Developmental Cell 11, 33–46, July, 2006 ª2006 Elsevier Inc. DOI 10.1016/j.devcel.2006.05.008
Endocytic Internalization in Budding Yeast
Requires Coordinated Actin Nucleation
and Myosin Motor Activity
Yidi Sun,1,2Adam C. Martin,1,2and David G. Drubin1,*
116 Barker Hall
Department of Molecular and Cell Biology
University of California, Berkeley
Berkeley, California 94720
Actin polymerization essential for endocytic internali-
zation in budding yeast is controlled by four nucle-
ation promoting factors (NPFs) that each exhibits a
unique dynamic behavior at endocytic sites. How
each NPF functions and is regulated to restrict actin
assembly to late stages of endocytic internalization
is not known. Quantitative analysis of NPF biochemi-
cal activities, and genetic analysis of recruitment and
regulatory mechanisms, defined a linear pathway in
which protein composition changes at endocytic sites
control actin assembly and function. We show that
yeastWASP initiates actinassemblyatendocytic sites
and that this assembly and the recruitment of a yeast
WIP-like protein by WASP recruit a type I myosin
with both NPF and motor activities. Importantly, type
I myosin motor and NPF activities are separable, and
both contribute to endocytic coat inward movement,
These results reveal a mechanism in which actin nu-
cleation and myosin motor activity cooperate to pro-
mote endocytic internalization.
Endocytosis is an intricate process, involving the or-
dered recruitment of endocytic adaptors, clathrin, and
actin cytoskeletal proteins to the plasma membrane in
both yeast and mammalian cells (Engqvist-Goldstein
diated actin filament nucleation underlies a transient
burst of actin polymerization, which coincides with en-
docytic membrane invagination and vesicle scission
(Benesch et al., 2005; Kaksonen et al., 2003; Martin
et al., 2005; Merrifield et al., 2002, 2004, 2005; Yarar
et al., 2005). In yeast, actin polymerization and organiza-
tion of actin filaments into a meshwork are required for
endocytic internalization (Engqvist-Goldstein and Dru-
bin, 2003). However, little is known about the mecha-
nism by which actin assembly facilitates endocytic ves-
icle formation. Furthermore, the Arp2/3 actin nucleation
machinery and type I myosins both localize to yeast en-
docytic sites (Jonsdottir and Li, 2004; Kaksonen et al.,
2003), but the relative contributions of nucleation and
myosin motor activity to endocytosis have yet to be
The Arp2/3 complex is activated by NPFs (Pollard and
Borisy, 2003). In yeast, at least three NPFs, Las17p
(WASP), Pan1p (Eps15), and Abp1p (mAbp1), perform
endocytosis-related functions (Figure 1A) (Engqvist-
Goldstein and Drubin, 2003). In addition, Vrp1p (WIP)
stimulates type I myosin NPF activity in vitro (Geli
et al., 2000; Lechler et al., 2001; Sirotkin et al., 2005).
Live-cell imaging demonstrated that each of these
behavior at cortical patches (Jonsdottir and Li, 2004;
docytic sites in budding yeast (Huckaba et al., 2004;
Kaksonen et al., 2003, 2005; Newpher et al., 2005).
Las17p and Pan1p are recruited to cortical patches
early, arriving w20 s before actin is detected (Kaksonen
et al., 2003). In contrast, Abp1p and Myo5p are late cor-
tical patch components, arriving around the same time
as actin and the Arp2/3 complex (Jonsdottir and Li,
2004; Kaksonen et al., 2003). Las17p and Myo5p remain
associated with the plasma membrane during endocytic
internalization, whereas Abp1p and Pan1p are internal-
ized. Despite differences in NPF behavior, genetic evi-
et al., 2001; Evangelista et al., 2000; Lechler et al., 2000;
Toshima et al., 2005). Therefore, it is important to under-
stand why such a high level of complexity in Arp2/3 reg-
ulation is necessary for endocytosis. Furthermore, be-
cause several NPFs arrive before actin assembly is
initiated, it is essential to determine how actin assembly
is restricted to the late stages of the endocytic internal-
Yeast type I myosins (Myo3p and Myo5p) localize to
endocytic sites and are essential for endocytic internal-
ization (Geli and Riezman, 1996; Goodson et al., 1996;
Jonsdottir and Li, 2004). Myo3p and Myo5p contain an
amino-terminal ATPase motor domain and a carboxy-
terminal Arp2/3 binding ‘‘CA’’ domain like those found
in NPFs (Figure 1A), suggesting that these myosins
may function in both myosin-based and nucleation-
based force generation. However, a predicted Myo3p
rigor mutant blocked cortical actin polymerization in
permeabilized cells (Lechler et al., 2000). Thus, it was
doubtful that the relative importance of motor function
and actin nucleation could be evaluated by separation-
of-function mutants. Analysis of purified, full-length
Myo3/5p is required to determine whether type I myosin
motor activity and NPF activity are functionally coupled.
In this study, we use biochemistry in parallel with
live-cell imaging and genetics to examine how indi-
vidual NPFs, and a type I myosin motor, function to
control endocytic actin assembly and to drive endocytic
Quantitative Biochemical Analysis of the NPF
Activities of Four Yeast Endocytic Proteins
We sought to directly compare the NPF activities of the
four yeast NPFs, but we first needed to establish that
budding yeast type I myosin has NPF activity. Vrp1p
stimulates the NPF activity of type I myosin fragments
containing the Arp2/3 binding ‘‘CA’’ domain in both
2These authors contributed equally to this work.
2001; Sirotkin et al., 2005). However, full-length type I
myosin containing both a motor and Arp2/3 binding do-
main has never been purified. Because NPF truncations
can have considerably different activity from the full-
length protein (Rodal et al., 2003; Rohatgi et al., 1999),
we purified full-length Myo5p and Vrp1p (Figure 1B). Al-
though full-length Myo5p was inefficient at activating
the Arp2/3 complex in a pyrene-actin assembly assay,
addition of Vrp1p resulted in a 10-fold increase in the
concentration of nucleated barbed ends, saturating at
w50 nM Vrp1p (Figures 1C and 1D). A Myo5p mutant
lacking the Arp2/3 binding domain (Myo5-CADp) failed
to activate the Arp2/3 complex in the presence of
Vrp1p, demonstrating that increased actin nucleation
is due to Arp2/3 complex activation by Myo5p/Vrp1p
sion yeast type I myosin (Sirotkin et al., 2005), full-length
Myo5p/Vrp1p promoted the formation of stable Arp2/3-
dependent y-branches (Movie S1), demonstrating that
Figure 1. Vrp1p Stimulates Arp2/3 Complex Activation by Myo5p
(A) Domain structure of yeast NPFs. TH1 and TH2, tail homology 1 and 2; SH3, Src homology 3; W, WASP homology 2; C, Central; A, acidic; PRD,
proline-rich domain; EH, Eps15 homology.
(B) Coomassie blue-stained gel of purified full-length Vrp1p, Myo5p, and Myo5p mutants.
and50 nM Vrp1p,when indicated.(D)Barbed end concentrations werecalculatedasdescribed intheExperimentalProcedures.Actin(1.5 mM), 5
nM Arp2/3, and 10 nM Myo5p were used with various Vrp1p concentrations. (E) Actin (1.5 mM, 10% pyrene labeled) was polymerized with 10 nM
Arp2/3 complex, 25 nM Vrp1p, and 15 nM Myo5p or Myo5p mutants, when indicated. (F) Yeast actin (1.5 mM, 5% pyrene-labeled rabbit muscle
actin) was polymerized with 2.5 nM Arp2/3 complex and the indicated NPF. Vrp1p (50 nM) was used with Myo5p. The slopes of polymerization
curves were measured at 30% polymerization and were normalized to the maximal rate of Las17p-induced polymerization.
