Mitosis-specific activation of LIM motif-containing protein kinase and roles of cofilin phosphorylation and dephosphorylation in mitosis.

Actin filament dynamics play a critical role in mitosis and cytokinesis. LIM motif-containing protein kinase 1 (LIMK1) regulates actin reorganization by phosphorylating and inactivating cofilin, an actin-depolymerizing and -severing protein. To examine the role of LIMK1 and cofilin during the cell cycle, we measured cell cycle-associated changes in the kinase activity of LIMK1 and in the level of cofilin phosphorylation. Using synchronized HeLa cells, we found that LIMK1 became hyperphosphorylated and activated in prometaphase and metaphase, then gradually returned to the basal level as cells entered into telophase and cytokinesis. Although Rho-associated kinase and p21-activated protein kinase phosphorylate and activate LIMK1, they are not likely to be involved in mitosis-specific activation and phosphorylation of LIMK1. Immunoblot and immunofluorescence analyses using an anti-phosphocofilin-specific antibody revealed that the level of cofilin phosphorylation, similar to levels of LIMK1 activity, increased during prometaphase and metaphase then gradually declined in telophase and cytokinesis. Ectopic expression of LIMK1 increased the level of cofilin phosphorylation throughout the cell cycle and induced the formation of multinucleate cells. These results suggest that LIMK1 is involved principally in control of mitosis-specific cofilin phosphorylation and that dephosphorylation and reactivation of cofilin at later stages of mitosis play a critical role in cytokinesis of mammalian cells.

During the cell division cycle, the morphology of cells is altered. In animal cells in culture, flat and adherent cells in interphase become spherical and weakly adherent in mitosis. At later stages of mitosis, cortical actin filaments are reorganized and recruited to the cleavage furrow, and an actomyosinbased contractile ring is formed around the equator of dividing cells, constricted to lead to cell cleavage, and disappears at the end of cytokinesis. Actin filament dynamics, reorganization, and redistribution play a principal role in these processes (1)(2)(3). Previous studies revealed the involvement of various actin-binding proteins and signaling proteins that regulate actin filaments in mitotic processes and cytokinesis (1)(2)(3), but mechanisms by which cells coordinately change shapes in response to cell cycle cues remain largely uncharacterized.
In addition to functions in actin cytoskeletal remodeling in interphase cells, cofilin/ADF seems to be involved in processes related to mitosis and cytokinesis. Cofilin/ADF is concentrated at the cleavage furrow and the midbody during cytokinesis of cultured mammalian cells (23) and cleavage of Xenopus fertilized eggs (24). Injection of the antibody that inhibits the activity of cofilin/ADF into Xenopus blastomeres blocks the cleavage of the blastomeres (24). In addition, Drosophila twinstar mutants, in which expression of the twinstar gene encoding a Drosophila cofilin/ADF ortholog is repressed, exhibit frequent failures in cytokinesis in larval neuroblasts and in testicular meiotic cells (25). These observations suggest that cofilin/ADF plays a critical role in cytokinesis. Furthermore, injection into Xenopus blastomeres of a nonphosphorylatable (constitutively active) form of cofilin/ADF blocks cytokinesis, but injection of wild-type cofilin/ADF that can be phosphorylated has no ap-parent effect (24), thus indicating that excess activity of cofilin/ ADF also prevents cytokinesis, and the proper control of cofilin/ ADF activity by phosphorylation and dephosphorylation is important for progression of mitosis. We thus assumed that LIM kinases play a role in mitosis and cytokinesis by phosphorylating and regulating the activity of cofilin/ADF.
We have now examined changes in the kinase activity of LIMK1 and the level of cofilin phosphorylation during the cell cycle, and we found that LIMK1 becomes activated and hyperphosphorylated in the early stages of mitosis and then gradually reverts to basal levels in late stages. Mitotic activation of LIMK1 is induced by a mechanism distinct from that of Thr-508 phosphorylation by ROCK or PAK. We also noted cell cycle-associated changes in the level of cofilin phosphorylation which are similar to those seen with LIMK1 activity. Ectopic expression of LIMK1 induces increases in the level of cofilin phosphorylation throughout the cell cycle and leads to the formation of multinucleate cells. We propose that LIMK1 plays an important role in regulating cofilin/ADF activity during mitosis and that cofilin dephosphorylation in the late stages of mitosis is critical for cytokinesis.

