The Rockefeller University Press $30.00
J. Cell Biol. Vol. 191 No. 3 453–462
Correspondence to Fumiko Toyoshima: firstname.lastname@example.org; or
Eisuke Nishida: email@example.com
Abbreviation used in this paper: CH, calponin homology.
Cell morphological changes during mitosis are accompanied by
dynamic rearrangements of the actin cytoskeleton. Actin fila-
ments are found beneath the cortical plasma membrane and in
the retraction fibers during early M phase, and also in the con-
tractile ring at the equatorial region during late M phase. The
cortical flow (Bray and White, 1988), which is generated by the
interaction of the cortical actin filaments with myosin, is shown
to be required for the proper centrosome separation and posi-
tioning (Rosenblatt et al., 2004). Moreover, the cortical actin
filaments play an important role in the orientation of the mitotic
spindle (Théry et al., 2005; Toyoshima and Nishida, 2007).
How the cortical actin filaments are rearranged at
the onset of M phase has remained unclear. Small GTPases
(Etienne-Manneville and Hall, 2002; Maddox and Burridge,
2003; Dao et al., 2009) have been shown to regulate cell round-
ing. Furthermore, dMoesin regulates the rearrangement of corti-
cal actin filaments during mitosis, which is important for cortical
stiffening (Carreno et al., 2008; Kunda et al., 2008). AIP and
cofilin are also involved in the cell rounding (Fujibuchi et al.,
2005). Although many players have been identified, the detailed
dynamics and mechanisms for actin rearrangements during
mitosis have not been fully elucidated.
Here, we find a novel phenomenon of actin assembly
dynamics during mitosis: formation of an amorphous actin
cluster and its revolving movement. Our analyses demonstrate
that Arp2/3 is essential for this dynamic actin cluster.
Results and discussion
An amorphous cluster of actin filaments is
formed and revolves during mitosis
To examine actin dynamics in living cells, we expressed a
calponin homology (CH) domain of utrophin fused to GFP
(GFP-UtrCH; Burkel et al., 2007; Woolner et al., 2008; Miller
and Bement, 2009), which binds to actin filaments and has been
used to visualize actin filaments in living cells. Time-lapse
observations in HeLa cells have unexpectedly revealed that an
amorphous cluster of GFP-UtrCH appears outside the nucleus
during prometaphase, and it moves around along the cell cor-
tex at a roughly constant speed until telophase (Fig. 1 A and
Video 1). This actin cluster underwent changes in its shape and
size during the revolving movement. Staining with phalloidin
and anti-actin antibody indicated that the cluster of GFP-UtrCH
consists of F-actin (Fig. 1 B). Observations in cells expressing
amorphous cluster of actin filaments appears during pro-
metaphase, revolves horizontally along the cell cortex
at a constant angular speed, and fuses into the contractile
ring after three to four revolutions. Cdk1 activity is re-
quired for the formation of this mitotic actin cluster and its
he actin cytoskeleton undergoes rapid changes in its
architecture during mitosis. Here, we demonstrate
novel actin assembly dynamics in M phase. An
revolving movement. Rapid turnover of actin in the fila-
ments takes place everywhere in the cluster and is also re-
quired for its cluster rotation during mitosis. Knockdown
of Arp3, a component of the actin filament–nucleating
Arp2/3 complex, inhibits the formation of the mitotic
actin cluster without affecting other actin structures. These
results identify Arp2/3 complex as a key factor in the
generation of the dynamic actin cluster during mitosis.
Revolving movement of a dynamic cluster of actin
filaments during mitosis
Masaru Mitsushima,1 Kazuhiro Aoki,2 Miki Ebisuya,1 Shigeru Matsumura,3 Takuya Yamamoto,1 Michiyuki Matsuda,2
Fumiko Toyoshima,3 and Eisuke Nishida1
1Department of Cell and Developmental Biology and 2Laboratory of Bioimaging and Cell Signaling, Graduate School of Biostudies, and 3Laboratory of Subcellular
Biogenesis, Institute for Virus Research, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan
© 2010 Mitsushima et al. This article is distributed under the terms of an Attribution–
Noncommercial–Share Alike–No Mirror Sites license for the first six months after the pub-
lication date (see http://www.rupress.org/terms). After six months it is available under a
Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license,
as described at http://creativecommons.org/licenses/by-nc-sa/3.0/).
