Actin binding of human LIM and SH3 Protein is regulated by cGMP- and cAMP-dependent Protein Kinase phosphorylation on Serine 146

cGMP- and protein (cAK) of PtK-2 and cGK confirmed phosphorylation of and F-actin of of the tips of extensions of the mutant S146D resulted in nearly complete relocalization to the cytosol and reduced migration of the cells. together, these that phosphorylation of LASP by cGK and cAK may be involved in cytoskeletal organization and motility. Here we report the identification of a specific substrate for cAK and cGK in intact human platelets using differential phosphoproteomic display of radiolabeled human platelets. The protein was identified as the LIM and SH3 domain protein (LASP) (27) by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). Phosphorylation of LASP at Ser-146 leads to a redistribution of the actin-bound protein from the tips of the cell membrane to the cytosol accompanied with a reduced cell migration. (RRD S efficient cGK phosphorylation human LASP Ser-146 is phosphorylated by cAK and cGK in . Using site-directed mutagenesis of all serine residues, we phosphorylation at Ser-99 and at Ser-61 in , although Ser-61 was phosphorylated in serine in position is only found in human and rabbit, while the corresponding amino acid in mouse and rat is an alanine. In contrast, Ser-99 and Ser-61 are conserved in all four and co-workers identified Ser-146 and Ser-99 as the major in vitro and in vivo phosphorylation sites of rabbit LASP by cAK (34). Phosphorylation of rabbit LASP at Ser-146 induced a Mr band shift that is absent in human LASP phosphorylated by cAK and cGK indicating differences in the structure of the two proteins. Further studies are underway to explore phosphorylation of LASP in the with F-actin-rich compartments, modulating platelet function.


SUMMARY
Various drugs that elevate cGMP levels and activate cGMP-dependent protein kinase (cGK) inhibit agonist-induced platelet activation. In the present study we identified the LIM and SH3 domain protein (LASP) that was recently cloned from human breast cancer cells

