Structural Features and Interfacial Properties of WH2,b-Thymosin Domains and Other Intrinsically DisorderedDomains in the Regulation of Actin CytoskeletonDynamics

Louis Renault,1* C�elia Deville,2 and Carine van Heijenoort2

1Laboratoire d’Enzymologie et Biochimie Structurales, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France2Laboratoire de Chimie et Biologie Structurales, Institut de Chimie des Substances Naturelles, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France

Received 10 June 2013; Revised 28 August 2013; Accepted 1 September 2013Monitoring Editor: Mikl�os Nyitrai

Many actin-binding proteins (ABPs) use complex multi-domain architectures to integrate and coordinate multi-ple signals and interactions with the dynamicremodeling of actin cytoskeleton. In these proteins,small segments that are intrinsically disordered in theirunbound native state can be functionally as importantas identifiable folded units. These functional intrinsi-cally disordered regions (IDRs) are however difficult toidentify and characterize in vitro. Here, we try to sum-marize the state of the art in understanding the struc-tural features and interfacial properties of IDRsinvolved in actin self-assembly dynamics. Recent struc-tural and functional insights into the regulation of wide-spread, multifunctional WH2/b-thymosin domains, andof other IDRs such as those associated with WASP/WAVE, formin or capping proteins are examined.Understanding the functional versatility of IDRs in actinassembly requires apprehending by multiple structuraland functional approaches their large conformationalplasticity and dynamics in their interactions. In manymodular ABPs, IDRs relay labile interactions with mul-tiple partners and act as interaction hubs in interdomainand protein–protein interfaces. They thus control multi-ple conformational transitions between the inactive andactive states or between various active states of multido-main ABPs, and play an important role to coordinatethe high turnover of interactions in actin self-assemblydynamics. VC 2013 Wiley Periodicals, Inc.

Key Words: actin cytoskeleton remodeling; actin-bindingproteins; WH2 (Wiskott-Aldrich syndrome protein Homology

domain 2) domains; ß-thymosin domains; intrinsically disor-

dered proteins; functional versatility; multifunctionality;

structural biology

Introduction

The actin cytoskeleton forms a very dynamic structuralnetwork that is constantly remodelled in eukaryotic

cells to control and coordinate multiple cellular processes,including cell polarity establishment and maintenance,polarized cell migration, cell adhesion, cytokinesis, intracel-lular transport, or intracellular pathogen infections. Thesecellular processes are driven by the tightly regulated self-assembly of actin monomers (G-actin) in polarized,double-stranded helical filaments (F-actin), formation ofhigher order F-actin structures, and their controlled disas-sembly [Winder and Ayscough, 2005; Pollard and Cooper,2009]. The dynamic behavior of the actin cytoskeleton andits involvement in many diverse cellular processes rely on alarge array of actin-binding proteins (ABPs) that work inconcert. Each ABP regulates elementary functions with G-or F-actin. ABPs regulate for example the nucleotide-binding and self-assembly properties of the cellular G-actinpool or actin filament nucleation, polarized elongation,length, stability, disassembly, cross-linking or anchoring.These multiple activities are spatiotemporally regulatedwithin the cells downstream of specific cellular cues to con-trol actin dynamics at restricted cellular locations. To adaptcell migration to diverse external microenvironments, thecortical actin cytoskeleton is for example intimately con-nected with the plasma membrane and signaling emanatingfrom plasma membrane receptors, curvature, compositionor dynamics [Bisi et al., 2013]. Changes in the actin cyto-skeleton organization are also controlled by changes in thetranscriptional expression profile of ABPs. Finally, post-translational regulations of ABPs or their partners bringadditional labile informations in proteins that are thoughtto drive rapid changes in actin dynamics [Campellone and

*Address Correspondence to: Louis Renault; Laboratoire d’Enzymo-logie et Biochimie Structurales, Centre de Recherche de Gif,CNRS, Gif-sur-Yvette, France. E-mail: renault@lebs.cnrs-gif.fr

Published online 22 October 2013 in Wiley Online Library(wileyonlinelibrary.com).

REVIEW ARTICLECytoskeleton, November 2013 70:686–705 (doi: 10.1002/cm.21140)VC 2013 Wiley Periodicals, Inc.

� 686

Welch, 2010; Hannappel, 2010; Hansen and Kwiatkowski,2013; Mendoza, 2013]. Many ABPs are therefore made ofmultiple functional folded domains by which they integrateand coordinate multiple cellular signals and interactions.Understanding all their interactions and coordination at themolecular level is arduous and requires to combine manyfunctional and structural investigations. In the last years, anadditional level of complexity has been highlighted in manystudies on multidomain ABP regulation: beyond identifia-ble folded domains, modular ABPs can contain smallregions or segments that are intrinsically disordered in theirnative state but essential for maintaining the ensemble ofinteractions and activities of the full-length protein. Thesefunctional intrinsically disordered segments/regions are dif-ficult to identify and characterize in vitro because they fre-quently have small sizes, high sequence variability anddisplay inherent instability, plasticity and dynamics in solu-tion. In addition they are especially sensitive to proteolysisand conducive to protein aggregations. Deciphering theirspecific structural properties and mechanisms of regulationsin modular ABPs represents nowadays a challenging butimportant task to understand the full regulation of manymultidomain protein architectures involved in actin cyto-skeleton remodeling. We overview here the hallmarks ofseveral cytoskeletal intrinsically disordered proteins (IDPs)or segments/regions (IDRs) involved in various importantfunctions of actin self-assembly dynamics.

About 10–35% of prokaryotic and 15–45% of eukaryo-tic proteins contain significant disorder, that is, long disor-dered regions of at least 30 amino acids [Uversky andDunker, 2010; Tompa, 2012]. Their higher presence ineukaryotes is correlated with their frequent involvement insignal transduction or transcriptional regulation. Many fur-ther investigations are required to understand this new classof proteins that defy the structure-function paradigm. Thedistinguishing feature of functional IDPs/IDRs is theirinability to fold into a unique and stable tertiary structurein solution and/or in the crowding environment of cells[Tompa, 2011]. Their native isolated state consists in adynamic ensemble of interconverting conformers. Althoughsome adopt well defined secondary or tertiary structuresupon binding to their partners, numerous IDPs appear tobe functional in complexes through retaining various degreeof dynamics and disorder, or adopting multiple conforma-tions in protein–protein interfaces, leading to the introduc-tion of fuzzy complexes [Fuxreiter, 2012; Fuxreiter andTompa, 2012]. The functional features of IDPs can be fur-ther summarized by adaptability in binding, often weak butspecific binding, and frequent regulation by post-translational modification or alternative splicing. The inter-est to understand deeply the molecular mechanisms ofIDP/IDR regulations at protein–protein interfaces is fur-ther accentuated by the involvement of numerous IDPs inmany human diseases, such as amyloidoses and neurodege-nerative diseases, but also cancer, cardiovascular disease,

and diabetes [Uversky, 2010]. What are the main functions,interfacial properties, and regulatory mechanisms of manyIDPs/IDRs involved in cytoskeleton remodeling, and theircommonalities and uniqueness compared with other IDPsremain open issues.

In this journal issue, Guharoy et al. provide a thoroughoverview on the abundance, functional diversity, and impli-cations in diseases of IDPs/IDRs involved in the regulationof the three major cytoskeletal systems of eukaryotic cells[Guharoy et al., 2013]. Here, we try to capture the state ofthe art in understanding the main structural features andinterfacial properties of IDRs involved in actin self-assembly dynamics. We outline more especially recentmolecular findings that analyze the regulation of WH2(WASP-homology 2) and b-thymosin (bT) domains inactin assembly, because they represent widespread, arche-typal actin-binding IDRs [Dominguez, 2009; Carlier et al.,2011]. Recent findings on IDRs that are associated with theregulations of nucleation promoting factors (NPFs) of theWASP/Scar/WAVE family, formin, or capping proteins(CPs) are also examined. We review what the main struc-tural and functional properties of these IDRs are when theywork in actin assembly as single domains, alone or in asso-ciation with other domains targeting actin or other ABPs,or as multiple repeats. Our analysis highlights how IDPs/IDRs involved in actin cytoskeleton remodeling are impor-tant and adapted for the multiple and rapid changes ininterdomain interfaces and protein–protein interactionsthat control the high turnover of interactions in actinassembly dynamics.

Elementary Properties andFunctional Versatility of IsolatedSingle Actin Cytoskeletal IDRs:Example of WH2/bT Domains

Among actin cytoskeletal IDRs/IDPs, widespread WH2and bT domains are among the shortest functional actin-binding modules known. They either build mono-domainIDPs or exist as single or repeated units in many multido-main proteins.

Variability of WH2/ßT IDRs and of TheirStructural Organizations in Proteins

Individual ß-thymosin domains were first identified as afamily of �5 kDa, heat stable polypeptide isoforms fromcalf thymus in the mid-1960’s, called ß-thymosins [Huffet al., 2001]. Within cells, ß-thymosins mainly interactwith actin monomers and influence many cellular func-tions, including cell migration or endothelial cell attach-ment and spreading [Sribenja et al., 2013]. Thymosin-ß4(Tß4), the most abundant ß-thymosin isoform in metazo-ans, is present at very high concentrations (300–500 mM)in platelets and lymphocytes, and is overexpressed in several

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 687 �

tumours. Full-length or truncated Tß4 acts also extracellu-larly with less understood mechanisms, for example to pro-mote dermal and corneal wound healing, or stimulatecoronary vasculogenesis [Sosne et al., 2010]. WH2 domainswere initially found in NPFs of the WASP and Scar/WAVEfamily that activate the Arp2/3 complex to generate branch-ing networks of filaments in many actin-based motile proc-esses involving membrane deformations (detailed below).WH2/ßT domains have been since identified or predictedas single or multiple units in a large number of modular sig-naling/regulatory proteins. The SMART database predicts

that almost 850 proteins in eukaryotes and few dozen inprokaryotes or viruses contain a variable number of possibleWH2/ßT domains (Fig. 1A). Like many IDRs, the latterdisplay highly variable sequences and a weak sequence sig-nature, which makes them difficult to identify or delimitproperly (Fig. 1B). Their sequences vary between 25 and55 amino acids [Paunola et al., 2002; Chereau et al., 2005;Carlier et al., 2011]. It is mainly identified by a central con-sensus actin-binding motif LKKT/V flanked by variable N-terminal and C-terminal extensions, with a N-terminal seg-ment of about 8–12 residues predicted to fold into an

Fig. 1. Hallmarks and distribution of ßT and WH2 domains in proteins, and structural basis for their functional versatility assingle domains bound to actin monomers. (A) Taxonomic distribution of proteins with ßT (gray labeling) or WH2 (black)domains in the SMART (Simple Modular Architecture Research Tool) database [Letunic et al., 2012]. The minimum and maximumnumber of repeats per protein found within each species is indicated in parenthesis. (B) The sequence homology of 30 ßT and30 WH2 sequences found in proteins from different species is displayed as a sequence logo for ßT (upper) and WH2 (lower)sequence pattern [Crooks et al., 2004]. The height of each amino acid letter is proportional to their relative frequency at that posi-tion in the sequence. Above are indicated the main sequence elements and how they interact with G-actin. (C) Conformational vari-ability of free G-actin, free and actin-bound bT. The latter correspond to full SAXS models of G-actin-bound CibD1 bT obtainedeither at low (sequestration function) or near physiological (profilin-like function) ionic strength [Didry et al., 2012]. Actin subdo-mains 1 and 3 (G-actin barbed face), and 2 and 4 (G-actin pointed face) are indicated on free G-actin. The intrinsic conformationalflexibility of free G-actin is shown by four overlaid G-actin conformations (PDB: 1HLU, 2VCP, 2PDB or SAXS model [Didryet al., 2012]), and the flexibility of free bT by the solution structure of free Tß9, which adopts a mostly helical folding in 40%1,1,1,3,3,3-hexafluoro-2-propanol (HFP) [Stoll et al., 1997], and different SAXS conformations of bound CibD1 bT, shown herewithout G-actin [Didry et al., 2012]. Arrows indicate to which end of the filament the monomer can associate with its barbed andpointed faces. (D) Structural basis for the functional versatility of single WH2/ßT domains in actin assembly. Functionally differentbT/WH2 domains differ mainly by distinct dynamics of their C-terminal half interactions with G-actin pointed face that inhibit orallow association with the most dynamic barbed-end (1) of filaments and can be controlled by a single residue in long WH2/ßT[Didry et al., 2012]. In contrast, static interactions between WH2/ßT N-terminal half and G-actin barbed face inhibit always associa-tion with filament pointed-end (-).

� 688 Renault et al. CYTOSKELETON

amphipatic a-helix of 2 to 3 turns (Fig. 1B). ßT share amore extended and conserved C-terminal half than WH2(Fig. 1B). Sequence identity within the separated ßT andWH2 domain families is �40 and �20%, respectively, anddrops to less than 15% when comparing sequences acrossthe families. In the last years, WH2/ßT domains/repeatshave emerged as multifunctional regulators of actin assem-bly dynamics but their exact contribution and regulatorymolecular mechanism in most multidomain ABPs remainpoorly understood or are still controversial [Dominguez,2009; Campellone and Welch, 2010; Carlier et al., 2011].

