Psidin, a conserved protein that regulates
protrusion dynamics and cell migration
Ji Hoon Kim,1Aeri Cho,1Hongyan Yin,1Dorothy A. Schafer,2,3Ghassan Mouneimne,4
Kaylene J. Simpson,4Kim-Vy Nguyen,4Joan S. Brugge,4and Denise J. Montell1,5
1Department of Biological Chemistry, Center for Cell Dynamics, Johns Hopkins School of Medicine, Baltimore, Maryland 21205,
University of Virginia, Charlottesville, Virginia 22904, USA;4Department of Cell Biology, Harvard Medical School, Boston,
Massachusetts 02115, USA
2Department of Biology, University of Virginia, Charlottesville, Virginia 22904, USA;
3Department of Cell Biology,
Dynamic assembly and disassembly of actin filaments is a major driving force for cell movements. Border cells in
the Drosophila ovary provide a simple and genetically tractable model to study the mechanisms regulating cell
migration. To identify new genes that regulate cell movement in vivo, we screened lethal mutations on
chromosome 3R for defects in border cell migration and identified two alleles of the gene psidin (psid). In vitro,
purified Psid protein bound F-actin and inhibited the interaction of tropomyosin with F-actin. In vivo, psid
mutations exhibited genetic interactions with the genes encoding tropomyosin and cofilin. Border cells
overexpressing Psid together with GFP-actin exhibited altered protrusion/retraction dynamics. Psid knockdown in
cultured S2 cells reduced, and Psid overexpression enhanced, lamellipodial dynamics. Knockdown of the human
homolog of Psid reduced the speed and directionality of migration in wounded MCF10A breast epithelial
monolayers, whereas overexpression of the protein increased migration speed and altered protrusion dynamics in
EGF-stimulated cells. These results indicate that Psid is an actin regulatory protein that plays a conserved role in
protrusion dynamics and cell migration.
[Keywords: cell migration; Drosophila; genetics; protrusion; actin, tropomyosin]
Supplemental material is available for this article.
Received January 4, 2011; revised version accepted February 11, 2011.
Cell motility is essential for many biological processes,
such as embryonic development, immune responses, and
wound healing. It is a complex, integrated process, requir-
ing changes in gene expression, signal transduction, mem-
brane organization, and the cytoskeleton (Ridley et al.
2003). The signals regulating cell migration need to be
precisely controlled because inappropriate cell migration
canresult in pathological conditionssuchasinflammation
or tumor invasion (Naora and Montell 2005).
The actin cytoskeleton provides the mechanical frame
for the force-generating machinery necessary for cells to
move. As they move, cells extend protrusions in the di-
rection of movement through the combination of actin
filament assembly and disassembly. Cellular adhesions
linkintracellular F-actin withthe extracellularsubstratum
and function as traction sites and mechanosensors. The
the traction force necessary for cells to move forward (for
review, see Ridley et al. 2003; Chhabra and Higgs 2007).
The leading edge of a migrating cell shows particularly
dynamic construction and destruction of the actin net-
work, facilitated by a variety of proteins (for review, see
Cooper and Schafer 2000; Pollard and Borisy 2003). For
filament depolymerization but also creates new barbed
ends, and thus stimulates actin polymerization when
monomer concentration is high (Ichetovkin et al. 2002;
Adrianantoandro and Pollard 2006). Behind the leading
edge, tropomyosindimerscoat the sides of actinfilaments,
protecting them from cofilin. Thus, the precise actin
dynamics in a particular cell type or a particular region
of a cell depend on the relative concentrations of actin-
binding proteins such as cofilin and tropomyosin. Even
though numerous proteins controlling actin dynamics
have been identified, it is unclear whether all of the genes
and proteins that contribute to actin dynamics and cell
motility have been identified, or whether some important
activities might have gone undetected.
Border cell migration in the Drosophila ovary provides
a well-developed genetic model system to address this
question. During Drosophila oogenesis, a group of six to
10 epithelial cells, the border cells, detaches from its
neighbors and migrates in between nurse cells to the
E-MAIL firstname.lastname@example.org; FAX (410) 955-5759.
Article published online ahead of print. Article and publication date are
online at http://www.genesdev.org/cgi/doi/10.1101/gad.2028611.
730 GENES & DEVELOPMENT 25:730–741 ? 2011 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/11; www.genesdev.org
oocyte (Fig. 1A–C; for review, see Rørth 2002; Montell
2003). Border cells extend cellular protrusions enriched
with actin during the migration (Murphy and Montell
1996; Fulga and Rørth 2002; Prasad and Montell 2007).
Furthermore, mutations in key actin-regulating proteins
such as the GTPase Rac, profilin, and cofilin cause border
and Montell 1996; Chen et al. 2001; Geisbrecht and
Montell 2004). Therefore, regulation of the actin cytoskel-
eton is critical for proper border cell migration.
In this study, we report the identification and charac-
terization of a novel F-actin-binding protein, Psidin (pro-
nounced ‘‘sigh-din’’ and abbreviated Psid). psid mutations
caused border cell migration defects in vivo. Altering Psid
expression affected protrusion dynamics in border cells
and Drosophila S2 cells. Moreover, altering the level of
expression of the human Psid homolog affected protru-
sive behavior of mammalian MCF10A cells, as well as the
speed and directionality of their movement. These results
suggest that Psid plays a conserved role in the regulation
of protrusion dynamics and cell migration.
Psid is required for border cell migration
In order to find mutations affecting border cell migration,
an EMS mutagenesis screen of the right arm of the third
chromosome was performed (Silver and Montell 2001).We
selected two mutant alleles, 55D4 and 85D1, which failed
to complement each other for lethality. In contrast to
control stage 10 egg chambers, in which border cells have
invariably reached the oocyte (Fig. 1C), border cells fre-
quently did not leave the anterior end of the egg chamber
at all in egg chambers containing clones of homozygous
mutant cells (Fig. 1D–F). Mutant border cells were speci-
fied normally because border cell-enriched proteins such
as Singed (SN) and SLBO were expressed at normal levels
result from a defect in cell fate determination. To quantify
the border cell migration defect, the extent of migration
was measured for border cell clusters in which all cells
were mutant (Fig. 1K). More than 80% of border cell
clusters mutant for the 55D4 allele failed to reach the
oocyte by stage 10, whereas 25% of clusters homozygous
for the 85D1 allele were defective.
Both alleles contained a lethal mutation that mapped to
deficiency mapping. One Piggybac insertion line, e02846,
failed to complement each EMS mutant for lethality. This
transposon is inserted in the second intron of CG4845
(Supplemental Fig. S1A). A previous study identified
CG4845 as a gene required for the innate immune
response and named the gene psid (Brennan et al. 2007).
Re-expression of Psid protein from a transgene in homo-
1K–N). In addition, the lethality of psid55D4/psid85D1flies
was rescued by actin-Gal4,UAS-psid (data not shown).
psid is predicted to encode a protein of 948 amino acids
(Supplemental Fig. S1). Both psid55D4and psid85D1con-
tained nonsense mutations: psid55D4at residue 471, and
psid85D1at residue 807 (Supplemental Fig. S1A). RT–PCR
and cDNA sequencing showed a single detectable tran-
script in the ovary, which encoded the same amino acid
sequence as predicted (data not shown). The predicted Psid
protein has two recognizable sequence motifs: a tetratrico-
peptide repeat (TPR) at its N terminus, and a pair of coiled-
coil motifs at its C terminus (Supplemental Fig. S1B). The
TPR motif is thought to mediate protein–protein interac-
tions in diverse protein families (Blatch and Lassle 1999;
D’Andrea and Regan 2003). Interestingly, the stronger
allele 55D4 causes truncation of the protein prior to both
coiled-coil domains, whereas the weaker allele 85D1
could, in principle, allow expression of a truncated protein
that would retain one coiled-coil domain. There is a single
psid homolog encoded in the genomes of most organisms,
including humans (Brennan et al. 2007; Smolikov et al.
egg chambers of the indicated stages showing border cell (yellow
arrows) development and migration to the oocyte. (D–N) Phe-
notype of psid55D4mutant clones in the absence (D–K) or
presence (K–N) of a rescuing transgene. Homozygous mutant
cells are labeled with GFP. (D–J) The border cell marker Singed
(SN) is labeled in red in D and G and white in F and J. E and
I show the GFP channel alone. (G–J) High-magnification views
of the mutant cluster shown in D–F. (K) Histogram summariz-
ing the migration defects found in the indicated numbers (n) of
egg chambers containing border cells mutant for two different
alleles of psid. (L–N) Rescue of the psid55D4phenotype. (L,M)
Homozygous mutant clones expressing GFP were generated
with UAS-psid transgene in the background (see the Materials
and Methods for details). Anti-Psid antibody staining is shown
in N. Arrowheads indicate the border between the oocyte and
nurse cells. Bars: C,D, 50 mm; G, 10 mm.
