The Ste20-like Kinase SLK Is Required for Cell Cycle Progression through G2*

We have previously shown that the Ste20-like kinase SLK is a microtubule-associated protein that can regulate actin reorganization during cell adhesion and spreading (Wagner, S., Flood, T. A., O'Reilly, P., Hume, K., and Sabourin, L. A. (2002) J. Biol. Chem. 277, 37685-37692). Because of its association with the microtubule network, we investigated whether SLK plays a role in cell cycle progression, a process that requires microtubule dynamics during mitosis. Consistent with microtubule association in exponentially growing cells, our results showed that SLK co-localizes with the mitotic spindle in cells undergoing mitosis. Expression of a kinase-inactive mutant or SLK small interfering RNAs inhibited cell proliferation and resulted in an accumulation of quiescent cells stimulated to re-enter the cell cycle in the G2 phase. Cultures expressing the mutant SLK displayed a normal pattern of cyclin D, E, and B expression but failed to down-regulate cyclin A levels, suggesting that they cannot proceed through M phase. In addition, these cultures displayed low levels of both phospho-H3 and active p34/cdc2 kinase. Overexpression of active SLK resulted in ectopic spindle assembly and the induction of cell cycle re-entry of Xenopus oocytes, suggesting that SLK is required for progression through G2 upstream of H1 kinase activation.

with adhesion markers during cell spreading and is intimately linked to the microtubule network (1).
The observation that SLK is associated with the microtubule network (1) and that it was shown to phosphorylate and activate Plk (33) prompted us to investigate the role of SLK in cell cycle progression. Consistent with microtubule association in exponentially growing cells, our results showed that SLK co-localizes with the mitotic spindle in cells undergoing mitosis. Down-regulation of SLK protein levels by siRNA knockdown or expression of a kinase-inactive mutant (K⌬C) inhibited cell proliferation and resulted in a G 2 /M accumulation of quiescent cells induced to re-enter the cell cycle by serum stimulation. Cultures expressing the mutant SLK displayed a normal pattern of cyclin D, E, and B expression but failed to down-regulate cyclin A levels, suggesting that they cannot proceed through M phase. Supporting this, cells expressing the inactive SLK mutant failed to induce histone H3 phosphorylation and to fully activate p34/Cdc2. Finally, overexpression of an activated kinase resulted in ectopic spindle formation and activation Plx1 in Xenopus extracts, suggesting that activated SLK induces cell cycle re-entry. Overall, our results suggested that SLK is required for progression through G 2 upstream of Cdc2.

EXPERIMENTAL PROCEDURES
Cell Lines, Culture, and Adenovirus Infection-The mouse fibroblast lines MEF-3T3 (MEF Tet-Off, C3018, Clontech) and C3H10T1/2 (ATCC number CCL-226) were used in all experiments. Similar results were obtained for both cell lines. The GFP-tubulin-expressing cells (LLCKP-1) were a kind gift from Patricia Wadsworth (34). Cell lines were maintained at 37°C with 5% CO 2 in Dulbecco's modified Eagle's medium (DMEM, BioWhittaker) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 50 g/ml penicillin, and 50 g/ml streptomycin. For cell cycle experiments, fibroblasts were arrested by 48-h incubations in 0.25% FBS-DMEM and released from quiescence by the addition of 20% FBS-DMEM. The epitope-tagged kinase-dead or activated versions of SLK used in these studies (HA-K⌬C or HA-Y⌬C) have been described previously (1) and consist of a carboxyl-terminal truncation (amino acids 1-373) with or without an ATP-binding site (Lys-63 3 Arg) mutation. To monitor the effect of kinase-deficient SLK on cell cycle kinetics, adenoviral vectors expressing HA-K⌬C or a ␤-galactosidase (LacZ) control were used to infect quiescent cultures. Cells were infected at a multiplicity of infection of 100 by the addition of the adenovirus directly to cells in 0.25% FBS-DMEM 16 h before the addition of 20% FBS-DMEM and analysis. To isolate fibroblast populations synchronized at M phase, the cultures were treated overnight with nocodazole (40 ng/ml, Sigma) and shaken off into PBS by aggressive tapping against a solid surface. Detached cells floating in PBS were then collected by centrifugation.
