Functional relationship among PLK2, PLK4 and ROCK2 to induce centrosome amplification

Abstract

The presence of more than 2 centrosomes (centrosome amplification) leads to defective mitosis and chromosome segregation errors, is frequently found in a variety of cancer types, and believed to be the major cause of chromosome instability. One mechanism for generation of amplified centrosomes is over-duplication of centrosomes in a single cell cycle, which is expected to occur when cells are temporarily arrested. There are a growing number of kinases that are critical for induction and promotion of centrosome amplification in the cell cycle-arrested cells, including Rho-associated kinase (ROCK2), Polo-like kinase 2 (PLK2) and PLK4. Here, we tested whether these kinases induce centrosome amplification in a linear pathway or parallel pathways. We first confirmed that ROCK2, PLK2 and PLK4 are all essential for centrosomes to re-duplicate in the cells arrested by exposure to DNA synthesis inhibitor. Using the centrosome amplification rescue assay, we found that PLK2 indirectly activates ROCK2 via phosphorylating nucleophosmin (NPM), and PLK4 functions downstream of ROCK2 to drive centrosome amplification in the arrested cells.

Abbreviations

Aph

aphidicolin

CDK2

Cyclin-dependent kinase 2

DN

dominant-negative

KD

kinase-dead

MEFs

mouse embryonic fibroblasts

NPM

Nucleophosmin

PLK2

Polo-like kinase 2

PLK4

Polo-like kinase 4

Puro

puromycin

ROCK2

Rho-(associated) kinase

wt

wild type

Introduction

The primary function of the centrosome is to organize microtubules. Especially, in mitosis, 2 centrosomes form spindle poles and direct the assembly of bipolar spindles, which is critical for equal distribution of chromosomes to daughter cells. Upon cytokinesis, each daughter cell inherits only one centrosome from a mother cell, and thus the centrosome must duplicate once in each cell cycle.^1,2 Centrosome duplication proceeds in coordination with other cell cycle events, and is a highly regulated process. If cells are temporarily arrested due to physiological stress, and the regulatory mechanism that limits to one round of centrosome duplication in a single cell cycle is disrupted, centrosomes continue to duplicate (centrosome re-duplication), leading to generation of ≥3 centrosomes (centrosome amplification).³ Centrosome amplification is a causal factor of aberrant mitosis and chromosome segregation errors, is commonly found in various types of cancers, and considered as the major causes of chromosome instability in cancer cells.^4,5

The centrosome consists of a pair of centrioles (duplication units of the centrosome) and surrounding amorphous pericentriolar material consisting of a number of different proteins. Duplication of the centrosome starts at the time of S-phase entry by physical separation of the paired centrioles, followed by formation of procentrioles close to each pre-existing centriole. Cyclin-dependent kinase 2 (CDK2)-cyclin E complex, a key initiator of DNA replication^6,7 was found to also govern the initiation of centrosome duplication.^8-11 Nucleophosmin (NPM) was identified as one of the targets of CDK2-cyclin E for initiation of centrosome duplication.¹² It has later been found that NPM is phosphorylated on Thr199 by CDK2-cyclin E,¹³ which increases the binding affinity of NPM to Rho-associated kinase (ROCK2).¹⁴ Acquisition of a high binding affinity to ROCK2 allows NPM to bind to centrosomally localized ROCK2.¹⁴ ROCK2 is one of the essential kinases that control initiation of centrosome duplication in the cycling cells and induce centrosome amplification in the cell cycle-arrested cells.^14,15 ROCK2 has a C-terminal auto-inhibitory domain, which folds back to interact with the N-terminal kinase domain, resulting in inhibition of the kinase activity. Although Rho small GTPase binds to ROCK2 to free the kinase domain from the auto-inhibitory domain, the Rho-binding appears only to prime for activation, because its binding results in ∼1.5-fold increase in the ROCK2 kinase activity.^16–18 Thus, further up-regulation of the kinase activity is critical for ROCK2 to function in certain biological events, including initiation of centrosome duplication. The NPM-binding was found to dramatically increases the kinase activity of ROCK2 (5 to10-fold), and this NPM-dependent “super-activation” of ROCK2 is essential for initiation of centrosome duplication in cycling cells and centrosome re-duplication in arrested cells.¹⁴

