RhoG and Cdc42 can contribute to Rac-dependent lamellipodia formation through WAVE regulatory complex-binding.

Full Te ms & Condi ions o access and use can be ound a h ps://www. and online.com/ac ion/jou nalIn o ma ion?jou nalCode=ksg 20 Small GTPases ISSN: 2154-1248 (P in ) 2154-1256 (Online) Jou nal homepage: h ps://www. and online.com/loi/ksg 20 RhoG and Cdc42 can con ibu e o Rac-dependen lamellipodia o ma ion h ough WAVE egula o y complex-binding Ma hias Schaks, He mann Dö ing, F ieda Kage, Anika S e en, Thomas Klünemann, Wul Blanken eld , The esia S adal & Klemens Ro ne To ci e his a icle: Ma hias Schaks, He mann Dö ing, F ieda Kage, Anika S e en, Thomas Klünemann, Wul Blanken eld , The esia S adal & Klemens Ro ne (2019): RhoG and Cdc42 can con ibu e o Rac-dependen lamellipodia o ma ion h ough WAVE egula o y complex-binding, Small GTPases, DOI: 10.1080/21541248.2019.1657755 To link o his a icle: h ps://doi.o g/10.1080/21541248.2019.1657755 © 2019 The Au ho (s). Published by In o ma UK Limi ed, ading as Taylo & F ancis G oup View supplemen a y ma e ial Published online: 26 Aug 2019. Submi you a icle o his jou nal A icle iews: 270 View ela ed a icles View C ossma k da a BRIEF REPORT RhoG and Cdc42 can con ibu e o Rac-dependen lamellipodia o ma ion h ough WAVE egula o y complex-binding Ma hias Schaks a,b , He mann Dö ing a,b , F ieda Kage a,b , Anika S e en b , Thomas Klünemann c , Wul Blanken eld c , The esia S adal b , and Klemens Ro ne a,b a Di ision o Molecula Cell Biology, Zoological Ins i u e, Technische Uni e si ä B aunschweig, B aunschweig, Ge many; b Cell Biology, Helmhol z Cen e o In ec ion Resea ch, B aunschweig, Ge many; c S uc u e and Func ion o P o eins, Helmhol z Cen e o In ec ion Resea ch, B aunschweig, Ge many ABSTRACT Cell mig a ion equen ly in ol es he o ma ion o lamellipodial p o usions, he ini ia ion o which equi es Rac GTPases signalling o he e open ame ic WAVE egula o y complex (WRC). While Rac- ela ed RhoG and Cdc42 can po en ly s imula e lamellipodium o ma ion, so a p esumed o occu by ups eam signalling o Rac ac i a ion, we show he e ha he la e can be bypassed by RhoG and Cdc42 gi en ha WRC has been a i icially ac i a ed. This e idence a ises om gene a ion o B16-F1 cells simul aneously lacking bo h Rac GTPases and WRC, ollowed by econs i u ion o lamellipodia o ma ion wi h speci ic Rho-GTPase and di e en ially ac i e WRC a ian combina ions. We conclude ha o ma ion o canonical lamellipodia equi es WRC ac i a ion h ough Rac, bu can possibly be uned, in addi ion, by WRC in e ac ions wi h RhoG and Cdc42. ARTICLE HISTORY Recei ed 2 Ap il 2019 Re ised 13 Augus 2019 Accep ed 15 Augus 2019 KEYWORDS A p2/3 complex; CRISPR/ Cas9; Rho-GTPase; ilopodium; lamellipodium; blebbing; s ess ib e; mig a ion; p o usion In oduc ion Cell mig a ion is essen ial o many physiological and pa hological p ocesses, such as emb yonic de elopmen , immuni y and me as asis [1–3]. P o usion o he plasma memb ane o enable cell mig a ion is commonly achie ed by he o ma ion o hin b anched a ays o ac in ilamen s, called lamellipodia [4,5]. Small GTPases o he Rac amily (i.e. Rac1, Rac2, Rac3 in mammals) a e essen ial o lamel- lipodia ini ia ion and main enance, a leas in pa , by ac i a ion and con inuous in e ac ion h ough wo binding si es on S a-1 (o i s o hologue PIR121) embedded in o he e open ame ic WAVE egula o y complex (WRC) [6,7]. The closes ela i es o Rac GTPases a e RhoG and Cdc42. Bo h ail o ini ia e lamellipodia in ib oblas s lack- ing Rac exp ession [6], bu can clea ly p omo e Rac- dependen lamellipodia o ma ion. This ac i i y is hough o de i e om c oss alk in ol ing dis inc Rac-GEF com- plexes [8–11]. The inabili y o RhoG and Cdc42 o ini ia e lamellipodia has hi he o been hough o be due o a lack o su icien in e ac ion wi h WRC. This is because bo h RhoG and Cdc42 show a bes a weak in e ac ion wi h S a- 1/PIR121 [6,12], and Cdc42, as opposed o Rac1, ails o ac i a e na i e WRC in i o [13]. Aside om S a-1 (o PIR121), WRC is composed o ou addi ional subuni s –WAVE2 (o i s pa alogues WAVE1/WAVE3), he S a-1/PIR121 in e ac o Nap1 (o Hem1 in he hema opoie ic sys em), Abi1 (o Abi2/Abi3) and HSPC300 [14–17]. The S a-1 subuni is ‘ ansinhibi - ing’ he A p2/3 complex-ac i a ing, so called WCA domain loca ed on he C- e minal end o WAVE p o eins, and Rac binding o S a-1 ou compe es his inhibi o y in e ac ion o elease he WCA domain, making i accessible o ac in and A p2/3 complex binding [18]. We ha e ecen ly shown ha he wo a o emen ioned Rac binding si es on S a-1/ PIR121 [18,19] a e essen ial o allos e ic ac i a ion o WRC in cells [7]. Howe e , in spi e o he p e iously p oposed sa e box model equi ing wo keys o allow o WRC ac i a ion o occu [19], we ha e ound su p isingly speci ic physiological unc ions o he wo si es in li e cells. Whe eas he low a ini y A si e is c ucial o ac i a ion in i o, he high a ini y D si e is con ibu ing o he e iciency o lamellipodial p o usion, bu by no means as impo an o WRC ac i a ion as he A si e. Aside om he appa en c i ical unc ion o Rac in WRC ac i a ion, i is less well es ablished i o i so o wha ex en Rac-WRC in e ac ions also d i e WRC ec ui men o and accumula- ion in he lamellipodium. Fo ins ance, we ha e also ound ha lamellipodia o ma ion can be ini ia ed, in p inciple, wi hou di ec WRC-Rac in e ac ions once WRC is en- de ed ac i e, assuming a leas ha in oduced, espec i e CONTACT Klemens Ro ne [email p o ec ed] Di ision o Molecula Cell Biology, Zoological Ins i u e, Technische Uni e si ä B aunschweig, B aunschweig, Ge many The supplemen al da a o his a icle can be accessed he e. SMALL GTPASES h ps://doi.o g/10.1080/21541248.2019.1657755 © 2019 The Au ho (s). Published by In o ma UK Limi ed, ading as Taylo & F ancis G oup This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion-NonComme cial-NoDe i a i es License (h p://c ea i ecommons.o g/licenses/by-nc- nd/4.0/), which pe mi s non-comme cial e-use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed, and is no al e ed, ans o med, o buil upon in any way. mu a ions o bo h A and D si es in o ac i e WRC abolished i s in e ac ion wi h Rac en i ely [7,19]. This is consis en wi h he ac ha dele ing he CAAX-box in Rac1, which is c ucial o plasma memb ane associa ion, does impai , bu no abolish lamellipodia o ma ion in Rac1 knockou (KO) ib oblas s [6]. In spi e o he absence o an unequi ocal, al e na i e mechanism o WRC ec ui men , hese da a sugges ha ac i a ionandlamellipodial a ge ing o WRC migh po en ially be sepa able. An example o such a sepa a ion clea ly cons i u es he ela ed GTPase Cdc42, which media es ac i a ion o i sdowns eame ec o s FMNL2 and −3 as p e equisi e o hei lamellipodial a ge - ing [20,21], which howe e can ully occu wi h ac i e a ian s comple ely lacking GTPase in e ac ion su aces ([22] and unpublished da a). The e o e, e ec o ec ui - men o lamellipodia is possible, in p inciple, wi hou engagemen o a gi en GTPase in spi e o i s es ablished ele ance in e ec o ac i a ion. He e, we ha e de eloped no el cell lines o compa e he capabili y o lamellipodia o ma ion by ac i a ed WRC in he absence e sus p esence o endogenous Rac GTPases. Resul s Rac –WRC in e ac ions a he plasma memb ane a e dispensable o lamellipodia o ma ion We sough o es i plasma memb ane inse ion o he p enyl g oup o Rac1 is equi ed o lamellipodia induc- ion in B16-F1 cells. Fo his, we exp essed cons i u i ely ac i e (Q61L), myc- agged Rac1 o an iden ical cons uc lacking he CAAX-box essen ial o C- e minal p enyla- ion in B16-F1 cells lacking Rac1/2/3 (clone#1 [7]). In analogy o ou p e iously published expe imen s employing ib oblas s gene ically dele ed o Rac1 [6], dele ion o he CAAX-box educed, bu did no abolish lamellipodia o ma ion in hese condi ions. Mo e speci- ically, hese s uc u es we e induced wi h o e all sligh ly educed equency, and he majo i y o hem ( oughly 60%) appea ed o be imma u e, acco ding o he p e- iously es ablished ca ego iza ion o unde de eloped lamellipodia [7]. In con as , he majo i y o cells exp es- sing ull leng h, cons i u i ely ac