Intracellular vesicle trafficking plays an essential role in mitochondrial quality control

Volume 29 Ap il 1, 2018 809 MBoC | ARTICLE In acellula esicle a icking plays an essen ial ole in mi ochond ial quali y con ol ABSTRACT The D osophila gene p oduc s Be 1, Slh, and CG10144, p edic ed o unc ion in in acellula esicle a icking, we e p e iously ound o be essen ial o mi ochond ial nucle- oid main enance. He e we show ha Slh and Be 1 coope a e o main ain mi ochond ial unc- ions. In hei absence, mi ochond ial con en , memb ane po en ial, and espi a ion became abno mal, accompanied by mi ochond ial p o eo oxic s ess, bu wi hou di ec e ec s on m DNA. Immunocy ochemis y showed ha bo h Slh and Be 1 a e localized a he Golgi, oge he wi h a p opo ion o Rab5-posi i e esicles. Some Be 1, as well as a iny amoun o Slh, co ac iona ed wi h highly pu i ied mi ochond ia, while li e-cell imaging showed coinci- dence o luo escen ly agged Be 1 wi h mos Lyso acke -posi i e and a small p opo ion o Mi o acke -posi i e s uc u es. This h ee-way associa ion was dis up ed in cells knocked down o Slh, al hough colocalized lysosomal and mi ochond ial signals we e s ill seen. Nei- he Slh no Be 1 was equi ed o global mi ophagy o endocy osis, bu p olonged Slh knockdown esul ed in G2 g ow h a es , wi h inc eased cell diame e . These e ec s we e sha ed wi h knockdown o be aCOP bu no o CG1044, Snap24, o Syn axin6. Ou indings implica e esicle so ing a he cis-Golgi in mi ochond ial quali y con ol. INTRODUCTION Mi ochond ia a e mul i unc ional, dynamic o ganelles esponsible o hund eds o biochemical eac ions. Al hough he mos impo - an unc ion o mi ochond ia is cellula ene gy conse a ion h ough oxida i e phospho yla ion (OXPHOS), hey a e also in ol ed in many o he p ocesses, including egula ion o calcium le els, bio- genesis o i on–sul u clus e s, s e ol, s e oid, and phospholipid bio- syn hesis, and egula ion o apop osis ( o e iew see Smi h e al., 2012). A heal hy mi ochond ial popula ion is c ucial o cellula su - i al and unc ioning, and his depends on se e al quali y con ol p ocesses (Suliman and Pian adosi, 2016). De ec s in hese p o- cesses p e en he clea ance o dys unc ional mi ochond ial com- ponen s, leading o he accumula ion o mi ochond ial damage and e en ually o me abolic ailu e, wi h pa hological consequences (Cenini and Voos, 2016; Liang and Kobayashi, 2016; Romanello and Sand i, 2016). Mi ochond ial quali y con ol mechanisms include he mi ochond ial un olded p o ein esponse (UPRm ; Haynes e al., 2013), he cycle o mi ochond ial usion and ission (Wai and Lange , 2016), mi ophagy (Lemas e s, 2014; Hamache -B ady and B ady, 2016), seques a ion in o endosomes (Hamme ling e al., 2017), and he ecen ly desc ibed deg ada ion o mi ochond ial-de i ed esi- cles (MDVs; Soubannie e al., 2012; Lemas e s, 2014; Sugiu a e al., 2014). The UPRm is ac i a ed upon mi ochond ial s ess caused by he accumula ion o un olded o mis olded p o eins. The mi ochond ia possess se e al chape ones and p o eases ha un old and hen Moni o ing Edi o Thomas D. Fox Co nell Uni e si y Recei ed: Oc 30, 2017 Accep ed: Jan 11, 2018 This a icle was published online ahead o p in in MBoC in P ess (h p://www .molbiolcell.o g/cgi/doi/10.1091/mbc.E17-10-0619) on Janua y 17, 2018. The au ho s decla e no con lic o in e es . †P esen add ess: CNR Ins i u e o Neu oscience and Depa men o Biomedical Sciences, Uni e si y o Padua, 35131 Padua, I aly. Au ho con ibu ions: M.G. codesigned he s udy, pe o med he expe imen s, and cow o e he manusc ip ; G.C. pe o med biochemical analyses o mi ochon- d ial unc ion; J.G.de C. pe o med subcellula ac iona ion and Wes e n blo ing; H.T.J. codesigned he s udy, compiled he igu es, and cow o e he manusc ip . *Add ess co espondence o: Howa d T. Jacobs ([email p o ec ed]). © 2018 Ge a ds e al. This a icle is dis ibu ed by The Ame ican Socie y o Cell Biology unde license om he au ho (s). Two mon hs a e publica ion i is a ail- able o he public unde an A ibu ion–Noncomme cial–Sha e Alike 3.0 Unpo ed C ea i e Commons License (h p://c ea i ecommons.o g/licenses/by-nc-sa/3.0). “ASCB®,” “The Ame ican Socie y o Cell Biology®,” and “Molecula Biology o he Cell®” a e egis e ed adema ks o The Ame ican Socie y o Cell Biology. Abb e ia ions used: ER, endoplasmic e iculum; FCCP, ca bonyl cyanide-4-( i luo o- me hoxy)phenylhyd azone; GFP, g een luo escen p o ein; m DNA, mi ochond ial DNA; RNAi, RNA in e e ence; TMRM, e ame hyl hodamine, me hyl es e . Mike Ge a dsa,b, Giuseppe Canninoa,†, Jose M. González de Cóza a, and Howa d T. Jacobsa,c,* aFacul y o Medicine and Li e Sciences and Tampe e Uni e si y Hospi al, FI-33014 Uni e si y o Tampe e, Finland; bMaas ich Cen e o Sys ems Biology (MaCSBio), Maas ich Uni e si y, 6229 ER Maas ich , The Ne he lands; cIns i u e o Bio