Drp1-mediated mitochondrial fission regulates calcium and F-actin dynamics during wound healing

RESEARCH ARTICLE D p1-media ed mi ochond ial ission egula es calcium and F-ac in dynamics du ing wound healing Susana Pon e 1 , La a Ca alho 1 , Ma ia Gaglia di 1 , Isabel Campos 2 , Paulo J. Oli ei a 3 and An onio Jacin o 1, * ABSTRACT Mi ochond ia adap o cellula needs by changes in mo phology h ough usion and ission e en s, e e ed o as mi ochond ial dynamics. Mi ochond ial unc ion and mo phology a e in ima ely connec ed and he dys egula ion o mi ochond ial dynamics is linked o se e al human diseases. In his wo k, we in es iga ed he ole o mi ochond ial dynamics in wound healing in he D osophila emb yonic epide mis. Mu an s o mi ochond ial usion and ission p o eins ail o close hei wounds, indica ing ha he egula ion o mi ochond ial dynamics is equi ed o wound healing. By li e- imaging, we ound ha loss o unc ion o he mi ochond ial ission p o ein Dynamin- ela ed p o ein 1 (D p1) comp omises he inc ease o cy osolic and mi ochond ial calcium upon wounding and leads o educed eac i e oxygen species (ROS) p oduc ion and F-ac in de ec s a he wound edge, culmina ing in wound healing impai men . Ou esul s highligh a new ole o mi ochond ial dynamics in he egula ion o calcium, ROS and F-ac in du ing epi helial epai . KEY WORDS: D p1, F-ac in, Calcium, Mi ochond ia, Mi ochond ial dynamics, Wound healing INTRODUCTION Mi ochond ia pe o m c i ical cellula unc ions such as ene gy p oduc ion, egula ion o calcium (Ca 2+ ), edox homeos asis and cell dea h (El-Ha ab and Scaglia, 2016). Mi ochond ial shape is con olled by an agonizing usion and ission e en s (Lewis and Lewis, 1914; Nunna i e al., 1997), desc ibed as mi ochond ial dynamics, which allow mi ochond ia o adap o cellula demands (Nunna i and Suomalainen, 2012). Dynamin- ela ed p o eins egula e mi ochond ial dynamics h ough hei GTPase ac i i y (Hoppins e al., 2007). Mi ochond ial ission is accomplished by Dynamin- ela ed p o ein 1 (D p1). Upon ac i a ion, D p1 is ec ui ed om he cy osol o he mi ochond ia, oligome izes and cons ic s his o ganelle un il i s di ision is achie ed (Bleaza d e al., 1999; Lab ousse e al., 1999; Smi no a e al., 2001; Yoon e al., 2001). Mi ochond ial usion equi es he me ging o bo h he ou e (OMM) and he inne mi ochond ial memb anes (IMM). Mi o usin 1 (M n1) and Mi o usin 2 (M n2) a e esponsible o OMM usion (Rojo e al., 2002), while Op ic a ophy 1 (Opa1) media es usion o he IMM (G ipa ic e al., 2004; Olichon e al., 2003). Regula ion o mi ochond ial dynamics is essen ial o de elopmen (Chen e al., 2003; Ishiha a e al., 2009; Wa e ham e al., 2007) and dys egula ion o i s machine y is implica ed in a wide ange o human diseases, including neu opa hies, ype II diabe es and cance (Ande son e al., 2018; Ranie i e al., 2013; Ro i a-Llopis e al., 2017). Howe e , he ole o mi ochond ial dynamics in o he con ex s, such as epi helial epai , is s ill la gely unknown. Wound healing in simple epi helia is cha ac e ized by he accumula ion o F-ac in and non-muscle myosin II (myosin) a he cell bounda ies ha ace he wound, o ming an ac omyosin cable ha con ac s and b ings cells oge he , he eby closing he hole (Bemen e al., 1999; Danjo and Gipson, 1998; Kieha e al., 2000; Xu and Chisholm, 2011). Addi ionally, wound healing in ol es cell c awling media ed by ac in p o usions (Ab eu-Blanco e al., 2012a; Ve boon and Pa khu s , 2015) and cellula ea angemen s (Ca alho e al., 2018; Razzell e al., 2014). Recen s udies sugges ha mi ochond ia migh be equi ed o issue epai by p oducing eac i e oxygen species (ROS). They ha e shown ha mi ochond ial ROS p omo e wound healing by egula ing F-ac in and myosin a he wound edge, ei he by ac ing on Rho GTPases (Muliyil and Na asimha, 2014; Xu and Chisholm, 2014) o on cell–cell junc ion emodelling (Hun e e al., 2018). In his wo k, we show ha he mi ochond ial dynamics machine y is essen ial o epai , as mu an s o hese p o eins ail o close epi helial wounds. In pa icula , he ission p o ein D p1 is equi ed o F-ac in accumula ion a he wound edge, o p ope cy osolic and mi ochond ial Ca 2+ dynamics and o ROS p oduc ion upon wounding. Ou wo k e eals a no el ole o mi ochond ial ission in egula ing ROS, Ca 2+ and F-ac in dynamics du ing epi helial epai . RESULTS Mi ochond ial dynamics p o eins a e equi ed o wound healing To es whe he he mi ochond ial dynamics machine y (Fig. 1A) is equi ed o epi helial epai , we pe o med a p e iously desc ibed wounding assay in he D osophila emb yonic epide mis (Campos e al., 2010). We lase -wounded la e-s age emb yos bea ing wild- ype and mu an alleles o mi ochond ial dynamics p o eins and assessed he wound-healing pheno ype by he pe cen age o non-healing wounds. Fig. 1A shows a scheme o mi ochond ial dynamics wi h all he es ed p o eins ep esen ed. Rega ding usion, we es ed ou mu an alleles and one miRNA o Mi ochond ial assembly egula o y ac o (Ma ,aD osophila M n homolog); as well as wo Opa1 mu