A new border cell migration mutant. (A–C) Wild-type
Psidin protein in protrusion dynamics and migration
GENES & DEVELOPMENT 731
2008). The fly and human proteins share an average of
31% amino acid identity (BLAST E value = 2e?117),
distributed through the whole protein length (Supple-
mental Fig. S1C). The sequence identity is higher (44%)
within the TPR domain.
Widespread psid expression in egg chambers
To determine the expression pattern of psid in the egg
chamber, we performed mRNA in situ hybridization and
antibody staining. An antisense RNA probe labeled germ-
line cells strongly. Border cells and outer follicle cells also
showed Psid mRNA expression (Fig. 2A,B). An antibody
against a C-terminal peptide labeled all cell types, consis-
tent with the mRNA expression, and the protein appeared
cytoplasmic (Fig. 2C,D). When we stained egg chambers
from slbo-Gal4; UAS-psid transgenic flies, the antibody
recognized the overexpressed protein in border cells, cen-
tripetal cells, and posterior follicle cells, as expected (Fig.
2E,F). In addition, psid mutant cells in mosaic clones
showed a decreased level of immunoreactivity compared
in FlyBase (http://flybase.org/reports/FBgn0243511.html)
indicates that, in the adult, Psid mRNA expression is
widespread and highest in ovary and testis, whereas in
larvae it is highest in the CNS and trachea.
psid mutations cause defects in multiple cell types
To investigate the function of the psid gene further, we
examined loss-of-function phenotypes caused by psid
mutations in additional egg chambercell types.Inaddition
to border cell migration defects, we observed multinucle-
ated cells in large clones within the follicular epithelium
(Supplemental Fig.S2A,B).This phenotypemay be due toa
defect in cytokinesis, which, like cell migration, requires
actin cytoskeleton dynamics. psid mutant follicle cells
also frequently formed multiple layers, particularly at the
posterior pole, instead of the monolayer foundinwild-type
egg chambers (Supplemental Fig. S2C,D). Multilayering of
the follicular epithelium has been described for mutations
affecting actin regulatory proteins, such as the Abelson
kinase (Baum and Perrimon 2001), or cellpolarity proteins,
such as Discs large (Goode and Perrimon 1997).
Disruption of psid function in nurse cells enabled us to
observe phenotypes related to the actin cytoskeleton more
nurse cells, phalloidin staining, which labels F-actin specifi-
cally, was brighter than in adjacent wild-type (GFP-positive)
predominantly at the cell cortex, which appeared irregular
and distorted relative to the smooth contours of wild-type
cells (Supplemental Fig. S2H). Overexpression of Psid in
nurse cells reduced phalloidin staining intensity compared
with controls stained in parallel (Supplemental Fig. S2I–L).
In cultured S2 cells, overexpression of Psid also resulted
in a reduction in the overall level of phalloidin staining
intensity and altered the distribution of F-actin (Supple-
mental Fig. S3). These findings indicate that Psid influ-
ences the level and organization of F-actin.
Effect of Psid on protrusion dynamics in S2 cells
To explore the effects of Psid on actin cytoskeleton
dynamics, we knocked down Psid expression in cultured
Drosophila S2 cells by RNAi (Fig. 3A) and evaluated the
effect on membrane ruffling and protrusion by time-lapse
microscopy. Control cells expressing EGFP-actin alone
showed, on average, eight membrane ruffles over a 10-min
interval (Fig. 3B,E; Supplemental Movie S1). In contrast,
one ruffle every 10 min (Fig. 3C,E; Supplemental Movie S2).
Cells overexpressing Psid exhibited an average of 17 dorsal
ruffles in 10 min (Fig. 3D,E; Supplemental Movie S3).
The dynamic region around the periphery of the cell is
referred to as the lamellipodium (Iwasa and Mullins
2007). We noticed that the lamellipodium was narrower
in Psid knockdown cells and wider following Psid over-
expression (Fig. 3F). The rate of rearward flow of EGFP-
actin within lamellipodia was also significantly lower in
Psid knockdown cells (Fig. 3G), suggesting reduced
lamellipodial actin dynamics.
To examine lamellipodial dynamics in more detail, we
carried out a kymography analysis (Fig. 3H–J). A kymo-
graph depicts the movement of the cell membrane at one
position over time and allows a number of parameters to
be quantified. Compared with the control, Psid knock-
Psid-overepressing cells protruded farther (Fig. 3H–K).
Moreover, Psid-overexpressing cells exhibited signifi-
cantly higher rates of protrusion and retraction compared
hybridization of late stage 9 egg chambers using antisense (A) or
sense (B) strand probes. (C) Anti-Psid antibody staining (green)
of a wild-type stage 9 egg chamber. (E) Overexpression of Psid
using slbo-Gal4 and UAS-psid transgenes. DAPI (blue) labels
DNA, and ARM (red) labels cell membranes. (G–I) Mosaic follicle
cells showing homozygous wild-type cells (GFP-positive, green in
G, white in H) and homozygous mutant (GFP-negative) cells.
(D,F,I) Same egg chambers as in C, E, and G, respectively,
showing Psid channel only. Bars: A,C,E, 50 mm; G, 10 mm.
Psid mRNA and protein expression. (A,B) In situ
Kim et al.
732GENES & DEVELOPMENT
with the control (Fig. 3L,M). Additionally, overexpressed
Psid-mRFP localized to areas of dynamic actin reorganiza-
tion at the cell edge that were visualized by the increase in
EGFP-actin using live-cell imaging (Fig. 3N,N9). In con-
trast, Tm1-mRFP was excluded from these areas at the
edge (Fig. 3O,O9). Tropomyosin stabilizes actin structures
inthe lamellum, the region behindthe lamellipodium,and
is absent from the lamellipodium (Ponti et al. 2004).
Together, these results indicate that Psid promotes lamel-
lipodial protrusion/retraction dynamics.
Altered protrusion dynamics in border cells
overexpressing Psid and EGFP-actin
To investigate Psid function in protrusion dynamics dur-
ing border cell migration, Psid was overexpressed together
with EGFP-actin, and border cell migration and morphol-
ogy were analyzed. Border cells normally extend and
retract actin-rich protrusions dynamically during their
migration (Prasad and Montell 2007). EGFP-actin is in-
corporated into actin filaments and labels the cytoplasm
and protrusions (Fulga and Rørth 2002). Border cells over-
expressing Psid and EGFP-actin at 29°C exhibited longer
protrusions than cells expressing EGFP-actin alone
(Fig. 4A,B; Table 1). Furthermore, the protrusions were
highly abnormal in morphology, as they were elongated,
irregular in length, of variable thickness, and even branched
(Fig. 4A,B; Supplemental Movie S4). Conversely, protru-
sions were shorter than the control in border cells hetero-
zygousfor the psid55D4allele (Table 1).Eventhoughneither
(A) Western blots of S2 cell lysates with or without dsRNA
treatment to knock down Psid expression. dsRNA1 and
dsRNA2 target the 59 untranslated region (UTR) and fifth exon
of Psid mRNA, respectively. a-Tubulin was used as a loading
control. (B–D) Fluorescence micrographs of EGFP-actin-express-
ing S2 cells. (B) A control cell expressing EGFP-actin alone.
