adf/cofilin binds phosphoinositides in a multivalent

Biophysical Journal Volume 98 May 2010 2327–2336 2327

brought to you by COREView metadata, citation and similar papers at core.ac.uk

provided by Elsevier - Publisher Connector

ADF/Cofilin Binds Phosphoinositides in a Multivalent Mannerto Act as a PIP2-Density Sensor

Hongxia Zhao,* Markku Hakala, and Pekka LappalainenInstitute of Biotechnology, University of Helsinki, Helsinki, Finland

ABSTRACT Actin-depolymerizing-factor (ADF)/cofilins have emerged as key regulators of cytoskeletal dynamics in cellmotility, morphogenesis, endocytosis, and cytokinesis. The activities of ADF/cofilins are regulated by membrane phospholipidPI(4,5)P2 in vitro and in cells, but the mechanism of the ADF/cofilin-PI(4,5)P2 interaction has remained controversial. Recentstudies suggested that ADF/cofilins interact with PI(4,5)P2 through a specific binding pocket, and that this interaction is depen-dent on pH. Here, we combined systematic mutagenesis with biochemical and spectroscopic methods to elucidate the phosphoi-nositide-binding mechanism of ADF/cofilins. Our analysis revealed that cofilin does not harbor a specific PI(4,5)P2-bindingpocket, but instead interacts with PI(4,5)P2 through a large, positively charged surface of the molecule. Cofilin interacts simul-taneously with multiple PI(4,5)P2 headgroups in a cooperative manner. Consequently, interactions of cofilin with membranesand actin exhibit sharp sensitivity to PI(4,5)P2 density. Finally, we show that cofilin binding to PI(4,5)P2 is not sensitive to changesin the pH at physiological salt concentration, although the PI(4,5)P2-clustering activity of cofilin is moderately inhibited at elevatedpH. Collectively, our data demonstrate that ADF/cofilins bind PI(4,5)P2 headgroups through a multivalent, cooperative mecha-nism, and suggest that the actin filament disassembly activity of ADF/cofilin can be accurately regulated by small changes inthe PI(4,5)P2 density at cellular membranes.

INTRODUCTION

The actin cytoskeleton underlies multiple cell biological

processes, such as motility, cytokinesis, endocytosis, phago-

cytosis, and intracellular signal transduction. Many human

diseases, including cancer and neuronal and cardiovascular

diseases, exhibit abnormalities in actin-dependent processes

such as cell division and adhesion, as well as in the motile

and invasive properties of a cell (1,2). The structure and

dynamics of the actin cytoskeleton are regulated both

spatially and temporally by a wide array of actin-binding

proteins (3,4). ADF/cofilins are among the most central

proteins that promote actin dynamics during cell motility

(5,6), cytokinesis (7), and endocytosis (8). These small

(molecular mass ¼ 15–20 kDa) proteins are present in all

eukaryotes and they bind both filamentous and monomeric

actin with high affinity, with a preference for ADP-actin

(9–11). ADF/cofilins alter the conformation of actin fila-

ments (12) and increase actin dynamics by accelerating the

dissociation of actin monomers from the filament pointed

end (13) and by severing actin filaments (14). In addition,

ADF/cofilins accelerate the dissociation of phosphate from

ADP-Pi filaments and promote the debranching of the

Arp2/3-nucleated actin filament network (15,16).

The activities of ADF/cofilins are tightly regulated by

phosphorylation, acidity, and interactions with other proteins

and phosphoinositides (17). Phosphoinositides are multi-

functional lipids that modulate many cellular events, such

Submitted November 22, 2009, and accepted for publication January 22,2010.

*Correspondence: [email protected]

Editor: Roberto Dominguez.

� 2010 by the Biophysical Society

0006-3495/10/05/2327/10 $2.00

as membrane traffic, intracellular signaling, cytoskeleton

organization, and apoptosis (18,19). ADF/cofilins bind

PI(4,5)P2, PI(3,4)P2, and PI(3,4,5)P3 with relatively high

affinity, but they do not interact with IP3, the inositol head-

group of PI(4,5)P2 (20,21). The actin-binding and F-actin

disassembly activities of ADF/cofilins are efficiently in-

hibited by PI(4,5)P2 both in vitro and in cells (20,22,23).

The PI(4,5)P2-binding of ADF/cofilins has also been re-

ported to be sensitive to pH (24).

Despite extensive studies, the molecular mechanism of the

PI(4,5)P2-interaction by ADF/cofilins has remained contro-

versial (21,24–28). Two peptides corresponding to residues

D9-T25 and P26-V36 at the N-terminus of chick cofilin

were shown to decrease the ability of PI(4,5)P2 to abrogate

cofilin-PI(4,5)P2 interaction, suggesting that these peptides

bind PI(4,5)P2 (28). Mutagenesis studies on budding yeast

and Acanthamoeba ADF/cofilins indicated that several

different regions of the protein interact with phosphoinositi-

des (21,25). However, in marked contrast to those works,

a recent NMR study of vertebrate cofilin suggested that the

protein binds PI(4,5)P2 through a specific binding pocket

and may also interact with the acyl chains of the lipids

(27). Thus, further studies were necessary to elucidate the

molecular mechanism of ADF/cofilin-PI(4,5)P2 interaction

and to reveal whether different ADF/cofilins bind phosphoi-

nositides through conserved or distinct mechanisms. In

addition, the mechanism by which ADF/cofilins can sense

relatively small changes in PI(4,5)P2 concentration at the

plasma membrane in cells, and why these proteins do not

interact with IP3, remained unknown.

doi: 10.1016/j.bpj.2010.01.046

https://core.ac.uk/display/82744144?utm_source=pdf&utm_medium=banner&utm_campaign=pdf-decoration-v1

mailto:[email protected]

2328 Zhao et al.

MATERIALS AND METHODS

Materials

Acrylamide was obtained from Sigma, and 1,6-diphenyl-1,3,5-hexatriene

(DPH) was obtained from Invitrogen. Bodipy-TMR-PI(4,5)P2 was purchased

from Echelon (Salt Lake City, Utah). 1-Palmitoyl-2-oleoyl-sn-glycero-3-phos-

phatidylcholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidyleth-

anolamine (POPE), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine

(POPS), 1-palmitoyl-2-(6,7-dibromo-stearoyl) phosphatidylcholine ((6,7)-

Br2- PC), 1-palmitoyl-2-(9,10-dibromostearoyl) phosphatidyl-choline ((9,

10)-Br2-PC), 1-palmitoyl-2-(11,12-dibromostearoyl) phosphatidylcholine

((11,12)-Br2-PC), and L-a-phosphatidylinositol-4,5-bisphosphate were pur-

chased from Avanti Polar Lipids (Alabaster, AL). Pyrenyl-actin was obtained

from Cytoskeleton Inc. (Denver, CO).

