Evolution of the metazoan mitochondrial replicase

E olu ion o he Me azoan Mi ochond ial Replicase Ma cos T. Oli ei a 1,2 , Jani Haukka 1 , and Lau ie S. Kaguni 1,3, * 1 Ins i u e o Biosciences and Medical Technology, Uni e si y o Tampe e, Finland 2 Depa amen o de Tecnologia, Faculdade de Cie ˆncias Ag a ´ ias e Ve e ina ´ ias, Uni e sidade Es adual Paulis a “Ju ´lio de Mesqui a Filho,” Jabo icabal, SP, B azil 3 Depa men o Biochemis y and Molecula Biology and Cen e o Mi ochond ial Science and Medicine, Michigan S a e Uni e si y *Co esponding au ho : E-mail: [email p o ec ed] Accep ed: Feb ua y 26, 2015 Abs ac The la ge numbe o comple e mi ochond ial DNA (m DNA) sequences a ailable o me azoan species makes i a good sys em o s udying genome di e si y, al hough li le is known abou he mechanisms ha p omo e and/o a e co ela ed wi h he e olu ion o his o ganella genome. By in es iga ing he molecula e olu iona y his o y o he ca aly ic and accesso y subuni s o he m DNA polyme ase, pol g, we sough o de elop mechanis ic insigh in o i s unc ion ha migh impac genome s uc u e by explo ing he ela ionships be ween DNA eplica ion and animal mi ochond ial genome di e si y. We iden i ied h ee e olu iona y pa e ns among me azoan pol gs. Fi s , a end owa d s abiliza ion o bo h sequence and s uc u e occu ed in e eb a es, wi h bo h subuni s e ol ing dis inc ly om hose o o he animal g oups, and acqui ing a leas ou no el s uc u al elemen s, he mos impo an o which is he HLH-3b(helix-loop-helix, 3 b-shee s) domain ha allows he accesso y subuni o homodime ize. Second, bo h subuni s o a h opods and unica es ha e becomesho e and e ol ed app oxima ely wice as apidly as hei e eb a e homologs. And hi d, nema odes ha e los he gene o he accesso y subuni , which was accompanied by he loss o i s in e ac ing domain in he ca aly ic subuni o pol g, and hey show he highes a e o molecula e olu ion among all animal axa. These indings co ela e well wi h he m DNA genomic ea u es o each g oup desc ibedabo e, and wi h hei modes o DNA eplica ion, al hough a subs an i e amoun o biochemical wo k is needed o d aw conclusi e links ega ding he la e . Desc ibing he pa allels be ween e olu ion o pol gand me azoan m DNA a chi ec u e may also help in unde s anding he p ocesses ha lead o mi ochond ial dys unc ion and o human disease- ela ed pheno ypes. Key wo ds: mi ochond ia, mi ochond ial DNA eplica ion, s uc u al e olu ion, mi ochond ial eplicase, pol g. In oduc ion Mi ochond ial DNA (m DNA) eplica ion is accomplished by he sole DNA polyme ase ound in animal mi ochond ia, DNA polyme ase g(pol g) ( e iewed in Kaguni 2004), which unc ions as pa o a la ge eplica ion machine y called he m DNA eplisome. The iden i y o all o he componen s o he m DNA eplisome is s ill unknown, bu biochemical s udies (Ko honen e al. 2003,2004;Oli ei a and Kaguni 2010, 2011) ha e shown ha a g oup o mi ochond ial p o eins can in e ac unc ionally o p omo e DNA syn hesis in i o, o ming he minimal m DNA eplisome: The m DNA helicase, also known as Twinkle in humans, unwinds he duplex DNA a he eplica ion o k; pol gca alyzes nascen DNA syn hesis on bo h he leading and lagging DNA s ands; and he mi ochond ial single-s anded DNA-binding p o ein (m SSB) coo dina es hei unc ions while binding and s abilizing he single-s anded DNA empla e (Ko honen e al. 2004;Oli ei a and Kaguni 2011). This scena io, as in mos DNA eplica ion e en s, equi es he p esence o sho RNA molecules o p im- ing o DNA syn hesis by he DNA polyme ase, which in animal mi ochond ia migh be achie ed by he ac ion o he mi o- chond ial RNA polyme ase (Wan ooij e al. 2008;Fus e e al. 2010;Reyes e al. 2013) and/o he newly iden i ied p imase P imPol (Ga cia-Gomez e al. 2013). Animal pol gis a he e o-oligome ic enzyme, in mos known cases: The ca aly ic co e, pol g-a(also known as POLG o