Protein Kinase D Activation by Mutations within Its Pleckstrin Homology Domain*

Protein kinase D (PKD) is a serine/threonine protein kinase that contains a cysteine-rich repeat sequence homologous to that seen in the regulatory domain of protein kinase C (PKC) and a catalytic domain with only a low degree of sequence similarity to PKCs. PKD also contains a pleckstrin homology (PH) domain inserted between the cysteine-rich motifs and the catalytic domain that is not present in any of the PKCs. To investigate the function of the PH domain in the regulation of PKD activity, we determined the kinase activity of several PKD PH domain mutants immunoprecipitated from lysates of transiently transfected COS-7 cells. Deletion of the entire PH domain (amino acids 429–557) markedly increased the basal activity of the enzyme as assessed by autophosphorylation (∼16-fold) and exogenous syntide-2 peptide substrate phosphorylation assays (∼12-fold). Mutant PKD proteins with partial deletions or single amino acid substitutions within the PH domain (e.g. R447C and W538A) also exhibited increased basal kinase activity. These constitutive active mutants of PKD were only slightly further stimulated by phorbol-12,13-dibutyrate treatment of intact cells. Our results demonstrate, for the first time, that the PKD PH domain plays a negative role in the regulation of enzyme activity.

Protein kinase D (PKD) is a serine/threonine protein kinase that contains a cysteine-rich repeat sequence homologous to that seen in the regulatory domain of protein kinase C (PKC) and a catalytic domain with only a low degree of sequence similarity to PKCs. PKD also contains a pleckstrin homology (PH) domain inserted between the cysteine-rich motifs and the catalytic domain that is not present in any of the PKCs. To investigate the function of the PH domain in the regulation of PKD activity, we determined the kinase activity of several PKD PH domain mutants immunoprecipitated from lysates of transiently transfected COS-7 cells. Deletion of the entire PH domain (amino acids 429 -557) markedly increased the basal activity of the enzyme as assessed by autophosphorylation (ϳ16-fold) and exogenous syntide-2 peptide substrate phosphorylation assays (ϳ12fold). Mutant PKD proteins with partial deletions or single amino acid substitutions within the PH domain (e.g. R447C and W538A) also exhibited increased basal kinase activity. These constitutive active mutants of PKD were only slightly further stimulated by phorbol-12,13-dibutyrate treatment of intact cells. Our results demonstrate, for the first time, that the PKD PH domain plays a negative role in the regulation of enzyme activity.
The newly identified PKD is a mouse serine/threonine protein kinase with distinct structural and enzymological properties (8). The catalytic domain of PKD is distantly related to Ca 2ϩ -regulated kinases and shows little similarity to the highly conserved regions of the kinase subdomains of the PKC family (9). Consistent with this, PKD does not phosphorylate a variety of substrates utilized by PKCs indicating that PKD is a protein kinase with distinct substrate specificity (8,10). However, the amino-terminal region of PKD contains a tandem repeat of cysteine-rich, zinc finger-like motifs that bind phorbol esters with high affinity (8,10). Immunopurified PKD is markedly stimulated by PDB or diacylglycerol in the presence of phosphatidylserine (10). The human protein kinase PKC (11,12) with 92% homology to PKD (extending to 98% homology in the catalytic domain) is also stimulated by phorbol esters and phospholipids (12). These in vitro results indicate that PKD/ PKC is a novel phorbol ester/diacylglycerol-stimulated protein kinase. Recently a new mechanism of PKD activation has been identified (13). Specifically, exposure of intact cells to biologically active phorbol esters and membrane-permeant diacylglycerol induces phosphorylation-dependent PKD activation via a PKC-dependent pathway (13). Thus, PKD can function either in parallel to or downstream of PKCs in signal transduction.
