Structure, inhibitor, and regulatory mechanism of Lyp, a lymphoid-specific tyrosine phosphatase implicated in autoimmune diseases

Xiao Yu, Jin-Peng Sun, Yantao He, Xiaoling Guo, Sijiu Liu, Bo Zhou, Andy Hudmon, and Zhong-Yin Zhang*

Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, 635 Barnhill Drive, Indianapolis, IN 46202

Edited by Arthur Weiss, University of California School of Medicine, San Francisco, CA, and approved October 25, 2007 (received for review July 5, 2007) The lymphoid-specific tyrosine phosphatase (Lyp) has generated generating a more active phosphatase that is more effective in enormous interest because a single-nucleotide polymorphism in the inhibiting T cell signaling than the wild-type enzyme (13). gene (PTPN22) encoding Lyp produces a gain-of-function mutant Given the strong association of the C1858T polymorphism with phosphatase that is associated with several autoimmune diseases, various autoimmune disorders and the elevated phosphatase ac- including type I diabetes, rheumatoid arthritis, Graves disease, and tivity associated with the resultant Lyp/R620W variant, Lyp repre- systemic lupus erythematosus. Thus, Lyp represents a potential target sents a potential target for a broad spectrum of autoimmune for a broad spectrum of autoimmune disorders. Unfortunately, no Lyp diseases. Small-molecule Lyp inhibitors may have therapeutic value inhibitor has been reported. In addition, little is known about the for treating these disorders. In addition, specific Lyp inhibitors will structure and biochemical mechanism that directly regulates Lyp be useful in delineating the mechanism of Lyp in T cell signaling, function. Here, we report the identification of a bidentate salicylic development, and differentiation. To aid the design of Lyp inhib- acid-based Lyp inhibitor I-C11 with excellent cellular efficacy. Struc- itors and gain insight into the regulatory mechanism for Lyp, we tural and mutational analyses indicate that the inhibitor binds both have determined the crystal structures of the Lyp PTP domain the active site and a nearby peripheral site unique to Lyp, thereby either alone or in complex with a selective small-molecule inhibitor. furnishing a solid foundation upon which inhibitors with therapeutic Structural analysis together with mutational studies led to the potency and selectivity can be developed. Moreover, a comparison of identification of molecular determinants that can be exploited for the apo- and inhibitor-bound Lyp structures reveals that the Lyp- 35 42 the acquisition of more potent and selective Lyp inhibitors. In specific region S TKYKADK , which harbors a PKC phosphorylation addition, the structures revealed a unique flexible region in Lyp, site, could adopt either a loop or helical conformation. We show that which harbors a consensus protein kinase C (PKC) phosphorylation Lyp is phosphorylated exclusively at Ser-35 by PKC both in vitro and site (Ser-35). We demonstrated that Lyp can be phosphorylated by in vivo. We provide evidence that the status of Ser-35 phosphoryla- PKC on Ser-35, which impairs the Lyp-mediated substrate dephos- tion may dictate the conformational state of the insert region and phorylation and signaling in T cell. This finding establishes a thus Lyp substrate recognition. We demonstrate that Ser-35 phos- mechanism whereby signaling through PKC may directly influence phorylation impairs Lyp’s ability to inactivate the Src family kinases the cellular processes regulated by Lyp. and down-regulate T cell receptor signaling. Our data establish a mechanism by which PKC could attenuate the cellular function of Lyp, Results and Discussion thereby augmenting T cell activation. Identification of a Lyp-Specific Inhibitor. Despite the promise of crystal structure ͉ enzyme regulation ͉ Lyp inhibitor ͉ phosphorylation Lyp-directed therapeutics for autoimmune diseases, no Lyp inhib- itor has been reported. Because PTPs share a high degree of structural conservation in the active site, i.e., the pTyr binding rotein tyrosine phosphorylation mediates multiple signal trans- pocket, designing active site-directed inhibitors with both high Pduction pathways that play key roles in innate and acquired affinity and selectivity for these enzymes is quite a challenge. immunity (1, 2). The level of tyrosine phosphorylation is controlled Fortunately, it has been shown that pTyr alone is not sufficient for by the coordinated action of protein tyrosine kinases (PTKs) and high-affinity binding and residues flanking the pTyr are important protein tyrosine phosphatases (PTPs). The importance of the PTKs for PTP substrate recognition (14). These studies point to a unique in the immune system are well recognized and widely appreciated. paradigm for PTP inhibitor design, namely bidentate ligands that However, the functional significance of the PTPs in regulating can engage both the active site and an adjacent less-conserved various immune responses is far from clear. The lymphoid-specific subpocket for enhanced affinity and selectivity. As an initial effort tyrosine phosphatase (Lyp) has received enormous attention be- to develop Lyp inhibitors, we screened an 80-member focused BIOCHEMISTRY cause of the finding that a single-nucleotide polymorphism (SNP) in the gene (PTPN22) encoding Lyp is associated with several library that was designed to bridge the PTP active site and an autoimmune diseases, including type I diabetes (3), rheumatoid adjacent peripheral site. The ability of the compounds to inhibit the arthritis (4, 5), Graves disease (6), and systemic lupus erythema- Lyp-catalyzed hydrolysis of p-nitrophenyl phosphate (pNPP) was tosus (7). Lyp is a 110-kDa protein consisting of an N-terminal PTP domain and a noncatalytic C-terminal segment with several Pro- Author contributions: X.Y. and J.