Sts-2 Is a Phosphatase That Negatively Regulates Zeta-associated Protein (ZAP)-70 and T Cell Receptor Signaling Pathways*

T cell activity is controlled in large part by the T cell receptor (TCR). The TCR detects the presence of foreign pathogens and activates the T cell-mediated immune reaction. Numerous intracellular signaling pathways downstream of the TCR are involved in the process of T cell activation. Negative regulation of these pathways helps prevent excessive and deleterious T cell responses. Two homologous proteins, Sts-1 and Sts-2, have been shown to function as critical negative regulators of TCR signaling. The phosphoglycerate mutase-like domain of Sts-1 (Sts-1PGM) has a potent phosphatase activity that contributes to the suppression of TCR signaling. The function of Sts-2PGM as a phosphatase has been less clear, principally because its intrinsic enzyme activity has been difficult to detect. Here, we demonstrate that Sts-2 regulates the level of tyrosine phosphorylation on targets within T cells, among them the critical T cell tyrosine kinase Zap-70. Utilizing new phosphorylated substrates, we demonstrate that Sts-2PGM has clear, albeit weak, phosphatase activity. We further pinpoint Sts-2 residues Glu-481, Ser-552, and Ser-582 as specificity determinants, in that an Sts-2PGM triple mutant in which these three amino acids are altered to their counterparts in Sts-1PGM has substantially increased activity. Our results suggest that the phosphatase activities of both suppressor of TCR signaling homologues cooperate in a similar but independent fashion to help set the threshold for TCR-induced T cell activation.

Within the immune system, T cells play an instrumental role in eliminating foreign pathogens. They do so either directly, by killing cells that harbor deleterious microbes, or indirectly, by secreting a variety of different cytokines that help orchestrate the responses of many other cell types (1,2). T cells become activated following stimulation of a surface receptor known as the T cell receptor (TCR). 3 Engagement of the TCR activates an array of intracellular signaling cascades (3). As part of the process to ensure T cells are only activated within the appropriate immune context, these signaling pathways are tightly regulated by a variety of overlapping mechanisms (4). Two proteins that belong to the suppressor of TCR signaling family, Sts-1 and Sts-2, have been identified as negative regulators of TCR signaling (5). Within T cells, the Sts proteins appear to regulate elements of intracellular signaling cascades that are proximal to the TCR, including the important T cell kinase Zap-70 and a number of unknown proteins that are rapidly modified by tyrosine phosphorylation and ubiquitination following TCR stimulation (6). The important role of Sts-1 and Sts-2 in controlling cellular responses is underscored by the hyper-reactivity of T cells that lack both proteins and the susceptibility to auto-immunity of mice in which both genes have been genetically ablated (7).
Sts-1 and Sts-2 are characterized by a unique tripartite structure composed of an N-terminal ubiquitin association (UBA) domain that can bind mono-and polyubiquitin (8), a central Src homology 3 (SH3) domain that interacts with proline-rich segments of other polypeptides, and a C-terminal PGM domain that has limited sequence homology to members of the phosphoglycerate mutase (PGM)/acid phosphatase (AcP) superfamily of enzymes. PGM/AcP enzymes possess phosphatase or phosphotransferase catalytic activity and are characterized by a common tertiary structure (9). They are also referred to as 2H-phosphatases because their catalytic mechanism relies on two signature histidines that are conserved in all members of the superfamily (10). Superfamily members include such diverse enzymes as phosphoglycerate mutase, fructose-2,6-bisphosphatase, prostatic acid phosphatase, and bacterial phosphatase SixA. Despite limited primary amino acid sequence homology to PGM family members, both Sts-1 PGM and Sts-2 PGM possess the critical catalytic residues and adopt the overall fold that characterizes all PGM enzymes (11,12).
T cells that lack the Sts proteins (Sts-1/2 Ϫ/Ϫ ) display hyperphosphorylation of the protein-tyrosine kinase Zap-70 following TCR stimulation (7). In addition, hyper-phosphorylation of a number of unknown ubiquitinated proteins that are gener-ated following TCR engagement is evident in T cells lacking Sts-1 and Sts-2 (6). These observations led to the hypothesis that Sts-1 and Sts-2 were intracellular phosphatases and that their substrates were tyrosine-phosphorylated components of TCR signaling pathways. Consistent with this model, the PGM domain of Sts-1 was demonstrated to have in vitro tyrosine phosphatase activity (11). In particular, it was shown to have the capacity to efficiently hydrolyze the nonspecific phosphatase substrate para-nitrophenyl-phosphate (pNPP), a Tyr(P) analogue, as well as tyrosine-phosphorylated peptides and proteins (13). In addition, Sts-1 was found to readily target numerous substrates within cells, including the tyrosine kinases Zap-70, Syk, and the EGF receptor (11,14,15). In contrast, Sts-2 PGM in vitro phosphatase activity toward pNPP was found to be 5 orders of magnitude weaker than the activity of Sts-1 PGM , despite the presence of all conserved catalytic residues. Additionally, Sts-2 PGM does not dephosphorylate tyrosine-phosphorylated peptides and proteins in vitro (12) nor does Sts-2 target tyrosine-phosphorylated substrates similar to Sts-1 when it is overexpressed in cells (14,16). The catalytic inefficiency of Sts-2 PGM has led to speculation that Sts-2 either has a very narrow substrate range or, alternatively, that it does not function as an intracellular phosphatase within T cells (5,11,17).
