Abstract
In the present study we examined involvement of nuclear protein tyrosine phosphatase TC-PTP in PRL-mediated signaling. TC-PTP could dephosphorylate signal transducer and activator of transcription 5a (STAT5a) and STAT5b, but the apparent dephosphorylation activity of TC-PTP was weaker than that of cytosolic PTP1B 30 min after PRL stimulation in transfected COS-7 cells, whereas both STAT5a and STAT5b were dephosphorylated to the same extent by recombinant TC-PTP and PTP1B in vitro. Tyrosine-phosphorylated STAT5 was coimmunoprecipitated with substrate trapping mutants of TC-PTP, suggesting that STAT5 is a specific substrate of TC-PTP. These observations were further extended in mammary epithelial COMMA-1D cells stably expressing TC-PTP. A time-course study revealed that dephosphorylation of STAT5 by TC-PTP was delayed compared with that by cytosolic PTP1B due to nuclear localization of TC-PTP throughout PRL stimulation in mammary epithelial cells. Endogenous β-casein gene expression and β-casein gene promoter activation in COS-7 cells were largely suppressed by TC-PTP wild type as well as catalytically inactive mutants, suggesting that stable complexes formed between STAT5 and TC-PTP in the nucleus. Taken together, we conclude that TC-PTP is catalytically competent with respect to dephosphorylation and deactivation of PRL-activated STAT5 in the nucleus.
PROTEIN-TYROSINE phosphatases (PTPs) are a large and structurally diverse family of enzymes, characterized by the consensus sequence of (I/V)HCXAGXXR. They are found in eukaryotes, prokaryotes, and viruses and can either antagonize or potentiate protein tyrosine kinase (PTK)-dependent signaling. PTPs have been shown to participate as either positive or negative regulators of signal transduction in a wide range of physiological processes, which include cellular growth and proliferation, migration, differentiation, and survival (1–3). Despite their important roles in such fundamental physiological processes, the mechanism by which PTPs exert their effects is often poorly understood.
PRL exhibits its activity through its cognate receptor and the activation of intracellular signaling molecules such as the Janus kinase 2 (JAK2) and signal transducers and activators of transcription 5 (STAT5). The PRL receptor, belonging to the hemopoietin receptor superfamily (4), does not possess intrinsic tyrosine kinase activity, but is constitutively associated with the cytoplasmic tyrosine kinase JAK2 (5–7). Upon ligand binding, the PRL receptor dimerizes, and JAK2 is activated through autophosphorylation on tyrosine residue (7). JAK2 then phosphorylates not only the PRL receptor, but also the transcription factors STAT5a and STAT5b, which then form homodimers, translocate to the nucleus, and specifically bind to the promoter regions of target genes, thus activating transcription (8, 9).
It has been demonstrated that STAT5 undergoes a rapid and transient activation and deactivation cycle through tyrosine phosphorylation upon cytokine stimulation (10). Because tyrosine phosphorylation is essential for PRL signaling, PTPs are believed to attenuate or block it and play a negative role. Although recent publications have shown that SH2-containing protein tyrosine phosphatase-2 (SHP-2) is involved in β-casein promoter activation in a positive manner (11, 12), dephosphorylation of the activated JAK2 and STAT5 through the PRL receptor and the involvement of the PTPs in a negative regulation remains to be elucidated.
More recently, we demonstrated that cytosolic PTP1B dephosphorylated and deactivated STAT5a and STAT5b in transfected COS-7 cells as well as in mammary epithelial COMMA-1D cells, thereby negatively regulating the PRL-mediated signaling pathway (13). As PTP1B and structurally highly related TC-PTP comprises a subfamily of cytosolic PTPs and TC-PTP was also shown to be expressed in mammary gland and mammary epithelial cells (14), in this study we examined the involvement of TC-PTP in the PRL-mediated signaling pathway. The data demonstrated that TC-PTP was also a potential negative regulator of PRL-mediated signal transduction by specifically dephosphorylating and deactivating STAT5 in nucleus. Our previous and current studies suggest that STAT5 might be cooperatively regulated in cytosol as well as in nucleus in mammary epithelial cells.
