Cofilin regulates actin filament dynamics by stimulating actin filament disassembly and plays a critical role in cytokinesis and chemotactic migration. Aip1 (actin-interacting protein 1), also called WDR1 (WD-repeat protein 1), is a highly conserved WD-repeat protein in eukaryotes and promotes cofilin-mediated actin filament disassembly in vitro; however, little is known about the mechanisms by which Aip1 functions in cytokinesis and cell migration in mammalian cells. In the present study, we investigated the roles of Aip1 in cytokinesis and chemotactic migration of human cells by silencing the expression of Aip1 using siRNA (small interfering RNA). Knockdown of Aip1 in HeLa cells increased the percentage of multinucleate cells; this effect was reversed by expression of an active form of cofilin. In Aip1-knockdown cells, the cleavage furrow ingressed normally from anaphase to early telophase; however, an excessive accumulation of actin filaments was observed on the contractile ring in late telophase. These results suggest that Aip1 plays a crucial role in the completion of cytokinesis by promoting cofilin-mediated actin filament disassembly in telophase. We have also shown that Aip1 knockdown significantly suppressed chemokine-induced chemotactic migration of Jurkat T-lymphoma cells, and this was blocked by expression of an active form of cofilin. Whereas control cells mostly formed a single lamellipodium in response to chemokine stimulation, Aip1 knockdown cells abnormally exhibited multiple protrusions around the cells before and after cell stimulation. This indicates that Aip1 plays an important role in directional cell migration by restricting the stimulus-induced membrane protrusion to one direction via promoting cofilin activity.
Actin cytoskeletal dynamics and reorganization play fundamental roles in a variety of cellular processes, including cell migration, cytokinesis, endocytosis and morphological change. During cytokinesis, an actomyosin-based contractile ring is formed and constricts at the equator of dividing cells before being disassembled at the end of cytokinesis . During cell migration, an F-actin-rich lamellipodial membrane protrusion is generated at the front of the cell and extends forward by actin filament assembly at the tip of the leading edge [2,3]. Actin filament dynamics and reorganization are regulated by a number of actin-binding proteins. Among them, cofilin is one of the most important regulators of actin filament dynamics, because it functions to sever and depolymerize actin filaments [4–6]. Cofilin is localized in the regions where actin filaments turn over rapidly, such as the cleavage furrow of dividing cells [7,8] and the lamellipodium of migrating cells [9,10]. Depletion or inactivation of cofilin results in a marked decrease in the actin monomer pool and an increase in filamentous actin structures within the cell and impairs cytokinesis and cell migration [4–6,11–18]. This indicates that cofilin plays a crucial role in cytokinesis and cell migration by promoting actin filament turnover through its actin filament-disassembling activity [17,18].
Cofilin activity is controlled in several ways, including phosphorylation and dephosphorylation [19–21], changes in pH , and association with phosphoinositides  and other regulatory proteins such as Aip1 (actin-interacting protein-1) and cyclase-associated protein [4–6]. Aip1, also called WDR1 (WD-repeat protein 1) in mammals, is a WD-repeat protein that is ubiquitously expressed in eukaryotes and has the potential to promote cofilin-mediated actin filament disassembly [5,6]. Genetic and biochemical studies have shown that Aip1 enhances actin filament disassembly only in the presence of cofilin, but has a negligible effect on actin dynamics on its own [24–27]. Two distinct mechanisms have been proposed for the action of Aip1 in promoting cofilin-mediated actin filament disassembly: Aip1 could cap the barbed ends of cofilin-severed actin filaments to prevent elongation and reannealing of severed filaments [28–30] and/or could actively sever cofilin-decorated actin filaments [31–35]. In addition, the cofilin-stimulating activity of Aip1 is further enhanced by co-operation with other actin-binding proteins, such as coronin or caspase 11 [36,37].
The cellular and biological functions of Aip1 have been investigated in several model organisms. For example, in budding yeast, the deletion of the Aip1 gene is viable, but results in a synthetic lethal phenotype when deleted in strains harbouring cofilin mutations [24,26,30,33]. In Dictyostelium, Aip1 is involved in endocytosis, cytokinesis and cell migration . In Caenorhabditis elegans, a null mutation in the Aip1 gene causes disorganized actin filament assembly in muscles of the body wall as well as defects in muscle contractile activity [34,38]. In Drosophila, mutations in the flare gene, which encodes Drosophila Aip1, cause F-actin accumulation and abnormal hair morphology . The genetic studies in these organisms suggest that Aip1 functionally interacts with cofilin and is involved in the regulation of actin filament dynamics and actin-dependent cellular processes in vivo by enhancing cofilin-mediated actin filament disassembly. In vertebrates, microinjection of Aip1 in Xenopus blastomeres causes cleavage arrest by disrupting the cortical accumulation of actin and cofilin . Recent studies have also shown that knockdown of Aip1 expression in mouse macrophage-like cells causes defects in cell migration , and knockdown of Aip1 in human osteosarcoma cells causes defects in mitotic cell rounding . Furthermore, mutations in the Aip1 gene in mice result in embryonic lethality, macrothrombocytopenia and autoinflammatory disease, depending on the level of gene disruption .
