Tyrosine Phosphatase Receptor Type γ Is a JAK Phosphatase and Negatively Regulates Leukocyte Activation

This information is current as Michela Mirenda, Lara Toffali, Alessio Montresor, Giovanni of October 1, 2021. Scardoni, Claudio Sorio and Carlo Laudanna J Immunol 2015; 194:2168-2179; Prepublished online 26 January 2015; doi: 10.4049/jimmunol.1401841 http://www.jimmunol.org/content/194/5/2168 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2015/01/23/jimmunol.140184 Material 1.DCSupplemental http://www.jimmunol.org/ References This article cites 46 articles, 16 of which you can access for free at: http://www.jimmunol.org/content/194/5/2168.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 1, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2015 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Protein Tyrosine Phosphatase Receptor Type g Is a JAK Phosphatase and Negatively Regulates Leukocyte Integrin Activation

Michela Mirenda,* Lara Toffali,*,† Alessio Montresor,*,† Giovanni Scardoni,† Claudio Sorio,* and Carlo Laudanna*,†

Regulation of networks depends on protein and phosphatase activities. Protein tyrosine of the JAK family have been shown to regulate integrin affinity modulation by chemokines and mediated homing to secondary lymphoid organs of human T lymphocytes. However, the role of protein tyrosine phosphatases in leukocyte recruitment is still elusive. In this study, we address this issue by focusing on protein tyrosine phosphatase receptor type g (PTPRG), a tyrosine phosphatase highly expressed in human primary monocytes. We developed a novel methodology to study the signaling role of receptor type tyrosine Downloaded from phosphatases and found that activated PTPRG blocks chemoattractant-induced b2 integrin activation. Specifically, triggering of LFA-1 to high-affinity state is prevented by PTPRG activation. High-throughput phosphoproteomics and computational analyses show that PTPRG activation affects the state of at least 31 signaling . Deeper examination shows that JAKs are critically involved in integrin-mediated monocyte adhesion and that PTPRG activation leads to JAK2 dephosphoryla- tion on the critical 1007–1008 phosphotyrosine residues, implying JAK2 inhibition and thus explaining the antiadhesive role of PTPRG. Overall, the data validate a new approach to study receptor tyrosine phosphatases and show that, by targeting JAKs, PTPRG downmodulates the rapid activation of integrin affinity in human monocytes, thus emerging as a potential novel critical http://www.jimmunol.org/ regulator of leukocyte trafficking. The Journal of Immunology, 2015, 194: 2168–2179.

ntegrin activation is a critical step in leukocyte recruitment regulators of integrin activation have been described, the mecha- (1). A number of protein and lipid kinases have been shown nisms regulating the on–off dynamics of kinase activity control- I to regulate the proadhesive signaling network triggered by ling integrin triggering by chemoattractants are unknown. Thus, chemoattractants (2). In this context, we have recently shown that considering the proadhesive role of JAKs, it is logical to expect protein tyrosine kinases (PTKs) of the JAK family are upstream that downmodulation of JAK activity could effect leukocyte ad-

transducers linking the chemokine receptor CXCR4 to the hier- hesion dynamics. This may suggest the possibility that specific by guest on October 1, 2021 archical activation of Rho and Rap small GTPases, thus control- protein tyrosine phosphatases could be involved in the overall ling integrin affinity upregulation and homing to secondary process of chemoattractant-induced leukocyte recruitment. lymphoid organs of T lymphocytes (3). Overall, although we are The protein tyrosine phosphatases (PTP) superfamily includes still far from a complete characterization of this complex signaling receptor-like PTP (RPTP) and nontransmembrane proteins for a mechanism, it is clear that leukocyte trafficking is under tight total of 38 coding (4, 5). Particularly, RPTPs display a modu- control of phosphorylating events. Notably, leukocyte adhesion lar structure including a variable extracellular region, consisting is a transient phenomenon showing oscillating dynamics, cycling of different domains possibly implicated in cell–cell and cell– between adhesion and de-adhesion states propaedeutic to cell matrix adhesive contacts, and an intracellular region commonly crawling, diapedesis, and chemotaxis. However, although negative shared with other components of the superfamily. The intracel- lular region is typically composed of two domains named D1 and D2 (intracellular domain [ICD]). The catalytic activity resides in *Division of General Pathology, Department of Pathology and Diagnostics, School the D1 domain, whereas the D2 domain is possibly involved in of Medicine, University of Verona, Verona 37134, Italy; and †Center for Biomedical Computing, University of Verona, Verona 37134, Italy substrate specificity, stability, and protein–protein interaction of Received for publication July 21, 2014. Accepted for publication December 30, 2014. RPTPs. The activity of RPTPs is controlled by a variety of mech- This work was supported by Italian Association for Cancer Research Grant IG 8690, anisms (5). Although not established for all RPTPs, the main reg- by Italian Ministry of Education, Universities and Research Grant PRIN 2009, as ulatory mechanism of RPTP activity consists of the reversible well as by the Nanomedicine Project of the University of Verona and the Fondazione transition from a homodimeric inactive form to a monomeric active Cariverona. form (6, 7). Oxidative stress is also an important regulatory mech- Address correspondence and reprint requests to Prof. Carlo Laudanna, Department of Pathology and Diagnostics, Division of General Pathology, University of Verona, anism of RPTPs, leading to homodimer stabilization and hence Strada le Grazie 8, Verona 37134, Italy. E-mail address: [email protected] to the inactivation of the itself (4). The online version of this article contains supplemental material. The lack of techniques allowing specific upmodulation of PTP Abbreviations used in this article: CPP, cell-penetrating peptide; eGFP, enhanced activity in primary cells has hampered investigating the regulatory GFP; HGNC, HUGO Nomenclature Committee; ICD, intracellular domain; implications of tyrosine phosphorylation/dephosphorylation turn- P1, Penetratin 1; PMN, polymorphonuclear cell; pNPP, p-nitrophenyl phosphate; PTK, protein ; PTP, protein tyrosine phosphatase; PTPRC, PTP type over under physiological conditions. Notably, it cannot be assumed C; PTPRG, PTP receptor type g; RPTP, receptor-like PTP; TKIP, tyrosine kinase a priori whether tyrosine dephosphorylation plays, at the cellular inhibitor peptide; WD, wedge domain. level, positive or negative regulatory roles, making unclear Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 which is the best approach to study tyrosine phosphatases under www.jimmunol.org/cgi/doi/10.4049/jimmunol.1401841 The Journal of Immunology 2169 physiological conditions. For instance, in a context where PTK ac- molecular mass, 3969.89 Da; isoelectric point, 12.02 (as calculated by tivity mediates a positive regulation, such as the role of JAKs in EMBOSS pepstats computation, http://www.ebi.ac.uk/Tools/seqstats/ integrin activation, an experimental approach based on chemical PTP- emboss_pepstats/). Peptide stock solutions (10 mM) in DMSO were kept at 280˚C and diluted in adhesion buffer immediately before the experiments. specific inhibitors or small interfering RNA silencing could be mis- All peptides were fully soluble in aqueous buffers. Standard treatment of leading. Alternatively, methods allowing selective activation of PTPs cells with peptides was for 1 h at 37˚C in 24- or 6-well plates. in primary cells are unavailable. To investigate the regulation of ty- PTPRG TAT-ICD, TAT-ICD_D1028A, and TAT–enhanced GFP (eGFP) rosine dephosphorylation in chemoattractant-triggered signal trans- fusion proteins were generated by using the pRSET-TATa expression vector, specifically designed to allow protein in-frame expression of TAT fusion duction in human leukocytes, we developed a novel approach based on proteins, as we previously described (10). Briefly, the cDNAs coding for the cell-penetrating peptides (CPPs) and allowing studying the effect of complete ICD of PTPRG (aa 797–1445) and for the complete eGFP sequence PTP receptor type g (PTPRG) activation in human primary monocyte were cloned into pRSET-TATa vector between BamHI and EcoRI sites. The ICD_D1028A cDNA was obtained by site-directed mutagenesis, wherein the adhesion. We demonstrate that specific activation of PTPRG tyro- 1028 sine phosphatase activity by means of two complementary CPP Asp codon was replaced with an Ala codon and verified by sequencing. Expression of recombinant TAT-ICD, TAT-ICD_D1028A, and TAT-eGFP fu- approaches leads to inactivation of LFA-1 high-affinity-state trig- sion proteins was controlled to allow purification under nondenaturing con- gering and mediated underflow cell arrest by chemoattractants. We ditions. Overnight cultures of the Escherichia coli strain BL21 (DE3) pLysS show that this effect is accompanied by a consistent remodeling of harboring recombinant plasmid expressing (His)6-TAT-ICD, (His)6-TAT- signaling phosphonetwork triggered by chemoattractants. More ICD_D1028A, and (His)6-TAT-eGFP were diluted in fresh Luria–Bertani medium containing ampicillin (50 mg/ml) and chloramphenicol (35 mg/ml) specifically, PTPRG activation leads to JAK2 dephosphorylation and andgrownat37˚CtoanOD600 of 0.6. Protein expression was induced inhibition, thus showing that JAKs are targets of PTPRG. These overnight at 16˚C with 0.1 mM isopropyl b-D-thiogalactoside. Harvested cells findings show that PTPRG is a novel critical regulator of leukocyte were resuspended in 1:50 culture volume of lysis buffer containing 20 mM tris, trafficking. More generally, the data describe a new approach suitable 500 mM NaCl, 10 mM imidazole [pH 7.3], and protease inhibitors without Downloaded from to study the functional interplay between tyrosine kinase and phos- EDTA. Samples were sonicated five times for 30 s at 4˚C and centrifuged for 15 min in a Sorval SS34 rotor at 12,500 rpm at 4˚C. After centrifugation, the phatase activities under physiological conditions. supernatants were 0.45 mm porous filtered and quantified by Bradford assay. BL21 lysates containing TAT-ICD_D1028A and TAT-eGFP were stored at Materials and Methods 220˚C. Lysates containing TAT-ICD protein were collected to the HisTrap HP Human primary cells and reagents (1 ml) charged with nickel ions. The clarified lysate was applied to the column