RESULTS
TC-PTP Dephosphorylates and Deactivates PRL-Activated STAT5a and STAT5b
In our recent publication we showed that cytosolic PTP1B dephosphorylated PRL-activated STAT5a and STAT5b (13). PTP1B is structurally highly related to TC-PTP, especially in PTP catalytic segment (Fig. 1). The cDNA sequence of only a C-terminally truncated form of TC-PTP with a molecular mass of 45 kDa has been available for mouse species (15, 16), whereas the protein sequence of the 48-kDa form of TC-PTP has been reported previously (17). The mouse 45-kDa TC-PTP was cloned by RT-PCR amplification, hemagglutinin (HA) epitope-tagged at its N-terminus, and examined for dephosphorylation of PRL-activated STAT5a and STAT5b.
Schematic Drawings of TC-PTP and PTP1B The identities of N-terminal region, PTP catalytic domain, and C-terminal region of the PTPs are indicated. The hydrophobic amino acid stretch of PTP1B is indicated (
).
Wild-type or catalytically inactive Cys/Ser or Asp/Ala forms of TC-PTP were cotransfected with PRL receptor and STAT5a or STAT5b into COS-7 cells. Thirty minutes after PRL stimulation, cells were lysed, and STAT5a or STAT5b was immunoprecipitated and subjected to immunoblotting with antiphosphotyrosine antibody. Upon coexpression of TC-PTP wild type, PRL-induced tyrosine phosphorylation of both STAT5a and STAT5b was abolished to approximately 20% and 30%, respectively, compared with that of mock transfectants (Fig. 2, A and B), whereas more than 90% of the proteins were dephosphorylated by PTP1B under the same conditions (13). Dephosphorylation of STAT5a and STAT5b was not observed when the cells were cotransfected with catalytically inactive Cys/Ser or Asp/Ala mutant of TC-PTP, suggesting that phosphatase activity of TC-PTP is essential for the dephosphorylation of STAT5. Comparable expression of TC-PTP wild type and mutants was confirmed by immunoblotting with anti-HA antibody (Fig. 2, lower panels).
Dephosphorylation of STAT5a and STAT5b by TC-PTP in Transfected COS-7 Cells A, COS-7 cells were cotransfected with expression plasmids for PRL receptor (1 μg), STAT5a or STAT5b (1 μg), and empty vector (mock) or each of HA-TC-PTP wild type, catalytically inactive Cys/Ser, and Asp/Ala mutants (2 μg for each). After serum starvation, cells were left untreated (−) or were stimulated (+) with PRL (5 μg/ml) for 30 min and lysed. STAT5a and STAT5b were immunoprecipitated, separated on SDS-PAGE, transferred to a nitrocellulose membrane, and probed with antiphosphotyrosine antibody (upper panels). The membrane was stripped and reprobed with anti-STAT5 antibody (middle panels). The expression of HA-tagged TC-PTP was assessed by immunoblotting (lower panels). B, The tyrosine phosphorylation level of STAT5a and STAT5b in A was densitometorically normalized. The phosphorylation level of STAT5a and STAT5b in mock transfectant stimulated with PRL was set as 100%. The mean and sd of three independent experiments are shown.
To confirm dephosphorylation action of TC-PTP on STAT5a and STAT5b, recombinant glutathione-S-transferase (GST) fusion proteins containing full-length TC-PTP were expressed in Escherichia coli and purified. GST-TC-PTP wild type exhibited catalytic activity against an artificial substrate pNPP, whereas C/S and D/A mutants showed no activity, and the phosphatase activity was comparable to that of PTP1B (data not shown). COS-7 cells that had been cotransfected with PRL receptor and STAT5a or STAT5b were stimulated with PRL, and phosphorylated STAT5a or STAT5b was immunoprecipitated. The indicated amounts of the recombinant GST-TC-PTP fusion proteins were added to the immune complexes and incubated at 37 C for 30 min. As clearly illustrated in Fig. 3A, the tyrosine phosphorylation level of STAT5a was reduced to approximately 50% by 1 μg GST-TC-PTP wild type, and incubation with 10 μg of the fusion protein resulted in complete dephosphorylation of STAT5a (upperpanels). In a similar manner, STAT5b was dephosphorylated by GST-TC-PTP wild type (lower panels). In both cases such dephosphorylation of STAT5 proteins by TC-PTP was indistinguishable from that by PTP1B (13), suggesting that TC-PTP and PTP1B can dephosphorylate STAT5 in a similar manner in vitro. Incubation of the immune complexes with empty GST and GST fused to catalytically inactive mutants of TC-PTP as well as GST-SHP-1 and -HSCF (data not shown) resulted in no reduction in tyrosine phosphorylation level of STAT5a and STAT5b.
