![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Molecular Genetics, The Weizmann Institute of Science, Rehovot 76100, Israel
Received 9 November 2006/ Returned for modification 26 December 2006/ Accepted 7 August 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
associates with microtubules in vivo and is inhibited by them in a noncompetitive manner. Microtubule-associated proteins, which interact strongly with microtubules in vivo, significantly increase binding of PTP to tubulin in vitro and further reduce phosphatase activity. Conversely, disruption of microtubule structures in cells reduces their association with PTP, alters the subcellular localization of the phosphatase, and increases its specific activity. Activation of the epidermal growth factor receptor (EGFR) increases the PTP-microtubule association in a manner dependent upon EGFR-induced phosphorylation of PTP at Y638 and upon microtubule integrity. These events are transient and occur with rapid kinetics similar to EGFR autophosphorylation, suggesting that activation of the EGFR transiently down-regulates PTP activity near the receptor by promoting the PTP-microtubule association. Tubulin also inhibits the tyrosine phosphatase PTP1B but not receptor-type PTPµ or the unrelated alkaline phosphatase. The data suggest that reversible association with microtubules is a novel, physiologically regulated mechanism for regulation of tyrosine phosphatase activity in cells. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The numbers of known tyrosine kinases and tyrosine phosphatases are similar and small compared to the numbers and complexities of their potential substrates (1). The apparent contradiction between this and the high degree of specificity that exists in signaling processes in vivo has made understanding the mechanisms by which PTP activity is regulated of paramount importance. PTP genes often produce several distinct protein products by alternative splicing or use of alternative promoters (3). At the protein level, PTP activity can be inhibited by dimerization (e.g., see references 22, 23, and 43) or by reversible oxidation of the key cysteine residue in the PTP catalytic domain (9, 44). Proteolysis and phosphorylation, which can affect subcellular localization, conformation, or the ability to bind other proteins can also activate or inhibit PTPs (e.g., see references 6, 13, 17, 48, and 49). Regulation of PTP activity by binding of extracellular ligands to receptor-type PTPs (RPTPs), which inhibits PTP activity and appears to be less common than among tyrosine kinases, has also been described (e.g., see reference 14). Despite these insights, major gaps still exist in our understanding of how PTP activity is regulated in a reversible manner and within the short time periods typical of signal transduction events.
The two major forms of PTP known today are RPTP and the non-receptor form (cyt-PTP), which differ only at their N termini and which are products of distinct promoters of the single Ptpre gene (10, 11, 31, 36). Other protein forms of PTP are p67 PTP and p65 PTP, whose production is regulated at the levels of transcription and posttranslational processing, respectively (17, 18). RPTP dephosphorylates c-Src and the related kinases c-Yes and c-Fyn at their C-terminal inhibitory tyrosine, thereby activating the kinases and assisting mammary tumor cells induced in vivo by Neu to maintain their transformed phenotype (16, 19). RPTP has also been suggested to downregulate insulin receptor signaling (2, 25, 28, 30). cyt-PTP is important for proper adhesion of osteoclasts to bone (7) and dephosphorylates and downregulates delayed-rectifier voltage-gated potassium channels in Schwann cells, thereby countering their activation by c-Src and c-Fyn. This effect correlates in vivo with delayed myelination of sciatic nerve axons in young PTP-deficient mice (33, 40, 41) and with delayed optic nerve myelination in mice expressing a "substrate-trapping" mutant of RPTP (29). cyt-PTP can also downregulate signaling mediated by the mitogen-activated protein kinase (42, 45) and JAK-STAT (37-39) pathways and is required for proper functioning of mouse macrophages (35).
In order to better understand at the molecular level the physiological roles and regulation of PTP activity in vivo, we searched for novel interactors with this phosphatase. We report here that PTP interacts in vivo with microtubules in a reversible manner. Binding is stabilized by microtubule-associated proteins (MAPs) and is regulated by physiological stimuli, such as specific phosphorylation of PTP following activation of the epidermal growth factor receptor (EGFR). Binding to tubulin inhibits PTP catalytic activity, suggesting that physiologically relevant signaling events may down-regulate PTP activity by promoting its transient association with microtubules. Our data indicate that tubulin also inhibits PTP1B but does not inhibit RPTPµ, suggesting that while association with tubulin is not universal among PTPs, it is not limited to PTP alone.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(10); RPTP (11), p67 PTP (18); S8M-cyt-PTP, S22M- cyt-PTP, and 
11-27-cyt-PTP (24); D245A cyt-PTP (33); D302A RPTP (16); and D1+12 and D2+12 (43). Y638F cyt-PTP was constructed by PCR-mediated site-directed mutagenesis and sequenced prior to use. The DAP4 cDNA was a gift from Adi Kimchi, The Weizmann Institute; the rat Kv2.1 cDNA was a gift from J. Barhanin and M. Lazdunski, and the EGFR cDNA was a gift from Yosef Yarden, The Weizmann Institute. All cDNAs were cloned in the pcDNA3 expression vector (Invitrogen, Carlsbad, CA) and contained a FLAG tag at their C termini. Cherry-labeled 
-tubulin cDNA was a gift of Benjamin Geiger, The Weizmann Institute. The following antibodies were used in this study: monoclonal anti--tubulin (clone DM1A; Sigma Chemical Co., St. Louis, MO), monoclonal anti-FLAG M2 (Sigma), monoclonal antiactin (clone AC-40; Sigma), monoclonal antiphosphotyrosine (clone PY99; Santa Cruz Biotechnology, Santa Cruz, CA); polyclonal anti-Kv2.1 (Alomone Laboratories, Jerusalem, Israel); monoclonal anti-PTP1B (BD Transduction Laboratories, Franklin Lakes, NJ); monoclonal anti-LIS1 (gift of Orly Reiner, The Weizmann Institute) (34); and monoclonal anti-Tau (Tau5; gift of Irith Ginzburg, The Weizmann Institute). All other reagents were purchased from Sigma. Cell culture and fluorescence microscopy. 293 cells, NIH 3T3 cells, and HeLa-JW cells were grown in Dulbecco's modified Eagle's medium (Invitrogen), supplemented with 10% (vol/vol) fetal calf serum (Invitrogen), 2 mM glutamine, 50 U/ml penicillin, and 50 µg/ml streptomycin. Jurkat cells were grown in RPMI 1640 medium (Sigma) supplemented with heat-inactivated fetal calf serum, glutamine, and antibiotics as described above. Cells were transfected using the calcium phosphate method (5). For fluorescence studies, cells were washed with phosphate-buffered saline (PBS), fixed in methanol at –20°C for 10 min, washed five times with PBS, and mounted with Fluoromount-G solution (Southern Biotech, Birmingham, AL). Cells were examined using a DeltaVision system (Applied Precision, Inc., Issaquah, WA) and Resolve3D software.