To more precisely define the role of Sts-2 in regulating TCR signaling pathways, we undertook an investigation into its specific functional and biochemical properties. Our results support a model in which Sts-2, along with Sts-1, is involved in the dephosphorylation and down-regulation of key components within the TCR signaling pathways.
Mice-The Sts mutations were back-crossed as heterozygote mutations through 10 generations of wild-type Bl/6 mice, following which Sts-1 Ϫ/Ϫ , Sts-2 Ϫ/Ϫ , and Sts-1/2 Ϫ/Ϫ mice were obtained by the appropriate matings. Mice were housed and bred under specific pathogen-free conditions according to institutional guidelines. Animal work was conducted under guidelines established and approved by the Stony Brook Institutional Animal Care and Use Committee.
Cell Culture, T Cell Stimulations, and Transient Transfections-Primary murine T cells were obtained from spleens dissected from wild-type, Sts-1 Ϫ/Ϫ , Sts-2 Ϫ/Ϫ , and Sts-1/2 Ϫ/Ϫ mice. To generate activated T cells, spleens were crushed in PBS containing 2% FBS; red blood cells were lysed by addition of ACK lysis buffer, pH 7.2, containing 150 mM NH 4 Cl, 1 mM KHCO 3 , and 0.1 mM EDTA, and debris was removed by straining through a 70-m filter (BD Biosciences). Total splenocyte cultures were established, and T cells were activated and expanded for 2 days in T cell medium (RPMI 1640 media containing 10% FCS, 10 mM HEPES, pH 7.0, 2 mM glutamine, 1 mM sodium pyruvate, 50 M ␤-mercaptoethanol, 0.1 mM nonessential amino acids, and the appropriate concentration of penicillin/streptomycin, all from Invitrogen) supplemented with 0.5 g/ml anti-TCR antibody 145-2C11 and 2 units/ml recombinant human IL-2. The T cells cultures were further expanded for 2 more days in fresh T cell media containing IL-2, collected by centrifugation, and utilized for biochemical analyses. Activated T cells that had been expanded in IL-2 for 4 days (2 ϫ 10 7 /point) were labeled on ice for 25 min with 20 g/ml biotinylated anti-CD3, anti-CD4, and anti-CD8. They were then washed with 2% FBS in PBS, incubated for 20 on ice with 10 g/ml streptavidin, placed in a 37°C waterbath for 2 min, and immediately lysed in cold lysis buffer (see below). Illustrated results are representative of multiple (Ն3) assays. HEK293 cells were cultured in DMEM (Invitrogen), supplemented with 10% FCS and penicillin/streptomycin. DNAs were transiently transfected into 293 cells by calcium phosphate-mediated transfection.
Cell Lysis, Immunoprecipitations, and Immunoblots-Cells were lysed in ice-cold lysis buffer containing 50 mM Tris, pH 7.6, 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1 mM Na 3 VO 4 , 1% Nonidet P-40 substitute Igepal TM CA-630 (United States Biochemical Corp.), 100 M PMSF, and Complete TM protease inhibitors (Roche Applied Science). Cell lysates were clarified by centrifugation and then either mixed with an equal volume of Laemmli sample buffer or processed for immunoprecipitation analysis. For immunoprecipitations, lysates were rotated at 4°C with specific antibody and 20 l of protein A-Sepharose 50% slurry (Sigma) for 1-2 h. Beads were washed three times with cold wash buffer (20 mM Tris, pH 7.6, 150 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40 substitute, 100 M PMSF, and 1 mM Na 3 VO 4 ), and bound proteins were eluted with Laemmli sample buffer and separated by SDS-PAGE. Proteins were transferred to either nitrocellulose (Whatman) or PVDF (Millipore) using a semidry transfer apparatus (Bio-Rad), blocked with 3% BSA in Tris-buffered saline (TBS, pH 8.0), incubated at 23°C for 1 h or 4°C overnight with specific antibody, washed with TBS, incubated for 1 h with the appropriate secondary antibody, and developed with the ODYSSEY Infrared Imaging System (LI-COR).