Arrows indicate ruffles. (C) A Psid knockdown cell. Note the
absence of ruffles. Bar, 10 mm. (D) A cell overexpressing (OE)
Psid. Note the more spread morphology and extra ruffles.
Magnification is the same in all three panels. (E–G) Bar graphs
showing quantification of the indicated parameters measured in
control, Psid knockdown, and Psid-overexpressing cells. (H–J)
Kymographs of EGFP-actin-expressing cells. A line was drawn
perpendicular to the cell surface; that line is shown for each
frame of a time-lapse movie and shows the dynamics of the cell
edge over time. The X-axis represents time. (K–M) Quantifica-
tion of edge dynamics from kymographic analysis (see the
Materials and Methods for details). Error bars represent SEMs.
(**) P < .0001. (N,O) Time-lapse images of an S2 cell over-
expressing RFP-tagged Psid or Tm1 (red in the top panels and
white in the bottom panels). Bars, 10 mm. N9 and O9 show
enlarged pictures corresponding to boxed areas in t = 4 pictures.
Time scale is shown in minutes.
Psid regulation on protrusion dynamics in S2 cells.
border cell protrusion dynamics. (A,B,D) Late stage 9 egg cham-
bers expressing EGFP-actin directly under the control of slbo
regulatory sequences. (A) Control of the indicated genotype. (B)
Psid overexpression in border cells. (C) Histogram summarizing
stage 10 migration defects in the indicated genotypes. (D) Egg
chamber heterozygous for Tm1Su(flw)4. Arrows indicate the tips of
long protrusions. In A, B, and D, GFP channel alone is shown.
Bar, 50 mm.
Effect of Psid and EGFP-actin overexpression on
Psidin protein in protrusion dynamics and migration
GENES & DEVELOPMENT733
Psid nor EGFP-actin resulted in a severe migration defect
on itsown,coexpression of Psid and EGFP-actin resulted in
dramatic inhibition of border cell movement (Fig. 4C).
To study the dynamic effects of Psid overexpression,
time-lapse imaging of migrating border cells was carried
out in the presence of EGFP-actin, with or without Psid
overexpression. During migration, border cells expressing
EGFP-actin alone exhibited dynamic extension and re-
traction of protrusions (Supplemental Movie S5). Pro-
trusions in the direction of migration are more numerous,
longer, and longer-lived than protrusions in other di-
rections (Prasad and Montell 2007). Border cells over-
expressing Psid also extended protrusions toward the
oocyte. However, Psid-overexpressing border cells did
not retract protrusions effectively, in contrast to control
cells (Supplemental Movie S6). Protrusions kept grow-
ing and became longer, thinner, and sometimes branched
and even fragmented, which we have never observed in
controls. Several observations indicated that this pheno-
type was not an indirect consequence of the inability of
the cells to move away from the anterior. First, wild-type
border cells typically extend and retract multiple pro-
trusions prior to moving away from the anterior, so
retraction does not require that the cells move forward.
Second, other perturbations that prevent forward move-
ment, such as expression of dominant-negative Kuz, do
not cause this phenotype. Third, we observed examples of
border cells overexpressing Psid and EGFP-actin in which
the cells detached from the anterior but still extended
extra-long and abnormal protrusions.
Biochemical properties of Psid
The effects on protrusion dynamics suggested that Psid
might interact directly with actin filaments (Cooper and
Schafer 2000; Ono 2007). To determine if Psid bound to
F-actin, we performed cosedimentation assays with puri-
fied Psid protein and found a fraction of Psidin protein
cosedimentedwithF-actin (Fig. 5A,B).To determineif Psid
directly influences actin polymerization or depolymeriza-
tion, we examined assembly of pyrenyl-actin from F-actin
seeds (FAS) or nucleated by Arp2/3 complex. Psid did not
alter the rate of actin assembly from FAS or de novo actin
nucleation by Arp2/3 complex; thus, it does not cap actin
filaments, sequester G-actin, or influence the rate of
filament elongation (Fig. 5D,E). Psid also did not directly
alter the actin filament-severing activity of cofilin when
assessed using pyrenyl-actin assembly from cofilin-treated
FAS (Fig. 5F) or by TIRF microscopy (data not shown).
The predicted coiled-coil domains in the C terminus
suggested that Psid protein might form multimers. To test
this hypothesis, we carried out coimmunoprecipitation
assays. Psid protein tagged with a V5 epitope coimmuno-
precipitated with HA-tagged protein and vice versa (Fig.
5C). This suggeststhatPsid likelyfunctionsasa multimer.
However, Psid did not bundle actin filaments (data not
shown), suggesting that multimers of Psid bind the sides
of F-actin but do not cross-link actin filaments.
Interactions between Psid and Tropomyosin
Since Psid bound actin filaments, we tested for genetic
interactions between Psid and other genes encoding actin
regulatory proteins. Psid was overexpressed in border
cells using slbo-Gal4 (without EGFP-actin) at 32°C in
otherwise wild-type flies, or in flies that were hetero-
zygous for mutations in genes encoding a variety of actin
regulatory proteins (Table 2). Whereas most mutations
showed little or no effect on border cell migration in this
assay, mutations in Tropomyosin1 (Tm1) caused border
cell migration defects in 27%–50% of stage 10 egg cham-
bers examined (Table 2). About 10% of Tm1PZ2299/+ or
Tm1ZCL0722/+ border cell clusters showed impaired migra-
tion, even in the absence of Psid overexpression (Table 2).
Moreover, Tm1Su(flw)4/+ border cell clusters expressing
EGFP-actin exhibited abnormal protrusions that resem-
bled those caused by overexpression of Psid and EGFP-
actin (Fig. 4D). Since reduction in tropomyosin concentra-
tion causes a similar phenotype as overexpression of Psid,
tropomyosin and Psid may exert opposing effects on actin
organization or dynamics.
We tested the simplest possibility, which was that Psid
might interfere with Tm1 binding to F-actin. We first
confirmed that Tm1 can cosediment with F-actin in an
F-actin pelleting assay (Fig. 5G). The presence of puri-
fied Psid protein caused 3.5-fold less Tm1 to pellet with
F-actin (Fig. 5H,I), providing a possible mechanistic expla-
nation for the observed genetic interaction.
These findings suggested that a balance between Tm1
and Psid activities is important for proper protrusion
dynamics and border cell migration. We predicted, there-
fore, that Tm1 mutants would exhibit migration defects.
Tm1 mutations are homozygous lethal, so we examined
mosaic egg chambers in which all border cells were
homozygous mutant for Tm1 (Fig. 6A–F). As predicted,
two different Tm1 alleles exhibited penetrant migration
defects (Fig. 6J). Since Psid and Tm1 exhibit antagonistic
effects in other assays, we tested whether reducing the
genetic dosage of psid would have an impact on the Tm1
Effect of Psid overexpression on border cell protrusion length
Number of clusters with abnormal protrusion/total
Number of protrusions per cluster
Average length of protrusions (mm)
Numbers in parentheses are the standard deviations.
Kim et al.
734GENES & DEVELOPMENT
migration defects. Strikingly, border cells simultaneously
migration defects than either mutant alone (Fig. 6G–J).
When tropomyosin binds actin filaments, it modifies the
ability of other proteins—such as cofilin, Arp2/3 complex,
and myosin—to interact with actin. For example, tropo-
myosin binding interferes with cofilin-mediated severing
of actin filaments (DesMarais et al. 2002). Cofilin is also
required for border cell migration (Chen et al. 2001).
