|Home | About | Journals | Submit | Contact Us | Français|
Precise signaling by the T cell receptor (TCR) is crucial for a proper immune response. To ensure that T cells respond appropriately to antigenic stimuli, TCR signaling pathways are subject to multiple levels of regulation. Sts-1 negatively regulates signaling pathways downstream of the TCR by an unknown mechanism(s). Here, we demonstrate that Sts-1 is a phosphatase that can target the tyrosine kinase Zap-70 among other proteins. The x-ray structure of the Sts-1 C-terminus reveals that it has homology to members of the phosphoglycerate mutase/acid phosphatase (PGM/AcP) family of enzymes, with residues known to be important for PGM/AcP catalytic activity conserved in nature and position in Sts-1. Point mutations that impair Sts-1 phosphatase activity in vitro also impair the ability of Sts-1 to regulate TCR signaling in T cells. These observations reveal a PGM/AcP-like enzyme activity involved in the control of antigen receptor signaling.
T cells detect foreign pathogens by recognition via the T cell receptor (TCR). Engagement of the TCR leads to T cell activation, and in recent years many of the signaling molecules that participate in this process have been identified and characterized (Singer and Koretzky, 2002). Less is known about the mechanisms that negatively regulate T cell activation. Two members of the Suppressor of TCR signaling family of proteins, Sts-1 and Sts-2, were recently described to play a role in negatively regulating signaling pathways downstream of the TCR (Carpino et al., 2004). Sts-1 was originally discovered in a screen for proteins that bind to a specific tyrosine phosphorylated peptide derived from the tyrosine kinase Jak2 (Carpino et al., 2002). Sts-1 and Jak2 were subsequently shown to interact, although the role of Sts-1 in Jak2 signaling has not been established. The Sts proteins were also identified by virtue of their ability to bind to the ubiquitin ligase Cbl (Kowanetz et al, 2004; Feshchenko et al, 2004).
The realization that Sts-1 and -2 have roles in regulating signaling pathways downstream of the T cell receptor emerged from an analysis of T cells derived from mice genetically engineered to lack the two proteins (Carpino et al., 2004). In particular, naïve Sts-1/2-/- T cells are hyper-sensitive to T cell receptor stimulation, exhibiting a pronounced increase in TCR-induced proliferation relative to wild-type cells. This phenotype is accompanied by a marked increase in cytokine production by Sts-1/2-/- T cells and significantly increased susceptibility of double knock-out mice to autoimmunity in a mouse model of multiple sclerosis. In addition, Sts-1/2-/- naïve T cells display enhanced phosphorylation and activation of Zap-70, a tyrosine kinase that plays an important role in relaying signals from the TCR. This latter finding suggests that the Sts proteins control the level of Zap-70 activation during T cell receptor engagement. However, while these studies identified a role for the Sts proteins in regulating signaling pathways downstream of the TCR, they did not address their intracellular targets or mechanisms of action.
Sts-1 and -2 are approximately 40% identical to one another, with both proteins having distinct expression patterns: Sts-1 appears to be ubiquitously expressed while Sts-2 is preferentially expressed in cells and tissues of the hematopoietic system (Wattenhofer et al., 2001). Both proteins are characterized by a unique multi-domain structure, with an N-terminal Ubiquitin-association (UBA) domain and a central Src-homology 3 (SH3) domain. Both UBA and SH3 domains are protein-protein interaction domains, the former binding mono- and poly-ubiquitin and the latter interacting with proline rich sequences (Hicke et al., 2005; Musacchio, 2002). In addition, the Sts C-terminus has passing resemblance to members of the phosphoglycerate mutase (PGM) family of enzymes (Carpino et al., 2002). PGM, an evolutionarily conserved enzyme that plays a critical role in glycolysis, lends its name to a family of enzymes that share a common structural fold and the ability to act as phosphatases and/or phospho-transferases. PGM family members also share a clear biochemical and structural relationship to members of the acid phosphatase (AcP) family of enzymes (Jedrzejas, 2000).