For mitotic spindle induction experiments, LLCKP-1 cells expressing GFP-tubulin were microinjected with HA-Y⌬C expression plasmid as described previously (1). Xenopus oocytes were prepared and injected with cRNA as described (35). Germinal vesicle breakdown (GVBD) was determined from at least 20 oocytes pooled from two animals 15 h following SLK injection or progesterone treatment. Western blot analysis for Myc-SLK was performed 15 h after injection.
Western Blotting, Immunoprecipitations, and in Vitro Kinase Assays-For protein analysis, the cells were lysed in modified radioimmune precipitation buffer as described previously (1). Equal amounts of extracts were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membranes (PerkinElmer Life Sciences), which were then probed with the following antibodies: anti-Cdc2 and anti-cyclins D, E, A, and B (Upstate Biotechnology). The SLK antibody was as described previously (30). Primary antibodies were detected using either a goat anti-rabbit IgG or a goat anti-mouse IgG horseradish peroxidase-labeled secondary antibody (Bio-Rad) and visualized using Western Lightning chemiluminescence reagent (PerkinElmer Life Sciences) and exposure to x-ray film. SLK immunoprecipitations and kinase assays were carried out essentially as described previously (31). Briefly, equal amounts of lysate were immunoprecipitated for 2 h at 4°C using 1 g of SLK antibody and 20 l of protein A-Sepharose 4 Fast Flow (Amersham Biosciences). Immunoprecipitates were then washed three times with NETN (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40) and once with kinase buffer (20 mM Tris-HCl (pH 7.5), 15 mM MgCl 2 , 10 mM NaF, 10 mM ␤-glycerophosphate, 1 mM orthovanadate). Kinase reactions (20 l) were then initiated by the addition of 5 Ci of [␥-32 P]ATP and incubated at 30°C for 30 min. These reactions were terminated by the addition of 4ϫ SDS loading buffer and then resolved by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and exposed to x-ray film. Membranes were then probed for SLK to assess immunoprecipitation efficiency. SLK autophosphorylation was used as an indicator of kinase activity.
Immunofluorescence and Flow Cytometry-Immunofluorescence studies were carried out by fixation of cells growing on glass coverslips in 4% paraformaldehyde for 10 min. The cells were then washed twice with PBS and incubated with primary antibodies for 1 h. The primary antibodies used in immunofluorescence studies were as follows: anti-␣tubulin (clone DM1, Sigma), anti-phospho-H3 (Ser-10) (Cell Signaling Technology), anti-␤-galactosidase (Promega), anti-HA (12CA5 or sc-805, Santa Cruz Biotechnology Inc.), and anti-SLK (30). Antibodies were detected with either anti-mouse or anti-rabbit antibodies conjugated to either fluorescein isothiocyanate or TRITC (Sigma). 4Ј,6-Diamidino-2-phenylindole (DAPI, 0.25 g/ml) was used to stain DNA, and the samples were visualized with a Zeiss Axioscope100 epifluorescence microscope equipped with the appropriate filters and photographed with a digital camera (Sony Corp. HB050) using the Northern Eclipse software package. Quantitative analysis for phospho-H3 (pH3) immunostaining was performed by visually scoring stained cells for both HA (or LacZ) and pH3. The data were graphed as double positive (HA or LacZ ϩ pH3) cells for each time point analyzed. At least 200 HA or LacZ-positive cells were scored for each time point. Cells analyzed by flow cytometry were trypsinized and washed once in 10% FBS-DMEM. Cells were then washed twice in PBS supplemented with 1 mM EDTA (PBSE) and then fixed in 1 ml of PBSE by the dropwise addition of 2 ml of 80% ethanol prechilled to Ϫ20°C. The samples were then stored at Ϫ20°C for a minimum of 2 h, washed once in PBSE, resuspended in DNA content staining buffer (1.1% citrate buffer, 10 g/ml propidium iodide, 1 mg/ml RNase), and incubated for 30 min at 37°C before being analyzed on a Beckman Coulter flow cytometer using the Expo 32 software package. BrdUrd pulse labeling was performed using a BrdUrd flow kit (BD Biosciences) according to the manufacturer's instructions. Briefly, 16 h before labeling, cultures were infected with either HA-K⌬C or LacZ adenovirus at a multiplicity of infection of 100 for 90 min in unsupplemented DMEM and then grown in 10% FBS-DMEM. The cultures were then labeled with BrdUrd for 1 h and collected at various times. The DNA content of BrdUrd positive cells was then analyzed flow cytometrically.
