PTEN regulates PLK1 and controls chromosomal stability during cell division

ABSTRACT

PTEN functions as a guardian of the genome through multiple mechanisms. We have previously established that PTEN maintains the structural integrity of chromosomes. In this report, we demonstrate a fundamental role of PTEN in controlling chromosome inheritance to prevent gross genomic alterations. Disruption of PTEN or depletion of PTEN protein phosphatase activity causes abnormal chromosome content, manifested by enlarged or polyploid nuclei. We further identify polo-like kinase 1 (PLK1) as a substrate of PTEN phosphatase. PTEN can physically associate with PLK1 and reduce PLK1 phosphorylation in a phosphatase-dependent manner. We show that PTEN deficiency leads to PLK1 phosphorylation and that a phospho-mimicking PLK1 mutant causes polyploidy, imitating functional deficiency of PTEN phosphatase. Inhibition of PLK1 activity or overexpression of a non-phosphorylatable PLK1 mutant reduces the polyploid cell population. These data reveal a new mechanism by which PTEN controls genomic stability during cell division.

KEYWORDS:

• Mitosis
• polyploidy
• PTEN; polo-like kinase 1
• phosphorylation
• chromosome stability

This article is referred to by:

Unmasking PTEN's nuclear functions

Introduction

PTEN is one of the most frequently mutated genes in human cancer, and chromosome instability (CIN) is often found in cells and tissues lacking PTEN function.^1-4 PTEN targets phosphatidylinositol-3,4,5-triphosphate (PIP3) and antagonizes the major survival pathway regulated by phosphatidylinositol 3-kinase (PI3K)/Akt.^5-7 Increasing evidence indicates that nuclear PTEN has fundamental functions in maintenance of chromosomal stability,^2,8-14 suggesting there are as yet unidentified mechanisms in the process of chromosome inheritance for which the function of PTEN is indispensable. It has been shown that PTEN plays an essential role in maintaining the structural integrity of individual chromosomes,² which supports recognition of PTEN as a guardian of the genome.^15,16 Although PTEN mutations and aneuploidy are concurrent events in tumorigenesis,^1,2,17,18 it remains elusive how PTEN controls numerical stability of chromosome transmission.

Aneuploidy is a common feature of cancer cells that is often preceded by an intermediate state of tetraploidy or polyploidy.^19-21 To explore the molecular mechanisms that bring about polyploidization, a genetic screen in yeast has identified 39 genes essential for the viability of polyploid yeast. Almost all of these genes are involved in mitosis, encoding components in mitotic spindles, sister chromatid cohesion or homologous recombination.²² Polo-like kinase 1 (PLK1) is the vertebrate homolog of Polo in Drosophila,²³ of Plo1 in fission yeast²⁴ and of Cdc5 in budding yeast.²⁵ PLK1 is an essential mitotic kinase that controls spindle bipolarity,²³ mitotic entry²⁶ and mitotic exit,²⁷ as well as cytokinesis.²⁸ Recent studies have demonstrated PLK1 functions in late mitosis in controlling cleavage furrow formation in anaphase cells.^29-31 Deregulation of PLK1 causes various forms of mitotic defects and leads to karyotypic instability, which may contribute to oncogenic transformation.³² For example, constitutively active Xenopus Plx1 induces formation of multiple nuclei³³; aberrant Cdc5 leads to premature separation of sister chromatids³⁴; and Drosophila polo mutants exhibit multinucleation due to failure of cytokinesis.³⁵ Moreover, there is significant correlation of PLK1 overexpression with tumor metastatic potential³⁶ and with prognosis in cancer patients.^37-40 PLK1 can be phosphorylated by protein kinase A, Cdc2,^41,42 and aurora A,^43,44 and phosphorylation of PLK1 at Thr210 has been identified as a major event in the activation of PLK1.^41,45

Here we report that PTEN protects the genome by preventing whole chromosome instability, which represents a novel function of PTEN in tumor suppression. In addition, the prototypical vertebrate polo kinase PLK1 emerges in this study as a protein substrate of PTEN phosphatase, and we show that PTEN is capable of physically interacting with PLK1 and dephosphorylating this critical mitotic kinase. On the basis of these findings PTEN may be characterized as a mitotic phosphatase that regulates PLK1 and prevents polyploidy. Our data emphasize the importance of the balance between mitotic kinases and phosphatases for the fidelity of chromosome inheritance during cell division.

