CDK1-mediated phosphorylation of EPLIN is crucial for mitotic spindle positioning Authors: Jia-Rong Fan, Yun Hsieh, Chien-Chen Lai, Hong-Chen Chen Published: 2026 ABSTRACT Correct anchoring of the mitotic spindle to the cell cortex is essential for proper cell division. However, how the architecture of the cell cortex is coordinately remodeled during mitotic progression remains poorly understood. Here, we show that depletion of EPLIN (an actin-bundling protein) causes spindle deviation and an arrest in metaphase, accompanied with abnormal cortical actin accumulation and blebbing. We identify the Ser203 of EPLIN as the major phosphorylation site for CDK1 in mitosis and demonstrate that this phosphorylation event is required for EPLIN to maintain proper actin cortex and correct spindle positioning. Moreover, EPLIN phosphorylation at Ser203 enhances its capability to reorganize branched actin filaments into astral structures in vitro, with which EPLIN may sequester actin bundles and the Arp2/3 complex, and thereby prevent excessive de novo actin polymerization. Thus, this study unveils an important role for EPLIN in spindle positioning via its function to maintain proper architecture of the actin cortex during mitosis. Introduction Symmetric spindle positioning ensures proper chromosome alignment, equal chromosome separation, and timely mitotic progression1–3. In contrast, mispositioning of the mitotic spindle possibly leads to spindle deviation, aneuploidy, and tumorigenesis. Correct orientation and positioning of the mitotic spindle are known to depend on the anchoring between the cell cortex and astral microtubules emanating from the centrosomes. Proper mitotic spindle orientation relies on polarized anchoring of force generators to the cell cortex by the evolutionarily conserved protein complex formed by the Gαi subunit of heterotrimeric G proteins, the Leucine-Glycine-Asparagine repeat protein (LGN), and the nuclear mitotic apparatus protein (NuMA) at the lateral cortex to ensure cortical anchoring of dynein-dynactin and astral microtubules and thereby planar alignment of the mitotic spindle2. The cell cortex, also known as the actin cortex, is a specialized layer (in the range of 100 to 1000 nm thick) of actin-rich network consisting of F-actin filaments, myosin motors, and actin-binding proteins underlying the inner face of the plasma membrane4. It is attached to the plasma membrane via membrane-anchoring proteins called ERM (ezrin, radixin and moesin) proteins and continuously remodeled by polymerization, depolymerization, and branching5. The cell cortex not only plays a central role in cell shape control, but also coordinates the progression of mitosis6,7. Both the initiation and progression of mitosis depend on the remodeling of the cell cortex. Upon entry into the metaphase, the cell cortex undergoes a remodeling, which generates cortical tension to drive mitotic cell rounding and allows the anchoring of the mitotic spindle. Although the mechanism of how the cell cortex is reorganized in metaphase remains elusive, emerging evidence suggests that cell cycle-dependent kinase 1 (CDK1) and various molecules known to regulate the actin cytoskeleton are involved. mDia 1 (Diaphanous-related formin 1), a member of the protein family called the formins, has potent actin nucleation and elongation activities8. The Arp2/3 complex nucleates actin filament branches on the sides of preexisting actin filaments9. Both mDia1 and the Arp2/3 complex are involved in the architecture of the mitotic cortex10. Additionally, CDK1 activates the accumulation of cortical myosin II (a motor that generates contractility) through Rho-associated kinase (ROCK) upon entry into mitosis, enabling cells to round up against mechanical confinement11,12. Subsequently, CDK1 maintains constant cortical tension by phosphorylating and thereby inhibiting the activity of formin, which prevents over-production of F-actin filaments and excessive cortical tension13. Besides, LIM kinase, a LIM-domain containing Ser/Thr protein kinase that is activated through phosphorylation by either ROCK or p21-activated kinase-1 (PAK1), which are downstream effectors of Rho and Rac, was shown to regulate cell cortex stability and precise spindle positioning via phosphorylating cofilin (an F-actin severing protein) during metaphase14. Consistently, depletion of cofilin increases cortical thickness and contractility in mitotic cells15. Importantly, cortical tension is determined not only by myosin II activity and F-actin length, but also crucially by the connectivity of the actin network (i.e., cross-linking and alignment of F-actin filaments), which is accomplished by various classes of actin bundling proteins to fulfill various roles in interphase cells16–18. Two actin bundlers, α-actinin and fascin, were shown to contribute to the control of mitotic rounding upon entry into mitosis19,20. EPLIN (Epithelial Protein Lost In Neoplasm) was first discovered and recognized as a putative tumor suppressor21,22. Its gene expression is down-regulated in many types of tumors and often associated with their malignant progression23. In humans, two EPLIN isoforms, α (600 amino acids) and β (759 amino acids), are expressed, which are generated from two promoters of the same gene and differ only in their N-terminus24,25. The tissue distribution and gene regulation of EPLIN-α and -β are different, but whether they perform distinct cellular functions is unclear24,26. Both EPLIN isoforms contain a central LIM domain (a protein-binding interface) and two flanking actin-binding regions, which thereby denote its gene name as Lima1 (LIM domain and actin-binding protein 1)26,27. While the function of its LIM domain is unclear, EPLIN is able to crosslink actin filaments into bundles, inhibit actin depolymerization, and reduce the nucleation of branched actin, as demonstrated by in vitro biochemical experiments27. The tumor-suppressive function of EPLIN is mostly attributed to its ability to regulate cell-cell adhesion and cell motility. EPLIN maintains adherent junctions by linking E-cadherin-catenin to F-actin, which prevents the initiation of cellular individualization and angiogenesis during epithelial-to-mesenchymal transition28. Moreover, EPLIN inhibits cell mobility by sustaining focal adhesions or terminating lamellipodial protrusions29–31. Upon stimulation by epidermal growth factor, EPLIN is phosphorylated at two serine sites in the N-terminal half (i.e., Ser203 in the α or Ser362 in the β) and C-terminal half (i.e., Ser445 in the α or Ser604 in the β) by ERK32 and dephosphorylated by CDC14A (a dual specificity protein phosphatase)33. The C-terminal phosphorylation of EPLIN induces its protein degradation, decreases its binding to actin bundles at adherent junctions, and relocates it from stress fibers towards lamellipodia, accompanied by increased cell mobility34,35. Moreover, increasing evidence also implicates that EPLIN is involved in other cellular activities, such as endocytosis36, cholesterol uptake37, cilia formation38, mitochondrial energetics39, and gene expression22,40. It should be noted that the expression of EPLIN is not confined to epithelial tissues. Instead, it is broadly expressed in various types of cells, particularly during early embryonic development. Thus, it has been suggested to play a fundamental role in cell division or tissue genesis39. Embryonic stem cells with EPLIN gene knockout exhibit reduced proliferation rate and give rise to smaller teratomas39. During mitosis, EPLIN is specifically enriched at the cell cortex in metaphase and then redistributed to the cleavage furrow and the contractile ring during cytokinesis41. EPLIN has been shown to be required for cytokinesis41. In addition, during oocyte meiosis, the spindle moves to the oocyte cortex, and the cortical actin accumulates to form an actin cap that overlays the spindle. EPLIN was found to accumulate on the actin cap, which plays a critical role in asymmetric division39. However, the role of EPLIN in metaphase for symmetrical cell division remains unexplored. In this study, we show that suppression of EPLIN expression causes spindle deviation and an arrest in metaphase, accompanied by abnormal cortical actin accumulation and blebbing. Moreover, we find that the phosphorylation of EPLIN at Ser203 by CDK1 is required for EPLIN to maintain proper actin cortex and precise spindle positioning in metaphase. Our results of biochemical assays suggest that EPLIN phosphorylation at Ser203 may increase its capability to inhibit Arp2/3-mediated actin polymerization and thereby ensure proper architecture of the actin cortex for spindle positioning. Results Depletion of EPLIN causes spindle deviation and an arrest in metaphase To examine the potential role of EPLIN in metaphase, EPLIN (both α and β isoforms) was depleted by the short-hairpin RNA (shRNA) approach in HeLa cells that predominantly express EPLIN α isoform (Fig. 1A). The HeLa cells were stably transfected with the plasmids encoding mCherry-histone 2B (H2B) and GFP-tubulin α to facilitate live cell imaging. We found that the depletion of EPLIN significantly prolonged the duration of mitosis (Fig. 1B, C), which mainly resulted from an arrest in metaphase (Fig. 1D). Our results of flow cytometry analysis also revealed