Polo-like kinase 1 (PLK1) is a widely conserved serine/threonine kinase that regulates progression of multiple stages of mitosis. Although extensive studies about PLK1 functions during cell division have been performed, it is still not known how PLK1 regulates myosin II activation at the equatorial cortex and ingression of the cleavage furrow. In this report, we show that an actin/myosin-II-binding protein, supervillin (SVIL), is a substrate of PLK1. PLK1 phosphorylates Ser238 of SVIL, which can promote the localization of SVIL to the central spindle and association with PRC1. Expression of a PLK1 phosphorylation site mutant, S238A-SVIL, inhibited myosin II activation at the equatorial cortex and induced aberrant furrowing. SVIL has both actin- and myosin-II-binding regions in the N-terminus. Expression of ΔMyo-SVIL (deleted of the myosin-II-binding region), but not of ΔAct-SVIL (deleted of actin-binding region), reduced myosin II activation and caused defects in furrowing. Our study indicates a possible role of phosphorylated SVIL as a molecular link between the central spindle and the contractile ring to coordinate the activation of myosin II for the ingression of the cleavage furrow.
Cytokinesis is the final step of cell division that physically disconnects two daughter cells. After the separation of chromosomes, bundles of antiparallel microtubules, called the central spindle, are organized, and an actomyosin-based contractile ring is assembled on the inner face of the plasma membrane at the division plane (Barr and Gruneberg, 2007; Glotzer, 2009). Constriction of the ring by myosin motor activity generates forces that drive the ingression of the cleavage furrow (Matsumura, 2005). Once the furrowing is complete, the ring is disassembled, and daughter cells are abscised by multi-protein complexes to complete cytokinesis. One of the major regulatory mechanisms to ensure the accurate progression of these processes is phosphorylation by mitotic kinases, such as polo-like kinase 1 (PLK1) and the Aurora-A and Aurora-B kinases. These kinases dynamically localize to the subcellular structures during mitosis and regulate bipolar spindle formation, microtubules/kinetochore interactions, formation and ingression of cleavage furrow, and abscission (Salaun et al., 2008; Taylor and Peters, 2008).
PLK1 is conserved in a wide range of species and is composed of the N-terminal kinase domain and the C-terminal polo-box domain (PBD) that associates with phosphorylated sequences. PLK1 is localized to different subcellular structures, such as centrosomes, kinetochores, the central spindle and the midbody, and regulates the progression of multiple stages of mitosis (Barr et al., 2004). Localization of PLK1 to the different subcellular structures is mediated by binding of PBD and phosphorylated proteins at the specific sites (Elia et al., 2003a; Elia et al., 2003b). In the early mitotic phase, PLK1 phosphorylates the centrosome- and kinetochore-localized proteins to regulate the formation of the mitotic spindle and the separation of chromosomes to the opposite poles. In the absence of PLK1, cells fail to establish a bipolar spindle and instead assemble a monopolar spindle, surrounded by chromosomes arranged in a circular fashion (Petronczki, et al., 2008). The critical role that PLK1 plays in the early phase of mitosis had previously hampered the elucidation of its role in the later stage of mitosis; however, the recent discovery of chemical inhibitors enabled the specific inactivation of PLK1 at anaphase and demonstrated an essential role for the initiation of cytokinesis (Brennan et al., 2007; Burkard et al., 2007; Petronczki et al., 2007; Santamaria et al., 2007). Central spindle-localized proteins, including MKLP1, MKLP2 and PRC1, are targets of PLK1, and the phosphorylated proteins associate with PBD to recruit PLK1 to the central spindle (Liu et al., 2004; Neef et al., 2003; Neef et al., 2007). Once at the central spindle, PLK1 phosphorylates MgcRacGAP to recruit ECT2 at the central spindle, which then activates RhoA at the equatorial cortex to initiate furrowing (Burkard et al., 2009; Wolfe et al., 2009; Yüce et al., 2005). Expression of a non-phosphorylated mutant of MgcRacGAP suppressed localization of RhoA as well as myosin II and F-actin at the equatorial cortex and inhibited furrowing. In addition to the furrowing, a recent report showed that PLK1 mediates the timing of abscission by phosphorylating the midbody-localized protein, Cep55 (Bastos and Barr, 2010). Although these findings have advanced our current understanding of how PLK1 controls cytokinesis, there seems to be more complicated regulatory networks mediated by PLK1 to coordinate cytokinesis. Thus, elucidation of the physiological role of other PLK1 substrates is essential to fully realize the regulatory mechanisms of cytokinesis.