Protein purification

Wild-type and mutant mouse cofilin-1 were expressed as glutathione-S-

transferase (GST)-fusion proteins and were enriched using glutathione

agarose beads as described previously (29). GST was cleaved with thrombin

(10 U/mL), and cofilins were purified with a Superdex-75 gel filtration

column (Pharmacia Biotech).

Circular dichroism measurements

Ultraviolet circular dichroism spectra from 250 to 190 nm were recorded with

a JASCO J-810 instrument. The spectra were measured using a 0.1-cm path

length quartz cell. The cofilin-1 concentration was 10 mM in 20 mM Hepes,

pH 7.5. The data are presented as the averages of five scans.

Vesicle cosedimentation assay

Lipids in desired concentrations were mixed and then dried under a stream of

nitrogen, and the lipid residue was subsequently maintained under reduced

pressure for at least 4 h. The dry lipids were then hydrated in 20 mM Hepes,

pH 7.5, 100 mM NaCl to obtain multilamellar vesicles. To obtain unilamel-

lar vesicles, vesicles were extruded through a polycarbonate filter (100-nm

pore size) using a mini extruder (Avanti Polar Lipids). Proteins and lipo-

somes were incubated at room temperature for 10 min and centrifuged at

100,000 rpm with a Beckman rotor (TLA 100) for 1 h. The final concentra-

tions of cofilin and liposomes were 2 mM and 500 mM, respectively, in

20 mM Hepes, pH 7.5 buffer (with or without 100 mM NaCl). Equal propor-

tions of supernatants and pellets were loaded for sodium dodecyl sulfate

polyacrylamide gel electrophoresis, and after electrophoresis the gels were

stained with Coomassie Blue. The intensities of the cofilin bands were quan-

tified with the use of the Quantity One program (Bio-Rad).

F-actin binding assays

Pyrene-labeled actin (20 mM) was polymerized for 30 min in F-buffer (2 mM

MgCl2, 1 mM ATP, and 100 mM KCl). After polymerization the actin fil-

ments were diluted at 1 mM in 20 mM Hepes, 100 mM NaCl, pH 7.5, for

fluorescence measurements. Liposomes with different PI(4,5)P2-densities

(buffer alone was used as control) were incubated with cofilin for 10 min

before they were added to the actin filament solution. The excitation and

emission wavelengths for pyrene were 365 and 406 nm, respectively.

Both the excitation and emission bandwidths were set at 10 nm. An F-actin

cosedimentation assay was carried out using rabbit muscle actin as described

previously (21).

Fluorescence spectroscopy experiments

All fluorescence measurements were performed in quartz cuvettes (3-mm

path length) with a Perkin-Elmer LS 55 spectrometer. Emission and excita-

tion band passes were set at 10 nm. Fluorescence anisotropy for DPH was

Biophysical Journal 98(10) 2327–2336

measured with excitation at 360 nm and emission at 450 nm. Tryptophan

(Trp) fluorescence was excited at 290 nm, and the emission spectra were re-

corded by averaging five spectra from 310 to 420 nm upon addition of

different concentrations of liposomes. Spectra were corrected for the contri-

bution of light scattering in the presence of vesicles. Bodipy-TMR-PI(4,5)P2

fluorescence was excited at 547 nm and the emission spectra were recorded

from 555 to 600 nm in the presence of different concentrations of cofilin-1.

Data were analyzed as described by Ward (30). The apparent dissociation

constant, Kd, was calculated by the following equation:

DF ¼ DFmax � ½PIP2�TKd þ ½PIP2�T

where DF is the fluorescence quenching at a given PIP2 concentration,

DFmax is the total fluorescence quenching of cofilin-1 saturated with PIP2,

and [PIP2]T is the total accessible concentration of PIP2. DFmax is estimated

by curve-fitting the binding data using the Hyperbol.fit program in Microcal

Origin. The stoichiometry of binding, p, was derived according to Stinson

and Holbrook (31).

For the acrylamide quenching assay, Trp was excited at 295 nm and the

data were analyzed according to the Stern-Volmer equation (32), F0/F ¼1 þ Ksv[Q], where F0 and F represent the fluorescence intensities in the

absence and the presence of the quencher (Q), respectively, and Ksv is the

Stern-Volmer quenching constant, which is a measure of the accessibility

of Trp to acrylamide. Collisional quenching of Trp by brominated phospho-

lipids (Br2-PCs) was introduced to assess the localization of this residue in

the bilayer (33).

RESULTS

Binding of cofilin to membranes is dependent onthe PI(4,5)P2-density

Cofilin-1 is the most abundant and ubiquitously expressed

ADF/cofilin in mammals, and the corresponding isoform

has been widely used to investigate interactions between

ADF/cofilins and PI(4,5)P2 (23,27). In this work, vesicle co-

sedimentation and fluorometric assays were applied to assess

the binding of mouse cofilin-1 to lipid membranes. The only

Trp of mouse cofilin-1 (Trp-104) locates at the hydrophobic

core of the protein (34). Our data show that upon PI(4,5)P2-

binding, this residue is exposed to a more hydrophilic envi-

ronment, which causes its fluorescence emission to decrease

(Fig. S1 in the Supporting Material). Quenching of Trp fluo-

rescence thus provides a tool for examining the cofilin-1-

PI(4,5)P2 interaction by a fluorometric method. Both vesicle

cosedimentation and Trp fluorescence assays showed weak

binding of cofilin-1 to vesicles containing only PC/PE/PS

(60:20:20). However, in the presence of PI(4,5)P2, the

binding of cofilin-1 to membranes was significantly

enhanced (Fig. 1, A and B). The affinity (Kd) of cofilin for

10% PI(4,5)P2 in the membrane was ~4 mM, as calculated

from Trp fluorescence assay (see Fig. 3 A).