that EPLIN depletion increased the ratios of cells in G2/M phases (Supplementary Fig. 1), supporting an arrest in mitosis. Besides, it also showed a higher subset of sub-G1 phase, suggesting that the prolonged metaphase arrest by EPLIN depletion may cause cell death. A small subset of the cell population with a karyotype > 2 N was noted (Supplementary Fig. 1), which may result from cytokinesis failure.Fig. 1Depletion of EPLIN causes spindle deviation and an arrest in metaphase.A HeLa cells stably expressing GFP-α-tubulin and mCherry-Histone 2B (H2B) were infected with lentiviruses encoding shRNAs specific to EPLIN (sh-EPLIN; clones #1 and #2) or luciferase (sh-Luc.). An equal amount of whole cell lysates was analyzed by immunoblotting with the indicated antibodies. B–D The cells as described in (A) were monitored by live cell imaging for 24 h. Representative images of the cells undergoing mitosis are shown (B). Scale bars, 10 μm. The duration of mitosis (C) and different mitotic phases (D) were measured. Prophase and prometaphase are respectively determined by chromosomal condensation and the accumulation of α-tubulin around released chromosomes after nuclear envelope breakdown. Metaphase is determined by the alignment of chromosomes and the attachment of bipolar spindles. Anaphase and telophase are determined as the separation of sister chromatids, then the reformation of the nuclear membrane and finally the decondensation of chromosomes into their interphase conformations. In (C), the total number of mitotic cells in each group is 75 (control), 70 (sh-Luciferase), 65 (sh-EPLIN #1), and 73 (sh-EPLIN #2) from 3 independent experiments. In (D), the total number of mitotic cells in each group is 70 (control), 23 (sh-Luciferase), 28 (sh-EPLIN #1), and 73 (sh-EPLIN #2) from 3 independent experiments. Abbreviations used are: Prophase (Pro.), Prometaphase (Prometa.), Metaphase (Meta.), Anaphase (Ana.), and Telophase (Telo.). E HeLa cells were infected with lentivirus encoding shRNAs specific to EPLIN (sh-EPLIN) or luciferase (sh-Luc.). To synchronize the cells in metaphase, the cells were arrested in G2 phase by treatment with the CDK1 inhibitor RO-3306 (10 μM) for 20 h and then released for 30 min into metaphase. The cells were fixed and stained for α-tubulin (red), EPLIN (green), and DNA (blue). Representative images acquired by epifluorescence microscopy are shown. White dashed lines mark the profile of the mitotic cells. Scale bars, 10 μm. F Schematic diagram illustrates how spindle deviation is quantified by measuring the difference in the “pole to cortex”. G, H The cells as described in (E) were synchronized in metaphase and stained for α-tubulin, EPLIN, and DNA. The difference in the “pole to cortex” (G) and the percentage of the cells with different extents of spindle deviation (H) were measured (n ≥ 140 cells). Data information: In (C, D), values (median ± interquartile range) are from three independent experiments. Values in G are presented as the box-whisker plot from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. Depletion of EPLIN causes spindle deviation and an arrest in metaphase. A HeLa cells stably expressing GFP-α-tubulin and mCherry-Histone 2B (H2B) were infected with lentiviruses encoding shRNAs specific to EPLIN (sh-EPLIN; clones #1 and #2) or luciferase (sh-Luc.). An equal amount of whole cell lysates was analyzed by immunoblotting with the indicated antibodies. B–D The cells as described in (A) were monitored by live cell imaging for 24 h. Representative images of the cells undergoing mitosis are shown (B). Scale bars, 10 μm. The duration of mitosis (C) and different mitotic phases (D) were measured. Prophase and prometaphase are respectively determined by chromosomal condensation and the accumulation of α-tubulin around released chromosomes after nuclear envelope breakdown. Metaphase is determined by the alignment of chromosomes and the attachment of bipolar spindles. Anaphase and telophase are determined as the separation of sister chromatids, then the reformation of the nuclear membrane and finally the decondensation of chromosomes into their interphase conformations. In (C), the total number of mitotic cells in each group is 75 (control), 70 (sh-Luciferase), 65 (sh-EPLIN #1), and 73 (sh-EPLIN #2) from 3 independent experiments. In (D), the total number of mitotic cells in each group is 70 (control), 23 (sh-Luciferase), 28 (sh-EPLIN #1), and 73 (sh-EPLIN #2) from 3 independent experiments. Abbreviations used are: Prophase (Pro.), Prometaphase (Prometa.), Metaphase (Meta.), Anaphase (Ana.), and Telophase (Telo.). E HeLa cells were infected with lentivirus encoding shRNAs specific to EPLIN (sh-EPLIN) or luciferase (sh-Luc.). To synchronize the cells in metaphase, the cells were arrested in G2 phase by treatment with the CDK1 inhibitor RO-3306 (10 μM) for 20 h and then released for 30 min into metaphase. The cells were fixed and stained for α-tubulin (red), EPLIN (green), and DNA (blue). Representative images acquired by epifluorescence microscopy are shown. White dashed lines mark the profile of the mitotic cells. Scale bars, 10 μm. F Schematic diagram illustrates how spindle deviation is quantified by measuring the difference in the “pole to cortex”. G, H The cells as described in (E) were synchronized in metaphase and stained for α-tubulin, EPLIN, and DNA. The difference in the “pole to cortex” (G) and the percentage of the cells with different extents of spindle deviation (H) were measured (n ≥ 140 cells). Data information: In (C, D), values (median ± interquartile range) are from three independent experiments. Values in G are presented as the box-whisker plot from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. Next, the cells were subjected to immunofluorescence staining for EPLIN, α-tubulin, and DNA. We noticed that EPLIN depletion caused a deviation of the mitotic spindle from the central axis of the cell (Fig. 1E and Supplementary Fig. 2). To quantify the deviation of the mitotic spindle, we measured the difference in the distance between two spindle poles to the cell cortex (Fig. 1F). The cells with an evident bipolar spindle were selected for the measurement. The cell boundaries were delineated by enhanced intensity of α-tubulin (Supplementary Fig. 2A). The median of the difference in the “pole to cortex” was 1.2 μm in the control HeLa cells and 3.2 μm in EPLIN-depleted cells (Fig. 1G). In ~50% of EPLIN-depleted cells, the difference in the “pole to cortex” was more than 3 μm (Fig. 1H). The spindle deviation was further confirmed by measuring the cells with two evident spindle poles (using γ-tubulin as a centrosomal marker) on the same microscopic focal plane (Supplementary Fig. 2B). Since the spindle pole (i.e., centrosome) is a relatively small organelle (1.32±0.27 μm in diameter at mitosis), this new criterion theoretically allows us to select cells with minimal spindle tilting. We found that using this method, the median spindle deviation was 0.9 μm in the control cells and 2.8 μm in EPLIN-depleted cells (Supplementary Fig. 2B). These results indicate that EPLIN is crucial for symmetric spindle positioning in metaphase. Furthermore, our live-cell imaging revealed that EPLIN depletion caused oscillatory spindle motion rather than a persistent off-center displacement (Supplementary Fig. 3), suggesting that the observed deviation may arise from a dynamic imbalance in cortical pulling forces and/or an unstable cortical anchoring of the mitotic spindle. The bilateral localization of the Gαi-LGN-NuMA complex at the mitotic cortex ensures cortical anchoring of mitotic spindle2. We found that NuMA was distributed at the polar areas of the mitotic spindle and the adjacent cell cortex, where it was co-localized with EPLIN (Supplementary Fig. 4A). EPLIN depletion did not affect NuMA’s protein level (Supplementary Fig. 4B) or its association with mitotic spindle (Supplementary Fig. 4C), but significantly (~40%) decreased the bilateral distribution of NuMA at the cell cortex (Supplementary Fig. 4C and D), supporting that the observed spindle deviation may at least partially result from an unstable cortical anchoring of the mitotic spindle. Phosphorylation of EPLIN at Ser203 by CDK1 is required for correct spindle positioning and mitotic progression The mobility of EPLIN proteins (both α and β isoforms) on the immunoblot was altered along with cell cycle progression (Fig. 2A). The bands of EPLIN proteins on the immunoblot apparently shifted up upon cellular synchronization in mitosis (M phase), which then shifted down upon mitotic exit to the G1 phase and reached to the lowest position in the S phase (Fig. 2A). Such a mobility-shift on the immunoblot suggests that EPLIN may undergo certain post-translational modifications during cell cycle progression. The result of the Phospho-tagTM staining indicated that EPLIN proteins from the HeLa cells synchronized in mitosis had a higher level of phosphorylation than those from asynchronous cells or the cells in the G1 phase (Fig. 2B).Fig. 2The Ser203 of EPLIN is the major phosphorylation site for CDK1 in mitosis.A HeLa cells remained asynchronous (Asyn.) or were synchronized in a specific phase of the cell cycle. To synchronize cells in M phase, the cells were treated with nocodazole at 200 ng/ml for 16 