To search for novel PLK1 substrates essential for cytokinesis, we performed a limited siRNA screen. In this screen, we found that an actin/myosin II-binding protein, supervillin (SVIL), was required for the completion of cytokinesis. SVIL consists of an N-terminal F-actin- and myosin-II-binding regions and a C-terminal region, similar to villin/gelsolin proteins, and is involved in cell spreading, cell–substrate attachment and matrix degradation (Bhuwania et al., 2012; Chen et al., 2003; Crowley et al., 2009; Fang et al., 2010; Pestonjamasp et al., 1997; Takizawa et al., 2006; Takizawa et al., 2007). A recent study has shown that SVIL associates with numerous cytokinesis-related proteins and was required for cytokinesis (Smith et al., 2010). However, how SVIL controls cytokinesis and the molecular mechanisms involved remain largely unknown. In this report, we show that PLK1 phosphorylates Ser238 of SVIL to induce association with PRC1 for the loading of SVIL to the central spindle. In addition, we show that SVIL functions as a scaffold that connects the central spindle and the contractile ring and promotes the activation of myosin II to guide the cleavage furrow in the restricted zone at the division plane.
Depletion of SVIL compromises cytokinesis
PLK1-mediated phosphorylation plays a crucial role in the progression of cytokinesis. We performed an siRNA screen to identify novel PLK1 substrates essential for the completion of cytokinesis. A previous study has identified a large number of proteins whose phosphorylation is increased in mitosis (Dephoure et al., 2008). Based on their results, we selected 88 genes whose products are phosphorylated in mitosis and have phosphorylation sites similar to the PLK1 consensus sites (supplementary material Table. S1). HeLa cells cultured in 24-well plates were transfected with each siRNA, and 72 hours later the cells were fixed and stained with Hoechst and FITC-labeled paclitaxel for tubulin. We performed quantitative analysis on the number of multinucleated cells because the depletion of proteins required for cytokinesis is known to induce multinucleated cells. In the first round of screen, the depletion of six genes (SQSTM1, CALCOCO1, TBC1D15, SEC61B, SVIL, UBAP2L) induced more than 10% multinucleated cells (Fig. 1A). In the second screen, we used two additional siRNAs for these six genes for further validation, and we found that the transfection of all four different siRNAs targeting supervillin (SVIL) increased the ratio of multinucleated cells more than 10% (Fig. 1A; supplementary material Fig. S1A). We generated polyclonal antibodies against SVIL to examine the efficiency of the siRNAs to knockdown the protein. A recent study showed that two isoforms of SVIL, which are isoform 1 and isoform 4, are expressed in non-muscle cells (Fang and Luna, 2013). Isoform 4 has additional 395 amino acids after Pro245 of isoform 1. The antibody detected two immunoreactive bands of ∼220 kDa and 250 kDa, which correspond to isoform 1 and isoform 4, respectively, and both bands were lost by the transfection of SVIL siRNAs (Fig. 1B). To confirm that the increase of multinucleated cells was induced by SVIL depletion, we carried out a rescue experiment. Major isoform expressed in HeLa cells was isoform 1 and depletion of isoform 4 alone did not induce multinucleated cells (supplementary material Fig. S1B), thus we used isoform 1 for the rescue experiment. HeLa cells were transfected with SVIL siRNA together with a plasmid encoding GFP-tagged either wild-type or siRNA-resistant SVIL, which contains silent mutations that confer resistance to the siRNA. Immunoblot analysis showed that SVIL siRNA effectively ablated both endogenously and exogenously expressed wild-type SVIL, whereas the expression of siRNA-resistant SVIL was not affected by the siRNA (Fig. 1C). The ratio of multinucleated cells was significantly reduced in cells transfected with siRNA-resistant SVIL compared to wild-type SVIL (Fig. 1D). These results indicate that the loss of SVIL function compromises cytokinesis. To gain further insight into SVIL function in cytokinesis, we examined the subcellular localization of SVIL during cytokinesis by immunofluorescence analysis using the anti-SVIL antibody. Diffuse localization of SVIL was observed in metaphase. In anaphase, SVIL accumulated in the region of the midzone of the microtubules of dividing cells, called the central spindle. As the cells progressed to cytokinesis, SVIL became concentrated at the midbody (Fig. 1E). These specific localizations were not observed in cells transfected with SVIL siRNA (supplementary material Fig. S1C). SVIL localization was clearly observed in cells depleted of isoform 4 alone, indicating that isoform 1 is localized to the central spindle (supplementary material Fig. S1D). We also checked the localization of exogenously expressed GFP–SVIL (isoform 1). Similar to the localization of endogenous SVIL, GFP–SVIL localized at the central spindle and at the midbody (supplementary material Fig. S1E). Faint localization of GFP–SVIL along the spindle was observed in metaphase, but this appears to be a non-specific signal because similar localization was observed in GFP-transfected cells (supplementary material Fig. S1E).