EXPERIMENTAL PROCEDURES
Plasmid Construction-To generate a pMYC-C1 vector containing the Myc epitope sequence (EQKLISEEDL), a pEYFP-C1 vector (CLON-TECH, Palo Alto, CA) was digested with NheI and BglII, and the oligonucleotides coding for Myc epitope peptide were inserted in place of the cDNA for yellow fluorescent protein (YFP). Expression plasmids coding for N-terminally Myc-tagged LIMK1 and LIMK2 were constructed by inserting full-length human LIMK1 and LIMK2 cDNAs (9,10), respectively, into the BglII and SacII sites of the pMYC-C1 vector. Plasmids coding for Myc-LIMK1(D460A) and Myc-LIMK1(T508V), in which Asp-460 and The-508 in Myc-LIMK1 were replaced by Ala and Val, respectively, were constructed, using a site-directed mutagenesis kit (CLONTECH). To construct the plasmid coding for Myc-LIM-K1(PK) containing amino acid residues 267-647, the cDNA fragment was amplified by PCR, using primers 5Ј-CTAGCTCGCCACCATG-GGATACCCATACGATGTTCCAGATTACGCTGGATCCAGATCTG-GGCCTGAGACCAGCCCC-3Ј and 5Ј-TCCCGCGGAGGAATCTGG-3Ј. The PCR-amplified fragments were digested with BglII and SacII and ligated into the BglII and SacII sites of the pMYC-C1 vector. The plasmid coding for LIMK1-YFP was constructed by doubly inserting the PCRamplified YFP cDNA into the XbaI and SalI sites of pMYC-C1-Myc-LIMK1. To construct the plasmid coding for PAK-AI (an autoinhibitory domain of PAK3, amino acid residues 78 -146), the cDNA fragment was PCR amplified, using primers 5Ј-CAGAGCGGCCGCATACGATTCATGT-GGGGTTT-3Ј and 5Ј-GTATGCGGCCGCACTTTTATCTCCTGATGTAA-3Ј, and PAK3 cDNA as a template (provided by Dr. H. Sumimoto, Kyushu University, Fukuoka, Japan), digested with NotI, and subcloned into the NotI site of pCAG vector (17). The authenticity of plasmids was confirmed by nucleotide sequence analysis.
Cell Culture, Synchronization, and Transfection-HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS). Cells were synchronized at beginning of the S phase using the double thymidine block method (26,27). In brief, cells were cultured in DMEM containing 10% FCS and 1 mM thymidine for 24 h, incubated in fresh DMEM containing 10% FCS for 8 h, and again cultured in thymidine-containing medium for 16 h. For cell cycle progression, cells synchronized at the beginning of S phase were then released for 6 -12 h in fresh DMEM containing 10% FCS. To synchronize cells at mitosis, cells were synchronized at the beginning of S phase by a single thymidine block and cultured in fresh DMEM containing 10% FCS for 6 h and then cultured for 6 h in the presence of 100 ng/ml nocodazole. For transient transfection experiments, HeLa cells were transfected using the Li-pofectAMINE method (Invitrogen), following the manufacturer's protocol, cultured in the thymidine-containing medium for 24 h, and released for 6 h in fresh DMEM containing 10% FCS. Nocodazole, a microtubule-depolymerizing agent, was added to the medium, and the culture was continued for another 6 h. Mitotic cells that rounded up and adhered weakly on the culture dish were collected selectively by the mechanical shake-off procedure, whereas the attached interphase cells were harvested by scraping them off. To prepare cells synchronized at later stages of mitosis, mitotic cells collected by nocodazole arrest were washed with phosphate-buffered saline (PBS) to remove nocodazole, suspended in DMEM containing 10% FCS, plated into poly-L-lysine-coated culture dishes or coverslips then in-FIG. 1. Mitosis-specific activation and phosphorylation of LIMK1. A, HeLa cells were synchronized at the beginning of S phase by double thymidine block and released for 6, 9, and 12 h. Cell cycle progression was analyzed by flow cytometry. B, HeLa cells were synchronized as in A. At the indicated times after release, cells were lysed, and endogenous LIMK1 was immunoprecipitated and subjected to in vitro kinase reaction, using His 6 -cofilin as a substrate. The reaction mixture was separated on 15 and 8% SDS-PAGE and transferred onto polyvinylidene difluoride membranes. Membrane from the 15% gel was analyzed using autoradiography to detect 32 P incorporation into cofilin (top panel) and by Amido Black staining (second panel from top). Membrane from the 8% gel was analyzed by immunoblotting with C-10 anti-LIMK1 antibody (third panel). The relative kinase activity of LIMK1 is shown as the mean Ϯ S.E. of three independent experiments, with the kinase activity of LIMK1 in cells at 6 h taken as 1.0 (bottom panel). The mean value of the activity is indicated under the bottom panel. C, HeLa cells were synchronized at early stages of mitosis by nocodazole treatment. Interphase (I) and mitotic (M) cells were selectively collected. The kinase activity and gel mobility of LIMK1 in interphase and mitotic cells were analyzed as in B. D, effects of phosphatase treatment on kinase activity and gel mobility of mitotic LIMK1. LIMK1 immunoprecipitated from lysates of interphase and mitotic cells was incubated with (ϩ) or without (Ϫ) calf intestinal alkaline phosphatase (CIP), and the kinase activity and gel mobility of LIMK1 were analyzed as in B. In C and D, the bottom panel shows the relative kinase activity of LIMK1, with the activity of interphase LIMK1 without phosphatase treatment taken as 1.0. cubated at 37°C to allow for cell cycle progression. At 0, 45, 90, 180 min after release from nocodazole arrest, the cells were harvested and subjected to analyses. For flow cytometry, cells were trypsinized, fixed with 70% methanol, stained with propidium iodide, and analyzed by a flow cytometer EPICS XL-MCL (Beckman Coulter). COS-7 cells were cultured in DMEM supplemented with 10% FCS. Cells were transiently transfected using LipofectAMINE. Three h after transfection, the medium was changed to normal cultivation medium. After an additional 36 h culture, cells were harvested and subjected to analyses.