The amino-terminal region of PKD also contains a PH domain which is not found in any of the PKCs. PH domains are molecular structures of approximately 120 amino acids with limited identity in sequence but similar three-dimensional structure (14 -19). These domains have been identified in a large number of signaling and cytoskeletal proteins (for review see Refs. 18 -21). It has been suggested that PH domains mediate intermolecular and/or intramolecular interactions like src homology domain 2 and 3, but their function and binding partners remain unclear. In some cases PH domains have been shown to bind phosphoinositides and their head groups or proteins such as the ␤␥ subunits of heterotrimeric G proteins (19,(22)(23)(24)(25). The integrity of the PH domain is critical for the activation and subcellular localization of many PH domaincontaining enzymes including Bruton's tyrosine kinase (14, 26 -28), ␤-adrenergic receptor kinase (25), and the serine/threonine kinase encoded by the proto-oncogene c-akt (29). These considerations prompted us to examine the function of the PH domain in the regulation of PKD activity.
In the present study we demonstrate that a PKD mutant lacking the entire PH domain exhibits high basal kinase activity. This active form of PKD binds phorbol esters as well as the wild type protein but is only slightly further activated by treatment with PDB in vivo. PKD mutants lacking part of the PH domain or with single amino acid substitutions within the PH domain also show high basal kinase activity. Our results indicate that the PH domain of PKD plays a negative role in the regulation of PKD kinase activity.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-COS-7 cells were cultured in DMEM supplemented with 10% fetal bovine serum at 37°C in a humidified atmosphere containing 10% CO 2 . Exponentially growing COS-7 cells, 40 -60% confluent, were transfected in serum-free medium by using 5 g of DNA from the various plasmids and 10 l of Lipofectin reagent (Life Technologies, Inc.) per 60-mm-diameter dish according to the * 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  protocol provided by the manufacturer. Briefly, Lipofectin (10 l) was diluted to 1 ml with Opti-MEM I medium (Life Technologies, Inc.), left for 30 min, and then mixed with the DNA previously diluted in 1 ml of the same medium. After 15 min, the volume of the DNA-Lipofectin mixture was increased to 2.5 ml with Opti-MEM I and overlaid onto rinsed (once with Opti-MEM I) COS-7 cells. The cultures were incubated at 37°C for 6 h, and the medium was then replaced with fresh DMEM containing 10% fetal bovine serum. The cells were used for experimental purposes 72 h later.
Deletion Mutants and Site-directed Mutagenesis-A PH domain deletion mutant was generated by direct-rapid mutagenesis of large plasmids using PCR (30) with rTth DNA polymerase XL with proofreading capability (GeneAmp XL PCR Kit, Perkin-Elmer). The PH domain deletion mutant of PKD (PKD⌬PH), lacking amino acids Val-429 to Gly-557, was made directly in PKD-cDNA cloned in pBluescript-SK (ϩ) using oligonucleotide primers starting upstream (reverse primer) and downstream (forward primer) from the desired deletion and also containing the unique restriction site AscI which is not present in wild type PKD (Table I, deletion mutant and Fig. 1, PKD⌬PH). PCR was performed by using a DNA Thermal Cycler (Perkin-Elmer) applying a short number of cycles (9 cycles) starting with a large amount of template DNA (1 g). The PCR parameters were optimized for the oligonucleotides used. After PCR, template DNA was eliminated by DpnI digestion. Amplified DNA was cut with AscI and ligated. The resulting deletion construct was then subcloned into the mammalian expression vector pcDNA3 (pcDNA3-PKD⌬PH) for transient expression in COS-7 cells.
Complete or partial sequences of the PH domain generated by PCR (see below) were reinserted into the AscI site of pcDNA3-PKD⌬PH. Primers carrying the AscI site (Table I, reinsertion mutants) were used to amplify the entire PH domain (pcDNA3-PKD⌬PHϩPH: reinsertion of Val-429 to Gly-557) or parts of the PH domain. These partial PH domain deletions included deletions of the carboxyl-terminal ␣-helix (pcDNA3-PKD⌬␣: reinsertion of Val-429 to Val-534, deletion of amino acids 535-557) and of the first four ␤-sheets of the ␤-barrel (pcDNA3-PKD⌬1-4␤: reinsertion of Ser-475 to Gly-557, deletion of amino acids 429 -474). These PCR products were digested with AscI and subcloned into pcDNA3-PKD⌬PH already digested with AscI. Fig.  1 shows a schematic representation of the different insertion mutants.