T H E J O U R N A L O F C E L L B I O L O G Y
JCB • VOLUME 191 • NUMBER 3 • 2010 454
Figure 1. An amorphous cluster of actin filaments revolves along the cell cortex. (A) Time-lapse images of HeLa cells expressing GFP-UtrCH and DsRed-
histone H1 during metaphase. GFP-UtrCH images were taken every 5 s, and images are shown at 30-s intervals. (B) Staining with anti-actin antibody (left)
or Alexa Fluor 546–phalloidin (right) of HeLa cells expressing GFP-UtrCH. (C) Staining with anti-actin antibody (left) or Alexa Fluor 546–phalloidin (right) of
Actin wave during mitosis • Mitsushima et al.
at all (unpublished data), Cdk1 inhibition by Ro3306 (Vassilev
et al., 2006) resulted in the disappearance of the actin cluster,
which suggests the requirement of Cdk1 activity for the actin
cluster dynamics. However, as the inhibition of Cdk1 activity
during prometaphase induces the cleavage furrow by promoting
M phase exit, the result might not necessarily demonstrate the
Cdk1 requirement for the actin dynamics. Then, to prevent the
cleavage furrow formation, we treated cells with nocodazole to
arrest cells in prometaphase. The treatment did not inhibit the
actin cluster formation and movement. In these cells, the addi-
tion of Ro3306 resulted in nearly complete disappearance of the
actin cluster within 20–30 min (Fig. 2 E and Video 2). As Cdk1-
dependent phosphorylation of various proteins (Fig. S2) as well
as several Cdk1-dependent events, such as chromosome con-
densation and cell rounding, also diminished within 15–30 min,
it is likely that Cdk1 activity is required for the actin cluster
formation and its revolving movement.
Cell rounding and cell substratum adhesion
are important for the actin cluster
formation and revolving movement
Time-lapse observations in various cell lines showed that for-
mation of a single amorphous actin cluster and its revolving
movement, which are similar to those in HeLa cells, occurred
during M phase in MCF-7, HepG2, and Cos1 cells (Video 3 and
unpublished data), whereas no obvious actin cluster formation
and movement occurred in NIH3T3, MDCK, and C3H10T1/2
cells (unpublished data). We noted that those cells that ex-
hibit these dynamic actin behaviors show almost complete cell
rounding during M phase, whereas the cell shape of those cells
that do not exhibit these actin behaviors is far from globular
even in M phase. Of note, these actin behaviors can only be
observed in M phase, when cells are rounding. Therefore, there
is a correlation between the extent of cell rounding and the oc-
currence of these dynamic actin behaviors in M phase. In good
agreement with this idea, in HaCaT, HEK293T, A431, and KB
cells, which show incomplete cell rounding during M phase, the
actin cluster formation and movement were also incomplete;
i.e., formation of a single actin cluster was not clearly observed,
but regions of higher densities of actin filaments were formed in
early M phase, and they moved around incompletely along the
cell cortex (Video 4 and unpublished data).
Next, we tested the possibility that cell rounding is suffi-
cient to induce the actin cluster dynamics. We cultured HeLa cells
in 3D by totally embedding the cells in a gel of reconstituted
basement membrane matrix. Under the 3D culture conditions,
HeLa cells are round throughout the cell cycle (Fig. 3 A). Time-
lapse observations of cells with GFP-UtrCH showed that some
heterogeneous distribution of actin filaments could be detected
even in interphase, but this heterogeneous distribution pattern
did not vary appreciably with time (Fig. 3, A and B, interphase).
GFP-actin have also revealed the formation of an amorphous
cluster of F-actin and its revolving movement during mitosis
(Fig. S1, A and B). Moreover, staining of control HeLa cells,
which do not express exogenous proteins, showed that an amor-
phous cluster, which is stained with both phalloidin and anti-
actin antibody, exists along the cell cortex during prometaphase
to anaphase, and that this actin cluster resembles, in its loca-
tion and shape, the actin cluster visualized with GFP-UtrCH or
GFP-actin (Fig. 1 C). Thus, the actin cluster, which is visual-
ized with GFP-UtrCH or GFP-actin, is not an artifact resulting
from their overexpression. Collectively, these results show that
an amorphous cluster of actin filaments is formed during early
prometaphase, and it revolves along the cell cortex until ana-
phase in HeLa cells.