i e (Q61L) Rac1 ha - bou ed ully de eloped lamellipodia (Figu e S1(a,b)). These da a sugges ha he same e ec s seen in Rac1 −/− ib oblas s [6] we e no cell- ype speci ic, and could no po en ially be explained by emnan s o Rac2 o −3 p o ein exp essed pe haps a unde ec able le els om espec i e genes no a ge ed in hese ib oblas s. In analogy, we assembled WRCs in S a-1/PIR121 KO cells (clone #3) ha bou ing a S a-1 a ian media ing con- s i u i e WRC ac i a ion, bu lacking unc ional Rac bind- ing si es (A + D si e WCA* [7], and Figu e S1(e)). In his cons uc , speci ic poin mu a ions abolished he binding o Rac o he A si e (C179R/R190D) and D si e (Y967A) [7,18,19], whe eas he WCA* mu a ion (L697D/Y704D/ L841A/F844A/W845A) in he WH2- and C- egion (W and C) con ac si es o S a-1 p e en ed he ‘ ansinhibi o y’ binding o S a-1 o he WCA domain o WAVE, which is hen eleased o ac i a ing A p2/3 complex [7,18]. We ound ha S a-1/PIR121 KO cells (clone #3) ha bou ing A + D si e WCA*-mu a ed WRC can s ill escue lamelli- podia o ma ion, albei a s ongly comp omised equency [7]. Once o med hough, and al hough comp omised, hese lamellipodia can s ill accumula e WRC a hei ips (Figu e S1(c), see also [7]), and media e con inuous p o- usion ha is less smoo h hough han wi h lamellipodia d i en by WRCs ha bou ing WT S a-1 (Figu e S1(d)). No su p isingly, a e age p o usion eloci ies o hese a e examples o comp omised lamellipodia we e educed as compa ed o hose media ed by wild ype S a-1 (Figu e S1 ( )). Toge he , all hese da ase s hus sugges ha al hough help ul, con inuous Rac-WRC in e ac ions a he plasma a e no absolu ely obliga o y o lamellipodium p o usion. Gene a ion o a cell line allowing u he dissec ion o he Rac-WRC signalling module Nex we asked whe he Rac is essen ial o WRC- media ed lamellipodia o ma ion solely because o i s essen ial unc ion in WRC ac i a ion o because o se - ing addi ional unc ions. To es his, we had o de elop cell sys ems in which endogenous WRC o Rac p o eins could be eplaced by ac i e a ian s o each o unc ional de iciency mu an s in a combina o ial ashion. In p e- ious wo k, we had es ablished cell lines lacking ei he Rac1/2/3 o unc ional WRC (S a-1/PIR121-null), in which o - a ge e ec s po en ially caused by CRISPR/ Cas9-media ed genome edi ing we e excluded by a combina ion o analysing mul iple, independen ly gen- e a ed clones as well as by escue o pheno ypes wi h exogenous Rac and S a-1, espec i ely [7]. In he cu en wo k, we ex ended his app oach o gene a e a no el cell line dis up ed o all i e genes (S a-1/PIR121+ Rac1/2/3 KO#3/11; Figu e 1(a)). The la e now allows deciphe - ing Rac/WRC signalling in mo e de ail. In hese cells, lamellipodia o ma ion is s ikingly dependen on exo- genous exp ession o bo h S a-1 and Rac1. While nei he exp ession o EGFP as con ol, EGFP-S a-1 no myc- Rac1L61 alone acili a ed lamellipodia o ma ion in hese cells, co- ans ec ion o EGFP-S a-1 and myc- Rac1L61 po en ly es o ed lamellipodia, indica ing he p esence o lamellipodia in hese cells o s ic ly equi e bo h Rac and S a-1 (Figu e 1(b,c)). No ably, Rac1L61 exp ession in hese WRC-de icien cells also caused plasma memb ane blebbing (Figu e 1(b) o 2M. SCHAKS ET AL. ep esen a i e image), eminiscen o ou p e ious obse a ions upon Rac mic oinjec ion upon WRC sub- uni knockdown [15]. This pheno ype was obus and occu ed a high equency (45 ± 13% in Rac1L61 exp essing cells s. 13 ± 12% in EGFP-exp essing con- ols), al hough a p ecise, mechanis ic unde s anding o Figu e 1. Gene a ion o S a-1/PIR121+ Rac1/2/3 KO cells. (a) Wes e n blo ing o dis inc cell lines o p obe o exp ession le els o endogenous S a-1/PIR121 and/o Rac GTPases. (b) Cell mo phologies o S a-1/PIR121+ Rac1/2/3 KO cells (clone #3/11) exp essing espec i e cons uc s, as indica ed. Panels in op ow show s ainings o he ac in cy oskele on wi h phalloidin, and middle and/o bo om ow images show luo escence o he same cells de i ed om ei he EGFP o an i-myc an ibody s ainings, as indica ed. (c) Quan i ica ion o lamellipodial pheno