Protein kinase C regulates pleckstrin by phosphorylation of sites adjacent to the N-terminal pleckstrin homology domain.

Pleckstrin is a substrate for protein kinase C in activated platelets that contains at its N and C termini two of the pleckstrin homology (PH) domains that have been proposed to mediate protein-protein and protein-lipid interactions. We have recently shown that pleckstrin can inhibit agonist-induced phosphoinositide hydrolysis and that this inhibition requires an intact N-terminal PH domain (residues 6 to 99). In the present studies, we have identified the sites of phosphorylation in pleckstrin and examined their contribution to pleckstrin function. In human platelets activated with thrombin or phorbol esters, and in COS-1 cells expressing pleckstrin, a combination of phosphopeptide analysis and site-directed mutagenesis shows that three residues in the intervening sequence between the two pleckstrin PH domains become phosphorylated: Ser, Thr, and Ser. Replacing all three of these sites with glycine decreased phosphorylation by >90% and reduced pleckstrin's ability to inhibit phosphoinositide hydrolysis by as much as 80%. Replacing the phosphorylation sites with alanine residues had a similar effect, while substitution with aspartate, glutamate, or lysine residues produced pleckstrin variants that were fully active even in the absence of phosphorylation. These results suggest that phosphorylation enhances pleckstrin's activity by introducing a cluster of charges into a region adjacent to, but not within, the N-terminal PH domain. This may have an allosteric effect on the N-terminal PH domain, regulating its interaction with other molecules necessary for the inhibition of phosphoinositide hydrolysis.

Pleckstrin is a substrate for protein kinase C in activated platelets that contains at its N and C termini two of the pleckstrin homology (PH) domains that have been proposed to mediate protein-protein and protein-lipid interactions. We have recently shown that pleckstrin can inhibit agonist-induced phosphoinositide hydrolysis and that this inhibition requires an intact N-terminal PH domain (residues 6 to 99). In the present studies, we have identified the sites of phosphorylation in pleckstrin and examined their contribution to pleckstrin function. In human platelets activated with thrombin or phorbol esters, and in COS-1 cells expressing pleckstrin, a combination of phosphopeptide analysis and site-directed mutagenesis shows that three residues in the intervening sequence between the two pleckstrin PH domains become phosphorylated: Ser 113 , Thr 114 , and Ser 117 . Replacing all three of these sites with glycine decreased phosphorylation by >90% and reduced pleckstrin's ability to inhibit phosphoinositide hydrolysis by as much as 80%. Replacing the phosphorylation sites with alanine residues had a similar effect, while substitution with aspartate, glutamate, or lysine residues produced pleckstrin variants that were fully active even in the absence of phosphorylation. These results suggest that phosphorylation enhances pleckstrin's activity by introducing a cluster of charges into a region adjacent to, but not within, the N-terminal PH domain. This may have an allosteric effect on the N-terminal PH domain, regulating its interaction with other molecules necessary for the inhibition of phosphoinositide hydrolysis.
Proteins that are important in signal transduction often contain discrete domains that mediate protein-protein interactions. Examples of this include Src homology domains 2 and 3 (SH2 and SH3) which interact with specific tyrosine-phosphorylated and proline-rich amino acid sequences, respectively (1,2). It has been proposed that the N and C termini of the hematopoietic protein, pleckstrin, are the prototypes for a new family of intermolecular interaction domains, referred to as pleckstrin homology or PH 1 domains (3,4). Within the past year, similar three-dimensional structures have been reported for the PH domains from ␤-spectrin, dynamin, and the N terminus of pleckstrin, supporting their status as a bona fide structural motif, despite large differences in their primary sequences (5)(6)(7)(8)(9)(10).
