Quantitative kinase and phosphatase profiling reveal that CDK1 phosphorylates PP2Ac to promote mitotic entry

Abstract

The reciprocal regulation of phosphoprotein phosphatases (PPPs) by protein kinases is essential to cell cycle progression and control, particularly during mitosis for which the role of kinases has been extensively studied. PPPs perform much of the serine/threonine dephosphorylation in eukaryotic cells and achieve substrate selectivity and specificity through the interaction of distinct regulatory subunits with conserved catalytic subunits in holoenzyme complexes. Using a mass spectrometry-based chemical proteomics approach to enrich, identify, and quantify endogenous PPP holoenzyme complexes combined with kinase profiling, we investigated the phosphorylation-dependent regulation of PPP holoenzymes in mitotic cells. We found that cyclin-dependent kinase 1 (CDK1) phosphorylated a threonine residue on the catalytic subunit of the phosphatase PP2A, which disrupted its holoenzyme formation with the regulatory subunit B55. The consequent decrease in the dephosphorylation of PP2A-B55 substrates promoted mitotic entry. This direct phosphorylation by CDK1 was in addition to a previously reported indirect mechanism, thus adding a layer to the interaction between CDK1 and PP2A in regulating mitotic entry.

Introduction

The division of a cell into two identical daughter cells is a highly regulated process that is primarily governed by post-translational protein modifications, most prominently protein phosphorylation (1, 2). The cyclin-dependent kinase 1 (CDK1), Polo-like kinase 1 (PLK1), and Aurora kinases A and B (AURKA and AURKB) and the counteracting phosphoprotein phosphatases (PPP) protein phosphatase 1 (PP1), protein phosphatase 2A (PP2A), and protein phosphatase 6 (PP6) are the primary regulators of mitosis. Activation of these mitotic kinases and inhibition of PPPs are essential for entry into mitosis (3–6). While kinase regulation and kinase substrates have been extensively studied in mitosis (7, 8), we are just starting to understand the roles of their cognate phosphatases.

PPPs are responsible for the majority of serine/threonine dephosphorylation in eukaryotic cells and share structurally related catalytic subunits that are among the most conserved proteins from yeast to human (9). PPPs achieve substrate selectivity and specificity through binding to regulatory subunits and formation of holoenzyme complexes (10–12). While the catalytic subunit of PP1 (PPP1CA or PP1c) forms heterodimers with regulatory subunits resulting in over 150–200 holoenzyme complexes (13), the catalytic subunits of PP2A subfamily (including PPP2CA or PP2Ac, PPP4C or PP4c and PPP6C or PP6c) form heterotrimers with one regulatory and one scaffolding protein. For instance, PP2Ac interacts with one of the four families of regulatory “B” proteins (PPP2R2/B55/PR55, PPP2R5/B56/PR61, PPP2R3/B72/PR72 and STRN/PR93/PR110/Striatin) and one scaffolding (PPP2R1A/A/PR65) subunit to form at least 100 different holoenzyme complexes (14–16).

While protein kinases obtain substrate specificity largely through the recognition of consensus amino acid sequences surrounding the phosphorylation site, PPPs recognize substrate through short linear motifs (SLiMs) away from the phosphorylation site (2). For instance, mitotic kinases CDK1, PLK1, AURKA and AURKB phosphorylate serine and threonine residues surrounded by specific consensus sequences, with CDK1 preferring proline-directed motifs, PLK1 preferring acidic motifs, and AURKA and AURKB preferring basic motifs. In contrast, PP1c recognizes a SLiM comprised of the amino-acid sequence “RVxF”. Besides substrates, the majority of regulatory subunits of PP1c also contain this motif (17). Binding of regulatory subunits can create a more selective substrate binding site or bring the phosphatase into the proximity of its substrates (17). PP2Ac in complex with B56 recognizes interacting proteins and substrates through the “LxxIxE” SLiM motif (18, 19).

In addition to their antagonistic action on substrates, protein kinases and phosphatases are subject of their own catalytic activity as well as substrates of each other and as a consequence, they control their own and each other’s activity. For instance, phosphorylation of the activation T-loops of AURKA and PLK1 are opposed by PP6 and PP1, respectively, to regulate their activities (20, 21). Conversely, phosphorylation of the catalytic subunit of PP1 by CDK1 inhibits its activity (22). Furthermore, the removal of an inhibitory phosphorylation site on CDK1 by the dual-specificity phosphatase CDC25 (23) as well as the phosphorylation-dependent inhibition of the phosphatase PP2A-B55 by ARPP19 are both essential to promote mitotic entry (24, 25). Phosphorylation in or near the SLiMs also impacts on substrate/regulatory subunit binding. In the case of the PP1c, phosphorylation in or around the “RVxF” motif reduces protein interaction with PP1c, in contrast, phosphorylation in or near the “LxxIxE” motif enhances the interaction of substrates with PP2A-B56 (18, 26). Thus, mutual control of kinases and phosphatases plays an essential role in the regulation of cell cycle phase transition.

Here, we used PPP and kinase profiling to globally investigate the interplay and counteraction of both enzyme classes in mitosis. We previously developed a chemical proteomics approach for the enrichment, identification, and quantification of endogenous PPP holoenzyme complexes that we refer to as PIB-MS (Phosphatase Inhibitor Beads combined with Mass Spectrometry) (27). PIB-MS can be easily implemented for investigations into PPP signaling without the need to incorporate affinity tags to PPP catalytic subunits, exogenously express PPPs, or the use of PPP-specific antibodies. PIB-MS reproducibly identifies and quantifies PP1, PP2A, PP4, PP5, and PP6 catalytic subunits as well as scaffolding and regulatory subunits. Notably, PIB-MS does not identify PPP catalytic subunits in complex with endogenous inhibitors that directly block the PPP active site, such as α4/IGBP1 (27). Thus, PIB-MS profiles a sub-proteome of PPP holoenzymes that are actively engaged in phosphorylation signaling.

Using PIB-MS with tandem mass tag (TMT) quantification, we quantitatively determine abundance and phosphorylation changes of PPP holoenzymes (also called the “PPPome”). Then we used affinity-based kinome profiling to identify kinases responsible for these phosphorylations. We found that combining PPP and kinome profiling can provide a comprehensive view of the dynamic changes in phosphorylation signaling networks and identify specific actions and counteractions between PPP and kinase family of enzymes. To elucidate the regulatory mechanisms underlying these changes, we compared PPP holoenzyme abundance and global protein abundance changes and discovered that CDK1 phosphorylates the catalytic subunit of PP2A (PP2Ac) on Thr³⁰⁴, which disrupts the formation of the PP2A-B55 holoenzyme. Our data further revealed that this phosphorylation was important for timely mitotic entry and exit through alteration of substrate phosphorylation levels.

Results

Quantitative PIB-MS reveals differences in the PPPome of mitotic and asynchronous cells

To identify differences in PPP holoenzyme composition between mitotic and asynchronous cells, we used a PIB-SILAC (stable isotope labeling with amino acids in cell culture) strategy (fig. S1A). Briefly, HeLa cells were either light-labeled and grown as asynchronous cultures or heavy-labeled and arrested in mitosis using Taxol. PPPs were enriched using phosphatase inhibitor beads (PIBs) and analyzed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) to identify differences in PIB binding between mitotic and asynchronous cell lysates. In this triplicate analysis, we identified and quantified 125 proteins known to be associated with PPPs across all samples (fig. S1B).

