MBoC | ARTICLE

Capping protein is essential for cell migration in

vivo and for filopodial morphology and dynamics
Shamim A. Sinnara, Susumu Antokub,*, Jean-Michel Saffina, Jon A. Cooperb, and Shelley Halpaina
a
Division of Biological Sciences, University of California, San Diego, and Sanford Consortium for Regenerative
Medicine, La Jolla, CA 92037; bFred Hutchinson Cancer Research Center, Seattle, WA 98109

ABSTRACT Capping protein (CP) binds to barbed ends of growing actin filaments and inhib- Monitoring Editor
its elongation. CP is essential for actin-based motility in cell-free systems and in Dictyostelium. Paul Forscher
Yale University
Even though CP is believed to be critical for creating the lamellipodial actin structure neces-
sary for protrusion and migration, CP’s role in mammalian cell migration has not been di- Received: Dec 19, 2013
rectly tested. Moreover, recent studies have suggested that structures besides lamellipodia, Revised: Apr 1, 2014
including lamella and filopodia, may have unappreciated roles in cell migration. CP has been Accepted: May 8, 2014
postulated to be absent from filopodia, and thus its role in filopodial activity has remained
unexplored. We report that silencing CP in both cultured mammalian B16F10 cells and in
neurons of developing neocortex impaired cell migration. Moreover, we unexpectedly ob-
served that low levels of CP were detectable in the majority of filopodia. CP depletion de-
creased filopodial length, altered filopodial shape, and reduced filopodial dynamics. Our re-
sults support an expansion of the potential roles that CP plays in cell motility by implicating
CP in filopodia as well as in lamellipodia, both of which are important for locomotion in many
types of migrating cells.

INTRODUCTION
Heterodimeric capping protein (CP) binds to actin filament barbed effectively push the membrane forward (Pollard and Borisy, 2003;
ends and thus inhibits further polymerization (Wear et al., 2003). In- Le Clainche and Carlier, 2008).
hibiting CP activity in live organisms and cells severely compromises In addition, capping of actin filaments maintains actin monomer
actin-based motility and lamellipodia formation (Hug et al., 1995; supply by preventing elongation of filaments that do not contribute
Rogers et al., 2003; Mejillano et al., 2004; Iwasa and Mullins, 2007). to leading-edge motility (Carlier and Pantaloni, 1997; Hu and
CP is thus thought to have a pivotal role in enabling leading-edge Papoian, 2010). These actin monomers may then increase overall
protrusion and net migration of cells (Pollard and Borisy, 2003; Le cell motility by either increasing actin filament elongation (Carlier
Clainche and Carlier, 2008). Specifically, CP activity is thought to and Pantaloni, 1997) or by increasing nucleation by the Arp2/3 com-
limit the growth of lamellipodial actin filaments nucleated by the plex (Akin and Mullins, 2008) in regions with a relatively lower con-
Arp2/3 complex, thus creating short and stiff filaments that can centration of CP. However, the requirement for CP in mammalian
cell migration has not been directly tested.
Recent work, moreover, has questioned the necessity of lamelli-
This article was published online ahead of print in MBoC in Press (http://www podia for cell migration (Gupton et al., 2005; Suraneni et al., 2012;
.molbiolcell.org/cgi/doi/10.1091/mbc.E13-12-0749) on May 14, 2014. Wu et al., 2012; Edwards et al., 2013). Fibroblast cells in which
*Present address: Department of Pathology and Cell Biology, Columbia University
Medical Center, 630 West 168th Street, New York, NY 10032. lamellipodium formation was inhibited following tropomyosin injec-
Address correspondence to: Shelley Halpain ([email protected]). tion moved faster than did control cells (Gupton et al., 2005). In
Abbreviations used: CP, capping protein; CP-ir, CP immunoreactivity; DIC, differ- addition, studies from two different laboratories have demonstrated
ential interference contrast; eGFP, enhanced green fluorescent protein; PBS, that cells lacking normal lamellipodia secondary to depletion of the
phosphate-buffered saline; RFP, red fluorescent protein; shRNA, short hairpin
RNA. Arp2/3 complex can migrate using a filopodia-based protrusion sys-
© 2014 Sinnar et al. This article is distributed by The American Society for Cell tem (Suraneni et al., 2012; Wu et al., 2012). Finally, the migration
Biology under license from the author(s). Two months after publication it is avail- defect in cells with silenced CARMIL protein could be substantially
able to the public under an Attribution–Noncommercial–Share Alike 3.0 Unported
Creative Commons License (http://creativecommons.org/licenses/by-nc-sa/3.0).
rescued without rescuing the concurrent lamellipodial defect, lead-
“ASCB®,” “The American Society for Cell Biology®,” and “Molecular Biology of ing the authors to suggest that lamellipodia are not essential for cell
the Cell®” are registered trademarks of The American Society of Cell Biology. migration in this setting (Edwards et al., 2013).

2152 | S. A. Sinnar et al. Molecular Biology of the Cell

FIGURE 1: shRNA-mediated silencing of CP induces morphological changes. (A) CP-ir is robustly decreased in silenced
B16F10 cells. Arrowhead indicates CP-silenced cell adjacent to a nontransfected cell. (B) Quantification of CP-ir
decrease in silenced cells. For each category, the average CP fluorescence intensity of transfected cells was normalized
to that of untransfected cells on the same coverslip that had been imaged and processed in an identical manner. EV,
empty vector; CTL, Scramble-transfected cells; KD, CP-silenced cells. A total of 20–40 cells were analyzed for each
condition; ***p < 0.001. (C) Filopodial density is increased in CP-depleted cells. A total of 47–69 cells per condition were
analyzed across three different experiments and include 1600–3200 filopodia/condition; ***p < 0.001. Controls include
eGFP-transfected (eGFP) and Scramble-transfected (CTL) cells. (D) A representative Scramble-transfected cell (left) and
a CP-depleted cell (right). Cells are stained with phalloidin to visualize F-actin. Scale bar: 10 μm.

