PCTK1 Regulates Integrin-Dependent Spindle Orientation via Protein Kinase A Regulatory Subunit KAP0 and Myosin X

Abstract

Integrin-dependent cell-extracellular matrix (ECM) adhesion is a determinant of spindle orientation. However, the signaling pathways that couple integrins to spindle orientation remain elusive. Here, we show that PCTAIRE-1 kinase (PCTK1), a member of the cyclin-dependent kinases (CDKs) whose function is poorly characterized, plays an essential role in this process. PCTK1 regulates spindle orientation in a kinase-dependent manner. Phosphoproteomic analysis together with an RNA interference screen revealed that PCTK1 regulates spindle orientation through phosphorylation of Ser83 on KAP0, a regulatory subunit of protein kinase A (PKA). This phosphorylation is dispensable for KAP0 dimerization and for PKA binding but is necessary for its interaction with myosin X, a regulator of spindle orientation. KAP0 binds to the FERM domain of myosin X and enhances the association of myosin X-FERM with β1 integrin. This interaction between myosin X-FERM and β1 integrin appeared to be crucial for spindle orientation control. We propose that PCTK1-KAP0-myosin X-β1 integrin is a functional module providing a link between ECM and the actin cytoskeleton in the ECM-dependent control of spindle orientation.

INTRODUCTION

Spindle orientation defines the axis of cell division and is important for asymmetric cell division, tissue morphogenesis, and organogenesis. Hertwig first identified cell shape as one of the determinants of spindle orientation (1). According to the Hertwig rule, cells divide along their long axis, which is considered the default mechanism for spindle orientation (2–4). Cells proliferating in culture dishes or in tissues in vivo, however, preferentially divide according to internal or external polarity cues that include asymmetrically localized polarity proteins as well as cell-cell and cell-extracellular matrix (ECM) adhesion (reviewed in references 4–8). Although in-depth studies have identified the mechanisms regulating spindle orientation within the context of cell-cell adhesion and the axis of cell polarity, the mechanisms underlying cell-ECM adhesion-dependent spindle orientation remain poorly characterized.

We have previously shown in nonpolarized adherent cells, such as HeLa cells, that integrin-mediated cell-ECM adhesion aligns the mitotic spindle along the plane of the ECM, ensuring both daughter cells remain attached to the ECM following cell division (9). This mechanism requires the actin cytoskeleton, astral microtubules, the microtubule plus-end-tracking protein EB1, the actin motor protein myosin X, and the plasma membrane phospholipid phosphatidylinositol 3,4,5-triphosphate (9–11). Following a genome-wide RNA-interference (RNAi) screen of human kinases, we identified ABL1 tyrosine kinase as a novel regulator of spindle orientation (12). We also identified PCTAIRE-1 kinase (PCTK1; also referred to as Cdk16) as a strong candidate regulator of spindle orientation.

PCTK1 to PCTK3 are highly conserved serine/threonine kinases that belong to the cyclin-dependent kinase (CDK) family of protein kinases (13, 14). These kinases are characterized by a highly conserved kinase domain that is closely related to that of other conventional CDKs, as well as N- and C-terminal extensions that are unique for each isoform (15). PCTK1 is expressed in many tissues and cell lines and is highly expressed in terminally differentiated tissues, including brain and testis (14, 16, 17). Although the activity of PCTK1 is known to be regulated during cell cycle progression (17), it is unclear whether PCTK1 participates in cell cycle-related events. Consistent with its high expression in terminally differentiated tissues, recent studies have highlighted non-cell cycle-related functions for PCTK1, including axonal vesicular transport in neurons of nematodes (18), neuronal cell differentiation (19, 20), intercellular vesicular transport (21, 22), and spermatogenesis (23). More recently, PCTK1 was shown to regulate myoblast migration during skeletal myogenesis (24).

In this report, we demonstrate a novel function for PCTK1 in the regulation of integrin-dependent spindle orientation. Phosphoproteomic analysis identified Ser83 of KAP0, a regulatory subunit of protein kinase A (PKA), as a PCTK1 target phosphorylation site that is required for spindle orientation control. This phosphorylation enhances the association of KAP0 with myosin X, a regulator of spindle orientation and positioning (9, 25). We further show that myosin X binds to β1 integrin in a KAP0-dependent manner, and that this is required for spindle orientation. Therefore, our study uncovers a novel cell cycle-related function for PCTK1 in the regulation of spindle orientation in mitosis and identifies the underlying mechanism by which this kinase exerts its function.

MATERIALS AND METHODS

Cell culture, reagents, and antibodies.

HeLa cells were cultured in Dulbecco's modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS). Cells were cultured on fibronectin and synchronized by a double-thymidine block. Forskolin (Sigma) was dissolved in dimethylsulfoxide (DMSO) and stored at room temperature (RT). H-89 (Sigma) was dissolved in deionized distilled water (DDW; 20 mM) and stored at 4°C. The antibodies used included mouse anti-PCTK1 (sc-53410; Santa Cruz), mouse anti-α-tubulin (T 6199; Sigma), rabbit anti-γ-tubulin (T 5192; Sigma), mouse anti-green fluorescent protein (anti-GFP; Clontech), mouse anti-KAP0 (610165; BD Transduction Laboratories), mouse anti-PKAα catalytic subunit (sc-28316; Santa Cruz), mouse anti-FLAG (F3165; Sigma), rabbit anti-FLAG (F7425; Sigma), mouse anti-Myc (sc-40; Santa Cruz), rabbit anti-phospho-CREB (9198; Cell Signaling), rabbit anti-CREB (9197; Cell Signaling), mouse anti-myosin X (sc-166720; Santa Cruz), and mouse anti-β1 integrin (MAB2000; Millipore).

siRNA experiments.