Antibodies-Rabbit anti-LIMK1 antibody (C-10) was raised against the C-terminal peptide of LIMK1, as described (10). An anti-cofilin antibody (COF-1) was prepared by immunizing rabbits with His 6 -cofilin, which was purified from lysates of Escherichia coli expressing it, using as nickel-nitrilotriacetic acid-agarose (Qiagen, Valencia, CA). Rabbit anti-P-cofilin antibody specific to the Ser-3-phosphorylated forms of cofilin and ADF was raised against the phosphopeptide acetyl-A(pS)GVAVSDC and purified, as described (28). Anti-Myc epitope monoclonal antibody (9E10) and anti-HA epitope monoclonal antibody (12CA5) were purchased from Roche Molecular Biochemicals.
Immunoprecipitation-Cells were washed three times with ice-cold PBS, suspended in the lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 1 mM dithiothreitol, 10 mM NaF, 20 mM ␤-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 2 g/ml pepstatin A), and incubated on ice for 30 min. After centrifugation, protein concentrations in cell lysates were determined in a Micro BCA protein assay (Pierce), and equal amounts of protein were precleared with protein A-Sepharose (Amersham Biosciences) at 4°C for 1 h. The supernatants were incubated overnight at 4°C with anti-LIMK1 (C-10) antibody or 9E10 anti-Myc antibody, and protein A-Sepharose. After centrifugation, the immunoprecipitates were washed three times with lysis buffer and used for in vitro kinase reaction and immunoblot analysis.
Immunoblot Analysis-For immunoblot analysis, cell lysates or immunoprecipitated proteins were separated on SDS-PAGE and transferred onto polyvinylidene difluoride membranes (Bio-Rad). The membrane was blocked overnight with 4% nonfat dry milk in PBS containing 0.05% Tween 20 and incubated for 2 h at room temperature with primary antibody diluted in PBS containing 1% nonfat dry milk and 0.05% Tween 20. After washing in PBS containing 0.05% Tween 20, the membrane was incubated with horseradish peroxidase-conjugated donkey anti-rabbit IgG or sheep anti-mouse IgG (Amersham Biosciences). Immunoreactive protein bands were visualized using an ECL reagent (Amersham Biosciences).
In Vitro Kinase Assay-The immunoprecipitates were washed three times with lysis buffer and then three times with kinase buffer (20 mM Hepes-NaOH, pH 7.2, 5 mM MgCl 2 , 5 mM MnCl 2 , 1 mM dithiothreitol, 10 mM NaF, 20 mM ␤-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 2 g/ml pepstatin A) and incubated for 1 h at 30°C in 30 l of kinase buffer containing 50 M ATP, 5 Ci of [␥-32 P]ATP (3,000 Ci/mmol, Amersham Biosciences) and 2 g of His 6cofilin. The reaction mixture was solubilized in Laemmli's sample buffer (50 mM Tris-HCl, pH 6.8, 10% glycerol, 1 mM dithiothreitol, 1% SDS, 0.002% bromphenol blue) for 5 min at 95°C, and aliquots were separated on SDS-PAGE, using 15 and 8% gels. Proteins were transferred onto polyvinylidene difluoride membranes. The membrane from a 15% gel was subjected to autoradiography to measure 32 P-labeled cofilin, using the BAS1800 Bio-Image Analyzer (Fuji Film, Tokyo, Japan), and Amido Black staining. The membrane from the 8% gel was analyzed by immunoblotting with the C-10 anti-LIMK1 antibody or the 9E10 anti-Myc antibody to detect LIMK. The kinase activity was normalized by dividing the radioactivity incorporated into cofilin by the immunoreactive density of LIMK estimated using a densitometer.
Phosphatase Treatment-Anti-LIMK1 immunoprecipitates were washed three times with lysis buffer without phosphatase inhibitors and then three times with phosphatase buffer (1 mM MgSO 4 and 100 mM Tris-HCl, pH 8.0) and treated for 1 h at 25°C with 20 units of calf intestinal alkaline phosphatase (Takara Biochemicals, Tokyo) in 30 l of phosphatase buffer, then the immunoprecipitates were washed three times with lysis buffer containing phosphatase inhibitors and three times with kinase buffer and subjected to in vitro kinase reaction, as described above.