Filopodia are thin, actin-based protrusions found in many construct to CP based on a previously published shRNA sequence
cell types, including migrating cells and neurons (Mattila and (Mejillano et al., 2004) and created a scrambled construct to use as
Lappalainen, 2008). They are integral for both chemosensing and control. In addition, we determined that a polyclonal antibody from
motility (Gupton and Gertler, 2007; Mattila and Lappalainen, 2008) Millipore (AB6017) recognizing the β2 subunit of CP is both sensitive
and are usually composed of 15–30 actin filaments (Small, 1981; and specific for immunostaining of multiple cell lines and primary
Lewis and Bridgman, 1992). Models of filopodial initiation posit that neuronal cultures. CP immunoreactivity (CP-ir) using this antibody
a relative absence of CP allows filopodia to initiate and grow increased dramatically with CP overexpression (Figure S1A) and de-
(Svitkina et al., 2003; Faix and Rottner, 2006). CP has been reported creased with CP silencing in B16F10 cells (see Figure 1A). In addi-
to be undetectable in filopodia (Svitkina et al., 2003), and most cur- tion, the signal decreased considerably with preadsorption by holo-
rent models for filopodial structure and function do not include a protein (Figure S1, B and C).
role for CP (Wear and Cooper, 2004; Gupton and Gertler, 2007; Transfection of B16F10 cells with CP shRNA effectively sup-
Mattila and Lappalainen, 2008; Faix et al., 2009). Interestingly, how- pressed CP activity. In a representative experiment, shRNA treat-
ever, some investigators have suggested that CP within filopodia ment for 5 d reduced CP-ir by an average of 80% in transfected cells
may have an important role in controlling filopodial dynamics compared with untransfected neighboring cells (Figure 1, A and B).
(Mallavarapu and Mitchison, 1999; Zhuravlev and Papoian, 2009). As reported previously (Mejillano et al., 2004), CP silencing led to a
In this study, we investigated the requirement of CP for mam- robust, more than twofold increase in filopodial density (Figure 1, C
malian cell migration in vitro and in vivo using a knockdown ap- and D).
proach. Moreover, we discovered the presence of CP in filopodia To test whether CP is required for mammalian cell migration, we
and explored its functional impact. We suggest that while a de- used time-lapse imaging to measure rates of cell migration in con-
crease in capping activity stimulates emergence of filopodia, CP trol and CP-depleted cells. The average cell migration rate of
activity within filopodia is nonetheless necessary for their proper Scramble-transfected cells was 0.68 μm/min (Figure 2A), a value
form and function. similar to those reported previously for a related cell type, B16F1
cells (Hotulainen et al., 2005; Yang et al., 2007). Silencing of CP re-
RESULTS duced rates of cell migration by more than half, from 0.68 μm/min
CP depletion impairs mammalian cell migration in vivo to 0.31 μm/min (Figure 2A). It also reduced the net area traversed by
To understand the role of CP in cell migration, we first used a silenc- cells over 6 h by 67% (Figure 2B). Representative tracks of Scramble-
ing strategy in B16F10 cells. We created a short hairpin RNA (shRNA) transfected and CP-depleted cells are shown in Figure 2C.

Volume 25 July 15, 2014 Capping protein in filopodia | 2153

 shown to be associated with sites of actin
assembly in the lamella (Schafer et al., 1998),
and we detected CP-ir in the lamella as well
(Figure 3B, arrow).
Interestingly, although CP’s role has been
studied most extensively at the leading
edge, the majority of CP-ir was not at the
leading edge but instead in the cell body
(Figures 1A and 3A); this distribution per-
sisted even after extracting soluble CP from
live cells with 1% Triton before fixation
(unpublished data). This pattern of immuno-
reactivity, in which the majority of the CP
signal is in the cell body, is similar to that ap-
pearing in published images of endogenous
CP in various cell types (Schafer et al., 1998;
Rogers et al., 2003). CP has been shown to
be an integral component of both the dy-
nactin complex (Schafer et al., 1994) and the
WASH complex (which associates with recy-
cling endosomes; Derivery et al., 2009).
Moreover, CP may have microtubule-associ-
ated functions (Davis et al., 2009; Bartolini
et al., 2012). Therefore CP may fulfill other
cellular roles besides leading-edge motility.
Unexpectedly, we also clearly detected
low levels of CP-ir in a punctate distribution
along filopodia of B16F10 cells (Figure 3, C,
D, and F, CTL). This was surprising because
CP has been postulated to be absent from
FIGURE 2: Depletion of CP impairs cell migration in vitro and in vivo. (A) Average migration filopodia (Svitkina et al., 2003; Faix and
speeds of Scramble-transfected (CTL) and CP-depleted (KD) B16F10 cells. Cells were visualized Rottner, 2006), and indeed, the relative ab-
by DIC imaging and tracked by their nuclei every 5 min for 6 h. (B) The average net area sence of CP is thought to be important for
traversed by Scramble-transfected and CP-depleted cells in 6 h. Net area traversed was filopodia to emerge from the dendritic actin
calculated for each cell by multiplying the greatest absolute difference between x-coordinates network (Svitkina et al., 2003; Wear and
by the greatest absolute difference between y-coordinates. For (A) and (B), 34–36 cells were
Cooper, 2004).
analyzed per condition; ***p < 0.001. (C) Representative migration tracks from Scramble-
We quantified CP-ir puncta along the
transfected and CP-depleted cells. (D) Neurons cotransfected with mCherry and CP shRNA
(right) have impaired migration to the cortical plate compared with neurons cotransfected with protruding portion of filopodia (since
mCherry and firefly luciferase shRNA (control; left). CP, cortical plate; SVZ/IMZ, subventricular puncta localization within the embedded
zone/ intermediate zone; VZ, ventricular zone. Scale bar: 100 μm. (E) Quantification of cortical portion is ambiguous). Out of 311 filopodia
migration. For each condition, fluorescence intensity from each region was normalized to total examined from B16F10 cells, 79% had at
fluorescence from all regions. Five animals from each category were analyzed; *p < 0.05. VZ+S/I, least one CP punctum, and 52% had at least
ventricular zone + subventricular/intermediate zone; CP, cortical plate. two puncta. On average, there were 2.4 CP
puncta detected per filopodia, correspond-
We next investigated the effects of CP depletion on the migra- ing to ∼1 punctum detected per micrometer length of filopodium.
tion of murine cortical neurons in vivo. After birth, cortical neurons The majority of immunopositive filopodia (91%) had CP puncta
migrate from the ventricular zone to the cortical plate (Bielas et al., along their length, but a large fraction of immunopositive filopodia
2004). Ninety-six hours after in utero electroporation, nearly half of (36%) also had CP puncta detectable at the presumptive tip.
control neurons, transfected with firefly luciferase shRNA, had To confirm whether filopodial CP-ir was a general phenomenon,
reached the cortical plate, but less than a third of CP-depleted neu- we examined another cell line (U2-OS osteosarcoma cells; Figure
rons had done so (Figure 2, D and E). Together these results demon- 3E), as well as growth cones in primary hippocampal neuronal cul-
strate that CP is indeed critical for migration of mammalian cells in tures (Figure S2). CP-ir was clearly detected in the filopodia of both
vitro and, more importantly, in vivo. Our results are consistent with a these cell types. Moreover, knockdown of CP by shRNA significantly
prior study in Dictyostelium (Hug et al., 1995) in which mutant cells decreased CP-ir within filopodia (Figure 3F) and decreased the frac-
underexpressing CP moved 35% slower than did control cells. tion of CP-positive filopodia (Figure 3G).

