The CLK Family Kinases, CLK1 and CLK2, Phosphorylate and Activate the Tyrosine Phosphatase, PTP-1B*

The protein-tyrosine phosphatase PTP-1B is an important regulator of intracellular protein tyrosine phosphorylation, and is itself regulated by phosphorylation. We report that PTP-1B and its yeast analog, YPTP, are phosphorylated and activated by members of the CLK family of dual specificity kinases. CLK1 and CLK2 phosphorylation of PTP-1B in vitro activated the phosphatase activity approximately 3–5-fold using eitherp-nitrophenol phosphate, or tyrosine-phosphorylated myelin basic protein as substrates. Co-expression of CLK1 or CLK2 with PTP-1B in HEK 293 cells led to a 2-fold stimulation of phosphatase activityin vivo. Phosphorylation of PTP-1B at Ser50 by CLK1 or CLK2 is responsible for its enzymatic activation. These findings suggest that phosphorylation at Ser50 by serine threonine kinases may regulate the activation of PTP-1B in vivo. We also show that CLK1 and CLK2 phosphorylate and activate the S. cerevisiae PTP-1B family member, YPTP1. CLK1 phosphorylation of YPTP1 led to a 3-fold stimulation of phosphatase activity in vitro. We demonstrate that CLK phosphorylation of Ser83 on YPTP1 is responsible for the activation of this enzyme. These findings demonstrate that the CLK kinases activate PTP-1B family members, and this phosphatase may be an important cellular target for CLK action.

Modification of proteins by phosphorylation is a rapid and reversible mechanism to control their function, and is central to many signal transduction pathways. While serine/threonine phosphorylation of proteins is a common post-translational modification, only a small proportion of proteins are phosphorylated on tyrosine residues. Protein-tyrosine kinases, which include many growth factor receptors, are important regulators of cellular responses (1)(2)(3)(4). Tyrosine phosphorylation may directly regulate enzyme activity, or it may direct the formation of large signaling complexes, which are essential for the transduction of signals throughout the cell. The levels of cellular protein tyrosine phosphorylation are governed by the combined actions of the tyrosine kinases and phosphatases. While the regulation of cellular tyrosine kinases has been extensively studied, comparatively little is known about the regulation of tyrosine phosphatases. Interestingly, like tyrosine kinases, the activity of tyrosine phosphatases is subject to regulation by both serine/threonine and tyrosine phosphorylation (5)(6)(7)(8)(9)(10)(11)(12)(13).
PTP-1B was the first tyrosine phosphatase to be isolated (14). While the regulation of PTP-1B activity in cells is poorly understood, it is known that phosphorylation of PTP-1B varies with the cell cycle and following treatment of cells with various stimuli, such as EGF, 1 okadaic acid, and phorbol esters (5, 14 -16). In vivo the phosphorylation of PTP-1B occurs on serine and tyrosine residues. In response to EGF stimulation of A431 cells, PTP-1B is phosphorylated at Tyr 66 by the EGF receptor, which leads to a 3-fold activation of PTP-1B (15). Moreover, evidence that PTP-1B phosphatase activity is regulated by serine phosphorylation is mounting. Treatment of cells with cAMP analogs or okadaic acid resulted in the serine phosphorylation of PTP-1B and a 4-fold stimulation of PTP-1B phosphatase activity (14). Previous studies have identified several serine phosphorylation sites within the C-terminal regulatory domain of PTP-1B (5); however, phosphorylation at these sites does not lead to alterations in phosphatase activity. Therefore, it is likely that heretofore unrecognized phosphorylation sites within the catalytic domain of PTP-1B exist, and that these sites are important for the in vivo regulation of PTP-1B phosphatase activity.
The CLK family kinases are an evolutionarily conserved group of dual specificity kinases, capable of phosphorylating protein substrates on serine, threonine, and tyrosine residues. The prototypic CLK family kinase member, CLK1, was initially identified through its ability to autophosphorylate on tyrosine residues (17,18). The family includes members from diverse species, including yeast, Drosophila, Arabidopsis, tobacco, mouse, rat, and human.
