Arabidopsis Phospholipase Dα1 Interacts with the Heterotrimeric G-protein α-Subunit through a Motif Analogous to the DRY Motif in G-protein-coupled Receptors*

Phospholipase D (PLD) and heterotrimeric G-protein both play important, diverse roles in cellular regulation and signal transduction. Here we have determined the physical interaction between plant PLD and the only canonical α-subunit (Gα) of the G-protein in Arabidopsis thaliana and the molecular basis for the interaction. PLDα1 expressed in either Escherichia coli or Arabidopsis was co-precipitated with Gα. PLDα1 contains a sequence motif analogous to the Gα-interacting DRY motif normally conserved in G-protein-coupled receptors. Mutation of the central Lys residue PLDK564A of this motif abolished the PLDα1-Gα binding, whereas mutation of the two flanking residues PLDE563A and PLDF565A decreased the binding. Addition of Gα to PLDα1 inhibited PLDα1 activity, whereas the PLDK564A mutation that disrupted the Gα-PLDα1 binding abolished the inhibition. GTP relieved the Gα inhibition of PLDα1 activity and also inhibited the binding between PLDα1 and Gα. Meanwhile, the PLDα1-Gα interaction stimulated the intrinsic GTPase activity of Gα. Therefore, these results have demonstrated the direct binding between Gα and PLDα1, identified the DRY motif on PLDα1 as the site for the interaction, and indicated that the interaction modulates reciprocally the activities of PLDα1 and Gα.

Phospholipase D (PLD), 1 which hydrolyzes phospholipids to phosphatidic acid and a head group, plays diverse roles in cellular metabolism and regulation. Plant PLD comprises a family of enzymes with different regulatory properties (1). Several Arabidopsis PLDs have been shown to display different requirements for Ca 2ϩ , polyphosphoinositides, and free fatty acids as well as varied substrate selectivity. Arabidopsis has at least 12 PLDs, of which PLD␣1 is most prevalent and responsible for the common plant PLD activity (1,2). PLD␣1 produces a majority of the phosphatic acid under several stress conditions, such as freezing and wounding (3,4). Suppression of PLD␣1 delayed abscisic acid (ABA)-promoted senescence (5), decreased wound-induced accumulation of jasmonic acid (4) and reactive oxygen generation (6), and increased freezing tolerance and water loss (3,7). These results show that the common plant PLD has multifaceted functions, including roles in metabolism and cell signaling, dependent on the nature and severity of the stress conditions.
The G␣ subunit of heterotrimeric G-proteins plays an important role in signal transduction. In animal systems, G␣ interacts with the upstream transmembrane G-protein-coupled receptors (GPCRs) and with the ␤-subunit. The binding of a ligand to a cognate receptor promotes the exchange of GDP for GTP on G␣, and the GTP-bound G␣ activates the downstream effector proteins (8). In addition, G␣ may interact with nonreceptor proteins to mediate signaling (9). Mammalian cells contain a number of G␣s that mediate many distinctive cellular functions (8). In contrast, the number of G␣s is very limited in plants; Arabidopsis has only a single canonical G␣ gene, GPA1 (10,11). Arabidopsis G␣-null mutants are impaired in several cellular processes, including cell division, certain ABA signaling steps in guard cells, and germination behaviors in response to glucose and hormones (12)(13)(14). These changes indicate that G␣ is involved in multiple cellular processes in plants and thus may interact with multiple effector proteins for different functions. The only protein reported to interact with G␣ in Arabidopsis is the cupin-domain protein AtPirin, which interacts with a CCAAT box-binding transcriptional factor (15). The mechanism by which plant G␣ interacts with its targets and the function of the interaction are unknown.
A role of G-proteins in regulating plant PLD has been proposed recently. Most of the studies have involved the use of potential G-protein activators and inhibitors, such as toxins, nonhydrolyzable guanine nucleotide analogues, and alcohols (16,17). In barley aleurone cells, a PLD␣-like activity has been suggested to be associated with a G-protein on the plasma membrane to mediate ABA signaling (17). Co-incubation of bacterially expressed tobacco PLD␣ with G␣ decreased the PLD␣ activity (18). It has also been shown that ablation of either G␣ or PLD␣1 in Arabidopsis affects ABA-mediated stomatal movement and increases plant water loss (7,12). These observations raise intriguing questions of whether PLD␣1 and G␣ directly interact with one another and, if so, what the molecular bases are for the interaction. Here we show that PLD␣1 binds to G␣ through a motif analogous to the DRY motif present in many GPCRs and that the binding modulates the activity of PLD␣1 and G␣.