Although null alleles of twinstar (tsr, the gene coding for
cofilin) are lethal, 58% of adult flies that are heterozygous
for two hypomorphic alleles, tsr1and tsrntf, surviveat 18°C
and exhibit border cell migration defects. To determine if
tsr andpsid interact genetically, we crossedpsid mutations
Psidin. A fraction from the peak of protein eluted from the Hi-TrapTM chelating HP column is shown. (B) F-actin pelleting assay. The
indicated proteins were incubated in the absence (?) or presence (+) of F-actin and then centrifuged at 150,000g. The supernatant (S) and
pellet (P) fractions were analyzed by PAGE and silver-stained. Psid protein is not detectable in the pellet in the absence of actin. In
contrast, in the presence of F-actin, a significant signal is present in the pellet, similar to the positive control a-actinin. There is no change
in the amount of BSA, the negative control, found in the pellet in the presence versus absence of F-actin. Arrowheads indicate the bands
corresponding to a-actinin, BSA, and Psid. Arrows indicate actin. (C) Immunoblots (IB) of immunoprecipitated (IP) HA- and V5-tagged Psid
proteins. Anti-HA and anti-V5 antibodies were used for immunoprecipitation and immunoblots, respectively. Changing antibodies for
immunoprecipitation and immunoblots showed the same result. Lysate (0.5%) used in the immunoprecipitation was loaded for the input.
(D) Psid does not affect the rate of actin polymerization from FAS. Seeded polymerization reactions contained 1 mM G-actin (5% pyrene-
labeled) and 3 mM FAS (20-mL aliquot), with or without 475 nM psidin, as indicated, in 20 mM imidazole (pH 7.0), 50 mM KCl, 2 mM
MgCl2, 1 mM EGTA, 0.2 mM ATP, and 0.1 mM DTT (MKEI-50 buffer). Fluorescence of pyrenyl-actin (excitation at 365 nm, emission at
386 nm) was monitored for 600 sec at 25°C. Psid was dialyzed in MKEI-50 prior to use. (E) Psid does not affect actin nucleation by Arp2/3
complex and the VCA domain of N-WASP. Actin polymerization reactions contained 24 nM Arp2/3 complex, 0.75 nM GST-VCA, and 1
mM G-actin, with and without 400 nM psidin, as indicated, in MKEI-50 buffer. Fluorescence of pyrenyl-actin was monitored for 600 sec at
25°C. (F) Psid does not affect actin filament severing by cofilin. Actin filament severing by cofilin was assessed in reactions containing 3
mM FAS (20 mL) incubated with either MKEI-50 buffer or Psid, as indicated for 4 min at room temperature. Cofilin (325 nM final
concentration) was added as indicated to the FAS-Psid mixture ,and the reaction was incubated for an additional 1 min prior to dilution
into 1 mM G-actin (5% pyrene-labeled) in MKEI-50. The fluorescence of pyrenyl-actin was monitored for 600 sec at 25°C. (G) F-actin
pelleting assay for purified GST-fused Tm1 protein. (H,I) Psid interferes with Tm1 binding to F-actin. A representative result of F-actin
pelleting assay of Tm1 in the absence or presence of Psid protein is shown in D. The fraction of Tm1 in the pellet was quantified from
Western blots such as that shown in D. The average of four different experiments is shown. Error bars indicate the standard deviations. In
D and E, each protein was visualized by Western blotting with anti-actin, anti-GST, and anti-Psid antibodies, respectively.
Biochemical analysis of purified Psid protein. (A) Coomassie Blue-stained SDS gel showing purified, recombinant His-tagged
Psidin protein in protrusion dynamics and migration
GENES & DEVELOPMENT735
intothe tsr hypomorphic background. Strikingly, tsr1/tsrntf
flies that were also heterozygous for psid were 100% lethal
(Supplemental Table S1). This was true for three differ-
ent psid alleles, making it unlikely that it was due to a
nonspecific genetic background effect. Similarly tsr1/tsrntf
was fully lethal in combination with a heterozygous mu-
tation in the cofilin activator slingshot (ssh) (Supplemental
Table S1). Thus, both biochemical and genetic interaction
data support a role for Psid in modulating actin dynamics.
Human Psid required for MCF10A cell migration
To determine whether the function of Psid is conservedin
mammalian cells, we examined the effect of silencing
C12orf30, the human homolog of Psid, in MCF10A breast
epithelial cells using siRNA. MCF10A cells are a good
model to examine collective, directional cell migration in a
wound healing assay (Simpson et al. 2008). After scratch/
wounding a confluent monolayer in EGF-treated cultures,
control cells move as a sheet, where cells at the edge of the
wound display significant protrusive activity and lead the
coordinated directional movement to close the wound
onstrated that C12orf30 knockdown dramatically affected
the speed and persistence of migration, resulting in a sig-
nificant suppression of wound closure (Fig. 7A,B; Supple-
mental Movie S8). Two distinct siRNA sequences showed
similar phenotypic effects, confirmingthe specificity of the
SMART pool (data not shown). C12orf30 overexpression,
on the other hand, induced the opposite effect, causing a
significant increase in both cell speed and persistence,
which resulted in faster closure of the wound (Fig. 7C;
of the monolayer also displayed obviously larger protru-
sions than control cells.
To further characterize the enhanced protrusive activ-
ity in cells overexpressing C12orf30, we examined EGF-
induced membrane protrusion dynamics (Supplemental
Fig. S4A; Supplemental Movies S11, S12). Cells were
starved for 4 h and acutely stimulated with EGF, and
membrane protrusions were recorded by time-lapse mi-
croscopy. C12orf30 overexpression increased the extent of
was consistent with the increase in migration speed and
These results show that Psid plays animportantrole in the
regulation of EGF-induced membrane protrusion and
migration of MCF10A cells.
Border cell migration serves as a genetically tractable
model system for the identification of genes controlling
cell motility. Numerous genes have been identified in
forward genetic screens for border cell migration mutants,
and major insights into the signalingpathways that govern
their specification, as well as the timing and direction of
regulatory proteins in border cell migration
Genetic interactions between Psid and actin
Genotype Migration defect
UAS-psid; slbo-Gal4; Tm1ZCL0722/+
UAS-psid; slbo-Gal4; Tm1Su(flw)4/+
UAS-psid; slbo-Gal4; Tm1PZ2299/+
UAS-psid; slbo-Gal4, +/tsr1
UAS-psid; slbo-Gal4; ssh1-63/+
UAS-psid; slbo-Gal4; wsp3/+
UAS-psid; slbo-Gal4; ketteJ4-48/+
UAS-psid; slbo-Gal4; rac1J11, rac22D/+
UAS-psid; slbo-Gal4, +/zip1D16
Migration defect was determined by counting the number of
stage 10 egg chambers showing the incomplete migration. (n)
Total number of stage 10 egg chambers counted in the indicated
(A–C) Defective migration of Tm1Su(flw)4mutant clones. Homo-
zygous mutant cells are labeled with GFP. (D–F) Defective
migration of Tm1ZCL0722mutant clones. (E) Homozygous mutant
cells are labeled with the absence of b-Galactosidase expression.
(F) Since Tm1ZCL0722is a GFP-Trap insertion, homo- or hetero-
zygous mutant cells are GFP-positive. (G–I) Normal migration of
Tm1ZCL0722, psid55D4double-mutant border cell clones. Homo-
zygous mutant cells are bGal-negative and GFP-positive. (J)
Histogram summarizing the migration defects. The indicated
numbers (n) mean numbers of egg chambers containing border
cells mutant for the indicated mutations. Tm1ZCL0722, psid55D4
double mutants showed less severe migration defects compared
with each single mutant. In A, D, and G, arrows indicate border
cell clusters and arrowheads indicate the border between the
oocyte and nurse cells. Bars, 50 mm.
Border cell migration phenotypes by Tm1 mutations.
Kim et al.