Cell Staining-HeLa cells were fixed in 4% paraformaldehyde in PBS for 20 min and permeabilized with absolute methanol for 10 min at Ϫ20°C. After blocking with 1% bovine serum albumin in PBS for 30 min, cells were stained with anti-P-cofilin antibody followed by staining with rhodamine-conjugated anti-rabbit IgG (Chemicon). 4,6-Diamidino-2-phenylindole (DAPI, Molecular Probes) was used for DNA staining. After washing with PBS, coverslips were mounted on a glass slide and images were obtained using a Leica DMLB fluorescence microscope.

Mitosis-specific Activation and Phosphorylation of LIMK1-
To investigate the role of LIMK1 during the cell cycle, we first examined the kinase activity of LIMK1 in different stages of the cell cycle. For cell cycle analysis, HeLa cells were synchronized at the beginning of S phase by double thymidine block. At different times after release from double thymidine block, cells were lysed, and endogenous LIMK1 was immunoprecipitated and subjected to in vitro kinase assay, using His 6 -cofilin as a substrate. Cell cycle progression was monitored by flow cytometry (Fig. 1A) and DAPI staining (data not shown). As shown in Fig. 1B, LIMK1 prepared from cells at 9 h after release from the thymidine block, when about 40% of the cells were in mitotic phase (as measured by DAPI staining), had 2.6-fold higher kinase activity than that of LIMK1 from cells at 6 h after release, when about 90% of the cells were in S phase (as determined using flow cytometry). At 12 h after release, about 80% of the cells had gone through cytokinesis and entered the G 1 phase, and here the kinase activity of LIMK1 reverted to the level seen in LIMK1 in cells at 6 h. Immunoblot analysis revealed that slow migrating bands of LIMK1 appeared at 9 h, although the total amount of LIMK1 remained unchanged throughout the cell cycle.
The gel migration shift and activation of LIMK1 became prominent when cells were more stringently synchronized at the early mitotic phase by treatment with the microtubuledepolymerizing agent, nocodazole (Fig. 1C). LIMK1 from cells in the early mitotic phase had a 5.8-fold higher kinase activity than in the interphase. Mitotic LIMK1 significantly retarded mobility on gel electrophoresis compared with interphase LIMK1 (Fig. 1C). To determine whether the activation and mobility shift of mitotic LIMK1 were related to the phosphorylation, LIMK1 was treated with calf intestinal alkaline phosphatase. This treatment abrogated the mobility shift and activation of mitotic LIMK1 (Fig. 1D). Effects of phosphatase were nil when phosphatase inhibitors were present (data not shown). These results suggest that the slow migrating bands are the phosphorylated forms of LIMK1 and that LIMK1 is specifically activated and phosphorylated in the mitotic phase.
Kinase Activity and Gel Mobility Shift of Kinase-inactive and N-terminally Truncated LIMK1 Mutants and LIMK2 in Mitosis-To search for the mechanism of mitosis-specific activation and gel mobility shift of LIMK1, Myc-tagged wild-type LIMK1 and its kinase-inactive mutant LIMK1(D460A), in which the catalytic residue Asp-460 is replaced by alanine, were transiently expressed in HeLa cells, and these cells were synchronized at the mitotic phase, using nocodazole. As shown in Fig.  2A, the transiently expressed Myc-LIMK1 in mitotic cells showed a gel mobility shift and a 6.4-fold increase in the kinase activity compared with Myc-LIMK1 in interphase cells. Thus, the transiently expressed Myc-LIMK1 behaves similarly to endogenous LIMK1. In both interphase and mitotic cells, Myc-LIMK1(D460A) immunoprecipitated from HeLa cells exhibited no cofilin phosphorylating activity, which suggests that Myc-LIMK1 immunoprecipitates were probably not contaminated with other cofilin-phosphorylating kinases. Similar to Myc-LIMK1, Myc-LIMK1(D460A) was retarded on gel electrophoresis, which means that autophosphorylation of LIMK1 may not be involved in mobility shift of mitotic LIMK1. Myc-LIMK2 transiently expressed in HeLa cells exhibited a 1.8-fold increase in the kinase activity in mitotic phase, but the gel mobility shift of Myc-LIMK2 was not visible after cells had been treated with nocodazole (Fig. 2B). Thus, it is likely that LIMK2 is at least in part involved in cofilin phosphorylation during mitosis.
To determine the region involved in the mitosis-specific phosphorylation and activation of LIMK1, an expression plasmid coding for an N-terminally truncated mutant of LIMK1, Myc-LIMK1(PK) (Fig. 2C), was constructed and transfected into HeLa cells. As reported (29), in interphase cells, LIMK1(PK) exhibited significant increases in kinase activity compared with the activity of wild-type LIMK1 (Fig. 2D). Although the LIMK1(PK) mutant was activated further by ROCK in in vitro experiments (17), it was not activated further in mitotic phase, and a gel mobility shift of mitotic LIMK1(PK) mutant was never observed (Fig. 2D). Thus, the N-terminal region of LIMK1 containing LIM and PDZ domains may be involved in the mitosis-specific mobility shift (phosphorylation) and activation of LIMK1. We attempted to show the gel mobility shift of the N-terminal fragment of LIMK1 in mitosis, but it was not successful because the protein band of the fragment expressed in HeLa cells was not detectable in mitotic cell lysates, probably because of its instability.