To differentiate between proteins that specifically bind to PIBs versus the non-specific background, we pretreated lysates with the PPP-specific inhibitor microcystin-LR to prevent PPP catalytic subunits and their associated proteins from binding to PIBs. We used tandem mass tag (TMT) quantification in a PIB-TMT strategy (Fig. 1A) to combine all conditions into a single LC-MS/MS analysis and improve coverage of proteins across all conditions. Asynchronous or mitotic HeLa cell lysates were either pre-treated with microcystin-LR or left untreated followed by enrichment of PPPome using PIBs and analyzed by LC-MS/MS (Fig. 1A). Using this approach, we were able to identify and quantify 214 proteins as specifically bound to PIBs, of which the binding of 35 proteins was decreased and that of 28 proteins was increased in a statistically significant manner in mitotic compared to asynchronous HeLa cell lysates (Fig. 1B and data file S1).

Fig. 1. PIB-TMT reveals quantitative differences between asynchronous and mitotic PPPome.

(A) Asynchronous or mitotic HeLa cell lysates were either treated with free microcystin-LR or left untreated, followed by enrichment of PPPome using PIBs. The PIB eluates from triplicate experiments were digested into peptides, labeled with tandem mass tags (TMT), combined, fractionated and analyzed by LC-MS/MS. (B) Volcano plot of proteins specifically bound to PIBs in asynchronous or mitotic lysates identified using PIB-TMT strategy. The proteins highlighted in red significantly increased and those in blue significantly decreased in their binding to PIBs in mitotic HeLa cells (fold-change mitotic/asynchronous being > or < respectively, P< 0.05, n=3 independent biological replicates). (C) Proteins identified by PIB-TMT strategy as either asynchronous-specific or mitosis-specific PPP binding proteins were validated by immunoblotting. (D) Scatter plot of the log₂ ratio of PIB binding between mitotic and asynchronous cells of proteins specifically bound to PIBs in PIB-TMT experiment compared to global changes in protein abundance to determine correlation between PIB binding and protein expression. n=3 independent biological replicates.

To validate changes in PPP holoenzyme composition and PPP interacting protein binding between mitotic and asynchronous HeLa cells, we performed Western blot analysis for selected proteins with increased or decreased binding to PIBs in mitosis, including the PP2A regulatory subunit B56δ, PLK1, the kinesin KIF18A, WD-repeat containing protein 92 (WDR92), the telomere-associated protein RIF1, and the serine/threonine-protein kinase LMTK2 (Fig. 1C), which confirmed the LC-MS/MS results.

Differences in PPP holoenzyme composition can be the result of differential protein abundances between cell cycle phases or through other regulatory mechanisms, for example, as described above for phosphorylation of PPP SLiM motifs. For instance, the protein abundance of mitotic regulators, including cyclin B (28), PLK1 (29), and Aurora kinase A (30) is known to differ between interphase and mitosis with many exhibiting increased translation in G2 and targeted degradation at the metaphase to anaphase transition by the anaphase-promoting complex/cyclosome (APC/C) (31, 32). Thus, differences observed in PPPome composition could be due to differences in protein abundance inputs. For PLK1 and the kinesin KIF18A, this was apparent in the Western blot analysis, which showed increased abundance not only in PIB eluates but also in the input mitotic lysates (Fig. 1B and 1C).

To globally and comprehensively compare changes in cellular protein abundance with changes in protein binding to PIBs, we performed global quantitative proteomics on asynchronous and mitotic HeLa cell lysates (fig. S2). Comparison of protein abundance and PIB binding changes between asynchronous and mitotic HeLa cells confirmed that differences in protein abundance correlate with differences in PIB binding for some PPP-associated proteins, including PLK1 and the kinesin KIF18A (Fig. 1D, and data file S2). However, for other PPP-associated proteins, changes in PIB binding are not due to differences in protein abundances. For instance, while RIF1 and LMTK2 exhibit comparable expression in asynchronous and mitotic HeLa cells, their binding to PIBs is significantly reduced in mitosis (Fig. 1, C and D, and data file S2). This suggests that mechanisms such as post-translational regulation by protein modifications might control the association of proteins with the phosphatases during cell cycle progression.

PhosPIB identifies mitotic phosphorylation sites on PPP regulatory subunits

To identify changes in the phosphorylation status of the PPPome, we searched again within our PIB-SILAC LC-MS/MS data (Fig. 1B) for phosphorylation sites and identified a total of 139 on PPP subunits and interacting proteins, of which 44 phospho-sites significantly increased while 14 significantly decreased in mitosis (fold-change mitotic/asynchronous being > or <, respectively, 1.5-fold; P < 0.05; data file S3). To increase the coverage of phosphorylation sites on PPPs, we performed phosphopeptide enrichment of PIB eluates, hereafter called the PhosPIB approach. Briefly, samples were processed as described above and after TMT-labeling and mixing, phosphopeptides were enriched from PIB eluates (Fig. 2A). Using this approach, we identified a total of 681 phosphorylation sites on 105 proteins specifically captured by PIBs. To account for changes in protein abundance between the asynchronous and mitotic conditions, we corrected phosphopeptide abundance by protein abundance. After this correction, 401 phosphorylation sites were significantly increased while 78 were significantly decreased in mitosis (Fig. 2B and data file S3). 321 of these phosphorylation sites were identified on PP1, 101 on PP2A, 40 on PP4, and 18 on PP6 catalytic or regulatory subunits (Fig. 2C).

Fig. 2. PhosPIB approach identifies differential phosphorylation sites on PPP regulatory proteins.

(A) Schematic of process to identify differential phosphorylation sites on PPP regulatory proteins: PIB eluates from asynchronous or mitotic HeLa cells were digested into peptides, TMT-labeled, and enriched for phosphorylated peptides using Fe-NTA columns prior to LC-MS/MS analysis. (B) Volcano plot of phosphorylation sites identified by phosphopeptide enrichment of PIB eluates from asynchronous or mitotic cells. Sites highlighted in blue significantly decreased and those in red significantly increased in mitotic HeLa PIB eluates (P<0.05, n=3 independent biological replicates). (C) Phosphorylation sites identified on PP1, PP2A, PP4 or PP6 catalytic or regulatory subunits that increased or decreased significantly (red or blue, respectively; P<0.05) in mitotic PIB eluates. (D) Enriched phosphorylation site motifs from phosphopeptides that significantly increased (P < 0.05, log₂ fold change of 1.5) in mitotic or asynchronous PIB eluates. (E) Significantly increased or decreased (P < 0.05) phosphorylation sites from mitotic PIB eluates were classified as either proline-directed, basic (basophilic) or acidic (acidophilic) per surrounding amino acids.

Among these phosphorylation sites, we identified Thr³²⁰ on the catalytic subunit of PP1α as increasing in mitosis (Fig. 2B and C, and data file S3). PP1α Thr³²⁰ was previously shown to be phosphorylated during mitosis by CDK1 and to inhibit PP1α phosphatase activity (22). Furthermore, we also identified other known mitotic phosphorylation sites on PPP subunits including Ser²²⁰⁵ in telomere-associated protein RIF1, Ser⁴⁷³ in myosin phosphatase targeting subunit 1 (MYPT1) and Ser⁶⁷⁴ and Ser⁶⁸⁴ in the kinesin KIF18A (20, 26, 33) (Fig. 2C, and data file S3). Ser⁴⁷³ in MYPT1 is phosphorylated by CDK1, generating a docking site for Plk1 interaction and is implicated in centrosome maturation and mitotic exit. We also identified Ser⁵⁰⁷ (Fig. 2C, and data file S3) in MYPT1, which functionally opposes Ser⁴⁷³, as one of the sites that significantly decrease in phosphorylation during mitosis (34). Together, this supports the notion that the PhosPIB approach can identify functionally important phosphorylation sites that are differentially regulated during cell cycle progression.