PLK1 phosphorylates SVIL at Ser238
We next examined phosphorylation of SVIL during mitosis. HeLa cells synchronized by double thymidine block were lysed at different stages of the cell cycle, and the electrophoretic mobility shift of SVIL was investigated using Phos-tag, which can enhance the mobility shift when proteins are phosphorylated. We measured the thickness of each SVIL band to determine the level of mobility shift. A clear increase in the thickness of SVIL band was observed when cyclin B1 expression level has started to decrease, suggesting that a mitotic kinase other than cyclin B1/Cdk1 is responsible for the phosphorylation (Fig. 2A). Analysis of the amino acid sequence of SVIL has revealed that there is one potential PLK1 phosphorylation site, Ser238 (Fig. 2B). Accumulating evidence has demonstrated that PLK1 can phosphorylate the serine or threonine residue in the consensus sequence ([E/D]×[S/T]Φ×[E/D]; Φ, a hydrophobic amino acid) (Nakajima et al., 2003). In addition, the proteomics analysis performed by Dephoure et al. detected S238 phosphorylation by mass spectrometry analysis (Dephoure et al., 2008). To further confirm whether PLK1 phosphorylates Ser238 of SVIL in the cells, we checked the mobility shift of SVIL in the presence of PLK1 expression. HA-tagged full length or deletion mutants of SVIL were transiently expressed with or without wild-type PLK1, and cell lysates were subjected to immunoblot analysis. We observed the mobility shift of aa1–345 SVIL in the presence of wild-type PLK1, but other deletion mutants did not show any shift (Fig. 2C). To confirm that the shift is due to the phosphorylation, we used λ-phosphatase. HA-tagged aa1–345 SVIL was expressed together with wild-type or kinase-dead PLK1 (K82M), and the immunoprecipitated SVIL fragment was treated with λ-phosphatase. The mobility shift induced by wild-type PLK1 was clearly eliminated by λ-phosphatase treatment (Fig. 2D). We next tested if the mobility shift of aa1–345 SVIL was due to the phosphorylation of Ser238. Wild-type or an S238A mutant of aa1–345 SVIL was expressed in the presence of either wild-type or kinase-dead PLK1. Immunoblot analysis showed a mobility shift of wild-type, but not S238A, aa1–345 SVIL in the presence of wild-type PLK1 (Fig. 2E). To confirm that Ser238 of SVIL was directly phosphorylated by PLK1, we performed an in vitro kinase assay. GST-fused wild-type or S238A aa228–256 of SVIL was purified from E. coli and mixed with recombinant PLK1. The reaction mixtures were separated by SDS-PAGE gel and subjected to autoradiography. Wild-type aa228–256 was phosphorylated by PLK1, whereas substitution at Ser238 abolished the phosphorylation (Fig. 2F). We also performed an in vitro kinase assay using full-length wild-type and S238A-SVIL. As shown in Fig. 2G, phosphorylation of S238A-SVIL was clearly reduced compared to that of wild-type SVIL. Phosphorylation of full-length S238A-SVIL indicates that there may be additional phosphorylation sites by PLK1. Although we failed to generate a specific antibody for phosphorylated Ser238, these results indicate that PLK1 directly phosphorylates SVIL at Ser238 in vivo and in vitro. We also examined the co-localization of SVIL and PLK1 during mitosis. Both SVIL and PLK1 were localized to the central spindle and midbody from the onset of anaphase to the completion of cytokinesis (supplementary material Fig. S2).