Roles of ROCK and PAK in Mitotic Activation of LIMK1-Previous studies revealed that LIMK1 is activated by protein kinases, ROCK and PAK, through phosphorylation at Thr-508 within the kinase domain of LIMK1 (15)(16)(17). We therefore asked whether ROCK and/or PAK is involved in the activation and mobility shift of LIMK1 in mitosis. To examine the role of ROCK in this regard, HeLa cells were treated with Y-27632, a specific inhibitor of ROCK (30), and the activity and gel mobility of mitotic LIMK1 were analyzed using in vitro kinase assay and immunoblotting. Gel retardation of mitotic LIMK1 was observed in the presence of Y-27632 (Fig. 3A). In vitro kinase assay revealed that the kinase activity of mitotic LIMK1 was only faintly reduced by Y-27632 treatment (Fig. 3A). Thus, a protein kinase(s) other than ROCK seems to be involved in the mitosis-specific mobility shift and activation of LIMK1. We next examined the role of PAK in mitosis-specific activation and mobility shift of LIMK1. As an inhibitor for PAK, we used an autoinhibitory domain of PAK (PAK-AI). As reported (15), expression of PAK-AI inhibited the activation of LIMK1 induced by RacV12, a constitutively active form of Rac (Fig. 3B), thereby indicating that PAK-AI has the potential to inhibit PAK activity in cultured cells. Because the coexpression of PAK-AI and LIMK1 into HeLa cells did not interfere with mobility shift or activation of LIMK1 in mitosis (Fig. 3C), PAK does not appear to have a role in this regard.
To determine further whether phosphorylation of Thr-508 by ROCK, PAK, or other protein kinases is involved in the mitotic mobility shift and activation of LIMK1, we expressed in HeLa cells LIMK1(T508V), in which Thr-508 is replaced by nonphosphorylatable valine, and analyzed its gel mobility and kinase activity in mitotic phase. Similar to wild-type LIMK1, LIMK1(T508V) in mitosis migrated slowly on the gel (Fig. 3D), thereby clearly indicating that phosphorylation at Thr-508 is not related to the mobility shift of mitotic LIMK1 and that phosphorylation of residue(s) other than Thr-508 is responsible for the gel retardation. As reported (31), replacing Thr-508 by Val significantly reduced the kinase activity of LIMK1 (Fig.  3D). However, kinase activity of the LIMK1(T508V) mutant increased about 3-fold in mitotic phase compared with that in interphase (Fig. 3D), which further suggests that the mitotic activation of LIMK1 is caused primarily by mechanisms other than Thr-508 phosphorylation. LIMK1 Is Activated in Early Stages of Mitosis-To examine further the role of LIMK1 in the mitotic phase, we analyzed changes in the kinase activity of LIMK1 during mitosis. Nocodazole-arrested mitotic cells were collected and released for 0, 45, 90, and 180 min to allow for cell cycle progression. This progression and cell synchronization were monitored by immunofluorescence microscopy for tubulin and DNA. Analysis of cell morphology and immunostaining indicated that almost all of the nocodazole-arrested cells (at zero time) were in prometaphase. At 45 min after release from nocodazole 70, 20, and 10% of the cells were in metaphase, prometaphase, and anaphase, respectively; at 90 min, 90 and 5% of the cells in telophase and anaphase, respectively; and at 180 min, 95% were in G 1 phase (data not shown). LIMK1 was immunoprecipitated from the cell populations at different stages of mitosis and subjected to immunoblot analysis and in vitro kinase assay (Fig. 4). As described above (Fig. 1B), the kinase activity of LIMK1 at 0 min increased about 5.3-fold compared with that of interphase LIMK1. At 45 min after nocodazole release, the kinase activity of LIMK1 was retained at a high level (about 4.8-fold higher than that of interphase LIMK1). LIMK1 activity declined to about 50% of the maximum level at 90 min after release and reverted to the basal level at 180 min. Immunoblot analysis revealed that slow migrating forms of LIMK1 were observed at 0 and 45 min after nocodazole release and continued for 90 min. The band of LIMK1 at 90 min was located halfway between the bands of interphase LIMK1 and mitotic LIMK1, which suggests that LIMK1 is phosphorylated on multiple sites in early stages of mitosis then gradually dephosphorylated in the late stage of mitosis. At 180 min, the gel mobility of LIMK1 was practically identical to that of interphase LIMK1. These results suggest that LIMK1 is phosphorylated on multiple sites and highly activated in early stages of mitosis (prometaphase and metaphase), then dephosphorylated, and reverts back to the basal activity in later stages of mitosis (telophase and cytokinesis).