STAT5 Is a Specific Substrate for TC-PTP A, COS-7 cells were cotransfected with expression plasmids for PRL receptor (2 μg) and STAT5a or STAT5b (2 μg) and stimulated with PRL (5 μg/ml) for 30 min after serum starvation. STAT5a and STAT5b were immunoprecipitated, washed with lysis buffer, and then subjected to an in vitro dephosphorylation assay as described in Materials andMethods. After termination of the incubation, proteins were separated by SDS-PAGE and analyzed with antiphosphotyrosine antibody (4G10). The same blot was reprobed with anti-STAT5 antibody after stripping. B, COS-7 cells were cotransfected with expression plasmids for PRL receptor (1 μg), STAT5a or STAT5b (1 μg), and empty vector (mock) or each of HA-TC-PTP wild type, catalytically inactive Cys/Ser, and Asp/Ala mutants (2 μg for each). After serum starvation, cells were stimulated with PRL (5 μg/ml) for 60 min and lysed in the absence (left panels) or presence (right panels) of vanadate. TC-PTP was immunoprecipitated with anti-HA antibody, separated on SDS-PAGE, transferred to a nitrocellulose membrane, and probed with antiphosphotyrosine antibody (first panelsfrom the top). The membrane was stripped and reprobed with anti-STAT5 antibody (second panels). Immunoprecipitation was ensured by immunoblotting with anti-HA antibody (third panels). Comparable expression of STAT5a and STAT5b in the total cell lysates (TCL) was confirmed by immunoblotting (fourth panels).
Direct dephosphorylation of STAT5 by TC-PTP was further confirmed by a coimmunoprecipitation study using substrate-trapping mutants of TC-PTP. COS-7 cells that had been cotransfected with PRL receptor, STAT5a or STAT5b, and TC-PTP wild type or catalytically inactive mutants were stimulated with PRL for 30 min after serum starvation and lysed in the presence or absence of vanadate, a potent PTP inhibitor, which binds the PTP catalytic domain and inhibits phosphotyrosine-dependent interaction with substrates. TC-PTP was immunoprecipitated with anti-HA antibody and subjected to immunoblotting using anti-phosphotyrosine antibody. As shown in Fig. 3B, in the absence of vanadate in cell lysates, tyrosine-phosphorylated 97-kDa protein was coimmunoprecipitated with Asp/Ala and, to a lesser extent, with the Cys/Ser mutant of TC-PTP, whereas such a tyrosine-phosphorylated band was scarcely detected in the immune complex of TC-PTP wild type (left panels). The membranes were stripped and reprobed with anti-STAT5 antibody. The tyrosine-phosphorylated 97-kDa band was demonstrated to be STAT5a and STAT5b by immunoblotting. On the other hand, in the absence of vanadate in cell lysates, such a tyrosine-phosphorylated 97-kDa band was not detected (right panels). Interestingly, STAT5 was also slightly present in the immune complexes of TC-PTP wild type regardless of the presence or absence of vanadate as well as in those of TC-PTP mutants in the presence of vanadate, suggesting some contribution of phosphotyrosine-independent interaction of TC-PTP with STAT5, as observed for PTP1B (13).
