Resource

Phenotypic and Interaction Profiling of the Human Identifies Diverse Mitotic Regulators

Graphical Abstract Authors Nicole St-Denis, Gagan D. Gupta, Zhen Yuan Lin, ..., Sally W.T. Cheung, Laurence Pelletier, Anne-Claude Gingras

Correspondence [email protected] (L.P.), [email protected] (A.-C.G.)

In Brief St-Denis et al. have systematically identified protein-protein interactions with phosphatases, important in cancer and other diseases. By combining this analysis with a functional screen, they uncover several tyrosine phosphatases with roles in cell division, providing potential targets for antimitotic cancer drugs.

Highlights d The interactome contains 1,335 interactions made by 140 phosphatases d Mitosis requires phosphatases of all evolutionary lineages, including several PTPs d PTPs play diverse roles in mitosis, notably in adhesion and centriole biogenesis d Combining complementary screens provides testable hypotheses about protein function

St-Denis et al., 2016, Cell Reports 17, 2488–2501 November 22, 2016 ª 2016 The Author(s). http://dx.doi.org/10.1016/j.celrep.2016.10.078 Cell Reports Resource

Phenotypic and Interaction Profiling of the Human Phosphatases Identifies Diverse Mitotic Regulators

Nicole St-Denis,1 Gagan D. Gupta,1 Zhen Yuan Lin,1 Beatriz Gonzalez-Badillo,1,3 Amanda O. Veri,2 James D.R. Knight,1 Dushyandi Rajendran,1 Amber L. Couzens,1 Ko W. Currie,2 Johnny M. Tkach,1 Sally W.T. Cheung,1 Laurence Pelletier,1,2,* and Anne-Claude Gingras1,2,4,* 1Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, ON M5G 1X5, Canada 2Department of Molecular Genetics, University of Toronto, Toronto, ON M5S 1A8, Canada 3Present address: Batchelor Children’s Institute, Miller School of Medicine, University of Miami, Miami, FL 33136, USA 4Lead Contact *Correspondence: [email protected] (L.P.), [email protected] (A.-C.G.) http://dx.doi.org/10.1016/j.celrep.2016.10.078

SUMMARY There are 150 in the encoding proteins containing domains consistent with protein dephos- Reversible phosphorylation is a fundamental regula- phorylation (reviewed in Duan et al., 2015). This capability tory mechanism, intricately coordinated by kinases evolved independently on at least four occasions, resulting in and phosphatases, two classes of enzymes widely families with unique catalytic mechanisms and varied substrate disrupted in human disease. To better understand specificity (reviewed in Moorhead et al., 2009). Serine and thre- the functions of the relatively understudied phospha- onine residues are mainly dephosphorylated by phosphoprotein tases, we have used complementary affinity purifica- phosphatase (PPP) family phosphatases, including PP1 and PP2A, and metallo-dependent protein phosphatases (PPM, tion and proximity-based interaction proteomics also known as PP2C). These families diverge in sequence but approaches to generate a physical interactome for converge structurally at the catalytic center (Das et al., 1996). 140 human proteins harboring phosphatase catalytic Several members of the aspartate-based phosphatases (also domains. We identified 1,335 high-confidence inter- known as haloacid dehalogenases [HADs]) act on serine or tyro- actions (1,104 previously unreported), implicating sine phosphorylated proteins, although other family members these phosphatases in the regulation of a variety of preferentially dephosphorylate small molecules (reviewed in Sei- cellular processes. Systematic phenotypic profiling fried et al., 2013). The majority of tyrosine dephosphorylation, of phosphatase catalytic and regulatory subunits re- however, is mediated through the actions of the largest phos- vealed that phosphatases from every evolutionary phatase family—the protein tyrosine phosphatases (PTPs). family impinge on mitosis. Using clues from the inter- This family includes the receptor and nonreceptor PTPs and actome, we have uncovered unsuspected roles for the dual-specificity phosphatases (DSPs) that can dephosphor- ylate both tyrosine and serine/threonine residues, and in some DUSP19 in mitotic exit, CDC14A in regulating micro- cases, nonprotein substrates, including phospholipids (reviewed tubule integrity, PTPRF in mitotic retraction fiber in Alonso and Pulido, 2016). integrity, and DUSP23 in centriole duplication. The In this study, we employed two complementary proteomics functional phosphatase interactome further provides approaches to identify interaction partners for each human pro- a rich resource for ascribing functions for this impor- tein phosphatase, leading to an interactome of 1,335 interac- tant class of enzymes. tions (1,104 previously unreported) among 1,051 proteins. Although this alone provides a number of promising avenues for follow-up study, combining proteomics data with phenotypic INTRODUCTION screens can help place these interactions within specific biolog- ical contexts. We employed this strategy to examine mitosis—a Reversible protein phosphorylation, catalyzed by the reciprocal complex process involving widespread phosphoproteome reor- actions of kinases and phosphatases, is a ubiquitous signal ganization (Olsen et al., 2010). Several kinases are crucial for transduction mechanism, central to the regulation of many bio- mitosis and are being investigated as targets for clinical interven- logical processes. Of these two classes of enzymes, phospha- tion (Manchado et al., 2012). However, although much work has tases are comparatively poorly studied, despite their recognized been done on understanding the roles of a handful of phospha- involvement in diseases such as Alzheimer’s, diabetes, and can- tases in mitotic regulation (reviewed in Qian et al., 2013), the roles cer (reviewed in Braithwaite et al., 2012; Gurzov et al., 2015; and (if any) of most other phosphatases in mitotic progression are Julien et al., 2011, respectively). largely unknown. To address this, we herein complement our

2488 Cell Reports 17, 2488–2501, November 22, 2016 ª 2016 The Author(s). This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). (legend on next page)