In this report, we identify and characterize an intrinsic phosphatase activity associated with the C-terminal PGM-like domain of Sts-1. In addition, we present the crystal structure of the C-terminal domain and compare its structure to that of phosphoglycerate mutase. Finally, we demonstrate that Sts-1PGM catalytic activity contributes to the ability of Sts-1 to negatively regulate signaling pathways downstream of the T cell receptor. Altogether, these results provide an important clue into the mechanism of action of Sts-1.
Although members of the PGM/AcP families of enzymes differ considerably in primary sequence and substrate specificity, they have several important features in common. These include a similar α/β core structure that consists of an inner parallel β sheet surrounded by a series of alpha helices, a catalytic core formed in part by a constellation of conserved active site residues, and a hypothesized two-step catalytic mechanism that involves the use of two conserved histidine residues (Bazan et al., 1989; Jedrzejas, 2000). The amino acid sequence Arg-His-Gly-Glu (‘RHGE’) has been identified as a signature pattern that correlates with membership in the PGM family, while the related sequence Arg-His-Gly/Asn (‘RHG/N’) correlates with membership in the AcP family (Vincent et al., 1992). The Arg-His residues within these motifs have been shown to play a significant role in catalysis. Structural and mutagenesis studies have also identified other conserved residues that are important for catalysis, including an additional conserved arginine and histidine (Jedrzejas, 2000). These two latter residues, along with the Arg-His of the signature motif, cluster together within the PGM/AcP active site.
The presence of the amino acid sequence 379RHGE382 within the C-terminal portion of Sts-1 suggests that this region of Sts-1 is evolutionary related to PGM/AcP enzymes. To investigate the relationship between Sts-1 and PGM/AcP family members, we evaluated the alignment of the Sts-1 C-terminus with a variety of enzymes belonging to the PGM/AcP super-family. The overall identity between Sts-1 and PGM/AcP enzymes is low, and there are large stretches of amino acids within all family members that do not align. Notably, however, the quartet of two arginines and two histidines that are known to be located in the active sites of all PGM/AcP enzymes align with identical residues within Sts-1 (Figure 1A). Thus, Sts-1 has four residues distributed within its primary sequence that suggest an evolutionary relationship to PGM/AcP family members. These same four residues are conserved in Sts-2 and all known Sts orthologs (data not shown).
Given that the C-terminal PGM domain of Sts-1 has primary sequence homology to enzymes that possess phosphatase activity, we determined if it had phosphatase activity. Sts-1PGM (Sts-1 residues Gly-369 – Glu-638) was expressed in E. coli as a 6 × His – tagged protein, purified to homogeneity (Figure S1), and tested for the ability to hydrolyze the phosphatase substrate para-nitrophenylphosphate, pNPP. A time course of Sts-1PGM (25 nM) – catalyzed pNPP hydrolysis at two different initial substrate concentrations is illustrated in Figure S1 These results demonstrate that recombinant Sts-1PGM is able to dephosphorylate pNPP with kinetics that resemble a Michaelis-Menten enzyme – catalyzed reaction, suggesting that Sts-1PGM has an intrinsic phosphatase activity. To gain further insight into the catalytic properties of Sts-1PGM, we performed the reaction at different substrate concentrations (50 μM – 20 mM). The initial reaction velocities at each substrate concentration were calculated and plotted as a function of substrate concentration (Figure 1B), and Lineweaver-Burke analysis was used to calculate reaction parameters. The Km value for pNPP of Sts-1PGM was determined to be approximately 1 mM, with a maximum velocity Vmax of 270 nmoles/min (turnover = 120 s-1). By comparison, the reported Km for pNPP of a recently characterized bacterial PGM is 3 mM, while that of human prostatic acid phosphatase is 0.7 mM (Rigden et al., 2001; Kilsheimer and Axelrod, 1957). Identical Sts-1PGM enzyme activity was observed in the presence or absence of a variety of metal ions, suggesting that the reaction does not require a metal ion co-factor (data not shown). Metal ion - independent catalysis is a property common to all PGM/AcP family members (Vincent et al., 1992).