echnology, FI-00014 Uni e si y o Helsinki, Finland 810 | M. Ge a ds e al. Molecula Biology o he Cell In a genomewide RNA in e e ence (RNAi) sc een o ac o s in- ol ed in m DNA eplica ion o copy-numbe main enance, based on he loss o mi ochond ial nucleoid signal in D osophila S2 cells (Fukuoh e al., 2014), we p e iously iden i ied se e al p o eins p o- posed o be in ol ed in esicle a icking. O hologues o wo o hese, Be 1 and Slh (homologue o yeas Sly1), a e conse ed com- ponen s o he syn axin 5-SNARE complex (Newman e al., 1990; Dasche e al., 1991; Nichols and Pelham, 1998; Xu e al., 2000; Zhang and Hong, 2001; Yamaguchi e al., 2002), which is equi ed o esicle a icking om he ER o he Golgi in he canonical sec e- o y pa hway, bu has no p e iously been implica ed in mi ochon- d ial unc ions. A hi d p o ein iden i ied in he sc een, CG10144, is a D osophila homologue o Vps8, a yeas p o ein in ol ed in esicle so ing o lysosomes (Chen and S e ens, 1996). The p e ious s udy (Fukuoh e al., 2014) ound no e idence o any di ec ole o hese p o eins in m DNA main enance. We he e o e emba ked on a wide s udy o elucida e hei oles in mi ochond ia. Ou indings implica e esicle so ing a he cis-Golgi as c ucial o mi ochond ial quali y con ol, and ul ima ely also o cell-cycle p og ession and a oidance o cellula senescence. RESULTS Knockdown o Slh o Be 1 impai s mi ochond ial unc ion A e e i ying he e icacy o RNAi-based knockdown o Slh, Be 1, and CG10144 in D osophila S2 cells (Figu e 1A), we e es ed he e ec s on m DNA. E en unde condi ions o p olonged knockdown (10 d), he e was no signi ican change in m DNA copy numbe , as measu ed by quan i a i e PCR (Figu e 1B). Because his me hod does no e eal sub le changes in m DNA in eg i y o opology, we implemen ed Sou he n blo s o bo h undiges ed and Ps 1-linea ized m DNA om con ol and knockdown cells (Figu e 1C). Knockdown o he h ee genes p oduced no change in he ep esen a ion o di - e en opological o ms o (undiges ed) m DNA (Figu e 1C, le mos ou lanes) and no appea ance o subgenomic agmen s ollowing linea iza ion, which would be in- dica i e o dele ions, nicking, o eplica ion de ec s (Figu e 1C, igh mos ou lanes). To in es iga e whe he hese p o eins a e in- ol ed in mo e gene al mi ochond ial unc- ions, we analyzed se e al key pa ame e s du ing knockdown, including mi ochond ial espi a ion (measu ed by oxygen consump- ion), memb ane po en ial (based on TMRM luo escence), cellula con en o mi ochon- d ia (Mi o acke G een luo escence), and exp ession o genes implica ed in he mi o- chond ial un olded p o ein esponse, m UPR (by quan i a i e e e se- ansc ip ion PCR [qRTPCR]). O e 5 d, knockdown o Slh e- sul ed in dec eased mi ochond ial espi a- ion (77% o con ol alues), while knock- down o Be 1 o CG10144 had no such e ec (Figu e 2A). P olonged Slh knockdown o 10 d p oduced a u he dec ease in es- pi a ion (55% o con ol alues), whe eas he smalle d op in espi a ion due o Be 1 o CG10144 knockdown s ill did no each s a- is ical signi icance. The dec eased mi o- chond ial espi a ion due o Slh knockdown was accompanied by inc eased TMRM luo escence (Figu e 2B) bu dec eased ei he e old o deg ade hese mis olded o un olded p o eins. De- ec s in his p ocess lead o he appea ance o p o ein agg ega es and ha e been associa ed wi h spas ic pa aplegia, a axia, and neu- odegene a i e diseases such as Alzheime ’s and Pa kinson’s dis- eases (A nould e al., 2015). Mi ochond ial usion and ission occu con inuously, se ing o isola e damaged compa men s and a ge hem o u no e . When mi ochond ia a e subjec o p olonged o ex eme s esses, he mi- ochond ial ne wo k can become highly agmen ed, leading o clea ance o damaged compa men s by Pa kin-media ed u no e pa hways, a e which he ne wo k can be es o ed (Ni e al., 2015). Mi ophagy has also been p oposed o emo e damaged mi ochon- d ial DNA (m DNA) molecules, p e en ing he p oduc ion o non- unc ional OXPHOS subuni s (Ca elli e al., 2015). Addi ionally, he mi ochond ial ne wo k can be hype used upon ansien s ess o p o ec he o ganelles om unnecessa y u no e o possibly o minimize o ganella damage by di iding he s ess among he pop- ula ion. De ec i e usion o ission has been associa ed wi h domi- nan op ic a ophy and Cha co –Ma ie–Too h disease (Züchne e al., 2004; Ama i-Bonneau e al., 2008; Yu-Wai-Man e al., 2009; Ryu e al., 2010; Nguyen e al., 2017). Recen ly, a new mechanism o clea ance o damaged mi ochon- d ial compa men s by MDVs was desc ibed (Soubannie e al., 2012). The e appea o be mul iple sub ypes o MDVs in ol ed in di e en pa hways o in e o ganella communica ion and quali y main enance (Sugiu a e al., 2014). In one o hese, MDVs con ain- ing lipids, speci ic p o eins, and p o ein complexes a e eleased om mi ochond ia and deli e ed o mul i esicula bodies o a - ge