Regulation of Phosphorylation Level and Distribution of PTP36, a Putative Protein Tyrosine Phosphatase, by Cell-Substrate Adhesion*

Recently we have cloned a putative protein tyrosine phosphatase, PTP36/PTPD2/pez, which possesses a domain homologous to the N-terminal half of band 4.1 protein. In mouse fibroblasts adhered to substrates, PTP36 was phosphorylated on serine residues. PTP36 was found to make complexes with serine/threonine kinase(s), which phosphorylated PTP36 in vitro. PTP36 was dephosphorylated rapidly when the cell-substrate adhesion was disrupted and it was phosphorylated again along with the reattachment of the cells to fibronectin. Rephosphorylation of PTP36 seemed to depend on actin polymerization since it was inhibited by cytochalasin D. The cell detachment also induced the translocation of PTP36 into the membrane-associated cytoskeletal fraction. Staurosporine and ML-9, which inhibited the phosphorylation of PTP36 in vivo, induced the translocation of PTP36 too. On the contrary, when the dephosphorylation of PTP36 was inhibited by okadaic acid, no translocation of PTP36 was induced by the cell detachment. These results demonstrate that the cell-substrate adhesion and cell spreading regulates the intracellular localization of PTP36 most likely through its phosphorylation and therefore, PTP36 may play important roles in the signal transduction pathway of cell-adhesion.

Recently, we have cloned a murine putative PTP, PTP36 (8). A human homologue, PTPD2/pez, was also reported (9,10).  (11). So far, five mammalian PTPs belong to the band 4.1 superfamily and they fall into three subgroups based on their structure. The first subgroup is composed of PTPH1 (12) and PTPMEG (13). The members of the second subgroup are PTPD1/PTP-RL10/rPTP2E (9,14,15) and PTP36/PTPD2/pez (8 -10). The third subgroup has so far only one member, PTP-BAS/hPTP1E/PTPL1/FAP-1 (16 -18). PTPs of the first and the second subgroup have one N-terminal band 4.1 homology domain and one C-terminal PTP domain separated by an intervening sequence. The intervening sequence of the first subgroup has a PDZ domain and a potential SH3binding site. Although there is poor homology, the intervening sequence of the second subgroup also has a potential SH3 or WW binding motif. Unlike these PTPs, PTP-BAS/hPTP1E/ PTPL1/FAP-1, a PTP of the third subgroup, has a band 4.1 homology domain in the middle of the molecule. It has five PDZ domains and interacts with a GTPase-activating protein for Rho (PARG1) (19) and the C terminus of human Fas (CD95) (20,21) through its PDZ domains.
Conservation of the band 4.1 superfamily PTP in Caenorhabditis elegans but not in yeast might suggest its fundamental roles in the multicellular organism. However, the biological roles of the band 4.1 superfamily PTP are largely unknown. Recently we have found that the overexpression of PTP36 in the HeLa cells induces changes in cytoskeletons, focal adhesions, cell-substrate adhesions, and cell growth (22). It is also reported that COS-7 cells overexpressing PTPMEG grow slower, reach confluence at a lower cell density, and make much fewer colonies in soft agar (23,24). These results suggest that the band 4.1 superfamily PTP might have a close relationship to the signal transduction of cell adhesion. However, it was not known if cell adhesion had any effect on the band 4.1 superfamily PTP.
Here we report the regulation of PTP36 by cell-substrate adhesion. Cell-substrate adhesion regulates the intracellular localization of PTP36 most likely through its phosphorylation. Therefore, PTP36 could play an important role in the signal transduction pathways of cell-substrate adhesion.
cin. To test the effects of various inhibitors, cells (1 ϫ 10 5 in 1 ml) were seeded in the 24-well tissue culture plate. The next day, cells were incubated with various inhibitors in DMEM containing 5% fetal calf serum and lysed directly on the wells by 50 l of SDS-PAGE sample buffer. For the cell-detachment experiments, cells were washed with phosphate-buffered saline (PBS) containing 5 mM EDTA and detached from the tissue culture dish by scraping. Alternatively, cells were detached by the incubation with trypsin (2.5 mg/ml) in PBS for no longer than 10 min. For the reattachment experiments, the trypsinized cells were washed by PBS containing 0.1% bovine serum albumin (BSA) and resuspended in DMEM containing 0.1% BSA. Then the cells were incubated in the precoated 24-well tissue culture plate at 37°C for 20 min. Floating cells were recovered by centrifugation, lysed, and mixed with the lysate from the attached cells. Each well of the plate was coated by incubation with 150 l of fibronectin (10 g/ml) or heattreated (80°C for 20 min) BSA (5 mg/ml) in PBS for 1 h at room temperature.