an alleles and wo he e oallelic combina ions. Conce ning mi ochond ial ission, we es ed h ee D p1 alleles and h ee Recei ed 7 Oc obe 2019; Accep ed 9 Ma ch 2020 1 CEDOC, Ch onic Diseases Resea ch Cen e , NOVA Medical School/Faculdade de Ciências Medicas, Uni e sidade NOVA de Lisboa, 1169-056 Lisboa, Po ugal. 2 Animal Pla o ms, Champalimaud Cen e o he Unknown, 1400-038 Lisboa, Po ugal. 3 CNC, Cen e o Neu oscience and Cell Biology, Uni e si y o Coimb a, UC Bio ech Building, 3060-197 Can anhede, Po ugal. *Au ho o co espondence ([email p o ec ed]) S.P., 0000-0002-4297-1265; L.C., 0000-0001-6350-3279; M.G., 0000-0001- 6411-0901; I.C., 0000-0002-1541-8552; P.J.O., 0000-0002-5201-9948; A.J., 0000- 0002-4193-6089 This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h ps://c ea i ecommons.o g/licenses/by/4.0), which pe mi s un es ic ed use, dis ibu ion and ep oduc ion in any medium p o ided ha he o iginal wo k is p ope ly a ibu ed. 1 © 2020. Published by The Company o Biologis s L d | Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open he e oallelic combina ions. We also es ed o he ission egula o s: Fission p o ein 1 (Fis1), which ac s as a ecep o o D p1 a he OMM (Losón e al., 2013), and Ganglioside-induced di e en ia ion associa ed p o ein 1 (GDAP1), whose unc ion is no well unde s ood (Hube e al., 2013). We obse ed h ee ypes o wound closu e pheno ypes: open, in e media e and closed wounds (Fig. 1B). Closed wounds a e iden i iable by a small melanized spo . Open wounds show a melanized ing a ound he hole. In he in e media e pheno ype, melaniza ion occu s in a la ge ci cula a ea bu a clea hole is absen , making i unce ain whe he he wound is open o closed. Con ol emb yos (w 1118 ) ha e an ou s anding capaci y o epi helial epai , as 94.7% o he wounds a e closed (Fig. 1C). Mu a ions in ei he mi ochond ial ission o usion genes inc eased he equency o open and in e media e wounds (Fig. 1C). As i is unclea whe he he in e media e wounds ep esen a closu e impai men o jus a melaniza ion de ec , we excluded hese wounds om he s a is ical analysis o he wound healing pheno ype. Compa ing only closed e sus open wounds, all mi ochond ial ission mu an s showed highe pe cen age o open wounds han con ols (Fig. 1D). Rega ding mi ochond ial usion, om he ou es ed Ma alleles and he miRNA, only Ma J showed an inc eased pe cen age o open wounds compa ed o con ols. Opa1 mu an s showed a signi ican wound closu e pheno ype (Fig. 1D). As we obse ed wound-closu e de ec s o mu a ed e sions o bo h usion and ission p o eins, hese da a sugges ha he egula ion o mi ochond ial dynamics is necessa y o wound healing. D p1 mu an s show delayed wound healing Mi ochond ial ission mu an s showed a mo e consis en wound healing pheno ype han usion mu an s. The e o e, we decided o explo e he ole o mi ochond ial ission in epi helial epai by ocusing on he unc ion o D p1. Fig. 1. Mi ochond ial dynamics p o eins a e equi ed o wound healing. (A) Scheme o he p o eins in ol ed in mi ochond ial dynamics used in he wounding assay sc een. (B) Rep esen a i e images o ha ching la ae, 16 h a e wounding, showing he h ee obse ed wound pheno ypes: closed, in e media e and open. Closed wounds p esen a small scab, while open wounds show a ing o melaniza ion a ound he hole. In e media e wounds ha e mo e melaniza ion han closed and open wounds bu no a clea hole. A owheads poin o he wound. Scale ba : 200 µm. (C) G aph o pe cen age o closed, in e media e and open wounds in con ols (w 1118 ) and mu an alleles o mi ochond ial dynamics p o eins. (D) G aph o pe cen age o open wounds in con ols and mu an alleles o mi ochond ial dynamics p o eins. Rega ding usion, all Opa1 alleles and he e oallelic combina ions showed inc eased pe cen age o open wounds compa ed o con ols; o Ma , only he Ma J mu a ion shows signi ican ly inc eased pe cen age o open wounds compa ed o con ols. All he es ed ission genes and he e oallelic combina ions showed highe pe cen age o open wounds compa ed o con ols. Fishe ’s exac es was used o es o signi ican di e ences be ween g oups. UAS-Ma miRNAi was exp essed unde he con ol o he da-Gal4 d i e . The g aph in D shows he same emb yos om C, excluding hose wi h in e media e wounds. ns, no signi ican (P>0.05), **P≤0.01, ***P≤0.001, ****P<0.0001. The numbe o emb yos o each condi ion is shown below he ba s in C and D. 2 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open To unde s and he ole o D p1 in wound healing, we used spinning-disk mic oscopy o image con ol and D p1 mu an emb yos exp essing GFP::Moesin (Kieha e al., 2000), an F-ac in ma ke , and ollowed he dynamics o closu e (Mo ie 1). Con ol emb yos accumula e F-ac in a he wound edge (Fig. 2A) and he wound a ea p og essi ely dec eases un il he hole is closed (Fig. 2A,F). Al hough he ini ial a ea was simila in bo h condi ions (Fig. 2D), D p1 mu an wounds ook on a e age 128±34 min o close, signi ican ly longe han con ols (56±17 min) (Fig. 2E). In milde cases, D p1 mu an