Myo5p/Vrp1p generates a branched actin array, similar
to Las17p (data not shown).
We next determined whether Vrp1p and Myo5p must
served tryptophan in the Src homology 3 (SH3) domain
of Myo5p (W1123S) abolishes Vrp1p binding (Geli
et al., 2000). We purified Myo5-W1123Sp and found
that direct binding of Vrp1p to Myo5p is required for
Myo5p’s full NPF activity (Figure 1E). In addition,
Myo5-W1123Sp and Vrp1p accelerated actin assembly
less than wild-type Myo5p and Vrp1p in the absence
of the Arp2/3 complex (Figure 1E). Deletion of both the
Myo5p’s ability to increase actin assembly in either the
presence or absence of the Arp2/3 complex, further
suggesting that both the SH3 and CA domains promote
actin assembly (Figure 1E).
Having established that Myo5p/Vrp1p functions as an
NPF in budding yeast, we next compared Myo5p/Vrp1p
NPF activity to that of other yeast NPFs with a homolo-
gous system of NPFs, actin, and Arp2/3 complex iso-
lated from yeast. We used 2.5 nM Arp2/3 complex to ac-
centuate differences between these NPFs. Under these
conditions, Pan1p and Abp1p had no detectable NPF
activity, whereas both Las17p and Myo5p/Vrp1p effi-
ciently activated the Arp2/3 complex (Figure 1F and
Figure S1). When we used a higher Arp2/3 complex con-
centration, Pan1p and Abp1p had detectable NPF activ-
ity (Figure S1), as reported previously (Duncan et al.,
2001; Goode et al., 2001; Toshima et al., 2005). Thus,
Las17p and Myo5p/Vrp1p have high NPF activity rela-
tive to Pan1p and Abp1p. The fact that the most potent
NPFs, Las17p and Myo5p, are not internalized during
endocytosis strongly suggests that actin filaments are
oriented primarily with their barbed ends facing the
plasma membrane during endocytic internalization, as
predicted by photobleaching studies (Kaksonen et al.,
A Linear Pathway of NPF Recruitment Precedes
Endocytic Vesicle Release
Budding yeast NPFs exhibit distinct dynamic behaviors
at endocytic sites (Jonsdottir and Li, 2004; Kaksonen
et al., 2003). Vrp1-GFP also had distinct dynamics
(Figure 2A), exhibiting a lifetime of w20 s, longer than
both Myo5-GFP (w10 s) and actin (w15 s), but shorter
than Las17p and Pan1p (w30–40 s). Similar to Las17p,
but unlike Pan1p or Abp1p, Vrp1-GFP remained nonmo-
tile throughout its lifetime (Figure 2A and Movie S2).
We nextexamined thespatiotemporal relationshipsof
all NPFs by using Myo5p as a reference (Figure 2B and
Movie S3). As summarized in Figure 2C, Las17p and
Pan1p appear at cortical patches first, followed by
Abp1p, Myo5p, and the Arp2/3 complex (Arc15p) (Jons-
dottir and Li, 2004; Kaksonen et al., 2003). Vrp1p ap-
peared at patches after Las17p but before Myo5p and
Abp1p (Figures 2B and 2C). Because extensive pro-
tein-protein interactions link Las17p, Myo5p, and
Vrp1p (Goode and Rodal, 2001), we tested whether
these proteins recruit each other to cortical patches. In
las17D cells, Vrp1-GFP no longer assembled into corti-
cal patches, demonstrating that Vrp1p localization re-
quires Las17p (Figure 2D), consistent with results in fis-
sion yeast (Sirotkin et al., 2005). In vrp1D cells, Las17p
still localized to patches (Figure S2A); however, Myo5p
cortical patch intensity was greatly reduced (Figure 2E).
Vrp1p appeared to recruit Myo5p via Myo5p’s SH3 do-
main, since Myo5-W1123S-GFP exhibited similar partial
mislocalization in the presence of wild-type Vrp1p (Fig-
ure 2E). Because the timing of Myo5-GFP recruitment
coincided with the initiation of actin assembly, it was
possible that actin filaments contribute to Myo5p local-
ization. Indeed, inhibition of actin assembly by Latrun-
culin A (LatA) resulted in partial loss of Myo5p from cor-
tical patches in wild-type cells (Figure S2B) (Anderson
et al., 1998) and complete loss of Myo5p in vrp1D cells
(Figure 2F). Overall, these results suggest a functional
hierarchy of protein recruitment to cortical patches in
which Las17p recruits Vrp1p, which, together with actin
filaments, recruits Myo5p. Because binding to Vrp1p is
required for Myo5p NPF activity, Myo5p is likely acti-
vated upon its recruitment to endocytic sites.
Myo5-GFP fluorescence was previously reported to
appear around the same time as Arp2-DsRed and to
peak immediately before the transition of Arp2-DsRed
from its slow-to-fast movement phases (Jonsdottir and
Li, 2004). Because actin patch fast movement likely rep-
resents movement of released endocytic vesicles (Kak-
sonen et al., 2003, 2005), Myo5p was proposed to play
a role in vesicle scission (Jonsdottir and Li, 2004). How-
ever, DsRed is an obligate tetramer (Baird et al., 2000),
thus the DsRed moiety may have altered actin and
Arp2/3 complex dynamics in those studies. Abp1p
tagged with monomeric RFP (Abp1-RFP) is a well-vali-
dated marker for actin (Kaksonen et al., 2005). There-
tion Myo5p might function, we simultaneously imaged
Myo5-GFP and Abp1-RFP with a high time resolution
of 1 s. Analysis of patch intensity demonstrated that
Myo5p peaked before Abp1p and the Arp2/3 complex
(Arc15p) and was no longer detectable during the transi-
tion from slow to fast movement (Figures 2B and 2C and
Movie S3). Recently, the peak fluorescence intensity for
a GFP-tagged yeast amphiphysin homolog, Rvs167p,
was shown to coincide with its one-step, inward move-
mentintothe cytoplasm, possibly representing avesicle
scission event (Kaksonen et al., 2005). Importantly,
Myo5p/Vrp1p, Las17p, and Pan1p all peaked before
the Rvs167p peak (Figures 2B and 2C), suggesting that
these NPFs function prior to vesicle scission.
To determine the stoichiometry of NPFs and their
binding partners in cortical patches, we used GFP-fluo-
(Figure2G,MoviesS5 andS6,andFigure S2C).Because
these proteins are dynamic, we measured the peak in-
tensity of each fusion protein at endocytic sites. Abp1p
was the most abundant cortical patch protein, appear-
ing to be 2-fold more abundant than the Arp2/3 complex
(Arc15p) and 6-fold more abundant than Las17p. Vrp1p
had a similar intensity to Las17p, consistent with its lo-
calization being dependent on Las17p. There was half
as much Vrp1p as Myo5p, suggesting that at most,
only half of the Myo5p is activated by Vrp1p at a given
time in vivo. This fraction is likely to be even lower since
(Anderson et al., 1998; Evangelista et al., 2000). These
stoichiometries agree with the observation that Myo5p
is only partially dependent on Vrp1p for cortical patch
The Mechanism of Actin’s Endocytic Function
Figure 2. Ordered Recruitment of NPFs and Their Regulators to Endocytic Sites
(A) Kymographs of single patches from movies of cells expressing the indicated GFP-tagged protein (one frame/s). Lifetimes 6 standard devi-
ations (SD) are indicated in parentheses.
(B) Montages, single channel or merged images, of single patches from two-color movies of cells expressing GFP-tagged and RFP-tagged pro-
teins. Time lapse between frames is 1 s.