Results

Loss of PTEN leads to spontaneous polyploidy and resistance to spindle disruption

Signature genetic alterations associated with Pten disruption include segmental chromosome aberrations such as translocations and chromosomal breakage.^2,46-48 PTEN-deficient cells also spontaneously exhibit aneuploidy,^1,2 reflecting whole chromosome instability. In order to determine whether Pten deletion can induce gross alteration of chromosome content, we compared cell cycle profiles and DNA ploidy status in Pten^+/+ and Pten^−/− mouse embryonic fibroblasts (MEFs). As shown in Fig. 1A, Pten^−/− MEFs accumulate either as a hypodiploid (<2N) or polyploid population (>4N), in association with a dramatic reduction in the cycling cell population (2N-4N) as compared with wild-type cells (62.2 ± 5.1% vs. 83.5 ± 1.1%, Fig. 1B). These data indicate that Pten depletion induces spontaneous formation of polyploid cells. Faulty spindle production from exogenous stimuli may exacerbate these spontaneous aberrations. To test this idea, we challenged Pten proficient and deficient cells with nocodazole, a microtubule-depolymerizing reagent, prior to evaluation of polyploidy status. The polyploid population in Pten null cells is further increased (51.9 ± 4.7% vs. 28.5 ± 2.2%, Fig. 1C), whereas the ploidy status remains much more stable in Pten^+/+ MEFs. The challenge of spindle disruption clearly exacerbates the already unstable karyotype in Pten null cells by further reducing the cycling population associated with aggravation of polyploidization. To determine how these polyploid cells respond to spindle perturbation in terms of ultimate cell fate, we monitored cell survival of Pten^+/+ and Pten^−/− MEFs in the presence of nocodazole. Nocodazole diminishes survival of Pten^+/+ cells to a significantly greater extent than that of Pten^−/− cells and Pten null cells appear to be more resistant to this agent (Fig. 1D). These results suggest that loss of Pten confers cellular resistance to spindle-disrupting drugs.

Figure 1. Pten disruption results in polyploidy and nocodazole resistance. (A and B) Flow cytometry analysis of ploidy distribution and DNA content in Pten^+/+ and Pten^−/− MEFs. Data are presented as mean ± SEM of 3 independent experiments (*, p < 0.05; as compared with Pten^+/+ cells with the same ploidy population). (C) Karyotypic profiles of Pten^+/+ and Pten^−/− MEFs with or without spindle disruption by nocodazole (1.5 μM). Data are summarized from 3 independent experiments and analyzed with the paired t-test. *, p < 0.05; **, p < 0.01; as compared with corresponding untreated or nocodazole treated Pten^+/+ cells. (D) Growth curves for Pten^+/+ and Pten^−/− MEFs in response to nocodazole (1.5 μM). (E) HeLa cells were transfected with pSuper/sh PTEN and PTEN expression was evaluated by Western blotting and compared with cells containing a control pSuper/sh control vector. (F) Immunofluorescent staining of microtubules using an anti-α-tubulin antibody and visualized with an FITC-conjugated species-matched secondary antibody. White arrowheads and red arrows indicate giant cells containing a single enlarged nucleus and multiple nuclei respectively. (G) A representative comparison of cell cycle and ploidy profiles in thymidine synchronized HeLa cells containing PTEN shRNA or scrambled shRNA, at a time point 12 hours after release from double-thymidine block (DTB). (H) Survival curves for PTEN-knockdown and control HeLa cells treated with 1.5 μM nocodazole for 24 h and 48 h.

To consolidate the data obtained from this mouse cell system, we knocked down PTEN in HeLa cells by PTEN shRNA (Fig. 1E). As compared with cells containing scrambled control shRNA, the effect of PTEN silencing is dramatic with generation of numbers of giant cells (Fig. 1F), including multinucleated cells and cells with a single enlarged nucleus. We used a double-thymidine block (DTB) synchronization approach to examine the cell cycle profile following PTEN knockdown. Flow cytometric analysis revealed a consistent increase of the polyploid population in cells containing PTEN shRNA at each time point during the 0∼16-hour monitoring process following DTB release. The most marked difference of polyploidy frequency between PTEN knockdown cells and control cells occurs following mitosis (12∼16 hours after DTB release). At the 12-hour time point, there is a 2.4-fold increase in polyploid DNA content in PTEN knockdown cells as compared with control cells (Fig. 1G). These results suggest that the aberrant chromosome complement associated with PTEN knockdown may result from chromosomal alterations generated during cell division. In evaluation of cell sensitivity to spindle pertubation, we found that PTEN knockdown cells are more resistant to nocodazole treatment (Fig. 1H). These data collectively demonstrate that PTEN deficiency results in cellular resistance to spindle damage, which may result from defective mitosis and polyploidization.

In order to determine whether PTEN deficiency causes mitotic defects, we employed time-lapse microscopy to monitor the dynamics of chromosome segregation and compare live cell progression through mitosis using cells expressing a GFP-labeled chromatin marker, histone H2B-GFP. Failure of chromosome segregation and/or cytokinesis is a prominent phenomenon in cells lacking PTEN. For example, cells containing PTEN shRNA may arrest in metaphase for hours without initiating anaphase, and chromosomes align normally on the metaphase plate but fail to segregate (Fig. S1, middle panel). PTEN knockdown cells often exhibit chromosome non-disjunction or cleavage furrow regression, leading to failure of cytokinesis (Fig. S1, lower panel). Therefore, deficiency of PTEN causes erroneous chromosome segregation or incomplete cytokinesis, leading to polyploidization.