Biochemical purification of tubulin and MAPs. Tubulin was purified from fresh calf brains as described previously (15). Tubulin was purified by three cycles of polymerization/depolymerization, followed by chromatography on a phosphocellulose column to remove MAPs and by a fourth cycle of polymerization/depolymerization. Purified tubulin was resuspended in buffer PM2: 100 mM PIPES [piperazine-N,N'-bis(2-ethanesulfonic acid)], 2 mM EGTA, 2 mM MgSO4 (pH 6.9). MAPs were eluted from the phosphocellulose column with 1 M KCl and desalted using a Vivaspin 20-ml centrifugal concentrator (Sartorius Vivascience GmbH, Goettingen, Germany). MAPs were diluted in PM2 buffer to a final concentration of 1.5 µg/µl.
Extraction of tubulin from mouse brain.
Brains from PTP-deficient (EKO) (33) or wild-type (WT) mice were homogenized in MES++ buffer (100 mM MES [morpholineethanesulfonic acid], 0.5 mM MgCl2, 0.1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol [DTT], 1% Nonidet P-40 [pH 7.4]) supplemented with the following protease inhibitors: 1 mM AEBSF [4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride], 40 µM bestatin, 15 µM E-64, 20 µM leupeptin, and 15 µM pepstatin (Sigma). The extracted material was centrifuged at 30 x 103 g and then at 100 x 103 g at 4°C to remove polymerized microtubules. In vitro polymerization of microtubules from the supernatant was performed in G-MES++ buffer (MES++ buffer supplemented with 4 M glycerol, 1 mM GTP, 0.4 mM DTT, 5 µM Taxol, and protease inhibitors) for 45 min at 37°C. Following centrifugation at 100 x 103 g at 37°C, the pellet containing polymerized microtubules was washed three times in MES++ buffer and was dissolved in 200 µl of buffer A (50 mM Tris-Cl [pH 7.5], 150 mM NaCl, 1% Nonidet P-40). The experiment shown in Fig. 1F was carried out in a similar manner, except that purified cyt-PTP was added prior to the polymerization step.
|
expressed in 293 cells were carried out as described previously (41). The purity and amounts of eluted material were determined by gel electrophoresis and silver staining. In some studies, endogenous tubulin or PTP1B was precipitated from 1-mg cell lysates prepared in buffer A using the relevant antibodies and protein A/G Sepharose beads (Santa Cruz). Tyrosine-phosphorylated Kv2.1 for use as a substrate of PTP was purified from 293 cells cotransfected with FLAG-tagged Kv2.1 and activated (Y527F) Src by anti-FLAG immunoprecipitation and elution with FLAG peptide as described previously (41). Sodium dodecyl sulfate (SDS) gel electrophoresis, blotting, and antibody hybridization were done as described previously (18).
Phosphatase and kinase activity assays.
Phosphatase activity was assayed in 96-well plates. Each well typically contained 0.1 µg of PTP in 200 µl of activity assay buffer (50 mM MES, 0.5 mg/ml bovine serum albumin, 0.5 mM DTT, and 10 mM para-nitrophenylphosphate [PNPP], pH 7.0). Under these conditions PTP activity is linear with time and amount of enzyme. Increasing amounts of purified bovine tubulin, bovine serum albumin (BSA; fraction V [MP Biomedicals, Solon, OH]) or actin (Sigma) were added, and activity was measured by monitoring the increase in absorption at 405 nm at room temperature for 1 h. The Km and Vmax for PTP were determined by Lineweaver-Burk analysis by measuring initial PTP activity in the presence of increasing concentrations of PNPP (0.6 to 15.0 mM) for 5 to 10 min, with or without added tubulin. In some studies, the phosphatase activity of 0.3 µg of calf intestinal alkaline phosphatase (CIP; New England Biolabs, Beverly, MA), 0.6 µg of the cytosolic domain of RPTPµ (Jena Biosciences GmbH, Jena, Germany), or 0.3 µg of endogenous PTP1B precipitated from 293 cells was measured similarly. In other studies, 25 ng of eluted phospho-Kv2.1 was incubated either alone or with 30 ng of purified cyt-PTP and with 25 to 50 µg purified bovine brain tubulin, actin, or BSA in PTP activity buffer at 32°C for 2 h. The reaction was terminated by addition of SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer and boiling. Phosphorylation of Kv2.1 was analyzed by SDS-PAGE and protein blotting with anti-pTyr antibodies. Src activity assay was performed using enolase according to reference 16.