The demonstration that individual WH2/ßT domainsare archetypes of functional IDPs were first provided byNMR studies [Safer et al., 1997; Domanski et al., 2004].NMR spectroscopy represents indeed the most powerfulstructural methods to analyze in solution the folding, con-formational plasticity and dynamics of IDPs/IDRs on alarge number of time scales in their native or bound states,or in a cellular crowded environment [McNulty et al.,2006; Bodart et al., 2008]. For IDRs involved in actinassembly regulation, detailed NMR analysis of actin confor-mations, and full characterization of complex interfaces arehowever limited by the quite large size of the system, theunavailability of (15N, 13C, 2H)-labelled actin, whichrequires to be expressed in eukaryotic systems, and the diffi-culty to avoid polymerization of actin at the concentrationsneeded (typically 300 mM) for a full NMR study. We illus-trate hereafter how the use of engineered chimeric proteins,actin-binding and polymerization assays, crystallographic,NMR, and SAXS (Small Angle X-ray Scattering) structuralapproaches and molecular dynamic simulations comple-ment each other to decipher the molecular regulatorymechanisms and functional versatility of IDPs/IDRs inactin assembly dynamics.

Functional Versatility of Individual WH2/ßTIDRs in Actin Assembly Controlled by FuzzyComplexes with G-actin

In solution, individual WH2/ßT domains characterized sofar as isolated domains, bind usually preferentially G-actin-ATP over G-actin-ADP, inhibit the nucleotide exchange inG-actin, and do not or weakly interact with filaments [Car-lier et al., 1993; Carlier et al., 1996; De La Cruz et al.,2000; Marchand et al., 2001; Mattila et al., 2003]. Theirbinding affinity for G-actin-ATP can be very sensitive tothe ionic strength but near physiological ionic conditionsthey usually form complexes of moderate affinity with equi-librium dissociation constants (Kd) in the mM range [DeLa Cruz et al., 2000; Didry et al., 2012]. Their complexeswith actin monomers inhibit the monomer association tofilament pointed ends, but can either associate or not withthe most dynamic barbed-ends of filaments to promote orinhibit their elongation. ß-thymosins (5 kDa) like Tß4sequester G-actin, ie their complexes can not polymerize ateither F-actin end. Most other WH2/ßT studied as individ-

ual domains in actin assembly, such as WASP and Scar1/WAVE1 WH2 domains [Egile et al., 1999; Higgs and Pol-lard, 1999], the 3 ßT repeats of Drosophila Ciboulot or itsisolated first ßT domain (CibD1) [Hertzog et al., 2002],enhance motility by directing polarized assembly, with vari-able association rate constants for the binding of their com-plexes to barbed-ends. Their complexes behave for filamentelongation as functional homolog of G-actin-ATP:profilincomplexes. Their sequestering or profilin-like functions inactin assembly do not rely on different binding affinities forG-actin [De La Cruz et al., 2000; Hertzog et al., 2002;Didry et al., 2012] or obvious differences in their sequencecomposition. Many biochemical and structural studies havebeen required to understand at the molecular level theinteractions of WH2/ßT with G-actin [Husson et al., 2010]and how the functional versatility between WH2/ßT iscoded in their highly variable sequence and achieved at amechanistic level in their complexes.

Circular dichroism data and NMR experiments usingisotopically labeled IDRs have first shown that Tß4, Tß9 orScar1 isolated WH2 are predominantly unfolded in theirfree native state in physiological buffers [Safer et al., 1997;Stoll et al., 1997; Domanski et al., 2004; Kelly et al.,2006]. Upon binding to G-actin-ATP, Tß4 folds com-pletely. Its secondary structure elements that are otherwisetransient are stabilized with G-actin and it displays a centralextended region flanked by two N- and C-terminal helices[Safer et al., 1997; Simenel et al., 2000; Domanski et al.,2004]. To complement these structural studies, many crys-tal structures of various WH2/ßT domains bound to G-actin have been obtained using different polymerizationinhibiting methods [Hertzog et al., 2004; Chereau et al.,2005; Aguda et al., 2006; Lee et al., 2007; Ducka et al.,2010; Rebowski et al., 2010; Didry et al., 2012]. Thesecrystal structures show that highly variable WH2/ßTdomains display a similar overall extended fold and bindingpath on G-actin with their N-terminal half but elusiveinteractions after their central LKKT/V motif. Their N-terminal amphipathic helix always caps the barbed face ofG-actin by binding into the hydrophobic cleft between itssubdomains 1 and 3, thus preventing association with theslow-growing pointed ends of filaments (Figs. 1C and 3E).Significant interactions with the pointed face of G-actinwere observed in a crystal structure only while using a chi-mera of a C-terminal segment of Tß4 fused to gelsolin seg-ment 1 in complex with G-actin [Irobi et al., 2004]. In thishybrid complex, Tß4 C-terminal residues (amino acid 30–39) fold into an a-helix that caps the pointed face of G-actin by binding between its subdomains 2 and 4. A fullsequestering structural model of Tß4 was therefore pro-posed to block/cap both the barbed and pointed face of G-actin with Tß4 N- and C-terminal helices, respectively.WH2/ßT domains with profilin-like function were pro-posed in contrast to conserve G-actin pointed face free incomplexes for association with filament barbed-ends. The

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 689 �

Tab

leI.

Mu

ltip

leIn

tera

cti

on

san

dR

ole

so

fID

Rs

inP

rote

ins

Re

pre

sen

tati

ve

of

Dif

fere

nt

Acti

vit

ies

inA

cti

nS

elf

-Ass

em

bly

Dyn

am

ics

Prot

eins

IDR(

s)Fu

nctio

nsre

late

dto

the

IDR(

s)Ro

leof

diso

rder

Refs

Tß

4,d.

mel

anog

aste

rC

ibou

lot,

WA

SP,

WA

VE

,M

IM,

WIP

,V

ASP

,...

Sin

gle

G-a

ctin

-bin

din

gW

H2/

ßT

sequ

este

rG

-act

in-A

TP

orp

ro-

mot

eit

su

nid

irec

tion

alas

sem

bly

onba

rbed

-en

d(p

rofi

lin

-lik

eac

tivi

ty)

Fu

zzy

com

ple

xes:

vari

able

loca

ld

ynam

ics

inG

-act

in-b

oun

dst

ates

con

trol

vari

able

asso

cia-

tion

pro

per

ties

wit

hba

rbed

-en

dan

dca

nbe

tun

edw

ith

ap

oin

tm

uta

tion

/mod

ifica

tion

[Did

ryet

al.,

2012

]

NP

Fs:

WA

SP,

N-W

ASP

,W

AV

E,

...

Mid

dle

B-G

BD

orm

ean

der

-Bre

gion

s,C

-ter

min

alV

/WH

2-C

A

B-G

BD

:bi

nd

ing

toP

IP2

/3,

Rh

o-G

TP

ase,

kin

ase/

SH2,

(V)C

AV

CA

:bi

nd

ing

toB

-G

BD

,G

-act

in-A

TP,

Arp

2/3

sign

alin

tegr

ator

,m

olec

ula

rh

ub,

mu

tual

lyex

clu

sive

inte

r-ac

tion

s,co

oper

ativ

ein

terd

o-m

ain

mid

dle

-to-

tail

auto

inh

ibit

ion

,co

oper

ativ

eal

lost

eric

acti

vati

ons

[Kim

etal

.,20

00;

Ch

enet

al.,

2010

]

mD

ia1–

3,IN

F2,

FM

NL

3,or

FM

N1–

2su

bfam

ilie

sof

form

ins

C-t

erm

inal

DA

D,

WH

2-D

AD

orF

SIN

-ter

m.

DID

bin

din

g,co

oper

-at

ivit

y/sy

ner

gyfo

rF

H2-

med

iate

dG

-act

inn

ucl

eati

onor

FH

1-F

H2-

med

iate

dfi

la-

men

tel

onga

tion

,W

H2-

DA

D:

fila

men

tse

veri

ng

inas

soci

atio

nw

ith

FH

2,F

SI:

Spir

e-K

IND

bin

din

g

sign

alin

tegr

ator

,m

utu

ally

excl

usi

vein

tera

ctio

ns,

inte

r-d

omai

nh

ead

-to-

tail

auto

inh

i-bi

tion

,d

ynam

icbi

nd

ing

wit

hG

-an

d/o

rF

-act

in,

FSI

:d

ynam

ictu

nin

gof

FH

2ac

tivi

tyw

ith

Spir

e

[Li

and

Hig

gs,

2005

;C

hh

abra

etal

.,20

09;

Gou

ldet

al.,

2011

;V

izca

rra

etal

.,20

11;

Bor

etal

.,20

12]

CP

(a:b

het

erod

imer

)b

-su

bun

itC

-ter

m.

ten

tacl

e(b

-te

nta

cle)

Fol

din

g&

bin

din

gto

barb

ed-

end

term

inal

acti

nre

quir

edfo

rti

ght

cap

pin

gof

CP

opti

mal

and

dyn

amic

bin

din

gof

CP

a:b

het

erod

imer

tofi

l-am

ent

barb

eden

d

[Tak

eda

etal

.,20

10]

CD

2AP,

CK

IP,

CA

RM

IL,.

..C

AR

MIL

-pep

tid

eu

nca

pC

Pa

:bh

eter

odim

erbo

un

dto

barb

ed-e

nd

allo

ster

icm

ech

anis

md

istu

rbin

gC

Pa

:bbe

nd

ing/

twis

tin

gbo

un

dto

barb

ed-e

nd

[Tak

eda

etal

.,20

11;

Kim

etal

.,20

12]

For

min

s(F

H1-

FH

2d

omai

ns)

FH

1d

omai

n:

rep

eats

ofp

roli

ne-

rich

mot

ifs

bin

din

gto

pro

fili

n

Rec

ruit

and

feed

the

FH

2-bo

un

dba

rbed

end

wit

hp

rofi

lin

:G-a

ctin

-AT

P

FH

1d

isor

der

stat

eof

fers

ala

rger

cap

ture

rad

ius

for

pro

fili

n:G

-act

in,

and

wit

hit

sn

on-r

edu

nd

ant

rep

eats

,sp

eed

sth

eel

onga

tion

rate

ofba

rbed

-en

ds

asso

ciat

edw

ith

FH

2-C

term

inal

dom

ain

s

[Sh

oem

aker

etal

.,20

00;

Pau

lan

dPo

llard

,20

08;

Cou

rte-

man

che

and

Polla

rd,

2012

]

Spir

e,C

obl,

Vop

F/V

opL

,JM

Y,N

-WA

SPW

H2/

ßT

rep

eats

sequ

este

ror

nu

clea

teG

-act

in-

AT

P/A

DP,

orp

rom

ote

G-

acti

nu

nid

irec

tion

alas

sem

bly,

seve

rfi

lam

ent,

cap

barb

ed-

end

,...

N-W

ASP

,JM

Y:

incr

ease

effi

cien

cyof

bran

ch-

ing

med

iate

dby

NP

F-A

rp2/

3

coop

erat

ive

&d

ynam

icbi

nd

ing

wit

hd

iffe

ren

tac

tin

tert

iary

&qu

ater

nar

yst

ruct

ure

s,re

pea

tsp

rom

ote

inte

ract

ion

sw

ith

seve

ral

acti

nsu

bun

its

offi

lam

ent

orm

ayas

sem

ble

vari

able

/dyn

amic

G-a

ctin

olig

omer

s

[Qu

inla

net

al.,

2005

;A

hu

jaet

al.,

2007

;B

osch

etal

.,20

07;

Hu

sson

etal

.,20

11;

Gau

cher

etal

.,20

12]

� 690 Renault et al. CYTOSKELETON

switch of function of ßT domains from inhibition to pro-motion of actin assembly were expected to depend on dif-ferent but illdefined interactions of their C-terminal regionwith actin subdomains 2 and 4, mediated by sequence var-iations widely distributed within their C-terminal half[Eadie et al., 2000; Hertzog et al., 2004; Irobi et al., 2004;Chereau et al., 2005; Aguda et al., 2006].

We have recently identified general structural basis forthe sequestering and profilin-like function of ßT and WH2in actin assembly [Didry et al., 2012]. The individualWH2/ßT of Tß4, CibD1, Caenorhabditis elegans Tet-rathymosinb containing 4 ßT domains [Van Troys et al.,2004], and human WASP-interacting protein (WIP) wereused as models of functionally or structurally differentWH2/ßT. Tß4 and CibD1 display more similar sequencesbut sequestering and profilin-like function, respectively,while Tß4 and WIP WH2 display only �15% sequencesimilarity. The segments in ßT that contribute the most totheir overall affinity with G-actin and their function inactin assembly were determined using different chimeras ofCibD1 ßT and Tb4. We combined mutational analysis ofthe domains with actin binding and polymerization assaysand several structural approaches, including protein crystal-lography, SAXS and NMR. Crystal structures were usefulto analyze at high resolution all static/stable interactions inactin:ßT complexes, including those that affect distantly thebehavior of disordered regions in complexes. SAXS andNMR have been instrumental to analyse the conforma-tional dynamics of ßT IDRs in physiological buffers, andto correlate the structural changes in their complexes withfunctional differences measured in biochemical assays. Thelow resolution SAXS structural technique revealed overallconformational changes in complexes between low (nonpolymerizing G-buffer) and physiological (polymerizing F-buffer) ionic strength that correspond to different functionsin actin assembly. The average conformations of flexibleregions in complexes were modeled using the solutionSAXS patterns of the complexes based on the uncompletecrystal structures of the complexes (Fig. 1C). NMR experi-ments using isotopically labeled ßT IDRs were essential toanalyze in detail at different ionic strengths the folding andconformational dynamics of ßT domains between their freeand actin bound states. Altogether the structural analysesdemonstrate that the entire sequence of all ßT can interactwith G-actin but with N-terminal static and C-terminaldynamic interactions. Single ßT domains with differentfunctions do not need to target alternative actin bindingsites via specific sequence variations in their central and C-terminal regions as previously thought, but to exhibit alter-native dynamics of their C-terminal half interactions withG-actin pointed face (Figs. 1C–D). At physiological ionicstrength these local interaction dynamics are primarily con-trolled by strong electrostatic interactions of a single residuealong their sequence. In Tß4, this key residue is Lys14 thatis located in the linker region between the N-terminal

TA

BL

EI.