EXPERIMENTAL PROCEDURES
Expression and Purification of Active Plant PLD␣ in Escherichia coli-Arabidopsis PLD␣1 cDNA was cloned previously in pBlue SK (19). The 2.4-kb PLD␣1 cDNA was amplified by polymerase chain reaction and cloned into the pGEM T-easy vector. The forward and reverse primers were 5Ј-GCGGATCCATGGCGCAGCATCTGTTGCACG and * This work was supported by grants from the National Science Foundation and the United States Department of Agriculture. This is contribution No. 04-112-J from the Kansas Agricultural Experimental Station. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
5Ј-CGGAGCTCTTAGGTTGTAAGGATTGGAGGC, respectively, and the bold letters mark the inserted BamHI and SacI sites. The PLD␣1 cDNA insert was digested with BamHI and SacI and ligated into the pET28(ϩ)a vector to produce PLD␣1 with 6 histidine residues fused at the N terminus. The recombinant plasmid was transformed into E. coli BL21(DE3). Expression of PLD␣1 was induced by 0.1 mM isopropyl-1thio-␤-D-galactopyranoside at room temperature for 12 h. After induction, bacteria were precipitated and lysed by sonication in phosphatebuffered saline plus 2 mM phenylmethanesulfonyl fluoride. Bacterial lysate was centrifuged at 12,000 ϫ g for 10 min. The resulting supernatant was incubated with Ni-affinity agarose beads for 1 h. The beads were pelleted by centrifugation at 500 ϫ g at 4°C and washed three times with a washing buffer containing 20 mM Tris-HCl, 0.5 M NaCl, and 20 mM imidazole at pH 8.0. PLD␣1 bound to the resin was eluted with an elution buffer with 1 M imidazole, Tris-HCl, and 0.5 M NaCl. In some cases, PLD␣1 bound to resins was used directly for assaying PLD␣1 activity. Concentration of the purified PLD␣1 was measured by the Bradford method using a Bio-Rad kit with bovine serum albumin (BSA) as a standard. The protein was stored in 20% glycerol at Ϫ20°C until use.
Expression and Purification of Active Plant G␣ in E. coli-An Arabidopsis G␣ cDNA clone was generously provided by Dr. Hong Ma (Pennsylvania State University, University Park, PA) (20). The 1.3-kb G␣ cDNA was amplified by PCR and cloned into the pGEM T-easy vector. The forward and reverse primers were 5Ј-GAATTCATGGGCT-TACTCTGCAGTAGAA and 5Ј-CTCGAGTCATAAAAGGCCAGCCTC-CAGTA, respectively, and the bold letters mark the inserted EcoRI and XhoI sites. The cDNA insert was digested with XhoI and EcoRI and ligated into pGEX-4T to produce G␣ with glutathione S-transferase (GST) fused at the N terminus. Junction of the GST-GPA1 fusion and full-length cDNA of G␣ was confirmed by sequencing. The recombinant plasmid was transformed into E. coli BL21(DE3) to express the GSTfused G␣ according to the procedure described previously (21).
Site-directed Mutagenesis of the DRY Motif in PLD␣1-Mutagenesis of the three codons in the putative DRY motif of PLD␣1 was performed using the QuikChange XL site-directed mutagenesis kit according to the manufacturer's instructions (Stratagene, La Jolla, CA). The three complementary reverse primers for mutating E563A, K564A, and F565A were as follows: 5Ј-GATTGAGAAAGGAGCGAAGTTCAGGGT-CTATGTTGTGG, 5Ј-GATTGAGAAAGGAGAGGCGTTCAGGGTCTAT-GTTGTGG, and 5Ј-GAGAAAGGAGAGAAGGCCAGGGTCTATGTTG-TGG, respectively. The wild-type PLD␣1 cDNA in pET28(ϩ)a served as the PCR template. The mutant clones were verified by DNA sequencing and were then transformed into BL21(DE3) for protein expression. Expression and purification of the mutated PLD␣1 proteins in E. coli were performed with the same procedure as that for the original PLD␣1.
