J Biol Chem, Vol. 274, Issue 31, 21515-21518, July 30, 1999

COMMUNICATION
Pleckstrin 2, a Widely Expressed Paralog of Pleckstrin Involved in Actin Rearrangement^*

Michael H. Hu, Eve Marie Bauman, Richard L. Roll, Newman Yeilding, and Charles S. Abrams

From the Department of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

We have identified a cDNA for pleckstrin 2 that is 39% identical and 65% homologous to the original pleckstrin. Like the original pleckstrin 1, this protein contains a pleckstrin homology (PH) domain at each end of the molecule as well as a DEP (Dishevelled, Egl-10, and pleckstrin) domain in the intervening sequence. A Northern blot probed with the full-length cDNA reveals that this homolog is ubiquitously expressed and is most abundant in the thymus, large bowel, small bowel, stomach, and prostate. Unlike pleckstrin 1, this newly discovered protein does not contain obvious sites of PKC phosphorylation, and in transfected Cos-7 cells, it is a poor substrate for phosphorylation, even after PMA stimulation. Cells expressing pleckstrin 2 undergo a dramatic shape change associated with actin rearrangement, including a loss of central F-actin and a redistribution of actin toward the cell cortex. Overexpression of pleckstrin 2 causes large lamellipodia and peripheral ruffle formation. A variant of pleckstrin 2 lacking both PH domains still had some membrane binding but did not efficiently induce lamellipodia, suggesting that the PH domains of pleckstrin 2 contribute to lamellipodia formation. This work describes a novel, widely expressed, membrane-associating protein and suggests that pleckstrin 2 may help orchestrate cytoskeletal arrangement.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

Pleckstrin homology, or PH,¹ domains are amino acid motifs that are capable of binding polyphosphoinositides and regulating protein function (1-6). Frequently, the binding of polyphosphoinositides to the PH domains within a protein localizes the molecules to the cell membrane (7). In addition, some PH domains may interact with other targets such as the / subunits of heterotrimeric G proteins (8-11) or protein kinase C (12-14). The structure of several PH domains complexed to inositol trisphosphate has been solved (15, 16), confirming a physical interaction between the inositol phosphate head-group and the positively charged face of the PH domain.

Pleckstrin 1 is a 40-kDa protein containing the prototypic PH domains at its amino and carboxyl termini. It was first described as a major substrate for protein kinase C (PKC) in platelets and leukocytes, and its phosphorylation has long been used as a marker for platelet activation. Although its function in vivo is unclear, heterologously expressed pleckstrin can affect second messenger-based signaling events mediated by phospholipase C, PI3K, and 5'-inositol phosphatases (17-19). Overexpression and microinjection studies suggest that pleckstrin 1 is membrane-localized, induces a shift of F-actin toward the cell cortex, and participates in the production of lamellipodia (3). These functions are tightly regulated by PKC-mediated phosphorylation of three residues (Ser¹¹³, Thr¹¹⁴, and Ser¹¹⁷) located near, but not within, the amino-terminal PH domain (24). Recently, a DEP (Dishevelled, Egl-10, and pleckstrin) domain has been described in pleckstrin 1, but the function of this motif is unknown (20).

Regulation of pleckstrin 1 is unique because it is the only described protein in which the PH domain is controlled by adjacent phosphorylation sites. In addition, tissue distribution studies show that pleckstrin 1 is restricted to blood cells. We now describe a new protein with homology to pleckstrin 1. In contrast to pleckstrin 1, pleckstrin 2 is found in a wide variety of tissues and is not an efficient substrate for PKC. In addition, it is bound to the cell membrane, and causes the production of large lamellipodia.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