PolGA), con ains bo h he 50–30DNA polyme ase and 30–50 exonuclease ac i i ies o he holoenzyme, whe eas he acces- so y subuni , pol g-b(o POLG2, PolGB), se es as a p ocessi - i y ac o , enhancing he in e ac ions be ween he holoenzyme and he DNA subs a e (Lewis e al. 1996; Wang e al. 1997;Ca odeguas e al. 1999;Lim e al. 1999). The ca aly ic co e is a membe o he amily A DNA polyme ase g oup, o which bac e ial DNA polyme ase I and he ca aly ic co e o bac e iophage T7 DNA polyme ase (gp5) also belong ( e iewed in Kaguni 2004); pol g-asha es GBE ßThe Au ho (s) 2015. Published by Ox o d Uni e si y P ess on behal o he Socie y o Molecula Biology and E olu ion. 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 p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed euse, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 943 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om subs an ial s uc u al and unc ional p ope ies wi h bo h Pol I and T7 Pol. No ably, Pol I is no he eplica i e DNA polyme ase in bac e ia. I is a single-subuni enzyme ha lacks he high ideli y and p ocessi i y o he holoenzyme o m o T7 Pol, which acqui es hese p ope ies as a esul o he associa ion o he ca aly ic co e wi h he bac e ial p o ein hio edoxin ha se es as i s accesso y subuni . Al hough he p esence o pol g as he mi ochond ial eplicase appea s o be conse ed among he me azoans (and o he euka yo es), impo an a ia ions in i s s uc u e ha e been desc ibed ha may impac he mode in which m DNA is eplica ed. I has long been known ha pol g o D osophila melanogas e , one o he bes s udied insec model o ganisms, is a he e odime comp ising one ca aly ic subuni and a single accesso y subuni (We ne e and Kaguni 1986;Olson e al. 1995;Wang and Kaguni 1999). On he o he hand, he human and mouse holoenzymes ha e a he - e o ime ic con o ma ion, consis ing o one pol g-aand a di- me ic pol g-b(Ca odeguas e al. 2001;Yakubo skaya e al. 2006;Lee e al. 2009). In e es ingly, he m DNA o he nem- a ode Caeno habdi is elegans appea s o be eplica ed by a machine y con aining only a single subuni pol g(only pol g-a) (B a ic e al. 2009;Addo e al. 2010), which esembles he ca aly ic co e in o he euka yo es, such as ha o he yeas Saccha omyces ce e isiae (Fou y 1989). The accesso y subuni o pol g, which indeed appea s o be p esen only in Me azoa, has a ema kable e olu iona y o igin, because amino acid sequence alignmen s, phylogene ic in e ences, and gene al p o ein s uc u e demons a e i s homology o class II aminoa- cyl- RNA syn he ases (Fan e al. 1999,2006;Ca odeguas e al. 2001;Wol and Koonin 2001). We sough o explo e he sequence and s uc u al di e si y o pol gin he animal kingdom, aking ad an age o he cu - en inc ease in nuclea genomes and ansc ip omes o which comple e sequences a e a ailable in public da abases. We e ie ed as many animal pol g-aand -bgene sequences as a e a ailable and pe o med in silico analyses o in e hei molecula e olu iona y his o y, aking in o accoun he sub- s an ial biochemical and s uc u al da a epo ed by ou g oup and o he s. He e we epo he oligome ic plas ici y o animal pol g, he inding o new s uc u al elemen s, and he dis inc a es o molecula e olu ion o di e en axa, which may e lec di e ences in he undamen al mechanisms o m DNA eplica ion. We discuss ou indings in he con ex o mi ochond ial genome di e si y, s uc u e, eplica ion, and e olu ion. Ma e ials and Me hods Sea ches o Animal pol g-aand -bHomologs and Mul iple Sequence Alignmen s TBLASTN sea ches (Al schul e al. 1990) in he NCBI (Na ional Cen e o Bio echnology In o ma ion) non edundan se- quence da abase we e pe o med using he ansla ed mRNA e e ence sequences om Homo sapiens (pol g-a, NM_001126131.1; pol g-b, NM_007215.3) and D. melano- gas e (pol g-a, NM_057473.3; pol g-b, FJ635829.1) as que ies. To e ie e sequences om Po i e