Although it is generally believed that PH domains will turn out to mediate intermolecular interactions and may be involved in membrane targeting, there is as yet no consensus on the specifics of this interaction. Both G ␤␥ (11) and phosphatidylinositol 4,5-bisphosphate (PIP 2 ) (12) have been proposed as potential partners for PH domains, and several recent reports have suggested that the PH domain of the ␤-adrenergic receptor kinase may interact individually (13) or simultaneously with both (14,15). Pleckstrin or p47 is a 40-kDa protein present in platelets and leukocytes that becomes phosphorylated when platelets are activated by agonists that directly or indirectly activate protein kinase C (16). Phosphoamino acid analysis and two-dimensional electrophoresis suggest that pleckstrin is phosphorylated heterogeneously on one or more serine and threonine residues, but not tyrosine residues (17). The function of pleckstrin has not been fully characterized, but we have shown previously that when transfected into COS-1 or HEK-293 cells pleckstrin can inhibit agonist-induced phosphoinositide hydrolysis initiated by G-protein-coupled receptors and growth factor receptors (18). In that model system, pleckstrin also inhibited the increase in inositol phosphate formation caused by the expression of a constitutively active variant of G q␣ , but had no effect on G ␣ -or G ␤␥ -mediated regulation of adenylyl cyclase. The inhibition of phosphoinositide hydrolysis required an intact N-terminal PH domain and was additive with that observed when mock-transfected cells were preincubated briefly with PMA, suggesting that it is independent of the phosphorylation of receptors, G-proteins, and phospholipase C known to occur under the same conditions (19 -22).
Given the ability of pleckstrin's PH domains to bind to lipid micelles containing PIP 2 (12), one possible explanation for these results is that pleckstrin inhibits phosphoinositide hydrolysis by binding to PIP 2 and that phosphorylation of pleckstrin by protein kinase C promotes this interaction. In this context, pleckstrin could play a role in the feedback regulation of phosphoinositide hydrolysis following activation of protein kinase C, limiting the duration of inositol 1,4,5-trisphosphate formation. Conceivably, pleckstrin might also interact with G ␤␥ , such an interaction has been demonstrated for GST fusion proteins containing pleckstrin's PH domains (23), 2 but this alone would not readily account for the observed ability of pleckstrin to inhibit phosphoinositide hydrolysis initiated by TrkA receptors and constitutively active G q␣ .
In the present studies, we have identified the sites of phosphorylation within pleckstrin and examined their relationship to pleckstrin's ability to inhibit agonist-induced phosphoinositide hydrolysis. The results show that: 1) pleckstrin is variably phosphorylated on a cluster of residues located near, but not within, the N-terminal PH domain, 2) elimination of any one of these sites does not alter the overall phosphorylation of the molecule, presumably because of compensatory increases in phosphorylation at the other sites, 3) phosphorylation of pleckstrin is required for maximal inhibition of phosphoinositide hydrolysis, and 4) the effects of phosphorylation may be due largely to the introduction of charged residues into the region between the two PH domains.

Construction of Pleckstrin Expression Vectors-
The expression plasmid encoding for full-length human pleckstrin was described previously (18). Pleckstrin variants containing mutations at Ser 113 , Thr 114 , and Ser 117 were generated using the technique of Landt et al. (24). All of the pleckstrin variants were cloned into pCMV5. The cDNA sequences of the cloned wild type pleckstrin and variants were all confirmed and in agreement with the published sequence of pleckstrin (25).
