SLP-76 Is a Direct Substrate of SHP-1 Recruited to Killer Cell Inhibitory Receptors*

Activation of immune system cells via antigen-, Fc-, or natural killer cell-triggering-receptor stimulation is aborted by co-engagement of inhibitory receptors. Negative signaling by killer cell inhibitory receptors and related receptors depends on the Src homology 2 (SH2)-containing protein tyrosine phosphatase SHP-1. Using a combination of direct binding and functional assays, we demonstrated that the SH2 domain-containing leukocyte protein 76 (SLP-76) is a specific target for dephosphorylation by SHP-1 in T cells and natural killer cells. Furthermore, we showed that tyrosine-phosphorylated SLP-76 is required for optimal activation of cytotoxic lymphocytes, suggesting that the targeted dephosphorylation of SLP-76 by SHP-1 is an important mechanism for the negative regulation of immune cell activation by inhibitory receptors.

Initiation, prevention, and termination of immune responses are governed by receptors that transduce signals from the exterior to the interior of hematopoietic cells. The proximal intracellular biochemical events initiated following stimulation of antigen, Fc-, and NK 1 cell triggering receptors include activation of Src and Syk family protein tyrosine kinases (PTKs), hydrolysis of membrane phospholipids, and mobilization of intracellular calcium (1)(2)(3)(4). This activation pathway is subject to negative regulation by inhibitory receptors. For example, class I major histocompatibility complex (MHC)-recognizing KIRs prevent NK cells and certain subpopulations of T cells from killing specific class I MHC-bearing target cells (5)(6)(7)(8).
KIRs belong to a large family of immune cell inhibitory receptors that contain a conserved immunoreceptor tyrosinebased inhibitory motif (ITIM) in their cytoplasmic signaling domains (9 -11). Tyrosine phosphorylation of the ITIM results in the recruitment of SH2 domain-containing phosphatases, such as the PTPs SHP-1 and SHP-2 or the SH2-containing inositol phosphatase SHIP (12)(13)(14)(15)(16). These two classes of phosphatases deliver distinct inhibitory signals. SH2-containing inositol phosphatase recruited to the inhibitory receptor Fc␥RIIB acts relatively late in the activation cascade by interfering with sustained calcium influx (17,18). In contrast, SHP-1 recruited to KIRs inhibits the earliest FcR-initiated signals, including tyrosine phosphorylation of the immunoreceptor tyrosine-based activation motif-containing -subunits of Fc␥RIIIA and the downstream effectors zeta-associated protein-70 (ZAP-70) and phospholipase C-␥ (12), and also inhibits the activation of p36 induced by natural killing (19). Due to the prominent role of protein tyrosine phosphorylation in early signal output from multisubunit immune recognition receptors, the field of candidate SHP-1 substrates is large. SHP-1-utilizing inhibitory receptors negatively regulate similar Src and Syk family PTK-dependent activation pathways in a variety of hematopoietic cell types (9 -11), however, suggesting that a common SHP-1 target may potentially exist. We therefore sought to identify a direct target of SHP-1 using KIR signaling as a model system. tor sequence. The 5Ј-oligonucleotide introduces an EcoRI restriction site and a glutamate to glutamine substitution at position 200 of SHP-1, outside the catalytic domain and at the junction with GST. The polymerase chain reaction-amplified product was ligated into the GST-encoding vector pGEX-KG. Generation and purification of the GST fusion protein were performed according to standard protocols. Full-length, agarose-bound SHP-1 was obtained from Upstate Biotechnology.
Cell Surface Phenotype Analysis, Cytotoxicity Assays, Vaccinia Virus Infection, Cell Stimulation, Immunoprecipitation, and Western Blotting-Protocols for analysis of cell surface phenotype, chromium release assays for cytotoxic activity, infection of cells with recombinant vaccinia viruses, stimulation of cells by antibody-mediated receptor cross-linking or with pervanadate, immunoprecipitation, and Western blotting have been previously published (12).
In Vitro Dephosphorylation Reactions-Following stimulation, cells were lysed in ice-cold lysis buffer containing 40 mM Hepes (pH 7.4), 100 mM NaCl, 1 mM EDTA, 5 mM iodoacetic acid (to inhibit cellular PTPs), 10% glycerol, 1% Triton X-100, and protease inhibitors. Following lysis, excess iodoacetic acid was quenched by the addition of 10 mM dithiothreitol. Full-length agarose-bound SHP-1 was added to cleared lysates in the presence or absence of the KIR ITIM-containing peptide (200 M) at 4°C. Following the dephosphorylation reaction, supernatants were separated by SDS-PAGE, transferred to membrane, and immunoblotted as indicated in the figure legends.