The sequences of the used short interfering RNAs (siRNAs) are the following (lowercase letters indicate overhang sequences): PCTK1-1, 5′-GGAGAUCAGACUGGAACAUtt-3′; PCTK1-2, 5′-GAUCUCCACUGAGGACAUCtt-3′; myosin X, 5′-GCGAAGACCGUGUACAACAtt-3′; integrin β1, 5′-GGAAUGCCUACUUCUGCACtt-3′. The siRNAs targeting human FAM129B, HNRNPA1L2, KAP0, LRRC16A, MKI67, NUCKS1, NUP210, PCYT1A, PHC3, RANBP2, RANBP3, RBBP6, RCC1, SNIP1, and TLE3 were purchased from Ambion (siRNA custom library). HeLa cells cultured on fibronectin-coated cover glasses were transfected with siRNA using Oligofectamine (Invitrogen), incubated for 4 h, and synchronized by a double-thymidine block. The expression levels were analyzed by Western blotting.

Plasmid construct and transfection.

pCMV-HA-WT-PCTK1, pCMV-HA-DN-PCTK1, and pMyrPalm_mEGFP_IRES_puro2b were purchased from Addgene (1999, 1998, and 21038) and subcloned into pEGFPC1. Full-length human KAP0 and myosin X cDNAs were purchased from KAZUSA (FHC01268 and FHC01122) and subcloned into pcDNA3-FLAG, pcDNA3-Myc, or pEGFPC1. Full-length human β1 integrin cDNA was purchased from NIBIO (clone HDG1180). GFP-PCTK1 siR was constructed using the following primer pair: 5′-GTAAGCTTACAGACAACCTTGTGGCACTCAAAGAAATAAGGCTGGAACAT-3′ and 5′-ATGTTCCAGCCTTATTTCTTTGAGTGCCACAAGGTTGTCTGTAAGCTTAC-3′. GFP-PCTK1 siR-WT, GFP-PCTK1 siR-KD, FLAG-KAP0 siR-WT, FLAG-KAP0 siR-S83A, FLAG-KAP0 siR-S83D, FLAG-β1 integrin siR-WT, and FLAG-β1 integrin siR-W775A were constructed by mutagenesis using a QuikChange site-directed mutagenesis kit (Stratagene). HeLa cells were transfected with plasmids by using Lipofectamine plus (Invitrogen) after the first release from a double thymidine block, incubated for 45 min, and washed into the fresh medium.

Spindle orientation assay.

Z-stack images (0.5 μm apart) of metaphase cells stained with anti-γ-tubulin antibodies and Hoechst were obtained. Images were subjected to our homemade ImageJ plug-in, which defines the position of the two poles of a spindle in a series of the z-stack images and measures the linear distance and the vertical distance between the two poles of a metaphase spindle. It also calculates the spindle angle by inverse trigonometric function (12). We judged that the spindle was properly oriented when the spindle angle was less than 20 degrees.

Cell staining.

For detection of α-tubulin and γ-tubulin, cells were fixed with 3.7% formaldehyde at 37°C for 10 min, followed by incubation with methanol at −20°C for 15 min. For detection of myosin X and KAP0, cells were fixed with 4% paraformaldehyde at RT for 20 min, followed by incubation with 0.2% Triton X-100 at RT for 10 min. For detecting β1 integrin, cells were incubated with anti-β1 integrin antibodies labeled with a Zenon Alexa Fluor 488- or 546-mouse IgG labeling kit (Molecular Probes) for 12 min at 37°C, washed with PBS, and then fixed. Cells were blocked with 3% bovine serum albumin (BSA) at 37 °C for 30 min, incubated with primary antibodies at 4°C overnight, washed, and incubated for 1 h with secondary antibodies (Alexa Fluor 488– or 546–goat anti-mouse or anti-rabbit IgG antibody [Molecular Probes]). For double staining of myosin X and KAP0, anti-myosin X antibodies were prelabeled with Zenon Alexa Fluor 546.

Western blotting and immunoprecipitations.