at 1 ml/min flow rate. After washing with binding buffer, TAT-ICD was eluted http://www.jimmunol.org/ Informed consent for this work was provided by the University of Verona; with binding buffer containing imidazole (500 mM), with a linear gradient the University of Verona Ethics Committee approved experimentation from 10 to 500 mM imidazole in 90 min at 1 ml/min flow rate. Protein pu- involving human primary cells. Human primary monocytes and poly- rification was monitored by evaluation of molecular mass on SDS-PAGE and morphonuclear cells were isolated from whole blood of healthy donors. by Western blot using an anti–His-(C-term)-HRP Ab (Invitrogen). Eluted Purity of monocyte preparation was evaluated by flow cytometry after proteins were concentrated on Amicon Ultra-15 centrifugal filter device staining with FITC-conjugated anti-CD14 Ab and was .95%. Isolated cells (Millipore) followed by a buffer exchange in storage buffer containing 100 mM were kept at 37˚C in standard adhesion buffer (PBS, 1 mM CaCl2,1mM arginine, 20 mM Tris, and 150 mM NaCl. Concentrated stock samples MgCl2, 10% FBS [pH 7.2]) and used within 1 h. FBS was from Lonza; of purified native TAT fusion proteins were stored at 220˚C. The tyrosine fMLF, human fibrinogen, ICAM-1, and E-selectin were from R&D Sys- phosphatase activity of the TAT-ICD preparations was systematically quan- tems; FITC-conjugated goat to mouse Ab was from Sigma-Aldrich; the tified by means of a p-nitrophenyl phosphate (pNPP) assay. Freshly purified mAb 327C detecting an LFA-1 epitope corresponding to high-affinity state TAT-ICD was dissolved in pNPP assay buffer (20 mM Tris, 10 mM DTT, and by guest on October 1, 2021 was provided by Dr. Kristine Kikly (Eli Lilly and Co.); tyrphostin AG490 2 mM pNPP) in the presence or absence of orthovanadate and the dephos- was from Sigma-Aldrich; WHI-P154 was from Calbiochem; the mono- phorylating activity was evaluated, after 30 min of incubation, at 405 nm by clonal anti-human PTPRG Ab was from R&D Systems; rabbit monoclonal a plate reader (Victor X5 multilabel plate reader, PerkinElmer). Standard anti-JAK2 (D2E12) was from Technology; rabbit mono- treatment of cells with 0.22-mm porous-filtered TAT-ICD was for 60 min at clonal anti–phospho-JAK2 (E132) was from EMD Millipore; anti-gp91phox 37˚C in 24- or 6-well plates. was provided by Dr. A.W. Segal (Department of Medicine, University College, London, U.K.) (8); 4-nitrophenyl phosphate disodium salt hexa- Membrane fraction isolation and pNPP assay hydrate was from Sigma-Aldrich. Freshly isolated monocytes and polymorphonuclear cells (PMNs) were Trojan CPP technology suspended in 100 mM PIPES, 12.5 mM EGTA, 300 mM KCl, 30 mM NaCl, and 35 mM MgCl2 (pH 7.3) (relaxation buffer) in the presence of proteases The control TATand Penetratin 1 (P1) peptides and the fusion P1-PTPRG wedge inhibitors. Cells were then sonicated twice (6 s, 15 W) and centrifuged at domain (P1-WD), P1-PTPRG wedge scrambled (P1-WD_scrambled), and P1- 2000 rpm for 10 min. The supernatants were collected and centrifuged at JAK2 (P1–tyrosine kinase inhibitor peptide [TKIP]) peptides were synthesized 28,000 rpm for 60 min. After centrifugation, the supernatants were dis- by GenScript. carded and the membrane pellets were dissolved in 20 mM Tris (pH 7.4). The P1-WD peptide (RQIKIWFQNRRMKWKKGKQFVKHIGELYS- Equal amounts of membrane fractions were added to pNPP buffer con- NNQHGFSEDFEEVQ) encompassed the complete P1 sequence (16 aa; taining 20 mM tris, 10 mM DTT, and 2 mM pNPP, after which DMSO and RQIKIWFQNRRMKWKK), an inserted glycine to allow flexibility of the peptides were added. After 30 min of incubation, levels of dephosphory- fusion peptide, and the computationally identified PTPRG WD sequence lation, in the presence or absence of orthovanadate, were measured at encompassing aa 831 to 856 of human PTPRG (26 aa; KQFVKHIGE- 405 nm by a plate reader (Victor X5 multilabel plate reader, PerkinElmer). LYSNNQHGFSEDFEEVQ) (see Results). Underlined is the amino acid sequence shared with the D1 phosphatase-active domain. The P1-WD pep- Western blot tide displayed the following properties: amino acids, 43, molecular mass, Cells were lysed in ice-cold 1% Nonidet P-40 buffer containing phosphatase 5396.14 Da; isoelectric point, 10.5714 (as calculated by EMBOSS pepstats inhibitors and complete protease inhibitor mixture (Roche). Total cell computation, http://www.ebi.ac.uk/Tools/seqstats/emboss_pepstats/). Resi- lysates were quantified by Bradford assay (Bio-Rad). Equal amounts of total due charge distribution and hydropathy analysis was performed with EM- lysates or of isolated membranes (see above) were subjected to 10% SDS- BOSS pepinfo computation (http://www.ebi.ac.uk/Tools/seqstats/emboss_ PAGE and then blotted. After incubation with anti-PTPRG, anti–gp91phox, pepinfo/). Accessible surface area (9) analysis was performed with the anti–phospho-JAK2, anti-JAK2, and HRP-linked secondary Abs (GE online algorithm ASAView (http://www.abren.net/asaview/). Crystallo- Healthcare), immunoreactive bands were visualized by ECL detection graphic rendering of PTPRG WD ( ID: 2nlk) was done (EMD Millipore). Intensities of band signals were quantified by densito- with PyMol software package (http://www.pymol.org). metric analysis (Quantity One, Bio-Rad) by using ImageQuant Las4000. The P1-TKIP peptide (RQIKIWFQNRRMKWKKGWLVFFVIFYFFR) encompassed the complete P1 sequence (16 aa), an inserted glycine to Static adhesion assay allow flexibility of the fusion peptide, and the TKIP sequence (12 aa; WLVFFVIFYFFR), specifically blocking JAK2 autophosphorylation (3). Human primary monocytes or PMNs were suspended at 5 3 106/ml in The P1-TKIP peptide displayed the following properties: amino acids, 29; standard adhesion buffer. Adhesion assays were performed on 18-well 2170 PTPRG NEGATIVELY REGULATES MONOCYTE ADHESION glass slides coated with 1 mg/ml human fibrinogen or with 1 mg/ml human (absorbance at 450 nm) with a microwell plate reader (Victor X5 multi- ICAM-1 in PBS; 20 ml cell suspension was added to the coated wells and label plate reader, PerkinElmer). stimulated for 1 min at 37˚C with 5 ml fMLF, 0.05 mM final concentration. After rapid washing, adherent cells were fixed in ice-cold 1.5% glutaral- Protein–protein interaction assay dehyde in PBS, and still images of adherent cells in 0.2-mm2 fields were 3 HIS-Select nickel affinity gel (30 ml; Sigma-Aldrich), previously washed acquired at 20 phase contrast magnification (numerical aperture, 0.40) three times with TBS and with bacterial lysis buffer, was incubated for 3 h with a charge-coupled device camera (ICD-42B; Ikegami) connected to an at 4˚C with 3 mg quantified supernatants expressing TAT-ICD_D1028A or inverted microscope (IX50; Olympus). Image acquisition and computer- TAT-eGFP under constant rotation. After 3 h, the nickel affinity gel was assisted enumeration of adherent cells were performed with ImageJ (Na- washed three times with monocyte lysis buffer (previously described). tional Institutes of Health). Each reaction, containing immobilized TAT-ICD_D1028A or TAT-eGFP in Underflow adhesion assay 30 ml HIS-Select nickel affinity gel, was incubated with 500 mg monocytic lysate treated with buffer (control) or with 50 nM fMLF for 3 h at 4˚C with Human primary monocytes were suspended at 1 3 106/ml in standard constant nutation. After incubation, beads were washed at least three times adhesion. Cell behavior in underflow conditions was analyzed with in monocyte lysis buffer and in TBS. Proteins were eluted from the beads the BioFlux 200 system (Fluxion Biosciences). Forty-eight–well plate in SDS sample buffer and resolved by SDS-PAGE and Western blot microfluidics were first cocoated overnight at room temperature with 2.5 analysis (13). mg/ml human E-selectin and 5 mg/ml human ICAM-1 in PBS. Immedi- ately before use, microfluidic channels were washed with PBS and then Statistical analysis coated with 10 nM fMLF in PBS for 3 h at room temperature. After ex- tensive washing of microfluidics with adhesion buffer, cells were fluxed at Statistical analysis was performed by calculating means and SD from 2 dynes/cm2 wall shear stress and the behavior of interacting cells was different experiments; significance was calculated by ANOVA followed by digitally recorded with a fast CCD video camera (25 frames/s). Single a Dunnett post hoc test versus control condition indicated in each graph. 2