The site-specific mutations in the PH domain of PKD resulting in single amino acid substitutions (R447C and W538A) were generated by overlap PCR using pBluescript SK(ϩ)-PKDcDNA (pBSK-PKD). For PKD-R447C, which alters a semi-conserved arginine located in the ␤2-sheet of the PH domain, a sequence upstream of PKDcDNA corresponding to pBSK and close to the polylinker and a sequence near the SphI site (at nucleotide 1754) within PKDcDNA were used as external forward and reverse primers together with internal reverse and forward primers, respectively, containing the residue substitution R447(AGG) to C447(TGC) ( Table I, single amino acid substitutions and Fig. 1 shows a scheme of the different mutants). After the second PCR reaction the amplified fragment was cut with XhoI and SphI. pBSK-PKD was also digested with these two restriction enzymes, and the wild type PKD-XhoI/SphI fragment was then replaced by the PCR product containing the desired mutation. For PKD-W538A, which alters the only invariant amino acid in the carboxyl-terminal ␣-helix of the PH domain, the same external forward primer and a reverse primer containing both the W538(TGG) to A538(GCG) mutation and the   (18). ␤1-␤7 indicate the seven ␤-sheets. PKD⌬PH is a deletion mutant that lacks amino acids 429 -557 encompassing the entire PH domain. PKD⌬PHϩPH is a construct where the PH domain was reinserted in the deletion mutant PKD⌬PH. PKD⌬␣1 and PKD⌬1-4␤ lack the carboxyl-terminal ␣-helix (amino acids 535-557) or the first amino-terminal four ␤-sheets (amino acids 429 -474), respectively. PKD-R447C carries an arginine to cysteine mutation at position 447 (present in ␤2), and PKD-W538A carries a tryptophan to alanine mutation at position 538 (present in the ␣-helix). The mutants were generated by PCR as described in detail under "Experimental Procedures" using the primers listed in Table I.
SphI-PKDcDNA site were used ( Table I). The PCR product was inserted into PKD using the same substitution strategy. Constructs were subcloned into the mammalian expression vector pcDNA3 (pcDNA3-PKD-R447C and pcDNA3-PKDW538A) for transient expression in COS-7 cells.
All mutant constructs were analyzed by DNA restriction enzyme analysis and sequencing (Sequenase version 2.0 DNA sequencing Kit, U. S. Biochemical Corp. and Amersham Corp.) and showed to contain the desired deletion or point mutations.
Autophosphorylation Assay-PKD autophosphorylation was determined in an in vitro kinase assay as described previously (13). Briefly, the immunoprecipitates were washed once with buffer A, twice with buffer B (buffer A minus Triton X-100), twice with kinase buffer (30 mM Tris-HCl, pH 7.6, 10 mM MgCl 2 ), and 20 l of PKD immune complexes were mixed with 20 l of kinase buffer containing 100 M final concentration of [␥-32 P]ATP (specific activity of 400 -600 cpm/pmol) for 10 min at 30°C. The reaction was then stopped by adding an equal volume of 2 ϫ SDS-PAGE sample buffer (1 M Tris-HCl, pH 6.8, 0.1 mM Na 3 VO 4 , 6% SDS, 0.5 M EDTA, 4% 2-mercaptoethanol, 10% glycerol) and analyzed by SDS-PAGE. The gels were dried and autoradiographs were scanned in a LKB Ultrascan XL densitometer, and the labeled band corresponding to autophosphorylated PKD or the different mutants was quantified using an Ultrascan XL internal integrator.
Exogenous Substrate Phosphorylation-The phosphorylation of various peptides by immunoprecipitated wild type PKD or mutants was carried out under the same conditions as the in vitro kinase assay adding a final concentration of 2.5 mg/ml syntide-2, epsilon-peptide, myelin basic protein, or histones. After 10 min at 30°C, the reaction was terminated by adding 100 l of 75 mM H 3 PO 4 and spotting 75 l of the supernatant on P-81 phosphocellulose paper, and free [␥-32 P]ATP was separated from the labeled substrate by washing four times (5 min each) in 75 mM H 3 PO 4 . The P-81 papers were dried, and radioactivity incorporated into the different peptides was determined by Cerenkov counting.