CP is detected in filopodia of multiple cell types CP depletion reduces filopodial length and changes
As expected, we clearly detected CP at the leading edge of migrat- filopodial shape
ing B16F10 mouse melanoma cells (Figure 3B, arrowhead). The Remarkably, filopodial length was significantly decreased by a
punctate pattern of CP-ir we observed is similar to that observed third in CP-depleted cells (Figures 1D and 4A). For measurement
previously by other investigators (Schafer et al., 1998; Rogers et al., of filopodial length, the entire actin bundle was quantified regard-
2003; Applewhite et al., 2007; Yang et al., 2007). CP has also been less of its position relative to the lamellipodium, since filopodia,

2154 | S. A. Sinnar et al. Molecular Biology of the Cell

 retraction fibers, and microspikes are dy-
namically interchangeable (Svitkina et al.,
2003). A frequency histogram comparing
shRNA-transfected and Scramble-trans-
fected cells shows that CP depletion results
in selective loss of longer filopodia and
therefore a higher proportion of shorter
filopodia (Figure 4B). These novel findings
imply that CP depletion directly or indi-
rectly reduces filopodial length.
Besides dramatically reducing filopodial
length, other effects of CP depletion on
filopodial morphology were apparent. First,
nearly the entire length of individual filopo-
dia in CP-depleted cells appeared to be
protruding beyond the cell margin (Figure 4,
C and D; see Materials and Methods for de-
tails on measurements). In contrast, filopo-
dia from Scramble-transfected cells often
had much of their length embedded within
the cell lamellipodium (Figure 4, C and D).
Second, the apparent shapes of filopodia
from CP-depleted cells, based on phalloidin
staining, were visibly altered (Figure 4, E and
F). More than 50% of the filopodia from
Scramble-transfected cells had a cone-like
or tapered appearance, with a smaller per-
centage having a more rod-like or uniform
appearance (Figure 4, E and F). However,
the majority of filopodia from shRNA-trans-
fected cells had a rod-like appearance
(Figure 4, E and F). In addition, a significant
fraction of filopodia in CP knockdown cells
had a “cattail” appearance, in which the
base was visibly thinner than the shaft and
tip regions (Figure 4E). This type of filopo-
dium was rarely seen in Scramble-transfected
cells. Of note, a similar filopodial morphol-
ogy (“club-like filopodia”) was described
with formin overexpression (a manipulation
expected to decrease relative capping activ-
ity; Yang et al., 2007). Collectively our results
demonstrate that CP activity is essential for
normal length and morphology of filopodia.
These effects of CP depletion on filopodial
length, shape, and F-actin concentration
(see next section) were confirmed by using
shRNA directed against a second sequence
of the CP β subunit (Figure S3; see Materials
and Methods for sequence information).

in filopodia from CP-silenced cells (KD)

compared with that in filopodia from
FIGURE 3: Endogenous CP is detected in filopodia of mammalian cells and filopodial CP-ir Scramble-transfected cells (CTL). Arrowheads
decreases with shRNA treatment. (A) B16F10 melanoma cell stained using antibody to indicate some of the CP-ir puncta. (G) The
endogenous CP. Note substantial CP signal in cell body. (B) A magnified view of the boxed proportion of CP-positive filopodia is reduced
region in (A). CP is present at the leading edge (arrowhead) and in the lamella (arrow). in CP-depleted cells compared with that in
(C) Enlarging and increasing the contrast of the boxed region in (B) shows that CP-ir is detected Scramble-transfected cells. For each
in filopodia (C1). Further magnification of the boxed region in (C1) shows that CP-ir is detected condition, a total of 250–445 filopodia from
both along the length (arrowhead) and at the presumptive tip (arrow) of filopodia (C2–C4). 13 to 20 cells across two independent
(D) An example of CP-ir in filopodia of a different B16F10 cell. (E) CP-ir is detected along experiments were analyzed; ***p < 0.001.
filopodia of U2-OS cells. Arrowheads in (D) and (E) indicate CP-ir. (F) The CP signal is decreased Scale bars: A, B, and F, 10 μm; C–E, 2 μm.