Cell Reports 17, 2488–2501, November 22, 2016 2489 physical interaction screens by RNAi screening to identify and The PP1 and PP2A catalytic subunits, which together con- study phosphatases with roles in mitotic progression, leading tribute the bulk of serine/threonine phosphatase activity in the us to ascribe functions to several tyrosine family phosphatases cell (reviewed in Virshup and Shenolikar, 2009), displayed the in the regulation of mitosis. highest interactivity (>20 interactions each; Figure 1A). As ex- pected, a large fraction of the PP1 and PP2A interactomes RESULTS consisted of known regulatory subunits, which are thought to provide specificity to the dephosphorylation process (reviewed Interaction Networks Reveal a Role for Phosphatases in in Shi, 2009; Figure S2). Generally, PPP family members also a Variety of Cellular Processes generated more high-confidence interactions (mean of 18) than Many phosphatases are regulated though protein-protein inter- the average phosphatase (mean of 6). There was no correlation actions. For example, PPP family phosphatases such as PP1 between the relative abundance of the phosphatase bait (as de- and PP2A are regulated through combinatorial binding to a tected by spectral counts) and the number of high-confidence in- myriad of regulatory subunits (reviewed in Shi, 2009) and teractions (Figure 1A). MAPK-dephosphorylating DSPs dock to their substrates Recently, our laboratories have adopted a proximity-based through conserved sequences (reviewed in Peti and Page, method, BioID (Roux et al., 2012), as a complement to AP-MS. 2013). Interaction proteomics should thus be a useful starting BioID consists of the fusion of an abortive biotin ligase (BirA*) point for exploring the function of these enzymes. Here, we to a bait expressed in mammalian cells; addition of biotin cata- defined a as any encoding one of lyzes its activation to biotinoyl-AMP by BirA* and allows covalent the four types of protein phosphatase catalytic domains labeling of lysines on proteins in the vicinity of the bait. Bio- (including inactive phosphatases), and compiled a list of 151 tinylated proteins are recovered by streptavidin affinity pull- genes of interest (Table S1; overlap with the DEPhOsphorylation down and identified by mass spectrometry. We previously Database [DEPOD] [Duan et al., 2015] is indicated). We gener- used BioID to detect phosphorylation-dependent interactions ated a library of Gateway entry clones and 3X-FLAG-tagged (Couzens et al., 2013), and for the identification of proximity destination vectors using the longest human isoform for each partners for components of relatively insoluble structures (Gupta phosphatase whenever possible to maximize the number of et al., 2015; Lambert et al., 2015). To survey the proximity putative interaction partners (see the Supplemental Experi- interaction landscape of phosphatases, and in particular of the mental Procedures). We next generated stable isogenic transmembrane receptor tyrosine phosphatases (RPTPs), we HEK293 Flp-In T-REx cell lines with tetracycline-inducible generated 60 stable cell lines with inducible expression of expression, resulting in the expression of 140 protein phospha- BirA*-tagged phosphatases, and performed BioID (Figure 1A; tases (Table S1; Figures S1A and S1B), 114 of which were pre- Table S3), revealing 2,358 proximity interactions at %1% FDR, viously shown to display endogenous mRNA expression in 44 of which were overlapping with the literature or our AP-MS HEK293 (Uhle´ n et al., 2015; Figure 1A; Table S1). We note data (Figures 1A and 1C). For the RPTPs, a dramatic gain in that the physical interactions for the were re- proximity interactions was observed (57 interactions by ported elsewhere (St-Denis et al., 2015) and are only included AP-MS, an average of 3 per RPTP; 1,083 interactions by BioID, here for completeness. Cycling cells expressing bait proteins an average of 60 per RPTP; all at an FDR % 1%; Figure 1A). It were induced for 24 hr before FLAG affinity purification coupled was, however, difficult to identify high-confidence preys that to mass spectrometry (AP-MS). Bait levels were monitored by exhibited a preference for one or a subset of phosphatases normalized spectral counts (see the Supplemental Experi- (i.e., that may mediate their specific biology) rather than mental Procedures; Figure 1A), and this correlated well with merely indicating plasma membrane localization. We therefore the level of selected baits detected by western blot (Fig- implemented a secondary Specificity score that evaluates ure S1A). Identified interactions with prey proteins were the quantitative enrichment of preys with specific baits against analyzed for significance using SAINTexpress (Teo et al., the same prey across all other baits profiled (see the Supple- 2014), with high-confidence interactors defined as those with mental Experimental Procedures and Table S3 for the bench- a false discovery rate (FDR) of %1% (across biological marking against literature interactions). Following application of duplicates; see Supplemental Experimental Procedures and the Specificity cut-off, we report 561 high-confidence specific Figure S1C for reproducibility metrics). This revealed 797 BioID interactions across the 60 baits profiled, 33 of which high-confidence interactions (Figures 1B and 1C), 242 (30%) overlapped with the literature (Figures 1B and 1C; Table S3). of which were previously reported in the BioGRID database We note several broad features for the phosphatase interac- (Chatr-Aryamontri et al., 2015; Figure 1C; Table S2). tome. Connections to kinases were exemplified by the numerous

Figure 1. The Human Phosphatase Interactome (A) Detailed interactions for 140 phosphatases identified by AP-MS and BioID. Phosphatases are organized by family and sequence similarity. Endogenous expression levels in HEK293 are from RNA-seq (gray bars). Relative bait levels for AP-MS (pale green) and BioID (pale orange) were determined by normalized spectral counts, and interactors that passed 1% FDR as assessed by SAINTexpress are shown for AP-MS (dark green) and BioID (dark orange). The legend indicates the value ranges; see Supplemental Experimental Procedures and Tables S1–S3 for details. (B) Network of the high-confidence interactions observed for the human protein phosphatases. For AP-MS, interactions %1% FDR are displayed. For BioID, the interactions displayed passed both the %1% FDR filter and the Specificity filter. (C) Overview of the overlap of our interactions with those reported in the literature (BioGRID). See also Figure S1.

2490 Cell Reports 17, 2488–2501, November 22, 2016 Figure 2. Selected Examples of Phosphatase Interactions (A) MAPK-interacting phosphatases. (B) 14-3-3 protein-interacting phosphatases. (C) PP2A-interacting phosphatases. For (A), (B), and (C), the data are represented as Cytoscape networks with each phosphatase family colored differently; see inset in (A) for other details. (D) Interactome for the tensin family highlighting interactions with the dystrophin complex (SNTB2, UTRN, DTNB, DMD, SNTB1); previously reported interactions are highlighted in green. (E) DUSP9 interacts with primase (PRIM1, PRIM2) and with the Elongator complex (ELP2, ELP3, IKBKAP). (F) STYX interacts with SKP1 and several F-box proteins, whereas STYXL1 recovers cytoskeletal components. (G) BioID reveals specific proximity interactions with exocytosis/trafficking components for PTPRH and PTPRJ; bold italics indicate that the reported interaction(s) satisfied both our FDR and specificity filter cutoffs. For (D)–(G), the data are represented as dot plots; see inset in (E) for details. See also Figures S2 and S3. phosphatases that associated with MAP kinases, as expected mentally determined (CDC25B, CDC25C, PTPN3, and SSH1) (Figure 2A). Phosphatases from several families also interacted or computationally predicted (PPM1H, TENC1) 14-3-3 binding with at least one of the phosphothreonine binding 14-3-3 pro- sites. We also detected several connections between phospha- teins (Figure 2B), including several phosphatases with experi- tases, notably for PP2A holoenzymes, which were recovered as

Cell Reports 17, 2488–2501, November 22, 2016 2491 Figure 3. Protein Phosphatases from All Evolutionary Families Play Roles in Mitosis (A) Schematic workflow of RNAi screening and validation process highlighting the number of hits. (B) Human protein phosphatases mitotic hits, by evolutionary family. Phylogenetic trees represent alignments of the longest isoforms of the full-length proteins. Phosphatases are labeled by official gene symbols. Colored circles, the size of which corresponds to the cumulative Z score from the mitotic index and high- resolution analyses, indicate primary and validated hits. (C) PPP family phosphatase subunit hits in the RNAi screen. See also Figure S4. preys with several unrelated phosphatases (Figure 2C). Interac- phatase to significantly interact with phosphorylase kinase, sug- tions point to putative roles for several phosphatases, e.g., ten- gesting a link to glycogen metabolism (Figure S3). To facilitate sins with the dystrophin, syntrophin, and synaptobrevin complex exploration, the entire dataset may be accessed via prohits- (Figure 2D); DUSP9 with DNA primase and components of the web.lunenfeld.ca. Taken together, AP-MS and BioID dramati- RNA Polymerase II-associated elongator acetyltransferase com- cally expand the catalog of physical interactions established plex (Figure 2E); and STYX with SKP1 and several F-box proteins by protein phosphatases. (Figure 2F; see additional examples in Figure S2). The BioID dataset also permitted the confirmation of previ- Mitosis Requires Phosphatases from All Evolutionary ously known interactions to specific phosphatases (e.g., GRB2 Lineages with PTPRA [den Hertog et al., 1994]; TRIO and the liprin PPFIA1 To identify phosphatases with potential roles in mitosis, we per- with PTPRF, PTPRS, and PTPRD [Debant et al., 1996; Pulido formed a high-resolution endoribonuclease-generated small et al., 1995; Figure S3]) and expanded the number of functional interfering RNA (esiRNA) screen (Kittler et al., 2007; Figure 3A). hypotheses for less characterized phosphatases, including Briefly, HeLa cells were transfected with esiRNAs targeting 228 PTPRH and PTPRJ in exocytosis (Figure 2G) and PTPRU in actin genes encoding 151 phosphatases and 77 known phosphatase regulation (Figure S3). Intriguingly, PTPRT was the only phos- regulatory subunits (Table S4). Note that we did not verify the