The biological functions of this family of proteins have remained elusive, but may play an important and evolutionarily conserved role in signal transduction within the cell. A critical role for the CLK family in development has been suggested by work on the Drosophila CLK homologue, DOA. Flies expressing low levels of the mutant DOA protein show marked neurologic abnormalities, and homozygosity for the DOA null allele is embryonically lethal (19).
Recent work on the murine CLK1 protein has begun to shed light on other physiological roles of the CLK family of kinases. Regulation of mRNA splicing is now recognized as a dynamic process, and one in which the CLK family of kinases may have an important function. CLK1 has been reported to bind to and phosphorylate serine/arginine-rich mRNA splicing factors on physiologically relevant sites in vitro (20,21). Moreover, Colwill et al. (20,22) demonstrated that overexpression of CLK1 in COS cells leads to the subcellular redistribution of serine/ arginine-rich proteins, and to alterations in mRNA splicing in vivo. Collectively, these data strongly suggest a role for the CLK family kinases in the regulation of mRNA splicing in vivo.
The CLK family kinases may also participate in intracellular signal transduction cascades. Myers et al. (23) showed that overexpression of CLK1 in the pheochromocytoma PC-12 cell line led to differentiation of these cells. Moreover, specific signal transduction intermediates were activated in these cells, including ERK1/2 and pp90 Rsk . Furthermore, immunocytochemical staining of 3T3 cells expressing human CLK3 demonstrated that the majority of immunoreactivity was present within the cytoplasm, and was less abundant in the nucleus. 2 Similarly, staining of endogenous CLK1 in PC12 cells found it to be mostly cytoplasmic as well. In agreement with a putative signaling role for the CLK kinases is the finding that ethylene stimulation of tobacco leaves stimulates the activity of the tobacco CLK family member, PK12 (24). These finding strongly suggest the existence of cytoplasmic targets for the CLK family kinases and their participation in intracellular signaling pathways.
We report here that PTP-1B and a yeast analog, YPTP1, are in vitro substrates for both CLK1 and CLK2. Moreover, phosphorylation of these two phosphatases by CLK1/CLK2 leads to their enzymatic activation in vitro. We have mapped the activation site within the catalytic domain of PTP-1B and show that it is important both for basal activity as well as enzymatic activation of PTP-1B. Furthermore, we show that co-expression of CLK1/CLK2 with PTP-1B leads to activation of PTP-1B in vivo.

MATERIALS AND METHODS
Phosphatase Mutants-Site-directed mutagenesis of hPTP-1B or YPTP1 was performed by polymerase chain reaction using 4-primer mutagenesis (25). The XL-1 Blue Escherichia coli strain was used as the host strain during mutagenesis. Two of the primers were anchored in the pGEX-KG sequence flanking the multi-cloning site: right primer, 5Ј-TCCGGTTCCCAACGATCAAGGCGAG; left primer, 5Ј-CCCAATGT-GCCTGGATGCGTTCCC.
Following polymerase chain reaction, the full-length proteins were digested with BamHI and SalI and sub-cloned into pGEXKG or in pEBG.
Eukaryotic Expression-HEK 293 cells were grown at 37°C in 5% CO 2 in Dulbecco's modified essential medium supplemented with 10% fetal bovine serum. Expression constructs were introduced into HEK293 cells (3 ϫ 10 6 cells) by electroporation using an Invitrogen Electroporator II apparatus. The cells were harvested 48 -60 h later in cold Lysis Buffer, and sonicated using a Kontes model ASI sonicator. After centrifugation, the GST fusion proteins were recovered from the supernatants either using glutathione-Sepharose beads or by immunoprecipitation using anti-PTP1B antibodies.
Protein Kinase Assays-Bacterially expressed recombinant CLK protein was incubated with substrates in Kinase Reaction Buffer (20 mM Tris, pH 7.4, 1 mM EGTA, 1 M ATP, 10 mM MgCl 2 , 2 mM MnCl 2 ). For assays requiring 32 P incorporation, the Kinase Reaction Buffer was supplemented with 10 Ci of [␥-32 P]ATP. Reactions were carried out at room temperature for 20 min and were stopped by the addition of 3ϫ Laemmli Sample Buffer, or diluted in Phosphatase Reaction Buffer and phosphatase activity monitored.