Expression of Recombinant Sts PGM Domains-Bacterial expression constructs of Sts-2 PGM and Sts-1 PGM were generated by excising the Sts PGM coding sequences from the pProEX-Htb vector (12,19) with BamHI and SalI and inserting them into the pMAL-c4X vector (New England Biolabs). Maltosebinding protein-Sts PGM fusion proteins were expressed in Escherichia coli K12 cells and purified by affinity chromatography using amylose resin columns following the manufacturer's suggested protocol (New England Biolabs). Briefly, a 1-liter culture of cells was grown to A 600 ϳ0.5, induced with 0.3 mM isopropyl 1-thio-␤-D-galactopyranoside for 3 h at 37°C, harvested by centrifugation, resuspended 1:5 in column buffer (20 mM Tris-HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA and 10 mM ␤-mercaptoethanol), and frozen overnight at Ϫ20°C, prior to lysis by sonication. The crude extract was loaded into an amylose resin column (15-ml bed volume) at a flow rate of 1 ml/min, and the fusion protein was eluted with 10 mM maltose in column buffer, collecting 3-ml fractions. Fractions containing the fusion proteins were identified by SDS-PAGE/Coomassie staining, pooled, dialyzed overnight in a 20 mM Tris-HCl buffer, pH 8.0, containing 25 mM NaCl. For further purification, the dialyzed sample was loaded into a DEAE-Sepharose ion-exchange column (6 ml bed volume) and step-eluted starting from 25 to 300 mM NaCl in 20 mM Tris-HCl buffer, pH 8.0. MBP-Sts PGM -containing fractions were pooled and concentrated using an Amicon Ultra-15 centrifugal filter device with a 5-kDa cutoff (Millipore).
Enzymes Assays-Continuous phosphatase assays (100 l of reaction volume) were conducted in 96-well flat-bottom plates (BD Biosciences) using a SpectraMax 190 Microplate reader (Molecular Devices) in buffer containing 25 mM HEPES, pH 7.2, or 50 mM MES, pH 6.0, 50 mM NaCl, 5 mM DTT, 2.5 mM EDTA, and 0.1 mg/ml BSA. Phosphate hydrolysis of pNPP (⑀, 17,800 M Ϫ1 cm Ϫ1 ) was monitored by measuring absorbance at 405 nm. Dephosphorylation of OMFP (⑀, 27,000 M Ϫ1 cm Ϫ1 ) and FDP (⑀, 34,000 M Ϫ1 cm Ϫ1 ) was monitored by measuring absorbance at 477 and 450 nm, respectively. Assays were conducted at 37°C, and no-enzyme controls were included as blanks. To determine kinetic parameters, substrate saturation curves were prepared from initial velocities using 150 nM Sts-2 PGM and 15 nM Sts-1 PGM with increasing substrate concentrations. The Michaelis-Menten constant K m and turnover rate k cat were determined using a dynamic curve fitting function in SigmaPlot 11 (Systat Software, Inc.) to a one-site saturation, ligand binding equation, End point assays were conducted to measure the phosphatase activity of FLAG-tagged Sts PGM transiently expressed in HEK293 cells. Briefly, FLAG-M2 immunoprecipitates immobilized on protein A-Sepharose beads were washed three times with assay buffer then incubated in 100 l of assay buffer-containing substrate. The beads were incubated at 37°C, and the reactions were allowed to proceed for the time specified in the figure legends. Reactions were stopped by incubating on ice for 5 min and collecting the supernatants to measure absorbance of hydrolysis products. Laemmli sample buffer was added to the beads to determine levels of FLAG-tagged Sts PGM by SDS-PAGE and immunoblotting. Assays were performed in triplicate, and immunoprecipitations from empty vector transfections were used as blanks.
Intracellular Cytokine Analysis-Activated T cells (2 ϫ 10 5 ) from wild-type and mutant mice were stimulated in triplicate with 0.1 g/ml plate-bound anti-CD3 (145-2C11, BD Biosciences). Following 6 h of stimulation, cells were treated for 2 h with brefeldin A and processed for intracellular IFN␥ staining using the Cytofix/Cytoperm Fixation/Permeabilization kit (BD Biosciences) according to the manufacturer's instructions. Analysis was conducted on a FACSCalibur.