2492 Cell Reports 17, 2488–2501, November 22, 2016 Table 1. Validated Phosphatases and Regulatory Subunits with Potential Roles in Mitosis Mitotic Abnormal Validation RNAi- Previous Other Index Spindle New Validation Resistant Hit In Literature Gene Name Aliases Family Z Score Z Score esiRNA siRNA Test Screen Evidence PPP2R1A PP2A-Aa, PR65A associated subunit (PPP) + + + + + + SBF2 MTMR13, DENND7B PTP family () + + + + + PPP1R14A CPI-17 associated subunit (PPP) + + + TPTE PTEN2, CT44 PTP family (PTEN/tensin) + + – + PPP2CA PP2Ac, PP2Aa PPP family + + + + + + PPP1R14D CPI17-like, GBP1 associated subunit (PPP) + – – + PPM1B PP2Cb PPM family + + + – PPP1R9A NRB1, Neurabin-1 associated subunit (PPP) + + + – PPM1A PP2Ca PPM family + + + – PPP3R2 Calcineurin subunit B associated subunit (PPP) + – – + PPP1R12A MYPT1 associated subunit (PPP) + + – + + + PPP1R7 SDS22 associated subunit (PPP) + – + + + PPP1R14C CPI17-like, KEPI associated subunit (PPP) + + + – PHACTR3 Scapinin, PPP1R123 associated subunit (PPP) + + + – + PPM1K PP2Cm, PP2Ck PPM family + + + – DUSP19 SKRP1, LMW-DUSP3 dual specificity PTP family + + + + + + PPP2R5D PP2A B’d, PP2A B56d associated subunit (PPP) + + + + + TPTE2 TPIP PTP family (PTEN/tensin) + + + + DUSP23 VHZ, LDP-3 dual specificity PTP family + + + + + CDC14A hCDC14 dual specificity PTP family – + – + + + PPP2R5E PP2A B’ε, PP2A B56ε associated subunit (PPP) – + – + + PTPRF LAR classical receptor PTP – + – + + + PPP1CA PP1A, PP1a PPP family – + + + + + PPP1R16B TIMAP, ANKRD4 associated subunit (PPP) – + + PPP1R1C IPP5 associated subunit (PPP) – + – + CTDSPL RBSP3, HYA22, SCP3 aspartate-based HAD –+ + + + family PPP2CB PP2CB, PP2Ab PPP family – + + – + RNGTT HCE, CAP1A dual specificity PTP family – + + – TNS3 TENS1 PTP family (PTEN/tensin) – + + EPM2A Laforin dual specificity PTP family – + – + PTPN23 HD-PTP classical nonreceptor PTP – + + PTPRE PTPε, RPTPepsilon classical receptor PTP – + – + PTPRT PTPr, RPTPrho classical receptor PTP – + – + + Hits are sorted by decreasing Mitotic Index Z score, then by decreasing Abnormal Spindle Z score in the primary screen. Hits that passed selected cutoffs in the primary and validation screens are indicated by ‘‘+,’’ whereas ‘‘–’’ indicates that they did not pass. Empty is not tested. Evidence of a mitotic role from similar mitotic screens or other literature sources is also indicated. See Table S4 for complete details. expression of these genes; however, 113 of the 140 phospha- abnormal mitosis). Of these, 27 affected mitotic progression, tase catalytic subunits we profiled as baits in our interactome 45 affected spindle assembly, and 24 affected both processes are reported to be endogenously expressed in HeLa cells, a (Figure S4C). 88% overlap with HEK293s (Table S1; Uhle´ n et al., 2015). Mitotic We validated 33 (69%) of these hits using at least one indices were calculated using DAPI and phospho-Histone H3 orthogonal silencing trigger, either siRNAs and/or nonoverlap- labels (Figure S4A). Additionally, high-resolution images were ping esiRNAs (Figures 3B and S4D). The validation data are manually annotated for the presence of abnormal metaphase summarized in Table 1 (details in Table S4) along with any structures, such as misaligned , mono/multipolar previously reported mitotic function for any of our validated spindles, centrosomal defects, etc. (Figure S4B; Table S4). In targets. For example, the PPP family was represented by total, the primary screen identified 48 candidate mitotic regula- PP1 and PP2A, both of which have multiple roles in mitosis tors (mean Z score of R2 for mitotic index and/or R2.5 for (De Wulf et al., 2009). Depletion of either catalytic PP1