For Western blot analysis, HeLa cells were lysed with the lysis buffer A (20 mM HEPES [pH 7.3], 12.5 mM β-glycerophosphate, 50 mM NaCl, 1.5 mM MgCl₂, 2 mM EGTA, 10 mM NaF, 0.5% Triton X-100, 2 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl fluoride, 1 mM Na₃VO₄, and 20 μg/ml aprotinin) and centrifuged at 13,000 rpm at 4°C for 15 min. The supernatants were subjected to Western blotting using the antibodies described above. For immunoprecipitation of FLAG-KAP0 and β1 integrin, M phase-synchronized cells were lysed with lysis buffer B (50 mM HEPES [pH 7.5], 2 mM EGTA, 2 mM MgCl₂, 12.5 mM β-glycerophosphate, 50 mM NaCl, 10% glycerol, 0.5% NP-40, 10 mM NaF, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na₃VO₄, 2 μg/ml aprotinin, and 100 μM okadaic acid) and centrifuged at 13,000 rpm at 4°C for 15 min. Anti-GFP antibody-conjugated beads (06083-05; Nacalai) were added to the supernatant, incubated at 4°C for 3 to 4 h, washed with lysis buffer B, and then subjected to immunoblotting. For immunoprecipitation of β1 integrin, mouse anti-β1 integrin antibodies (4 μg) were added to the supernatant and incubated at 4°C for 1 h. Protein G-Sepharose beads (GE Healthcare), preincubated with 3% BSA-phosphate-buffered saline (PBS), were added to the mixtures, incubated at 4°C for 2 to 3 h, washed three times with wash buffer (50 mM HEPES [pH 7.5], 2 mM EGTA, 2 mM MgCl₂, 12.5 mM β-glycerophosphate, 150 mM NaCl, 0.5% NP-40, 10 mM NaF, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na₃VO₄, 2 μg/ml aprotinin, and 100 μM okadaic acid), and then subjected to Western blotting.

Phosphoproteomic analysis.

Cell extracts were digested with trypsin and labeled with ¹³CD₂O for Luci siRNA- or HA-PCTK1-WT-transfected cells and with ¹²CH₂O for PCTK1 siRNA- or HA-PCTK1-DN-transfected cells. After desalting, labeled peptides were subjected to a titanium dioxide (TiO₂)-based phosphopeptide enrichment (26, 27), followed by nano-scale liquid chromatography-tandem mass spectrometry (nanoLC-MS/MS). NanoLC-MS/MS analyses were conducted with a TripleTOF 5600 mass spectrometer (ABSciex, Foster City, CA) equipped with an UltiMate 3000 RSLCnano pump (Thermo Fisher Scientific, Bremen, Germany) and a HTC-PAL autosampler (CTC Analytics, Zwingen, Switzerland). Reprosil-Pur C18-AQ materials were packed into a self-pulled needle (150-mm length, 100-μm inner diameter, 6-μm opening). The injection volume was 5 μl, and the flow rate was 500 nl/min. The mobile phases consisted of 0.5% acetic acid (phase A) and 0.5% acetic acid in 80% acetonitrile (phase B). A three-step linear gradient of 5 to 10% B in 5 min, 10 to 40% B in 60 min, and 40 to 100% B in 5 min, followed by 100% B for 10 min, was employed. Spray voltages of 2,300 V were applied. The mass scan ranges were m/z 300 to 1,500, and the top 10 precursor ions were selected in each MS scan for subsequent MS/MS scans. Peptides and proteins were identified by Mascot v2.3 (Matrix Science, London, United Kingdom) against UniProtKB/Swiss-Prot with a precursor mass tolerance of 20 ppm, a fragment ion mass tolerance of 0.1 Da, and strict trypsin specificity, allowing for up to 2 missed cleavages. Cysteine carbamidomethylation was set as a fixed modification, and methionine oxidation was allowed as a variable modification. Dimethylation of N termini and ε-amino groups of lysine and phosphorylation of serine, threonine, and tyrosine were set as variable modifications. Peptide quantitation was performed using Mass Navigator (Mitsui Knowledge Industry, Tokyo, Japan), based on the integrated peak areas, and the heavy- and light-labeled peptide ratio (H/L ratio) was calculated for individual runs.

RESULTS

PCTK1 regulates β1 integrin-dependent spindle orientation.

When HeLa cells are cultured on ECM fibronectin, mitotic spindles are aligned in parallel to the plane of the ECM in a β1 integrin-dependent manner (9). Our previous genome-wide RNAi screen of kinases identified PCTK1 as a strong candidate regulator of spindle orientation (Fig. 1A) (12). To examine the requirement of PCTK1 in spindle orientation control, PCTK1 was depleted in synchronized HeLa cells using two independent, nonoverlapping siRNAs (Fig. 1B). Metaphase spindles, where chromosomes align along the metaphase plate, were properly aligned in parallel to the substratum in control cells transfected with luciferase siRNA (Luci si) but were misoriented significantly in PCTK1-depleted cells (Fig. 1C). The average spindle angle, the angle between the axis of a metaphase spindle and that of the substrate surface (9), was significantly increased in PCTK1-depleted cells compared to that in control cells (Fig. 1D). To further evaluate the spindle orientation phenotype in cells, we judged that the spindle was properly oriented when the spindle angle was less than 20 degrees. Statistical analysis for the percentage of metaphase cells with properly oriented spindles, as well as for the average spindle angle, were performed in later investigations. A GFP-tagged, siRNA-resistant form of wild-type PCTK1 (GFP-PCTK1 siR-WT), but neither kinase-defective PCTK1 (GFP-PCTK1 siR-KD) containing the Lys194-to-Arg mutation nor GFP alone, restored normal spindle orientation in PCTK1-depleted cells (Fig. 1E and F). These results indicate that PCTK1 regulates spindle orientation in HeLa cells and that the kinase activity of PCTK1 is required for this regulation.

FIG 1.