Regulation of Zap-70 Tyrosine Phosphorylation by Sts-2-To assess the individual roles of Sts-1 and Sts-2 in regulating
Zap-70 activation, we utilized T cells derived from wild-type, Sts-1 Ϫ/Ϫ , Sts-2 Ϫ/Ϫ , and Sts-1/2 Ϫ/Ϫ mice (supplemental Fig. S1). Cells were stimulated and lysed, and total cytoplasmic proteins were separated by gel electrophoresis. The levels of phosphorylation of the three Zap-70 tyrosines, Tyr-493, Tyr-319, and Tyr-292, were then determined using phospho-specific antibodies. Each of these tyrosines has been shown to play unique functional roles; Tyr-493 is located in the activation loop of the Zap-70 kinase domain, and its phosphorylation plays a positive regulatory role; Tyr-319 is also thought to play a positive signaling role through its ability to recruit PLC␥; and Tyr-292 appears to play a negative regulatory role, although the mechanism is not completely understood (20,21). The data in Fig. 1 (A, C, and E) demonstrates differences in the relative levels of Zap-70 site-specific tyrosine phosphorylation in T cells derived from wild-type and mutant mice. Although there was no statistically significant difference between phosphorylation levels observed within Sts-2 Ϫ/Ϫ and wild-type T cells, there was a clear difference between Sts-1 Ϫ/Ϫ and Sts-1/2 Ϫ/Ϫ cells. Thus, although the absence of Sts-2 alone had no readily apparent effect on Zap-70, maximal hyper-phosphorylation at distinct Zap-70 tyrosine phosphorylation sites required the absence of both Sts-2 and Sts-1. This suggests that each Sts protein functions as a separate and distinct negative regulator of Zap-70 signaling, with Sts-2 having a weaker effect than Sts-1. That they do not function equivalently is suggested by the fact that the absence of Sts-1 consistently produced higher levels of relative phosphorylation than the absence of Sts-2 (see quantification, Fig. 1, B, D, and F). Whether this is due to differences in the intrinsic phosphatase catalytic properties of the different Sts PGM domains or the involvement of Sts domains or other factors is currently unclear (see below).
Regulation of Tyrosine-phosphorylated, Ubiquitinated Proteins by Sts-2-In addition to targeting Zap-70 for dephosphorylation and down-regulation, the Sts proteins have been shown to be instrumental in down-modulating T cell proteins that are dually modified by both tyrosine phosphorylation and ubiquitination following engagement of the TCR (6). In particular, Sts-1/2 Ϫ/Ϫ T cells display strikingly elevated levels of dually modified proteins following TCR stimulation. The extent to which each Sts protein individually contributes to the down-regulation of tyrosine-phosphorylated, ubiquitinated polypeptides is unclear. To assess the individual contributions of Sts-1 and Sts-2, we compared levels of ubiquitinated proteins in phosphotyrosine immunocomplexes isolated from stimulated Sts-1 Ϫ/Ϫ or Sts-2 Ϫ/Ϫ T cells to those obtained from Sts-1/ 2 Ϫ/Ϫ T cells. Compared with the latter, T cells isolated from Sts singly deficient mice accumulated significantly reduced levels of proteins that were dually modified by tyrosine phosphorylation and ubiquitination (Fig. 2). As in the case of Zap-70 phosphorylation levels, the absence of Sts-1 produced a stronger effect than the absence of Sts-2. In particular, the level of dually modified proteins observed in Sts-2 Ϫ/Ϫ T cells was considerably less than the level observed in Sts-1 Ϫ/Ϫ T cells.
Regulation of Downstream Signaling Pathways by Sts-2-The absence of both Sts proteins in T cells leads to hyper-activation of downstream signaling pathways, a result that likely stems from deregulation of TCR proximal kinases such as Zap-70 (7). Although the individual contributions of each Sts protein have not been evaluated, the results outlined above suggest that both Sts proteins play distinct roles in the regulation of downstream signaling pathways. To determine the contribution of each Sts protein to the activation of downstream signaling cascades, we compared both the activation of calcium signaling pathways and the activation of MAPK in T cells derived from wild-type, Sts-1 Ϫ/Ϫ , Sts-2 Ϫ/Ϫ , and Sts-1/2 Ϫ/Ϫ mice. The enzyme PLC␥1 is an essential mediator of intracellular calcium fluxes because it generates key secondary messengers such as inositol 1,4,5-trisphosphate (22). PLC␥1 is activated, in part, by tyrosine phosphorylation that is thought to be induced by TCR proximal kinases (23). To evaluate the individual contributions of Sts-1 and Sts-2 to the state of PLC␥1 activation, lysates derived from unstimulated and stimulated T cells were prepared, and the levels of activated PLC␥1 were assessed by Western analysis utilizing a phosphorylation-specific antibody directed against PLC␥1 Tyr-783. The results mirrored those obtained in the analysis of Zap-70 phosphorylation. Specifically, Sts-1 Ϫ/Ϫ and Sts-2 Ϫ/Ϫ T cells each displayed levels of phosphorylation on PLC␥1 Tyr-783 that were elevated over those observed in wildtype cells but less than the level observed in Sts-1/2 Ϫ/Ϫ T cells. In addition, loss of Sts-2 alone had less of an effect on PLC␥1 activation than loss of Sts-1. Downstream of PLC␥1, levels of MAPK activation in stimulated Sts-1 Ϫ/Ϫ and Sts-2 Ϫ/Ϫ T cells were found to be greater than in wild-type cells but less than that observed in Sts-1/2 Ϫ/Ϫ cells (Fig. 3B). These data suggest that Sts-1 and Sts-2 contribute distinct, nonredundant negative regulatory functions within T cells to modulate the levels of activation of downstream calcium signaling pathways and MAPK signaling.   