Cell Counts, Cloning, and Transfections-To monitor cell proliferation, the cells were counted after infection on day 0 at a multiplicity of infection of 100 as described above. Cell populations were trypsinized and scored by trypan blue exclusion over time. Cell counts were performed in triplicates in three independent experiments. For antisense SLK expression, a 5Ј 300-bp SLK fragment was cloned into the pEMSV-puro expression vector in the reverse orientation. The antisense SLK plasmid and the puromycin control were then transfected into MEF-3T3 cells using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's protocol and selected in puromycin (100 g/ml) over a 2-week period. Following selection, stable transfectants were visualized using CYTO-QUIK staining (Fisher Health Care) followed by several PBS washes. SLK Smart Pools siRNAs were obtained from Dharmacon against the following murine SLK target sequences: 5Ј-GGTTGAGAT-TGACATATTA-3Ј, in addition to a scrambled siRNA (5Ј-GATAATT-TATGGATGTGAC-3Ј). Control siRNA consisted of the siCONTROL (Dharmacon; 5Ј-UAGCGACUAAACACAUCAAUU-3Ј), having no perfect match to known human or mouse sequences. All siRNAs were transfected using the Transit-TKO reagent (Mirrus Corp.) according to the manufacturer's instructions.

SLK Co-localizes with the Mitotic Spindle and Is Regulated during the
Cell Cycle-We have previously shown that a proportion of SLK is associated with the microtubule network of exponentially growing fibroblasts (1). During immunofluorescence studies involving the co-localization of SLK to the microtubule network in asynchronous cultures, we observed rare patterns of SLK and ␣-tubulin co-localization that resembled the mitotic spindle. To further investigate the possibility that SLK might co-localize with the mitotic spindle, we performed confocal microscopy. Although some SLK was found outside of the mitotic spindle, confocal analysis of DAPI-stained cells, in combination with anti-SLK and anti-␣-tubulin, shows that most of the SLK staining co-localized with tubulin during metaphase (Fig. 1, A and B). The observation that SLK is associated with ␣-tubulin, a central component of the   DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 mitotic machinery, suggests a role for SLK during mitosis. Alternatively, SLK may be required for spindle assembly.

SLK Is Required for Cell Cycle Progression
To further investigate the potential role of SLK in cell cycle progression, its activity was monitored throughout the different phases of the cell cycle using synchronized cell populations. Serum-starved cultures were released from G 0 by the addition of serum, collected at various times, and monitored for SLK activity and DNA content by kinase assays and fluorescence-activated cell sorter analysis, respectively. Fig.  2A displays the cell cycle phase as determined by flow cytometric measurements of DNA content in these synchronized populations following serum stimulation. After 24 h of serum stimulation, a marked and consistent 3-4-fold increase in SLK kinase activity (Fig. 2B) was observed when ϳ60 -70% of the cells entered the G 2 /M compartment, as determined by fluorescence-activated cell sorter analysis. Similarly, an increase in SLK activity during G 2 has been reported previously, albeit to a much lesser extent (ϳ1.3-fold) (33). Interestingly, as the cells exited M phase and re-entered G 1 , a marked reduction in kinase activity was observed. The total levels of SLK protein were found to be unaffected throughout the time course. These results suggested that SLK kinase activity is up-regulated at a time point at which the vast majority of the cells display a 4N DNA content.