Phosphoamino acid analysis and phosphopeptide analysis were performed as described (27). The hydrolyzed phosphoamino acid or the phosphopeptides digested by TPCK-trypsin (Worthington, NJ) was resolved in two dimensions on the TLC plates (catalog number 5716, Merck, Darmstadt) using the Multiphor II horizontal electrophoresis system (Amersham Pharmacia Biotech).
Dephosphorylation of PTP36 in Vitro-The phosphorylated PTP36 was immunoprecipitated with 21-4 mAb from 3T3/P36-8 cells. The dephosphorylation of PTP36 in vitro was done as described (28  mM sodium pyrophosphate. After PTP36 was immunoprecipitated with 21-4 followed by MAR 18.5, the sample was washed once with the kinase buffer (25 mM Tris-HCl, pH 7.4, 4 mM MgCl 2 , 1 mM CaCl 2 ). For the kinase reaction, kinase buffer was added to 12.5 l and the reaction was initiated by the addition of 1 l of 62.5 mM [␥-32 P]ATP (5 Ci). One-fifth of the precipitated sample was used for a reaction. The reaction was terminated after incubation at 25°C for 30 min. In the preliminary experiment, the phosphorylation level of PTP36 increased linearly up to 30 min. For the reprecipitation of 32 P-labeled PTP36, the phosphorylated samples were incubated in 50 l of 0.5% SDS, 50 mM Tris-HCl, pH 8.0, at 95°C for 3 min, diluted with 350 l of RIPA buffer and used for the immunoprecipitation with 21-4 mAb.
To test the possibility that MLCK could phosphorylate PTP36 directly, PTP36 was immunoprecipitated from the cell lysate prepared with RIPA buffer and used as a substrate. The immunoprecipitated PTP36 was incubated in 25 l of the kinase buffer containing 2.8 nM MLCK, 10 nM calmodulin, 100 M [␥-32 P]ATP at 25°C for 30 min. A weak phosphorylation of PTP36 was detectable without MLCK probably because a small amount of kinase(s) was coprecipitated even in the RIPA buffer. However, there was no increase in the phosphorylation level of PTP36 whereas 22% (145060 cpm) of the total count was incorporated into the MLC peptide (KKRPQRTSNVFS, Peninsula Laboratories, used at 24 mM) in the same assay condition.
In Gel Kinase Assay-Protein kinases were detected in SDS-PAGE gels by a modification of the method of Ferrell and Martin (37). The immunoprecipitates containing kinases were denatured and separated by SDS-PAGE. Following electrophoresis, SDS was removed by soaking the gels in 50 mM Tris-HCl, pH 7.5, 0.1% (v/v) Nonidet P-40, 3 mM dithiothreitol for 1 h at room temperature. For the denaturation, gels were soaked in 50 mM Tris-HCl, pH 8.3, 7 M guanidine-HCl, 3 mM dithiothreitol, 2 mM EDTA for 1 h at room temperature. The proteins were then allowed to renature by soaking the gel in 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 2 mM dithiothreitol, 2 mM EDTA, 1% (w/v) BSA, 0.1% Nonidet P-40 overnight at 4°C. In situ phosphorylation was performed by soaking the gel in 50 ml of 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 2 mM MnCl 2 , 400 Ci of [␥-32 P]ATP (3000 Ci/mmol) for 3 h at room temperature. Unincorporated ATP was removed by washing overnight at room temperature in 50 mM Tris-HCl, pH 7.5, containing 20 g of Dowex 2 ϫ 8 -50 anion exchange resin, changing the buffer three to four times.
Subcellular Fractionation-3T3/P36-8 cells were fractionated as described (29) with some modifications. Cells were suspended in the hypotonic buffer A containing 10 mM Tris-HCl, pH 7.4, 0.2 mM MgCl 2 , 5 mM KCl, 1 mM Na 3 VO 4 , 5 mM sodium pyrophosphate, 1 mM EDTA, and 1 mM PMSF, and homogenized using Dounce tissue homogenizer until over 90% cells were disrupted, then sucrose was added to 250 mM. The homogenate was centrifuged at 1,000 ϫ g for 5 min at 4°C yielding the crude nuclear pellet (P1) and the supernatant 1 (S1). S1 was centrifuged at 12,500 ϫ g for 15 min at 4°C yielding the crude mitochondrial pellet (P2) and the supernatant 2 (S2). S2 was centrifuged at 100,000 ϫ g for 45 min at 4°C yielding the crude microsomal pellet (P3) and the cytosolic fraction (S3). P2 and P3 were lysed in buffer A containing 250 mM sucrose and 1% Triton X-100, and separated by centrifugation at 100,000 ϫ g for 45 min at 4°C yielding the pellet (P) and the supernatant (S).