areas of 0.2 mm were recorded for at least 120 s. Interactions of 40 ms or Downloaded from longer were considered significant and scored. Cells that remained firmly adherent for at least 1 s were considered fully arrested. Cells arrested for at Results least 1 s and then detached, or for 10 s and then remained adherent, were Identification and functional characterization of the PTPRG scored separately and plotted as independent groups. Interacting cell WD behaviors were automatically quantified with BeQuanti (http://fluxionbio. com/bioflux/). To develop a methodology allowing modulation of PTPRG en- zymatic activity in primary leukocytes, we exploited some prop- Measurement of LFA-1 affinity state erties of RPTPs. Although variability may occur among the http://www.jimmunol.org/ Cells suspended in standard adhesion buffer at 2 3 106/ml were stimulated different RPTP subgroups (5), the inhibitory receptor homo- under stirring at 37˚C for 1 min with 50 nM fMLF in the presence of 10 dimerization has been suggested to involve, under physiological mg/ml 327C mAb (reporter for fully extended conformation epitope cor- conditions, a region, called WD, partially conserved between the responding to a high-affinity state). After rapid washing, the cells were stained with FITC-conjugated secondary Ab and analyzed by cytofluo- different RPTPs and involved in intracellular domain interaction rimetric quantification. and consequent enzymatic inhibition of the D1 catalytic domain (6, 14). Notably, previous data suggest a unique capability of High-throughput phosphoproteomics and computational PTPRG intracellular domains to undergo in vitro dimerization analysis (15), thus suggesting that PTPRG is possibly expressed on cell by guest on October 1, 2021 To analyze the phosphoproteome we took advantage of the comprehensive plasma membrane as a phosphatase-inactive homodimer. There- high-throughput service developed by Kinexus (http://www.kinexus.ca). fore, we reasoned that interference by steric hindrance with the ∼ . The Kinexus Ab microarray includes 500 pan-specific and 350 WD could potentially lead to PTPRG monomerization and/or phosphosite-specific Abs per chip, in duplicate spots. The analysis covers a variety of selected cell signaling proteins regulating cell proliferation, catalytic domain separation, thus freeing the intrinsic tyrosine stress, apoptosis, adhesion, secretion, motility, and other biological pro- phosphatase activity by removing intermolecular D2 versus D1 cesses (see http://www.kinexus.ca/pdf/infopackage_kinex.pdf for a com- domain inhibition (15). Bioinformatics analysis based on com- plete list of identified proteins). Monocytes were treated with 50 mMP1or parison with previously characterized WDs in PTPRA (14), PTP P1-WD peptides for 60 min at 37˚C and then stimulated with 50 nM fMLF for 3 min. Cells were immediately lysed according to Kinexus protocol and receptor type C (PTPRC) (6), LAR (16), and PTPRM (16), per- the samples were sent to Kinexus for microarray analysis. Data, provided formed with BLASTp (http://www.ebi.ac.uk/Tools/sss/ncbiblast/) by e-mail, were previously normalized to control cells and statistically and PTP-dedicated online resources (http://ptp.cshl.edu), identi- filtered by Kinexus to identify the most significant best hits (array data are fied the PTPRG WD as a juxtamembrane 26-aa sequence en- ∼ available at http://dp.univr.it/ laudanna/LCTST/downloads/downloads-2/ compassing aa 831–856, with the last 11 aa belonging to the D1 index.html). The analysis was then focused on phosphotyrosine-specific Abs. Network inference analysis was performed within the Cytoscape 3.1 phosphatase-active domain (Fig. 1A). The PTPRG WD displays bioinformatics environment by using previously established interatomic a helix-turn-helix structure (Fig. 1B) and a high solvent accessi- datasets (http://dp.univr.it/∼laudanna/LCTST/downloads/index.html) compiled bility, as evidenced by residue analysis and hydropathy calculation from public databases. The compiled protein–protein interatomic dataset and by accessible surface area (9) calculations (see Materials and was manually curated to remove self-loops and duplicate edges; different protein IDs were normalized to HUGO Committee Methods), thus suggesting the capability of this region to support (HGNC) standard IDs. The set of 31 identified proteins whose tyrosine protein interactions. A fusion P1-WD Trojan peptide (17, 18) was phosphorylation status was affected by PTPRG activity was used as a then generated (see Materials and Methods) to allow overstepping bioinformatics probe allowing subnetwork reconstruction. Directed node cell plasma membrane, thus investigating the biological activity of centrality indexes were computed with the Cytoscape plug-in Centiscape the PTPRG WD peptide in primary cells. We first tested the 2.1 (11). Network modular decomposition was executed with the Cyto- scape plug-in MCODE (12). Literature data mining was performed by biochemical activity of the fusion peptide by performing in vitro multiple database interrogations, mainly based on PhosphoSite (http:// assays of phosphatase activity on purified membranes isolated www.phosphosite.org/), Phospho.ELM (http://phospho.elm.eu.org), Phos- from human primary monocytes compared with membranes iso- phoNet (http://www.phosphonet.ca), and KEA (http://amp.pharm.mssm. lated from human primary PMNs. Western blot analysis showed edu/lib/kea.jsp). that PTPRG is expressed in monocytes but absent in PMNs JAK2 activation assay (Fig. 1C), as also previously suggested (19) (see also http:// JAK2 activation was determined with ELISA JAK2 (pYpY1007/1008) humanproteomemap.org). Notably, Western blot of total monocyte (Invitrogen). Briefly, cells were stimulated with agonist and lysed follow- lysates reveals a double PTPRG band absent in the membrane ing the manufacturer’s protocol; levels of phospho-JAK2 were quantified fraction, likely due to differential PTPRG posttranslational mod- The Journal of Immunology 2171 Downloaded from http://www.jimmunol.org/ FIGURE 1. The P1-WD peptide triggers a specific tyrosine phosphatase activity in human monocyte membranes. (A) Diagram of PTPRG protein domains. (B) Crystallographic renderings of PTPRG D1 catalytic and wedge (red) domains (14–16). (C) Western blot of PTPRG expression in monocytes and PMNs. PTPRG expression was assessed in total cell lysates and membrane fractions with a mouse monoclonal anti-PTPRG Ab. Whole-cells lysates were probed with rabbit anti- Ab, whereas membrane fraction lysates were probed with rabbit anti-gp91phox as loading controls; a representative experiment of three independent experiments is shown. (D) pNPP assay on human monocyte and (E) PMN membranes. Isolated membranes were dissolved in pNPP buffer and treated with DMSO, with the indicated concentrations of P1-WD or with 200 mM P1 or P1-WD_scrambled peptides. Values are absorbance at 405 nm shown as percentage fold increase (FI) over the blank; mean values of three experiments are shown; error bars are SDs. *p , 0.05, ***p , 0.001 versus vehicle. by guest on October 1, 2021 ification, such as glycosylation (20). In contrast, other membrane (10, 21, 22). Indeed, treatment with P1-WD led to a consistent markers, such as CD45 (PTPRC), gp91phox (NADPH oxidase reduction of arrested cells accompanied by a corresponding in- component), and CD18 were correspondingly expressed in crease of rolling cells, as expected. This finding was confirmed in monocytes as well as in PMNs (not shown). The P1-WD peptide LFA-1 affinity assays by using a specific reporter mAb detecting elicited in a dose-dependent manner a consistent phosphatase a fully extended LFA-1 conformational epitope corresponding to activity in membrane isolated from monocytes (Fig. 1D). In heterodimer high-affinity state. Pretreatment with P1-WD led to contrast, control P1 and P1-WD_scrambled peptides were mini- a very consistent inhibition of LFA-1 triggering to high-affinity mally effective. Orthovanadate treatment completely prevented state by fMLF (Fig. 1D). In the assays, treatment with the P1 or phosphatase activity, supporting the triggering of a tyrosine P1-WD_scrambled peptides was completely ineffective, thus phosphatase activity (not shown). Furthermore, and very impor- supporting the specificity and excluding toxic effects of the P1- tantly, the P1-WD peptide did not elicit a consistent phosphatase WD peptide. Importantly, PMNs were completely unaffected by activity in membrane isolated from PMNs (Fig. 1E). Altogether, P1-WD treatment (not shown). Taken together, these data strongly these data corroborate our hypothesis, suggesting that the P1-WD suggest that activated PTPRG acts as a negative regulator of peptide may act as an activator of PTPRG tyrosine phosphatase signaling mechanisms controlling integrin activation and mediated activity, possibly by promoting PTPRG monomerization and D1 adhesion in human primary monocytes. To confirm these findings domain catalytic activity. with a complementary approach, we exploited a different Trojan Activation of PTPRG inhibits LFA-1 affinity triggering and peptide approach based on a TAT fusion protein technology we mediated adhesion by chemoattractants in human primary previously developed (10). A fusion recombinant protein was monocytes generated encompassing the complete PTPRG intracellular We then investigated the functional effect of PTPRG activation on phosphatase domains (ICD, aa 846–1407), including both D1 and integrin activation and mediated adhesion by chemoattractants D2 domains (Fig. 1A), fused, at the N-terminus, to the Trojan in human primary monocytes. Treatment of monocytes with the 13-aa TAT sequence, capable of transporting biologically active P1-WD peptide prevented in a dose-dependent manner fMLF- proteins into the cell (17). PTPRG TAT-ICD fusion protein triggered static adhesion to fibrinogen (Fig. 2A) as well as to retained tyrosine phosphatase activity, as verified in in vitro ICAM-1 (Fig. 2B), thus suggesting blockade of Mac1 (CD11b/ phosphatase assays in the absence and presence of orthovana- CD18), P150/95 (CD11c/CD18), and LFA-1 (CD11a/CD18) b2 date (Fig. 3A), thus showing the presence of soluble, catalytically integrin activation. The P1-WD peptide also prevented monocyte active, TAT-ICD monomers. We then investigated the functional arrest on ICAM-1 in underflow adhesion assays (Fig. 1C), a con- effect of PTPRG TAT-ICD on integrin activation and mediated dition dependent on rapid LFA-1 triggering to a high-affinity state adhesion by chemoattractants. Monocyte treatment with TAT-ICD 2172 PTPRG NEGATIVELY REGULATES MONOCYTE ADHESION