Cell Reports 17, 2488–2501, November 22, 2016 2493 (PPP1CA) or PP2A (PPP2CA/B) subunits or several known ing that DUSP19 may inhibit the progression of late mitosis. Cells scaffolding, regulatory, or inhibitory proteins led to defects depleted of DUSP19 also displayed defects in late mitosis, re- in mitosis (Figure 3C). Several of these hits have been sulting in cell death or multinucleated daughter cells (Figure 4F; well documented, e.g., PPP1R7 (SDS22) targets PP1 to the Movie S2). The nonspecific mitotic spindle defects and modest kinetochores where it antagonizes Aurora B activity (Posch increase in mitotic index (Table S4) detected in the initial screen et al., 2010), whereas PPP1R12 (myosin phosphatase target are likely due to continued division of these multinucleated subunit; MYPT1) enables PP1 to counter PLK1 activity at daughter cells. centromeres (Matsumura et al., 2011). PP1 regulatory subunits Depletion of CDC14A, a DSP with to the not previously associated with mitosis included PPP1R9A, yeast mitotic phosphatase Cdc14p (Li et al., 1997), resulted PHACTR3, and several members of the PPP1R14 family of in abnormal spindles and defects in alignment PP1 inhibitory proteins (PPP1R14A/C/D; note that PPP1R14A at the metaphase plate (Figure 4G; Table S4), suggesting that preferentially inhibits MYPT1-PP1 [Eto et al., 2004] consistent it affects mitotic spindle microtubule organization. CDC14A with a mitotic function). Non-PPP serine/threonine phosphatase AP-MS revealed interactions with components of the recently hits included the PPM phosphatase PPM1A and the HAD-like discovered 3M complex, which acts in microtubule stabilization phosphatase CTDSPL. Strikingly, 13 of the validated mitotic (Figure 4H; Li et al., 2014; Yan et al., 2014). Similarly to regulators were members of the protein tyrosine phosphatase CDC14A, depletion of 3M complex components resulted in family, suggesting that phosphatases from all evolutionary altered microtubule dynamics, chromosome congression fail- families are involved in the regulation of mitotic cell division ure, and mitotic arrest (Yan et al., 2014) and like the 3M com- (Figure 3B). plex, CDC14A localized to the centrosome (Figure S5), and affected cell shape throughout the cell cycle (Figure 4I). PTP Family Phosphatases Are Implicated in Mitotic Although phosphoregulation of the 3M complex has yet to be Processes investigated, it is noteworthy that all 3M components To confirm the surprising identification of many PTPs family interacting with CDC14A contain identified proline-directed members in our screen, we selected six of them for further inves- phosphoserine residues (Dephoure et al., 2008; Kim et al., tigation: CDC14A, DUSP19, DUSP23, PTPRF, PTPRT, and 2012), suggesting that a proline-directed CDC14 phospha- SBF2. To confirm the specificity of the phenotypes, HeLa Flp- tase (Gray et al., 2003) might regulate the complex through In T-REx cells harboring siRNA-resistant inducible variants of dephosphorylation. each of these PTPs were generated. In each case, siRNA-medi- ated depletion resulted in increased mitotic defects, which were The Receptor Tyrosine Phosphatase PTPRF Is a Crucial prevented with concurrent expression of the RNAi-resistant PTP Component of Mitotic Retraction Fibers (Figure 4A). PTPRF (also known as LAR [leukocyte antigen-related]) is a The phenotypes observed upon depletion of many of the transmembrane RPTP whose depletion led to multipolar spin- phosphatases were similar: abnormal spindle morphology, with dles and/or chromosome misalignment (Figure 5A). PTPRF, notable numerical and structural centrosomal abnormalities which is important for the development of the nervous system coincident with mitotic arrest. However, when the results from (Chagnon et al., 2004), contains several the phenotypic screens were considered in the context of the (CAM)-like domains in its extracellular portion (O’Grady et al., interaction data (Figure 4B), more precise roles for four of these 1994) and regulates focal adhesions (Sarhan et al., 2016; PTP phosphatase hits began to emerge. Serra-Page` s et al., 1995; Streuli et al., 1990). We observed that For example, DUSP19 (also known as SKRP1 or LMWDSP3) is GFP-PTPRF localized to the plasma membrane as expected an atypical DSP previously implicated in JNK signaling (Zama (Figure S6A), but strikingly also to long extracellular fibers during et al., 2002). Unlike several other DSPs in our dataset (Fig- mitosis (Figure 4B) where it exhibited a punctate pattern, resem- ure 2A), DUSP19 did not interact with JNK isoforms, consistent bling beads on a string (Figure 4B, zoomed panels). These fibers with the fact that it lacks the MAP kinase docking motif (Patter- were similar to retraction fibers, actin-rich mitotic remnants of son et al., 2009). Instead, DUSP19 interacted with catalytic interphase focal adhesions that maintain attachment between (PPP2CA), scaffolding (PPP2R1A), and regulatory (PPP2R5D mitotic cells and the extracellular matrix, and control mitotic and PPP2R5E) subunits of PP2A, all of which were also validated spindle positioning in response to extracellular cues (Fink hits in our screen (Figure 3C). PPP2R5D and PPP2R5E have et al., 2011; The´ ry et al., 2007). The mitotic phenotypes observed recently been shown to recruit PP2A to the anaphase central upon PTPRF depletion resembled loss of the retraction fiber spindle (Bastos et al., 2014), which controls the assembly of component mitotic spindle positioning (MISP) (Maier et al., the actin-rich cleavage furrow (Fededa and Gerlich, 2012). As ac- 2013; Zhu et al., 2013). Consistent with a specific role in retrac- tins were also prominent DUSP19 interactors, we hypothesized tion fiber formation, depletion of PTPRF resulted in the loss of that DUSP19 may play a role in anaphase. Live-cell imaging of retraction fibers from metaphase cells (Figure 5C) but caused HeLa cells stably expressing RFP-Histone 2B and GFP-Tubulin no obvious changes in interphase actin cytoskeleton orga- revealed no significant difference in the duration of early mitotic nization (Figure S6B), suggesting that the effect is specific to events (from nuclear envelope breakdown [NEB] to anaphase retraction fibers. Loss of retraction fibers was prevented by onset) in DUSP19 siRNA-treated cells (Figure 4D; Movies S1 siRNA-resistant PTPRF-GFP expression (Figure 4D). Little is and S2), but these cells progressed through late mitosis more known about the composition of retraction fibers, although rapidly than control cells (Figure 4E; Movies S1 and S2), suggest- actin-associated focal adhesion proteins seem important (Maier

2494 Cell Reports 17, 2488–2501, November 22, 2016 Figure 4. PTP Family Phosphatases Play Roles in Mitosis (A) HeLa Flp-In T-REx cells with stable expression of indicated siRNA-resistant GFP-PTPs were treated with their respective PTP RNAi, with or without tetracycline to induce PTP expression. Cells were analyzed for mitotic defects. Data represent the mean of three independent ex- periments ± SE. ns, not significant; **p < 0.01; ***p < 0.0001. (B) Strategies to combine physical and functional screens to annotate mitotic hits: phosphatase hits that have interactors with known mitotic function are prioritized for functional analysis. (C) DUSP19 interactions with actin and PP2A. See legend inset for details. (D and E) HeLa cells with stable expression of GFP- Tubulin and RFP-Histone 2B were treated with control or DUSP19 siRNA and subjected to live-cell imaging. For individual cells, the time from nuclear envelope breakdown (NEB) to anaphase (D) and from anaphase to completion of abscission (E) was measured. Results show one of three replicate experiments. ns, not significant; **p < 0.01. (F) Representative still images from cells in (D) and (E). Numbers in bottom left indicate minutes elapsed between adjacent panels. Scale bars, 20 mm. In DUSP19 images, timing relates to cell 1. (G) Sample images of mitotic HeLa cells, treated with control or CDC14A siRNA. Scale bars, 10 mm. (H) CDC14A interactions with 3M complex com- ponents. See legend inset in (C) for details. (I) Sample images of interphase HeLa cells, treated with control or CDC14A siRNA. Scale bar, 10 mm. See also Figure S5 and Movies S1 and S2.

DUSP23 Interacts with PLK4 and Regulates Centriole Duplication The atypical DSP DUSP23 (also known as VHZ; Alonso et al., 2004; Wu et al., 2004)is a 16-kDa protein comprising essentially only a PTP domain. AP-MS identified a sin- gle high-confidence prey: the serine/thre- onine kinase PLK4, a master regulator of centriole duplication (Habedanck et al., 2005; Figure 6A). DUSP23 depletion re- sulted in a marked increase in abnormal mitoses (Figure 6B), with increased monopolar and multipolar spindles (Fig- ures 6C and 6D), and fewer than four cen- trioles (Figures 6D and 6E). In a sensitized centriole duplication assay (Kleylein- Sohn et al., 2007), DUSP23 RNAi inhibited Myc-PLK4-induced centriole overduplica- et al., 2013; Zhu et al., 2013). PTPRF BioID preys were signifi- tion in S-phase-arrested U-2 OS cells to levels comparable to cantly enriched for proteins associated with cell-cell junctions the depletion of two core centriolar components, CEP120 and ( [GO]: 0005911; p = 2.04EÀ12), focal adhesions CEP135 (Comartin et al., 2013; Kleylein-Sohn et al., 2007; Mah- (GO: 0005925; p = 7.07EÀ08), and the actin cytoskeleton (GO: joub et al., 2010)(Figure 6F). Like PLK4, DUSP23 RNAi resulted

0015629; p = 3.38EÀ04; Figure 4E). These interactors could in an accumulation of S- and G2/M-phase cells (Figure S7A). have important roles on retraction fibers, therefore regulating The decrease in centrioles observed upon DUSP23 deple- mitotic spindle organization. tion was partially prevented by expression of siRNA-resistant

Cell Reports 17, 2488–2501, November 22, 2016 2495 Figure 5. PTPRF Is a Crucial Component of Extracellular Mitotic Retraction Fibers (A) Sample images of mitotic HeLa cells, treated with siRNA, and stained to visualize the mitotic machinery. Scale bars, 10 mm. (B) HeLa Flp-In T-REx cells with stable, inducible expression of PTPRF-GFP were treated with control or PTPRF siRNA. Zoomed images represent the boxed region in merged images. Scale bars, 10 mm. (C) HeLa cells treated as in (A) were stained for phalloidin and endogenous PTPRF. Scale bars, 10 mm. (D) HeLa Flp-In T-REx cells with induced expression of siRNA-resistant PTPRF-GFP were treated with siRNA and analyzed for the presence of retraction fibers. Data represent the mean of three independent experiments ± SE. *p < 0.05, **p < 0.01. (E) PTPRF BioID preys with an FDR of %1% and a project frequency of %50%. Preys were sorted into GO terms using the Panther Overrepresentation Test. See also Figure S6.

