by white letters on a black background. The boundaries of the N- terminal variable domain, S/T kinase domain, junction domain, CaM- like domain, and C-terminal tail are indicated by arrows. Four EF- hand motifs (EFh) and five potential myristoylation sites (ML1 -5) are also shown. The matched sequences of seven peptides obtained in the first sequencing of the purified 58-kD ABA-stimulated CDPK (natural ACPK1 protein, see Fig. 7) by TMS are marked by Peptide 1 to Peptide 7 underneath the GmS-CDPKa sequence, respectively. The degenerate oligonucleotides corresponding to the conserved sequences of the peptide 1 in kinase domain (forward primer) and peptide 5 in junction domain (reverse primer) were used for cloning the putative ACPK1. The underlined sequences of ACPK1 are the matched sequences obtained in the second sequencing of the natural ACPK1 protein by TMS after the cDNA coding for the putative ACPK1 was isolated. B, Tree depicting phylogenetic relationships based on alignment of deduced amino acid sequences among top 11 of CDPKs having higher identity with ACPK1 and the other three ABA-responsive CDPKs, NtCDPKI (Yoonetal., 1999), and AtCPK10 and AtCPK 30 (Sheen, 1996). Percentages of amino acid sequence identity are also indicated. The corresponding accession numbers or locus tag numbers for the genes encoding these CDPKs are given in the section "Accession Numbers" in "Materials and Methods."

ACPK1 has a lower amino acid sequence identity (53%) with the three previously identified ABA-responsive CDPKs (Fig. 5B; Sheen, 1996; Yoon et al, 1999). The analysis of phylogenetic relationships among the closet homologs of ACPKl and the three ABA-responsive CDPKs based on amino acid sequence alignment shows that ACPKl may be classified as a different group from the three previously identified ABA-responsive CDPKs (Fig. 5B).

Genomic Southern Analysis of ACPJCl

Genomic Southern-blot analysis was done with a probe corresponding to full-length ACPKl cDNA. Genomic DNA was digested with BamHl, Kpnl, EcoRI, and HindIII restriction enzymes. The probe generated two hybridizing bands in BamHl or Kpnl digestion, and four bands in EcoRI or HindIII digestion (Fig. 6). Since there is a catalytic site within the cDNA sequence of ACPKl for both EcoRI and HmdIII, these results suggest that the grape genome contains two copies or more than one copy of ACPKl-related genes.

Characterization of the Products of the ACPKl Gene and Up- Regulation of Both Its Products and Transcripts by ABA

To confirm that ACPXl encodes an active CDPK whose expression and activity are stimulated by ABA, the full-length ACPKl (ACPK1^sup L^) and the truncated form specific to ACPKl (residues 1-40; ACPKl- N^sup 40^) containing only the N-terminal variable region and its adjacent portion in the kinase domain (as mentioned above) were expressed in Escherichia coli as fusion proteins. The ACPK1^sup L^ protein showed Ca^sup 2+^-dependent autophosphorylation and histone- phosphorylating activities (Fig. 7A), and also Ca^sup 2+^-dependent electrophoretic mobility shift in SDS-PAGE, shown by its autophosphorylation (Fig. TB). Both the autophosphorylation and histone-phosphorylating activities of the ACPK1^sup L^ were inhibited by W7 but not by W5, and these activities were apparently not stimulated by CaM (Fig. 7C). Two-dimensional, thin-layer chromatography of the hydrolyzed autophosphorylated ACPK1^sup L^ showed that ^sup 32^P-labeled spots corresponded to the positions of phospho-Ser and phospho-Thr (Fig. TD), suggesting that ACPKl possesses Ser/Thr kinase activity. Taken together, these data confirmed the nature of ACPK1 as an active CDPK.

Figure 6. Southern-blot analysis of the ACPKl gene. A 10-/Ag portion of grape genomic DNA was digested with BamHl, Kpn1, EcoR1 and HindIII, electrophoresed in a 0.8% agarose gel, and transferred onto a nylon membrane. The membrane was hybridized with the ^sup 32^P-labeled full-length cDNA of ACPKI. The DNA markers (M) are indicated at the left of the section in kb.