Kinase-deficient SLK Inhibits Proliferation-To investigate the potential role of SLK in proliferation, an expression vector bearing an antisense SLK fragment was transfected into MEF-3T3 fibroblasts and subjected to puromycin selection. Stable clones were visualized after 14 days using Cyto-Quick stain (Fisher). As shown in Fig. 2C, antisense SLK-transfected cultures reproducibly displayed a marked reduction in colony numbers when compared with the pEMSV-puro control vector. Furthermore, expansion and Western blot analysis of the resulting antisense SLK clones showed that they did not down-regulate SLK, suggesting that the antisense RNA was not expressed in those clones (not shown).
As for antisense SLK expression, infection of MEF-3T3 cells with an adenoviral vector carrying a kinase-dead truncation of SLK (SLK 1-373K63R, termed HA-K⌬C) suppressed cell proliferation as determined by dye exclusion cell counts (Fig. 3D). Taken together, these results indicate that SLK activity is required for cell proliferation and that our truncated kinase-dead version, HA-K⌬C, is able to interfere with SLKdependent pathways.
SLK Is Required for Progression through G 2 -To further investigate the mechanism by which K⌬C inhibits proliferation, quiescent MEF-3T3 cultures were infected with a K⌬C-expressing adenovirus or control LacZ virus, serum-stimulated, and subjected to DNA content analysis over time. As shown in Fig. 3, A and B, cultures expressing LacZ proceeded through the cell cycle and exited the G 2 /M compartment by 32 h following serum stimulation. In contrast, K⌬C-expressing cultures displayed a marked increase in the proportion of cells in the G 2 /M compartment, suggesting that K⌬C expression delays progression through the G 2 or M phase by inducing a specific block, or alternatively, a failure to exit M phase. Supporting these data, flow tracking of BrdUrd-pulsed exponentially growing K⌬C-infected cultures showed that they proceed through the G 2 /M compartment with delayed kinetics when compared with LacZ-infected cells. Taken together, these results suggested that SLK is required for normal progression through the G 2 or M phase.
The levels of the various cyclins have been demonstrated to fluctuate during the cell cycle (reviewed in Ref. 5). Cyclin D has been observed to be induced prior to S phase and to remain elevated in proliferating cells. Cyclin E is transiently up-regulated at the G 1 /S boundary, whereas cyclins A and B are induced at the S/G 2 boundary and down-regulated at the onset and the end of M phase, respectively. Therefore, to better define the cell cycle block induced by kinase-deficient SLK, control and K⌬C-infected cultures were surveyed for cyclin expression over time following release from G o . The infection efficiency was found to be typically between 70 and 80% for both viruses, and the HA-K⌬C protein was observed to be expressed at levels that were similar to endogenous SLK (data not shown). Our results show that, relative to actin, cyclin D was slightly up-regulated following serum stimulation in both the control and the K⌬C-infected cultures, suggesting that they re-entered the cell cycle (Fig. 4A). Similarly, cyclin E levels were up-regulated 16 h following stimulation, when a significant proportion of the cells entered S phase ( Fig. 2A), and down-regulated thereafter, suggesting that both cultures entered and exited S phase with similar kinetics. In addition, both cultures induced cyclin B expression at ϳ8 -16 h. Similarly, both cultures induced cyclin A at the G 1 /S transition. However, only controlinfected cultures showed a marked and reproducible down-regulation at 32 h, suggesting that K⌬C-expressing cells fail to down-regulate cyclin A. Cyclin A degradation has been shown to occur during prometaphase, at the onset of mitosis (36,37). Therefore, our observations suggested that the K⌬C-induced cell cycle block occurs in G 2 , prior to M phase. One possibility is that the cell cycle block induced by K⌬C fails to activate specific signals required for cyclin A degradation (36,37). Alternatively, SLK may be required for the direct activation of these checkpoints or proteolysis of cyclin A.