PCTK1 regulates spindle orientation in a kinase activity-dependent manner. (A) An RNA-mediated interference screen of human kinases identified PCTK1 as a strong candidate for a spindle orientation regulator. Details for the screening were described previously (12). In brief, we performed a two-step screening. In the 1st screening, we used the Silencer Human Kinase siRNA Library (AM80010V3; Ambion), targeting 719 human kinase and kinase-related genes. In the 2nd screening, we rescreened the 28 candidate genes from the 1st screening. The statistical analysis was done by one-sided Mann-Whitney's U test, and the Bonferroni procedure was used to correct for multiple testing. The single-tail traditional significance level of 0.05 was corrected to be 9.0 × 10⁴ (0.05/56, where 56 is the number of siRNAs used in the 2nd screen). The results of statistical analysis shows that PCTK1-1 si and PCTK1-2 si reached statistical significance in the both 1st and 2nd screenings. (B) Western blot analysis for PCTK1 and control α-tubulin in the cells transfected with Luci siRNA or two kinds of PCTK1 siRNAs (PCTK1-1 si and PCTK1-2 si). (C) The X-Z projections of metaphase cells treated as described for panel B and stained with anti-γ-tubulin antibodies and Hoechst. (D) Distribution of (histogram) and average (inset, right) spindle angles in the cells transfected with the indicated siRNAs. Data are means ± standard errors of the means (SEM) from three different experiments; n > 50/experiment; *, P < 0.05 by Dunnett's multiple-comparison test. (E) Western blot analysis for GFP-PCTK1, PCTK1, and control α-tubulin in the cells transfected with Luci siRNA or PCTK1-1si together with GFP alone, a GFP-tagged siRNA-resistant form of wild-type PCTK1 (GFP-PCTK1 siR-WT), or a kinase-defective form of PCTK1 (GFP-PCTK1 siR-KD). (F) Distribution of (histogram) and average (inset, right) spindle angles and the percentage of metaphase cells with properly oriented spindles (inset; left) in the cells transfected with the indicated siRNA together with the indicated plasmids as described in panel E. Means ± SEM from five different experiments; n = 51/experiment; *, P < 0.05 by t test.

Phosphoproteomic analysis identifies KAP0 as a downstream target of PCTK1 in spindle orientation control.

To identify downstream targets of PCTK1 involved in spindle orientation, phosphoproteomic analysis was performed on cell extracts from M phase-arrested HeLa cells that had been transfected with either Luci siRNA, PCTK1-1 siRNA, HA-tagged wild-type PCTK1 (HA-PCTK1-WT), or a dominant-negative form of PCTK1 (HA-PCTK1-DN) containing the Asp304-to-Asn mutation (28, 29) (Fig. 2A). Tryptic digests from Luci siRNA- and HA-PCTK1-WT-transfected cells were labeled with ¹³CD₂O (heavy label), while digests from PCTK1-1 siRNA- and HA-PCTK1-DN-transfected cells were labeled with ¹²CH₂O (light label). The heavy-labeled peptides from the Luci siRNA-transfected cells then were mixed at a ratio of 1:1 with the light-labeled peptides from the PCTK1 siRNA-transfected cells. Likewise, heavy-labeled peptides from HA-PCTK1-WT-transfected cells were mixed with light-labeled peptides from HA-PCTK1-DN-transfected cells. We identified 296 heavy-labeled phosphopeptides whose peak area was more than twice the size of that found for their light-labeled counterparts. We have recently developed a method for the large-scale identification of phosphorylation sites for profiling protein kinase selectivity (30). In this method, cellular proteins are dephosphorylated with phosphatase, phosphorylated with the target kinase, and digested with Lys-C/trypsin. The resulting phosphopeptides are enriched using TiO₂-based hydroxy acid-modified metal oxide chromatography (HAMMOC) and subsequently subjected to LC-MS/MS (30). Using this method, we identified 81 in vitro-phosphorylation sites for PCTK1 in HeLa cell extracts (see Table S1 in the supplemental material) and extracted the Ser-Pro-Pro/Ser-Pro-Lys amino acid sequence as a target motif for PCTK1 phosphorylation (Fig. 2B). Among the 296 phosphopeptides identified, 15 phosphorylation sites from 15 proteins conformed to the Ser-Pro-Pro/Ser-Pro-Lys motif (see Table S2). To identify proteins required for spindle orientation control, each of these 15 candidate proteins was depleted in HeLa cells using two independent siRNAs. We identified four genes (kap0, pcyt1a, tle3, and nucks1) whose downregulation of expression by either siRNA resulted in a spindle misorientation phenotype compared with that of control Luci siRNA-transfected cells (Fig. 2C, color bars). The phosphorylation site of KAP0 lies at Ser83 (Fig. 2D). Phosphorylation of this residue was significantly decreased in cells transfected with either PCTK1 siRNA or HA-PCTK1-DN compared with cells transfected with the respective Luci siRNA or HA-PCTK1-WT controls (Fig. 2E). Additionally, the same site was identified as an in vitro phosphorylation site for PCTK1 (see Table S1). Furthermore, phosphorylation of the same site was greatly upregulated in the M phase-arrested cells compared with the G₁/S phase-arrested cells (Fig. 2F). These results prompted us to investigate a role for KAP0 Ser83 phosphorylation in spindle orientation control.

FIG 2.