a, Placozoa, Cnida ia, Mollusca, and Hemicho da a species, complemen- a y HMMR3 BLAST (Basic Local Alignmen Sea ch Tool) sea ches (Eddy 2011) we e pe o med (h p:// oolki .lmb.uni- muenchen.de/hmme 3, las accessed Augus 2014), ollowed by BLAST sea ches agains he Ensembl Me azoa da abase (h p://me azoa.ensembl.o g, las accessed Augus 2014), al- lowing he inclusion o missing exons. Mos sequences om Nema oda species we e e ie ed om genomic sca olds wi h no gene models deposi ed in he 959 Nema ode Genomes da abank (h p://www.nema odes.o g/nema odegenomes, las accessed Augus 2014), ollowed by manual p ocessing o he exons. A Py hon sc ip u ilizing unc ions o he Biopy hon lib a y (Cock e al. 2009) was de eloped o ex ac he coding sequences and o he speci ic in o ma ion om he o iginal iles e ie ed om he di e se da abanks. The sc ip can be p o ided by he au ho s upon eques . Highly di e gen sequences we e es ed o ue o hology by using hem as que ies o BLAST sea ches agains he human genome da a- base and by p edic ing possible mi ochond ial localiza ion using he Ta ge P se ice (Emanuelsson e al. 2007). The com- ple e se o sequences e ie ed is shown in supplemen a y able S1,Supplemen a y Ma e ial online. Because o he o e ep esen a ion by mammalian and D osophila sequences, his da a se was educed be o e he alignmen s we e pe o med o diminish bias in in e p e a ion o he esul s. The mul iple amino acid sequence alignmen s we e pe o med wi h he so wa e MAFFT (Ka oh and Toh 2008), using he G-INSi and E-INSi algo i hms o he pol g-a and -bsequences, espec i ely, and a e shown in supplemen- a y igu es S1 and S2,Supplemen a y Ma e ial online. Phylogene ic In e ences To c ea e he inpu iles o he phylogene ic in e ences, he so wa e PAL2NAL (Suyama e al. 2006) was used o con e he amino acid sequence alignmen s in o codon-based nucle- o ide alignmen s, which we e hen con e ed o NEXUS o ma . The phylogene ic ees we e in e ed using he Bayesian algo i hm buil in he so wa e M Bayes, e sion 3.2.2 (Ronquis e al. 2012). The consensus ee, un wi h 200,000 cycles o pol g-aand 1 million cycles o pol g-b, was se o he 50-majo i y ule, gamma a ia ion was ex- pec ed among he si es and Gene alized Time Re e sible model was used; o he pa ame e s we e kep as de aul . Modeling o P o ein S uc u e The s uc u e o he diso de ed egions in he human pol g-a c ys allog aphy da a (PDB accession numbe 3IKM, chain A) was p edic ed wi h he so wa e I-TASSER (Bazzoli e al. 2011), using de aul pa ame e s. I-TASSER was also used Oli ei a e al. GBE 944 Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om wi h de aul pa ame e s o model he whole s uc u e o pol g-a om C. elegans (using he human pol gc ys al s uc u e 3IKM:A), and o pol g-b om S ongylocen o us pu pu a us, Ciona in es inalis, and T ichoplax adhae ens (using he human pol g-bc ys al s uc u e 2G4C:A). D osophila melanogas e pol g-bs uc u e was modeled wi h he so wa e MODELLER, e sion9.12(Sali and Blundell 1993), using he 2G4C:A ile as empla e and he mul iple sequence alignmen (MSA) shown in supplemen a y igu e S2,Supplemen a y Ma e ial online, as a pa ame e . The selec ed models we e e alua ed by he Z- sco es, he disc e e op imized p o ein ene gy alues, and he esidue e o plo (SwissP o websi e, h p://swissmodel. expasy.o g/wo kspace/index.php? unc= ools_s uc u e- assessmen 1, las accessed Augus 2014). S uc u es and models we e analyzed and igu es we e p oduced using Pymol (www.pymol.o g, las accessed Augus 2014). Resul s Ou sea ches o public genomic sequence da abases ound non edundan sequences o pol g-aand -bo 62 and 52 animal species, espec i ely (supplemen a y able S1, Supplemen a y Ma e ial online). The esul ing da a se is o e - ep esen ed by sequences om insec and e eb a e (espe- cially mammalian) species due o he bias in he da abases. Animal g oups in he P