TC-PTP Is a Potential Negative Regulator in PRL Receptor-Mediated Signaling in Mammary Epithelial Cells
To demonstrate more physiological relevance of TC-PTP in PRL-mediated signaling, TC-PTP was introduced into mammary epithelial COMMA-1D cells by a retroviral infection system. TC-PTP cDNA was ligated into a retroviral vector and introduced into mammary epithelial cells. Cells were selected in G418-supplemented cell culture medium and then directly used for subsequent experiments. The polyclonal cells expressing PTP1B wild type were also included as a control. Nearly the same amounts of HA-tagged TC-PTP and HA-tagged PTP1B were expressed in the cells (Fig. 4A). Cells were serum-starved and lysed at the various time points indicated after PRL stimulation. Endogenous STAT5 was immunoprecipitated, followed by immunoblotting with antiphosphotyrosine antibody. Five minutes after PRL stimulation, tyrosine phosphorylation of STAT5 did not differ in mock and TC-PTP infectants, whereas only faint signal was detected in PTP1B wild type-expressing cells at the same time point (Fig. 4, B and C). The phosphorylation level of STAT5 in TC-PTP wild type-expressing cells was significantly less than that in mock transfectants or cells expressing TC-PTP mutants 10 min after PRL stimulation, and this became more obvious after 30 min. More than 90% of STAT5 was dephosphorylated after 40-min PRL stimulation in TC-PTP wild type-expressing cells, which was nearly the same as in PTP1B-expressing cells at the same time point. On the other hand, in the cells expressing inactive Cys/Ser and Asp/Ala mutants of TC-PTP, the phosphorylation level of STAT5 was not significantly different from that in mock infectants at all time points examined. Phosphorylation of JAK2 upon PRL stimulation was not affected by overexpressing TC-PTP wild type as well as catalytically inactive forms of TC-PTP (Fig. 4D).
TC-PTP Is a Potential Negative Regulator in PRL Receptor-Mediated Signaling Pathway in Mammary Epithelial COMMA-1D Cells A, COMMA-1D cells were retrovirally infected with HA-TC-PTP wild type (WT), Cys/Ser, or Asp/Ala mutant and selected in the cell culture medium supplemented with G418 (1 mg/ml). Polyclonal clones for each as well as PTP1B WT-expressing clone (13 ) were lysed, and aliquots were immunoblotted with anti-HA antibody. B, COMMA-1D cells expressing TC-PTP or PTP1B were stimulated with PRL (5 μg/ml) for the indicated time after serum starvation. STAT5 was immunoprecipitated, followed by immunoblotting with antiphosphotyrosine antibody (left panels). The same membranes were reprobed with anti-STAT5 antibody (right panels). C, The tyrosine phosphorylation level of STAT5 was densitometorically normalized. The phosphorylation level of STAT5 in mock transfectant stimulated with PRL for 40 min was set at 100%. The mean and sd of three independent experiments are shown. D, COMMA-1D cells expressing TC-PTP were stimulated with PRL (5 μg/ml) for 30 min after serum starvation. JAK2 was immunoprecipitated, followed by immunoblotting with antiphosphotyrosine antibody (upper panel). The same membrane was reprobed with anti-JAK2 antibody (lower panel). E, COMMA-1D cells expressing TC-PTP or PTP1B were treated with PRL (5 μg/ml) and hydrocortisone (0.1 μm) for 48 h. RNA was prepared, and RT-PCR amplification forβ -casein and GAPDH was carried out as previously described (13 ) (upper panels). The expression level of β-casein gene was normalized to that of GAPDH. The expression level of β-casein gene in mock infectants treated with lactogenic hormones was set as 100%. The mean and sd of three independent experiments are shown. F, COS-7 cells were cotransfected with expression plasmids for PRL receptor (1 μg), STAT5a or STAT5b (1μ g), β-casein gene promoter-luciferase (1 μg), and empty vector (mock) or each of HA-TC-PTP wild type, catalytically inactive C/S, and D/A mutants (2 μg). A β-galactosidase gene (0.4 μg) was also included to normalize for transfection efficiency. Cells were induced with PRL for 15 h (+, odd lanes) or were left untreated (−, even lanes) and then lysed for enzymatic assay. Luciferase activity was represented as the fold induction to that of mock transfectant without PRL induction. Data are shown as the mean ± sd of three independent experiments.