Other kinases were also found to be essential for centrosome duplication and re-duplication. For instance, depletion of Polo-like kinase 2 (PLK2) results in failure to undergo centrosome duplication in cycling cells, and centrosome amplification in arrested cells.¹⁹ Moreover, PLK2, when overexpressed, promotes centrosome duplication and re-duplication in the kinase activity- and centrosome localization-dependent manners.²⁰ NPM was identified as one of the targets of PLK2 to drive centrosome duplication; PLK2 phosphorylates NPM on Ser4, which is essential for centrosomes to re-duplicate in the arrested cells.²¹ It was further found that PLK2 drives centrosome re-duplication in the PLK4-dependent manner.²⁰ PLK4 is another member of the Polo-like kinase family, and has been found as an essential kinase for centrosome duplication and re-duplication.^22-24

In this study, using the centrosome amplification rescue assay, we examined whether ROCK2, PLK2, and PLK4 induce centrosome re-duplication (amplification) in cells arrested by exposure to DNA synthesis inhibitors in a linear pathway or parallel pathways. We found that PLK2 functions in the upstream of ROCK2 via phosphorylating NPM, and ROCK2 functions in the upstream of PLK4 to drive centrosome amplification.

Results and Discussion

Silencing of PLK2, PLK4, ROCK2 all inhibits centrosome amplification in the cells arrested by exposure to a DNA synthesis inhibitor

Recent studies have shown the critical roles of 3 kinases, PLK2, PLK4 and ROCK2, to control centrosome duplication in cycling cells and promote centrosome amplification in arrested cells.^14,19-24 Silencing of each kinase alone either blocks or delays initiation of centrosome duplication in cycling cells, and inhibits centrosome re-duplication in arrested cells. We decided to examine the potential functional cross-talk among these 3 kinases to induce centrosome amplification in cells arrested by exposure to aphidicolin (Aph), a DNA synthesis inhibitor. When cells are exposed to DNA synthesis inhibitors, cells are arrested in centrosome duplication permissive state (i.e., S and G2 phases), and centrosomes continue to duplicate, resulting in generation of amplified centrosomes. However, if cells retain wild-type (wt) p53, p53 is up-regulated in response to the genotoxic stress associated with the inhibitors as well as the stress associated with the cell cycle arrest itself. The upregulated p53 then transactivates p21 CDK inhibitor, which in turn inhibits CDK2-cyclin E (or cyclin A), a critical kinase complex for induction of centrosome re-duplication. Thus, cells with wt p53 fail to undergo centrosome re-duplication in the arrested cells, and inversely cells lacking functional p53 undergo efficient amplification of centrosomes.³ Because our experiments required the cells that undergo centrosome re-duplication at a high frequency when arrested by exposure to Aph, we decided to use p53-null primary mouse embryonic fibroblasts (MEFs).

We first tested the effect of silencing of PLK2, PLK4 and ROCK2 on centrosome re-duplication in the Aph-arrested p53-null MEFs. The cells were transfected with the siRNA sequence targeting each of these kinases. Each kinase was silenced to <10% of the normal level (Fig. 1A). These cells were further exposed to Aph for 48 h, and their centrosome profiles were determined by immunostaining of γ-tubulin, a centrosome marker²⁵ (Fig. 1C; representative immunostaining images are shown in Fig. 1B). In the control cells, ∼70% of cells contained amplified centrosomes. In contrast, cells silenced for PLK2, PLK4 or ROCK2 all showed ∼40% frequency of centrosome amplification. Because 20–30% of p53-null MEFs already contain amplified centrosomes, and ∼10% of cells undergo centrosome amplification during the transfection period, ∼40% frequency of centrosome amplification in this assay is equated with the near complete block of centrosome re-duplication. Thus, as shown previously,^14,19,23 depletion of PLK2, PLK4 and ROCK2 all results in failure to re-duplicate centrosomes in the Aph-arrested cells.

Figure 1.