Protein-tyrosine Phosphatase Assays-The activity of YPTP1 and PTP-1B were assayed by hydrolysis of p-nitrophenol phosphate (PNPP). The phosphatases were incubated in Phosphatase Reaction Buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, 2 mM dithiothreitol, 1 mM PNPP) for 2-20 min at 37°C. The reaction was stopped with 0.2 N NaOH, and the absorbance at 410 nm was measured. The reactions were run in triplicate.
The activity of PTP-1B was also assayed using radiolabeled tyrosinephosphorylated MBP as a substrate. The MBP was radiolabeled by incubation with the tyrosine kinase GST-FER in Kinase Reaction Buffer. The phosphatases were incubated in Assay Buffer (25 mM Hepes, 1 mM dithiothreitol, and 1 mM EDTA, pH 7.5) plus the indicated concentrations of the tyrosine-phosphorylated MBP for 2-5 min. The assays were stopped in Charcoal Stop Mix (30% charcoal, 6% diatomaceous earth, 0.9 M HCl, 90 mM NaPPO 4 , 2 mM NaPO 4 ). Liberated phosphate in the supernatants was determined by Cerenkov counting using a Beckman LS 3801 scintillation counter.
Phosphoamino Acid Analysis-Radiolabeled phosphatases were separated by SDS-PAGE, and transferred to polyvinylidene difluoride membrane, and visualized by autoradiography. The protein bands were excised from the membrane and subjected to acid hydrolysis as described by Kamps (26).
Tricine-SDS-Polyacrylamide Electrophoresis-32 P-Labeled proteins were proteolytically digested as described by Luo et al. (27). The 32 Plabeled phosphopeptides were resolved in one-dimension using Tricine-SDS-electrophoresis essentially as described by Schagger and von Jagow (28). The protocol was modified by use of 24% acrylamide separation gels. The gels were run at 100 mV for 18 h at 4°C. The gels were immediately dried and the phosphopeptides visualized by autoradiography.
Two-dimensional Phosphopeptide Mapping-The 32 P-labeled phosphopeptides were resolved in the first dimension by electrophoresis at pH 1.9 on TLC plates as described (29). The plates were dried, and then subjected to ascending chromatography in the second dimension using a buffer composed of isobutanol, pyridine, acetic acid, and water (75: 15:50:60). The plates were dried, and the phosphopeptides were visualized by autoradiography.

CLK1 and CLK2 Phosphorylate Human PTP-1B in Vitro-In
experiments initially designed to test the phosphorylation dependence of CLK1 activity in vitro, we observed that when PTP-1B was incubated with CLK1, PTP-1B became highly phosphorylated. We subsequently found that recombinant, constitutively active CLK1 or CLK2 phosphorylated PTP-1B (Fig.  1A) in vitro. Phosphoamino acid analysis revealed that CLK1 and CLK2 phosphorylated PTP-1B exclusively on serine residues (Fig. 1B). In order to determine whether CLK phospho-rylation altered PTP-1B activity, the phosphatase was preincubated in the presence or absence of either CLK1 or CLK2 for 20 min in vitro. Subsequent in vitro phosphatase assays demonstrated an approximate 5-fold activation of phosphatase activity of CLK1 or CLK2-treated PTP-1B ( Fig. 2A). Incubation of PTP-1B with increasing amounts of CLK protein led to a corresponding increase in activation of PTP-1B (Fig. 2B). We conclude that phosphorylation of PTP-1B by CLK1 and CLK2 activates the phosphatase.