Western Blot Analysis-For Western blot analysis of either PKD or PKD mutants expressed in transfected COS-7 cells, 50 g of total protein from cell lysates were mixed with the same volume of 4 ϫ SDS-PAGE sample buffer, boiled for 10 min, and analyzed by SDS-PAGE followed by transfer to Immobilon membranes at 100 V, 0.4 A at 4°C for 4 h using a Bio-Rad transfer apparatus. The transfer buffer composition was 200 mM glycine, 25 mM Tris, 0.01% SDS, and 20% CH 3 OH. Membranes were blocked for 1 h at room temperature in 5% non-fat dried milk in PBS, pH 7.2, and incubated for 3 h with PA-1 antiserum (1:500 dilution) in PBS containing 3% non-fat dried milk. Immunoreactive bands were visualized using either 125 I-labeled protein A (0.1 Ci/ml) and autoradiography or horseradish peroxidase-conjugated anti-rabbit IgG and subsequent enhanced chemiluminescence detection.
PDB Binding to COS-7 Cells-[ 3 H]PDB binding to intact COS-7 cells was performed as described previously (31). Briefly, cultures were washed twice with DMEM and incubated with binding medium (DMEM containing 1 mg/ml bovine serum albumin and 10 nM [ 3 H]PDB) at 37°C for 30 min. The cultures were then rapidly washed at 4°C with PBS containing 1 mg/ml bovine serum albumin and lysed with NaOH/SDS, and bound radioactivity was determined by scintillation counting. Nonspecific binding was determined in the presence of 10 M unlabeled PDB.
Materials-[␥-32 P]ATP (370 MBq/ml), 125 I-labeled protein A (15 mCi/ ml), and enhanced chemiluminescence reagents were from Amersham Int. (United Kingdom). PDB was obtained from Sigma. The inhibitor GF I was from LC Laboratories. Protein A-agarose was from Boehringer Mannheim. Other items were from standard suppliers or as indicated in the text.

Expression of PKD⌬PH in COS-7 Cells-
The region comprising nucleotides 1408 -1796 in the cDNA encoding PKD (corresponding to amino acids 429 -557 coding for the PH domain) was deleted and the mutated PKD, lacking the entire PH domain, was subcloned into the mammalian expression vector pcDNA3. The resulting construct (PKD⌬PH) was transiently transfected into COS-7 cells. To examine the expression of PKD⌬PH, lysates from these cells were analyzed by Western blotting with the PA-1 antibody directed against the carboxyl terminus of PKD. As can be seen in Fig. 2 (left), PKD⌬PH migrated in SDS-PAGE gels with an apparent molecular mass of ϳ96 kDa instead of 110 kDa as expected after deletion of the 128 amino acids corresponding to the PH domain. The level of immunoreactive PKD⌬PH was comparable to that of PKD (Fig.  2, left). In addition, cells transfected with either pcDNA3-PKD or pcDNA3-PKD⌬PH showed a similar increase (5-6-fold) in specific [ 3 H]PDB binding as compared with that obtained in COS-7 cells transfected with the vector alone (Fig. 2, right). Thus, the level of expression of PKD⌬PH, judged either by immunoblotting or [ 3 H]PDB binding, is similar to that of PKD in transiently transfected COS-7 cells.
Deletion of the PH Domain Causes PKD Activation-To ex- amine the function of the PH domain in the regulation of PKD activity, COS-7 cells transiently transfected with pcDNA3-PKD or pcDNA3-PKD⌬PH were treated with or without 200 nM PDB for 10 min and then lysed and immunoprecipitated with the PA-1 antibody. The immune complexes were incubated with [␥-32 P]ATP and analyzed by SDS-PAGE, autoradiography, and scanning densitometry to determine the level of PKD autophosphorylation. In agreement with previous results (10,13), PKD isolated from unstimulated cells had low catalytic activity that was markedly activated by PDB stimulation of intact cells. In striking contrast, PKD⌬PH exhibited a high level of basal catalytic activity (16-fold increase compared with unstimulated PKD) which was not affected by treatment with PDB (Fig. 3A).