Co

nti

nu

ed

Prot

eins

IDR(

s)Fu

nctio

nsre

late

dto

the

IDR(

s)Ro

leof

diso

rder

Refs

JMY,

MR

TF

-Aor

Ph

actr

1tr

ansc

rip

tion

alco

fact

ors

rep

eats

ofW

H2/

ßT

orof

RP

EL

-mot

ifs

G-a

ctin

and

imp

orti

n-a

bin

d-

ing,

G-a

ctin

-con

cen

trat

ion

-se

nsi

ng

mec

han

ism

regu

lati

ng

nu

clea

rlo

cali

zati

onfo

rtr

ansc

rip

tion

sign

alin

tegr

ator

,m

utu

ally

excl

usi

vein

tera

ctio

ns,

coop

er-

ativ

eG

-act

inbi

nd

ing

[Mou

ille

ron

etal

.,20

12;

Zu

cher

oet

al.,

2012

]

WH

2:W

ASP

Hom

olog

y2

dom

ain

,b

T:

b-t

hym

osin

dom

ain

,N

PF

s:N

ucl

eati

ng

Pro

mot

ing

Fact

ors,

FH

1:F

orm

in-H

omol

ogy

1,B

:p

olyb

asic

regi

on,

GB

D:

GT

Pas

e-B

ind

ing

Dom

ain

,V

CA

:V

erp

ro-

lin

-hom

olog

y/W

H2,

Cen

tral

,A

cid

icre

gion

s,D

AD

:D

iaau

tore

gula

tory

dom

ain

,F

SI:

For

min

-Sp

ire

Inte

ract

ing,

RP

EL

:ac

tin

-bin

din

gID

Rs

wit

hco

rese

quen

ceR

Pxx

xEL

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 691 �

amphipatic helix and the central consensus LKKT/V motifand provides a salt bridge with the highly conservedGlu334 of actin subdomain 3. Thus, a single salt bridgewith G-actin located just upstream (in Tß4/10/15 ßT) or

downstream (in WIP WH2) of their LKKT/V motif caninduce a sequestering function in actin assembly with dif-ferent and distantly related ßT or WH2 sequences. Theremoval of this strong electrostatic interaction between

Fig. 2.

� 692 Renault et al. CYTOSKELETON

WH2/ßT and G-actin by a point mutation in Tß4 ßT orWIP WH2 on their key residue (K14Q/A in Tß4, R54Nin WIP) is sufficient to reverse their function into aprofilin-like function in actin assembly or to weaken theirinteraction with G-actin so much that it abolishes their pri-mary function in actin assembly. Similarly, CibD1 ßT canbe switched into an efficient sequestering ßT following asingle point mutation that introduces a salt bridge with G-actin next to the LKKT/V motif like in Tß4 ßT. Theseresults open perspectives for elucidating the multiple func-tions of both ßT and WH2 domains in other modular pro-teins [Didry et al., 2012].

IDR Regulation by PostranslationalModifications

WH2/ßT:G-actin complexes therefore represent fuzzy com-plexes in which the inherent disorder of WH2/ßT is par-tially conserved and functional in actin assembly (Table I).These results expand the variety of molecular mechanismsby which IDPs can fulfill different functions followingsmall sequence modifications that affect their electrostaticinteractions in complexes [Dyson and Wright, 2005;Tompa et al., 2005]. IDPs involved in cell signaling andtranscription regulation recruit structurally dissimilar pro-teins and regulate different activities by adopting alternativeconformations within their complexes. These alternativeconformations are driven by the inherent structural plastic-ity of IDPs combined with the remodeling of their primarystructure by reversible phosphorylations [Fabrega et al.,2003; Dyson and Wright, 2005; Tompa et al., 2005]. Here,the intrinsic structural disorder of WH2/ßT domains leadsto a novel mode of functional versatility, in which the disor-dered protein does not regulate multiple targets, but indu-ces different functions with the same target by varying thedynamics of their C-terminal interactions with smallsequence differences. The activity of these IDRs in actinassembly may therefore be switched by single reversiblesequence modifications like by other IDPs. In that regardß-Thymosins undergo posttranslational modifications

whose biological relevance is mostly unknown [Hannappel,2010]. In a human proteomic study a fraction of Tß4 andTß10 was found to be acetylated on the e-amino group ofLys14 [Choudhary et al., 2009; Hannappel, 2010]. Byabolishing the salt bridge with Glu334, acetylation ofLys14 can be predicted to decrease significantly the affinityfor G-actin [Didry et al., 2012] and to switch the sequester-ing factors Tß4/10 into factors promoting unidirectionalactin assembly at filament barbed-ends. Future experimentsare required to challenge this possibility and to understandbetter how posttranslational modifications may impact invivo the functions of WH2/ßT IDRs. Single phosphoryla-tions play important roles in the regulation of many otherunrelated IDRs found in multidomain ABPs. This is illus-trated in the following section with recent molecularinsights into the regulation of modular ABPs that acceleratede novo filament nucleation, a key activity in actin assembly.

Interfacial Properties of IndividualIDRs Functionally Coupled withOther Neighbouring Domains inModular ABPs

In multidomain proteins, IDRs are involved in many mac-romolecular interfaces and work in association with otheradjacent disordered or folded domains. IDRs associatedwith the regulation of nucleating or CPs illustrate well howthese segments control and coordinate multiple labile inter-actions at interdomain and protein–protein interfaces inmodular ABPs.

Actin nucleation is the key rate-limiting step in spontane-ous actin polymerization, and tight regulation of this pro-cess is critical to ensure that actin filaments form rapildy atthe right time and place in cells. To date, three main differ-ent protein machineries have been proposed to nucleatenew actin filaments: Arp2/3 complex activated by NPFs ofthe WASP/Scar/WAVE family, formin with profilin, andproteins containing nucleating WH2 repeats [Dominguez,2009; Campellone and Welch, 2010]. The main relevant in

Fig. 2. Domain organization of WASP, WAVE1 and different formin subfamilies, and structural models describing how theirstructured or intrinsically disordered domains regulate the protein inactive and active states. (A) Modular domain architectureof human WASP, WAVE1 and different formins. Partially/fully IDR (dashed red lines), and autoinhibiting intramolecular interactions(dashed black lines) are indicated above and below each protein, respectively. A, acidic region; B, poly-basic region; C, central region;CC, coil-coiled region; DAD, Diaphanous autoinhibitory-like domain; DD, Dimerization-domain; DID, Diaphanous inhibitory-likedomain; FH1/2, Formin-homology 1/2 domain; FSI, formin–Spire interaction domain; GBD, GTPase-binding domain; PIP2,PtdIns(4,5)P2, PtdIns-4,5-bisphosphate; PRD, Proline-rich domain; SH2/3, Src –homology 2/3 domain; SHD, Scar/WAVE homol-ogy domain; V, WH2; WH1, WASP homology domain 1; (B) In WASP autoinhibited state, its middle B-GBD and C-terminal C-AIDRs fold and bind with each other to make inaccessible the V-C-A regions to G-actin and Arp2/3. (C) WAVE1/2 is autoinhibitedby intra- and intermolecular interactions within a complex of 5 proteins (Sra1, Nap1, HSPC300, Abi2 and WAVE1/2). In inactiveWAVE1, its middle meander-B IDRs (yellow trace), located between its N-terminal structured SHD and unfolded PRD, integrate allthe functional intramolecular inhibitory contacts with its C-terminal V-C-A IDRs. (D) In mDia, INF2, FMNL-3 or FMN subfamilyof formin, the homodimeric G-/F-actin-binding FH2 domain is surrounded in each protomer by a FH1 and C-terminal region thatare both at least partially disordered. Their C-terminal IDRs control both interdomain head-to-tail autoinhibition (DAD/FSI withN-term. GBD-DID/DID-like) and different interactions in the active state of formin in actin assembly (with G- and/or F-actin, orSpire KIND).

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 693 �

vivo functions of the latter remain however complex anddebated as we will see in the last section [Renault et al.,2008; Qualmann and Kessels, 2009; Carlier et al., 2011].All these machineries use different combinations of foldedand disordered domains/proteins, and supposedly differentmolecular mechanisms of nucleation.

IDRs Found in NPFs of the WASP/Scar/WAVEFamily Regulate Labile Cooperative andAntagonist Interactions

NPFs of the WASP/Scar/WAVE family are variable multi-domain ABPs predicted to be extensively disordered (Fig.2A) [Guharoy et al., 2013]. Working as membrane-boundstimulus responsive ABPs, they elicit cycles of filamentbranching that generate a dendritic array of actin filamentsin many processes involving membrane deformations, suchas lamellipodia protrusions in migrating cells, phagocytosis,dendritic spine activity, spatial organization of the Golgi orpathogen infections [Campellone and Welch, 2010; Rott-ner et al., 2010]. The prevailing model of regulation ofWASP/Scar/WAVE ABPs is as follows. They exist in cells inconformational equilibria between inactive and active states.They are recruited on membrane and/or activated by phos-phoinositide lipids (PIP2 or PIP3), phosphorylations, SH2or SH3 (Src Homology 2/3) domain-containing proteins,or the active membrane-associated GTP-bound form ofRho small GTPases (Cdc42, Rac1,. . .). This leads to thedisruption of their autoinhibited conformation in an indi-vidual or cooperative manner, and make accessible theirconserved C-terminal constitutively active moiety, conven-tionally called V-C-A (Figs. 2B and 2C) [Campellone andWelch, 2010; Padrick and Rosen, 2010]. The C-terminalV-C-A region is composed of at least 3 IDRs of about 20–35 amino acids that are unfolded by themselves [Panchalet al., 2003; Kelly et al., 2006]: one to three WH2 domains(WH2/V), followed by a connector/central domain (C)and an acidic short extension (A). In this V-C-A region,each WH2/V domain can bind one G-actin molecule andis expected to deliver it to the actin-related Arp2 and/orArp3 subunits of the 220kDa Arp2/3 complex, while theC-terminal CA domains bind one molecule of Arp2/3 com-plex (Fig. 2B). The resulting ternary V-C-A-actin-Arp2/3complex associates with a preexisting filament after struc-tural rearrangements in Arp2/3 complex to form thebranched junction (Fig. 2C). Many of the molecular mech-anisms involved in the branching reaction remain elusive,but few facets of the NPF IDR regulations have been high-lighted at different stages of the process.