Cosedimentation and fluorometric assays carried out using

vesicles with different PI(4,5)P2 densities (0–20%) revealed

that cofilin binding is sensitive to the spatial PI(4,5)P2

density in the membrane (Fig. 1, A and B). The PI(4,5)P2

density-dependent binding curve showed a typical sigmoidal

shape indicative of cooperative binding. The half-maximal

binding in both vesicle cosedimentation and fluorometric

A

B

C

FIGURE 1 Cofilin-1 is a PI(4,5)P2-density sensor. (A) A vesicle cosedi-

mentation assay shows that the binding of cofilin-1 to membranes is depen-

dent on the PI(4,5)P2 density. The sigmoidal binding isotherm demonstrates

that cofilin-1 binds PI(4,5)P2 cooperatively when the density of PI(4,5)P2 in

the membrane is augmented. POPE and POPS were included at 20%. PIP2

was included at 0%, 2%, 4%, 6%, 8%, 10%, and 20%, and the POPC

concentration was varied accordingly. For 0% PIP2, vesicles composed of

POPC/POPE/POPS (6:2:2) were used. The final concentrations of cofilin-

1 and liposomes were 2 mM and 500 mM, respectively, in 20 mM Hepes,

pH 7.5 buffer. Each data point represents the mean of triplicate measure-

ments, with the error bars indicating 5SD. (B) Membrane binding of cofi-

lin-1 was also examined by quenching of Trp fluorescence. The percentage

of quenching shows a sigmoidal PI(4,5)P2-binding isotherm, revealing that

cofilin-1 binds PI(4,5)P2 cooperatively and that the interaction is dependent

on PI(4,5)P2 density. The cofilin-1 concentration was 1 mM in 20 mM

Hepes, 100 mM NaCl, pH 7.5, and total lipid concentration was 140 mM.

Lipid composition was as described in panel A. (C) PI(4,5)P2 inhibits the

binding of cofilin-1 to actin filaments, and inhibition is PI(4,5)P2-density

dependent. Pyrene-labeled actin (20 mM) was polymerized for 30 min in

F-buffer (2 mM MgCl2, 1 mM ATP, and 100 mM KCl). The actin filaments

Interaction of ADF/Cofilin with PIP2 2329

binding assays was observed at ~6% PI(4,5)P2 (Fig. 1, Aand B). Of note, the F-actin binding activity of cofilin-1, as

detected by steady-state quenching of pyrenyl-actin fluores-

cence by cofilin-1, also displayed a similar dependence on

PI(4,5)P2 density (Fig. 1 C). The inhibition curve was

sigmoidal, showing a sharp threshold between 4% and 9%

PI(4,5)P2. Thus, interaction of cofilin-1 with membranes

and inhibition of its F-actin binding activity are dependent

on the spatial PI(4,5)P2 density in the membrane.

Cofilin clusters PI(4,5)P2 molecules at themembrane surface

Because the interaction of cofilin with membranes is depen-

dent on the PI(4,5)P2 density, we next investigated whether

cofilin could simultaneously bind multiple PI(4,5)P2 mole-

cules in the membrane. For this purpose, PI(4,5)P2 labeled

with bodipy-TMR at the sn-1 position of the acyl chains

was applied for a fluorometric PI(4,5)P2-clustering assay.

This fluorescent probe has highly superimposable absorption

and emission spectra, and exhibits self-quenching properties

when two or more molecules are brought together in close

proximity. We experimentally tested whether cofilin-1 could

induce self-quenching of bodipy-TMR-PI(4,5)P2 when the

fluorescent lipid was incorporated into the vesicles at

a concentration below the level at which self-quenching

occurs without proteins (35). Of importance, strong quench-

ing of bodipy-TMR-PI(4,5)P2 fluorescence was detected

upon cofilin binding, indicating an increased proximity of

neighboring bodipy-TMR-PI(4,5)P2 molecules sequestered

by cofilin-1 (Fig. 2 A). The degree of bodipy-TMR-

PI(4,5)P2 fluorescence quenching was similar to that previ-

ously reported for I-BAR domains, which are efficient

PI(4,5)P2-clustering molecules (36). The stoichiometry of

the binding derived from the graph indicates that ~2,6

PI(4,5)P2 molecules are bound per one cofilin molecule.

Therefore, this assay demonstrates that cofilin simultaneously

binds multiple PI(4,5)P2 molecules and induces their clus-

tering in the membrane.

Cofilin-1 interacts with PI(4,5)P2 throughelectrostatic interactions and does not insert intothe lipid bilayer

When the PI(4,5)P2-clustering assay was carried out at

different NaCl concentrations, we noticed that the lateral

sequestration of PI(4,5)P2 decreased as the salt concentration

were diluted to 1 mM in 20 mM Hepes, 100 mM NaCl, pH 7.5, for fluores-

cence measurements. Liposomes with different PI(4,5)P2 densities (buffer

alone was used as control) were incubated with cofilin-1 for 10 min before

addition of the actin filament solution. Pyrene fluorescence was quenched

upon cofilin-1 binding to actin filaments. PI(4,5)P2 containing lipsomes in-

hibited the binding of cofilin-1 to actin filaments, as detected by the

decreased quenching of pyrene fluorescence. The final cofilin-1 concentra-

tion in the assay was 0.2 mM.


A B

C D

FIGURE 2 Cofilin simultaneously

binds multiple PI(4,5)P2 molecules on

the membrane surface and does not

insert into the lipid bilayer. (A) Quench-

ing of bodipy-TMR-PI(4,5)P2 demon-

strates that cofilin-1 sequesters

PI(4,5)P2 on the membrane, and this

activity is salt-sensitive. The lipid com-

position in this assay was POPC/POPE/

POPS/PIP2/bodipy-TMR-PI(4,5)P2 (50:

20:20:9.5:0.5) and the total lipid

concentration was 40 mM. Salt inhibits

PI(4,5)P2 clustering by cofilin-1, sug-

gesting electrostatic interactions be-

tween cofilin-1 and PI(4,5)P2. (B) A

vesicle cosedimentation assay reveals

that the cofilin-1-PI(4,5)P2 interaction

is inhibited in the presence of high

salt, indicative of electrostatic interac-

tions between cofilin-1 and PI(4,5)P2.