Phosphoproteomic analysis identified KAP0 as a downstream target of PCTK1 in spindle orientation control. (A) Western blot analysis for PCTK1 and control α-tubulin in the cells transfected with Luci siRNA, PCTK1-1 siRNA, HA-PCTK1-WT, or HA-PCTK1-DN. (B) Flow diagram of the phosphoproteomic analysis to identify PCTK1 substrates in the M phase-synchronized HeLa cells. (C) Average spindle angles in the cells transfected with the indicated siRNAs. (D) Results of phosphoproteome analysis of the extracts of the Luci siRNA-transfected cells shown in panel A. The annotated MS/MS spectrum of EDEIpSPPPPNPVVK is shown. (E) Normalized peak areas of the extracted ion current chromatogram of EDEIpSPPPPNPVVK in the cells transfected with PCTK1-1 siRNA or HA-PCTK1-DN relative to that in the cells transfected with Luci siRNA or HA-PCTK1-WT, respectively. (F) Extracted ion current chromatograms of EDEIpSPPPPNPVVK from the immunoprecipitated FLAG-KAP0, which was expressed in the cells arrested in M phase or G₁/S phase.

Ser83 phosphorylation of KAP0 is necessary for correct spindle orientation.

KAP0, also known as PRKAR1A, is a type 1α (RIα) regulatory subunit of the cyclic AMP (cAMP)-dependent protein kinase A (PKA) (31–33). The inactive PKA holoenzyme is a tetrameric protein complex consisting of two catalytic subunits and one regulatory subunit dimer. Binding of cAMP to the regulatory subunits results in the dissociation and activation of the catalytic subunits (34, 35). KAP0 contains a dimerization/docking domain within its N-terminal region and two cAMP-binding domains within its C terminus (Fig. 3A). Our phosphoproteomic analysis identified Ser83 within the dimerization/docking domain of KAP0 as a PCTK1 target phosphorylation site (Fig. 2D). When KAP0 was depleted in HeLa cells by either of two independent siRNAs, spindles were significantly misoriented (Fig. 3B and C), demonstrating the requirement for KAP0 in spindle orientation control. Additionally, expression of a FLAG-tagged, siRNA-resistant form of wild-type KAP0 (FLAG-KAP0 siR-WT) and a phosphomimetic mutant of KAP0 (FLAG-KAP0 siR-S83D) but not a nonphosphorylatable KAP0 mutant (FLAG-KAP0 siR-S83A) restored normal spindle orientation in KAP0-depleted cells (Fig. 3D and E). This indicates that KAP0 Ser83 phosphorylation is required for spindle orientation control. Furthermore, expression of KAP0-S83D but not KAP0-S83A rescued defects in spindle orientation induced by PCTK1 depletion (Fig. 3F and G). These results indicate that PCTK1 regulates spindle orientation through phosphorylation of KAP0 on Ser83.

FIG 3.

Ser83 phosphorylation of KAP0 is required for spindle orientation. (A) A schematic representation of KAP0 and its phosphorylation site (Ser83) by PCTK1. (B) Western blot analysis for KAP0 and control α-tubulin in cells transfected with Luci siRNA or two kinds of KAP0 siRNAs (KAP0-1 si and KAP0-2 si). (C) Distribution of (histogram) and average (inset; right) spindle angles and the percentage of metaphase cells with properly oriented spindles (inset; left) in the cells transfected with the indicated siRNAs as described for panel B. Results are means ± SEM from three different experiments; n = 51/experiment; P < 0.05 (*) and P < 0.01 (**) by Dunnett's multiple-comparison test. (D) Western blot analysis for KAP0, FLAG-KAP0, and control α-tubulin in the cells transfected with Luci siRNA or KAP0-1 siRNA together with vector alone, FLAG-KAP0 siR-WT, FLAG-KAP0 siR-S83A, or FLAG-KAP0 siR-S83D. (E) Distribution of (histogram) and average (inset, right) spindle angles and the percentage of metaphase cells with properly oriented spindles (inset; left) in the cells transfected with the indicated siRNA together with the indicated plasmids as described for panel D. Results are means ± SEM from three different experiments; n = 51/experiment; P < 0.05 (*) and P < 0.01 (**) by Dunnett's multiple-comparison test. (F) Western blot analysis for PCTK1, FLAG-KAP0, and control α-tubulin in the cells transfected with Luci siRNA or PCTK1-1 siRNA together with vector alone, FLAG-KAP0 siR-WT, FLAG-KAP0 siR-S83A, or FLAG-KAP0 siR-S83D. (G) Distribution of (histogram) and average (inset; right) spindle angles and the percentage of metaphase cells with properly oriented spindles (inset; left) in the cells transfected with the indicated siRNA together with the indicated plasmids as described for panel F. Results are means ± SEM from four different experiments; n = 51/experiment; P < 0.05 (*) and P < 0.01 (**) by t test.

KAP0 regulates spindle orientation in a PKA catalytic activity-independent manner.