CLK1 and CLK2 Phosphorylate Serine 50 and Serines 242/ 243 on Human PTP-1B-One-dimensional phosphopeptide mapping of PTP-1B was utilized to investigate which serine residues CLK1 and CLK2 phosphorylated in vitro. The phosphopeptide maps demonstrate that CLK1 and CLK2 phosphorylated PTP-1B at similar sites, as evidenced by the detection of identical phosphopeptides (Fig. 3). Examination of the primary sequence of PTP-1B revealed multiple sites conforming to the CLK family consensus phosphorylation sequence (R/K-X-R/K-X-R/K-X-S-X-X-R). 3 These data and the size of the phosphopeptides allowed identification of Ser 50 as a likely site of CLK phosphorylation. Substitution of an alanine at the Ser 50 site (S50A) by site-directed mutagenesis diminished CLK1 and CLK2 phosphorylation of PTP-1B by approximately 90%, indicating that this was the principal site of phosphorylation (Fig.  4A). To further establish that this residue was phosphorylated, we generated a threonine substitution at Ser 50 (S50T). Following incubation of the S50T PTP-1B mutant with CLK1 or CLK2, phosphoamino acid analysis revealed the presence of phosphothreonine residues on the S50T mutant (Fig. 4B). Sub-  3. CLK1 and CLK2 phosphorylate PTP-1B on identical peptides. Recombinant PTP-1B was phosphorylated in vitro by either constitutively active CLK1 or CLK2. The 32 P-labeled PTP-1B protein was separated by SDS-PAGE and transferred to nitrocellulose. The PTP-1B band was excised and subjected to digestion with trypsin, chymotrypsin, or a combination of trypsin and chymotrypsin. Tricine-SDS electrophoresis was used to separate the resulting peptides and the phosphopeptides visualized by autoradiography. stitution of threonine for serine slightly decreases the affinity of the CLKs for this site compared with the native enzyme, as evidenced by the presence of equal amounts of phosphoserine and phosphothreonine on the S50T mutants. Moreover, proteolytic maps of the S50T mutant show increased phosphorylation on other phosphopeptides (see below), as well as on the phosphopeptide containing Thr 50 (data not shown). We conclude that both CLK1 and CLK2 phosphorylate PTP-1B principally at Ser 50 .
The S50A mutant was phosphorylated by CLK1 and CLK2, albeit at lower levels than the wild-type enzyme, suggesting the existence of additional CLK phosphorylation sites on PTP-1B. Phosphopeptide maps of S50A mutants phosphorylated by CLK1 and CLK2 were consistent with Ser 242 or Ser 243 as the second site of CLK phosphorylation on PTP-1B. Substitution of alanine at positions 242/243 only modestly diminishes CLK1 and CLK2 phosphorylation of PTP-1B (data not shown). Moreover, CLK2 poorly phosphorylates, and CLK1 does not phosphorylate the triple mutant, S50A/S242A/S243A PTP-1B (Fig.  4C). Following incubation with CLK1 or CLK2, phosphoamino acid analysis on the S242T/S243T PTP-1B mutant revealed phosphothreonine residues (Fig. 4B and data not shown). These data show that, although CLK1 and CLK2 directly phosphorylate PTP-1B on both Ser 50 and Ser 242 /Ser 243 , the preferred CLK phosphorylation site is Ser 50 , as it is preferentially phosphorylated at an approximate ratio of 9:1 over the Ser 242 / Ser 243 site.
Serine 50 Is Important for Catalytic Activity of PTP-1B-We tested whether the mutation of Ser 50 affected the phosphatase activity of PTP-1B. The basal activities of the wild-type, S50T, and S50A PTP-1B proteins were analyzed using in vitro phosphatase assays. The S50A mutants have significantly diminished basal phosphatase activity toward the PNPP substrate, compared with the wild-type enzyme, while the S50T mutants have wild-type level phosphatase activity (Fig. 5). CLK1 and CLK2 activated the phosphatase activity of S50T and wild-type PTP-1B in vitro (Fig. 6). However, the S50A mutant was resistant to activation by either CLK1 or CLK2 (Fig. 6). These data strongly suggest that Ser 50 is the phosphorylation site on PTP-1B responsible for CLK-induced stimulation of phosphatase activity.