736GENES & DEVELOPMENT
border cell movement, have been described (for review, see
Rørth 2002; Montell 2003). Two genome-wide expression
profiles for migrating border cells have also been reported
(Borghese et al. 2006; Wang et al. 2006). Although border
cell migration has been extensively characterized, satura-
tion for border cell migration mutants has not yet been
reached. In particular, our understanding of the regulation
of actin dynamics in migrating border cells is incomplete.
In the leading edge of migrating cells in culture, the
turnover ofactin filaments occursrapidly anddynamically
(Pollard and Borisy 2003). Actin filament growth by the
addition of profilin-associated G-actin to barbed ends is
the major force for membrane protrusion. Just behind the
leading edge, actin filaments are severed and depolymer-
ized by actin-disassembling factors such as cofilin. This
process is also important for the continued rapid growth of
filaments because dissociation of actin subunits replen-
ishes the pool of monomeric actin required for further
polymerization, and filament severing provides new barbed
ends, which serve as substrates for rapid polymerization
(DesMarais et al. 2005), particularly when the G-actin
concentration is high. Further inside the cell, in the region
known as the lamellum, longer and more stable actin
filaments, together with myosin and tropomyosin, form
a contractile network, which provides traction via focal
adhesionsthat linkthe F-actinstress fiberstothe substrate
(Ponti et al. 2004; Gupton et al. 2005).
It is not clear to what extent cells migrating in a three-
dimensional environment share this precise organization
of the actin cytoskeleton, and it is more difficult to probe
cytoskeletal dynamics in molecular detail in vivo. How-
ever, protrusion dynamics are clearly important in vivo as
wellas in vitro. Migrating bordercells,for example,extend
and retract long protrusions enriched in actin filaments
both cofilin and profilin to move normally (Verheyen and
Cooley 1994; Chen et al. 2001). They move in response
to chemotactic growth factors, which stimulate receptor
tyrosine kinases that are highly related to mammalian
chemoattractant receptors. Therefore, the mechanisms
governing border cell migration are likely to share much
in common with the movement of mammalian cells, and
the genetic tractability of the system offers the possibility
of identifying new molecules. Multiple lines of evidence
presented here—including loss-of-function and gain-of-
function experiments, live imaging, and protein interac-
tions—demonstrate that the conserved protein Psid binds
and promotes migration not only in border cells, but also
in mammalian cells.
psid phenotypes suggest a function in actin dynamics
In each cell type studied, the loss-of-function and gain-of-
In border cells and MCF10A cells, loss of psid caused
impaired migration. In S2 cells and MCF10A cells, altered
protrusion dynamics were evident upon Psid depletion.
In mutant nurse cells, F-actin levels were elevated and
cortical actin was irregular and disorganized, whereas
pression (OE) on wound closure in MCF10A cells. (A) The
paths of five individual cells from the wound edge were tracked
for control and C12orf30 knockdown cells. (B,C) Quantifica-
tion of cell speed (microns per minute) and persistence of cells
at the wound edge in control cells versus those transfected
with the siRNA SMART pool (B) or MCF10A cells infected
with the pBabe-C12orf30 overexpression construct (C). Values
are means 6 SEM. (D) Schematic representation of the com-
plementary distributions and proposed antagonistic functions
of Psidin and tropomyosin in a migrating cell. Psidin in the
lamellipodium inhibits tropomyosin, promoting more dynamic
assembly and disassembly. In contrast, tropomyosin is concen-
trated in the more stable lamellum, where it stabilizes actin
Effects of C12orf30 knockdown (KD) and overex-
Psidin protein in protrusion dynamics and migration
GENES & DEVELOPMENT737
nurse cells overexpressing Psid showed a reduction in
phalloidin staining. These findings suggest more specifi-
cally that Psid normally has a negative effect on filament
stability, at least in this cell type.
The Psid overexpression phenotype in border cells also
suggests defective actin dynamics. In EGFP-actin-express-
ing cells, Psid overexpression caused protrusions to grow
exceptionally long and thin and to fail to retract. This
phenotype is very different from that caused by over-
expression of Enabled, which promotes addition of actin
monomers to free barbed ends. When Enabled is overex-
short, fine protrusions form all over the cluster (Gates et al.
2009). The Psid overexpression phenotype did, however,
resemble that reported previously for border cells mutant
for spaghetti squash, a Drosophila homolog of nonmuscle
myosin II light chain, which also show extremely long
protrusions (Fulga and Rørth 2002). We observed a similar
effect in tropomyosin heterozygotes. It is easy to imagine
that, in the absence of adequate tropomyosin or myosin,
actin filaments behind the leading edge are less stable and
contractility is reduced, thus impeding retraction of pro-
trusions and forward translocation of the cell body. Thus,
Psid overexpression resembles inadequate activity of myo-
sin and tropomyosin. The phenotype of cells overexpress-
ing the human homolog of Psid also resembled that of cells
with reduced myosin II (Vicente-Manzanares et al. 2007) or
myosin light chain kinase (Simpson et al. 2008).
Another contributing factor to the Psid overexpression
phenotype could be cofilin, since psid and tsr also interact
genetically. Cofilin severs actin filaments, which can
provide new barbed ends and, somewhat counterintui-
tively, stimulate actin polymerization, particularly when
the G-actin concentration is high. This effect may con-
tribute to the generation of long protrusions when Psid is
overexpressed and could explain why this effect is dra-
matic only when GFP-actin is also overexpressed. We did
not detect a direct stimulation of cofilin’s effect on actin
in vitro however, suggesting that it is more likely the
Psid-mediated antagonism of Tm that enhances cofilin,
rather than a direct effect.
The activities of purified Psid protein in vitro provide
a plausible mechanism for the observedeffects in vivo(Fig.
7D). Psid has at least three biochemical activities: self-
association, F-actin binding, and antagonizing Tm1 asso-
ciation with F-actin. These activities suggest that Psid
forms dimers or higher-order multimers, which bind actin
filaments. In contrast with Tm-decorated filaments, Psid-
decorated filaments remain permissive for Arp2/3 com-
plex-mediated actin filament nucleation and severing by
cofilin. The observations in S2 cells that Psid knockdown
resembles the cofilin knockdown phenotype and that Psid
overexpression resembles that of tropomyosinknockdown
(Iwasa and Mullins 2007) are consistent with a model in
which Psid antagonizes the action of tropomyosin. How-
ever, this may not be the only function of Psid.
The protein most related to Psid in Saccharomyces
cerevisiae is Mdm20, which associates with NatB, an
N-acetyltransferase enzyme. The precise function of the
Mdm20 subunit is unknown. In budding yeast, NatB
acetylates tropomyosin and actin, strengthening their
interaction (Polevoda et al. 2003; Singer and Shaw
2003). Interestingly, mdm20 mutants lack the normal
actincablesthatrunfrom mothercell tobud,aphenotype
that can be suppressed by specific mutations in actin or
Likethe workinyeast, the observed geneticinteractions
in Drosophila also implicate Psid in actin dynamics. How-
ever, it is difficult to compare the results directly, since the
genetic interactions were carried out differently. In yeast,
specific amino acid substitutions in tropomyosin and in
actin were found in a genome-wide screen for suppressors
of loss-of-function mutations in mdm20. We were not able
to do the same experiment, since we do not have the same
tropomyosin or actin alleles in flies. The genetic interac-
tion in Drosophila showed that partial loss of function of
tropomyosin enhanced the effect of Psid overexpression,
and that reduction in psid gene dosage partially suppressed
the migration defects of Tm1 mutant border cells.
It remains to be seen whether the yeast protein can
interact directly with F-actin or whether Psid functions as
identity between the yeast and Drosophila proteins is low
(13% identical; 31% similar), there may also be some
differences in their functions. Alternatively, these proteins
may be multifunctional. If Psid does function as part of
an N-acetyltransferase complex, its ability to bind F-actin
could, in principle, help localize the enzyme complex to
actin filaments in proximity to key substrates such as
actin and tropomyosin. The Drosophila gene CG14222
encodes a protein that is likely to be the homolog of the
catalytic subunit of the yeast N-acetyltransferase com-
plex; however, no mutant alleles are available.