Volume 25 July 15, 2014 Capping protein in filopodia | 2155

 it was clear that the phalloidin staining of individual filopodia was
also dramatically increased in CP-depleted cells. Quantification of
phalloidin intensities demonstrated a greater than twofold increase
in F-actin concentration globally in CP-depleted cells (Figure 5C)
and within individual filopodia of CP-depleted cells (Figure 5D).
These data suggest that filopodia from CP-depleted cells have a
greater number of F-actin filaments than do those from Scramble-
transfected cells. In other words, decreased capping leads to greater
actin polymerization. These results are consistent with those in Dic-
tyostelium (Hug et al., 1995), in which cells underexpressing CP had
an increase in F-actin. CP silencing has also been shown to increase
total actin levels in both Dictyostelium (Hug et al., 1995) and mam-
malian cells (Edwards et al., 2013).

CP depletion reduces filopodial dynamics

To understand whether CP depletion has consequences for filopo-
dial function, we examined filopodial behavior using time-lapse
imaging. The average protrusion rate for Scramble-transfected
filopodia was 95 nm/s, a value similar to that reported previously
for B16F1 cells (Yang et al., 2007; Figure 6A and Supplemental
Video S1). However, filopodia from CP-depleted cells had signifi-
cantly reduced protrusion rates, on average only 61 nm/s (Figure
6A and Video S2). Rates of filopodial shortening were similar in
control cells and CP-depleted cells (Figure 6A). In addition, filopo-
dia from shRNA-transfected cells spent significantly less time
growing and more time pausing (Figure 6B and Videos S1 and S2).
Thus we conclude that CP depletion reduces filopodial dynamics
and could therefore have functional consequences for cell-based
activities that require proper filopodial function.

DISCUSSION
CP is an essential player in creating the actin architecture of lamelli-
podia, which in turn has been thought necessary for cell protrusion
and migration (Pollard and Borisy, 2003; Le Clainche and Carlier,
2008). CP is necessary for actin-based motility in vitro (Loisel et al.,
1999) and for migration in vivo in Dictyostelium (Hug et al., 1995).
Moreover, mammalian cells depleted of CP have a decreased lamel-
FIGURE 4: CP depletion reduces filopodial length and changes lipodial protrusion rate (Mejillano et al., 2004). However, recent work
filopodial morphology. (A) Quantification of filopodial length in has questioned the need for normal lamellipodial dynamics in actual
Scramble-transfected (CTL) and CP-knockdown (KD) cells. Length of cell migration (Gupton et al., 2005; Suraneni et al., 2012; Wu et al.,
the entire actin bundle was measured (as indicated in (D) by the yellow 2012; Edwards et al., 2013). In addition, the requirement for CP in
bracket), as described in Svitkina et al. (2003), and averaged per cell. A mammalian cell migration has not been established.
total of 47–69 cells per condition were analyzed across three different We report here that depleting CP impairs mammalian cell migra-
experiments and include 1600–3200 filopodia/condition; ***p < 0.001.
tion in vitro and in vivo. Therefore CP is necessary for proper loco-
(B) Frequency histogram comparing the lengths of Scramble-
motion. CP’s role in migration could result from its proper function in
transfected and shRNA-transfected cells. (C) Knockdown of CP
increases the fraction of filopodial length that protrudes beyond the the lamellipodia or in other cell compartments. For example, some
cell margin. Filopodial values were averaged per cell, and 600–1150 investigators have proposed that the lamella, the region behind the
filopodia from 30 to 35 cells across three independent experiments lamellipodium, drives net cell displacement (Ponti et al., 2004; Gup-
were analyzed; ***p < 0.001. (D) A representative Scramble-transfected ton et al., 2005). CP is present in the lamella (see Figure 3, A and B)
(left) and CP-knockdown (right) cell. Note that a greater portion of and has been shown to associate with sites of actin assembly in this
each individual filopodium is embedded within the lamellipodium in the region (Schafer et al., 1998). Thus CP function in the lamella may be
Scramble-transfected cell. (E) Categories of filopodial shapes found in vital for cell motility.
Scramble-transfected and CP-depleted cells. See text for category In addition, cells without normal lamellipodia are postulated
descriptions. (F) CP knockdown alters the apparent shape of filopodia.
to use a filopodia-based system to locomote (Suraneni et al., 2012;
A total of 325–355 filopodia from two independent experiments were
Wu et al., 2012). Though CP has been posited to be absent from
analyzed for each group; ***p < 0.001. Scale bar: 2 μm.
filopodia (Svitkina et al., 2003; Mejillano et al., 2004; Faix and
CP depletion increases cellular and filopodial F-actin Rottner, 2006), we find that CP is detectable in filopodia from
concentration multiple mammalian cell types and is required for normal filopodial
Strikingly, knockdown of CP caused a significant increase in F-actin form and dynamics (see following discussion). Filopodia have im-
concentration inside cells, as measured by phalloidin staining (Figure portant sensing and adhesion roles in cell migration (Mattila and
5A). This increased staining was especially evident at cell margins at Lappalainen, 2008; Arjonen et al., 2011), and it is possible that CP’s
low magnification. At higher magnification (Figure 5, A, inset, and B), role in filopodia is also necessary for proper cell migration.

2156 | S. A. Sinnar et al. Molecular Biology of the Cell

FIGURE 5: F-actin concentration is increased in CP-depleted cells. (A) F-actin concentration as determined by phalloidin
staining is visibly increased in a CP-knockdown cell (left cell in image) compared with an untransfected cell (right cell,
indicated by arrowhead). Inset is displayed with lower contrast to illustrate that much of the phalloidin intensity at the
cell margin of the CP-depleted cell is associated with short filopodia. (B) Enlarged image of filopodia in a different
Scramble-transfected cell (CTL, left) and CP-knockdown cell (KD, right). These two phalloidin-stained cells were imaged
under identical conditions and are displayed with identical brightness and contrast settings. (C) Mean phalloidin
intensity is increased globally in CP-knockdown cells. A total of 50–80 cells from two independent experiments were
analyzed; ***p < 0.001. Phalloidin intensities were normalized to the mean intensity of Scramble-transfected cells.
(D) Mean phalloidin intensity is increased in individual filopodia from CP-knockdown cells. A total of 180–340 filopodia
from the same two independent experiments as in (C) were analyzed; ***p < 0.001. Phalloidin intensities were
normalized to the mean intensity of filopodia from Scramble-transfected cells. Scale bars: A, 20 μm; B, 5 μm.