Polo-like kinase 1 is a PTEN-associated protein during mitosis

The mitotic and karyotypic abnormalities following PTEN depletion suggest that PTEN may be physically involved in mitotic control. In order to identify potential cellular factors in the PTEN mitotic pathway, we used a biochemical purification approach to isolate proteins that specifically interact with PTEN. A stable HeLa cell line expressing FLAG-HA double-tagged PTEN (Fig. 2A) was employed for tandem affinity purification of potential PTEN-associated proteins from mitosis-enriched cells. Mass spectrometric analysis of a ∼70 k_D band identified polo-like kinase 1 (PLK1) as a potential PTEN-interacting protein (Figs. 2B and S2).

Figure 2. PLK1 is a PTEN-associated mitotic protein. (A) Expression of FH-PTEN and endogenous PTEN in FH-PTEN-expressing HeLa cells. Western blot analysis of HeLa cells containing FH-PTEN or a control vector with an anti-PTEN (upper panel) or anti-FLAG (lower panel) antibody. (B) Identification of Plk1 as a component of PTEN-associated protein complexes by pull-down assay. PTEN-complexes from HeLa/FLAG-HA-PTEN cells and a control elution from HeLa cells containing an empty vector were separated on SDS-PAGE following tandem affinity purification. A ∼70-kDa band was analyzed by mass-spectrometry and revealed PLK1 as a potential PTEN-associated protein. (C) Direct interaction of PTEN with PLK1 in vitro. Sf9 cells were infected with indicated recombinant baculovirus for 72 hours and subsequently harvested for protein purification using either an Ni-NTA agarose column or an anti-FLAG M2 affinity gel. Purified His-tagged proteins were incubated with purified FLAG-PLK1 protein followed by precipitation with Ni-NTA beads. Samples were then subjected to analysis of FLAG-PLK1 expression (upper panel). His-tagged PTEN and control protein in these reactions were shown in the lower panel. (D) Physical association of endogenous PTEN and PLK1. HeLa cells as well as MEFs with or without Pten deletion were subjected to nocodazole treatment (50 nM, 8 h) followed by immunoprecipitation with a rabbit anti-PTEN antibody and subsequent evaluation for PLK1 and PTEN by Western blotting. An isotype-matched mouse IgG was used as a control for immunoprecipitation and ExactaCruz reagents were used to reduce the interference of IgG bands. (E) Cell cycle-dependent PTEN-PLK1 association during cell division. HeLa cells were synchronized by DTB and released after different times as indicated. The interaction between PTEN and PLK1 was examined by PTEN immunoprecipitation for detection of PTEN-associated PLK1 by Western analysis. Cyclin B expression in synchronized pre-IP samples is shown to indicate mitotic entry and exit.

In order to validate the pull-down results, we performed an in vitro binding assay using purified recombinant FLAG-tagged PLK1 (FLAG-PLK1) and His-tagged PTEN (His-PTEN). As expected, FLAG-PLK1 associates with His-PTEN but not with an irrelevant protein His-control (Fig. 2C), indicating that PTEN can directly interact with PLK1. We next examined the in vivo interaction between PTEN and PLK1. As shown in Fig. 2D, endogenous PLK1 in HeLa cells is detectable in PTEN immunoprecipitates, but not in immunoprecipitates of a species-matched non-specific IgG (lane 2 vs. lane 3). Similarly, physical association of Pten with Plk1 can be found in Pten^+/+, but not in Pten^−/− MEFs (Fig. 2D, lane 5 vs. lane 8). These data suggest that PTEN is physically associated with PLK1.

To further delineate the relationship between PTEN and PLK1, we examined their interaction in the cell cycle particularly during mitosis. DTB synchronization and co-immunoprecipitation experiments reveal that PTEN primarily interacts with PLK1 during mitosis (Fig. 2E). This interaction peaks when cells enter mitosis (8 h following DTB) and subsequently decreases as cells exit mitosis. These results depict a dynamic interaction between PTEN and PLK1, which implies they have a functional interplay in mitosis. Therefore, it was next imperative to determine the mechanism by which PTEN interacts with PLK1 in the control of chromosome inheritance during cell division.

PTEN deficiency results in PLK1 hyperphosphorylation

PLK1 is activated by phosphorylation on residue Thr210 within its activation loop.⁴⁹ In order to delineate a signaling link between PTEN and PLK1, we first examined the phosphorylation status of Plk1 in Pten-deficient cells, and found that the level of Plk1 phosphorylation at Thr210 is markedly increased in Pten^−/− MEFs as compared to Pten^+/+ MEFs in both the presence and absence of nocodazole treatment (Fig. 3A). We also observed an elevation of the total level of Plk1 expression in Pten null cells, which may result from impairment of APC-Cdh1-mediated Plk1 degradation.⁸ Nevertheless, the expression change is less significant as compared with that of phosphorylation levels, suggesting that Pten may directly target PLK1 to regulate its phosphorylation.