In WASP/N-WASP or WAVE proteins, the IDRs regu-late different conformational equilibria between the NPFinactive and active states. The archetypal members of thefamily, WASP and N-WASP, are autoinhibited by interdo-main interactions. Their full autoinhibitory interfacesremain unresolved but an important autoinhibitorymiddle-to-tail arrangement has been identified using a

small domain fusion between the middle GTPase-bindingdomain (GBD) and C-terminal C region of WASP (Fig.2B) [Kim et al., 2000]. The GBD IDR is like V-C-Alargely unstructured by itself in its free state [Kim et al.,2000]. In this minimal reconstituted autoinhibited state ofthe NPF, WASP GBD and C region interact with eachother to form a small compact folded assembly consistingof a short b-hairpin and five a-helices (Fig. 3A) thatoccludes partly Cdc42-GTP and Arp2/3 binding interfaceson each functional ID segments (Fig. 4) [Kim et al., 2000;Panchal et al., 2003; Hemsath et al., 2005; Kelly et al.,2006]. The resulting coupled folding and binding of thetwo distant GBD and C IDRs display a modest affinity cor-responding to a Kd of �1mM [Peterson et al., 2004]. Acti-vators relieve WASP/N-WASP autoinhibition allostericallyby interacting with the NPF middle segments (B, GBD,and/or PRD IDRs, Fig. 2B), while the weak affinity of theC/CA:Arp2/3 interaction (Kd>200mM) is inefficient aloneto compete with the autoinhibited GBD:C interface [Kellyet al., 2006]. Binding of Cdc42-GTP to the middle GBDhas thus been observed to cause a dramatic conformationalchange of the inactive GBD structure, resulting in disrup-tion of the hydrophobic core of the GBD:C contacts andrelease of the C-terminal C region (Figs. 3B, 3C, and 4)[Kim et al., 2000]. The GBD also acts as a signal integrator.Indeed, several WASP/N-WASP activators including PIP2,Cdc42-GTP, and phosphorylation-mediated SH2 domaininteract directly with or next to the GBD and activateWASP/N-WASP synergistically [Campellone and Welch,2010; Padrick and Rosen, 2010]. A cluster of basic residuesknown as the polybasic (B) region is located at the GBDN-terminus (Fig. 2A) and provides an electrostatic steeringmechanism of Cdc42-GTP recognition [Hemsath et al.,2005]. B is also essential for multiple PIP2 headgroup bind-ing and WASP/N-WASP membrane recruitment (Fig. 4).In the inactive state, the distant B and A regions arebrought into close proximity by the GBD:C interaction.They may interact with each other in an ordered or disor-dered mode and extend the middle-to-tail interdomainautoinhibition to a larger B-GBD:C-A functional interface.NPFs are additionnally activated or inhibited followingphosphorylations in response to various stimuli, associatedfor example with growth factors or substrate adhesion[Padrick and Rosen, 2010; Mendoza, 2013]. In WASP/N-WASP, these local modifications target residues locateddirectly within their GBD or C IDRs or close to their auto-inhibitory interface. Src-family kinases phosphorylate aconserved tyrosine sidechain within WASP/N-WASP GBD(Y291/Y256), and the casein kinase II (CKII) two serines atthe end of their C region (S483S484/S484S485) (Fig. 4).WASP/N-WASP phosphorylations at Y291/Y256 activatethe NPFs in a two-step process. In the autoinhibitedGBD:C structure, Y291 is buried in one of the helix form-ing the five helical arrangement of the inactive interfaceformed by the two IDRs. After phosphorylation, a binding

� 694 Renault et al. CYTOSKELETON

motif surrounding Y291 and including at least 5 to 7 resi-dues [Huang et al., 2008] is expected to be recognized bythe SH2 domains of Src-family kinases (Fig. 4). Binding ofSH2 should promote and stabilize an extended conforma-tion of the motif within the GBD, incompatible with theautoinhibited GBD:C interface. Altogether these data showthat WASP/N-WASP IDRs can integrate multiple interac-tions (Figs. 2B and 4) and conformations (Figs. 3A–3C),and, consequently, control cooperative middle-to-tail auto-inhibition and synergistic activation of the NPFs. Theseinterfacial and regulatory properties of WASP/N-WASPIDRs seem to be conserved with Scar/WAVE/WASH IDRs.

Scar/WAVE/WASH proteins display a domain organiza-tion slightly different from WASP/N-WASP (Fig. 2A) and

are autoinhibited within a pentameric WAVE/WASH regu-latory complex (WRC) (Fig. 2C). The crystal structure of aminimal WRC that contains half of WAVE1 sequence hasbeen recently solved in its inactive state, revealing a largerfunctional autoinhibitory interface of the NPFs than forWASP [Chen et al., 2010]. To solve by crystallography thepentameric and multidomain WRC of 380 kDa, Chen andcoworkers reconstituted a functional minimal recombinantWRC optimized for structural homogeneity after trunca-tions of WAVE1 PRD and Abi2 PRD-SH3 flexible regions.The region of WAVE1 between its N-terminal structuredSHD (Scar/WAVE Homology Domain) and middle poly-basic (B) region is predicted to contain significant disorder(Fig. 2A). It has been called the meander region because

Fig. 3. Different known conformations of IDRs in WASP and WAVE1 inactive and active states and the mutually exclusivestructural interfaces they control between the two states. (A) In WASP autoinhibited state, its middlle GBD (blue ribbon/surfaceand yellow ribbon) and C-terminal C (red ribbon) IDRs fold with each other to form a small compact folded assembly, whichoccludes partly Cdc42-GTP and Arp2/3 binding interfaces [Kim et al., 2000]. The side chains of C residues involved in hydrophobicand polar interactions at the interface are shown in green (L466, V467, L470, V473, Q475 and R477). (B) Activated, extendedWASP-GBD (yellow/blue) bound to Cdc42-GTP (brown surface), with WASP-GBD shown in the same orientation as in the rightpanel of (A) [Abdul-Manan et al., 1999]. Only part of the C-terminal fragment of the GBD (in blue) is depicted since it unfoldsupon binding of the GBD to Cdc42-GTP. (C) Superimposition of the two different conformations of WASP GBD bound to C(orange) or Cdc42-GTP (yellow). (D) The autoinhibited interface of WAVE1 (green ribbon) in the crystal structure of the minimalWAVE1 regulatory complex (WRC) [Chen et al., 2010]. The surfaces of Sra1, Nap1, HSPC300, and Abi2 are represented in cyan,white, pink, and orange, respectively. Disordered or missing (WAVE1 PRD) regions in the minimal WRC crystal structure are shownwith green dashed lines. WAVE1 extended meander region (a2–a6) is shown in light green, and ordered basic residues of the Bregion in blue. The side chains of phosphorylation sites that are sufficient to induce allosterically the V-C-A release, leading to WRCactivation, are shown in red. (E) The extended conformation of WAVE2 WH2 domain (in red) bound to G-actin-ATP (white sur-face) [Chereau et al., 2005]. Actin subdomains 1 to 4 are indicated. Disordered regions are shown with dashed lines. The alternativeconformation of WAVE1 WH2 domain in its inhibited state within the WRC is superimposed in green with an orientation veryclose to the one displayed in (D). The two structural interfaces with either Sra1 (cyan surface in D) or G-actin (here) are mutuallyexclusive.

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 695 �

within the inactive WRC, it adopts an extended, meander-ing folded conformation in contact of Sra1 protein. Themeander region contains WAVE1 a-helices a2 to a6, withB starting at a6 C-terminus (Fig. 3D) [Chen et al., 2010].The consecutive middle meander-B regions appear to coor-dinate many interactions in the WRC, like the B-GBDIDRs in WASP/N-WASP. In the inactive WRC structure,the C-terminal V-C-A region remains 40% disordered,with A being fully disordered. The V/WH2 domain ofWAVE1 adopts an helical fold stabilized by both intra- andintermolecular interactions, with WAVE1 a2 helix andSra1 protein, respectively (Fig. 3D). This interface occludesthe WH2 residues which contribute most to G-actin inter-

face (Figs. 3E and 4) [Chereau et al., 2005; Gaucher et al.,2012]. In parallel, the C region of WAVE1 folds into anamphipatic a-helix as in WASP GBD:C. This folding isalso stabilized by both intramolecular interactions withWAVE1 a6 helix and intermolecular interactions with Sra1protein (Fig. 3D). This interface masks C residues whichhave been shown by NMR to undergo a “disorder-to-ordertransition” upon Arp2/3 binding (Fig. 4) [Panchal et al.,2003; Kelly et al., 2006]. Finally, the disordered acidic (A)region at WAVE1 C-terminus is clearly positionned in the3D-interface close to the middle phosphatidylinositol-binding B region (Figs. 2C and 3D). Ordered or disorderedlong range electrostatic contacts between the two distant

Fig. 4. The B-GBD IDRs of WASP act as a molecular node or hub in the multidomain NPF by integrating many interactionsthat are mutually exclusive or cooperative. The amino acid sequences of the B-GBD and V-C-A IDRs from human WASP, N-WASP and WAVE1, or of the N-WASP binding IDR of E. coli EspFu are shown with different labellings that underline residuesdeeply involved in involved in different interdomain and protein–protein interfaces. Secondary structures and interface residues areextracted from the solution or crystal structures obtained in [Abdul-Manan et al., 1999; Kim et al., 2000; Chereau et al., 2005;Cheng et al., 2008, 2010; Gaucher et al., 2012]. Only residues burying a solvent-accessible surface area> 30 A2 in the above struc-tures are labeled. Arp2/3 interacting residues of WASP were identified by NMR line broadening experiments [Panchal et al., 2003].The interaction boundaries that are not clearly delimited are shown in dashed lines. The B-GBD IDRs of WASP/N-WASP control 2interdomain middle-to-tail autoinhibitory interactions, ie with C and possibly A IDRs (GBD:C and B:A), and 4 active interactions,ie with phosphoinositide lipids (B:PIP2), Cdc42-GTP (GBD:Cdc42-GTP), Src family kinases to phosphorylate Y291, and finally Srcfamily SH2 domains after Y291 phosphosphorylation (GBD C-term.:SH2). The inactive and active interactions are mutually exclu-sive, while active B:PIP2, GBD:Cdc42-GTP and GBD C-term.:SH2 interactions are cooperative. In enterohaemorrhagic E. coli infec-tion, the GBD IDR is targeted by the bacterial effector EspFu (GBD:EspFu) that outcompetes autoinhibited interactions and cellularactivators upon a tight binding to GBD, leading to a kind of N-WASP constitutive activation.

� 696 Renault et al. CYTOSKELETON

IDRs (B and A) may contribute to WAVE overall autoinhi-bition in the full-length WRC. Altogether the middlemeander and B regions integrate therefore all the functionalintramolecular inhibitory contacts with the C-terminal V-C-A (Fig. 2C). Interactions of the V-C-A, and perhaps thestructural assembly of the entire meander:V-C-A element,appear also to be highly cooperative, as perturbation ofeither the V- or C-region contacts in Sra1 or in WAVE1meander region produces WRC activity near that of the iso-lated V-C-A [Chen et al., 2010]. Mapping the binding areaof WAVE activators on the inactive WRC interface showsadditionally that the meander-B IDRs integrate or relaymany allosteric signals that release the V-C-A contacts andactivate synergistically the WRC. Rac1-GTP binds the inac-tive WRC with only a modest affinity (Kd �1–2 mM)[Chen et al., 2010] and seems to cooperate with other fac-tors for efficient membrane recruitment and activation ofthe WRC in actin assembly such as the small GTPase Arf1[Koronakis et al., 2011] or phosphophoinositide lipids.Rac1-GTP binding area has been mapped on Sra1 in apatch directly adjacent to a4-a6 of WAVE1 meanderregion and possibly on WAVE1 a6 as well (Fig. 3D), andthe binding of Rac1-GTP competes with the sequestrationof V-C-A in WRC [Chen et al., 2010]. Interactions ofRac1 appear therefore to trigger conformational changes inthe meander region and/or its contact site on Sra1. Simi-larly, the meander region is modulated by phosphorylationsof Src, Abl kinases or Cdk5 at WAVE1 Tyr125 and Tyr151and Thr138, respectively [Chen et al., 2010; Mendoza,2013]. These residues are involved in inter- or intramolecu-lar contacts of the meander inactive conformation in WRCbut away from the V-C-A sequestering interface (Fig. 3D).A single phospho-mimetic mutant of either Tyr125 orTyr151 in WRC is however sufficient to activate the com-plex both in vitro and in vivo [Chen et al., 2010]. Otherphosphorylations on WAVE proteins that activate or inhibitthe WRC [Mendoza, 2013] are localized on the neighbour-ing IDRs which are absent or disordered in the crystalstructure, ie the PRD or A region. Finally, the conformationor stability of helix a6 C-terminus in the inactive WRCcan be as well modified by the binding of phosphoinositidelipids to WAVE B region, the latter being essential formembrane recruitment of the WRC and formation oflamellipodia in cells [Oikawa et al., 2004]. Altogether thedata suggest that the entire meander-B:V-C-A elementfunctions as a cooperative labile folding block in WRC tocontrol reversible interdomain middle-to-tail autoinhibi-tion, and allosteric and synergistic activation, like the B-GBD:C-A IDR interface in WASP/N-WASP.

In conclusion, the IDRs of V-C-A-containing NPFsbehave as efficient signal integrators. Their bindings ofmoderate affinity allow the control of multiple conforma-tional equilibria that can be easily displaced by many part-ners in physiological or pathological processes. Duringbacterial intracellular infections, many bacteria subvert the

actin cytoskeleton of host cells by hijacking Arp2/3 medi-ated nucleation. These pathogens inject either their ownconstitutively active V-C-A-containing proteins [Rottneret al., 2010], or unrelated proteins [Cheng et al., 2008;Aitio et al., 2012] or small molecules [Guenin-Mace et al.,2013] that disrupt the labile autoinhibited state of WASP/Scar/WAVE proteins. Enterohaemorrhagic Escherichia colibacteria for example translocate EspFu into host cells dur-ing their intestinal colonization, to induce formation ofactin-rich “pedestals” beneath the bacteria bound to humanintestinal epithelial cells. EspFu contains 6 repeats of a 27-residue hydrophobic segment and a 20-residue proline-richsegment (Fig. 4) and is disordered in its native state [Aitioet al., 2012]. The N-terminal hydrophobic segment of eachrepeat activates N-WASP by mimicking structurally theautoinhibitory C:GBD interaction and competitively dis-placing C thanks to a higher affinity for GBD in the nano-molar range (Fig. 4) [Cheng et al., 2008]. The analysis ofall the above structural interfaces involved in physiologicalor pathological processes show overall that the structuralplasticity and dynamics of NPF IDRs allow to accommo-date both autoinhibited compact and active extended con-formations, and to drive variable and specificconformations with different targets. These alternativeinteractions can be efficiently regulated by single phosphor-ylations. The IDRs act as interaction hubs in these modularABPs (Table I). In WASP/N-WASP for example, the con-tinuous B and GBD IDRs coordinate six physiological andone pathological interfaces that are mutually exclusive orcooperative and regulate the NPF inactive and active states(Fig. 4). Additional interactions likely occurs, that will beuncovered when the NPF full autoinhibitory interfaces ormore complete active interfaces will be analyzed. The shortadjacent binding sites of physiological activators with B-GBD may work cooperatively by stabilizing altogether the90 amino acid segment in an extended conformation, eachadditional binding making the conformational transitiontowards the B-GBD:C-A autoinhibitory compact foldingless reversible. Future studies will delineate the conforma-tional dynamics of these mutliple interfaces involving sev-eral IDRs, and how, after V-C-A release, the inherentflexibility of IDRs in the resulting ternary V-C-A-actin-Arp2/3 complex may be important as well for promotingall the dynamic molecular arrangements required in thebranching process.