Phosphorylation and Dephosphorylation of Cofilin during Mitosis-Because LIMK1 specifically phosphorylates cofilin at Ser-3, we next analyzed the level of cofilin phosphorylation during the cell cycle. The level of cofilin phosphorylation was first measured using immunoblots on two-dimensional gel electrophoresis and an anti-cofilin antibody. As shown in Fig. 5A, FIG. 5. Immunoblot analyses of changes in the level of cofilin phosphorylation during the cell cycle. HeLa cells were synchronized using nocodazole. Lysates from interphase cells (I) and cells released for 0, 45, 90, and 180 min after removal of nocodazole were separated by two-dimensional gel electrophoresis (A) or one-dimensional SDS-PAGE on a 15% gel (B). In A, gels were analyzed by immunoblotting with an anti-cofilin antibody. Arrowheads indicate the positions of P-cofilin and cofilin. The ratio of P-cofilin in total cofilin during the cell cycle is indicated in the bottom panel. In B, gels were analyzed by immunoblotting with an anti-P-cofilin antibody (top panel) or an anti-cofilin antibody (bottom panel). Anti-P-cofilin antibody specifically recognizes P-cofilin and P-ADF (28).

FIG. 6. Immunofluorescence analyses of cell cycle-associated changes in levels of cofilin phosphorylation. Asynchronized HeLa cells
were stained with an anti-P-cofilin antibody (top panels) and DAPI for DNA (bottom panels). Arrowheads indicate representative cells in prometaphase, metaphase, anaphase, and telophase, respectively, as monitored using DAPI staining. Scale bar, 10 m.  (first and third horizontal panels). Arrowheads indicate the cells expressing LIMK1-YFP or LIMK1(D460A)-YFP, in interphase, metaphase, telophase, and later telophase, respectively. Scale bar, 10 m. the Ser-3-phosphorylated cofilin (P-cofilin) comprises only a small percent of the total amount of cofilin in interphase HeLa cells. The ratio of P-cofilin increased in cells in the early stage of mitosis, reached the maximum level (55%) at 45 min after release from nocodazole arrest, and then gradually decreased as cells entered into telophase (at 90 min) and underwent cytokinesis to enter into the G 1 phase (at 180 min). Similar results were obtained using one-dimensional gel immunoblot analysis and an anti-P-cofilin antibody that specifically recognizes the P-cofilin and P-ADF (28) (Fig. 5B, top panel), in which the total expression level of cofilin remained unchanged throughout the cell cycle (Fig. 5B, bottom panel).
The level of cofilin phosphorylation during the cell cycle was examined further, using asynchronized HeLa cells immunostained with an anti-P-cofilin antibody. When cell cycle progression of asynchronized HeLa cells was monitored by cell morphology and DAPI staining, a significant increase in the level of cofilin phosphorylation was observed in cells at prometaphase, metaphase, and anaphase but not in cells at telophase compared with findings with surrounding interphase cells (Fig.  6). These results provide further evidence that cofilin becomes highly phosphorylated from prometaphase to metaphase then gradually dephosphorylated in the later stage of mitosis. P-cofilin was localized diffusely in the cytoplasm throughout mitosis (Fig. 6), whereas cofilin was concentrated in the cleavage furrow in the late stages of mitosis (23). Because the phosphorylation of Ser-3 inhibits the actin binding activity of cofilin, these results suggest that dephosphorylation of cofilin in the late stages of mitosis correlates with the accumulation of cofilin in the cleavage furrow in these stages.
Overexpression of LIMK1 Induces Multinucleate Cells-Significant changes in the level of cofilin phosphorylation during mitosis suggest a role for cofilin phosphorylation and dephosphorylation in cell division. To determine whether the dephosphorylation of cofilin in the later phase of mitosis is required for cytokinesis, we examined the effects of LIMK1 overexpression on cell division. For this we constructed an expression plasmid coding for LIMK1-YFP fusion protein and then transfected it into HeLa cells. Immunofluorescence analysis using an anti-Pcofilin antibody revealed that cells expressing LIMK1-YFP showed a significant increase in the level of cofilin phosphorylation throughout the cell cycle, including telophase and interphase, compared with findings in surrounding nontransfected cells (Fig. 7, top two horizontal panels). Thus, overexpression of LIMK1 facilitated increases in levels of cofilin phosphorylation in all stages of the cell cycle, including the later stages of mitosis. In contrast, in cells expressing kinase-inactive LIMK1(D460A)-YFP fusion protein (Fig. 7, bottom two lower panels) or control YFP (data not shown), the level of P-cofilin changed in the cell cycle-dependent manner, similar to changes seen in nontransfected cells (see Fig. 6), with the highest level of P-cofilin in metaphase and the low level in telophase and interphase. Expression of LIMK1(D460A)-YFP reduced the level of P-cofilin accumulation in metaphase (compare lower two panels in Fig. 7 with Fig. 6), but the inhibitory effect was incomplete, which may be attributed to the LIMK2 activity in mitosis.