In the present study, we have examined the role of Aip1 in cytokinesis and the chemotactic migration of human cell lines by knocking down the expression of Aip1 using siRNA (small interfering RNA). Knockdown of Aip1 in HeLa cells induced the aberrant accumulation of F-actin near the contractile ring in late telophase and impaired cytokinesis, leading to an increase in the number of multinucleate cells. Knockdown of Aip1 in Jurkat leukaemic T-cells suppressed their chemotactic migration in response to SDF-1 (stromal cell-derived factor-1) and induced multiple membrane protrusions before and after cell stimulation. The ectopic expression of an active cofilin mutant significantly blocked the inhibitory effects of the Aip1 knockdown on cytokinesis and the chemotactic response. Our results indicate that Aip1 plays crucial roles in both cytokinesis and chemotactic migration of mammalian cells by promoting cofilin-mediated actin filament disassembly.
SDF-1 was purchased from PeproTech (London, U.K.). The rabbit polyclonal antibody against cofilin (COF-1) was prepared as described previously . Mouse monoclonal antibodies against the Myc epitope (9E10) and β-actin (AC-15) were purchased from Roche Diagnostics and Sigma respectively.
Plasmids coding for Myc- or YFP (yellow fluorescent protein)-tagged Aip1 were constructed by subcloning PCR-amplified human Aip1 cDNA into the pMYC-C1  or pEYFP-C1 (Clontech) expression vectors. Plasmids for YFP–actin, YFP–cofilin(S3A) and CFP–H2B [CFP (cyan fluorescent protein)-tagged histone H2B] were constructed as described previously [15,16]. The siRNA plasmids targeting human Aip1 and cofilin were generated using the pSUPER vector . The 19-base targeting sequences were as follows: 5′-GGAGCACCTGTTGAAGTAT-3′ (Aip1 siRNA-1), 5′-AGTGCGTCATCCTAAGGAA-3′ (Aip1 siRNA-2) and 5′-GGAGGATCTGGTGTTTATC-3′ (cofilin). As a control, we used a non-targeting sequence, 5′-GAATGTTGTGGTGGCTGCC-3′, which does not exist in the human genome. In order to construct expression plasmids coding for the sr-Aip1 (siRNA-resistant Aip1), three bases in the 19-base target sequence in the YFP–Aip1 or Myc–Aip1 plasmid were mutated (GGAACATCTCTTGAAGTAT) by using a QuikChange® site-directed mutagenesis kit (Stratagene).
Immunoblot analysis was performed as described previously .
Cell culture, synchronization and transfection
HeLa cells were cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% (v/v) FBS (fetal bovine serum). Cells were transfected with plasmids using Lipofectamine™-2000 (Invitrogen), according to the manufacturer's protocol. To analyse the distribution of YFP–Aip1, HeLa cells that had been co-transfected with YFP–Aip1 and CFP–H2B were cultured in DMEM containing 10% FBS and 1 mM thymidine for 36 h, and were then released for 12 h in fresh DMEM containing 10% FBS. For staining the mitotic cells, HeLa cells that had been transfected with plasmids were initially cultured in DMEM containing 10% FBS for 48 h and synchronized by incubating in DMEM containing 10% FBS and 1 mM thymidine for 36 h, followed by fresh DMEM containing 10% FBS for 6 h, DMEM containing 10% FBS and 0.3 μM nocodazole for 6 h, fresh DMEM containing 10% FBS for 2–3 h, and finally cells were fixed and stained. Jurkat human leukaemia T-cells were maintained in RPMI 1640 medium containing 10% FBS. For transfection, approx. 107 cells were mixed with plasmids in 400 μl of electroporation medium (RPMI 1640 medium containing 20% FBS and 25 mM Hepes, pH 7.4) before being electroporated at 280 V and 725 μF using a Gene Pulser II (Bio-Rad). After electroporation, the cells were cultured for 72 h in RPMI 1640 medium supplemented with 10% FBS. The transfection efficiency of the Jurkat cells was >90%, as assessed by transfection of the YFP plasmid.