Our findings regarding the immunolocalization of CP within filaments may persist and thus become easily detected. Indeed, in
filopodia and the effects of CP silencing yield new insights into time-lapse imaging of knockdown cells, we sometimes observed
filopodial structure. Given that CP is known to bind only to barbed uniform filopodia morphing into cattail filopodia, indicating a
ends of actin filaments (Wear et al., 2003), the punctate distribution temporal relationship between uniform and cattail filopodia
of CP along the length of filopodia (as opposed to CP localizing (unpublished data). Our results suggest that CP within filopodia may
exclusively to filopodial tips) suggests that actin filaments within directly influence filopodial morphology.
filopodia are not necessarily of uniform length (Figure 7). Instead, it
suggests that individual filaments may grow to different lengths be-
fore being stochastically capped by CP. Filopodia in which relatively
few filaments extend the entire length from base to tip have a ta-
pered appearance, whereas filopodia that contain actin filaments of
equal lengths appear to be uniform. In control cells, most filopodia
are tapered, with a smaller but significant fraction having a uniform
morphology. In cells depleted of CP, decreased capping of individ-
ual filaments within filopodia would allow all of the filaments to grow
uniformly. Thus there is a dramatic increase in the proportion of
uniform filopodia, offset by a concomitant decrease in the propor-
tion of tapered filopodia.
Reducing CP also induces the emergence of filopodia that have
a thicker tip and a thinner base, which we have named “cattail”
filopodia. One explanation for this morphology is that some filopo-
FIGURE 6: CP depletion reduces filopodial dynamics.
dial actin filaments may normally sever or depolymerize from their
(A) Quantification of filopodial protrusion and retraction rates. CTL,
pointed ends and begin to treadmill (Figure 7). In control filopodia, Scramble-transfected cells; KD, CP-knockdown cells. (B) Filopodia
CP caps these treadmilling filaments, and retrograde flow moves from CP-depleted cells spend more time pausing. For each condition
them out of filopodia; hence, they are transient structures and not in (A) and (B), 51–56 filopodia from 13 to 14 time-lapse sequences
frequently detected. Fewer than 1% of control filopodia fell into this were analyzed; (A) ***p < 0.001; (B) **p < 0.01. See Materials and
category (Figure 4F). In CP-depleted filopodia, however, these Methods for details on quantification.

Volume 25 July 15, 2014 Capping protein in filopodia | 2157

 diffusion from the base to the growing filopodial tip (Mogilner
and Rubinstein, 2005). Moreover, CP is thought to be essential for
maintaining an adequate supply of actin monomers for actin-based
motility (Carlier and Pantaloni, 1997; Loisel et al., 1999; Pollard and
Borisy, 2003). Thus we suggest that inhibiting CP activity indirectly
reduces filopodial length due to depletion of actin monomer.
Why does a greater part of individual filopodia in CP-depleted
cells extend beyond the cell margin (Figure 4, C and D)? A possible
explanation takes into account relative rates of retrograde flow,
which exists in filopodia as well as in lamellipodia (Mallavarapu and
Mitchison, 1999). A reduced rate of F-actin polymerization at filopo-
dial tips in CP-depleted cells (as evidenced by a decreased protru-
sion rate) would be expected to reduce retrograde flow within
filopodia. Thus a greater fraction of the filopodial length protrudes
beyond the lamellipodia. Another explanation could be that,
because lamellipodial protrusion is disrupted with CP depletion
(Mejillano et al., 2004), the lamellipodium does not extend as far
with respect to the filopodia as it does in control cells.
Why are filopodia in CP-depleted cells less dynamic than those
in control cells? One possibility is that CP depletion indirectly re-
duces filopodial dynamics by limiting actin monomer supply. Limit-
ing levels of G-actin reduce polymerization at filopodial tips and
thus decrease protrusion rate. Consequently, filopodia spend less
time growing and more time pausing. Interestingly, theoretical work
has suggested that stochastic capping of individual actin filaments
is necessary for macroscopic filopodial dynamics (Zhuravlev and
Papoian, 2009), and this mechanism may also contribute to our ob-
servation of reduced filopodial dynamics in CP-depleted cells.
Consistent with other studies, we observe a dramatic increase in
filopodial density upon inhibition of capping activity (Figure 1). We
therefore agree with other investigators that a relative decrease in
capping activity seems necessary for filopodia to emerge (Svitkina
FIGURE 7: A model for how CP in filopodia influences filopodial et al., 2003; Mejillano et al., 2004). We suggest, however, that low
shape. (A) We propose that stochastic capping of individual actin levels of CP within filopodia are essential for normal filopodial mor-
filaments within filopodia results in heterogeneity of filament lengths phology and dynamic behavior. It is perhaps not so surprising that
and consequently a tapered appearance (left). In filopodia with CP is detectable in filopodia: other actin-capping proteins, such as
reduced capping activity, all filaments span the entire length of the gelsolin and Eps8, are already known to be important components
filopodia, giving the filopodium a uniform appearance (middle). In of axonal filopodia (Gallo, 2013).
CP-depleted cells, cattail filopodia are present, in which there is a Moreover, anti-capping proteins such as Ena/VASP (Lebrand
greater density of actin filaments at the tip of the filopodia than at the
et al., 2004; Mattila and Lappalainen, 2008) and the diaphanous
base (right). (B) A possible mechanism for the creation of cattail
filopodia. In cells depleted of CP, the majority of filopodia are of the
formins (Yang et al., 2007; Faix et al., 2009) are present in filopodia
uniform variety (left). Actin filaments that sever or begin to and are thought to have important roles in filopodial form and func-
depolymerize from their pointed ends (middle) are normally capped tion. These studies have shown that both excessive capping and
and cleared away by retrograde flow. However, decreased capping anti-capping activity perturb filopodial morphology and dynamics.
within filopodia from CP-depleted cells results in these filaments We therefore suggest that the anti-capping activity of Ena/VASP
treadmilling and persisting (right). and formins within filopodia may be balanced by the capping activ-
ity of capping protein to produce normal filopodia. In sum, our data
Why are filopodia in CP-depleted cells remarkably shorter than suggest that the filopodial functions of CP, in addition to other po-
those in control cells? This observation may seem counterintuitive if tential roles like microtubule regulation and membrane trafficking
one assumes that a decrease in capping allows actin filaments to (Davis et al., 2009; Derivery et al., 2009; Bartolini et al., 2012), may
polymerize for longer periods and attain greater lengths. However, contribute to the critical role of CP in cell migration in vivo.
unchecked polymerization may consume a limiting cellular factor,
such as monomeric G-actin. CP silencing induced a 2.5-fold increase MATERIALS AND METHODS
in filopodial density (Figure 1C). Moreover, individual filopodia from Reagents
CP-silenced cells probably contain an increased number of actin fila- Plasmids. We created an shRNA construct to CPβ based on the
ments, as evidenced by their twofold increase in phalloidin intensity T1 sequence described in Mejillano et al. (2004; CCTCAGCGAT­
(Figure 5). CTGATCGAC). We cloned this sequence into pSuper vector
The burst of actin polymerization evidenced by increased filopo- (a generous gift from T. Wittmann, University of California, San
dial density and increased F-actin content in individual filopodia con- Francisco) that contained the polymerase-III H1-RNA promoter
ceivably depletes cellular G-actin and limits filopodial length. Indeed, (Brummelkamp et al., 2002). We also created a Scrambled sequence
mathematical modeling has suggested that growth of filopodia (GCACCCGTCTTCAACAGGT) and cloned it into p-Super. We
composed of more than ∼30 filaments would be limited by G-actin additionally used an shRNA directed against a second sequence in