The assay was carried out under the

same conditions (excluding the varia-

tion in the salt concentration) as in

Fig. 1 A. Each data point represents

the mean of three independent assays,

with the error bars indicating 5SD. A

t-test was used to analyze p-values

(*** p < 0.001, ** p < 0.01). (C) Cofi-

lin-1 does not display detectable effects

on DPH anisotropy, suggesting that co-

filin-1 does not insert into the lipid

bilayer upon membrane binding. The I-BAR domain of missing-in-metastasis protein (36) was used as a positive control in this assay. The lipid composition

in the assay was POPC/POPE/POPS/PIP2 (5:2:2:1, DPH is incorporated into liposomes at a molar ratio of 1:500) and the total lipid concentration was 40 mM.

The assay was carried out in 20 mM Hepes, 100 mM NaCl, pH 7.5. (D) Trp fluorescence of cofilin-1 is not affected by lipids that are brominated at different

positions along the acyl chains, indicating that cofilin-1 does not interact with the acyl chains of the lipids. The lipid composition in this assay was POPC/

POPE/POPS/PIP2/brominated phosphatidylcholine (2:2:2:1:3). The cofilin-1 concentration was 1 mM.

2330 Zhao et al.

was increased (Fig. 2 A). We further examined the salt

sensitivity of the cofilin-PI(4,5)P2 interaction by vesicle co-

sedimentation and fluorometric binding assays. The results

revealed that the cofilin-PI(4,5)P2 interaction is indeed salt-

sensitive (Figs. 2, A and B, and 3 A). These data indicate

that the interaction between cofilin and PI(4,5)P2-rich

membranes is mediated by electrostatic interactions.

However, a previous study suggested that cofilin may

penetrate into the lipid bilayer upon PI(4,5)P2 binding (27).

To examine whether cofilin-1 is capable of inserting into

the membrane bilayer, we monitored the steady-state fluores-

cence anisotropy of a membrane probe, DPH. DPH locates at

the hydrophobic core of a lipid bilayer without affecting the

physical properties of the membranes. It can thus be applied

to monitor changes in the rotational diffusion of acyl chains

in the membrane interior (37). In contrast to the I-BAR

domain of missing-in-metastasis protein, which inserts into

the lipid bilayer (36), cofilin-1 did not induce detectable

changes in the DPH anisotropy (Fig. 2 C). We further inves-

tigated membrane insertion of cofilin-1 by quenching Trp

fluorescence using lipids that are brominated at different

positions along the lipid acyl chains and can thus be used

to estimate the depth of Trp insertion into the bilayer (36).


Consistent with the DPH anisotropy data, no further fluores-

cence quenching of Trp-104 was observed by means of the

brominated lipids (Fig. 2 D). Together, these data provide

evidence that cofilin-1 binds solely PI(4,5)P2 headgroups

at the membrane surface without inserting into the hydro-

phobic core of the lipid bilayer.

The microenvironment of Trp-104 of cofilin-1 became

more hydrophilic upon PI(4,5)P2 binding because its fluores-

cence was significantly quenched (Fig. 3 A). We further

examined the change in the microenvironment of Trp-104

due to PI(4,5)P2 binding by performing assays with an

aqueous quencher acrylamide. Collisional quenching of

Trp fluorescence by acrylamide provides information con-

cerning the accessibility of Trp to this quencher (38).

Maximum quenching occurs when Trp is freely accessible

to the quencher. However, when Trp is protected by ligand

binding, fluorescence quenching by acrylamide is expected

to decrease. The Trp fluorescence intensity was quenched

by acrylamide in a concentration-dependent manner in the

absence and presence of liposomes (Fig. 3 B). Of interest,

the value for Ksv was significantly smaller in the presence

of liposomes (Ksv¼ 0.69 vs. Ksv¼ 2.3 in the absence of lipo-

somes), suggesting that Trp-104 was protected by membrane

A

B

FIGURE 3 Trp fluorescence assay shows that cofilin-1 binds PI(4,5)P2,

and salt inhibits the interaction. (A) PI(4,5)P2 binding quenches the Trp fluo-

rescence of cofilin-1, suggesting that the environment of Trp-104 is altered

due to the binding. Salt inhibits the cofilin-1-PI(4,5)P2 interaction, as de-

tected by the decreased quenching of Trp fluorescence. The final concentra-

tion of cofilin-1 in this assay was 1 mM and the lipid composition was POPC/

POPE/POPS/PIP2 (5:2:2:1). (B) Stern-Volmer plots for the quenching of

cofilin-1 Trp fluorescence by acrylamide in buffer and in the presence of

liposomes. Quenching of Trp fluorescence by acrylamide was diminished

in the presence of liposomes, indicating that Trp-104 is protected by lipid

binding. The final concentrations of cofilin-1 and liposomes were 1 mM

and 140 mM, respectively, in 20 mM Hepes, 100 mM NaCl, pH 7.5.


binding and became less accessible to the aqueous quencher.

Together, these data suggest that Trp-104 is located at the

bilayer interface close to the polar PI(4,5)P2 headgroups

when cofilin-1 binds to PI(4,5)P2-rich membranes.

Identification of the PI(4,5)P2-binding site ofcofilin-1

Several techniques, including NMR and native gel electro-

phoresis, have been applied to map the PI(4,5)P2-binding

site of ADF/cofilins. However, both methods have technical

limitations, which may explain the discrepancies in the

results. The NMR chemical shift experiments were carried

out with water-soluble di-C8 forms of PI(4,5)P2, whereas

the native gel electrophoresis assays were performed under

nonphysiological ionic conditions (21,27). To unambigu-

ously identify the PI(4,5)P2-binding site of ADF/cofilins,

we carried out a systematic mutagenesis of mouse cofilin-1

combined with cosedimentation and fluorometric PI(4,5)P2-

clustering assays. Because our analysis demonstrated that the

interaction between cofilin-1 and PI(4,5)P2 is mainly electro-

static, we mutated nine positively charged clusters of mouse

cofilin-1 to alanines and glutamates, and all of the mutants

were properly folded (Fig. 4 and Fig. S2). The data revealed

that the positively charged clusters K95, K96; K112, K114;