Protein Extraction from Arabidopsis-Fully expanded leaves of Arabidopsis thaliana (Columbia) plants were frozen with liquid N 2 and homogenized with a buffer containing 50 mM Tris-HCl (pH 7.5), 80 mM KCl, 2 mM EDTA, 5 mM dithiothreitol, and 2 mM phenylmethanesulfonyl fluoride. The homogenate was centrifuged at 10,000 ϫ g at 4°C for 15 min, and the resultant supernatant was used as protein extracts for co-precipitation. Protein concentration in the supernatant was measured by the Bradford method using a Bio-Rad kit with BSA as a standard.
PLD␣ Pulldown by GST-G␣-Agarose Beads and Immunoblotting-Purified GST-G␣-agarose beads were incubated with plant protein extracts; bacterially expressed PLD␣1; or PLD␣1 mutants in a co-precipitation buffer containing 50 mM Tris-HCl (pH 8.0), 200 mM NaCl, 2 mM MgCl 2 , 5 mM CaCl 2 , and protease inhibitors (5 g each of aprotinin, leupeptin, and antipain) in a total volume of 150 l. After incubation at 4°C for 3 h, beads were pelleted by centrifugation at 500 ϫ g and washed three times with the co-precipitation buffer containing 0.01% Triton X-100 at 4°C. Pulldown beads were assayed for PLD␣1 activity and/or were subjected to 10% SDS-PAGE, followed by immunoblotting with antibodies specific to PLD␣1. PLD␣1 polyclonal antibodies were raised in rabbit against the 13 C-terminal amino acid residues as described previously (5). PLD bands on blots were visualized by staining alkaline phosphatase activity conjugated to a second antibody (goat against rabbit immunoglobulin). The procedures for SDS-PAGE, protein blotting, and band detection were described previously (5).
PLD␣1 Activity Assays-PLD activity was assayed by using 1,2dipalmitoyl-3-phosphatidyl-[methyl-3 H]choline as a substrate. Two different methods were used. One is the common plant PLD activity assay FIG. 1. Expression and purification of Arabidopsis G␣ and PLD␣1 in E. coli. A, purified GST-G␣, His-PLD␣1, and GST resolved by a 10% SDS-PAGE gel and stained with Coomassie Blue. GST and His tag were fused to the N terminus of Arabidopsis G␣ and PLD␣1, respectively. B, GTPase activity of expressed G␣. Hydrolysis of GTP was assayed by measuring P i release as a function of time and amounts of purified GST-G␣.

FIG. 2. Binding of G␣ with PLD␣1 expressed in E. coli.
A, co-precipitation of PLD␣1 with GST-G␣. Top, immunoblotting of PLD␣1 that was co-precipitated with GST-G␣. GST-G␣ (0.05 mol) was used to precipitate bacterially expressed His-PLD␣1 (0.15 mol). GSTϩPLD␣1, PLD␣1 precipitated with GST-bound beads (molar equivalent to the GST-G␣-bound beads). Total PLD␣1, the starting PLD␣ used for precipitation. G␣ϩpET, bacterial lysate harboring an empty pET vector precipitated by G␣. Bottom, PLD␣1 activity in the co-precipitates corresponding to those in the immunoblot above. PLD activity was expressed as nmol of choline released/min/mg of starting proteins. B, effect of different molar ratios of G␣ to PLD␣1 on the binding. Purified GST-G␣ (0.05 mol) was mixed with 0, 0.025, 0.05, and 0.1 mol of purified His-PLD␣1 in a binding buffer followed by pulldown with glutathione beads. Top, immunoblotting of PLD␣1 that was co-precipitated with GST-G␣. Middle, Coomassie Blue staining of GST-G␣ in the precipitates resolved by 10% SDS-PAGE. Bottom, PLD␣1 activity in the G␣ precipitates, expressed as nmol of choline released/min from the same amount of precipitates.
involving the use of phosphatidylcholine (PC), 25 mM Ca 2ϩ , and 0.3 mM SDS (22). This assay is economical and specific for PLD␣1 and thus was used for most PLD␣1 assays unless otherwise stated. When the effect of G␣ binding on PLD activity was measured, however, the presence of SDS and high levels of Ca 2ϩ might have interfered with the PLD-G␣ interaction. Thus, a second PLD assay was used, which contained mixed lipid vesicles composed of 3.6 mol of phosphatidylethanolamine, 0.32 mol of PIP 2 , and 0.22 mol of PC in the presence of 100 M CaCl 2 . This assay, referred to as the PC-PIP 2 method, was performed according to a procedure described previously (23).