Subsequently, we determined whether a high level of basal PKD⌬PH activity could also be demonstrated using an exogenous substrate. The synthetic peptide syntide-2 (13, 32, 33) has been identified as an efficient substrate for the catalytic domain of PKD and for the full-length PKD (8,10,13). As shown in Fig. 3B, PKD⌬PH, unlike PKD, displayed high basal syntide-2 kinase activity (8.5-fold increase) that was only slightly further enhanced by PDB stimulation of intact cells. These results corroborated that deletion of the PH domain of PKD leads to a constitutively active state of this enzyme.
The catalytic domain of PKD or the full-length enzyme efficiently phosphorylate syntide-2, but they phosphorylate myelin basic protein inefficiently and they do not phosphorylate histones or a peptide based on the pseudosubstrate motif of PKC⑀ which is a substrate for all members of the PKC family (34,35). As shown in Table II, this characteristic substrate specificity was not altered by deletion of the PH domain from PKD.
Given that PKD is activated by PDB in intact cells through a PKC-dependent pathway (13), we considered the possibility that the active state of PKD⌬PH is also mediated through PKC. To examine this possibility, we assessed the effect of the PKC inhibitor GF I (also known as bisindolylmaleimide I or GF 109203X; Ref. 36) on the activity of PKD⌬PH immunoprecipitated from lysates of COS-7 cells pretreated with or without PDB. Exposure to 3.5 M GF I for 90 min did not affect the high activity of PKD⌬PH recovered from either control or PDBtreated cells as shown by autophosphorylation and syntide-2 phosphorylation assays (Fig. 4). In contrast, an identical treatment (3.5 M GF I for 90 min) markedly inhibited the activation of PKD induced by PDB. These results indicate that the active state of PKD⌬PH is not mediated by PKC.
Partial Deletions and Single Amino Acid Substitutions within the PH Domain Also Lead to PKD Activation-The preceding results suggest that deletion of the PKD PH domain leads to an active form of PKD. To substantiate this conclusion, we analyzed the activity of a set of PKD mutants containing partial deletions or single amino acid substitutions within the PH domain. The partially deleted PKD mutants were generated by reinserting the PH domain sequence of PKD containing partial deletions into PKD⌬PH to produce PKD⌬␣ and PKD⌬1-4␤. As a control, the entire sequence of the PH domain was also reinserted into PKD⌬PH (PKD⌬PHϩPH). Fig. 1 shows a schematic representation of the different mutants. COS-7 cells transiently transfected with these mutants were incubated in the absence or presence of PDB for 10 min, and kinase activity was determined after immunoprecipitation by measuring either syntide-2 phosphorylation or autophosphorylation. The expression of these mutants was verified by Western blotting (Fig. 5).
When the complete PH sequence was reinserted into PKD⌬PH, the basal kinase activity of the new construct PKD⌬PHϩPH was markedly reduced as compared with the unstimulated activity of PKD⌬PH. Treatment with PDB of intact cells transiently transfected with PKD⌬PHϩPH induced marked kinase activation. The level of syntide-2 phosphorylation by PKD⌬PHϩPH activated by PDB was comparable to that obtained in the wild type PKD (Fig. 5B). In contrast, the partial deletion mutants (i.e. PKD⌬␣, PKD⌬1-4␤) showed high basal activity as judged either by syntide-2 phosphorylation or autophosphorylation assays (Fig. 5B). Treatment with PDB of the cells transfected with these mutants caused only a small further increase in kinase activity as compared with PKD or PKD⌬PHϩPH.