ypes. EGFP-S a-1 WCA* deno es a cons uc ende ing WRC cons i u i ely ac i e due o mu a ions elie ing he au oinhibi o y in e ac ion o S a-1 wi h he C- e minal WCA- domain (hence WCA* o ac i e WCA) o WAVE [7,18]. Lamellipodial ac in ne wo ks ha we e gene ally small, na ow, i egula o displayed mul iple u les we e de ined as ‘“imma u e lamellipodia”’, as opposed o egula , ully de eloped lamellipodia. n gi es numbe o cells analysed, da a co espond o a i hme ic means ± SEM om a leas h ee independen expe imen s. S a is ical signi icance was assessed o di e ences be ween pe cen ages o cells wi h ‘no lamellipodia’pheno ype. ***p < 0.001 ( wo-sample, wo-sided - es ). SMALL GTPASES 3 he phenomenon is cu en ly lacking. To es i cons i- u i e ac i a ion o WRC would be su icien o igge lamellipodia o ma ion in he absence o Rac, we exp essed EGFP- agged S a-1 WCA*. Su p isingly, how- e e , EGFP- agged S a-1 WCA*, lacking he need o Rac-media ed ac i a ion o WRC was s ill incapable o d i ing lamellipodia o ma ion in his cell line and con- di ions. I compa ing his esul wi h cons i u i ely ac i e WRCs comp omised in Rac binding, bu gi ing ise o ine icien lamellipodia o ma ion in S a-1/ PIR121-KO cells ([7] and Figu e S1(c,d)), wo heo e ical explana ions o his disc epancy a e hinkable: Fi s ly, he A + D si e-mu a ed, ac i e WRC used in Figu e S1 can s ill ine icien ly bind o endogenous Rac p o eins p esen in hese cells. Al e na i ely, Rac possesses addi- ional, essen ial unc ions absen in cells ha bou ing ac i e WRC bu lacking endogenous Rac p o eins (as in Figu e 1(b,c)). Fu u e expe imen s will ha e o dis in- guish be ween hese wo possibili ies. Rac- ela ed Rho GTPases ail o ac i a e WRC, bu can subs i u e o Rac in he p esence o ac i a ed WRC Up o his da e, he li e a u e con ains con lic ing esul s conce ning he ele ance o he closes ela- i es o Rac GTPases in mammals, in pa icula RhoG, bu o a ce ain ex en also Cdc42. In spi e o p ominen s udies es ablishing unc ions o RhoG in signalling complexes ope a ing ups eam o Rac [10–12], RhoG has also al eady been concluded o con ibu e o ib oblas mig a ion independen o Rac ac i a ion [23], al hough i has emained unclea how ha migh occu mechanis ically [24]. O no e, and again in ull acco dance wi h ou p e iously pub- lished ib oblas da a [6], ou Rac1/2/3-de icien B16- F1 melanoma line ailed en i ely o o m lamellipodia e en upon exp ession o cons i u i ely ac i e RhoG (Figu e 2(a,b)). Iden ical esul s we e ob ained wi h o e exp essed, cons i u i ely ac i e Cdc42 bu no Rac1, which obus ly es o ed lamellipodia o ma ion (Figu e 2(a,b)), as expec ed [7]. Aside om he incapabili y o Cdc42 o induce lamellipodia, we ound a p ominen induc ion o s ess ib es in hese condi ions (Figu e 2(a); 89 ± 2% o ans ec ed cells), which was no seen wi h RhoG (2 ± 3%). Such an ac i i y is adi ionally s ill mos ly a ibu ed o RhoA/B/C ac i i y in he li e a u e, and hus no ollowed up u he in he con ex o he cu en s udy. Howe e , since ou p e ious e o s allowed us o expe imen ally sepa a e Rac-media ed WRC ac i a ion om o he po en ial unc ions ([7] and see abo e), we wonde ed whe he we migh –in analogy o Rac –be able o es ablish connec ions be ween RhoG o Cdc42 ac i i ies and WRC unc ion independen o Rac-media ed WRC ac i a ion. Fo his, we explo ed RhoG- o Cdc42-d i en ac in Figu e 2. Rac- ela ed GTPases RhoG and Cdc42 ail o induce lamellipodia in he absence o Rac exp ession. (a) Cell mo phologies o Rac1/2/3 KO cells (clone #1) exp essing EGFP- agged GTPases, as indica ed. (b) Quan i ica ion o lamellipodial pheno ypes was pe o med as desc ibed o Figu e 1(c). n gi es numbe o cells analysed, da a co espond o a i hme ic means ± SEM om a leas h ee independen expe imen s. S a is ical signi icance was assessed o di e ences be ween pe cen ages o cells wi h ‘no lamellipodia’pheno ype. ***p < 0.001 ( wo-sample, wo-sided - es ). 