FIGURE 2. The P1-WD peptide inhibits LFA-1 affinity triggering and mediated adhesion by che- moattractants in human primary monocytes. Static adhesion assays on (A) fibrinogen or (B) ICAM-1 are shown. Human monocytes were treated with buffer (control), the indicated concentrations of P1-WD, or with 50 mM P1 or P1-WD_scrambled peptide for 1 h at 37˚C; adhesion was stimulated with 50 nM fMLF for 60 s; mean values of six experiments in triplicate are shown; errors bars are SDs. (C) Underflow adhesion assay. Human monocytes were treated with buffer (control), 50 mM P1-WD, P1, or P1-WD_scrambled peptides for 1 h at 37˚C. Shown are percentages of rolling and arrested cells for the indicated times over total interacting cells. Values are means of five experi- ments; errors bars show SDs. (D) LFA-1 high–af-

finity-state detection. Human monocytes were Downloaded from treated with 50 mM P1-WD, P1, or P1-WD_scram- bled peptides for 1 h at 37˚C and stimulated with 50 nM fMLF for 60 s. LFA-1 high-affinity-state conformer was detected by cytofluorimetric anal- ysis with the 327C Ab. Shown are percentages of fold increases over the control; mean values of

five experiments in triplicate are shown; errors bars http://www.jimmunol.org/ are SDs. **p , 0.01, ***p , 0.001 versus CTRL or P1. CTRL, control.

strongly prevented, in a dose-dependent manner, static adhesion investigate the influence of PTPRG on signaling tyrosine phos- to fibrinogen (Fig. 3B) as well as to ICAM-1 (Fig. 3C). TAT-ICD phonetworks. Thus, we focused the phosphoproteomics analysis also prevented monocyte arrest on ICAM-1 in underflow adhesion on the effect of P1-WD. We also focused on fMLF-potentiated by guest on October 1, 2021 assays (Fig. 3D). As above, this finding was finally confirmed in signaling networks, because this directly correlates with integrin LFA-1 affinity assays. Pretreatment with TAT-ICD led to a rather triggering and mediated rapid adhesion. PTPRG activation by the consistent inhibition of LFA-1 triggering to a high-affinity state by P1-WD peptide affected the tyrosine phosphorylation of several fMLF (Fig. 1E). In the assays, treatment with the TAT control signaling molecules. Data analysis identified 31 molecules whose peptide was completely ineffective. Altogether, these data show phosphorylation was modified in a statistically significant manner the coherence of the combined CPPs approach as a novel meth- (Table I). Interestingly, 16 proteins showed reduced tyrosine odology suitable to investigate the functional effect of PTPRG phosphorylation, possibly indicating a direct PTPRG targeting tyrosine phosphatase activity upregulation in primary cells. More activity. In contrast, 15 proteins showed an increased tyrosine importantly, the data clearly suggest that PTPRG is a novel neg- phosphorylation. This may suggest an indirect and more complex ative regulator of signaling mechanisms triggered by chemo- effect, possibly mediated by PTPRG-sensitive, concurrent, sig- attractants and controlling integrin activation and mediated naling mechanisms, thus indicating that PTPRG, beside a recruitment of human monocytes. dephosphorylating activity on direct targets, may also generate High-throughput phosphonetwork analysis reveals the complex nonlinear effects mediated by more complex signaling cascades. influence of PTPRG activity on chemoattractant signaling Literature data mining suggests that the functional activity of a subgroup of 16 proteins, including ABL1, BMX, BTK, CDK2, The previous data imply PTPRG as a downregulator of chemo- attractant signaling. To identify PTPRG targets in the context of CTTN, DAB1, ITGB1, JAK2, KDR, KIT, LIMK1, MET, fMLF signaling, we employed a high-throughput Ab array tech- PDGFRB, SHC1, SRC, and VCL, is inhibited by PTPRG acti- nology. Notably, although the two Trojan peptide approaches vation. Among these, inhibition of ABL1, BMX, BTK, DAB1, provided fully coherent functional results, the P1-WD approach is ITGB1, JAK2, KDR, KIT, LIMK1, MET, PDGFRB, SHC1, and potentially more specific than the TAT-ICD–based one. Indeed, the VCL correlates with tyrosine dephosphorylation. In contrast, SRC catalytic domains of many RPTPs appear extremely promiscuous inhibition correlates with hyperphosphorylation of the inhibitory 530 419 with respect to substrate specificity (15). Moreover, sequence Tyr residue and with dephosphorylation of the activatory Tyr . alignment shows more variability in the WDs versus D1/D2 Moreover, CDK2 and CTTN inhibition correlates with a hyper- 15 470 domains, which appear more conserved (see http://ptp.cshl.edu). phosphorylation of the inhibitory Tyr and Tyr , respectively. Consistently, the P1-WD did not elicit a relevant tyrosine phos- In contrast, a subgroup of 13 proteins, including BLNK, DOK2, phatase activity in membrane isolated from PMNs, which do ex- ERBB2, GRIN2B, INSR, PDGFRA, PRKCD, PXN, STAT1, press CD45 (PTPRC) but not PTPRG. These considerations may STAT2, STAT3, STAT5A, and ZAP70, appears to be activated suggest that modulation of PTPRG tyrosine phosphatase activity by PTPRG activity. Among these, activation of BLNK, DOK2, by means of P1-WD peptide may be a more specific approach to ERBB2, GRIN2B, INSR, PDGFRA, PRKCD, PXN, STAT1, The Journal of Immunology 2173 Downloaded from