Depletion of KAP0 had only a minor effect on the expression level of the PKA catalytic subunit (Fig. 3B). Because Ser83 is located within the dimerization/docking domain of KAP0 (Fig. 3A), we then examined whether Ser83 phosphorylation affected KAP0 dimerization or its capacity to bind PKA catalytic subunits. FLAG-tagged and Myc-tagged pairs of KAP0-WT, KAP0-S83A, and KAP0-S83D were coexpressed in all possible combinations in metaphase-synchronized HeLa cells, and then the cell extracts were subjected to immunoprecipitation with FLAG antibodies. Results demonstrated that each variant of KAP0 (WT, S83A, or S83D) was equally capable of immunoprecipitating all other forms of the protein (Fig. 4A). Additionally, endogenous PKA catalytic subunits coimmunoprecipitated with KAP0-WT, KAP0-S83A, and KAP0-S83D (Fig. 4B). These results indicate that phosphorylation of Ser83 of KAP0 has no effect on its dimerization or its capacity to bind PKA catalytic subunits. We then hypothesized that the spindle misorientation phenotype observed for KAP0-depleted cells was the result of an increase in the activity of PKA catalytic subunits, which may exist in the form of an active monomer in these cells due to the absence of KAP0. To examine this hypothesis, we measured the activation of PKA, which can be monitored by the state of CREB phosphorylation, in control cells and KAP0-depleted cells. Results demonstrated that depletion of KAP0 had only a minor effect on the activity of PKA in synchronized cells (Fig. 4C), suggesting the existence of counterpart proteins to KAP0 within the cells to regulate PKA activity. We then treated cells with the selective PKA inhibitor H-89 (36). The forskolin-induced activation of PKA was inhibited in a dose-dependent manner by H-89, with maximal inhibition achieved at a concentration of 20 μM (Fig. 4D). Additionally, the PKA activity was dramatically decreased by 20 μM H-89 in synchronized cells depleted of KAP0 (Fig. 4C). However, treatment of KAP0-depleted cells with 20 μM H-89 had no effect on the spindle misorientation phenotype in these cells (Fig. 4E and F), indicating that this defect is not a result of PKA activation. Therefore, KAP0 regulates spindle orientation by a mechanism that is independent of PKA activity.

FIG 4.

PKA catalytic activity is not involved in spindle orientation. (A) FLAG-KAP0 WT, FLAG-KAP0 S83A, or FLAG-KAP0 S83D was coexpressed with the indicated form of Myc-tagged KAP0 (WT, S83A, or S83D) in metaphase-synchronized HeLa cells. Total lysates were subjected to immunoprecipitation (IP) with anti-FLAG antibodies and analyzed by Western blotting with anti-FLAG and anti-Myc antibodies. (B) FLAG-KAP0 WT, FLAG-KAP0 S83A, or FLAG-KAP0 S83D was expressed in metaphase-synchronized HeLa cells. Total lysates were subjected to immunoprecipitation with anti-FLAG antibodies and analyzed by Western blotting with anti-FLAG and anti-PKA catalytic subunit antibodies. (C) HeLa cells transfected with the indicated siRNAs were synchronized by a double thymidine block. H-89 (20 μM) was added to the medium 9.5 h after the release. Total lysates were Western blotted with anti-phospho-Ser133-CREB (p-CREB) and CREB antibodies. (D) Total lysates of HeLa cells treated with forskolin (30 μM) for 10 min, followed by incubation with the indicated concentrations of H-89 for 30 min, were Western blotted with anti-p-CREB and CREB antibodies. (E) Western blot analysis for KAP0 and control α-tubulin in metaphase-synchronized cells transfected with Luci siRNA or KAP0-1 siRNA and treated with H-89 or left untreated. H-89 was added to the cell culture medium (20 μM) 9.5 h after the release of a double thymidine block, followed by incubation for an additional 30 min. (F) Distribution (histogram) and averages (inset; right) of spindle angles and the percentage of metaphase cells with properly oriented spindles (inset; left) in the cells transfected with the indicated siRNA and either left untreated or treated with H-89 (20 μM) for 30 min as described for panel C. Results are means ± SEM from three different experiments; n = 51/experiment; n.s., not significant; P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***) by Tukey's test.

Ser83-phosphorylated KAP0 interacts with the FERM domain of myosin X.

We next sought to determine the mechanism by which KAP0 regulates spindle orientation. KAP0 previously has been shown to bind the band 4.1/ezrin/radixin/moesin (FERM) domain of myosin VIIa in a yeast two-hybrid system (37). However, our reverse transcription-PCR and Western blot analysis revealed that myosin VIIa expression is detectable at only very low levels in HeLa cells (unpublished observations). Myosin constitutes a diverse superfamily of actin-based motor proteins (38–40). All myosin proteins have head (motor) domains, which contain ATP and actin-binding sites, and tail domains, which are structurally unique and which confer class-specific functions. The tail domains of both myosin VIIa and myosin X contain a FERM domain as well as a myosin tail homology 4 (MyTH4) domain (Fig. 5A) (41–43). We have previously shown that myosin X plays a role in spindle orientation in HeLa cells by remodeling the actin cytoskeleton during mitosis (9). Additionally, myosin X is required for nuclear anchoring to the cell cortex in Xenopus oocytes (25). To examine the possibility that KAP0 binds to the FERM domain of myosin X, we performed immunoprecipitation assays in M phase-synchronized HeLa cells coexpressing a GFP-tagged FERM domain of myosin X (GFP-MyoX-FERM) along with either FLAG-KAP0 WT, FLAG-KAP0 S83A, or FLAG-KAP0 S83D. In these experiments, both FLAG-KAP0 WT and FLAG-KAP0 S83D coimmunoprecipitated with GFP-MyoX-FERM, whereas FLAG-KAP0 S83A did not (Fig. 5B). Additionally, while FLAG-KAP0 S83D coimmunoprecipitated with GFP-MyoX-FERM, it did not do so with a GFP-tagged MyTH4 domain of myosin X (GFP-MyoX-MyTH4) (Fig. 5C). These results indicate that KAP0 binds to the FERM domain of myosin X in a Ser83 phosphorylation-dependent manner.