When HeLa cells were transfected with plasmids coding for LIMK1-YFP and cultured for 48 h (a period sufficient for one or two cycles of cell division), then stained with DAPI, 28% of cells expressing LIMK1-YFP became multinucleate (Fig. 8, A and  B). Considering that this percentage was seen after one (or two) cycles of cell division, more than 44% (or 60%) of LIMK1-YFPexpressing cells failed to divide. In contrast, for cells expressing the LIMK1(D460A)-YFP or the YFP control vector, only a small percentage of cells had such multinucleate phenotypes (Fig. 8,  A and B). These results suggest that cofilin phosphorylation by ectopic expression of LIMK1 blocked cytokinesis and that cofilin dephosphorylation at the later stage of mitosis plays a critical role in cytokinesis.

DISCUSSION
Cofilin/ADF plays an essential role in the dynamics of actin filaments by depolymerizing and severing actin filaments (4 -6). In addition to the roles of cofilin/ADF in cell motility and morphological changes in interphase cells, it is also implicated in mitosis and cytokinesis (23)(24)(25). In Drosophila twinstar mutants, in which expression of a Drosophila cofilin/ADF gene is decreased, there are defects in aster migration/separation during prophase/prometaphase of meiotic divisions of spermatocytes as well as defects in cytokinesis in mitotic and meiotic cells; large actin aggregates are detected in association with centrosomes in mature primary spermatocytes, and during anaphase and telophase aberrantly large and misshaped actin structures appear at the site of contractile ring formation. These structures fail to disassemble at the end of telophase (25). These in vivo analyses suggest that cofilin/ADF plays important roles both in centrosome movement in early stages of mitosis and in contractile ring formation, constriction, and/or disassembly during later stages of mitosis/meiosis and cytokinesis. The role of cofilin/ADF in cytokinesis is also suggested by observations that cofilin/ADF accumulates at the contractile ring and midbody during later stages of mitosis in Xenopus eggs and in mammalian cells undergoing cytokinesis (23,24) and that inhibition of cofilin/ADF activity in Xenopus blastomeres by injection of an inhibitory antibody blocks cytokinesis yet without inhibiting nuclear division (24). In addition, the role of cofilin/ADF in centrosome migration and/or spindle formation in early stages of meiosis is suggested by the observation that cofilin/ADF inactivation by overexpression of LIMK in progesterone-treated Xenopus oocytes inhibits formation of the white maturation spot, which indicates entry into meiosis, and impairs the organization and migration of the microtubule organizing center and transient microtubule array (32). Thus, cofilin/ADF probably plays a role in centrosome movement and contractile ring formation and/or disassembly by regulating actin filament dynamics and reorganization, but precise mechanisms remain to be determined.
In the present study, we noted changes in the level of cofilin phosphorylation during the cell cycle; it gradually increases in early stages of mitosis, peaks at metaphase, and then gradually decreases in telophase and cytokinesis, findings consistent with data that in Xenopus eggs cofilin/ADF is hyperphospho-rylated at metaphase II arrest and dephosphorylated after fertilization (24). Because cofilin/ADF activity is depressed by phosphorylation at Ser-3, the content of active cofilin/ADF is high in interphase and at the onset of mitotic phase, lowers from prometaphase to metaphase, and augments during telophase and cytokinesis. Cell cycle-associated changes in the level of cofilin phosphorylation seem to correlate with functional roles of cofilin/ADF in centrosome migration and separation in the entry stage of mitosis and in contractile ring formation and/or disassembly in later stages of mitosis and cytokinesis. We found that ectopic expression of LIMK1 increases the level of cofilin phosphorylation throughout the cell cycle and induces production of multinucleate cells, an event probably caused by aberrant phosphorylation and inactivation of cofilin/ADF in later stages of mitosis. Thus, dephosphorylation and reactivation of cofilin/ADF in later stages of mitosis are probably critical for cytokinesis. Because S3D-cofilin, with replacement of Ser-3 by Asp and which mimics the phosphorylated form of cofilin, does not accumulate on the contractile ring in late stages of mitosis, dephosphorylation is also required to localize cofilin/ADF onto the cleavage furrow where it functions (33). Additionally, injection into Xenopus blastomeres of a nonphosphorylatable (constitutively active) mutant of cofilin/ADF destroys the contractile ring and regresses the cleavage furrow (24); hence, excess activity of cofilin/ADF is probably harmful for cell division. Thus, the cofilin/ADF activity must be controlled precisely during the cell cycle by phosphorylation and dephosphorylation for cells to proceed to mitosis and cytokinesis.