Figure 7. Characterization of the products of ACPK1 gene and up- regulation of both its products a\nd transcripts by ABA. A, The purified recombinant protein (10 g) of the full-length ACPK1 (ACPK1^sup L^) was subjected to SDS-PAGE and then assayed for its ingel autophosphorylation (a) and histone-phosphorylating activity (b) in the absence (-Ca^sup 2+^) or presence (+Ca^sup 2+^) of Ca^sup 2+^. B, Ca^sup 2+^-dependent electrophoretic mobility shift of the ACPK1^sup L^, shown by its in-gel autophosphorylation; -Ca^sup 2+^ and +Ca^sup 2+^ indicate the absence and presence of Ca^sup 2+^, respectively. C, Neither the in-gel autophosphorylation (a) nor histone-phosphorylating activity (b) is activated by CaM, whereas both activities are inhibited by CaM antagonist W7 but not by W5. CK, Control. D, Phosphoamino acid analysis of the autophosphorylated ACPK1^sup L^. γ-^sup 32^P-labeled ACPKI^sup L^ was hydrolyzed with HCI and subjected to two-dimensional thin-layer electrophoresis. The positions of phospho-Ser (p-Ser), phospho-Thr (p-Thr), and phospho-Tyr (p-Tyr) are indicated. E, Specificity of the anti-ACPK1-N^sup 40^ serum to ACPKI. The 58-kD ACPK1 immunosignal was detected in the microsomes of grape mesocarp (Grape) by both the anti-ACPK1^sup L^ (indicated by AL) and -ACPK1- N^sup 40^ (indicated by AN) sera. In the microsomes of apple flesh prepared with the same procedures as described in "Materials and Methods" for grape mesocarp, two immunosignals having molecular mass of about 60 and 67 kD, respectively, were detected by the antiserum against CaM-like domain of soybean CDPKa (indicated by AS) or anti- ACPK1^sup L^ serum (AL), but not by anti-ACPK1-N^sup 40^ serum (AN). F, ABA up-regulates gene expression, autophosphorylation, and kinase activity of ACPK1. Berry tissues were incubated in the medium containing different concentrations of ()-ABA from O to 80 m (indicated by ABA applied; unit: m), and after the incubation, the concentrations of ABA present in the treated tissues (ABA in tissue; unit: m) were determined by radioimmunoassay as described in "Materials and Methods." Total RNA and microsomes were extracted from the treated tissues, respectively. Total RNA (20 g), transferred on a nylon membrane after electrophoresed in an agarose gel, was hybridized with the ^sup 32^P-labeled cDNA fragment corresponding to the N-terminal variable domain of ACPK1 to detect specifically mRNA level of ACPK1 (ACPK1 mRNA). rRNA indicates loading control of the RNA samples stained with ethidium bromide. Total microsomal proteins (50 g) were immunoprecipitated with 3 g anti-ACPK1-N^sup 40^ serum protein. The immunoprecipitated proteins were subjected to SDS-PACE and then to the assays for detecting the levels of ACPK1 protein (ACPK1 protein, blotted with anti-ACPK1- N^sup 40^ serum), autophosphorylation (ACPK1 Autophosphorylation), or histone-phosphorylating kinase activity (ACPK1 Kinase activity) as described in "Materials and Methods."

The antisera against either ACPK1^sup L^ or ACPK1-N^sub 40^ were prepared in rabbit. Both the anti-ACPK1^sup L^ and -ACPK1-N^sup 40^ sera recognized the ACPK1 signal from grape berry (Fig. 7E). The anti-ACPK1^sup L^ serum can recognize other putative members of CDPK family as both this antiserum and the antiserum against the CaM- like domain of soybean CDPKa detected from apple fruit (Mains domestica) two immunosignals having molecular mass of about 60 and 67 kD, respectively, but the anti-ACPK1-N^sup 40^ serum could not (Fig. 7E). This showed that the anti-ACPK1-N^sup 40^ serum is specific to ACPK1. The same in vivo incubation assays as used above when testing the ABA-induced 58-kD kinase stimulation were performed to confirm the action of ABA on CDPK1, but the specifically immunoprecipitated ACPKl protein by anti-ACPK1-N^sup 40^ serum was used for the phosphorylation assays, and also ACPK1 mRNA was probed with ^sup 32^P-labeled cDNA fragment corresponding to the variable domain of ACPK1 to assess the ACPK1 expression. The results showed that ABA stimulated the ACPK1 kinase in all levels of the gene transcription, mRNA translation and kinase autophosphorylation, and catalytic activities in a dose-dependent manner (Fig. 7F). It is noteworthy that in the immuno- and enzyme assays of ACPK1, the immunoprecipitated proteins with the anti-ACPK1L serum gave substantially the same results as those with the ACPK1-specific anti- ACPK1-N^sup 40^ serum (data not shown), suggesting that ACPK1 may be a major CDPK having a predominant abundance in grape mesocarp as was observed above (Fig. 2A). All these assays based on molecular cloning of ACPK1 consistently confirmed the above-mentioned results from the biochemical approaches (Fig. 2A).