cell and the context in which TCR stimulation occurs (24). T cells lacking the Sts proteins have been shown to secrete dramatically increased levels of several cytokines, including IFN␥ relative to wild-type T cells (7). To determine the contribution of each Sts protein to the TCR-induced up-regulation of IFN␥ expression, we evaluated levels of IFN␥ induced in T cells derived from wild-type, Sts-1 Ϫ/Ϫ , Sts-2 Ϫ/Ϫ , and Sts-1/2 Ϫ/Ϫ mice following TCR stimulation. As illustrated in Fig. 4, both Sts-1-deficient and Sts-2-deficient T cells expressed levels of IFN␥ that were significantly less than the levels produced by Sts-1/2 Ϫ/Ϫ T cells. The consistent production of greater IFN␥ levels in Sts-1 Ϫ/Ϫ T cells relative to levels in Sts-2 Ϫ/Ϫ T cells paralleled the consistent increase in activation of different signaling molecules in Sts-1 Ϫ/Ϫ cells relative to Sts-2 Ϫ/Ϫ cells. As with the latter results, the levels of IFN␥ production by the mutant T cells support the notion that Sts-1 and Sts-2 contribute distinct negative regulatory functions within T cell signaling pathways.
Phosphatase Activity of Sts-2-We have recently shown that the in vitro phosphatase activity of Sts-2 PGM toward pNPP is 5 orders of magnitude weaker that that of Sts-1 PGM , despite conservation of all catalytic residues (12). We argued that a possible explanation for this difference is the nature of the substrate rather than a lack of intrinsic phosphatase activity. To deter- mine whether other molecules might be hydrolyzed with greater efficiency by Sts-2 PGM , we evaluated phosphorylated fluorescein analogues as possible substrates. Phosphorylated fluorescein compounds have been developed as readily available tools to study the activity of protein-tyrosine phosphatases (25,26). Fig. 5A illustrates that recombinant Sts-2 PGM dephosphorylates O-methylfluorescein phosphate (OMFP) and FDP, and it does so with significantly greater efficiency than it hydrolyzes pNPP. Notably, 20-fold lower Sts-2 PGM concentrations are needed to achieve measurable reaction velocities with OMFP and FDP than with pNPP, resulting in higher specific activities for the fluorescein analogues. For example, as estimated from assays in which the concentrations of pNPP and OMFP were 0.5 mM and FDP was 0.2 mM, the Sts-2 PGM specific  05-1 mM). C, substrate saturation curves for Sts-2 PGM against pNPP, OMFP, and FDP. D, OMFP hydrolysis activity of wild-type Sts-2 PGM or Sts-2 PGM active site mutants (H366C, R448A, and H551A). FLAG-tagged PGM domains were expressed in HEK293 cells, immunoprecipitated with M2 anti-FLAG antibody, and immobilized on protein A-Sepharose beads. The phosphatase assay was performed by adding 0.5 mM OMFP to the beads and incubating at 37°C for 1 h. Assays were conducted in triplicate, and representative protein levels for each reaction were identified with FLAG antibodies. E, inhibition of Sts-2 PGM phosphatase activity by vanadate. Activity of recombinant Sts-2 PGM (150 nM) against OMFP (0.5 mM) was measured in the presence of increasing amounts of orthovanadate (1.56 -50 mM). Rates were calculated from the slopes of the time course of the reaction and presented as ratios relative to the uninhibited (no vanadate) reaction.
Further analysis into the hydrolysis of OMFP by Sts-2 PGM demonstrated that the reaction follows Michaelis-Menten kinetics in which increases in substrate concentration result in faster initial reaction velocities (Fig. 5B). The phosphatase reaction was performed at multiple substrate concentrations (12.5 M to 1 mM), and the initial velocity at each substrate concentration was plotted (Fig. 5C). As expected, the estimated specificity constants (k cat /K m ) for the Sts-2 PGM -catalyzed hydrolysis of FDP and OMFP were found to be greater than Sts-2 PGM hydrolysis of pNPP (see Table 2). This is due both to significantly lower K m values and to significantly greater k cat values for the reactions involving the fluorescein compounds. The finding that OMFP is a better substrate than pNPP is not unprecedented. Previous work on Cdc25B, a dual specificity phosphatase, also demonstrated that OMFP is a better substrate than pNPP with an 810-fold greater specificity constant (27).