ROCK2, PLK2 and PLK4 are all essential for centrosome re-duplication in the Aph-arrested p53−/− MEFs. We used the pSUPER.puro plasmid system to introduce siRNA sequences targeting ROCK2, PLK2 and PLK4 into cells. This system allows selection of cells that are successfully transfected with the siRNA sequences by puromycin (Puro)-selection. (A) The cells were co-transfected with the siRNA sequence targeting each of these kinases and GFP-plasmid (as a transfection marker), and selected with Puro-treatment for 48 h. For controls, the randomized siRNA sequences were used. The lysates prepared from the transfected cells were immunoblotted with anti-ROCK2 (a), anti-PLK2 (b) and anti-PLK4 (c) antibodies. (B and C) The cells silenced for ROCK2, PLK2 and PLK4 were treated with Aph for 48 h, and the percent of cells that contain ≥3 centrosomes were determined by co-immunostaining with anti-GFP and anti-γ-tubulin antibodies. Cells were also stained for DNA with DAPI. Immunostaining of GFP identifies the cells successfully transfected with the siRNA sequences. The representative immunostaining images are shown in (B). The panels e, j, o, t show the magnified images of the areas indicated in panels b, g, l, q, respectively. The arrowheads point to centrosomes. Scale bar; 10 μm. The percent of cells with amplified centrosomes are shown in the graph (C) as means ± SE from 3 independent experiments. For each experiment, >200 cells were examined. **P < 0.01.

Both PLK2 and PLK4 can rescue the ROCK2-silenced cells to undergo centrosome amplification during the Aph-induced arrest

The study shown in Figure 1 demonstrated that PLK2, PLK4 and ROCK2 are all essential for induction of centrosome amplification in the Aph-arrested cells, making it possible to address the question of whether these kinases operate in the linear pathway or independently from each other to drive centrosome amplification by the rescue experiment; one of the 3 kinases is silenced, and cells will be tested whether ectopic expression of other 2 kinases can rescue the failure to undergo centrosome amplification. We first tested whether PLK2 and PLK4 can rescue the ROCK2-silenced cells. The ROCK2-silenced cells were pre-treated with Aph for 16 h, and transfected with either PLK2 or PLK4. After confirming that both PLK2 and PLK4 were expressed at comparable levels in the control and ROCK2 siRNA-transfected cells (Fig. 2A), the transfected cells were further exposed to Aph, and their centrosome profiles were determined (Fig. 2C; representative immunostaining images are shown in Fig. 2B). Both PLK2 and PLK4 successfully rescue the ROCK2-silenced cells to re-duplicate centrosomes, suggesting that both PLK2 and PLK4 may be downstream of ROCK2 to drive centrosome re-duplication in the Aph-arrested cells.

Figure 2.

Both PLK2 and PLK4 can rescue the ROCK2-silenced cells to undergo centrosome re-duplication during the Aph-induced arrest. (A) p53−/− MEFs were transfected with either pSUPER.puro-ROCK2 or control vector with randomized sequences, and selected with Puro for 48 h. Cells were pretreated with Aph for 16 h, then transfected with GFP-PLK2, GFP-PLK4, or control GFP- plasmid. The lysates prepared from the transfected cells were immunoblotted with anti-GFP antibody. (B and C) The transfected cells were further incubated in the presence of Aph for 48 h, and the centrosome profiles of the GFP-positive cells were determined by co-immunostaining with anti-GFP and anti-γ-tubulin antibodies. The representative immunostaining images are shown in (B). For images shown, we specifically selected the areas with cells with successfully transfected with GFP-PLK2 or -PLK4 (GFP-positive) and not transfected (GFP-negative). The magnified images of the areas indicated in panels b, f, j, n are shown on the right, in which the arrowheads point to centrosomes. Scale bar; 10 μm. The percent of cells with amplified centrosomes are shown in the graph (C) as means ± SE from 3 independent experiments. For each experiment, >200 cells were examined. **P < 0.01. It should be noted that the frequency of centrosome amplification in the control cells increased 5–10% more compared with the results shown in Figure 1. The reason is that prior to transfection of GFP-PLK2 and -PLK4, cells were pre-treated with Aph for 16 h (no pre-treatment in the experiment shown in Fig. 1), and centrosome amplification was induced to some extent during the pre-treatment period. This also applies to the experiments shown in Fig. 3 and 4.