PLK1 is essential for the localization of SVIL
Phosphorylation by mitotic kinases plays important roles in the localization of mitosis-related proteins. If PLK1-mediated phosphorylation is required for the localization of SVIL to the central spindle, the suppression of PLK1 activity should result in an aberrant localization of SVIL in anaphase. To specifically inhibit the activation of PLK1 during cytokinesis, HeLa cells were arrested in prometaphase by nocodazole treatment and released in the presence of MG132 for 1 hour. Cells were then released with or without the PLK1 inhibitor BI2536 and immunostained. SVIL and PLK1 were localized at the central spindle in anaphase; however, accumulation of both proteins at the central spindle was reduced in the presence of the PLK1 inhibitor (Fig. 3A). We next tested if localization of PLK1 at the central spindle was required for SVIL localization. PRC1, a protein essential for the organization of central spindle (Jiang et al., 1998; Mollinari et al., 2002), is phosphorylated by PLK1 during cytokinesis and the phosphorylation is crucial for the loading of PLK1 at the central spindle (Neef et al., 2007). To specifically inhibit the localization of PLK1, we used ST4AA-PRC1 in which alanines were substituted for the PLK1 phosphorylation sites (Ser577, Thr578, Ser601, Thr602). Consistent with the previous finding (Neef et al., 2007), PLK1 localization to the central spindle was reduced by the overexpression of ST4AA-PRC1 (Fig. 3B). In the presence of ST4AA-PRC1, SVIL localization at the central spindle was clearly diminished (Fig. 3B). These results indicate that PLK1 activation, as well as localization, is required for the recruitment of SVIL to the central spindle.
SVIL associates with PRC1
Although the SVIL localization to the central spindle was dependent on PLK1, we could not detect the direct interaction of these proteins. We speculated that additional factors that directly interact with SVIL were required for the recruitment of SVIL to the central spindle. As a result of extensive analysis, we identified PRC1 as a candidate SVIL binding partner. SVIL and PRC1 colocalize from anaphase to the completion of cytokinesis (supplementary material Fig. S3A). To confirm the association between these two proteins, GFP–SVIL was expressed with or without HA–PRC1 in HEK293T cells, and the cell lysates were immunoprecipitated by anti-HA antibody. GFP–SVIL was precipitated only in the presence of HA–PRC1, indicating the association of the two proteins in the cells (supplementary material Fig. S3B). The specific association was also detected by immunoprecipitation using anti-GFP antibody (supplementary material Fig. S3B). We next examined which region of SVIL was responsible for the interaction. GFP-tagged deletion mutants of SVIL were transiently expressed in the HEK293T cells together with HA–PRC1 and then immunoprecipitated with anti-HA antibody. As shown in Fig. 4A, aa676–1009 of SVIL was responsible for the interaction with PRC1. We also determined the region of PRC1 required for the binding to SVIL. PRC1 has a central spindle targeting site in the N-terminus, PLK1-phosphorylation sites in the C-terminus, and a microtubule-associated region in between (Liu et al., 2009; Mollinari et al., 2002). We found that aa1–200 of PRC1 was the critical site for the interaction with SVIL (Fig. 4B). To confirm the direct interaction of these two proteins, a pull-down assay was performed. HA-tagged aa676–1009 SVIL was produced by in vitro translation and mixed with GST-fused aa1–200 PRC1 bound to glutathione agarose beads, and then the beads were subjected to immunoblot to probe for HA. The pull-down assay clearly showed the direct interaction of these two proteins (Fig. 4C). Finally, we examined the association of endogenous proteins during mitosis. HeLa cells were arrested by double thymidine block, released and lysed at different time points. The cell lysates were immunoprecipitated with an anti-SVIL antibody and blotted for PRC1 and PLK1. Expression of cyclin B1, PRC1, AuroraB and PLK1 indicates that the cells started to enter mitosis ∼8 hours after the release. As shown in Fig. 4D, both PRC1 and PLK1 were in a complex with SVIL during mitosis.
PLK1-mediated phosphorylation of SVIL at Ser238 regulates interaction with PRC1
The above findings indicating the interaction of SVIL with PRC1, and the requirement of SVIL phosphorylation for the localization at the central spindle led us to speculate that this phosphorylation event may affect its association with PRC1. To confirm this hypothesis, we examined the association of PRC1 and phosphorylated SVIL. We first expressed HA–SVIL (wild-type and S238A) together with either wild-type or kinase-dead PLK1 and purified phosphorylated or non-phosphorylated HA–SVIL by immunoprecipitation. The precipitates were mixed with lysates made from HeLa cells that transiently expressed GFP–PRC1, and the proteins bound to HA–SVIL were affinity precipitated. The precipitates were then subjected to immunoblot analysis to probe for GFP–PRC1. As shown in Fig. 5A, the phosphorylation of SVIL by PLK1 significantly increased the amount of co-precipitated PRC1, but the substitution of alanine for Ser238 to abolished the increase in PRC1 precipitation by PLK1 treatment. To further validate this result, the binding of PRC1 to SVIL with a mutation at Ser238 to aspartic acid was examined. HA-tagged wild-type, S238A- and S238D-SVIL were expressed in HeLa cells together with GFP–PRC1, and the cell lysates were immunoprecipitated with anti-HA antibody. Immunoblot analysis of the precipitates revealed increased binding of S238D-SVIL to PRC1 compared to wild-type and S238A-SVIL (Fig. 5B). These results indicate that phosphorylation at Ser238 enhances the binding of SVIL to PRC1. We also tested if PLK1-mediated phosphorylation of PRC1 could affect the binding of two proteins. PLK1-mediated phosphorylation sites of PRC1 were replaced with either alanine (ST4AA-PRC1) or aspartic acid (ST4DD-PRC1), and the interaction with SVIL was examined. Both ST4AA-PRC1 and ST4DD-PRC1 interacted with SVIL at similar levels to wild-type PRC1 (Fig. 5C). We next examined the localization of S238A- and S238D-SVIL. Both wild-type and S238D-SVIL were localized to the central spindle, whereas the accumulation of S238A to the central spindle was significantly diminished (Fig. 5D). To further confirm the localization, we measured fluorescence intensity across the cells in anaphase. As shown in Fig. 5E, accumulation of S238A to the central spindle was clearly disrupted. These results indicate that PLK1-mediated phosphorylation at Ser238 promotes the association of SVIL with PRC1 and stabilizes SVIL localization at the central spindle.