Interaction between tubulin and PTP.
Eluted cyt-PTP (0.1 µg) was incubated with 20 µg purified bovine tubulin, either with or without 15.5 µg eluted bovine MAPs, for 1 h at room temperature. PTP was immunoprecipitated with anti-FLAG antibodies, and associated tubulin was detected by protein blotting. To examine the interaction in cells following EGFR stimulation, endogenous or exogenous EGFR was stimulated in 293 cells with 100 ng/ml of EGF (Sigma) or vehicle (PBS). Cells were lysed in buffer A supplemented with 0.5 mM sodium vanadate at various time points; FLAG-tagged cyt-PTP was immunoprecipitated and examined by protein blotting. Detection of cyt-PTP phosphorylation following EGFR stimulation was done in a similar manner using buffer A containing 0.5 mM sodium pervanadate.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
binds tubulin in vivo.
In order to identify proteins that interact with PTP in cells, we expressed the nonreceptor form of the phosphatase, cyt-PTP, in 293 cells and examined proteins that coprecipitated with it. In this study, we utilized D245A cyt-PTP, a "substrate-trapping" mutant of PTP that is virtually inactive but retains the ability to bind substrates and interacting molecules in a cellular context (12). SDS-PAGE analysis of D245A cyt-PTP immunoprecipitates revealed a coprecipitating protein with a mass of approximately 55 kDa that was subsequently identified by mass spectrometry as -tubulin (not shown). Further immunoprecipitation experiments confirmed that cyt-PTP as well as RPTP bound -tubulin (Fig. 1A). Both WT and D-to-A trapping mutants of cyt-PTP and RPTP bound -tubulin similarly (Fig. 1A); this result, therefore, does not shed light on whether or not tubulin is a substrate of PTP, although it does indicate that a complex that contains both molecules exists. Binding was specific since a nonrelated protein, DAP4, did not bind -tubulin and since actin, a second cytoskeletal protein of high abundance, did not coprecipitate with PTP (Fig. 1A and B). The association between PTP and -tubulin was observed also in HeLa and in A431 cells, as well as when the experiment was performed in reverse by precipitating -tubulin (see Fig. 3B) (data not shown). Similar results were obtained also when binding to ß-tubulin was examined (not shown); subsequent experiments were therefore performed with antibodies against -tubulin alone. In separate experiments, comparison of the fraction of FLAG-tagged cyt-PTP molecules that coprecipitated with a given fraction of precipitated tubulin molecules in 293 cells indicated that 10 to 15% of cyt-PTP molecules and 0.3 to 0.4% of tubulin molecules participate in the PTP-tubulin interactions. The last result agrees with the expression level of tubulin in these cells, which is significantly higher than that of cyt-PTP.
|
and -tubulin was examined. Accordingly, endogenous cyt-PTP coprecipitated with -tubulin from Jurkat T cells (Fig. 1C). In separate experiments, mouse brain lysate, which had been cleared of polymerized microtubules by centrifugation (Fig. 1D, lanes 1 and 2), was incubated at 37°C in the presence of GTP and glycerol to promote assembly of nonpolymerized tubulin dimers into microtubules. Following their collection by centrifugation, examination of the newly assembled microtubules (Fig. 1D, lanes 3 and 4) revealed that significant amounts of endogenous RPTP had coassembled and copurified with the precipitated microtubules. Similar results were obtained when binding of endogenous RPTP to microtubules was examined in mammary tumor cells induced in vivo in mice by expression of an activated Neu transgene (Fig. 1E). Tau, a well-known MAP, was also enriched significantly in the precipitated microtubule fraction (Fig. 1D). Similar binding was observed when exogenous FLAG-tagged cyt-PTP that had been purified from overexpressing cells was added to depolymerized tubulin obtained from mice genetically lacking PTP prior to polymerization in vitro (Fig. 1F).
The interaction between PTP and tubulin was reconstituted in vitro using purified cyt-PTP and tubulin. These studies used tubulin from bovine brain that had undergone three cycles of assembly/disassembly and was stripped of tightly associated MAPs by phosphocellulose column chromatography. Tubulin was first incubated with cyt-PTP, cyt-PTP was then immunoprecipitated via its FLAG tag, and the presence of associated -tubulin was examined by protein blotting. As seen in Fig. 1G, PTP-bound -tubulin was readily detected under these conditions. Interestingly, binding was strengthened significantly when MAPs, which were previously removed from tubulin during its purification, were added to the incubation mixture (Fig. 1G). Neither -tubulin nor PTP was detected in the MAP fraction. We conclude, therefore, that the tubulin-cyt-PTP interaction is stabilized significantly by one or more components of the MAP fraction, which are physiological interactors of microtubules. In all, these results indicate that stable interactions exist between microtubules and the major forms of PTP, RPTP and cyt-PTP.
cyt-PTP colocalizes with microtubules in cells.