4M. SCHAKS ET AL. cy oskele on emodelling in cells exclusi ely ha bou - ing ac i e WRC bu lacking endogenous Rac GTPases (see Figu e 1). And indeed, co- ans ec ion o ac i e (Q61L), myc- agged RhoG o Cdc42 oge he wi h S a-1 WCA* led o pa ial escue o lamellipodia o ma ion in S a-1/PIR121+ Rac1/2/3 KO cells (clone #3/11) (Figu e 3(a,c) o ep esen a i e images and quan i a ions, espec i ely). While RhoG was able o induce imma u e lamellipodiainmo e han 40% o ans ec ed cells and e en occasional, ully de eloped lamellipodia, Cdc42 caused lamellipodia o ma iononlyinasmallsub ac iono cells(see Figu e 3(a), as e isks, and Figu e 3(c) o quan i ica- ion). Al hough he low equency o Cdc42-media ed induc ion o (a leas imma u e) lamellipodia in cells co-exp essing S a WCA* could no be igno ed, i was no signi ican ly di e en s a is ically om cells co- exp essing cons i u i ely ac i e Cdc42 wi h wild ype S a-1 (no shown). This esul was ob ained wi h wo-sided, wo-sample - es , bu when using Fishe ’s exac es also app op ia e o compa ing wo expe imen al g oups (Cdc42 co-exp ession wi h S a-1 WT e sus WCA*) alling in o wo ca ego ies (wi h o wi hou lamellipodia), he di e ence be ween hem u ned ou o be s a is ically signi ican (p = 0.023). Mo eo e , bo h RhoG and Cdc42 we e capable, in p inciple, o d i ing accumula ion o S a- 1 WCA* a p o usion si es gi en ha lamellipodia we e o med (Figu e 3(b)). Speci ic S a-1 WCA* en ichmen upon co-exp ession o bo h RhoG and Cdc42 was also con i med by linescan analyses o espec i e images (Figu e 3(b)). In con as , co- an ec ion o S a-1 WCA* wi h myc- agged, cons i- u i ely ac i e RhoA (G14V) ailed o s imula e lamellipodia o lamellipodia-like s uc u es in all cells analysed (96 cells om 3 independen expe i- men s; Figu e 3(a,c) o quan i ica ion). This was consis en wi h he lack o any accumula ion o EGFP- agged S a-1 WCA* a he pe iphe y o hese cells (Figu e 3(b), igh panel), in spi e o cons i u- i ely ac i e RhoA clea ly being unc ional in his case, as e idenced by he expec ed, p ominen s imu- la ion o s ess ib es (Figu e 3(a), igh panels). As al eady men ioned in he con ex o s a is ical analysis, he abili y o RhoG and Cdc42 o induce lamellipodia in hese expe imen s was s ic ly depen- den on cons i u i e WRC ac i i y in his cell line ( h ough using S a-1 WCA*), as bo h RhoG and Cdc42 ailed o induce lamellipodia when co- exp essed wi h WT-S a-1 (Figu e 3(c)). Finally, addi- ional mu a ion o he wo Rac binding si es on con- s i u i ely ac i e WRC comple ely p e en ed (in case o Cdc42) o a leas s ongly inhibi ed (in case o RhoG) he induc ion o lamellipodia in hese condi- ions (Figu e 3(a,c)), indica ing ha he ac ion o RhoG/Cdc42 may be explained by di ec , (Rho- GTPase binding su ace-dependen ) in e ac ion wi h WRC. In line wi h his, immunop ecipi a ion expe i- men s showed clea ly de ec able in e ac ions o RhoG and Cdc42 wi h S a-1 WCA*, albei somewha weake pe haps han obse ed o Rac1 (Figu e 3(d)). Assuming ha he desc ibed, po en ial RhoG/Cdc42 - WRC in e ac ions may be di ec , we also wonde ed whe he hese GTPases migh p e e en ially a ge he A o he D si e p e iously es ablished as in e ac ion su aces on WRC wi h Rac1 [19]. Due o he low equency o Cdc42-media ed lamellipodia o ma ion in his assay (Figu e 3), we ocused on RhoG in hese expe imen s compa ed o Rac1 used as con ol. We again co- ans ec ed ou cell line lacking endogenous WRC plus all Rac-GTPases wi h he indi idual A and D si e mu an s o WCA* S a-1 and cons i u i ely ac i e Rac1 (as con ol) e sus RhoG. While mu a ing ei he he A o D si e caused signi ican impai men o lamellipodia o ma ion induced by ac i e Rac1, albei o dis inc easons, as expec ed ( o compa ison o analogous, bu no iden ical expe imen s, see e [7].), mu a ing he A si e did no cause a s a is ically sig- ni ican educ ion o RhoG-media ed lamellipodia o - ma ion in his assay. In con as , ab oga ion o D si e unc ion caused a se e e educ ion o lamellipodia (Figu e S2(a)), i ually iden ical o he le el obse ed upon simul aneous mu a ion o bo h si es (Figu e 3 (c)). F om his, we concluded ha he po en ial, pe - haps di ec WRC-RhoG in e ac ion is la gely media ed by he D si e o S a-1. All