Mapping of Phosphorylation Sites-Human platelets isolated by differential centrifugation from blood anticoagulated with ACD were washed in HEPES/Tyrode's buffer (129 mM NaCl, 8.9 mM NaHCO 3 , 2.8 mM KCl, 0.8 mM KH 2 PO 4 , 56 mM dextrose, 10 mM HEPES, pH 7.4, 0.8 mM MgCl 2 ) with 1 mM prostaglandin E 1 . A total of 1 ϫ 10 10 platelets were resuspended in 1 ml of HEPES/Tyrode's buffer with 250 Ci of orthophosphate. After incubation at 30°C for 1 h, 10 ml of HEPES/ Tyrode's buffer with 1 M prostaglandin E 1 was added, and the platelet suspension was centrifuged at 500 ϫ g for 15 min. The platelets were then resuspended in HEPES/Tyrode's buffer, stimulated with 50 nM PMA at room temperature for 5 min, and then lysed with ice-cold 2 ϫ lysis buffer (1 ϫ lysis buffer contains 1% Triton, 10 mM Tris, pH 7.6, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 0.1% aprotinin, 1 mM vanadate, and 25 g/ml leupeptin). After lysis of the platelets, the pleckstrin was immunoprecipitated with rabbit anti-pleckstrin sera 354 raised against a recombinant protein corresponding to residues Glu 104 -Asp 233 of human pleckstrin (18). In the studies with COS-1 cells transiently expressing pleckstrin, the cells were incubated in phosphate-free media with 1 mCi/ml 32 P i for 3.5 h, stimulated with 50 nM PMA at 37°C for 10 min, and then mixed with 2 ϫ lysis buffer before the pleckstrin was immunoprecipitated.
CNBr mapping was performed on immunoprecipitated pleckstrin which was gel-purified by SDS-polyacrylamide gel electrophoresis. A gel slice containing the phosphopleckstrin was mixed with 50 mM ammonium carbonate (pH 8.5), 0.1% SDS, 1% ␤-mercaptoethanol. Pleckstrin contained in the eluate was passed through glass wool, trichloroacetic acid-precipitated, washed twice with a 50:50 mixture of cold ethanol:ether, and then vacuum-dried. The sample was resuspended in 30 l of 50 mg/ml CNBr in 70% formic acid and incubated at room temperature for 1 h. At this point, the digest was lyophilized with 1 ml of distilled water and again vacuum-dried. When noted, the CNBr fragments were further digested by incubating them overnight in the dark at room temperature in 30 l of 10 mg/ml iodosobenzoate (IBZO) in 80% acetic acid with 4 M guanidine HCl (26). All samples were then lyophilized with 1 ml of distilled water, vacuum-dried an additional three times, and fractionated on a Tricine gel (27).
Transfection of COS-1 Cells and Thrombin-induced Inositol Phosphate Production-This was performed as described previously (18). In brief, two 100-mm tissue culture plates of COS-1 cells were transfected using DEAE-dextran (28) with the human thrombin receptor in pCMV5 (a gift from Mark Stinski, University of Iowa) plus either wild type pleckstrin, variant pleckstrin, or an equivalent amount of empty plasmid. Twenty-four hours after transfection, the cells were trypsinized and reseeded into eight 60-mm tissue culture plates. [ 3 H]Inositol (4 Ci/ml, ICN) was added to six of the plates, then all of the cells were incubated at 37°C for an additional 18 h. The unlabeled cells were used to assess protein expression. The labeled cells were used to measure [ 3 H]inositol phosphate formation. Forty eight hours after transfection, the six [ 3 H]inositol-labeled plates were divided into three sets of dupli-cates which were extracted with perchloric acid either under resting condition, after stimulation by thrombin (2 units/ml) for 45 min, or after preincubation with 50 -100 nM PMA for 5 min followed by stimulation with thrombin, all in the presence of 20 mM LiCl. The neutralized extracts were applied to Dowex 1 columns, which were washed sequentially with 5 mM inositol and 5 mM sodium tetraborate, 60 mM ammonium formate. Total [ 3 H]inositol phosphate was eluted with 0.1 M formic acid plus 1.5 M ammonium formate and quantitated by scintillation counting. Equal levels of thrombin receptor expression in transfected cells was confirmed by flow cytometric analysis using ATAP2, a peptidedirected monoclonal antibody that recognizes the human thrombin receptor (29). Pleckstrin expression was assessed by immunoblotting cell lysates with rabbit anti-pleckstrin antibody.