and K125, K126, K127 of cofilin-1 are most critical for

PI(4,5)P2-binding (Fig. 5, A and C), because substituting

these residues with either alanines or glutamates resulted in

statistically significant defects in the vesicle cosedimentation

assay (Fig. 5, A and C). In addition, residues K132, H133

and K144, R146 appear also to be located at the binding

interface, because substitution of these residues by gluta-

mates significantly reduced binding to PI(4,5)P2-rich vesi-

cles (Fig. 5 C). However, substitution of K132, H133 and

K144, R146 clusters by alanines resulted in only small

defects in the vesicle cosedimentation assay, and these

were not statistically significant. This suggests that K132,

H133 and K144, R146 make only a minor contribution to

PI(4,5)P2 binding (Fig. 5 A). Of importance, the bodipy-

TMR-PI(4,5)P2 quenching assay demonstrated that the

same residues that have a central role in PIP2 binding also

play an important role in PI(4,5)P2 clustering (Fig. 5, Band D). Together, these data show that the PI(4,5)P2-binding

site of mouse cofilin-1 is a large, positively charged surface

located on one face of the molecule (Fig. 6 A).

Modulation of cofilin-PI(4,5)P2 interaction by pH

It was recently reported that phosphoinositide binding of

ADF/cofilin is pH-sensitive, with diminished binding at

a high pH. His-133, which is located at the previously

proposed PIP2-binding pocket of cofilin, was shown to

play a central role in the pH sensitivity of phosphoinositide

binding by cofilin (24). Because our data revealed that

ADF/cofilins do not contain a specific phosphoinositide-

binding pocket, but instead interact with multiple PI(4,5)P2

headgroups through a large, positively charged interface

(see above), we reexamined the pH dependency of phosphoi-

nositide binding by ADF/cofilins. Both the Trp fluorescence

and vesicle cosedimentation assays revealed that at physio-

logical ionic conditions, PI(4,5)P2 binding of wild-type cofi-

lin or a cofilin mutant (K132A, H133A) were not affected by

changes in the pH (Fig. 7, A and B, and Fig. S3). However, in

the absence of NaCl, PI(4,5)P2 binding was pH-sensitive for

wild-type cofilin and to a somewhat lesser extent for the

K132A, H133A mutant (Fig. 7, C and D). Although no pH

sensitivity was detected for cofilin binding to PI(4,5)P2 by

Trp fluorescence quenching or vesicle cosedimentation

assays at physiological salt, the PI(4,5)P2-clustering activity

of wild-type cofilin displayed a clear defect at elevated pH

(Fig. 8 A). At identical conditions, the K132A, H133A

mutant showed no pH dependency in its ability to cluster

PI(4,5)P2 (Fig. 8 B), suggesting that His-133 indeed contrib-

utes to the pH sensitivity of the PI(4,5)P2-clustering activity

of cofilin-1.


FIGURE 4 Multiple sequence alignment of selected

ADF/cofilins. The residues mutated in mouse cofilin-1 in

this study are indicated by lines above the sequences. The

mutated residues that displayed severe (red) or moderate

(yellow) defects in the PI(4,5)P2-binding assays are high-

lighted. The conserved Trp, Trp-104, is also highlighted

(black).

2332 Zhao et al.

DISCUSSION

By applying a combination of biochemical, biophysical, and

mutagenesis methods, we were able to reveal several new

important aspects concerning the mechanism by which

ADF/cofilin family proteins bind to and are regulated by

phosphoinositides. We found that 1), ADF/cofilins simulta-

neously bind multiple PI(4,5)P2 headgroups in a cooperative

manner; 2), ADF/cofilins act as sensors of spatial PI(4,5)P2

density at the membrane; 3), ADF/cofilins do not interact

A

C D

B

TMR-PI(4,5)P2 quenching experiment with cofilin-1 mutants (glutamate substit

(K95, K96; K112, K114; and K125, K126K, 127) on PI(4,5)P2 binding. Each d


with the acyl chains of lipids or insert into the membrane

bilayer; 4), ADF/cofilins do not contain a specific phosphoi-

nositide-binding pocket, but instead interact with PI(4,5)P2

headgroups by unspecific electrostatic interactions through

a large, positively charged interface of the protein; and 5),

PI(4,5)P2-clustering activity of ADF/cofilins is diminished

at elevated pH.

Our work reveals a large, positively charged interface of

cofilin that simultaneously interacts with multiple PI(4,5)P2

headgroups, and thus provides an explanation for earlier

FIGURE 5 Determination of the PI(4,5)P2-

binding site of cofilin-1. (A) Vesicle cosedimen-

tation assay for examining the effects of various

cofilin-1 mutants (alanine substitution) on its

PI(4,5)P2-binding activity. This assay suggests

that the positively charged clusters K95, K96;

K112, K114; and K125, K126, K127 play

important roles in PI(4,5)P2 binding by cofi-

lin-1. The assay was carried out under the

same conditions described in Fig. 1 A. Each

data point represents the mean value of three

measurements, with the error bars indicating

5SD. A t-test was applied to analyze p-values

(** p < 0.01). (B) A bodipy-TMR-PI(4,5)P2

quenching experiment to examine the effects

of various cofilin-1 mutants (alanine substitu-

tion) for PI(4,5)P2 clustering. This assay re-

vealed the importance of the three positively

charged clusters (K95, K96; K112, K114; and

K125, K126, K127) in PI(4,5)P2 interactions

with cofilin-1. The lipid composition in the

assay was POPC/POPE/POPS/PIP2 (5:2:2:1)

and the total lipid concentration was 40 mM.

Each data point represents the mean of triplicate

measurements. (C) The vesicle cosedimentation

assay with cofilin-1 mutants (glutamate substi-

tution) suggests that K95, K96; K112, K114;

and K125, K126, K127 are the most critical

residues in PI(4,5)P2 binding. Each data point

represents the mean of triplicate measurements,

with the error bars indicating 5SD. A t-test

was applied to analyze p-values (*** p< 0.001,

** p < 0,01, * p < 0,05). (D) In similarity to

the vesicle cosedimentation assay, the bodipy-

ution) also revealed the importance of the same positively charged clusters

ata point represents the mean of triplicate measurements.