o os omia clade, o he han A h opoda and Nema oda, we e mos o en absen , excep o one pol g-asequence om Mollusca. None heless, we ob ained sequences om key species o basal g oups o Me azoa, such as Po i e a, Placozoa and Cnida ia, and basal and sis e g oups o Cho da a, such as Tunica a, Cephalocho da a, Hemicho da a and Echinode ma a, which a e c ucial o he analyses desc ibed below. We excluded se e al sequences o mammalian and D osophila species om ou pol g-aand -bda a se s o bal- ance he axa ep esen a ion, and also a ew o he BLAST esul s because hey con ained only pa ial gene sequences (see supplemen a y able S1,Supplemen a y Ma e ial online). Supplemen a y igu e S1,Supplemen a y Ma e ial online, shows he alignmen o 43 pol g-aaminoacidse- quences plus he ou g oup ( he sequence om S. ce e isiae pol g-a, accession numbe NM_001183750.1). The alignmen o 35 pol g-bamino acid sequences plus he ou g oup ( he glycyl- RNA syn he ase sequence om he bac e ium The mus he mophilus, accession numbe AJ222643.1) is shown in supplemen a y igu e S2,Supplemen a y Ma e ial online. The numbe o pol g-bsequences e ie ed was lowe han ha o pol g-amainly because o he absence o he pol g-b gene in he genome o nema ode species, as discussed below. In addi ion, we we e unable o ind comple e pol g-bgene sequences o he po i e an Amphimedon queenslandica, he mollusk C assos ea gigas, and he c us acean Daphnia pulex. Conside ing ha he pol g-asequences om hese species do ha e a po en ial o o m an accesso y-in e ac ing de e minan (AID) s uc u e (see below), ou ailu e o ind he co espond- ing pol g-bmigh be because o he cu en low co e age o hei genome/ ansc ip ome sequences. A summa y o ou mos in e es ing indings is p esen ed in igu e 1, along wi h schema ics o he ca aly ic and accesso y subuni polypep- ides. We iden i ied speci ic ea u es o all sequences e- ie ed o indica e ha ou da a se mos likely consis s o ue o hologs o he ca aly ic and accesso y subuni s o he mi ochond ial eplicase, and no andom genes coding o o he amily A DNA polyme ases o aminoacyl- RNA syn he- ases, espec i ely: 1) Conse ed ac i e si e mo i s in he exo- nuclease (Exo I–III) and polyme ase (Pol A–C) domains o pol g- a( e iewed in Kaguni 2004)(supplemen a y ig. S1, Supplemen a y Ma e ial online) ha a e sha ed among amily A DNA polyme ases; 2) conse ed pol g-a-speci ic se- quences in he space egion and polyme ase domains o pol g-a(Lewis e al. 1996;Lec enie e al. 1997;Kaguni 2004); 3) conse ed sequence o he AID subdomain o pol g-a(excep in nema odes), which is a pol g-a-speci ic ea u e o me a- zoans ( e iewed in Kaguni 2004) (see below); 4) conse ed hyd ophobic esidues in he C- e minal egion o pol g-b ha a e mos ele an o he in e ac ions wi h he pol g-aAID, and a e ound almos exclusi ely in pol g-bs(Fan e al. 1999; Ca odeguas e al. 2001;Kaguni 2004;Lee e al. 2009); and 5) absence in pol g-bo he esidues equi ed o dime iza ion o class II aminoacyl- RNA syn he ase (Logan e al. 1995;A nez e al. 1999). Dis inc Ra es o E olu iona y Changes in pol g-aand -b Sequences Phylogene ic in e ences wi h he pol g-anucleo ide sequences using Bayesian analysis ( ig. 2) ep oduced mode a ely well he cu en ly accep ed ela ionships among animal axa (Philippe e al. 2009), wi h ew excep ions, and p o ided impo an indings ega ding he e olu ion o he gene. Fi s , i is clea ha he sequences om nema odes ha e e ol ed a a highe a e han hose om insec species, which in u n ha e accu- mula ed mo e subs i u ions han he sequences om e e- b a es. Second, Tunica a, which is ep esen ed only by he sequence om Ci. in es inalis (Ascidiacea: Cionidae), was g ouped wi hin he pol g-asequences om A h opoda. Finally, he b anches o he pol g-asequences om Deu e os omia ( e eb a es and sis e g oups), excluding Ci. in es inalis, and om he mollusk C . gigas we e as sho as