FIG 5.

Phospho-Ser83 KAP0 binds to myosin X FERM domain. (A) Schematic representations of myosin VIIa (MyoVIIa) and myosin X (MyoX). a.a., amino acid. (B) FLAG-KAP0 WT, FLAG-KAP0 S83A, or FLAG-KAP0 S83D was coexpressed with GFP alone or with the GFP-tagged myosin X FERM domain (GFP-MyoX-FERM) in metaphase-synchronized HeLa cells. Total lysates were subjected to immunoprecipitation with anti-GFP antibodies and analyzed by Western blotting with anti-GFP and anti-FLAG antibodies. (C) FLAG-KAP0 S83D was coexpressed with GFP alone, GFP-MyoX FERM, or GFP-MyoX MyTH4 in metaphase-synchronized HeLa cells. Total lysates were subjected to immunoprecipitation with anti-GFP antibodies and analyzed by Western blotting with anti-GFP and anti-FLAG antibodies. (D) Western blot analysis for MyoX and control α-tubulin in metaphase-synchronized cells transfected with Luci siRNA or MyoX siRNA. (E) Images of metaphase cells transfected with Luci siRNA, MyoX siRNA, KAP0-1 siRNA, or PCTK1-1 siRNA and stained with anti-MyoX antibodies (green) and phalloidin (red). Images in boxed areas are enlarged. (F) Images of metaphase cells stained with anti-MyoX (red) and anti-KAP0 (green) antibodies. endo, endogenous. (G) Images of metaphase cells transfected with vector alone, FLAG-KAP0 WT, FLAG-KAP0 S83A, or FLAG-KAP0 S83D and stained with anti-MyoX (red) and anti-FLAG (green) antibodies.

We next investigated the subcellular localization of myosin X during mitosis. As HeLa cells round up during mitosis, they remain connected to the ECM via retraction fibers, fibrous structures filled with actin filaments (44). Immunolabeling of endogenous myosin X revealed punctate signals at the tips and along the length of retraction fibers in control metaphase cells (Fig. 5E, Luci si). These signals were diminished in cells depleted of myosin X by RNAi (Fig. 5D and E, MyoX si), confirming the specificity of the antibody. In cells depleted of KAP0 or PCTK1, a faint myosin X signal still could be observed at the tips of retraction fibers, but immunolabeling along the length of these fibers was greatly reduced (Fig. 5E, KAP0-1si and PCTK1-1si). Additionally, retraction fibers appeared disorganized and less uniform in the KAP0- or PCTK1-depleted cells. Moreover, endogenous KAP0 was closely associated with myosin X at the tips and along the retraction fibers in metaphase cells (Fig. 5F). Similar results were obtained with FLAG-KAP0 WT and FLAG-KAP0 S83D but not with FLAG-KAP0 S83A (Fig. 5G). These results demonstrate that phosphorylation of Ser83 regulates KAP0 binding to myosin X and suggest that KAP0 regulates the functions of myosin X at retraction fibers.

KAP0 enhances myosin X-β1 integrin interaction to regulate spindle orientation.

Myosin X previously has been shown to bind β1 integrin via its FERM domain, providing a link between integrins and the actin cytoskeleton during filopodium formation (45). Because myosin X localizes to retraction fibers in a KAP0-dependent manner (Fig. 5E), we hypothesized that KAP0 enhances myosin X-β1 integrin binding to promote the coupling of integrins to the actin cytoskeleton during retraction fiber formation. In support of this hypothesis, immunolabeling revealed β1 integrin at the tips and along the length of retraction fibers in control metaphase cells that express GFP-MyrPalm (46) to visualize plasma membrane/retraction fibers, whereas in cells depleted of KAP0 or PCTK1, signals of β1 integrin on these fibers were greatly reduced (Fig. 6A), just like that of myosin X (Fig. 5E). Additionally, myosin X was colocalized with β1 integrin on retraction fibers in control cells, whereas in cells depleted of KAP0 or PCTK1, this colocalization was barely detected (Fig. 6B). Moreover, endogenous β1 integrin coimmunoprecipitated with both GFP-MyoX-FERM (Fig. 6C) and endogenous myosin X (Fig. 6D) far more efficiently in control cells than in KAP0-depleted cells following synchronization in metaphase. Furthermore, the known defects in spindle orientation observed following depletion of β1 integrin (9) could be rescued following expression of an siRNA-resistant, FLAG-tagged form of wild-type β1 integrin (Fig. 6E and F, FLAG-Int siR-WT) but not following expression of the β1 integrin-W775A (Fig. 6E and F, FLAG-Int siR-W775A) mutant that is defective in myosin X binding (45). These results demonstrate that Ser83-phosphorylated KAP0 promotes the interaction between the FERM domain of myosin X and β1 integrin and that this association is required for β1 integrin-dependent spindle orientation.

FIG 6.