2158 | S. A. Sinnar et al. Molecular Biology of the Cell

the CPβ subunit, GCACGCTGAATGAGATCTA. Both the construct Olympus IX70 inverted microscope mounted with a Yokogawa CSU-
containing this target sequence (T2) and one containing a mismatch X1 Spinning disk using either a 60× or 20× oil objective or a 20× dry
control sequence (T2*), were generous gifts from T. Svitkina air objective. Maximum projection images were acquired for all
(University of Pennsylvania) and have been published previously analyses, except for those related to cell migration, for which single-
(Mejillano et al., 2004). Enhanced green fluorescent protein (eGFP)- plane images containing the cell nucleus were obtained. For live
tagged versions of the α2 and β2 subunits of CP as well as CP holo- imaging, samples were kept at 35°C in a humid chamber containing
protein were generous gifts from D. Schafer (University of Virginia). 5% CO2. For movies of filopodial dynamics, images were acquired
Red fluorescent protein (RFP)-LifeAct was a generous gift from every 5 s for 2 min. For cell migration experiments, cells were im-
R. Truant (McMaster University, Canada). aged using differential interference contrast (DIC) imaging every 5
min for 6 h. Cell tracking was accomplished by tracking the position
Antibodies. A polyclonal antibody against the CP β2 subunit was of the nucleus in individual consecutive time-lapse images using the
purchased from Millipore (AB6017; Billerica, MA) and was used at a Manual Tracking plug-in in ImageJ.
dilution of 1:200 in B16F10 and U2-OS cells and a dilution of 1:50 in For measurement of filopodial length, the entire actin bundle
primary hippocampal neurons. Secondary antibodies (Alexa Fluor was quantified regardless of its position relative to the lamellipo-
488, 568, or 647) and Alexa Fluor 568 or 647-phalloidin were dium, since filopodia, retraction fibers, and microspikes are dy-
purchased from Life Technologies (Carlsbad, CA) and used at a namically interchangeable (Svitkina et al., 2003). Measurements
dilution of 1:500. were based on the methods of Vignjevic et al. (2006). For the frac-
tional length protruding beyond the cell margin (Figure 4C), the
Cell culture length protruding beyond the cell edge was divided by the total
B16F10 cells were a generous gift from S. Dowdy (University of length. The segmented line function of ImageJ was used to manu-
California, San Diego), and U2-OS cells were a generous gift from ally measure all filopodia. Lamellipodial lengths of Scramble-trans-
G. Yeo (University of California, San Diego). Both cells lines were cul- fected cells were also manually measured using ImageJ. CP-de-
tured in DMEM supplemented with 10% fetal bovine serum. Primary pleted cells often had abnormal lamellipodia, and so for these
hippocampal neurons were cultured as described (Calabrese and cells, the perimeter of the cell region containing filopodia was
Halpain, 2005). measured.
B16F10 cells were transfected with Scramble or shRNA to CP for Filopodial morphologies were classified qualitatively from phal-
5 d using Lipofectamine 2000 (Life Technologies) in 24-multiwell loidin images as follows: “tapered,” base detectably wider than tip;
plates. For live-cell imaging experiments, cells were cotransfected “uniform,” similar width along entire length, with base and tip not
with RFP-LifeAct to visualize the actin cytoskeleton. After 4 d, cells detectably different; “cattail,” base detectably thinner than tip;
were trypsinized and replated onto glass coverslips that had been “other,” filopodial-like protrusions not readily classified. A second
sequentially precoated with poly-l-ornithine (0.1 mg/ml overnight; investigator blinded to the treatment analyzed a subset of the
Sigma-Aldrich, St. Louis, MO) and laminin (5 μg/ml for 3–4 h; Life images using these criteria and obtained strikingly similar results.
Technologies) in preparation for imaging. For analysis of filopodial dynamics (Figure 6), 2-min time-lapse
sequences of Scramble-transfected or shRNA-transfected cells
Immunostaining cotransfected with RFP-LifeAct were analyzed. Lengths of individual
For antibody staining, cells on glass coverslips were fixed in 3.7% filopodia were measured every 5 s until the filopodium disappeared
formaldehyde in phosphate-buffered saline (PBS) with 8% sucrose or merged with a neighboring filopodium. Total growth (or total
(wt/vol) for 20 min at 37°C. Cells were permeabilized in 0.2% Triton shrinkage) in micrometers was divided by total time spent growing
in PBS for 5 min, blocked with 2% bovine serum albumin for 45 min, (or shrinking) in seconds to obtain growth (or shrinkage) rate. The
incubated with primary antibody for 1 h at room temperature, fraction of total time spent growing, shrinking, or pausing was also
washed with PBS, and incubated with secondary antibody and la- calculated by considering a filopodium to be pausing if its length
beled phalloidin for 45 min at 37°C. After being washed, coverslips did not change by more than 2 pixels between frames.
were mounted with Aquamount (Thermo Fisher Scientific, Lafayette, For cortical migration experiments, sections of fetal brain were
CO), dried overnight, and imaged. imaged using a Nikon E800 scope and 10× objective. The fluores-
cence of mCherry-expressing neurons in each cortical region
In utero electroporation (ventricular zone, intermediate zone/subventricular zone, and cor-
In utero microinjection and electroporation was performed as previ- tical plate) was quantified using ImageJ and normalized to total
ously described (Simó et al., 2010), in accordance with procedures fluorescence in all regions.
approved by the Fred Hutchinson Cancer Research Center Institu- Statistical analysis was done using GraphPad Prism. Student’s
tional Animal Care and Use Committee. Briefly, timed pregnant CD1 two-tailed unpaired t test or one-way analysis of variance (with
mice (Charles River) were anesthetized on embryonic day (E) 14.5, Tukey’s postcomparison test) was used for all statistical analyses, ex-
and the uterine horns were exposed. Approximately 2 μg of a lenti- cept for comparison of cortical migration (Figure 2), for which a one-
viral vector DNA, containing an H1 promoter driving shRNA against tailed t test was used, and for filopodial morphologies (Figure 4), for
CPβ or firefly luciferase (as a control) and a CAG promoter driving which the binomial test was used.
mCherry fluorescent protein, was microinjected into the lateral
ventricles and introduced into the progenitor cells lining the lateral ACKNOWLEDGMENTS
ventricle by electroporation. Development was allowed to continue We thank members of the Halpain lab for advice and for providing
until E18.5. Fetal brains were sectioned on a cryostat and imaged. hippocampal cell cultures. We especially thank Soroosh Aidun for
assistance with quantitative image analyses. We acknowledge
Image acquisition and analysis Steven Dowdy, Dorothy Schafer, Tatyana Svitkina, Ray Truant,
Images for all fixed and live-cell analyses (except for cortical Torsten Wittmann, and Gene Yeo for providing reagents. We
migration; see discussion later in this section) were acquired on an also thank Andres Santos (University of California, San Diego) for