To examine whether full-length Sts-1 has a similar phosphatase activity as recombinant Sts-1PGM, 293T cells were transfected with either an empty expression vector or an expression vector designed to express Flag-tagged Sts-1. Then, cell lysates were prepared and proteins were immuno-precipitated with anti-Flag monoclonal antibodies. Anti-Flag antibodies precipitated a pNPP phosphatase activity from Sts-1-transfected cells that was absent from cells transfected with control vector (Figure 1C, left). To test whether endogenous Sts-1 has a similar catalytic activity, it was immuno-affinity purified from primary murine splenocytes and tested for phosphatase activity. Similar to transfected Sts-1, endogenous Sts-1 demonstrated pNPP phosphatase activity that was not present in control antibody precipitates (Figure 1C, right). In contrast to Sts-1, immunoaffinity purified Sts-2 displayed substantially reduced in vitro phosphatase activity towards pNPP. Specifically, when both Sts-1 and Sts-2 were precipitated from Jurkat cells and tested for pNPP phosphatase activity, Sts-1-mediated phosphate hydrolysis was evident within 10 minutes, while Sts-2 displayed an activity after a 3 hour reaction that was only modestly above the activity in control precipitates (Figure 1D). Recombinant Sts-2PGM displayed in vitro pNPP phosphatase activity, although it also was significantly reduced relative to Sts-1PGM activity (Figure 1E). These results demonstrate that both Sts-1 and Sts-2 have intrinsic in vitro pNPP hydrolytic activity, although that of Sts-1PGM is greater than Sts-2 PGM.
Given that Sts-1PGM was able to hydrolyze a small molecule mimic of phosphotyrosine, we next examined whether it was able to dephosphorylate phosphorylated tyrosine. Sts-1PGM dephosphorylated a pTyr-containing peptide but was not able to dephosphorylate peptides containing pSer or pThr (Figure 2A). To test whether Sts-1PGM could dephosphorylate a tyrosine phosphorylated protein, an activated form of the Src tyrosine kinase that is constitutively tyrosine phosphorylated was expressed in cells, isolated by immunoprecipitation, and used as a substrate in a Sts-1PGM-mediated dephosphorylation reaction. Sts-1PGM de-phosphorylated tyrosine phosphorylated Src in vitro in a dose dependent manner, with an activity that was roughly one order of magnitude less than PTP1B. In contrast, a mutant of Sts-1PGM lacking one of the potential active site histidines displayed no phosphatase activity in our in vitro assay (Figure 2B). To further investigate the protein tyrosine phosphatase activity of Sts-1, the ability of Sts-1 to suppress intracellular Src-induced tyrosine phosphorylation was examined. Co-expression of Sts-1 with constitutively active Src in 293T cells resulted in significantly reduced protein tyrosine phosphorylation, while an inactive Sts-1 mutant was unable to suppress Src-induced intracellular phosphorylation (Figure 2C). Because the recruitment of Sts-1 into the activated EGF receptor complex has been reported (Kowanetz et al., 2004), we also assessed the ability of Sts-1 to mediate dephosphorylation of the EGFR. Similar to the effects of Sts-1 on Src-induced tyrosine phosphorylation, expression of Sts-1 was able to suppress EGF receptor tyrosine phosphorylation (see Figure 2D).
Collectively, our data demonstrate that the Sts proteins possess an intrinsic phosphatase activity localized to their C-terminal domains. Although most PGM/AcP family members are known to target small molecule substrates rather than tyrosine phosphorylated proteins, our results are consistent with the finding that another family member, prostatic acid phosphatase (PAcP), can dephosphorylate tyrosine phosphorylated proteins. Indeed, the latter enzyme has been demonstrated to target the receptor tyrosine kinase c-ErbB-2 oncoprotein, with decreased expression of hPAcP in human prostate cancer cells correlating with increased cellular tyrosine phosphorylation and increased tumorigenicity (Meng and Lin, 1998; Lin et al., 2001). Interestingly, in our in vitro assay utilizing pNPP as a substrate, Sts-2 is significantly less active than Sts-1, despite the presence of similar catalytic residues within its PGM domain. This observation suggests that Sts-1 and Sts-2 are functionally distinct. However, the need to delete both proteins from T cells in order to observe a significant physiological phenotype also suggests that they have overlapping functions. Precedence for homologous but functionally distinct enzymes having both independent and overlapping functions within T cells can be found within several other enzyme families, including tyrosine kinases (Zap-70 and Syk, Cheng et al, 1997) and ubiquitin ligases (c-Cbl and Cbl-b, Naramura et al, 2002).