wounds closed a a slowe a e (Fig. 2B,F). In o he cases ( h ee ou o 13 D p1 mu an emb yos), he pheno ype was s onge ; al hough he wound con ac ed o abou 40 min pos -wounding (mpw), i s a ea began o inc ease again un il 120–130 mpw (Fig. 2C,F). A e his expansion phase, wounds con ac ed again, and in one case i was almos closed by he end o imaging (Fig. 2C 180 mpw, F). We quan i ied he wound a ea o con ol and D p1 mu an s in he i s 30 mpw and ound signi ican di e ences in he i s minu es a e wounding (4 mpw and 10 mpw o mild and s ong condi ions, espec i ely) (Fig. 2G). Ou esul s sugges ha inhibi ion o mi ochond ial ission impai s wound closu e. We nex asked whe he he induc ion o mi ochond ial ission has he opposi e e ec , accele a ing he wound healing p ocess. A e alida ing ha he o e exp ession o D p1 leads o induc ion o mi ochond ial ission, esul ing in a agmen ed mi ochond ial ne wo k (Fig. S1), we compa ed con ols and emb yos o e exp essing D p1 (UAS-D p1) and exp essing he F-ac in ma ke mChe y::Moesin (Milla d and Ma in, 2008) unde he con ol o a ubiqui ous d i e (da-Gal4) and ollowed he wound closu e dynamics o e ime (Fig. 2H,I). We ound no signi ican di e ences in ei he he ime o wound closu e (Fig. 2J) o in he wound a ea o e ime (Fig. 2K). This sugges s ha inc eased ission does no ha e an impac on wound closu e dynamics. These esul s show ha , while D p1 o e exp ession has no e ec on wound closu e, D p1 loss-o - unc ion impai s wound healing, sugges ing ha mi ochond ial ission is necessa y o wound epai egula ion. Wounding induces no majo changes in mi ochond ial mo phology Ou p e ious esul s sugges ha mi ochond ial ission is equi ed o p ope wound healing, so we wonde ed whe he wounding igge s changes in mi ochond ial mo phology, owa ds a mo e agmen ed mi ochond ial ne wo k. To analyse mi ochond ial mo phology, we used emb yos exp essing mi ochond ia (EYFP::mi o, Lajeunesse e al., 2004) and memb ane (PLCγPH::ChFP, He sz e g e al., 2013) ma ke s and compa ed con ol and D p1 mu an emb yos. D p1 mu an s showed longe mi ochond ia han con ols bo h be o e and upon wounding (Fig. 3A,B). Mi ochond ial mo phology quan i ica ion con i med ha he mi ochond ial leng h (Fig. 3C) was highe in D p1 mu an s han in con ols, while he numbe o b anches was simila (Fig. 3D), bo h be o e and upon wounding. In con ol emb yos, wounding led o a educ ion in he numbe o mi ochond ial b anches (Fig. 3D) bu he o e all mi ochond ial leng h was una ec ed (Fig. 3C), sugges ing ha wounding does no lead o majo mi ochond ial mo phology changes. Rega ding localiza ion and apicobasal dis ibu ion o mi ochond ia inside he cell, we did no obse e majo di e ences be ween con ol and D p1 mu an s, ei he be o e o upon wounding (Fig. 3A–Bi, XZ and YZ sec ions). Ou esul s sugges ha , al hough wounding does no s ongly in luence mi ochond ial mo phology, an elonga ed mi ochond ial ne wo k such as ha seen in D p1 mu an s is de imen al o wound healing. D p1 mu an s ha e F-ac in de ec s du ing wound closu e Al hough cells can compensa e o he loss o he ac omyosin cable (Ducuing and Vincen , 2016), his s uc u e is one o he main d i ing o ces o wound healing (Zulue a-Coa asa and Fe nandez- Gonzalez, 2017). The e o e, we checked whe he he wound healing pheno ype in D p1 mu an s was associa ed wi h ac omyosin cable de ec s. We imaged con ol and D p1 mu an emb yos exp essing GFP::Moesin (Kieha e al., 2000) and Zip::GFP (Lye e al., 2014) o compa e hei F-ac in and myosin le els. Bo h con ols and D p1 mu an emb yos accumula ed F-ac in (Fig. 4A,B) and myosin (Fig. 4C,D) a he wound edge. Howe e , F-ac in le els we e lowe in D p1 mu an s when compa ed o con ols (Fig. 4E). We ound no signi ican di e ences in myosin le els be ween D p1 mu an and con ol emb yos (Fig. 4F). These esul s sugges ha he wound healing pheno ype in D p1 mu an s migh be caused by de ec s in F-ac in bu no in myosin le els. The o ma ion o he ac omyosin cable depends on emodelling o he adhe ens junc ions (AJs) (Ab eu-Blanco e al., 2012a; Ca alho e al., 2014; Hun e e al., 2015; Ma subayashi e al., 2015). A e wounding, he AJ p o ein E-cadhe in (E-cad) is down egula ed a he cell bounda ies acing he wound, emaining only a he la e al junc ions o leading-edge cells. To es whe he he F-ac in de ec s obse ed in D p1 mu an s we e associa ed wi h E-cad emodelling de ec s, we imaged con ol and D p1 mu an emb yos exp essing ubi-E-cad::GFP (Oda and Tsuki a, 1999) and mChe y::Moesin (Milla d and Ma in, 2008) be o e and upon wounding. We obse ed no signi ican di e ences in E-cad le els o con ol and D p1 mu an emb yos, ei he be o e o a e wounding (Fig. S2). In summa y, we p opose ha D p1 egula es F-ac in dynamics du ing wound closu e, independen ly o AJs emodelling. D p1 mu an s ha e al e ed cy osolic and mi ochond ial calcium dynamics The i s signal o be de ec ed upon wounding is an in acellula Ca 2+ bu s (An unes e al., 2013; Razzell e al., 2013; Sammak e al., 1997; Xu and Chisholm, 2011; Cla