Assay of GTPase Activity of E. coli-expressed GPA1-GTPase activity assays were performed using the EnzChek® phosphate assay kit (Molecular Probes, Eugene, OR). The rate of phosphate released from GTP by G␣ was monitored spectrophotometrically at A 360 nm according to the manufacturer's instructions. Briefly, the assay mixture contained different amounts of GST-G␣, 10 l of 0.2 mM GTP or GTP at indicated concentrations, 0.2 ml of 2-amino-6-mercapto-7-methylpurine ribonucleotide, 10 l (1 unit) of purine nucleotide phosphorylase, and 0.72 ml of reaction buffer (50 mM Tris-HCl, pH 7.6, and 10 mM NaCl) in a total volume of 1 ml in a cuvette. After mixing, the reaction was started with addition of 10 mM MgCl 2 . Absorbance changes were recorded every 5 min. The amount of phosphate released from GTP was calculated based on a standard curve that was determined with known amounts of KPO 4 in the same manner as described above.

G␣ Binds to PLD␣1 Expressed in E. coli and Arabidopsis-
G␣ and PLD␣1 were expressed in E. coli as GST-fused and His-tagged proteins, respectively (Fig. 1A). Both PLD␣1 and G␣ are catalytically active as indicated by the presence of their respective activities (Figs. 1B and 2A). When bacterial lysate harboring His-PLD␣1 was incubated with the lysate containing GST-G␣, PLD␣1 was pulled down together with GST-G␣ by FIG. 3. Binding of G␣ with PLD␣1 from Arabidopsis leaf extracts. A, immunoblotting of PLD␣1 co-precipitated with G␣. Increasing amounts of purified GST-G␣ that were bound to beads were added to plant extracts (PE; 200 g/reaction) to precipitate PLD␣1. As a control, GST-bound beads (GST) were added to plant extracts and precipitated in the same manner. Precipitates pulled down with GST-G␣ were subjected to 8% SDS-PAGE followed by blotting with PLD␣1 antibody. The PLD␣1 band was made visible by staining alkaline phosphatase activity conjugated to a second antibody. B, PLD␣1 activity in the GST-G␣ precipitates corresponding to those in panel A. PLD activity was expressed as nmol of choline released/min/mg of starting plant proteins.
FIG. 4. The DRY motif of PLD␣1 and the amino acid residues involved in the PLD␣1 binding to G␣. A, schematic presentation of various domain structures of PLD␣1. C2 is a Ca 2ϩ -dependent phospholipids-binding fold. HKD1 and HKD2 are the duplicated catalytic motifs. B, alignment of the DRY motif sequences between PLD␣1 and the chicken GPCR rhodopsin. C, PLD␣1 activity in wild type (WT) and the DRY motif mutants E563A, K564A, and F565A. Proteins expressed and purified from E. coli were assayed for PLD activity. D, immunoblotting of wildtype and mutated PLD␣s with PLD␣1 antibodies before G␣ precipitation with glutathione beads. E, immunoblotting of wild-type and mutated PLD␣s in G␣ precipitates from the same amounts of starting PLD␣ proteins (0.2 mol; see panel D). F, PLD activity in the G␣ precipitates, expressed as nmol of choline released/ min/mg of starting PLD␣ proteins. glutathione beads (Fig. 2A). The presence of PLD␣1 in the precipitates was measured by immunoblotting with PLD␣1specific antibodies and assaying PLD␣1 activity ( Fig. 2A). PLD␣1 was not pulled down with GST, indicating that the co-precipitation resulted from the presence of G␣ but not GST or the beads. To estimate the stoichiometry of the interaction, G␣ and PLD␣1 were purified and co-incubated at different molar ratios (Fig. 2B). Increasing the ratio of PLD␣1 to G␣ from 0.5 to 1 led to a doubling of the PLD co-precipitated with G␣, whereas an increase in the molar ratio to 2:1 did not lead to a further increase in PLD␣ complexed with G␣. These results indicate that PLD␣1 binds to G␣ in a 1:1 ratio.