Like all PGM family members, Sts-2 PGM is characterized by a unique ␣/␤ core structure that consists of an inner parallel ␤ sheet surrounded by a series of ␣ helices and a catalytic core formed in part by several conserved active site residues, including His-366, Arg-448, and His-551, that cluster together (12). His-366 is the proposed nucleophile in the phosphatase reaction and is likely transiently phosphorylated during the catalytic reaction (28). To confirm that Sts-2 PGM -mediated dephosphorylation of OMFP represented a bona fide 2H phosphatase reaction, His-366, Arg-448, and His-551 were individually mutated to non-functional, and the activities of the resulting single point mutants were assessed. As shown in Fig. 5D, each mutation inactivated Sts-2 PGM , consistent with the hypothesis that the phosphatase activity of Sts-2 PGM is likely an intrinsic property derived from evolutionarily conserved active site residues common to all members of the PGM/AcP family. This model is strengthened by the observation that orthovanadate, a commonly used phosphotyrosine phosphatase inhibitor, can inhibit the dephosphorylation of OMFP by Sts-2 PGM (Fig. 5E). X-ray crystallographic studies have indicated that vanadate binds to the active site of Sts-2 PGM and interacts noncovalently with the nucleophilic residue His-366 (28).
Differences between Sts-2 and Sts-1 Catalytic Efficiency-As alluded to above, our prior work has demonstrated that Sts-2 PGM phosphatase activity toward the substrate pNPP is close to 5 orders of magnitude less than that of Sts-1 PGM . Additionally, although Sts-1 PGM can readily dephosphorylate tyrosinephosphorylated peptides and proteins in vitro, Sts-2 PGM -mediated hydrolysis of peptide and polypeptide substrates in vitro is difficult to detect (supplemental Fig. S2) (12). Thus, Sts-2 PGM is significantly less active relative to Sts-1 PGM . An important question is whether this is due to the weak intrinsic phosphatase activity of Sts-2 PGM or does the choice and nature of the substrate matter? To address this, we compared the abilities of Sts-2 PGM and Sts-1 PGM enzymes to hydrolyze FDP and OMFP. Relative phosphatase activities utilizing either bacterially expressed recombinant Sts PGM domains ( Fig. 6A and Table 1) or Sts PGM domains expressed in eukaryotic cells (Fig. 6B) or endogenous Sts proteins obtained from T cells (Fig. 6C) were evaluated. In each case, Sts-2 was found to be significantly less active than Sts-1. For example, the specific activity (0.5 mM substrate) of recombinant Sts-2 PGM toward OMFP was 72.3 nmol/min/mg, whereas the activity of recombinant Sts-1 PGM was 15,705 nmol/min/mg (Table 1). This represents an ϳ200fold difference between the two phosphatases in hydrolytic activity toward the fluorescein-based compounds. In stark contrast, Sts-2 PGM exhibited a 6000-fold difference in activity toward pNPP relative to Sts-1 PGM . Interestingly, Sts-1 PGM displayed a similar activity as protein-tyrosine phosphatase 1B (PTP1B).
We have hypothesized that the difference in hydrolytic activity between the two Sts PGM enzymes stems from differences in noncatalytic, nonconserved active residues (12). Such residues could affect both the orientation and/or movements of critical catalytic residues or influence the manner in which phosphorylated substrates are positioned within the active site. Recent structural and mutagenesis analysis suggests that the nonconserved Sts-2 residues Glu-481, Ser-552, and Ser-582 are among the amino acids that contribute to the dramatic differences in Sts-2 phosphatase activity between Sts1-and Sts-2 (12). Specifically, these three amino acids, unlike their counterparts in Sts-1 (Val-495, Ala-566, and Tyr-596) are within hydrogen bond distances of critical catalytic residues and so might restrict their movement during catalysis (Fig. 7A). For example, Glu-481 makes a stabilizing hydrogen bond interaction with Arg-448, a basic residue that is thought to be involved in correctly positioning the negatively charged phosphate moiety of the substrate (27). Additionally, Glu-481 results in a negative charge situated close to acidic Glu-476, which is a conserved residue whose counterpart in other PGM family members has been proposed to act as a general acid/base during the catalytic reaction. As a result, Glu-476 is forced into a different conformation within the Sts-2 PGM active site than its counterpart in the active site of Sts-1 PGM . To evaluate whether Sts-2 residues Glu-481, Ser-552, and Ser-582 are among the residues responsible for the weaker phosphatase activity of Sts-2, we generated the VAY (E481V, S552A, S582Y) triple mutant of Sts-2 PGM in  which each amino acid was altered to its counterpart in Sts-1.