Multi-Functionality of IDRs Found in Forminsat the C-terminus of Their Main CatalyticActin-Binding FH2 Domain

Unlike Arp2/3 and VCA containing NPFs, formin proteinsnucleate new filaments that are unbranched. They thusassemble diverse cellular structures, including the cytoki-netic contractile ring, polarized actin cables, stress fibersand filopodia [Goode and Eck, 2007; Campellone andWelch, 2010]. Based on their different modular

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 697 �

architectures, formins have been classified into at least sevenmain subclasses in metazoans. The modular architecture offour representative subfamilies is reported in Fig. 2A. Thebasic catalytic units that identify formins in actin assemblyis related with the presence of a formin homology 1 (FH1)and 2 (FH2) domain. Both are located in the C-terminalpart of formins while other domains are more variable andadapted to different cell cues (Fig. 2A). Several formin sub-families contain for example N-terminally a GBD IDR thatcontributes to their activation on membrane upon bindingto activated Rho GTPases. FH2 domains form flexibledonut-like dimers that can accelerate de novo filamentnucleation from G-actins by stabilizing two G-actinbetween their arms in an arrangement mimicking the lateralactin:actin contacts of filaments and by providing smallerbinding interfaces with two other monomers per FH2 pro-tomer. FH2 dimers also bind to filament barbed-end andprotect it from barbed-end CPs that stop elongation. Infull-length proteins, FH2 dimers can translocate with thebarbed-end as it elongates, acting as processive elongatingfactors. Additional high-resolution structural studies arerequired to understand the translocation of formins or thecompetition between formin and barbed-end CPs at themost dynamic end of actin filaments. The crystal structuresof G-actin-bound FH2 dimers [Otomo et al., 2005;Thompson et al., 2013] and the cryo-electron microscopymodel of the heterodimeric CP bound to F-actin barbedend [Narita et al., 2006] suggest that free filament barbed-ends cannot accommodate a simultaneous binding of for-min and CP because filament-bound FH2 dimer or CPa=b heterodimer are expected to display overlapping bind-ing sites for the terminal actin barbed-faces. Intrinsicallydisordered FH1 domains are composed of multipleprofilin-binding polyproline segments. They catch anddeliver G-actin:profilin complexes to the FH2-boundbarbed-end to promote formin-mediated processive elonga-tion of growing barbed-ends. Recent biochemical and struc-tural data suggest that the catalytic machinery of differentformin subclasses is tripartite because it requires as wellIDRs located at the C-terminus of their FH2. These IDRs,which can be unrelated between formin subclasses, shareseveral functional properties with the IDRs of V-C-A-containing NPFs by regulating multiple interactions andconformational transitions between formin inactive andactive states (Table I).

IDRs at FH2 C-terminus regulate first interdomainhead-to-tail autoinhibition. The two subclasses ofDiaphanous-related formins (mDia1–3) and forminhomology domain proteins (FHOD1/3) are activated byRho family GTPases and contain a C-terminal Diaphanousautoregulatory domain (DAD) [Campellone and Welch,2010] identified as an IDR in disorder prediction server(Fig. 2A). Their N-terminus contains a GBD and Diapha-nous inhibitory domain (DID) that participate in autoinhi-bition. Their C-terminal DAD IDR binds the GBD–DID

to inhibit the actin polymerizing activity of the FH2domain (Fig. 2D, left panel) [Wallar et al., 2006]. Thisinhibitory interaction is disrupted by the binding of RhoGTPases to the GBD–DID [Lammers et al., 2005] andresults in activation of the C-teminal FH2 in vitro [Li andHiggs, 2005]. Structural comparisons of mDia GBD–DIDbound to either a DAD or to Rho GTPases indicate thatbinding of Rho and DAD is mutually exclusive, and theirbinding sites in the GBD–DID partially overlap explainingthe molecular basis for the allosteric activation [Campelloneand Welch, 2010]. Importantly, mDia activation onGTPase binding in vitro is incomplete, raising the possibil-ity that additional cellular factors may coalesce upon theautoinhibited GBD-DID interface for full activation like inV-C-A-containing NPFs. In that matter, the DAD ofFHOD1 or mDia2 is phosphorylated by ROCK (rho-kinase). This relieves the formin autoinhibition in vitro andpromotes FHOD1-mediated stress fibre assembly in cells[Takeya et al., 2008], or mDia2-mediated actin polymeriza-tion and smooth muscle cell-specific gene transcription[Staus et al., 2011]. Founding mammalian formins(FMN1/2 or Drosophila melanogaster Cappucino (Capu)),FMNL-3 or INF2 (Inverted Formins) formins display a dif-ferent overall modular architecture, lacking an N-terminalGBD (except for FMNL-3) and ending either with an FSI(Formin-Spire Interacting) IDR for FMN, or with WH2-DAD IDRs for FMNL-3 and INF2 (Fig. 2A). Both FMNFSI [Bor et al., 2012] and INF2 WH2-DAD [Chhabraet al., 2009] IDRs at their FH1-FH2 C-terminus areinvolved in interdomain head-to-tail autoinhibition with anunrelated N-terminal DID.

In addition to their role in interdomain autoinhibition,these IDRs control as well multiple interactions in forminactive states. They appear to be required for the catalyticnucleating and/or processive activity of their FH2 on fila-ment barbed-ends, and are sometimes involved in alterna-tive interactions with other ABPs (Fig. 2D). The DAD ofmDia1 interacts with G-actin and works in concert withthe FH2 to enhance nucleation without affecting the rate offilament elongation, and independently from the FH1domain (Fig. 2D, middle panel) [Gould et al., 2011]. TheWH2 and DAD containing C-terminus of INF2 [Chhabraand Higgs, 2006] or FMNL-3 [Heimsath and Higgs, 2012]act in synergy with their FH2 domain to accelerate actinpolymerization from monomers. Additionally, the C-terminal region of INF2 is required for actin filament sever-ing, while the C-terminal region of FMNL-3 enhances theotherwise modest severing activity of FMNL-3-(FH1-FH2)(Fig. 2D, right panel). The isolated C-terminus of FMNL-3 including its WH2-DAD IDRs binds both monomersand filament barbed-ends [Heimsath and Higgs, 2012].Finally, the C-terminal FSI of FMN-subfamily forminscooperates with their FH2 domain to accelerate actin poly-merization [Vizcarra et al., 2011], and interacts as well withthe N-terminal KIND (Kinase homology domain) of Spire

� 698 Renault et al. CYTOSKELETON

[Vizcarra et al., 2011; Zeth et al., 2011]. The molecularbases for these multiple regulations in actin assembly arenot yet understood. It remains to be determined if and howthe C-terminal FSI interacts with G- and/or F-actin in pres-ence of the FH2. The regulation with Spire is even morecomplex since Spire N-terminal half contains four consecu-tive WH2 domains next to its KIND domains that alto-gether display multiple activities in actin assembly withboth G- and F-actin as described below. As isolateddomain, the interaction of the N-terminal KIND of Spirewith the FMN FSI inhibits the acceleration of actin poly-merization induced by FMN FH2 and C-terminaldomains, while a longer N-terminal construct of Spire thatincludes its KIND and 4 consecutive WH2 domains,enhances in contrast synergistically the FMN-induced actinpolymerization [Quinlan et al., 2007; Vizcarra et al.,2011]. Future studies are required to elucidate the complexmultidomain regulation and synergy between Spire andFMN nucleating factors at filament barbed end, whichinvolve multiple IDRs in the two ABPs.

Adaptability and Allostery of IDRs Associatedwith the Capping Regulation of CP

Structurally unrelated CPs like the CP [Zwolak et al., 2010]or Eps8 [Hertzog et al., 2010] use C-terminal IDRs to regu-late their barbed-end capping activity. The regulatory molec-ular mechanisms of CP have been lately more extensivelycharacterized, including inhibitory mechanisms by disor-dered CARMIL peptides. CP facilitates the actin turnover inthe formation of various actin structures, including Arp2/3-mediated dendritic arrays, by reducing the overall number ofelongating barbed-ends. It is an a:b heterodimer. Its twofolded subunits of �30 kDa form a stable pseudosymmetricstructure with the shape of a mushroom, with a stalk and cap

(Fig. 5). NMR and crystallographic data have shown that theC-terminal �30 amino acids of the b-subunit, called the b-tentacle, are disordered [Zwolak et al., 2010; Takeda et al.,2011]. CP binds to the barbed-end of actin filaments withhigh affinity (Kd �0.1 nM) and terminates assembly anddisassembly at the end, i.e. caps the filament.

Several structures of CP, including a low resolution EMmodel of CP bound to filament [Narita et al., 2006], areavailable to approach the different steps of the capping reac-tion and the role of local disorder in CP at the atomic scale(Table I). Structural comparisons and molecular dynamicsimulations revealed that CP is composed of two rigiddomains (CP-L and CP-S) that undergo an intrinsic twist-ing motion relative to each other but do not directly corre-spond to the subunit interface [Takeda et al., 2010; Kimet al., 2012]. The current models describe the barbed-endcapping as following [Narita et al., 2006; Kim et al., 2010;Takeda et al., 2010, 2011; Kim et al., 2012]. CP first asso-ciates with filament barbed-end by the cap surface involvingsignificantly the a-subunit C-terminus (Fig. 5). CP thenfirmly binds to the barbed-end by two independent proc-esses. (1) The interaction with filament barbed-end inducesa particular bending and twisting motion of the cap. (2)The b-tentacle IDR can eventually reach the hydrophobiccleft between subdomain 1 and 3 of the terminal actin inorder to fold and bind to the barbed-end. In this interface,the IDR is therefore important for a dynamic and optimalbinding to the barbed-end. Its deletion induces a two-foldreduction of CP affinity for barbed-ends [Kim et al., 2010].Local disorder in the b subunit C-terminus may as well beimportant to accommodate the conceivable conformationaldynamics of the two last actin subunits at barbed-ends. Inthat regard all folded proteins that cap filament barbed-ends such as Gelsolin, twinfilin, or formin, are similarly

Fig. 5. Structural basis for the regulation of two unrelated IDRs associated with the capping activity of CP at filamentbarbed-end. The CP is an intrinsically twisting, pseudosymmetric a:b heterodimer with the shape of a mushroom, including a stalkcontaining a and b N-terminii and a cap containing their C-terminii. CP binds to F-actin barbed-end with its “mushroom cap” andthe Cterminal IDR of the b-subunit (b-tentacle) is important in the dynamic and optimal binding to the filament barbed-end (step2). In contrast, disordered CARMIL peptides perturb allosterically the overall optimal bending and twisting of CP heterodimerinduced by its barbed-end binding. CARMIL IDRs bind to CP “mushroom stalk” across its two CP-L and CP-S twisting domains[Takeda et al., 2010; Kim et al., 2012].

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 699 �

composed of folded domains including or connected byflexible or disordered linker/regions.

Recent structural data on CP bound to several inhibitoryCARMIL peptides illustrate further how disorder partici-pates as well in the allosteric inhibition of CP. The modularCARMIL protein facilitates the dissociation of CP from thebarbed-end (uncapping activity) mainly via a conservedshort CP-binding motif (CPB) of �20 amino acids presentin the otherwise unrelated proteins like CD2AP and CKIP-1. The CPB is surrounded by flanking regions that contrib-ute to the overall binding affinity to CP and extend theinteractions more significantly on the peptide C-terminalside. These CARMIL peptides are intrinsically disordered.Several crystal structures show that CARMIL peptides wraparound the stalk of the mushroom-shaped CP at a site dis-tant from the actin-binding interface on the top of themushroom cap (Fig. 5) [Hernandez-Valladares et al., 2010;Takeda et al., 2010, 2011]. Molecular dynamic simulationssuggest that CARMIL peptides inhibit CP allosterically byaltering the conformational twisting flexibility between CP-L and CP-S from the mushroom stalk binding [Takedaet al., 2010, 2011; Kim et al., 2012]. The CPB-binding issufficient to induce this allosteric mechanism, while the C-terminal flanking region in CD2AP and CARMIL appearonly to increase the overall peptide stability and affinitywith CP, suggesting a decoupling between the regionsresponsible for binding specificity (stronger contribution offlanking regions than CPB) and function (e.g., CPB)[Takeda et al., 2011]. The flexible conformation of theCARMIL peptides was also proposed to be potentiallyimportant for intrusion into the mobile binding grooveacross the two twisting rigid domains CP-L and CP-S[Takeda et al., 2011].