To gain insight into the catalytic activity associated with Sts-1PGM, we solved its three-dimensional structure by X-ray crystallography. The crystal structure was solved by the single wavelength anomalous dispersion (SAD) method using the anomalous signal present in Se-Met Sts-1PGM crystals at the peak wavelength data (Dauter et al., 2002). The model was built in an electron density map calculated to 2.1 Å resolution and refined against native data collected to 1.82 Å resolution to crystallographic residuals Rcryst/Rfree of 20.3/24.2%. The final model has good stereochemistry with only Ala-564 of all three chains in generously allowed (, ψ) values despite clear electron density (see Materials and Methods). Statistics on data collection were reported elsewhere (Kleinman et al., 2006). Table I summarizes data refinement statistics, with additional discussion on the structure solution found in the Supplemental Materials. In the unit cell, Sts-1PGM exists as a dimer (Figure 3). This result is consistent with size exclusion data indicating that Sts-1PGM is a dimer in solution (Kleinman et al., 2006) and in vivo analysis indicating that Sts dimerizes via its PGM domain (Kowanetz et al., 2004). Interestingly, the Sts-1PGM mode of dimerization differs considerably from that of the closely related E. coli PGM (Supplemental Materials, Figure S2).
The structure of Sts-1PGM complements our enzymatic analysis, and confirms that Sts-1PGM belongs to the PGM/AcP super-family of enzymes. In particular, the Sts-1PGM monomer has an α/β structure similar to that adopted by all members of the super-family (Jedrzejas, 2000). Architecturally, a central 7-stranded β sheet forms the core of the molecule (Figure 3). It is surrounded by 8 α-helices (α1 to α8), with a C-terminal β-strand located outside the protein core that makes strong interactions with a neighboring molecule. All β-strands are parallel with the exception of β6. The 379RHGE382 signature pattern is located at the carboxyl end of the central β sheet structure, and the two arginines and two histidines of Sts-1 that are homologous to PGM/AcP catalytic residues extend into a semi-barrel – shaped depression (17 Å in length by 8 Å in diameter) present at this location on the surface of Sts-1PGM. Based on overall structural homology to PGM/AcP enzymes, this cavity is likely the Sts-1PGM catalytic pocket.
The closest known structural homologue to the Sts-1PGM monomer is E. coli phosphoglycerate mutase (ecPGM), with a calculated RMS deviation after superposition of 156 Cαs of 1.98 Å (for primary sequence alignment, see Figure S3). However, despite their structural similarities, the two enzymes display some notable topological differences. In particular, three regions that shape the catalytic pocket deviate significantly between Sts-1PGM and ecPGM. The regions termed Inserts 1 and 2 (red and green in Figure 4A) contribute residues that line the catalytic pocket, and in Sts-1PGM they extend outward and angle away from the active site to a greater degree than in ecPGM. Additionally, the Sts-1PGM C-terminus (green) forms an extended loop that protrudes away from the central core of the enzyme, while the C-terminus of ecPGM folds over the active site. These unique features of Sts-1PGM are likely to have important functional implications with regard to its interaction with substrate (see below).
The structural similarity between Sts-1PGM and ecPGM extends to the active site residues within the catalytic pocket. For example, in ecPGM, His-10 has been identified as the nucleophile during the ecPGM-catalyzed reaction, and four other residues (Arg-9, Arg-61, Glu-88, and His-183) are all thought to help stabilize the negatively charged phosphate that is being hydrolyzed (Bond et al., 2001). Identical residues within Sts-1 (Arg-379, His-380, Arg-462, Glu-490, and His-565) adopt an identical configuration (Figure 4B), suggesting that they are involved in the dephosphorylation reaction. To ascertain whether the shallow cavity surrounding these residues is the active site of Sts-1PGM, a Sts-1PGM crystal was soaked in a phosphate buffer and the structure of the complex at 2.6 Å resolution was determined (for statistics on data collection, see Table S1) . A clear density in which a phosphate molecule was built was evident in the difference Fourier electron density map. As shown in Figure 4C, the side chains of Arg-379, His-380, Arg-383, Arg-462, Glu-490, His-565, and the main chain amino group of Ala-566 all stabilize the phosphate moiety. The phosphate molecule replaces three water molecules that were found close to His-380 in the phosphate-free structure without disturbing the overall structure of the cavity (Figure S4). Thus, this solvent exposed cavity can stabilize a phosphate group and is likely the active site of Sts-1PGM.