Time-lapse observations of 146 mitotic cells with GFP-
UtrCH demonstrated that every cell exhibited the formation
of an amorphous actin cluster, which always revolved during
M phase, and that the plane of revolving movement, for the
most part (>90%), was oriented parallel to the substrate surface;
i.e., the actin cluster revolved horizontally (Fig. 1 D, left). Once
the cluster started to revolve, it did not change the direction of
movement in most cells; the direction was about half clockwise
and half counterclockwise (Fig. 1 D, right). In rare occasions
(13%; both in Fig. 1 D, right), however, changes in direction
occurred during the movement. A spatiotemporal representation
of the actin cluster movement (Figs. 1 E and S1 C) demonstrates
a revolving movement with constant angular velocity in each
cell, and its Fourier transformation (Fig. 1 F and S1 D) indi-
cates that the mean value of frequency = 0.0026 ± 0.00062 Hz
(period of 404 ± 117 s, n = 57; Fig. 1 G). The angular velocity
varied slightly from cell to cell, but was not highly dependent on
the cell size (correlation coefficient, 0.361; Fig. 1 H). Thus,
we could see approximately three to four revolutions during
Cdk1 activity is required for the actin
cluster formation and revolving movement
Phalloidin staining in cells with no exogenous proteins has
shown that an amorphous cluster of actin filaments is de-
tected during prometaphase to anaphase, but not in G2 phase,
prophase, telophase, or G1 phase (Fig. 2, A and C). This is in
good agreement with observations with GFP-UtrCH, which
show that an amorphous actin cluster appears during early
prometaphase and disappears during anaphase to telophase
(Fig. 2, B and C). A close examination of time-lapse images
suggests that the amorphous cluster fuses into the contractile
ring during telophase (Fig. 2 D). Thus, this actin dynamic is
an M phase–specific event. We then examined the effect of the
addition of the inhibitors of mitotic kinases during prometaphase
on the actin cluster revolving movement. Although inhibitors
of Aurora B and Plk-1 did not affect the actin cluster dynamics
HeLa cells. Arrowheads indicate an amorphous cluster of actin filaments. (D) Direction of the revolving movement of the amorphous actin cluster (n = 146).
(E) Spatiotemporal representation of the revolving movement of the amorphous actin cluster of A. Intensities of GFP-UtrCH signals in areas between the
0.70–0.75 radius away from the center of the cell were averaged and plotted. (F) Fourier transformation of E. (G) Histogram of the period of the revolv-
ing movement of the cluster (n = 57). (H) Relationship between the radius of cells and the period of the revolving movement of the actin cluster (n = 57).
Bars: (A) 10 µm; (B and C) 5 µm.
JCB • VOLUME 191 • NUMBER 3 • 2010 456
Figure 2. Cdk1 activity is required for the actin cluster formation and revolving movement. (A) Staining with phalloidin of HeLa cells at the indicated
phases (G2 to post-mitosis phase). Arrowheads, the cluster of actin filaments. (B) Time-lapse images of HeLa cells expressing GFP-UtrCH and DsRed-histone
H1 from G2 to post-mitosis phase. Images were taken every 3 min. Arrowheads, the cluster of actin filaments. (C) Red bars represent the percentages of
457 Actin wave during mitosis • Mitsushima et al.
cells containing the cluster of actin filaments stained with phalloidin (n = 50 at each phase). Blue bars represent the percentages of cells exhibiting a revolv-
ing actin cluster visualized by GFP-UtrCH (n = 50). (D) Close examination of time-lapse images of HeLa cells expressing GFP-UtrCH and DsRed-histone H1.
Asterisks indicate the region of the cleavage furrow. Images were taken every 30 s and images are shown at 1.5-min intervals. Arrowheads indicate the
actin cluster. (E) Nocodazole-treated, prometaphase-arrested HeLa cells expressing GFP-UtrCH and DsRed-histone H1 were treated with 20 µM Ro3306
immediately after time 0. Images were taken every 2 min. Arrowheads, the actin cluster. Bars: (A, D, and E) 10 µm; (B) 5 µm.
Figure 3. Cell rounding and cell–substratum adhesion are important for the revolving movement of the actin cluster. (A) Time-lapse images of HeLa cells ex-
pressing GFP-UtrCH and DsRed-histone H1 during interphase (top) and M phase (bottom) under the 3D condition. Images were taken every 1 min or 30 s,
respectively. The pictures of interphase or M phase, at every 1 min or 2 min, respectively, are shown. Arrowheads indicate the actin cluster. Bar, 10 µm.
(B) Spatiotemporal representation of GFP signals of A. (C) Confocal images of cross sections of living HeLa cells expressing GFP-UtrCH.
JCB • VOLUME 191 • NUMBER 3 • 2010 458
During M phase, however, a relatively large amorphous cluster
of actin filaments, within which there were several regions of
higher densities of actin filaments, appeared outside the nucleus
and moved around along the cell cortex, similar to the amor-
phous actin cluster under the 2D conditions (Fig. 3, A and B,
M phase; and Video 5). The orientation of the plane of revolv-
ing movement under the 3D conditions, however, shows a ran-
dom distribution from parallel to vertical, and it varies even
during the revolving movement. These results indicate that cell
rounding alone is unable to induce the actin cluster formation
and revolving movement, and that cell substratum adhesion
should be important for determining the orientation of the plane
of revolving movement, as these results can be interpreted as
the plane of revolving movement being oriented parallel to the
cell substratum adhesion surface. Remarkably, x-y, y-z, and x-z
images of HeLa cells with GFP-UtrCH cultured in normal 2D
conditions demonstrate that densities of actin filaments in the
cluster become higher in the bottom part (Fig. 3 C), and this
may suggest that most of actin filaments in the cluster originate
from the bottom part of the cell, which adheres to the substratum.