Cloning of Mammalian Expression Vectors and Northern Blotting-- Identification of a pleckstrin-homologous clone was done using the frame search option of the Embl-Heidelberg Bioccelerator and using the original human pleckstrin as the query. Multiple EST clones containing the murine pleckstrin-2 cDNA were obtained from Genome Systems (St. Louis, MO), and a fragment of cDNA corresponding to the largest open frame was generated by PCR with the following primers: gc ggc aag ctt tta agc gta gtc tgg tac gtc gta tgg gta tgt tag ctt ctt gat agc and cg cca gaa ttc cca gct gct gag agg agt tgc ctg aga gtg accttt gca tctgcc tgt cca gcc agc atg gag gac ggc gtg ctc. The primers incorporated HindIII and EcoRI restriction sites that facilitated cloning into pCMV5, a 5'-untranslated region that allowed efficient expression, and a carboxyl-terminal HA (YPYDVPDYA) epitope. A plasmid directing the expression of HA epitope-tagged double PH domain deletion variant (6-102, 249-352) was generated by PCR. The sequences of all clones were fully confirmed. The human multiple tissue-specific Northern blot was purchased from CLONTECH (Palo Alto, CA).

Transfection and Immunoprecipitation-- 100-mm plates containing Cos-7 cells were transfected using the Ca²⁺ phosphate method. 24 h after transfection, the cells were shocked with 10% glycerol and analyzed again 24 h later. Immunoprecipitation was performed by lyzing cells in boiling 1% SDS, 10 mM Tris, pH 7.2, and 2 mM EDTA for 15 min, followed by clarification at 13,000 × g for 15 min. The supernatant was diluted in immunoprecipitation buffer (1% Triton X-100, 1% deoxycholic acid, 0.1% SDS, 158 mM NaCl, 10 mM Tris, pH 7.2, 5 mM Na-EDTA) containing 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, and 0.1% aprotinin. The lysates were precleared of nonspecific proteins by three washes with pansorbin, followed by immunoprecipitation with murine HA-11 anti-HA antibody (Babco, Berkeley, CA).

In Vivo Phosphorylation-- 24 h after transfection, 100-mm plates containing COS-7 cells transfected with pleckstrin-1, pleckstrin-2, or empty vector were each split into triplicate 60-mm plates. 24 h later, two of each set of three plates were incubated for 2 h with phosphate-free medium and 1 mCi/ml ³²P-labeled free orthophosphate. PMA was then added at a concentration of 50 nM to one plate of each pair to stimulate PKC activity. After a 30-min incubation, the cells were lysed and the HA epitope tagged proteins were immunoprecipitated as described above. The precipitated proteins were then fractionated by SDS-PAGE and exposed on autoradiography film. In parallel, the nonradioactive plate of each triplicate was lysed, immunoprecipitated with anti-HA antibody, fractionated by SDS-PAGE, and immunoblotted with anti-HA antibody.

Immunofluorecence-- This was performed as described previously (3).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

Pleckstrin 2 Is a Widely Expressed Pleckstrin Paralog-- When platelets are activated by agonists such as thrombin, pleckstrin 1 is phosphorylated by protein kinase C, and then regulates polyphosphoinositide second messenger formation and induces actin rearrangement (3, 17, 18). To look for pleckstrin homologs, we used the original human pleckstrin 1 as the query, and searched for pleckstrin-homologous clones using the frame search option of the Embl-Heidelberg Bioccelerator. This search revealed several potential murine EST clones including: AA051380, AI385784, AA798712, AA403397, and AA008011. Further analysis indicated that these clones were overlapping and represented the same transcript. The mRNA contains an open reading frame of 1107 nucleotides and an ATG codon at positions 54-56 that conforms to Kozak's rules for translation initiation (21). The amino acid sequence predicted from the open reading frame is 353 residues in length, terminating with a TGA codon at nucleotide positions 1113-1115. An alternative splice form was also present that was missing amino acids 293-312.

A homology search revealed that the translational product was highly related to human pleckstrin 1. As shown in Fig. 1A, the translational product is 39% identical and 65% homologous over the full length of the protein. Like pleckstrin 1, it contains an amino- and a carboxyl-terminal PH domain plus an intervening DEP domain. When aligned with the corresponding regions in the original pleckstrin, the amino-terminal PH domain is 35% identical and 61% homologous, the DEP domain is 33% identical and 66% homologous, and the carboxyl-terminal PH domain is 57% identical and 82% homologous. Because of the strong primary sequence and organizational homology to pleckstrin 1, we have named this clone pleckstrin 2. 