Other Materials-Human ␣-thrombin (Ϸ3000 units/mg) was kindly provided by Dr. J. Fenton (New York State Dept. of Health, Albany, NY). Recombinant full-length pleckstrin was expressed in bacteria. The nucleotide sequence for pleckstrin was generated by reverse transcriptase-polymerase chain reaction of HL-60 total cellular RNA, and then cloned into a derivative of PET-11b. This plasmid directs the expression of the pleckstrin utilizing T7 polymerase and inserts a six histidine peptide tag at the N terminus of the expressed protein, allowing it to be purified over a nickel-agarose column. Stoichiometry of in vitro phosphorylated pleckstrin was performed using the Life Technologies, Inc. Protein Kinase C Assay System (Catalog No. 13161-013) and purified protein kinase C␣ (Life Technologies, Inc.) as described by the manufacturer.

RESULTS AND DISCUSSION
Phosphopeptide Mapping the Sites of Pleckstrin Phosphorylation-Previous studies have shown that pleckstrin is phosphorylated exclusively on serine and threonine residues (17). Two-dimensional electrophoresis of 32 P-labeled pleckstrin from platelets and of purified pleckstrin shows multiple species with different isoelectric points and specific activities, consistent with heterogeneous phosphorylation at multiple sites (17). In keeping with this conclusion, the predicted amino acid sequence of pleckstrin includes six potential sites for phosphorylation by protein kinase C, which prefers serine and threonine residues adjacent to basic residues (25). Three of these sites are within the N-terminal PH domain (Ser 40 , Ser 43 , and Thr 73 ). The other three are located in the intervening sequence between the two PH domains (Ser 113 , Thr 114 , and Ser 117 ) in a region ( 107 QKFARKSTRRSIRL 120 ) that closely resembles the protein kinase C pseudosubstrate site, RFARKGSLRQ (see Fig.  1) (25,30). To determine which of these potential sites actually becomes phosphorylated, pleckstrin was immunoprecipitated from platelets that had been labeled with [ 32 P]PO 4 and stimulated with PMA. The immunoprecipitate was digested with CNBr, which cleaves after methionine residues to produce 9 predicted fragments that were fractionated by gel electrophoresis and autoradiography (Fig. 1). The predominant radioactive fragment had an apparent molecular mass of approximately 8 kDa that corresponds to residues Phe 67 -Met 133 . In some experiments, two additional phosphorylated fragments were also observed. When present, these bands always contained Ͻ10% of the total incorporated 32 P. The first was a 19-kDa fragment corresponding to residues Ser 197 -Ala 344 . The second was a 10-kDa fragment corresponding to residues Ile 48 -Met 133 that was produced by incomplete cleavage at Met 66 .
Of the six potential sites for phosphorylation, four are located in the 8-kDa Phe 67 -Met 133 CNBr fragment: Thr 73 , Ser 113 , Thr 114 , and Ser 117 (Fig. 1). To determine which of these four sites becomes phosphorylated, the 32 P-labeled CNBr fragments were incubated with IBZO, which cleaves after tryptophan residues (26). IBZO is predicted to liberate two fragments of 3 kDa (Phe 67 -Trp 92 ) and 5 kDa (Val 93 -Met 133 ) from the 8-kDa CNBr fragment. As is shown in the right lane of Fig. 1A, only the 5-kDa fragment contained 32 P, effectively limiting the potential phosphorylation site(s) to residues Ser 113 , Thr 114 , and Ser 117 . Although not formally excluded by this part of the analysis, the only other serine and threonine residues within the 5-kDa CNBr/IBZO fragment are Thr 123 and Ser 132 , neither of which are predicted sites for phosphorylation by protein kinase C.