PLK2 and ROCK2 fail to rescue the PLK4-silenced cells to promote centrosome amplification during the Aph-induced arrest

We next tested whether ROCK2 and PLK2 could rescue the PLK4-silenced cells to re-duplicate centrosomes during the Aph-induced arrest. Because ROCK2 has the C-terminal auto-inhibitory domain, which folds back to interact with the N-terminal kinase domain, resulting in inhibition of the kinase activity, we used the ROCK2 mutant whose C-terminal autoinhibitory domain was deleted (ROCK2/CAT)²⁶ instead of wt ROCK2 for this study. The PLK4-silenced cells were pre-treated with Aph, and transfected with either ROCK2/CAT or PLK2. Confirming that ROCK2/CAT as well as PLK2 were expressed at comparative levels in the control and PLK4-silenced cells (Fig. 3A and C), the transfected cells were further exposed to Aph, and their centrosome profiles were determined (Fig. 3B and D). PLK2 failed to rescue the PLK4-silenced cells to re-duplicate centrosomes, which is consistent with the previous study reported by Cizmecioglu et al., showing that PLK2 drives centrosome amplification in the PLK4-dependent manner in the arrested cells.²⁰ Similarly, ROCK2/CAT also failed to rescue the PLK4-silenced cells to re-duplicate centrosomes. To exclude the possibility that silencing of PLK4 might have led to unrecoverable damages that abrogate the centrosome re-duplication potential, we transfected the PLK4 that were genetically engineered to be resistant to the siRNA used for silencing of PLK4 (denoted as PLK4*). The introduction of PLK4* could restore the ability of the PLK4-silenced cells to undergo centrosome amplification (Fig. 3B and D), indicating that the PLK4-silenced cells retain the ability to undergo centrosome amplification during the Aph-induced arrest. Thus, PLK4 appears to promote centrosome re-duplication in the Aph-arrested cells downstream of PLK2 and ROCK2.

Figure 3.

ROCK2 and PLK2 fail to rescue the Aph-arrested PLK4-silenced cells to undergo centrosome re-duplication. (A) p53−/− MEFs were transfected with pSUPER.puro-PLK4 or control vector with randomized sequences, and selected with Puro for 48 h. The transfected cells were then pre-treated with Aph for 16 h and transfected with either GFP-ROCK2/CAT or -PLK4* (PLK4 siRNA-resistant mutant). The lysates prepared from the transfected cells were immunoblotted with anti-GFP antibody. (B) The transfected cells described in (A) were further incubated in the presence of Aph for 48 h, and the centrosome profiles of the GFP-positive cells were determined. The results are shown in the graph as means ± SE from 3 independent experiments. For each experiment, >200 cells were examined. **P < 0.01. (C) Cells transfected with the PLK4 siRNA described above were pre-treated with Aph, and transfected with either GFP-PLK2 or -PLK4*. The lysates prepared from the transfected cells were immunoblotted with anti-GFP antibody. (D) The transfected cells described in (C) were incubated in the presence of Aph for 48 h. Then, the centrosome profiles of the GFP-positive cells were determined, and the results are shown in the graph as means ± SE from 3 independent experiments. For each experiment, >200 cells were examined. **P < 0.01.

Both ROCK2 and PLK4 can rescue the PLK2-silenced cells to undergo centrosome amplification during the Aph-induced arrest

We next tested whether ROCK2 and PLK4 can rescue the PLK2-silenced cells to re-duplicate centrosomes during the Aph-induced arrest. The PLK2-silenced cells were pre-treated with Aph, and transfected with either ROCK2/CAT or PLK4. After confirming that ROCK2/CAT and PLK4 were expressed at comparative levels in the control and PLK2-silenced cells (Fig. 4A and C), the transfected cells were further exposed to Aph, and their centrosome profiles were determined. PLK4 successfully rescued the PLK2-silenced cells to re-duplicate centrosomes, which is consistent with the finding described in Figure 3, indicating that PLK4 functions downstream of PLK2 to induce centrosome amplification in the Aph-arrested cells (Fig. 4D). Interestingly, although PLK2 could rescue the ROCK2-silenced cells (Fig. 2), ROCK2 could also rescue the PLK2-silenced cells to re-duplicate centrosomes (Fig. 4B). Because we used the ROCK2 mutant deleted for the auto-inhibitory region (ROCK2/CAT) in the experiment, to exclude the possibility that the use of the mutant might have influenced to the results, we also tested using wt ROCK2. Wt ROCK2 requires being primed for activation by the Rho small GTPase binding. The binding of the active form of Rho (GTP-bound Rho: Rho-GTP) releases the kinase domain from the negative interaction with auto-inhibitory region.^17,18 We found that wt ROCK2 could rescue the PLK2-silenced cells to induce centrosome amplification during the Aph-induced arrest, although slightly less efficiently than the ROCK2/CAT mutant, which likely reflected the amount of available Rho-GTP in the cells (Supplementary Fig. S1). Indeed, co-transfection of the constitutively GTP-bound form of RhoA (RhoA-L63: Gln63 to Leu)²⁷ rescued the PLK2-silenced cells at the level comparable with the ROCK2/CAT (Fig. S1). Thus, ROCK2 and PLK2 can rescue the depletion of the other to induce centrosome amplification during the Aph-induced arrest, suggesting the potential functional interaction between PLK2 and ROCK2.