To demonstrate the association of G␣ with native PLD␣1, the purified G␣ was incubated with Arabidopsis leaf protein extracts followed by precipitation with glutathione beads. Native PLD␣1 was co-precipitated with G␣ as measured by immunoblotting and PLD␣1-specific activity assays (Fig. 3). As a control, GST beads gave negligible binding to PLD␣1 from plant extracts. The amount of PLD␣1 co-precipitated with G␣ increased as the amount of G␣ was increased (Fig. 3), indicating that the G␣ and PLD␣1 binding is dose-dependent.
PLD␣1 Contains a DRY-like Motif That Is Responsible for Its Binding to G␣-To determine the molecular basis for the PLD␣1 and G␣ interaction, the PLD␣1 sequence was analyzed for the presence of sequences with similarity to known Gprotein-interacting motifs. At amino acid residues 562-586, the sequence of PLD␣1 is highly similar to the DRY motif that has been found in more than 200 GPCRs (Fig. 4, A and B). This motif in GPCRs is located at the cytosolic juncture of the third transmembrane domain and is thought to associate with G␣ of heterotrimeric G-proteins (24,25). The DRY motif consists of a core triplet of amino acids, D-R-Y, and also a highly conserved hydrophobic region, VYVVV, located immediately downstream. The hydrophobic region is completely conserved in PLD␣1 (Fig.  4B) and is conservatively modified in eight other Arabidopsis PLDs. The corresponding triplet amino acids in PLD␣1 are conservatively substituted to EKF, which lies between the two duplicated, catalytic HKD motifs (Fig. 4A). Conservative substitutions of the DRY amino acid residues are allowable. For example, receptors in chicken and bullfrog both exhibit E-R-F (Fig. 4B).
The EKF residues were mutated individually to Ala to determine the involvement of each of the residues in G␣ binding. All of the three mutated PLD␣1s displayed the PLD activity comparable with that of wild-type PLD (Fig. 4C), indicating that changing any of the residues to Ala results in no major alteration of PLD catalytic activity. GST-G␣ was then co-incubated with the same amounts of wild-type and mutated PLD␣1s (Fig. 4D), followed by precipitation of G␣ with glutathione beads. Virtually no PLD␣1 K564A was co-precipitated as measured by the lack of PLD␣1 protein (Fig. 4E) and activity (Fig. 4F) in the G␣ precipitates. The amount of mutated PLD E563A and PLD F565A co-purified with G␣ was more than 3-fold lower than that of wild-type PLD␣1 (Fig. 4, E and F). These results indicate that the residue Lys 564 is essential for PLD␣1 to interact with G␣, whereas the flanking Glu 563 and Phe 565 are important for enhancing this interaction.
Association of G␣ with PLD␣1 Decreases PLD␣1 Activity-To test the effect of the G␣ binding on PLD␣1 activity, purified G␣ was co-incubated with wild-type and mutated PLD␣1s at approximately a 1:1 molar ratio, followed by assaying PLD␣1 activity (Fig. 5). Addition of G␣ decreased the wild-type PLD␣1 activity by more than 3-fold, whereas addition of the same amount of an unrelated protein, BSA, had no effect, indicating that the G␣ inhibition is not a nonspecific protein effect (Fig.  5A). The G␣ inhibition of PLD␣1 E563A or PLD␣1 K565A was not as strong as that of wild-type PLD␣1, and the lower degree of inhibition is consistent with the decreased G␣ binding exhibited by these mutated PLDs. G␣ had the least inhibitory effect on PLD␣1 K564A (Fig. 5A), which exhibited very little binding to G␣ (Fig. 4, E and F). The extent of G␣ inhibition of PLD␣1 is conversely correlated with the binding ability of the various PLD␣1s to G␣, indicating that the binding is required for the G␣ inhibition of PLD␣1 activity.