hese esul s p omp ed us o compa e by addi- ional in silico analyses he pu a i e binding su aces o RhoG and Cdc42 wi h he D si e o S a-1, which also cons i u es he high a ini y binding si e o Rac GTPases [19]. Sequence alignmen s e ealed ha wo (incaseo RhoG)o h ee(incaseo Cdc42)amino acids di e ed om Rac1 in he pu a i e binding in e ace (Figu e S2(b)). Howe e , nei he o hese esidues caused signi ican changes in elec os a ic su ace po en ials (Figu e S2(c)). RhoA, on he con- a y, showed ob ious di e ences in su ace elec o- s a ic po en ial in he pu a i e binding in e ace, pa icula ly caused by he glu amine o aline sub- s i u ion a posi ion 33 in RhoA, he analogous sub- s i u ion o which in Rac1 (E31V) appa en ly in e e ed wi h p ope lamellipodia o ma ion [7,25]. Toge he , hese da a sugges ha bo h RhoG and Cdc42 can speci ically in e ac wi h WRC, in p inci- ple, and in a physiologically ele an manne once WRC has been ac i a ed by Rac1. SMALL GTPASES 5 Figu e 3. Induc ion o lamellipodia by RhoG o Cdc42 in conjunc ion wi h ac i a ed WRC, bu in he absence o Rac exp ession. (a) Rep esen a i e images o S a-1/PIR121+ Rac1/2/3 KO cells (clone #3/11) exp essing EGFP-S a-1 WCA* oge he wi h myc- agged, small GTPases, as indica ed. T ans ec ed cells iden i ied by myc-s aining (as e isks) we e analysed conce ning hei cell mo phologies ( op ow, ac in ilamen s aining wi h phalloidin) and he capabili y o accumula e ac i e WRC (S a-1 WCA*) a pe iphe al lamellipodia (b). Measu emen s shown in b we e pe o med along line scans as shown in he images p o ided in a (g een lines). (c) Quan i ica ion o lamellipodial pheno ypes, done as desc ibed o Figu e 1(c). n gi es numbe o cells analysed, and di e en ially colou ed columns a e a i hme ic means ± SEM om a leas h ee independen expe imen s. S a is ical signi icance was assessed o di e ences be ween he pe cen ages o cells wi h ‘no lamellipodia’pheno ype. ***p < 0.001 ( wo-sample, wo-sided - es ). (d) S a-1/PIR121 KO cells (clone # 3) we e co- ans ec ed wi h mChe y-S a-1 WCA* and ei he EGFP o cons i u i ely ac i e (Q61L), EGFP- agged small GTPases, lysed and subjec ed o immunop ecipi a ion agains EGFP. No e he comple e absence o immunop ecipi a ion o S a-1 upon co-exp ession o EGFP alone (nega i e con ol), and signi ican in e ac ions o S a-1 WCA* wi h Rac-1, RhoG and Cdc42. (e) Model o how RhoG and Cdc42 migh egula e WRC and lamellipodia o ma ion. 6M. SCHAKS ET AL. Discussion Al hough puzzling ini ially, we ha e p e iously ound and con i med he e using dis inc expe imen al sys ems ha Rac may no ha e o associa e wi h he plasma memb ane in o de o ac i a e o ec ui WRC du ing lamellipodia ini ia ion and main enance. Ye , and no inconsis en wi h his iew, Rac GTPases emain o be obliga o y o bo h WRC ac i a ion and lamellipodia o ma ion, as cells lacking bo h endogenous Rac GTPases and WRC can only o m lamellipodia upon addi ional, speci ic manip- ula ion. To ou su p ise, we es ablish o he i s ime he e ha lamellipodia o ma ion is possible, in p inciple, wi hou Rac GTPases, gi en ha cells lack he need o WRC ac i a ion (because i is al eady ac i a ed o does no need o be ac i a ed), and ha hey o e -exp ess cons i u i ely ac i e a ian s o ei he RhoG o Cdc42. In o he wo ds, o e exp ession o RhoG, and o a lesse ex en Cdc42 can cause he accumula ion o cons i u i ely ac i e WRC e en in he comple e absence o endogenous Rac GTPases, p esumably causing WRC-dependen A p2/3 complex ac i a ion. This sugges s ha aside om p e iously es ablished signalling c oss alk be ween Rac GTPases and RhoG/Cdc42 [8–10], he la e may di ec ly con ibu e o he main enance and/o ac i i y o WRC a p o uding lamellipodia edges once Rac has managed o ac i a e indi idual WRCs. I is hus hinkable ha RhoG and Cdc42 may di ec ly suppo Rac signalling by pa ially aking o e Rac unc ions conce ning WRC posi ioning i espec i e o WRC ac i a ion. Consis en wi h his iew, we ha e p e iously ound a e age u no e imes o WRC subuni s a he p o- uding plasma memb ane ha ce ainly i he hypo h- esis o