With either pNPP (wild type: K m ϭ 2 mM, k cat ϭ 0.03 s Ϫ1 ; VAY: K m ϭ 0.59 mM, k cat ϭ 1.3 s Ϫ1 ) or OMFP (Fig. 7B), the VAY mutant was significantly more active than wild-type Sts-2 PGM . Thus, mutating nonconserved residues of Sts-2 PGM to their counterparts in Sts-1 PGM substantially increases the activity of The Sts-2 PGM reaction was allowed to proceed for 50 min, and the Sts-1 PGM and PTP1B reactions were stopped after 5 min. B, FLAG-tagged wild-type Sts-2 PGM and Sts-1 PGM were expressed in HEK293 cells and immunoprecipitated using anti-FLAG antibody, and phosphatase activity was assessed with 0.5 mM pNPP, 0.5 mM OMFP, and 0.2 mM FDP. The reactions were allowed to proceed as in A. Assays were conducted in triplicate, and representative protein levels for each reaction are illustrated (right). C, endogenously expressed Sts-2 and Sts-1 were immunoprecipitated from mouse primary T cells using rabbit polyclonal antibodies specific for each Sts protein, immobilized on protein A-Sepharose beads, and assayed for phosphatase activity with 0.5 mM OMFP at 37°C for 75 min (Sts-2) or 10 min (Sts-1). Values obtained from precipitation reactions in which no specific antibodies were added served as blank values and were subtracted from values obtained in the experimental samples. Assays were conducted in triplicate, and representative protein levels for each reaction were identified with anti-Sts-1 and anti-Sts-2 antibodies (right).
Sts-2 PGM toward two model substrates. This suggests that Glu-481, Ser-552, and Ser-582 might be key specificity determinants that regulate the interaction of Sts-2 PGM with its substrate(s).

DISCUSSION
Our results firmly establish the distinct role of Sts-2 as a negative regulator of TCR signaling, acting through a mechanism that involves dephosphorylation of key signaling components such as Zap-70. Because of the extraordinarily poor catalytic efficiency of Sts-2 PGM toward all previously tested artificial substrates, it has long been unclear whether Sts-2 could act as an intracellular phosphatase. However, that Sts-2 is a catalytically active phosphatase is placed on firm ground by the clear display of Sts-2 PGM enzyme activity toward OMFP and FDP. At the present time, it is not clear why Sts-2 PGM prefers these two latter molecules over pNPP. It has been noted that OMF has a substantially lower pK a than p-nitrophenol, making OMFP generally a more reactive substrate (27). It is also possible that the relatively larger size and orthogonal planes of the fluorescein-based compounds ensure a more energetically favorable orientation of the phosphate moiety within the Sts-2 PGM active site than occurs with the smaller phenyl ring of pNPP. If so, it might suggest that substrate conformation(s) is more critical for Sts-2 PGM than for the more promiscuous Sts-1 PGM . Whether this indicates a more limited subset of in vivo substrates for Sts-2 than Sts-1 is currently unclear. Nonethe-less, by highlighting the phosphatase activity associated with the Sts-2 C-terminal PGM domain, our results suggest a mechanistic basis for how the Sts homologues cooperate in a similar fashion to help set the threshold for TCR-induced T cell activation. Given the shared domain structure of Sts-1 and Sts-2, a mechanism of action for both proteins is postulated to involve PGM domain phosphatase effector activity acting in concert with the unique protein binding properties of their respective UBA and SH3 domains. Therefore, we propose that a catalytically active PGM phosphatase domain is necessary for the ability of Sts-2 to act as an intracellular phosphatase that regulates both Zap-70 tyrosine phosphorylation and the level of phosphorylation within the population of TCR-induced tyrosinephosphorylated, ubiquitinated proteins. Whether Sts-2 phosphatase activity targets Zap-70 directly or indirectly is currently unclear, given our inability to observe Sts-2 PGM -mediated dephosphorylation of Zap-70 in vitro (supplemental Fig. S2).