FIGURE 3. The TAT-ICD fusion protein inhibits LFA-1 affinity triggering and mediated adhesion by chemoattractants in human primary monocytes. (A) pNPP assay with 50 mM TAT control peptide or with 50 mM TAT-ICD in the presence or absence of 200 mMNa3VO4. Substrate dephosphorylation was http://www.jimmunol.org/ detected by reading the absorbance at 405 nm after 30 min of incubation. Shown is percentage fold increase over the pNPP buffer; mean values of seven experiments in triplicate are shown; errors bars are SDs. Static adhesion assays on (B) fibrinogen or (C) ICAM-1 are shown. Human monocytes were treated with buffer (control) or the indicated concentrations of TAT-ICD or TAT peptide for 1 h at 37˚C and stimulated with 50 nM fMLF for 60 s; mean values of five experiments in triplicate are shown; errors bars are SDs. (D) Underflow adhesion assay. Human monocytes were treated with buffer (control), 50 mM TAT peptide, or 0.5 mM TAT-ICD for 1 h at 37˚C. Shown are percentages of rolling and arrested cells for the indicated times over total interacting cells; mean values of four experiments are shown; errors bars are SDs. (E) LFA-1 high-affinity-state detection. Human monocytes were treated with buffer (control), 50 mM TAT peptide, or 0.5 mM TAT-ICD for 1 h at 37˚C and stimulated with 50 nM fMLF for 60 s. LFA-1 high-affinity-state conformer was detected by cytofluorimetric analysis with the 327C Ab. Shown are percentages of fold increases over the control; mean values of six experiments are shown; errors bars are SDs. **p , 0.01, ***p , 0.001 versus CTRL or TAT. CTRL, control. by guest on October 1, 2021 STAT2, STAT3, and STAT5A correlates with tyrosine hyper- PTK2 appear as direct or indirect PTPRG targets, possibly result- phosphorylation. Moreover, activation of ZAP70 correlated with ing in inhibition (SRC, KIT, ABL1, BTK, CDK2, JAK2, KDR, dephosphorylation of the inhibitory Tyr292. For two proteins, PDGFRB) or activation (ZAP70, PDGFRA) of their functional PTK2 and EGFR, functional interpretation of phosphotyrosine activities. status was ambiguous. Network inference analysis shows that the 31 identified proteins generate a single network-connected com- JAK2 mediates LFA-1 affinity triggering and mediated ponent, with no isolated nodes, suggesting that PTPRG may affect adhesion by chemoattractants in human primary monocytes the function of an integrated signaling macromodule controlling The high-throughput phosphoproteomics analysis showed that many cellular functions (Fig. 4A). Graph topological analysis PTPRG activation leads to JAK2 protein tyrosine kinase dephos- performed by applying mathematical algorithms computing net- phorylation (Fig. 4, Table I). These data attracted our attention work node centrality indexes, categorizing proteins by topological because JAKs regulate, in human primary T lymphocytes, LFA-1 relevance (11, 23–26), and predicting functional significance (27) and VLA-4 activation by chemokines and mediated trafficking to shows that SRC appears by far the most connected and central secondary lymphoid organs (3). Thus, JAK2 dephosphorylation by signaling protein in the network, followed by KIT, PTK2, ITGB1, PTPRG could explain, at least partially, the antiadhesive role of STAT1, EGFR, PDGFRB, ABL1, and BTK (Supplemental PTPRG on monocytes. However, the role of JAKs in integrin Table I), suggesting that, in the context of chemoattractant sig- activation in primary monocytes is unknown. Furthermore, the naling, Src family kinases are possibly main functional targets capability of fMLF to trigger JAKs activation was never of PTPRG activity. Network modular decomposition (28–30) addressed. Thus, we first set out to define the role of JAK2 in allowed isolating two modules, prevalently including activated fMLF-triggered rapid adhesion in monocytes. Time course (Fig. 4B) or inhibited proteins (Fig. 4C). Thus, the two modules experiments, either performed with ELISA assay (Fig. 5A) or may potentially represent a more specific and structured context of Western blot (Fig. 5B, 5C), showed that JAK2 is effectively functional influence of PTPRG on chemoattractant signaling in phosphorylated and activated in monocytes stimulated by fMLF. human monocytes. Notably, the graph edge directionality also To evaluate whether JAK2 mediates fMLF-triggered rapid adhe- highlights the signaling flow in the modules, thus facilitating the sion in human monocytes, we tested the capability of three JAK inference of protein reciprocal functional influence. inhibitors, AG490, WHI-P154, and P1-TKIP peptide, to block Overall, the combined analysis shows that PTPRG activation adhesion triggering. All of the inhibitors completely prevented may determine a complex remodeling of chemoattractant-triggered fMLF-induced tyrosine autophosphorylation and activation of signaling network. The protein kinases SRC, KIT, ABL1, ZAP70, JAK, whereas the P1 control peptide was ineffective (Fig. 5D, 5E). BTK, CDK2, JAK2, KDR, PDGFRA, PDGFRB, PRKCD, and Importantly, the three JAK inhibitors blocked fMLF-induced ad- 2174 PTPRG NEGATIVELY REGULATES MONOCYTE ADHESION

Table I. Summary of Kinexus high-throughput phosphoproteomics data

HGNC Protein Functional Effect Symbols p-Site % p-Site % p-Site % p-Site % of P1-WD ABL1 Y412 240 Inhibition BLNK Y84 19 Activation BMX Y40 27 Inhibition BTK Y223 258 Inhibition CDK2 Y15 282 Inhibition CTTN Y470 21 Inhibition DAB1 Y198 234 Inhibition DOK2 Y142 9 Activation EGFR Y1068 213 Y1148 216 Y1173 68 Inhibition/inhibition/activation ERBB2 Y1248 34 Activation GRIN2B Y1474 29 Activation INSR Y999 58 Y1189/Y1190 27 Activation/activation ITGB1 Y783 211 Inhibition JAK2 Y1007/Y1008 233 Inhibition KDR Y1054 237 Y1054 + Y1059 22 Inhibition KIT Y703 260 Y730 247 Inhibition LIMK1 Y507 210 Inhibition MET Y1003 247 Inhibition Downloaded from PDGFRA Y742 133 Y754 86 Activation PDGFRB Y716 236 Inhibition PRKCD Y313 189 Activation PTK2 Y397 30 Y577 260 Y576 65 Y861 211 Activation/inhibition/activation/inhibition PXN Y118 144 Y31 19 Activation/activation SHC1 Y349 + Y350 210 Inhibition SRC Y419 26 Y530 94 Inhibition/inhibition http://www.jimmunol.org/ STAT1 Y701 38 Activation STAT2 Y690 61 Activation STAT3 Y705 182 Activation STAT5A Y694 50 Activation VCL Y821 220 Inhibition ZAP70 Y292 246 Y315 + Y319 12 Activation/activation The table shows the 31 identified proteins whose tyrosine phosphorylation is affected by PTPRG activation. Values reported in % columns are percentage increase or decrease of protein tyrosine phosphorylation upon fMLF triggering in P1-WD–treated versus P1-treated monocytes. For some proteins (EGFR, INSR, KDR, KIT, PDGFRA, PTK2, PXN, SRC, and ZAP70) multiple phosphotyrosine residues are detected. From left to right, columns are HGNC protein symbols (in alphabetical order), phosphosites (p-Sites), percentage changes of phosphorylation (induced by P1-WD), and the putative functional effect, inferred from literature data mining. EGFR and PTK2 are highlighted in bold type because the functional effect could not be unambiguously inferred. by guest on October 1, 2021 hesion to both fibrinogen (Fig. 6A, 6B) and ICAM-1 (Fig. 6B), phosphorylation may increase the binding affinity for PTPRG. Al- thus implying a role for JAKs in b2 integrin activation. JAK in- together, these data are qualitatively and quantitatively in keeping hibition also prevented rapid arrest on ICAM-1 in underflow with the high-throughput Ab array analysis and demonstrate that assays (Fig. 6C). Finally, JAK inhibitors blocked LFA-1 triggering JAK2 is a direct PTPRG target, thus indicating that the antiadhesive to high-affinity state (Fig. 6D). Taken together, these data show effect of PTPRG involves JAK2 tyrosine dephosphorylation. that JAK2 mediates fMLF-triggered signal transduction leading to human primary monocyte integrin activation and adhesion, thus implying JAKs as general upstream transducers controlling Discussion proadhesive chemoattractant signaling in human leukocytes. In this study we examined the role of RPTPs in the regulation of leukocyte recruitment. We focused the analysis on PTPRG and on JAK2 is a PTPRG target human primary monocytes triggered to LFA-1 high-affinity-state Finally, we wanted to confirm, with different techniques, that JAK2 upregulation by the FPR1 ligand fMLF. To investigate the func- is effectively a direct PTPRG target. First, we show in ELISA tional effect of PTPRG activation, we developed a CPP-based assays that in monocytes treated with the P1-WD peptide (Fig. 7A) approach allowing testing in primary cells the impact of PTPRG or with TAT-ICD (Fig. 7B) and then stimulated with fMLF, JAK2 activity on chemoattractant-triggered signaling networks. The tyrosine phosphorylation is effectively reduced, thus indicating results can be summarized as follows: 1) a combined CPP-based inhibition of the JAK2 tyrosine kinase activity by PTPRG acti- approach is suitable to study the regulatory role of PTPRG acti- vation. Second, we confirmed these data by Western blot analysis. vation on signal transduction in primary cells; 2) PTPRG appears Indeed, also in this case, TAT-ICD and P1-WD almost completely expressed in primary monocytes as an inactive homodimer, whose prevented fMLF-induced tyrosine phosphorylation of JAK2, catalytic activity can be upregulated by interfering with the putative whereas the control P1, P1-WD_scrambled, and TAT peptides PTPRG WD; 3) PTPRG activation inhibits chemoattractant- were completely ineffective (Fig. 7C, 7D). Finally, we performed induced LFA-1 affinity triggering and mediated adhesion in a protein–protein interaction assay based on a substrate trapping human primary monocytes; 4) PTPRG activation consistently af- approach by using the phosphatase-defective PTPRG D1028A fects the structure of chemoattractant-triggered phosphonetworks; mutant (31). This approach showed JAK2 direct interaction 5) JAK2 mediates LFA-1 activation and mediated adhesion in with TAT-ICD_D1028A, but not with the negative control human monocytes; and 6) JAK2 is a direct target of PTPRG ty- TAT-eGFP. Interestingly, we also observed an increased JAK2 rosine phosphatase activity, and this correlates with the anti- binding upon fMLF treatment, suggesting that the JAK2 tyrosine adhesive role of PTPRG. The Journal of Immunology 2175 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 4. Network reconstruction and modular decomposition of PTPRG-influenced protein tyrosine phosphonetwork. (A) The unique connected network component generated by the binary interactions between the 31 identified PTPRG targets. No isolated nodes are present. In blue are the 15 proteins whose signaling activity is inhibited by PTPRG activity; in red are the 14 proteins whose signaling activity is activated by PRPRG activity. The activation state of PTK2 and EGFR was not inferred. Edges between nodes are directed and indicate the direction of signaling flow within the network, thus sug- gesting the direction of functional influence between the identified proteins. (B and C) Topological clusters of interacting nodes obtained by means of network modular decomposition; module (B) (4.6 clustering score) and module (C) (5.5 clustering score) prevalently include activated or inhibited proteins, respectively. In blue are proteins whose signaling activity is inhibited by PTPRG activity; in red are proteins whose signaling activity is activatedby PRPRG activity.