(C) Alignment of averaged patch intensity measurements of NPFs and their regulators. The data were averaged from at least three patches with
one-color movies of GFP-taggedendocytic proteins. Traces were aligned by time separating intensity peaks in two-colormovies for 15 patches.
Dotted lines show the distance of Sla1p (blue) and Abp1p (red) movement.
(D) Vrp1-GFP localization depends on Las17p. Vrp1-GFP was expressed in wild-type and las17D cells.
(E) Myo5-GFP localization depends partially on Vrp1p. Myo5-GFP was expressed in wild-type and vrp1D cells. Myo5-W1123S-GFP was ex-
pressed in otherwise wild-type cells.
(F) Myo5-GFP localization requires both Vrp1p and F-actin. Abp1-RFP (left) and Myo5-GFP (right) were coexpressed in vrp1D cells, and either
DMSO (top) or 200 mM LatA (bottom) was added for 10 min at 25ºC.
(G) Cortical patch protein stoichiometries. Peak GFP fluorescence from at least 15 patches was averaged for each fusion protein. Fluorescence
was normalized to Abp1-GFP fluorescence. See Figure S2 for further details. All scale bars are 1 mm.
recruitment. Because the total NPF concentration is
three times greater than that of the Arp2/3 complex,
only a fraction of these NPFs can interact with the
Arp2/3 complex at a given time.
Distinct NPF Functions during Endocytic
To dissect the functions of individual NPFs in the endo-
complex binding domains of Las17p (las17-WCAD),
Pan1p(pan1-WAD),andMyo5p (myo5-CAD).These mu-
tants exhibited wild-type protein expression levels and
localization (data not shown). Because Myo3p and
Myo5p function redundantly (Geli and Riezman, 1996;
Goodson et al., 1996), analysis of myo5-CAD was per-
formed in a myo3D background. Mutant cells express-
ing a component of the endocytic coat module, Sla1-
GFP, and a marker for actin, Abp1-RFP, were imaged
to visualize endocytosis in real time. In wild-type cells,
Sla1-GFP containing patches are joined by Abp1-RFP
shortly before both of these proteins disappear (Kakso-
nen et al., 2003). NPF mutations had specific effects on
the timing of Abp1p recruitment. The las17-WCAD mu-
tant delayed actin polymerization at endocytic sites,
represented by the delay in Abp1-RFP recruitment
(Figure 3A and Movie S7). Neither pan1-WAD, myo5-
CAD myo3D, nor pan1-WAD myo5-CAD myo3D delayed
actin polymerization, suggesting that Las17p is suffi-
cient to initiate actin polymerization at endocytic sites
(Figure 3A and Movies S7 and S8). However, when com-
bined with las17-WCAD, both pan1-WAD and myo5-
CAD myo3D further delayed actin assembly initiation
(Figure 3A). Thus, Pan1p and Myo3/5p NPF activity are
important for initiating actin assembly only in the ab-
sence of Las17p.
In wild-type cells, Abp1p recruitment accompanies
the slow inward movement of both Sla1p and Abp1p
away from the plasma membrane (Kaksonen et al.,
2003). This slow Sla1-GFP movement likely reflects en-
docytic membrane invagination (Kaksonen et al., 2003,
2005). Surprisingly, although las17-WCAD caused a sig-
inward movement was only slightly lower than that in
wild-type cells (Figure 3A and Movie S7). Furthermore,
most Sla1-GFP patches in pan1-WAD and pan1-WAD
las17-WCAD mutant cells showed inward movement
tant defective in Las17p and Pan1p NPF activity shows
onlyamodest defectin a-factorinternalization (Toshima
et al., 2005). In contrast, the frequency of Sla1-GFP in-
ward movement was significantly reduced in the
myo5-CAD myo3D mutant (Figure 3A). The myo3D
myo5D and vrp1D mutants blocked Sla1-GFP inward
movement (Figure 3A and Movie S9), further supporting
the conclusion that Myo5p/Vrp1p NPF activity is impor-
tant for endocytic internalization. Strikingly, the las17-
WCAD myo5-CAD myo3D triple mutant was completely
defective in Sla1-GFP inward movement (Figure 3A and
Movie S8), demonstrating that although Myo5p/Vrp1p is
most important, Las17p and Myo5p/Vrp1p NPF activi-
ties both contribute to endocytic internalization. The
myo3D triple mutant likely resulted from a defect in actin
nucleationsince Abp1-RFPfluorescence wasgreatlyre-
duced at endocytic sites (Figure 3A and Movie S8). The
timing of actin assembly in the las17-WCAD myo5-CAD
myo3D mutant was indistinguishable from that in the
las17D mutant (Figure 3A). This is probably because
the las17D mutant causes Vrp1p mislocalization, which
subsequently affects Myo5p localization and NPF activ-
ity. Together, these data demonstrate that the two most
potent NPFs, Las17p and Myo5p/Vrp1p, are most im-
Regulation of Actin Nucleation Activity
As shown above, Las17p initiates actin assembly at en-
docytic sites. However, Las17p is present in cortical
patches for w20 s before actin is detected (Kaksonen
et al., 2003). Sla1p may inactivate Las17p until the ap-
we tested whether other SH3-containing proteins can
activate Las17p NPF activity in the presence of Sla1p.
We found that purified Bzz1p (Figure S3), a syndapin-
like protein with two SH3 domains, relieved Las17p inhi-
bition by Sla1-SH3p in a dose-dependent manner (Fig-
ures 3B and 3C). Relief of Sla1p’s inhibition of Las17p
required Bzz1p’s SH3 domains (data not shown). There-
fore, this result reveals a novel mechanism for the acti-
vation of a WASP-family protein, in which inhibition of
NPF activity by one SH3-containing protein is relieved
by another SH3-containing protein. To investigate
whether Bzz1p regulates Las17p NPF activity in vivo,
we examined Bzz1-GFP dynamics in live cells. Bzz1-
GFP remained associated with the plasma membrane
during endocytic internalization, similar to Las17p, but
exhibited a lifetime of w17 s (Figure 2A and Movie S4).
Two-color imaging demonstrated that Bzz1-GFP was
recruited to cortical patches immediately before actin
polymerization was initiated (Figures 2B and 2C and
Movie S3), consistent with Bzz1p relieving Las17p inhi-
bition by Sla1p. Although initiation of actin polymeriza-
tion was normal in the bzz1D mutant, actin polymeriza-
tion was delayed in bzz1D vrp1D cells (data not
shown), suggesting that Bzz1p contributes to actin as-
terized SH3-containing Las17p interacting proteins may
also function with Bzz1p to activate Las17p in vivo (Ma-
dania et al., 1999; Tong et al., 2002). In total, these data
suggest that Las17p activity is regulated by the protein
composition of the patch, which changes dynamically.
to prevent excessive polymerization, which may impair
proteins (Sekiya-Kawasaki et al., 2003; Toshima et al.,
2005). Bbc1p inhibits Las17p NPF activity in vitro (Rodal
et al., 2003), and bbc1D cells exhibit exaggerated Sla1p
and Abp1p inward movement, with distances of up to
four times as far into the cytoplasm (Kaksonen et al.,
2005). Because Bbc1p dynamics were indistinguishable
from those ofMyo5p (Figure 2B), and Myo5p/Vrp1p NPF
activity appeared most important for Sla1-GFP slow
movement, we determined whether Bbc1p also inhibits
Myo5p/Vrp1p Arp2/3 complex activation. Bbc1p in-
hibited Myo5p/Vrp1p activity in a dose-dependent man-
ner, similar to Las17p (Figure 3D). To determine whether
the enhanced Sla1-GFP inward movement observed in
The Mechanism of Actin’s Endocytic Function
Figure 3. NPF Functions during Endocytic Internalization
(A) Kymographs of individual patches from two-color movies of cells expressing Sla1-GFP, Abp1-RFP, and the indicated NPF mutant(s). Kymo-
graphs were aligned with Abp1-RFP appearance (dotted line). Means of the frequency of Sla1-GFP inward movement (%) and Sla1-GFP lifetime
before Abp1-RFP appearance (s) 6 SD are indicated to the right. At least 45 patches from three cells were analyzed to calculate the frequency of
Sla1-GFP inward movement.