PTP36 Is a 130-kDa Protein Phosphorylated on Serine
PTP36 was immunoprecipitated from cell lysate, dephosphorylated with the alkaline phosphatase in vitro, and analyzed by immunoblotting. The dephosphorylated PTP36 migrated faster ( Fig. 1, C, lane 4 and D, lane 2) than the untreated (phosphorylated) PTP36 ( Fig. 1, C, lanes 2 and 3 and D, lane 1). The change in migration was not observed in the presence of phosphatase inhibitors (Fig. 1D, lanes 3 and 4).
Regulation of the Phosphorylation Level of PTP36 Through Cell-Substrate Adhesion-When the cell-substrate adhesion of 3T3/P36-8 cells was disrupted by scraping, the less phosphorylated and faster migrating form of PTP36 was detectable within 5 min ( Fig. 2A). This dephosphorylation was completely blocked by preincubating the cells with okadaic acid, an inhibitor of serine/threonine phosphatases (Fig. 2E, lane 4). Similar results were obtained using wild type 3T3 cells (Fig. 2F). Cell detachment by trypsinization also induced dephosphorylation (Fig. 2B, lane 2). The dephosphorylation was reversible since PTP36 was phosphorylated again along with the reattachment of the cells to the fibronectin-coated wells (Fig. 2C, lanes 3 and  5). The rephosphorylation was induced by the cell spreading on the poly-L-lysine-coated wells but was much slower (Fig. 2C,  lanes 4 and 6). Neither cell spreading nor rephosphorylation was induced when cells were incubated in the BSA-coated wells (Fig. 2D, lane 3). The phosphorylation was inhibited by cytochalasin D, an inhibitor of actin polymerization (Fig. 2D, lane  5), but not by colchicine, a microtubule-disrupting agent (Fig.  2D, lane 6).
Association of PTP36 with Serine/Threonine Kinases, Which Phosphorylate PTP36 -To study the kinases involved in phos-phorylation of PTP36, we first examined whether PTP36 makes complexes with serine/threonine kinases. 3T3/P36-8 cells were lysed in a buffer containing 1% Triton X-100. Then, PTP36 was immunoprecipitated with an anti-PTP36 mAb . The immune complex kinase assay was performed and as shown in Fig. 3A, PTP36 was phosphorylated in vitro by the coprecipitated kinase(s). To confirm that the major 32 P-labeled 130-kDa band contained PTP36, PTP36 in the sample was reprecipitated after removing associated molecules by denaturation with SDS (Fig. 3A, lane 4).
To visualize the coprecipitated kinases, in gel kinase assay was performed. The immunoprecipitates were separated by SDS-PAGE, renatured in gel, and then kinases were detected by 32 P incorporation. Two kinase bands, 160 and 88 kDa in size, were observed (Fig. 3B, lane 2) and the phosphoamino acid analysis showed that serine residues were phosphorylated in these bands (data not shown).
Phosphorylation sites of PTP36 in vivo and in vitro were compared by the phosphotryptic peptide mapping. Although several spots in vitro were phosphorylated weakly (Fig. 4B,  arrowheads), almost all the spots in vivo (Fig. 4A) were detectable in vitro after a prolonged exposure (Fig. 4D).
The Effects of Various Inhibitors on the Phosphorylation of PTP36 in Vitro and in Vivo-To characterize the kinase(s) further, various inhibitors were added to immune complex kinase reactions. Staurosporine, a broad spectrum inhibitor of protein kinases, was found to inhibit phosphorylation of PTP36 in vitro (Fig. 5, A, lane 2, and C). Other inhibitors, including ML-9 (MLCK inhibitor), calphostin C (protein kinase C inhibitor), KN-62 (calcium calmodulin-dependent kinase II inhibitor), H-89 (protein kinase A inhibitor), genistein (tyrosine kinase inhibitor), and W-7 (calmodulin inhibitor), revealed little effect (Fig. 5A).