HeLa cells were fixed in 4% formaldehyde in PBS for 20 min and permeabilized with 0.2% Triton X-100 in PBS for rhodamine–phalloidin staining, or with absolute methanol for cofilin staining for 10 min at −20 °C. After blocking with 1% BSA in PBS, cells were stained with rhodamine–phalloidin, DAPI (4′,6-diamidino-2-phenylindole) or an anti-cofilin antibody (COF-1). Rhodamine-conjugated anti-rabbit IgG (Chemicon, Temecula, CA, U.S.A.) was used as the second antibody. After washing with PBS, coverslips were mounted on a glass slip and images were obtained using fluorescence microscopy (model DMLB; Leica). Jurkat cells were suspended in RPMI 1640 medium containing 25 mM Hepes (pH 7.4) and 0.2% BSA, incubated for 20 min at 37 °C, then plated on coverslips and allowed to attach for 5 min at 37 °C. Attached cells were left unstimulated or were stimulated with 5 nM SDF-1 for 20 min and fixed with 3.7% (w/v) paraformaldehyde. The cells were permeabilized and stained using the HeLa cell protocol described above. Stacked fluorescent images were obtained using a laser scanning confocal microscope (LSM510; Carl Zeiss) equipped with a PL Apo (NA 1.4) ×63 oil immersion objective lens (Carl Zeiss).
For time-lapse imaging, Jurkat cells that had been electroporated with YFP–actin and siRNA plasmids were cultured for 72 h in RPMI 1640 medium supplemented with 10% FBS, and then suspended in RPMI 1640 medium containing 25 mM Hepes (pH 7.4) and 0.2% BSA. Cells were plated on to a 35 mm glass-bottom dish and incubated for 5 min to attach them to the dish. Then cells were stimulated with SDF-1, and time-lapse images of the stacked optical sections were collected every 10 s for 11 min using a laser scanning confocal microscope and objective lens, as described above. Kymograph analyses were done with ImageJ (http://rsb.info.nih.gov/ij/). To generate kymographs, a 0.42 μm×14 μm (3×100 pixels) rectangular region was drawn outwardly from the cell centre and the image information was compiled sequentially from each frame of the time-lapse movie. The frequency of lamellipodial protrusions for 10 min after SDF-1 stimulation was counted from the kymograph images. Protrusions with the peak height over 10 pixels (1.4 μm) in the kymograph were counted.
Transwell culture chambers (pore size: 5 μm; Costar) were used for the chemotactic cell migration assays. The lower wells were filled with 600 μl of medium (RPMI 1640 containing 0.5% BSA and 25 mM Hepes, pH 7.4) with or without 5 nM SDF-1. Jurkat cells (2×105 cells) suspended in 100 μl of the medium were loaded on to the upper wells. After incubation for 3 h at 37 °C, the cells that had migrated into the lower wells were counted.
Knockdown of Aip1 increases the percentage of multinucleate cells and this effect is significantly blocked by the overexpression of cofilin
In order to examine the role of Aip1 in mammalian cell cytokinesis, we constructed two distinct siRNA plasmids targeting human Aip1 (Aip1 siRNA-1 and -2) and analysed their effects in HeLa cells. The effect of these siRNAs on the expression of Aip1 was analysed by co-transfecting HeLa cells with the plasmid for Myc-tagged Aip1 as well as the plasmid for the control or Aip1 siRNAs. Immunoblot analysis revealed that both Aip1 siRNA plasmids effectively suppressed the expression of Myc–Aip1 (Figure 1A). To examine the effects of Aip1 knockdown on cytokinesis, HeLa cells were co-transfected with the YFP plasmid and the plasmid for the control or Aip1 siRNA with a molar ratio of 1:9. Cells were cultured for 96 h (a period sufficient for one or two cycles of cell division after knockdown of Aip1) and stained with DAPI (Figure 1B) and then the percentage of multinucleate cells in YFP-positive cells was determined (Figure 1C). While only a small percentage (<4%) of cells that had been transfected with the control siRNA plasmid exhibited multinucleate phenotypes, 14–15% of the cells that were transfected with each Aip1 siRNA plasmid became multinucleate (Figure 1C). The plasmid coding for YFP-tagged sr-Aip1 has three silent mutations in the siRNA target sequence of Aip1 cDNA. Co-transfection of this YFP–sr-Aip1 plasmid with the Aip1 siRNA significantly reduced the percentage of multinucleate cells (Figure 1D), indicating that the multinucleate phenotype of Aip1-knockdown cells can be attributed to suppression of the expression of endogenous Aip1. These results suggest that Aip1 plays a critical role in cytokinesis in HeLa cells. To examine whether Aip1 contributes to cytokinesis by regulating cofilin activity, HeLa cells were co-transfected with the plasmid for Aip1 siRNA and the plasmid for YFP or YFP-tagged cofilin(S3A), a constitutively active form of cofilin, in which the inhibitory phosphorylation residue (Ser-3) was replaced by alanine residue. After culture for 96 h, the number of multinucleate cells in YFP-positive cells was determined. Co-transfection of YFP–cofilin(S3A) with Aip1 siRNA significantly decreased the percentage of multinucleate cells, compared with cells co-transfected with Aip1 siRNA and control YFP (Figure 1E). Thus Aip1 appears to be involved in cytokinesis by enhancing cofilin activity.