GFP-DUSP23, demonstrating the specificity of the phenotype are both important for centriole duplication. Taken together, (Figure 6G). Expression of a DUSP23 phosphatase-dead mutant the data suggest that loss of DUSP23 activity leads to decreased (C95S), however, did not prevent centriole loss. Intriguingly, centriole duplication, which results in spindle polarity defects DUSP23 also possesses phosphotransferase ability in vitro— upon progression into mitosis. i.e., instead of releasing inorganic phosphate after dephosphor- DUSP23 interacts with PLK4, suggesting that DUSP23 acts ylation, DUSP23 can transfer phosphate to another phosphory- early in the centriole duplication pathway. Using three-dimen- latable substrate (Kuznetsov and Hengge, 2013). A DUSP23 sional structured illumination microscopy (3D-SIM), we exam- point mutant with normal phosphatase activity, but deficient in ined early and late markers of centriole duplication in S-phase in vitro phosphotransferase activity (P68V; Kuznetsov and Myc-PLK4-overexpressing cells (Comartin et al., 2013). Deple- Hengge, 2013) also failed to prevent centriole loss, suggesting tion of DUSP23 resulted in centrioles lacking the late daughter that dephosphorylation and transphosphorylation by DUSP23 marker Centrin (CETN), and devoid of the typical ‘‘florette’’

2496 Cell Reports 17, 2488–2501, November 22, 2016 Figure 6. DUSP23 Regulates Centriole Duplication through PLK4 (A) PLK4 is the only DUSP23 interactor identified by AP-MS with a FDR of %1%. (B) HeLa cells were treated with siRNA and analyzed for mitotic spindle defects. Data in (B), (C), (E), (F), and (G) represent the mean of three independent ex- periments ± SE. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.0001. (legend continued on next page)

Cell Reports 17, 2488–2501, November 22, 2016 2497 arrangement of overduplicated centrioles (Kleylein-Sohn et al., can provide clues regarding their functions (in mitosis or other 2007). Importantly, DUSP23 inhibited the recruitment of cellular processes). Our results suggest that, in contrast to CEP135 to the procentriole daughter assembly region, which is PPP family members and the MAP kinase phosphatases, the an early step that involves CEP135 forming a ring coincident majority of phosphatases do not establish multiple stable pro- with the PLK4-labeled mother ring structure (Comartin et al., tein-protein interactions, a phenomenon that appears to be 2013; Figure 6H). These phenotypes were similar to centrioles evolutionarily conserved (Breitkreutz et al., 2010). Yet, several depleted of STIL, a key target of PLK4 kinase activity and an early of our human phosphatases interacted with components of the regulator of centriole biogenesis (Arquint et al., 2015; Moyer phosphorylation machinery—kinases, phosphobinding proteins, et al., 2015; Figure 6H). Interestingly, DUSP23 depletion did and other phosphatases (Figure 2). This is consistent with the in- not affect the recruitment of STIL to the mother centriole in terconnectivity detected for yeast kinases and phosphatases Myc-PLK4 cells (Figure 6I), a very early event in the initiation of (Breitkreutz et al., 2010), and later for human CMGC family ki- centriole duplication (Comartin et al., 2013; Moyer et al., 2015). nases (Varjosalo et al., 2013). Our findings therefore accentuate However, the normal progression of STIL localization from just the complexity of phosphorylation-based signal transduction, peripheral to the mother (early) to the base of the daughter whereby modifying enzymes can act in phosphoregulatory cas- centriole (late) during centriole daughter assembly and elonga- cades, and cross talk to other enzymes may add robustness to tion was arrested (Figure 6I). signaling systems (Levy et al., 2010). Regulation of PLK4 levels is achieved via phosphoregulation Our findings extend other excellent existing resources for (Cunha-Ferreira et al., 2009; Rogers et al., 2009). DUSP23 deple- generic interactions (BioGRID [Chatr-Aryamontri et al., 2015], tion resulted in increased HA-PLK4 in cells (Figure S7B). HA- IntAct [Orchard et al., 2014], etc.) by systematically assessing PLK4 intensity at the centriole was also increased upon nearly all human phosphatase catalytic subunits under the DUSP23 depletion (Figures S7C and S7E). We also monitored same conditions by AP-MS, permitting comparative assess- PLK4 phospho-S305 intensity (as a surrogate for PLK4 kinase ments (through semiquantitative measurements and statistical activity; Mason et al., 2014) in cells expressing HA-PLK4 evaluation of each interaction). We note that, although we have and depleted of DUSP23. The amount of associated PLK4 identified many stable interactions, our study was not designed phospho-S305 label did not increase proportionately, resulting to identify substrates for phosphatases: phosphatases establish in a significantly lower ratio of phospho-S305 staining to total short-lived interactions with their substrates, which are typically PLK4 staining at the centriole (Figures S7D and S7E). This sug- difficult to capture by biochemical enrichment (AP-MS). It is gests that loss of DUSP23 results in decreased centriolar PLK4 also not clear that the dynamics of interactions for kinases and activity. Contrary to previous work (Tang et al., 2010), in our phosphatases are compatible with BioID (this study and other hands, DUSP23 did not localize to centrioles (Figure S7F). unpublished data from the A.-C.G. laboratory), in contrast to in- DUSP23 did not detectably interact with any centriolar proteins teractions involving direct phospho-binding modules (Couzens other than PLK4 (Table S2), suggesting an early function for et al., 2013) or ubiquitin ligase substrates (Coyaud et al., 2015). DUSP23 in the centriole duplication cycle, potentially through As such, it is not surprising that the overlap with phosphatase- the modulation of PLK4 activity. substrate interactions reported in DEPOD (Duan et al., 2015), to our knowledge the most comprehensive database of such re- DISCUSSION lationships, is very low, with 21/838 annotated relationships de- tected here. Clearly, more work will be needed—likely including We have used a combined physical and functional screening global phosphoproteomics approaches (reviewed in Gingras approach to uncover cellular roles for protein phosphatases, and Wong, 2016)—to fully understand the signaling events with a focus on the critical process of mitosis. Our high-resolu- downstream of each phosphatase. tion RNAi screen identified phosphatases from every evolu- Globally, our work identifies a variety of phosphatase protein- tionary family, and every subgroup of the PTP family, as mitotic protein interactions and underscores how much there is to be regulators. Individual PTPs play distinct roles in mitotic progres- learned about these enzymes. Although we have performed sion, including centriole duplication (DUSP23), microtubule spin- initial follow-up experiments on selected phosphatases in dle stability (CDC14A), cell-matrix attachment (PTPRF), and mitosis, other hits in our phenotypic screen warrant further inves- mitotic exit (DUSP19). tigation as will deciphering the role of the hundreds of additional Our proteomic screening also identified previously unreported interactors, in other cellular process. To assist the community in interactions for a variety of phosphatases, and these interactions the continued exploration of our data-rich resource, we have

(C) HeLa cells were treated as in (B). Metaphase cells were analyzed for polarity defects. (D) Sample cells from (C). Scale bars, 10 mm. (E) HeLa cells were treated as in (B). Metaphase cells were analyzed for the number of centrioles. (F) U2 O-S cells with stable, inducible expression of Myc-PLK4 were treated with the indicated siRNAs for 48 hr before concurrent S-phase arrest with aphidicolin and Myc-PLK4 induction for 24 hr. Cells were analyzed for centriole overduplication. (G) HeLa Flp-In T-REx cells were treated with siRNA with concurrent induction of siRNA-resistant versions of GFP-DUSP23. Metaphase cells were analyzed for centriole number. Lysates were also generated and subjected to western blot analysis.