A

B

C

FIGURE 6 A schematic model for the interaction of cofilin with

PI(4,5)P2-containing membranes. (A) Space-filling model of the molecular

surface of human cofilin-1. The most critical residues in PI(4,5)P2 binding

are colored in red and other positively charged residues contributing to

PI(4,5)P2 binding are in orange. The mutated residues that did not display

effects on PI(4,5)P2 binding are in green. The view on the right represents

a 90� clockwise rotation around the long axis of the protein. (B) G-actin

and F-actin binding sites of cofilin. Residues that are critical for ADF/cofilin

interactions with G-actin and F-actin are in yellow (21,34,41). Residues that

are critical for interaction with F-actin, but not G-actin, are in cyan (42–44).

The view on the right represents a 90� clockwise rotation around the long

axis of the protein. (C) A schematic model for the interaction of cofilin-1

with PI(4,5)P2-rich membranes. The residues that are critical for PI(4,5)P2

binding are indicated in red and PI(4,5)P2 headgroups are indicated in

blue. Trp-104 is indicated in magenta. Cofilin-1 simultaneously binds

multiple PI(4,5)P2 headgroups through its large, positively charged inter-

face, and thus induces clustering of PI(4,5)P2 on the membrane.


studies demonstrating that ADF/cofilins do not bind IP3 with

detectable affinity (20,21). Because cofilin binding to

membranes displays a sharp sensitivity to PI(4,5)P2 density,

cofilin is not expected to bind soluble IP3 molecules that are

not aligned into lateral clusters. Of importance, as previously

proposed for N-WASP (39), ADF/cofilins can function as

PI(4,5)P2-density sensors at the plasma membrane. Thus,

our work provides a biochemical explanation for the results

of previous cell biology studies in which a relatively small

decrease in the plasma membrane PI(4,5)P2 concentration

upon EGF stimulation efficiently activated ADF/cofilin in

carcinoma cells (22). It is also important to note that the

binding threshold of cofilin occurs at 3–8% PI(4,5)P2, which

is similar to that reported recently for N-WASP and close to

the estimated physiological PI(4,5)P2 density at the plasma

membrane (39,40). Thus, ADF/cofilins bind PI(4,5)P2 head-

groups in a multivalent manner and sense the density of

phosphoinositides at the plasma membrane. Therefore, the

F-actin disassembly activity of ADF/cofilins can be effi-

ciently regulated by small changes in the PI(4,5)P2 concen-

tration or spatial distribution of PI(4,5)P2 at the plasma

membrane.

Our mutagenesis studies demonstrate that the PI(4,5)

P2-binding site of cofilin-1 overlaps extensively with both

G-actin and F-actin binding sites of ADF/cofilin (17,41)

(see also Fig. S4), which explains why the actin-binding

activities of ADF/cofilins are inhibited by PI(4,5)P2

(Fig. 6). However, in contrast to the previous NMR study

performed with water-soluble di-C8 forms of PI(4,5)P2

(27), our work revealed no specific PI(4,5)P2-binding

pocket(s) in cofilin-1. Although neutralization of the three

positively charged clusters significantly decreased the

PI(4,5)P2-binding and -clustering affinities of cofilin-1,

none of the mutants completely abolished PI(4,5)P2 binding.

The lack of specific phosphoinositide-binding pocket in

ADF/cofilins also explains the lack of phosphoinositide

specificity in this family of actin-binding proteins. It is

important to note that the PI(4,5)P2-binding residues identi-

fied here for mouse cofilin-1 are relatively well conserved in

all eukaryotic ADF/cofilins, and that all ADF/cofilins

contain a similar positively charged interface that is also crit-

ical for actin binding (Figs. 4 and 6). Furthermore, a similar

positively charged interface in yeast cofilin was previously

suggested to contribute to phosphoinositide binding (21).

Therefore, the PI(4,5)P2-binding mechanism identified here

for mouse cofilin-1 likely represents a general PI(4,5)P2-

binding mechanism of all ADF/cofilins.

In our Trp fluorescence and vesicle cosedimentation anal-

yses, the cofilin-1 interaction with membranes did not display

a detectable pH dependency at physiological salt concentra-

tion, although PI(4,5)P2 clustering was inhibited at elevated

pH (Figs. 7 and 8, and Fig. S3). Thus, it is possible that the

previously reported pH sensitivity of the cofilin-PI(4,5)P2

interaction is detectable only at nonphysiological low-salt

conditions. It is also important to note that a mutation in the

previously identified pH-sensor residue H133 (24) resulted

only in relatively small defects in PI(4,5)P2 binding and clus-

tering. Thus, H133 and its neighboring residues do not form


A

C D

B

FIGURE 7 The pH dependency of

cofilin-1 binding to PI(4,5)P2. (A and

B) PI(4,5)P2 binding quenches the Trp

fluorescence of cofilin-1. Based on this

assay, the binding of wild-type cofilin-

1 (A) and mutant K132A, H133A (B)

to PI(4,5)P2-rich membranes is not pH-

sensitive under physiological ionic

conditions. The final concentration of

cofilin-1 in this assay was 1 mM in a total

volume of 100 mL of 20 mM Hepes, 100

mM NaCl, pH 7.5. (C and D) PI(4,5)P2

binding of wild-type cofilin-1 (C) is

sensitive to changes in pH in the absence

of NaCl. However, the binding of

mutant K132A, H133A to PI(4,5)P2-

rich membranes was less sensitive to

changes in pH in the absence of NaCl

(D). The final concentration of cofilin-

1 was 1 mM in a total volume of 100

mL of 20 mM Hepes, pH 7.5. In both

assays the lipid composition was

POPC/POPE/POPS/PIP2 (5:2:2:1).

2334 Zhao et al.

an important PI(4,5)P2-binding pocket as previously

proposed (27). Furthermore, the corresponding histidine is

not conserved in invertebrate, plant, or yeast ADF/cofilins.

Thus, protonation-deprotonation of H133 is unlikely to

have a general role in regulating ADF/cofilin-PI(4,5)P2 inter-

actions in cells.