Far Western Blotting-Following stimulation, cells were lysed in buffer containing 10 mM Tris (pH 7.4), 50 mM NaCl, 5 mM EDTA, 50 mM NaF, 30 mM Na 4 P 2 O 7 , 1% Triton X-100, and protease and phosphatase inhibitors. Cleared lysates were separated by SDS-PAGE and transferred to membrane. After blocking in 0.5% bovine serum albumin, the membrane was probed with the purified GST-SHP-1 (C453S) fusion protein (5 g/ml) in a far Western blot analysis. The bound fusion protein was detected by probing with the anti-KT3 epitope tag monoclonal antibody.

RESULTS
In Vitro Substrate Selectivity of the SHP-1 Catalytic Domain-To determine whether SHP-1 displays substrate selectivity, we subjected a pool of NK cell phosphotyrosine-containing proteins to dephosphorylation by SHP-1 in vitro. The catalytic activity of SHP-1 is regulated by an intramolecular association of the amino-terminal SH2 domains with the carboxyl-terminal catalytic portion of the molecule (27,28). Binding of the SHP-1 SH2 domains to phosphotyrosine-containing proteins or peptides increases the catalytic activity of the enzyme (13,29). Peptides corresponding to the cytoplasmic ITIMcontaining region of the p70 KIR NKB1 were synthesized (24). The tyrosine-phosphorylated form (pY1) of this peptide activates SHP-1 in vitro, whereas the nonphosphorylated form (Y1) does not (13,29). The pY1 peptide therefore mimics the activation of SHP-1 that occurs upon binding to tyrosine-phosphorylated KIR. NK cells stimulated with the PTP inhibitor pervanadate ( Fig. 1A) or via FcR cross-linking (Fig. 1B) were lysed, and the lysates were incubated in vitro with full-length, GSH agarose-bound SHP-1. Significantly, activation of SHP-1 by addition of the phospho-ITIM-containing peptide resulted in selective, time-dependent dephosphorylation of proteins of around 50, 70, and 116 kDa, whereas little dephosphorylation of these proteins was observed in the presence of the nonphosphorylated peptide. Consistent with a previous report (19), some proteins, notably a p36 phosphotyrosyl protein, were dephosphorylated by SHP-1 even in the absence of the activating pY1 peptide (Fig. 1B). Particularly striking, however, is the large number of phosphotyrosyl proteins not efficiently dephosphorylated by SHP-1. These results suggest that, in vivo, in-hibitory receptor-associated SHP-1 may not act promiscuously, but rather may target certain phosphotyrosyl proteins for dephosphorylation. This method does not permit identification of the dephosphorylated proteins, but rather serves to focus the search to proteins of certain molecular masses.
Catalytically inactive tyrosine phosphatases can be used to identify substrates by direct binding (30 -32). We constructed a GST fusion protein containing the catalytic domain of enzymatically inactive SHP-1 (C453S). The SH2 domains of SHP-1 were removed because of their ability to limit accessibility of the catalytic site to substrate. To examine the substrate specificity of the SHP-1 catalytic domain, we used this fusion protein as a probe in far Western blots. This method allows analysis of direct protein-protein interactions and has been used successfully by Black and Bliska (31) to show that p130Cas is the substrate of the Yersinia PTP YopH. Lysates of unstimulated or pervanadate-stimulated NK cells were separated by SDS-PAGE and probed with the SHP-1 C453S catalytic domain fusion protein. A single protein of around 70 kDa showed pervanadate-inducible reactivity with the fusion protein (Fig. 2). Reprobing of the membrane with an anti-phosphotyrosine monoclonal antibody revealed the vast number of phosphotyrosyl proteins present, underscoring the high degree of specificity displayed by the SHP-1 catalytic domain for the single . Cell lysates were subjected to in vitro dephosphorylation by agarose-bound, full-length SHP-1 in the presence of a peptide containing the KIR ITIM that is either nonphosphorylated (Y1) or in which the tyrosine residue is phosphorylated (pY1). At the indicated times, SHP-1 was pelleted by centrifugation, and the supernatant was added to sample buffer. Following separation by SDS-PAGE, phosphotyrosine-containing proteins were detected by anti-phosphotyrosine immunoblot analysis. The arrowhead indicates the protein specifically dephosphorylated by pY1-activated SHP-1. Equal protein loading was confirmed by reprobing the membrane in B with ZAP-70-specific antiserum. Molecular masses are indicated on the left in kDa. protein recognized.