Identification of the Phosphorylation Sites by Site-directed Mutagenesis-To determine whether the candidate sites are, in fact, phosphorylated, COS-1 cells were transfected with wild type pleckstrin or pleckstrin in which the sites were individually or collectively replaced with glycine residues. Two days after transfection, the cells were labeled with [ 32 P]PO 4 and incubated briefly with PMA. As predicted by the CNBr/IBZO phosphopeptide analysis, simultaneous replacement of all three sites (denoted 3 Gly in the figures) caused a Ͼ90% decrease in 32 P incorporation into pleckstrin ( Fig. 2A). This decrease was not due to reduced expression or immunoprecipitation of the mutant (Fig. 2B). Interestingly, mutation of any one of these sites individually had little or no effect on 32 P incorporation ( Fig. 2A), while replacement of the sites two at a time resulted in diminished, but not absent, phosphorylation (data not shown). Since CNBr maps of the pleckstrin variants in which the sites were mutated individually were identical with those derived from wild type pleckstrin, the mutations appeared not to cause the phosphorylation of additional site(s) outside the Phe 67 -Met 133 CNBr fragment (Fig. 2C). In contrast, there was no phosphorylation apparent in the 8-kDa Phe 67 -Met 133 CNBr fragment from the 3 Gly pleckstrin variant, even when 4 times as much of the protein was applied to the electrophoresis gel (Fig. 2C, second lane from left). Taken together, these results suggest that phosphorylation occurs primarily at one or more of the three sites that were also identified by the CNBr/IBZO digest. To determine the stoichiometry of phosphorylation, purified recombinant pleckstrin was incubated with purified protein kinase C ␣ in the presence of [ 32 P]ATP. Digestion of the resultant phosphoprotein with CNBr produced an 8-kDa fragment identical with that isolated from PMA-stimu-lated 32 P-labeled platelets. The stoichiometry of phosphate incorporation was approximately 2:1 (data not shown).
Previous studies have shown that phosphorylated pleckstrin is comprised of multiple isoforms that can be resolved by twodimensional electrophoresis (17). Together with the present results, this suggests that pleckstrin normally becomes phosphorylated on residues Ser 113 , Thr 114 , and Ser 117 near, but not within, the N-terminal PH domain and that no one site is predominant. Since elimination of any one of these sites has no effect on overall phosphorylation, there is apparently a compensatory increase in phosphorylation at the other sites, which is also consistent with the observation that the stoichiometry of phosphorylation of recombinant pleckstrin is 2:1 rather than 3:1. An alternative interpretation of the data, which is that phosphorylation actually occurs at unidentified sites elsewhere in the molecule and is indirectly affected by mutagenesis of Ser 113 , Thr 114 , and/or Ser 117 is less likely when the results of the mutagenesis studies are combined with those from the CNBr/IBZO double digest of 32 P-labeled platelet pleckstrin. The minor CNBr fragment that at times becomes phosphorylated, Ser 197 -Ala 344 , accounts for Ͻ10% of total phosphorylation and does not show increased phosphorylation when Ser 113 , Thr 114 , and Ser 117 are individually mutated.
Phosphorylation and Function-Previous reports have shown that a brief exposure to the phorbol ester, PMA, will inhibit agonist-induced phosphoinositide hydrolysis in a variety of cells, including platelets (31)(32)(33). This effect is seen in cells that express pleckstrin as well as those that do not and has been attributed variously to the phosphorylation of receptors, G ␣ and/or phospholipase C (19 -22). We have shown previously that when expressed in COS-1 or HEK-293 cells pleckstrin will inhibit phosphoinositide hydrolysis initiated by thrombin receptors, angiotensin-II AT 1A receptors, muscarinic M1 receptors, and ␣ 2A receptors, as well as by TrkA receptors

FIG. 1. CNBr and IBZO digest of [ 32 P]pleckstrin from platelets.
A, pleckstrin that had been immunoprecipitated from 32 P-labeled platelets with an anti-pleckstrin antibody was digested with CNBr alone or with CNBr followed by iodosobenzoate (IBzo). The digest was electrophoresed on a Tricine polyacrylamide gel and analyzed by autoradiography. B, a map of pleckstrin highlighting the CNBr and IBZO cleavage sites, as well as the locations of the PH domains and potential sites of phosphorylation.