Thus, cell–substratum adhesion should be important for the
actin cluster formation and revolving movement.
A rapid actin turnover in the actin cluster
The mitotic amorphous actin cluster undergoes changes in its
shape and size continuously (Fig. 1 A), which suggests that
actin polymerization and depolymerization actively occur in
the cluster. Treatment with the F-actin–depolymerizing drugs
latrunculin B and cytochalasin D or the F-actin–stabilizing drug
jasplakinolide resulted in the rapid cessation of the revolving
movement and the disappearance of the actin cluster (Fig. S3).
This indicates that actin polymerization and depolymerization
reactions are required for generating the mitotic actin cluster
and movement. We then performed a FRAP experiment. After
photobleaching in the actin cluster, a rapid fluorescent recovery
of GFP-actin was observed with a half-life (1/2) of 10.6 ± 5.9 s
(Fig. 4, A and B; and Video 6), which indicates a rapid turnover
of actin in the cluster. The fluorescent intensity was restored to
almost the original level (Fig. 4 B). Of note, we did not observe
any forward or backward movement of the bleached area in the
migrating actin cluster (Fig. 4 A). These results suggest that the
actin cluster movement is not a movement of stable actin fila-
ments but is generated by a rotational movement of the site with
the high actin polymerization activity, which may locate in the
cell cortex on the cell–substratum adhesion surface. In other
words, a zone, in which actin polymerization and depolymer-
ization occur actively, should revolve in the bottom part of the
cell, generating a wavelike movement of an amorphous actin
cluster during M phase, like an audience raising their hands to
generate a wave in a baseball stadium (Fig. 4 C).
Arp2/3 complex is essential for the actin
cluster formation and revolving movement
Our results thus indicated the importance of actin-nucleating
activity in the mitotic actin cluster dynamics. To identify a fac-
tor responsible for the actin cluster formation, we performed
RNAi knockdown of actin-nucleating factors (Pollard, 2007;
Chesarone and Goode, 2009). Although knockdown of mDia1,
mDia2, or spire had no effect on the actin cluster (unpublished
data), knockdown of Arp3, a component of Arp2/3 complex,
completely inhibited the actin cluster formation (Fig. 5, A and B).
Consistent with previous studies (Steffen et al., 2006; Gomez
et al., 2007), Arp3 knockdown resulted in concomitant reduction
of Arp2 and ArpC2, two other components of Arp2/3 com-
plex (Fig. 5 A). Two different siRNA sequences against Arp3,
which down-regulated Arp3, gave essentially the same results.
Moreover, expression of Arp3res, which is resistant to the
siRNA but encodes for wild-type Arp3 protein, in the Arp3
siRNA-transfected cells resulted in the recovery of the actin clus-
ter formation and movement (Fig. 5, C and D). Remarkably,
knockdown of Arp3 did not significantly affect the cortical actin
structures and retraction fibers (Fig. 5 E), which indicates that
the Arp2/3 complex is specifically involved in the generation of
the revolving actin cluster during mitosis.
We then determined the subcellular localization of Arp3.
Although dense staining in the cytoplasm with anti-Arp3 anti-
body hindered detection of specific localization of Arp3 in ordi-
narily prepared samples, the patchy staining of Arp3, which
roughly coincided with the amorphous actin cluster and the
cortical structures, was clearly detected in permeabilized cells
(Fig. 5 F). The area showing higher densities of the patchy spots
of Arp3 staining coincided with the amorphous actin cluster area,
which was revealed by GFP-UtrCH (Fig. 5 G). The patchy Arp3
staining disappeared in Arp3 siRNA-treated cells (Fig. 5 F),
confirming the specificity of Arp3 staining. Moreover, nei-
ther patchy staining nor colocalization with the actin cluster was
observed in the staining of the same samples with an antibody
against Arp1, which has been shown to localize in the spindle
(Fig. 5 F; Clark and Meyer, 1999). In addition, the Arp3 siRNA
did not affect the staining with the anti-Arp1 antibody, con-
firming the specificity of the Arp3 siRNA. These results suggest
that Arp3 specifically resides in the mitotic actin cluster.
How does the mitotic actin cluster revolve? Our analyses
suggest a model in which a zone, where actin polymerization
and depolymerization occur actively, revolves in waves along
the cell cortex (Fig. 4 C). Because actin polymerization and de-
polymerization occur actively everywhere in the cluster, but not
in specific sites such as the forward or backward area of the
direction of revolving movement, the unidirectional actin poly-
merization may not be a driving force. The cell cortex has been
shown to flow from the polar regions of the cell to the equator
during M phase. This is called “cortical flow,” and was shown to
be inhibited by inhibition of myosin II (Rosenblatt et al., 2004).