k e al., 2009). This Ca 2+ inc ease egula es many wound-healing s eps, including ac omyosin cable o ma ion (An unes e al., 2013; Xu and Chisholm, 2011). Mi ochond ia a e known egula o s o Ca 2+ homeos asis (Finkel e al., 2015; Gio gi e al., 2008; Rizzu o e al., 2012), so we asked whe he he F-ac in de ec s obse ed upon D p1 loss-o - unc ion could esul om impai ed Ca 2+ dynamics. We imaged emb yos exp essing he GCaMP6 Ca 2+ senso (Chen e al., 2013) and measu ed Ca 2+ le els be o e and upon wounding (Mo ie 2). As p e iously desc ibed (Razzell e al., 2013), wounding induces a d ama ic and ansien inc ease in cy osolic Ca 2+ (cy Ca 2+ ) le els in he cells a ound he wound, ha p opaga es in a wa e-like manne eaching abou wo o i e cell laye s away om he wound, depending on he wound size (Fig. 5A). In D p1 mu an emb yos, he cy Ca 2+ bu s was less p onounced han in con ols (Fig. 5B,C). Mo eo e , he a ea in which Ca 2+ inc ease was obse ed was signi ican ly educed in D p1 mu an s compa ed o con ols (Fig. 5D), sugges ing ha impai ing D p1 unc ion a ec s no only Ca 2+ le els bu also he in e cellula Ca 2+ p opaga ion. Mi ochond ia can up ake Ca 2+ om he cy osol, he eby modula ing cy Ca 2+ (Szabadkai and Duchen, 2008). As mi ochond ial mo phology in luences mi ochond ial Ca 2+ (mi Ca 2+ ) le els (Bianchi e al., 2006; Ge encse and Adam-Vizi, 3 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open Fig. 2. D p1 emb yos show delayed wound healing. (A–C) Maximum Z p ojec ions o he epide mis o con ol (A), D p1 mild (B) and D p1 s ong (C) mu an emb yos exp essing an F-ac in ma ke (GFP::Moesin) du ing wound closu e. In D p1 mild mu an s (B) wounds close slowe han in con ols (compa e B wi h A). In D p1 s ong mu an s (C), al hough he wound con ac s in he i s 30–40 mpw, i hen s a s o expand (see 60–120 mpw). La e on, he wound con ac s again and by 180 mpw i is almos closed. (D) G aph o a e age ini ial wound a ea in con ol and D p1 mu an emb yos (s ong and mild). (E) G aph o wound closu e ime in con ol and D p1 mu an emb yos. Al hough he ini ial wound a ea o con ol and D p1 mu an s is simila (D), D p1 mu an s ake longe o close hei wounds (E). Unpai ed - es wi h Welch’s co ec ion was pe o med o es o signi ican di e ences be ween g oups in D and E. ns, no signi ican (P>0.05), ****P≤0.0001. (F) G aph o a e age wound a ea in con ol, D p1 mild and D p1 s ong mu an s o e ime. D p1 mild mu an wounds close slowe han con ols. D p1 s ong mu an wounds ini ially con ac bu s a o expand a e 40 mpw. A 120–130 mpw wounds s a o con ac again. (G) G aph o a e age wound a ea in con ol, D p1-mild and D p1-s ong mu an s in he i s 30 mpw, co esponding o he g ey egion in F. Signi ican di e ences be ween con ol and D p1 mu an s s a a 4 mpw in D p1-mild mu an s and a 10 mpw in D p1-s ong mu an s. A wo-way ANOVA wi h a Tukey’s co ec ion o mul iple compa isons was used o es o signi ican di e ences be ween g oups in G. As e isks (*) e e o con ol and D p1-mild mu an s’compa isons. Numbe signs (#) e e o con ol and D p1-s ong mu an compa isons. Dashed lines depic an in e al o poin s in which he compa ison be ween g oups gi es he same deg ee o s a is ical signi icance, gi en by he symbols abo e. #, P≤0.05, **; ##, P≤0.01; ***P≤0.001; ****P≤0.0001. E o ba s ep esen s.e.m. Numbe o emb yos pe condi ion is shown in each g aph. (H,I) Maximum Z p ojec ions o he epide mis o con ol (H) and D p1-o e exp essing (UAS-D p1) (I) emb yos exp essing an F-ac in ma ke (mChe y::Moesin) ubiqui ously unde he con ol o he da- Gal4 d i e du ing wound closu e. The wound-closu e dynamics a e simila be ween he wo g oups. (J) G aph o wound closu e ime in con ol and UAS-D p1 emb yos. Unpai ed - es wi h Welch’s co ec ion was pe o med o es o signi ican di e ences be ween g oups. (K) G aph o a e age wound a ea in con ol and UAS-D p1 emb yos o e ime. No signi ican di e ence was ound be ween con ol and D p1-o e exp essing emb yos, nei he in he ime o wound closu e no he wound closu e dynamics. A wo-way ANOVA wi h a Sidak co ec ion o mul iple compa isons was used o es o signi ican di e ences be ween g oups. ns, no signi ican (P>0.05). E o ba s ep esen s.e.m. Numbe o emb yos pe condi ion is shown in each g aph. Scale ba s: 20 µm. 4 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open 2005; Szabadkai e al., 2004), we examined con ol and D p1 mu an emb yos exp essing a mi ochond ia- a ge ed GCaMP3 Ca 2+ senso (mi o::GCaMP3, Lu as e al., 2012) be o e and upon wounding (Mo ie 3). Simila o wha was de ec ed o cy Ca 2+ ,we obse ed an inc ease in mi Ca 2+ a ound he wound in bo h con ol (Fig. 5E) and D p1 mu an (Fig. 5F) epide mis. Quan i ica ion o mi Ca 2+ in ensi y showed a educed esponse upon wounding in D p1 mu an emb yos compa ed o con ols (Fig. 5G, 0 mpw). This educ ion is no as d ama ic as ha seen o cy Ca 2+ , which may be due o he di e en sensi i i y o he cy Ca 2+ senso . No di e ences we e ound in he a ea o inc eased mi Ca 2+ (Fig. 5H), sugges ing ha only he cy Ca 