ed deg ada ion. The exac con ibu ion o MDVs o mi ochon- d ial quali y con ol is s ill unclea , as is he mechanism whe eby damaged componen s a e selec ed o and a icked in o his u n- o e pa hway. FIGURE 1: Slh, Be 1, o CG10144 knockdown does no a ec m DNA copy numbe , opology, o in eg i y. (A) Rela i e RNA le els o he indica ed genes, no malized agains RpL32 RNA and hen eno malized agains he alues o un ea ed S2 cells, and (B) ela i e m DNA copy numbe (based on qPCR), also no malized agains he alues o un ea ed S2 cells. Means ± SD, signi ican di e ences om con ol cells (S uden ’s es wi h Bon e oni co ec ion, n ≥ 4 biological eplica es) deno ed by * (p < 0.01) and # (p < 0.05). (C) Sou he n blo o m DNA, ± diges ion wi h Ps I, om un ea ed S2 cells o cells ea ed wi h dsRNAs agains he indica ed genes, o he numbe o days shown. The main opological o ms o m DNA a e indica ed as OC (open ci cles), L (linea s), and SC (supe coiled ci cles). Volume 29 Ap il 1, 2018 Vesicle a icking and mi ochond ia | 811 ing epi ope agging. The V5 epi ope ag did no dis u b he a - ge ing o p e iously cha ac e ized mi ochond ial o cy osolic p o- eins (Supplemen al Figu e S1A). Howe e , immunocy ochemis y did no con incingly de ec colocaliza ion o V5- agged Be 1 o Slh wi h mi ochond ia (Figu e 3, A and B, panels a–c and j–l; Table 1). Ins ead, bo h p o eins we e ound o colocalize wi h he cis- Golgi ma ke GM130 (Figu e 3, A and B, panels d– and m–o; Table 1), as well as wi h app oxima ely hal o all esicles bea ing he ea ly endosome ma ke Rab5 (Figu e 3, A and B, panels g–i and p– ; Table 1). To in es iga e his issue u he , we de eloped a subcellula ac- iona ion p ocedu e o S2 cell lysa es ha yields highly pu i ied cy oplasmic (i.e., pos mi ochond ial) and mi ochond ial ac ions. Using Wes e n blo ing, we hen s udied he associa ion o Be 1-V5 and Slh-V5 in ac ions om ansien ly ans ec ed cells. The cis-Golgi ma ke GM130 was p esen in he pos mi ochond ial (cy oplasmic) ac ion, bu was unde ec able in he highly pu i ied mi ochond ial ac ion (Figu e 3C), whe eas Cox4 was abundan in he la e and only ain ly isible in he cy oplasmic ac ion when o e exposed (Figu e 3C). Howe e , p obing wi h he V5 an ibody e ealed ha some Be 1-V5 ou inely co ac iona ed wi h highly pu- i ied mi ochond ia, as did a iny amoun o Slh-V5 (Figu e 3C). No e ha o e exp ession o V5- agged Be 1 o Slh p oduced no change in he in acellula dis ibu ion o mo phology o mi ochond ia (Sup- plemen al Figu e S1B), and labeling pa e ns we e homogeneous ac oss many cells, a guing ha he obse ed colocaliza ions we e no due o an o e exp ession a i ac . Mi o acke G een luo escence pe cell (Figu e 2C). Knockdown o Be 1 o CG10144 o 5 d had only mino e ec s on hese pa ame e s (Figu e 2, B and C). Howe e , p olonged knockdown o Be 1, up o 10 d, also esul ed in signi ican ly inc eased TMRM luo escence pe cell (Figu e 2D, Supplemen al Table S1), while p olonged knock- down o Slh showed an e en ual es o a ion o he le el o Mi o- acke G een luo escence o con ol alues (Figu e 2E). The impo - ance o bo h Slh and Be 1 in mi ochond ial unc ions was u he suppo ed by a ma ked inc ease in he exp ession o Hsp22 and a i- ous mi ochond ial p o eases (Figu e 2F), which a e conside ed o be indica o s o mi ochond ial p o eo oxic s ess (Chan and McQuib- ban, 2013; Haynes e al., 2013; Mo ow e al., 2016) Thei exp ession was inc eased u he by combined knockdown o Slh and Be 1 (Figu e 2F), al hough e ec s on memb ane po en ial and mi ochon- d ial con en we e simila o hose p oduced by knockdown o Slh alone (Figu e 2, D and E). Finally, knockdown o ei he Slh o Be 1 p oduced a ma ked inc ease in he exp ession o he o he (Figu e 1A), while CG10144 exp ession was almos unchanged upon knock- down o Slh o Be 1 o ice e sa (Figu e 1A). These esul s indica e an impo an ole o he SNARE-binding p o ein Slh and he SNARE p o ein Be 1 in main aining mi ochond ial unc ions and sugges ha hey may ac in a common pa hway, while CG10144 likely has a di e en ole. Be 1 and Slh localize o he Golgi To gain mo e insigh in o how Be 1 and Slh a ec mi ochond ia, we s udied hei subcellula localiza ion in D osophila S2 cells, us- FIGURE 2: Slh knockdown impai s mi ochond ial unc ions. (A) Mi ochond ial O2 consump ion, no malized agains cell numbe , (B) TMRM luo escence pe cell, (C) Mi o acke G een luo escence pe cell, (D, E) ime cou se o changes in (D) TMRM luo escence and (E) Mi o acke G een luo escence pe cell, and (F) ela i e RNA le els o he indica ed genes, no malized agains RpL32 RNA; all eno malized agains he alues o un ea ed S2 cells, in cells ea ed wi h dsRNAs agains he indica ed genes o he numbe o days shown. (A–C, F) Means ± SD, signi ican di e ences om con ol cells (S uden ’s es wi h Bon e oni co ec ion, n ≥ 4 biological eplica es) deno ed by * (p < 0.01) and # (p < 0.05). (D, E) Means o ime cou ses ( o sou ce da a, including SD and s a is ical analysis, see Supplemen al Table S1); no e ha y axes do no commence a 0. 