However, inhibition of myosin II did not inhibit the actin cluster
revolution (unpublished data), which suggests that neither corti-
cal flow nor myosin II plays a role in the mitotic actin cluster
dynamics. Therefore, elucidation of molecular mechanisms of
the mitotic actin cluster dynamics should await further studies.
Murthy and Wadsworth (2008) found that a movement of
wavelike changes in GFP-actin fluorescence can be observed in
LLC-Pk1 cells only when astral microtubules are disrupted by
nocodazole. As we can clearly detect the revolving actin cluster
in both intact cells and nocodazole-treated cells, whether the same
mechanisms underlie these two apparently similar phenomena
459Actin wave during mitosis • Mitsushima et al.
actin cluster, which is generated by Arp2/3-mediated actin nu-
cleation, plays a role in controlling cell division. However, there
was also a report showing that depletion of the Arp2/3 complex
has no effect on cytokinesis (Bompard et al., 2008). The next
challenge is to identify physiological roles of the revolving
movement of the mitotic actin cluster.
Materials and methods
Cytochalasin D was obtained from Sigma-Aldrich, and latrunculinB, jasplakino-
lide, nocodazole, and Ro-3306 were obtained from EMD. The following
antibodies were used: anti-actin (AC15), anti–-tubulin (DM1A), and anti-
Arp3 (FMS338; Sigma-Aldrich); anti-cyclinB1 and anti-cMyc (9E10; Santa
Cruz Biotechnologies, Inc.) monoclonal antibodies; and anti-Arp1 (Sigma-
Aldrich), anti-Arp2 (Santa Cruz Biotechnologies, Inc.), and anti-p34-Arc/
ARPC2 (Millipore) rabbit polyclonal antibodies. For actin filament staining, we
used Alexa Fluor 488–phalloidin or Alexa Fluor 546–phalloidin (Invitrogen).
Human -actin and CH domain of human utrophin (Burkel et al., 2007)
were amplified from the HeLa cDNA library by PCR and subcloned into
should be examined in future studies. A filamentous actin mesh
was observed in starfish oocytes (Lénárt et al., 2005). A cloud of
dynamic actin filaments (Li et al., 2008) or filamentous actin
meshes (Azoury et al., 2008) were also found in mouse oocytes.
Although these mouse structures were shown to depend on
formin2, the revolving actin cluster in this study depends on
Arp2/3, but not on mDia1 and -2, other formin family members.
It will be interesting to see whether these structures share com-
mon molecular mechanisms.
What is the physiological role of the revolving movement
of the mitotic actin cluster? Arp2/3 has been shown to be re-
quired for cell division in yeast (Winter et al., 1997), and a wsp1
(Wiskott–Aldrich syndrome protein [WASP] is an Arp2/3 acti-
vator) mutant of C. elegans shows a partial cytokinesis defect
(Withee et al., 2004). Moreover, the forced activation of the
Arp2/3 complex was reported to cause the delay in M phase
progression and the increase in the number of multinucleic cells
(Moulding et al., 2007). Thus, it is possible that the revolving
Figure 4. Actin dynamics in the actin cluster. (A) Time-lapse confocal images of HeLa cells expressing GFP-actin during prometaphase. Images were taken
every 7.3 s. At time 0, photobleaching was performed in the area shown by the red circle. Bar, 5 µm. (B) Quantification of FRAP analysis (n = 30). Error
bars indicate SD. (C) A model of the amorphous actin cluster movement during mitosis. A region (pink) in the bottom part of the cell indicates a zone in
which actin polymerization and depolymerization occur actively. Red, actin filaments; yellow, the direction of the revolving movement.
JCB • VOLUME 191 • NUMBER 3 • 2010 460
Figure 5. The Arp2/3 complex is essential for the actin cluster formation and revolving movement. (A) Arp3 siRNAs efficiently down-regulate the Arp2/3
complex. (B) Time-lapse images of HeLa cells expressing GFP-UtrCH and DsRed-histone H1. Top, control (Luc siRNA); bottom, Arp3 siRNA. Images were
taken every 3 s and images are shown at 30-s intervals. Spatiotemporal representation is also shown. (C) Time-lapse images of HeLa cells expressing
461Actin wave during mitosis • Mitsushima et al.
that Cdk1 activity is required for the actin cluster formation and its revolv-
ing movement. Video 3 shows the revolving movement of the amorphous
cluster of actin filaments during M phase in an MCF-7 cell expressing
GFP-UtrCH and DsRed-histone H1 during M phase. Video 4 shows the
revolving movement of the amorphous cluster of actin filaments during
M phase in a HaCaT cell expressing GFP-UtrCH and DsRed-histone H1
during M phase. Video 5 shows the revolving movement of the amorphous
cluster of actin filaments in a HeLa cell expressing GFP-UtrCH and DsRed-
histone H1 under the 3D conditions during M phase. Video 6 shows the
actin turnover in the amorphous cluster of actin filaments in HeLa cells
expressing GFP-actin during M phase. Online supplemental material is avail-
able at http://www.jcb.org/cgi/content/full/jcb.201007136/DC1.