(B) Bzz1p relieves Las17p inhibition by Sla1-SH3p. Actin (1 mM, 5% pyrene labeled) was polymerized with 10 nM Arp2/3 complex, 10 nM Las17p,
250 nM Sla1-SH3p, and 150 nM Bzz1p, when indicated.
(C) Dose dependence of Bzz1p activation of Las17p. Actin (2 mM, 5% pyrene labeled), 20 nM Arp2/3 complex, 20 nM Las17p, and 250 nM Sla1-
SH3p were used.
(D) Bbc1p inhibits Myo5p/Vrp1p induced actin assembly. Percent inhibition was calculated with the reduction in the concentration of nucleated
barbed ends. Actin (1.5 mM) was polymerized with 10 nM Arp2/3, 15 nM Myo5p, and 25 nM Vrp1p or 5 nM Arp2/3 and 10 nM Las17p.
(E) Abp1p inhibits Myo5p/Vrp1p induced actin assembly. Actin (1.5 mM) was polymerized with 5 nM Arp2/3, 5 nM Myo5p, and 25 nM Vrp1p or 2.5
nM Arp2/3 and 5 nM Las17p.
(F) Kymographs from two-color movies of bbc1D cells expressing Sla1-GFP, Abp1-RFP, and the indicated NPF mutant(s). The mean distance of
Sla1-GFP inward movement 6 SD was calculated from five patches for each strain.
or Myo5p/Vrp1p NPF activity, we crossed the bbc1D
mutant with the NPF mutants. The enhanced Sla1-GFP
inward movement in bbc1D cells was specifically sup-
pressed by myo5-CAD myo3D and vrp1D, but not by
las17-WCAD or pan1-WAD (Figure 3F and Movie S10).
These results suggest that misregulation of Myo5p/
Vrp1p NPF activity contributes to the exaggerated
cells. Therefore, Myo5p/Vrp1p regulation by Bbc1p ap-
pears to affect the duration of endocytic coat inward
In addition to Bbc1p, Abp1p can also inhibit Arp2/3
complex activation by Las17p in vitro (D’Agostino and
Goode, 2005). This inhibition was not specific to
Las17p, as Abp1p also inhibited Myo5p/Vrp1p activity
in a dose-dependent manner (Figure 3E). Therefore,
to function as a general NPF inhibitor, possibly limiting
thelevel ofnucleation thatoccursatanendocytic patch.
Myo5p’s Motor Activity Functions Independently
from Its NPF Activity
So far, our study has focused on the importance of actin
nucleation during endocytosis. The type I myosin motor
is another ‘‘force generator’’ that may also contribute to
endocytic internalization. Because Myo3/5p motor ac-
tivity was proposed to affect actin polymerization in per-
meabilized cells (Lechler et al., 2000), a critical question
is whether Myo5p motor activity can be distinguished
from its actin assembly activity. To address this ques-
tion, we first purified full-length Myo5p and Myo5p mo-
tor mutants to test their motor activities in vitro (Fig-
ure 4A). Mutation of glycine 132 to arginine (G132R)
changes a conserved residue of the nucleotide binding
region and is predicted to be a rigor mutant (Lechler
et al., 2000). Phosphorylation by p21 activated kinases
(PAKs) is proposed to regulate yeast type I myosin mo-
tor activity (Wu et al., 1996). Therefore we also mutated
the phosporylated serine (S357) of the PAK consensus
tive). In addition, we purified Myo5p that completely
lacked the motor domain (motorD).
Using an actin filament gliding assay, we assessed
Myo5p motor activity by measuring the speed of actin
filament translocation across a Myo5p-coated coverslip
(Figure 4B and Movie S11). Wild-type Myo5p moved
rabbit actin filaments at a speed of 0.15 6 0.03 mm/s,
similar to Acanthamoeba type I myosin (Zot et al.,
1992). Similar gliding speeds were obtained with yeast
actin filaments (0.17 6 0.04 mm/s). These observations
demonstrate that yeast type I myosin has motor activity.
Actin filament translocation was dependent on the mo-
tor domain as Myo5-motorDp and the Myo5-G132Rp
mutants failed tomove actin filaments (Figure 4B).In ad-
dition, Myo5-S357Ap and Myo5-S357Dp reduced the
speed of actin filament movement by 80% and 60%, re-
spectively (Figure 4B). Because both the S357A and
S357D mutants were defective in actin filament gliding,
it is not clear whether the S357D mutation fully mimics
To determine whether Myo5p motor activity is re-
quired for its NPF activity, we investigated whether
Myo5p motor mutants activate the Arp2/3 complex
with the pyrene-actin assembly assay. All Myo5p motor
mutants, including the motor deletion, showed indistin-
guishable NPF activity from the wild-type protein
(Figure 4C). Furthermore, Myo5-CADp exhibited normal
binding region of Myo5p is not required for motor func-
tion. Therefore, the motor and NPF functions of Myo5p
Myo5p Motor Activity Is Required for Endocytic Coat
We next determined whether Myo5p motor activity is re-
quired for endocytosis. We integrated the myo5-G132R,
myo5-S357A, and myo5-S357D mutations into the yeast
genome, replacing the endogenous copy of MYO5.
Myo5p motor mutant proteins were expressed at wild-
type levels (data not shown). Although myo5-G132R,
myo5-S357A, and myo5-S357D grew normally in the
presence of MYO3, myo5-G132R, myo5-S337A, and to
a lesser extent, myo5-S357D, growth was temperature
sensitive when combined with the myo3D mutant
(Figure S4). Examination of GFP-tagged Myo5p mutants
demonstrated that Myo5-G132R was recruited to corti-
cal patches with similar efficiency to the wild-type pro-
tein (Figure 5A and Movie S12). In contrast, Myo5-
S357A and Myo5-S357D were present in cortical
patches at reduced levels (Figure 5A and Movie S12).
Because F-actin is important for Myo5p patch localiza-
tion, these mutants may decrease Myo5p affinity for F-
actin. Myo5p lifetime at patches was increased for all
three mutants, with Myo5-G132R being the longest lived
and Myo5-S357D being the shortest lived (Figure 5B).
Thus, Myo5p lifetimes are inversely correlated with the
in vitro motor activities of these mutants (Figure 4B).
To determine how Myo5p motor activity affects endo-
cytic internalization, we analyzed Sla1-GFP and Abp1-
RFP dynamics in myo5-G132R myo3D, myo5-S357A
myo3D, and myo5-S357D myo3D mutant cells. Sla1-
GFP inward movement was defective in Myo5p motor
mutants,suggesting that Myo5p motor activity is essen-
tial for plasma membrane invagination (Figure 5C and
Movie S13). In addition, internalization of a marker for
fluid phase endocytosis, Lucifer yellow, was defective
in all three motor mutants, similar to the myo5D
myo3D mutant (Figure 5D). Mislocalization of Myo5-
S357Ap, and Myo5-S357Dp, is unlikely to be solely re-
sponsible for their endocytic defects because the
myo5-W1123S mutant, which was also partially mislo-
calized, was only partially defective in Sla1-GFP and Lu-
cifer yellow internalization (Figures 5A–5D). Further-
more, Myo5-G132R is present at normal levels in
corticalpatches. Importantly, Abp1-RFP isefficientlyre-
cruited to Sla1-GFP patches in the motor mutants, pro-
viding strong evidence that Myo5p motor activity is not
required for cortical actin polymerization in vivo
(Figure 5C). Together, these data suggest that Myo5p
motor activity, independent of Myo5p actin nucleation
activity, is required for endocytic internalization.