Knockdown of Aip1 increases the number of multinucleate cells
Knockdown of Aip1 induces an excessive accumulation of actin filaments and an abnormal contractile ring
In order to investigate the role of Aip1 in cytokinesis, we examined changes in cell morphology and actin cytoskeletal organization in Aip1-knockdown cells. When asynchronized HeLa cells that had been co-transfected with YFP and Aip1 siRNA were fixed and stained with rhodamine-labelled phalloidin to detect actin filaments, Aip1-knockdown (YFP-positive) cells exhibited thicker actin fibres, compared with the surrounding non-transfected (YFP-negative) cells (Figure 2A, top panels). In addition, Aip1-knockdown cells occasionally induced membrane blebbing (Figure 2A, bottom panels). Quantitative analyses revealed that knockdown of Aip1 significantly increased the percentages of the cells with thick stress fibres and membrane blebs, compared with control siRNA cells (Figure 2A, right-hand panels). Similar phenotypes were also observed in cofilin-depleted cells [11,13,17] or LIM-kinase (LIM motif-containing protein kinase, where LIM is an acronym of the three gene products Lin-11, Isl-1 and Mec-3)-overexpressing cells, in which cofilin is excessively phosphorylated and inactivated [43,44]. These results indicate that Aip1 functions to promote cofilin activity in HeLa cells and that knockdown of Aip1 decreases the actin filament-disassembling activity of cofilin. We next examined the effects of Aip1 knockdown on the actin cytoskeletal organization in mitotic HeLa cells. HeLa cells that had been co-transfected with YFP and the control or Aip1 siRNA were synchronized with nocodazole, co-stained with rhodamine–phalloidin and DAPI, and the morphology and actin cytoskeletal organization of YFP-positive cells were analysed. In contrast with the reported phenotypes for U2OS osteosarcoma cells , Aip1-knockdown cells, similar to the results observed in control siRNA cells, normally rounded up from the substratum and became spherical in metaphase (Figures 2Ba and 2Bd), and then the cleavage furrow formed and constricted in anaphase and early telophase (Figures 2Bb and 2Be). However, in late telophase, actin filaments at the cleavage furrow accumulated excessively and asymmetrically in Aip1-knockdown cells, compared with those in control cells (Figures 2Bc and 2Bf), and the cleavage furrow subsequently regressed to produce binucleate cells (Figure 2Bg). Quantitative analysis showed that Aip1-knockdown cells exhibited the increased ratio of the cells with asymmetric cleavage furrow, compared with control siRNA cells (Figure 2B, histogram). These observations suggest that Aip1 plays a critical role in the completion of cytokinesis by regulating actin filament disassembly in the contractile ring in the final stage of cytokinesis.
Knockdown of Aip1 induces an excessive actin filament assembly and an abnormal contractile ring
Aip1 is diffusely localized in the cytoplasm during mitosis
In order to examine the subcellular localization of Aip1 in HeLa cells during mitosis, YFP–Aip1 was co-expressed with CFP–H2B, and the localization of YFP–Aip1 was analysed using fluorescence microscopy. Similar to our observations with control YFP, YFP–Aip1 was diffusely distributed in the cytoplasm in both metaphase and telophase, and no accumulation was observed near the cleavage furrow (Figure 3A).