This work was supported by grants from the Ministry of Education, Culture,
Sports, Science and Technology of Japan (to E. Nishida and F. Toyoshima) and
a grant from Takeda Science Foundation (to E. Nishida).
Submitted: 23 July 2010
Accepted: 4 October 2010
Azoury, J., K.W. Lee, V. Georget, P. Rassinier, B. Leader, and M.H. Verlhac.
2008. Spindle positioning in mouse oocytes relies on a dynamic mesh-
work of actin filaments. Curr. Biol. 18:1514–1519. doi:10.1016/j.cub
Ballestrem, C., B. Wehrle-Haller, and B.A. Imhof. 1998. Actin dynamics in
living mammalian cells. J. Cell Sci. 111:1649–1658.
Beli, P., D. Mascheroni, D. Xu, and M. Innocenti. 2008. WAVE and Arp2/3
jointly inhibit filopodium formation by entering into a complex with
mDia2. Nat. Cell Biol. 10:849–857. doi:10.1038/ncb1745
Bompard, G., G. Rabeharivelo, and N. Morin. 2008. Inhibition of cytokinesis by
wiskostatin does not rely on N-WASP/Arp2/3 complex pathway. BMC
Cell Biol. 9:42. doi:10.1186/1471-2121-9-42
Bray, D., and J.G. White. 1988. Cortical flow in animal cells. Science. 239:883–
Burkel, B.M., G. von Dassow, and W.M. Bement. 2007. Versatile fluorescent
probes for actin filaments based on the actin-binding domain of utrophin.
Cell Motil. Cytoskeleton. 64:822–832. doi:10.1002/cm.20226
Carreno, S., I. Kouranti, E.S. Glusman, M.T. Fuller, A. Echard, and F. Payre.
2008. Moesin and its activating kinase Slik are required for cortical stabil-
ity and microtubule organization in mitotic cells. J. Cell Biol. 180:739–
Chesarone, M.A., and B.L. Goode. 2009. Actin nucleation and elongation fac-
tors: mechanisms and interplay. Curr. Opin. Cell Biol. 21:28–37. doi:10
Clark, I.B., and D.I. Meyer. 1999. Overexpression of normal and mutant
Arp1alpha (centractin) differentially affects microtubule organization
during mitosis and interphase. J. Cell Sci. 112:3507–3518.
Dao, V.T., A.G. Dupuy, O. Gavet, E. Caron, and J. de Gunzburg. 2009. Dynamic
changes in Rap1 activity are required for cell retraction and spreading
during mitosis. J. Cell Sci. 122:2996–3004. doi:10.1242/jcs.041301
Etienne-Manneville, S., and A. Hall. 2002. Rho GTPases in cell biology. Nature.
Fujibuchi, T., Y. Abe, T. Takeuchi, Y. Imai, Y. Kamei, R. Murase, N. Ueda, K.
Shigemoto, H. Yamamoto, and K. Kito. 2005. AIP1/WDR1 supports mi-
totic cell rounding. Biochem. Biophys. Res. Commun. 327:268–275.
Gomez, T.S., K. Kumar, R.B. Medeiros, Y. Shimizu, P.J. Leibson, and D.D.
Billadeau. 2007. Formins regulate the actin-related protein 2/3 complex-
independent polarization of the centrosome to the immunological syn-
apse. Immunity. 26:177–190. doi:10.1016/j.immuni.2007.01.008
Kunda, P., A.E. Pelling, T. Liu, and B. Baum. 2008. Moesin controls cortical
rigidity, cell rounding, and spindle morphogenesis during mitosis. Curr.