ACPK1 Expression Is Fruit Specific and Altered Accordantly with Endogenous ABA Concentrations during Fruit Development

Northern-blot analysis showed that ACPK1 expressed only in the grape berry covering two essential portions of the fleshy mesocarp and seeds with much higher abundance of ACPK1 mRNA in the mesocarp than in seeds, but ACPK1 mRNA was undetectable in roots, young stems, or leaves (Fig. 8A). The amounts of mRNA and protein of ACPK1 as well as its autophosphorylation and histone-phosphorylating activities were all altered with changes in the endogenous ABA concentrations in the grape mesocarp during fruit development at different stages: I (early growth stage), II (first rapid growth phase), III (lag growth phase), and FV (second rapid growth phase or ripening stage; Fig. 8B).

ACPK1 Is Localized in Both Plasma Membranes and Chloroplasts

Transient expression of the green fluorescent protein (GFP)- ACPK1 fusion protein in chloroplast-free epidermis cells of onion (Allium cepa) showed that the fusion protein was targeted to the cell periphery, likely the plasma membrane (Fig. 9A, 1 and 2). A further assay of the transient expression of the fusion protein in Arabidopsis protoplasts showed that the fusion protein was localized to both plasma membranes and chloroplasts (Fig. 9B, 14). The localization of the enzyme in grape mesocarp by immunogold technique with the specific antiserum against ACPK1-N40 showed clearly the double-localized ACPK1 in both plasma membranes and chloroplasts (Fig. 9, C-E). No substantial immunosignal was detected in the controls without antiserum or using the preimmune serum instead of antiserum (data not shown), indicating the specificity and reliability of the immunolocalization.

The ACPK1-specific anti-ACPK1-N^sup 40^ serum recognized the ACPK1 in the total membrane proteins (microsomes), purified plasma membrane, and endomembrane fractions, but not in the soluble fraction, whereas the antiserum against the full-length ACPK1 (ACPK1^sup L^) recognized in these membrane fractions not only the 58-kD kinase ACPK1 but also a signal having higher molecular mass than ACPK1 (Fig. 9F). This immunosignal may be the 64-kD phosphoprotein detected in the in vitro phosphorylation of the total membrane proteins (Fig. 2A7 a), which may be a heavily modified form of 58-kD ACPK1 or a different protein kinase whose content is probably too low to be detectable in the crude extracts of membrane fractions (microsomes) but was sufficiently enriched in the purified membranes to be immunodetected by the antiACPK1^sup L^ serum. These data from irnmunoblotting further demonstrated the specificity of the anti-ACPK1-N^sup 40^ serum to ACPK1 and confirmed the plasma membraneassociated nature of ACPK1. The signal of ACPK1 immunodetected in the endomembrane fraction (Fig. 9F) should be the chloroplast-associated ACPK1. It is finally noteworthy that neither the transmembrane domains nor signal to target ACPK1 to plasma membranes or plastids were predicated for this kinase with several programs (data not shown). So, ACPK1 may be myristoylated in its N- terminal region for targeting to the membranes (Dammann et al., 2003 and refs. therein).