FIG. 2. Phosphorylation mutants in vivo.
COS-1 cells were transfected with wild type pleckstrin or pleckstrin variants in which the 3 potential sites were individually (S113G, T114G, and S117G) or collectively (3 Gly) mutated to glycine. A shows the relative phosphorylation of the pleckstrin variants when the cells were labeled with [ 32 P]PO 4 and stimulated with 50 nM PMA. The expressed pleckstrin was immunoprecipitated, fractionated by SDS-polyacrylamide gel electrophoresis, and quantitated by a PhosphorImager. B shows a rabbit anti-pleckstrin immunoblot demonstrating equivalent immunoprecipitation of the wild type and variant pleckstrin. C shows that the phosphorylated CNBr fragments produced from wild type and variant pleckstrin are similar in size to each other and to those produced from platelet pleckstrin. Note that the amount of protein applied to lane 2 of C is 4 times that applied to the other lanes. Despite this, there was no phosphorylation on the major 8-kDa CNBr fragment in the triple glycine mutant. and a constitutively active variant of G q␣ , while having no effect on the regulation of cAMP formation (18). Since agonists that stimulate phosphoinositide hydrolysis cause pleckstrin to become phosphorylated, we next examined the impact of phosphorylation on pleckstrin's ability to inhibit inositol phosphate formation.
To accomplish this, COS-1 cells were transfected with thrombin receptors and either wild type pleckstrin or the pleckstrin variant described above in which all three sites of phosphorylation were mutated to glycine. In the absence of PMA, wild type pleckstrin inhibited thrombin-induced [ 3 H]inositol phosphate formation by 46% (Fig. 3). As we have shown previously, PMA in the absence of pleckstrin, or pleckstrin in the absence of PMA, each inhibited thrombin-induced phosphoinositide hydrolysis by 40 -50%. The combination of pleckstrin plus PMA was additive, inhibiting thrombin-induced [ 3 H]inositol phosphate formation by 91%. Keeping in mind that thrombin, like PMA, causes the phosphorylation of pleckstrin, we found that replacing residues Ser 113 , Thr 114 , and Ser 117 with glycine (3 Gly) reduced pleckstrin's ability to inhibit phosphoinositide hydrolysis by at least two-thirds in either the presence or absence of PMA (Fig. 3B). This was not due to a decrease in the level of pleckstrin expression, which was the same for the glycine variant as it was for wild type pleckstrin (Fig. 3C). Replacing the sites of phosphorylation with another neutral amino acid, alanine, had a similar effect (data not shown). These results suggest that phosphorylation is required for max-imal inhibition of thrombin-induced phosphoinositide hydrolysis by pleckstrin, but also show that some inhibition can occur even in the absence of phosphorylation.
Finally, to assess the importance of any one of the phosphorylation sites, COS-1 cells were transfected with the pleckstrin variants in which the phosphorylation sites were mutated to glycine individually or in pairs. The results that were obtained mirrored the effects on phosphorylation: the single-site mutants behaved like wild type pleckstrin and the double-site mutants had an intermediate effect (data not shown). These results suggest that the phosphorylation of at least two of the sites is required for maximal pleckstrin activity and that this phosphorylation can occur on any two of the three sites.
Replacement of Phosphorylation Sites with Charged Residues-Since the pleckstrin expressed in the transfected COS-1 cells becomes phosphorylated when the cells are stimulated with thrombin, we next examined whether replacing the sites of phosphorylation with charged, rather than neutral, amino acids would mimic the effects of phosphorylation. Pleckstrin variants in which all three sites were mutated to either positively charged lysine residues or negatively charged glutamate residues were expressed in COS-1 cells and compared with wild type pleckstrin and the glycine-mutated pleckstrin for their ability to inhibit thrombin-induced phosphoinositide hydrolysis. In contrast to the "uncharged" glycine variants, the "charged" variants were at least as active as wild type pleckstrin (compare Figs. 3 and 4), showing that the effects of phosphorylation can be mimicked by introducing positive or negative charges into the regulatory region adjacent to the N-terminal PH domain. Similar results were obtained when these sites were replaced with aspartate residues (data not shown). Since pleckstrin expressed in COS-1 cells is almost maximally phosphorylated after thrombin stimulation, the charged variants, unlike the glycine and alanine variants, possess activity roughly comparable to wild type phosphopleckstrin.