The interaction between cyt-PTP and microtubules was visualized in HeLa cells expressing fluorescently tagged cyt-PTP and -tubulin. Cells were fixed and permeabilized in cold methanol, which removes most soluble cytosolic molecules including nonbound cyt-PTP. cyt-PTP and tubulin colocalized throughout the cytosol in long fibrous structures (Fig. 2A to C and G to I); cyt-PTP was detected also in the nucleus, in agreement with previous studies (24). Treatment of cells with nocodazole disrupted the cellular microtubule array and led to a diffuse, punctate cytosolic localization pattern for tubulin (Fig. 2E). Importantly, nocodazole treatment disrupted localization of cyt-PTP (Fig. 2D and F), confirming that cyt-PTP was associated with microtubules prior to nocodazole treatment. The qualitative nature of this experiment precluded quantitative conclusions about the effect of nocodazole on the PTP-microtubule interaction. In order to address the last point, we expressed cyt-PTP in 293 cells and disrupted their microtubules with nocodazole or colchicine. Coprecipitation of cyt-PTP with -tubulin was reduced in treated cells (Fig. 2J), indicating that disruption of the microtubule array reduces binding of PTP. Stabilization of microtubules with taxol enhanced this interaction in some experiments (e.g., Fig. 2J) but not in others and requires further studies. Disruption of actin cytoskeletal components with cytochalasin D did not affect the PTP-tubulin interaction. Residual binding of PTP after disruption of microtubular structures, which is evident by immunofluorescence (Fig. 2D to F) and biochemically (Fig. 2J), may arise from incomplete depolymerization of microtubules or from weaker binding of cyt-PTP to nonpolymerized tubulin heterodimers. Both explanations are consistent with cyt-PTP binding preferentially microtubules in cells.
|
play major roles in binding tubulin.
Throughout the above studies, full-length cyt-PTP consistently bound more -tubulin than p67 PTP, a naturally occurring form of cyt-PTP that lacks the 27 N-terminal residues (18) (Fig. 3A and B). This observation suggested that the N-terminal domain of cyt-PTP mediates at least part of the ability of this PTP to bind -tubulin. Further studies indicated that deletion of the first eight amino acids of cyt-PTP did not affect association between cyt-PTP and -tubulin (Fig. 3B). In contrast, deletion of residues 1 to 22 or 11 to 27 or deletion of a larger portion of cyt-PTP beginning at residue 13 and extending to the C terminus of the first PTP domain (construct D2+12) reduced binding to levels observed with p67 PTP (Fig. 3B and C). These results suggest that at a minimum, residues between positions 13 and 22 participate in binding of cyt-PTP to -tubulin. These residues do not contain sequence motifs known to regulate binding to tubulin heterodimers or to microtubules; they do, however, include several charged residues and a potential phosphorylation site that may regulate binding indirectly.
In order to examine the role of the second, inactive (D2) PTP domain of cyt-PTP in binding -tubulin, we examined the tubulin-binding abilities of cyt-PTP lacking this domain (Fig. 3A). Removal of the D2 PTP domain greatly reduced binding of cyt-PTP to -tubulin (Fig. 3C, left panel), attesting to its central role in mediating this interaction. Binding of Y638F cyt-PTP, which lacks a prominent phosphorylation site at the C terminus of D2, was reduced significantly (Fig. 3D). This indicates that the presence of Y638 plays an important, but not exclusive, role in mediating the contribution of D2 to the PTP-tubulin interaction. Participation of several domains of PTP molecules, including the D2 domain, in intermolecular associations of PTPs has been noted before (e.g., see reference 22).
Tubulin inhibits PTP activity in a specific manner.
In order to examine in vitro the consequences of the interaction between tubulin and cyt-PTP we determined the relative amounts of both proteins in cells and tissues. For this purpose, protein lysates from several sources were analyzed by protein blotting. Membranes were probed with antibodies against -tubulin or PTP; the absolute amount of either protein in each sample was estimated by comparison to the signal generated by known amounts of purified tubulin or cyt-PTP that were analyzed in parallel. As expected, a marked excess of -tubulin was detected in all samples, with the -tubulin/PTP molar ratio (calculated per -tubulin monomer) ranging from approximately 10:1 in Neu-induced mouse mammary tumor cells particularly rich in RPTP to 1,315:1 in murine embryo fibroblasts (Fig. 4A).
|
may affect either molecule or other molecules that interact with them. In order to begin addressing this issue, we examined the catalytic activity of cyt-PTP in vitro in the presence of purified tubulin. Addition of tubulin in excesses of 100-, 500-, and 1,000-fold versus cyt-PTP reduced PTP activity to 79.9%, 42.2%, and 25.4% of activity in the absence of tubulin, respectively (Fig. 4B). Heat-treated tubulin did not inhibit cyt-PTP (Fig. 4B). Inhibition was observed also when purified tubulin was added to purified samples of RPTP or p67 PTP (Fig. 4B and results not shown), indicating that tubulin inhibits activity of all the major forms of PTP in vitro under the conditions used. In agreement with its weak binding to tubulin, activity of cyt-PTP lacking the D2 domain (construct D1+12 [Fig. 3C]) was only slightly inhibited by tubulin. Kinetic analysis of cyt-PTP activity in the presence of purified tubulin indicated that tubulin is a noncompetitive inhibitor of the phosphatase. Addition of 40 µg purified tubulin to 0.1 µg purified cyt-PTP did not affect the Km of cyt-PTP for PNPP, which remained at 4.57 ± 0.47 mM, but did reduce the maximal rate of the dephosphorylation reaction from 2.47 ± 0.35 to 0.86 ± 0.13 µmol/min/mg enzyme (mean ± standard error [SE]; n = 3; P = 0.0128 by Student's t test). The ability of tubulin to inhibit PTP in vivo is likely to be stronger than measured here given the more significant amounts of microtubules and the presence of MAPs in cells, both of which increase the PTP-tubulin interaction. These results also suggest that tubulin is not a substrate of cyt-PTP and does not inhibit the phosphatase by competing against other substrates, in agreement with significant involvement of regions outside the D1 PTP domain in binding tubulin (Fig. 3).