To determine what kinases might be responsible for these phosphorylation modifications, we generated motif logos of sites that were significantly increased in asynchronous or mitotic cells (35). This revealed that phosphorylation sites increased in mitosis were enriched for proline-directed motifs (Fig. 2D). To further explore this, we asked how many sites had either a proline-directed, a basic, or an acidic phosphorylation site motif. We found that 77% of mitotic phosphorylation sites compared to 49% of asynchronous phosphorylation sites had a proline-directed motif, pointing to a role of—for instance—cyclin-dependent kinases (a family with proline-directed consensus motifs) in the phosphorylation of the PPPome (Fig. 2E).

Kinome profiling identifies differences in the kinase expression in asynchronous and mitotic cells

To experimentally determine which protein kinases are active and expressed in both states, we reviewed our global proteomics dataset and identified and quantified 86 kinases (fig. S2), which is a significant underrepresentation of the human kinome. Indeed, many important cell cycle-related kinases are of low abundance and are not detected using global quantitative proteomics. To increase our coverage of such kinases, we performed kinome profiling using multiplexed kinase-inhibitor conjugated beads (MIBs) (36). MIBs are a mixture of sepharose beads with covalently immobilized selective and broad pan-kinase inhibitors, including UNC8088A, PP58, purvalanol B, UNC2147A, VI-16832 and CTx-0294885 (Fig. 3A). MIB enrichment of kinases was coupled with quantitative TMT labeling to profile the differential abundance of the human kinome in mitotic cells (Fig. 3B). Using this approach, we identified and quantified a total of 280 protein kinases with two or more peptides that belong to 7 human protein kinase superfamilies as depicted in a phylogenetic tree (Fig. 3C). Out of the total 280 kinases identified, 14 protein kinases significantly increased in abundance in mitotic cells while 51 significantly decreased in abundance in mitotic cells (Fig. 3D and data file S4). Comparison of kinase binding to MIBs (data file S4) and protein abundances (data file S2) revealed that in many instances—for example: PLK1, AURKA, AURKB, and WEE1—the observed increase or decrease in kinase binding to MIBs in mitosis was due to their differential protein abundance in mitotic and asynchronous cells (fig. S3). Among the protein kinases which significantly increase in abundance in mitotic cells, only CDK1 had a reported preference for the phosphorylation of proline-directed consensus motifs (fig. S4) (37). Thus, it is likely that CDK1 is the main kinase phosphorylating and potentially regulating the differential association of proteins with PPPs in mitosis.

Fig. 3. MIB-TMT approach uncovers the differential kinome between asynchronous and mitotic cells.

(A) Illustration showing the multiplexed inhibitor beads (MIBs), which are sepharose beads covalently attached to linker-adapted small molecule kinase inhibitors. (B) Schematic of process with which MIBs were used to enrich for kinases from asynchronous or mitotically arrested HeLa cells: MIB eluates from three replicates were digested with trypsin, labeled with TMT reagents, combined, fractionated and analyzed by LC-MS/MS. (C) Phylogenetic kinome tree depicting the protein kinase super families of all 280 human kinases identified in the MIB-TMT experiment. (D) Plot showing log₂ fold changes of all kinases that significantly increased or decreased (P < 0.05) in mitotic vs asynchronous MIB eluates. Data are means ± S.E.; n=3 biological replicates.

PP2Ac is phosphorylated at Thr³⁰⁴ during mitosis by the CDK1-cyclin B complex

To determine the functional significance of our findings, we focused on threonine at residue 304 (Thr³⁰⁴) in the catalytic subunit of PP2Ac that increased in mitosis (Fig. 2B and C, and data file S3). This increase was not due to changes in PP2Ac protein abundance (data file S2). It was previously shown that a phosphomimetic mutation of Thr³⁰⁴ reduced the binding of B55 and B56α, but not B56δ, regulatory subunits to PP2Ac, whereas a phospho-ablating mutation showed equal or greater binding than wild-type PP2Ac (38–40). However, the biological context in which PP2Ac Thr³⁰⁴ is phosphorylated, the cellular function of this phosphorylation, and the upstream kinase responsible for it are not currently known.

To comprehensively and quantitatively determine the effects of the PP2Ac Thr³⁰⁴ phospho-ablating (T304A) and phospho-mimetic (T304D) mutations on PP2A holoenzyme composition, we generated HEK293T cells stably expressing PP2Ac WT, T304A, and T304D expression construct. We used these cell lines to perform FLAG pull-downs of PP2Ac WT, T304A, and T304D and analyzed them by LC-MS/MS (data file S5). Our results confirmed previous findings that the phospho-mimetic PP2Ac T304D mutant displayed significantly reduced binding to all B55 isoforms (Fig. 4A), while the phospho-ablated PP2Ac T304A mutant increased in binding to all B55 regulatory subunits (Fig. 4B). The PP2Ac T304D mutant also showed significantly reduced binding to certain members of the B56 family (α, β and ε), but not others (δ and γ) compared to PP2Ac WT, suggesting different isoforms of the B56 regulatory protein family might have different mechanisms of binding to the catalytic PP2A subunit (data file S5). Interestingly, PP2R3A and PP2R3B but not PP2R3C signficantly increase in their interaction with PP2Ac-T304D compared to PP2Ac-WT (data file S5). We confirmed some of these observations by immunoblotting using B55α, B56δ and FLAG antibodies (Fig. 4C).

Fig. 4. PP2Ac is phosphorylated at Thr³⁰⁴ during mitosis and regulates PP2A holoenzyme composition.

(A and B) Relative quantification of iBAQ abundances, by LC-MS/MS, of B55 family regulatory proteins in complex with FLAG-tagged WT or mutant [T304D (A) or T304A (B)] PP2Ac purified by FLAG-M2 resin from HEK293T cells. Data are means ± S.E. from n=3 biological replicates. * P < 0.05 by two-tailed Student’s t-test. (C) Immunoblotting for FLAG, B56δ and B55α in FLAG eluates from cells expressing FLAG-tagged PP2Ac WT, T304A or T304D; n=3. (D and E) FLAG-M2 affinity pulldown and LI-COR immunoblotting analysis in asynchronous or mitotically arrested HEK293T cells stably expressing FLAG-tagged WT or T304A PP2Ac to validate the phospho-specific Thr³⁰⁴ PP2Ac antibody. Data are means ± S.E. from n=4 independent blots. * P < 0.05 by two-tailed Student’s t-test. (F) Representative in vitro kinase assay using bacteria-purified GST-PP2Ac and the CDK1-cyclin B complex; n≥2. (G and H) LI-COR Western blotting and analysis of FLAG-PP2Ac eluates from HEK293T cells stably expressing FLAG-tagged PP2Ac were left asynchronous (Async), synchronized in mitosis (Mitotic) and treated with calyculin A (100 nM for 30 min), flavopiridol (10 μM for 15 min) or RO-3306 (10 μM for 15 min). Data are means ± S.E., n=3 independent replicates; * P < 0.05 by two-tailed Student’s t-test. (I) Immunoblotting for pThr³⁰⁴ and total PP2Ac using LI-COR in PIB-mediated PP2Ac pull-downs from G1/S-arrested or mitotically arrested HeLa cells. Data is representative of n=3 experiments.