(H and I) Myc-PLK4 U-2 OS cells were transfected with siRNA for 64 hr with Myc-PLK4 induction starting after 24 hr. Cells were arrested in G2 with sequential overnight treatments with aphidicolin and RO3306. Cells were imaged by 3D-SIM. Scale bars, 500 nm. See also Figure S7.

2498 Cell Reports 17, 2488–2501, November 22, 2016 deposited the mass spectrometry data to the public repository is the Canada Research Chair in Centrosome Biogenesis and Function. MassIVE (http://massive.ucsd.edu/ProteoSAFe/static/massive. N.S.-D. was supported by a postdoctoral fellowship from the Canadian Insti- jsp) and also present our dataset in its entirety on a dedicated tutes of Health Research. website (prohits-web.lunenfeld.ca). Received: July 25, 2016 Revised: September 3, 2016 EXPERIMENTAL PROCEDURES Accepted: October 19, 2016 Published: November 22, 2016 Physical Interactome Mapping Generation of FLAG- and BirA*-tagged phosphatase cell lines and mass spec- trometric analysis was performed essentially as described (Couzens et al., REFERENCES 2013), and significant interactions were scored using SAINTexpress (Teo et al., 2014). Alonso, A., and Pulido, R. (2016). The extended human PTPome: a growing tyrosine phosphatase family. FEBS J. 283, 1404–1429. Phenotypic Screening Alonso, A., Burkhalter, S., Sasin, J., Tautz, L., Bogetz, J., Huynh, H., Bremer, esiRNAs were designed and generated as in Kittler et al. (2005). HeLa cells M.C.D., Holsinger, L.J., Godzik, A., and Mustelin, T. (2004). The minimal essen- were screened for mitotic phenotypes essentially as in Lawo et al. (2009). tial core of a cysteine-based protein-tyrosine phosphatase revealed by a novel 16-kDa VH1-like phosphatase, VHZ. J. Biol. Chem. 279, 35768–35774. Cell Biology and Biochemistry Arquint, C., Gabryjonczyk, A.-M., Imseng, S., Bo¨ hm, R., Sauer, E., Hiller, S., Cell culture, transfection, immunofluorescence, fixed and live-cell imaging, Nigg, E.A., and Maier, T. (2015). STIL binding to Polo-box 3 of PLK4 regulates and western blot analysis were performed as previously described (St-Denis centriole duplication. eLife 4, e07888. et al., 2015). Centriole duplication assays and 3D-SIM were performed as described in Comartin et al. (2013). All reagents generated and materials Bastos, R.N., Cundell, M.J., and Barr, F.A. (2014). KIF4A and PP2A-B56 form a used are summarized in Table S1. spatially restricted feedback loop opposing Aurora B at the anaphase central spindle. J. Cell Biol. 207, 683–693. Accessing the Data Braithwaite, S.P., Stock, J.B., Lombroso, P.J., and Nairn, A.C. (2012). Protein The accession numbers for mass spectrometry data reported in this paper are phosphatases and Alzheimer’s disease. Prog. Mol. Biol. Transl. Sci. 106, MassIVE: MSV000079889, MSV000079891, and MSV000079890. The mass 343–379. spectrometry data are also associated with ProteomeXchange: PXD004525, Breitkreutz, A., Choi, H., Sharom, J.R., Boucher, L., Neduva, V., Larsen, B., Lin, PXD004527, and PXD004526. Data are also available in a searchable format Z.-Y., Breitkreutz, B.-J., Stark, C., Liu, G., et al. (2010). A global protein kinase at prohits-web.lunenfeld.ca. and phosphatase interaction network in yeast. Science 328, 1043–1046. See also the Supplemental Experimental Procedures. Chagnon, M.J., Uetani, N., and Tremblay, M.L. (2004). Functional significance of the LAR receptor protein tyrosine phosphatase family in development and SUPPLEMENTAL INFORMATION diseases. Biochem. Cell Biol. 82, 664–675.

Supplemental Information includes Supplemental Experimental Procedures, Chatr-Aryamontri, A., Breitkreutz, B.-J., Oughtred, R., Boucher, L., Heinicke, seven figures, four tables, and two movies and can be found with this article S., Chen, D., Stark, C., Breitkreutz, A., Kolas, N., O’Donnell, L., et al. (2015). online at http://dx.doi.org/10.1016/j.celrep.2016.10.078. The BioGRID interaction database: 2015 update. Nucleic Acids Res. 43, D470–D478. AUTHOR CONTRIBUTIONS Comartin, D., Gupta, G.D., Fussner, E., Coyaud, E´ ., Hasegan, M., Archinti, M., Cheung, S.W.T., Pinchev, D., Lawo, S., Raught, B., et al. (2013). CEP120 and Conceptualization, N.S.-D, L.P., and A.-C.G; Methodology, N.S.-D., G.D.G, SPICE1 cooperate with CPAP in centriole elongation. Curr. Biol. 23, 1360– L.P., and A.-C.G.; Investigation, N.S.-D., G.D.G., Z.Y.L., B.G.-B., A.O.V., 1366. J.D.R.K., D.R., A.L.C., K.W.C., J.M.T., and S.W.T.C.; Data Curation, N.S.-D., Couzens, A.L., Knight, J.D.R., Kean, M.J., Teo, G., Weiss, A., Dunham, W.H., G.D.G., and A.-C.G.; Writing – Original draft, N.S.-D. and A.-C.G.; Writing – Re- Lin, Z.-Y., Bagshaw, R.D., Sicheri, F., Pawson, T., et al. (2013). Protein interac- view and Editing, N.S.-D., G.D.G., J.D.R.K., A.O.V., A.L.C., L.P., and A.-C.G.; tion network of the mammalian Hippo pathway reveals mechanisms of kinase- Supervision, L.P. and A.-C.G; Project Administration, L.P. and A.-C.G; and phosphatase interactions. Sci. Signal. 6, rs15. Funding Acquisition, L.P. and A.-C.G. Coyaud, E., Mis, M., Laurent, E.M.N., Dunham, W.H., Couzens, A.L., Robitaille, ACKNOWLEDGMENTS M., Gingras, A.-C., Angers, S., and Raught, B. (2015). BioID-based Identifica- tion of Skp Cullin F-box (SCF)b-TrCP1/2 E3 Ligase Substrates. Mol. Cell. We thank Jianping Zhang and Guomin Liu for website design (prohits-web. Proteomics 14, 1781–1795. lunenfeld.ca) and maintenance; Brett Larsen and Monika Tucholska for expert Cunha-Ferreira, I., Rodrigues-Martins, A., Bento, I., Riparbelli, M., Zhang, W., advice on mass spectrometry experiments; and Marilyn Goudreault, Christine Laue, E., Callaini, G., Glover, D.M., and Bettencourt-Dias, M. (2009). The SCF/ Holley, and Jessica Cuppage for technical assistance. We also thank Fred Slimb ubiquitin ligase limits centrosome amplification through degradation of Robinson, Sylvain Meloche, Antje Gohla, Erich Nigg, and Tak Mak for the SAK/PLK4. Curr. Biol. 19, 43–49. kind gift of reagents; Christopher Go for sharing unpublished BioID results Das, A.K., Helps, N.R., Cohen, P.T., and Barford, D. (1996). Crystal structure of for membrane proteins; Benjamin Neel and Wolfgang Peti for helpful discus- the protein serine/threonine phosphatase 2C at 2.0 A resolution. EMBO J. 15, sions; and Brian Raught and Jean-Philippe Lambert for critical reading of 6798–6809. this manuscript. This work was supported by the Government of Canada through Genome Canada and the Ontario Genomics Institute (OGI-088 to De Wulf, P., Montani, F., and Visintin, R. (2009). Protein phosphatases take the A.-C.G.; and OGI-097 to A.-C.G. and L.P.), the Canadian Institutes of Health mitotic stage. Curr. Opin. Cell Biol. 21, 806–815. Research (FDN 143301 to A.-C.G.; MOP-123468 and MOP-130507 to L.P.), Debant, A., Serra-Page` s, C., Seipel, K., O’Brien, S., Tang, M., Park, S.H., and and the Natural Sciences and Engineering Research Council of Canada Streuli, M. (1996). The multidomain protein Trio binds the LAR transmembrane (RGPIN355644 to L.P.). We also acknowledge the generous support of the tyrosine phosphatase, contains a protein kinase domain, and has separate Krembil Foundation (to L.P.). A.-C.G. is the Canada Research Chair in Func- rac-specific and rho-specific guanine nucleotide exchange factor domains. tional Proteomics and the Lea Reichmann Chair in Cancer Proteomics. L.P. Proc. Natl. Acad. Sci. USA 93, 5466–5471.