The kinase activity of LIMK1 changes significantly during the cell cycle, the highest activity being in prometaphase and metaphase, which is similar to changes in levels of cofilin phosphorylation. However, the level of cofilin phosphorylation peaks at metaphase, whereas the kinase activity of LIMK1 in prometaphase is slightly higher than that in metaphase. This difference is probably the result of differences in cofilin phosphatase activity during the cell cycle because the level of this phosphorylation is probably regulated by the balance between activities of cofilin kinase and phosphatase. Niwa et al. (34) recently identified a novel family of protein phosphatases, termed "Slingshot," which specifically dephosphorylates cofilin/ ADF at Ser-3 in vitro and in vivo (34). Determination of changes in the activity of Slingshot phosphatase during the cell cycle will be important to elucidate further the mechanisms by which cofilin/ADF activity is regulated during the cell cycle. Although testicular protein kinase 1, which contains a protein kinase domain similar to those of LIMKs, also phosphorylates cofilin/ADF specifically at Ser-3 (28), the kinase activity of TESK1 did not change during the cell cycle. 2 We show in this study that the kinase activity of LIMK2 increases slightly in mitosis. Expression of a kinase-negative LIMK1 did not completely inhibit P-cofilin accumulation in metaphase, which may suggest that LIMK2 also contributes to some extent to the mitosis-specific cofilin phosphorylation. Thus, it appears that LIMK1 and in part LIMK2 are responsible for the phosphorylation of cofilin/ADF during mitosis.
Previous studies revealed that LIMK1 is activated by ROCK and PAK, downstream effectors of Rho and Rac, respectively, through phosphorylation of Thr-508 within the PK domain of LIMK1 (15)(16)(17). However, the present study revealed that the mitotic activation of LIMK1 is not inhibited by Y-27632 (a specific inhibitor of ROCK) or by PAK-AI (an autoinhibitory peptide for PAK), which suggests that neither the Rho-ROCK nor the Rac-PAK signaling pathway is involved in the mitotic activation of LIMK1. Although Rho and ROCK are activated during mitosis and were shown to be critical for cytokinesis (3,27,(35)(36)(37), the time courses of activation of LIMK1 and Rho during mitosis significantly differ; activation of LIMK1 peaks at prometaphase and metaphase, whereas activation of Rho peaks in telophase (27). These findings further suggest that mitotic activation of LIMK1 is independent of the Rho-ROCK signaling pathway.
We have found that the LIMK1(T508V) mutant, in which Thr-508 is replaced by a nonphosphorylatable Val residue, is also retarded on gel electrophoresis and activated in mitotic cells, which further suggests that mitotic activation of LIMK1 is caused by a mechanism distinct from the phosphorylation of Thr-508. Interestingly, mitosis-specific activation and gel mobility shift of LIMK1 are abolished by deleting the N-terminal half region of LIMK1. Thus, the N-terminal region may be involved in the mitotic activation and phosphorylation of LIMK1. In this respect, we showed previously that the Nterminal LIM domain regulates the kinase activity of LIMK1 negatively by direct interaction with the C-terminal PK domain (29). It may be that mitotic LIMK1 is activated by phosphorylation of the N-terminal region, by which the conformational change is induced to activate the C-terminal catalytic domain by releasing it from restraint of the N-terminal region.
Because we detected at least two distinct retarded bands of LIMK1 during mitosis, mitotic LIMK1 is probably phosphorylated on multiple sites. The protein kinase(s) that is responsible for mitotic activation of LIMK1 has yet to be identified. Potential candidates for LIMK1-activating protein kinases include Cdc2 kinase, mitogen-activated protein kinase (MAPK), or other kinases that are activated in early stages of mitosis (38,39). Human LIMK1 carries in the N-terminal half region nine (Ser/Thr)-Pro sequences, and these are canonical phosphorylation sites catalyzed by Cdc2 kinase and MAPK. Indeed, both Cdc2 kinase and MAPK were able to phosphorylate LIMK1, but neither Cdc2 kinase nor MAPK induced gel mobility shift or activation of LIMK1 in in vitro experiments. 2 Furthermore, mitotic LIMK1 did not immunoreact significantly with MPM-2 antibody, which specifically recognizes phospho-Ser/Thr-Pro sequences (40). 2 Thus, it is most likely that protein kinase(s) other than Cdc2 kinase and MAPK are involved in the mitotic activation and the gel mobility shift of LIMK1, although it is still possible that these kinases have a partial role in the mitotic activation of LIMK1. To define the molecular basis of the cell cycle-associated activation of LIMK1, in future studies we will search for the responsible protein kinases and the residues phosphorylated on mitotic LIMK1.
By way of summary we obtained evidence for the mitosisspecific activation and hyperphosphorylation of LIMK1. The mitotic activation of LIMK1 is apparently not related to the Rho-ROCK or Rac-PAK pathway. These findings suggest a novel mechanism of LIMK1 activation and novel cellular functions of LIMK1 and cofilin in the mitotic phase, in addition to their functions in cell motility and morphology in interphase.