Next, we wanted to determine the upstream kinase responsible for PP2A phosphorylation at amino acid residue Thr³⁰⁴. Thr³⁰⁴ is a proline-directed phosphorylation site, and our kinome analysis suggested CDK1 as a likely candidate. To test if CDK1 is indeed the upstream kinase responsible for PP2Ac Thr³⁰⁴ phosphorylation, we raised a phospho-specific antibody against a phosphorylated version of the C-terminal peptide of PP2Ac (CPHVTRRpTPDYFL). To validate the specificity of this antibody for the phosphorylation site, we immunoprecipitated FLAG-tagged PP2Ac WT or PP2Ac T304A mutant from HEK293T cell lysates and performed Western blots. The phospho-specific antibody only recognized the wild-type FLAG-tagged PP2Ac and not the mutated PP2Ac-T304A (Fig. 4D). Using this phospho-specific antibody, we confirmed that phosphorylation of Thr³⁰⁴ of PP2Ac was increased in mitosis (Fig. 4, D and E). Next, we performed in vitro kinase assays using CDK1-cyclin B complex and GST-tagged, phosphatase-dead PP2Ac (PP2Ac H118N mutant). Western blotting with the anti-pThr³⁰⁴ PP2Ac antibody demonstrated that CDK1-cyclin B phosphorylates PP2Ac in vitro (Fig. 4F). To determine if CDK1 phosphorylates PP2Ac on Thr³⁰⁴ in cells, we inhibited CDK activity using flavopiridol or RO-3306 in mitotically arrested cells expressing FLAG-tagged PP2Ac. Treatment with flavopiridol or RO-3306 strongly reduced Thr³⁰⁴ levels as determined by Western blotting of FLAG immunoprecipitations using the pThr³⁰⁴ PP2Ac antibody. As shown before, phosphorylation of Thr³⁰⁴ on PP2Ac increased in mitotically arrested cells compared to asynchronous cells (Fig. 4G, H). This phosphorylation was further increased upon the inhibition of PPP activity by calyculin A (Fig. 4G, H), suggesting this site is dynamically turned over in cells by the opposing actions of serine/threonine phosphatases and CDK1. To determine if this phosphorylation event also occurs on endogenous PP2Ac, we used PIBs to pull-down endogenous PP2Ac from asynchronous or mitotically-arrested cells and used the phospho-specific antibody against Thr³⁰⁴ to detect phosphorylation. As for FLAG-PP2Ac, we see increased phosphorylation of Thr³⁰⁴ in endogenous PP2Ac in mitotic cells (Fig. 4I).

Phosphorylation of PP2Ac at Thr³⁰⁴ regulates mitotic timing and PP2A-B55–dependent dephosphorylation at mitotic exit

Inhibition of PP2A-B55 is necessary to promote mitotic entry and to maintain the mitotic state of cells. It has been shown that the depletion of PP2A-B55 accelerates mitotic entry and arrests cells indefinitely in mitosis (41, 42). Phosphorylation of PP2Ac at Thr³⁰⁴ abrogates binding to B55 family regulatory subunits based on the analysis of the phospho-mimetic PP2Ac T304D mutant (Fig. 4, A and C) (38–40). Thus, we sought to test if this phosphorylation of PP2Ac affects mitotic entry and exit timing. We depleted endogenous PP2Ac using siRNAs against both α and β isoforms of PP2Ac in HeLa cells expressing siRNA-resistant forms of either WT, T304D mutant or T304A mutant, Myc-tagged PP2Ac (Fig. 5A and fig. S5A). Next, we determined the percentage of cells entering mitosis as determined by cell rounding and nuclear envelope breakdown and the duration of mitosis from that breakdown to anaphase onset by time-lapse microscopy (Fig. 5B). As previously reported (41, 42), HeLa cells depleted of endogenous PP2Ac showed accelerated mitotic entry in mitosis compared to control cells (Fig. 5C). This was rescued by the expression of a siRNA-resistant version of PP2Ac WT (Fig. 5C). Notably, HeLa cells expressing a siRNA-resistant version of the phospho-ablated PP2Ac T304A were significantly delayed in mitotic entry compared to PP2Ac WT expressing cells as indicated by the smaller percentage of cells entering mitosis. Conversely, Hela cells expressing a siRNA-resistant phospho-mimetic PP2Ac T304D mutant entered mitosis more readily (Fig. 5C). When we investigated the duration of time required by the cells to progress from nuclear envelope breakdown to anaphase, we found that wild-type HeLa cells took a median time of 50 minutes, while HeLa cells transfected with PP2Ac siRNAs were significantly delayed (median time 460 minutes). Expression of siRNA-resistant PP2Ac WT or PP2Ac T304A rescued this delay, while the expression of PP2A T304D delayed mitotic progression compared to PP2Ac WT expressing cells (Fig. 5D).

Fig. 5. Phosphorylation of PP2Ac regulates mitotic timing and PP2A-B55–dependent dephosphorylation at mitotic exit.

(A) Representative blotting for PP2Ac, Myc, GFP and tubulin in HeLa-Flp-In T-REx cells that were knocked down of endogenous α and β isoforms of PP2Ac and doxycycline-induced to express siRNA-resistant Myc-tagged WT PP2Ac. (B) Live cell imaging of cells expressing doxycycline-inducible Myc-PP2Ac as they go through mitosis. (C) Bar plot of the percentage of cells (depleted endogenous PP2Ac and expressing siRNA-resistant WT or Thr³⁰⁴-mutant PP2Ac) that entered mitosis, as determined by nuclear envelope breakdown and cell rounding. Data are mean ± SD from at least 80 cells per condition in a representative of three independent experiments; ns: non-significant, *P<0.05 by Mann-Whitney U-test. (D) Time from nuclear envelope breakdown to mitotic exit of cells depleted of endogenous PP2Ac and expressing either WT or Thr³⁰⁴-mutant PP2Ac. Each dot represents a single cell; red line indicates the median time. A representative result from three independent experiments is shown; ns: non-significant, *P<0.05 by Mann-Whitney U-test. (E) Nocodozaole-treated (mitotically arrested) HeLa cells that had been depleted of endogenous PP2Ac and doxycycline-induced to express WT, T304A or T304D PP2Ac, were pre-treated with MG132 (30 μM for 30 mins) then split and treated with flavopiridol (20 μM) to induce mitotic exit and collected every 5 mins for immunoblotting analysis as indicated. (F) Immunofluorescent imaging and analysis of TPX2 and tubulin staining in cells expressing doxycycline-induced Myc-PP2Ac T304D and control or PP2Ac shRNA. Data are relative intensity of TPX2 at telophase bridges compared to tubulin intensity, calculated from at least 20 cells imaged per condition over three independent experiments; *P<0.05. Scale bar, 10 μm. (G) Model depicting the role of PP2Ac Thr³⁰⁴ phosphorylation in regulating PP2A holoenzyme composition.