Diffuse distribution of Aip1 and the effects of Aip1 knockdown on cofilin localization in HeLa cells
Localization of cofilin in Aip1-knockdown cells
To investigate further the role of Aip1 in cytokinesis, we examined the distribution of cofilin in Aip1-knockdown cells. When asynchronized HeLa cells were co-transfected with YFP and the control or Aip1 siRNA plasmids and stained with an anti-cofilin antibody, cofilin aberrantly accumulated on fibrous structures in Aip1-knockdown cells but localized diffusely in the cytoplasm in control siRNA cells (Figure 3B, upper panels). To determine whether the fibrous structures decorated with cofilin are actin fibres, HeLa cells were co-transfected with CFP–actin and Aip1 siRNA and the localization of actin and cofilin was analysed by CFP fluorescence and immunostaining with an anti-cofilin antibody. As shown in the lower panels in Figure 3(B), cofilin co-localized well with CFP–actin in Aip1 siRNA cells, thus indicating that knockdown of Aip1 causes aberrant localization of cofilin on actin stress fibres and cortical actin filaments. We next examined the effects of Aip1 siRNA on the localization of cofilin in mitotic HeLa cells. HeLa cells that had been co-transfected with CFP–H2B and the control or Aip1 siRNA plasmids were synchronized with nocodazole and the cofilin distribution was analysed with an anti-cofilin antibody (Figure 3C). Cofilin was diffusely distributed in the cytoplasm of both control and Aip1 siRNA cells in metaphase and accumulated on the contractile ring in late telophase; however, the localization of cofilin in late telophase differed between control and Aip1 siRNA cells. In control siRNA cells, cofilin was localized both in the cytoplasm and on the contractile ring. In contrast, in Aip1 siRNA cells, cofilin concentrated on the contractile ring and the cell cortex (Figure 3C). The line scan analyses across the control and Aip1-knockdown cells clearly showed the increase in the intensity of cofilin immunofluorescence in the cortical region of Aip1 siRNA cells, compared with control siRNA cells (Figure 3C, bottom panels). Thus the Aip1 knockdown appears to increase the localization of cofilin on cortical actin filaments.
Knockdown of Aip1 suppresses SDF-1-induced chemotaxis of Jurkat cells and this effect is significantly blocked by the overexpression of cofilin
To examine the role of Aip1 in the chemotactic response of mammalian cells, we analysed the effect of the Aip1 knockdown on SDF-1-induced chemotactic migration of Jurkat cells. Jurkat cells are known to express the SDF-1 receptor [CXCR4 (CXC chemokine receptor 4)] and exhibit chemotactic migration towards SDF-1 [16,45]. Jurkat cells were transfected with the siRNA plasmid targeting Aip1 or cofilin and cultured for 72 h before their chemotactic migration towards SDF-1 was analysed using Transwell culture chambers. SDF-1 was added to the lower chamber to analyse the chemotactic response, whereas it was not added for control experiments. As previously reported , knockdown of cofilin markedly reduced the chemotactic response of Jurkat cells towards SDF-1, compared with the cells that were transfected with the control siRNA plasmid (Figure 4A). Knockdown of Aip1 by either of two distinct siRNA constructs also suppressed chemotactic cell migration significantly, compared with control siRNA cells (Figure 4A). These results suggest that Aip1 is involved in the SDF-1-induced chemotactic response of Jurkat cells. Inhibition of the chemotactic response by Aip1 knockdown was significantly blocked by co-transfection with the Myc-tagged sr-Aip1 plasmid (Figure 4B), thus indicating that Aip1 siRNA inhibited the chemotactic response by suppressing endogenous Aip1 expression. To examine whether Aip1 knockdown inhibits chemotactic migration by regulating cofilin activity, Jurkat cells were co-transfected with Aip1 siRNA and YFP–cofilin(S3A), cultured for 72 h and subjected to the chemotaxis assay. Expression of cofilin(S3A) significantly blocked the inhibitory effect of Aip1 knockdown on the chemotactic response of Jurkat cells, which suggests that cofilin acts downstream of Aip1 (Figure 4C). To further examine whether cofilin and Aip1 function in the same pathway, we analysed the combined effects of cofilin and Aip1 double knockdowns on Jurkat cell migration. As shown in Figure 4(D), the combination of cofilin and Aip1 knockdowns did not additively decrease the number of migrating cells. Together, these results suggest that Aip1 contributes to the SDF-1-induced chemotactic migration of Jurkat cells by promoting cofilin activity.