812 | M. Ge a ds e al. Molecula Biology o he Cell FIGURE 3: Be 1 and Slh localize o mul iple cellula memb anes, based on epi ope agging. Immunocy ochemis y o he indica ed ma ke s, Cox4 (mi ochond ia), GM130 (cis-Golgi), Rab5 (ea ly endosomes), in cells ansien ly ans ec ed wi h (A) Be 1-V5 o (B) Slh-V5, as shown. Images we e op imized on he mic oscope o b igh ness and con as , bu ha e no been manipula ed in any o he way, apa om he addi ion o panel labels as e e ed o in he ex . (C) Wes e n blo s o p o ein ex ac s om Be 1-V5 and Slh-V5 ans ec ed cells, p obed as indica ed. T— o al cellula ex ac , 20 μg, C, cy oplasmic (pos mi ochond ial) ac ion, 20 μg; M, highly pu i ied mi ochond ial ac ion, shown a inc easing dilu ion (20, 10, 4 μg). Ex apola ed molecula weigh s (kDa) o he de ec ed bands, in e ed om he mig a ion o molecula weigh ma ke s, shown alongside. No e ha D osophila Cox4, like ha o mammals, uns on SDS–12% PAGE gels a ∼15 kDa; i s N- e minal p ocessing is unknown. Slh-V5 is ∼74 kDa bu appea s o mig a e close o 80 kDa on SDS–PAGE. Blo images we e op imized o b igh ness and con as , c opped, and esized o cla i y, bu no manipula ed in any o he way. T ans ec ed cons uc Be 1-V5 Slh-V5 (i) Cells cos ained o V5 and Rab5 Numbe o indi idual cells analyzed 24 20 Numbe o V5-posi i e s uc u es pe cell 18 ± 5 22 ± 9 Numbe o Rab5-posi i e s uc u es pe cell 34 ± 11 29 ± 8 % o V5-posi i e s uc u es also posi i e o Rab5 96 ± 11 83 ± 16 % o Rab5-posi i e s uc u es also posi i e o V5 52 ± 10* 60 ± 14* (ii) Cells cos ained o V5 and GM130 Numbe o indi idual cells analyzed 24 16 Numbe o GM130-posi i e s uc u es pe cell 25 ± 5 28 ± 4 % o V5-posi i e s uc u es also posi i e o GM130 100 100 % o GM130-posi i e s uc u es also posi i e o V5 100 100 (iii) Cells cos ained o V5 and Cox4 Numbe o indi idual cells analyzed 23 21 Numbe o V5-posi i e s uc u es pe cell 26 ± 6 26 ± 8 % o V5-posi i e s uc u es also posi i e o Cox4 0 0 *No signi ican ly di e en , S uden ’s es wi h Bon e oni co ec ion, p > 0.05. TABLE 1: Subcellula localiza ion o Be 1 and Slh based on epi ope agging and ma ke colocaliza ion. Volume 29 Ap il 1, 2018 Vesicle a icking and mi ochond ia | 813 Figu e S2). These esul s suppo he idea ha Be 1 and Slh oge he play a ole in he in e ac ion o mi ochond ia and lysosomes and, aking accoun o loss o mi ochond ial unc ionali y when ei he is knocked down, suppo he idea ha his in luences he speci ici y o mi ochond ial quali y con ol o u no e . Slh is no equi ed o mi ophagy o endocy osis The obse a ion ha Slh is equi ed o he no mal in acellula local- iza ion o mi ochond ial ma e ial associa ed wi h lysosomes sug- ges s ha i may be equi ed o one o he p e iously s udied mi o- chond ial u no e pa hways. To es whe he Slh is needed o he well-cha ac e ized p ocess o mi ophagy, we ea ed con ol and Slh-knockdown cells wi h he uncouple FCCP (5 μM), which induces global mi ophagy, as e ealed by he comple e colocaliza ion o mi- ochond ia and lysosomes wi hin 2 h o such ea men . Slh knock- down made no di e ence o his p ocess: comple e colocaliza ion o Mi o acke G een and Lyso acke Red signal was obse ed, bo h in con ol and in Slh knockdown cells (Figu e 6A). Following he ecen iden i ica ion o an al e na e pa hway o mi ochond ial u no e in ol ing seques a ion in o endosomes (Hamme ling e al., 2017), we also es ed whe he Be 1 and Slh a e needed o endocy osis, using an assay based on up ake o he non- memb ane-pe mean pH-sensi i e dye pH odo Red Dex an. Knockdown o Slh o Be 1 esul ed in no sys ema ic dis up ion o he endocy ic pa hway, based on his assay (Figu e 6B). Knockdown o Be 1 and Slh p oduces g ow h a es and inc eased cell size A e 5 d o knockdown, ewe cells we e p esen in wells ans ec ed wi h double-s anded RNA (dsRNA) agains Be 1 o Slh, compa ed wi h con ol wells (Figu e 7A). To s udy whe he knockdown o Be 1 and Slh leads o a dec eased g ow h a e o o inc eased cell dea h, he o al numbe o cells was moni o ed in each well o e 10 d o cul u e. While a clea inc ease in cell numbe was e iden o con ol S2 cells and o cells knocked down using an ine dsRNA a ge ed Taken oge he , hese indings indica e ha bo h p o eins a e loca ed in close p oximi y o he Golgi, and in s uc u es ha also con ain endocy ic and mi ochond ial componen s, al hough no he cis-Golgi ma ke GM130. Because Be 1-V5 and Slh-V5 we e bo h ound o colocalize mos ly wi h GM130, we in es iga ed he e ec o hei knockdown on Golgi s uc u e. In bo h cases, he e was a signi ican inc ease in Golgi size (Figu e 4, A and B), while he num- be o Golgi s uc u es was unchanged (Figu e 4, A and C). The inding