Figure 4.

Both ROCK2 and PLK4 can rescue the PLK2-silenced cells to undergo centrosome re-duplication during the Aph-induced arrest. (A) p53−/− MEFs were transfected with pSUPER.puro-PLK2 or control vector with randomized sequences, and selected with Puro for 48 h. Cells were then pre-treated with Aph for 16 h, and transfected with GFP-control vector or GFP-ROCK2/CAT. The lysates prepared from the transfected cells were immunoblotted with anti-GFP antibody. (B) The transfected cells described in (A) were further incubated in the presence of Aph for 48 h. The centrosome profiles of the GFP-positive cells were determined, and the results are shown in the graph as means ± SE from 3 independent experiments. For each experiment, >200 cells were examined. **P < 0.01. (C) Cells transfected with PLK2 siRNA described above were pre-treated with Aph, and transfected with either GFP-PLK4 or control siRNA sequences. The lysates prepared from the transfected cells were immunoblotted with anti-GFP antibody. (D) The transfected cells described in (C) were incubated in the presence of Aph for 48 h. The centrosome profiles of the GFP-positive cells were determined, and the results are shown in the graph as means ± SE from 3 independent experiments. For each experiment, >200 cells were examined. **P < 0.01.

Ser4 phosphorylation of NPM by PLK2 accounts for the abilities of PLK2 and ROCK2 to mutually rescue for induction of centrosome amplification during the Aph-induced arrest

ROCK2 can rescue the PLK2-silenced cells (Fig. 4), and PLK2 can rescue the ROCK2-silenced cells to drive centrosome amplification during the Aph-induced arrest (Fig. 2). To find the explanation for these seemingly contradicting observations, we focused on the NPM that is involved in diverse cellular functions, including ribosome assembly, pre-rRNA and mRNA processing, centrosome duplication, DNA replication, nucleocytoplasmic protein trafficking, mitotic spindle assembly, to name a few. NPM is believed to participate in these diverse cellular functions through its molecular chaperoning activity as well as activity to control target proteins via direct binding.²⁸ It has recently been shown that PLK2 phosphorylates Ser4 of NPM, and this phosphorylation is critical for re-duplication of centrosomes in arrested cells.²¹ We have previously shown that NPM physically interacts with ROCK2, which leads to “super-activation” of ROCK2, and the NPM-binding is critical for ROCK2 to drive centrosome amplification.¹⁴ Moreover, Thr199 of NPM is phosphorylated by CDK2-cyclin E (as well as cyclin A),¹³ which dramatically increases the binding affinity of NPM to ROCK2.¹⁴ Although unphosphorylated NPM is able to bind to and super-activate ROCK2, the Thr199 phosphorylation plays a critical role in NPM-ROCK2 interaction in vivo where the protein concentrations are limited. We thus decided to investigate the possible relationship between the PLK2-mediated phosphorylation on Ser4 and CDK2-mediated phosphorylation on Thr199 of NPM. We first performed the in vitro kinase assay using bacterially purified NPM as a substrate and immunopurified CDK2-cyclin E as well as either immunopurified wt PLK2 (PLK2/WT) or kinase-dead PLK2 mutant (PLK2/KD) (Fig. 5A). NPM was efficiently phosphorylated on Ser4 in the reaction with PLK2/WT (lane 2, 2nd panel), but failed to be phosphorylated in the reaction with PLK2/KD (lane 1, 2nd panel). Importantly, NPM was more efficiently phosphorylated (∼4-fold) on Thr199 by CDK2-cyclin E in the reaction with PLK2/WT (lane 2, 1st panel) than that with PLK2/KD (lane 1, 1st panel). Thus, PLK2-mediated phosphorylation on Ser4 of NPM facilitates the CDK2-mediated phosphorylation on Thr199. We next tested whether this is the case in vivo. Cells were co-transfected with HA-tagged CDK2 and either FLAG-tagged wt NPM (NPM/WT) or Ser4 non-phosphorylatable NPM mutant (Ser4→Ala; S4A). Wt NPM was efficiently phosphorylated on Thr199 (Fig. 5B, lane 2, 1st panel), while the phosphorylation of the NPM/S4A mutant on Thr199 was barely detectable (lane 3, 1st panel). Moreover, co-expression of PLK2 further increased the level of Thr199 phosphorylation of wt NPM, but not the NPM/S4A mutant (lane 4 and 5, 1st panel). Thus, Ser4 phosphorylation by PLK2 facilitates Thr199 phosphorylation by CDK2 in vivo as well.