Complex Multifunctional Regulationsof Intrinsically Disordered Repeatsin ABPs

The preponderance of repeats in intrinsically disorderedsequences is high and may represent a way to expand IDPfunctions in evolution [Tompa, 2003]. Repeats can be eitherfunctionally equivalent or nonredundant, or introduce novelfunctions that emerge as the repeat number exceeds a thresh-old level. Similarly, many ABPs contain diverse modularorganizations of repeated IDRs as emphasized for WH2/bTdomains in Fig. 1A. Their molecular mechanisms of regula-tions remain largely poorly understood.

Nonredundant Repeats of FH1 PolyprolineStretches Work in Synergy in Formin to Speedthe Elongation Rate of FH2-AssociatedBarbed-Ends

In formin, the FH1 domain in presence of profilin speedsthe elongation of individual barbed-ends associated with

FH1-FH2-Cterminus domains, often above the diffusion-limited rate of elongation of free barbed-ends [Paul andPollard, 2009]. FH1 domains of most formins lie N-terminal to their FH2 domain and are predicted to bemostly disordered except for multiple discrete stretches(often between 2 and 8) of contiguous proline residues,that can transiently form type-II polyproline helices. Eachpolyproline track displays individually weak affinity forfree or actin-bound profilin, with Kd values varying from�5 to 20 mM for stretches that include 13 to 20 prolinesto >1000 mM for stretches including only 5 to 8 prolines[Perelroizen et al., 1994; Petrella et al., 1996; Courte-manche and Pollard, 2012]. Transfer of actin monomersto the barbed-end of actin filament is established by dif-fusion and closure of FH1 loops filled with profilin:G-actin on the FH2-bound barbed-end (Fig. 2D, middlepanel). This enables direct addition of actin monomers tothe FH2-bound barbed-end and may promote transloca-tion of one of the FH2 protomer while the full FH2dimer remains processively attached to the end. Recentdata show that in presence of profilin, the rates of elonga-tion of filaments associated with the FH1-FH2-Cterminusof yeast formin Bni1p increased with the total number ofFH1 polyproline tracks [Paul and Pollard, 2008]. How-ever, each FH1 polyproline track contributes differentlyand is nonredundant along the FH1 sequence [Paul andPollard, 2008; Courtemanche and Pollard, 2012]. Theircontribution to the barbed-end elongation rate varyaccording to their distance in the FH1 sequence to theFH2, and to their binding strength for profilin:actin.Their distance to the FH2 domain determines the volumeexplored by diffusing profilin-actin bound to a polypro-line track before contacting the barbed-end. The affinityof each polyproline track in the FH1 domain of Bni1pand possibly of other formin for profilin-actin is tunedrelative to its distance to the FH2 domain [Courtemancheand Pollard, 2012]. Near the FH2 domain, FH1 polypro-line tracks contain a smaller number of prolines in orderto display lower affinity for profilin, which favors rapidand dynamic G-actin transfer to the barbed-end uponFH1 loop closure. Far from the FH2, they contain incontrast a larger number of prolines to display higheraffinity for profilin (slower dissociation rates) and to com-pensate for the slower loop closure rates. In conclusion,the individual weak affinity of these IDRs allows drivingdynamic interactions at the elongating barbed-end. TheFH1 disorder state may provide a larger capture radiusfor profilin:G-actin leading to faster on-rates for bindingas it has been proposed for IDRs involved in DNA bind-ing [Pontius, 1993; Shoemaker et al., 2000]. The repeatsspeed the overall elongation mediated by formin, and thespecificity of each repeats along the FH1 sequence opti-mizes G-actin transfer efficiency with the number ofrepeats or the length of the flexible loop relative to itsdistance to the FH2 (Table I).

� 700 Renault et al. CYTOSKELETON

Non-Redundant Repeats of WH2/ßT IDRsCooperate in Several ABPs to Regulate theDynamics of G- and/or F-Actin Assembly/Disassembly

Some modular organizations containing repeats of actin-binding WH2/ßT IDRs extend significantly the panoply offunctions regulated by single WH2/ßT domains in actinassembly (Table I) [Renault et al., 2008; Husson et al.,2010; Carlier et al., 2011]. The specificities and synergy ofmultiple ßT domains were highlighted in C. elegans Tet-raThymosinß that contains 4 consecutive ßT domains andbinds multiple G-actins in vitro [Van Troys et al., 2004].Mutants with only one repeat intact or that have only onerepeat rendered unfunctional or less active were analyzed inG/F-actin-binding and polymerization assays. Some repeatswere thus shown to exhibit specificity for either G-actin(repeats 1 and 4), F-actin (repeat 3) binding or both (repeat2). Together the four repeats cooperate to display the fullbiochemical activity of TetraThymosinß in G- and F-actinassembly regulation, including its desequestering ability inpresence of Tß4 [Van Troys et al., 2004]. Spire [Quinlanet al., 2005], Cordon-bleu [Ahuja et al., 2007], V-C-A-containing JMY [Zuchero et al., 2009] or the pathogenVopF/VopL effectors [Liverman et al., 2007; Tam et al.,2007] contain different modular organizations with 3 to 4WH2/ßT repeats that accelerate de novo filament nucleationfrom pure G-actin. They have been then proposed to repre-sent a new subfamily of actin nucleators. The specificitiesthat confer nucleation for these repeats but not in others arenot yet well understood. In Spire, the 4 contiguous WH2bind cooperatively 4 G-actin with a Kd of �0.15 mM4

[Bosch et al., 2007] but are functionally not redundant fornucleation. Spire last two WH2 are more determinant fornucleation than its first two domains [Quinlan et al., 2005].In Cobl containing 3 WH2, nucleation of pure G-actinrelies mainly on the combination of its first WH2 associatedwith a predicted disordered Lys-rich region, located directlyN-terminal to the WH2 [Husson et al., 2011]. Severalobservations suggest that nucleation may not represent theprimary in vivo function of these repeats [Renault et al.,2008; Carlier et al., 2011]. First, compared to other knownactin nucleators like formin, WH2 repeats of Spire or Cobldisplay a relatively weak nucleating efficiency in vitro in viewof their affinity for G-actin, suggesting that only a very smallportion of their complexes with G-actin monomers behaveas nuclei. Their complexes appear to exist in solution inequilibrium with other main functional states that regulateother activities such as profilin-like and/or sequestering activ-ities, including sequestration of G-actin-ADP after filamentdepolymerization [Bosch et al., 2007; Husson et al., 2011].Secondly, the nucleating activity of Spire or Cobl WH2repeats is abolished in vitro when G-actin is bound to profi-lin [Bosch et al., 2007; Husson et al., 2011] and in vivo,Spire regulates actin dynamics in many developmental proc-esses in concert with FMN-subfamily formin and profilin

[Manseau and Schupbach, 1989; Dahlgaard et al., 2007;Pfender et al., 2011]. An important clue to understand therole of WH2/ßT repeats in ABPs may therefore rely on theircapacity to interact jointly with the side and/or barbed-endof filaments, which is otherwise weak with individual WH2.The 4 WH2 of Spire bind to filament barbed-ends with agood affinity, i.e. with an equilibrium dissociation constantof approximately 0.010 mM [Bosch et al., 2007]. The 4WH2 repeats of Spire, or first two WH2 domains of Coblassociated with their N-terminal Lysine-rich motif, modulateboth the dynamics of assembly and disassembly of filaments.They regulate filament severing with different efficiencies,and Spire regulates also barbed-end capping in presence ofprofilin [Bosch et al., 2007; Carlier et al., 2011; Hussonet al., 2011]. Similarly, the presence of two consecutive WH2in the V regions of N-WASP introduce a capacity to interactwith filament barbed-end [Gaucher et al., 2012], and ahigher branching efficiency in N-WASP V-C-A than inWASP or WAVE V-C-A [Yamaguchi et al., 2000]. Finally, inthe V-C-A of JMY, the function of its three consecutiveWH2 repeats is further extended by integrating an overlapingimportin-a-binding nuclear-localizing sequence [Zucheroet al., 2012]. The mutually exclusive interactions between G-actin and importin-a=b bindings to the repeats regulateJMY cytoplasmic or nuclear localization, and consequentlythe nuclear transcription induced by the protein with G-actin cytoplasmic concentration and binding (Table I). Thisis similar to the regulation of RPEL-motifs that are othersmall, intrinsically disordered G-actin-binding domains pres-ent as repeats in several transcriptional co-factors such asMAL/MRTF-A or Phactr1 [Mouilleron et al., 2008, 2012].In conclusion, the intrinsic conformational dynamics of thesesmall actin binding domains allow them in repetition tointeract cooperatively with multiple G-actin-ATP/ADP and/or with the side or barbed-end of filaments. They can thusact as versatile or multifunctional regulators of the G-actinpool, and/or of the dynamics of assembly and disassembly offilaments, regulating multiple key activities in actin assemblysuch as G-actin nucleation, unidirectional assembly orsequestration, filament severing or capping. What are themain activities relevant in vivo for each modular organiza-tions with WH2/ßT repeats in presence of other ABPs, suchas FMN-subfamily formin and profilin for Spire, remainmajor issues to be solved. These investigations will requireto understand better what are the specificities at the level ofeach individual WH2/ßT, the synergy of their differentassociations, and the exact number of WH2/ßT which sup-port each of their activities in actin assembly. For decipher-ing the molecular bases of their versatility ormultifunctionality, it will be essential to analyze differentconformations and the dynamics of their complexes whichmay include local disorder and fuzzy complexes like it hasbeen observed at the level of their elementary functionalunit with a single ßT bound to one G-actin [Didry et al.,2012].

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 701 �

Concluding Remarks

The structural and functional properties of IDRs involvedin actin cytoskeletal remodeling begin only to emerge inthe light of few complementary studies that dissect andapproach their conformational dynamics in simplifiedmolecular and functional models of multimodular ABPsand actin assembly dynamics. These studies suggest thatactin cytoskeletal IDRs are especially designed to relaymany labile interactions in ABPs (Table I).

IDRs of ABPs share important properties of IDRsinvolved in signal transduction or transcription regulation.Thanks to their inherent structural plasticity and dynamics,they can accommodate both autoinhibited compact andactive extended conformations, and adopt variable confor-mations with different and structurally unrelated targets(Fig. 3). By adopting extended conformations, they canbuild highly specific interfaces with shorter sequences thanstructured globular domains [Gunasekaran et al., 2003].The folding upon binding processes provides high specific-ity with relatively low affinity due to the entropic cost asso-ciated with the disorder-to-order transition [Dyson andWright, 2005]. Small but highly specific functional seg-ments are adapted for relaying reversible and transientinteractions at macromolecular interfaces in actin self-assembly dynamics. Disorder may be partially preservedand functional in their bound states [Didry et al., 2012].Intrinsic conformational flexibility facilitates access toenzymes that introduce or remove post-translational modi-fications. In bound states, intrinsic conformational dynam-ics and moderate binding strengths enable completeremodelings of macromolecular interfaces after single post-translational modifications like phosphorylations (WAVE1meander region, WASP/N-WASP/WAVE1 C-A, forminFHoD1/mDia2 DAD,. . .) or possibly lysine acetylations(Tb4/10).

The conformational variability and adaptability of IDRsand their regulations via diverse post-translational modifica-tions allow them to act as interaction hubs in ABPs. Theirintrinsically disordered sequence can integrate multipleoverlapping binding sites of different domains and/or pro-teins and regulate many mutually exclusive interactions(Fig. 4). They thus regulate reversible switches between theinactive and active states of modular ABPs, as it has beenhighlighted here for V-C-A containing NPFs or differentformin subfamilies (Fig. 2). In these ABPs, one or severaldistant IDRs control interdomain head/middle-to-tail auto-inhibition and consequently allosteric activation. Theantagonistic interactions induced by overlapping bindingsites on IDRs may also be between different partners in theactive states of ABPs (Figs. 2, 4, and Table I). In macromo-lecular interfaces, multiple consecutive IDRs or long IDRsthat can integrate distinct nonoverlapping binding regions,regulate cooperative interactions with their targets. WASP/N-WASP B-GBD IDRs or WAVE1 meander-B IDRs inte-

grate many activating signals/interactions that each modifiesthe conformational equilibria between the inactive andactive states of the IDRs and thus synergize or cooperate toactivate the NPF. Similarly, multiple WH2/bT domains inSpire or Cordon-Bleu introduce an efficient regulation ofthe dynamics of assembly and/or disassembly of filamentsby a cooperative binding of several flexible actin-bindingIDRs with several actin subunits of filaments [Bosch et al.,2007; Renault et al., 2008; Carlier et al., 2011; Hussonet al., 2011]. Finally, IDRs display often poor sequence sig-nature in correlation with the fact that their sequence con-tain less folding constraints than intrinsically structureddomains. The multiple regulations we have highlightedhere for IDRs in modular ABPs further explain why knownfunctional IDRs like actin-binding WH2/bT display highlyvariable sequences between ABPs. By controlling the auto-inhibited states of modular proteins and recognizing severalpartners in their active states, the functional sequence ofeach IDR is specifically engineered in each ABP to permit afine-tuning of multiple intra- and intermolecular interac-tions. Overall, the numerous and labile interactions ofIDRs in modular ABPs contribute to explain how multido-main proteins can integrate and coordinate efficiently mul-tiple signals and interactions with high turnover in actinself-assembly dynamics. A great deal of work remains to bedone to further elucidate the full panoply of IDR regula-tions, in a larger number of ABPs, and in more completeand integrated systems.