Although our previous reports identified the critical T cell tyrosine kinase Zap-70 as a potential downstream target of Sts activity (7,11), the individual contributions of each Sts protein to the regulation of Zap-70 had not been established. Our current data indicate that Sts-2 has a distinct regulatory role, a finding that is underscored by the observation that the combined loss of both Sts proteins (Sts-1/2 Ϫ/Ϫ ) has a greater effect on the hyper-activation of signaling pathways than the loss of FIGURE 7. Nonconserved residues adjacent to the active site contribute to differences between Sts-2 PGM and Sts-1 PGM catalytic activity. A, ribbon and surface representation of the Sts-2 PGM dimer with a zoom in on the active site. The four invariant residues (two His and two Arg) of the catalytic core are shown in black. The nonconserved residues between Sts-2 PGM and Sts-1 PGM (Glu-481, Ser-552, and Ser-582) are labeled and shown in red. These residues are conserved in Sts-1 as Val-495, Ala-566, and Tyr-596, respectively. Figure was prepared with PyMol. B, Sts-2 PGM VAY triple mutant was generated by site-directed mutagenesis, altering Sts-2 PGM residues (Glu-481, Ser-552, and Ser-582) for the corresponding residues found in Sts-1 PGM (Val, Ala, and Tyr, respectively). Phosphate hydrolysis activities of recombinant Sts-2 PGM VAY, wild-type Sts-2 PGM , and wild-type Sts-1 PGM toward OMFP were compared. The reactions contained 0.5 mM OMFP and 100 nM Sts-2 PGM , 100 nM Sts-2 PGM VAY, or 10 nM Sts-1 PGM and were performed in triplicate.
Sts-1 alone. Interestingly, the effects of Sts-2 deficiency are not always readily apparent in cells that express Sts-1, suggesting that Sts-1 can to a limited extent compensate for the loss of Sts-2. Thus, the present data illustrate that Sts-1 and Sts-2 each target Zap-70 in separate but overlapping fashions.
Judging by the extent of Zap-70 hyper-phosphorylation in Sts-1 Ϫ/Ϫ and Sts-2 Ϫ/Ϫ T cells, the in vivo phosphatase activity of Sts-2 is less than Sts-1. Additionally, the population of proteins that are dually modified by tyrosine phosphorylation and ubiquitination following TCR stimulation that have been identified as likely Sts targets also appear to be targeted by Sts-1 and Sts-2 in separate but overlapping manners, with Sts-2 also appearing to be less efficient in vivo. It is noteworthy that in vitro the two PGM domains exhibit markedly different enzymatic activity, with weak Sts-2 PGM phosphatase activity standing in stark contrast to the potent activity of Sts-1 PGM . The direct correlation between the in vitro catalytic efficiency of each PGM phosphatase domain and the phosphorylation levels observed in mutant T cells implies a clear functional relevance for the disparate enzymatic activities as measured in vitro associated with each Sts protein. However, it is important to keep in mind that the manner in which the Sts-1 and Sts-2 interact with their intracellular substrate(s) is not fully established. Therefore, it is possible that the N-terminal UBA and SH3 protein interaction domains of each Sts protein contribute to their unique in vivo functional properties.
To date, evidence for the role of Sts-2 as an intracellular phosphatase with negative regulatory functions has been limited. Initially, the apparent absence of Sts-2 PGM activity raised the possibility that Sts-2 might function not as a phosphatase but rather as a naturally occurring, T cell-specific dominant inhibitor of Sts-1 phosphatase functions (17). Other proposed cellular functions of Sts-2 are thought to be mediated by domains other than the PGM domain. Instead, these functions appear to require the nonenzymatic UBA and SH3 domains. For example, a role for inhibiting tyrosine kinase receptor endocytosis by preventing endocytic adaptor proteins access to ubiquitinated receptor complexes is thought to be mediated both by the UBA domain interacting with mono-ubiquitinated receptor and the SH3 domain interacting with the E3 ubiquitin ligase/adaptor protein, Cbl (29). In addition, a role promoting caspase-independent apoptosis via interaction with apoptosis-inducing factor was demonstrated to involve primarily the UBA and SH3 domains together (30).
Aside from this study, there is additional compelling evidence in support of a role for Sts-2 as an intracellular phosphatase. This evidence includes the remarkably close structural similarity between Sts-2 PGM and the strongly active phosphatase Sts-1 PGM (11,12), as well as the evolutionary conservation of all critical catalytic residues in all Sts-2 homologues. It is worth noting that several members of the class I cysteine-based protein-tyrosine phosphatase family of phosphatases are functionally similar to Sts-2 in that they have little or no measurable in vitro phosphatase activity on peptide substrates yet are thought to act in vivo as tyrosine phosphatases (31)(32)(33). Given the interesting difference between the in vivo and in vitro observations, we speculate that Sts-2 PGM enzymatic activity is regu-lated in vivo by a mechanism that is currently unclear. Unlike protein-tyrosine phosphatases, which have a prominent structural element known as the WPD loop whose movements adjacent to the active site affect catalytic activity (34), to date the PGM domain of Sts-2 appears to have no intrinsic structural component that regulates its enzymatic activity. It is tempting to speculate that the in vivo mechanism of regulation of Sts-2 PGM phosphatase activity is a specific property of T cells, as these are the cells within which Sts-2 is almost exclusively expressed. Addressing the unique requirement for Sts-2 expression in T cells and how Sts-2 PGM phosphatase activity is regulated in the context of its ability to negatively regulate TCR signaling is a focus of ongoing research.