-P.S. contributed equally to this work; X.Y., J.-P.S., Y.H., rich motifs (8, 9). Lyp belongs to a subfamily of PTPs, which include and Z.-Y.Z. designed research; X.Y., J.-P.S., Y.H., S.L., B.Z., and A.H. performed research; X.G. PTP-PEST (PTPN12), PTP-HSCF/BDP1 (PTPN18), and Lyp/PEP contributed new reagents/analytic tools; X.Y., J.-P.S., Y.H., and Z.-Y.Z. analyzed data; and (PTPN22) (10). Biochemical studies suggest that Lyp inhibits T cell Z.-Y.Z. wrote the paper. activation, likely through dephosphorylation of the T cell receptor The authors declare no conflict of interest. (TCR)-associated Lck and ZAP-70 kinases (9, 11, 12). Interestingly, This article is a PNAS Direct Submission. the disease-causing SNP (a C-to-T substitution at position 1858 in Data deposition: The atomic coordinates have been deposited in the Protein Data Bank, the coding region of Lyp) produces an amino acid substitution www.pdb.org (PDB ID codes 2QCJ and 2QCT). (R620W) within the first Pro-rich region in the C terminus, thereby *To whom correspondence should be addressed. E-mail: [email protected]. impairing Lyp binding to the Src homology 3 (SH3) domain of Csk This article contains supporting information online at www.pnas.org/cgi/content/full/ (3, 4). Moreover, it has been shown that the autoimmune- 0706233104/DC1. predisposing variant of Lyp is actually a gain-of-function mutation, © 2007 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0706233104 PNAS ͉ December 11, 2007 ͉ vol. 104 ͉ no. 50 ͉ 19767–19772 Downloaded by guest on September 29, 2021 mational flexibility may be important for Lyp substrate/ligand O recognition. A comparison with published PTP structures using DALI (16) HO O reveals that the Lyp structure is most similar to the catalytic domain N of PTP1B with a Z score of 33.3 and a rmsd of 1.9 Å for all C␣ pairs. N N HO N H In general, the core elements of Lyp superimpose well with those O of PTP1B (Fig. 2A). However, several surface loops surrounding Fig. 1. The chemical structure of I-C11. the active site display substantial differences between the Lyp structures and that of PTP1B. For example, Lyp has a longer loop between ␣1 and ␤1 (residues 75–80). Both the WPD loop and the assessed at pH 7 and 25°C. Compound I-C11 (Fig. 1) was identified loop between ␣3 and ␤8 (residues 216–222) of Lyp assume con- as the most potent and selective Lyp inhibitor from the library with formations that are different from those in PTP1B. Finally, the an IC50 of 4.6 Ϯ 0.4 ␮M. Kinetic analysis indicated that I-C11 is a most striking difference between Lyp and PTP1B resides in the Lyp-specific insert (residues 35–42), which is absent in other PTP reversible and competitive inhibitor for Lyp with a Ki of 2.9 Ϯ 0.5 ␮M. To examine the specificity of I-C11 for Lyp, its inhibitory subfamilies. These structural variations in the surface loops sur- activity toward a panel of PTPs including cytosolic PTPs, PTP1B, rounding the active site may contribute to the distinct PTP sub- SHP2, HePTP, PTP-Meg2, and FAP1, the receptor-like PTPs, strate/ligand binding specificity. CD45, LAR, and PTP␣, and the dual specificity phosphatase VHR, was determined. As shown in Table 1, 1-C11 is reasonably selective Molecular Basis of Lyp Inhibition by I-C11. Fig. 3A depicts the overall Ͼ structure of Lyp showing electron density for I-C11 contoured at 2.0 for Lyp, exhibiting, with one exception, 7-fold selectivity against ␴ Ϫ all PTPs examined. The sole exception is PTP1B, which is more in the simulated annealing Fo Fc OMIT map. The structure related to Lyp. A 2.6-fold preference for Lyp was observed in the reveals that I-C11 interacts with both the active site and a nearby latter instance. peripheral site. Consistent with the ability of salicylic acid deriva- tives to serve as effective pTyr surrogates (17, 18) and the observed competitive mode of Lyp inhibition by I-C11, the benzofuran Crystal Structures of Lyp. To map the architecture of Lyp and determine the molecular basis for Lyp inhibition by I-C11, we salicylic acid moiety occupies the Lyp active-site pocket. The remarkable potency and selectivity of I-C11 for Lyp are the results crystallized the PTP domain of Lyp (residues 1–294) with and of numerous specific interactions (Fig. 3B). The benzofuran sali- without I-C11. The 3D structures of Lyp alone or bound to I-C11 cylic acid engages in both polar and hydrophobic interactions with were solved by molecular replacement using the coordinates of the Lyp active site. The carboxylic acid of the inhibitor forms PTP1B (15) as a search model and refined to 3.0- and 2.8-Å hydrogen bonds with the main-chain amide of Ala-229, the side resolution, respectively. The details of