Taken together, our results show that ADF/cofilins can act

as PI(4,5)P2-density sensors that bind and stabilize PI(4,5)P2

clusters at the plasma membrane (Fig. 6 C). Because the

large, positively charged phosphoinositide-binding site over-

laps with the G-actin/F-actin binding sites of ADF/cofilins,

the activity of these proteins is inhibited at PI(4,5)P2-rich

membranes. Thus, a relatively small increase in PI(4,5)P2

density at the plasma membrane can significantly inhibit

ADF/cofilin-induced filament severing and depolymeriza-

tion. Furthermore, an increase in PI(4,5)P2 density will

lead to the activation of N-WASP-Arp2/3-promoted actin

filament assembly (39). These biochemical results are in

agreement with recent findings that an increase in

PI(4,5)P2 concentration at the plasma membrane promotes

actin filament assembly in mammalian cells (45,46). It is

important to note that the binding of ADF/cofilins to

PI(4,5)P2-rich membranes appears to be transient, and their

dissociation from the membranes can be further facilitated

by PI(4,5)P2 hydrolysis (22,23). Therefore, unspecific,

multivalent electrostatic interactions of ADF/cofilins with

phosphoinositide-rich membranes appear to be ideally suited

for down-regulating the activity of this protein.


SUPPORTING MATERIAL

Four figures are available at http://www.biophysj.org/biophysj/supplemental/

S0006-3495(10)00211-0.

We thank Maria Vartiainen, Juha Saarikangas, and Elena Kremneva for crit-

ical readings of the manuscript. Anna-Liisa Nyfors, and Riitta Kivikari

(Department of Applied Chemistry and Microbiology, University of Hel-

sinki) are acknowledged for providing technical assistance and access to

facilities, respectively.

This study was supported by grants from the Academy of Finland (to H.Z.

and P.L.), and the Sigrid Juselius Foundation and Biocentrum Helsinki

(to P.L.).

REFERENCES

1. Bamburg, J. R., and O. P. Wiggan. 2002. ADF/cofilin and actindynamics in disease. Trends Cell Biol. 12:598–605.

2. Friedl, P., and D. Gilmour. 2009. Collective cell migration in morpho-genesis, regeneration and cancer. Nat. Rev. Mol. Cell Biol. 10:445–457.

3. Pollard, T. D. 2003. The cytoskeleton, cellular motility and the reduc-tionist agenda. Nature. 422:741–745.

4. Paavilainen, V. O., E. Bertling, ., P. Lappalainen. 2004. Regulation ofcytoskeletal dynamics by actin-monomer-binding proteins. Trends CellBiol. 14:386–394.

5. Zebda, N., O. Bernard, ., J. S. Condeelis. 2000. Phosphorylation ofADF/cofilin abolishes EGF-induced actin nucleation at the leadingedge and subsequent lamellipod extension. J. Cell Biol. 151:1119–1128.

6. Kiuchi, T., K. Ohashi, ., K. Mizuno. 2007. Cofilin promotes stimulus-induced lamellipodium formation by generating an abundant supply ofactin monomers. J. Cell Biol. 177:465–476.

http://www.biophysj.org/biophysj/supplemental/S0006-3495(10)00211-0

http://www.biophysj.org/biophysj/supplemental/S0006-3495(10)00211-0

A

B

FIGURE 8 The pH dependency of PI(4,5)P2 clustering by cofilin-1. (A

and B) The bodipy-TMR-PI(4,5)P2 quenching assay reveals that PI(4,5)P2

clustering by wild-type cofilin-1 (A) displays moderate pH sensitivity,

whereas the K132A, H133A mutant (B) shows no detectable pH sensitivity

at physiological ion strength. The final concentration of liposomes was 40

mM in a total reaction volume of 100 mL in 20 mM Hepes, 100 mM

NaCl, pH 7.5. The lipid composition was POPC/POPE/POPS/PIP2/Bod-

ipy-TMR-PIP2 (50:20:20:9.5:0.5). The K132A, H133A mutant showed no

pH dependency in its ability to cluster PI(4,5)P2, suggesting that His-133

indeed contributes to the pH sensitivity of the PI(4,5)P2-clustering activity

of cofilin-1.


7. Abe, H., T. Obinata, ., J. R. Bamburg. 1996. Xenopus laevis actin-de-polymerizing factor/cofilin: a phosphorylation-regulated protein essen-tial for development. J. Cell Biol. 132:871–885.

8. Lappalainen, P., and D. G. Drubin. 1997. Cofilin promotes rapid actinfilament turnover in vivo. Nature. 388:78–82.

9. Maciver, S. K., and A. G. Weeds. 1994. Actophorin preferentially bindsmonomeric ADP-actin over ATP-bound actin: consequences for celllocomotion. FEBS Lett. 347:251–256.

10. Carlier, M. F., and D. Pantaloni. 1997. Control of actin dynamics in cellmotility. J. Mol. Biol. 269:459–467.

11. Blanchoin, L., and T. D. Pollard. 1999. Mechanism of interaction ofAcanthamoeba actophorin (ADF/cofilin) with actin filaments. J. Biol.Chem. 274:15538–15546.

12. Galkin, V. E., A. Orlova, ., E. H. Egelman. 2003. ADF/cofilin use an

intrinsic mode of F-actin instability to disrupt actin filaments. J. CellBiol. 163:1057–1066.

13. Carlier, M. F., V. Laurent, ., D. Pantaloni. 1997. Actin depolymeriz-

ing factor (ADF/cofilin) enhances the rate of filament turnover: implica-

tion in actin-based motility. J. Cell Biol. 136:1307–1322.

14. Andrianantoandro, E., and T. D. Pollard. 2006. Mechanism of actin fila-

ment turnover by severing and nucleation at different concentrations of

ADF/cofilin. Mol. Cell. 24:13–23.

15. Blanchoin, L., T. D. Pollard, and R. D. Mullins. 2000. Interactions of

ADF/cofilin, Arp2/3 complex, capping protein and profilin in remodel-

ing of branched actin filament networks. Curr. Biol. 10:1273–1282.

16. Chan, C., C. C. Beltzner, and T. D. Pollard. 2009. Cofilin dissociates

Arp2/3 complex and branches from actin filaments. Curr. Biol.19:537–545.

17. Ono, S. 2007. Mechanism of depolymerization and severing of actin

filaments and its significance in cytoskeletal dynamics. Int. Rev. Cytol.258:1–82.

18. Niggli, V. 2005. Regulation of protein activities by phosphoinositide

phosphates. Annu. Rev. Cell Dev. Biol. 21:57–79.

19. Di Paolo, G., and P. De Camilli. 2006. Phosphoinositides in cell regu-

lation and membrane dynamics. Nature. 443:651–657.