As dephosphorylation of STAT5 leads to transcriptional inactivation of PRL-responsive genes, the expression level of β-casein gene was investigated by a semiquantitative RT-PCR strategy. As shown in Fig. 4E, β-casein gene expression was largely suppressed in the cells expressing TC-PTP wild type, which was comparable in the cells expressing PTP1B wild type. Surprisingly, overexpression of Cys/Ser as well as Asp/Ala mutants of TC-PTP also suppressed β-casein gene expression, and the expression level in Asp/Ala mutant-expressing cells was less and indistinguishable from that in TC-PTP wild type-expressing cells, whereas overexpression of PTP1B mutants exhibited apparently no effect on β-casein gene expression (13). This suggests that whereas TC-PTP wild type has the capacity to dephosphorylate STAT5, the TC-PTP mutants would bind to the phosphotyrosine on the STAT5 through mutated catalytic domains, thus competing with the SH2 domain of the STAT5 and interfering with binding to β-casein gene promoter sequence.
These results were further confirmed by studying the effect of TC-PTP on PRL-induced transcriptional activation of the β-casein gene promoter. TC-PTP was transfected into COS-7 cells together with PRL receptor, STAT5a or STAT5b, and the β-casein gene promoter-luciferase construct. A β-galactosidase gene was also included to normalize for transfection efficiency. Luciferase activity was determined in extracts from cells left untreated or stimulated with PRL. As shown in Fig. 4F, transcriptional induction of β-casein gene promoter was completely suppressed, when TC-PTP wild type was coexpressed (lanes 4 and 12). Consistent with suppression of endogenous β-casein gene expression, coexpression of catalytically inactive forms of TC-PTP suppressed transcriptional activation of the β-casein gene promoter (lanes 6, 8, 14, and 16).
Overexpression of TC-PTP Does Not Inhibit Nuclear Translocation of STAT5, but Accelerates Its Export Back to Cytosol
Next, subcellular localization of STAT5 was examined in TC-PTP-expressing COMMA-1D cells. At the indicated time points after PRL stimulation, cells were lysed, and cytosolic and nuclear fractions were prepared. As shown in Fig. 5, until 30 min after PRL stimulation, subcellular localization of STAT5 in TC-PTP-expressing cells was similar to that in mock infectants, where the amounts of cytosolic STAT5 protein gradually reduced after PRL stimulation and, conversely, the amounts of nuclear protein increased. However, significant amounts of STAT5 were detected in the cytosolic fraction of TC-PTP wild type-expressing cells, and conversely nuclear STAT5 was reduced 40 min after PRL stimulation, whereas TC-PTP mutant exhibited no effect on the subcellular localization of STAT5 at the same time point, suggesting that nuclear dephosphorylation of STAT5 by TC-PTP accelerated its export back to cytosol. In PTP1B-expressing cells, STAT5 was retained in the cytosol throughout the time points after PRL stimulation. TC-PTP wild type as well as catalytically inactive mutants were localized mostly (>90% as determined densitometorically) in nucleus, whereas PTP1B mostly (>90%) in cytosol throughout PRL stimulation. Faint bands in the nuclear fractions for PTP1B and in the cytosolic fractions for TC-PTP were always detected, possibly due to experimental limitation for subcellular fractionation protocol used.
Subcellular Localization of STAT5 after PRL Stimulation COMMA-1D cells expressing TC-PTP or PTP1B were stimulated with PRL for the indicated time after serum starvation. Cells were lysed, and cytoplasmic (C) and nuclear (N) fractions were prepared as described in Materials and Methods. An equivalent of the cytoplasmic and nuclear fractions was separated by SDS-PAGE and immunoblotted with anti-STAT5 antibody. After stripping, the same membranes were reprobed with anti-HA antibody for localization of HA-TC-PTP or PTP1B.
DISCUSSION
As the findings of JAK-STAT pathway involved in many cytokines, including GH and PRL, many efforts were focused on the positive regulatory mechanisms. Like many signaling pathways, JAK-STAT signal transduction is balanced by tyrosine phosphorylation and dephosphorylation. Deactivation of tyrosine-phosphorylated STAT proteins involves both tyrosine dephosphorylation and nuclear export back to the cytoplasm for a subsequent cycle of activation/inactivation (18, 19), whereas these processes have been largely unknown.