PP2A-B55 holoenzymes have been implicated in the ordered dephosphorylation of CDK1 substrates during mitotic exit (43). To determine if the inability of the phospho-mimetic T304D mutant to form PP2A-B55 holoenzymes affects the dephosphorylation of B55-dependent substrates, we quantified the extent of dephosphorylation of CDK1 substrates during mitotic exit by Western blot. HeLa cells expressing siRNA-resistant forms of PP2Ac WT or PP2Ac T304D were depleted of endogenous PP2Ac, arrested in mitosis, collected by mitotic shake-off, and treated with the CDK1 inhibitor flavopiridol to force an exit from mitosis as previously described for B55-depleted cells (44). The timing of CDK1 substrate dephosphorylation during mitotic exit was monitored using a phospho-CDK substrate (pTPXK) antibody. Upon mitotic arrest, HeLa cells expressing PP2Ac-T304D had a two-fold or 100% increase in total CDK substrate phosphorylation compared to PP2Ac-WT expressing cells (Fig. 5E). Furthermore, a comparison of CDK substrate phosphorylation levels during mitotic arrest and at 10 min upon induction of mitotic exit revealed a decrease by 83% and 70% in PP2Ac-WT and PP2A-T304D expressing cells, respectively, (Fig. 5E). This indicates reduced PP2A function in cells expressing T304D, possibly due to the abrogation of PP2A-B55 holoenzyme formation.

Disruption of PP2A-B55 holoenzyme function alters the localization of a mitotic spindle assembly factor TPX2 (targeting protein for Xklp2) (43). Recruitment of TPX2 to the central spindle and midbody in anaphase and telophase requires B55-dependent dephosphorylation. In the absence of B55, TPX2 is directly recruited to the nucleus (43). Upon depletion of endogenous PP2Ac and expression of PP2Ac T304D, we found that TPX2 was imported into the nucleus and not to the central spindle and midbody (Fig. 5F), suggesting impaired PP2A-B55 function. In contrast, in cells depleted of endogenous PP2Ac and expressing either PP2Ac WT or T304A, TPX2 localized to the central spindle and midbody normally (fig. S5B).

Together, our data suggest that PP2A Thr³⁰⁴ phosphorylation by CDK1 disrupts PP2A-B55 holoenzyme formation and thereby phosphatase activity in mitosis. This inhibition results in increased CDK1-dependent phosphorylation levels of substrates that maintain the mitotic state and prevent the transition into anaphase (Fig. 5G).

Discussion

In this study, we combined “PPP-ome” and kinome profiling to investigate phosphoprotein phosphatase regulation by kinases in asynchronous and mitotic HeLa cells. By combining both approaches, we were able to investigate the reciprocal regulation and dynamic changes of both classes of enzymes in asynchronous and mitotic states. While the role of kinases in phosphorylation signaling is well established and appreciated, the contribution of PPPs is still emerging. With PIB-MS, this frontier is now within our reach. In contrast to affinity purification approaches, PPPome and kinome profiling by inhibitor beads does not rely on antibodies or tagging of PPP subunits or kinases and allows for profiling of endogenous phosphatases and kinases. Thus, this approach is widely applicable to any type of cell or tissue. By combing PIB-MS with phosphopeptide enrichment, we were also able to identify phosphorylation sites on PPP regulatory subunits and their interacting proteins to investigate the mechanisms which govern the formation of these complexes. For instance, in the case of the PP1 regulatory protein RIF1, binding to PP1c has been shown to be regulated by phosphorylation of Ser²²⁰⁵ within the PP1 binding SLiM (²²⁰²RRVSF²²⁰⁶) during mitosis (26). Using our approach, we were able to validate the abrogation of the RIF1-PP1 holoenzyme complex as shown by a decrease in RIF1 binding to PIBs in mitosis, which is accompanied by increased phosphorylation of Ser²²⁰⁵ of RIF1. This highlights that PhosPIB is a powerful approach to identify phosphorylations on PPP regulatory proteins, which play a critical role in PPP holoenzyme formation.

PP2A holoenzyme complexes are critical in regulating cell cycle progression (45). The activity of each PP2A holoenzyme complex is strictly regulated at mitotic entry and mitotic exit (25, 41, 46–48). Using PPP and kinase profiling, we were able to identify an important regulatory mechanism that links PP2A and CDK1 and contributes to the regulation of mitotic progression. Previous studies have shown that PP2Ac can be post-translationally modified at the C-terminus by phosphorylation and methylation (39, 49, 50). Structurally, C-terminal phosphorylation of Thr³⁰⁴ has been predicted to disrupt the interaction of PP2Ac and its regulatory subunits, and mutational studies have confirmed this (16, 38–40, 51). However, until now, we lacked an understanding of the signaling context in which phosphorylation occurs, the upstream kinases responsible for this phosphorylation, and the functional implications of these modifications. Through our experiments here, we have shown that PP2Ac Thr³⁰⁴ is phosphorylated in mitosis and that this phosphorylation requires CDK1-cyclin B activity. CDK1-cyclin B has been shown to regulate mitotic activity of PP2A-B55 indirectly through the activation of the kinase Greatwall (GWL) (52, 53). Upon activation in mitosis, CDK1 and GWL itself phosphorylate and activate GWL, which in turn phosphorylates its substrates α-endosulfine (ENSA) and cyclic adenosine monophosphate–regulated phosphoprotein 19 (ARPP19), converting them into potent and specific inhibitors of PP2A-B55 holoenzyme complex promoting mitotic entry (24, 25, 52).

We believe that direct phosphorylation of PP2Ac at Thr³⁰⁴ by CDK1- cyclin B complex provides an additional layer of temporal regulation to reduce PP2A-B55 activity upon CDK1-cyclin B activation at the G₂/M transition of cell cycle to promote mitotic entry. Temporal regulation by phosphorylation can also be observed at mitotic exit, where timely reactivation of PP2A-B55 is important to dephosphorylate mitotic CDK1 substrates (43, 47). We have shown here that dephosphorylation of CDK1 substrates during mitotic exit is disrupted in cells expressing the PP2Ac-T304D, which is unable to bind B55 subunits. This is further confirmed by the mislocalization of the PP2A-B55 substrate TPX2 to the nucleus instead of the midbody at anaphase. This suggests that reactivation of the PP2A-B55 complex at mitotic exit is not only dependent on CDK1-cyclin B inactivation but also involves the dephosphorylation of Thr³⁰⁴ of PP2Ac. This suggests a possible feedback mechanism with contribution from other serine/threonine phosphatases which can be inhibited by calyculin A. Indeed, this type of phosphatase feedback regulation has been shown before in fission yeast where, upon the decline of CDK1-cyclin B activity, PP1 reactivates itself, PP2A-B55, and PP2A-B56 in a phosphatase relay that coordinates PPP activities during mitotic exit (54).

Faithful segregation of chromosomes in mitosis involves strict spatial and temporal regulation of proteins. Phosphorylation-dependent signaling networks have largely been studied from the perspective of kinases, but we still lack knowledge of signaling networks from the protein phosphatase standpoint. The work presented here describes an advancement in the development of new strategies for the interrogation of phosphoprotein phosphatase networks, adding to our understanding of phosphorylation signaling in mitosis.

Materials and Methods

Cell culture and reagents

HEK293T and HeLa cells were cultured in Dulbecco’s modified Eagle’s media (Cellgro Mediatech, Inc., Manassas, VA) with 10% fetal bovine serum (Hyclone, Logan, Utah) and penicillin-streptomycin (100 U ml⁻¹ and 100 μg ml⁻¹, respectively; Cellgro Mediatech, Inc.) at 37°C in a humidified incubator with 5% CO₂. For SILAC experiments, HeLa cells were grown in ‘light’ (containing 100 mg/L ¹²C₆,¹⁴N₄-arginine and 100 mg/L ¹²C₆,¹⁴N₂-lysine, SIGMA) or ‘heavy’ (containing 100 mg/L ¹³C₆,¹⁵N₄-arginine and 100 mg/L ¹³C₆,¹⁵N₂-lysine, Cambridge Isotope Laboratories) SILAC media (Fischer Science) for at least six population doublings. After amino acid incorporation, light-labeled cells were used as asynchronous control and heavy-labeled cells were synchronized in mitosis using 2 mM thymidine (SIGMA) for 22 hours, and a 3-hour release, followed by Taxol (Sigma) arrest for 16 hours.