To determine whether the elevated NPF activity pres-
ent in a bbc1D mutant can suppress the Myo5p motor
mutant internalization defect, we combined Myo5-
G132R-GFP with the myo3D and bbc1D mutants and
examined Abp1-RFP inward movement. Myo5-G132R-
GFP mutant cells were completely defective in the
The Mechanism of Actin’s Endocytic Function
inward movement of Abp1-RFP patches, which nor-
ure 5E and Movie S14). Thus, even in the presence of el-
evated NPF activity, Myo5p motor activity is required for
internalization. This result further suggests that Myo5p
motor activity is not required for actin polymerization,
but perhaps plays a direct role promoting membrane
Myo5p Motor Activity Is Required for Actin Network
Las17p and Myo5p/Vrp1p NPF activity and Myo5p mo-
tor activity are both required for endocytic internaliza-
tion (Figure 6A). The mechanism by which these activi-
ties promote internalization needs to be investigated.
Wild-type actin patches are smaller than the resolving
power of a fluorescence microscope, making it impos-
sible to observe the structure, organization, and orien-
tation of actin filaments by this technique. However, de-
letion of Sla2p, a protein essential for endocytosis,
results in elongated actin tails that originate from a
stable complex of NPFs, endocytic adaptors, and cla-
thrin (Kaksonen et al., 2003; Newpher et al., 2005). Pho-
tobleaching studies indicated that actin subunits are
added near the plasma membrane and are treadmilled
back into the cytoplasm (Kaksonen et al., 2003), as is
expected to occur in wild-type patches. Therefore, we
used the sla2D mutant as a model in vivo system to
Figure 4. Myo5p Functions as a Molecular Motor
(A) Coomassie blue-stained gel of purified full-length Myo5p and Myo5p motor mutants.
(B)Actinfilamentgliding assaydemonstrating Myo5p motor activity. Rhodamine phalloidin-labeledactin filaments werebound toMyo5p coated
coverslips, and the gliding reaction was initiated with ATP. Single frames from a 1 min movie are presented (two frames/s). Stars indicate the
starting positions of individual filament ends. Paths of individual actin filaments are shown in color. The mean speed 6 SD is shown above
each mutant. Scale bar = 2.5 mm.
(C) Myo5p motor activity is independent of Myo5p/Vrp1p NPF activity. Actin (1.5 mM, 10% pyrene labeled) was polymerized with 10 nM Arp2/3
complex, 25 nM Vrp1p, and 15 nM Myo5p or Myo5p mutants, when indicated.
examine how NPF and Myo5p motor activities affect
We expressed either Myo5-GFP, Myo5-CAD-GFP (in
a myo3D), or Las17-WCAD-GFP in sla2D mutant cells
also expressing Abp1-RFP to label actin tails. Similar
to Las17-GFP, Myo5-GFP localized to the heads of actin
comet tails (Figure 6B). Interestingly, actin tails still
formed in both the Myo5-CAD-GFP and Las17-WCAD-
and Movie S15). A similar phenotype for sla2D tails was
lower nucleation activity and impair endocytic internali-
zation (Martin et al., 2005). Therefore, reducing actin po-
lymerization efficiency does not block the formation of
Figure 5. Myo5p Motor Activity Is Required for Endocytic Coat Inward Movement
(A) Cells expressing GFP-tagged Myo5p and Myo5p mutants.
(B) Myo5p mutant lifetime in a myo3D mutant background. Error bars represent SD.
(C) First image (top) and kymograph representation (bottom) from two-color movies of cells expressing Sla1-GFP, Abp1-RFP, and the indicated
Myo5p mutant. Means for the frequency of Sla1-GFP inward movement 6 SD were calculated from at least 45 patches from three cells.
(D) Fluorescence microscopy analysis of Lucifer yellow endocytic uptake. Cells were incubated with Lucifer yellow for 2 hr at 25ºC.
GFP and Abp1-RFP in wild-type or a myo3D mutant background, respectively. All scale bars are 1 mm.
The Mechanism of Actin’s Endocytic Function
sla2D actin tails. In contrast, Myo5-G132R-GFP (in
a myo3D) resulted in a striking block in actin tail forma-
tion in sla2D cells (Figure 6C and Movie S16). Actin poly-
merization still occurred at Myo5-G132R-GFP sites;
however, instead of actin filaments flowing away from
the plasma membrane in tails that project into the cyto-
plasm, patch-like actin structures that colocalized with
Myo5-G132R-GFP were formed (Figure 6C). Similar re-
sults were also obtained with the Myo5-S357A mutant
(data not shown). These results suggest that Myo5p
motor activity is required for actin filament retrograde
flow away from the plasma membrane.
A Pathway for Endocytic Actin Assembly
and Force Production
Actin and actin-associated proteins arrive at endocytic
sites in a highly defined temporal order in both yeast
and mammalian cells (Merrifield, 2004). Although bud-
ding yeast NPFs are recruited to endocytic sites with
distinct timing and have distinct motility behaviors,
cytosis was not clear. Combining live-cell imaging of en-
docytosis with genetics and biochemical experiments
using full-length proteins has enabled us to propose
ners function and how they are recruited to, and regu-
lated at, endocytic sites in budding yeast (Figure 7).
Las17p (WASP) arrives at endocytic sites w20 s before
actin assembly is initiated (Kaksonen et al., 2003).
Las17p serves as a scaffold that assembles key compo-
nents of the actin machinery for endocytic internaliza-
tion. Initially, Sla1p, and possibly other early endocytic
proteins, maintain Las17p in an inactive state (Rodal
et al., 2003). Vrp1p (WIP) and Bzz1p (syndapin) are sub-
sequently recruited to endocytic sites by Las17p (this
study; Sirotkin et al., 2005; Soulard et al., 2002). Bzz1p,
and possibly other SH3-containing proteins, relieves
Las17p inhibition by Sla1p, initiating actin nucleation at
endocytic sites. Vrp1p and F-actin then recruit Myo5p
(type I myosin) to endocytic sites, and Vrp1p activates
brane invagination, which is most dependent on Myo5p/
Vrp1p NPF activity. Concurrent with membrane invagi-
nation, Abp1p and Bbc1p negatively regulate Las17p
and Myo5p/Vrp1p NPF activity (this study; D’Agostino
and Goode, 2005; Rodal et al., 2003). In particular,
Bbc1p inhibition of Myo5p/Vrp1p NPF activity regulates
the duration of coat protein inward movement. This reg-
ulation presumably fine-tunes the level of actin nucle-
ation at endocytic sites, ensuring the proper size and
architecture of the actin meshwork.
Las17p and Myo5p/Vrp1p NPF activity and Myo5p
motor activity are required for endocytic internalization.
We speculate that actin nucleation and myosin motor
activity cooperate to generate forces that drive mem-
brane invagination. We showed that the NPFs that re-
main nonmotile at the plasma membrane, Las17p and
Myo5p/Vrp1p, have the highest NPF activity, which
Figure 6. Myo5p Motor Activity Is Required
for Actin Network Retrograde Flow
(A) Summary of in vivo and in vitro pheno-
types of Myo5p mutants. T.S., temperature
sensitive; P.M., plasma membrane; Cyto.,
(B) Actin tail formation in sla2D cells express-
ing Abp1-RFP and either Myo5-GFP, Myo5-
CAD-GFP, or Las17-CAD-GFP. First image
(top) and kymograph representation (bottom)
from two-color movies (one frame/s).