References

Abdul-Manan N, Aghazadeh B, Liu GA, Majumdar A, Ouerfelli O,Siminovitch KA, Rosen MK. 1999. Structure of Cdc42 in complexwith the GTPase-binding domain of the ’Wiskott-Aldrich syn-drome’ protein. Nature 399(6734):379–383.

Aguda AH, Xue B, Irobi E, Preat T, Robinson RC. 2006. Thestructural basis of actin interaction with multiple WH2/beta-thymo-sin motif-containing proteins. Structure 14(3):469–476.

Ahuja R, Pinyol R, Reichenbach N, Custer L, Klingensmith J, KesselsMM, Qualmann B. 2007. Cordon-bleu is an actin nucleation factorand controls neuronal morphology. Cell 131(2):337–350.

Aitio O, Hellman M, Skehan B, Kesti T, Leong JM, Saksela K,Permi P. 2012. Enterohaemorrhagic Escherichia coli exploits a tryp-tophan switch to hijack host f-actin assembly. Structure 20(10):1692–1703.

Bisi S, Disanza A, Malinverno C, Frittoli E, Palamidessi A, Scita G.2013. Membrane and actin dynamics interplay at lamellipodia lead-ing edge. Curr Opin Cell Biol 25(5):565–573.

Bodart JF, Wieruszeski JM, Amniai L, Leroy A, Landrieu I,Rousseau-Lescuyer A, Vilain JP, Lippens G. 2008. NMR observa-tion of Tau in Xenopus oocytes. J Magn Reson 192(2):252–257.

Bor B, Vizcarra CL, Phillips ML, Quinlan ME. 2012. Autoinhibi-tion of the formin Cappuccino in the absence of canonical autoin-hibitory domains. Mol Biol Cell 23(19):3801–3813.

Bosch M, Le KH, Bugyi B, Correia JJ, Renault L, Carlier MF.2007. Analysis of the function of Spire in actin assembly and itssynergy with formin and profilin. Mol Cell 28(4):555–568.

� 702 Renault et al. CYTOSKELETON

Campellone KG, Welch MD. 2010. A nucleator arms race: cellularcontrol of actin assembly. Nat Rev Mol Cell Biol 11(4):237–251.

Carlier MF, Jean C, Rieger KJ, Lenfant M, Pantaloni D. 1993.Modulation of the interaction between G-actin and thymosin beta4 by the ATP/ADP ratio: possible implication in the regulation ofactin dynamics. Proc Natl Acad Sci USA 90(11):5034–5038.

Carlier MF, Didry D, Erk I, Lepault J, Van Troys ML,Vandekerckhove J, Perelroizen I, Yin H, Doi Y, Pantaloni D. 1996.Tbeta 4 is not a simple G-actin sequestering protein and interactswith F-actin at high concentration. J Biol Chem 271(16):9231–9239.

Carlier MF, Husson C, Renault L, Didry D. 2011. Control of actinassembly by the WH2 domains and their multifunctional tandemrepeats in Spire and Cordon-Bleu. Int Rev Cell Mol Biol 290:55–85.

Cheng HC, Skehan BM, Campellone KG, Leong JM, Rosen MK.2008. Structural mechanism of WASP activation by the enterohae-morrhagic E. coli effector EspF(U). Nature 454(7207):1009–1013.

Chen Z, Borek D, Padrick SB, Gomez TS, Metlagel Z, Ismail AM,Umetani J, Billadeau DD, Otwinowski Z, Rosen MK. 2010. Struc-ture and control of the actin regulatory WAVE complex. Nature468(7323):533–538.

Chereau D, Kerff F, Graceffa P, Grabarek Z, Langsetmo K,Dominguez R. 2005. Actin-bound structures of Wiskott-Aldrichsyndrome protein (WASP)-homology domain 2 and the implica-tions for filament assembly. Proc Natl Acad Sci USA 102(46):16644–16649.

Chhabra ES, Higgs HN. 2006. INF2 Is a WASP homology 2motif-containing formin that severs actin filaments and acceleratesboth polymerization and depolymerization. J Biol Chem 281(36):26754–26767.

Chhabra ES, Ramabhadran V, Gerber SA, Higgs HN. 2009. INF2is an endoplasmic reticulum-associated formin protein. J Cell Sci122(Pt 9):1430–1440.

Choudhary C, Kumar C, Gnad F, Nielsen ML, Rehman M,Walther TC, Olsen JV, Mann M. 2009. Lysine acetylation targetsprotein complexes and co-regulates major cellular functions. Science325(5942):834–840.

Courtemanche N, Pollard TD. 2012. Determinants of ForminHomology 1 (FH1) domain function in actin filament elongationby formins. J Biol Chem 287(10):7812–7820.

Crooks GE, Hon G, Chandonia JM, Brenner SE. 2004. WebLogo:a sequence logo generator. Genome Res 14(6):1188–1190.

Dahlgaard K, Raposo AA, Niccoli T, St Johnston D. 2007. Capuand Spire assemble a cytoplasmic actin mesh that maintains micro-tubule organization in the Drosophila oocyte. Dev Cell 13(4):539–553.

De La Cruz EM, Ostap EM, Brundage RA, Reddy KS, SweeneyHL, Safer D. 2000. Thymosin-beta(4) changes the conformationand dynamics of actin monomers. Biophys J 78(5):2516–2527.

Didry D, Cantrelle FX, Husson C, Roblin P, Moorthy AM, PerezJ, Le Clainche C, Hertzog M, Guittet E, Carlier MF and others.2012. How a single residue in individual beta-thymosin/WH2domains controls their functions in actin assembly. EMBO J 31(4):1000–1013.

Domanski M, Hertzog M, Coutant J, Gutsche-Perelroizen I,Bontems F, Carlier MF, Guittet E, van Heijenoort C. 2004. Cou-pling of folding and binding of thymosin beta4 upon interactionwith monomeric actin monitored by nuclear magnetic resonance. JBiol Chem 279(22):23637–23645.

Dominguez R. 2009. Actin filament nucleation and elongation fac-tors–structure-function relationships. Crit Rev Biochem Mol Biol44(6):351–366.

Ducka AM, Joel P, Popowicz GM, Trybus KM, Schleicher M,Noegel AA, Huber R, Holak TA, Sitar T. 2010. Structures of actin-bound Wiskott-Aldrich syndrome protein homology 2 (WH2)domains of Spire and the implication for filament nucleation. ProcNatl Acad Sci USA 107(26):11757–11762.

Dyson HJ, Wright PE. 2005. Intrinsically unstructured proteinsand their functions. Nat Rev Mol Cell Biol 6(3):197–208.

Eadie JS, Kim SW, Allen PG, Hutchinson LM, Kantor JD, ZetterBR. 2000. C-terminal variations in beta-thymosin family membersspecify functional differences in actin-binding properties. J Cell Bio-chem 77(2):277–287.

Egile C, Loisel TP, Laurent V, Li R, Pantaloni D, Sansonetti PJ,Carlier MF. 1999. Activation of the CDC42 effector N-WASP bythe Shigella flexneri IcsA protein promotes actin nucleation byArp2/3 complex and bacterial actin-based motility. J Cell Biol146(6):1319–1332.

Fabrega C, Shen V, Shuman S, Lima CD. 2003. Structure of anmRNA capping enzyme bound to the phosphorylated carboxy-terminal domain of RNA polymerase II. Mol Cell 11(6):1549–1561.

Fuxreiter M. 2012. Fuzziness: linking regulation to protein dynam-ics. Mol Biosyst 8(1):168–177.

Fuxreiter M, Tompa P. 2012. Fuzzy complexes: a more stochasticview of protein function. Adv Exp Med Biol 725:1–14.

Gaucher JF, Mauge C, Didry D, Guichard B, Renault L, CarlierMF. 2012. Interactions of isolated C-terminal fragments of neuralWiskott-Aldrich syndrome protein (N-WASP) with actin and Arp2/3 complex. J Biol Chem 287(41):34646–34659.

Goode BL, Eck MJ. 2007. Mechanism and function of formins inthe control of actin assembly. Annu Rev Biochem 76:593–627.

Gould CJ, Maiti S, Michelot A, Graziano BR, Blanchoin L, GoodeBL. 2011. The formin DAD domain plays dual roles in autoinhibi-tion and actin nucleation. Curr Biol 21(5):384–390.

Guenin-Mace L, Veyron-Churlet R, Thoulouze MI, Romet-Lemonne G, Hong H, Leadlay PF, Danckaert A, Ruf MT,Mostowy S, Zurzolo C and others. 2013. Mycolactone activation ofWiskott-Aldrich syndrome proteins underpins Buruli ulcer forma-tion. J Clin Invest 123(4):1501–1512.

Guharoy M, Szabo B, Martos SC, Kosol S, Tompa P. 2013. Intrin-sic structural disorder in cytoskeletal proteins. Cytoskeleton (Hobo-ken) in press; DOI: 10.1002/cm.21118.

Gunasekaran K, Tsai CJ, Kumar S, Zanuy D, Nussinov R. 2003.Extended disordered proteins: targeting function with less scaffold.Trends Biochem Sci 28(2):81–85.

Hannappel E. 2010. Thymosin beta4 and its posttranslational mod-ifications. Ann NY Acad Sci 1194:27–35.

Hansen MD, Kwiatkowski AV. 2013. Control of actin dynamics byallosteric regulation of actin binding proteins. Int Rev Cell MolBiol 303:1–25.

Heimsath EG, Jr., Higgs HN. 2012. The C terminus of forminFMNL3 accelerates actin polymerization and contains a WH2domain-like sequence that binds both monomers and filamentbarbed ends. J Biol Chem 287(5):3087–3098.

Hemsath L, Dvorsky R, Fiegen D, Carlier MF, Ahmadian MR.2005. An electrostatic steering mechanism of Cdc42 recognition byWiskott-Aldrich syndrome proteins. Mol Cell 20(2):313–324.

Hernandez-Valladares M, Kim T, Kannan B, Tung A, Aguda AH,Larsson M, Cooper JA, Robinson RC. 2010. Structural

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 703 �

characterization of a capping protein interaction motif defines afamily of actin filament regulators. Nat Struct Mol Biol 17(4):497–503.

Hertzog M, van Heijenoort C, Didry D, Gaudier M, Coutant J,Gigant B, Didelot G, Preat T, Knossow M, Guittet E et al. 2004. Thebeta-thymosin/WH2 domain; structural basis for the switch frominhibition to promotion of actin assembly. Cell 117(5):611–623.

Hertzog M, Yarmola EG, Didry D, Bubb MR, Carlier MF. 2002.Control of actin dynamics by proteins made of beta-thymosinrepeats: the actobindin family. J Biol Chem 277(17):14786–14792.

Hertzog M, Milanesi F, Hazelwood L, Disanza A, Liu H, PerladeE, Malabarba MG, Pasqualato S, Maiolica A, Confalonieri S, et al.2010. Molecular basis for the dual function of Eps8 on actindynamics: bundling and capping. PLoS Biol 8(6):e1000387.

Higgs HN, Pollard TD. 1999. Regulation of actin polymerizationby Arp2/3 complex and WASp/Scar proteins. J Biol Chem 274(46):32531–32534.

Huang H, Li L, Wu C, Schibli D, Colwill K, Ma S, Li C, Roy P,Ho K, Songyang Z, et al. 2008. Defining the specificity space ofthe human SRC homology 2 domain. Mol Cell Proteomics 7(4):768–784.

Huff T, Muller CS, Otto AM, Netzker R, Hannappel E. 2001.beta-Thymosins, small acidic peptides with multiple functions. Int JBiochem Cell Biol 33(3):205–220.

Husson C, Cantrelle FX, Roblin P, Didry D, Le KH, Perez J,Guittet E, Van Heijenoort C, Renault L, Carlier MF. 2010. Multi-functionality of the beta-thymosin/WH2 module: G-actin sequestra-tion, actin filament growth, nucleation, and severing. Ann NY AcadSci 1194:44–52.

Husson C, Renault L, Didry D, Pantaloni D, Carlier MF. 2011.Cordon-Bleu uses WH2 domains as multifunctional dynamizers ofactin filament assembly. Mol Cell 43(3):464–477.

Irobi E, Aguda AH, Larsson M, Guerin C, Yin HL, Burtnick LD,Blanchoin L, Robinson RC. 2004. Structural basis of actin seques-tration by thymosin-beta4: implications for WH2 proteins. EMBOJ 23(18):3599–3608.

Kelly AE, Kranitz H, Dotsch V, Mullins RD. 2006. Actin binding tothe central domain of WASP/Scar proteins plays a critical role in theactivation of the Arp2/3 complex. J Biol Chem 281(15):10589–10597.

Kim AS, Kakalis LT, Abdul-Manan N, Liu GA, Rosen MK. 2000.Autoinhibition and activation mechanisms of the Wiskott-Aldrichsyndrome protein. Nature 404(6774):151–158.