Receptor homodimerization is widely recognized as the inhib- activity only in membranes isolated from PTPRG-expressing itory mechanism of RPTP catalytic activity. Depending on the monocytes and, accordingly, the P1-WD peptide was active only RPTP subtype, proposed models either suggest a constitutive on monocyte adhesion, showing no effect in PMNs. Moreover, at homodimerization under physiological conditions, with ligand the cellular level, P1-WD and TAT-ICD had the same effects on binding leading to dimer dissociation and release of the intrinsic cell adhesion and on LFA-1 affinity triggering by fMLF. Finally, phosphatase activity, or a ligand-induced dimerization, thus leading both P1-WD and TAT-ICD induced JAK2 tyrosine dephosphory- to phosphatase activity inhibition (5, 32). Data obtained in dif- lation on the same amino acid residues, and its effect on cell ferent cellular contexts suggest a role for a juxtamembrane WD in adhesion fully correlates with the effect of JAK inhibitors. Alto- homodimerization (5). In the present study, we applied a CPP- gether, these data suggest a model in which PTPRG is constitu- based approach to investigate the function of PTPRG in primary tively expressed on monocyte plasma membrane as a homodimer leukocytes. The P1-WD peptide activated a tyrosine phosphatase with the WD involved in catalytic domain blockade. Notably, 2176 PTPRG NEGATIVELY REGULATES MONOCYTE ADHESION Downloaded from

FIGURE 5. JAK2 PTK is activated by fMLF stimulation. Monocytes were treated with buffer (Resting) or with 50 nM fMLF for the indicated times. (A) Time course of JAK2 activation by phospho-JAK2 ELISA assay; mean values of 3 experiments in duplicate are shown; errors bars are SDs. (B) Time course C of JAK2 activation by Western blot analysis (representative experiment of three). ( ) Western blot immuoreactive band quantification; mean values of the http://www.jimmunol.org/ three experiments are shown; the relative ratio of the band intensity of phospho-JAK2 was normalized to the level of JAK2 band intensities. (D) Effect of JAKs inhibitors on JAK2 activation. Monocytes were treated with buffer (Resting and control), 100 mM AG490, 40 mM P1 or P1-TKIP peptides, or 200 mM WHI-P154 for 1 h at 37˚C and stimulated with 50 nM fMLF for 180 s. Cell lysates were probed with anti–phospho-JAK2 and with anti-JAK2; a repre- sentative experiment of three is shown. (E) Western blot immuoreactive band quantification; the relative ratio of the band intensity of phospho-JAK2 was normalized to the level of JAK2 band intensities; mean values of three experiments are shown. ***p , 0.001 versus Resting, CTRL, or P1. CTRL, control. among the various RPTPs, PTPRG has the highest enzymatic obtained under fully physiological conditions, we hypothesize that activity (15), thus suggesting the importance of tight regulation by the juxtamembrane PTPRG WD may assume a regulatory rele-

dimerization. Although previous data show a peculiar tendency of vance once the molecule is expressed in the lipid membrane en- by guest on October 1, 2021 PTPRG catalytic domains to dimerize (15), in keeping with our vironment. The importance of the lipid environment is also model, structural data obtained in a cell-free system challenged suggested by the capability of triphosphorylated nucleosides to the importance of the WD (15). This discrepancy could reflect upregulate PTPRG activity only in the context of cell membranes differences in the experimental setting. Because our data were (33). This was also previously conjectured for other RPTPs (15).

FIGURE 6. JAK2 PTKs control LFA-1 high- affinity-state triggering and mediated monocyte adhesion by fMLF. Static adhesion assays on (A) fibrinogen or (B) ICAM-1 are shown. Human monocytes were treated with buffer (control), 40 mM P1 or P1-TKIP peptides, 100 mMAG490,or200 mM WHI-P154 for 1 h at 37˚C and stimulated with 50 nM fMLF for 60 s; mean values of six experi- ments in triplicates are shown; errors bars are SDs. (C) Underflow adhesion assay. Human monocytes were treated with buffer (control), 100 mM AG490, 40 mM P1 or P1-TKIP peptides, or 200 mMWHI- P154for1hat37˚C.Shownarepercentagesof rolling and arrested cells for the indicated times over total interacting cells; mean values of five experi- ments are shown; errors bars are SDs. (D)LFA-1 high-affinity-state detection. Human monocytes were treated with buffer (control), 40 mMP1orP1-TKIP peptides, 100 mMAG490,or200mM WHI-P154 for 1 h at 37˚C and stimulated with 50 nM fMLF for 60 s. LFA-1 high-affinity-state conformer was de- tected by cytofluorimetric analysis with 327C Ab. Shown are percentages of fold increases over the control; mean values of five experiments are shown; errors bars show SDs. **p , 0.01, ***p , 0.001 versus CTRL or P1. CTRL, control. The Journal of Immunology 2177 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 7. JAK2 is direct substrate of PTPRG tyrosine phosphatase activity. Phospho-JAK2 ELISA assays on human monocytes treated with (A)buffer (control), 50 mM P1-WD, or P1 or P1-WD_scrambled peptides or (B) with buffer (control), 0.5 mM TAT-ICD, or 50 mM TAT peptide for 1 h at 37˚C and stimulated with 50 nM fMLF for 90 s are shown. Shown are percentages of fold increases over the control; mean values of three experiments in triplicates are shown; errors bars are SDs. (C) Effect of TAT-ICD and P1-WD on JAK2 activation evaluated by Western blot analysis. Monocytes were treated with buffer (Resting, control), 0.5 mM TAT-ICD, 50 mM TAT peptide, 50 mM P1-WD, P1, or P1-WD_scrambled peptides for 1 h at 37˚C and stimulated with 50 nM fMLF for 180 s. Total lysates were probed with anti–phospho-JAK2 or with anti-JAK2; a representative experiment of three is shown. (D) Western blot immunoreactive band quantification; the relative ratio of the band intensity of phospho-JAK2 was normalized to the level of JAK2 band intensities; mean values of three experiments are shown. (E) JAK2 interacts with TAT-ICD_D1028A. Immobilized TAT-ICD_D1028A and TAT-eGFP were incubated with total cell lysates derived from monocytes treated with buffer or with 50 nM fMLF. JAK2 bound to the affinity resin was detected by Western blot with anti-JAK2 Ab. (F) Western blot immunoreactive band quantification; mean values of three experiments are shown; errors bars are SDs. ***p , 0.001 for resting TAT-ICD_D1028A versus resting TAT-eGFP, ***p , 0.001 for fMLF 50 nM TAT-ICD_D1028A versus fMLF 50 nM TAT-eGFP. **p , 0.01, ***p , 0.001 versus CTRL, TAT, or P1. CTRL, control. 2178 PTPRG NEGATIVELY REGULATES MONOCYTE ADHESION