Figure 3. PLK1 is a protein substrate of PTEN phosphatase. (A and B) Pten depletion induces Plk1 phosphorylation. Pten^+/+ and Pten^−/− MEFs cultured with or without nocodazole (50 nM, 8 h, A), as well as various tissues from mice with wild-type Pten or heterozygous Pten deletion (B), were analyzed for Plk1 phosphorylation and abundance by immunoblotting with site-specific phospho-Plk1 (Thr210) and total Plk1 antibodies. β-Actin was used as a loading control. (C) PTEN directly and dose-dependently dephosphorylates PLK1. Sf9-expressed His-PTEN protein was purified using a Ni-NTA agarose column. FLAG-PLK1 expressed in Sf9 cells was immunoprecipitated with anti-FLAG M2 beads and incubated with increasing amounts of His-PTEN proteins, followed by Western blot analysis of PLK1 phosphorylation. The same blot was probed with PLK1 antibody to show PLK1 levels loaded in each lane. (D) Ectopic PTEN suppresses PLK1 kinase activity. Phosphorylation of NudC, a substrate of PLK1 kinase, was analyzed by immunoblotting in Pten^−/− MEFs containing ectopic PTEN in the presence and absence of PLK1^T210D, an enzymatically active form of PLK1.

In mice, decreasing Pten dosage correlates with increasing tumor susceptibility,^50,51 suggesting sufficient levels of Pten are required for suppressing tumorigenesis. In order to determine whether Pten haploinsufficiency also affects PLK1 phosphorylation, we examined the status of PLK1 phosphorylation in various tissues from mice with heterozygous Pten deletion. As expected, PLK1 is phosphorylated in various Pten^+/− tissues including lung, bone marrow and thymus, as compared with wild-type tissues from littermate control mice (Fig. 3B). These results suggested that disruption of Pten is associated with spontaneous hyperphosphorylation of PLK1.

PTEN directly dephosphorylates PLK1 and reduces its kinase activity

Because PTEN regulates cellular processes through its dual phosphatase activity,^52,53 PTEN may employ its protein phosphatase function to dephosphorylate and inactivate PLK1. To determine whether PTEN can directly dephosphorylate PLK1 in vitro, we employed a phosphatase assay to evaluate PLK1 phosphorylation following incubating purified PLK1 protein with increasing amounts of His-tagged PTEN. As shown in Fig. 3C, PLK1 phosphorylation is reduced by PTEN in a dose-dependent manner. In contrast, the total level of PLK1 remains relatively stable although a slight reduction is noticeable in the presence of very high concentrations of PTEN. These data suggest that PLK1 can serve as a direct substrate of PTEN phosphatase and that exhaustive dephosphorylation of PLK1 by PTEN may result in a reduction of the PLK1 total expression level.

Since PLK1 phosphorylation at the T210 residue is required for its kinase activity, one would expect that PTEN regulation of PLK1 phosphorylation might affect its downstream targets. Nuclear distribution gene C (NudC) is a substrate of PLK1 kinase that mediates PLK1 function in cytokinesis,⁵⁴ which may also serve as a target of the PTEN-PLK1 signaling axis in preventing cytokinesis failure and polyploidy. We therefore chose NudC to assess how NudC phosphorylation is influenced by the interaction between PTEN and PLK1. To this end, we constructed a phospho-mimetic PLK1 plasmid, PLK1^T210D, to imitate enzymatically active PLK1.⁴⁹ NudC phosphorylation was then examined in Pten null MEFs transfected with PTEN and PLK1^T210D individually and in combination. Pten null cells express a high basal level of NudC phosphorylation, likely due to PLK1 activation. Additional PLK1 activation by ectopic PLK1^T210D further elevates NudC phosphoryolation only slightly. Nevertheless, NudC phosphorylation is reduced by ectopic PTEN, both alone and in combination with PLK1^T210D (Fig. 3D). These results suggest that PTEN can antagonize the kinase activity of both endogenous and enzymatically active PLK1, likely through a dephosphorylation mechanism.