2+ p opaga ion is a ec ed. To u he in es iga e he impac o mi ochond ial ission impai men in wound closu e, we es ed how ano he componen o he ission machine y egula es wound healing. We knocked down Fis1, a D p1 ecep o (Losón e al., 2013), by ubiqui ously exp essing RNAi agains Fis1 and analysing i s e ec on Ca 2+ and F-ac in dynamics upon wounding (Fig. S3). Bo h he cy Ca 2+ and mi Ca 2+ bu s s upon wounding we e educed in Fis1 RNAi- exp essing emb yos compa ed o con ols (Fig. S3A,B,C,E,F,G), al hough no e ec on Ca 2+ p opaga ion ac oss he epide mis was de ec ed (Fig. S3D,H). On he o he hand, no signi ican di e ences in F-ac in le els we e ound be ween Fis1 knockdown and con ol emb yos (Fig. S3I–K). These esul s suppo he hypo hesis ha mi ochond ial ission egula es Ca 2+ dynamics du ing wound closu e. D p1 mu an s show educed mi ochond ial ROS p oduc ion upon wounding Ca 2+ -dependen ROS p oduc ion upon wounding has been shown o egula e he F-ac in cy oskele on and he wound healing esponse bo h in Caeno habdi is elegans and D osophila (Hun e e al., 2018; Xu and Chisholm, 2014). Mi ochond ia a e known sou ces o ROS (Mu phy, 2009) and hei abili y o p oduce hem can be egula ed by mi ochond ial dynamics. F agmen a ion o he mi ochond ial ne wo k is associa ed wi h inc eased ROS p oduc ion and inhibi ion o mi ochond ial ission can educe oxida i e s ess (Galloway e al., 2012). As ou esul s show ha D p1 egula es Ca 2+ and F-ac in dynamics du ing wound healing, we hypo hesized ha ROS p oduc ion migh be he link be ween hese wo wound closu e e en s. To in es iga e whe he D p1 mu an s ha e al e ed ROS p oduc ion, we imaged con ol and D p1 mu an emb yos exp essing a gene ically-encoded a iome ic mi ochond ial g een luo escen p o ein ha shi s i e e sibly o ed luo escence when oxidized (Lake e al., 2014). Upon wounding, we obse ed an inc ease in he ed signal, an indica o o ROS p oduc ion, bo h in con ol and in D p1 mu an s (Fig. 6A,B). To quan i y ROS p oduc ion, we calcula ed he ed: g een luo escence in ensi y a io and obse ed ha he ROS le els p oduced upon wounding we e signi ican ly lowe in D p1 mu an s compa ed o con ol emb yos (Fig. 6C). Ou esul s show ha impai ing mi ochond ial ission leads o educed mi ochond ial ROS p oduc ion in esponse o wounding. Fig. 3. Wounding does no induce majo changes in mi ochond ial mo phology. (A–Bi) Maximum Z p ojec ions o he epide mis o con ol (A,Ai) and D p1 (B,Bi) mu an emb yos exp essing ubiqui ous mi ochond ial (EYFP::mi o, g een) and memb ane (PLCγPH::ChFP, magen a) ma ke s. XZ and YZ sec ions a e shown below and on he igh , espec i ely. Inse s show a zoom o he dashed egion o he espec i e image. Scale ba : 10 µm. Inse scale ba : 5 µm. (C) G aph o a e age numbe o b anches in con ol and D p1 mu an s, be o e and upon wounding. Con ol and D p1 mu an s show simila numbe s o mi ochond ial b anches. Wounding leads o a educ ion o b anching in con ols bu no in D p1 mu an s. (D) G aph o a e age mi ochond ial leng h in con ol and D p1 mu an s, be o e and upon wounding. D p1 mu an mi ochond ial ne wo k has inc eased leng h, compa ed o con ols, bo h be o e and a e wounding. Mi ochond ia om wounded epide mis show a simila leng h compa ed o unwounded, in bo h con ol and D p1 mu an s. A Mann-Whi ney U es was used o es o signi ican di e ences be ween g oups. ns, no signi ican (P>0.05); *P=0.0275; ***P=0.0001; ****P<0.0001. n(con ol)=22 cells om nine emb yos, n(D p1)=29 cells om 12 emb yos. E o ba s ep esen s.d., bw, be o e wounding; mpw, minu es pos wounding. 5 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open The Rho GTPase e ec o Pkn is down egula ed a he wound edge in D p1 mu an s To u he in es iga e he mechanisms h ough which mi ochond ial ission egula es he wound-healing esponse, we assessed whe he known F-ac in modula o s we e dys egula ed in D p1 mu an s. F-ac in dynamics and myosin con ac ili y a he wound edge ha e been shown o be modula ed by membe s o he Rho amily o GTPases, Rho1, Rac and Cdc42 (Ve boon and Pa khu s , 2015). Mo eo e , he egula ion o hese GTPases is media ed by Ca 2+ and ROS igge ed upon wounding (So o e al., 2013; Xu and Chisholm, 2014). Rho and i s e ec o s coo dina e ac omyosin cable o ma ion and con ac ili y, while Rac and Cdc42 a e mo e impo an o o m F-ac in p o usions ha coo dina e cell mig a ion a he leading edge and he inal kni ing o he epi helium a he end o closu e (Ab eu- Blanco e al., 2012b; Ve boon and Pa khu s , 2015). As we obse ed de ec s in he F-ac in accumula ion a he wound edge in D p1 mu an s, we decided o ocus on Rho e ec o s: Rho kinase (Rok), Diaphanous (Dia) and P o ein kinase N (Pkn). Rok ac i a es he myosin egula o y ligh chain, di ec ly by phospho yla ion o by inac i a ion o myosin phospha ases, hus p omo ing ac omyosin con ac ili y (Kimu a e al., 1996; Ueda e al., 2002). Dia is a o min ha p omo es he polyme iza ion o unb anched F-ac in (Na umiya e al., 1997). Pkn has been implica ed in he egula ion o cell mig a