is consis en wi h bo h p o eins playing a ole in Golgi unc- ion, hough hei knockdown does no lead o comple e loss o Golgi. Slh is equi ed o he colocaliza ion o mi ochond ia, lysosomes, and Be 1 The iden i y and subcellula dis ibu ion o Be 1 and Slh (Figu e 3) sugges in ol emen in esicle so ing, whe eas he e ec s o hei knockdown (Figu e 2) indica e, in addi ion, ha hey a e needed o p e en mi ochond ial p o eo oxic s ess, ha is, he accumula ion o damaged mi ochond ial componen s. Taken oge he , hese ob- se a ions sugges ha Be 1/Slh-associa ed esicles could be in- ol ed in he u no e o mi ochond ia-de i ed ma e ial, and spe- ci ically in i s selec i e a ge ing o lysosomes o deg ada ion. To es his idea, we used li e-cell imaging o cells exp essing a Be 1- BFP (blue luo escen p o ein) usion p o ein, combined wi h s ain- ing o mi ochond ia and lysosomes espec i ely wi h Mi o acke G een and Lyso acke Red (Figu e 5), In con ol cells, a majo i y o he s uc u es posi i e o bo h Mi o acke G een and Lyso acke Red we e also posi i e o Be 1-BFP (Figu e 5, A, panels a–d, and B). In con as , knockdown o Slh almos comple ely abolished his co- localiza ion (Figu e 5, A, panels e–h, and B). Colocaliza ion o Mi o- acke G een and Lyso acke Red was s ill obse ed o an ex en simila o ha in con ol cells, bu hese s uc u es we e no longe posi i e o Be 1-BFP, which was seen in sepa a e oci. The pa e n o colocaliza ion o Be 1-BFP wi h Lyso acke Red and Mi o acke G een was no dis u bed upon CG10144 knockdown (Supplemen al FIGURE 4: Be 1 and Slh knockdown esul in Golgi enla gemen . (A) Immunocy ochemis y o GM130 (cis-Golgi ma ke ) o cells knocked down by dsRNA ea men o 5 d, a ge ed on he indica ed genes, wi h zoomed image showing ypical cells. (B) Su ace a ea, calcula ed using ImageJ so wa e, and (C) numbe o Golgi esicles pe cell, based on such images (≥60 cells in each case). Means ± SD; #, * indica e s a is ically signi ican di e ences om con ol (con) cells (S uden ’s es wi h Bon e oni co ec ion, p < 0.05 and p < 0.01, espec i ely). Images we e op imized on he mic oscope o b igh ness and con as , bu ha e no been manipula ed in any o he way, apa om he addi ion o panel labels as e e ed o in he ex . No e ha an ∼60% inc ease in su ace a ea ep esen s a wo old change in olume. 814 | M. Ge a ds e al. Molecula Biology o he Cell DISCUSSION Be 1 and Slh, p o eins al eady known o unc ion in he sec e o y pa hway, we e p e- iously iden i ied in a genomewide RNAi sc een o no el m DNA eplica ion ac o s (Fukuoh e al., 2014). He e we showed ha hey a e no di ec ly in ol ed in m DNA me- abolism, bu localized a he Golgi, whe e hey appea o unc ion unexpec edly in mi- ochond ial quali y con ol. The Golgi is conside ed as he cen al clea inghouse o he cell, allowing c oss- alk be ween he se- c e o y, endosomal, and au ophagic ma- chine y. He e we ex end his concep o mi- ochond ia, p o iding e idence ha esicle a icking a he Golgi in luences mi ochon- d ia and may ac as a su eillance sys em o de ec i e mi ochond ial componen s. Slh and Be 1 coope a e o main ain mi ochond ial quali y Be 1 and Slh, D osophila homologues o p o eins o he syn axin 5-SNARE complex, we e ound o colocalize in S2 cells wi h Golgi ma ke GM130, as expec ed om p e ious s udies showing ha hey a e e- qui ed o p o ein sec e ion (Ba d e al., 2006), consis en wi h hei canonical ole in o he o ganisms. Unexpec edly, RNAi- induced Slh de iciency led o p og essi ely dec eased mi ochond ial con en , accom- panied by dec eased mi ochond ial espi- a ion and inc eased pe -cell TMRM luo- escence, usually conside ed indica i e o mi ochond ial memb ane po en ial. We no e ha , despi e he inc ease in cell size (Figu e 7C), when he inc eased TMRM lu- o escence (Figu e 2, B and D) is no malized agains mi ochond ial con en (Figu es 2C and 7E), i emains con- sis en ly abo e con ol le els. In addi ion, Slh knockdown caused inc eased exp ession o genes media ing mi ochond ial p o eo oxic s ess, no ably he majo p o e- ases (m-AAA, i-AAA, and ho-7) and he s ess-induced chape one Hsp22, implica ing Slh in he main enance o mi ochond ial p o ein quali y. Be 1 knockdown p oduced simila , hough less d as ic e ec s, pa ially addi i e wi h hose o Slh knockdown. Dec eased espi a- ion, associa ed wi h he ac i a ion o ATP-dependen in ami ochon- d ial p o eases and chape ones, may engende mi ochond ial ATP deple ion, igge ing elec ogenic ATP impo o main ain mi ochon- d ial unc ions, and accoun ing o ele a ed memb ane po en ial. Mi ochond ial damage accumula ion accompanied by de- c eased mi ochond ial con en implies a ailu e o mi ochond ial biogenesis, dys egula ed mi ochond ial u no e , o bo h. The h ee-way associa ion o Be 1-BFP wi h lysosomal and mi ochon- d ial ma ke s, which was dis up ed by Slh knockdown, s ongly