hose o he sequences om basal animal g oups, such as Po i e a (A. queenslandica), Placozoa (T. adhae ens), and Cnida ia (Nema os ella ec ensis). In combina ion, his may indica e ha mos deu e os ome pol g-as e ained mo e an- ces al cha ac e s, whe eas pol g-asequences om nema- odes, a h opods (especially insec s), and unica es ha e di e ged conside ably om he o iginal enzyme. Bayesian phylogene ic analysis using pol g-bnucleo ide se- quences ( ig. 3) also shows ha he insec p o eins ha e E olu ion o Animal pol gGBE Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 945 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om e ol ed a a much highe a e han e eb a es. The a e o subs i u ions obse ed o he sequences om mos Deu e os omia species again ma ches ha o he sequences om basal animal g oups, sugges ing ha pol g-balso e ained mo e ances al cha ac e s o hese g oups. Mo eo e , he sequence om he unica e Ci. in es inalis again g ouped wi h hose om A h opoda, indica ing a e- ma kable esemblance and hei signi ican di e gence om FIG.1.—Schema ics o animal ( e eb a e) pol g-aand -bsequence and s uc u e. (A) Rep esen a ion o he amino acid sequence o he pol g-aand -b polypep ides, showing he p o ein domains, subdomains, conse ed mo i s, and new mo i s p oposed in his wo k. Fo pol g-a: NTD, he N- e minal domain o which no unc ional da a a e a ailable; Exo, he exonuclease domain esponsible o he 30–50exonuclease ac i i y ha edi s misinco po a ed nucleo ides and inc eases he ideli y o DNA syn hesis se e al-hund ed old; AID, he accesso y-in e ac ing de e minan subdomain ha p o ides he p ima y con ac s be ween he ca aly ic co e and he accesso y subuni ; IP, he in insic p ocessi i y subdomain ha con ibu es o he abili y o he ca aly ic co e o polyme ize mul iple nucleo ides in a single enzyme binding cycle; Pol, he DNA polyme ase domain esponsible o he 50–30DNA polyme ase ac i i y; T, he bipa i e humb subdomain (acco ding o Lee e al. 2009) ha con ibu es o empla e–p ime DNA binding; boxes I–III, he conse ed mo i s Exo I–III ha o m he 30–50exonuclease ac i e si e; boxes A–C, he conse ed mo i s Pol A–C ha comp ise he 50–30polyme ase ac i e si e ( o mo e de ailed desc ip ions o each o hese ea u es, see Kaguni [2004] and Eu o e al. [2011]; blue box, he new e eb a e Exo mo i ; ed box, he egion absen in he nema ode species; and o ange box, he new e eb a e IP mo i [see he ex o de ails]. Fo pol g-b, he domain designa ions ollow ha desc ibed by [Fan and Kaguni 2001;Fan e al. 2006]. G een box, he e eb a e dime iza ion [HLH-3b] in e ace; cyan box, he e eb a e M loop [see ex o de ails]; and b own boxes, he conse ed hyd ophobic egions o pol g-ain e ac ion [acco ding o Lee e al. 2009]). All s uc u al elemen s a e ep esen ed o scale, excep mo i s Exo I–III and Pol A–C. (B) Su ace ep esen a ion o he c ys al s uc u e o he human pol gapo-holoenzyme (3IKM; Lee e al. 2009), highligh ing he in e ac ions among he subuni s and pol g-a unc ional domains. S uc u es a e colo ed as shown in (A); da k g ay, dis al pol g-b.(C) Su ace ep esen a ion o he c ys al s uc u e o he human pol gapo-holoenzyme (3IKM; Lee e al. 2009), highligh ing he mo i s and domains iden i ied in his s udy. The s uc u es a e colo ed asshownin(A)and(B). Oli ei a e al. GBE 946 Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om he ances al accesso y subuni polypep ide. We we e unable o ind any pol g-bcoding sequence om species o nema- odes, as discussed below. The as e olu iona y a es obse ed o he pol g-aand -b genes in insec s, unica es, and nema odes appea o ollow a gene al endency desc ibed o he majo i y o nuclea - encoded genes om hese g oups (Li 1997;Mi e a e al. 2005;D osophila 12 Genomes Conso ium e al. 2007; Denoeud e al. 2010). In e es ingly, ou p elimina y analyses o he genes encoding m SSB and m DNA helicase show sub- s