To further confirm that Arg-379, His-380, Arg-462, and His-565 play a role in Sts-1PGM catalytic activity, we mutated each of these residue to alanine and assessed the catalytic activities of the resulting mutants. Each mutation led to reduced in vitro catalytic activity of Sts-1PGM, in the context of both full-length Sts-1 (Figure 4D and recombinant Sts-1PGM (Figure S5). Importantly, mutation of the Sts-1PGM active site residues did not alter the stability of Sts-1, nor did it alter the ability of Sts-1 PGM to dimerize (immunoblot analysis and gel filtration analysis respectively, data not shown). Combined, our structural and biochemical data are consistent with a model in which the cluster of basic residues located at the carboxyl end of the central β sheet structure form the active site of Sts-1.
The active site cleft of Sts-1PGM is considerably broader than that of ecPGM, and its catalytic pocket is more solvent exposed and accessible than that of the latter enzyme (Figure 4E). To some extent, this may account for the ability of Sts-1PGM to accommodate large macromolecular substrates such as tyrosine phosphorylated proteins. As a whole, PGM/AcP enzymes are known to differ widely in their ability to recognize and bind diverse substrates, with some enzymes having limited and well-defined substrate specificity and others interacting with an array of different substrates. For example, the only known substrate of Fru-2,6-Pase is Fru-2,6-P2, with a reported Km of 4 nM (Lee et al., 1996). In contrast, PGM utilizes 3-phosphoglycerate, 2-phosphoglycerate, and 2, 3-diphosphoglycerate as substrates, with respective Km's of 100-300 μM, <100 μM, and 0.5-0.8 μM (Winn et al., 1981). One of the most promiscuous of the acid phosphatases appears to be PAcP, with an ability to dephosphorylate a wide range of substrates in vitro including ATP, ADP, pyrophosphate, and large macromolecules such as phosphopeptides and phosphoproteins (Jakob et al., 2000). The open configuration of the Sts-1PGM catalytic pocket suggests that it might have a wide promiscuity with regard to different substrates. Clearly, Sts-1 can dephosphorylate a wide variety of tyrosine phosphorylated peptides and proteins (Figures 2A, 2B, and and5B5B below, and data not shown). However, it should also be noted that when a variety of small molecules were tested as potential Sts-1PGM substrates, none of the compounds tested were dephosphorylated in vitro by Sts-1PGM (see Figure S6). This latter result suggests that while Sts-1 can act as a broadly specific protein tyrosine phosphatase, it does not have the broad substrate specificity characteristic of some AcP enzymes.
The substrate specificity of enzymes can also be greatly influenced by protein interaction domains found within the same polypeptide. Such interaction domains can either increase the local concentration of potential substrates or target the enzyme to specific subcellular compartments. In addition, they can regulate enzyme activity. For example, a wide variety of domains that are thought to influence catalytic activity, sub-cellular localization and substrate specificity are found within different protein tyrosine phosphatases (Alonso et al., 2004). Sts-1 is unique among the PGM/AcP enzymes in possessing modular protein interaction domains. Specifically, a ubiquitin association (UBA) motif and an SH3 domain are juxtaposed along with the Sts-1 PGM domain in the context of full-length Sts-1. UBA domains have been shown to bind directly to ubiquitin, thereby allowing them to interact with proteins that are ubiquitinated, and SH3 domains are known to bind proline rich sequences. While these protein interaction modules are thought to regulate the functions of many signaling proteins, their exact role in Sts-1 function is unknown. They do not appear to play a role in the intra-molecular regulation of Sts-1 phosphatase activity, as mutations that inactivate the UBA and SH3 domains respectively do not alter catalytic activity (Figure 4F). However, they may be involved in the subcellular localization of Sts-1, or the interaction of Sts-1 with other protein binding partners, thereby influencing Sts-1 substrate specificity.