In vivo, staurosporine also suppressed phosphorylation of PTP36 when it was added to 3T3/P36-8 cells (Fig. 5, B, lane 2,  and C). Unexpectedly, ML-9, which did not inhibit the phosphorylation in vitro, suppressed phosphorylation in vivo (Fig.  5B, lanes 8). ML-7 (50 M), another inhibitor of MLCK, also inhibited the phosphorylation (data not shown). Furthermore, W-7, a calmodulin inhibitor, revealed the inhibitory effect (Fig.  5B, lane 6). This makes sense since MLCK is a Ca 2ϩ -calmodulindependent kinase. However, no direct phosphorylation of PTP36 in vitro was detectable when purified MLCK and calmodulin were added to immunoprecipitated PTP36 as described under "Experimental Procedures." Thus, MLCK might regulate the phosphorylation level of PTP36 indirectly. It should be noted that a relatively high amount of ML-9 (100 M) is required to completely inhibit the phosphorylation of PTP36. Further study is necessary to determine if the target of ML-9 is MLCK.
Redistribution of PTP36 Induced by Cell Detachment-To study the effects of the phosphorylation of PTP36, we compared the subcellular distribution of the phosphorylated and dephosphorylated PTP36 in attached and detached 3T3/P36-8 cells. We first performed the subcellular fractionation experiments. 3T3/P36-8 cells grown on the tissue culture dish were either left untreated or detached by scraping and incubated in suspension (Fig. 6A). The membrane fractionation was done as described under "Experimental Procedures." Cell homogenate was centrifuged at 1,000 ϫ g and separated in the pellet (P1; crude nuclear fraction) and the post-nuclear supernatant (S1). S1 was further separated by the centrifugation at 12,500 ϫ g followed by centrifugation at 100,000 ϫ g yielding the pellets (P2 and P3) and the final supernatant (S3). The phosphorylated PTP36 in the attached cells was found almost exclusively in the S3 cytosolic fraction (Fig. 6A, lane 5) while the dephos- phorylated PTP36 in the detached cells was relatively enriched in the P3 membrane fraction (Fig. 6A, lane 9).
The dephosphorylated PTP36 in the P2 and P3 fractions was resistant to solubilization by Triton X-100 (Fig. 6A, lanes 11  and 13). This suggested the association of the dephosphorylated PTP36 with cytoskeletons. To confirm the association, the cytoskeletal fractionation was performed. Cells were lysed in the cytoskeleton stabilizing buffer (CSK buffer) containing Tri-ton X-100 and separated at 4°C by centrifugation. In this separation condition, the actin-based microfilaments and the intermediate filaments were stabilized and recovered in the 14,000 ϫ g pellet, whereas the microtubules were destabilized and recovered in the supernatant (Fig. 6B). Actin monomers were recovered in the supernatant fraction too. As expected, the phosphorylated PTP36 was found in the cytosolic supernatant fraction (Fig. 6B, lane 3) whereas the dephosphorylated FIG. 4. Phosphopeptide analysis of PTP36 phosphorylated in vivo and in vitro. 32 P-Labeled PTP36 phosphorylated in vivo or in vitro was prepared as described in the legends of Figs. 1 and 3, separated by SDS-PAGE, and electrotransferred to a polyvinylidene difluoride membrane. After digestion by TPCK-trypsin, phosphopeptides were separated in two dimensions on a TLC plate. As a first dimension, electrophoresis at pH 1.9 was performed horizontally, with the anode on the left. Then, chromatography was performed vertically in 37.5% 1-butanol, 25% pyridine, and 1. PTP36 was enriched in the pellet fraction containing the polymerized actin and vimentin (Fig. 6B, lane 5). Collectively, we conclude that cell detachment induces dephosphorylation and translocation of PTP36 from cytosol to the membrane-associated cytoskeletons. The translocation of the endogenous PTP36 was confirmed in the wild type (untransfected) 3T3 cells (Fig.  6F).
The translocation of PTP36 was further confirmed by the immunofluorescence microscopy experiments. Unfortunately, 21-4 mAb failed to detect PTP36, probably because the epitope was masked in the cells. Thus, the localization of the HAepitope-tagged PTP36 (PTP36-HA) and the PTP36 fused to green fluorescent protein (PTP36-GFP) (22) was studied. As expected, PTP36-HA and PTP36-GFP distributed homogeneously in the cytoplasm in the attached cells (Fig. 7, b and d), while they were enriched in the region near the cytoplasmic membrane in the detached cells (Fig. 7, a and c).