h; in G1 phase, the cells were treated with nocodazole at 200 ng/ml for 16 h and then released for 4 h; in S phase, the cells were treated with thymidine at 1 mM for 24 h; G2 phase, the cells were treated with RO-3306 at 10 μM for 20 h. An equal amount of whole-cell lysate was analyzed by immunoblotting with the indicated antibodies. Note that EPLIN proteins alter their motility on the immunoblot along with the progression of the cell cycle. B HeLa cells remained asynchronous (Asyn.) or were synchronized in M phase or G1 phase. EPLIN was immunoprecipitated (IP) from an equal amount of whole cell lysates (WCL) with anti-EPLIN antibody. The immunocomplexes were fractionated by SDS-polyacrylamide gel electrophoresis, and the gel was stained with Phospho-TagTM phosphoprotein gel stain, according to the manufacturer’s instructions, or analyzed by immunoblotting (IB) with anti-EPLIN. An equal amount of WCL was analyzed by immunoblotting with anti-cyclin B1 or anti-actin. C The scheme diagram illustrates the serine phosphorylation sites on EPLIN α in mitosis identified by mass spectrometry. D HeLa cells were transfected with the vector alone (FLAG) or the plasmids encoding FLAG-EPLIN α WT or the mutants for 24 h. The cells remained asynchronous (Asyn.) or were synchronous in M phase. An equal amount of whole cell lysates (WCL) was analyzed by immunoblotting with the indicated antibodies. Note that the position of the S203A mutant on the immunoblot is lower than that of the WT and other mutants. E HeLa cells were transfected with the vector alone (-) or the plasmids encoding FLAG-EPLIN α (WT or S203A) for 24 h. The cells remained asynchronous (Asyn.) or were synchronized in M phase. FLAG-EPLIN was immunoprecipitated (IP) with anti-FLAG antibody and analyzed by immunoblotting (IB) with anti-phospho-Ser/Thr antibody (p-S/T) or anti-EPLIN. An equal amount of whole cell lysates was analyzed by immunoblotting with anti-cyclin B1. The relative level of the Ser/Thr phosphorylation of FLAG-EPLIN α (WT vs. S203A) in M phase was measured. F HeLa cells were transfected with the vector alone (-) or the plasmids encoding FLAG-EPLIN α (WT or S203A) for 24 h. FLAG-EPLIN was immunoprecipitated (IP) with anti-FLAG antibody and analyzed by immunoblotting (IB) with a customized antibody specific to S203-phosphorylated EPLIN (EPLIN pS203). G HeLa cells remained asynchronous (Asyn.) or were synchronized in M or G1 phase. An equal amount of whole cell lysates was analyzed by immunoblotting with the indicated antibodies. H The in vitro kinase assay was carried out in the presence of recombinant CDK1/cyclin B1 proteins and p-Nitrobenzyl mesylate (PNBM; 1 mM) using purified recombinant T7-tagged N-terminal region (aa. 1–230) of EPLIN α (T7-EPLIN-α-N; WT and S203A) as the substrate, as described in Materials and Methods. The extent of T7-EPLIN-α-N phosphorylation by CDK1 was analyzed by immunoblotting with antibodies specific to Thiophosphate ester or T7-epitope tag. The relative level of the Thiophosphate ester of T7-EPLIN-α-N (WT vs. S203A) was measured. Data information: In (E, H), values (mean ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. The Ser203 of EPLIN is the major phosphorylation site for CDK1 in mitosis. A HeLa cells remained asynchronous (Asyn.) or were synchronized in a specific phase of the cell cycle. To synchronize cells in M phase, the cells were treated with nocodazole at 200 ng/ml for 16 h; in G1 phase, the cells were treated with nocodazole at 200 ng/ml for 16 h and then released for 4 h; in S phase, the cells were treated with thymidine at 1 mM for 24 h; G2 phase, the cells were treated with RO-3306 at 10 μM for 20 h. An equal amount of whole-cell lysate was analyzed by immunoblotting with the indicated antibodies. Note that EPLIN proteins alter their motility on the immunoblot along with the progression of the cell cycle. B HeLa cells remained asynchronous (Asyn.) or were synchronized in M phase or G1 phase. EPLIN was immunoprecipitated (IP) from an equal amount of whole cell lysates (WCL) with anti-EPLIN antibody. The immunocomplexes were fractionated by SDS-polyacrylamide gel electrophoresis, and the gel was stained with Phospho-TagTM phosphoprotein gel stain, according to the manufacturer’s instructions, or analyzed by immunoblotting (IB) with anti-EPLIN. An equal amount of WCL was analyzed by immunoblotting with anti-cyclin B1 or anti-actin. C The scheme diagram illustrates the serine phosphorylation sites on EPLIN α in mitosis identified by mass spectrometry. D HeLa cells were transfected with the vector alone (FLAG) or the plasmids encoding FLAG-EPLIN α WT or the mutants for 24 h. The cells remained asynchronous (Asyn.) or were synchronous in M phase. An equal amount of whole cell lysates (WCL) was analyzed by immunoblotting with the indicated antibodies. Note that the position of the S203A mutant on the immunoblot is lower than that of the WT and other mutants. E HeLa cells were transfected with the vector alone (-) or the plasmids encoding FLAG-EPLIN α (WT or S203A) for 24 h. The cells remained asynchronous (Asyn.) or were synchronized in M phase. FLAG-EPLIN was immunoprecipitated (IP) with anti-FLAG antibody and analyzed by immunoblotting (IB) with anti-phospho-Ser/Thr antibody (p-S/T) or anti-EPLIN. An equal amount of whole cell lysates was analyzed by immunoblotting with anti-cyclin B1. The relative level of the Ser/Thr phosphorylation of FLAG-EPLIN α (WT vs. S203A) in M phase was measured. F HeLa cells were transfected with the vector alone (-) or the plasmids encoding FLAG-EPLIN α (WT or S203A) for 24 h. FLAG-EPLIN was immunoprecipitated (IP) with anti-FLAG antibody and analyzed by immunoblotting (IB) with a customized antibody specific to S203-phosphorylated EPLIN (EPLIN pS203). G HeLa cells remained asynchronous (Asyn.) or were synchronized in M or G1 phase. An equal amount of whole cell lysates was analyzed by immunoblotting with the indicated antibodies. H The in vitro kinase assay was carried out in the presence of recombinant CDK1/cyclin B1 proteins and p-Nitrobenzyl mesylate (PNBM; 1 mM) using purified recombinant T7-tagged N-terminal region (aa. 1–230) of EPLIN α (T7-EPLIN-α-N; WT and S203A) as the substrate, as described in Materials and Methods. The extent of T7-EPLIN-α-N phosphorylation by CDK1 was analyzed by immunoblotting with antibodies specific to Thiophosphate ester or T7-epitope tag. The relative level of the Thiophosphate ester of T7-EPLIN-α-N (WT vs. S203A) was measured. Data information: In (E, H), values (mean ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. Indeed, our results of mass spectrometry revealed that six serine residues (Ser203, Ser214, Ser215, Ser331, Ser445, and Ser450) of endogenous EPLIN α are phosphorylated during mitosis (Fig. 2C and Supplementary Fig. 5). To examine which of these residues is the major phosphorylation site of EPLIN in mitosis, the corresponding Ser residues of FLAG epitope-tagged EPLIN α (FLAG-EPLIN α) were substituted with Ala. The substitution of Ser203, but not on other Ser residues, abolished the mitosis-induced mobility shift of FLAG-EPLIN α on the immunoblot (Fig. 2D), suggesting that Ser203 may be the major phosphorylation site of EPLIN α in mitosis. The S203A substitution also abolished the mitosis-induced phosphorylation of FLAG-EPLIN α, as detected by immunoblotting with a commercially available anti-phospho-Ser/Thr antibody (anti-p-S/T) (Fig. 2E). Next, a customized antibody specific to S203-phosphorylated EPLIN (anti-EPLIN pS203) was generated (Fig. 2F). Using this antibody, we demonstrated that endogenous EPLIN α was phosphorylated at Ser203 during mitosis, but much less in the G1 phase (Fig. 2G). Moreover, Ser203 is within a minimal consensus motif for CDK1 phosphorylation. Indeed, purified CDK1/Cyclin B1 complexes were capable of phosphorylating T7 epitope-tagged the N-terminal region (aa. 1-230) of EPLIN α (T7-EPLIN α-N) in vitro (Fig. 2H). The S203 A substitution significantly (~50%) decreased the phosphorylation of the T7-EPLIN α-N construct by CDK1 in vitro (Fig. 2H). These results together suggest that Ser203 of EPLIN α may be the major phosphorylation site for CDK1 in mitosis. To examine the significance of EPLIN Ser203 phosphorylation in mitosis, FLAG-EPLIN α and its mutants (S203A and S203D) were reintroduced into EPLIN-depleted HeLa cells (Fig. 3A). FLAG-EPLIN α and the S203D mutant (the phosphomimetic mutant), but not the S203A mutant, restored the mitosis duration (Fig. 3B, C) and symmetric spindle positioning (Fig. 3D–F) in EPLIN-depleted cells. These results together suggest that phosphorylation of EPLIN at Ser203 is important for symmetric spindle positioning in metaphase.Fig. 3Phosphorylation of EPLIN at Ser203 may be required for proper spindle positioning and mitotic progression.A HeLa cells stably expressing GFP–α-tubulin and mCherry–H2B were infected with lentiviruses encoding shRNAs specific to EPLIN (sh-EPLIN) or luciferase (sh-Luc.). FLAG–EPLINα WT and the S203 mutants were transfected and exogenously expressed in the absence of endogenous EPLIN (sh-EPLIN/FLAG–EPLINα). Whole-cell lysates obtained from technical replicates conducted in parallel were analyzed by immunoblotting with the indicated antibodies. B, C HeLa cells, as described in (A), were monitored by live cell