implies u no e as a key a ge , while lea ing open he p ecise mechanism. One possibili y is ha Be 1 and Slh a e in some way in ol ed in so ing and dis inguishing be ween unc ional and dys unc ional mi ochond ial componen s, o example, ia he ec- ogni ion o damage ma ke s (McLelland e al., 2014). Slh knock- down was no accompanied by a la ge inc ease in colocaliza ion o Lyso acke and Mi o acke signals, implying ha he accumula ion agains GFP, cell numbe did no change subs an ially a e Be 1 o Slh was knocked down (Figu e 7A). FACS analysis o p opidium io- dide–s ained cells showed ha signi ican ly mo e cells we e in G2 phase a e 10 d o Slh knockdown han o con ol cells, indica ing a dis u bed cell cycle (Figu e 7B). The accumula ion o cells in G2 was accompanied by an inc ease in cell size (Figu e 7, C and D). No e ha an inc ease in cell diame e o 45% equa es o a ebling o cell olume. Knockdown o Be 1 again p oduced simila bu quan i a i ely less d ama ic e ec s, while knockdown o CG10144 had no e ec on cell g ow h, size, o cell-cycle s a us (Figu e 7, A–C). These esul s indica e ha de iciency o Be 1 o Slh (bu no CG10144) leads o cell g ow h a es , p edominan ly in he G2 phase o he cell cycle, wi h concomi an ly inc eased cell size. Because Slh and Be 1 ha e p e iously been in e ed o unc ion in p o ein sec e ion (Ba d e al., 2006), we es ed whe he RNAi agains o he genes known o be in ol ed in esicle a icking in he sec e o y pa hway would p oduce he same e ec s on mi ochond ial p o eos asis and cell g ow h and size as RNAi agains Be 1 and Slh. P olonged knockdown o βCOP, which media es esicle anspo om he ER o he cis-Golgi, had an e ec simila o ha seen wi h Slh knockdown (Figu e 8, A–D; compa e wi h Figu es 7A, 2F, and 7, C and D, espec i ely), bu knockdown o SNAP-24 and Syx6, bo h o which unc ion in pos -Golgi esicle a icking (Niemeye and Schwa z, 2000; Malsam and Söllne , 2011) did no (Figu e 8, A–D). FIGURE 5: Slh is equi ed o Be 1 coassocia ion wi h lysosomal and mi ochond ial componen s. (A) Rep esen a i e li e-cell images o cells ans ec ed wi h Be 1-BFP, ei he alone (con ol S2 cells, con, panels a–d) o wi h concomi an knockdown o Slh (panels e–h), cos ained wi h Lyso acke Red and Mi o acke G een, oge he wi h (inse ) zoomed images a highe magni ica ion. Whi e a ows indica e esicles posi i e o Mi o acke G een and Lyso acke Red, which, in con ol cells, bu no Slh knockdown cells, we e mos ly posi i e o Be 1-GFP. Images we e op imized on he mic oscope o b igh ness and con as , bu ha e no been manipula ed in any o he way, apa om he addi ion o panel labels and a ows as e e ed o he e and in he ex . (B) Quan i a ion (%) o s uc u es cos ained wi h Mi o acke G een and Lyso acke Red, which we e also posi i e o Be 1-BFP, based on inspec ion o all isibly luo escen s uc u es in eigh cells o each class analyzed. con, con ol S2 cells ans ec ed wi h Be 1-BFP; Slh KD, cells also knocked down o Slh. * deno es a s a is ically signi ican di e ence, p < 0.001, S uden ’s es . Volume 29 Ap il 1, 2018 Vesicle a icking and mi ochond ia | 815 Al hough Slh and Be 1 a e pa o he sec e o y machine y in D osophila (Ba d e al., 2006), as in o he o ganisms, ou inding ha hey also colocalize a he Golgi wi h a p opo ion o Rab5-posi i e esicles sugges ed ha hey migh also play a ole in he endocy ic pa hway. This aised he possibili y ha hei e ec on mi ochond ia could be indi ec , a ising, o example, om de ec i e cellula ans- po o i on. Mo eo e , a sepa a e s udy has ecen ly implica ed en- dosomes in an al e na i e pa hway o mi ochond ial u no e (Ham- me ling e al., 2017). MDVs ha e been shown o associa e wi h he e ome -complex componen s Vps35and Vps26 (B aschi e al., 2010), which media e esicle anspo om endosomes o he Golgi. Mo eo e , Pa kinson’s disease–associa ed mu an s o VPS35 a e known o p oduce mi ochond ial dys unc ion (Wang e al., 2016), and Vps35 in e ac s gene ically wi h pa kin in D osophila (Malik e al., 2015). Howe e , ou indings (Figu e 6B) demons a e ha Slh and Be 1 a e no equi ed o endocy osis in S2 cells. Colo- caliza ion wi h Rab5 is he e o e simply a e lec ion o he ole o he Golgi as he cell’s cen al clea inghouse o esicles in ansi . Why we e esicle- anspo genes picked up in a sc een o m DNA main enance ac o s? Exac ly why Be 1 and Slh we e iden i ied as posi i es in he o iginal genomewide sc een emains unclea . Thei knockdown did no a - ec m DNA con en o in eg i y (Figu e 1). The mechanism by which Pico G een accumula es in mi ochond ia is no known, bu likely e- qui es a mi ochond ial memb ane po en ial. Al hough memb ane po en ial was inc eased when Slh o Be 1 was knocked down, i s hypo hesized main enance by elec ogenic ATP impo , accompa- nied by dec eased espi a ion, implies ha in ami ochond ial pH o mi ochond ial damage did no induce an inc eased amoun o u no e , bu a he a change in i s selec i i y. The colocaliza ion o Be 1 wi h a subse o mi ochond ia, based bo h on subcellula ac- iona ion (Figu e 3D) and on li e-cell imaging (Figu e 5), plus i s dis- up ion by Slh knockdown (Figu e 5), sugges s ha Be 1 may se e as a ag o mi ochond ia o MDVs des ined o u no e . We hy- po hesize ha Slh may be needed o he docking o Be 1- agged mi ochond ial esicles a he Golgi and hei onwa d so ing o lyso- somes, and ha in he absence o his p ocess, mi ochond ial in e - ac ion wi h lysosomes is la gely nonselec i e. The syn axin-5 SNARE complex and mi ochond ial u no e pa hways In mammalian cells, he syn axin-5 SNARE complex plays a ole in au ophagy as well as in p o ein sec e ion (Renna e al., 2011). This has been a ibu ed o he need o an e og ade anspo o ma u- a ion o lysosomal p o eases in he Golgi. In yeas , SNARE and SNARE-binding mu an s a ec ing ER- o-Golgi a icking, including Be 1 and he Slh homologue Sly1, a e de icien in mac oau opaghy (Tan e al., 2013), and COPII esicles ha e been sugges ed o sup- ply memb anes o au ophagosome o ma ion (Tan e al., 2013). Howe e , ou da a (Figu e 6A) indica e clea ly ha Be 1 and Slh a e no equi ed in D osophila S2 cells o global mi ophagy, induced by agen s ha comp omise mi ochond ial memb ane po en ial (Na end a e al., 2008). The mi ochond ia- ela ed p ocess dis up ed by Be 1 o Slh knockdown in S2 cells mus he e o e be mo e spe- ci ic. The o he molecula playe s in ol ed, and hei ela ionship wi h he a ious mi ochond ial u no e pa hways so a desc ibed, emain o be elucida ed. FIGURE 6: Slh is no equi ed o mi ophagy o endocy osis. (A) Li e-cell imaging o con ol S2 cells and cells knocked down o Slh, ea ed o 2 h wi h 5 μM FCCP and s ained wi h Mi o acke G een and Lyso acke Red. The wo p obes o e lap almos comple ely in bo h cases, indica ing ha Slh knockdown does no in e e e wi h global mi ophagy. (B) Visualiza ion o acidi ied la e endosomes and endolysosomes in un ea ed S2 cells and in cells knocked down o Be 1 o Slh o ea ed wi h an ine dsRNA a ge ed on GFP. Red luo escence due o he pH-sensi i e nonpe mean dex an-conjuga ed dye pH odo Red Dex an is isible in all cells. 816 | M. Ge a ds e al. Molecula Biology o he Cell dis up ed o cell-cycle p og ession by RNAi agains Rho-1 GTPase (Roge s and Roge s, 2008). Links be ween mi ochond ial dys unc- ion and senescence ha e been epo ed p e iously (Passos e al., 2006; Ziegle e al., 2015), and a ecen s udy showed mi ochond ial dys unc ion o be causal (Wiley e al., 2016). In iguingly, his was also associa ed wi h a cha ac e is ic de ec in he sec e o y pa hway. Howe e , we canno exclude he possibili y ha he senescence-like pheno ype obse ed he e has no hing di ec ly o do wi h mi ochon- d ia. The inding should he e o e be ea ed wi h cau ion. Ne e - heless, i is clea ly an impo an consequence o Be 1 o Slh knockdown. The senescence-associa ed sec e o y pheno ype (SASP) is a common hallma k o senescence ha , a leas in mammalian cells, in ol es he sec e ion o p oin lamma o y cy okines and o he sig- naling molecules, as well as p o eases and o he ac o s ha modi y he ex acellula ma ix (Coppé e al., 2010). The unde lying mecha- nisms a e poo ly unde s ood, bu a e belie ed o in ol e ansc ip- ional changes dic a ed by ch oma in al e a ions. He e, we induced a senescencelike pheno ype by al e ing he p ope ies o key com- ponen s o he sec e o y pa hway, a he han p oducing sec e ome changes h ough senescence induc ion. This suppo s he idea o senescence as a p og am a he han a haphaza d se ies o mal- adap i e changes. Fu he mo e, i s eng hens he iew ha esicle so ing a he cis-Golgi is a c ucial p ocess o main aining cellula homeos asis and iabili y, linking bo h p o ein sec e ion and may be highly abno mal. This could unde lie he loss o Pico G een signal. Al e na i ely, i may e lec an al e ed s uc u e o nucleoids caused by p o eo oxic s ess. The ole o Be 1 and Slh in mi ochond ial quali y con ol appea s o be sha ed wi h a leas one o he p o ein in ol ed in esicle a - icking a he cis-Golgi, βCOP (Figu e 8). In con as , p o eins ha pe o m o he oles in he ans-Golgi ne wo k, including o he SNAREs, such as SNAP-24 and Syx-6, do no appea o be in ol ed in mi ochond ial quali y main enance (Figu e 8). CG1044 alls in o he same ca ego y, because i s o hologue in yeas , Vps8, is e- qui ed o he e en ion o se e al la e-Golgi memb ane p o eins in he Golgi appa a us, and o he co ec so ing o he acuola (lysosomal) ca boxypep idase Y (Chen and S e ens, 1996). I is also a membe o he CORVET complex, in ol ed in memb ane e he ing in la e endosomes and hei in e ac ion wi h lysosomes (Balde haa e al., 2013; Solinge and Spang, 2013). This ne e heless aises he simila ques ion o why CG10144 was ep