PTEN inhibits PLK1 in a phosphatase-dependent manner

To further validate the role of PTEN phosphatase activity in suppression of PLK1 phosphorylation and in protection against polyploidization, we examined the level of Plk1 phosphorylation in Pten null cells with ectopic expression of wild-type PTEN or 2 phosphatase-deficient PTEN mutants. PTEN is well known to antagonize the PI3K-mediated lipid pathway.⁵ In addition, PTEN can also act as a protein phosphatase to dephosphorylate protein substrates.^55,56 To distinguish its activities in suppressing the lipid or the protein pathways, 2 PTEN mutants have been commonly used. PTEN^C124S lacks both lipid and protein phosphatase activity,⁵⁷ whereas PTEN^G129E is lipid phosphatase deficient but retains protein phosphatase activity.⁵⁸ Plk1 phosphorylation is markedly elevated in Pten null cells, as compared with Pten^+/+ MEFs (Fig. 4A). Introduction of wild-type PTEN significantly reduces Plk1 phosphorylation, as compared with Pten^−/− cells transfected with a control vector. Reduction of Plk1 phosphorylation mirrors the change in Akt phosphorylation (Fig. 4B), suggesting that Plk1 serves as a target for Pten in the nucleus in a manner similar to that in which Akt acts as a signaling target of Pten in the cytoplasm. However, phosphorylation of Plk1 at Thr210 remains at a high level in cells transfected with the PTEN mutant PTEN^C124S, which is deficient in both protein and lipid phosphatase activity (Fig. 4A, lane 5). In contrast, the lipid phosphatase-deficient PTEN mutant PTEN^G129E still suppresses Plk1 phosphorylation to an extent comparable to wild-type PTEN (Fig. 4A, lane 6). It is of particular interest to compare the phosphorylation changes in Plk1 (Fig. 4A) with those of Akt (Fig. 4B) in response to ectopic expression of different mutant forms of PTEN. Although both PTEN^C124S and PTEN^G129E lack ability to inhibit Akt phosphorylation (Fig. 4B), the lipid phosphatase-deficient form of PTEN, PTEN^G129E, retains its ability to suppress Plk1 phosphorylation (Fig. 4A). These data indicate that suppression of Plk1 phosphorylation by PTEN requires its protein phosphatase activity.

Figure 4. PTEN reduces PLK1 phosphorylation, PLK1 stability and polyploidy in a phosphatase-dependent manner. (A) Pten^−/− MEFs were transfected with vectors encoding FLAG-tagged wild-type PTEN or the phosphatase-deficient PTEN mutants PTEN^C124S and PTEN^G129E, and subjected to immunoprecipitation with anti-Plk1 antibody followed by evaluation for phospho-PLK1 (Thr210). The expression of PLK1 and β-actin was then evaluated by re-blotting with corresponding antibodies. (B) Cell lysates identical to those in (A) prior to immunoprecipitation were processed for immunoblot analysis of phospho-Akt (Ser473). Endogenous and ectopic FLAG-PTEN expression was evaluated using anti-PTEN and anti-FLAG monoclonal antibodies respectively. Equal protein loading was demonstrated by re-probing the same blot with anti-β-actin antibody. (C) Pten null MEFs were co-transfected with FLAG-tagged PLK1 and different forms of PTEN (wild-type or phosphatase-deficient mutants) prior to analysis of FLAG-PLK1 expression. (D) Pten^+/+ MEFs as well as Pten^−/− MEFs containing wild-type PTEN, or one of the 2 phosphatase-deficient mutants (PTEN^C124S and PTEN^G129E) were cultured for 48 h followed by flow cytometric analyis of ploidy status and analyzed with the paired t-test for statistical significance. *, p < 0.05; **, p < 0.01. (E) Pten^−/− MEFs containing wild-type PTEN, PTEN^C124S or PTEN^G129E were treated with 1.5 μM nocodazole for 24 h and subjected to survival analysis by trypan blue exclusion. *, p < 0.05.

Phosphorylation is associated with protein stability during cell division.⁵⁹ In agreement with this notion, we often observed a concomitant change of PLK1 expression in Pten-deficient cells that corresponds to its phosphorylation state (see Figs. 3A, 3C and 4A). We therefore hypothesized that dephosphorylation of PLK1 by PTEN may render a reduction of PLK1 stability. To test this hypothesis, we co-transfected Pten null cells with FLAG-tagged PLK1 and different forms of PTEN with or without phosphatase activity prior to analysis of FLAG-PLK1 expression. As shown in Fig. 4C, the expression of FLAG-PLK1 is significantly diminished by wild-type PTEN but not by PTEN^C124S, the mutant lacking both protein and lipid phosphatase activity. The PTEN^G129E mutant lacking lipid phosphatase activity, however, retains the capability of reducing FLAG-PLK1 expression (Fig. 4C). These results support that PTEN reduces the expression level of ectopic FLAG-PLK1 in a manner similar to PTEN suppression of PLK1 phosphorylation (Fig. 4A), suggesting that PLK1 phosphorylation plays an important role in maintaining its protein stability. These data highlight a phosphatase-dependent regulation of PLK1 by PTEN leading to a successive inhibition of PLK1 phosphorylation and stability.

Phosphatase function of PTEN is essential for prevention of polyploidization and restoration of sensitivity to spindle drugs

In order to determine whether differences in PTEN phosphatase activity are reflected in phenotypic changes, we investigated ploidy status in Pten^−/− cells ectopically expressing PTEN with differing phosphatase activity (wild-type PTEN, PTEN^C124S or PTEN^G129E). As shown in Fig. 4D, ectopic expression of wild-type PTEN as well as the lipid-phosphatase-deficient PTEN mutant (PTEN^G129E) results in a dramatic reduction of the polyploid population. However, the PTEN^C124S mutant which lacks both lipid and protein phosphatase activity fails to cause such a reduction.