imaging for 24 h. Representative images of the cells undergoing mitosis are shown (B). Scale bars, 10 μm. The duration of mitosis (n ≥ 71 cells) was measured (C). Values (median ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. D FLAG-EPLIN α WT or the S203 mutants were reintroduced into EPLIN-depleted HeLa cells. Control HeLa cells, EPLIN-depleted HeLa cells (sh-EPLIN), and those with re-expression of FLAG-EPLIN α (sh-EPLIN/FLAG-EPLIN α) were synchronized in metaphase and stained for α-tubulin (red), EPLIN (green) and DNA (blue). Representative images acquired using a Carl Zeiss Apotome2 microscope system are shown. White dashed lines mark the profile of the cells. Scale bars, 10 μm. E, F The cells as described in (D) were synchronized in metaphase and stained for epifluorescence microscopy. The difference in the “pole to cortex” (E) and the percentage of the cells with different extents of spindle deviation (F) were measured (n ≥ 34 cells). The values in E are presented as the box-whisker plot from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. Phosphorylation of EPLIN at Ser203 may be required for proper spindle positioning and mitotic progression. A HeLa cells stably expressing GFP–α-tubulin and mCherry–H2B were infected with lentiviruses encoding shRNAs specific to EPLIN (sh-EPLIN) or luciferase (sh-Luc.). FLAG–EPLINα WT and the S203 mutants were transfected and exogenously expressed in the absence of endogenous EPLIN (sh-EPLIN/FLAG–EPLINα). Whole-cell lysates obtained from technical replicates conducted in parallel were analyzed by immunoblotting with the indicated antibodies. B, C HeLa cells, as described in (A), were monitored by live cell imaging for 24 h. Representative images of the cells undergoing mitosis are shown (B). Scale bars, 10 μm. The duration of mitosis (n ≥ 71 cells) was measured (C). Values (median ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. D FLAG-EPLIN α WT or the S203 mutants were reintroduced into EPLIN-depleted HeLa cells. Control HeLa cells, EPLIN-depleted HeLa cells (sh-EPLIN), and those with re-expression of FLAG-EPLIN α (sh-EPLIN/FLAG-EPLIN α) were synchronized in metaphase and stained for α-tubulin (red), EPLIN (green) and DNA (blue). Representative images acquired using a Carl Zeiss Apotome2 microscope system are shown. White dashed lines mark the profile of the cells. Scale bars, 10 μm. E, F The cells as described in (D) were synchronized in metaphase and stained for epifluorescence microscopy. The difference in the “pole to cortex” (E) and the percentage of the cells with different extents of spindle deviation (F) were measured (n ≥ 34 cells). The values in E are presented as the box-whisker plot from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. EPLIN phosphorylation at Ser203 is required for it to maintain the proper structure and contractility of the actin cortex Precise spindle positioning relies on the proper architecture of the cell cortex1. Since endogenous EPLIN was enriched at the cell cortex of mitotic cells (Fig. 3D), we assumed that EPLIN may be involved in organizing cortical architecture to ensure proper cortical tension. Indeed, EPLIN depletion caused an abnormal cortical actin accumulation in mitotic cells (Fig. 4A and Supplementary Fig. 6), leading to an increase in the intensity of F-actin at the cell cortex (Fig. 4B). Re-expression of FLAG-EPLIN α and the S203D mutant, but not the S203A mutant, into EPLIN-depleted cells was able to partially rescue the defect in the cortical F-actin during mitosis (Fig. 4A, B).Fig. 4EPLIN depletion-induced spindle deviation is accompanied by increased cortical contractility and blebbing.A, B FLAG-EPLIN α WT or the S203 mutants were reintroduced into EPLIN-depleted HeLa cells. HeLa cells (control), EPLIN-depleted HeLa cells (sh-EPLIN), and those with re-expression of FLAG-EPLIN α were synchronized in metaphase. The cells were stained for F-actin (red, by phalloidin), FLAG epitope (green), α-tubulin (white), and DNA (blue). Representative images acquired using epifluorescence microscopy are shown (A). Scale bars, 10 μm. In B, the relative fluorescence intensity of F-actin at the cell peripheral region with a width of 3 μm was measured (n ≥ 36 cells) using the Zeiss ZEN2 software. C, D HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) were synchronized in metaphase and stained for MCL pS19 (green), F-actin (red, by phalloidin), α-tubulin (white), and DNA (blue). Representative images acquired using a Carl Zeiss LSM880 confocal microscope imaging system are shown (C). The arrow points to a cortical bleb. Scale bars, 10 μm. The relative fluorescence intensities of MLC pS19 and F-actin were measured (n ≥ 152 cells) using the Zeiss ZEN2 software (D). E An equal amount of whole cell lysates from HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) was analyzed by immunoblotting with the indicated antibodies. F–H HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) were transfected with the plasmids encoding GFP-Histone 2B (H2B) and Cherry-UtrCH (Utrophin calponin homology domain; to label F-actin) and then subjected to live cell imaging for 24 h in the presence or absence of Y27632 (0.5 μM). Representative images of the cells in metaphase are shown (F). The arrow points to a cortical bleb. Scale bars, 5 μm. The percentage of the cells (n ≥ 135) forming cortical blebs during mitosis is counted (G). The duration of mitosis (n ≥ 120 cells) was measured (H). I, J HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) were synchronized in metaphase and then treated with or without 0.5 μM Y27632 for 2 h before fixation. The cells were stained for MCL pS19 (green), F-actin (red, by phalloidin), α-tubulin (white) and DNA (blue). Representative images acquired using a Carl Zeiss LSM880 confocal microscope imaging system are shown (I). Scale bars, 10 μm. The difference in the pole to cortex (n ≥ 112 cells) was measured (J). Data information: In D, G and J, values (means ± SD) are from three independent experiments. Values in B and H are presented as the box-whisker plot from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. EPLIN depletion-induced spindle deviation is accompanied by increased cortical contractility and blebbing. A, B FLAG-EPLIN α WT or the S203 mutants were reintroduced into EPLIN-depleted HeLa cells. HeLa cells (control), EPLIN-depleted HeLa cells (sh-EPLIN), and those with re-expression of FLAG-EPLIN α were synchronized in metaphase. The cells were stained for F-actin (red, by phalloidin), FLAG epitope (green), α-tubulin (white), and DNA (blue). Representative images acquired using epifluorescence microscopy are shown (A). Scale bars, 10 μm. In B, the relative fluorescence intensity of F-actin at the cell peripheral region with a width of 3 μm was measured (n ≥ 36 cells) using the Zeiss ZEN2 software. C, D HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) were synchronized in metaphase and stained for MCL pS19 (green), F-actin (red, by phalloidin), α-tubulin (white), and DNA (blue). Representative images acquired using a Carl Zeiss LSM880 confocal microscope imaging system are shown (C). The arrow points to a cortical bleb. Scale bars, 10 μm. The relative fluorescence intensities of MLC pS19 and F-actin were measured (n ≥ 152 cells) using the Zeiss ZEN2 software (D). E An equal amount of whole cell lysates from HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) was analyzed by immunoblotting with the indicated antibodies. F–H HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) were transfected with the plasmids encoding GFP-Histone 2B (H2B) and Cherry-UtrCH (Utrophin calponin homology domain; to label F-actin) and then subjected to live cell imaging for 24 h in the presence or absence of Y27632 (0.5 μM). Representative images of the cells in metaphase are shown (F). The arrow points to a cortical bleb. Scale bars, 5 μm. The percentage of the cells (n ≥ 135) forming cortical blebs during mitosis is counted (G). The duration of mitosis (n ≥ 120 cells) was measured (H). I, J HeLa cells (control) and those with EPLIN depletion (sh-EPLIN) were synchronized in metaphase and then treated with or without 0.5 μM Y27632 for 2 h before fixation. The cells were stained for MCL pS19 (green), F-actin (red, by phalloidin), α-tubulin (white) and DNA (blue). Representative images acquired using a Carl Zeiss LSM880 confocal microscope imaging system are shown (I). Scale bars, 10 μm. The difference in the pole to cortex (n ≥ 112 cells) was measured (J). Data information: In D, G and J, values (means ± SD) are from three independent experiments. Values in B and H are presented as the box-whisker plot from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. The abnormal cortical actin accumulation induced by EPLIN depletion was associated with increased cortical contractility in mitotic cells, as manifested with increased phosphorylation of myosin II light chain at Ser19 (MLC pS19) (Fig. 4C–E). In addition, cortical blebs were occasionally observed in EPLIN-depleted cells by immunofluorescence staining (Fig. 4C). Next, the dynamics of the cortical blebbing were monitored by live cell imaging. EPLIN depletion apparently induced dynamic formation of cortical blebs in ~57% of the cells during metaphase (Fig. 4F, G and Supplementary Fig. 7). It has