ion, bu how i exe s i s unc ion is no ully unde s ood (Lachmann e al., 2011; Lim e al., 2004). We imaged emb yos exp essing GFP- agged e sions o he Rok, Dia and Pkn (Fig. 7). Fo he analysis o Rok and Pkn, we compa ed con ols wi h D p1 mu an emb yos, while o Dia we used RNAi o knockdown D p1. Ou o he h ee Rho1 e ec o s, only Pkn showed a signi ican educ ion in i s accumula ion a he wound edge in D p1 mu an s when compa ed o con ols (Fig. 7D–F). The accumula ion o Rho1 e ec o s a he wound edge is a consequence o hei ac i a ion by Rho1, so ou esul s sugges ha D p1 loss-o - unc ion leads o a educ ion in Rho1 ac i i y upon wounding. In summa y, we ha e iden i ied mi ochond ial-dynamics p o eins as no el emb yonic wound-healing egula o s. Ou da a show ha inhibi ion o mi ochond ial ission by D p1 loss-o - unc ion leads o de ec s in Ca 2+ , ROS and F-ac in dynamics upon wounding, culmina ing in wound-healing impai men . DISCUSSION Epi helial issues a e c i ical o p o ec us om he ex e nal en i onmen (Lowe and Ande son, 2015). Unde s anding how Fig. 4. D p1 mu an s show ac in de ec s du ing wound closu e. (A–D) Maximum Z p ojec ions o he epide mis o con ol (A,C) and D p1 (B,D) mu an emb yos exp essing an F-ac in (GFP::Moesin) (A,B) and a Myosin (Zip::GFP) (C,D) ma ke be o e and a e wounding. Images a e pseudo-colo ed wi h a g adien o luo escence in ensi y, anging om blue (low) o yellow (high). Al hough no di e ences be ween con ols and D p1 mu an s a e e iden be o e wounding, D p1 mu an emb yos accumula e less F-ac in a he wound edge han con ols (compa e A,B). Myosin accumula ion a he wound edge seems simila be ween con ol and D p1 mu an emb yos (compa e C,D). Scale ba : 20 µm. (E) G aph o a e age F-ac in in ensi y a he cell co ex be o e wounding and a he wound edge. F-ac in le els a e signi ican ly educed in D p1 mu an s a 10 and 20 mpw. (F) G aph o a e age Myosin in ensi y a he cell co ex be o e wounding and a he wound edge. No signi ican di e ences we e ound be ween con ol and D p1 mu an s. A wo-way ANOVA wi h a Sidak co ec ion o mul iple compa isons was used o es o signi ican di e ences be ween g oups in E and F. Only signi ican di e ences (P≤0.05) a e ep esen ed. *P<0.05. E o ba s ep esen s.e.m. Numbe o emb yos pe condi ion is shown in each g aph. a.u., a bi a y uni s; bw, be o e wounding; mpw, minu es pos wounding. 6 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open epi helial issues d i e e icien wound epai is o he upmos impo ance o he biomedical ield. In his s udy, we used a model o simple epi helial wound closu e, he emb yonic epide mis o he ui ly, o unco e he ole o mi ochond ial dynamics in he wound-healing esponse. Li le is known abou how mi ochond ia con ibu e o issue epai , besides hei in ol emen in he p oduc ion o ROS, which in u n can egula e he wound-healing p ocess (Hun e e al., 2018; Sanchez e al., 2018; Xu and Chisholm, 2014). Some da a suppo he conclusion ha mi ochond ial dynamics can egula e cell mig a ion, an essen ial p ocess in wound epai (Ko e al., 2017; Zhao e al., 2013), bu whe he his is ele an o emb yonic wound epai had ne e been add essed. We s a ed by pe o ming a gene ic sc een o unde s and whe he he mi ochond ial dynamics machine y is equi ed o wound closu e. Mi ochond ial dynamics is media ed by la ge GTPases, namely D p1, ha media es mi ochond ial ission; and Opa1 and M ns (Ma in D osophila) ha con ol he usion o he mi ochond ial memb anes. Ou o he es ed ou Ma alleles and miRNA-media ed knockdown, only Ma J mu an s showed an inc eased numbe o open wounds compa ed o con ols. Mos o he Ma alleles ha e no been cha ac e ized, so i is unclea why only Ma J mu an s showed a wound-healing pheno ype, being possible ha his is an unspeci ic e ec o his allele. I is also concei able ha he pe du ance o ma e nal con ibu ion masks he e ec s o Ma loss-o - unc ion in he case o he emaining Ma alleles. On he o he hand, bo h Opa1 mu an alleles showed a signi ican wound-healing pheno ype, a ou ing he hypo hesis ha mi ochond ial usion is necessa y o p ope emb yonic wound healing. Fu u e s udies ocusing on mi ochond ial usion a e impo an o complemen ou esul s. Rega ding mi ochond ial Fig. 5. D p1 mu an s show al e ed cy osolic and mi ochond ial Ca 2+ dynamics. (A,B) Maximum Z p ojec ions o he epide mis o con ol (A) and D p1 (B) mu an emb yos exp essing a cy osolic Ca 2+ senso (GCaMP6 ) be o e and a e wounding. Bo h con ol and D p1 mu an cells a ound he wound d ama ically inc ease cy osolic Ca 2+ le els immedia ely upon wounding (0 mpw). In ensi y e u ns o p e-wound le els a e 15 min. Ca 2+ le els and a ea o cells ha espond o he wound a e lowe in D p1 mu an s (B, 0 mpw) compa ed o con ols (A, 0 mpw). (C) G aph o cy osolic Ca 2+ in ensi y shows ha cy osolic Ca 2+ is lowe in D p1 mu an s compa ed o con ols in he i s 2.5 mpw. (D) G aph o a e age a ea o ele a ed cy osolic Ca 2+ shows ha he Ca 2+ bu s a ea is lowe in D p1 mu an s compa ed o con ols om 0 o 1 mpw. (E,F) Maximum Z p