con inuous ac i a ion e en s o A p2/3 complexes, media ed by indi idual WRC uni s in a Rho GTPase binding-dependen ashion [7]. Slow u no e and molecula c owding o indi idual WRCs a he memb ane likely con ibu es o e icien A p2/3 ac i a ion a hese si es (see also [26,27]), las , no leas because e icien A p2/3 complex ac i a ion was p e- iously demons a ed o in ol e simul aneous engage- men o wo WCA domains [28–30]. The e o e, i seems plausible ha e icien A p2/3 complex ac i a ion a he lamellipodium ip coincides wi h WRC clus e - ing, wi h he la e being a ec ed, a leas in pa , by GTPase signalling. The desc ibed, di ec con ibu ion o RhoG and Cdc42 o WRC-media ed ac in emodelling is also ound o occu in spi e o a bes weak o en i ely non- speci ic in e ac ions o RhoG and Cdc42 wi h wild ype WRC [6,12], as we show he e ha he si ua ion changes d ama ically i cons i u i ely ac i e WRC is used, which can be p ecipi a ed qui e e icien ly by bo h Rac- ela ed GTPases. Mo eo e , we also ound ha a leas he in e ac ion o RhoG wi h WRC is la gely media ed h ough he D si e o S a-1 in WRC, and no he A si e (Figu e S2(a)), es ablished p e iously o be c u- cial o WRC ac i a ion by Rac1 [7]. This hen could explain he comple e ailu e o RhoG (and Cdc42) o ac i a e WRC in he absence o endogenous Rac GTPases (Figu es 2 and 3(c)). All hese conside a ions a e also con i med by sequence alignmen s and s uc- u al conside a ions conce ning e ec o in e ac ion su aces p esen on Rac1, RhoG and Cdc42 e sus RhoA (Figu e S2). Fu u e s uc u al s udies will be needed o solidi y he hypo hesis o di ec in e ac ions and explain why –i con i med o be di ec – he compa ably obus in e ac ion o RhoG and Cdc42 wi h S a-1 obse ed he e canno occu wi h inac i e WRC and/o ansla e in o WRC ac i a ion. Al hough he binding e iciency o RhoG and Cdc42 o ac i e WRC appea ed compa able in immunop ecipi a- ion expe imen s, he e was clea ly measu able di e ences be ween he e iciency o he ou pu esponse (lamellipo- dia in his case), he p ecise easons o which emain o be de e mined. Ye , he as majo i y o bo h RhoG and Cdc42-dependen lamellipodia o med in hese condi- ions (absence o endogenous Rac GTPases) we e s ill imma u e, clea ly illus a ing he ele ance o Rac in o - ma ion and u no e o hese s uc u es beyond i s es ab- lished, essen ial unc ion in WRC ac i a ion. The no el cell sys em lacking endogenous Rac and WRC p o eins (B16-F1 S a-1/PIR121+ Rac1/2/3 KO clone #3/11) also ha bou s he po en ial o emphasizing p e iously men ioned, bu less well s udied pheno ypes caused by ac i e, small GTPases, indica i e o he commonly es ablished, bu ch onically unde es ima ed complexi y a ising om Rho-GTPase c oss alk [31,32]. Fo ins ance, wi hou WRC, Rac ailed o induce lamel- lipodia, bu no in equen ly caused plasma memb ane blebbing much less common o B16-F1 cells exp essing endogenous WRC and hus capable o lamellipodia o ma ion. Simila obse a ions we e p e iously epo ed o o e exp ession o he Rac e ec o loop mu an F37A [33], which we now conside o be impai ed in d i ing lamellipodia o ma ion as a esul o comp omised WRC in e ac ion [7]. Whe he his blebbing ac i i y a ises om in insic, Rac-speci ic ea- u es (i.e. equi ing di ec Rac-e ec o binding) o is amo e indi ec esul o c oss alk o RhoA/B/C signal- ling emains unclea . Mo eo e , ou p elimina y obse - a ions also indica ed ha Cdc42 ac i i ies can p ominen ly unnel in o con ac ile s ess ib e o ma- ion i Rac signalling (o lamellipodia o ma ion) is missing. This could likely be media ed h ough signal- ling o MRCK kinases, p e iously shown o coope a e SMALL GTPASES 7 wi h Rho-ROCK signalling [34–36], which hi he o appea ed less p e alen in ou cell sys ems wi hou comp omised lamellipodia o ma ion (see [37,38]). As he lamellipodia esponse seen wi h he Cdc42 – (ac i e) WRC combina ion was much less p ominen han seen o he RhoG –(ac i e) WRC couple, i is emp ing o specula e ha he s ess ib e induc ion pheno ype men ioned abo e migh in e e e wi h mo e obus lamellipodia o ma ion. This migh be