the crystals and structure chains of Cys-227 and Cys-129, and charge–charge interactions with solution are summarized in [supporting information (SI) Table 3]. Arg-233 and Lys-138. The adjacent hydroxyl group makes addi- While this work was in progress, coordinates of the PTP domains tional hydrogen bonds with Glu-133, which further strengthens of Lyp and BDP1 were deposited in the Protein Data Bank (ID polar interactions with the active site. In addition to the polar codes 2P6X and 2OC3, respectively. Like other PTPs, the Lyp PTP interactions, the benzofuran ring participates in aromatic–aromatic domain adopts a compact ␣ϩ␤ structure comprising a central ␤ ␣ stacking interactions with Tyr-60 in the pTyr binding loop and van eight-stranded -sheet surrounded by six -helixes on one side and der Waals contacts with the aliphatic side chains of Gln-274 in the two ␣-helixes on the other (Fig. 2A). The PTP signature motif 226 233 Q loop, Ala-229 and Ser-228 in the P loop, and Lys-138 in the loop (H CSAGCGR ) forms a loop (P loop) at the base of the between ␤3 and ␤4. Besides interactions with the Lyp active site, the active-site pocket. The overall structures of Lyp in the ligand free triazolidin ring in the linker makes van der Waals contacts with and bound forms are quite similar, with significant conformational side-chain atoms (C␦,O␧1, and N␧2) of Gln-274. In addition, the differences confined to the WPD loop (residues 193–204), which distal naphthalene ring in I-C11 binds a peripheral site defined by contains the catalytically important general acid Asp-195, and an Phe-28, Leu-29, and Arg-33 (Fig. 3B). Interestingly, these residues, ␣ Ј ␣ insert (residues 35–42) between 2 and 1 that is unique to the Lyp which are located in ␣2Ј, border a binding pocket equivalent to the subfamily of PTPs (Fig. 2). While the WPD loop exits in a half-open second aryl phosphate-binding site previously identified in PTP1B conformation in the apo-Lyp structure, it is fully open in the (19). This pocket (Fig. 2), with Arg-254 and Gly-259 (Arg-266 and ligand-bound form to accommodate the inhibitor in the active site. Ser-271 in Lyp) in the bottom and surrounded by Arg-24, Met-258, Interestingly, the Lyp-specific insert can either adopt a loop con- and Gln-262 (Lys-32, Pro-270, and Gln-274 in Lyp), is important for formation in the apo-structure or a helical conformation in the PTP1B substrate recognition (20, 21) and has been targeted for Lyp⅐I-C11 complex (Fig. 2A). As will be shown below, this confor- PTP1B inhibitor development (22). To further investigate the structural basis for Lyp inhibition by I-C11, we mutagenized several amino acids in Lyp, including Table 1. Selectivity of compound I-C11 against a panel of PTPs Cys-129, Lys-138, Phe-28, Leu-29, and Arg-33, that were implicated PTP IC50, ␮M in I-C11 binding and evaluated the effect on Lyp activity and inhibitor binding affinity. As shown in Table 2, no significant Lyp 4.6 Ϯ 0.4 Ϯ differences in the kinetic parameters for pNPP hydrolysis were PTP1B 11.8 1.8 observed for the wild-type and mutant Lyps, indicating that the SHP2 36.0 Ϯ 3.8 Ϯ mutations did not perturb the catalytic site’s integrity. In support of HePTP 32.1 3.2 the structural observations, replacement of Cys-129 and Lys-138 PTP-Meg2 31.8 Ϯ 4.6 Ϯ with an Ala reduced Lyp’s affinity for I-C11 by 1.8- and 4.3-fold, FAP-1 36.8 8.8 respectively. Similarly, removal of the side chain at Phe-28, Leu-29, VHR 103 Ϯ 27 Ϯ and Arg-33 also led to a reduction in I-C11 binding affinity, with a CD45 186 14 7.6-, 2.6-, and 2.7-fold increase in the IC values for F28A, L29A, LAR No inhibition at 100 ␮M 50 ␣ ␮ and R33A, respectively. Collectively, our structural and mutational PTP No inhibition at 100 M analyses of the interactions between Lyp and I-C11 showed that the All measurements were made by using pNPP as a substrate at pH 7.0, 25°C, benzofuran salicylic acid occupies the active site, whereas the distal and ionic strength of 0.15 M. naphthalene ring makes hydrophobic interactions with a region

19768 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0706233104 Yu et al. Downloaded by guest on September 29, 2021 Fig. 2. Structural features of Lyp. (A) Superposition of the structures of apo (blue), I-C11 bound Lyp (green), and PTP1B (red; Protein Data Bank ID code 1G1H). (Right) Superposition of the overall structures. Besides the WPD loop, the most significant difference is the Lyp-specific insert (S35TKYKADK42) between ␣2Ј and ␣1. This region can exist either as an extension of ␣2Ј in the Lyp⅐I-C11 complex or a loop in the native structure. (Left) Superposition of the residues corresponding to the ‘‘second aryl phosphate-binding site’’ of PTP1B and the Lyp-specific insert region. Residues in apo-Lyp, Lyp⅐I-C11 complex, and PTP1B are depicted in blue, green, and red, respectively. (B) Sequence alignment of the Lyp-specific insert and ␣5-Q-loop with corresponding regions from PTP-PEST, BDP1, LAR, PTP␣, SHP2, and PTP1B. The Lyp-specific insert is highlighted in red with the PKC phosphorylation site marked by a star. Residues involved in defining the second aryl phosphate-binding site are shown in blue. Lyp residues involved in binding the distal naphthalene ring in I-C11 are shown in green.