Next, we examined whether single amino acid substitutions of the PH domain were sufficient to increase the basal activity of PKD. We mutated arginine 447 to cysteine (PKD-R447C) localized in the second ␤-sheet because comparable mutations in the PH domains of Bruton's tyrosine kinase (14,26,27) and Akt (29,37) interfered with their function. Tryptophan 538 was mutated to alanine (PKD-W538A) because this residue is highly conserved in the ␣-helix of most known PH domains. These two mutants exhibited higher basal activity when compared with wild type PKD either in syntide-2 phosphorylation or autophosphorylation assays (Fig. 6B). The increases in the basal syntide-2 kinase activity of PKD-R447C and PKD-W538A were 3.3-and 8-fold, respectively. The high constitutive activity of PKD-W538A was comparable to that of PKD mutants carrying partial or complete deletions of the PH domain. DISCUSSION Although the tertiary structure of several PH domains, some bound to ligands, has been determined, the function of this domain remains incompletely understood (18 -21). The PH domains of a variety of enzymes have been mutated or deleted in an effort to elucidate their function. Recent studies with Akt (23,29,37), G protein-coupled receptor kinase (38), Bruton's tyrosine kinase (14, 26 -28), ␤-adrenergic receptor kinase (25), the Ras exchange factor Sos (39 -42), and phospholipase C-␦ (43) have demonstrated that the PH domain plays an important role in the regulation of the enzyme activity. Most studies have shown that the integrity of the PH domain is necessary for  The newly identified PKD contains a PH domain inserted between the second cysteine-rich motif and the kinase catalytic domain (9). The presence of the PH domain distinguishes PKD from all known members of the PKC family. In the present study, we have deleted or mutated the PH domain of PKD and analyzed the regulatory properties of the resulting PKD mutants.
In striking contrast to most previous studies showing that functional PH domains are necessary for enzyme activation, our results demonstrate that partial or complete deletions of the PH domain of PKD (see Fig. 5) resulted in PKD mutants that exhibit a high level of basal kinase activity. Single amino acid substitutions (e.g. R447C and W538A) within the PH domain also promoted an activated state of PKD. The R447C mutation was tested because a comparable mutation (R28C) in the PH domain of Bruton's tyrosine kinase is responsible for the induction of X-linked agammaglobulinemia in humans and X-linked immunodeficiency in CBA/N mice (14,26,27), and a similar mutation in the serine/threonine kinase Akt abolished activation of this enzyme by platelet-derived growth factor (29,37). The W538A mutation alters the only invariant amino acid in the carboxyl-terminal ␣-helix and is conserved in all known PH domains (18,20). Interestingly, the W538A mutation was as effective as partial or complete deletions of the PH domain in promoting an active state of PKD.
Recent results demonstrated that treatment of intact cells with PDB, cell-permeant diacylglycerols, or bryostatin induces rapid PKD activation that was maintained during cell disruption and immunoprecipitation (13,44). Several lines of evidence including the use of PKC inhibitors and cotransfection of PKD with constitutively activated mutants of PKC⑀ and - (13) indicate that PKD can be activated by phosphorylation in intact cells through a PKC-dependent signal transduction pathway. Our results demonstrate that PKD carrying deletions of the PH domain or the single amino acid substitution W538A within the ␣-helix of the PH domain were only slightly stimulated further by treatment of the cells with PDB, implying that PKD rendered active by PH domain mutation is already fully activated. We conclude that the integrity of the PH domain is critical for maintaining unstimulated PKD in a state of low catalytic kinase activity.
The low basal activity of many protein kinases is maintained by the interaction between an autoinhibitory domain located within the enzyme with its catalytic site, thereby preventing the binding of substrates (45). For example, all members of the PKC family from yeast to human PKCs possess an autoinhibitory motif that is located upstream of the first cysteine-rich domain (46). In contrast, PKD does not contain a pseudosubstrate region in a comparable position (9). Although it is conceivable that PKD is also regulated by autoinhibition, as many other regulated protein kinases, the putative autoinhibitory region has not been identified yet. In view of our results, it is tempting to speculate that the PH domain of PKD has an inactivating function by functioning as an autoinhibitory domain. This suggests a novel role of PH domains in enzyme regulation. In the context of this model, partial or complete deletions of the PH domain of PKD or single amino acid substitutions within this domain should stabilize an active conformation of PKD as, in fact, it is shown by the mutational analysis presented here. Alternatively, our results cannot exclude the possibility that the PH domain of PKD could bind an inhibitory ligand(s) that is released by allosteric stimulation or by phosphorylation-dependent activation induced by treatment of intact cells with PDB. However, PH domain ligands such as phosphoinositides and ␤␥ subunits of G proteins promote correct subcellular localization and enzyme activation rather than inhibition of enzyme activity (18 -21). Future studies should attempt to distinguish between these alternative models. Regardless of the precise mechanism(s), our results demonstrate, for the first time, that the PH domain of PKD plays a negative role in the regulation of PKD kinase activity.