View larger version (48K): [in this window] [in a new window]   Fig. 1.   A, alignment of human pleckstrin 1 with murine pleckstrin 2. Identical amino acids are marked with an asterisk (*), and similar amino acids are denoted with a caret (); this shows that alignment between human pleckstrin 1 and murine pleckstrin 2 reveals 39% identity and 65% homology on the amino acid level. Shown in boldface are the sites of phosphorylation of the original human pleckstrin and the homologous residues in pleckstrin 2. The PH and DEP domains are underlined by solid or dotted lines, respectively. B, detection of pleckstrin 2 mRNA. Northern blots of various human tissues were hybridized with 32P-labeled antisense DNA from the open reading frame of murine pleckstrin 2 and autoradiographed. PBL, peripheral blood leukocytes.

Further analysis of the EST data bank demonstrated several partial human pleckstrin 2 clones including AA226122, AA143492, and AA152444. The amino acid sequence predicted from the long open reading frame from human pleckstrin 2 is completely identical to the amino acid sequence of murine pleckstrin 2 and is 39% identical to the original human pleckstrin. Thus, although the original pleckstrin and pleckstrin 2 are highly homologous, they are clearly different proteins.

Despite overall organizational homology, there is little similarity between the regulatory region of pleckstrin 1 (amino acids 107-117) and the comparable region in pleckstrin 2. This region in pleckstrin 1 contains three sites of phosphorylation, Ser¹¹³, Thr¹¹⁴, and Ser¹¹⁷, preceded by a pseudo-substrate site for PKC composed of basic amino acids. Pleckstrin 2 does not contain the highly basic region and possesses only a single potential phosphorylation site comparable with Ser¹¹⁷.

Previous studies indicated that the pleckstrin 1 transcript and pleckstrin 1 protein are present in platelets and various leukocytes (22, 23). To determine the tissue distribution of pleckstrin 2, we hybridized mRNA isolated from a variety of normal human tissues with a probe representing the open reading frame of pleckstrin 2. As shown in Fig. 1B, a 1.7-kb transcript was readily detected in the thymus, prostate, testis, ovary, small bowel, and large bowel. In other Northern blots, expression was apparent in the stomach, bone marrow, kidney, and uterus (not shown). Preliminary experiments using reverse transcription-PCR of RNA from several murine tissues reveal that both the long and short splice variants are present, although the distribution of the two isoforms does vary (data not shown). In contrast to the 1.7-kb transcript of pleckstrin 2, a previous report showed that the pleckstrin 1 transcript was 3.0 kb (22). This different tissue distribution and transcription size provide additional evidence that pleckstrin 2 is a distinct protein from pleckstrin 1.

Pleckstrin 2 Is Not an Efficient Substrate of PKC-- Pleckstrin 1 is a substrate of multiple isoforms for PKC, and its phosphorylation has been used as an early marker of platelet activation. We have mapped three sites of phosphorylation of pleckstrin 1 (Ser¹¹³, Thr¹¹⁴, and Ser¹¹⁷) that are located near its amino-terminal PH domain (24). Although pleckstrin 2 does have a single serine at a position homologous to Ser¹¹⁷ in the original pleckstrin, pleckstrin 2 lacks the consensus PKC pseudo-substrate site. We tested the ability of pleckstrin 2 to become phosphorylated in vivo. Cos-7 cells were transfected with epitope-tagged HA-pleckstrin 1 or HA-pleckstrin 2, labeled with ³²P_i, and stimulated with PMA. The HA epitope-tagged proteins were then immunoprecipitated and analyzed for incorporation of radioactive P_i. As shown on the left panel of Fig. 2, considerable amounts of HA-pleckstrin 1 and HA-pleckstrin 2 were expressed and immunoprecipitated. However, as seen on the right panel of Fig. 2, whereas pleckstrin 1 incorporated substantial amounts of ³²P, pleckstrin 2, even in the presence of PMA, incorporated little label. In several experiments, a faint phosphoprotein could sometimes be detected in the pleckstrin 2 immunoprecipitates, suggesting that phosphorylation of pleckstrin 2 might occur but is inefficient under the conditions of our experiment. However, we cannot exclude the possibility that pleckstrin 2 is phosphorylated by a kinase other than PKC, one that is not active in our transfected cells. There was no difference in ³²P incorporation when the shorter splice variant of pleckstrin 2 was used instead of the full-length pleckstrin 2 isoform (data not shown).