close to the second aryl phosphate-binding site. It is worth noting Km of 0.59 Ϯ 0.07 ␮M. These values are within the normal range that many of the residues that are in direct contact with I-C11, observed for PKC substrates (24, 25), indicating that Lyp/Ser-35 is especially those in the peripheral site, are unique to Lyp, which an efficient substrate for PKC␦. We should note that the corre- provide a structural basis for Lyp inhibitor selectivity. The atomic- sponding Ser-39 in PTP-PEST could also be phosphorylated by level information on the architecture of Lyp furnishes a solid PKC (26). To show that Lyp is phosphorylated at Ser-35 in vivo, foundation for the design of more potent and selective Lyp-based Jurkat T cells transfected with Lyp, Lyp/S35E, and Lyp/S35A were small-molecule therapeutics. treated with phorbol 12-myristate 13-acetate. Western blot analysis with a phospho-specific antibody that recognizes PKC substrates The Lyp-Specific Insert Harbors a PKC Phosphorylation Site. Although with the motif (R/K)X(phosphoS)(Hyd)(R/K) revealed that PKC Lyp is implicated as a negative regulator of TCR signaling, little is activation in T cells leads to Lyp phosphorylation (Fig. 4C). known about how Lyp is regulated. Interestingly, sequence analysis revealed that residues K32RQSTKYK39 in Lyp correspond to a Potential Role of Ser-35 Phosphorylation in Lyp Substrate Recogni- consensus PKC phosphorylation motif, (K/R)XX(S/T)X(K/R) tion. The observation that Lyp is phosphorylated exclusively at (23). To determine whether Lyp can be phosphorylated by PKC, we Ser-35 by PKC raises the possibility that the activity of Lyp may be incubated the Lyp catalytic domain with recombinant PKC␦ in the regulated by Ser-35 phosphorylation. Remarkably, the PKC phos- BIOCHEMISTRY presence of [␥-32P] ATP. Lyp phosphorylation was monitored by phorylation site Ser-35 is located within the Lyp-specific insert the incorporation of radioactive ␥ phosphate into the protein and (S35TKYKADK42) that exists in two discrete conformations in the found to be time- and dose-dependent (Fig. 4A). In addition, Lyp apo- and ligand-bound Lyp structures (Fig. 2A). In our apo-Lyp phosphorylation was saturable with a stoichiometry of 1.08 Ϯ 0.11 structure, the insert assumes a loop conformation projecting out mol of phosphate/mol Lyp. The PKC␦-catalyzed Lyp phosphory- from the protein and making crystal packing contacts with residues lation was also examined by electrospray mass spectrometry. Under Tyr-94, Met-245, Lys-248, and Leu-252 in the neighboring Lyp. In the same conditions, an increase of 81 Da in mass was observed for the Lyp⅐I-C11 complex, residues S35TKYKADK42 adopt a helical phospho-Lyp, consistent with a single residue being phosphorylated conformation and become part of ␣2Ј. The Lyp-specific insert is (data not shown). To identify the site of phosphorylation, we also helical in the apo-Lyp structure (Protein Data Bank ID code determined the effect of S35E or T36E mutation on the PKC␦- 2P6X) crystallized at a higher pH (7.9 vs. 5.9), indicating confor- catalyzed Lyp phosphorylation. As shown in Fig. 4B, although mational flexibility in this region. In the helical form, Ser-35 faces Lyp/T36E could still be phosphorylated by PKC␦, replacement of the interior of the structure and makes a H-bond with the conserved Ser-35 by Glu abolished Lyp phosphorylation. The results suggest Arg-266 at the bottom of the second aryl-phosphate binding pocket. that PKC␦ phosphorylates Lyp exclusively on Ser-35. We further Moreover, the side chain of Tyr-38 participates in favorable edge to demonstrated that the PKC␦-mediated Lyp phosphorylation fol- face stacking interactions with the phenyl rings of Tyr-44 and Ϫ1 lows Michaelis–Menten kinetics, with a kcat of 10.6 Ϯ 0.4 min and Tyr-66. These interactions are important for the stabilization of the