(C) First image (left) and kymograph repre-
sentation (right) from a two-color movie of
sla2D myo3D cells expressing Abp1-RFP
and Myo5-G132R-GFP. All scale bars are
further supports the conclusion based on photobleach-
ing studies that actin nucleation occurs preferentially at
the plasma membrane in wild-type cells (Kaksonen
et al., 2003, 2005). The likely actin filament orientation
at endocytic sites indicates that a type I myosin associ-
ated with these filaments would walk toward the plasma
membrane. Myo5p motor activity may contribute to
membrane invagination by translocating actin filaments
associated with endocytic coat proteins away from the
plasma membrane into the cytosol. Sla2p and Pan1p
may link actin filaments to the endocytic coat (Baggett
et al., 2003; Henry et al., 2002; Sun et al., 2005; Toshima
et al., 2005; Wendland and Emr, 1998). This model for
actin-mediated endocytic membrane invagination is
supported by the observation that Myo5p mutants de-
fective inmotor activity fail to translocate actin filaments
offtheplasmamembrane insla2Dcells.Inaddition, type
I myosin motor activity may be required for the proper
organization and/or orientation of actin filaments at en-
docytic sites such that actin polymerization effectively
generates force. Further investigation is needed to clar-
Clathrin-mediated endocytosis in yeast and mamma-
liancells isfar moresimilar than had previously been ap-
tin machinery, including N-WASP, WIP, and syndapin,
share homology to yeast endocytic proteins, and in
both instances, these proteins play a role in clathrin-
mediated endocytosis (Kaksonen et al., 2005; Merrifield,
2004; Newpher et al., 2005). Type I myosins are also im-
portant for endocytosis and other membrane-related
processes in a variety of organisms (Osherov and May,
2000). Thus, the mechanisms proposed here are likely
to be widely conserved.
Distinct Roles for NPFs during Endocytosis
Our biochemical studies showed that Las17p and
Myo5p/Vrp1p have significantly higher NPF activity
than Pan1p and Abp1p, suggesting that the different
NPFs may have distinct roles during endocytosis. The
difference in NPF activities may reflect the fact that
Pan1p and Abp1p lack typical WH2 (W) and Central (C)
domains, which are important for Arp2/3 complex acti-
vation by WASP (Pollard and Borisy, 2003).
Each yeast NPF is a modular protein. Our analysis of
NPF-related endocytic function therefore depended on
ing domains. NPFs with the greatest NPF activity in vitro
were the most important for endocytic function in vivo.
Importantly, we identified unique endocytic functions
for the NPFs, including those with similarly potent NPF
activities. Las17p NPF activity is required to efficiently
initiate actin assembly at endocytic sites. Myo5p/
Vrp1p NPF activity is most important during endocytic
coat inward movement. NPFs also appeared to have
could help initiate endocytic actin polymerization in the
absence of Las17p. In addition, Las17p and Myo5p/
Vrp1p NPF activities cooperated to promote endocytic
coat inward movement, which likely represents mem-
brane invagination. Because blocking invagination pre-
vents the analysis of subsequent endocytic steps, new
strategies will be needed to determine whether actin
nucleation also plays a role during vesicle scission.
In contrast to the other yeast NPFs, deletion of Abp1p
(Kaksonen et al., 2005). The Abp1p ‘‘A’’ domains, which
bind to the Arp2/3 complex, have been shown to be re-
quired for Las17p inhibition (D’Agostino and Goode,
2005). This suggests that Abp1p may functionally com-
pete with other NPFs for Arp2/3 complex binding.
Because Abp1p localization at endocytic sites is depen-
dent on F-actin, Abp1p is possibly involved in a negative
feedback loop to restrain actin nucleation after actin
assembly is initiated.
Regulation of Actin Assembly by Dynamic Changes
in Patch Composition
Yeast NPFs contain SH3 or proline-rich domains that
mediate a complex array of protein-protein interactions
(Goode and Rodal, 2001). Our results suggest that
Figure 7. A Model for How Actin Assembly Is
Regulated and Produces Force during Endo-
See Discussion for description.
The Mechanism of Actin’s Endocytic Function
dynamic changes in the protein composition of endo-
cytic sites, mediated by SH3/proline-rich domain inter-
actions, play a key role in spatio-temporal regulation of
We identified acritical mechanism for theactivation of
a WASP-family protein in which the SH3 domains of
Bzz1p (a yeast syndapin-like protein) relieve Las17p in-
hibition by the SH3 domains of Sla1p. Consistent with
this activity, Bzz1p appears at endocytic sites immedi-
ately before actin polymerization is detected. Bzz1p
also contains an N-terminal F-BAR/EFC domain that tu-
Tsujita et al., 2006). Therefore, Bzz1p is a conserved en-
docytic component that may link actin assembly to
membrane curvature. Not all SH3-containing proteins
are antagonistic with Sla1p because another SH3-con-
taining protein, Bbc1p, further inhibits Las17p in the
presence of Sla1-SH3p (Rodal et al., 2003). In addition
to Bzz1p, Bbc1p, and Sla1p, at least four other unchar-
acterized SH3-containing proteins interact with Las17p
(Madania et al., 1999; Tong et al., 2002). Analysis of their
combinatorial effects on Las17p nucleation activity
in vitro, as well as their appearance at endocytic sites
in vivo, will be required to fully elucidate the mecha-
nisms of Las17p regulation.
In contrast to Las17p, it appears that Myo5p’s NPF
activity is activated upon its recruitment to endocytic
sites. Vrp1p cooperates with actin filaments to recruit
Myo5p to cortical patches via its SH3 domain. Further-
more, Vrp1p binding to Myo5p’s SH3 domain activates
Myo5p NPF activity in vitro. The proline-rich domain of
Bbc1p also binds to Myo5p’s SH3 domain (Mochida
et al., 2002) and inhibits Myo5p/Vrp1p NPF activity. We
found that Myo5p/Vrp1p NPF activity, but not Las17p
activity, is required for the exaggerated movement of
endocytic structures in a bbc1D mutant. Because
Bbc1p and Vrp1p both bind to Myo5p’s SH3 domain,
Bbc1p possibly regulates Myo5p NPF activity by com-
peting with Vrp1p for Myo5p binding. Based on our
quantitative analysis of the relative abundance of these
proteinsincortical patches, Bbc1p and Myo5pare pres-
ent in similar amounts, whereas Vrp1p is present at
about half the concentration of these proteins. The
bbc1D mutant may result in a higher proportion of
Myo3/5p being activated by Vrp1p. Bbc1p may cooper-
ate with Abp1p to restrain actin nucleation at endocytic
files of genetic interactions, suggesting that they share
a common function that is important for efficient endo-
cytic internalization (Mochida et al., 2002).
Type I Myosin Motor Activity Is Required
for Endocytic Internalization but Not for Cortical
Previously, Myo3/5p motor activity was implicated in
cortical actin assembly in permeabilized yeast cells
(Lechler et al., 2000). Therefore, whether Myo5p’s motor
ation was not clear. We conclusively demonstrated that
the Myo5p motor and Arp2/3 complex activation activi-
ties are separable in vitro. Furthermore, actin polymeri-
zation efficiently occurred at endocytic sites in motor
mutants in vivo and motor mutant defects could not be
rescued by increasing NPF activity with the bbc1D mu-
tant. Significantly, Myo5p motor mutants blocked endo-
cytic coat inward movement, the same step blocked by
NPF mutants defective in Arp2/3 activation. Therefore,
we propose that motor and NPF activities function to-
gether to promote endocytic membrane invagination.