The G␣ inhibition of PLD activity was also observed with proteins from Arabidopsis (Fig. 5B). When purified GST-G␣ was co-incubated with Arabidopsis leaf protein extracts, G␣ decreased ϳ60% of PLD␣1 activity, whereas co-incubation with the same molar amount of GST or BSA had no effect on PLD activity. This result indicates that the inhibition is caused by G␣, not by GST or agarose beads present in the solution. The extent of inhibition for the leaf PLD was smaller than that of the bacterially expressed PLD␣1, which could be caused by having other PLDs in plant extracts, as all Arabidopsis PLDs are active under the PC-PIP 2 assay conditions (1,2). In addition, because the molar amount of PLD␣1 in plant extracts was difficult to determine, the presence of excessive PLD␣1 could also contribute to less inhibition of PLD by G␣.
Guanine Nucleotides Affect the G␣-PLD␣1 Binding and PLD␣1 Activity-G␣ can bind GDP and GTP; it is GDP-bound in the resting state, whereas an exchange of GDP for GTP is associated with the activation of G␣. To determine whether guanine nucleotides affect the G␣-PLD␣1 interaction, purified PLD␣1 and G␣ were co-incubated at a 1:1 molar ratio in the presence of different concentrations of the hydrolysis-resistant analogues GDP␤S and GTP␥S. GDP␤S gave a dose-dependent inhibition, whereas GTP␥S yielded a dose-dependent stimulation of PLD␣1 activity (Fig. 6A). When GDP␤S or GTP␥S was added to the DRY motif mutants and G␣ complexes, the activity of PLD␣1 K564A was not affected significantly (Fig. 6, B and C; GST versus K564A). Compared with that of wild-type PLD␣1 and PLD␣1 K564A , GDP␤S and GTP␥S had an intermediate inhibitory and stimulatory effect, respectively, on the activity of PLD␣1 E563A and PLD␣1 F565A (Fig. 6, B and C; wildtype (WT) versus E563A and F565A). Binding results showed that PLD␣1 K564A was unable to bind to G␣ and that PLD␣1 E563A and PLD␣1 F565A had decreased binding (Fig. 4E). Thus, these nucleotide effects on PLD activity are in agreement with the G␣ binding abilities of these PLDs, indicating that guanine nucleotides modulate PLD activity through their effects on the G␣ and PLD␣1 binding.
The effect of guanine nucleotides on the G␣-PLD␣1 binding was further determined using Arabidopsis leaf extracts (Fig.  7). Consistent with the results shown in Figs. 1-6, G␣ was able to bind to PLD␣1 without any added guanine nucleotide. Adding GDP␤S to plant protein extracts increased the amounts of PLD␣1 co-purified with G␣ as measured by immunoblotting and assaying PLD␣1 activity in the co-precipitates (Fig. 7). Conversely, GTP␥S decreased the G␣-PLD␣1 binding in a dosedependent manner. The inhibition of G␣-PLD␣1 binding by Gpp(NH)p, another guanosine triphosphate nucleotide analogue, was even greater than that by GTP␥S; almost no PLD␣1 was co-precipitated with G␣ at 20 M Gpp(NH)p. This difference could be attributed to the possibility that the Gpp(NH)p is more stable than GTP␥S (Fig. 7). These results indicate that the binding of GTP to G␣, thus the activation of G␣, decreases the interaction between G␣ and PLD␣1, whereas G␣ in the resting, presumably GDP-bound state binds to PLD.
PLD␣1 Binding Stimulates GTPase Activity-To determine potential reciprocal effects of the G␣-PLD␣1 interaction, the intrinsic GTPase activity of G␣ was determined in the presence of wild-type PLD␣1 or the non-G␣-binding mutant PLD␣1 K564A . Wild-type PLD␣1 increased the GTPase activity by about 35%, whereas PLD␣1 K564A had no stimulatory effect on the GTPase activity (Fig. 8A). This result indicates that binding of PLD␣ to G␣ is required for the PLD␣1 stimulation of G␣-mediated hydrolysis of GTP. The kinetics of the GTPase activity was assessed by varying the substrate GTP concentrations in the presence or absence of PLD␣1 (Fig. 8B). The FIG. 7. Effects of guanine nucleotide analogues on the G␣ binding to PLD␣ from Arabidopsis. A, immunoblotting of Arabidopsis PLD␣1 co-precipitated with GST-G␣ in the presence of 0, 5, 10, and 20 M GDP␤S, GTP␥S, or Gpp(NH)p. Arabidopsis leaf extracts (50 g of proteins from 10,000 ϫ g supernatant) were co-incubated with or without the nucleotide analogues followed by precipitation with GST-G␣ beads (0.05 mol of GST-G␣). A portion of the precipitates was subjected to 10% SDS-PAGE and immunoblotting, and another was assayed for PLD activity. B, PLD␣1 activity in the GST-G␣ precipitates. PLD activity was expressed as nmol of choline released/min/mg of starting plant proteins.  (PLDK564A). B, GTPase activity of G␣ as a function of increasing the substrate GTP concentrations in the presence or absence of wildtype PLD␣1. Equal molar amounts of bacterially expressed, purified PLD␣1 and G␣ (0.17 mol each) were co-incubated for the assays. Hydrolysis of GTP was assayed by measuring P i release by GST-G␣. The control was the reaction with GST-G␣ in the absence of GTP.