Yu et al. PNAS ͉ December 11, 2007 ͉ vol. 104 ͉ no. 50 ͉ 19769 Downloaded by guest on September 29, 2021 Fig. 4. Lyp is phosphorylated at Ser-35 by PKC. (A)(Upper) Four micrograms of (His)6-tagged Lyp catalytic domain was incubated with 40 ng of PKC␦ with or without [␥-32P]-ATP at 25°C for 20 min. (Lower) Four micrograms of Lyp was incubated with 40 ng of PKC␦ with [␥-32P]-ATP at 30°C and quenched at different time points. (B) Identification of the phosphorylation site in Lyp. Four micrograms of (His)6-tagged wild-type, S35E, and T36E mutant Lyps were incubated with or without 40 ng of PKC␦, at 30°C for 40 min. The phosphor- Fig. 3. Structure of Lyp in complex with I-C11. (A) Overall structure of Lyp ylation of Lyp was analyzed by autoradiography, and total Lyp protein was showing electron density of I-C11 contoured at 2.0 ␴ in the F Ϫ F simulated o c detected with anti-(His) antibody. (C) Lyp phosphorylation by PKC in Jurkat annealing OMIT map. (B) Detailed interactions between Lyp and I-C11. Hy- 6 cells, analyzed by an anti-phospho(Ser) PKC substrate antibody. drogen bonds and charge–charge interactions are depicted as black or red dotted lines, respectively. Residues involve in hydrophobic interactions were shown with a cutoff distance of 5 Å. Ser-35 Phosphorylation Impairs Lyp-Mediated Src Dephosphorylation. To examine the possibility that Lyp activity may be regulated by helical conformation of the Lyp-specific insert. Thus, phosphory- PKC, we initially studied the effect of Ser-35 phosphorylation on lation at Ser-35 may destabilize the helical conformation of the Lyp Lyp-catalyzed hydrolysis of a small-molecule substrate pNPP. No insert region. Interestingly, in the helical conformation, residues significant changes in the kcat and Km values were observed for the Phe-28, Lys-32, Ser-35, and Lys-39 form part of the second aryl- pNPP reaction catalyzed by phospho-Lyp (Table 2). Thus, Ser-35 phosphate binding pocket, which, in the case of PTP1B, has been phosphorylation does not affect the intrinsic Lyp phosphatase shown to be important for substrate recognition (20, 21). Given activity. We then assessed the effect of Lyp phosphorylation on the potential importance of the helical conformation to Lyp more physiologically relevant substrates. Previous studies suggest substrate recognition and the role of Ser-35 in maintaining this that Lyp is a potent negative regulator immediately downstream of helical conformation, we hypothesized that Lyp phosphorylation TCR (27) and identify the Src family kinase Lck as a cellular by PKC at Ser-35 may regulate the Lyp-mediated substrate substrate for Lyp (9, 12). We were unable to prepare sufficient dephosphorylation. quantities of recombinant phospho-Lck for kinetic analysis. As an alternative, we turned to Src, which has almost identical residues surrounding the regulatory tyrosine phosphorylation sites and Table 2. Kinetic parameters and IC50 values for the wild-type shares Ͼ60% of overall sequence identity with Lck. and mutant Lyps The activity of the Src family kinase is regulated by phosphory-

IC50 for lation at two distinct tyrosine residues. Autophosphorylation of Ϫ1 Lyp kcat,S Km,mM I-C11, ␮M Tyr-416 (Tyr-394 in Lck) in the kinase domain is required for Src activation. In contrast, phosphorylation of Tyr-527 (Tyr-505 in Lck) WT 0.70 Ϯ 0.03 4.2 Ϯ 0.4 4.6 Ϯ 0.4 in the C-terminal tail by Csk inactivates Src because of an intramo- pLyp35 0.64 Ϯ 0.03 4.7 Ϯ 0.3 48 Ϯ 7.4 lecular pTyr-SH2 interaction (28, 29). We prepared doubly phos- F28A 0.67 Ϯ 0.07 4.0 Ϯ 0.4 35 Ϯ 3.8 phorylated Src and measured Src dephosphorylation by purified Ϯ Ϯ Ϯ L29A 0.56 0.05 3.9 0.3 11.8 0.9 Lyp, pLyp35, and Lyp/S35E by using phosphospecific pSrc416 and Ϯ Ϯ Ϯ R33A 0.61 0.03 3.8 0.2 12.5 2.9 pSrc527 antibodies. As shown in Fig. 5A, the level of pSrc416 Ϯ Ϯ Ϯ S35E 0.59 0.8 4.5 0.6 21 3.1 decreased by Ͼ75% when Src was incubated with Lyp for 5 min, Ϯ Ϯ Ϯ T36E 0.37 0.03 2.4 0.2 2.8 0.2 whereas no significant change was detected in pSrc527. This result Ϯ Ϯ Ϯ S35E/T36E 0.89 0.04 6.9 0.6 250 45 is consistent with the finding that Lyp negatively modulates the Ϯ Ϯ Ϯ C129S 0.73 0.6 2.4 0.2 8.1 0.9 activity of Lck (9, 11, 27). Strikingly, Ͻ20% of pSrc416 was Ϯ Ϯ Ϯ K138A 0.81 0.07 2.8 0.3 19 2.5 hydrolyzed by pLyp35 or Lyp/S35E under the same conditions. All measurements were made by using pNPP as a substrate at pH 7.0, 25°C, Time-course analysis showed that the rate of pSrc416 dephosphor- and ionic strength of 0.15 M. ylation was 8-fold slower for the pLyp35-catalyzed reaction (Fig.