PTEN deficiency not only induces ploidy aberration but also conveys resistance to spindle disruption (Figs. 1D and 1H). If polyploidy were functionally linked to cellular resistance to spindle disruption, one would expect that amelioration of karyotypic instability would restore cellular sensitivity to spindle poison. As expected, a significantly greater cell death response to nocodazole is induced by ectopic expression of wild-type PTEN as well as by PTEN^G129E, but not by PTEN^C124S (Fig. 4E). Therefore, PTEN's function in sensitizing cells to spindle disruption is dependent upon its protein phosphatase activity, in a manner parallel to PTEN's effect on ploidy status. These results indicate that reconstitution of PTEN expression in Pten null cells can restore the ability of PTEN phosphatase to suppress Plk1 phosphorylation and in consequence, reduce polyploidy and restore cellular sensitivity to spindle disruption. The phosphatase activity of PTEN is therefore necessary not only for counterbalancing PLK1 phosphorylation and stability but also for maintaining normal chromosomal content and cellular responsiveness to spindle damage.

Constitutive activation of PLK1 mimics PTEN deficiency to induce polyploidy whereas inhibition of Plk1 ameliorates karyotypic instability

In order to investigate the functional interplay between PTEN and PLK1, we employed 2 strategies to determine how PLK1 phosphorylation contributes to an aberrant chromosome complement. First, we sought to determine whether abnormal PLK1 activation mimics PTEN deficiency and how the phospho-mimicking mutant PLK1^T210D affects ploidy status and cellular sensitivity to spindle poison. Ectopic expression of constitutively activated PLK1 results in significantly increased numbers of multinucleated cells (Figs. 5A and 5B), indicating that PLK1^T210D induces polyploidization. Moreover, the phospho-mimicking PLK1^T210D endows cells with resistance to spindle disruption by nocodazole (Fig. 5C). All of these phenotypic features induced by PLK1^T210D mirror our observations in PTEN-deficient cells (Fig. 1), suggesting that PLK1 hyperphosphorylation may mediate generation of polyploidy following PTEN depletion. To consolidate these observations, we next investigated whether chemical inhibition of PLK1 activity reduces polyploidy in Pten-deficient cells. We first utilized BI 2536, a specific Plk1 inhibitor ^60,61, to block Plk1 activity in Pten null cells. As shown in Fig. 5D, the polyploid population in Pten null cells is increased by nocodazole, and this effect can be blocked by pretreatment with BI 2536. Moreover, overexpression of a non-phosphorylatable PLK1^T210A mutant in Pten null cells can ameliorate the abnormal karyotype, as demonstrated by an increase in the cycling diploid cell population and reduction in numbers of polyploid cells in both the presence and absence of nocodazole treatment (Fig. 5E). These data indicate that inhibition of PLK1 can correct polyploidy in Pten null cells and restore normal karyotypic characteristics. Our data thus suggest that deregulation of PLK1 activity is responsible for polyploidization in Pten-deficient cells. Therefore, the functional interplay between a potent tumor suppressor phosphatase (PTEN) and a key mitotic kinase (PLK1) may serve to prevent polyploidization and maintain chromosomal stability.

Figure 5. Constitutively active PLK1 induces multinucleation whereas inhibition of PLK1 mitigates polyploidy. (A) HeLa cells were co-transfected with GFP-labeled histone H2B as well as PLK1^T210D or a control expression vector, followed by immunofluorescent microscopic analysis of cell morphology. Arrows indicate multinucleated cells. (B) HeLa cells with or without PLK1^T210D (n ≥ 400) were analyzed in randomly selected regions and the percentage of multinucleated cells was plotted on the histogram. **, p < 0.01. (C) Cells with or without PLK1^T210D were treated with different doses of nocodazole and cell survival was examined by trypan blue exclusion at various times during a 3-day course. *, p < 0.05; **, p < 0.01; as compared with control cells analyzed on the same day. (D) Pten^−/− MEFs were pre-treated with BI 2536 (100 nM, 30 min) prior to nocodazole treatment (1.5 μM) for 48 h. Cells were evaluated for polyploidization by flow cytometry. *, p < 0.05. (E) Pten^−/− MEFs transiently transfected with PLK1^T210A or an empty plasmid were used as controls and were treated with nocodazole (1.5 μM) for 48 h followed by ploidy analysis. *, p < 0.05; **, p < 0.01.

Discussion

In this study, we demonstrate that PTEN protects the genome against polyploidization through regulation of the mitotic kinase PLK1. Our data highlight the phosphatase function of PTEN that is required for maintaining the optimal phosphorylation and stability of PLK1 to prevent its overactivation. The fact that constitutively active PLK1 induces polyploidy whereas inhibition of PLK1 phosphorylation reduces the frequency of polyploidization suggests that aberrant PLK1 phosphorylation mediates polyploidization as a result of PTEN deficiency. These results corroborate that PTEN is essential for the maintenance of whole chromosome stability by tuning the balance of mitotic phosphorylation. Our findings thus reveal a mechanism for cell division that is based on the interplay of PTEN phosphatase and PLK1 kinase, and this mechanism is essential for faithful genetic transmission.