It has been reported that IL-2-activated STAT5 is dephosphorylated by a cytoplasmic phosphatase SHP-2 in cytosol (10), although the phosphatase was shown to be involved in PRL-mediated signaling in a positive fashion (11, 12). PRL-activated STAT5 was not dephosphorylated by SHP-2 and was, instead, efficiently dephosphorylated by PTP1B in cytosol in reconstituted COS-7 cells as well as in mammary epithelial COMMA-1D cells (11–13). Although PTP1B was the first one that could dephosphorylate PRL-activated STAT5 in vivo, we could not rule out the possibility that other cytoplasmic, nuclear, or membrane-spanning PTPs were involved in PRL signaling, because most of the PTPs identified were also shown to be down-regulated in lactating mammary glands (14). Simply assumed from the suppressed gene expression of most of the PTPs in lactating mammary gland, PTP1B as well as other PTPs might cooperatively contribute to dephosphorylation of STAT5 in their respective subcellular compartments.
In the present study we showed that a nuclear phosphatase, TC-PTP, could clearly dephosphorylate PRL-activated STAT5a and STAT5b, but the degree of dephosphorylation activity appeared to be weaker than that of PTP1B 30 min after PRL stimulation in transfected COS-7 cells (Fig. 2) (13) as well as in COMMA-1D mammary epithelial cells (Fig. 4), although both PTPs could dephosphorylate STAT5 to the same extent in vitro (Fig. 3A). This was reasoned by the data that tyrosine phosphorylation of STAT5 in TC-PTP-expressing cells were scarce and indistinguishable from those in PTP1B-expressing cells 40 min following PRL stimulation (Fig. 4, B and C), suggesting that nuclear translocation of STAT5 was necessary for its dephosphorylation by TC-PTP, and therefore, further time was required for its efficient dephosphorylation by the phosphatase (Figs. 4 and 5).
Dephosphorylated STAT5 protein in nucleus should be exported back to cytosol for subsequent recycling. In COMMA-1D cells expressing TC-PTP wild type, most of STAT5 was retained in nucleus 30 min after PRL stimulation, although approximately 80% of the protein was dephosphorylated compared with mock infectants (Fig. 4, B and C). Further incubation resulted in nearly complete dephosphorylation of STAT5 (Fig. 4, B and C) and appearance of cytosolic STAT5 concomitant with reduction in nuclear STAT5, but still a significant amount of STAT5 was present in nucleus (Fig. 5). This apparent time lag might reflect a complex mechanism for STAT5 nuclear export back through unidentified molecules, which might also be regulated by TC-PTP.
Most of ectopically expressed PTP1B (>90%) was localized in cytosol, whereas TC-PTP (>90%) in nucleus in mammary epithelial cells regardless of PRL stimulation (Fig. 5). Although all the data in our present and previous studies were obtained by overexpression study, we suggest that such dual localization of the two different inhibitory PTPs guarantees proper regulation of post-PRL receptor signaling by dephosphorylating and deactivating STAT5 in vivo. PTP1B-null mice have been available (20, 21), whereas the use of TC-PTP null mice has been limited, because they die between 3 and 5 wk of age (22). A greater physiological relevance of PTP1B and TC-PTP in mammary epithelial cells could be clarified using the cells isolated from the gene-disrupted mice. However, based on our findings that both TC-PTP and PTP1B are potent inhibitors of PRL signaling, no phenotype might be observed when the cells have a single gene disruption, possibly due to biological redundancy.
Localization of nuclear TC-PTP was unchanged throughout PRL stimulation (Fig. 5), whereas it has been reported that nuclear TC-PTP, which is the same one focused in the present study, translocated to cytosol upon EGF stimulation and inhibited the EGF-dependent activation of PI3K and PKB/Akt (23, 24), suggesting that localization of nuclear TC-PTP is differentially regulated by individual ligand stimulation. Nuclear localization of TC-PTP throughout PRL stimulation suggests that dimerized STAT5 through its phosphotyrosine and SH2 domains is attacked by TC-PTP possibly in a competitive manner, leading to the formation of a stable complex between the molecules, and that TC-PTP does not function as a chaperon for nuclear translocation of STAT5 from cytosol.