Michaelis constant K m and the maximal rate V max were derived by the Lineweaver-Burk plot. With or without PLD␣1, the K m for GTP was 182 nM. The V max was 27.0 and 36.4 pmol/mg of protein/min in the presence or absence of PLD␣1, respectively. These data suggest that the PLD␣1 promotes the turnover number of the GTPase activity without affecting the affinity of G␣ for GTP. DISCUSSION Results of this study provide evidence for the direct physical interaction between PLD␣1 and G␣. Moreover, this study has identified the specific DRY motif on PLD␣1 responsible for the binding (Fig. 3). The DRY motif is present in PLD␣-related PLDs from different plant species (26), indicating that this G␣ regulation is conserved for the common plant PLD. The DRY motif is normally found on GPCRs and is important in coupling the receptor and G-protein activation, as mutation of this motif results in a variety of modified phenotypes, including constitutive, moderate, and null receptor activities (24,25). This raises the question of whether the DRY motif on plant PLDs competes with plant GPCRs from the receptor-G␣ interaction. However, almost nothing is known about the coupling of G␣ to GPCR in plants (10,11). The only reported putative Arabidopsis GPCR is GCR1 (27), but its direct interaction with G␣ has yet to be demonstrated. In addition, our sequence analysis indicates that Arabidopsis GCR1 does not contain the DRY motif. It has also been stated that the loss-of-function mutant of GCR1 (gcr1) showed no shared phenotypes with that of G␣ (gpa1) and that G␣ might not be coupled to GCR1 in the cell (11). Thus, the study on the role of the PLD-G␣ interaction in the receptor-G␣ interaction awaits identification of a receptor interacting with G␣ in plants.
It is unlikely that PLD␣1 itself serves as a canonical G␣coupled receptor protein because of its lack of transmembrane domains. Rather, PLD␣1 may function as an intracellular regulator of G-proteins. A number of regulator proteins have been identified in other organisms, which include GTPase-activating protein (GAP) and guanine nucleotide exchange factors (8,9). However, proteins with such functions have not been reported in plants (10,11). The present results indicate that PLD␣1 has some GAP activity and that the GAP activity requires the G␣-PLD␣1 binding (Fig. 8A). The GTPase-activating effect of PLD␣1 is relatively small, and this could be due to the absence of an activated GPCR that is required for the function of some GAPs (28). Kinetic analysis suggests that PLD␣1 stimulates GTPase activity via increasing the turnover number without affecting the GTP affinity to the enzyme (Fig. 8B). One way of increasing turnover is to promote the exchange between GDP and GTP. In addition, other modes of regulation of G␣ by PLD␣1 should also be considered. PLD␣1 has been proposed to translocate between cytosol to membranes in plants (1,4), and it might be possible that the PLD␣1-G␣ binding modulates the intracellular location and thus the activity of G-protein.
The effect of the PLD␣1-G␣ interaction is reciprocal; while it stimulates the GTPase activity of G␣, the interaction inhibits PLD␣1 activity (Fig. 5). Results from the mutagenesis of the DRY motif have indicated that the direct binding is required for the effect on both G␣ and PLD␣1. Whereas the DRY motif on PLD␣1 is essential to the binding, the state of G␣ activation also determines the PLD␣1-G␣ interaction. PLD␣1 is able to bind to G␣ in the resting or GDP-bound state but is unable to bind GTP-bound G␣. A previous study with bacterially expressed tobacco G␣ and PLD␣ also indicates that GTP␥S can decrease the G␣ inhibition of PLD␣ activity (18). The dissociation of G␣ from PLD␣1 is likely to result from a conformational change of G␣ triggered by the exchange of GDP for GTP and the activation of G␣.