Figure 5.

PLK2-mediated phosphorylation of NPM on Ser4 facilitates the phosphorylation of NPM on Thr199 by CDK2. (A) In vitro kinase reaction was performed with bacterially purified 6× His-tagged NPM and immuno-purified GFP-wt PLK2 or kinase-dead mutant (KD) in the kinase reaction buffer for 30 min, then immuno-purified CDK2-cyclinE were added to the reaction mixtures, and further incubated for 30 min. The reaction mixtures were resolved by SDS-PAGE, and stained with Ponceau S for visualization of NPM bands. The reaction mixtures were also immunoblotted with anti-phospho-Thr199-NPM, -phospho-Ser4-NPM, and -GFP antibodies. (B) p53−/− MEFs were co-transfected with HA-CDK2 and either FLAG-NPM/WT or NPM/S4A mutant. The co-transfection was also performed with inclusion of GFP-PLK2/WT. The lysates prepared from the transfected cells were subjected to immunoprecipitation with anti-FLAG antibody, and the immunoprecipitates were blotted with anti-phospho-T199-NPM and -FLAG antibodies. The lysates (5% of the amounts used for immunoprecipitation) were also immunoblotted with anti-FLAG, -HA, -GFP, and −β-actin antibodies as references. (C) A schematic model of the linear pathway involving ROCK2, PLK2 and PLK4 to drive centrosome amplification in the cell cycle-arrested cells.

The findings described above provide a reasonable explanation for the abilities of PLK2 and ROCK2 to override the depletion of each other in respect to centrosome amplification in the arrested cells. The siRNA-mediated silencing does not completely deplete a target protein, but always leave the residual amounts (5–10% of the normal level). NPM binds to ROCK2, and NPM-binding is critical for ROCK2 to promote centrosome re-duplication. Although nascent form of NPM can binds to ROCK2, phosphorylation on Thr199 by CDK2 increases the binding affinity of NPM to ROCK2. We found that phosphorylation of NPM on Ser4 by PLK2 facilitates the phosphorylation of NPM on Thr199. Thus, when PLK2 is overexpressed, more NPM proteins are phosphorylated on Ser4, which results in an increase in the amount of Thr199-phosphorylated NPM. The increase in the concentration of NPM with a high binding affinity to ROCK2 can effectively binds to ROCK2, and thus is able to compensate the reduced level of ROCK2 in the ROCK2-silenced cells. Therefore, overexpression of PLK2 can rescue the ROCK2-silenced cells to undergo centrosome amplification. In support of this prediction, overexpression of PLK2 fails to rescue the cells with ROCK2 nearly completely inhibited by the potent ROCK specific chemical inhibitor to undergo centrosome amplification during the Aph-induced arrest (Fig. S2), indicating that PLK2 requires the activity of ROCK2 to drive centrosome amplification. In contrast, in the PLK2-silenced cells, phosphorylation of NPM on Ser4 is largely abrogated. Because Ser4 phosphorylation is pre-requisite for Thr199 phosphorylation by CDK2, depletion of PLK2 results in failure of NPM to be phosphorylated on Thr199. However, because the nascent form of NPM can still bind to ROCK2, although not efficiently,¹⁴ the presence of excess ROCK2 allows the non-phosphorylated NPM to manage to bind to ROCK2 under a physiological condition, and thus overexpression of ROCK2 can compensate the reduced concentration of Thr199-phosphorylated NPM. In support, it has previously been shown that overexpression of ROCK2 can rescue the cells inactivated for the CDK2 kinase activity to undergo centrosome amplification during the Aph-induced arrest.²⁹