explained pe haps by he widely accep ed and long- s anding an agonism be ween p o usion- (lamellipo- dia) e sus con ac ili y-dependen (s ess ib es) p o- cesses [39,40]. Fu u e expe imen s will ha e o e eal whe he his assump ion is co ec . Wha e e he case, ou da a add o he iew ha sys ema ic gene a ion and side-by-side compa ison o Rho GTPase and e ec o knockou s in he same pa - en al cell line will con inue o un old mechanis ic insigh s in o he in icacies o Rho signalling and c oss- alk ele an o ac in emodelling p ocesses. Ma e ials and me hods Cell cul u e B16-F1 cell line was pu chased om ATCC (CRL-6323, sex:male). B16-F1 de i ed S a-1/PIR121 KO cells (clone #3), as well as Rac1/2/3 KO cells (clone #1) we e as desc ibed [7]. B16-F1 cells and de i a i es we e cul u ed in DMEM (4.5 g/l glucose; In i ogen), supplemen ed wi h 10% FCS (Gibco), 2 mM glu amine (The mo Fishe Scien i ic) and penicillin (50 Uni s/ml)/s ep omycin (50 µg/ml) (The mo Fishe Scien i ic). B16-F1 cells we e ou inely ans ec ed in 35 mm dishes (Sa s ed ), using 0.5 µg DNA in o al and 1 µl Je P ime o con ols, and 1 µg DNA in o al and 2 µl Je P ime o B16-F1-de i ed knockou cells. A e o e nigh ans ec ion, cells we e pla- ed on o acid-washed, laminin-coa ed (25 µg/ml) co e slips and allowed o adhe e o a leas 5 hou s p io o analysis. DNA cons uc s pEGFP-C1 and –C2 ec o s we e pu chased om Clon ech Inc. (Moun ain View, CA, USA). pEGFP- C2-S a-1, and de i ed mu an cons uc s (i.e. WCA*, A si e WCA*, D si e WCA*, A + D si e WCA*) we e desc ibed p e iously [7] and co espond o he splice a ian CYFIP1a, sequence AJ567911. mChe y- agged S a-1 WCA* was gene a ed by swapping EGFP wi h mChe y, kindly p o ided by D . Roge Tsien (Uni e si y o Cali o nia a San Diego, La Jolla, Cali o nia, USA) using NheI/Bs GI es ic ion si es. pRK5-myc-Rac1L61 and pRK5-myc-RhoAV14 we e kindly p o ided by Alan Hall and Lau a Machesky (CRUK Bea son Ins i u e, Glashow, UK). Cdc42L61 (placen al iso o m) was syn hesized by Eu o ins Genomics and cloned in o pRK5-myc and pEGFP- C1 ec o s. pEGFP-C1-Rac1L61 and pRK5-myc- RhoGL61 we e as desc ibed [6,7]. Fo gene a ion o pEGFP-C1-RhoGL61, he co esponding DNA ag- men immobilized om pRK5-myc-RhoGL61 wi h BamHI/EcoRI was liga ed in o pEGFP-C1 ec o diges ed wi h BglII/EcoRI. pRK5-myc-Rac1L61- ΔCAAX was gene a ed by si e di ec ed mu agenesis using 5ʹ-GAGGAAGAGAAAATGACTGCTGTTGTA AGTC-3ʹas o wa d p ime . The ideli y o all con- s uc s was e i ied by sequencing. CRISPR/Cas9-media ed genome edi ing B16-F1 cells lacking unc ional CYFIP1 andCYFIP2 genes, as well as Rac1, Rac2 and Rac3 genes we e gene a ed by ea ing S a-1/PIR121 KO cells (clone #3) wi h pSpCas9 (BB)-2A-Pu o (PX459) ec o s a ge ing Rac1, Rac2 and Rac3 genes. Speci ically, cells we e co- ans ec ed wi h plasmids a ge ing ATGCAGGCCATCAAGTGTG (Rac1/2) and ATGCAGGCCATCAAGTGCG (Rac3) genomic egions as desc ibed [7]. A e pu omycin selec- ion o ans ec ed cells (3 days), cells we e ex ensi ely dilu ed and a ew days la e , mac oscopically isible colo- nies picked, o ob ain single cell-de i ed clones. De i ed cell clones al eady lacking S a-1/PIR121 we e sc eened o he addi ional absence o Rac exp ession by Wes e n Blo ing (see Figu e 1). Wes e n blo ing P epa a ion o whole cell lysa es was pe o med essen- ially as desc ibed [7]. Wes e n blo ing was ca ied ou using s anda d echniques. P ima y an ibodies used we e S a-1/PIR121 [15], Rac1/3 (23A8, Me ck), Rac2 [6], GAPDH (6C5, Calbiochem) and GFP (clones 7.1 and 13.1, Roche). HRP-conjuga ed seconda y an ibodies we e pu chased om In i ogen. Chemiluminescence signals we e ob ained upon incuba ion wi h ECL™P ime Wes e n Blo ing De ec ion Reagen (GE Heal hca e), and we e eco ded wi h ECL Chemocam image (In as, Goe ingen, Ge many). Immunop ecipi a ion Fo EGFP-immunop ecipi a ion expe imen s, S a-1/ PIR121 KO cells (clone #3) co-exp essing EGFP alone o EGFP- agged a ian s o cons i u i ely ac i e (Q61L) GTPases oge he wi h mChe y- agged S a-1 WCA* we e lysed wi h lysis bu e (1% T i on X-100, 8M. SCHAKS ET AL.