After centrifugation at 14,000 ϫ g, a substantial amount of dephosphorylated PTP36 was found in the supernatant fraction (Fig. 6B, lane 6). However, most of PTP36 in the supernatant fraction was insoluble in Triton X-100 and was precipitated by centrifugation at 100,000 ϫ g (data not shown). This might suggest that some of the dephosphorylated PTP36 translocated to the microdomains of cellular membranes that are insoluble in Triton X-100 (32). However, PTP36 was not copurified with caveolin, a marker protein for the microdomain (Fig.  6C).
Effects of Various Inhibitors on the Redistribution of PTP36 -The results described above suggest that cell detachment induces the translocation of PTP36 through its dephosphorylation. In good accordance with this notion, the reduction of the phosphorylation level of PTP36 by staurosporine and ML-9 resulted in the translocation of PTP36 without cell detachment (Fig. 6D, lanes 7 and 9). Moreover, preincubation of  6 -10). The total homogenate (T) was centrifuged at 1,000 ϫ g and was separated in the pellet (P1) and the supernatant (S1). S1 was further separated by centrifugation at 12,500 ϫ g followed by centrifugation at 100,000 ϫ g yielding P2, P3, and S3 (upper panel). P2 and P3 were incubated in the buffer containing 1% Triton X-100 and centrifuged at 100,000 ϫ g yielding pellet (P) and supernatant (S) (lanes 11-14). B, cytoskeletal fractionation. 3T3/P36-8 cells on dish (lanes 1-3) or cells scraped and incubated in suspension for 15 min (lanes 4 -6) were lysed in the cytoskeleton stabilizing buffer (CSK-buffer) containing phosphatase inhibitors and 0.15% Triton X-100 at 4°C. The total lysate (T) was centrifuged at 14,000 ϫ g yielding the pellet (P) and the supernatant (S). PTP36, actin, vimentin, and tubulin were visualized by the immunoblotting experiments as described under "Experimental Procedures." The lower panel shows the silver staining of the each fraction. C, separation of Triton X-100-insoluble membrane microdomains. Attached (lanes 1-5) and detached (lanes 6 -10) cells were lysed in the presence of 1% Triton X-100 and phosphatase inhibitors, and separated by sucrose gradient ultracentrifugation. D, induction of translocation by staurosporine and ML-9. 3T3/P36-8 cells preincubated for 30 min without (lanes 2, 5, and 6) or with 50 nM staurosporine (lanes 3, 7, and 8) or with 100 M ML-9 (lanes 4, 9, and 10) were lysed in the buffer containing phosphatase inhibitors and 1% Triton X-100, and separated (lanes 5-10) by centrifugation at 100,000 ϫ g yielding the pellet (P) and supernatant (S). The mixture of the phosphorylated and dephosphorylated PTP36 was shown as a control (lane 1). E, inhibition of translocation by okadaic acid. The attached 3T3/P36-8 cells preincubated for 60 min with (lanes 3, 4, 7, and 8) or without (lanes 1, 2, 5, and 6) 1 M okadaic acid were either untreated (lanes [1][2][3][4]  the cells with okadaic acid completely blocked not only the dephosphorylation but also the translocation of PTP36 induced by cell detachment (Fig. 6E, lane 8). Cell detachment and okadaic acid revealed little effect on the distribution of Src kinase (Fig. 6E, lower panel).
The translocation was observed following the dephosphorylation of PTP36 in vitro. When the cell lysates were prepared from the attached cells in the absence of phosphatase inhibitors, PTP36 was dephosphorylated in vitro. The dephosphorylated PTP36 was recovered in the Triton X-100 insoluble fraction (Fig. 6G, lane 2). Taken collectively, it is most likely that the phosphorylation level of PTP36 regulates its subcellular localization.

DISCUSSION
The phosphorylation of band 4.1 superfamily PTPs, PTPH1 and PTPMEG, by serine/threonine kinase(s) has been reported (24,33). However, no physiological regulation of the phosphorylation level of the PTPs was known. Here we report the phosphorylation of PTP36 and its complex formation with serine/threonine kinases. What is more important, cell adhesion has a profound regulatory effect on the phosphorylation level and subcellular localization of PTP36.
The kinase that phosphorylates PTP36 in vivo remains to be identified. We found kinase activities coprecipitated with PTP36. Several lines of evidence suggest that the coprecipitated kinase(s) phosphorylate PTP36 in vivo too. For example, coprecipitated kinase(s) were inhibited by staurosporine, which inhibited the in vivo phosphorylation in similar dose-response relationships. Moreover, comparing the phosphopeptide maps, almost all the spots phosphorylated in vivo were also detectable in vitro. It should be noted, however, several spots were phosphorylated weakly in vitro. The PTP36 used as a substrate in the immune complex kinase assay was a mixture of phosphorylated and dephosphorylated PTP36. Thus, it is possible that some sites have been already phosphorylated in vivo and there is less room left for the incorporation of 32 P in vitro. Another explanation is that there are several kinases that phosphorylate PTP36 on different sites and some of the kinases are coprecipitated less well.