Further studies indicated that the effect of tubulin was specific, since addition of 100-, 500-, or 1,000-fold excesses of actin or of BSA did not inhibit cyt-PTP (Fig. 4C). Addition of actin actually increased cyt-PTP activity somewhat, possibly due to increased effective concentrations of both cyt-PTP and PNPP due to molecular exclusion effects of the added actin. Additional studies indicated that a different tyrosine phosphatase, PTP1B, is also inhibited in vitro by tubulin, but that a third PTP, the receptor-type RPTPµ, and alkaline phosphatase, which is not a PTP, are not inhibited (Fig. 4C). RPTPm contains a C-terminal tyrosine residue embedded in a sequence distinct from that of PTP; it does not contain a sequence similar to the N-terminal sequence of cyt-PTPe. Binding of microtubules is therefore not unique to PTP within the PTP family and needs to be examined for individual PTPs.
We next examined whether modulation of the interaction between cyt-PTP and microtubules affects phosphatase activity. Treatment of cells with nocodazole reduced binding of cyt-PTP to tubulin (Fig. 2J), an effect that would be expected to increase cyt-PTP activity. In order to test this, we treated 293 cells expressing cyt-PTP with nocodazole, immunoprecipitated the phosphatase, and assayed its activity. cyt-PTP isolated from cells treated with nocodazole was 20.3% more active than the enzyme from untreated cells (Fig. 5A). Since only a fraction of cyt-PTP molecules bind tubulin, the data in Fig. 5A, which were obtained using total cellular cyt-PTP, most likely underestimate the effect of nocodazole on cyt-PTP activity in cells. Importantly, this experiment demonstrates that association with microtubules can affect the activity of cyt-PTP isolated from cells and that its effect is not limited to experiments in which purified components are combined in vitro.
|
and tubulin in vitro (Fig. 1G), suggesting that MAPs may further reduce cyt-PTP activity by stabilizing the PTP-tubulin interaction. In agreement, addition of 25 µg purified tubulin to 0.1 µg purified cyt-PTP reduced cyt-PTP activity to 85.1% of cyt-PTP activity in the absence of tubulin, while the presence of MAPs decreased activity further to 72.3% (Fig. 5B). In the absence of tubulin, MAPs did not inhibit cyt-PTP activity (not shown), consistent with MAPs reducing cyt-PTP activity by promoting its interaction with tubulin.
We next examined the effect of purified tubulin on the in vitro activity of cyt-PTP towards a physiological substrate—the voltage-gated potassium channel Kv2.1. Activation of Kv2.1 by depolarization of the cell membrane can be increased by phosphorylation at Y124 by c-Src or c-Fyn; cyt-PTP counters this effect by dephosphorylating Kv2.1 mainly at Y124 (33, 40, 41). Kv2.1 was coexpressed with activated (Y527F) Src in 293 cells, and phosphorylated Kv2.1 (pKv2.1) was immunoprecipitated and purified. Equal amounts of pKv2.1 were incubated with cyt-PTP in the presence or absence of a 250- or 500-fold molar excess of purified tubulin, actin, or BSA. The remaining pKv2.1 was detected by blotting the purified Kv2.1 with antiphosphotyrosine antibodies. In agreement with the data obtained using PNPP (Fig. 4), the presence of tubulin, but not of actin or BSA, inhibited dephosphorylation of pKv2.1 by cyt-PTP (Fig. 5C). Finally, cyt-PTP activates Src (16), suggesting that disrupting the PTP-microtubule interaction may further increase Src activity in cells. Treatment of 293 cells expressing WT Src with nocodazole reduced Src activity to 55.2%±13.3% of Src activity in nontreated cells (mean ± SE; P = 0.0266 by paired t test; n = 5 to 6). Expression of cyt-PTP increased Src activity slightly to 110.2% ± 18.6%, while treatment of cyt-PTP-expressing cells with nocodazole increased Src activity further to 160.3% ± 31.7% (mean ± SE; P = 0.0318 by paired t test; n = 6 [Fig. 5D]). Increased Src activity in the presence of PTP in cells following microtubule disruption is consistent with microtubules inhibiting cyt-PTP in vivo, although further studies are required to determine whether this is caused exclusively by breakdown of the PTP-microtubule interaction.
EGFR-mediated phosphorylation at Y638 drives cyt-PTP to bind tubulin in vivo.
Previous studies have shown that activation of the EGFR results in phosphorylation of cyt-PTP within minutes in a temporal pattern that follows closely autophosphorylation of the EGFR itself. This phenomenon occurs when either exogenous (41) or endogenous (Fig. 6A) EGFR is examined. Interestingly, EGFR signaling also resulted in increased association between cyt-PTP and -tubulin that followed a very similar time course (Fig. 6B). Basal association between cyt-PTP and tubulin was detected in the cells studied (not visible in the exposure shown in Fig. 6B). Association between tubulin and cyt-PTP increased significantly as early as 1 min following addition of EGF to the cells; the association peaked at 2.5 and 5 min and receded afterwards. Disruption of microtubules with colchicine reduced basal association between cyt-PTP and tubulin and prevented its increase following stimulation with EGF (Fig. 6C).
|
phosphorylation, and association between -tubulin and cyt-PTP suggests that phosphorylation of cyt-PTP may regulate its ability to bind tubulin. A prime candidate site for EGFR-mediated phosphorylation of cyt-PTP is its C-terminal Y638, whose presence supports the PTP-tubulin interaction (Fig. 3D). Y638 is analogous to Y695 of RPTP and to Y789 of the related RPTP, whose phosphorylation is crucial for PTP-mediated activation of c-Src (4, 49). In agreement, Y638F cyt-PTP was not phosphorylated even following activation of exogenous EGFR, which phosphorylates WT cyt-PTP and promotes its association with microtubules more strongly than the endogenous EGFR (Fig. 7A) (data not shown). This result indicates that EGFR-mediated phosphorylation of cyt-PTP is highly specific and affects this single tyrosine from among the 22 tyrosine residues present in this phosphatase.