SVIL is required to confine the cleavage furrow in the limited zone
Central spindle-localized proteins play critical functions in the initiation and progression of furrowing; thus, we examined furrow ingression in the absence of SVIL by time lapse microscopy. To monitor furrowing, we used human mammary epithelial cells, MCF10A cells, which constitutively expressed GFP-tagged myosin regulatory light chain (RLC) and DsRed-tagged histone 2B (H2B). MCF10A cells expressed only isoform 1 and siRNA transfection efficiently depleted SVIL expression (supplementary material Fig. S4A). In addition, SVIL is localized to the central spindle in MCF10A cells (supplementary material Fig. S4B). In this analysis, we noticed a marked difference in the furrowing process between control and SVIL-depleted cells. In control siRNA-transfected cells, the cleavage furrow was restricted to the narrow region of the division plane until the completion of ingression (Fig. 6A). In contrast, the furrow became significantly broadened during ingression in SVIL-knockdown cells (Fig. 6A). Measurement of the ratio of the furrow length to the length between the polar cortexes of the two separating daughter cells revealed a clear widening of the furrow in SVIL-knockdown cells (Fig. 6B). To determine whether the phosphorylation of Ser238 was required to confine the furrow in the limited region, we generated MCF10A cells that constitutively expressed siRNA-resistant GFP-tagged wild-type, S238A- and S238D-SVIL. Cells were transfected with SVIL siRNA to deplete endogenous protein, and cell division was monitored by time-lapse microscopy. The furrow became elongated in S238A-SVIL-expressing cells but not in wild-type and S238D-SVIL-expressing cells (Fig. 6C,D). F-actin and myosin II are major components of the contractile ring, and SVIL has been reported to interact with F-actin and the S2 domain of the myosin II heavy chain (Chen et al., 2003). We tested if SVIL association with myosin II or F-actin was required for the proper ingression of the furrow. We first confirmed the region of SVIL responsible for binding to F-actin and myosin II in cells. SVIL was able to associate with polymerized actin, but not with non-polymerized actin, in the cells (supplementary material Fig. S5A). Immunoprecipitation analysis demonstrated that regions of aa346–675 and aa1–174 were critical for binding to F-actin and myosin II, respectively (supplementary material Fig. S5B,C). MCF10A cells that constitutively expressed either siRNA-resistant GFP–ΔAct-SVIL (deleted of aa346–675) or GFP–ΔMyo-SVIL (deleted of aa1–174) were generated (Fig. 6E), and cell division was observed in the absence of endogenous SVIL. Although GFP–ΔAct-SVIL-expressing cells showed normal ingression that was similar to control cells, the broadening of the furrow was clearly observed in GFP–ΔMyo-SVIL cells (Fig. 6F). These results indicate that PLK1-mediated phosphorylation and interaction with myosin II is required for SVIL to confine the furrow in the restricted zone at the division plane.