oducibly sco ed as posi- i e in he o iginal sc een. Unlike ha o Be 1 and Slh, i s knockdown p oduced no e ec on any o he mi ochond ial pa ame e s es ed he e. A link be ween mi ochond ial u no e and senescence Slh, Be 1, o βCOP knockdown esul ed also in G2 p oli e a ion a - es and inc eased cell size. These a e hallma ks o cells unde going senescence (Rodie and Campisi, 2011) and a e also seen in S2 cells FIGURE 7: Slh and Be 1 knockdown en ains cell g ow h de ec s. (A) G ow h cu es o con ol S2 cells and cells ea ed con inuously o 10 d, wi h dsRNAs a ge ed agains he indica ed genes, commencing 5 d a e ans ec ion. Fo sou ce da a, including SD and s a is ical analysis, ollowing no maliza ion o cell numbe on day 5, see Supplemen al Table S1. (B) Cell-cycle analysis o con ol S2 cells and cells knocked down o 10 d by dsRNAs a ge ed on he indica ed genes. (C) Mic og aphs o con ol S2 cells and cells a e 5 d o ea men wi h dsRNAs a ge ed agains he indica ed genes. Scale ba s: 20 μm. (D) Cell diame e o con ol S2 cells and cells knocked down o he indica ed genes (means ± SD o 52 cells o each class). #, * deno e signi ican di e ences om he co esponding da a class o con ol cells (S uden ’s es wi h Bon e oni co ec ion, p < 0.05 o 0.001, espec i ely). Volume 29 Ap il 1, 2018 Vesicle a icking and mi ochond ia | 817 ans ec ion, exp ession o V5- agged Be 1 and Slh was induced by adding CuSO4 o a inal concen a ion o 100 μM. Two days a e in- duc ion, cells we e ixed and s ained as p e iously desc ibed (Roge s and Roge s, 2008; including pla ing on ConA) using mouse an i-V5 (Li e Technologies) and abbi an i-COXIV, an i-Rab5, o an i-GM130 (Abcam) as p ima y an ibodies and AlexaFluo 568 goa an i-mouse Immunoglobulin G (IgG) (H+L) and goa an i- abbi AlexaFluo 488 IgG (H+L) (Li e Technologies) as seconda y an ibodies, and imaged by con ocal mic oscopy. Golgi size and numbe we e calcula ed us- ing a spo -a ea calcula ion wi h ImageJ. Fo li e-cell imaging o Be 1, he coding sequence o BFP was cloned in o he pMT-V5-Be 1 con- s uc be ween he V5 and he His ags. Cells we e ans ec ed wi h he inal cons uc using FuGene, and exp ession was induced 1 d a e ans ec ion by adding CuSO4 o a inal concen a ion o 100 μM. Two days la e , cells we e ans e ed o ConA-coa ed slides as desc ibed p e iously (Roge s and Roge s, 2008), s ained wi h 25 nM Mi oT acke G een FM and 50 nM LysoT acke Red DND-99 (The mo Fishe Scien i ic), and imaged by con ocal imaging. In acellula localiza ion by subcellula ac iona ion and Wes e n blo ing Cells we e pla ed, ans ec ed, and induced as o immunocy o- chemis y, bu on a la ge scale (s a ing om 5 × 107 cells in 100 ml mi ochond ial damage o cell a e, issue emodeling, and aging, and sugges ing a possible “mi ochond ial checkpoin ” o cell-cycle p og ession. MATERIALS AND METHODS S2 cell cul u e D osophila S2 cells (In i ogen) we e cul u ed unde s anda d condi- ions in Schneide ’s medium (Sigma). Knockdown was es ablished by ea men wi h gene-speci ic dsRNAs. B ie ly, dsRNA was gene - a ed om T7-labeled PCR p oduc s using he MEGAsc ip T7 an- sc ip ion ki (The mo Scien i ic). Fo 5-d knockdown expe imen s, cells we e seeded a a densi y o 106 cells/ml and ea ed wi h 8 μg/ ml dsRNA hal an hou a e pla ing. An addi ional 8 μg/ml dsRNA was added a e 3 d. Fo 10-d knockdown, 105 cells/ml we e seeded and ea ed wi h 2 μg/ml dsRNA hal an hou la e and wi h 8 μg/ml dsRNA a days 3, 5, and 8. Addi ionally, he medium was e eshed a e 5 d. Immunocy ochemis y and li e-cell imaging Be 1 and Slh we e cloned in ame wi h he V5- ag in he pMT-V5/His plasmid (Li e Technologies), which con ains an inducible me allo hio- nein p omo e . S2 cells we e ans ec ed using FuGene (P o- mega) acco ding o he manu ac u e ’s ins uc ions. One day a e FIGURE 8: Pos -Golgi componen s o he sec e o y pa hway a e no needed o main ain mi ochond ial quali y and cell g ow h. (A) Rela i e cell numbe , compa ed wi h con ol cells, a e 5 d o dsRNA ea men agains he indica ed genes. See also Supplemen al Figu e S3. (B) Rela i e le el o Hsp22 RNA, no malized agains RpL32 RNA and eno malized agains he alues o un ea ed S2 cells, a e dsRNA ea men agains he indica ed genes. Because he e we e so ew emaining iable cells a e 5 d o βCOP knockdown, RNA le els we e assessed a e 2 d o knockdown. Means ± SD; signi ican di e ences om con ol cells (S uden ’s es wi h Bon e oni co ec ion, n ≥ 4 biological eplica es) deno ed by * (p < 0.05). (C) Mic og aphs o S2 cells a e 5 d o ea men wi h dsRNAs a ge ed agains he indica ed genes. Scale ba s: 20 μm. (D) Cell diame e o con ol (S2) cells and cells knocked down o he indica ed genes (means ± SD o 52 cells o each class). * deno es signi ican di e ences om he co esponding da a class o con ol cells (S uden ’s es wi h Bon e oni co ec ion, p < 0.001).