i u ion a es ha di e om hose o he pol g-aand -b genes among animal g oups (Oli ei a MT, Haukka J, Kaguni LS, unpublished da a). Thus, i appea s ha he m SSB and m DNA helicase genes may ha e a ypical e olu iona y con- s ain s, bu his equi es u he alida ion. New Mo i in he Exonuclease Domain o Ve eb a e pol g-a: Implica ions o DNA Binding The MSA iden i ied an inse ion o app oxima ely 17 amino acid esidues be ween mo i s Exo II and III o he pol g-a exonuclease domain (H320–A336 in humans; A308–A309 in D. melanogas e ) ha is p esen in all species o Ve eb a a, and possibly o he Deu e os omia species (excep Ci. in es inalis; ig. 4A). Simila ly long egions in he same posi ion a e p esen in Da. pulex (A h opoda: C us acea), Oscheius ipulae (Nema oda: Rhabdi ida), C . gigas (Mollusca: Bi al ia), and A. queenslandica (Po i e a: Demospongiae), bu hese ha e li le sequence conse a ion and may ep esen independen inse ion e en s in o he pol g-agenes o hese species. The inse ion is se e al esidues downs eam o he o ien e , a s uc u al module ha is highly conse ed wi hin euka yo es, and which has been epo ed o coo dina e he balance be ween he polyme ase and exonu- clease unc ions (Szczepanowska and Fou y 2010). Loca ing he new e eb a e Exo mo i in he c ys al s uc- u e o he human pol g-a(PDB: 3IKM; Lee e al. 2009) e ealed ha he inse ion is pa o a diso de ed egion (K319–S344) o which we ha e no s uc u al in o ma ion, bu which may indica e a lexible domain in ol ed in ansien in e ac ions. Modeling he missing esidues in he human pol g-as uc u e based upon he seconda y s uc u e p edic ion algo i hm buil in o he I-TASSER se e esul ed in wo sho alpha helices connec ed by a sho loop, which o ien his elemen owa d he DNA binding cle o he enzyme ( ig. 4B). In pa icula , esidues H320, K327, K331, and K335 a e in close p oximi y o he mino g oo e o he p imed DNA empla e modeled on o he pu a i e DNA bind- ing cle o pol g-a(Eu o e al. 2011). We pos ula e ha he FIG.2.—Bayesian phylogene ic in e ence o animal pol g-anucleo ide sequences. The ou g oup sequence used was he pol g-a(mip-1) om he yeas S. ce e isiae (accession numbe NM_001183750.1). The 50% majo i y- ule consensus ee was in e ed using M Bayes 3.2, as desc ibed unde Ma e ials and Me hods. Bayesian pos e io p obabili y alues a e indica ed o almos all nodes. The scale ba indica es subs i u ions pe si e. E olu ion o Animal pol gGBE Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 947 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om new Exo mo i may ep esen a module ha se es o enhance DNA binding in e eb a e pol gs. New Mo i in he IP Subdomain o Ve eb a e pol g-a: Implica ions o pol g-bIn e ac ions The space egion o pol g-a, which connec s he N- e minal exonuclease domain o he C- e minal polyme ase domain, con ains in insic p ocessi i y (IP) and AID subdomains ( ig. 1). As hei names sugges , hese subdomains p o ide s uc u al pla o ms o suppo ing bo h he in insic p oces- si i y o pol g-aaloneand heenhancedp ocessi i yo he holoenzyme, espec i ely (Lee e al. 2009). Ou MSA also iden- i ied an inse ion o 30 amino acid esidues on a e age in he IP subdomain (E692–R722 in humans; L640–S641 in D. mel- anogas e ) ha is p esen consis en ly in all species o Ve eb a a ( ig. 5A). O he deu e os ome species, such as S . pu pu a us (Echinode ma a) and B anchios oma flo idae (Cephalocho da a), and some o he noninsec animal species also ha e inse ions o a ying sizes in his posi ion. Howe e , he low sequence simila i y does no p o ide enough e idence o homology among he new e eb a e IP mo i and he egions o which i aligns in hese o he animals. In ac , he N- e minal egion o his mo i in e eb a es is highly con- se ed, bu i s C- e minus shows high a iabili y. Again, he sequence om Ci. in es inalis esembles signi ican ly hose o insec species due o he lack o any amino acid esidues in his egion. In summa y, his elemen is a dis inc and conse ed, de i ed ea u e in e eb a e pol g-a; o o he me azoan g oups, he e is no clea indica ion o i s e olu iona y cons ain s. The new IP mo i was localized in a egion o he human pol g-ac ys al s uc u e o which mos o he esidues we e also diso de ed (G674–R709) (Lee e al. 2009). Al hough he C- e minus o he mo i con ains a esidue mu a ed in some cases o human pa ien s wi h Alpe s disease (R722H) and FIG.3.