20. Yonezawa, N., E. Nishida, ., H. Sakai. 1990. Inhibition of the interac-

tions of cofilin, destrin, and deoxyribonuclease I with actin by phosphoi-

nositides. J. Biol. Chem. 265:8382–8386.

21. Ojala, P. J., V. Paavilainen, and P. Lappalainen. 2001. Identification of

yeast cofilin residues specific for actin monomer and PIP2 binding.

Biochemistry. 40:15562–15569.

22. van Rheenen, J., X. Song, ., J. S. Condeelis. 2007. EGF-induced PIP2

hydrolysis releases and activates cofilin locally in carcinoma cells.

J. Cell Biol. 179:1247–1259.

23. Leyman, S., M. Sidani, ., M. Van Troys. 2009. Unbalancing the phos-

phatidylinositol-(4,5) bisphosphate-cofilin interaction impairs cell steer-

ing. Mol. Biol. Cell. 20:4509–4523.

24. Frantz, C., G. Barreiro, ., D. L. Barber. 2008. Cofilin is a pH sensor for

actin free barbed end formation: role of phosphoinositide binding.

J. Cell Biol. 183:865–879.

25. Van Troys, M., D. Dewitte, ., C. Ampe. 2000. The competitive inter-

action of actin and PIP2 with actophorin is based on overlapping target

sites: design of a gain-of-function mutant. Biochemistry. 39:12181–

12189.

26. Yonezawa, N., Y. Homma, ., E. Nishida. 1991. A short sequence

responsible for both phosphoinositide binding and actin binding activi-

ties of cofilin. J. Biol. Chem. 266:17218–17221.

27. Gorbatyuk, V. Y., N. J. Nosworthy, ., G. F. King. 2006. Mapping the

phosphoinositide-binding site on chick cofilin explains how PIP2 regu-

lates the cofilin-actin interaction. Mol. Cell. 24:511–522.

28. Kusano, K., H. Abe, and T. Obinata. 1999. Detection of a sequence

involved in actin-binding and phosphoinositide-binding in the N-

terminal side of cofilin. Mol. Cell. Biochem. 190:133–141.

29. Vartiainen, M. K., T. Mustonen, ., P. Lappalainen. 2002. The three

mouse actin-depolymerizing factor/cofilins evolved to fulfill cell-type-

specific requirements for actin dynamics. Mol. Biol. Cell. 13:183–194.

30. Ward, L. D. 1985. Measurement of ligand binding to proteins by fluo-

rescence spectroscopy. Methods Enzymol. 117:400–414.

31. Stinson, R. A., and J. J. Holbrook. 1973. Equilibrium binding of nico-

tinamide nucleotides to lactate dehydrogenases. Biochem. J. 131:

719–728.

32. Eftink, M. R., and C. A. Ghiron. 1976. Exposure of tryptophanyl resi-

dues in proteins. Quantitative determination by fluorescence quenching

studies. Biochemistry. 15:672–680.

33. Ladokhin, A. S. 1999. Analysis of protein and peptide penetration into

membranes by depth-dependent fluorescence quenching: theoretical

considerations. Biophys. J. 76:946–955.


2336 Zhao et al.

34. Pope, B. J., K. M. Zierler-Gould, ., L. J. Ball. 2004. Solution structure

of human cofilin: actin binding, pH sensitivity, and relationship to actin-

depolymerizing factor. J. Biol. Chem. 279:4840–4848.

35. Blin, G., E. Margeat, ., C. Picart. 2008. Quantitative analysis of the

binding of ezrin to large unilamellar vesicles containing phosphatidyli-

nositol 4,5 bisphosphate. Biophys. J. 94:1021–1033.

36. Saarikangas, J., H. Zhao, ., P. Lappalainen. 2009. Molecular mecha-

nisms of membrane deformation by I-BAR domain proteins. Curr.

Biol. 19:95–107.

37. Zaritsky, A., A. H. Parola, ., H. Masalha. 1985. Homeoviscous adapta-

tion, growth rate, and morphogenesis in bacteria. Biophys. J. 48:

337–339.

38. De Kroon, A. I., M. W. Soekarjo, ., B. De Kruijff. 1990. The role of

charge and hydrophobicity in peptide-lipid interaction: a comparative

study based on tryptophan fluorescence measurements combined with

the use of aqueous and hydrophobic quenchers. Biochemistry.

29:8229–8240.

39. Papayannopoulos, V., C. Co, ., W. A. Lim. 2005. A polybasic motif

allows N-WASP to act as a sensor of PIP(2) density. Mol. Cell. 17:

181–191.


40. McLaughlin, S., J. Wang, ., D. Murray. 2002. PIP(2) and proteins:interactions, organization, and information flow. Annu. Rev. Biophys.Biomol. Struct. 31:151–175.

41. Paavilainen, V. O., E. Oksanen, ., P. Lappalainen. 2008. Structure ofthe actin-depolymerizing factor homology domain in complex withactin. J. Cell Biol. 182:51–59.

42. Lappalainen, P., E. V. Fedorov, ., D. G. Drubin. 1997. Essential func-tions and actin-binding surfaces of yeast cofilin revealed by systematicmutagenesis. EMBO J. 16:5520–5530.

43. Pope, B. J., S. M. Gonsior, ., A. G. Weeds. 2000. Uncoupling actinfilament fragmentation by cofilin from increased subunit turnover.J. Mol. Biol. 298:649–661.

44. Ono, S., A. McGough, ., A. G. Weeds. 2001. The C-terminal tail ofUNC-60B (actin depolymerizing factor/cofilin) is critical for maintain-ing its stable association with F-actin and is implicated in the secondactin-binding site. J. Biol. Chem. 276:5952–5958.

45. Rozelle, A. L., L. M. Machesky, ., H. L. Yin. 2000. Phosphatidylino-sitol 4,5-bisphosphate induces actin-based movement of raft-enrichedvesicles through WASP-Arp2/3. Curr. Biol. 10:311–320.

46. Yamamoto, M., D. H. Hilgemann, ., H. L. Yin. 2001. Phosphatidyli-nositol 4,5-bisphosphate induces actin stress-fiber formation andinhibits membrane ruffling in CV1 cells. J. Cell Biol. 152:867–876.

adf/cofilin binds phosphoinositides in a multivalent

Documents