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Phosphorylation of pleckstrins in vivo. Cos-7 cells were transfected with HA epitope-tagged pleckstrin 1 or pleckstrin 2. The left panel shows an anti-HA immunoblot demonstrating roughly equivalent expression and immunoprecipitation of both pleckstrins. The right panel shows the relative phosphorylation of the pleckstrins when the cells were labeled with 32P-PO4 with, and without, stimulation with 50 nM PMA. The expressed pleckstrin was immunoprecipitated, fractionated by SDS-PAGE, and autoradiographed.

Also as shown in Fig. 2, the mobility of pleckstrin 2 during PAGE was faster than predicted, and it was faster than the mobility of comparably sized pleckstrin 1. Because the DNA sequencing of pleckstrin 2 was confirmed on several occasions, it is probable that the difference in apparent size reflects an alternative folding or post-translational processing of the protein. Thus, although pleckstrin 2 is highly homologous to pleckstrin, there are fundamental differences in the regulation of these proteins.

Pleckstrin 2 Is Membrane-associated and Contributes to Lamellipodia Formation-- We previously reported that pleckstrin 1 associates with the plasma membrane in a phosphorylation-dependent fashion (3). To determine whether pleckstrin 2 also associates with the cell membrane, Cos-7 cells were transfected with either wild type pleckstrin 1, wild type, full-length pleckstrin 2, the shorter splice variant of pleckstrin 2, or a mutant form of pleckstrin 2 lacking both PH domains. As shown in Fig. 3, full-length pleckstrin 1 and pleckstrin 2 were membrane-associated. However, there was a difference between the two proteins. Cells overexpressing pleckstrin 1 were flat with some villous projections, and pleckstrin 1 was uniformly distributed over the entire surface membrane including lamellipodia. In contrast, pleckstrin 2-expressing cells had more microvilli and large lamellipodia with ruffle formation. Both types of pleckstrin were predominantly membrane-bound. In contrast, cells expressing the shorter form of pleckstrin 2 were rounder but still had large lamellipodia. The majority of the expressed shorter isoforms of pleckstrin 2 appeared on the membranes of both the cell body and the nucleus but not on the membrane of the lamellipodia. Thus, pleckstrin 1 and both forms of pleckstrin 2 all contribute to lamellipodia formation, but in contrast to the original pleckstrin, pleckstrin 2 induces more microvilli and ruffles.

View larger version (87K): [in this window] [in a new window]   Fig. 3.   Intracellular localization of wild type and variant forms of pleckstrin 2. Indirect immunofluorescence was performed on Cos-7 cells transiently transfected with HA-tagged pleckstrin (Plk) variants, using the 12CA5 monoclonal antibody against the HA epitope. Cells were simultaneously stained with rhodamine-conjugated phalloidin. Each panel was photographed at 20×. Arrows mark transiently transfected cells. These results show that pleckstrin 1 and pleckstrin 2 target ligands that are found in the cell membrane and induce actin rearrangement.

Because of marked ruffle formation, we examined the effect of pleckstrin 2 on actin distribution. As shown by phalloidin staining in Fig. 3, mock-transfected Cos-7 cells exhibited stress fiber formation with thick central F-actin. In contrast, the actin in cells expressing either splice form of pleckstrin 2 redistributed toward a cortical pattern. When taken with the changes in cell morphology, this suggests that pleckstrin 2 causes redistribution of actin within cells.

The Role of the PH Domains within Pleckstrin 2-- Pleckstrin 1, like other PH domain-containing proteins, targets the cell membrane by a PH domain-dependent interaction. We hypothesized that the association of pleckstrin 2 with the cell membrane would require its PH domains. To test this hypothesis, we expressed a pleckstrin 2 variant lacking both PH domains and examined its intracellular distribution. The pleckstrin 2 mutant was uniformly distributed throughout the cell including the nucleus and the surface membrane (Fig. 3). At this point, we are not able to discern whether this distribution is a consequence of overexpression. If the mutant lacking both PH domains does associate with the cell membrane, it implies that pleckstrin 2 contains a membrane-targeting motif in the region between its two PH domains. It is notable that the region between the two PH domains is composed largely of the DEP homology motif. The morphology and actin staining of cells expressing this mutant also suggest that the PH domains of pleckstrin 2 are required for its ability to efficiently induce lamellipodia formation.

	CONCLUSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

Although pleckstrin 1 has long been known as a substrate of protein kinase C in hematopoietic cells, until now there has never been a known non-hematopoietic paralog. In that context, the present studies demonstrate that: 1) like the original pleckstrin, pleckstrin 2 contains a PH-DEP-PH organization; 2) its transcript is most abundant in the thymus, stomach, large and small bowels, and prostate; 3) in contrast to the original pleckstrin, it is not efficiently phosphorylated by PKC; and 4) it induces lamellipodia and ruffle formation. These observations raise a number of issues, including the mechanism by which pleckstrin 2 induces actin redistribution and the mechanism by which pleckstrin 2 is regulated.

The first issue is the mechanism of pleckstrin 2-mediated actin changes. The original pleckstrin also induces cytoskeletal reorganization, although it tends to contribute to less lamellipodia. In the case of the original pleckstrin, many of its effects can be inhibited with a dominant-negative Rac and are independent of PI3K.² Whether pleckstrin 2-mediated cytoskeletal changes involve Rac is currently unknown as is the role of other polyphosphoinositide-binding proteins such as gelson or villin. The observation that the pleckstrin 2 truncated variants of both PH domains may still bind to the cell membrane was an unexpected finding. Whether the association of this mutant with the cell membrane is mediated by its DEP domain is an area for further study, as are the circumstances that allow different regions of pleckstrin 2 to direct localization to different structures.

The second issue is the mechanism by which pleckstrin 2 is regulated and the related issue of the identity of its potential binding partners. Pleckstrin 1 is phosphorylated by multiple isoforms of PKC, and its ability to induce actin reorganization is regulated by that phosphorylation. Pleckstrin 2 is not efficiently phosphorylated by PKC, and thus it requires an alternative means of regulation. During the preparation of this work, we became aware of the results by E. Skolnik and co-workers (25) who used a genetic screen for proteins that bind PI3K products. By this assay, both the amino- and carboxyl-terminal PH domains of pleckstrin 2 bind D3-containing phosphoinositides. It therefore seems reasonable to speculate that pleckstrin 2 is regulated by PI3,4P₂ or PIP₃ and may in turn be an effector for PI3K. We are currently investigating whether the intracellular distribution of pleckstrin 2 is affected by PI3K-mediated signaling and whether pleckstrin 2 moderates any PIP₃-dependent signaling events such as the activation of Akt, PKC, or mitogen-activated protein kinase.

	ACKNOWLEDGEMENTS

The authors thank Dr. Toby Gibson (EMBL Heidelberg) for the initial alignments on the Embl-Heidelberg Bioccelerator and Drs. Joel S. Bennett and Lawrence F. Brass for their helpful comments on this manuscript.

	FOOTNOTES

^* These studies were supported in part by Grants P50 HL54500 and P01 HL40387 (to C. S. A.) from the National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Hematology-Oncology Div., University of Pennsylvania, Basic Research Bldg. 2/3, Rm. 912, 421 Curie Blvd., Philadelphia, PA 19104. Tel.: 215-898-1058; Fax: 215-573-7400; E-mail: abramsc@mail.med.upenn.edu.

² A. D. Ma and C. S. Abrams, submitted for publication.

	ABBREVIATIONS

The abbreviations used are: PH domain, pleckstrin homology domain; DEP, Dishevelled, Egl-10, and pleckstrin; EST, express sequence tag; kb, kilobase (pairs); PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction; PIP₃, phosphatidylinositol 3,4,5-trisphosphate; PKC, protein kinase C; PMA, phorbol 12-myristate 13-acetate; PI3K, phosphatidylinositol 3-kinase; HA, hemagglutinin antigen.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES