If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Canada Research Chair in Endothelial Cell Biology at the University of Alberta. To whom correspondence should be addressed: Dept. of Medicine, University of Alberta, CSB 11-132, Edmonton, Alberta T5R 0B7, Canada. Tel.: 780-407-8944; Fax: 780-407-7771;
* This work was supported by National Institutes of Health Grant DK50764, by the Canadian Institutes of Health Research (Grant MOP-64184), and by a grant from the Alberta Ingenuity Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
TIMAP (TGF-β1 inhibited, membrane-associated protein) is a prenylated, endothelial cell-predominant protein phosphatase 1 (PP1c) regulatory subunit that localizes to the plasma membrane of filopodia. Here, we determined whether phosphorylation regulates TIMAP-associated PP1c function. Phosphorylation of TIMAP was observed in cells metabolically labeled with [32P]orthophosphate and was reduced by inhibitors of protein kinase A (PKA) and glycogen synthase kinase-3 (GSK-3). In cell-free assays, immunopurified TIMAP was phosphorylated by PKA and, after PKA priming, by GSK-3β. Site-specific Ser to Ala substitution identified amino acid residues Ser333/Ser337 as the likely PKA/GSK-3β phosphorylation site. Substitution of Ala for Val and Phe in the KVSF motif of TIMAP (TIMAPV64A/F66A) abolished PP1c binding and TIMAP-associated PP1c activity. TIMAPV64A/F66A was hyper-phosphorylated in cells, indicating that TIMAP-associated PP1c auto-dephosphorylates TIMAP. Constitutively active GSK-3β stimulated phosphorylation of TIMAPV64A/F66A, but not wild-type TIMAP, suggesting that the PKA/GSK-3β site may be subject to dephosphorylation by TIMAP-associated PP1c. Substitution of Asp or Glu for Ser at amino acid residues 333 and 337 to mimic phosphorylation reduced the PP1c association with TIMAP. Conversely, GSK-3 inhibitors augmented PP1c association with TIMAP-PP1c in cells. The 333/337 phosphomimic mutations also increased TIMAP-associated PP1c activity in vitro and against the non-integrin laminin receptor 1 in cells. Finally, TIMAP mutants with reduced PP1c activity strongly stimulated endothelial cell filopodia formation, an effect mimicked by the GSK-3 inhibitor LiCl. We conclude that TIMAP is a target for PKA-primed GSK-3β-mediated phosphorylation. This phosphorylation controls TIMAP association and activity of PP1c, in turn regulating extension of filopodia in endothelial cells.
) that includes PP2A, PP4-PP7, and calcineurin (PP2B). Ser/Thr phosphatases counteract essentially every signal involving Ser/Thr kinases, and like the kinases their activity is tightly regulated. Because of extreme sequence conservation (>90%) among PP1c isoforms, substrate- and location-specific actions are not inherent in PP1c itself but dictated by more than 50 PP1c regulatory (targeting) subunits (
). Currently, it is not known whether TIMAP can act as a myosin light chain phosphatase-targeting subunit. However, we have previously shown that TIMAP associates with the plasma membrane of endothelial cell filopodia, where it interacts with, and serves to dephosphorylate the non-integrin laminin receptor 1 (LAMR1) (
In this study, we sought evidence for regulation of TIMAP/PP1c by phosphorylation. We find that TIMAP is phosphorylated by PKA and, after PKA priming, by GSK-3β. Furthermore, phosphorylation by GSK-3β regulates both the association of PP1c with TIMAP and its activity in vitro and in cells. Mutations of TIMAP that reduce its associated PP1c activity markedly change the organization of endothelial cell filopodia. The data indicate that TIMAP function is regulated by PKA/GSK-3β-mediated phosphorylation and that TIMAP may function in the formation of endothelial cell filopodia.
Materials—All chemicals and enzymes were of reagent grade and purchased from Sigma-Aldrich Canada, unless otherwise specified. Chicken anti-TIMAP antibody was custom-prepared by Avis Laboratories (Tigard, OR) (
) and maintained in RPMI 1640 medium supplemented with 20% calf serum, 2 μg/ml heparin, 8.0 μg/ml bovine brain extract (Cambrex BioScience, Walkersville, MD), and penicillin/streptomycin. To assess GEN filopodia formation, the cells were plated sparsely on laminin-1-coated coverslips after transient transfection with TIMAP mutants and assessed 48 h later. MDCK and COS7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and penicillin/streptomycin. In some experiments GEN cells were treated with LiCl (5, 10, or 20 mm) for up to 4 h. Cells were fixed with 1% paraformaldehyde in PBS for 10 min. The plasma membrane was stained with 1 μg/ml Alexa Fluor 594-conjugated wheat germ agglutinin (Invitrogen). The density of filopodia was assessed by counting projections from cells that extended >2 μm from their base at the plasma membrane. For each condition and time point, between 25 and 50 cells were counted in each of three replicate wells. Statistical analysis was by analysis of variance.
TIMAP Constructs, Mutants, and Transfection—The full-length cDNA of bovine TIMAP (GenBank™ accession number AF362909) was cloned into the pBlueScript SKII(+) (pBS-bTIMAP). A cDNA, TIMAPWT encoding the full open reading frame, was PCR-amplified from pBS-bTIMAP with primers 5′-CGCAGGTACCACCATGGGGCCAGCCACGTGGACCTGCTGACC-3′ and 5′-CCGGAATTCCTAGGAGATGCGGCAGCAGCCAT-3′, in which KpnI and ECoR I (underlined) are integrated for cloning and one Kozak consensus sequence (bold) is added to enhance expression in mammalian cells. The PCR product was directly cloned into pcDNA3.1/V5-His-TOPO (Invitrogen). Orientation was confirmed by restriction enzyme digestion. Fidelity was verified by full insert sequencing. This clone, designated pcDNA3.1-TIMAPWT, was used to generate several TIMAP mutants using the QuikChange II site-directed mutagenesis kit (Stratagene). EGFP epitope-tagged TIMAP mutants were produced similarly, using the pEGFP-TIMAPWT construct previously described (
) as template. Primers used for specific mutations are listed in Table 1. Briefly, PCR was carried out in a 50-μl reaction using 20 ng of the pcDNA3.1-TIMAPWT template, 100 ng of each primer carrying the mutation, 10 nmol of dNTPs, and 2.5 units of PfuUtral DNA polymerase in polymerase reaction buffer. For PCR, initial denaturation was at 95 °C for 30 s, followed by 12 cycles of 95 °C for 30 s, 55 °C for 1 min, and 68 °C for 8 min. Two microliters of the DpnI-treated PCR reaction was used to transform XL1-Blue supercompetent cells. Each clone was verified by DNA sequencing.
For all transient transfections, cells were trypsinized and replated to achieve 80-90% confluency 24 h later. The COS7 cells were then transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. The GEN cells were transfected with the FuGENE 6 reagent (Roche Applied Science).
Immunoprecipitation and Immunoblotting—Cells were washed with ice-cold PBS twice and scraped from culture plates into cold immunoprecipitation (IP)-lysis buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 1.0% Nonidet P-40, 0.5% sodium deoxycholate, and Complete protease inhibitors from Roche Applied Science). To protect phosphorylated sites, 30 mm sodium fluoride, 40 mm β-glycerophosphate, 20 mm sodium pyrophosphate, 1 mm sodium orthovanadate, and 100 nm calyculin A were included in lysis buffer. Cells were homogenized on ice and then centrifuged at 17,000 × g for 30 min at 4 °C. The supernatants were incubated with chicken anti-TIMAP antibodies, chicken anti-LAMR1 antibodies (
), or equal amounts of control chicken IgY for 1 h at 4 °C. Goat anti-chicken IgG-coated beads (Aves) were then added followed by incubation overnight at 4 °C. The beads were sedimented at 1500 rpm for 5 min at 4 °C, washed twice with IP wash buffer I (cold lysis buffer), twice with IP wash buffer II (50 mm Tris-HCl, pH 7.5, 500 mm NaCl, 0.1% Nonidet P-40, 0.05% sodium deoxycholate), and once with IP wash buffer III (10 mm Tris-HCl, pH 7.5, 0.1% Nonidet P-40, 0.05% sodium deoxycholate). The beads were then re-suspended in 2× Laemmli buffer and boiled for 10 min. The supernatants were fractionated on 8% or 10% SDS-PAGE gels and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA). Blots were probed with various antibodies, and detection was achieved with ECL chemiluminescent substrate (GE Healthcare, UK).
Metabolic Labeling—Upon reaching 80-90% confluency, cells were washed three times with serum-free, phosphate-free Dulbecco's modified Eagle's medium (Sigma), and then incubated in the same phosphate-free medium for 1 h. The medium was then replaced with fresh phosphate-free Dulbecco's modified Eagle's medium containing 400 μCi/ml [32P]H3PO4 (PerkinElmer Life Sciences) and incubated for 5 h. During the last 30 min of incubation, 100 nm calyculin A (Upstate, Waltham, MA) was added in some experiments. In experiments examining the dephosphorylation of LAMR1, cells were labeled with [35S]methionine, as previously described (
). Cells were washed once with ice-cold PBS and lysed in cold IP-lysis buffer containing complete proteinase and phosphatase inhibitors. The immunoprecipitates were washed five times as described, except that all IP wash buffers also contained 20 nm okadaic acid, and IP wash buffer III contained 0.1% SDS. Beads were then suspended in 2× Laemmli buffer, boiled for 5 min, followed by SDS-PAGE. Proteins in the gels were transferred to PDVF membranes, and subjected to autoradiography. Subsequent immunoblotting with polyclonal rabbit anti-TIMAP antibody served to detect loading.
In Vitro Dephosphorylation by Recombinant PP1cα—TIMAP immunoprecipitates were prepared from [32P]H3PO4 metabolically labeled GEN cells that had been treated with 100 nm calyculin A during the last 30 min of the labeling period. To preserve TIMAP phosphorylation, cell lysates and TIMAP precipitates were prepared in the presence of phosphatase inhibitors. They were then washed five times with IP wash buffer without phosphatase inhibitors and divided into four aliquots. The samples were incubated in 500 μl of PP1c reaction buffer (50 mm Tris-HCl, pH 7.0, 0.1 mm Na2EDTA, 5 mm dithiothreitol, 0.01% Brij 35, 1 mm MnCl2) for 30 min at 37 °C, either in buffer alone or with 0.8 unit of recombinant PP1cα in the absence or presence of the PP1c-specific inhibitor-2 (4 μg, New England Biolabs) or 600 nm calyculin A. TIMAP was then eluted from the beads with 2× Laemmli buffer and subjected to autoradiography and Western blot analyses.
In Vitro Kinase Assay—TIMAP was immunoprecipitated from GEN, MDCK, or COS7 cells transfected with TIMAP as described, except that phosphatase inhibitors were omitted from the lysis and wash buffers. The TIMAP immobilized on beads was washed once with the appropriate kinase reaction buffer. The beads were then incubated with 500 units of casein kinase II (New England Biolabs, reaction buffer: 20 mm Tris-HCl, 50 mm KCl, 10 mm MgCl2 pH 7.5), 100 ng of GSK-3α (Upstate, reaction buffer: 8 mm MOPS, pH 7.0, 0.2 mm EDTA), 500 units of GSK-3β (New England Biolabs, reaction buffer: 20 mm Tris-Cl, 10 mm MgCl2 5 mm dithiothreitol, pH 7.5), 2500 units of PKA catalytic subunit (New England Biolabs, reaction buffer: 50 mm Tris-Cl, 10 mm MgCl2), or 12.5 ng of protein kinase C catalytic fragment (Biomol, reaction buffer: 40 mm MES, pH 6.0, 1 mm EGTA, 10 mm MgCl2) all in the presence of 200 μm ATP and 5 μCi of [γ-32P]ATP (final specific activity, 800 μCi/μm) at 30 °C for 60 min in a volume of 30 μl. The immunoprecipitates were washed once with IP-lysis buffer I, containing phosphatase inhibitors. TIMAP was eluted in 50 μl of 2× Laemmli buffer and boiled for 10 min.
For primed GSK-3 in vitro phosphorylation studies, the TIMAP immunoprecipitates were incubated with 2500 units of PKA for 5 h at 30 °C in PKA reaction buffer containing 10 μm unlabeled ATP. The samples were then heated at 65 °C for 20 min to inactivate PKA, washed once with GSK-3 reaction buffer, and divided into three or four portions, as appropriate. The aliquots were then incubated for 30 min at 30 °C with buffer alone or GSK-3α or -3β in GSK-3 reaction buffer containing 5 μm ATP and 10 μCi [γ-32P]ATP. The immunoprecipitates were washed once with cold IP-wash buffer I containing phosphatase inhibitors, and proteins were eluted from the beads with 50 μl of 2× Laemmli buffer. Samples were then subjected to SDS-PAGE (8% gel), blotted onto a polyvinylidene difluoride membrane, and subjected to autoradiography and Western blot analyses.
PP1c Activity in Vitro—TIMAP-associated PP1c activity was measured using the small fluorogenic substrate 6,8-difluoro-4-methylumbelliferyl phosphate (Molecular Probes, Eugene, OR), which, when cleaved by PP1c, generates fluorescent 6,8-difluoro-4-methylumbelliferyl with excitation/emission maxima at 358/452 nm. TIMAP immunoprecipitates were prepared from the COS7 cells transfected with vector alone, TIMAPWT, or various TIMAP mutants. Each immunoprecipitate was divided into two portions, one to quantify PP1c activity, the other for Western blot analysis. The immunoprecipitates were incubated with 50 μm 6,8-difluoro-4-methylumbelliferyl phosphate in 100 μl of PP1c reaction buffer for 90 min at 37 °C. The fluorescence was measured at 460 nm every 5 min using a Fluoroskan Ascent fluorometer (ThermoLabsystems, Ramsey, MN). Recombinant PP1cα (0.05 unit) and buffer alone served as a positive or negative control, respectively. Addition of calyculin A (250 nm) to the same samples was used to show specificity. Background, determined from samples prepared identically from vector-transfected cells was subtracted.
TIMAP Is Phosphorylated in Living Cells—When endogenous TIMAP was immunoprecipitated from [32P]orthophosphate-labeled GEN cells, a 32P-labeled band with a molecular mass of ∼65 kDa was observed (Fig. 1A). Similarly, in MDCK cells that lack endogenous TIMAP, 32P was incorporated into TIMAP that was overexpressed, and then immunoprecipitated. That the 32P-labeled ∼65-kDa band represents TIMAP was confirmed by immunoblot analysis with a rabbit polyclonal anti-TIMAP antibody (Fig. 1A). Therefore, both endogenous and overexpressed TIMAP are phosphorylated in cells.
PKA Phosphorylates TIMAP in Vivo and in Vitro—In silico analysis of the TIMAP aa sequence with NetPhos2.0 (
) predicted 32 possible phosphorylation sites (21 Ser, 6 Thr, and 5 Tyr). Two prominent potential phosphorylation clusters were predicted in the intervals from residues 324 to 352 and 505 to 534 of TIMAP. In silico analysis with MacVector 7.2.2 (Accelrys, San Diego, CA) suggested that TIMAP may be a substrate for casein kinase II, GSK-3, PKA, and protein kinase C (data not shown).
When MDCK cells forced to overexpress TIMAPWT were treated with the PKA inhibitor H89 and/or the GSK-3 inhibitor LiCl during the last 30 min of 32P metabolic labeling, incorporation of 32P into TIMAP was significantly and reproducibly reduced compared with TIMAP from cells not exposed to inhibitors (Fig. 1B). Densitometric analysis showed that TIMAP phosphorylation was reduced by 21 ± 2.9, 27 ± 3.4, and 53 ± 2.6 compared with control (mean ± S.D., n = 3, p < 0.01) in the presence of H89, LiCl, or both, respectively. Not shown, the protein kinase C inhibitor calphostin C (200 nm), the phosphatidylinositol 3-kinase inhibitor wortmannin (100 nm), and the tyrosine kinase inhibitor genistein (12 μm) had no discernable effect on 32P incorporation by TIMAP in MDCK cells. Hence, PKA and GSK-3 were considered to be candidate kinases for TIMAP phosphorylation.
TIMAP was immunopurified from both GEN and MDCK cells to determine whether it can serve as a substrate in vitro for those kinases predicted by in silico analysis. Not shown, when phosphatase inhibitors were omitted from the IP-lysis and wash buffers, the 32P label was lost from TIMAP during the IP process. Therefore, phosphatase inhibitors were omitted from TIMAP immunoprecipitation buffers for the in vitro phosphorylation studies. Shown in Fig. 1C, no phosphorylation of TIMAP was observed in vitro in the absence of added kinase, indicating that no active, TIMAP-directed kinase was co-precipitated with TIMAP. Significant phosphorylation in vitro was observed in the presence of PKA, but not with casein kinase II, protein kinase C, GSK-3α, or GSK-3β (Fig. 1C). The in vitro phosphorylation of immunopurified TIMAP was significantly reduced by the PKA inhibitor H89 (Fig. 1D), in keeping with a specific PKA-mediated phosphorylation.
GSK-3β, Not-3α, Phosphorylates TIMAP after a PKA Priming Phosphorylation—For most GSK-3 substrates, efficient phosphorylation by GSK-3 requires a priming phosphorylation by other kinases (
). The lack of TIMAP phosphorylation by GSK-3α and GSK-3β in vitro (Fig. 1C) could therefore be due to the absence of a priming phosphorylation. Because PKA strongly phosphorylates TIMAP in vitro (Fig. 1C), it seemed plausible that PKA might serve as a priming kinase for GSK-3α or GSK-3β. As shown in Fig. 1E, when immunoprecipitated TIMAP was treated with GSK-3β in vitro, in the presence of [32P]ATP, TIMAP was phosphorylated, as long as this step was preceded by a PKA-priming phosphorylation, done with unlabeled ATP. By contrast, GSK-3α did not phosphorylate PKA-primed TIMAP. Also, no 32P incorporation was observed when TIMAP was incubated with [32P]ATP in buffer lacking GSK-3 kinases, indicating that the phosphorylation observed in the presence of GSK-3β was not due to residual activity of the heat-inactivated PKA (Fig. 1E). Thus, GSK-3β but not GSK-3α can phosphorylate TIMAP in vitro, but only after PKA priming.
Identification of One GSK-3β Phosphorylation Site—A potential GSK-3β phosphorylation site (Ser333) was identified by MacVector 7.2.2. The GSK-3 consensus sequence Ser/Thr-X-X-X-Ser/Thr-P (
) contained within this region of TIMAP is Ser333-Arg-Arg-Thr-Ser337, where Ser337 would serve as the potential PKA priming site, and Ser333 as the potential GSK-3β phosphorylation site (Fig. 2A). A deletion mutant of TIMAP, lacking residues 324-381, containing this site was created and transfected transiently into COS7 cells. This deletion reduced TIMAP phosphorylation to nearly undetectable levels (Fig. 2B). Single or dual Ser to Ala mutations were then introduced at residues 333 (TIMAPS333A) and 337 (TIMAPS337A) and both 333 and 337 (TIMAPS333A/S337A). Immunoprecipitation of the transiently expressed TIMAPS333A, TIMAPS337A, or TIMAPS333A/S337A from COS7 cells that had been metabolically labeled with [32P]orthophosphate showed that these mutations significantly blunt phosphorylation of TIMAP when compared with TIMAPWT (Fig. 2C). These mutations did not alter in vitro PKA-mediated phosphorylation of TIMAP substantially (Fig. 2D), indicating that other PKA-sensitive sites exist. Nonetheless, the Ser to Ala mutations at aa residues 333 and/or 337 completely abolished in vitro GSK-3β-mediated phosphorylation of TIMAP (Fig. 2E). These findings suggest strongly that TIMAP is phosphorylated by GSK-3β at the Ser333 residue, with Ser337 likely serving as the PKA priming site.
Disruption of the PP1c Consensus Motif Alters PP1c Binding and Activity—It was predicted that disruption of the KVSF (aa 63-66) PP1c binding motif (Fig. 3A) would alter PP1c association (
). It also seemed plausible that phosphorylation of TIMAP at aa residue Ser65 could regulate the PP1c association. Site-directed mutagenesis was used to produce three TIMAP mutants: TIMAPS65A, TIMAPF66A, and TIMAPV64A/F66A, which were then transiently expressed in COS7 cells. Shown in Fig. 3B, PP1c co-immunoprecipitated with TIMAPWT and TIMAPS65A, but little if any PP1c was found in TIMAPF66A and TIMAPV64A/F66A immunoprecipitates. TIMAP-associated phosphatase activity, readily detected in immunoprecipitates of TIMAPWT, was essentially abolished in the TIMAPV64A/F66A mutant (Fig. 3C, p < 0.01, analysis of variance). These results are consistent with previous evidence that the valine and phenylalanine residues within the PP1c binding motif of PP1c regulatory subunits are necessary for the association with PP1c (
). The finding that the substitution of Ser with Ala at aa residue 65 did not alter the PP1c association suggests either that phosphorylation at this site does not alter PP1c association, or that TIMAPWT was not phosphorylated at Ser65 under conditions examined here. Not shown, 32P incorporation into TIMAP was not altered by the Ser to Ala substitution at aa residue 65.
Auto-dephosphorylation of TIMAP by Its Associated PP1c—Phosphorylation of endogenous TIMAP in GEN cells or overexpressed TIMAP in COS7 cells was strongly augmented when the cells were exposed to the phosphatase inhibitor calyculin A during the last 30 min of [32P]orthophosphate labeling (Fig. 4A), consistent with TIMAP de-phosphorylation by endogenous phosphatase(s). To determine whether TIMAP itself can serve as a PP1c substrate, endogenous TIMAP was purified from GEN cells that had been metabolically labeled with [32P]orthophosphate (Fig. 4B). Four equal portions of the immunoprecipitate were incubated with buffer alone, or with recombinant PP1cα in the absence or presence of the PP1c inhibitor II (I-2) or calyculin A. We observed that TIMAP was de-phosphorylated by PP1cα in vitro and that PP1c inhibitors reduced this de-phosphorylation (Fig. 4B). We next determined whether TIMAP could be de-phosphorylated by TIMAP-associated PP1c. When TIMAPWT, TIMAPF66A, or TIMAPV64A/F66A were transiently expressed in COS7 cells followed by [32P]orthophosphate metabolic labeling and immunoprecipitation, the TIMAPF66A and TIMAPV64A/F66A mutants were hyper-phosphorylated compared with TIMAPWT (Fig. 4C). These findings indicate that the association of PP1c with TIMAP brings phosphorylated sites within TIMAP into sufficiently close proximity of the PP1c catalytic domain to result in their de-phosphorylation.
We next determined whether GSK-3β-mediated phosphorylation of TIMAP is subject to regulation by TIMAP-associated PP1c. To this end, COS7 cells were transfected with constitutively active or inactive GSK-3β in the presence of TIMAPWT, TIMAPS333AS337A, TIMAPV64A/F66A, or a TIMAP mutant in which both the PP1c binding motif and the GSK-3β site were mutated: TIMAPV64A/F66A/S333AS337A. Although no GSK-3β-dependent phosphorylation of TIMAPWT or TIMAPS333AS337A was detected, TIMAPV64A/F66A was significantly more phosphorylated in the presence of constitutively active GSK-3β than with inactive GSK-3β (Fig. 4D). In two separate experiments, TIMAPV64A/F66A phosphorylation was 26 and 47% greater in the presence of active, compared with inactive GSK-3β. Hence, the reduced association of PP1c with TIMAP unmasked GSK-3β-mediated TIMAP phosphorylation in cells. By contrast, phosphorylation of the TIMAPV64A/F66A/S333AS337A mutant was similar in the presence of constitutively active compared with inactive GSK-3β (Fig. 4D). Therefore, TIMAP-associated PP1c may dephosphorylate the Ser333/Ser337 PKA/GSK-3β site directly, or it may indirectly unmask the PKA/GSK-3β-mediated phosphorylation.
Phosphorylation of TIMAP by GSK-3β Regulates Both PP1c Activity and Association with TIMAP—To determine whether PP1c association with TIMAP or TIMAP-associated PP1c activity are altered when the Ser333/Ser337 site is phosphorylated, Ser at residues 333 and 337 was replaced by aspartic acid (TIMAPS333D/S337D) or glutamic acid (TIMAPS333E/S337E), to mimic phosphorylation. Compared with TIMAPWT and TIMAPS333A/S337A, much less PP1c was co-immunoprecipitated with TIMAPS333D/S337D and TIMAPS333E/S337E (Fig. 5A). Nonetheless, TIMAP-associated PP1c activity was consistently greater in the TIMAPS333D/S337D and TIMAPS333E/S337E mutants when compared with TIMAPWT and TIMAPS333A/S337A (Fig. 5B, p < 0.01, analysis of variance). These findings suggest that phosphorylation of TIMAP at aa residues Ser333/Ser337 reduces PP1c association with TIMAP. Nevertheless the same phosphomimic mutation increases the activity of TIMAP-associated PP1c.
In COS7 cells, three distinct GSK-3 inhibitors, alsterpaullone, LiCl, and 4-benzyl-2-methyl-1,2,4-thiadiazolidine-3,5-dione (TDZD-8) enhanced PP1c co-immunoprecipitation with overexpressed TIMAPWT, consistent with inhibition of the TIMAP-PP1c association by GSK-3β-mediated phosphorylation of TIMAP. Similarly, TDZD-8 enhanced the PP1c association with endogenous TIMAP in GEN cells (Fig. 5C). In addition, in COS7 cells treated for 30 min prior to harvest with calyculin A less PP1c co-immunoprecipitated with TIMAP compared with that from cells not treated with calyculin A (Fig. 5D). Taken together, these findings indicate that the association of PP1c with TIMAP is reduced when TIMAP is phosphorylated by GSK-3β.
Phosphomimic Mutations at aa Residues Ser333/Ser337 Augment TIMAP/PP1c-mediated De-phosphorylation of LAMR1 in Cells—To determine whether the changes in the TIMAP-associated PP1c activity observed in vitro also influence its activity in cells, TIMAPWT, TIMAPV64A/F66A, TIMAPS333A/S337A, and the phosphomimic TIMAPS333E/S337E were transiently expressed in COS7 cells. The cells were then metabolically labeled with [32P]orthophosphate or [35S]methionine. The LAMR1, previously shown to bind TIMAP and to be a substrate for TIMAP in cells (
) was then immunoprecipitated followed by autoradiography and Western blot analysis for co-precipitated TIMAP (Fig. 5E). Phosphorylation of LAMR1 was significantly reduced in cells transfected with TIMAPS333E/S337E (Fig. 5F) consistent with enhanced TIMAPS333E/S337E associated PP1c activity. The 35S labeling showed the presence of both TIMAP and LAMR1 in the LAMR1 immunoprecipitates, with essentially equal amounts precipitated for the different conditions.
TIMAP-associated PP1c Activity Regulates the Structure of Endothelial Cell Filopodia—TIMAP is known to localize predominantly to endothelial cell filopodia, where it co-localizes with its substrate, LAMR1 (
)) or the GSK-3β phosphorylation sites were mutated would result in reorganization of endothelial cell filopodia. GFP-TIMAPWT, GFP-TIMAPV64A/F66A, GFP-TIMAPS333A/S337A, GFP-TIMAPANK4(-), GFP-TIMAPS333D/S337D, or GFP-TIMAPS333E/S337E were therefore overexpressed in GEN cells (Fig. 6A). In GEN cells expressing GFP-TIMAPV64A/F66A, filopodia were significantly longer than those observed in cells transfected with GFP-TIMAPWT (147 ± 16 versus 91 ± 11 μm, mean ± S.E., n = 25 cells each, p = 0.005, Student's t test). The length of filopodia in the other mutants was not different from GFP-TIMAPWT. In GEN cells, the number of filopodia was also altered depending on the TIMAP mutant expressed. Filopodial numbers per cell were 2.59 ± 0.29 (n = 26 cells) in GFP-TIMAPWT-transfected cells, 4.17 ± 0.30 (n = 34 cells) in GFP-TIMAPV64A/F66A-transfected cells, and 4.22 ± 0.27 (n = 26 cells) in GFP-TIMAPS333A/S337A-transfected cells (mean ± S.E., p = 0.004 versus TIMAPWT, Student's t test). In GEN cells expressing GFP-TIMAPS333D/S337D or GFP-TIMAPS333E/S337E (Fig. 6A, panels v and vi, respectively) filopodia number was similar to that observed in cells expressing TIMAPWT. Not shown are similar findings for filopodia number and length obtained in COS7 cells. In GEN cells expressing TIMAPANK4(-), a mutant that cannot interact with, and fails to dephosphorylate LAMR1 (
), the number of filopodia per cells was 4.23 ± 0.36, n = 34 cells (p = 0.001, versus TIMAPWT, Student's t test). Finally, in GEN cells treated with the GSK-3 inhibitor LiCl, a significant increase in filopodia was observed 2 and 4 h after addition of LiCl (Fig. 6, B and C). It therefore seems likely that the PP1c-targeting subunit TIMAP regulates filopodia function in endothelial cells and that phosphorylation of TIMAP by GSK-3β controls this effect.
This study demonstrates phosphorylation-mediated regulation of the endothelium-restricted PP1c-targeting subunit TIMAP. Like its closest family member MYPT3 (
), TIMAP can be phosphorylated by PKA. In addition, a PKA-primed GSK-3β phosphorylation site not conserved in MYPT3 (Fig. 7A) is found within the central phosphorylation cluster of TIMAP. Phosphorylation at this site reduces the association between TIMAP and PP1c and enhances PP1c activity. TIMAP-associated PP1c in turn serves to auto-dephosphorylate TIMAP. Based on evidence obtained from PP1c-deficient and phosphomimic TIMAP mutants, TIMAP phosphorylation controls PP1c activity against the metastasis-associated LAMR1, currently the only known target of TIMAP (
). In addition, mutants of TIMAP with reduced PP1c activity markedly enhanced the formation of filopodia in endothelial cells. We therefore conclude that TIMAP is a PKA/GSK-3β-regulated PP1c targeting subunit involved in the regulation of endothelial cell filopodia extension.
A large body of previous work indicates that the physiological function of MYPT family members is extensively regulated by phosphorylation (
), as well as changes in associated PP1c activity. It was already known that a central cluster of potential phosphorylation sites is highly conserved between TIMAP and MYPT3 and that PKA can phosphorylate MYPT3 (
) identified aa residues Ser340 and Ser353 (corresponding to TIMAP Ser337 and Ser350, respectively) as the preferred PKA phosphorylation sites of MYPT3. In keeping with those observations for MYPT3, we found that endogenous TIMAP in endothelial cells as well as overexpressed TIMAP is phosphorylated, that this phosphorylation can be inhibited by H89, that deletion of the central phosphorylation cluster dramatically reduces TIMAP phosphorylation, and that PKA can phosphorylate TIMAP in vitro (Figs. 1 and 2).
GSK-3-mediated phosphorylation usually requires a priming phosphorylation 4 residues upstream of the GSK-3 site (
). We identified a potential GSK-3 site not observed in MYPT3, just upstream of the conserved (Ser337) PKA site. After priming with PKA, TIMAP was phosphorylated in vitro by GSK-3β, but not by GSK-3α, and this phosphorylation was abolished by site-directed substitution of Ala for Ser at aa residues 333 and/or 337. Reduced phosphorylation in cells of TIMAP mutants in which Ser333 and/or Ser337 were substituted by Ala (Fig. 2C) and inhibition of TIMAP phosphorylation by LiCl (Fig. 1B) furthermore suggested that these sites can be phosphorylated in living cells.
Because TIMAP binds PP1c, it was of interest to determine whether TIMAP is subject to auto-dephosphorylation by its associated PP1c. As expected (
), mutation of the KVSF motif (TIMAPV64A/F66A) dramatically reduced PP1c binding and abolished TIMAP-associated PP1c activity (Fig. 3, B and C). In keeping with dephosphorylation of TIMAP by its associated PP1c, TIMAPV64A and TIMAPV64/F66A were hyperphosphorylated in cells as compared with TIMAPWT (Fig. 4C). That TIMAP-associated PP1c is directed against the Ser333/Ser337 PKA/GSK-3β sites is strongly suggested by the finding that TIMAPV64/F66A but not TIMAPV64/F66A/S333A/S337A is phosphorylated by co-expressed, constitutively active GSK-3β.
If the GSK-3β-mediated phosphorylation of TIMAP is to be functionally important, it seemed that the association of PP1c with TIMAP or the level of PP1c activity might be altered by phosphorylation of TIMAP at Ser333/Ser337. TIMAP mutants in which acidic residues at the 333/337 sites were introduced to mimic phosphorylation co-precipitated much less PP1c from cells than wild-type TIMAP or mutants in which Ala was substituted at the 333/337 sites. These experiments left open the possibility that phosphorylation by PKA at Ser337 could be sufficient for the regulation of PP1c association with TIMAP, and/or that substitution of acidic residues produces a non-physiological conformational change. However, we also observed that inhibition of GSK-3 in cells increased PP1c co-immunoprecipitation with endogenous and overexpressed TIMAPWT. We therefore conclude that TIMAP phosphorylation by GSK-3β reduces its association with PP1c.
The in vitro activity of TIMAP-associated PP1c was also greater in the phosphomimic mutants compared with TIMAPWT or TIMAPS333A/S337A (Fig. 5B). Similarly, the TIMAP/PP1c activity in cells toward LAMR1 was also greater for the phosphomimic mutants compared with TIMAPWT or TIMAPS333A/S337A (Fig. 5, E and F).
We interpret these findings to indicate that phosphorylation of TIMAP at aa residues Ser333/Ser337 not only alters the TIMAP-PP1c association but also PP1c activity. Further studies will be required to define the mechanism by which phosphorylation at the Ser333/Ser337 site can simultaneous diminish the TIMAP-PP1c association and increase PP1c activity. It is possible that the Ser333/Ser337 site acts as an auto-inhibitory domain that blocks the active site of PP1c in the unphosphorylated state. The data are also consistent with the possibility that TIMAP inhibits its associated PP1c activity, as is the case for Mypt 3 (
), and that phosphorylation at the Ser333/Ser337 site promotes dissociation of PP1c from TIMAP rendering it more active.
The findings that TIMAPV64/F66A but not TIMAPWT was phosphorylated by co-expressed active GSK-3β (Fig. 4D) together with the finding of enhanced PP1c activity associated with the two 333/337 phosphomimic mutants of TIMAP (Fig. 5B) is consistent with the possibility that phosphorylation of TIMAP at the PKA/GSK-3β-sensitive site is subject to rapid auto-dephosphorylation, although this effect of TIMAP-associated PP1c could also be indirect.
In endothelial cells, TIMAP associates with the plasma membrane of filopodia (
) and that PP1c associates with LAMR1 in the presence, but not in the absence, of TIMAP. Here we find that in cells LAMR1 is less phosphorylated in the presence of the more active PKA/GSK-3β phosphomimic TIMAP than in the presence of TIMAPWT. Whether TIMAP/PP1 can directly dephosphorylate LAMR1 when its Ser333/Ser337 residues are phosphorylated or whether the effect is indirect, awaits experiments to determine whether purified LAMR1 is indeed a substrate for phosphorylated TIMAP/PP1c in vitro. Nonetheless, the work indicates that LAMR1 is a likely target of TIMAP/PP1c in endothelial cells.
In endothelial cells transfected with GFP-TIMAPV64A/F66A, a mutant in which TIMAP activity is reduced, very long filopodia, not observed in cells transfected with TIMAPWT, formed. The number of filopodia in cells transfected with TIMAPS333A/S337A, a mutant with reduced PP1c activity, or with the TIMAPANK4(-), which cannot interact with LAMR1, was also increased. Similarly, nonspecific inhibition of GSK-3 with LiCl enhanced filopodia formation in GEN cells. By contrast, in endothelial cells transfected with the TIMAPS333D/S337D or TIMAPS333E/S337E phosphomimic mutants, filopodia were not longer, nor more abundant than those in control cells. We therefore propose that TIMAP-associated PP1c activity is regulated by GSK-3β and that this activity alters the phosphorylation status of LAMR1 and the extension of filopodia (Fig. 7B). If LAMR1 is the intermediate that regulates TIMAP/PP1c-dependent filopodia extension, its mechanism of action is as yet obscure.
In summary, this study shows that the endothelium-predominant PP1c regulatory subunit TIMAP is phosphorylated in vivo and in vitro by PKA/GSK-3β, that its associated PP1c regulates TIMAP phosphorylation, and that the PKA/GSK-3β-mediated phosphorylation regulates both PP1c association and activity. The finding that TIMAP, known to localize to endothelial cell filopodia, regulates filopodial length raises the intriguing possibility that TIMAP may target its activity specifically to endothelial cell protrusions involved in migration, angiogenesis, and transmigration of leukocytes.
The GSK-3β plasmids were generously provided by Dr. Frank McCormick and Vivianne W. Ding (University of California at San Francisco Cancer Research Center).