Figure 8. Expression of ACPK1 in different tissues and during fruit development. A, Expression of ACPK1 gene in different tissues. Total RNA (20 g) from root (R), young stem (St), leaf (L), flesh (F), or seed (S) of berry, transferred on a nylon membrane after electrophoresed in an agarose gel, was probed with the ^sup 32^P- labeled cDNA fragment corresponding to the N-terminal variable domain of ACPK1 to detect specifically mRNA level of ACPK1 (ACPK1 mRNA). rRNA indicates the RNA samples stained with ethidium bromide. B, Expression of ACPK1 during berry development. Total RNA from berry tissues was probed as in A with the 32P-labeled cDNA fragment corresponding to the N-terminal variable domain of ACPK1 to detect specifically mRNA level of ACPK1 (ACPK1 mRNA). rRNA, RNA samples stained with ethidium bromide. Total microsomal proteins (100 g) extracted from berry tissues were immunoprecipitated with 5 g anti- ACPK1-N40 serum protein. The immunoprecipitated proteins were subjected to SDS-PACE and then to the assays for detecting the levels of ACPK1 protein (ACPK1 protein, blotted with anti-ACPK1- N^sup 40^ serum), autophosphorylation (ACPK1 Autophosphorylation), or histone-phosphorylating kinase activity (ACPK1 Kinase activity) as described in "Materials and Methods." The concentrations of ABA in the berry tissues (columnar figures, ABA content in berry) were determined by radioimmunoassay (values in the columnar figures are means of five replications SE). I, II, III, and IV indicate the sampling dates that are the early stage of berry growth (10 d after anthesis, stage I), first rapid growth phase (stage II, 10th-30th d after anthesis), lag growth phase (stage III, 30th-50th d after anthesis), and second rapid growth phase or ripening stage (stage IV, 50th-80th d after anthesis).

ACPK1 Positively Regulates PM H+-ATPaSe in Vitro

We previously reported that ABA activates plasma membrane (PM H+- ATPase) in apple fruit (Peng et al., 2003). The similar results were obtained in grape berry (data not shown). To assess whether ACPK1 is possibly involved in this ABA-induced event, we first showed \the presence of a 97-kD plasma membrane-localized H+-ATPase in grape mesocarp by both immunogold subcellular localization (Fig. 10, A and B) and immunoblotting of the subcellular fractions (Fig. 10C). This plasma membrane colocalization of the ACPKl kinase with PM H+- ATPase may facilitate possible interactions between the two proteins. A preincubation of the microsomal proteins with the recombinant full-length ACPK1 kinase ACPK1^sup L^ significantly enhanced the activity of PM H+-ATPase in an ACPK1^sup L^ dose- dependent manner with the activity being doubled in the peaks of the activation (Fig. 10D). The controls of the precipitation in the ACPK1^sup L^ free-or denatured ACPK1^sup L^-containing medium showed no significant effect on PM H+-ATPase (Fig. 10D), indicating that the ACPK1-related effects were specific. Removal of free Ca^sup 2+^ from the preincubation medium abolished the PM H+-ATPase activation by ACPK1 (Fig. 10D), showing that the PM H+-ATPase activation was Ca^sup 2+^ dependent. This is consistent with the nature of the action of a CDPK.

It has been reported that the posttranslational modification of PM H+-ATPase involves the C-terminal auto-inhibitory domain of the enzyme, and cleavage of the C-terminal autoinhibitory domain by mild trypsin treatment or genetic deletion activates the enzyme with its pH optimum shifted to more alkaline (Palmgren et al., 1990, 1991; Rasi-Caldognov et al., 1993; Sekler et al., 1994; Regenberg et al., 1995). In this experiment, the activation of PM H+-ATPase by ACPK1 was enhanced when the medium pH was shifted from 6.5 to 7.0 (Fig. 10E), and a mild treatment of the microsomal proteins by trypsin also induced a significant PM H+-ATPase activation, which was comparable to the ACPKl treatment (Fig. 10F). The ACPK1 treatment did not induce any additional activating effect on PM H+-ATPase in the trypsin-treated microsomes (Fig. 10F). These data suggest that ACPK1 regulates PM H+-ATPase by acting on its C-terminal autoinhibitory domain.

Figure 9. Subcellular localization of ACPK1. A1 and A2, Transient expression of ACPK1-GFP fusion protein in the epidermis cells of onion. The fusion protein is present in the cell outlines, shown by the ACPK1-CFP fluorescence image (A1) and merged image (A2) of laser- scanning confocal microscopy. B1 to B4, Transient expression of ACPK1-GFP fusion protein in the protoplast of Arabidopsis shows that the fusion protein is localized in both the plasma membranes and protoplasts. The laser-scanning confocal microscopy images are the ACPK1-CFP fluorescence (B1), chlorophyll autofluorescence (B2), bright field (83), and merged (B4) images. C, Immunogold labeling of ACPK1 in the grape berry tissue with anti-ACPK1-N^sup 40^ serum (antiserum against the N-terminal 40 amino acids of ACPK1) specific to ACPK1. Immunogold particles are distributed in plasma membranes (PM) and chloroplast (ChI) but not in other cellular compartments such as cell wall (CW), cytosol (Cyt), and vacuole (V). The arrows indicate the presence of the immunogold particles along the plasma membranes. D, An amplified portion of the image C, displaying clearly the chloroplast localization of the immunogold particles representing ACPKI. E, Blow up of the boxed-in area in C. Arrows indicate the immunogold particles localized to plasma membranes. F, immunoblotting in different subcellular fractions with either anti- ACPK1^sup L^ serum (antiserum against the full-length ACPK1; a) or anti-ACPK1-N^sup 40^ serum (b). The fractions of soluble portion (lanes S, 40 g protein), microsomes (lanes M, 20 g protein), plasma membranes (lanes PM, 20 g protein), and endomembranes (lanes M, 20 g protein) were separated on a 12% SDS-polyacrylamide gel and then subjected to immunoblotting. The molecular masses of protein standards are shown at the left of the sections in kD.