Cell Reports 17, 2488–2501, November 22, 2016 2499 den Hertog, J., Tracy, S., and Hunter, T. (1994). Phosphorylation of receptor Li, Z., Pei, X.-H., Yan, J., Yan, F., Cappell, K.M., Whitehurst, A.W., and Xiong, protein-tyrosine phosphatase alpha on Tyr789, a binding site for the SH3- Y. (2014). CUL9 mediates the functions of the 3M complex and ubiquitylates SH2-SH3 adaptor protein GRB-2 in vivo. EMBO J. 13, 3020–3032. survivin to maintain genome integrity. Mol. Cell 54, 805–819. Dephoure, N., Zhou, C., Ville´ n, J., Beausoleil, S.A., Bakalarski, C.E., Elledge, Mahjoub, M.R., Xie, Z., and Stearns, T. (2010). Cep120 is asymmetrically local- S.J., and Gygi, S.P. (2008). A quantitative atlas of mitotic phosphorylation. ized to the daughter centriole and is essential for centriole assembly. J. Cell Proc. Natl. Acad. Sci. USA 105, 10762–10767. Biol. 191, 331–346. Duan, G., Li, X., and Ko¨ hn, M. (2015). The human DEPhOsphorylation data- Maier, B., Kirsch, M., Anderhub, S., Zentgraf, H., and Kra¨ mer, A. (2013). The base DEPOD: a 2015 update. Nucleic Acids Res. 43, D531–D535. novel actin/focal adhesion-associated protein MISP is involved in mitotic spin- Eto, M., Kitazawa, T., and Brautigan, D.L. (2004). Phosphoprotein inhibitor dle positioning in human cells. Cell Cycle 12, 1457–1471. CPI-17 specificity depends on allosteric regulation of protein phosphatase-1 Manchado, E., Guillamot, M., and Malumbres, M. (2012). Killing cells by target- by regulatory subunits. Proc. Natl. Acad. Sci. USA 101, 8888–8893. ing mitosis. Cell Death Differ. 19, 369–377. Fededa, J.P., and Gerlich, D.W. (2012). Molecular control of animal cell cyto- Mason, J.M., Lin, D.C.-C., Wei, X., Che, Y., Yao, Y., Kiarash, R., Cescon, D.W., kinesis. Nat. Cell Biol. 14, 440–447. Fletcher, G.C., Awrey, D.E., Bray, M.R., et al. (2014). Functional characteriza- Fink, J., Carpi, N., Betz, T., Be´ tard, A., Chebah, M., Azioune, A., Bornens, M., tion of CFI-400945, a Polo-like kinase 4 inhibitor, as a potential anticancer Sykes, C., Fetler, L., Cuvelier, D., and Piel, M. (2011). External forces control agent. Cancer Cell 26, 163–176. mitotic spindle positioning. Nat. Cell Biol. 13, 771–778. Matsumura, F., Yamakita, Y., and Yamashiro, S. (2011). Myosin phosphatase- Gingras, A.-C., and Wong, C.J. (2016). Proteomics approaches to decipher targeting subunit 1 controls chromatid segregation. J. Biol. Chem. 286, 10825– new signaling pathways. Curr. Opin. Struct. Biol. 41, 128–134. 10833. Gray, C.H., Good, V.M., Tonks, N.K., and Barford, D. (2003). The structure of Moorhead, G.B.G., De Wever, V., Templeton, G., and Kerk, D. (2009). Evolu- the cell cycle protein Cdc14 reveals a proline-directed protein phosphatase. tion of protein phosphatases in plants and animals. Biochem. J. 417, 401–409. EMBO J. 22, 3524–3535. Moyer, T.C., Clutario, K.M., Lambrus, B.G., Daggubati, V., and Holland, A.J. Gupta, G.D., Coyaud, E´ ., Gonc¸ alves, J., Mojarad, B.A., Liu, Y., Wu, Q., Gheir- (2015). Binding of STIL to Plk4 activates kinase activity to promote centriole atmand, L., Comartin, D., Tkach, J.M., Cheung, S.W.T., et al. (2015). assembly. J. Cell Biol. 209, 863–878. A dynamic protein interaction landscape of the human centrosome-cilium O’Grady, P., Krueger, N.X., Streuli, M., and Saito, H. (1994). Genomic organi- interface. Cell 163, 1484–1499. zation of the human LAR protein tyrosine phosphatase gene and alternative Gurzov, E.N., Stanley, W.J., Brodnicki, T.C., and Thomas, H.E. (2015). Protein splicing in the extracellular fibronectin type-III domains. J. Biol. Chem. 269, tyrosine phosphatases: molecular switches in metabolism and diabetes. 25193–25199. Trends Endocrinol. Metab. 26, 30–39. Olsen, J.V., Vermeulen, M., Santamaria, A., Kumar, C., Miller, M.L., Jensen, Habedanck, R., Stierhof, Y.-D., Wilkinson, C.J., and Nigg, E.A. (2005). The Polo L.J., Gnad, F., Cox, J., Jensen, T.S., Nigg, E.A., et al. (2010). Quantitative phos- kinase Plk4 functions in centriole duplication. Nat. Cell Biol. 7, 1140–1146. phoproteomics reveals widespread full phosphorylation site occupancy during mitosis. Sci. Signal. 3, ra3. Julien, S.G., Dube´ , N., Hardy, S., and Tremblay, M.L. (2011). Inside the human cancer tyrosine phosphatome. Nat. Rev. Cancer 11, 35–49. Orchard, S., Ammari, M., Aranda, B., Breuza, L., Briganti, L., Broackes-Carter, F., Campbell, N.H., Chavali, G., Chen, C., del-Toro, N., et al. (2014). The Kim, S.J., DeStefano, M.A., Oh, W.J., Wu, C.C., Vega-Cotto, N.M., Finlan, M., MIntAct project–IntAct as a common curation platform for 11 molecular inter- Liu, D., Su, B., and Jacinto, E. (2012). mTOR complex 2 regulates proper turn- action databases. Nucleic Acids Res. 42, D358–D363. over of receptor substrate-1 via the ubiquitin ligase subunit Fbw8. Mol. Cell 48, 875–887. Patterson, K.I., Brummer, T., O’Brien, P.M., and Daly, R.J. (2009). Dual- specificity phosphatases: critical regulators with diverse cellular targets. Kittler, R., Heninger, A.-K., Franke, K., Habermann, B., and Buchholz, F. Biochem. J. 418, 475–489. (2005). Production of endoribonuclease-prepared short interfering RNAs for gene silencing in mammalian cells. Nat. Methods 2, 779–784. Peti, W., and Page, R. (2013). Molecular basis of MAP kinase regulation. Pro- tein Sci. 22, 1698–1710. Kittler, R., Pelletier, L., Heninger, A.-K., Slabicki, M., Theis, M., Miroslaw, L., Poser, I., Lawo, S., Grabner, H., Kozak, K., et al. (2007). Genome-scale RNAi Posch, M., Khoudoli, G.A., Swift, S., King, E.M., Deluca, J.G., and Swedlow, profiling of cell division in human tissue culture cells. Nat. Cell Biol. 9, 1401– J.R. (2010). Sds22 regulates aurora B activity and microtubule-kinetochore in- 1412. teractions at mitosis. J. Cell Biol. 191, 61–74. Kleylein-Sohn, J., Westendorf, J., Le Clech, M., Habedanck, R., Stierhof, Pulido, R., Serra-Page` s, C., Tang, M., and Streuli, M. (1995). The LAR/PTP Y.-D., and Nigg, E.A. (2007). Plk4-induced centriole biogenesis in human cells. delta/PTP sigma subfamily of transmembrane protein-tyrosine-phosphatases: Dev. Cell 13, 190–202. multiple human LAR, PTP delta, and PTP sigma isoforms are expressed in a tissue-specific manner and associate with the LAR-interacting protein LIP.1. Kuznetsov, V.I., and Hengge, A.C. (2013). New functional aspects of the atyp- Proc. Natl. Acad. Sci. USA 92, 11686–11690. ical protein tyrosine phosphatase VHZ. Biochemistry 52, 8012–8025. Lambert, J.-P., Tucholska, M., Go, C., Knight, J.D.R., and Gingras, A.-C. Qian, J., Winkler, C., and Bollen, M. (2013). 4D-networking by mitotic phospha- tases. Curr. Opin. Cell Biol. 25, 697–703. (2015). Proximity biotinylation and affinity purification are complementary ap- proaches for the interactome mapping of chromatin-associated protein com- Rogers, G.C., Rusan, N.M., Roberts, D.M., Peifer, M., and Rogers, S.L. (2009). plexes. J. Proteomics 118, 81–94. The SCF Slimb ubiquitin ligase regulates Plk4/Sak levels to block centriole Lawo, S., Bashkurov, M., Mullin, M., Ferreria, M.G., Kittler, R., Habermann, B., reduplication. J. Cell Biol. 184, 225–239. Tagliaferro, A., Poser, I., Hutchins, J.R., Hegemann, B., et al. (2009). HAUS, the Roux, K.J., Kim, D.I., Raida, M., and Burke, B. (2012). A promiscuous biotin 8-subunit human Augmin complex, regulates centrosome and spindle integ- ligase fusion protein identifies proximal and interacting proteins in mammalian rity. Curr. Biol. 19, 816–826. cells. J. Cell Biol. 196, 801–810. Levy, E.D., Landry, C.R., and Michnick, S.W. (2010). Cell signaling. Signaling Sarhan, A.R., Patel, T.R., Cowell, A.R., Tomlinson, M.G., Hellberg, C., Heath, through cooperation. Science 328, 983–984. J.K., Cunningham, D.L., and Hotchin, N.A. (2016). LAR protein tyrosine phos- Li, L., Ernsting, B.R., Wishart, M.J., Lohse, D.L., and Dixon, J.E. (1997). A family phatase regulates focal adhesions through CDK1. J. Cell Sci. 129, 2962–2971. of putative tumor suppressors is structurally and functionally conserved in hu- Seifried, A., Schultz, J., and Gohla, A. (2013). Human HAD phosphatases: mans and yeast. J. Biol. Chem. 272, 29403–29406. structure, mechanism, and roles in health and disease. FEBS J. 280, 549–571.