CDK1 inhibitor experiments were performed in HEK293T cells arrested in mitosis using thymidine/taxol block followed by pre-incubation with 20 μM MG132 (Selleck Chemicals) for 30 min. These cells were treated with either flavopiridol (SIGMA; 5 μM) or RO-3306 (Tocris; 5 μM), for 15 min and collected for further experiments. Mitotic HEK293T cells were treated with calyculin A (LC Laboratories); final concentration 100 nM, for 30 min.

Antibodies

Antibodies used in the study: PP1 (sc-7482, Santa Cruz Biotechnology), PP2A (gift from Dr. Egon Ogris, Medical University of Vienna), B56δ (5687S, Cell Signaling Technology), KIF18A (sc-390600, Santa Cruz Biotechnology), WDR92 (sc-376734, Santa Cruz Biotechnology), RIF1 (A300–569A, Bethyl Laboratories, Inc.), LMTK2 (A394–270A-T, Bethyl Laboratories, Inc.), pCDK1 substrate pTPxK (14371S, Cell Signaling Technology), B55α (sc-365282, Santa Cruz Biotechnology), and FLAG (SIGMA). The phospho-Thr³⁰⁴ antibody was generated in rabbits using the peptide CPHVTRR(Tphos)DYFL coupled to BSA and KLH using the N-terminal Cys. The antibody was affinity-purified by first using a non-phosphorylated form of the peptide followed by enrichment using the phosphorylated peptide and used at a dilution of 1:100. Primary antibodies were used at 1:1000 dilution with overnight incubation at 4°C, unless stated otherwise. For detection using LICOR, IRDye® 800CW goat anti-rabbit and IRDye® 680RD goat anti-mouse were used at 1:5000 dilution for 30 min. All Western blots signals were detected by either Chemi DOC XRS (Bio-Rad) or ODYSSEY® CLx (LI-COR).

PIB pull-downs and TMT labeling

PIBs were generated as described before (27). Asynchronous or mitotic cells were collected, lysed in lysis buffer (50 mM Tris pH 7.5 (SIGMA), 500 mM NaCl (SIGMA), 5 mM beta-glycerophosphate (SIGMA), 0.5% (w/v) Triton X-100 (SIGMA), 0.1 mM DTT (SIGMA), and one EDTA-free protease inhibitor tablet (Roche) per 10 mL of lysis buffer) and incubated with PIBs (15 mg lysate with 100 μl PIBs per replicate) for 2 h at 4°C rotating end over end. To determine the specific PIB binding by competition with free microcystin-LR (MCLR; Millipore), asynchronous or mitotic cell lysates were pre-incubated with 100 nM MCLR for 30 mins at 4°C before incubating with PIBs. The PIBs were washed three times with lysis buffer and eluted with 1% (w/v) SDS overnight at room temperature.

In case of PIB-SILAC, the eluates collected from light (asynchronous) and heavy (mitotic) cells were mixed, precipitated using 20% tri-chloroacetic acid (TCA), washed twice with acetone (Burdick & Jackson, Muskegon, MI) and digested overnight in 25 mM ammonium bicarbonate with trypsin (Promega) for mass spectrometric analysis.

In case of PIB-TMT, PIB pull-downs were performed in triplicates each for asynchronous, mitotic and asynchronous with MCLR lysates and in duplicate for mitotic with MCLR lysate to process the experiment as one TMT-11plex. PIB eluates were TCA precipitated and digested with trypsin in 166 mM HEPES (SIGMA), pH 8.5 overnight at 37°C. Individual TMT reagent aliquots (0.8 mg) (ThermoFisher Scientific) were resuspended in 85 μl dry acetonitrile, out of which 2 μl reagent was added to each digested peptide aliquot and vortexed. After 1 hour at room temperature, the reactions were quenched with 3 μl of 500 mM ammonium bicarbonate solution for 10 minutes, mixed into one 11-plex, acidified with 20% trifluorocetic acid, and desalted using OASIS HLB C₁₈ desalting plate. The desalted multiplex was dried by vacuum centrifugation. Phosphopeptides were enriched in the mixed TMT PIB sample using High-Select™ Fe-NTA Phosphopeptide Enrichment Kit according to manufacturer’s protocol (ThermoFisher Scientific).

MIB pull-down

MIBs were generated as previously described (36). Asynchronous or mitotic cells were collected, lysed in lysis buffer (50 mM Tris pH 7.5 (SIGMA), 1M NaCl (SIGMA), 5 mM beta-glycerophosphate (SIGMA), 0.5% (w/v) Triton X-100 (SIGMA), and one EDTA-free protease inhibitor tablet (Roche) per 10 mL of lysis buffer). Lysates were passed over MIB columns, washed with lysis buffer and modified lysis buffer (50 mM Tris pH 7.5 (SIGMA), 150 mM NaCl (SIGMA), 0.1% SDS (SIGMA)). MIBs were eluted at 95°C for 15 min twice. Eluates were processed as described for PIBs.

Stage-tip fractionation

Basic reverse phase fractionation of TMT-labelled PIB and MIB eluates were performed as described in (55). Briefly, stage tip columns were prepared using C18 solid phase extraction disks (Empore) and pre-equilibrated using 100% acetonitrile followed by 100 mM ammonium bicarbonate, pH 8.0. The TMT-labelled peptide sample was resuspended in 5% acetonitrile with 0.1% formic acid and loaded onto the column. The sample was fractionated by eluting with 50 μl of elution buffer with increasing acetonitrile concentrations (10%, 15%, 20%, 25% and 50% acetonitrile in 100 mM ammonium bicarbonate, pH 8.0). The eluates were dried by vacuum centrifugation and desalted again using C₁₈ solid phase stage tips before LC-MS/MS analysis.