Biol. 18:91–101. doi:10.1016/j.cub.2007.12.051
pcDNA-EGFP at BamHI and EcoRI sites as described previously (Ballestrem
et al., 1998). Arp3 was subcloned from pEGFP-Arp3 (Morita et al., 2007),
which was a gift from K. Sobue (Osaka University, Suita, Osaka, Japan),
into pcDL-SR-myc or pcDNA3-HA. Silent mutations in the siRNA-targeting
region of Arp3 were introduced by PCR-based mutagenesis using
the following primers: forward, 5-AGGAGTCAGCAAAAGTCGGGGAC-
CAAGCTCAAAGGAGGG-3; and reverse, 5-CCCTCCTTTGACCTTG-
The siRNA for human Arp3 and spire, mDia1, and mDia2 were designed
as described previously (Unsworth et al., 2004; Beli et al., 2008; Morel
et al., 2009). As a control, the siRNA for lusiferase (5-CUUACGCUGAGUA-
CUUCGATT-3) was used. HeLa cells were transfected with annealed siRNA
using Lipofectamine 2000 (Invitrogen), incubated overnight, washed with
fresh medium, and synchronized by a double-thymidine block. The expres-
sion levels of Arp3 and mDia1 were analyzed by immunoblotting, and
spire and mDia2 expression levels were analyzed by RT-PCR using the
following primer sets; spire forward, 5-CCAAGAGCGGCAGTACAA-
CCC-3; spire reverse, 5-GAGGCTCCGTGGCCGTGGTGG-3; mDia2
forward, 5-GAAAGAATTATGAGTGAGGAG-3; and mDia2 reverse,
Cell staining, image analysis, and time-lapse observations
Synchronized HeLa cells in M phase were fixed in methanol at 20°C for
5 min or with 3.7% formaldehyde at room temperature for 10 min. They
were then permeabilized with 0.2% Triton X-100/PBS at room temperature
for 10 min, washed three times with PBS, blocked with 3% BSA/PBS, and
immunostained with each antibody. For Arp3 and Arp1 staining, cells
were preextracted in 0.5% Triton X-100 in PHEM buffer with 5 µM taxol for
1 min, and fixed in methanol at 20°C for 5 min. Retraction fibers were
stained with Alexa Fluor 488–conjugated Phalloidin as described previously
(Mitchison, 1992). Images were acquired using the DeltaVision optical sec-
tioning system (Applied Precision) equipped with an inverted microscope
(IX71; Olympus), a 60× 1.4 NA oil objective lens (Olympus), a charge-
coupled device camera (CoolSNAP HQ Monochrome; Photometrics), and
SoftWoRx software (Applied Precision). Image deconvolution was performed
using SoftWoRx software. For live imaging, we used optical sectioning sys-
tems (DeltaVision) with a temperature-controlled and motorized stage. HeLa
cells were transfected with GFP-UtrCH and DsRed-histone H1, which enabled
us to monitor actin filaments and chromosome states, respectively. During the
acquisition of the time-lapse images, cells were grown in the DME with
20 mM Hepes, pH 7.3, in a glass-bottomed dish (Iwaki) coated with human
fibronectin (Sigma-Aldrich). Cell culture in 3D gel was performed as de-
scribed previously (Toyoshima and Nishida, 2007). In brief, HeLa cells with
GFP-UtrCH and DsRed-histone H1, synchronized by a double-thymidine
block, were trypsinized immediately after the release from a double-thymidine
block. 10,000 cells were suspended in 100 ml of Matrigel Basement Mem-
brane Matrix (BD) and subjected to the time-lapse observations.
We used a confocal laser microscope (FV1000-D; Olympus). HeLa cells
were transfected with GFP-actin and DsRed-histone H1. After a double-
thymidine block, M phase synchronized HeLa cells were observed with a
60× oil-immersion objective lens (UPLSAPO 60× O, NA 1.35) at wave-
lengths of 488 nm (GFP) and 559 nm (DsRed). Photobleaching was per-
formed with a high-powered laser with a wavelength of 488 nm.
Online supplemental material
Fig. S1 shows the revolving movement of the amorphous actin cluster in
HeLa cells. Fig. S2 shows the effects of the Cdk1 inhibitor Ro3306 on an
M phase event. Fig. S3 shows that actin polymerization and depolymer-
ization reactions are required for generating the mitotic actin cluster and
movement. Video 1 shows the revolving movement of the amorphous actin
cluster in a HeLa cell expressing GFP-UtrCH during M phase. Video 2 shows
GFP-UtrCH and DsRed-histone H1. Top, control (Luc siRNA); middle, Arp3 siRNA1; bottom, Arp3 siRNA1 + myc-Arp3res. Arrowheads indicate the actin
cluster. (D) Percentages of cells showing the revolving movement of the actin cluster. Control (Luc siRNA), 100% (n = 22); Arp3 siRNA1, 26.3% (n = 19);
Arp3 siRNA1 + myc-Arp3res, 73.3% (n = 15); Arp3 siRNA1 + HA-Arp3res, 70.6% (n = 17). (E) Staining with phalloidin of HeLa cells transfected with
luciferase siRNA (left), Arp3 siRNA1 (middle), and Arp3 siRNA2 (right). The bottom sections show the retraction fibers, and the middle sections show the
cortical actin structures and the actin cluster, which is indicated by a white circle with a yellow arrowhead. (F) Double staining with anti-Arp3 (green) and
anti-Arp1 (red) antibodies of HeLa cells transfected with luciferase siRNA (left) and Arp3 siRNA1 (right). Arrowheads, the patchy staining of endogenous
Arp3. (G) Staining with anti-Arp3 antibody (red) of HeLa cells expressing GFP-UtrCH. Arrowheads, the actin cluster. Bars, 10 µm.