In order to determine whether the CLKs could activate PTP-1B toward a protein substrate, wild-type PTP-1B, the S50A mutant, and activated PTP-1B were assayed using tyrosine-phosphorylated MBP as a substrate. CLK2 phosphorylated PTP-1B exhibited a 3-fold increase in activity relative to the wild-type PTP-1B (Fig. 7). Furthermore, the S50A mutant possessed approximately 20% of the activity of the wild-type enzyme. Phosphorylation of PTP-1B by CLK2 led to a decrease in the K m by 3-fold (Table I). Interestingly, substitution of alanine for Ser 50 led to 5-fold increase in the K m over wild-type; however, the V max for the S50A mutant was roughly half that of the wild-type enzyme. The change in the K m of the PTP-1B for substrate following phosphorylation by the CLKs is consistent with the observed increase in activity of PTP-1B for protein and synthetic substrates. The changes observed in the S50A mutant are concordant with the view that Ser 50 is an important determinant of the substrate binding pocket conformation. Our results demonstrate that phosphorylation of PTP-1B at Ser 50 enhances substrate binding to the enzyme.
CLK1 and CLK2 Can Activate PTP-1B in Vivo-The effect of CLK1 and CLK2 on PTP-1B in vivo was investigated by overexpressing these proteins in HEK293 cells. GST-tagged fulllength CLK1 or CLK2 were co-transfected with GST-tagged PTP-1B into HEK293 cells. Glutathione-Sepharose was used to precipitate the tagged proteins from the transfected cells. In vitro phosphatase assays were performed on the precipitated proteins. Co-expression of CLK1 with PTP-1B activated PTP-1B 2-fold in vivo (Fig. 8). Similar results have been obtained by co-expression of CLK2 with PTP-1B (data not shown). In a series of similar experiments, we co-expressed CLK1 or CLK2 with untagged-PTP-1B in NIH 3T3 cells and observed a 2-3-fold increase in phosphatase activity in immunoprecipitates of PTP-1B (data not shown). The 2-fold activation of PTP-1B in vivo by CLK1 is significantly lower than that observed in vitro. However, this is likely a consequence of the lower enzymatic activity of the full-length CLK1, relative to the constitutively active, truncated CLK1 employed in the in vitro studies. These data demonstrate that PTP-1B is regulated by the CLK family kinases in vivo as well as in vitro.
CLK1 and CLK2 Phosphorylate YPTP1 on Ser 83 -We investigated whether the CLKs could also activate other phosphatases related to PTP-1B. Yeast protein-tyrosine phosphatase, YPTP1, a Saccharomyces cerevisiae PTP-1B analog (30), was phosphorylated and enzymatically activated in vitro by CLK1 and CLK2 (Figs. 9A and 10). We therefore investigated which residues on YPTP1 the CLKs phosphorylated. We used the CLK consensus phosphorylation sequence to search the aligned sequences of YPTP1 and PTP-1B for potential phosphorylation sites. In YPTP1, Ser 83 was found to closely conform to the consensus phosphorylation site of CLK1. This serine residue was mutated to alanine (S83A YPTP1) to test whether this serine was phosphorylated by the CLKs. We also substituted a serine for the invariant catalytic cysteine, Cys 252 , creating a catalytically inactive YPTP1 mutant (C252S YPTP1). Two-dimensional tryptic peptide mapping was performed on CLK1 phosphorylated wild-type, S83A, and C252S YPTP1. The tryptic maps of YPTP1 and C252S YPTP1 produced several major phosphopeptides and a number of minor phosphopeptides (Fig.  9B). Importantly, the tryptic map of S83A YPTP1 showed the  specific loss of a single phosphopeptide (Fig. 9B). A peptide map was produced from a mixture of all three phosphorylated forms of YPTP1 (wild-type, S83A, and C252S YPTP1), indicating that identical sites were phosphorylated in all thee forms of YPTP1 (Fig. 9B). The loss of a single phosphopeptide in the S83A mutant demonstrates that Ser 83 was phosphorylated by CLK1. The identities of the other major phosphorylation sites on YPTP1 