LC-MS/MS analysis

PIB pulldowns were analyzed on a Q-Exactive Plus quadrupole Orbitrap mass spectrometer (ThermoFisher Scientific) equipped with an Easy-nLC 1000 (ThermoFisher Scientific) and nanospray source (ThermoFisher Scientific). Peptides were resuspended in 5% methanol / 1% formic acid and loaded on to a trap column (1 cm length, 100 μm inner diameter, ReproSil, C₁₈ AQ 5 μm 120 Å pore (Dr. Maisch, Ammerbuch, Germany)) vented to waste via a micro-tee and eluted across a fritless analytical resolving column (35 cm length, 100 μm inner diameter, ReproSil, C18 AQ 3 μm 120 Å pore) pulled in-house (Sutter P-2000, Sutter Instruments, San Francisco, CA) with a 45 minute gradient of 5–30% LC-MS buffer B (LC-MS buffer A: 0.0625% formic acid, 3% ACN; LC-MS buffer B: 0.0625% formic acid, 95% ACN). The Q-Exactive Plus was set to perform an Orbitrap MS1 scan (R=70K; AGC target = 1e6) from 350 – 1500 m/z, followed by HCD MS2 spectra on the 10 most abundant precursor ions detected by Orbitrap scanning (R=17.5K; AGC target = 1e5; max ion time = 50ms) before repeating the cycle. Precursor ions were isolated for HCD by quadrupole isolation at width = 1 m/z and HCD fragmentation at 26 normalized collision energy (NCE). Charge state 2, 3, and 4 ions were selected for MS2. Precursor ions were added to a dynamic exclusion list +/− 20 ppm for 15 seconds. Raw data were searched using COMET (release version 2014.01) in high resolution mode(56) against a target-decoy (reversed) (57) version of the human proteome sequence database (UniProt; downloaded 1/2020) with a precursor mass tolerance of +/− 1 Da and a fragment ion mass tolerance of 0.02 Da, and requiring fully tryptic peptides (K, R; not preceding P) with up to three mis-cleavages. Static modifications included carbamidomethyl cysteine and variable modifications included oxidized methionine. For SILAC experiments, heavy lysine (+ 8.01420 Da) and arginine (+ 10.00827 Da) were also allowed as variable modifications. Searches were filtered using orthogonal measures including mass measurement accuracy (+/− 3 ppm), Xcorr for charges from +2 through +4, and dCn targeting a <1% FDR at the peptide level. Quantification of LC-MS/MS spectra was performed using MassChroQ(58) and the iBAQ method (59).

TMT quantitative proteomics

HeLa cells were left unsynchronized, synchronized either in G₁/S phase using double thymidine block or in mitosis using thymidine/taxol block in biological triplicate. Cells were collected and lysed in lysis buffer (8 M urea and 50 mM Tris pH 8.1 with protease inhibitors) in the presence of protease inhibitors. After lysis, a small aliquot was removed for BCA (Pierce) analysis, and the remaining lysate was reduced with 5 mM DTT at 50°C for 30 min followed by alkylation with 15 mM iodoacetamide in dark for 1 h. The lysates were then diluted fivefold with 25 mM Tris pH 8.1 and digested overnight with trypsin (1:100 w/w) at 37°C. The digests were desalted using C₁₈ solid-phase extraction cartridges (ThermoFischer Scientific) and an aliquot of each of the desalted eluates corresponding to 40 μg of peptide digest was dried by vacuum centrifugation in separate tubes. For TMT-labeling, acetonitrile to a final concentration of 20% in 166 mM HEPES was added and peptides were transferred to dried, individual TMT 11-plex reagent (ThermoFisher Scientific), vortexed, and mixed with the reagent. After 1 hour at room temperature, each digest was quenched with 5 μl of 500 mM ammonium bicarbonate solution for 10 min, combined, diluted threefold with 0.1% TFA in water, and desalted using C₁₈ solid-phase extraction cartridges (ThermoFisher Scientific). The desalted multiplexes were dried by vacuum centrifugation, resuspended in 60 μl of 3% ACN/0.1% TFA in water, and separated on a pentafluorophenyl (PFP) analytical column [XSelect HSS PFP XP column, Waters; 100 A, 2.5 mm inner diameter (ID), 150 mm in length] into 48 fractions by gradient elution [buffer A (3% ACN, 0.1% TFA); buffer B (95% ACN, 0.1% TFA)] with 5% B, 0 to 1 min; 5 to 11% B, 1 to 2 min; 11 to 47% B, 2 to 60 min; 47 to 100% B, 60 to 61 min; 100% B, 61 to 69 min; 100 to 5% B, 69 to 70 min; 5% B, 70 to 110 min, all at a flow rate of 0.15 ml/min. The 48 fractions were reduced into 24 fractions and dried by vacuum centrifugation.

Each fraction from TMT proteomics was analyzed using an Easy LC-1000 (Proxeon) and Orbitrap Fusion(60) (ThermoFisher Scientific) LC-MS/MS platform across a 2-hour gradient from 3% acetonitrile/0.0625% formic acid to 37% acetonitrile/0.0625% formic acid. The Orbitrap Fusion was operated in data-dependent, SPS-MS3 quantification mode (61, 62) wherein an Orbitrap MS1 scan was taken (scan range = 350 – 1200 m/z, R = 120K, AGC target = 3e5, max ion injection time = 100ms). Followed by data-dependent Orbitrap trap MS2 scans on the most abundant precursors for 3 seconds. Ion selection; charge state = 2: minimum intensity 2e5, precursor selection range 650–1200 m/z; charge state 3: minimum intensity 3e5, precursor selection range 525–1200 m/z; charge state 4 and 5: minimum intensity 5e5). Quadrupole isolation = 0.7 m/z, R = 30K, AGC target = 5e4, max ion injection time = 80ms, CID collision energy = 32%). Orbitrap MS3 scans for quantification (R = 50K, AGC target = 5e4, max ion injection time = 100ms, HCD collision energy = 65%, scan range = 110 – 750 m/z, synchronous precursors selected = 5).

TMT-labeled PIB and MIB fractionated samples were analyzed on an Orbitrap Fusion Lumos mass spectrometer (ThermoFisher Scientific) equipped with an Easy-nLC 1200 (ThermoFisher Scientific). Peptides were resuspended in 8% methanol / 1% formic acid across a column (45 cm length, 100 μm inner diameter, ReproSil, C₁₈ AQ 1.8 μm 120 Å pore) pulled in-house across a 2 h gradient from 3% acetonitrile/0.0625% formic acid to 37% acetonitrile/0.0625% formic acid. The Orbitrap Lumos was operated in data-dependent, SPS-MS3 quantification mode^53,54 wherein an Orbitrap MS1 scan was taken (scan range = 350 – 1250 m/z, R = 120K, AGC target = 2.5e5, max ion injection time = 50ms), followed by data-dependent Orbitrap MS2 scans on the most abundant precursors for 2 seconds. Ion selection; charge state = 2: minimum intensity 2e5, precursor selection range 650–1250 m/z; charge state 3: minimum intensity 3e5, precursor selection range 525–1250 m/z; charge state 4 and 5: minimum intensity 5e5). Quadrupole isolation = 1 m/z, R = 30K, AGC target = 5e4, max ion injection time = 55ms, CID collision energy = 35%). Orbitrap MS3 scans for quantification (R = 50K, AGC target = 5e4, max ion injection time = 100ms, HCD collision energy = 65%, scan range = 100 – 500 m/z, synchronous precursors selected = 5).

The raw data files from both Orbitrap Fusion and Orbitrap Fusion Lumos were searched using COMET with a static mass of 229.162932 on peptide N-termini and lysines and 57.02146 Da on cysteines, and a variable mass of 15.99491 Da on methionines and 79.96633 Da on serines, threonines and tyrosines against the target-decoy version of the human proteome sequence database (UniProt; downloaded 1/2020) and filtered to a <1% FDR at the peptide level. Quantification of TMT intensities was performed using in house developed software (63). Briefly, peptide-spectrum match (PSM) channel intensities were recorded as the sum of total ion intensity in +/− 5 ppm windows on each TMT reporter ion m/z value. PSM-level intensities were summed into peptide or protein level intensities.