PLK1 is regulated in a meticulous manner to ensure optimal levels of expression and activity over various stages of the cell cycle. Although PLK1 activation by upstream kinases has been well studied,^43,44 how PLK1 is dephosphorylated remains obscure. Our study demonstrates that PTEN phosphatase directly interacts with PLK1 and inhibits its activity by dephosphorylation. Identification of PLK1 as a molecular target of the phosphatase PTEN advances current knowledge of PTEN function beyond its role as an antagonist of the PI3K/Akt pathway. The functional interplay between PTEN and PLK1 thus forms a control mechanism for maintaining genomic stability during cell division.

PLK1 controls spindle bipolarity ²³ and precise progression of mitosis from entry ²⁶ to exit.²⁷ By regulating such a versatile mitotic kinase, PTEN may also play multiple roles in various aspects of mitotic control. In addition, PTEN may regulate PLK1 through different mechanisms. It has been reported that PTEN regulates the APC-CDH1 complex to reduce PLK1 protein stability.⁸ Moreover, elevated expression of PLK1 in PTEN-depleted cancer cells can confer the tumorigenic competence.⁶² Interestingly, PTEN function in APC-CDH1-mediated regulation of PLK1 protein levels appears to be phosphatase-independent and involved in cellular senescence.⁸ Therefore, PTEN may utilize different mechanisms to regulate PLK1 in multiple cellular processes. Distinct mechanisms point to the same direction, e.g. PTEN suppresses PLK1 to ensure optimal expression levels and kinase activity of PLK1 during the cell cycle. PTEN phosphatase can directly inhibit PLK1 phosphorylation to reduce its stability whereas it can also influence PLK1 indirectly by targeting the APC-CDH1 complex. More interestingly, PTEN and PLK1 can mutually regulate each other as PLK1 can phosphorylate PTEN to regulate glucose metabolism⁶³ as well as mitotic timing.⁶⁴ The bidirectional interplay between PTEN and PLK1 may represent a signaling axis that plays a critical role in modulating cancer evolution. As the phospho-mimicking PLK1 mutant imitates PTEN deficiency and causes polyploidy, it is therefore intriguing to investigate whether PLK1 hyperphosphorylation may lead to tumorigenesis in future studies.

PTEN phosphatase may have different protein substrates necessary for its mitotic function. PLK1 may be only one among many nuclear PTEN targets that act to regulate the complicated process of cell division. For example, we have recently found that PTEN uses its protein phosphatase activity to regulate a critical mitotic motor protein (unpublished data). This in turn suggests PTEN acts as a calibrator of protein phosphorylation to maintain optimal activity of multiple mitotic proteins for normal well-orchestrated mitosis. It is of interest that the major cytoplasmic substrate of the PTEN lipid phosphatase PIP3 has recently been found in the midbody acting to control cytokinesis,⁶⁵ which argues PTEN is involved in the same process. As such, future efforts may uncover additional PTEN targets that together constitute a surveillance mechanism for ensuring mitotic fidelity.

Spindle perturbation is currently in use in clinical oncology as an anti-mitotic strategy which aims at activation of the spindle assembly checkpoint to cause mitotic arrest and subsequent cell death.⁶⁶ However, this strategy has limits as cancer cells may escape this type of cell death.⁶⁷ Our data similarly suggest that cells lacking PTEN may acquire a mechanism defective in the control of chromosome segregation resulting in polyploidy, which confers cellular resistance to anti-mitotic spindle drugs. Establishment of this novel link between functional deficiency of PTEN and resistance to spindle drugs provides a theoretical framework from which novel therapeutic approaches may be devised. For example, there is potential for restoration of cellular sensitivity to spindle drugs by inhibition of PLK1, and such a strategy could enhance chemotherapeutic efficiency by overcoming drug resistance caused by PTEN deficiency. Knowledge gained from this study will provide a new perspective for further mechanistic studies to identify additional intracellular targets for PTEN itself and PTEN-related drug development.

Materials and methods

Cells, tissues, antibodies, chemicals and plasmids

Pten^+/+ and Pten^−/− mouse embryo fibroblasts (MEFs) of less than 10 passages were used for phenotypic analyses. HeLa cells were from the American Type Culture Collection. Sf9 insect cells were from EMD Biosciences. Mammalian cells were cultured in MEM and Sf9 cells were cultured in Grace's insect medium, both supplemented with 10% fetal bovine serum (FBS). Monoclonal anti-PTEN antibodies were purchased from Santa Cruz biotechnology and Cascade Bioscience. A polyclonal PTEN antibody was generated by immunizing rabbits with a purified full-length PTEN protein. Anti-PLK1 and anti-phospho-PLK1 (Thr210) antibodies were from BioLegend. Nocodazole and BI 2536 were purchased from Sigma and AxonMedchem respectively.

shRNA and plasmids

The pSUPER RNAi system (Oligoengine Inc.) was used to construct PTEN-specific shRNA expression plasmids and scrambled control shRNA. The expression vectors containing FLAG or FLAG-HA tags were gifts from W. Gu.⁶⁸ The PLK1 expression plasmid was created by ligating the full-length coding region of human PLK1 into the expression vector with an N-terminal FLAG tag. Two PLK1 mutants, PLK1T^210D and PLK1^T210A, were generated by changing codon Thr210 from ACC to GAC or GCC respectively by using the QuikChange site-directed mutagenesis kit (Strategene) according to the manufacturer's protocol. Wild-type PTEN and phosphatase-deficient PTEN mutants, PTEN^C124S and PTEN^G129E, have been described previously.² A pFastBac1 expression vector (Invitrogen) was used to construct plasmids expressing PTEN and control proteins with His-tag.