Plasmids and Strains
Plasmids and yeast strains used in this study are listed in Tables S1
and S2, respectively. C-terminal GFP and RFP tags were integrated
by homologous recombination, as described previously (Kaksonen
et al., 2005). Growth of Myo5-GFP myo3D, Vrp1-GFP, and wild-type
strains were indistinguishable, demonstrating that the tagged pro-
teins are functional. Myo5 mutants were generated as described
in the Supplemental Experimental Procedures. All strains were
grown in standard rich media (YPD) at 25ºC, unless otherwise
Myo5p, Vrp1p, Las17p, Sla1-SH3p, Bzz1p, Bbc1p, and Abp1p were
overexpressed and purified by previously described procedures
(Rodal et al., 2003), with the following exceptions. Full-length
Myo5p was purified in a buffer containing 20 mM HEPES (pH 7.5),
1 mM EGTA, 5 mM MgCl2, 1 mM ATP, and 100 mM KCl. Myc-tagged
Myo5p bound to anti-myc-coated beads was washed with buffer
containing 1 M KCl to remove contaminating Las17p. Cmd1p,
Myo5p’s light chain, copurified with Myo5p, as detected by Western
blotting (data not shown). Myo5p was stored at 280ºC in 20 mM
HEPES (pH 7.5), 1 mM EGTA, 0.1 mM MgCl2, 0.1 mM ATP, 50 mM
KCl, and 5% glycerol. To purify Vrp1p, a Mono Q column (Amersham
Pharmacia) was used to remove contaminating Las17p and actin.
HA-tagged Arp2/3 complex, TAP-tagged Pan1p, rabbit skeletal
muscle actin, and yeast actin were purified as described previously
(Martin et al., 2005; Toshima et al., 2005). Protein concentrations
were determined by using SYPRO dye (Molecular Probes) with
BSA as a standard.
MgCl2(final concentration). Pyrene-actin assembly was monitored
with a Fluoromax 2 fluorometer (Jobin-Yvon Horiba) or a Victor 3
plate reader (Perkin Elmer) with 355 nm and 410 nm filters. Barbed
end concentrations nucleated by the Arp2/3 complex were calcu-
lated with the rate of actin polymerization at 50% polymerization
and a k+of 8.7 mM21s21, as described previously (Rodal et al.,
2003). Rabbit skeletal muscle actin was used unless otherwise
The Lucifer yellow uptake assay (Duncan et al., 2001) and live-cell
imaging (Kaksonen et al., 2005) were performed as described previ-
ously. All imaging studies were performed at approximately 25ºC.
Image analysis was performed with ImageJ (http://rsb.info.nih.
gov/ij/) (Kaksonen et al., 2003).
Myosin Gliding Assay
Myo5p (200 nM) was bound to a nitrocellulose-coated coverslip for
4 min, and the coverslip was subsequently blocked with 0.5% BSA
for 1 min. Rhodamine-phalloidin (Molecular Probes) stabilized actin
filaments (160 nM) were bound to Myo5p on the coverslip in the ab-
sence of ATP. The motility reaction was initiated by adding ATP
(2 mM) and was immediately visualized by using an Olympus IX-71
fluorescence microscope, with a 1003/NA 1.4 objective, and an
Orca-ER camera. The reaction buffer contained 25 mM Imidazole
(pH 7.0), 25 mM KCl, 4 mM MgCl2, 1 mM EGTA, 1 mM DTT,
18 mg/ml catalase, 0.1 mg/ml glucose oxidase, 3 mg/ml glucose,
and 2 mM ATP.
Supplemental Data include additional Experimental Procedures,
figures, tables, and movies and are available at http://www.
We thank A. Rodal, J. Toshima, M. Sekiya-Kawasaki, S. Carroll,
C. Toret, M.Kaksonen, andL. Lee forproviding strains andplasmids
used in this study. We also thank J. Toshima for purified Pan1p and
yeast actin. Finally, we thank M. Welch, L. Lee, V. Okreglak, C. Toret,
C.X. Zhang, I. LeBlanc, and M. Kaksonen for critically reading the
manuscript. This work was supported by National Institutes of
Health grants GM42759 and GM50399 to D.G.D. and National Insti-
tutes of Health Shared Instrumentation Grant S10-RR019290 to
Received: November 29, 2005
Revised: May 3, 2006
Accepted: May 16, 2006
Published: July 10, 2006
Anderson, B.L., Boldogh, I., Evangelista, M., Boone, C., Greene,
L.A., and Pon, L.A. (1998). The Src homology domain 3 (SH3) of
a yeast type I myosin, Myo5p, binds to verprolin and is required
for targeting to sites of actin polarization. J. Cell Biol. 141, 1357–
Baggett, J.J., D’Aquino, K.E., and Wendland, B. (2003). The Sla2p
talin domain plays a role in endocytosis in Saccharomyces cerevi-
siae. Genetics 165, 1661–1674.
Baird, G.S., Zacharias, D.A., and Tsien, R.Y. (2000). Biochemistry,
mutagenesis, and oligomerization of DsRed, a red fluorescent pro-
tein from coral. Proc. Natl. Acad. Sci. USA 97, 11984–11989.
Benesch, S., Polo, S., Lai, F.P., Anderson, K.I., Stradal, T.E., Weh-
land, J., and Rottner, K. (2005). N-WASP deficiency impairs EGF in-
ternalization and actin assembly at clathrin-coated pits. J. Cell Sci.
D’Agostino, J.L., and Goode, B.L. (2005). Dissection of Arp2/3 com-
plex actin nucleation mechanism and distinct roles for its nucle-
ation-promoting factors in Saccharomyces cerevisiae. Genetics
Duncan, M.C., Cope, M.J., Goode, B.L., Wendland, B., and Drubin,
D.G. (2001). Yeast Eps15-like endocytic protein, Pan1p, activates
the Arp2/3 complex. Nat. Cell Biol. 3, 687–690.
Engqvist-Goldstein, A.E., and Drubin, D.G. (2003). Actin assembly
and endocytosis: from yeast to mammals. Annu. Rev. Cell Dev.
Biol. 19, 287–332.
Evangelista, M., Klebl, B.M., Tong, A.H., Webb, B.A., Leeuw,T., Leb-
erer,E.,Whiteway,M.,Thomas, D.Y., andBoone, C.(2000). Arole for
myosin-I in actin assembly through interactions with Vrp1p, Bee1p,
and the Arp2/3 complex. J. Cell Biol. 148, 353–362.
Geli, M.I., and Riezman, H.(1996).Role of type Imyosins in receptor-
mediated endocytosis in yeast. Science 272, 533–535.
Geli, M.I., Lombardi, R., Schmelzl, B., and Riezman, H. (2000). An in-
tact SH3 domain is required for myosin I-induced actin polymeriza-
tion. EMBO J. 19, 4281–4291.
Goode, B.L., and Rodal, A.A. (2001). Modular complexes that regu-
late actin assembly in budding yeast. Curr. Opin. Microbiol. 4, 703–
Goode, B.L., Rodal, A.A., Barnes, G., and Drubin, D.G. (2001). Acti-
vation of the Arp2/3 complex by the actin filament binding protein
Abp1p. J. Cell Biol. 153, 627–634.
Goodson, H.V., Anderson, B.L., Warrick, H.M., Pon, L.A., and Spu-
dich, J.A. (1996). Synthetic lethality screen identifies a novel yeast
myosin Igene(MYO5): myosin Iproteinsarerequired forpolarization
of the actin cytoskeleton. J. Cell Biol. 133, 1277–1291.