Conclusion-In summary, these observations demonstrate that pleckstrin is principally phosphorylated by protein kinase C on a cluster of two serine and one threonine residues located adjacent to, but not within, the N-terminal PH domain. The stoichiometry of phosphorylation is approximately 2:1, and elimination of any one of the three potential sites appears to result in a compensatory increase in phosphorylation of the other two. The various combinations of phosphorylation at two out of a total of three potential sites accounts in part for the multiplicity of phosphorylated pleckstrin isoforms previously demonstrated by Haslam and co-workers (17) using isoelectric focusing. Inhibiting pleckstrin's ability to undergo phosphorylation by replacing the target serine and threonine residues with neutral glycine or alanine residues greatly limits pleckstrin's ability to inhibit thrombin-induced phosphoinositide hydrolysis, while replacing the phosphorylation sites with charged lysine, glutamate, or aspartate residues mimics the effects of phosphorylation. Although the present studies focused on thrombin receptors, previous results obtained with other G-protein-coupled or growth factor receptors that couple to phospholipase C ␤ and phospholipase C ␥ show that this phenomenon is not limited to thrombin-initiated phosphoinositide hydrolysis (18).
These results raise the question of how phosphorylation regulates pleckstrin's activity. It is conceivable that the phosphorylation of pleckstrin affects its cellular location, in the process affecting its proposed interaction with PIP 2 . However, preliminary studies in platelets and pleckstrin-transfected COS-1 cells indicate that this is not the case. 3 We have previously shown that the N-terminal PH domain of pleckstrin is critical for its ability to inhibit phosphoinositide hydrolysis. Since the N-terminal PH domain and the 3 sites of phosphorylation are close to each other in the linear sequence, one alternative explanation is that the first portion of the intervening sequence between the PH domains obstructs the access of ligands to binding sites within the N-terminal PH domain. Under this model, phosphorylation relieves this obstruction, allowing the N-terminal PH domain to interact with other molecules necessary for the inhibition of phosphoinositide hydrolysis. Alternatively, the first part of the intervening sequence may undergo a conformational change upon phosphorylation that affects necessary spatial relations between the N-terminal and C-terminal PH domains or an as-yet-unidentified accessory protein.
The PH domain from the N-terminal PH domain of pleckstrin, like that from dynamin and ␤-spectrin, possesses a highly polarized electrostatic potential. It is possible that this asymmetry of charges contributes to an interaction between the N-terminal PH domain and the charged residues surrounding the protein kinase C phosphorylation site. However, regardless of which model ultimately proves to be correct, phosphorylation appears to affect pleckstrin activity by altering the charge distribution in a region proximal to the N-terminal PH domain. Somewhat surprisingly, the effects of phosphorylation could be mimicked by replacing the phosphorylation sites with either a negatively charged glutamate residue or a positively charged lysine residue. This suggests that it is the presence of charged residues in this region of pleckstrin that is important and not just the negative charge carried by the phosphate groups.
Finally, what is the role of pleckstrin in cells that normally contain it, such as platelets and other blood cells? The rapid phosphorylation of pleckstrin is one of the hallmarks of platelet activation and is thought to occur when diacylglycerol and Ca 2ϩ activate protein kinase C. This suggests that phosphorylated pleckstrin could play a regulatory role, helping to limit the duration of agonist-induced phospholipase C activity and perhaps preventing premature platelet activation. Whether it does so in blood cells remains to be demonstrated.