Cell growth, cell cycle profile and ploidy analysis

Cells were plated so that 50% confluence was reached on the second day followed by continuing culture or nocodazole treatment with or without 30-min pretreatment with BI 2536 for an additional 48 h. Cells were then harvested as single cell suspension. Cell growth was analyzed with trypan blue exclusion. For cell cycle and ploidy profiles, cells were fixed with ice-cold 75% ethanol prior to staining with 4 μg/ml propidium iodide in PBS containing 0.1 mg/ml RNase A. DNA content of cells was analyzed using a Becton Dickinson FACScan flow cytometer, and the cell cycle distribution and ploidy status were determined using CellQuest Pro software. Three independent experiments were analyzed with the paired t-test and p < 0.05 was considered statistically significant.

Immunofluorescence, confocal microscopy and live cell imaging

Cells were fixed for immunofluorescence staining, confocal microscopy and live cell imaging. Exponentially growing cells were fixed with 4% formaldehyde in PHEM (60 mM Pipes, 25 mM Hepes, 10 mM EGTA, and 2 mM MgCl₂, pH 6.9) and permeablized for immunostaining. Cell morphology was examined by staining microtubules and centrosomes with a monoclonal anti-α-tubulin (Sigma) antibody. Dynamics of cell cycle progression through mitosis were analyzed by time-lapse imaging using a Leica TCS SP5 confocal microscope. Image acquisition was performed with a 63× objective enclosed in a humidified incubation chamber with 5% CO₂ at 37°C. Images were collected every 1.5–2.5 min for 5–20 hours. Images were analyzed by 3-dimensional maximum projection with LAS AF software.

FLAG-HA pull-down and mass spectrometry

To purify PTEN-associated mitotic proteins for mass spectrometric analysis, HeLa cells stably expressing FLAG-HA-tagged PTEN and control cells containing an empty FLAG-HA vector were treated with nocodazole (100 nM) for 8 hours. Nuclear fractions of cell extracts were subjected to sequential affinity purification using anti-FLAG M2 and anti-HA agarose beads and eluted with excessive FLAG and HA peptides. The final eluted materials were resolved by SDS-PAGE and a ∼70-kDa band in PTEN complexes was excised for mass spectrometric analysis.

In vitro protein-protein interaction

Purified His-PTEN and FLAG-PLK1 were incubated in vitro for analysis of their direct interaction. Recombinant proteins FLAG-PLK1, His-PTEN and a His-tagged irrelevant protein, His-Control, were expressed in asynchronous Sf9 cells for 72–96 hours and purified with anti-FLAG M2 beads (Sigma) or Ni-NTA beads (Qiagen). Purified FLAG-PLK1 (0.5 μg) was incubated with His-PTEN (0.5 μg) or His-Control (0.5 μg) for 30 min at room temperature. Protein complexes were then immunoprecipitated with Ni-NTA beads and subjected to SDS-PAGE for detection of FLAG-PLK1 using anti-FLAG antibody.

Dephosphorylation assay in vitro and in vivo

FLAG-PLK1 protein was expressed in Sf9 cells and isolated by immunoprecipitation using anti-FLAG M2 beads. Sf9-expressed His-PTEN protein, as well as 2 mutant forms of PTEN, C124S and G129E, were purified and added to the FLAG immunoprecipitates and incubated at room temperature for 30–60 min. Samples were washed and subjected to Western blot analysis of PLK1 phosphorylation. For in vivo dephosphorylation assay, Pten^−/− MEFs were transfected with FLAG-PTEN, FLAG-PTEN^C124S, FLAG-PTEN^G129E or an empty control vector. The level of PLK1 phosphorylation was evaluated by immnoblotting with a site-specific anti-phospho-PLK1 (Thr201) antibody.

Abbreviations

APC	=	anaphase-promoting complex
CIN	=	chromosome instability
DTB	=	double-thymidine block
NudC	=	Nuclear distribution gene C
PI3K	=	phosphatidylinositol 3-kinase
PIP3	=	phosphatidylinositol-3,4,5-triphosphate
PLK1	=	polo-like kinase 1
PTEN	=	phosphatase and tensin homolog
shRNA	=	short hairpin RNA

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

This work was supported by NIH grants R01 CA133008 to Y.Y. and R01 GM100478 to W.H.S.