Knockdown of Aip1 suppresses the chemotactic migration of Jurkat cells towards SDF-1
Knockdown of Aip1 induces multiple membrane protrusions in Jurkat cells before and after SDF-1 stimulation
To investigate the mechanisms by which Aip1 siRNA impaired Jurkat cell chemotaxis, alterations in actin filament assembly and cell morphology were analysed by rhodamine–phalloidin staining before and after SDF-1 stimulation. Jurkat cells that had been transfected with the plasmids for control, Aip1 or cofilin siRNA were left unstimulated or were stimulated for 20 min with SDF-1. Cells were then fixed and stained with rhodamine–phalloidin to visualize actin filaments. Most of the control siRNA-transfected cells exhibited a round and symmetrical morphology before SDF-1 stimulation and generated a single lamellipodial membrane protrusion on one side of the cell after SDF-1 stimulation (Figure 5A). In contrast, the cells that were transfected with Aip1 or cofilin siRNA exhibited aberrant F-actin assembly and multiple membrane protrusions around the cells both before and after SDF-1 stimulation (Figure 5A). Quantitative analyses revealed that the total numbers of the cells with single and multiple lamellipodial membrane protrusion(s) before SDF-1 stimulation were 20, 48 and 68%, for control, Aip1 and cofilin siRNA cells respectively (Figure 5B). Furthermore, the percentages of the cells with single and multiple lamellipodial protrusion(s) after SDF-1 stimulation were 71 and 21% for control siRNA cells, 46 and 44% for Aip1 siRNA cells and 24 and 64% for cofilin siRNA cells (Figure 5C). Thus the number of the multipolar cells significantly increased by the knockdown of Aip1 or cofilin. These results suggest that both Aip1 and cofilin play critical roles in maintaining the round morphology of Jurkat cells before SDF-1 stimulation and in restricting the lamellipodial membrane protrusion to one direction after SDF-1 stimulation.
Aip1- or cofilin-knockdown cells aberrantly induce multiple membrane protrusions before and after SDF-1 stimulation
Time-lapse analysis of the lamellipodial membrane protrusions in Aip1-knockdown cells
Knockdown of cofilin generates multiple lamellipodial membrane protrusions and remarkably suppresses the motion of the protrusions even after SDF-1 stimulation . We therefore examined the effect of Aip1 knockdown on the motility of the lamellipodial membrane protrusions. Jurkat cells that had been co-transfected with the plasmid for control, Aip1 or cofilin siRNA and the plasmid for YFP–actin were analysed by time-lapse fluorescence microscopy (Figure 6A and Supplementary Movies S1–S3 at http://www.BiochemJ.org/bj/414/bj4140261add.htm). Control siRNA cells maintained a round morphology before SDF-1 stimulation, but after SDF-1 stimulation YFP–actin quickly accumulated in the cell periphery and multiple membrane protrusions were generated around the cell (Figure 6A and Supplementary Movie S1). These protrusions moved rapidly and were then converted into a single lamellipodial protrusion by 10 min after stimulation. In contrast, cofilin-knockdown cells generated multiple large protrusions before and after SDF-1 stimulation, and these protrusions moved only slowly, compared with those in control cells (Figure 6A and Supplementary Movie S3). These observations suggest that cofilin-mediated actin remodelling plays a critical role in the dynamic motion of the lamellipodial membrane protrusion(s) and the formation of a single protrusion in response to SDF-1 stimulation. Aip1-knockdown cells also produced membrane protrusions in multiple directions before and after SDF-1 stimulation; however, these protrusions moved more dynamically compared with those in cofilin-knockdown cells (Figure 6A and Supplementary Movie S2). To quantify the effects of cofilin and Aip1 knockdowns on lamellipodial dynamics, we used kymograph analysis (Figure 6B). Kymographs showed that the lamellipodium rapidly protruded and retracted in control or Aip1-knockdown cells, but it moved slowly in cofilin-knockdown cells. When we calculated the average frequency of protrusions for 10 min after SDF-1 stimulation, cofilin-knockdown cells showed significantly less frequent protrusions than control or Aip1-knockdown cells (Figure 6C). Collectively, these results suggest that the knockdown of Aip1 is less effective than the knockdown of cofilin in the inhibition of lamellipodial dynamics.
Time-lapse analyses of the lamellipodial membrane protrusions in Aip1- or cofilin-knockdown cells
Aip1 is diffusely localized in Jurkat cells before and after SDF-1 stimulation
We examined the subcellular localization of Aip1 in Jurkat cells before and after SDF-1 stimulation by expressing YFP–Aip1 and analysing its distribution by confocal fluorescence microscopy (Figure 7A). YFP–Aip1 was diffusely distributed throughout the cytoplasm in Jurkat cells in a pattern similar to that observed with YFP alone. No accumulation of YFP–Aip1 was detected in the SDF-1-induced lamellipodial membrane protrusions in Jurkat cells.