been shown that activation of cortical contractility via the RhoA/ROCK pathway induces blebs in metaphase12,42. Accordingly, the treatment with the ROCK inhibitor Y27632 diminished the formation of cortical blebs in EPLIN-depleted cells (Fig. 4F, G and Supplementary Fig. 7). Y27632 also restored the mitosis duration (Fig. 4H) and spindle positioning (Fig. 4I, J) in EPLIN-depleted cells. As expected, Y27632 suppressed the increased Ser19 phosphorylation of MLC in EPLIN-depleted cells (Fig. 4I). These results together suggest that phosphorylation of EPLIN at Ser203 is required for it to maintain proper structure and contractility of the cortical F-actin to ensure correct spindle positioning. EPLIN phosphorylation at Ser203 enhances its capability to inhibit Arp2/3-mediated actin polymerization in vitro The results in Fig. 4 raise an intriguing possibility that EPLIN may have an inhibitory effect on F-actin polymerization at the mitotic cell cortex. To examine this possibility, in vitro pyrene-actin polymerization assays were performed in the presence or absence of purified EPLIN. The histidine epitope-tagged EPLIN α (His-EPLIN) proteins were expressed and purified in our laboratory (Supplementary Fig. 8). In the absence of Arp2/3, EPLIN WT, S203A, and S203D inhibited actin polymerization to a similar extent (Fig. 5A, B). To better characterize the inhibition of the in vitro actin polymerization by His-EPLIN proteins, the relative maximum of the in vitro actin polymerization slope was calculated from the time-course experiments with different concentrations of His-EPLIN proteins. Our results showed that His-EPLIN proteins were able to inhibit in vitro actin polymerization in a dose-dependent manner, yet without significant differences among the WT, S203A mutant, and S203D mutant (Fig. 5B and Supplementary Fig. 9A).Fig. 5EPLIN phosphorylation at Ser203 may enhance its capability to inhibit Arp2/3-mediated actin polymerization by organizing astral actin structures.A, B The in vitro actin polymerization assays were carried out with pyrene-label actin in the presence or absence of purified His-EPLIN proteins (30 nM), as described in Materials and Methods. The elution buffer for His-tagged protein purification was used as a control. The fluorescence intensity of the reaction was monitored at 1-min intervals for 70 min using a spectrofluorometer (A). A.U., arbitrary units. The relative maximum polymerization slope was calculated (B). C, D The in vitro Arp2/3-mediated actin polymerization assays were carried out with pyrene-label actin in the presence of the Arp2/3 complex and the WASP-VCA domain, as described in Materials and Methods. His-EPLIN proteins (30 nM) were added to the assay to evaluate their effect on the Arp2/3-mediated actin polymerization. The elution buffer for His-tagged protein purification was used as a control. The fluorescence intensity of the reaction was monitored at 1-min intervals for 70 min using a spectrofluorometer (C). A.U., arbitrary units. The relative maximum polymerization slope was calculated (D). (E) The products of the Arp2/3-mediated actin polymerization assays as described in (C) were stained with phalloidin (red) and anti-EPLIN antibody (green), and then visualized under an epifluorescence microscope. Note that His-EPLIN S203D reorganizes short branched actin filaments into astral actin structures (arrowhead), whereas His-EPLIN S203A mainly induces long filamentous F-actin (arrow). The areas selected by the white frame are enlarged. The further enlarged spokes (from the dashed white frame) show that the S203D is associated with the F-actin spokes at ~0.48-μm intervals. Scale bars: 10 μm (overviews), 1 μm (zoom-ins) and 0.5 μm (enlarged spokes). The green fluorescence intensity in the images of the control “-“ was enhanced to visualize F-actin; the red fluorescence intensity in the images of “EPLIN S203A” was enhanced to visualize spokes. F–I Images were quantified from (E) to determine the number of asters (in (F), n ≥ 13 random images), the number of spokes of an aster (in (G), n ≥ 15 asters), the length (H) and width (I) of the spoke (n ≥ 138 spokes, from n ≥ 15 asters). Data information: In (B, D), values (means ± SD) are from three independent experiments. Values in (F, G, H and I) are presented as the box-whisker plot. Statistical significance of differences is assessed with Student’s t-test. EPLIN phosphorylation at Ser203 may enhance its capability to inhibit Arp2/3-mediated actin polymerization by organizing astral actin structures. A, B The in vitro actin polymerization assays were carried out with pyrene-label actin in the presence or absence of purified His-EPLIN proteins (30 nM), as described in Materials and Methods. The elution buffer for His-tagged protein purification was used as a control. The fluorescence intensity of the reaction was monitored at 1-min intervals for 70 min using a spectrofluorometer (A). A.U., arbitrary units. The relative maximum polymerization slope was calculated (B). C, D The in vitro Arp2/3-mediated actin polymerization assays were carried out with pyrene-label actin in the presence of the Arp2/3 complex and the WASP-VCA domain, as described in Materials and Methods. His-EPLIN proteins (30 nM) were added to the assay to evaluate their effect on the Arp2/3-mediated actin polymerization. The elution buffer for His-tagged protein purification was used as a control. The fluorescence intensity of the reaction was monitored at 1-min intervals for 70 min using a spectrofluorometer (C). A.U., arbitrary units. The relative maximum polymerization slope was calculated (D). (E) The products of the Arp2/3-mediated actin polymerization assays as described in (C) were stained with phalloidin (red) and anti-EPLIN antibody (green), and then visualized under an epifluorescence microscope. Note that His-EPLIN S203D reorganizes short branched actin filaments into astral actin structures (arrowhead), whereas His-EPLIN S203A mainly induces long filamentous F-actin (arrow). The areas selected by the white frame are enlarged. The further enlarged spokes (from the dashed white frame) show that the S203D is associated with the F-actin spokes at ~0.48-μm intervals. Scale bars: 10 μm (overviews), 1 μm (zoom-ins) and 0.5 μm (enlarged spokes). The green fluorescence intensity in the images of the control “-“ was enhanced to visualize F-actin; the red fluorescence intensity in the images of “EPLIN S203A” was enhanced to visualize spokes. F–I Images were quantified from (E) to determine the number of asters (in (F), n ≥ 13 random images), the number of spokes of an aster (in (G), n ≥ 15 asters), the length (H) and width (I) of the spoke (n ≥ 138 spokes, from n ≥ 15 asters). Data information: In (B, D), values (means ± SD) are from three independent experiments. Values in (F, G, H and I) are presented as the box-whisker plot. Statistical significance of differences is assessed with Student’s t-test. The cell cortex is known to contain parallel actin filaments and branched actin networks10. The Arp2/3 complex nucleates and cross-links actin filaments to form branched actin networks, and its activity is stimulated by the C-terminal region of the Wiskott-Aldrich syndrome protein (WASP), called the VCA domain43,44. As reported previously45, the in vitro actin polymerization was significantly enhanced in the presence of purified VCA domain and the Arp2/3 complex (Fig. 5C). More importantly, His-EPLIN proteins (WT, S203A, and S203D) were able to inhibit the Arp2/3-mediated actin polymerization in a dose-dependent manner (Supplementary Fig. 9B). Among them, the S203D mutant possessed the strongest potency in this regard (Fig. 5C, D and Supplementary Fig. 9B). EPLIN phosphorylation at Ser203 promotes its capability to form astral actin structures in vitro To understand how EPLIN inhibits the Arp2/3-mediated actin polymerization, the products of the in vitro Arp2/3-mediated actin polymerization assays were stained with phalloidin and visualized under a fluorescent microscope. As described previously46, the Arp2/3-mediated actin polymerization mainly generated short-branched actin filaments (Fig. 5E). Surprisingly, the architecture of the F-actin products of the Arp2/3-mediated actin polymerization was markedly changed by His-EPLIN proteins (S203A and S203D). The S203A mainly induced long filamentous F-actin (Fig. 5E and Supplementary Fig. 10). In contrast, the S203D apparently increased the formation of astral actin structures with radial spokes of bundled F-actin emanating from the core (Fig. 5E and Supplementary Fig. 10). The S203D was more potent (~5.2-fold) than the S203A to organize such actin asters (Fig. 5F). Next, these “in vitro actin asters” formed by His-EPLIN was further characterized. The actin asters formed by the S203D had more F-actin spokes than those formed by the S203A (13 spokes versus 8.5 spokes) (Fig. 5G). In addition, the length of the F-actin spokes formed by the S203D was longer than that formed by the S203A (4.4 μm versus 3.1 μm) (Fig. 5H), while the thickness (0.3 μm) of the F-actin spokes formed by the S203A and S203D was similar (Fig. 5I). Moreover, the S203D was more apparent than the S203A to localize at the spokes of the astral structures. Notably, the S203D was associated with the F-actin spokes at ~0.48-μm intervals (Fig. 5E, the enlarged spokes). Although the intensity of EPLIN S203A and S203D was positively correlated with the intensity of F-actin at the spokes, the slopes of their regression lines were different (Supplementary Fig. 11). The difference in the slope reflected that at the same amount of EPLIN proteins, S203D had a higher extent of F-actin association than S203A did. These results suggest that the S203D may be more potent than the S203A to bundle F-actin into the spokes of the actin asters. In consistent with this assumption, we showed that the S203D possessed a higher F-actin-bundling activity than the S203A did, as measured by F-actin co-sedimentation assays (Fig. 6).Fig. 6EPLIN phosphorylation at Ser203 may enhance its F-actin bundling activity.A The in vitro actin polymerization assay was carried out in the presence of His-EPLIN S203A or S203D for 70 min. The reaction products were centrifuged at 8000 × g for 20 min. The supernatant (S) and pellet (P) were fractionated by SDS-PAGE and stained with Coomassie blue to visualize actin or analyzed by immunoblotting with anti-His antibody. The actin-bundling activity was analyzed by measuring the extent of actin in the pellet fraction relative to the control (i.e., the reaction without His-EPLIN). B The in vitro Arp2/3-mediated actin polymerization assays were carried out in the presence of His-EPLIN S203A or S203D for 20 min. The products of the reaction were subjected to centrifugation at 8000 × g for 20 min, and then the pellets were analyzed by immunoblotting with the indicated antibodies. The relative level of co-sedimentation of actin and His-EPLIN (S203A vs S203D) was measured. Values in (A, B) (means ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. EPLIN phosphorylation at Ser203 may enhance its F-actin bundling activity. A The in vitro actin polymerization assay was carried out in the presence of His-EPLIN S203A or S203D for 70 min. The reaction products were centrifuged at 8000 × g for 20 min. The supernatant (S) and pellet (P) were fractionated by SDS-PAGE and stained with Coomassie blue to visualize actin or analyzed by immunoblotting with anti-His antibody. The actin-bundling activity was analyzed by measuring the extent of actin in the pellet fraction relative to the control (i.e., the reaction without His-EPLIN). B The in vitro Arp2/3-mediated actin polymerization assays were carried out in the presence of His-EPLIN S203A or S203D for 20 min. The products of the reaction were subjected to centrifugation at 8000 × g for 20 min, and then the pellets were analyzed by immunoblotting with the indicated antibodies. The relative level of co-sedimentation of actin and His-EPLIN (S203A vs S203D) was measured. Values in (A, B) (means ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. It has been shown that EPLIN inhibits the Arp2/3 complex-mediated actin polymerization in vitro27 and interacts with the Arp2/3 complex for terminating membrane protrusions during cell migration30. In this study, we found that Arp 3 was co-localized with His-EPLIN S203A and S203D at the core of the in vitro actin asters (Fig. 7A). However, the intensity of Arp 3 was higher at the core of the actin asters formed by the S203D than by the S203A (Fig. 7B). Importantly, the intensity of Arp 3 at the core of the actin asters was positively correlated with the intensity of the S203D, but not the S203A (Fig. 7C). Experiments to this point, we assumed that the phosphorylation of EPLIN at Ser203 may promote its ability to reorganize F-actin into the astral structures, with which it sequesters Arp2/3 and thereby prevents excessive de novo Arp2/3-mediated actin polymerization. Indeed, we found that the Arp2/3 inhibitor CK-66647 rescued the defects in the cortical F-actin and spindle positioning in EPLIN-depleted cells (Fig. 7D, E). Taken together, our results suggest that the mitotic defects induced by EPLIN depletion may be mostly attributed to excessive Arp2/3-mediated actin polymerization.Fig. 7EPLIN phosphorylation at Ser203 may enhance its capability to sequester Arp2/3 at the core of in vitro actin asters.A Purified His-EPLIN S203A or S203D proteins (60 nM) were used in the in vitro pyrene-actin polymerization assay in the presence of Arp2/3 complex and WASP-VCA domain. After 20 min of reaction, the product was stained for F-actin (by phalloidin, green), EPLIN (red) and Arp3 (white). The images were taken by an epifluorescence microscope. The cores of astral structures were enlarged from the white squares. Scale bars, 10 μm. B, C The intensities of Arp3 and EPLIN at the astral core (a circle of 2.074 μm2) were analyzed from images as described in (A). The levels of Arp3 at the cores of asters were shown in (B) (S203A, n = 21; S203D, n = 78). The intensities of EPLIN and Arp3 at the cores were analyzed in the correlation analysis by Excel, as shown in (C). D, E Control HeLa cells and those with EPLIN-depletion (sh-EPLIN) were synchronized in metaphase and then treated with (+) or without (-) the Arp2/3 inhibitor CK-666 (50 μM) for 2 h. The cells were fixed and stained for F-actin (red, by phalloidin), α-tubulin (white), and DNA (blue). Representative images acquired using epifluorescence microscopy are shown (D). Scale bars, 10 μm. The difference in the “pole to cortex” (n ≥ 104 cells) was measured (E). Values (means ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. EPLIN phosphorylation at Ser203 may enhance its capability to sequester Arp2/3 at the core of in vitro actin asters. A Purified His-EPLIN S203A or S203D proteins (60 nM) were used in the in vitro pyrene-actin polymerization assay in the presence of Arp2/3 complex and WASP-VCA domain. After 20 min of reaction, the product was stained for F-actin (by phalloidin, green), EPLIN (red) and Arp3 (white). The images were taken by an epifluorescence microscope. The cores of astral structures were enlarged from the white squares. Scale bars, 10 μm. B, C The intensities of Arp3 and EPLIN at the astral core (a circle of 2.074 μm2) were analyzed from images as described in (A). The levels of Arp3 at the cores of asters were shown in (B) (S203A, n = 21; S203D, n = 78). The intensities of EPLIN and Arp3 at the cores were analyzed in the correlation analysis by Excel, as shown in (C). D, E Control HeLa cells and those with EPLIN-depletion (sh-EPLIN) were synchronized in metaphase and then treated with (+) or without (-) the Arp2/3 inhibitor CK-666 (50 μM) for 2 h. The cells were fixed and stained for F-actin (red, by phalloidin), α-tubulin (white), and DNA (blue). Representative images acquired using epifluorescence microscopy are shown (D). Scale bars, 10 μm. The difference in the “pole to cortex” (n ≥ 104 cells) was measured (E). Values (means ± SD) are from three independent experiments. Statistical significance of differences is assessed with Student’s t-test. Discussion In this study, we unveil an important role for EPLIN in symmetric spindle positioning during mitosis. Based on our results, we propose a model (Fig. 8) that, at the onset of mitosis, the phosphorylation of EPLIN at Ser203 by CDK1 promotes its capability to bundle F-actin and reorganize it into specialized actin structures, which may be analogous to the in vitro actin asters composed of EPLIN and the Arp2/3 complex (Fig. 5E). Such actin asters sequester F-actin bundles from the turnover of actin filaments and thereby shut down the supply of G-actin for de novo actin polymerization. In the meantime, the interaction of EPLIN with the Arp2/3 complex prevents excessive Arp2/3-mediated actin polymerization in mitotic cells. In the absence of EPLIN, no such astral actin structures are formed, which unleashes the Arp2/3 complex to promote branched F-actin networks and causes abnormal cortical actin accumulation. The excessive cortical actin is accompanied by increased myosin II contractility and cortical blebbing (Fig. 4), thereby leading to spindle deviation.Fig. 8A hypothetical model depicts how EPLIN maintains the proper architecture of the actin cortex for spindle positioning.Phosphorylation of EPLIN at Ser203 by CDK1 promotes its capability to reorganize branched actin filaments into astral structures by bundling F-actin and interacting with the Arp2/3 complex. Such actin asters may sequester F-actin bundles from the turnover of actin filaments and thereby shut down the supply of G-actin for de novo actin polymerization. In the meantime, the sequestration of the Arp2/3 complex in the actin asters prevents excessive Arp2/3-mediated actin polymerization in mitotic cells. In the absence of EPLIN, no such actin asters are formed, which unleashes the Arp2/3 complex to promote branched F-actin networks, and causes abnormal cortical actin accumulation and aberrant cortical contractility, leading to cortical blebs and spindle deviation. A hypothetical model depicts how EPLIN maintains the proper architecture of the actin cortex for spindle positioning. Phosphorylation of EPLIN at Ser203 by CDK1 promotes its capability to reorganize branched actin filaments into astral structures by bundling F-actin and interacting with the Arp2/3 complex. Such actin asters may sequester F-actin bundles from the turnover of actin filaments and thereby shut down the supply of G-actin for de novo actin polymerization. In the meantime, the sequestration