SHP-1 Recruited to KIR Blocks Syk-generated Signaling-Multisubunit immune recognition receptors, such as FcR, are thought to initiate signaling through the sequential activation of Src family PTKs followed by Syk family PTKs. Additionally, Syk plays a key role in the generation of NK cell-mediated natural killing (33). To determine whether the inhibitory effect of KIR/SHP-1 engagement on this activation process is located upstream or downstream of Syk activation, we used a chimeric receptor system. Cross-linking of a chimeric receptor consisting of the CD16 extracellular region and CD7 transmembrane region containing the PTK Syk (CD16:7:Syk) in the cytoplasmic tail results in Src family PTK-independent activation of T cells (26). 2 If SHP-1 acts upstream of Syk, then KIR engagement should not affect CD16:7:Syk-initiated cellular activation. We found, however, that co-engagement of KIR completely inhibited CD16:7:Syk-triggered, T cell-mediated cytotoxicity (Fig. 3). The inhibition was reversed by overexpression of dominantnegative, catalytically inactive (C453S) SHP-1 (data not shown). No inhibition of CD16:7:Syk-initated signaling was observed following cross-linking of another cell surface protein, LFA-1 (data not shown). Thus, KIR-associated SHP-1 is able to inhibit lymphocyte activation by acting at or downstream of the point of Syk family PTK activation.
SHP-1 Catalytic Domain Interacts Directly with SLP-76 -Based on these results, we utilized a genetic model to determine if the protein recognized in the far Western blot is either Syk or ZAP-70. A protein of identical molecular mass to that observed from NK cell lysates was detected in lysates obtained from pervanadate-stimulated Jurkat T cells ( Fig. 4A and data not shown). Jurkat E6 cells do not express Syk (34), and therefore the protein is not Syk. Furthermore, the protein is also present in lysates obtained from a ZAP-70-deficient Jurkat subclone, and overexpression of either Syk or ZAP-70 did not result in the appearance of additional reactive proteins (Fig.  4A). Important to the interpretation of these results is that tyrosine phosphorylation of SLP-76 following pervanadate stimulation remained intact in the ZAP-70-deficient P116 Jurkat clone (data not shown), although CD3 cross-linking-induced tyrosine phosphorylation of SLP-76 is reduced in the P116 cell line (21). Thus, by genetic criteria, the inducible ϳ70-kDa phosphotyrosyl protein recognized by the SHP-1 catalytic domain is neither Syk nor ZAP-70. We therefore considered downstream effectors of Syk that are in the 70-kDa range and that would be likely SHP-1 substrates.
Immunoprecipitation of SLP-76 from lysates of pervanadatestimulated NK cells revealed that the protein recognized by the GST SHP-1 (C453S) catalytic domain in far Western analysis was SLP-76 (Fig. 4B). Although a small degree of association between the SHP-1 (C453S) catalytic domain and non-tyrosinephosphorylated SLP-76 was consistently observed (Fig. 4B, left  lane), the association was clearly augmented by tyrosine phosphorylation of SLP-76 (Fig. 4B, second lane from left, and see Fig. 4D). Control immunoprecipitations of Syk and ZAP-70 showed no reactivity with the SHP-1 catalytic domain despite significant tyrosine phosphorylation of both kinases. Lck, another heavily tyrosine-phosphorylated PTK, was also not recognized by the SHP-1 catalytic domain in this assay (data not shown). Far Western blots with control GST fusion proteins demonstrated no reactivity with SLP-76 (data not shown). Immunodepletion of SLP-76 from whole cell extracts selectively removed the reactive protein (Fig. 4C). The mutated form of SLP-76 in which the three critical tyrosines are replaced by phenylalanine (3Y-F) was not tyrosine-phosphorylated following antigen-receptor stimulation of T cells (Fig. 4D) or FcR stimulation of NK cells (data not shown). More importantly, the SHP-1 catalytic domain did not interact with SLP-76 (3Y-F) (Fig. 4D), suggesting that these functionally critical residues are direct targets of SHP-1-dependent dephosphorylation. The activation-inducible association of the SHP-1 catalytic domain with SLP-76 contrasts with the potential constitutive interaction between these two proteins (42). KIR/SHP-1 Dephosphorylates SLP-76, a Key Regulator of Lymphocyte Cytotoxicity-Although SLP-76 is known to regulate T helper cell antigen receptor signaling and cytokine production (23, 36 -41), a role for SLP-76 in cytotoxic lymphocytes has not been established. Significantly, we found that overexpression of wild-type SLP-76 enhanced CD3-initiated cytotoxic T lymphocyte-mediated killing, whereas overexpression of SLP-76 (3Y-F) did not (Fig. 5A). Similar functional studies in NK cells were attempted, but the level of SLP-76 overexpression achieved in NK cells was minimal relative to that in T cells. Therefore, we next asked whether SLP-76 is involved in NK cell Fc␥RIIIA-mediated signaling and whether it is subject to the SHP-1-mediated KIR inhibitory signal. Indeed, FcR ligation resulted in the tyrosine phosphorylation of SLP-76 in NK cells. Co-engagement of KIR completely inhibited FcRinitiated SLP-76 tyrosine phosphorylation, whereas ligation of another NK cell surface receptor, CD56, did not (Fig. 5B). Furthermore, overexpression of a catalytically inactive form of SHP-1 (C453S) reversed the KIR-mediated inhibition of SLP-76 tyrosine phosphorylation, whereas overexpression of wild-type SHP-1 did not (Fig. 5B). These results demonstrate that SLP-76 acts in the pathway involved in generation of lymphocyte cytotoxic activity that is subject to SHP-1-dependent, KIR-mediated inhibitory signaling.