Diffuse distribution of Aip1 and the effects of Aip1 knockdown on cofilin distribution in Jurkat cells
Localization of cofilin in Aip1-knockdown cells before and after SDF-1 stimulation
We next examined the localization of cofilin in Aip1-knockdown Jurkat cells. Cells that had been transfected with the control or Aip1 siRNA plasmid were left unstimulated or were stimulated for 20 min with SDF-1, before being fixed and stained with an anti-cofilin antibody. In control siRNA cells, cofilin accumulated in the lamellipodial membrane protrusion that was induced by SDF-1 stimulation (Figure 7B, upper panels). In Aip1 siRNA cells, multiple lamellipodial protrusions were generated both before and after SDF-1 stimulation, and cofilin accumulated in these membrane protrusions, in a similar manner to that observed in control siRNA cells (Figure 7B, lower panels). These results suggest that cofilin accumulates in the lamellipodial membrane protrusions, independently of Aip1.
In the present study, we provided evidence that Aip1 plays an important role in cytokinesis by positively regulating cofilin activity. We also showed that in Aip1-silenced HeLa cells, the cleavage furrow ingressed normally from anaphase to early telophase, but an excessive and asymmetric accumulation of actin filaments was observed near the contractile ring in late telophase. Similar phenotypes were observed in cofilin-depleted cells from several model organisms. For example, depletion of cofilin in Drosophila cells induced the aberrant accumulation of actin filaments and formation of a misshaped contractile ring in telophase, leading to frequent failure of cytokinesis [11,13]. In mammalian cells, the knockdown of cofilin or overexpression of a phosphatase-dead form of Slingshot (a cofilin-phosphatase that activates cofilin) induced the accumulation of F-actin on the contractile ring and the production of multinucleate cells [14,15,17]. These results suggest that cofilin activity is required for the accurate disassembly of actin filaments in the contractile ring in late telophase. The aberrant accumulation of actin filaments in cofilin- or Aip1-depleted cells probably inhibits the ring contraction and the abscission of the cell into two daughter cells, thus leading to the failure of cytokinesis. Consequently, Aip1 appears to play a critical role in cytokinesis, particularly in the completion of cytokinesis, by promoting the cofilin-mediated disassembly of actin filaments in the contractile ring in the final stage of cell division. A recent study showed that Aip1 knockdown in U2OS human osteosarcoma cells suppressed mitotic cell rounding, resulting in the cell flattening during mitosis . In contrast, we found that mitotic cell rounding was normal by Aip1 knockdown in HeLa cells.
We also provided evidence that Aip1 contributes to the chemotaxis of Jurkat cells through enhancing cofilin activity. Inhibition of the chemotactic response by knocking down Aip1 was also reported during fMLP (N-formylmethionyl-leucyl-phenylalanine)-mediated chemotaxis of J774 macrophage-like cells . In this study, we have shown that knocking down Aip1 induced changes in the actin filament organization and cell morphology in Jurkat cells. Phalloidin staining revealed that while most control cells have a round and symmetric morphology in the absence of SDF-1, Aip1-knockdown cells abnormally produced multiple lamellipodial membrane protrusions even before SDF-1 stimulation. When exposed to SDF-1, control cells produced multiple F-actin-rich lamellipodial protrusions around the cell in the initial stages of the cell response (1–2 min after stimulation), which then converted into a single lamellipodium by 10 min. In contrast, Aip1-knockdown cells retained multiple lamellipodial protrusions for up to 20 min after SDF-1 stimulation. These results suggest that Aip1 is required for maintaining the round morphology of the non-stimulated Jurkat cells and producing the single lamellipodium and polarized cell morphology after cell stimulation. As described previously [16,46] and shown in the present study, knockdown of cofilin also induced multiple lamellipodial protrusions around the cells before and after stimulation of Jurkat cells and MTC mammary adenocarcinoma cells. It is therefore likely that the Aip1 knockdown causes a decrease in cofilin-mediated actin filament disassembly in the cells and, as a result, leads to an excessive actin filament assembly, a decrease in actin filament turnover in the lamellipodia and a failure in the remodelling of multiple lamellipodia into a single lamellipodium. The inability to form a single lamellipodium in the direction of cell migration probably causes the inhibition of the chemotactic migration of Jurkat cells. Although both Aip1- and cofilin-knockdown cells generated multiple membrane protrusions, the cofilin knockdown exhibited more severe phenotypes in F-actin accumulation in the lamellipodia, the motility of lamellipodia and chemotactic migration towards SDF-1. Time-lapse analyses and kymographs showed that the dynamic movement of the lamellipodial protrusions was markedly lost in cofilin-knockdown cells, whereas it was almost retained in Aip1-knockdown cells. These differences in phenotype are probably due to the fact that cofilin still exists in Aip1 siRNA cells and has its own basal activity to promote actin filament disassembly, whereas in cofilin siRNA cells, cofilin expression is directly suppressed. Taken together, our results suggest that Aip1 plays a crucial role in chemokine-induced directional cell migration by enhancing actin filament dynamics and the formation of a single lamellipodium in the direction of cell migration via its action to promote cofilin activity.