Drosophila psid was shown previously to be required
for clearance of bacterial infections (Brennan et al. 2007).
Mutant phagocytes lacking Psid were able to engulf
bacteria but not degrade them. This may reflect a re-
quirement for normal actin dynamics and cytoskeletal
organization for the movements and/or functions of
intracellular phagosomes, or, alternatively, that Psid
plays multiple roles inside cells.
Conservation of function in mammalian cells
One reason to carry out genetic screens in Drosophila is
to identify genes and proteins with conserved functions
in mammals that have not been identified using other
approaches. Therefore, it is of interest that C12orf30, the
human homolog of Psid, exhibited loss-of-function and
gain-of-function phenotypes related to cell migration
speed and persistence as well as protrusion dynamics in
MCF10A mammary epithelial cells. Live-imaging studies
reveal that C12orf30 knockdown cells fail to migrate
directionally into a scratch wound, even though they
exhibit some mobility within the monolayer. Moreover,
membrane ruffling at the leading edge was greatly re-
duced in the knockdown cells, which correlates well with
the knockdown phenotype in S2 cells. Conversely, over-
expression of C12orf30 caused increased protrusion and
delayed retraction relative to control MCF10A cells
Kim et al.
738GENES & DEVELOPMENT
following EGF stimulation. Together, these studies sug-
gest a conserved role for Psid and C12orf30 in protrusion/
retraction dynamics and cell migration, and that all of the
proteins that regulate actin dynamics and cell migration
have not yet been identified.
Materials and methods
Drosophila culture and crosses were performed following stan-
dard procedures at 25°C, except where indicated. To map the psid
locus, the original psid85D1line was crossed to the deficiency
kit for 3R (Bloomington Stock Center) and scored for lethality.
psid85D1and psid55D4alleles were meiotically mapped following
the border cell migration phenotype with respect to the recessive
markers curved (86D1-4), stripe (90D2-7), and claret (99B8-10).
The piggybac line PBac[e02846] failed to complement psid85D1
To negatively mark mosaic clones in egg chambers, FRT82B,
psid/TM3 flies were crossed to hsp70-FLP; FRT82B, ubGFPnlsor
hsp70-FLP; FRT82B, arm-lacZ. To positively mark mosaic clones
in border cell clusters, c306-Gal4; FRT82B, psid/TM3 flies were
crossed to hsp70-FLP, UAS-mCD8GFP; FRT82B, tub-Gal80. c306-
Gal4; FRT82Bline was used as a control. Dissection of ovaries was
performed 7 or 8 d after heat shock. Tm1 mutant mosaic clones
were generated in the same way using FRT82B, Tm1/TM3 flies.
To overexpress Psid in the egg chamber, the Gal4/UAS system
was used (Brand and Perrimon 1993). The following Gal4 drivers
were used: slbo-Gal4 for border cell expression, triple-Gal4 for
germline expression, and actin-Gal4 for ubiquitous expression.
Flies with the designated genotypes were incubated overnight at
29°C or 32°C before the dissection of ovaries. For the rescue of
lethality caused by psid mutations, UAS-psid was expressed by
actin-Gal4 in psid55D4/psid85D1flies.
Immunohistochemistry and immunofluorescence
Ovary dissection was performed in Schneider’s medium (GIBCO)
supplemented with 10% FBS (Sigma). Ovary fixation and staining
with antibodies, phalloidin, and DAPI was performed as described
previously (Bai et al. 2000). The primary antibodies used were
mouseanti-Singed (1:50;SN7C),mouseanti-Armadillo (1:100;N2
7A1, Developmental Studies Hybridoma Bank), mouse or rabbit
anti-GFP (1:2000; Molecular Probes), mouse anti-HA (1:1000;
Santa Cruz Biotechnology), and rabbit anti-Psid (1:1000). Alexa
Fluor-conjugated goat anti-mouse or anti-rabbit IgG antibodies
were used as the secondary antibodies (Molecular Probes). In situ
hybridization was performed with sense or antisense probes
against psid mRNA as described previously (Wang et al. 2006).
For S2 cell immunostaining, cells were plated on coverslips
coated with 0.5 mg/mL Concanavalin-A (Sigma) for 1 h and fixed
with 4% formaldehyde for 10 min. After washing three times in
PBT (0.1% Triton X-100), primary and secondary antibodies were
incubated for 1 h in PBT. Phalloidin and DAPI were incubated
with secondary antibodies.
Generation of a polyclonal antibody and expression
A peptide corresponding to the amino acid sequence 799–812 of
Psid protein (ESNGIDGLWKRRGQ) was used as an antigen for
the polyclonal antibody production in rabbits (Genescript).
To make a Psid expression construct for generating transgenic
fly lines, a full-length cDNA of the psid gene from EST clone
AT25164 was subcloned into pUAST vector or pUASp vector
(Rørth 1998). For germline transformation, each construct was
injected in a w1118 embryo according to standard procedures
For Psid expression in S2 cells, a full-length psid cDNA was
amplified by PCR, with EST clone AT25164 as a template, and
subcloned in-frame into pMT/V5-HisB (Invitrogen), pUAST-HA,
or pUAST-mRFP vector.
To make Tm1 constructs for expression in S2 cells or bacteria,
a full-length cDNA of an isoform of the Tm1 gene was amplified
by PCR from EST clone LD11194 and subcloned in-frame into
pUAST-mRFP or pGEX-5X-3 vector.
S2 cell culture and RNAi
S2 cells were cultured in Schneider’s medium (GIBCO) supple-
mented with 10% FBS (Sigma). Transfection was performed using
Effectene (Qiagen). To generate a stable line to express the Psid
protein, pMT/V5-HisB containing full-length psid cDNA was
cotransfected with pBS-Puro. Stable line selection and mainte-
nance was performed as described previously (Benting et al. 2000).
In order to knock down Psid expression in S2 cells, RNAi
2008). To generate templates for in vitro transcription, two
different regions of psid cDNAwere amplified by PCR and cloned
into pGEM-Teasyvector (Promega). Sense and antisense RNAs
were transcribed with T7 RNA polymerase (Ambion) and then
hybridized overnight. S2 cells were cultured in SF 900 medium
(GIBCO) containing 10 mg/mL dsRNA, which was exchanged
every 24 h. Cells were harvested after day 6 and analyzed by
Western blot and live imaging.
Live imaging of border cells and S2 cells
Live-imaging experiments of border cell migration were per-
formed as described previously (Prasad et al. 2007). Briefly,
ovaries were dissected in Schneider’s medium (GIBCO) supple-
mented with 10% FBS (Sigma), 0.63 penicillin/streptomycin
(GIBCO), and 0.2 mg/mL insulin (Sigma). Egg chambers were
mounted on a 50-mm Petriperm plate (Greiner Bio) and covered
with a 22-mm coverslip (Fisher).
For live-imaging experiments of S2 cell membrane dynamics,
a S2 cell line expressing EGFP-fused actin was used (Rogers et al.
2003). EGFP-actin was expressed with 1 mM CuSO4induction
for 3–6 h, then S2 cells were plated on coverslips coated with 0.5
mg/mL Concanavalin-A for 1 h. Coverslips were mounted on a
50-mm Petriperm plate for imaging.
Time-lapse images were taken using an Axiovision MRm
camera on a Zeis Axioplan 2 microscope at room temperature.
Z-stack images were integrated into a maximal intensity pro-
jection (MIP) image to generate final images.