Having established the existence of Sts-1PGM phosphatase activity, we next sought to determine its functional significance. The Sts proteins have been implicated in the negative regulation of T cell receptor signaling (Carpino et al., 2004). Indeed, the hyper-phosphorylation and hyper-activation of Zap-70, a tyrosine kinase that plays a critical signaling role downstream of the T cell receptor, is a prominent biochemical defect within Sts-1/2 - deficient naïve T cells. This latter phenomenon suggests that Sts-1 might directly regulate key signaling proteins downstream of the TCR. To determine if Sts-1 could target Zap-70, phosphorylated Zap-70 was isolated by immunoprecipitation and utilized as a substrate in a Sts-1PGM-mediated dephosphorylation reaction. Sts-1PGM was able to dephosphorylate Zap-70 in a dose-dependent manner (Figure 5A). Sts-1PGM was also able to dephosphorylate a wide spectrum of proteins that become tyrosine phosphorylated following TCR stimulation (Figure 5B), an observation that strengthens the notion that the Sts-1 catalytic pocket can accommodate a wide variety of different tyrosine phosphorylated proteins. The size and shape of the Sts-1PGM active site cleft is consistent with this possibility.
The effect of over-expressing wild-type Sts-1 vs. catalytically-inactive Sts-1 in T cells was then assessed. To generate a Sts-1 mutant in which phosphatase activity was completely abrogated, both active site histidines were mutated to alanine (Sts-1H380,565A, see Figure S7). Expression of wild-type Sts-1 in Jurkat cells resulted in significantly decreased intracellular tyrosine phosphorylation following TCR stimulation, as compared to expression of vector control. In contrast, expression of Sts-1H380,565A failed to reduce intracellular protein tyrosine phosphorylation (Figure 5C). This result suggests that Sts-1 might negatively regulate TCR signaling pathways by targeting TCR substrates such as Zap-70 for dephosphorylation.
To further investigate the role of Sts-1 catalytic activity in regulating signaling pathways downstream of the TCR, a retroviral reconstitution assay was developed in which Sts-1 cDNA is expressed in primary T cells that lack the Sts proteins. Sts-1/2-/- activated T cells are hyper-responsive to T cell receptor stimulation, and reconstitution of mutant T cells with Sts-1 cDNA significantly reduces the proliferative response (Figure 6A). Therefore, we compared the proliferative responses of Sts-1/2-/- T cells infected with retrovirus that expressed either wild-type Sts-1 or mutant forms of Sts-1 that lack phosphatase activity. As illustrated previously (Figure 4D), point mutation of the arginines and histidines within the Sts-1 active site reduces Sts-1 in vitro phosphatase activity. Figure 6B illustrates that reconstitution of Sts-1/2-/- T cells with three different catalytically impaired isoforms of Sts-1 have a proliferative response that is significantly greater than T cells reconstituted with wild-type Sts-1. These results support the hypothesis that Sts-1 phosphatase activity plays a role in the ability of Sts-1 to suppress TCR signaling.
At the present time, it is not clear how the Sts-1 UBA and SH3 domains cooperate with the C-terminal phosphatase domain to negatively regulate TCR signaling. That they play functionally important roles is suggested by the observation that UBA and SH3 domain – inactivating mutations impair Sts-1's ability to reduce the hyper-proliferative response of Sts-1/2-/- T cells (Figure 6 C&D). As alluded to above, the UBA and SH3 domains may be involved in the intracellular localization of Sts-1 or they may contribute to Sts-1 substrate specificity. Another possibility is that they contribute to Sts-1 function in a manner that is independent of Sts phosphatase activity. Given the apparent differences in catalytic activity between Sts-1 and Sts-2, and the need to functionally eliminate both family members in order to observe a dramatic effect on T cells, this latter hypothesis is also attractive.