To further assess the regulation of PM H+-ATPase by ACPK1, we partially purified PM H+-ATPaSe, which was tested by SDS-PAGE and immunoblotting (Fig. 10G, lanes 1 and 2). The assay of in vitro phosphorylation of the partially purified fraction in an ACPK1-free medium showed no phosphorylated protein band in the SDS-PAGE gel (Fig. 10G, lane 3), indicating that there was no phosphorylating activity in this partially purified fraction. Using this fraction, we showed that the PM H+-ATPase was phosphorylated by ACPK1 (Fig. 10G, lane 5). Removal of free Ca^sup 2+^ from the in vitro phosphorylation medium completely abolished the PM H+-ATPase phosphorylation (Fig. 10G, lane 4), indicating the Ca^sup 2+^ dependence of this phosphorylation event. These data suggest that ACPKl activates PM H+-ATPase by phosphorylating the enzyme.

Figure 10. Presence of plasma membrane-localized H+-ATPase in grape berry and positive regulation of its activity by ACPK1. A, Immunogold labeling of PM H+-ATPase in the grape mesocarp with antiserum against Arabidopsis PM H+-ATPaSe AHA3, showing numbers of immunogold particles along the plasma membranes. Cyt, Cytosol; CW, cell wall; V, vacuole. B, Blow up of the boxed-in area in A, displaying clearly the presence of numbers of H+-ATPase molecules visualized by gold particles along the plasma membranes. C, A 97-kD (indicated by arrow) immunosignal was detected on SDS- polyacrylamide gel by antiserum against Arabidopsis PM H+-ATPaSe AHA3 in both the microsomes (M) and plasma membrane fraction (PM), but not in either the endomembrane (EM) or soluble fraction (S). D, ACPK1 activates PM H+-ATPase in a dose-dependent manner. The recombinant ACPK1^sup L^ protein was added to the microsomes in the presence (M+ACPK1+Ca^sup 2+^) or absence (M+ACPK1+ECTA) of free Ca^sup 2+^ for a preincubation before measuring PM H+-ATPase activities as described in "Materials and Methods." The assay with the ACPK1^sup L^-free microsomes preincubated in the presence (M+Ca^sup 2+^, Control 1) or absence (M + EGTA) of free Ca^sup 2+^ were taken as controls. The same amount of the microsomal proteins denatured by boiling was used instead of the ACPK1^sup L^ protein as another control (M + denatured ACPK1 + Ca^sup 2+^, Control 2). E, The same assay as described above in M + ACPK1 + Ca^sup ++^ in D, but with the PM H+-ATPase activities measured in different medium pH. The microsomes only (M) or the microsomes together with ACPK1^sup L^ protein (M + ACPK1) was used for the assay. F, The same assay was done as described above in M + ACPK1 + Ca^sup ++^ in D, but the PM H+-ATPase activities were measured at pH 7.0, and the trypsin-treated microsomes only (ts-M) or the trypsin-treated microsomes together with ACPK1^sup L^ protein (ts-M + ACPK1) was used for the assay where the nontreated microsomes (M) or the nontreated microsomes in the presence of ACPKI^sup L^ protein was used as controls. G, ACPK1 phosphorylates PM H+-ATPase. The molecular masses of protein standards are shown at the left of the sections in kD. The partially purified PM H+-ATPaSe fraction was subjected to SDS-PAGE (lane 1), and an immunosignal of about 97 kD (lane 2) was recognized from the separated protein bands by the antiserum against Arabidopsis PM H+-ATPase AHA3, which is the grape PM H+-ATPase. The in vitro phosphorylation in the medium containing the partially purified PM H+-ATPase fraction and ACPK1^sup L^ protein showed that the PM H+-ATPase was phosphorylated (lane 5, indicated by ATPase with arrow). ACPK1^sup L^ with arrow in lane 5 indicates the autophosphorylated GST-tagged ACPK1^sup L^ protein having a molecular mass of about 85 kD. The in vitro phosphorylation assay of the partially purified PM H+-ATPase fraction in the ACPKI^sup L^-free medium (lane 3) was taken as a control, and the same assay as described for lane 5 but in the absence of free Ca^sup 2+^ was taken as another control (lane 4).