S2 cell kymographs were generated from time-lapse sequences
(one frame/10 sec over 10 min) of S2 cells expressing GFP-actin
along 1-pixel-wide lines oriented perpendicular to the cell periph-
ery and spaced in a radial pattern around each cell (12–17
kymographs were generated per cell; five to seven cells were
analyzed for each condition). Each protrusion and retraction was
marked with a straight line along its leading or trailing edge,
respectively. Slopes of these lines were used to calculate the
velocities of protrusion and retraction, and projections of the lines
along the X-axis (time) or Y-axis (distance) were used to calculate
the persistence and distance, respectively, of protrusions (Hinz
et al. 1999). Similarly, the trajectories of GFP-actin-enriched
structures within the lamellipod were tracked to measure rates
of retrograde flow of F-actin.
Psidin protein in protrusion dynamics and migration
GENES & DEVELOPMENT739
The width of the lamellipodium was measured from summed
Z-projections of the time-lapse sequences wherein the most
dynamic region appears as a bright band encircling the cell at
its periphery. The width of the bright band was measured at ;20
locations per cell in a radial pattern around each cell; a minimum
of four cells was measured for each condition.
Psid coimmunoprecipitation and protein purification
For coimmunoprecipitation, HA- or V5-tagged Psid proteins
were expressed in S2 cells. Cells were harvested, washed in cold
PBS once, and lysed in the lysis buffer (20 mM Tris at pH 8. 0, 150
mM NaCl, 1 mM EDTA, 10% glycerol, 1% NP40, 1:500 protease
inhibitor cocktail [Sigma]). Mouse anti-HA antibody and mouse
anti-V5 antibody were used for immunoprecipitation and im-
To purify recombinant Psid protein from a S2 stable line, a 200-
mL culture containing 5 3 106cells per milliliter was induced
overnight with 1 mM CuSO4. Cells were harvested and washed in
PBS once. Cells were lysed in the binding buffer (5 mM imidazole,
20 mM Tris-HCl at pH 8. 0, 0. 2% Triton X-100, 150 mM KCl, 5%
glycerol, 5 mM 2-mercaptoethanol, 1:500 protease inhibitor cock-
tail [Sigma], 1 mM PMSF) for 1 h at 4°C. Cell lysate was cleared by
centrifugation. The supernatant was flowed through Hi-TrapTM
Chelating HP column (GE Healthcare) charged with NiCl2, and
Psid protein was eluted with imidazole.
To purify GST-fused Tm1 protein from bacteria, protein
expression was induced with 0.8 mM IPTG overnight at room
temperature. Cells were harvested and lysed by sonication in the
GST-binding buffer (25 mM Tris at pH 7.5, 150 mM KCl, 1 mM
EDTA, 1:500 protease inhibitor cocktail). After centrifugation,
the supernatant was collected and incubated with pre-equili-
brated glutathione-agarose resin overnight at 4°C. Tm1 protein
was eluted with 10 mM glutathione in GST-binding buffer.
In vitro actin biochemical assays
F-actin cosedimentation was performed according to the manu-
facturer’s protocol (Cytoskeleton). Briefly, 2 mM purified Psid, 0.5
mM F-actin prepared freshly for 1 h in F buffer (5 mM Tris-HCl at
pH 8. 0, 0. 2 mM CaCl2, 50 mM KCl, 2 mM MgCl2, 1 mM ATP).
For Psid inhibition on Tm1 binding to F-actin, 0.05–1 mM Psid
and 0.5 mM Tm1 were coincubated with 4 mM F-actin for 1 h in
Airfuge (Beckman-Coulter), and supernatants and pellets were
analyzed by SDS-PAGE and silver staining or Western blotting.
Actin polymerization assays using pyrenyl-actin were performed
in MKEI-50 buffer (20 mM imidazole, 1 mM EGTA, 2 mM MgCl2,
1 mMDTT, 50 mM KCl) with 3 mM FAS (20-mL aliquot) and 1 mM
G-actin (5% pyrene-labeled), with or without Psid, in a final
volume of 200 mL. Reactions to monitor actin nucleation by
Arp2/3 complex contained 24 nM Arp2/3 complex and 0.75 nM
GST-VCA, with and without added Psid. To assess actin filament
severing by cofilin, a 20-mL aliquot of 3 mM FAS was incubated
of 325 nM cofilin for 1 min; the reaction was subsequently diluted
into 1 mM G-actin (5% pyrene-labeled) to monitor the rate of actin
assembly from the FAS. Severing of filaments in the FAS results
in increased rates of actin assembly compared with FAS not in-
cubated with cofilin. For all assays, fluorescence of pyrenyl-actin
(excitation at 365 nm, emission at 386 nm) was monitored for 600
sec at room temperature. Actin was purified from rabbit muscle
and gel-filtered on a Sephacryl S-200 HR column (GE Healthcare)
(Spudich and Watt 1971). Pyrene-labeled actin was prepared as
described (Bryan 1986). Arp2/3 complex was purified from bovine
calf thymus (Higgs et al. 1999).
Mammalian cell culture, siRNA transfection,
and virus infection
MCF-10A cells were cultured as described (http://brugge.med.
harvard.edu/protocols). Cells were transfected at 30% confluence,
targeting C12orf30 (SMARTpool, catalog no. M-014530-00; individ-
ual siRNAs, catalog nos. D-014530-01 and D-014530-03). siRNAs
were transfected at 50 nM final concentration with DharmaFECT3
lipid, as recommended by the supplier (Thermo Fisher Scientific).
pBabe-C12orf30 or control pBabe constructs were transfected
collected after 24 h of transfection. MCF10A cells were infected
with virus suspension and cultured in selection medium for 2 wk
Wound healing and area change studies
For time-lapse video microscopy, C12orf30 knockdown cells, at
48 h post-transfection, and C12orf30 overexpression cells were
wounded using a p200 pipette tip and imaged at 5-min intervals
for 20 h using a 203 ELWD objective on a Nikon TE2000E
automated inverted microscope at 37°C. Imaging began ;45 min
dishes for 24 h, then trypsinized and replated onto glass coverslips
for a further 24 h. Cells were starved (complete starvation) for
4 h in HBSS medium containing 10 mM HEPES and 0.3% BSA.
Individual cells were imaged at 10-sec intervals for a total of 20
min, with EGF stimulation at 2 min after imaging (final concen-
tration of 25 ng/mL). Cells were imaged using a 203 DIC 0.75NA
objective on a Nikon TE2000E inverted microscope. Area changes
of individual cells were traced every minute by ImageJ, and all
values are normalized over the corresponding areas at time 0.
We thank Chang-Hun Lee for his technical support for biochem-
ical experiments. This work was supported by R01 GM73164 (to
D.J.M.), the Cell Migration Consortium U54-GM064346 (to
J.S.B.), and R01 GM67222 (to D.A.S.).
Adrianantoandro E, Pollard TD. 2006. Mechanism of actin
filament turnover by severing and nucleation at different
concentrations of ADF/cofilin. Mol Cell 24: 13–23.
Bai J, Uehara Y, Montell DJ. 2000. Regulation of invasive cell
behavior by Taiman, a Drosophila protein related to AIB1,
a steroid receptor coactivator amplified in breast cancer. Cell
Baum B, Perrimon N. 2001. Spatial control of the actin cytoskel-
eton in Drosophila epithelial cells. Nat Cell Biol 3: 883–890.
Benting J, Lecat S, Zacchetti D, Simons K. 2000. Protein expres-
sion in Drosophila Schneider cells. Anal Biochem 278: 59–68.
Blatch GL, Lassle M. 1999. The tetratricopeptide repeat: a struc-
tural motif mediating protein-protein interactions. Bioessays
Borghese L, Fletcher G, Mathieu J, Atzberger A, Eades WC, Cagan
RL, Rørth P. 2006. Systematic analysis of the transcriptional
switch inducing migration of border cells. Dev Cell 10: 497–508.
Brand AH, Perrimon N. 1993. Targeted gene expression as a
means of altering cell fates and generating dominant pheno-
types. Development 118: 401–415.
Kim et al.