ojec ions o he epide mis o con ol (E) and D p1 mu an (F) emb yos exp essing a mi ochond ial Ca 2+ senso (mi o::GCaMP3) be o e and a e wounding. Wounding igge s an inc ease in mi ochond ial Ca 2+ le els in bo h con ol and D p1 mu an cells a ound he wound (E,F a 0 mpw). (G) G aph o mi ochond ial Ca 2+ in ensi y in con ol and D p1 mu an s. D p1 mu an s ha e a educed mi ochond ial Ca 2+ bu s a 0 mpw, compa ed o con ols. (H) G aph o a e age a ea o ele a ed mi ochond ial Ca 2+ in con ols and D p1 mu an emb yos. No signi ican di e ences we e ound be ween con ol and D p1 mu an s. Images a e pseudo-colo ed wi h a g adien o luo escence in ensi y, anging om blue (low) o yellow (high). Dashed lines show he wound bounda ies. Scale ba : 20 µm. A wo-way ANOVA wi h a Sidak co ec ion o mul iple compa isons was used o es o signi ican di e ences be ween g oups in C, D, G and F. Only signi ican di e ences a e ep esen ed: *P≤0.05, **P≤0.01, ****P≤0.0001. E o ba s ep esen s.e.m. Numbe o emb yos pe condi ion is shown in each g aph. bw, be o e wounding; mpw, minu es pos wounding. 7 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open ission, D p1 KG03815 , a loss o unc ion allele, showed he s onges pheno ype compa ed wi h he o he D p1 mu an alleles. No much de ail is a ailable abou how he di e en mu a ions a ec D p1 unc ion. Howe e , all di e en he e oallelic combina ions esul ed in an inc ease in open wounds, which indica es ha he D p1 loss- o - unc ion pheno ype is consis en . O he playe s o mi ochond ial ission, Fis1 and Gdap1, a e also equi ed o wound healing, as hei mu a ion leads o an inc eased numbe o unhealed wounds. Al oge he , hese da a s ongly implica e mi ochond ial ission in he egula ion o epi helial epai . We ound ha he wound-closu e dynamics o D p1 mu an s is signi ican ly a ec ed, as hese emb yos ake mo e han wo imes longe o close he wounds han con ol emb yos. We obse ed wo deg ees o wound closu e pheno ypes: a mild pheno ype, cha ac e ized by a slowe wound closu e a e han in con ols; and a s ong pheno ype, in which he wound a ea expands. I is unclea why some wounds p esen his expansion phase, bu a simila pheno ype has been desc ibed in mu an s o a componen o he in e eb a e occluding junc ions (OJs). OJ dis up ion leads o de ec s in he ac omyosin cable, cellula shapes and ea angemen s as well as in issue mechanical p ope ies (Ca alho e al., 2018). I would be in e es ing o explo e he link be ween mi ochond ial dynamics and OJs in wound healing in u u e s udies. A ecen s udy in he C. elegans epide mis showed ha inc eased mi ochond ial ission, ei he by impai ing usion o by d ug- media ed ission induc ion, led o accele a ed wound closu e (Fu e al., 2020). In e es ingly, we obse ed no di e ence in wound- healing dynamics be ween D p1-o e exp essing emb yos and con ols. Mo eo e , we obse ed no majo di e ences in mi ochond ial mo phology be ween wounded and unwounded emb yos, while in C. elegans wounding igge s mi ochond ial ission (Fu e al., 2020). I seems ha mi ochond ial dynamics can ha e di e en e ec s on wound healing, and ice e sa, depending on he con ex . In he case o he D osophila emb yonic epide mis, an elonga ed mi ochond ial ne wo k, such as obse ed in D p1 mu an s, is p ejudicial o wound healing, while inc easing he agmen a ion o mi ochond ia has no impac on he wound-closu e a e. We hus p opose ha he e is an op imal mi ochond ial mo phology ha a ou s p ope wound closu e. To unde s and he mechanism h ough which mi ochond ial ission egula es wound healing, we cha ac e ized he known wound-healing e en s in D p1 mu an s, such as he o ma ion o he ac omyosin cable a he wound edge (Kieha e al., 2000; Wood e al., 2002; Xu and Chisholm, 2011). D p1 mu an s show de ec s in he F-ac in accumula ion a he wound edge, which may be he cause o he obse ed wound-healing impai men . The o ma ion o he ac omyosin cable is known o depend on se e al ac o s, such as an in acellula Ca 2+ inc ease upon wounding (An unes e al., 2013; Razzell e al., 2013; Xu and Chisholm, 2011), he emodelling o he AJs (Ab eu-Blanco e al., 2012a; Ca alho e al., 2014; Hun e e al., 2015), ROS p oduc ion (Hun e e al., 2018; Xu and Chisholm, 2014) and he ac i a ion o he Rho amily o GTPases and hei a ge s (B ock, 1996; Ve boon and Pa khu s , 2015; Wood e al., 2002). In addi ion o he es ablished cy Ca 2+ bu s , we also obse ed a apid inc ease in mi Ca 2+ le els upon wounding, consis en wi h wha has been seen in C. elegans wound epai (Xu and Chisholm, 2014), sugges ing ha , simila ly o he cy Ca 2+ bu s , his is a conse ed esponse o issue inju y. Rema kably, bo h wound- induced cy Ca 2+ and mi Ca 2+ bu s s we e educed upon D p1 loss- o - unc ion, sugges ing ha mi ochond ial ission s ongly impac s on Ca 2+ dynamics du ing wound closu e. I is well es ablished ha mi ochond ia can ake up Ca 2+ om he cy osol, he eby modula ing cy Ca 2+ le els (Szabadkai and Duchen, 2008). Howe e , he ela ionship be ween cy Ca 2+ and mi Ca 2+ in he con ex