DISCUSSION

Identification of an ABA-Stimulated CDPK, ACPK1, from Grape Berry

Numerous studies have demonstrated that calcium is involved in ABA signal transduction (for review, see Finkelstein et al., 2002; Himmelbach et al., 2003; Fan et al., 2004). However, many of the Ca^sup 2+^-binding sensory proteins as the components of the ABA- signaling pathway remain to be elucidated. In this study, we identified and purified a 58-kD ABA-stimulated CDPK from the mesocarp of grape berries (Figs. 1-4), designated ACPK1. Based on sequencing of the two-dimensional electrophoresis-purified ACPKl, we isolated a cDNA clone encoding a putative CDPK (Fig. 5). This cDNA clone was demonstrated to encode ACPK1 essentially by two mutually linked approaches. One is the second time of sequencing of the purified natural ACPK1 protein, showing that the predicted amino acid sequence from the cDNA clone matches the sequence of the natural ACPK1 protein (Fig. 5, see also Supplemental Fig. 1; Supplemental Table I), and another is the assay of immunoblotting where the purified 58-kD ACPK1 kinase was specifically immunorecognized with the anti-ACPK1-N^sup 40^ serum specific to the full-length putative ACPK1 cDNA-encoded protein (Fig. 4). The analysis of the recombinant full-length protein expressed in E. coli further demonstrated that the isolated ACPK1 cDNA clone indeed encodes an active number of CDPK family (Fig. 7, A-D), and the further assays based on the molecular cloning of ACPKl revealed that ABA stimulates the ACPKl kinase in both protein (amounts and activities) and mRNA levels (Fig. 7F), confirming the ABA-induced effects of stimulating the ACPK1 kinase.

Biological Significance of ABA-Induced ACPK1 Kinase Stimulation

The features of the ACPK1 kinase stimulation by ABA are of importance to possible biological functions of ACPK1. The dose dependence of the ACPK1 stimulation induced by a physiological range of ABA concentrations (<3.7 M) within the tissues (Figs. 3 and 7) is suggestive of action of endogenous ABA on ACPK1. For this action of endogenous ABA, further supporting evidence was provided by correlating the developmental changes of grape berries intheir ABA concentrations with their ACPK1 expression and enzyme activities (Fig. 8B). The dose dependence of the ABA-induced ACPK1 stimulation essentially exhibited a saturation-like phenomenon with a slight decline of the ABA-induced effects with the overphysiological higher levels of ABA (>3.7 M) resulting from excessively applied ABA (20- 80 M; Figs. 2 and 7). This phenomenon of saturation, difficult to explain, may be due to a possible CDPK turnover mechanism involving a degradation pathway induced by excessive expression of the enzyme as observed in tobacco NtCDPK2 by Romeis et al. (2001) or probably to a decline of ABA-signaling pathway resulting from a typical saturation-binding nature of ABA to its putative receptor in this fruit (Zhang et al., 1999).

ABA functions in mediating both rapid responses such as stomatal apertures and slow events involving gene expression to regulate growth and development (Finkelstein et al., 2002; Himmelbach et al., 2003; Fan et al., 2004). The rapid stimulation of the ACPK1 kinase in 15 min by ABA without increase of the ACPK1 protein (Fig. 2B) suggests involvement of ACPK1 in mediating some rapid ABA- responsive events, while a slow and steady state of the ABA- stimulated ACPK1 activities paralleling de novo ACPK1 synthesis (Fig. 2B) implicates roles of ACPK1 in regulating developmentrelated ABA-signaling pathway. Although, in the rapid phase of the ACPK1 stimulation by ABA, ABA could probably act on posttranslational modification of the ACPK1 protein, the ABA signal must require a living cell-dependent mechanism to be transduced to its target ACPK1, indirectly acting on the enzyme (Fig. 3C). The pharmacological assays showed that influx of apoplasmic Ca^sup 2+^ into cytoplasm may be involved in this event of ABA signal transduction (Fig. 3B) by promoting direct binding of Ca^sup 2+^ to ACPK1 or activating a more complex mechanism to stimulate ACPK1. However, an artificial influx of Ca^sup 2+^ created by Ca^sup 2+^ ionophore A23187 could not enhance the ABA-induced effects, and it could not otherwise replace ABA to stimulate ACPK1 (Fig. 3D), implying that more complex signal cascades than a single Ca^sup 2+^ signal may be involved in the ACPK1 stimulation.