are currently unknown. Ser 83 Is Essential for CLK1 Activation of YPTP1-We tested whether substitution of an alanine at Ser 83 would effect the phosphatase activity of YPTP1. In vitro phosphatase assays were performed on YPTP1 proteins that had been incubated in the absence or presence of CLK1. Mutation of Ser 83 to alanine resulted in a nearly 50% reduction in the basal activity of YPTP1 (Fig. 10). Moreover, S83A YPTP1 proteins are resistant to activation by CLK1 in vitro (Fig. 10). We conclude from these data that CLK phosphorylation of YPTP1 at Ser 83 activates the phosphatase activity of YPTP1. DISCUSSION The CLK family kinases were initially identified on the basis of their ability to autophosphorylate on tyrosine residues. Subsequent analysis of the CLK kinases showed them to be members of the growing class of kinases termed dual-specificity kinases, capable of phosphorylating substrates on serine, thre-onine, and tyrosine residues. Work from several laboratories has suggested an important role for the CLK kinases in regulation of mRNA splicing in vivo (20 -22, 31). Furthermore, the Arabidopsis CLK family member AFC1 is capable of regulating transcription in vivo (32). However, we have recently demonstrated that the majority of cellular CLK protein is located in the cytoplasm, 2 suggesting the existence of non-nuclear targets for the CLK kinases. Moreover, a role for the CLK family kinases in signaling cascades has been suggested by several findings. Overexpression of CLK1 in PC12 cells caused the differentiation of these cells into a neuronal phenotype (23). Analysis of these cells showed that CLK1 expression activated elements of the mitogen-activated protein kinase signaling cascade, including ERK1/ERK2 and pp90 Rsk . Although CLK1 caused the activation of the ERKs and pp90 Rsk , the mechanism through which CLK stimulated these activities is unclear, as we have ruled out direct phosphorylation of these molecules by CLK1 (data not shown).
We report here the identification of a direct non-nuclear target of the CLKs, the tyrosine phosphatase, PTP-1B. Serendipitously, in the course of studying CLK1 activation in vitro, it was noted that CLK1 was capable of phosphorylating PTP-1B in vitro. We have subsequently demonstrated that both CLK1 and CLK2 are capable of activating PTP-1B in vitro and in vivo. Similarly, a yeast PTP-1B family member, YPTP1, was also FIG. 8. CLK1 activates PTP-1B in vivo. HEK 293 cells were cotransfected with expression constructs for GST-tagged PTP-1B and GST-tagged full-length CLK1 or the pEBG parent vector (as a control). Forty-eight hours after electroporation, the HEK 293 cells were lysed and the GST-tagged PTP-1B was isolated using GST-Sepharose beads. A, in vitro phosphatase assays were performed on the precipitated phosphatases from control (untransfected), PTP-1B only, and PTP-1B plus CLK1-transfected HEK293 cells. B, following the phosphatase assay, the beads were suspended in Laemmli Sample Buffer, and the bound proteins were separated by SDS-PAGE, and transferred to polyvinylidene difluoride membrane. The membrane was probed with the anti-PTP-1B antibody, FG6, to visualize expressed PTP-1B. The arrowhead indicates the PTP-1B protein band.
FIG. 9. The yeast phosphatase, YPTP1, is phosphorylated on Ser 83 by CLK1. A, wild-type (WT), C252S, and S83A YPTP1, or CLK1 alone (Auto) were phosphorylated by CLK1 in vitro. The phosphorylated proteins were separated by SDS-PAGE, transferred to nitrocellulose, and visualized by autoradiography. B, the YPTP1 bands were excised and digested with trypsin. Two-dimensional phosphopeptide maps of each of the digested proteins, or a mixture of all three proteins (Mixture) are shown. Electrophoresis (pH 1.9) and ascending chromatography were performed in the direction indicated. The arrowhead indicates the peptide containing Ser 83 . phosphorylated and activated by CLK1 and CLK2 in vitro.