PP2Ac mutagenesis and PP2Ac mutant purification

Human PP2Ac (β isoform) was cloned into p3XFLAG-CMV10 vector using HindIII and ECoRI restriction sites and sequence verified. Thr³⁰⁴ in PP2Ac was mutated to either the phospho-mimetic D or the phospho-null A mutant using Quick Change Lightning Kit (Agilent). All mutations were sequence verified. HEK293T cells were transfected with p3XFLAG-CMV10-PP2Ac WT, T304D or T304A and selected with G418. Cells expressing either PP2Ac WT, T304D or T304A mutants were collected and lysed in lysis buffer (50 mM Tris pH 7.5, 250 mM NaCl and 0.5% (w/v) Triton X-100). PP2Ac complexes from the lysates were purified using FLAG M2 affinity gel (SIGMA) for 2 hours at 4°C. The beads were washed three times with lysis buffer and eluted with 3XFLAG peptide (final concentration 150 ng μl⁻¹). The eluates were boiled with 2X laemelli buffer for Western blot analysis or TCA precipitated and digested with trypsin overnight for mass spectrometric analysis.

GST-PP2Ac protein purification from E.Coli

Human PP2Ac (α isoform) was into pGEX-6P using EcoRI and NotI restriction sites. After sequence verification, pGEX-6P-PP2Ac was retransformed into BL21 (DE3) pLys E.Coli for protein expression. Colonies were grown in LB medium containing 0.1 mg ml⁻¹ ampicillin and 1 mM MnCl₂ at 37°C until they reached an OD₆₀₀ of 0.6 and were then induced with 1 mM IPTG at 18°C for 16 hours. The cells were collected and snap frozen at −80°C until ready for purification. For protein purification, the cell pellets were lysed in GST lysis buffer (50 mM Tris pH 8.1, 1 M NaCl, 1 mM DTT, 0.5% Triton X-100, 10 mM MgCl₂, 1 mM MnCl₂ and 1 protease inhibitor tablet per 10 mL buffer (Mini-Complete, Roche), precleared and incubated with prewashed Glutathione-Sepharose beads overnight at 4°C. The beads were washed 3 times with GST lysis buffer and eluted with GST elution buffer (50 mM Tris-HCl pH 8.1, 1 M NaCl, 50 mM reduced glutathione, 0.5 mM DTT and 1 mM MnCl₂) at pH 8.5 and dialyzed overnight at 4°C against dialysis buffer (50 mM Tris pH 8.1, 150 mM NaCl, 1 mM DTT, and 1 mM MnCl₂).

In-vitro kinase assays with Cdk1-cyclin B complex

CDK1-cyclin B complex was purified from insect cells as described previously (64). 100 ng CDK1-cylin B complex was incubated with 500 ng of GST-PP2Ac in kinase reaction buffer (30 mM Tris pH 7.5, 100 mM NaCl, 10 mM MgCl₂, 1 mM EGTA and 10% glycerol) with or without ATP at 28°C overnight. The reactions were boiled with 2X laemelli buffer for Western blotting.

Generation of siRNA-resistant PP2Ac cell lines

Full-length PP2Ac (α isoform) was amplified with Myc-tag in the forward primer from cDNA and cloned into BamHI and NotI sites of pcDNA5/FRT/TO/IRES2EGFP. siRNA sequence s10958 (Ambion Silencer Select) was used against α isoform of PP2Ac and s10961 (Ambion Silencer Select) was used against β isoform of PP2Ac. siRNA-resistant PP2Ac lines were generated using site-directed mutagenesis with the forward primer: 5’- CAGTTACACTGCTTGTGGCCTTGAAAGTTCGTTACCGTGAACG −3’ and reverse primer 5’- CGTTCACGGTAACGAACTTTCAAGGCCACAAGCAGTGTAACTG −3’. Stable HeLa cell lines expressing these constructs under the control of doxycycline-inducible promoter were generated as described previously (65).

Live cell imaging

Live-cell analysis was performed on a Deltavision Elite system using a × 40 oil objective with a numerical aperture of 1.35 (GE Healthcare). The DeltaVision Elite microscope was equipped with a CoolSNAP HQ2 camera (Photometrics). Cells were seeded in eight-well Ibidi dishes (Ibidi). Knockdown of PP2Ac was induced by two overnight transfections of the relevant RNAi oligo in RNAiMAX (InVitrogen) 48 and 24 hours prior to analysis. Expression of the various PP2Ac subunits were induced 24 hours prior to filming with 5 or 10 ng/ml doxycyclin. Cells were synchronized in G1 phase by a double thymidine treatment protocol. Before filming, the media was changed to Leibovitz’s L-15 (Life Technologies). Appropriate channels were recorded for the times indicated.

Mitotic exit assay

HeLa cells expressing siRNA-resistant Myc-tagged PP2Ac WT, T304D or T304A mutants were depleted of endogenous PP2Ac using siRNA and arrested in mitosis with nocodazole (100 ng ml⁻¹) for 16 hours. Mitotic cells were pre-incubated with 30 μM MG132 for 30 min, followed by collection by mitotic shake-off. Cells were resuspended in PBS containing 30 μM MG132 and divided into equal aliquots for 4 time points. Flavopiridol (20 μM) was added to induce mitotic exit after which cells were collected by boiling in 2X lamelli buffer at every 5 min interval. These samples were resolved on SDS-PAGE gel for western blotting. Western blots were quantified by taking the sum total of intensity for each lane in Image lab software (BioRad) and normalized to Myc-PP2Ac WT at the 0-hour time point.

Immunofluorescence

HeLa cells expressing siRNA-resistant Myc-tagged PP2Ac T304D mutant was depleted of endogenous PP2Ac using siRNA as described above. The cells were fixed using 3 % formaldehyde, permeabilized with PBS containing 0.1 % Triton X-100 (PBST), followed by staining with antibodies to either tubulin (SIGMA) or TPX2 (a gift from Dr. Duane Compton, Dartmouth College) for 2 hours at room temperature. DNA was stained using DAPI. Images were collected using ZEISS LSM 880 confocal laser scanning microscope.

Data analysis and statistics

For a protein to be considered for further analysis, it had to be identified with total peptide (TP) count > 1 in at least one replicate. PIB-TMT experiments were normalized based on total sum tmt intensities of all PPP catalytic subunits. PIB-specific binding was established by comparison with MCLR-treated samples. For phosphorylation site analysis, only the sites on proteins considered specific in the protein PIB-TMT experiment were considered for further analysis. Phosphopeptide intensities were log₂ transformed and corrected based on the protein TMT intensities for each channel to account for protein abundance differences. P-values were calculated using a two-tailed Student’s t-test in Perseus software (66). FDR correction was performed using Benjamini-Hochberg method in Perseus software (66). For Fig. 2C, statistically significant phospho-sites (P<0.05) with a localization score of 0.95 or above were considered for analysis.

Supplementary Material

Supplementary Material

Fig. S1. PIB-SILAC strategy to determine differences in PPPome of asynchronous and mitotic cells.

Fig. S2. Quantitative TMT labeling strategy to determine global protein abundance changes.

Fig. S3. MIB kinase profiling is correlated with kinase abundances.

Fig. S4. Kinases consensus motifs.

Fig. S5. Inducible expression of WT or mutant Myc-PP2Ac.

SM table S5

Data file S5. PP2A regulatory proteins identified in FLAG eluates from cells expressing FLAG-tagged WT or Thr³⁰⁴-mutant PP2Ac.

SM table S3

Data file S3. Phosphopeptides identified on proteins that specifically bind PIBs in PIB-SILAC and PIB-TMT approaches.

SM table S4

Data file S4. Kinases profiled in asynchronous and mitotic HeLa cells using MIBs.

SM table S1

Data file S1. List of specific proteins identified in PIB pull-downs.

SM table S2

Data file S2. Abundances of proteins specifically identified in PIB pull-downs in asynchronous, G1/S or mitotically arrested cells.