A complex signaling network, consisting of at least 67 molecules The high-throughput phosphoproteomics analysis showed that (2), allows adaptation of circulating leukocytes to variable envi- PTPRG may affect the tyrosine phosphorylation of several sig- ronments, thus ensuring the completion of the migration process naling proteins. Besides JAK2, which was more deeply analyzed in response to chemoattractants. Notably, most of the studies de- owing to its known role in integrin affinity triggering (3), many scribe proadhesive signaling events, even when antiadhesive sig- tyrosine kinases appear to be influenced by PTPRG activation. nals likely have an equally relevant role in leukocyte recruitment, The combined phosphoproteomics and network topological anal- for example, during the on–off turnover of integrin activation ysis suggest a focus on SRC, KIT, PTK2, EGFR, PDGFRB, necessary for cell motility and transmigration. To date, few sig- ABL1, and BTK. Among these, SRC, KIT, PDGFRB, ABL1, and naling molecules have been shown to negatively regulate leuko- BTK appear unambiguously inhibited by PTPRG. In the context cyte integrin activation; these include PKA (34), RHOH (35), of LFA-1 activation, it is reasonable to exclude SRC (37) and also CDC42 (10, 36), SRC (37–39), H-RAS (40), ILK (41), FAM65B PDGFRB, which do not mediate chemoattractant signaling. Taken (42), and CAPN2 (43), with only CDC42 investigated according together, these observations suggest that KIT, ABL1, and BTK to the previously formalized four criteria (2). In the present study, may be involved in LFA-1 affinity modulation by chemo- fully compliant with these criteria, we show that PTPRG is a novel attractants, perhaps concurrently cooperating with JAKs. Inter- negative regulator of LFA-1 high-affinity-state triggering and estingly, ABL1 was already shown to be a direct PTPRG target mediated arrest by chemoattractants in human primary monocytes. (45). A role for these kinases in LFA-1 affinity modulation is Notably, PTKs of the JAK and SRC families have a regulatory role unknown, and it will be of interest to investigate their involvement in LFA-1 affinity triggering, with JAKs showing a positive role in leukocyte trafficking. (3), whereas SRCs possibly have a negative role (37). In our In conclusion, our study describes a novel approach to inves- context, SRC appears inhibited by PTPRG activation (Table I), tigate the regulatory role of PTPRG in primary cells, based on Downloaded from thus making it unlikely that the antiadhesive effect of PTPRG is activation of the tyrosine phosphatase activity. By exploiting this mediated by SRC activation. Then, the most obvious possibility approach, we show that PTPRG is a JAK phosphatase and nega- was that PTPRG could interfere with JAK activity. We tested this tively regulates chemoattractant signaling to LFA-1 affinity up- hypothesis by first showing that JAKs, as previously demonstrated regulation in human primary monocytes, thus suggesting a criti- in primary T lymphocytes (3), are critically involved in LFA-1 cal role for PTPRG in immune system regulation. Besides affinity triggering also in monocytes. Following this, the combined chemoattractant-triggered signaling leading to monocyte integrin http://www.jimmunol.org/ phosphoproteomics analysis showed that PTPRG activity induces activation, our study suggests that PTPRG could affect many JAK2 dephosphorylation on the adjacent 1007–1008 tyrosine signaling pathways also in other cellular contexts in which a dys- residues. Because tyrosine 1007 and 1008 are sites of JAK2 regulated tyrosine kinase activity possibly exists, for example, in autophosphorylation critical to kinase activity (44), this suggests neoplastic cells (28, 46). Notably, many neoplastic cells modulate that JAK2 is directly dephosphorylated and inhibited by PTPRG. the expression of PTPRG (19). In this context, PTPRG has been This result is also in keeping with the observation that PTPRG recently proposed as a tumor suppressor gene involved in the contains a secondary substrate-binding pocket (15) capable of pathogenesis of CML (45). Furthermore, KIT and ABL1, PTPRG interacting with proteins containing adjacent phosphotyrosines, targets, are . Thus, it will be of great relevance to ex- by guest on October 1, 2021 such as JAK2 upon fMLF triggering. Importantly, this was fully ploit our CPP-based approach to study the role of PTPRG in confirmed by the capability of the substrate-trapping PTPRG cancer development and progression. mutant (TAT-ICD_D1028A) to bind with higher efficiency tyro- sine phosphorylated JAK2 (Fig. 7E, 7F). Notably, previous data Acknowledgments obtained by means of genetic deletion suggested that JAKs could We are grateful to Dr. Kristine Kikly (Eli Lilly and Company) for providing be targets of CD45 (PTPRC), although not in the context of the anti–LFA-1 mAbs 327C and 327A. We thank the Center for Biomed- chemoattractant signaling (9). To our knowledge, our data, based ical Computing of the University of Verona for providing the Victor X5, on CPP technology, provide the first direct demonstration in pri- BioFlux, and computational facilities. mary leukocytes that JAKs are target of PTPRG-induced de- phosphorylation, clearly indicating that the antiadhesive role of Disclosures PTPRG is, at least partially, mediated by direct JAK2 blockade. The authors have no financial conflicts of interest. Considering that PTPRG expression appears mainly restricted to monocytes and B cells, with no expression in PMNs or other leukocytes (19) (see also http://humanproteomemap.org), our data References 1. Ley, K., C. Laudanna, M. I. Cybulsky, and S. Nourshargh. 2007. Getting to the show a cell-specific mechanism of downmodulation of LFA-1 site of inflammation: the leukocyte adhesion cascade updated. Nat. Rev. Immu- affinity triggering and mediated cell recruitment by PTPRG. No- nol. 7: 678–689. tably, although we circumvented the constitutive inhibition with 2. Montresor, A., L. Toffali, G. Constantin, and C. Laudanna. 2012. Chemokines and the signaling modules regulating integrin affinity. Front. Immunol. 3: 127. the CPP approach, a relevant, still open question concerns the 3. Montresor, A., M. Bolomini-Vittori, L. Toffali, B. Rossi, G. Constantin, and modality of PTPRG activation under physiological conditions, C. Laudanna. 2013. JAK tyrosine kinases promote hierarchical activation of Rho such as during monocyte adhesion in the microcirculation. The and Rap modules of integrin activation. J. Cell Biol. 203: 1003–1019. 4. Tonks, N. K. 2006. Protein tyrosine phosphatases: from genes, to function, to extracellular segment of PTPRG includes one disease. Nat. Rev. Mol. Cell Biol. 7: 833–846. domain and one fibronectin type 3 domain. One interesting pos- 5. Tonks, N. K. 2013. Protein tyrosine phosphatases—from housekeeping sibility is that extracellular ligands may interact with the PTPRG to master regulators of signal transduction. FEBS J. 280: 346–378. 6. Majeti, R., A. M. Bilwes, J. P. Noel, T. Hunter, and A. Weiss. 1998. fibronectin type 3 domain, a motif commonly found in adhesion Dimerization-induced inhibition of receptor protein tyrosine phosphatase func- molecules, in turn dissociating the homodimer and thus freeing the tion through an inhibitory wedge. Science 279: 88–91. 7. Jiang,G.,J.denHertog,J.Su,J.Noel,J.Sap,andT.Hunter.1999.Dimerizationinhibits phosphatase activity (32). It is tempting to speculate that under the activity of receptor-like protein-tyrosine phosphatase-a. Nature 401: 606–610. these conditions PTPRG, once activated, could provide a stop 8. Mazzi, P., M. Donini, D. Margotto, F. Wientjes, and S. Dusi. 2004. IFN-g signal to monocyte recruitment, thus behaving as an important induces gp91phox expression in human monocytes via C- dependent phosphorylation of PU.1. J. Immunol. 172: 4941–4947. mechanism of immune system modulation. This interesting hy- 9. Irie-Sasaki, J., T. Sasaki, W. Matsumoto, A. Opavsky, M. Cheng, G. Welstead, pothesis will be the focus of further investigation. E. Griffiths, C. Krawczyk, C. D. Richardson, K. Aitken, et al. 2001. CD45 is The Journal of Immunology 2179