Collectively, our results demonstrate an intrinsic phosphatase activity associated with the Sts C-terminal PGM domain, and indicate that Sts might function as a protein tyrosine phosphatase. In particular, our results implicate Sts-1PGM catalytic activity in the negative regulation of signaling pathways downstream of the T cell receptor. As such, the Sts proteins are the only known members of the PGM/AcP family to antagonize TCR signaling. It is well known that reversible protein tyrosine phosphorylation plays a key role in TCR signaling, and T cells are known to express more than an average number of protein tyrosine kinases and phosphatases (Mustelin et al., 2005). Interestingly, while most known PTPs contain modular binding domains as part of their structure, none are known to possess UBA or SH3 domains (Alonso et al., 2004). Thus, the juxtaposition of these two unique domains with protein tyrosine phosphatase catalytic activity in the context of full-length Sts-1 suggests that Sts-1 operates within an intracellular signaling niche that is separate from other PTPs. Currently, we favor a model in which the combined functions of the Sts proteins and members of the PTP family act to negatively regulate TCR signaling pathways by dephosphorylating key signaling molecules. In this model, Sts-1 is recruited to its substrates via SH3 and UBA domain interactions. For example, both the ubiquitination of the activated TCR complex (Wiedemann et al., 2005) and the recruitment of Cbl to the TCR complex might promote the interaction of Sts-1 with substrates that either are found within the activated TCR complex or are themselves targets of proximal TCR signaling pathways.
To fully understand the role of Sts-1 in modulating TCR signaling pathways, it will be important both to delineate the manner in which Sts-1 interacts with its in vivo substrates and determine how its catalytic activity is regulated. In addition, the manner in which Sts-1 cooperates with Sts-2 to negatively regulate T cell activation is an avenue of continuing investigation. The data presented herein, and by others, suggest that although they are homologues, Sts-1 and -2 each have distinct biological properties. For example, the binding properties of the two Sts UBA domains are significantly different (Hoeller et al, 2006). Additionally, the substantially reduced ability of Sts-2 relative to Sts-1 to hydrolyze pNPP in vitro despite the conservation of key catalytic residues suggests that Sts-1 and -2 could each act on a unique set of substrates as well as synergize in dephosphorylating a common subset of intracellular substrates. As evidenced by the need to delete both proteins from T cells in order to observe a significant physiological phenotype, this latter subset of common substrates is hypothesized to be involved in setting the threshold of TCR-induced T cell activation. Finally, because various auto-immune pathologies are associated with aberrant T cell activation, it is important to note that the Sts catalytic domain might be an attractive target for therapies designed to modulate the dynamics of TCR signaling.
293T cells were cultured in DMEM (Mediatech) supplemented with 10% FCS, 2 mM glutamine and penicillin (50 IU/ml)/streptomycin (50 μg/ml). Jurkat cells were cultured in RPMI (Mediatech) supplemented identically. Sts-1/2-/- mice have been described (Carpino et al., 2004). Animal work was conducted under guidelines of the Stony Brook Institutional Animal Care and Use Committee (IACUC). All Sts-1 mutants were generated by PCR and sequenced to confirm absence of additional mutations. Constitutively active Src cDNA was generated by replacing sequence coding for the last 15 amino acids of c-Src with sequence coding for the Flag epitope. EGF was purchased from Invitrogen, and human anti-CD3 (UCHT1) was purchased from Ancell Corp. (Bayport, MN). Primary antibodies: anti-Sts-1 (Carpino et al., 2002), anti-Flag M2 (Sigma), and anti-phosphotyrosine 4G10 (Upstate Biotechnology); secondary antibodies: Alexa Fluor® 680-conjugated goat anti-mouse (Molecular Probes) and IRDye800-conjugated goat anti-rabbit secondary antibodies (Rockland).