During interphase, cytosolic MPF is kept inactive by inhibitory phosphorylation of cdc2 on Thr-14 and Tyr-15 by Myt1 and Wee1, respectively (9 -11). Activation of this complex is triggered by the Cdc25C phosphatase through cdc2 dephosphorylation of Thr-14 and Tyr-15 (12)(13)(14). In agreement with elevated cyclin A levels, K⌬C-infected cultures did not significantly up-regulate Cdc2 activity as evidenced by the high levels of Cdc2 tyrosine 15 phosphorylation (Fig. 4B). These results strongly suggest that K⌬C induces a cell cycle block prior to mitotic entry, in G 2 . Supporting this, K⌬C-positive cells were not found to assemble mitotic spindles 24 h after stimulation, at which point 60% of the cells displayed 4N DNA content (data not shown).
Chromosome condensation initiated in early G 2 (38) is accompanied by the hyperphosphorylation of histone H1 (24) and phosphorylation of H3 (25) on Ser-10 (26). To further refine the cell cycle block induced by K⌬C expression, adenovirus-infected cultures were stained for both K⌬C expression and phospho-H3. Double immunostaining of serumstimulated fibroblast cultures at 24 h (Fig. 5, A-F) shows that H3 phosphorylation was markedly reduced in K⌬C-expressing cells in comparison with control infected cultures. Although the proportion of HA-and phospho-H3-positive cells slightly increased at 28 and 32 h after stimulation, their number was significantly lower than control cells, suggesting that K⌬C-expressing cells are delayed in early G 2 .
SLK Expression Induces Spindle Formation and Mitotic Entry-Our results showed that expression of a kinase-inactive SLK is sufficient to induce an early G 2 cell cycle block in cycling fibroblasts, suggesting that an SLK-dependent pathway is required during G 2 for progression into mitosis. To determine whether SLK plays a central role in G 2 and to rule out potential nonspecific effects by K⌬C overexpression, exponentially growing fibroblasts were transfected with an SLK siRNA pool and subjected to DNA content analysis. Transfection of SLK siRNAs downregulated SLK protein levels by 80 -90% within 48 h (Fig. 6A). No effect was observed in the siRNA control samples. DNA content analysis 48 h following siRNA transfection showed that SLK knockdown resulted in a marked G 2 /M accumulation (92% 4N DNA content; Fig. 6B), an inhibition of proliferation, and increased cyclin A levels (data not shown). These data further support a role for SLK in cell cycle progression and rule out potential nonspecific effects by K⌬C overexpression. Supporting these results, microinjection of activated SLK (SLK 1-373 ) in GFPtubulin labeled cells induced ectopic mitotic spindles in the injected cells within 3-6 h, ultimately resulting in cell death (Fig. 6C) (1). Similarly, injection of the activated SLK, but not kinase-dead, mRNA in Xenopus oocytes resulted in the hyperphosphorylation of Plx1 and GVBD without progesterone induction, suggesting that the injected eggs re-entered the cell cycle as for progesterone-treated oocytes (Fig.  6D). Interestingly, expression of a kinase-dead version followed by progesterone treatment did not inhibit cell cycle re-entry and GVBD (not shown), suggesting that alternative pathways, such as the mitogen-activated protein kinase (MAPK) pathway, may be sufficient to promote maturation (39)

DISCUSSION
Cell cycle progression is regulated by complex signaling networks involving post-translational modification, gene expression, and cytoskeletal reorganization. Progress through the various phases involves the activation of key factors and is monitored by various checkpoint proteins. We have previously isolated an Ste20-like kinase termed SLK that is involved in cytoskeletal reorganization (1,30,31). Interestingly, a fraction of SLK protein is also observed to associate with the microtubule network in spreading and exponentially growing cells (1).
Here we have shown that SLK also associates with the microtubule network at mitosis, suggesting that it plays a role in cell cycle progression. To investigate this, the activity and expression of SLK was evaluated throughout the cell cycle, and the effect of a kinase-defective SLK on cell cycle progression was assessed.