o wound healing is no clea . Inju y igge s Ca 2+ in lux om he ex acellula en i onmen (An unes e al., 2013; Razzell e al., 2013; Xu and Chisholm, 2011). The ele a ed cy Ca 2+ le els induce Ca 2+ elease om he endoplasmic e iculum (ER) media ed by he inosi ol-3-phospha e (IP3) ecep o (IP3R), ollowed by p opaga ion o Ca 2+ and IP3 o neighbou ing cells h ough gap Fig. 6. D p1 mu an emb yos show educed mi ochond ial ROS p oduc ion upon wounding. (A,B) Maximum Z p ojec ions o he wound egion o con ol (A) and D p1-mu an (B) emb yos exp essing he mi ochond ial ROS senso Mi oTime ubiqui ously, be o e and upon wounding (0 mpw). This epo e gene encodes a p o ein ha i e e sibly changes i s luo escence spec um om g een o ed upon oxida ion. Images show he g een and ed channels o each emb yo, as well as he ed:g een a io a e image p ocessing. Red:g een a io images a e pseudo-colo ed wi h a g adien o luo escence in ensi y, anging om blue (low) o yellow (high). Scale ba : 10 µm. (C) G aph o he a e age ed:g een a io o con ol and D p1 mu an emb yos, be o e and a e wounding (0 mpw). Red:g een a io is a measu e o ROS le els. P e-wound ROS le els a e simila be ween con ol and D p1 mu an s. Wounding inc eases ROS le els, bo h in con ols and D p1 mu an s bu his inc ease is lowe in D p1 mu an emb yos compa ed o con ols. A wo-way ANOVA wi h a Sidak co ec ion o mul iple compa isons was used o es o signi ican di e ences be ween g oups. ns, no signi ican , *P=0.0170, **P=0.0049, ****P<0.0001. E o ba s ep esen s.e.m. Numbe o emb yos pe condi ion is shown in he g aph. 8 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open junc ions (Na ciso e al., 2015; Razzell e al., 2013; Res epo and Basle , 2016). In o he cellula con ex s, mi ochond ia localize close o he ER, o ming Ca 2+ signalling mic odomains. Ca 2+ up ake by mi ochond ia educes he cy Ca 2+ le els close o he open ER channels (local cy Ca 2+ ), p e en ing hei Ca 2+ -dependen inac i a ion. By con olling ER Ca 2+ channels ac i i y, mi Ca 2+ up ake a ec s global cy Ca 2+ (Billups and Fo sy he, 2002; Rizzu o e al., 2012). Ou esul s lead us o specula e ha in D p1 mu an s he Ca 2+ bu e ing capaci y o mi ochond ia is comp omised, leading o IP3R inhibi ion and lowe global cy Ca 2+ le els. This could also a ec he Ca 2+ wa e p opaga ion, as less Ca 2+ and/o IP3 would c oss gap junc ions. Indeed, D p1 mu an s showed a educ ion in he a ea o Ca 2+ inc ease, which may indica e ha he Ca 2+ wa e p opaga ion is a ec ed. Mi ochond ial shape, numbe and dis ibu ion can a ec hei con ac s wi h he ER and impac on mi Ca 2+ p opaga ion. Based on p e ious s udies, he inhibi ion o D p1 wi h consequen elonga ed mi ochond ial ne wo k should a ou he p oximi y wi h he ER and acili a e mi Ca 2+ up ake (Cie i e al., 2018; Szabadkai e al., 2006). He e, we obse e he opposi e e ec , sugges ing ha he ole o D p1 on mi Ca 2+ egula ion may be con ex dependen . Fu he wo k is needed o unde s and how D p1 egula es mi ochond ial Ca 2+ up ake in he D osophila epide mis. The wound-induced Ca 2+ bu s igge s he p oduc ion o mi ochond ial ROS, which hen egula es he ac omyosin cable o ma ion, ei he by ac i a ing RHO-1 o by egula ing AJ emodelling (Hun e e al., 2018; Xu and Chisholm, 2014). Ou esul s show ha he absence o unc ional D p1 leads o educed Fig. 7. D p1 mu an s show educed accumula ion o Pkn a he wound edge, bu no o Rok o Dia. (A,B) Maximum Z p ojec ions o con ol (A) and D p1-mu an (B) emb yos exp essing he Rok::GFP, be o e wounding and a 10 and 30 mpw. (C) G aph o he a e age Rok::GFP in ensi y a he wound edge in con ols and D p1 mu an s a 10 and 30 mpw, no malized o p e-wound le els. No signi ican di e ences in Rok accumula ion a he wound edge we e ound be ween con ols and D p1 mu an s. (D,E) Maximum Z p ojec ions o con ol (D) and D p1 mu an s (E) exp essing he Dia::GFP, be o e wounding and a 20 mpw. (F) G aph o he a e age Dia::GFP in ensi y a indi idual cell junc ions a he wound edge in con ol and D p1-mu an emb yos a 20 mpw, no malized o hei espec i e in ensi y be o e wounding. No signi ican di e ences in Dia localiza ion a he wound edge we e ound be ween con ols and D p1 mu an s. (G,H) Maximum Z p ojec ions o con ol (G) and D p1-mu an (H) emb yos exp essing he Pkn-GFP a 10 and 30 mpw. (I) G aph o he a e age Pkn-GFP in ensi y a he wound edge in con ols and D p1 mu an s a 10 and 30 mpw, no malized o backg ound le els. D p1 mu an s show a educed accumula ion o Pkn a he wound edge compa ed o con ols. E o ba s ep esen s.e.m. A wo-way ANOVA wi h a Sidak co ec ion o mul iple compa isons was used o es o signi ican di e ences be ween g oups in C and I. A Wilcoxon es was used o es o signi ican di e ences be ween g oups in F. The numbe o emb yos pe condi ion is shown in C and I. In F, n(Con ol)=82 junc ions om nine emb yos and n(D p1)=99 cell junc ions om en emb yos. Scale ba s: 20 µm. mpw, minu es pos wounding. 9 RESEARCH ARTICLE Biology Open (2020) 9, bio048629. doi:10.1242/bio.048629 Biology Open