Exogenous DNA was introduced into 293T cells by calcium-phosphate transfection and into Jurkat T cells by retroviral transduction. For the latter, a GFP-blasticidin fusion protein downstream of an IRES element within the retroviral vector (Ueki and Hayman, 2003) allowed for flow cytometric sorting of infected cells and selection in media containing blasticidin (20 μg/ml). For precipitations, cells were lysed in cold lysis buffer (Carpino et al, 2004). Lysates were clarified by centrifugation, rotated at 4°C with specific antibody and 20 μl Protein A Sepharose (PAS) 50% slurry for 1-2 hr. Beads were either tested for phosphatase activity (see below) or boiled in Laemmli sample buffer, following which bound proteins were separated by SDS-PAGE. Proteins were then transferred to nitrocellulose by semi-dry transfer (BioRad), membranes were blocked with 3% BSA in TBS (Tris buffered saline, pH 8.0), incubated at 4°C overnight with specific antibodies, washed with TBS, incubated with the appropriate secondary antibody and developed with the ODYSSEY Infrared Imaging System (LI-COR).
25 nM recombinant Sts-1PGM ([Sts-2PGM] = 250 nM) was incubated with pNPP (Sigma) in reaction buffer (25 mM HEPES, pH 7.2, 50 mM NaCl, 5 mM DTT, 2.5 mM EDTA at 37°C. Aliquots were removed, quenched in stop buffer (13% potassium phosphate), diluted with TE, and the OD405nm was measured. Illustrated results are representative of multiple assays. Reaction velocities during the initial linear phase of the reaction were utilitzed for Lineweaver-Burke analysis. For other small molecule substrates, phosphate hydrolysis was detected by malachite green. To detect phosphatase activity of precipitated proteins, PAS beads were washed three times with cold lysis buffer and twice with reaction buffer. pNPP was added at the indicated concentrations, and the reaction proceeded at 37°C in a final volume of 100 μl for 10-20 min. Supernatants were quenched with stop buffer, diluted, and the OD405nm was measured. Following the reactions, levels of bound Sts-1 were evaluated by SDS-PAGE/immunoblotting. Phosphorylated Src bound to PAS beads was also utilized as a substrate, as were total tyrosine phosphorylated proteins precipitated from stimulated Jurkat T cells that were eluted from the PAS by incubation with phenyl phosphate for 15 min at 4°C.
See Supplemental Materials. The coordinates of Sts-1PGM alone or in complex with phosphate are deposited in the Protein Data Bank with accession numbers 2H0Q and 2IKQ, respectively.
T cells were obtained from wild-type and Sts-1/2-/- mice. In brief, dissected spleens were crushed in PBS containing 2% FCS, red blood cells were lysed by addition of ACK lysis buffer (pH 7.2), and debris was removed by straining through a 70 uM filter (Becton Dickinson). Splenocytes were cultured for two days and then a bicistronic retroviral vector expressing GFP downstream of an IRES was used to express wild-type or mutant Sts-1 in activated T cells (Morrigl et al., 1999). T cells were spin-infected 2X in the presence of polybrene (8 μg/ml) and RetroNectin (Takara), allowed to grow 24 hours in the presence of IL2, sorted by flow cytometry, and an equal number of GFP+ cells were assayed for their ability to respond to platebound anti-CD3: 5 × 104 - 1 × 105 T cells were placed in individual wells of a round bottom 96 well plate that had been coated with anti-CD3e, cultured for 24 hours, labeled with 1 μCi/well [3H]-thymidine for 8-14 hr, lysed, and DNA was TCA-precipitated onto nitrocellulose filters. [3H]-Thy. incorporation was evaluated with a 1900 TR scintillation counter (Packard). Jurkat T cells were stimulated as described (Carpino et al., 2004).
The authors would like to thank Neena Carpino, Holly Kleinmann, Kevin Lau, and Malgosia Skowron for expert technical assistance. We also thank Jim Ihle and Jorge Benach for support, Nick Tonks for purified PTP1B enzyme, Todd Miller for pTyr peptides, and N. Ueki and M. Hayman for the EGFP-blasticidin cDNA. Special thanks to D. Bar-Sagi, N. Reich, and M. Hayman for helpful discussions and comments on the manuscript. This work was supported by Stony Brook University (N.C), The National Science Foundation (MCB-0316600, N.N), and The American Heart Association (0235522N, N.N). The National Synchrotron Light Source is supported by the Department of Energy and the National Institutes of Health, and beamline X26C is supported in part by Stony Brook University and the Research Foundation of New York. B. F. was supported by a Medical Scientist Training Grant.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.