STAT5a and STAT5b were dephosphorylated by recombinant TC-PTP and were coimmunoprecipitated with catalytically inactive forms, so-called substrate-trapping mutants of TC-PTP (Fig. 3), indicating that STAT5a and STAT5b are specific substrates for not only PTP1B (13), but also TC-PTP. Roughly estimated, 30% and 40% of phosphorylated STAT5 were coimmunoprecipitated with Cys/Ser and Asp/Ala mutant of TC-PTP, respectively (data not shown). On the other hand, phosphorylated STAT5 could be precipitated only upon using excess amounts of GST fusion proteins of PTP1B substrate-trapping mutants (13), but not specific antibody (data not shown), suggesting that STAT5 is a more preferred in vivo substrate for TC-PTP. Partially inconsistent with coimmunoprecipitation data, substrate-trapping mutants of TC-PTP suppressed endogenous β-casein gene expression in COMMA-1D epithelial cells (Fig. 4E) as well asβ -casein gene promoter activation in COS-7 cells to the similar extent as TC-PTP wild type. This might be in part explained by the fact that coimmunoprecipitation data do not necessarily reflect the in vivo situation, possibly due to experimental limitation and/or that TC-PTP might also contribute to deactivation of STAT5 through to date unidentified mechanisms.
EGF receptor and insulin receptor have also been identified to be specific and common substrates of TC-PTP (23) and PTP1B (25), whereas a distinct set of proteins was coprecipitated with the PTPs (23) despite the extensive similarity between TC-PTP and PTP1B catalytic domains (72%; Fig. 1). It might be interesting to examine whether sequence-dependent and/or conformational similarities among tyrosine-phosphorylated STAT5, EGF receptor, and insulin receptor exist for dephosphorylation of the proteins by TC-PTP and PTP1B. In addition to STAT5, it has been reported that the PRL stimulation resulted in tyrosine phosphorylation of STAT1 and STAT3 (26). Whether TC-PTP as well as PTP1B are involved in negative regulation of other JAK-STAT pathways is currently being studied in our laboratory.
The TC-PTP is an intracellular nontransmembrane phosphatase that was originally cloned from a human T cell cDNA library (27), but is now known to be expressed in many tissues. TC-PTP contains a conserved catalytic domain and a noncatalytic C-terminal segment that varies in size and function as a result of alternative splicing (28). Two splice variants differing only in their extreme C-termini are expressed. The 48-kDa form of human TC-PTP contains a 34-residue hydrophobic tail as PTP1B, which is replaced by a hydrophilic 6-residue sequence in the 45-kDa form. The 48-kDa form of TC-PTP localizes to the endoplasmic reticulum (29, 30), whereas under basal conditions the 45-kDa form is localized in the nucleus due to the presence of a bipartite nuclear localization sequence (15, 28, 30–32). In this study we examined the involvement of the 45-kDa form of mouse TC-PTP in PRL-mediated signaling, because it was actually expressed in mammary epithelial cells (14), and the cDNA sequence for the 48-kDa form has been unavailable. Although the 45-kDa form of mouse TC-PTP was shown to be active in nucleus and dephosphorylate STAT5, it remains uncertain at present whether the 48-kDa form of mouse TC-PTP locates in ER and dephosphorylates STAT5 in mammary epithelial cells.
Recently, Wang et al. (33) reported that a small amphipathic α-helical region was required for proteasome-dependent turnover of the tyrosine-phosphorylated STAT5a, and accordingly, truncation of the C-terminal region of STAT5a resulted in prolonged tyrosine phosphorylation, suggesting that the C-terminal small region is involved in tyrosine dephosphorylation. We examined the dephosphorylation activity of PTP1B and TC-PTP on the truncated forms of STAT5a and STAT5b in transfected COS-7 cells, but the degree of dephosphorylation was indistinguishable from that of STAT5a and STAT5b wild type. Furthermore, PTP1B and TC-PTP dephosphorylation activity was insensitive to proteasome inhibitor MG132 (Aoki, N., et al. unpublished observations). Therefore, we still cannot rule out the possibility that other phosphatases might be involved in dephosphorylation of STAT5.
In conclusion, we demonstrated nuclear dephosphorylation and deactivation of STAT5 by TC-PTP. Our previous and current studies suggest that PRL-activated STAT5 is cooperatively regulated by PTPs in both cytosol and nucleus.