a JAK phosphatase and negatively regulates cytokine receptor signalling. Nature 29. Segal, E., M. Shapira, A. Regev, D. Pe’er, D. Botstein, D. Koller, and 409: 349–354. N. Friedman. 2003. Module networks: identifying regulatory modules and their 10. Bolomini-Vittori, M., A. Montresor, C. Giagulli, D. Staunton, B. Rossi, condition-specific regulators from data. Nat. Genet. 34: 166– M. Martinello, G. Constantin, and C. Laudanna. 2009. Regulation of conformer- 176. specific activation of the integrin LFA-1 by a chemokine-triggered Rho signaling 30. Cho, Y. R., and A. Zhang. 2010. Identification of functional hubs and modules by module. Nat. Immunol. 10: 185–194. converting interactome networks into hierarchical ordering of proteins. BMC 11. Scardoni, G., M. Petterlini, and C. Laudanna. 2009. Analyzing biological net- Bioinformatics 11(Suppl. 3): S3. work parameters with CentiScaPe. Bioinformatics 25: 2857–2859. 31. Blanchetot, C., M. Chagnon, N. Dube´, M. Halle´, and M. L. Tremblay. 2005. 12. Bader, G. D., and C. W. Hogue. 2003. An automated method for finding mo- Substrate-trapping techniques in the identification of cellular PTP targets. lecular complexes in large protein interaction networks. BMC Bioinformatics 4: Methods 35: 44–53. 2. 32. Burridge, K., S. K. Sastry, and J. L. Sallee. 2006. Regulation of cell adhesion by 13. Nguyen, T. N., and J. A. Goodrich. 2006. Protein-protein interaction assays: protein-tyrosine phosphatases. I. Cell-matrix adhesion. J. Biol. Chem. 281: eliminating false positive interactions. Nat. Methods 3: 135–139. 15593–15596. 14. Bilwes, A. M., J. den Hertog, T. Hunter, and J. P. Noel. 1996. Structural basis for 33. Sorio, C., J. Mendrola, Z. Lou, S. LaForgia, C. M. Croce, and K. Huebner. 1995. inhibition of receptor protein-tyrosine phosphatase-a by dimerization. Nature Characterization of the receptor protein tyrosine phosphatase gene product 382: 555–559. PTPg: binding and activation by triphosphorylated nucleosides. Cancer Res. 55: 15. Barr, A. J., E. Ugochukwu, W. H. Lee, O. N. King, P. Filippakopoulos, I. Alfano, 4855–4864. € P. Savitsky, N. A. Burgess-Brown, S. Muller, and S. Knapp. 2009. Large-scale 34. Laudanna, C., J. J. Campbell, and E. C. Butcher. 1997. Elevation of intracellular structural analysis of the classical human protein tyrosine phosphatome. Cell cAMP inhibits RhoA activation and integrin-dependent leukocyte adhesion in- 136: 352–363. duced by chemoattractants. J. Biol. Chem. 272: 24141–24144. 16. Xie, Y., S. M. Massa, S. E. Ensslen-Craig, D. L. Major, T. Yang, M. A. Tisi, 35. Cherry, L. K., X. Li, P. Schwab, B. Lim, and L. B. Klickstein. 2004. RhoH is V. D. Derevyanny, W. O. Runge, B. P. Mehta, L. A. Moore, et al. 2006. Protein- required to maintain the integrin LFA-1 in a nonadhesive state on lymphocytes. tyrosine phosphatase (PTP) wedge domain peptides: a novel approach for in- Nat. Immunol. 5: 961–967. hibition of PTP function and augmentation of protein-tyrosine kinase function. J. 36. Kempf, T., A. Zarbock, C. Widera, S. Butz, A. Stadtmann, J. Rossaint, Biol. Chem. 281: 16482–16492. M. Bolomini-Vittori, M. Korf-Klingebiel, L. C. Napp, B. Hansen, et al. 2011. 17. Madani, F., S. Lindberg, U. Langel, S. Futaki, and A. Gra¨slund. 2011. Mecha-

GDF-15 is an inhibitor of leukocyte integrin activation required for survival after Downloaded from nisms of cellular uptake of cell-penetrating peptides. J. Biophys. 2011: 414729. myocardial infarction in mice. Nat. Med. 17: 581–588. 18. Wang, F., Y. Wang, X. Zhang, W. Zhang, S. Guo, and F. Jin. 2014. Recent 37. Giagulli, C., L. Ottoboni, E. Caveggion, B. Rossi, C. Lowell, G. Constantin, progress of cell-penetrating peptides as new carriers for intracellular cargo de- C. Laudanna, and G. Berton. 2006. The Src family kinases Hck and Fgr are livery. J. Control. Release 174: 126–136. dispensable for inside-out, chemoattractant-induced signaling regulating b 19. Vezzalini, M., A. Mombello, F. Menestrina, A. Mafficini, M. Della Peruta, C. van 2 integrin affinity and valency in neutrophils, but are required for b integrin- Niekerk, M. Barbareschi, A. Scarpa, and C. Sorio. 2007. Expression of trans- 2 mediated outside-in signaling involved in sustained adhesion. J. Immunol. 177: membrane protein tyrosine phosphatase g (PTPg) in normal and neoplastic 604–611. human tissues. Histopathology 50: 615–628. 38. Nakata, Y., B. Tomkowicz, A. M. Gewirtz, and A. Ptasznik. 2006. Integrin in-

20. Steentoft, C., S. Y. Vakhrushev, H. J. Joshi, Y. Kong, M. B. Vester-Christensen, http://www.jimmunol.org/ hibition through Lyn-dependent cross talk from CXCR4 chemokine receptors in K. T. Schjoldager, K. Lavrsen, S. Dabelsteen, N. B. Pedersen, L. Marcos-Silva, + et al. 2013. Precision mapping of the human O-GalNAc glycoproteome through normal human CD34 marrow cells. Blood 107: 4234–4239. SimpleCell technology. EMBO J. 32: 1478–1488. 39. Zou, J. X., Y. Liu, E. B. Pasquale, and E. Ruoslahti. 2002. Activated SRC on- 21. Constantin, G., M. Majeed, C. Giagulli, L. Piccio, J. Y. Kim, E. C. Butcher, and cogene phosphorylates R-ras and suppresses integrin activity. J. Biol. Chem. 277: 1824–1827. C. Laudanna. 2000. Chemokines trigger immediate b2 integrin affinity and mobility changes: differential regulation and roles in lymphocyte arrest under 40. Weber, K. S., G. Ostermann, A. Zernecke, A. Schro¨der, L. B. Klickstein, and flow. Immunity 13: 759–769. C. Weber. 2001. Dual role of H-Ras in regulation of lymphocyte function 22. Giagulli, C., E. Scarpini, L. Ottoboni, S. Narumiya, E. C. Butcher, G. Constantin, antigen-1 activity by stromal cell-derived factor-1a: implications for leukocyte and C. Laudanna. 2004. RhoA and z PKC control distinct modalities of LFA-1 transmigration. Mol. Biol. Cell 12: 3074–3086. activation by chemokines: critical role of LFA-1 affinity triggering in lympho- 41. Friedrich, E. B., S. Sinha, L. Li, S. Dedhar, T. Force, A. Rosenzweig, and cyte in vivo homing. Immunity 20: 25–35. R. E. Gerszten. 2002. Role of integrin-linked kinase in leukocyte recruitment. J.

23. Jeong, H., S. P. Mason, A. L. Baraba´si, and Z. N. Oltvai. 2001. Lethality and Biol. Chem. 277: 16371–16375. by guest on October 1, 2021 centrality in protein networks. Nature 411: 41–42. 42. Rougerie, P., Q. Largeteau, L. Megrelis, F. Carrette, T. Lejeune, L. Toffali, 24. Gu, Z., J. Liu, K. Cao, J. Zhang, and J. Wang. 2012. Centrality-based pathway B. Rossi, M. Zeghouf, J. Cherfils, G. Constantin, et al. 2013. Fam65b is a new enrichment: a systematic approach for finding significant pathways dominated by transcriptional target of FOXO1 that regulates RhoA signaling for T lymphocyte key genes. BMC Syst. Biol. 6: 56. migration. J. Immunol. 190: 748–755. 25. Scardoni, G., A. Montresor, G. Tosadori, and C. Laudanna. 2014. Node inter- 43. Svensson, L., A. McDowall, K. M. Giles, P. Stanley, S. Feske, and N. Hogg. ference and robustness: performing virtual knock-out experiments on biological 2010. Calpain 2 controls turnover of LFA-1 adhesions on migrating networks: the case of leukocyte integrin activation network. PLoS ONE 9: T lymphocytes. PLoS ONE 5: e15090. e88938. 44. Feng, J., B. A. Witthuhn, T. Matsuda, F. Kohlhuber, I. M. Kerr, and J. N. Ihle. 26. Raman, K., N. Damaraju, and G. K. Joshi. 2014. The organisational structure of 1997. Activation of Jak2 catalytic activity requires phosphorylation of Y1007 in protein networks: revisiting the centrality-lethality hypothesis. Syst. Synth. Biol. the kinase activation loop. Mol. Cell. Biol. 17: 2497–2501. 8: 73–81. 45. Della Peruta, M., G. Martinelli, E. Moratti, D. Pintani, M. Vezzalini, 27. De Franceschi, L., G. Scardoni, C. Tomelleri, A. Danek, R. H. Walker, A. Mafficini, T. Grafone, I. Iacobucci, S. Soverini, M. Murineddu, et al. 2010. H. H. Jung, B. Bader, S. Mazzucco, M. T. Dotti, A. Siciliano, et al. 2012. Protein tyrosine phosphatase receptor type g is a functional tumor suppressor Computational identification of phospho-tyrosine sub-networks related to gene specifically downregulated in chronic myeloid leukemia. Cancer Res. 70: acanthocyte generation in neuroacanthocytosis. PLoS ONE 7: e31015. 8896–8906. 28. Hanahan, D., and R. A. Weinberg. 2011. Hallmarks of cancer: the next gener- 46. Hanahan, D., and R. A. Weinberg. 2000. The hallmarks of cancer. Cell 100: ation. Cell 144: 646–674. 57–70.