SVIL controls activation of myosin II at the cleavage furrow
Previous studies have reported that the association of the N-terminus of SVIL with myosin II can induce the phosphorylation of RLC at Ser19 to activate myosin II for contraction (Takizawa et al., 2007). Activation of myosin II at the cleavage furrow is required for the promotion of constriction of the contractile ring (Asano et al., 2009; Dean and Spudich, 2006; Komatsu et al., 2000); thus, we speculated that the aberrant furrowing in SVIL-knockdown cells was associated with reduced activity of myosin II. We first examined global myosin II activation during cytokinesis by immunoblot using an antibody specific for phosphorylated RLC at Ser19. HeLa cells transfected with siRNAs were nocodazole-arrested, released and lysed at different time points. Phosphorylation of RLC was increased 1 hour after the release and declined with the progression of mitosis in control cells. In contrast, RLC phosphorylation was delayed and prolonged in SVIL-depleted cells (Fig. 7A), which suggests that SVIL is associated with RLC phosphorylation during cytokinesis. We then performed immunofluorescence analysis to examine whether SVIL depletion affected the activation of myosin II at the cleavage furrow in anaphase. Accumulation of phosphorylated RLC at the cleavage furrow was clearly diminished in SVIL-knockdown cells compared to the control cells (Fig. 7B). Localization of other components of the contractile ring, such as F-actin and Anillin, was not affected by the SVIL depletion (Fig. 7B). We next tested if SVIL association with myosin II was required for the myosin II activation at the cleavage furrow. HeLa cells that constitutively expressed siRNA-resistant GFP-tagged ΔAct-SVIL and ΔMyo-SVIL were established, and endogenous SVIL was suppressed by siRNA transfection. Both GFP–ΔMyo-SVIL and GFP–ΔAct-SVIL localized to the central spindle similar to wild-type SVIL (Fig. 7C). Localization of phosphorylated RLC was suppressed in GFP–ΔMyo-SVIL-expressing cells but not in GFP–ΔAct-SVIL-expressing cells (Fig. 7C). Finally, we determined whether the phosphorylation of SVIL by PLK1 was required for myosin II activation. HeLa cells that expressed siRNA-resistant GFP-tagged wild-type, S238A- and S238D-SVIL were depleted of endogenous SVIL by siRNA transfection and immunostained for phosphorylated RLC. Although both S238A- and S238D-SVIL could associate with F-actin and myosin II (supplementary material Fig. S6A,B), the accumulation of phosphorylated RLC was clearly diminished in S238A cells compared to wild-type and S238D-expressing cells (Fig. 7D). We also examined phosphorylation of RLC during cytokinesis in S238A cells by immunoblot analysis. Similar to the SVIL-knockdown cells (Fig. 7A), phosphorylation of RLC in S238A cells was delayed and prolonged (supplementary material Fig. S7). These results suggest that SVIL localized at the central spindle by PLK1 phosphorylation promotes the accumulation of active myosin II at the equatorial cortex to coordinate furrowing (Fig. 7E).
PLK1 is dynamically localized to the centrosomes, kinetochores and central spindle and phosphorylates various proteins at these distinct subcellular structures to ensure the accurate progression of cell division. Recent studies using small chemical PLK1 inhibitors revealed that PLK1 activity is not only essential for the initiation of furrowing but also for the continuous furrow ingression once constriction has started (Santamaria et al., 2007). PLK1-mediated phosphorylation of MgcRacGAP induces RhoA activation in the initiation of furrowing, but how PLK1 maintains ingression remains elusive. Recent study showed the large scaffold protein, SVIL, was required for the ingression of the furrow for the completion of cytokinesis (Smith et al., 2010). In this study, we presented a novel regulatory mechanism for the ingression of the furrow that was mediated by PLK1 phosphorylation of SVIL. Phosphorylation of SVIL at Ser238 by PLK1 was required for the accumulation of SVIL at the central spindle. Exogenous expression of nonphosphorylatable SVIL suppressed the activation of myosin II at the equatorial cortex and induced significant broadening of the cleavage furrow. Our results indicate that PLK1 utilizes SVIL for the activation of myosin II and for the confinement of the furrow in the restricted zone at the division plane for the completion of furrow ingression.
The coordinated action of the central spindle and the contractile ring is essential for the progression of furrowing to complete cytokinesis. In the absence of the central spindle organization, accumulation of RhoA at the equatorial cortex is disrupted, and the cells are no longer able to assemble the contractile ring to initiate furrowing. An organized central spindle is also required to maintain the furrowing because the depolymerization of microtubules during ingression induces delocalization of the furrow components (Straight et al., 2003). Thus, continuous communication of the central spindle and the contractile ring is crucial for the progression of furrowing. Recent studies have revealed an important molecular link between these two subcellular structures in Drosophila cells. RacGAP50C, a Drosophila homolog of MgcRacGAP, associated with Anillin, a critical component of the contractile ring, during cytokinesis (D'Avino et al., 2008; Gregory et al., 2008). The association of these two proteins is crucial for the completion of furrowing. We showed that the cells expressing ΔMyo-SVIL, which is defective in association with myosin II, exhibited significant broadening of the furrow and the suppression of myosin II activation. This result suggests that SVIL association with myosin II is also a molecular link between the central spindle and contractile ring to coordinate proper furrowing. A number of cytokinesis-related proteins have been reported to associate with SVIL. For example, SVIL can associate with the spindle-localized protein, KIF14, and the cortex-localized protein, EPLIN (Smith et al., 2010). Although further analysis is needed, SVIL may play a critical role as a hub protein to connect multiple components to coordinate cytokinesis progression.