To determine whether KIR-associated SHP-1 selectively dephosphorylates SLP-76, we subjected lysates of pervanadatestimulated Jurkat cells to an in vitro dephosphorylation reaction. Immunoprecipitation of SLP-76 or ZAP-70 from these treated lysates revealed that SHP-1 activated by the phospho-ITIM-containing peptide (pY1) selectively dephosphorylated SLP-76 (Fig. 5C). Although some dephosphorylation of SLP-76 was observed in the presence of the nonphosphorylated peptide, ZAP-70 was not dephosphorylated under any of the conditions tested. Taken together, these results support a model in which SHP-1 recruited to tyrosine-phosphorylated KIR ITIMs interrupts immune cell activation by the selective dephosphorylation of the critical signaling molecule, SLP-76. DISCUSSION The multisubunit immune recognition receptors (T cell antigen receptor, B cell antigen receptor, Fc⑀RI, and Fc␥RIIIA) are each subject to negative regulation by SHP-1-utilizing inhibitory receptors (12,(43)(44)(45)(46)(47)(48). Each of these activating receptor systems relies not only on the immunoreceptor tyrosine-based activation motif-containing receptor subunits (CD3, FcR␥, Ig␣, and Ig␤) but also on Src and Syk family PTKs (1-4). Importantly, the recently cloned NK cell-activating receptor subunit DAP12 also bears an immunoreceptor tyrosine-based activation motif, suggesting that DAP12 may couple to similar signal transduction pathways (49). We therefore investigated the hypothesis that the direct target of SHP-1-dependent inhibitory signaling is a phosphotyrosyl protein common to several activation pathways.
Using both in vitro and functional assays, we identified SLP-76 as a direct substrate for SHP-1 recruited to inhibitory receptors. SLP-76 plays a key role in signaling by the T cell WR strain vaccinia virus or with recombinant vaccinia virus expressing either FLAG-tagged wild-type SLP-76 or FLAG-tagged SLP-76 (3Y-F). Following infection, the cells were either unstimulated or stimulated by cross-linking of the T cell antigen receptor for 1 min or by pervanadate treatment for 5 min. Following stimulation, cleared lysates were subjected to immunoprecipitation with anti-FLAG monoclonal antibody and analyzed by far Western blotting; the arrow indicates SLP-76. The membrane was sequentially reprobed with anti-phosphotyrosine mAb (4G10) and anti-SLP-76 serum that preferentially recognizes non-tyrosine-phosphorylated SLP-76. Importantly, the three tyrosine residues critical to the function of SLP-76 are also critical for recognition of SLP-76 by the catalytic domain of SHP-1. Of particular note is the fact that these tyrosine residues reside in the following motifs: FEED-DYESP, EDDGDYESP, and EDDADYEPP (35). The tyrosine residues recognized by a variety of protein tyrosine phosphatases are contained in similar motifs with several acidic residues in the -1 to -5 (relative to Tyr) position and a proline residue in the ϩ3 position (30 -32, 50 -52). Thus, based on structural criteria, SLP-76 is a likely SHP-1 substrate.