19770 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0706233104 Yu et al. Downloaded by guest on September 29, 2021 Fig. 5. Dephosphorylation of pSrc416 by Lyp. (A) Lyp prefers pSrc416 over pSrc527, and Ser-35 phosphorylation or S35E substitution impairs Lyp-catalyzed hydrolysis of pSrc416. (B) Time course of pSrc416 dephosphorylation by Lyp (■), Fig. 6. Effect of Lyp and the phosphorylation mimicking Lyp/S35E on T cell S35E (Œ), and phospho-Lyp (F). Doubly phosphorylated Src (200 nM) was incu- activation. (A) Effects on T cell signaling. Jurkat T cells were transfected with bated with 10 nM Lyp, S35E, or phospho-Lyp. The dephosphorylation of pSrc416 the indicated plasmids and treated with medium or anti-CD3 antibody for 5 or pSrc527 was monitored by Western blotting with specific antibodies. min. The level of Myc-tagged Lyp protein, Lck phosphorylation at both Tyr-394 and Tyr-505, total Lck protein, phosphorylated ERK1/2, and total ERK1/2 protein were detected by specific antibodies. (B) Effect on TCR-mediated 5B). Apparently, Ser-35 phosphorylation inhibits Lyp activity to- transcriptional activation. Activation of an NFAT/AP-1-driven luciferase re- ward pSrc416. Similar to pLyp35, mutation of Ser-35 to Glu had no porter gene in Jurkat T cells cotransfected with the indicated plasmids was measured after treatment with medium (empty bars) or anti-CD3 antibody effect on the Lyp-catalyzed pNPP hydrolysis (Table 2) but impeded (filled bars) for 6 h. The activity of cotransfected Renilla luciferase was used for pSrc416 dephosphorylation by 4-fold (Fig. 5B). Consequently, we normalization. The data are presented as mean Ϯ SD. conclude that Glu-35 mimics the effect of Ser-35 phosphorylation. Our results with Src highlight the importance of using physiologi- cally relevant substrates in mechanistic characterization of the family kinases and PKC plays a role in augmenting T cell activation PTPs. by inhibiting Lyp. The loss-of-function phenotype associated with pLyp35 or Lyp/ Phosphorylation at Ser-35 Suppresses Lyp Function in T Cell. To S35E may result from a phosphorylation-induced structural change 35 42 further investigate the role of Ser-35 phosphorylation as a potential in the Lyp-specific insert (S TKYKADK ), which in the helical regulatory mechanism for Lyp function, we measured the status of conformation forms part of a presumed substrate recognition Lck phosphorylation in human Jurkat T cells expressing either pocket (the second aryl phosphate-binding site). Because the distal naphthalene ring in I-C11 also binds the same area, one would Myc-Lyp or Myc-Lyp/S35E, which functions as constitutively Ser-35 predict that structural alterations at Ser-35 (phosphorylation or phosphorylated Lyp (Fig. 6A). Lck activation is an early event in phosphorylation mimicking substitutions) should also reduce I-C11 TCR signaling, and stimulation of T cell with anti-CD3 antibody for binding affinity. Indeed, the IC50 values of I-C11 for pLyp35, 5 min led to a dramatic increase in Lck Tyr-394 phosphorylation. Lyp/S35E, and S35E/T36E are 10.4-, 4.6-, and 54-fold higher than As expected, overexpression of Lyp resulted in a 2.4-fold decrease that of the wild-type Lyp (Table 2). These results further support in pLck394 phosphorylation, consistent with a negative role for Lyp the notion that Ser-35 phosphorylation obliterates a major sub- in T cell signaling. In accord with the reduced capacity of Lyp/S35E strate-binding element in Lyp and thus interferes with Lyp substrate to dephosphorylate pSrc416, we found that Lyp/S35E also failed to recognition. BIOCHEMISTRY dephosphorylate pLck394 inside the cell. To provide further evidence that Ser-35 phosphorylation impairs Inhibition of Lyp by I-C11 Enhances TCR Signaling. Given the observed Lyp’s ability to down-regulate TCR signaling, we next examined two selectivity of I-C11 for Lyp, we proceeded to evaluate its ability to major pathways downstream of Lck, the Ras-Raf-MAPK module inhibit Lyp inside the cell. Jurkat T cells expressing either wild-type important for T cell proliferation, and the nuclear factor of acti- Lyp or the inhibitor insensitive Lyp/S35E were incubated with 20 ␮ vated T cells (NFAT) and activator protein-1 (AP-1), two critical M I-C11 for 1 h and subsequently treated with or without transcription factors involved in TCR-induced IL-2 production (13, anti-CD3 antibody for 5 min. As expected, I-C11 increased the TCR-stimulated Lck394 and ERK1/2 phosphorylation by 1.8- and 18). In agreement with the effect on Lck, a 3.5-fold decrease in 2.9-fold, respectively, in T cells expressing wild-type Lyp (Fig. 7). ERK1/2 activity was observed in the Lyp cells when compared with Again, Lyp/S35E was inefficient in down-regulating TCR signaling. the vector control, whereas no appreciable change in ERK1/2 No appreciable changes in pLck394 and pERK1/2 were observed phosphorylation was observed as a result of Lyp/S35E overexpres- when the Lyp/S35E-expressing cells were treated with the same sion (Fig. 6A). In addition, wild-type Lyp could reduce the TCR- concentration of I-C11. induced NFAT/AP-1 transcriptional activation by nearly 3-fold, whereas Lyp/S35E was unable to attenuate NFAT/AP-1 activation Conclusions (Fig. 6B). Taken together, the results support the conclusion that We have identified a salicylic acid-based Lyp inhibitor from a Ser-35 phosphorylation abrogates Lyp’s ability to inactivate the Src focused library that binds with micromolar affinity and is relatively

Yu et al. PNAS ͉ December 11, 2007 ͉ vol. 104 ͉ no. 50 ͉ 19771 Downloaded by guest on September 29, 2021 (OKT3) was from eBioscience. All other reagents were obtained from Sigma.