Taken all together, we propose the following model for the functional relationship among PLK2, PLK4 and ROCK2 to promote centrosome amplification in the cell cycle-arrested cells (Fig. 5C). PLK2 phosphorylates NPM on Ser4, which promotes phosphorylation of NPM on Thr199 by CDK2-cyclin E (or cyclin A). NPM acquires a high binding affinity to ROCK2 by Thr199 phosphorylation, which leads to facilitation of NPM to interact with ROCK2 at centrosomes. ROCK2 then either directly or indirectly acts on PLK4, which in turn acts on its targets, leading to induction of centrosome re-duplication. Several potential targets of PLK4 have been recognized, including CP110 and centriolar protein SAS-6.²³ The activity of SCF-FBXW5 E3-ubiquitin ligase, which controls centrosome duplication by degradation of SAS-6, has also been shown to be regulated by PLK4-mediated phosphorylation of FBXW5.³⁰ In addition, PLK4 physically interacts with and phosphorylates GCP6, a core component of the γ-tubulin ring complex, and PLK4-mediated phosphorylation of GCP6 has been shown to be important for centrosomes to duplicate.³¹ In this model, the Thr199-phosphorylation of NPM is depicted as the sole function of CDK2 in induction of centrosome amplification during the arrest. However, CDK2-cyclin E is expected to target multiple effectors, including Mps1,³² to induce initiation of centrosome duplication in cycling cells and re-duplication in arrested cells. Indeed, the study reported by Habedanck et al. has shown that overexpression of PLK4 fails to induce centrosome amplification when CDK2 was inhibited by expression of dominant-negative (DN) CDK2.³³ Interestingly, as mentioned earlier, in our previous study examining the relationship between ROCK2 and CDK2 in induction of centrosome amplification in the arrested cells, we found that overexpression of ROCK2 could induce centrosome amplification in the cells expressing DN CDK2.²⁹ The discrepancy between these 2 studies may be explained by the difference in the experimental procedures employed. In the study by Habedanck et al., PLK4 and DN CDK2 were directly introduced into cycling cells, and cells were exposed to Aph. ROCK2 is known to function in cytokinesis as well, and overexpression of ROCK2 in cycling cells interferes with proper cytokinesis. Because of this reason, in our previous study (as well as the present study), cells were necessary to be arrested in S-phase by Aph exposure prior to transfection of ROCK2 and DN CDK2. It is possible that the other CDK2-associated events critical for centrosome re-duplication may be completed during the pre-arresting period, which allows re-duplication of centrosomes.

One important question remained to be addressed is whether the linear pathway involving PLK2, PLK4 and ROCK2 that drives centrosome re-duplication in the arrested cells also operates in the initiation of centrosome duplication in the cycling cells. The physiological states of arrested cells and cycling cells are apparently different, and thus, although it is assumed that centrosome duplication in cycling cells and re-duplication in arrested cells are controlled by many common proteins and mechanisms, it is also expected that there are some differences between them.³⁴ Because PLK2, PLK4 and ROCK2 are all essential for initiation of centrosome duplication in cycling cells,^14,19,22,23 the pathway involving PLK2, PLK4 and ROCK2 to promote centrosome re-duplication in arrested cells shown in this study likely operates in the regulation of centrosome duplication in cycling cells.

Materials and Methods

Cell culture, plasmids and transfection

p53^−/− MEFs were maintained in DMEM supplemented with 10% fetal bovine serum, penicillin (100 U/ml) and streptomycin (100 µg/ml) in an atmosphere containing 10% CO₂ at 37°C. We used a pSUPER vector system for the expression of the plasmid-based siRNA. The PLK2 sequence corresponding to its cDNA 1474–1492 (5′-GAG CAG CTG AGC ACG TCC T-3′)¹⁹ was used as the siRNA. For PLK4, the sequence corresponding to its cDNA 2305–2323 (5′-GGT AAT ACT AGT TCA CCT A-3′)²³ was used. The siRNA sequence for ROCK2 is described in the previous study.¹⁴ For the control, a randomized sequence was inserted into the vector.

For transfection of siRNA sequences, to maximize the transfection efficiency, we used the pSUPER plasmid harboring a puromycin (Puro)-resistance gene. After transfection, successfully transfected cells were selected by exposure to Puro for 48 h.