INTRODUCTION
The activation process of human platelets and vessel wall -platelet interactions are tightly regulated under physiological conditions and are often impaired in thrombosis, arteriosclerosis, hypertension, and diabetes.
Platelet activation can be inhibited by a variety of agents, including aspirin and Ca 2+ antagonists as well as cGMP-and cAMP-elevating agents such as NO and prostaglandin I 2 (for review see Ref. 1). The inhibitory effects of cGMP and cAMP are mediated by cAMP-dependent protein kinases types I and IIβ and by cGMP-dependent protein kinase Iβ (cAK I, cAK IIβ, and cGK Iβ, respectively), representing the major forms of cyclic nucleotide-dependent protein kinases in human platelets (2,3).
The molecular mechanisms of platelet inhibition by cGMP signaling downstream of cGK activation are only partially understood. In cGK-deficient mice cGMP-mediated inhibition of platelet aggregation is impaired (4). To date, only a few substrates for cAK and cGK have been identified and characterized in human platelets. The 22-kDa small GTP-binding protein rap 1b is phosphorylated by cAK and cGK in intact platelets (5, this study). Phosphorylation of rap 1b is associated with translocation of the protein from the membrane to the cytosol (6).
The vasodilator-stimulated phosphoprotein VASP is another major substrate of cAK and cGK in human platelets (7). Its three phosphorylation sites are phosphorylated with different specificities by these two kinases (8). VASP phosphorylation is thought to be involved in the negative regulation of integrin α IIb b III (9). Experiments in vitro revealed reduced F-actin binding and actin polymerization of phosphorylated VASP (10). Two recent studies investigated RhoA-mediated myosin light chain (MLC) activation and its contribution to platelet aggregation and secretion, showing that cGK phosphorylates RhoA and counteracts the phosphorylation of myosin light chain (MLC) through activation of MLC phosphatase (11,12). We previously identified heat shock protein 27 (Hsp27) as a substrate for cGK in intact platelets (13). Phosphorylation of Hsp27 by cGK reduced the stimulatory effect of MAPKAP kinase 2-phosphorylated Hsp27 on actin polymerization. There is also evidence that at least part of the inhibitory response mediated by cGK depends on phosphorylation of the thromboxane receptor (14,15) and the IP 3 -receptor (16).
Several other proteins have been reported to be phosphorylated in response to cGK activation either in vitro or in intact cells, including cGMP-specific phosphodiesterase (2), MLC kinase (17), an IP 3 receptor-associated cGMP kinase substrate (18), Na + /K + -ATPase (19), cysteine rich protein 2 (20), MEKK1 (21), and endothelial NO synthase (22). None of these proteins, however, have been established as a downstream target of cGK in platelets.
Here we report the identification of a specific substrate for cAK and cGK in intact human platelets using differential phosphoproteomic display of radiolabeled human platelets. The protein was identified as the LIM and SH3 domain protein (LASP) (27) by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). Phosphorylation of LASP at Ser-146 leads to a redistribution of the actin-bound protein from the tips of the cell membrane to the cytosol accompanied with a reduced cell migration.
All other chemicals, reagents, and solvents of the highest purity available were purchased from Sigma (Deisenhofen, Germany).
cGK Iα and the catalytic subunit of cAK type II were purified from bovine lung and bovine heart, respectively (23). cGK Iβ and cGK II were expressed in and purified from the baculovirus-Sf9 cell system (24).
Isolation of plateletsFreshly donated blood from healthy volunteers (50 ml) was collected in acid-citrate dextrose and centrifuged for 10 min at 300 x g to yield platelet-rich plasma (PRP). PRP was centrifuged for 20 min at 500 x g and the platelet pellet was resuspended and washed once in an isotonic buffer containing 10 mM Hepes (pH 7.4), 137 mM NaCl, 2.7 mM KCl, 5.5 mM glucose and 1 mM EDTA at a density of 1 x 10 9 cells/ml. After resuspension, platelets were allowed to rest at 37°C for 15 min.
32 P-Labeling of plateletsPlatelet preparation was carried out essentially as described above. After washing, 1 ml of platelets at a concentration of 1 x 10 9 /ml was incubated with 500 µCi [ 32 P]orthophosphate (HCl-free) for 1.5 h at 37°C. Platelets were then centrifuged at 500 x g for 7 min and resuspended in 1 ml isotonic buffer. Aliquots of 100 µl (corresponding to 200 µg protein) were treated with 500 µM 8pCPT-cGMP for 30 min at 37°C. After stimulation, platelets were briefly centrifuged (500 x g for 3 min) to yield a pellet. After equilibration in 50 mM Tris, pH 8.9, 6 M urea, 30% (w/v) glycerol, 2% (w/v) SDS, gels were immediately applied to a vertical 10% SDS gel without a stacking gel. Electrophoresis was carried out at 8°C with a constant current of 40 mA per gel. The gels of radioactively labeled platelet proteins were fixed in 30% ethanol, 10% acetic acid and exposed. Radioactive spots visualized by autoradiography were excised.
Mass-SpectrometryGel pieces were washed sequentially for 10 min in tryptic digestion buffer (10 mM NH 4 HCO 3 ) and digestion buffer: acetonitrile 1:1. These steps were repeated three times and led to a shrinking of the gel. It was reswollen with 2 µl protease solution (trypsin at 0.05 µg/µl) in digestion buffer and incubated overnight at 37°C.
Analysis of the resulting peptides was carried out using a nano-HPLC system coupled directly to an ESI-iontrap mass spectrometer equipped with a custom-built nanoelectrospray ion source (LCQ" Classic, Thermo Finnigan, San Jose, USA). Fifteen µl of 5% Expression of GST-LASP fusion proteinsRecombinant LASP and LASP mutants S61D, S146D, and S61/146D were expressed in E.coli as GST fusion proteins using pGEX-4T1.
Expression and purification of the GST fusion proteins were performed according to the manufacturer's protocol. Removal of GST from LASP was achieved by digestion with thrombin overnight at 4°C. Purity was analyzed by examination of Coomassie-stained SDSpolyacrylamide gels.
LASP polyclonal antibody generationTo generate a polyclonal antibody specific for LASP, recombinant human GST-LASP that had been expressed and affinity-purified from bacteria was injected into New Zealand rabbits (Immunoglobe, Himmelstadt, Germany).
Immunoreactive serum was affinity-purified against LASP protein coupled to a HiTrap-NHS-activated affinity column according to the manufacturer's instructions.
Western Blot Analysis of LASPCell extracts were resolved by 10% SDS-PAGE. After blotting on nitrocellulose membrane and blocking with 3% nonfat dry milk in 10 mM Tris, pH 7.5, 100 mM NaCl, 0.1% (w/v) Tween 20, the membrane was first incubated with the polyclonal antibody raised against LASP (1:16,000) followed by incubation with horseradish peroxidase-coupled goat anti-rabbit IgG (1:5000) and detection by ECL.
In vitro phosphorylation of LASPLASP and its mutants S61A, S146A, and S61/146A (0. Proteins were separated by SDS-PAGE on 10% gels. Incorporation of 32 P was visualized by autoradiography. ImmunofluorescenceFor immunofluorescence microscopy, transfected PtK2 cells grown on glass chamber slides were fixed in 4% (w/v) paraformaldehyde in PBS, permeabilized with 0.1% (w/v) Triton X-100 in PBS, and then stained with affinity-purified LASP antibody (1:1000) followed by secondary Cy3-labeled anti-rabbit antibody.