Phosphorylation of RLC is a prerequisite for the activation of myosin II and essential for the execution of cytokinesis. Previous studies reported that the inhibition of myosin II activation by exogenous expression of non-phosphorylatable RLC induced a delay or distorted furrowing (Asano et al., 2009; Komatsu et al., 2000). PLK1 has been reported to contribute to the phosphorylation of RLC. Rho-associated kinase (ROCK) phosphorylates RLC to initiate constriction of the contractile ring, and PLK1 can phosphorylate ROCK to promote catalytic activity during cytokinesis (Lowery et al., 2007). A recent study showed that PLK1 phosphorylates the energy-sensing AMP-activated protein kinase (AMPK) to promote phosphorylation of RLC at the furrow (Vazquez-Martin et al., 2012). In addition to these pathways, our study showed that PLK1 can regulate myosin II activation by phosphorylating SVIL. These results demonstrate that PLK1 utilizes redundant pathways to promote myosin II activation for the constriction of the cleavage furrow. Currently, we do not know how the phosphorylated SVIL induces myosin II activation during cytokinesis. RhoA at the equatorial cortex activates protein kinases such as ROCK and Citron kinases for myosin II activation. SVIL at the central spindle may promote activation of RhoA-mediated pathways for RLC phosphorylation.
In summary, we showed that PLK1 phosphorylates SVIL at Ser238 to promote its association with PRC1 for localization at the central spindle. Once at the central spindle, SVIL association with myosin II induces the activation of myosin II at the cleavage furrow and confines the furrow in the limited region. SVIL can associate with additional mitotic spindle- and contractile ring-localized proteins; thus, elucidation of the physiological role of SVIL will provide greater insight into the molecular mechanisms involved in cytokinesis progression.
Materials and Methods
Cells, antibodies and chemicals
HeLa cells were cultured in DMEM (WAKO, Osaka Japan) and supplemented with 10% fetal bovine serum (EQUITEC, Hendra, Australia). The human mammary epithelial cells, MCF10A, were cultured in DMEM F12 (WAKO) supplemented with 5% horse serum (GIBCO, Carlsbad, CA). To generate an anti-SVIL antibody, aa99–160 of SVIL fused with GST was produced in bacteria, and recombinant protein was purified using glutathione agarose beads (Sigma-Aldrich, St. Louis, MO). The protein was mixed with Freund's adjuvant (Sigma-Aldrich) and injected into a rabbit four times every 2 weeks. To purify the anti-SVIL antibody, we used HiTrap NHS-activated HP columns (GE Healthcare BioScience, Uppsala, Sweden) coupled with recombinant GST-SVIL (aa99–160). Other antibodies were purchased from the following manufacturers: anti-α-tubulin antibodies, Sigma-Aldrich; anti-PLK1 antibody, Cell Signaling (Danvers, MA); anti-PRC1 antibodies, Santa Cruz (Santa Cruz, CA); anti-GFP antibody, Nacalai Tesque (Tokyo, Japan). Rhodamine-conjugated phalloidin was obtained from Invitrogen (Carlsbad, CA). PLK1 inhibitor BI2536 was purchased from Selleck Chemicals (Houston, TX), CDK1 inhibitor roscovitine from Santa Cruz, MG132 from Enzo Life Sciences (Farmingdale, NY) and nocodazole and thymidine from Sigma-Aldrich.
Human SVIL was amplified by PCR from a HeLa cDNA library. Full-length SVIL was cloned into a pQCXIP retrovirus vector with a GFP or HA tag on the N-terminus (Clontech, Mountain View, CA). Point mutations were introduced by PCR-based site-directed mutagenesis.
A library of siRNAs was purchased from Invitrogen. The name of the genes and sequences of siRNAs for each gene are shown in supplementary material Table S1. HeLa cells were cultured in 24-well plates and transfected with 20 nM of each siRNA with Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. Seventy-two hours later, the cells were fixed with 4% paraformaldehyde and stained with FITC-labeled paclitaxel (Invitrogen) and Hoechst for visualization. Images were taken using an IX71 fluorescence microscope (Olympus, Tokyo, Japan). The ratio of multinucleated cells was evaluated from five randomly selected fields.