Volume 25 July 15, 2014 Capping protein in filopodia | 2159

statistical advice and Alex Mogilner (University of California, Davis) Lebrand C, Dent EW, Strasser GA, Lanier LM, Krause M, Svitkina TM, Borisy
for valuable input and critical reading of the manuscript. This work GG, Gertler FB (2004). Critical role of Ena/VASP proteins for filopodia
formation in neurons and in function downstream of netrin-1. Neuron
was supported by National Institutes of Health grants MH087823 42, 37–49.
(S.H.) and NS080194 (J.C.); S.A.S. was the recipient of a National Le Clainche C, Carlier M-F (2008). Regulation of actin assembly associated
Science Foundation Graduate GK–12 STEM fellowship. with protrusion and adhesion in cell migration. Physiol Rev 88, 489–513.
Lewis AK, Bridgman C (1992). Nerve growth cone lamellipodia contain two
populations of actin filaments that differ in organization and polarity.
REFERENCES J Cell Biol 119, 1219–1243.
Akin O, Mullins RD (2008). Capping protein increases the rate of actin- Loisel TP, Boujemaa R, Pantaloni D, Carlier M-F (1999). Reconstitution of
based motility by promoting filament nucleation by the Arp2/3 complex. actin-based motility of Listeria and Shigella using pure proteins. Nature
Cell 133, 841–851. 401, 613–616.
Applewhite DA, Barzik M, Kojima S, Svitkina TM, Gertler FB, Borisy GG Mallavarapu A, Mitchison T (1999). Regulated actin cytoskeleton assembly
(2007). Ena/VASP proteins have an anti-capping independent function in at filopodium tips controls their extension and retraction. J Cell Biol 146,
filopodia formation. Mol Biol Cell 18, 2579–2591. 1097–1106.
Arjonen A, Kaukonen R, Ivaska J (2011). Filopodia and adhesion in cancer Mattila PK, Lappalainen P (2008). Filopodia: molecular architecture and
cell motility. Cell Adhes Migration 5, 421–430. cellular functions. Nat Rev Mol Cell Biol 9, 446–454.
Bartolini F, Ramalingam N, Gundersen GG (2012). Actin-capping protein Mejillano MR, Kojima S, Applewhite DA, Gertler FB, Svitkina TM, Borisy GG
promotes microtubule stability by antagonizing the actin activity of (2004). Lamellipodial versus filopodial mode of the actin nanomachinery:
mDia1. Mol Biol Cell 23, 4032–4040. pivotal role of the filament barbed end. Cell 118, 363–373.
Bielas S, Higginbotham H, Koizumi H, Tanaka T, Gleeson JG (2004). Cortical Mogilner A, Rubinstein B (2005). The physics of filopodial protrusion.
neuronal migration mutants suggest separate but intersecting pathways. Biophys J 89, 782–795.
Annu Rev Cell Dev Biol 20, 593–618. Pollard TD, Borisy GG (2003). Cellular motility driven by assembly and disas-
Brummelkamp TR, Bernards R, Agami R (2002). A system for stable sembly of actin filaments. Cell 112, 453–465.
expression of short interfering RNAs in mammalian cells. Science 296, Ponti A, Machacek M, Gupton S, Waterman-Storer C, Danuser G (2004).
550–553. Two distinct actin networks drive the protrusion of migrating cells.
Calabrese B, Halpain S (2005). Essential role for the PKC target MARCKS in Science 305, 1782–1786.
maintaining dendritic spine morphology. Neuron 48, 77–90. Rogers SL, Wiedemann U, Stuurman N, Vale RD (2003). Molecular require-
Carlier M-F, Pantaloni D (1997). Control of actin dynamics in cell motility. ments for actin-based lamella formation in Drosophila S2 cells. J Cell
J Mol Biol 269, 459–467. Biol 162, 1079–1088.
Davis DA, Wilson MH, Giraud J, Xie Z, Tseng H-C, England C, Herscovitz Schafer DA, Gill SR, Cooper JA, Heuser JE, Schroer TA (1994). Ultrastruc-
H, Tsai L-H, Delalle I (2009). Capzb2 interacts with β-tubulin to regu- tural analysis of the dynactin complex: an actin-related protein is a com-
late growth cone morphology and neurite outgrowth. PLoS Biol 7, ponent of a filament that resembles F-actin. J Cell Biol 126, 403–412.