Kinetics and Inhibition of Lyp-Catalyzed Substrate Dephosphoryla- tion. Initial rate measurements for the Lyp-catalyzed pNPP hydro- lysis in the absence and presence of small-molecule inhibitors were determined as described (15). All assays were carried out at 25°C in 50 mM 3,3-dimethylglutarate (pH 7.0) buffer, containing 1 mM DTT and 1 mM EDTA, with an ionic strength of 0.15 M adjusted with NaCl. Recombinant Src protein phosphorylated at both Tyr-416 and Tyr-527 was used as a physiological substrate for Lyp. The Lyp-catalyzed Src dephosphorylation was carried out under the same conditions used for pNPP. The reaction was quenched Fig. 7. Inhibition of Lyp by I-C11 enhances T cell signaling. Jurkat T cells by adding 1 mM pervanadate and the SDS buffer. The extent of expressing either Lyp or Lyp/S35E were treated with 20 ␮m IC-11 followed by the reaction was analyzed by Western blot and quantitated by anti-CD3 stimulation. The effects of IC-11 on T cell signaling were evaluated densitometry. with specific antibodies. Cell Culture, Transfection, Immunoblotting, and Luciferase Assay. Jurkat T cells were grown at 37°C under an atmosphere of 5% CO2 specific among the PTPs tested. Crystal structures of Lyp solved in in RPMI medium 1640 supplemented with 10% FBS. Full-length the presence and absence of the I-C11 reveal that the inhibitor Lyp and Lyp/S35E mutant were subcloned into the pcDNA4/ interacts with both the catalytic site and a cluster of unique resides mycHis plasmid, and the resulting vectors were introduced into adjacent to the active site. I-C11 is a Lyp specific inhibitor that Jurkat T cells by electroporation. Forty-eight hours after transfec- exhibits cellular activity, and our studies provide a platform to tion, the cells were treated with 5 ␮g/ml anti-CD3 antibody (OKT3; generate future inhibitors that display therapeutic potency and eBioscience) or medium for 5 min. Subsequently, cells were lysed selectivity. Further comparison of the apo- and I-C11-bound Lyp in 50 mM Tris⅐HCl (pH 7.5), 150 mM NaCl, 10% glycerol, 1% structures indicate that the Lyp-specific insert S35TKYKADK42 Nonidet P-40, 50 mM NaF, 10 mM pyrophosphate, 5 mM iodoac- could adopt either a loop or helical conformation. Interestingly, the etate, 1 mM sodium orthovanadate, 1 mM PMSF, and the protease Lyp-specific insert harbors a PKC phosphorylation site, raising the inhibitor mixture. Cell lysates were subjected to SDS/PAGE and possibility that the conformation of the insert region may be transferred electrophoretically to nitrocellulose membrane, which was immunoblotted by appropriate antibodies followed by incuba- controlled by phosphorylation. Biochemical analyses show that tion with horseradish peroxidase-conjugated secondary antibodies. in vitro in vivo Ser-35 is phosphorylated by PKC both and .We The luciferase assay was carried out as described (31). In general, provide evidence that the status of Ser-35 phosphorylation dictates 1 ϫ 107 cells were transfected by electroporation with 2 ␮g of the the conformational state of the insert region and thus Lyp substrate NFAT/AP-1-luc plasmid, 50 ng of the Renilla-TK luciferase plas- recognition. We further demonstrate that Ser-35 phosphorylation mid, and full-length Lyp plasmids or pcDNA4 vector. Forty-eight impairs Lyp’s ability to inactivate the Src family kinases and hours after transfection, cells were stimulated with OKT3 (5 ␮g/ml) down-regulate TCR signaling. Given the importance of PKC in or left untreated for 6 h. Dual luciferase activity was measured TCR signaling (30), our data suggest a unique mechanism by which according to Promega’s instruction, and NFAT/AP-1-luciferase PKC could negatively regulate the cellular function of Lyp, thereby activity was normalized by Renilla activity. augmenting T cell activation. Details on expression and purification of Lyp catalytic domain, crystallization, data collection, structure determination, Lyp phos- Materials and Methods phorylation by PKC, and inhibition by I-C11 in Jurkat cells are Materials. pNPP was purchased from Fluke. [␥-32P]-ATP was from provided in SI Text. Perkin–Elmer. The monoclonal anti-Myc antibody was from Up- state Biotechnology. Anti-Src, anti-Src/pY416, and anti-Src/pY527 We thank Jim Hurley (National Institutes of Health, Bethesda) for the baculovirus pGEX-PKC␦ expression vector, Robert Stahelin for advice antibodies were from Biosource Interantional. Polyclonal anti- with PKC assays, and Millie Georgiadis for assistance with crystallo- ERK1/2, anti-phospho-ERK1/2, and anti-phospho(Ser) PKC sub- graphic data analysis. This work was supported by National Institutes of strate antibodies were purchased from Cell Signaling. Anti-CD3 Health Grant CA69202.