The sequences of the siRNAs used to suppress SVIL expression were 5′-GAGAACAAGGGAAUGUUGAGAGAAU-3′ (siRNA-1) and 5′-GGAGGUGAUGAAGCCAGAUGAUGAU-3′ (siRNA-2). The sequence of the control siRNA targeting luciferase was 5′-CUUACGCUGAGUACUUCGATT-3′. The cells were transfected with 20 nM siRNA using Lipofectamine RNAiMAX (Invitrogen).
Generation of stable cell lines
HeLa and MCF10A cells that constitutively expressed each protein were established by retrovirus infection. HEK293T cells were transfected with the pQCXIP retroviral vector that encoded each cDNA in combination with the pVPack-GP and pVPack-Ampho vectors (Stratagene, La Jolla, CA) using Lipofectamine 2000 (Invitrogen). Forty-eight hours after transfection, the supernatants were added to the cells with 2 µg/ml polybrene (Sigma-Aldrich), and the infected cells were selected with 1 µg/ml puromycin for 3 days.
To synchronize cells by double thymidine block, the cells were incubated with 2 mM thymidine for 24 hours, washed three times with PBS and released into fresh medium for 8 hours. The cells were again incubated in the presence of 2 mM of thymidine for 15 hours, released into fresh medium and lysed at the different time points. For mitotic synchronization, the cells were treated with 40 ng/ml nocodazole for 13 hours. Mitotic arrested cells were collected by shaking and were subsequently released into fresh medium.
The cells grown on 35 mm glass-bottom dishes (IWAKI, Tokyo, Japan) were transfected with siRNAs using Lipofectamine RNAiMAX (Invitrogen). Twenty-four hours after transfection, the cells were placed on the heat stage (37°C) of an IX81 microscope (Olympus). Phase-contrast and fluorescence images of live cells were collected at the indicated intervals and processed using Meta imaging software v6.1 (Molecular Devices, Sunnyvale, CA).
The cells were washed once with cold PBS and lysed in TNE buffer (50 mM Tris-HCl PH 7.4, 150 mM NaCl, 0.1% NP-40, 1 mM PMSF) for 15 minutes on ice and centrifuged at 15,000 r.p.m. for 20 minutes to obtain clear cell lysates. The cell lysates were mixed with the indicated primary antibodies for 60 minutes at 4°C and then with protein-A– or protein-G−Sepharose beads for an additional 60 minutes. The beads were washed three times with TNE buffer and suspended in sample buffer.
In vitro translation and GST pull-down assay
HA-tagged aa676–1009 SVIL was in vitro translated using TNT SP6 Quick coupled Transcription/Translation system (Promega, Madison, WI) following the manufacturer's protocol. Five microliters of translated protein was incubated with 10 µg of GST protein or GST fusion protein for 1 hour at 4°C. The beads were subjected to SDS–PAGE and immunoblotting using anti-HA antibodies.
Nocodazole-arrested and released cells were seeded on the poly-D-lysine-coated glass coverslips and fixed with cold methanol/acetone (50:50) or 4% paraformaldehyde. The cells were blocked with 7% FBS in PBS for 30 minutes and then incubated with primary antibodies. After washing with PBS, the cells were incubated with FITC-conjugated anti-rabbit antibody (Invitrogen) or Rhodamine-conjugated anti-mouse antibody (Invitrogen). Images were acquired using a laser scanning confocal microscope FV1000 (OLYMPUS, Tokyo, Japan).
In vitro kinase assay
Purified GST or GST-fused aa228–256 SVIL (3 µg) was incubated with 0.5 µg of recombinant active-PLK1 (Millipore, Billerica, MA) in kinase buffer (50 mM Hepes pH 7.5, 10 mM MgCl2, 1 mM EGTA, 1 mM DTT, 250 µM ATP) and 1 µCi of [γ-32P]ATP. The reaction mixtures were incubated at 30°C for 30 minutes in 50 µl and resolved on SDS-PAGE gels. The gels were dried and subjected to autoradiography.
We thank the members of the Division of Cancer Biology for helpful discussions and technical assistance.
H.H. designed and performed experiments. T.H., E.A. and H.K. helped to perform siRNA screen. S.I., M.M., A.N., T.W. and M.H. helped to interpret the data. T.S. supervised the project. All authors contributed to writing the manuscript.
This research was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan [grant numbers 22570182 to T.S., 23791591 to H.H. and 23107010 to A.N.].