R.T., Riezman, H., and Lemmon, S.K. (2002). Scd5p and clathrin
function are important for cortical actin organization, endocytosis,
and localization of sla2p in yeast. Mol. Biol. Cell 13, 2607–2625.
Huckaba, T.M., Gay, A.C., Pantalena, L.F., Yang, H.C., and Pon, L.A.
(2004). Live cell imaging of the assembly, disassembly, and actin ca-
ble-dependent movement of endosomes and actin patches in the
budding yeast, Saccharomyces cerevisiae. J. Cell Biol. 167, 519–
Itoh,T., Erdmann, K.S., Roux, A., Habermann, B., Werner, H., and De
Camilli, P. (2005). Dynamin and the actin cytoskeleton cooperatively
regulate plasma membrane invagination by BAR and F-BAR pro-
teins. Dev. Cell 9, 791–804.
Jonsdottir, G.A., and Li, R. (2004). Dynamics of yeast Myosin I: evi-
dence for a possible role in scission of endocytic vesicles. Curr.
Biol. 14, 1604–1609.
Kaksonen, M., Sun, Y., and Drubin, D.G. (2003). A pathway for asso-
ciation of receptors, adaptors, and actin during endocytic internali-
zation. Cell 115, 475–487.
for the clathrin- and actin-mediated endocytosis machinery. Cell
Lechler, T., Shevchenko, A., and Li, R. (2000). Direct involvement of
yeast type I myosins in Cdc42-dependent actin polymerization. J.
Cell Biol. 148, 363–373.
A two-tiered mechanism by which Cdc42 controls the localization
and activation of an Arp2/3-activating motor complex in yeast. J.
Cell Biol. 155, 261–270.
Madania, A., Dumoulin, P., Grava, S., Kitamoto, H., Scharer-Brod-
romyces cerevisiae homologue of human Wiskott-Aldrich syndrome
protein Las17p interacts with the Arp2/3 complex. Mol. Biol. Cell 10,
Martin, A.C., Xu, X.P., Rouiller,I., Kaksonen, M., Sun, Y.,Belmont, L.,
of Arp2 and Arp3 nucleotide-binding pocket mutations on Arp2/3
complex function. J. Cell Biol. 168, 315–328.
single sites of endocytosis. Trends Cell Biol. 14, 352–358.
Merrifield, C.J., Feldman, M.E., Wan, L., and Almers, W. (2002). Im-
aging actin and dynamin recruitment during invagination of single
clathrin-coated pits. Nat. Cell Biol. 4, 691–698.
Merrifield, C.J., Qualmann, B., Kessels, M.M., and Almers, W. (2004).
Neural Wiskott Aldrich Syndrome Protein (N-WASP) and the Arp2/3
complex are recruited to sites of clathrin-mediated endocytosis in
cultured fibroblasts. Eur. J. Cell Biol. 83, 13–18.
Merrifield, C.J., Perrais, D., and Zenisek, D. (2005). Coupling be-
tween clathrin-coated-pit invagination, cortactin recruitment, and
membrane scission observed in live cells. Cell 121, 593–606.
Mochida, J., Yamamoto, T., Fujimura-Kamada, K., and Tanaka, K.
(2002). The novel adaptor protein, Mti1p, and Vrp1p, a homolog of
Wiskott-Aldrich syndrome protein-interacting protein (WIP), may
antagonistically regulate type I myosins in Saccharomyces cerevi-
siae. Genetics 160, 923–934.
Newpher, T.M., Smith, R.P., Lemmon, V., and Lemmon, S.K. (2005).
In vivo dynamics of clathrin and its adaptor-dependent recruitment
to the actin-based endocytic machinery in yeast. Dev. Cell 9, 87–98.
Cell Motil. Cytoskeleton 47, 163–173.
Pollard, T.D., and Borisy, G.G. (2003). Cellular motility driven by as-
sembly and disassembly of actin filaments. Cell 112, 453–465.
Rodal, A.A., Manning, A.L., Goode, B.L., and Drubin, D.G. (2003).
Negative regulation of yeast WASp by two SH3 domain-containing
proteins. Curr. Biol. 13, 1000–1008.
Rohatgi, R., Ma, L., Miki, H., Lopez, M., Kirchhausen, T., Takenawa,
T., and Kirschner, M.W. (1999). The interaction between N-WASP
and the Arp2/3 complex links Cdc42-dependent signals to actin as-
sembly. Cell 97, 221–231.
Sekiya-Kawasaki, M., Groen, A.C., Cope, M.J., Kaksonen, M., Wat-
son, H.A., Zhang, C., Shokat, K.M., Wendland, B., McDonald, K.L.,
McCaffery, J.M., and Drubin, D.G. (2003). Dynamic phosphoregula-
tion of the cortical actin cytoskeleton and endocytic machinery re-
vealed by real-time chemical genetic analysis. J. Cell Biol. 162,
Sirotkin, V., Beltzner, C.C., Marchand, J.B., and Pollard, T.D. (2005).
Interactions of WASp, myosin-I, and verprolin with Arp2/3 complex
The Mechanism of Actin’s Endocytic Function
during actin patch assembly in fission yeast. J. Cell Biol. 170, 637– Download full-text
Soulard, A., Lechler, T., Spiridonov, V., Shevchenko, A., Li, R., and
Winsor, B. (2002). Saccharomyces cerevisiae Bzz1p is implicated
with type I myosins in actin patch polarization and is able to recruit
actin-polymerizing machinery in vitro. Mol. Cell. Biol. 22, 7889–7906.
Sun, Y., Kaksonen, M., Madden, D.T., Schekman, R., and Drubin,
D.G. (2005). Interaction of Sla2p’s ANTH domain with PtdIns(4,5)P2
is important for actin-dependent endocytic internalization. Mol.
Biol. Cell 16, 717–730.
Tong, A.H., Drees, B., Nardelli, G., Bader, G.D., Brannetti, B., Cas-
tagnoli, L., Evangelista, M., Ferracuti, S., Nelson, B., Paoluzi, S.,
et al. (2002). A combined experimental and computational strategy
to define protein interaction networks for peptide recognition mod-
ules. Science 295, 321–324.
Toshima, J., Toshima, J.Y., Martin, A.C., and Drubin, D.G. (2005).
Phosphoregulation of Arp2/3-dependent actin assembly during re-
ceptor-mediated endocytosis. Nat. Cell Biol. 7, 246–254.
Tsujita, K., Suetsugu, S., Sasaki, N., Furutani, M., Oikawa, T., and
Takenawa, T. (2006). Coordination between the actin cytoskeleton
and membrane deformation by a novel membrane tubulation do-
main of PCH proteins is involved in endocytosis. J. Cell Biol. 172,
Wendland, B., and Emr, S.D. (1998). Pan1p, yeast eps15, functions
as a multivalent adaptor that coordinates protein-protein interac-
tions essential for endocytosis. J. Cell Biol. 141, 71–84.
Wu, C., Lee, S.F., Furmaniak-Kazmierczak, E., Cote, G.P., Thomas,
D.Y., and Leberer, E. (1996). Activation of myosin-I by members of
the Ste20p protein kinase family. J. Biol. Chem. 271, 31787–31790.
Yarar, D., Waterman-Storer, C.M., and Schmid, S.L. (2005). A dy-
namic actin cytoskeleton functions at multiple stages of clathrin-
mediated endocytosis. Mol. Biol. Cell 16, 964–975.
Zot, H.G., Doberstein, S.K., and Pollard, T.D. (1992). Myosin-I moves
actin filaments on a phospholipid substrate: implications for mem-
brane targeting. J. Cell Biol. 116, 367–376.