Phospho-Ser83 KAP0 enhances the interaction between myosin X-FERM and β1 integrin to regulate spindle orientation. (A) Images of GFP-MyrPalm (green)-expressing metaphase cells transfected with Luci siRNA, KAP0-1 siRNA, or PCTK1-1 siRNA and stained with anti-β1 integrin (red) antibodies. (B) Images of metaphase cells transfected with Luci siRNA, KAP0-1 siRNA, or PCTK1-1 siRNA and stained with anti-MyoX (red) and anti-β1 integrin (green) antibodies. (C) HeLa cells were transfected with Luci siRNA or KAP0-1 siRNA together with GFP alone or GFP-MyoX-FERM and synchronized in metaphase. Total lysates were subjected to immunoprecipitation with anti-β1 integrin antibodies and analyzed by Western blotting with anti-KAP0, anti-GFP, and anti-β1 integrin antibodies. (D) HeLa cells were transfected with Luci siRNA or KAP0-1 siRNA and synchronized in metaphase. Total lysates were subjected to immunoprecipitation with anti-β1 integrin antibodies and analyzed by Western blotting with anti-MyoX, anti-KAP0, and anti-β1 integrin antibodies. (E) Western blot analysis for β1 integrin, FLAG-tagged β1 integrin, and control α-tubulin in metaphase-synchronized cells transfected with Luci siRNA or β1 integrin siRNA together with vector alone, FLAG-β1 integrin siR-WT, or FLAG-β1 integrin siR-W775A. (F) Distribution of (histogram) and average (inset; right) spindle angles and the percentage of metaphase cells with properly oriented spindles (inset; left) in the cells transfected with the indicated siRNA together with the indicated plasmids as described for panel E. Results are means ± SEM from three different experiments; n = 51/experiment. P < 0.05 (*) and P < 0.01 (**) by t test. (G) A model of the PCTK1-KAP0-myosin X-β1 integrin pathway that provides a link between ECM and the retraction fibers in the ECM-dependent control of spindle orientation.

DISCUSSION

As cells round up during mitosis, the ECM remains connected to the intracellular actin cytoskeleton via retraction fibers (44). These fibers have been proposed to segregate cortical components (47) and to exert forces on the mitotic cell body. This in turn induces variations in the dynamics of cortical actin, leading to the alignment of the mitotic spindle with ECM polarity cues (48). However, the molecular nature of the mechanism linking ECM and the actin cytoskeleton during spindle orientation has remained unclear. Our results demonstrate that PCTK1-KAP0-myosin X-β1 integrin is the functional module that provides this mechanism based on the following five observations: first, PCTK1-mediated phosphorylation of KAP0 on Ser83 is required for spindle orientation; second, myosin X binds to phospho-Ser83 KAP0; third, KAP0 is required for the interaction between myosin X and β1 integrin at retraction fibers; fourth, the interaction between myosin X and β1 integrin is required for normal spindle orientation; and fifth, KAP0 associates with myosin X on retraction fibers. Myosin X is known to bind microtubules and regulate spindle positioning in Xenopus oocytes (25), which could raise the hypothesis that myosin X anchors microtubules to retraction fibers to align the spindles in adherent cells. However, inconsistent with this hypothesis, we did not see microtubules along the retraction fibers (data not shown), suggesting that myosin X exerts its function on spindle orientation via a microtubule-independent mechanism. We propose that myosin X binds to β1 integrin in a phospho-Ser83 KAP0-dependent manner at the tips of retraction fibers, which links the ECM with the intercellular actin cytoskeleton to stabilize the retraction fibers (Fig. 6G). This would lead to the accumulation of cortical components on the cell body at the end of these fibers, which would frame cortical cues for spindle orientation. The molecular nature of the mechanisms that capture microtubules at the end of the retraction fibers, which possibly include dynein/dynactin complexes (10, 11), should be clarified in future studies.

PCTK1 activity is low during G₀/G₁ phase and increases during S and G₂ phases (17). Phosphoproteomic analysis shows that KAP0 Ser83 phosphorylation is significantly higher in M phase-arrested control cells than in PCTK1 knockdown cells (Fig. 2), indicating that PCTK1 remains active during mitosis. It should be noted that PCTK1 activity is suppressed by PKA-dependent phosphorylation of Ser119 and Ser153 (19). Phosphorylation of Ser119 generates a 14-3-3 biding site, while phosphorylation of Ser153 inhibits association of PCTK1 with cyclin Y, and both of these phosphorylation events inhibit the function of PCTK1 (19, 23). It has been reported that downregulation of PKA is implicated in the initiation of mitosis (49), and at metaphase-to-anaphase transition, PKA is upregulated to stimulate exit from mitosis (50). Therefore, it is possible that the downregulation of PKA activity during mitosis ensures the activation of PCTK1; consequently, it is necessary for PCTK1-dependent spindle orientation control. Our observation that PCTK1 regulates spindle orientation through KAP0, a regulatory subunit of PKA, via a mechanism that is independent of PKA catalytic activity is consistent with this idea (Fig. 4C, E, and F).

PCTK1 knockout mice developed normally, but male mice are infertile owing to defects in spermatogenesis. Given the growing evidence for the importance of spindle orientation in spermatogenesis in Drosophila melanogaster and mammals (51–55), it would be interesting to examine the possible role of PCTK1-KAP0-mediated spindle orientation control in spermatocyte differentiation.