|
at Y638 is important for cyt-PTP-tubulin binding, we examined this association in 293 cells following stimulation of their endogenous EGFR. As in previous experiments, the association between WT cyt-PTP and -tubulin increased transiently during the first few minutes following EGFR stimulation. In contrast, no increase in binding of -tubulin was observed in cells expressing Y638F cyt-PTP (Fig. 7B). cyt-PTP autodephosphorylates strongly (4), a finding that precluded isolation of cyt-PTP and measuring its activity while still phosphorylated and associated with microtubules immediately following EGFR activation. Nevertheless, results presented in this study indicate that cyt-PTP is inhibited by associating with microtubules in cells and that this association is influenced by transient phosphorylation of cyt-PTP at Y638 that occurs in vivo following physiological activation of the EGFR. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and that this association inhibits activity of PTP. The association between microtubules and the two major forms of PTP, RPTP and cyt-PTP, can be detected biochemically when the endogenous proteins are examined as well as when PTP is expressed in a variety of cells. Furthermore, tubulin and cyt-PTP colocalize in microtubule structures, and the organizations of both proteins are affected strongly and similarly when microtubules are disrupted. Importantly, actin, another major cytoskeletal protein, does not associate with PTP under the same conditions, attesting to the specificity of this interaction.
The tubulin-PTP interaction can also be recapitulated in vitro using purified proteins. A striking finding here is that the interaction is strengthened significantly in the presence of MAPs (Fig. 1G). The finding that a third protein component that is known to bind microtubules physiologically stabilizes the PTP-tubulin interaction further supports its physiological context. This finding also suggests that the interaction between PTP and microtubules may be indirect, accounting in part for how PTP and PTP1B, which are structurally distinct, can both be influenced by microtubules. Finally, the requirement for MAPs and absence of PTP from the MAP fraction itself indicate that the phosphatase is not among the proteins that interact most tightly with microtubules, further suggesting that this interaction may be subject to dynamic regulation by physiological processes.
Analysis of the structural requirements of this interaction sheds light on its regulation and further supports its physiological relevance. The role of the D2 PTP domain is crucial, since deleting it significantly reduces the interaction. Phosphorylation of Y638 in the D2 domain of cyt-PTP (Y695 in RPTP) plays a significant, but not exclusive, role in supporting the interactions mediated by D2. The role of the short N-terminal region of cyt-PTP is also striking, since removal of residues 13 to 22 of this domain decreases binding by about 50%. Several tubulin-binding sequence motifs have been described to date, such as the imperfect 18-residue repeats of MAP2 and Tau (21, 27), multiple copies of the short KKEE/I/V motif found in MAP1B (32), or the PX6EX4L motif found in dynein heavy chain, kinesin, and related molecules (46). None of these motifs are present in cyt-PTP, suggesting that cyt-PTP binds tubulin via an uncharacterized domain or that binding is indirect. As indicated above, the latter possibility is consistent with enhancement of the cyt-PTP-tubulin interaction by MAPs. The first 27 residues of cyt-PTP contain 9 positively charged residues and 6 serine/threonine potential phosphorylation sites; of these, 4 charged residues and 1 serine are located between positions 13 and 22. Further study is required to determine if any of these residues participate in indirect binding of tubulin. The precise molecular mechanism by which tubulin binding inhibits PTP is not clear at the present time. The noncompetitive nature of this inhibition suggests that bound tubulin or MAPs may alter the conformation of PTP, thus reducing its activity. Complete accounting of this effect will be aided by identification of the relevant MAP(s) and determination of the structure of the PTP-MAP/tubulin complex.
Binding of PTP to tubulin may affect the phosphatase, microtubules, or other proteins that associate with microtubules. Limited available information indicates that tubulin is phosphorylated in B lymphocytes by the tyrosine kinase Syk and following stimulation of Jurkat T cells with alpha interferon (47). In other studies, c-Fes phosphorylated tubulin and promoted tubulin polymerization in vitro (26). The tyrosine phosphatase PTP-BL localizes to centrosomes during interphase and metaphase and to the spindle midzone during anaphase, and elevated PTP-BL expression leads to defects in cytokinesis (20). CDC14B, a dual-specificity phosphatase, binds and stabilizes microtubules in vitro and in interphase cells (8). We detected basal phosphorylation of tubulin, although this was not affected by expression of PTP (not shown).
In this study, we focused on the effect of binding on PTP activity, showing that presence of tubulin inhibits cyt-PTP activity in vitro when both proteins are present in ratios in the range found in cells and tissues. Importantly, tubulin inhibited PTP activity also towards Kv2.1 and towards Src, both of which are physiological substrates of the phosphatase. The finding that tubulin inhibits PTP activity while similar amounts of actin or BSA do not do so argues strongly that tubulin has a unique effect on the phosphatase. Similarly, the ability of tubulin to inhibit the tyrosine phosphatase PTP1B but not RPTPm or the unrelated alkaline phosphatase suggests that the effect of tubulin is not common to all types of phosphatases and should be examined on a case-by-case basis within the PTP family. Of note, PTP1B does not contain the N-terminal and D2 domains of cyt-PTP, which play important roles in binding tubulin. This suggests that the mechanisms by which both PTPs interact with tubulin are distinct.