Phosphorylation of LASP in intact human platelets treated with the cGK-specific stimulus
8pCPT-cGMPTo identify substrates of cGK in intact human platelets, cells were labeled with [ 32 P]orthophosphate, stimulated with 500 µM of the specific cGMP-dependent protein kinase activator 8pCPT-cGMP, and proteins of the resulting platelet lysate were separated by two-dimensional gel electrophoresis. The 2D-phosphoproteomes resulting from this experiment (Fig. 1) demonstrate the phosphorylation/dephosphorylation of several proteins after stimulation with 8pCPT-cGMP. For identification of these proteins, the spots were excised from the gel, digested with trypsin, and the resulting peptides were analyzed by ESI-MS/MS. Spot 2a-c was identified as the vasodilator-stimulated phosphoprotein (VASP), a well known substrate of cGK in human platelets (8) with multiple phosphorylation isoforms (26). Spot 3 was characterized as rap 1b, a low molecular weight GTP-binding protein.
Phosphorylation of rap 1b has been observed in nitric oxide-stimulated human platelets through stimulation of guanylyl cyclase and activation of the cGMP-dependent protein kinase   Fig. 2, LASP is clearly a substrate of cGK Iβ and cAK, the two isoforms known to be present in human platelets.

Identification of LASP phosphorylation sites in vitroHuman LASP contains two cAK
consensus motifs for serine phosphorylation: Ser-99 (KGFS) and Ser-146 (RRDS) (28). In the present study, phosphorylation of amino acids was determined by mass spectrometry. A complete trypsin digest of phosphorylated LASP was fractionated on a nano-HPLC connected online to an ion trap mass spectrometer. The resulting MS/MS-spectrum of the phosphopeptide with the sequence QpSFTM ox VADTPENLR is presented in Fig. 3. The bion series shows many signals with a loss of -98 Da typical for phosphoserine-and phosphothreonine-containing peptides (29), leading to the unequivocal identification of Ser-61 as phosphorylation site for cGK. In addition, the predicted Ser-146 was identified in a second phosphopeptide (data not shown). However, no phosphorylation of Ser-99 was observed. The same results were obtained with LASP phosphorylated by cAK.
In vitro phosphorylation of LASP and LASP mutantsTo study the specificity of phosphorylation, the identified serine phosphorylation sites Ser-61 and Ser-146 were mutated to alanine (S61A, S146A and S61/146A). The purified, recombinant proteins (wildtype, single mutants, and double mutant) were incubated with the cGK isoforms Iα, Iβ, and II and the catalytic subunit of cAMP-dependent protein kinase in the presence of [γ-32 P]ATP.
Incorporation of phosphate was observed after 30 min with each of the four kinases, albeit at different levels (Fig. 4). cGK Iβ and cAK, the two isoforms present in human platelets, caused the highest phosphate incorporation. As expected, no phosphorylation was observed for the double mutant S61/146A. In a control experiment, VASP, a well known substrate for cAK and cGK (8), was equally phosphorylated by each of the four kinases (Fig. 4).
In vivo phosphorylation of LASP by cGK Iβ and cAK at Ser-146To evaluate the role of LASP as an in vivo substrate of cGK Iβ and cAK (the two isoforms present in human platelets), LASP deficient PtK-2 cells were transfected with wild-type LASP or the mutants S61A and S146A, and cGK Iβ simultaneously or with either protein alone. The cells express cAK endogenously. Phosphorylation of the proteins by cGK Iβ or cAK was analyzed after stimulation of the cells with 8pCPT-cGMP or forskolin, respectively, followed by immunoprecipitation. Wild-type LASP and the mutant S61A showed identical in vivo phosphorylation (Fig. 5). In contrast, no phosphorylation was observed with the S146A mutant, indicating that in vivo Ser-146 is the only cGK Iβ phosphorylation site in LASP. In the absence of cGK Iβ no phosphorylation was detected. Similar results were obtained for cAK phosphorylation of LASP (data not shown).

Expression of LASP in various cell lines and tissuesThe expression of human LASP in
different cell types was studied using a rabbit polyclonal antibody raised against GST-tagged human LASP. Western blot analyses of cell extracts from several human tissues and from different cell lines showed LASP expression in human platelets, brain, heart, kidney, lung, liver, fibroblasts, smooth muscle cells (SMC) and endothelial cells (HUVEC) and various cell lines ( Fig. 6A and B).

Phosphorylation of Ser-146 reduces binding of LASP to F-actinIn earlier studies by
Schreiber et al. (27) as well as in our experiments, the filamentous expression pattern of LASP suggested that the protein is colocalized with F-actin. To test whether this association might be affected through phosphorylation of LASP, we performed F-actin/LASP cosedimentation experiments with wild-type LASP and the phosphorylation-mimicking mutant S146D, because in vitro phosphorylation by cAK and cGK would also phosphorylate Ser-61. In the absence of F-actin, LASP was exclusively located in the soluble fraction, whereas in the presence of F-actin, about half of the LASP protein was found in the pellet (Fig. 7). However, using LASP S146D (mimicking the phosphorylation by cAK and cGK in vivo), two thirds of the protein remained in the supernatant (Fig. 7, left panel). In control experiments, the actin binding protein α-actinin (positive control) cosedimented almost completely with F-actin, whereas >95% of BSA (negative control) remained in the soluble fraction (Fig. 7, right panel). These results suggested that upon phosphorylation of Ser-146, LASP loses its ability to bind to F-actin.

Phosphorylation-dependent redistribution of LASP in PtK2 cells.
In view of these results, we investigated whether the intracellular localization of LASP is directly affected by phosphorylation. PtK2 cells, which express no detectable amount of endogenous LASP (Fig.   6B), were transiently transfected with expression vectors encoding either wild-type LASP, LASP mutant S146A, or LASP mutant S146D. Forty-eight hours after transfection, cells were prepared for immunofluorescence. Wild-type LASP and LASP S146A were predominantly present in the tips of cell membrane extensions and at cell-cell contacts where it co-localizes with F-actin (Fig. 8A, B, D and E) and. However, double staining analysis with the LASP antibody and Oregon green phalloidin for F-actin staining revealed no colocalization with actin stress fibers ( Fig. 8D and 8E). In contrast, the phosphorylationmimicking mutant LASP S146D was found predominantly in the cytosol (Fig. 8C). The specificity of the staining was controlled with preadsorbed LASP antibody, which showed no immunofluorescence (data not shown). Redistribution of LASP to the cytosol reduces cell motility. Since LASP is prominently present within focal contacts and the leading edges of the cell membrane we wondered whether the protein might be involved in cell motility. Therefore we tested migration of PtK2 cells transiently expressing wild-type LASP or LASP mutant S146D in a modified Boyden Chamber system. Cells were seeded in the upper chamber of a transwell polycarbonate membrane and after 4 h cells migrated through the porous membrane were counted.