One intriguing question raised from the PLD␣1-G␣ interaction is whether the PLD␣1-G␣ interaction modulates the association of G␣ with G␤␥. According to the mammalian heterotrimeric G-protein paradigm, G␣ in the GDP-bound resting state associates with G␤␥, whereas G␣ in the GTP-bound activated state dissociates from G␤␥ (8). Some activators of Gprotein signaling proteins identified in other organisms have been shown to bind to the resting or GDP-bound form of G␣ to release G␤␥ (28,29). The present results demonstrate that PLD␣1-G␣ interaction occurs with or without added GDP, and thus, PLD␣1 binds to the resting state, GDP-bound G␣. One canonical G␤ gene and two G␥ genes have been identified in Arabidopsis (30,31). The association has been demonstrated for G␤ and G␥ subunits (31), but not for G␣ and G␤␥. Further studies will be needed to identify the condition under which G␣ interacts with G␤␥ and then to determine whether the PLD␣1 binding regulates the association and/or dissociation between G␣ with G␤␥. Possibilities remain also that the association of G␣ with G␤␥ may affect the G␣-PLD␣1 interaction.
Another question relevant to the understanding of the PLD-G␣ interaction is the specificity of the G␣ interaction with various PLDs. Except for two PLDs and PLD␥2, nine of the 12 PLDs in Arabidopsis have the DRY motif-like sequences (data not shown) and thus can potentially interact with G␣. Different patterns of temporal and spatial expression and distribution have been noted for some PLDs (32)(33)(34), which could be important determinants for the G␣ interaction with different PLDs in different cells and timing. In addition, individual PLDs exhibit different requirements for Ca 2ϩ , PIP 2 , free fatty acids, and lipid environments for activity (1,33,35,36). Thus, the other cellular factors might also modulate the PLD-G␣ interaction and specificity. It should be noted that analyses of PLD␣1-depleted (7) and G␣-null (12) Arabidopsis have shown some shared phenotypic alterations, such as retardation of stomatal movement and increasing plant water loss. It is likely that the PLD␣1-G␣ interaction documented in this study underlies a structural and molecular basis for their involvement in specific plant signaling pathways.
Based on "Results" and "Discussion," a working model is proposed for the role and further investigation of the PLD␣1-G␣ interaction (Fig. 9). PLD␣1 binds to the resting state or GDP-bound G␣ through the DRY motif. The PLD␣1-G␣ binding hinders PLD␣1 activity and may also promote dissociation of G␤␥ from G␣. In addition, the PLD␣1-G␣ binding FIG. 9. A proposed model depicting the PLD␣1-G␣ binding and the functions of the interaction. PLD␣1 binds to the resting state, GDP-bound G␣ through the EKF residues in the DRY motif. The PLD␣1-G␣ binding decreases PLD␣1 activity, and GTP decreases the binding and restores PLD␣1 activity. Reciprocally, the PLD binding to G␣ promotes GTPase activity, possibly via promoting the exchange of GDP for GTP and also the dissociation of G␤␥ from G␣. The binding of GTP to and the activation of G␣ promote dissociation of PLD␣1-G␣ complexes. Thus, the association and dissociation between PLD and G␣ form a positive loop to stimulate the function of both PLD and Gproteins. The upward and downward arrows next to PLD indicate a decrease and increase in activity, respectively. The scale of drawing is not proportional to the size of the proteins.
stimulates GTPase activity, possibly via promoting the exchange of GDP for GTP. At the same time, the binding of GTP to G␣ promotes dissociation of the PLD␣1-G␣ complexes and thus restores PLD␣1 activity. The released PLD␣1 will be able to bind to GDP-bound G␣ again, and another cycle begins. Thus, the association and dissociation between PLD and G␣ would form a positive loop to stimulate the activity of PLD␣1 and G␣, and this dynamic process would regulate reciprocally the cellular function of both PLD and G-proteins. Further studies on elucidating this novel signaling process and identifying upstream regulators and downstream effectors of the interaction will facilitate the understanding of the cellular roles of PLDs and heterotrimeric G-protein in plant signal transduction.