MATERIALS AND METHODS
Materials, Antibodies, and Plasmid Constructs
Ovine PRL was obtained from Sigma (St. Louis, MO). Polyclonal antibodies to STAT5 (C-17), recognizing both mSTAT5a and mSTAT5b, HA epitope (Y-11), and JAK2 (M-126), were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Monoclonal antiphosphotyrosine antibodies (4G10) were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Protein A-Sepharose beads used for immunoprecipitation were obtained from Amersham Pharmacia Biotech (Little Chalfont, UK).
Mouse TC-PTP (S52655) was amplified by RT-PCR using the primer set: 5′-CTG-TCC-GCT-GTG-GTA-GTT-CC-3′ (nucleotides 22–41) and 5′-GCT-GCA-GAA-TAG-TCT-CAA-GT-3′ (nucleotides 1220–1239). HA-tagging to TC-PTP at its N-terminal was performed by PCR amplification using 5′-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCC-CAG-ACT-ACG-CTT-CGG-CAA-CCA-TCG-AGC-GG-3′ and the above- mentioned antisense primer. All of the PCR products were cloned into a mammalian expression vector, pTargeT vector (Promega Corp., Madison, WI), and confirmed by sequencing on both strands. The HA-tagged TC-PTP mutants containing a cysteine to serine alteration at position 216 and an aspartic acid to a alanine at position 182 were generated using oligonucleotide primers 5′-CCG-CAC-TGC-TAT-GGA-TCA-3′ and 5′-AAC-CCC-AAA-AGC-TGG-CCA-GGT-3′, respectively, as previously described (34). The mutation was confirmed by DNA sequencing.
Expression plasmids for mouse PRL receptor (pCMX-PL1), mouse STAT5a (pXM-mSTAT5a), and STAT5b (pXM-mSTAT5b) were provided by Dr. B. Groner (Institute for Experimental Cancer Research, Freiburg, Germany).
Cell Culture, Transfection, Cell Lysis, Subcellular Fractionation, and Western Blotting
COS-7 and COMMA-1D cells were maintained in DMEM containing 10% FCS and transfected as previously described (35). After stimulation with PRL (5 μg/ml) for the indicated time, cells were lysed, followed by immunoprecipitation and Western blotting with the respective antibodies or by biochemical cell fractionation as previously described (13).
Retrovirus-Mediated Gene Delivery
HA-tagged TC-PTP was ligated into pLXSN retroviral vector (CLONTECH Laboratories, Inc., Palo Alto, CA) via EcoRI site and introduced into Pheonix ecotropic packaging cells. COMMA-1D cells were infected with the retrovirus-containing culture medium and then selected in the presence of G418 (1 mg/ml) for 2 wk. To eliminate clonal deviation, G418-resistant polyclonal cells were used for subsequent experiments.
In Vitro Dephosphorylation Assay
GST fusion proteins containing full-length TC-PTP were constructed as follows. Full-length TC-PTP was PCR amplified using a primer set of 5′-ATGAATTCTCGGCAACCATGGAGCGG-3′ and 5′-GCT-GCA-GAA-TAG-TCT-CAA-GT-3′ with pTargeT-HA-TC-PTPs as templates, and the resultant products were digested with EcoRI and ligated into pGEX-5X-1 vector (Amersham Pharmacia Biotech) through the same cloning site. GST fusion proteins were purified on glutathione-Sepharose beads and eluted with neutralized glutathione. Enzymatic activities of the GST fusion proteins were determined using para-nitrophenyl phosphate, as described previously (36). STAT5a and STAT5b immune complexes prepared from PRL-treated COS-7 cells that had been cotransfected with PRL receptor and STAT5a were processed and incubated with indicated GST fusion proteins as previously described (13).
Acknowledgments
We thank Dr. Berund Groner for provision of expression plasmids.
Abbreviations:
- GST,
Glutathione-S-transferase;
- HA,
hemagglutinin;
- JAK,
Janus kinase;
- PTK,
protein tyrosine kinase;
- SHP,
SH2-containing protein tyrosine phosphatase-2;
- STAT,
signal transducer and activator of transcription.