In support of the above results, treatments that affect binding of cyt-PTP to microtubules cause the expected changes in cyt-PTP activity. Accordingly, addition of MAPs purified from bovine brain, which increase binding of cyt-PTP to tubulin, significantly increased the inhibitory effect of tubulin on activity of purified cyt-PTP (Fig. 5B). These results indicate that interactions with other proteins may modify further the effect of tubulin on PTP activity in vivo. In other experiments, disruption of the cellular microtubule network by nocodazole treatment increased activity of cyt-PTP purified from these cells (Fig. 5A) and increased activation of Src by cyt-PTP in the cells (Fig. 5D). Importantly, the last results show that PTP is inhibited by its interaction with tubulin not only in vitro but also in intact cells. We note that nocodazole-induced mitotic arrest of cells increased activity of the related RPTP and increased its ability to activate Src (50) in a manner suggested to be linked with phosphorylation of the juxtamembrane residues S180 and S204 of RPTP (48). Although the effect of nocodazole on RPTP activity is consistent with results presented in the present study, the experimental systems are not identical. In particular, nocodazole-induced mitotic arrest requires significantly longer treatment of cells with nocodazole (10 h [50] versus 45 min in the present study); furthermore, S180 is not conserved in PTP, and replacement of the residue in PTP analogous to S204 with an aspartic acid residue did not affect PTP activity towards Src (4). It remains, therefore, to be determined whether the molecular mechanisms that regulate phosphatase activity are the same in both cases.
Binding of cyt-PTP to tubulin is affected by physiological signals in a reversible manner and with rapid kinetics typical of signaling processes. Activation of the EGFR results in phosphorylation of cyt-PTP specifically at its C-terminal Y638, an event that is required for increased binding of the phosphatase to tubulin in this context. Increased binding and phosphorylation of cyt-PTP occur rapidly after receptor activation and follow closely the time course of receptor autophosphorylation. Since association with microtubules reduces activity of cyt-PTP in cells (Fig. 5A), this result makes it tempting to speculate that increased binding of cyt-PTP to microtubules during the initial stages of EGFR activation reduces phosphatase activity and removes it from the immediate vicinity of the receptor. The dependence on specific phosphorylation of cyt-PTP downstream of the EGFR should ensure that cyt-PTP binds microtubules in concert with a true physiological signal and in a manner most likely limited to the vicinity of the activated receptor. At later time points, phosphorylation of cyt-PTP and its binding to microtubules are reduced, restoring the conditions that existed prior to EGFR stimulation. This latter process is most likely driven by the normal down-regulation of EGFR signaling and by auto- and trans-dephosphorylation of cyt-PTP (4). We note that the receptor type of PTP, RPTP, binds microtubules constitutively in mammary tumor cells induced in mice by Neu. RPTP is constitutively phosphorylated at Y695 (Y638 of cyt-PTP) in these cells (4), providing another correlation between phosphorylation of PTP at its C-terminal tyrosine downstream of a (constitutively active) growth factor receptor and increased binding of microtubules. RPTP activates Src and supports the transformed phenotype of these tumor cells (16), suggesting that the fraction of RPTP molecules that bind microtubules in this system may perform a distinct role. The precise mechanism by which phosphorylation of PTP promotes its binding to microtubules is unknown at present. It is possible that phosphorylation promotes association of cyt-PTP with additional proteins that are needed to stabilize the interaction with microtubules, in agreement with promotion of this interaction by MAPs. Further studies are required to address this issue.
In summary, this study suggests that association of PTP with microtubules is a novel mechanism for regulation of PTP activity that is regulated by physiological signals in a dynamic and reversible manner. We note that our data do not rule out other possible roles for the tubulin-PTP association, such as regulating microtubule dynamics or providing access for PTP to other molecules that also bind microtubules. The data suggest that the effect of microtubules is not limited to PTP but may affect other tyrosine phosphatases, thereby providing an additional mechanism for physiological regulation of phosphatase activity.
|
ACKNOWLEDGMENTS |
|---|
This study was supported by a grant from the Israel Science Foundation, founded by the Israel Academy of Sciences and Humanities.
|
FOOTNOTES |
|---|
Published ahead of print on 20 August 2007.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Andersen, J. N., A. Elson, R. Lammers, J. Romer, J. T. Clausen, K. B. Moller, and N. P. Moller. 2001. Comparative study of protein tyrosine phosphatase-epsilon isoforms: membrane localization confers specificity in cellular signalling. Biochem. J. 354:581-590.[CrossRef][Medline]
3. Andersen, J. N., P. G. Jansen, S. M. Echwald, O. H. Mortensen, T. Fukada, R. Del Vecchio, N. K. Tonks, and N. P. Moller. 2004. A genomic perspective on protein tyrosine phosphatases: gene structure, pseudogenes, and genetic disease linkage. FASEB J. 18:8-30.
4. Berman-Golan, D., and A. Elson. 2007. Neu-mediated phosphorylation of protein tyrosine phosphatase epsilon is critical for activation of Src in mammary tumor cells. Oncogene [E-pub ahead of print.] doi:10.1038/sj.onc.1210505.
5. Chen, C. A., and H. Okayama. 1988. Calcium phosphate-mediated gene transfer: a highly efficient transfection system for stably transforming cells with plasmid DNA. BioTechniques 6:632-638.[Medline]
6. Chen, M., S. C. Chen, and C. J. Pallen. 2006. Integrin-induced tyrosine phosphorylation of protein-tyrosine phosphatase-alpha is required for cytoskeletal reorganization and cell migration. J. Biol. Chem. 281:11972-11980.
7. Chiusaroli, R., H. Knobler, C. Luxenburg, A. Sanjay, S. Granot-Attas, Z. Tiran, T. Miyazaki, A. Harmelin, R. Baron, and A. Elson. 2004. Tyrosine phosphatase epsilon is a positive regulator of osteoclast function in vitro and in vivo. Mol. Biol. Cell 15:234-244.