Our results showed that SLK co-localizes with the mitotic spindle during M phase and that its kinase activity is up-regulated as synchronized fibroblast cultures enter the G 2 /M compartment. Expression of an  DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 antisense SLK construct or a kinase-deficient mutant inhibited cell proliferation. Flow cytometric analysis of these populations showed that SLK is required for progression through the G 2 /M compartment and that cells expressing a kinase-inactive SLK fail to down-regulate cyclin A. Furthermore, we have observed that cultures expressing an inactive SLK show reduced histone H3 phosphorylation and p34 Cdc2 activation. Supporting this, overexpression of activated SLK induced spindle formation and cell cycle re-entry in Xenopus oocytes. Overall, our data suggested that SLK is required upstream of Cdc2 and that interfering with SLK-dependent pathways leads to cell cycle arrest early in the G 2 phase of the cell cycle.

SLK Is Required for Cell Cycle Progression
An interpretation of the data presented here indicated that SLK is required during G 2 , at a point after cyclin A expression and before Cdc2 activation, for progression into mitosis. A Xenopus LOK/Stx10 homolog, xPolo-like kinase kinase (xPlkk1), has been demonstrated to activate Plx1 (22). Interestingly, SLK has been demonstrated to phosphorylate and activate Plk in a mammalian system (33). However, we and others 4 have not observed phosphorylation and activation of mammalian Plk1 by SLK in in vitro kinase assays or transient transfections (not shown).
An evolutionary conserved relationship appears to exist between mammalian and amphibian Ste20-like kinases, the polo-like kinase family, and the Cdc25C phosphatase. The human homolog of the Ste20-like kinase LOK, Stx10, has been shown to associate with and phosphorylate Plk1 (40), and Plk3 has been shown to initiate the nuclear translocation of Cdc25C by phosphorylating Cdc25C on serine 191 (41). In Xenopus, the polo-like kinase, Plx1, has been shown to activate Cdc25C (21), an event that has not been unarguably proven in a mammalian system and may be species-specific. If Plk requires a signal from SLK to activate Cdc25C in a mammalian system, one would predict a G 2 /M arrest, similar to the findings presented here. This would result in a Cdc25C protein that is deficient in all of the post-translational modifications required to initiate mitosis. However, the questionable validity and specificity of phospho-specific Cdc25C antibodies has made the identification of the various forms of the phospho-Cdc25C difficult and inconclusive (not shown). We have previously shown that SLK can induce actin depolymerization and cell death in various cell lines (1,31). One possibility is that SLK overexpression in cycling cells induces a deregulated mitotic entry, bypassing cell cycle controls, resulting in actin breakdown and death. Alternatively, SLK-mediated cytoskeletal reorganization may be required for, or trigger, checkpoint activation, G 2 progression, and mitotic entry.
Overall, we have shown that SLK is a component of the mitotic spindle and that its activity is required in G 2 , upstream of Cdc2, for efficient progression into mitosis. The characterization of the molecular components of SLK-dependent signaling pathways now awaits the identification of its substrates and binding proteins. Identical results were obtained with a SLK scrambled siRNA (not shown). A marked down-regulation of SLK at 50 nM of siRNA resulted in the accumulation of the cells in the G 2 /M compartment, further supporting a requirement for SLK for progression through G 2 . C, exponentially growing LLCPK-1 fibroblasts expressing GFP-tubulin were microinjected with an activated form of SLK (HA-tagged; amino acids 1-373). All (n ϭ 25) injected cells (arrowheads), detected by anti-HA staining, displayed ectopic mitotic spindles when expressing activated SLK (panels I and III; merged fluorescence), suggesting that SLK induces mitotic entry. No spindle formation was observed when cells were injected with the kinase-dead version (panels II and IV). D, Xenopus oocytes injected with the same form of activated SLK re-entered the cell cycle, as evidenced by GVBD and the shift in the molecular weight of Plx1, indicative of phosphorylation (pPlx1), suggesting that SLK can activate mitotic entry in Xenopus eggs. As a control, activation of Plx1 by progesterone (lane P) was used (lane C ϭ untreated). Expression of kinase-dead SLK (mSLK-KD) could not induce oocyte maturation. The GVBD data represent the average of four independent experiments. Western blot analysis shows the expression of the Myc-tagged SLK protein.