JCB • VOLUME 191 • NUMBER 3 • 2010 462
Lénárt, P., C.P. Bacher, N. Daigle, A.R. Hand, R. Eils, M. Terasaki, and J.
Ellenberg. 2005. A contractile nuclear actin network drives chromosome
congression in oocytes. Nature. 436:812–818. doi:10.1038/nature03810
Li, H., F. Guo, B. Rubinstein, and R. Li. 2008. Actin-driven chromosomal motility
leads to symmetry breaking in mammalian meiotic oocytes. Nat. Cell
Biol. 10:1301–1308. doi:10.1038/ncb1788
Maddox, A.S., and K. Burridge. 2003. RhoA is required for cortical retraction
and rigidity during mitotic cell rounding. J. Cell Biol. 160:255–265.
Miller, A.L., and W.M. Bement. 2009. Regulation of cytokinesis by Rho GTPase
flux. Nat. Cell Biol. 11:71–77. doi:10.1038/ncb1814
Mitchison, T.J. 1992. Actin based motility on retraction fibers in mitotic PtK2
cells. Cell Motil. Cytoskeleton. 22:135–151. doi:10.1002/cm.970220207
Morel, E., R.G. Parton, and J. Gruenberg. 2009. Annexin A2-dependent poly-
merization of actin mediates endosome biogenesis. Dev. Cell. 16:445–
Morita, T., T. Mayanagi, T. Yoshio, and K. Sobue. 2007. Changes in the balance
between caldesmon regulated by p21-activated kinases and the Arp2/3
complex govern podosome formation. J. Biol. Chem. 282:8454–8463.
Moulding, D.A., M.P. Blundell, D.G. Spiller, M.R. White, G.O. Cory, Y.
Calle, H. Kempski, J. Sinclair, P.J. Ancliff, C. Kinnon, et al. 2007.
Unregulated actin polymerization by WASp causes defects of mitosis
and cytokinesis in X-linked neutropenia. J. Exp. Med. 204:2213–2224.
Murthy, K., and P. Wadsworth. 2008. Dual role for microtubules in regulating
cortical contractility during cytokinesis. J. Cell Sci. 121:2350–2359.
Pollard, T.D. 2007. Regulation of actin filament assembly by Arp2/3 complex and
formins. Annu. Rev. Biophys. Biomol. Struct. 36:451–477. doi:10.1146/
Rosenblatt, J., L.P. Cramer, B. Baum, and K.M. McGee. 2004. Myosin II-
dependent cortical movement is required for centrosome separation
and positioning during mitotic spindle assembly. Cell. 117:361–372.
Steffen, A., J. Faix, G.P. Resch, J. Linkner, J. Wehland, J.V. Small, K. Rottner,
and T.E. Stradal. 2006. Filopodia formation in the absence of functional
WAVE- and Arp2/3-complexes. Mol. Biol. Cell. 17:2581–2591. doi:10
Théry, M., V. Racine, A. Pépin, M. Piel, Y. Chen, J.B. Sibarita, and M. Bornens.
2005. The extracellular matrix guides the orientation of the cell division
axis. Nat. Cell Biol. 7:947–953. doi:10.1038/ncb1307
Toyoshima, F., and E. Nishida. 2007. Integrin-mediated adhesion orients the
spindle parallel to the substratum in an EB1- and myosin X-dependent
manner. EMBO J. 26:1487–1498. doi:10.1038/sj.emboj.7601599
Unsworth, K.E., M. Way, M. McNiven, L. Machesky, and D.W. Holden. 2004.
Analysis of the mechanisms of Salmonella-induced actin assembly dur-
ing invasion of host cells and intracellular replication. Cell. Microbiol.
Vassilev, L.T., C. Tovar, S. Chen, D. Knezevic, X. Zhao, H. Sun, D.C. Heimbrook,
and L. Chen. 2006. Selective small-molecule inhibitor reveals critical
mitotic functions of human CDK1. Proc. Natl. Acad. Sci. USA. 103:
Winter, D., A.V. Podtelejnikov, M. Mann, and R. Li. 1997. The complex con-
taining actin-related proteins Arp2 and Arp3 is required for the
motility and integrity of yeast actin patches. Curr. Biol. 7:519–529.
Withee, J., B. Galligan, N. Hawkins, and G. Garriga. 2004. Caenorhabditis
elegans WASP and Ena/VASP proteins play compensatory roles in
morphogenesis and neuronal cell migration. Genetics. 167:1165–1176.
Woolner, S., L.L. O’Brien, C. Wiese, and W.M. Bement. 2008. Myosin-10 and
actin filaments are essential for mitotic spindle function. J. Cell Biol.