of the Arp2/3 complex in the actin asters prevents excessive Arp2/3-mediated actin polymerization in mitotic cells. In the absence of EPLIN, no such actin asters are formed, which unleashes the Arp2/3 complex to promote branched F-actin networks, and causes abnormal cortical actin accumulation and aberrant cortical contractility, leading to cortical blebs and spindle deviation. Since the cortical blebbing is dynamic and the positions of the blebs are constantly changed during mitosis (Supplementary Fig. 7), it is possible that the cortical tension of mitotic cells is disturbed by an excessive and irregular cortical contractility. When the cell cortex is in such chaos, the Gαi-LGN-NuMA complex may no longer be able to accurately localize at the bilateral sides of the cell cortex (Supplementary Fig. 4), leading to an unstable cortical anchoring of the mitotic spindle. Therefore, the observed spindle deviation may arise from an excessive and irregular cortical contractility, which then disturbs bilateral cortical localization of the Gαi-LGN-NuMA complex, leading to an unstable cortical anchoring of the mitotic spindle. Haviv et al.46 reported that Fascin induces actin asters in the presence of Arp2/3. Like EPLIN, Fascin is an F-actin bundling protein. These in vitro actin asters described by us and by Haviv et al.46 share very similar astral morphology, which has radial spokes of bundled F-actin emanating from the core. EPLIN and Fascin are associated with the F-actin spokes, while the Arp2/3 complexes are accumulated at the core. Moreover, Fritzsche et al.48 reported that actin asters (2.4–5.0 μm) actually exist at the cell cortex of HeLa cells using super-resolution microscopy. These cortical actin asters are further interconnected to form a network via their spokes. Although the Arp2/3 complexes are certainly involved in the cortical actin asters48, it remains unclear whether the in vitro actin asters formed by EPLIN and Fascin represent the analog found at the cell cortex of HeLa cells. EPLIN is already known to be phosphorylated at two serine sites in the N-terminal half (i.e., Ser203 in the α or Ser362 in the β) and C-terminal half (i.e., Ser445 in the α or Ser604 in the β) by ERK upon stimulation by epidermal growth factor32. It has been shown that the C-terminal phosphorylation of EPLIN induces its protein degradation, decreases its binding to actin bundles at adherent junctions, and relocates it from stress fibers towards lamellipodia34,35. Since EPLIN is known to inhibit the Arp2/3 complex-mediated actin polymerization in vitro27 and interact with the Arp2/3 complex for terminating membrane protrusions during migration30, EPLIN seemingly acts as a negative regulator for cell migration by suppressing Arp2/3-mediated lamellipodial formation. Compared to the C-terminal phosphorylation, the significance of the N-terminal phosphorylation (i.e., Ser203 in the α or Ser362 in the β) of EPLIN is less characterized. In this study, we showed that Ser203 of EPLIN α is phosphorylated by CDK1 in mitosis (Fig. 2), indicating that EPLIN Ser203 can be phosphorylated by different Ser/Thr protein kinases (i.e., ERK and CDK1) during different phases of the cell cycle. Since the depletion of EPLIN causes blebbing and abnormal enrichment of the actin cortex of mitotic cells (Fig. 4), it indicates that EPLIN is required for maintaining the proper architecture of the cortical actin network of mitotic cells by preventing excessive F-actin polymerization. In accordance with the previous study27. We found that EPLIN inhibits the Arp2/3 complex-mediated actin polymerization in vitro and, more importantly, the phosphomimetic S203D mutant is more potent than the S203 A mutant in this regard (Fig. 5). Therefore, our study suggests that the N-terminal phosphorylation of EPLIN at Ser203 by ERK in interphase or CDK1 in mitosis may be crucial for it to inhibit the Arp2/3 complex-mediated actin polymerization and thereby regulate lamellipodia in interphase and the actin cortex in mitosis. Since proper architecture and contractility of the mitotic cell cortex coordinate the progression of mitosis, multiple mechanisms involving various classes of molecules are evolutionarily conserved to ensure proper architecture of the cortical actin cytoskeleton during mitosis. In particular, the assembly of the actin cortex during mitosis is constrained by phosphorylation-dependent regulation of several key actin regulators, such as EPLIN (an actin bundler), cofilin (an actin severer), and DIAPH1 (an actin nucleator). Similar to the consequences caused by EPLIN depletion (this study), cofilin depletion causes blebbing and abnormal cortical actin accumulation15. Moreover, LIM kinase-mediated cofilin phosphorylation during mitosis is required for precise spindle positioning14. DIAPH1 (also known as mDia1) is a member of the actin nucleator formin family of proteins, which is required for actin stress fiber formation49 and maintenance of the cortical force during mitotic cell rounding50. CDK1-mediated phosphorylation of DIAPH1 prevents cortical actin accumulation, which coordinates cortical tension maintenance and spindle assembly checkpoint inactivation at anaphase onset13. The puzzle of how these kinases and actin regulators coordinate the dynamic remodeling of the cortical actin cytoskeleton during mitosis remains to be solved. The tumor-suppressive function of EPLIN has been mostly attributed to its ability to regulate cell-cell adhesion and cell motility. Besides, EPLIN has been shown to be required for cytokinesis41. In this study, we propose that EPLIN serves as a guardian for spindle positioning in metaphase, which further supports EPLIN as a versatile tumor suppressor. Thus, downregulation of EPLIN could cause not only disruption of cell-cell adhesion and increased cell motility, but also spindle deviation and aneuploidy, which are often associated with malignant progression. However, whether downregulation of EPLIN clinically correlates with increased cell cortical contractility and spindle deviation in cancer cells, as well as whether the inhibition of cell cortical contractility serves as a therapeutic strategy for targeting EPLIN-downregulated cancer cells, awaits future studies. Moreover, EPLIN has been shown to be important for various cellular activities, such as endocytosis36, intestinal cholesterol-absorption37, cilia formation38, and mitochondrial function39. Whether EPLIN participates in those cellular activities through its regulation of cortical actin and membrane dynamics remains to be studied. Methods Antibodies and reagents The anti-EPLIN (sc-136399, for IF 1:150 and for WB 1:1000) and anti-cyclin B1 (sc-245, 1:1500) antibodies were purchased from Santa Cruz Biotechnology. The anti-β-actin (A1978, 1:8000), anti-FLAG (F1804, for IF 1:500 and for WB 1:5000), anti-MLC (M4401, 1:1200) antibodies were purchased from Sigma-Aldrich. The anti-MLC pS19 antibody (for IF #3671 1:150 or for WB #3675 1:1200) was purchased from Cell Signaling. The anti-T7 antibody (695223, 1:5000) was purchased from Novagen. The anti-thiophosphate ester (ab92570, 1:5000) and anti-phospho-Ser/Thr (ab17464, 1:1000) antibodies were purchased from Abcam. The anti-EPLIN pS203 (customized, 1:500) and anti-α-tubulin (GTX112141, for IF 1:200) antibodies were purchased from GeneTex. The anti-α-tubulin (T6199, for WB 1:6000) antibody was purchased from Sigma-Aldrich. The anti-Cyclin A antibody (BF683, 1:1000) was purchased from Millipore. The anti-Arp3 antibody (PA5-80307, 1:100), anti-NuMA (PA5-17293, for IF 1:200 and for WB 1:1000) and phalloidin (conjugated with Alexa Fluor-546, R415, 1:300) were purchased from Invitrogen. Phos-Tag™ Phosphoprotein Gel Stain (P005A) was purchased from ABP Biosciences. Thymidine (6060) was purchased from Calbiochem. PNBM (p-nitrobenzyl mesylate, ab138910) was purchased from Abcam. Ni-SepharoseTM 6 fast flow beads (GE17-5318-01) were purchased from GE Healthcare. RO-3306 (ALX-270-463) and Y27632 (ALX-270-333) were purchased from Enzo Life Sciences. Nocodazole (M1404) and CK-666 (182515) were purchased from Sigma-Aldrich. Cyclin B1/CDK1 proteins (14-450) were purchased from Millipore. General Actin Buffer (G Buffer, BSA01-010), pyrene-labeled muscle actin (pyrene-actin, AP05), rabbit muscle actin protein (AKL95), Arp2/3 protein complex (RP01P), and WASP-VCA Domain-GST protein (VCG03) were purchased from Cytoskeleton. Plasmids To construct the plasmid pCMV-3tag-3A-EPLIN-α for expressing FLAG-tagged-EPLIN-α, the EPLIN-α cDNA was amplified from pEGFP-C1-EPLIN-α (Addgene, #40947) by polymerase chain reaction (PCR) and then cloned into pCMV-3Tag-3A using Not I and BamH I sites. The mutagenesis of pCMV-3tag-3A-EPLIN-α for S203A, S214A, S215A, S331A, S445A, S450A, and S203D was performed using the Quickchange site-directed mutagenesis kit (Stratagene) and the desired mutations were confirmed by DNA sequencing. The pET21a-(+) vector carries an N-terminal T7•Tag® sequence plus an optional C-terminal. His•Tag® sequence. To construct the plasmid pET21a(+)-EPLIN-α for expressing His-tagged EPLIN, the EPLIN-α cDNAs (WT, S203A, and S203D) obtained from pCMV-3tag-3A-EPLIN-α WT, S203A or S203D were subcloned into the pET21a(+) vector using Not I and Xho I sites. To construct the plasmid pET21a(+)-EPLIN-α−Ν for expressing