Kim T, Cooper JA, Sept D. 2010. The interaction of capping pro-tein with the barbed end of the actin filament. J Mol Biol 404(5):794–802.

Kim T, Ravilious GE, Sept D, Cooper JA. 2012. Mechanism forCARMIL protein inhibition of heterodimeric actin-capping protein.J Biol Chem 287(19):15251–15262.

Koronakis V, Hume PJ, Humphreys D, Liu T, Horning O, JensenON, McGhie EJ. 2011. WAVE regulatory complex activation bycooperating GTPases Arf and Rac1. Proc Natl Acad Sci USA108(35):14449–14454.

Lammers M, Rose R, Scrima A, Wittinghofer A. 2005. The regula-tion of mDia1 by autoinhibition and its release by Rho*GTP.EMBO J 24(23):4176–4187.

Lee SH, Kerff F, Chereau D, Ferron F, Klug A, Dominguez R.2007. Structural basis for the actin-binding function of missing-in-metastasis. Structure 15(2):145–155.

Letunic I, Doerks T. Bork P. 2012. SMART 7: recent updates tothe protein domain annotation resource. Nucleic Acids Res 40:D302–D305.

Li F, Higgs HN. 2005. Dissecting requirements for auto-inhibitionof actin nucleation by the formin, mDia1. J Biol Chem 280(8):6986–6992.

Liverman AD, Cheng HC, Trosky JE, Leung DW, Yarbrough ML,Burdette DL, Rosen MK, Orth K. 2007. Arp2/3-independentassembly of actin by Vibrio type III effector VopL. Proc Natl AcadSci USA 104(43):17117–17122.

Manseau LJ, Schupbach T. 1989. cappuccino and spire: two uniquematernal-effect loci required for both the anteroposterior and dorso-ventral patterns of the Drosophila embryo. Genes Dev 3(9):1437–1452.

Marchand JB, Kaiser DA, Pollard TD, Higgs HN. 2001. Interac-tion of WASP/Scar proteins with actin and vertebrate Arp2/3 com-plex. Nat Cell Biol 3(1):76–82.

Mattila PK, Salminen M, Yamashiro T, Lappalainen P. 2003.Mouse MIM, a tissue-specific regulator of cytoskeletal dynamics,interacts with ATP-actin monomers through its C-terminal WH2domain. J Biol Chem 278(10):8452–8459.

McNulty BC, Young GB, Pielak GJ. 2006. Macromolecular crowd-ing in the Escherichia coli periplasm maintains alpha-synuclein dis-order. J Mol Biol 355(5):893–897.

Mendoza MC. 2013. Phosphoregulation of the WAVE regulatorycomplex and signal integration. Semin Cell Dev Biol 24(4):272–279.

Mouilleron S, Guettler S, Langer CA, Treisman R, McDonald NQ.2008. Molecular basis for G-actin binding to RPEL motifs fromthe serum response factor coactivator MAL. EMBO J 27(23):3198–3208.

Mouilleron S, Wiezlak M, O’Reilly N, Treisman R, McDonaldNQ. 2012. Structures of the Phactr1 RPEL domain and RPELmotif complexes with G-actin reveal the molecular basis for actinbinding cooperativity. Structure 20(11):1960–1970.

Narita A, Takeda S, Yamashita A, Maeda Y. 2006. Structural basisof actin filament capping at the barbed-end: a cryo-electron micros-copy study. EMBO J 25(23):5626–5633.

Oikawa T, Yamaguchi H, Itoh T, Kato M, Ijuin T, Yamazaki D,Suetsugu S, Takenawa T. 2004. PtdIns(3,4,5)P3 binding is neces-sary for WAVE2-induced formation of lamellipodia. Nat Cell Biol6(5):420–426.

Otomo T, Tomchick DR, Otomo C, Panchal SC, Machius M,Rosen MK. 2005. Structural basis of actin filament nucleation andprocessive capping by a formin homology 2 domain. Nature433(7025):488–494.

Padrick SB, Rosen MK. 2010. Physical mechanisms of signal inte-gration by WASP family proteins. Annu Rev Biochem 79:707–735.

Panchal SC, Kaiser DA, Torres E, Pollard TD, Rosen MK. 2003. Aconserved amphipathic helix in WASP/Scar proteins is essential foractivation of Arp2/3 complex. Nat Struct Biol 10(8):591–598.

Paul AS, Pollard TD. 2008. The role of the FH1 domain and pro-filin in formin-mediated actin-filament elongation and nucleation.Curr Biol 18(1):9–19.

Paul AS, Pollard TD. 2009. Review of the mechanism of processiveactin filament elongation by formins. Cell Motil Cytoskeleton66(8):606–617.

Paunola E, Mattila PK, Lappalainen P. 2002. WH2 domain: a small,versatile adapter for actin monomers. FEBS Lett 513(1):92–97.

Perelroizen I, Marchand JB, Blanchoin L, Didry D, Carlier MF.1994. Interaction of profilin with G-actin and poly(L-proline). Bio-chemistry 33(28):8472–8478.

Peterson JR, Bickford LC, Morgan D, Kim AS, Ouerfelli O,Kirschner MW, Rosen MK. 2004. Chemical inhibition of N-WASP

� 704 Renault et al. CYTOSKELETON

by stabilization of a native autoinhibited conformation. Nat StructMol Biol 11(8):747–755.

Petrella EC, Machesky LM, Kaiser DA, Pollard TD. 1996. Struc-tural requirements and thermodynamics of the interaction of pro-line peptides with profilin. Biochemistry 35(51):16535–16543.

Pfender S, Kuznetsov V, Pleiser S, Kerkhoff E, Schuh M. 2011.Spire-type actin nucleators cooperate with Formin-2 to drive asym-metric oocyte division. Curr Biol 21(11):955–960.

Pollard TD, Cooper JA. 2009. Actin, a central player in cell shapeand movement. Science 326(5957):1208–1212.

Pontius BW. 1993. Close encounters: why unstructured, polymericdomains can increase rates of specific macromolecular association.Trends Biochem Sci 18(5):181–186.

Qualmann B, Kessels MM. 2009. New players in actin polymeriza-tion–WH2-domain-containing actin nucleators. Trends Cell Biol19(6):276–285.

Quinlan ME, Heuser JE, Kerkhoff E, Mullins RD. 2005. Drosoph-ila Spire is an actin nucleation factor. Nature 433(7024):382–388.

Quinlan ME, Hilgert S, Bedrossian A, Mullins RD, Kerkhoff E.2007. Regulatory interactions between two actin nucleators, Spireand Cappuccino. J Cell Biol 179(1):117–128.

Rebowski G, Namgoong S, Boczkowska M, Leavis PC, Navaza J,Dominguez R. 2010. Structure of a longitudinal actin dimerassembled by tandem w domains: implications for actin filamentnucleation. J Mol Biol 403(1):11–23.

Renault L, Bugyi B, Carlier MF. 2008. Spire and Cordon-bleu:multifunctional regulators of actin dynamics. Trends Cell Biol18(10):494–504.

Rottner K, Hanisch J, Campellone KG. 2010. WASH, WHAMMand JMY: regulation of Arp2/3 complex and beyond. Trends CellBiol 20(11):650–661.

Safer D, Sosnick TR, Elzinga M. 1997. Thymosin beta 4 bindsactin in an extended conformation and contacts both the barbedand pointed ends. Biochemistry 36(19):5806–5816.

Shoemaker BA, Portman JJ, Wolynes PG. 2000. Speeding molecu-lar recognition by using the folding funnel: the fly-casting mecha-nism. Proc Natl Acad Sci USA 97(16):8868–8873.

Simenel C, Van Troys M, Vandekerckhove J, Ampe C, DelepierreM. 2000. Structural requirements for thymosin beta4 in its contactwith actin. An NMR-analysis of thymosin beta4 mutants in solu-tion and correlation with their biological activity. Eur J Biochem267(12):3530–3538.

Sosne G, Qiu P, Goldstein AL, Wheater M. 2010. Biological activ-ities of thymosin {beta}4 defined by active sites in short peptidesequences. FASEB J 24(7):2144–2151.

Sribenja S, Wongkham S, Wongkham C, Yao Q, Chen C. 2013.Roles and mechanisms of beta-thymosins in cell migration and can-cer metastasis: an update. Cancer Invest 31(2):103–110.

Staus DP, Taylor JM, Mack CP. 2011. Enhancement of mDia2activity by Rho-kinase-dependent phosphorylation of the diapha-nous autoregulatory domain. Biochem J 439(1):57–65.

Stoll R, Voelter W, Holak TA. 1997. Conformation of thymosinbeta 9 in water/fluoroalcohol solution determined by NMR spec-troscopy. Biopolymers 41(6):623–634.

Takeda S, Minakata S, Koike R, Kawahata I, Narita A, KitazawaM, Ota M, Yamakuni T, Maeda Y, Nitanai Y. 2010. Two distinctmechanisms for actin capping protein regulation–steric and alloste-ric inhibition. PLoS Biol 8(7):e1000416.

Takeda S, Koike R, Nitanai Y, Minakata S, Maeda Y, Ota M.2011. Actin capping protein and its inhibitor CARMIL: howintrinsically disordered regions function. Phys Biol 8(3):035005.

Takeya R, Taniguchi K, Narumiya S, Sumimoto H. 2008. Themammalian formin FHOD1 is activated through phosphorylationby ROCK and mediates thrombin-induced stress fibre formation inendothelial cells. EMBO J 27(4):618–628.

Tam VC, Serruto D, Dziejman M, Brieher W, Mekalanos JJ. 2007.A type III secretion system in Vibrio cholerae translocates a formin/spire hybrid-like actin nucleator to promote intestinal colonization.Cell Host Microbe 1(2):95–107.

Thompson ME, Heimsath EG, Gauvin TJ, Higgs HN, Kull FJ.2013. FMNL3 FH2-actin structure gives insight into formin-mediated actin nucleation and elongation. Nat Struct Mol Biol20(1):111–118.

Tompa P. 2003. Intrinsically unstructured proteins evolve by repeatexpansion. Bioessays 25(9):847–855.

Tompa P. 2011. Unstructural biology coming of age. Curr OpinStruct Biol 21(3):419–425.

Tompa P. 2012. Intrinsically disordered proteins: a 10-year recap.Trends Biochem Sci 37(12):509–516.

Tompa P, Szasz C, Buday L. 2005. Structural disorder throws newlight on moonlighting. Trends Biochem Sci 30(9):484–489.

Uversky VN. 2010. Targeting intrinsically disordered proteins inneurodegenerative and protein dysfunction diseases: another illustra-tion of the D(2) concept. Expert Rev Proteomics 7(4):543–564.

Uversky VN, Dunker AK. 2010. Understanding protein non-fold-ing. Biochim Biophys Acta 1804(6):1231–1264.

Van Troys M, Ono K, Dewitte D, Jonckheere V, De Ruyck N,Vandekerckhove J, Ono S, Ampe C. 2004. TetraThymosinbeta isrequired for actin dynamics in Caenorhabditis elegans and acts viafunctionally different actin-binding repeats. Mol Biol Cell 15(10):4735–4748.

Vizcarra CL, Kreutz B, Rodal AA, Toms AV, Lu J, Zheng W,Quinlan ME, Eck MJ. 2011. Structure and function of the interact-ing domains of Spire and Fmn-family formins. Proc Natl Acad SciUSA 108(29):11884–11889.

Wallar BJ, Stropich BN, Schoenherr JA, Holman HA, Kitchen SM,Alberts AS. 2006. The basic region of the diaphanous-autoregulatory domain (DAD) is required for autoregulatory inter-actions with the diaphanous-related formin inhibitory domain. JBiol Chem 281(7):4300–4307.

Winder SJ, Ayscough KR. 2005. Actin-binding proteins. J Cell Sci118(Pt 4):651–654.

Yamaguchi H, Miki H, Suetsugu S, Ma L, Kirschner MW,Takenawa T. 2000. Two tandem verprolin homology domains arenecessary for a strong activation of Arp2/3 complex-induced actinpolymerization and induction of microspike formation by N-WASP.Proc Natl Acad Sci USA 97(23):12631–12636.

Zeth K, Pechlivanis M, Samol A, Pleiser S, Vonrhein C, KerkhoffE. 2011. Molecular basis of actin nucleation factor cooperativity:crystal structure of the Spir-1 kinase non-catalytic C-lobe domain(KIND)*formin-2 formin SPIR interaction motif (FSI) complex. JBiol Chem 286(35):30732–30739.

Zuchero JB, Coutts AS, Quinlan ME, Thangue NB, Mullins RD.2009. p53-cofactor JMY is a multifunctional actin nucleation factor.Nat Cell Biol 11(4):451–459.

Zuchero JB, Belin B, Mullins RD. 2012. Actin binding to WH2domains regulates nuclear import of the multifunctional actin regu-lator JMY. Mol Biol Cell 23(5):853–863.

Zwolak A, Fujiwara I, Hammer JA, 3rd, Tjandra N. 2010. Struc-tural basis for capping protein sequestration by myotrophin (V-1). JBiol Chem 285(33):25767–25781.

CYTOSKELETON Roles of Intrinsically Disordered Domains in Actin Dynamics 705 