Whether the effect of ABA on ACPK1 is specific to the physiological active form of ABA is fundamental in establishing the function of ABA. We showed that the ABA-induced ACPK1 stimulation is stereo specific to physiological active (+)-ABA, while other inactive structurally similar isomers (-)-ABA or trans-ABA are ineffective (Fig. 3A). Also, other phytohormones were shown to be ineffective in the ACPK1 stimulation (Fig. 3B). These findings of stereo specificity and exclusivity of ABA in stimulating ACPK1 support further the specific biological relevance of the ABA- induced ACPK1 stimulation.

It has been documented that PM H+-ATPase could be phosphorylated in its C terminus in a calciumdependent manner (Schaller and Sussman, 1988; Camoni et al., 1998; Lino et al, 1998; De Nisi et al., 1999; Rutschmann et al., 2002), but complex positive or negative regulatory mechanisms to control its activity may occur (Lino et al., 1998; De Nisi et al., 1999; Kinoshita and Shimazaki, 1999; Sze et al., 1999; Morsomme and Boutry, 2000; Palmgren, 2001). We showed that ACPK1 is a plasmalemma and chloroplast/plastid double- localized protein (Fig. 9), suggesting that it may function both in the plasma membranes and intracellularly. It is particularly interesting that the activity of plasma membrane-colocalized H+- ATPase is up-regulated by ACPK1 in vitro most likely through phosphorylation of the PM H+-ATPase by ACPK1 (Fig. 10), showing that ACPK1 may be positively involved in ABA-signaling pathway where ABA stimulates PM H+-ATPase in grape berry.

What possible functions may ACPKl have in regulating fruit development? ABA has been believed to regulate development of fleshy fruits, especially grape berry, essentially by promoting assimilate accumulation and triggering fruit ripening (Coombe, 1992; Rock and Quatrano, 1995; Wayne and John, 1996; Opaskornkul et al., 1999; Peng et al., 2003; Pan et al., 2005). The activation of plasma membrane ATPase by ACPKl (Fig. 10) suggests that ACPKl may be involved in ABA- stimulated assimilate accumulation by improving energization of the fruit cells to promote plasmalemma H+-ATPase-powered active uptake (Palmgren, 2001). The expression of ACPK1 exclusively in grape berry including the mesocarp and seeds (Fig. 8) appears to be associated with its possible specificity of biological function to this fruit. The expression pattern of ACPK1 during fruit development (Fig. 8) seems to be of particular significance. The lag growth phase stage III is a critical period when intracellular structural changes and active metabolism occur in spite of relatively lag appearance of berry growth, and this has been believed to prepare the onset of ripening (stage IV, see Fig. 8) during which nearly the totality of storage assimilates accumulates and fruit quality and yield are formed (Coombe, 1992; Zhang et al., 1997). The higher levels of ACPK1 in both its protein and mRNA amounts and enzyme activities in accordance with higher ABA concentrations (Fig. 8B) may be connected with its possible involvement in regulating postripening, triggering onset of ripening, and promoting rapid accumulation of carbon reserves during these two phases. It is difficult, however, to explain the high level of ACPK1 at stage I, an early stage after anthesis when berries begin growing, as the biological significance of the peak level of ABA at this stage remains currently unexplainable. Getting an insight into how ACPK1 works in fruit development will be of interest to understand ABA signal transduction in plants.

MATERIALS AND METHODS

Plant Materials and Chemicals

Grape (Vitis vinifera Vitis Lubrusca L. cv Kyoho) berries and other organs were sampled from a commercial vineyard in the western suburb of Beijing. We divided the sampling period of grape berries into four stages, i.e. early stage of berry growth (10

Source: Plant Physiology

More News in this Category