GFP-PLK2 and GFP-PLK4 were generated by inserting cDNAs of PLK2 (a gift from Dr. Wafik El-Deiry, University of Pennsylvania School of Medicine) and of PLK4 (a gift from Dr. Hiroyuki Mano, Jichi Medical University) into the pEGFP-c1 vector. HA-CDK2 is a gift of Dr. Mark Alexandrow (H. Lee Moffitt Cancer Center). The ROCK2/CAT mutant is described previously.¹⁴ The Flag-tagged protein plasmids were constructed using the p3× FlagCMV7.1 vector.

Antibodies

The primary antibodies used in this study were: anti-ROCK2 (610623, BD Biosciences), anti-HA (12CA5, Roche), anti-GFP (7.1/13.1, Roche, sc-8334, Santa Cruz), anti-β-actin (AC-74, Sigma-Aldrich), anti-FLAG (M2 and F7425, Sigma-Aldrich), anti-PLK2 (sc-25421, Santa Cruz), anti-PLK4 (sc-3258S, Cell Signaling), anti-phospho-Ser4-NPM (D19C1, Cell signaling), and anti-phospho-Thr199-NPM (#3541, Cell Signaling) antibodies.

Immunoblot analysis

Cells were lysed in lysis buffer [1% SDS, 1% NP-40, 50 mM Tris (pH8.0), 150 mM NaCl, 4 mM Pefabloc SC, 2 μg/ml leupeptin, 2 μg/ml aprotinin]. The lysates were briefly sonicated, boiled for 5 min, and cleared by centrifugation at 20,000 g at 4°C for 10 min. The lysates were denatured at 95°C for 5 min in sample buffer [2% SDS, 10% glycerol, 60 mM Tris (pH6.8), 5% β-mercaptoethanol, 0.01% bromophenol blue], resolved by SDS-PAGE and transferred to Immobilon-P sheets (Millipore Corp., Bedford, MA, USA). The blots were incubated in blocking buffer [(5% (w/v) non-fat dry milk in Tris-buffered saline + 0.2% Tween 20 (TBST)] for 1 h at room temperature, and then incubated with primary antibody for 16 h at 4°C, rinsed in TBST, and incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. The blots were rinsed in TBST, and the antibody-antigen complex was visualized by Pierce ECL Western Blotting Substrate Kit (Thermo Scientific, Rockford, IL, USA) or SuperSignal West Femto Maximum Sensitivity Substrate Kit (Thermo Scientific, Rockford, IL, USA).

For immunoblot analysis of PLK4, because PLK4 is a highly unstable protein, rapidly degraded by the ubiquitin-dependent proteolytic machinery,³¹ to visualize the PLK4 band clearly, cells were pre-treated with MG132 proteasome inhibitor (10 μM) for 4 h prior to preparation of the lysates for immunoblotting. For all immunoblotting analyses, the lysates were immunoblotted with anti-β-actin antibody as a loading control.

Indirect immunofluorescence

Cells were washed in PBS, and fixed with cold methanol for 20 min. The cells were incubated with 10% normal goat serum in PBS for 1 h at room temperature, and probed with primary antibodies for 2 h. The antibody-antigen complexes were detected with either Alexa Fluor 488- or Alexa Fluor 594-conjugated goat antibody (Molecular Probes, Eugene, OR, USA). To detect GFP-fusion proteins, Alexa Fluor 488-conjugated rabbit polyclonal antibody was used. The cells were also counterstained for DNA with DAPI, and then examined under a fluorescence microscope (Zeiss Automated Upright/Inverted Fluorescent Microscope).

In vitro kinase assay

In vitro kinase assay was performed by incubating immuno-purified kinases and bacterially purified recombinant 6× His-tagged NPM in a reaction buffer [40 mM Tris-HCl (pH7.5), 20 mM MgCl₂, 0.1 mg/ml BSA, 50 μM DTT, and 50 μM ATP] at 30°C for 30 min. The reaction mixtures were resolved by SDS-PAGE, and phosphorylation of NPM was detected by immunoblotting with anti-phospho-Ser4-NPM and anti-phospho-Thr199-NPM antibodies.

Statistical analysis

The data were analyzed by one-way analysis of variance, followed by Scheffe's F test. Difference with P < 0.05 was considered as significant.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank J. Johnson and A. Kasprzak for technical support for microscopy.

Funding

This work was supported by the grant GM087328 from the National Institute of Health to K. F.

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website.