PTP-1B has been shown to be one of the major tyrosine phosphatase activities within cells (14). Its activity and phosphorylation varies with the cell cycle and following stimulation with various cellular stimuli. However, the exact role of PTP-1B in the cell is not understood. Identified cellular targets for PTP-1B include the activated EGF receptor, the insulin receptor, and several integrins, suggesting that PTP-1B acts within cells to antagonize receptor driven signaling pathways (33)(34)(35)(36). PTP-1B activity is at least partially controlled by regulation of intracellular compartmentalization, as it is localized to the endoplasmic reticulum by its C-terminal regulatory domain (37). Moreover, prolonged treatment of HeLa cells with insulin or 12-O-tetradecanoylphorbol-13-acetate leads to the alternative splicing of the PTP-1B mRNA, giving rise to a C-terminally truncated protein (38). This C-terminal truncation may be important in altering the subcellular localization of the enzyme. However, it is now apparent that phosphorylation of PTP-1B can directly control levels of activity of this phosphatase. PTP-1B activity is stimulated following EGF stimulation of A431 cells and phorbol ester treatment of HeLa cells (14,15). PTP-1B appears to be regulated by multiple signaling pathways, as evidenced by discrete phosphorylation events following a variety of cellular stimuli. Treatment of cells with cAMP analogs leads to elevation in PTP-1B activity by 4-fold, while EGF stimulation of A431 cells leads to a 3-fold stimulation of phosphatase activity. However, these mechanisms drive this elevation in PTP-1B differentially, as cAMP promotes serine phosphorylation of PTP-1B while EGF stimulates the tyrosine phosphorylation of PTP-1B, suggesting the existence of multiple activating phosphorylation sites within the catalytic domain of PTP-1B. The first identified activating phosphorylation site on PTP-1B was Tyr 66 , which is directly phosphorylated by the EGF receptor (15). However, many cellular stimuli which activate PTP-1B lead only to serine phosphorylation of the enzyme, suggesting that an important regulatory serine phosphorylation site exists (5,14). The previously identified sites of serine phosphorylation have been localized to the Cterminal regulatory domain of PTP-1B. Significantly, phosphorylation at these C-terminal sites has not been demonstrated to be responsible for enzyme activation, suggesting that the activating phosphorylation sites may lie within the catalytic domain (5).
We observed that the phosphorylation of PTP-1B by CLK1 or CLK2 led to an approximately 5-fold stimulation of phosphatase activity. Phosphoamino acid analysis demonstrated that these enzymes phosphorylated PTP-1B on serine residues only. We determined that CLK1 and CLK2 phosphorylate PTP-1B on two sites within the catalytic domain, Ser 50 and Ser 242 / Ser 243 . Mutagenesis of Ser 242 /Ser 243 did not alter phosphatase activity nor did it affect the ability of CLK1 or CLK2 to activate PTP-1B, indicating this is not a regulatory phosphorylation site. However, substitution of an alanine at Ser 50 significantly reduced the basal level of phosphatase activity of PTP-1B and the mutant phosphatase no longer activable by CLK1 or CLK2. Thus, phosphorylation of Ser 50 is responsible for the observed activation of PTP-1B by CLK1 and CLK2.
Analysis of crystallographic data has shown that Ser 50 lies near the substrate-binding pocket of PTP-1B (39). Furthermore, Sarmiento et al. (40) have recently shown that the three residues most responsible for determining the substrate specificity of PTP-1B are Tyr 46 , Arg 47 , and Asp 48 . The authors showed that mutation of these residues altered the K m of these mutant enzymes. The dramatically reduced catalytic activity of the Ala 50 mutant is consistent with the premise that the Ser 50 residue is important for appropriate conformation of the substrate-binding pocket. Phosphorylation of this site may alter the characteristics of the binding pocket, and thereby lead to the activation of the phosphatase by shifting the binding pocket into a more open (active) state. Consistent with this hypothesis is the localization of Tyr 66 in this same area of the protein.
Moreover, alignment of YPTP1 with PTP-1B shows that the activating phosphorylation at Ser 83 is also in this same region of the phosphatase molecule. Thus, there may be a general effect of phosphorylation near the substrate-binding pocket that serves to activate the PTP-1B family of phosphatases. Importantly, co-expression of CLK1 or CLK2 with PTP-1B in HEK293 cells stimulated the phosphatase activity 2-3-fold over basal levels. This study demonstrates that the CLK family kinases regulate cellular phosphatases. FIG. 10. YPTP1 mutated at Ser 83 is resistant to CLK1-mediated activation. The phosphatase activity of wild-type (WT), S83A, and C252A YPTP1 after incubation in the absence or presence of CLK1 was tested. Wild-type YPTP1, S83A YPTP1, or a catalytically inactive form of YPTP1, C252S YPTP1, were incubated in the presence or absence of CLK1 for 20 min and then assayed for phosphatase activity using PNPP as a substrate.