8. Cho, H. P., Y. Liu, M. Gomez, J. Dunlap, M. Tyers, and Y. Wang. 2005. The dual-specificity phosphatase CDC14B bundles and stabilizes microtubules. Mol. Cell. Biol. 25:4541-4551.
9. den Hertog, J., A. Groen, and T. van der Wijk. 2005. Redox regulation of protein-tyrosine phosphatases. Arch. Biochem. Biophys. 434:11-15.[CrossRef][Medline]
10. Elson, A., and P. Leder. 1995. Identification of a cytoplasmic, phorbol ester-inducible isoform of protein tyrosine phosphatase epsilon. Proc. Natl. Acad. Sci. USA 92:12235-12239.
11. Elson, A., and P. Leder. 1995. Protein-tyrosine phosphatase epsilon. An isoform specifically expressed in mouse mammary tumors initiated by v-Ha-ras OR neu. J. Biol. Chem. 270:26116-26122.
12. Flint, A. J., T. Tiganis, D. Barford, and N. K. Tonks. 1997. Development of "substrate-trapping" mutants to identify physiological substrates of protein tyrosine phosphatases. Proc. Natl. Acad. Sci. USA 94:1680-1685.
13. Frangioni, J. V., A. Oda, M. Smith, E. W. Salzman, and B. G. Neel. 1993. Calpain-catalyzed cleavage and subcellular relocation of protein phosphotyrosine phosphatase 1B (PTP-1B) in human platelets. EMBO J. 12:4843-4856.[Medline]
14. Fujikawa, A., D. Shirasaka, S. Yamamoto, H. Ota, K. Yahiro, M. Fukada, T. Shintani, A. Wada, N. Aoyama, T. Hirayama, H. Fukamachi, and M. Noda. 2003. Mice deficient in protein tyrosine phosphatase receptor type Z are resistant to gastric ulcer induction by VacA of Helicobacter pylori. Nat. Genet. 33:375-381.[CrossRef][Medline]
15. Fygenson, D. K., E. Braun, and A. Libchaber. 1994. Phase diagram of microtubules. Phys. Rev. E 50:1579-1588.[CrossRef]
16. Gil-Henn, H., and A. Elson. 2003. Tyrosine phosphatase-epsilon activates Src and supports the transformed phenotype of Neu-induced mammary tumor cells. J. Biol. Chem. 278:15579-15586.
17. Gil-Henn, H., G. Volohonsky, and A. Elson. 2001. Regulation of protein-tyrosine phosphatases alpha and epsilon by calpain-mediated proteolytic cleavage. J. Biol. Chem. 276:31772-31779.
18. Gil-Henn, H., G. Volohonsky, H. Toledano-Katchalski, S. Gandre, and A. Elson. 2000. Generation of novel cytoplasmic forms of protein tyrosine phosphatase epsilon by proteolytic processing and translational control. Oncogene 19:4375-4384.[CrossRef][Medline]
19. Granot-Attas, S., and A. Elson. 2004. Protein tyrosine phosphatase epsilon activates Yes and Fyn in Neu-induced mammary tumor cells. Exp. Cell Res. 294:236-243.[CrossRef][Medline]
20. Herrmann, L., T. Dittmar, and K. S. Erdmann. 2003. The protein tyrosine phosphatase PTP-BL associates with the midbody and is involved in the regulation of cytokinesis. Mol. Biol. Cell 14:230-240.
21. Himmler, A., D. Drechsel, M. W. Kirschner, and D. W. Martin, Jr. 1989. Tau consists of a set of proteins with repeated C-terminal microtubule-binding domains and variable N-terminal domains. Mol. Cell. Biol. 9:1381-1388.
22. Jiang, G., J. den Hertog, and T. Hunter. 2000. Receptor-like protein tyrosine phosphatase homodimerizes on the cell surface. Mol. Cell. Biol. 20:5917-5929.
23. Jiang, G., J. den Hertog, J. Su, J. Noel, J. Sap, and T. Hunter. 1999. Dimerization inhibits the activity of receptor-like protein-tyrosine phosphatase-alpha. Nature 401:606-610.[CrossRef][Medline]
24. Kraut, J., G. Volohonsky, H. Toledano-Katchalski, and A. Elson. 2002. Nuclear localization of non-receptor protein tyrosine phosphatase epsilon is regulated by its unique N-terminal domain. Exp. Cell Res. 281:182-189.[CrossRef][Medline]
25. Lacasa, D., N. Boute, and T. Issad. 2005. Interaction of the insulin receptor with the receptor-like protein tyrosine phosphatases PTPalpha and PTPepsilon in living cells. Mol. Pharmacol. 67:1206-1213.
26. Laurent, C. E., F. J. Delfino, H. Y. Cheng, and T. E. Smithgall. 2004. The human c-Fes tyrosine kinase binds tubulin and microtubules through separate domains and promotes microtubule assembly. Mol. Cell. Biol. 24:9351-9358.
27. Lewis, S. A., D. H. Wang, and N. J. Cowan. 1988. Microtubule-associated protein MAP2 shares a microtubule binding motif with tau protein. Science 242:936-939.
28. Moller, N. P., K. B. Moller, R. Lammers, A. Kharitonenkov, E. Hoppe, F. C. Wiberg, I. Sures, and A. Ullrich. 1995. Selective down-regulation of the insulin receptor signal by protein-tyrosine phosphatases alpha and epsilon. J. Biol. Chem. 270:23126-23131.
0.0043 by the Mann-Whitney test). Heat-treated (Heat inact.) tubulin did not inhibit cyt-PTP