DISCUSSION
As an approach to identify novel substrates of cGMP-dependent protein kinase, we analyzed cGK-mediated protein phosphorylation in intact human platelets using two-dimensional gel electrophoresis. In addition to the previously known substrates VASP and rap 1b, we identified the LIM and SH3 domain protein (LASP) as a novel substrate for cGK and cAK. LASP consists of an N-terminal zinc-binding LIM domain, followed by two actin binding sites and a Src homology region 3 (SH3) at the C-terminal end (30). The human LASP gene was previously cloned and identified from a human breast cancer cDNA library (31). It was mapped to human chromosome 17 q12-q21 and was shown to be amplified and overexpressed in breast tumors (32). LASP is expressed in all human tissues tested including platelets, brain, heart, kidney, lung, liver, endothelial cells, smooth muscle cells, and fibroblasts. Northern blot analysis of murine LASP revealed a constant expression of the protein during embryogenesis from day 7.5 to day 18.5 with various levels in all adult tissues, which is consistent with an essential role for LASP in basic cellular function (32).
Immunofluorescence analysis of LASP subcellular distribution showed that the protein colocalizes with F-actin at focal adhesion plaques and membrane edges in mouse cardiac fibroblasts and rat mesangial cells (unpublished results). These results confirmed earlier observations by Schreiber et al. who found LASP at peripheral cell extensions in individual epithelial cancer cells (27). Experiments performed using PtK2 cells transfected with wild-type LASP also demonstrated that the protein is colocalized with F-actin at membrane extensions, although not along intracellular stress fibers. In PtK2 cells, however, the LASP mutant S146D, which simulates phosphorylation at Ser-146, accumulates in the cytoplasm when transiently expressed, suggesting that phosphorylation of human LASP by cAK and cGK regulates the intracellular localization of the protein.
Recently, it was shown that the cAK-dependent acid secretory agonists histamine and forskolin induce a rapid sustained rise in LASP phosphorylation in rabbit gastric parietal cells, and this increase is closely correlated with 15 by guest on March 24, 2020 http://www.jbc.org/ Downloaded from the acid secretory response (28). In parallel, LASP redistributes from a predominantly cortical location to a region surrounding the intracellular canaliculus, which is the site of active HCl secretion (35). Mutation of the two major cAK phosphorylation sites in rabbit LASP, Ser-99 and Ser-146, to alanine appears to block this recruitment (34).
The function of LASP in living cells seems to be complex and cell type specific. The localization of LASP to the part of active membrane extension in addition to our observations of a reduced cell migration after phosphorylation and relocalization to the cytosol indicates a prominent role for LASP in cell movement -either by interacting directly with actin and promoting actin polymerization or by acting as a scaffolding molecule recruiting other motility proteins to the tips of the cells involved in the organization of the cytoskeleton. In gastric parietal cells, LASP was identified to bind to dynamin, a large GTPase involved in vesicular fission and control of membrane trafficking in the H + /K + -ATPase pathways at the apical membrane (36).
Apart from LASP, cGK phosphorylates the vasodilator-stimulated phosphoprotein VASP (8), a protein that has been implicated in the regulation of actin dynamics and associated processes such as cell adhesion and motility by its ability to associate with F-actin, profilin, zyxin, and vinculin (37). In platelets, VASP phosphorylation seems to be involved in the negative regulation of the integrin ± IIb b III (9). Actually, in vitro phosphorylation of VASP reduces F-actin binding (10), however, in contrast to LASP, phosphorylation of VASP plays no obvious role in subcellular targeting (38).
As a newly identified signaling protein within the cGMP and cAMP pathway, the specific function of LASP is still under investigation. Future experiments will address the question of whether there are platelet-specific binding partners for LASP and determine whether LASP might be a phosphorylation-dependent molecular switch that regulates the interaction of other proteins with F-actin-rich compartments, thereby modulating platelet function.        Immunofluorescence microscopy of the subcellular distribution of wild-type LASP and the mutants S146A and S146D in PtK-2 cells. Cells were fixed with paraformaldehyde and permeabilized, and LASP was immunostained with the LASP polyclonal antibody and Cy3-labeled goat anti-rabbit secondary antibody. Actin was stained with oregon green phalloidin. Wild-type LASP (A, D) and LASP S146A (B) are predominantly present at the leading edges and focal contacts of cells (indicated by arrows) while the LASP mutant S146D is mainly localized in the cytosol (C). The co-localisation of LASP and F-actin is demonstrated by double-staining in D and E (arrows) while no LASP binding to actin stress fibers (arroheads in E) is observed.