1. Mustelin T, Abraham RT, Rudd CE, Alonso A, Merlo JJ (2002) Front Biosci 7:d918–d969. 16. Holm L, Sander C (1996) Methods Enzymol 266:653–662. 2. Palacios EH, Weiss A (2004) Oncogene 23:7990–8000. 17. Sarmiento M, Wu L, Keng Y-F, Song L, Luo Z, Huang Z, Wu G-Z, Yuan AK, Zhang Z-Y 3. Bottini N, Musumeci L, Alonso A, Rahmouni S, Nika K, Rostamkhani M, MacMurray J, (2000) J Med Chem 43:146–155. Meloni GF, Lucarelli P, Pellecchia M, et al. (2004) Nat Genet 36:337–338. 18. Liang F, Huang Z, Lee S-Y, Liang J, Ivanov MI, Alonso A, Bliska JB, Lawrence DS, 4. Begovich AB, Carlton VE, Honigberg LA, Schrodi SJ, Chokkalingam AP, Alexander HC, Mustelin T, Zhang Z-Y (2003) J Biol Chem 278:41734–41741. Ardlie KG, Huang Q, Smith AM, Spoerke JM, et al. (2004) Am J Hum Genet 75:330–337. 19. Puius YA, Zhao Y, Sullivan M, Lawrence DS, Almo SC, Zhang Z-Y (1997) Proc Natl Acad 5. Carlton VE, Hu X, Chokkalingam AP, Schrodi SJ, Brandon R, Alexander HC, Chang M, Sci USA 94:13420–13425. Catanese JJ, Leong DU, Ardlie KG, et al. (2005) Am J Hum Genet 77:567–581. 20. Zhang Z-Y, Walsh AB, Wu L, McNamara DJ, Dobrusin EM, Miller WT (1996) J Biol Chem 6. Smyth D, Cooper JD, Collins JE, Heward JM, Franklyn JA, Howson JM, Vella A, Nutland 271:5386–5392. S, Rance HE, Maier L, et al. (2004) Diabetes 53:3020–3023. 21. Salmeen A, Andersen JN, Myers MP, Tonks NK, Barford D (2000) Mol Cell 6:1401– 7. Kyogoku C, Langefeld CD, Ortmann WA, Lee A, Selby S, Carlton VE, Chang M, Ramos 1412. P, Baechler EC, Batliwalla FM, et al. (2004) Am J Hum Genet 75:504–507. 22. Zhang S, Zhang Z-Y (2007) Drug Discovery Today 12:373–381. 8. Cohen S, Dadi H, Shaoul E, Sharfe N, Roifman CM (1999) Blood 93:2013–2024. 23. Pearson RB, Kemp BE (1991) Methods Enzymol 200:62–81. 9. Gjorloff-Wingren A, Saxena M, Williams S, Hammi D, Mustelin T (1999) Eur J Immunol 24. Fujise A, Mizuno K, Ueda Y, Osada S, Hirai S, Takayanagi A, Shimizu N, Owada MK, 29:3845–3854. 10. Alonso A, Sasin J, Bottini N, Friedberg I, Friedberg I, Osterman A, Godzik A, Hunter T, Nakajima H, Ohno S (1994) J Biol Chem 269:31642–31648. Dixon J, Mustelin T (2004) Cell 117:699–711. 25. Herget T, Oehrlein SA, Pappin DJ, Rozengurt E, Parker PJ (1995) Eur J Biochem 233:448–457. 11. Cloutier J-F, Veillette A (1999) J Exp Med 189:111–121. 26. Garton AJ, Tonks NK (1994) EMBO J 13:3763–3771. 12. Wu J, Katrekar A, Honigberg LA, Smith AM, Conn MT, Tang J, Jeffery D, Mortara K, 27. Hasegawa K, Martin F, Huang G, Tumas D, Diehl L, Chan AC (2004) Science 303:685–689. Sampang J, Williams SR, et al. (2006) J Biol Chem 281:11002–11010. 28. Xu W, Harrison SC, Eck MJ (1997) Nature 385:595–602. 13. Vang T, Congia M, Macis MD, Musumeci L, Orru V, Zavattari P, Nika K, Tautz L, Tasken 29. Sicheri F, Moarefi I, Kuriyan J (1997) Nature 385:602–609. K, Cucca F, et. al. (2005) Nat Genet 37:1317–1319. 30. Trushin SA, Pennington KN, Carmona EM, Asin S, Savoy DN, Billadeau DD, Paya CV 14. Zhang Z-Y (2002) Annu Rev Pharmacol Toxicol 42:209–234. (2003) Mol Cell Biol 23:7068–7081. 15. Sun J-P, Fedorov AA, Lee SY, Guo XL, Shen K, Lawrence DS, Almo SC, Zhang Z-Y (2003) 31. Alonso A, Bottini N, Bruckner S, Rahmouni S, Williams S, Schoenberger SP, Mustelin T J Biol Chem 278:12406–12414. (2004) J Biol Chem 279:4922–4928.

19772 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0706233104 Yu et al. Downloaded by guest on September 29, 2021