We showed that Aip1 knockdown caused the aberrant accumulation of F-actin and unusual localization of cofilin on thick actin fibres in interphase HeLa cells. These observations are consistent with the previously reported observations in yeast and C. elegans where mutations in the Aip1 gene induced the formation of thick actin fibres and the alteration of cofilin localization on these F-actin structures [30,34]. It is likely that the actin filament-disassembling activity of cofilin is weakened in Aip1-knockdown cells; therefore actin filaments are stabilized and the cofilin that is present on the actin filaments is retained for an extended period of time. In HeLa cells, cofilin is localized diffusely throughout the cytoplasm in metaphase in both control and Aip1-knockdown cells; this is probably because cofilin is highly phosphorylated and loses its F-actin binding ability in metaphase of the cell cycle [14,47]. In late telophase, cofilin is dephosphorylated and concentrated on to the contractile ring in both control and Aip1-knockdown cells; however, while cofilin is localized in the cytoplasm and on the contractile ring in control cells, it accumulates in both the contractile ring and the cell cortex in Aip1-knockdown cells. Again, this could be due to the decrease in cofilin activity in Aip1-knockdown cells, which may lead to the stabilization of cofilin-decorated actin filaments in the cell cortex. Aip1 was distributed diffusely in the cytoplasm and its co-localization with F-actin or cofilin was not observed in either HeLa or Jurkat cells, although Aip1 has the potential to bind to F-actin and cofilin in biochemical studies [25,26]. The diffuse distribution of Aip1 may suggest that Aip1 can disassemble actin filaments very rapidly, once it associates with cofilin-bound actin filaments.
Whereas cofilin promotes actin filament disassembly by severing and depolymerizing actin filaments in vitro, it is also required for actin filament assembly in the cell. Two distinct models have been proposed to explain the role of cofilin in actin filament assembly in vivo. The first model proposes that cofilin contributes to actin filament assembly by disrupting actin filaments and thereby supplying actin monomers for polymerization [3,17,18]. We and other investigators have shown that cofilin is involved in producing an abundant pool of actin monomers in the cytoplasm and that these monomers are used for actin polymerization in the cell [17,18], hence supporting this model. An alternative model is that cofilin can promote actin filament assembly by severing actin filaments and creating free barbed ends that are used as nucleation sites for actin polymerization [48,49]. This model is possible under the conditions that actin monomers are sufficiently available and the cofilin-created barbed ends remain uncapped. As Aip1 has the potential to cap the barbed ends of cofilin-severed actin filaments, it can block the elongation or reannealing of cofilin-severed actin filaments. Thus, in cells that express high levels of Aip1, it is implausible that cofilin mediates actin filament assembly by creating the nucleation sites for polymerization. We have also shown that the severing activity, rather than the depolymerizing activity, of cofilin plays a dominant role in increasing the actin monomer pool in the cell . The ability of Aip1 to cap the barbed ends of cofilin-severed actin filaments could ensure the cofilin-induced actin filament disruption (by severing of actin filaments and subsequent breakdown of F-actin fragments to actin monomers) and seems to play an important role in the regulation of the balance of cofilin-mediated actin filament disassembly and assembly in the cell.
In conclusion, we have provided evidence that Aip1 plays critical roles in mammalian cell cytokinesis and chemotactic migration by promoting cofilin activity. Aip1 regulates the actin filament dynamics of the contractile ring and is crucial for completing cell abscission at the final stage of cytokinesis. Aip1 is also required for maintaining the round cell morphology of nonstimulated Jurkat cells and contributes to SDF-1-induced directional cell migration by restricting the lamellipodial protrusion to one side of the cell. Further studies on the mechanisms regulating Aip1 activity will be important to understand the spatiotemporal control of cofilin activity and actin cytoskeletal dynamics during cytokinesis and polarized cell migration.
This study was supported by a grant-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
cyan fluorescent protein
CFP-tagged histone H2B
Dulbecco's modified Eagle's medium
fetal bovine serum
stromal cell-derived factor-1
small interfering RNA
yellow fluorescent protein
These authors contributed equally to this work.