740 GENES & DEVELOPMENT
Brennan CA, Delaney JR, Schneider DS, Anderson KV. 2007. Download full-text
Psidin is required in Drosophila blood cells for both phago-
cytic degradation and immune activation of the fat body. Curr
Biol 17: 67–72.
Bryan J. 1986. Isolation of fascin, an actin-bundling protein, and
SU45, an actin-severing/capping protein from sea urchin
eggs. Methods Enzymol 134: 13–23.
Chen J, Godt D, Gunsalus K, Kiss I, Goldberg M, Laski FA. 2001.
Cofilin/ADF is required for cell motility during Drosophila
ovary development and oogenesis. Nat Cell Biol 3: 204–209.
Chhabra ES, Higgs HN. 2007. The many faces of actin: matching
assembly factors with cellular structures. Nat Cell Biol
Cooper JA, Schafer DA. 2000. Control of actin assembly and
disassembly at filament ends. Curr Opin Cell Biol 12: 97–103.
D’Andrea LD, Regan L. 2003. TPR proteins: the versatile helix.
Trends Biochem Sci 28: 655–662.
DesMarais V, Ichetovkin I, Condeelis J, Hitchcock-DeGregori SE.
2002. Spatial regulation of actin dynamics: a tropomyosin-
free, actin-rich compartment at the leading edge. J Cell Sci
DesMarais V, Macaluso F, Condeelis J, Bailly M. 2005. Synergistic
interaction between the Arp2/3 complex and cofilin drives
stimulated lamellipod extension. J Cell Sci 117: 3499–3510.
Fulga TA, Rørth P. 2002. Invasive cell migration is initiated by
guided growth of long cellular extensions. Nat Cell Biol 4:
Gates J, Nowotarski SH, Yin H, Mahaffey JP, Bridges T, Herrera C,
Homem CC, Janody F, Montell DJ, Peifer M. 2009. Enabled
and Capping protein play important roles in shaping cell
behavior during Drosophila oogenesis. Dev Biol 333: 90–107.
Geisbrecht ER, Montell DJ. 2004. A role for Drosophila IAP1-
mediated caspase inhibition in Rac-dependent cell migra-
tion. Cell 118: 111–125.
Goode S, Perrimon N. 1997. Inhibition of patterned cell shape
change and cell invasion by Discs large during Drosophila
oogenesis. Genes Dev 11: 2532–2544.
Gupton SL, Anderson KL, Kole TP, Fischer RS, Ponti A, Hitchcock-
DeGregori SE, Danuser G, Fowler VM, Wirtz D, Hanein D, et al.
2005. Cell migration without a lamellipodium: translation of
actin dynamics into cell movement mediated by tropomyosin.
J Cell Biol 168: 619–631.
Higgs HN, Blanchoin L, Pollard TD. 1999. Influence of the
C terminus of Wiskott-Aldrich syndrome protein WASp and
the Arp2/3 complex on actin polymerization. Biochemistry
Hinz B, Alt W, Johnen C, Herzog V, Kaiser HW. 1999. Quantifying
lamella dynamics of cultured cells by SACED, a new com-
puter-assisted motion analysis. Exp Cell Res 251: 234–243.
Ichetovkin I, Grant W, Condeelis J. 2002. Cofilin produces newly
polymerized actin filaments that are preferred for dendritic
nucleation by the Arp2/3 complex. Curr Biol 12: 79–84.
Iwasa JH, Mullins RD. 2007. Spatial and temporal relationships
between actin-filament nucleation, capping, and disassem-
bly. Curr Biol 17: 395–406.
Montell DJ. 2003. Border-cell migration: the race is on. Nat Rev
Mol Cell Biol 4: 13–24.
Murphy AM, Montell DJ. 1996. Cell type-specific roles for Cdc42,
Rac, and RhoL in Drosophila oogenesis. J Cell Biol 133: 617–
Naora H, Montell DJ. 2005. Ovarian cancer metastasis: in-
tegrating insights from disparate model organisms. Nat Rev
Cancer 5: 355–366.
Ono S. 2007. Mechanism of depolymerization and severing of
actin filaments and its significance in cytoskeletal dynamics.
Int Rev Cytol 258: 1–82.
Polevoda B, Cardillo T, Doyle TC, Bedi GS, Sherman F. 2003.
Nat3p and Mdm20p are required for function of yeast NatB
Na-terminal acetyltransferase and of actin and tropomyosin.
J Biol Chem 278: 30686–30697.
Pollard TD, Borisy GG. 2003. Cellular motility driven by assem-
bly and disassembly of actin filaments. Cell 112: 453–465.
Ponti A, Machacek M, Gupton SL, Waterman-Storer CM,
Danuser G. 2004. Two distinct actin networks drive the
protrusion of migrating cells. Science 305: 1782–1786.
Prasad M, Montell DJ. 2007. Cellular and molecular mecha-
nisms of border cell migration analyzed using time-lapsed
live-cell imaging. Dev Cell 12: 997–1005.
Prasad M, Jang AC, Starz-Gaiano M, Melani M, Montell DJ. 2007.
A protocol for culturing Drosophila melanogaster stage 9 egg
chambers for live imaging. Nat Protoc 2: 2467–2473.
Ridley AJ, Schwarts MA, Burridge K, Firtel RA, Ginsberg MH,
Borisy G, Parsons JT, Horwitz AR. 2003. Cell migration:
integrating signals from front to back. Science 302: 1704–1709.
Rogers SL, Rogers GC. 2008. Culture of Drosophila S2 cells and
their use for RNAi-mediated loss-of-function studies and
immunofluorescence microscopy. Nat Protoc 3: 606–611.
Rogers SL, Wiedemann U, Stuuman N, Vale RD. 2003. Molec-
ular requirements for actin-based lamella formation in
Drosophila S2 cells. J Cell Biol 162: 1079–1088.
Rørth P. 1998. Gal4 in the Drosophila female germline. Mech
Dev 78: 113–118.
Rørth P. 2002. Initiating and guiding migration: lessons from
border cells. Trends Cell Biol 12: 325–331.
Silver DL, Montell DJ. 2001. Paracrine signaling through the
JAK/STAT pathway activates invasive behavior of ovarian
epithelial cells in Drosophila. Cell 107: 831–841.
Simpson KJ, Selfors LM, Bui J, Reynolds A, Leake D, Khvorova
A, Brugge JS. 2008. Identification of genes that regulate
epithelial cell migration using an siRNA screening approach.
Nat Cell Biol 10: 1027–1038.
Singer JM, Shaw JM. 2003. Mdm20 protein functions with Nat3
protein to acetylate Tpm1 protein and regulate tropomyosin-
actin interactions in budding yeast. Proc Natl Acad Sci 100:
Smolikov S, Schild-Pru ¨fert K, Colaia ´covo MP. 2008. CRA-1
uncovers a double-strand break-dependent pathway promot-
ing the assembly of central region proteins on chromosome
axes during C. elegans meiosis. PLoS Genet 4: e1000088. doi:
Spudich JA, Watt S. 1971. The regulation of rabbit skeletal
muscle contraction. I. Biochemical studies of the interaction
of the tropomyosin-troponin complex with actin and the
proteolytic fragments of myosin. J Biol Chem 246: 4866–4871.
Verheyen EM, Cooley L. 1994. Profilin mutations disrupt multiple
actin-dependent processes during Drosophila development.
Development 120: 717–728.
Vicente-Manzanares M, Zareno J, Whitmore L, Choi CK, Horwitz
AF. 2007. Regulation of protrusion, adhesion dynamics, and
polarity by myosins IIA and IIB in migrating cells. J Cell Biol
Wang X, Bo J, Bridges T, Dugan KD, Pan TC, Chodosh LA,
Montell DJ. 2006. Analysis of cell migration using whole-
genome expression profiling of migratory cells in the
Drosophila ovary. Dev Cell 10: 483–495.
Psidin protein in protrusion dynamics and migration
GENES & DEVELOPMENT 741