e1000208. Schafer DA, Welch MD, Machesky LM, Bridgman PC, Meyer SM, Cooper
Derivery E, Sousa C, Gautier JJ, Lombard B, Loew D, Gautreau A (2009). JA (1998). Visualization and molecular analysis of actin assembly in living
The Arp2/3 activator WASH controls the fission of endosomes through a cells. J Cell Biol 143, 1919–1930.
large multiprotein complex. Dev Cell 17, 712–723. Simó S, Jossin Y, Cooper JA (2010). Cullin 5 regulates cortical layering
Edwards M, Liang Y, Kim T, Cooper J (2013). Physiological role of the by modulating the speed and duration of Dab1-dependent neuronal
interaction between CARMIL1 and capping protein. Mol Biol Cell 24, migration. J Neurosci 30, 5668–5676.
3047–3055. Small J (1981). Organization of actin in the leading edge of cultured cells:
Faix J, Breitsprecher D, Stradal TEB, Rottner K (2009). Filopodia: complex influence of osmium tetroxide and dehydration on the ultrastructure of
models for simple rods. Int J Biochem Cell Biol 41, 1656–1664. actin meshworks. J Cell Biol 91, 695–705.
Faix J, Rottner K (2006). The making of filopodia. Curr Opin Cell Biol 18, Suraneni P, Rubinstein B, Unruh JR, Durnin M, Hanein D, Li R (2012). The
18–25. Arp2/3 complex is required for lamellipodia extension and directional
Gallo G (2013). Mechanisms underlying the initiation and dynamics of neu- fibroblast cell migration. J Cell Biol 197, 239–251.
ronal filopodia: from neurite formation to synaptogenesis. Int Rev Cell Svitkina TM, Bulanova EA, Chaga OY, Vignjevic DM, Kojima S, Vasiliev JM,
Mol Biol 301, 95–156. Borisy GG (2003). Mechanism of filopodia initiation by reorganization of
Gupton SL, Anderson KL, Kole TP, Fischer RS, Ponti A, Hitchcock-DeGregori a dendritic network. J Cell Biol 160, 409–421.
SE, Danuser G, Fowler VM, Wirtz D, Hanein D (2005). Cell migration Vignjevic D, Kojima S-i, Aratyn Y, Danciu O, Svitkina T, Borisy GG (2006).
without a lamellipodium translation of actin dynamics into cell move- Role of fascin in filopodial protrusion. J Cell Biol 174, 863–875.
ment mediated by tropomyosin. J Cell Biol 168, 619–631. Wear MA, Cooper JA (2004). Capping protein: new insights into mechanism
Gupton SL, Gertler FB (2007). Filopodia: the fingers that do the walking. and regulation. Trends Biochem Sci 29, 418–428.
Sci Signaling 2007, re5. Wear MA, Yamashita A, Kim K, Maéda Y, Cooper JA (2003). How cap-
Hotulainen P, Paunola E, Vartiainen MK, Lappalainen P (2005). Actin-depo- ping protein binds the barbed end of the actin filament. Curr Biol 13,
lymerizing factor and cofilin-1 play overlapping roles in promoting rapid 1531–1537.
F-actin depolymerization in mammalian nonmuscle cells. Mol Biol Cell Wu C, Asokan S, Berginski M, Haynes E, Sharpless N, Griffith J, Gomez
16, 649–664. S, Bear J (2012). Arp2/3 is critical for lamellipodia and response to
Hu L, Papoian GA (2010). Mechano-chemical feedbacks regulate actin mesh extracellular matrix cues but is dispensable for chemotaxis. Cell 148,
growth in lamellipodial protrusions. Biophys J 98, 1375–1384. 973–987.
Hug C, Jay PY, Reddy I, McNally JG, Bridgman PC, Elson EL, Cooper JA Yang C, Czech L, Gerboth S, Kojima S-i, Scita G, Svitkina T (2007). Novel
(1995). Capping protein levels influence actin assembly and cell motility roles of formin mDia2 in lamellipodia and filopodia formation in motile
in Dictyostelium. Cell 81, 591–600. cells. PLoS Biol 5, e317.
Iwasa JH, Mullins RD (2007). Spatial and temporal relationships between Zhuravlev PI, Papoian GA (2009). Molecular noise of capping protein bind-
actin-filament nucleation, capping, and disassembly. Curr Biol 17, ing induces macroscopic instability in filopodial dynamics. Proc Natl
395–406. Acad Sci USA 106, 11570–11575.

2160 | S. A. Sinnar et al. Molecular Biology of the Cell