2500 Cell Reports 17, 2488–2501, November 22, 2016 Serra-Page` s, C., Kedersha, N.L., Fazikas, L., Medley, Q., Debant, A., and Uhle´ n, M., Fagerberg, L., Hallstro¨ m, B.M., Lindskog, C., Oksvold, P., Mardino- Streuli, M. (1995). The LAR transmembrane protein tyrosine phosphatase glu, A., Sivertsson, A˚ ., Kampf, C., Sjo¨ stedt, E., Asplund, A., et al. (2015). Pro- and a coiled-coil LAR-interacting protein co-localize at focal adhesions. teomics. Tissue-based map of the human proteome. Science 347, 1260419. EMBO J. 14, 2827–2838. Varjosalo, M., Keskitalo, S., Van Drogen, A., Nurkkala, H., Vichalkovski, A., Ae- Shi, Y. (2009). Serine/threonine phosphatases: mechanism through structure. bersold, R., and Gstaiger, M. (2013). The protein interaction landscape of the Cell 139, 468–484. human CMGC kinase group. Cell Rep. 3, 1306–1320. St-Denis, N., Gupta, G.D., Lin, Z.Y., Gonzalez-Badillo, B., Pelletier, L., and Virshup, D.M., and Shenolikar, S. (2009). From promiscuity to precision: pro- Gingras, A.-C. (2015). Myotubularin-related proteins 3 and 4 interact with tein phosphatases get a makeover. Mol. Cell 33, 537–545. polo-like kinase 1 and centrosomal protein of 55 kDa to ensure proper abscis- sion. Mol. Cell. Proteomics 14, 946–960. Wu, Q., Li, Y., Gu, S., Li, N., Zheng, D., Li, D., Zheng, Z., Ji, C., Xie, Y., and Mao, Y. (2004). Molecular cloning and characterization of a novel dual-specificity Streuli, M., Krueger, N.X., Thai, T., Tang, M., and Saito, H. (1990). Distinct func- phosphatase 23 gene from human fetal brain. Int. J. Biochem. Cell Biol. 36, tional roles of the two intracellular phosphatase like domains of the receptor- 1542–1553. linked protein tyrosine phosphatases LCA and LAR. EMBO J. 9, 2399–2407. Tang, J.P., Tan, C.P., Li, J., Siddique, M.M., Guo, K., Chan, S.W., Park, J.E., Yan, J., Yan, F., Li, Z., Sinnott, B., Cappell, K.M., Yu, Y., Mo, J., Duncan, J.A., Tay, W.N., Huang, Z.Y., Li, W.C., et al. (2010). VHZ is a novel centrosomal Chen, X., Cormier-Daire, V., et al. (2014). The 3M complex maintains microtu- phosphatase associated with cell growth and human primary cancers. Mol. bule and genome integrity. Mol. Cell 54, 791–804. Cancer 9, 128. Zama, T., Aoki, R., Kamimoto, T., Inoue, K., Ikeda, Y., and Hagiwara, M. (2002). Teo, G., Liu, G., Zhang, J., Nesvizhskii, A.I., Gingras, A.-C., and Choi, H. (2014). Scaffold role of a mitogen-activated protein kinase phosphatase, SKRP1, for SAINTexpress: improvements and additional features in Significance Analysis the JNK signaling pathway. J. Biol. Chem. 277, 23919–23926. of INTeractome software. J. Proteomics 100, 37–43. Zhu, M., Settele, F., Kotak, S., Sanchez-Pulido, L., Ehret, L., Ponting, C.P., The´ ry, M., Jime´ nez-Dalmaroni, A., Racine, V., Bornens, M., and Julicher,€ F. Go¨ nczy, P., and Hoffmann, I. (2013). MISP is a novel Plk1 substrate required (2007). Experimental and theoretical study of mitotic spindle orientation. for proper spindle orientation and mitotic progression. J. Cell Biol. 200, Nature 447, 493–496. 773–787.

Cell Reports 17, 2488–2501, November 22, 2016 2501