T7-tagged EPLIN N-terminal region, the cDNA encoding the N-terminal (a.a. 1-230) of EPLIN-α was amplified by PCR and subcloned into the pET21a(+) vector using Not I and Xho I sites. The plasmid pCS2(+)-mCherry-UtrCH (Utrophin calponin homology domain) and pEGFP-H2B were kindly provided by Dr. Hsiao-Hui Lee (National Yang Ming Chiao Tung University, Taiwan) and Dr. Hsiu-Ming Shih (Academia Sinica, Taiwan), respectively. Cell lines and culture conditions The human cell lines used in this study were obtained from the American Type Culture Collection and periodically examined for mycoplasma contamination. HeLa (human cervical carcinoma) and HEK293T (human embryonic kidney) cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum and antibiotics. Cells were incubated in humidified conditions at 37 °C with 5% CO2. Transient transfection was performed using Lipofectamine 2000 (Invitrogen). Lentivirus-mediated gene knockdown The lentiviral expression system was obtained from the National RNAi Core Facility (Academia Sinica, Taiwan). For the lentivirus production, HEK293T cells were co-transfected with 0.25 μg pMD.G, 2.25 μg pCMV-ΔR8.91, and 2.5 μg pLKO.1-shRNA-Luciferase (TRCN0000072249, 5’-GCGGTTGCCAAGAGGTTCCAT-3’) or pLKO.1-shRNA-LIMA1 (TRCN0000149941, 5’-GTCTCTGAATTGGTCGAGTTT-3’) through Lipofectamine 2000. After three days, the medium with the viral particles was collected and stored at −80 °C. Cells were infected with lentiviruses for 24 h and subsequently selected with puromycin (2 μg/mL) for 4 days before further analysis. Immunofluorescence staining, microscopy, and image analysis Cells were fixed with phosphate-buffered saline containing 4% paraformaldehyde at room temperature for 15 min. Cells were permeabilized with 0.5% Triton X-100 at room temperature for 20 min. Cells were stained with primary antibodies for 1 h and then incubated with Alexa Fluor 488- or 546-conjugated secondary antibodies or phalloidin for 1 h. Cells were mounted on the slides with mounting medium (anti-Fade Dapi-Fluoromount-G, Southern Biotech). The images (2048 ×2048 pixels) were acquired using an upright fluorescence microscope (Axio imager, M2 Apotome2 system, Carl Zeiss), which was equipped with 63x and 100x oil-immersion objective lenses (1.4 NA, Plan Apochromat) and a camera (ORCA-Flash4.0 V2; Hamamatsu). For Fig. 4C, I and Supplementary Fig. S6, images were acquired using a Carl Zeiss LSM880 confocal microscope imaging system with a Zeiss Plan-Apochromat 63x/NA 1.4 oil immersion objective. Images were analyzed by the ZEN2 software (Carl Zeiss). The representative images were cropped by Photoshop CS6 (Adobe) and assembled by Illustrator CS6 (Adobe). Live cell imaging Cells were incubated in a micro-cultivation system with temperature and CO2 control devices (Carl Zeiss) and monitored on an inverted microscope (Axio Observer; Carl Zeiss) using the EC Plan-NEOFLUAR 20 x objective (for Figs. 1B and 3B) or 63 x objective (for Fig. 4F and Supplementary Fig. 7). Images were captured every 3 min for 24 h using a digital camera (ORCA-Flash4.0 V2; Hamamatsu) and were processed by the ZEISS ZEN2 image software. Flow cytometry Cells were treated with trypsin, washed with PBS and fixed with ethanol (70%, at −20 °C for 1 h). Subsequently, cells were treated with Triton X-100 (0.1%), RNase A (10 μg/ml), and stained with propidium iodide (5 μg/ml). Flow cytometry was carried out using a Becton Dickinson FACSCaliburTM flow cytometer with the FlowJo-V10 software. The primary scatter gate was to isolate the intact cell, and the singlet gate was to remove cell clumps and doublets. The cell cycle histogram was obtained from the final gated single-cell population. Immunoblotting and immunoprecipitation To prepare whole cell lysates, cells were lysed on ice in NP-40 buffer (1% NP-40, 137 mM NaCl, 100 nM NaF, 0.2 mM Na3VO4, 10% glycerol, 20 mM Tris, pH 7.4) and the protein concentration was measured using the Bradford protein assay reagent (Bio-Rad). For immunoprecipitation, whole cell lysates (200 μg) were incubated with anti-FLAG or anti-EPLIN antibody and protein A-Sepharose beads (GE Healthcare) at 4 °C for 3 h. The beads were washed with NP-40 buffer for three times and resuspended in SDS sample buffer for immunoblotting analysis. For immunoblotting, samples were subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. The nitrocellulose membranes were probed with the indicated primary antibodies and subsequently with secondary antibodies conjugated to horseradish peroxidase or IR-780/680 iodide fluorescence dye, which were then scanned by the luminescence imaging system (LAS-4000; Fujifilm) or the Odyssey® CLx Imaging System (LI-COR Biosciences), respectively. Mass spectrometry HeLa cells were treated with nocodazole (200 ng/ml) for 16 h and then subjected to immunoprecipitation with anti-EPLIN antibody. The immunocomplexes were fractionated by SDS-PAGE and stained with Coomassie blue. EPLIN proteins were then removed from SDS-PAGE, and their phosphorylation sites were identified by nanoscale capillary liquid chromatography/tandem mass spectrometry (LC-MS/MS) analysis using an LC system (UltiMate Capillary; LC Packings) coupled to a quadrupole time-of-flight mass spectrometer (QSTAR; Applied Biosystems). A nano-electrospray interface was used for LC-MS/MS analysis, and the ionization (2.0-kV ionization potential) was performed with a coated nano-LC tip. For data acquisition, it was performed by automatic Information Dependent Acquisition (Applied Biosystems). Then, the Information Dependent Acquisition automatically finds the most intense ions in a time-of-flight MS spectrum, and subsequently performs an optimized MS/MS analysis on the selected ions. Finally, the ion spectra generated by nano–LC-MS/MS were searched against NCBI databases for exact matches using the MASCOT search program (Matrix Science). Expression and purification of His-EPLIN and T7-EPLIN-N-terminal region The pET21a-(+) plasmids encoding His-EPLIN-α protein or T7-EPLIN-α-N-terminal region were transformed into E. coli (BL21), which were then cultivated in LB medium at 37 °C until OD600 = 0.4−0.5. The expression of His-EPLIN or T7-EPLIN proteins was induced with IPTG (0.4 mM) at 22 °C for 6 h. The bacteria were collected by centrifugation, suspended in binding buffer (10 mM imidazole, 20 mM sodium phosphate, 0.3 M NaCl, pH 7.4), and lysed by sonication (QSONICA Q700). After centrifugation, the supernatants were incubated with Ni-SepharoseTM 6 Fast Flow beads (at 4 °C for 3 h). The beads were washed with wash buffer (20 mM imidazole, 20 mM sodium phosphate, 0.3 M NaCl, pH 7.9) and the bound proteins were harvested by elution buffer (100 mM imidazole, 20 mM sodium phosphate, 0.3 M NaCl, pH 8.0). In vitro kinase assay Purified T7-EPLIN-α-N WT or S203A proteins (0.5 μg) were incubated with 10 ng CDK1/cyclin-B1 in the kinase buffer (50 mM Tris, pH 7.4, 10 mM MgCl2, and 2 ng/μl ATPγS) at 25 °C for 30 min. Subsequently, PNBM (1 mM) was added to the reaction to yield thiophosphate esters on phosphorylated EPLIN proteins at 25 °C for 1 h. The reaction was terminated by SDS sample buffer and analyzed by immunoblotting with anti-thiophosphate ester antibody. In vitro actin polymerization assay Pyrene-labeled actin was prepared in G Buffer (10 mM Tris, pH 7.5, 0.2 mM CaCl2, 0.2 mM ATP, 1 mM DTT). For the in vitro actin polymerization assay, pyrene-labeled actin was diluted to a final concentration of 0.8 μM in polymerization buffer (2 mM MgCl2, 50 mM KCl, 1 mM ATP, final concentrations) to initiate spontaneous actin polymerization at 27 °C. For the in vitro Arp2/3-mediated actin polymerization assay, the Arp2/3 complex (10 nM, final concentration) and the WASP-VCA domain (400 nM, final concentration) were mixed with pyrene-actin (0.8 μM, final concentration) in polymerization buffer to initiate the reaction at 27 °C. To examine the effect of EPLIN on the actin polymerization, the in vitro actin polymerization assay and the Arp2/3-mediated actin polymerization assay were carried out in the presence of purified His-EPLIN (at 15 nM, 30 nM, or 60 nM). The elution buffer for His-tagged protein purification was used as a control. The fluorescence intensity of polymerized pyrene-actin was monitored for 70 min (with 1-min interval) by a fluorescence spectrophotometer (TECAN Infinite 200 PRO, excitation at 360 ± 9 nm and emission at 420 ± 20 nm). Statistics and reproducibility Densitometric quantitation of immunoblot images was performed using an Odyssey® CLx Imaging System (LI-COR Biosciences). The immunofluorescence images were quantified by the ZEN2 software (Carl Zeiss). Data were obtained from three independent experiments, and the number of cells counted in the experiments is described in the figure legends. Data analysis was performed using Excel or GraphPad Prism. The two-tailed Student’s t-test was used to determine whether the differences between experimental values were significant. P < 0.05 was considered statistically significant. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Transparent Peer Review file Supplementary Information Description of Additional Supplementary Files Supplementary Data Reporting Summary Transparent Peer Review file Supplementary Information Description of Additional Supplementary Files Supplementary Data Reporting Summary