The kinases which phosphorylate PTPH1 and PTPMEG in vivo have not been identified either. It was reported that protein kinase A, protein kinase C, mitogen-activated protein kinase, and C-TAK1 phosphorylate PTPH1 in vitro (33,34). However, these kinases may not be responsible for the in vivo phosphorylation of PTP36 in view of their size and sensitivity to various kinase inhibitors.
The mechanism that regulates the phosphorylation level of PTP36 is largely unknown. The phosphorylation of PTP36 is induced by the cell attachment to fibronectin more efficiently than the cell spreading on the poly-L-lysine-coated wells. Thus, the cell adhesion through integrins might not be essential but an important regulation mechanism. We found that ML-9, an inhibitor of MLCK, reduced the phosphorylation level of PTP36 in vivo. It was reported that the inhibitors of MLCK promoted disassembly of actin stress fibers and focal adhesions (35,36). Moreover, cytochalasin D, an inhibitor of actin polymerization, reduced the rephosphorylation of PTP36. Collectively, we speculate that actin-cytoskeleton might be involved in the regulation.
Based on the subcellular fractionation and the immunofluorescence microscopy experiments, we conclude that the dephosphorylated PTP36 is enriched in the membrane-associated cytoskeletal fraction. This distribution pattern is similar to that of PTPMEG, which is localized primarily to the membrane and the cytoskeletal fractions (24). Not all the localizations of PTP36 are clarified completely. For example, while the dephosphorylated PTP36 is not copurified with Triton X-100-insoluble membrane microdomains, it is still possible that PTP36 may dissociate from the microdomain during the purification process. Further study is necessary to answer such a question.
The role of the phosphorylation in the regulation of PTPH1 and PTPMEG is largely unknown. Little change in the enzymatic activity was observed after the phosphorylation of PTPH1 (33). Intriguingly, it was reported that PTPH1 made a complex with 14-3-3␤ in a phosphorylation-dependent manner (34). This might suggest that the phosphorylation of PTPH1 regulates its association with 14-3-3␤. However, no physiological stimulation that affects the phosphorylation level of PTPH1 has been reported. We demonstrated in this study that cell-substrate adhesion regulates the phosphorylation and the localization of PTP36. The localization seems to be regulated through the phosphorylation level of PTP36 since: 1) okadaic acid, an inhibitor of serine/threonine phosphatase, completely blocks the translocation induced by cell detachment and 2) the reduction of the phosphorylation level by staurosporine and ML-9 is sufficient to induce translocation. Furthermore, dephosphorylation of PTP36 in vitro is accompanied by the translocation. Taken collectively, it is most likely that the phosphorylation level of PTP36 regulates its subcellular localization. For more definitive demonstration, identification and specific inactivation of the kinase and/or identification and mutagenesis of the phosphorylation sites on PTP36 might be necessary.
The physiological functions of PTP36 remain to be determined. Recently, we established a HeLa cell line in which PTP36 overexpression was inducible (22). There was little endogenous PTP36 detectable in the parental HeLa cells. When the overexpression of PTP36 was induced, we observed: 1) morphological changes, 2) a decrease in the actin stress fibers, 3) a decrease in the focal adhesions, 4) a reduction in the cell adhesion to collagen, and 5) a decrease in cell growth (22). Interestingly, unlike in the HeLa transfectants, we failed to detect any convincing effect of the overexpression in the 3T3/ P36-8 cells. 2 One fascinating explanation for the apparent discrepancy is that PTP36 overexpressed in 3T3 cells is attenuated by sequestration from its targets. In 3T3 cells adhered to the substrates, PTP36 is phosphorylated and localized to the cytosol. In contrast, PTP36 in the HeLa transfectant is largely in the dephosphorylated form 2 and, as expected, PTP36 is localized near the plasma membrane (22). Collectively, our current working hypothesis is as follows. Detachment from 2 M. Ogata and T. Hamaoka, unpublished observation. substrates in part of the cell induces the dephosphorylation and translocation of PTP36 from cytosol to cytoskeleton, where it plays roles in the regulation of cytoskeleton, cell adhesion, and cell growth.