Interestingly, the tyrosine residues in Syk and ZAP-70 (Tyr 341 and Tyr 315 , respectively, based on porcine Syk and human ZAP-70 numbering) that mediate binding to phospholipase C-␥ and Vav also appear to reside in PTP recognition motifs (53)(54)(55)(56). The recently identified p36 phosphotyrosyl protein, linker of activated T cells (LAT), also contains a tyrosine (Tyr 127 ) in a PTP recognition motif, but Tyr 127 does not have an identified docking function (57). One study (19) suggests that SHP-1 may dephosphorylate p36 in vitro, albeit in an ITIMindependent fashion. Although our GST SHP-1 C453S catalytic domain recognized neither Syk nor ZAP-70, a similar SHP-1 (C453S) fusion protein has been used to isolate ZAP-70 from Jurkat lysates (58). Interpretation of such results is difficult, however, due to the possibility that the observed interaction is indirect. The far Western method we utilized does allow analysis of direct protein-protein interaction, and, in this assay, the SLP-76 catalytic domain did not interact directly with Syk or ZAP-70. One caveat, however, is that protein denaturation during electrophoresis may prevent binding in a far Western analysis. Thus, we cannot rule out the possibility that Syk, ZAP-70, linker of activated T cells (LAT), or other molecules could serve as additional in vivo substrates of SHP-1, particularly molecules known to transduce signals from cytokine receptors regulated by SHP-1 (59,60).
SLP-76 is known to be activated by Syk family kinases (35)(36)(37)(38)(39)(40). Consistent with these observations, we have demonstrated here that the SHP-1-dependent KIR inhibitory signal acts at or downstream of Syk activation. Thus, in addition to meeting the structural criteria expected of a potential SHP-1 substrate, SLP-76 fulfills the functional criteria expected for SHP-1-dependent inhibitory receptor signaling pathways.
Particularly important is the fact that overexpression of a wild-type form of SLP-76 augments lymphocyte responses, whereas SLP-76 (3Y-F) has no effect (Fig. 5A) (37,41). One model that is consistent with these findings is that endogenous SLP-76 exists in a relatively stable signaling complex prior to cell stimulation. Thus, overexpressed SLP-76 (Y3-F) is both unable to remove endogenous, wild-type SLP-76 from this complex and is unable to generate signals independently-it does not act as a dominant-negative. In contrast, overexpressed wild-type SLP-76 may provide a core for the formation of new, signaling-competent complexes and therefore augment lymphocyte responsiveness. In such a model, SLP-76 activation is a rate-limiting step in lymphocyte activation. Selective targeting of such a rate-limiting step by SHP-1-utilizing inhibitory receptors would consequently be an effective strategy for inhibiting immune cell activation.
When considered in conjunction with our previous demonstration that SHP-1 recruited to KIR inhibits the tyrosine phosphorylation of , ZAP-70, and phospholipase C-␥ (12), the current study raises important questions regarding how SLP-76 and potential homologues function to regulate immune cell activation. A model in which several PTKs and their sub-FIG. 5. SLP-76 augments lymphocyte-mediated cytotoxicity and is the target of inhibitory receptor-associated SHP-1. A, the human CD8 ϩ T cell clone KT8 was infected with either control WR strain vaccinia virus or with recombinant vaccinia virus encoding wildtype SLP-76 or SLP-76 (Y3-F). The infected cells were used as effectors in a reverse antibody-dependent cell-mediated cytotoxicity assay against the FcR ϩ target cell P815 coated with a monoclonal antibody specific for CD3 (OKT3). Results are expressed as lytic units. B, NK cell clones (10 7 /sample) expressing the KIR identified by the monoclonal antibody GL183 were infected with control vaccinia virus or with vaccinia virus encoding wild-type SHP-1 or catalytically inactive SHP-1 (C453S). The cells were then stimulated by cross-linking the FcR alone or in combination with CD56 (control) or KIR as indicated. SLP-76 was immunoprecipitated from lysed cells, and proteins were analyzed by immunoblot with anti-phosphotyrosine (antibody 4G10) and then with SLP-76-specific antiserum. vacc infx, vaccinia infection. C, Jurkat cells (5 ϫ 10 6 /sample) were stimulated with pervanadate. Cleared lysates were subjected to in vitro dephosphorylation as described in Fig. 1. Following a 60-min dephosphorylation reaction, the supernatants were halved, and either SLP-76 or ZAP-70 was immunoprecipitated. Blotting was performed as already described. The arrow indicates the position of SLP-76. strates are physically or functionally coordinated by SLP-76 is consistent with our results. SHP-1-dependent inhibitory signaling may thus act by targeting SLP-76, a central integrator of immune cell activation pathways.