—Bayesian phylogene ic in e ence o animal pol g-bnucleo ide sequences. The ou g oup sequence used was he glycyl- RNA syn he ase om he bac e ial species The mus he mophilus (accession numbe AJ222643.1). The 50% majo i y- ule consensus ee was in e ed using M Bayes 3.2, as desc ibed unde Ma e ials and Me hods. Bayesian pos e io p obabili y alues a e indica ed o almos all nodes. The scale ba indica es subs i u ions pe si e. Oli ei a e al. GBE 948 Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om possibly implica ed in DNA binding (Eu o e al. 2011), model- ing he missing 36 amino acids esul ed in a helix- u n-helix s uc u e loca ed in close p oximi y o a loop a he end o he Middle domain o he p oximal pol g-bp o ome ( ig. 5B). Pa o his loop in he p oximal pol g-b, which o ou knowledge has no been desc ibed p e iously and is he ea e called he “M loop” (T357–K364 in humans; D238–H239 in D. melano- gas e ), is again almos exclusi e o e eb a es ( ig. 5A), FIG.4.—Iden i ica ion o a new Exo mo i in e eb a e pol g-a, po en ially implica ed in p ime – empla e DNA binding. (A) Amino acid sequence alignmen indica es he p esence o he ex a esidues (boxed) be ween pol g-amo i s Exo II and III in all species o Ve eb a a, and a ew o he animal g oups, including o he deu e os ome species. These esidues a e diso de ed in he c ys al s uc u e o he human pol gapo-holoenzyme (3IKM; Lee e al. 2009). (B) S uc u al model o he human esidues indica ed in (A) showing hei p oximi y o he p ime – empla e DNA binding cle . The le panel shows he DNA binding cle s uc u e wi hou he new e eb a e Exo mo i , as i appea s in he PDB da a ile 3IKM. P ime – empla e DNA binding o pol g-awas modeled by Eu o e al. (2011) and he o ien e module (see ex o de ails) is shown as desc ibed by Szczepanowska and Fou y (2010). Colo s a e as in igu e 1. E olu ion o Animal pol gGBE Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 949 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om al hough Echinode ma a (S . pu pu a us), Cnida ia (N. ec en- sis), and Placozoa (T. adhae ens) do possess se e al esidues in his same egion. A i m co ela ion be ween he conse ed new IP mo i in pol g-aand he M loop o pol g-boccu s only o e eb a e species; he absence o bo h s uc u al elemen s is i mly co ela ed only o insec s and unica es. Un o una ely, wi hou a be e axa ep esen a ion, we a e unable o conclude whe he bo h elemen s ha e been los in insec s and unica es o gained in e eb a es. Dime iza ion o pol g-b S uc u al and biochemical da a ha e documen ed ha he human and mouse pol g-bs o m homodime s in solu ion and ha he homodime o m associa es wi h pol g-a o cons i u e a unc ional he e o ime ic holoenzyme (Ca odeguas e al. 2001;Fan e al. 2006;Yakubo skaya e al. 2006;Lee e al. 2009). The majo pol g-bdime iza ion in e ace is p o ided by he HLH-b3 [helix-loop-helix, 3 b-shee s] domain (H133–R182 in humans; N63–Q65 in D. melanogas e ), which is p esen consis en ly ac oss all e eb a e species ( ig. 6A). In e es ingly, he echinode m S . pu pu a us and he placozoan T. adhae- ens also ha e amino acid esidues ha could po en ially old in o a pa ial HLH-b3 s uc u e. We es ed his hypo hesis by modeling he s uc u e o pol g-b om hese wo species and ha om D. melanogas e and Ci. in es inalis (con ol species ha lack comple ely he HLH-b3 elemen ). Nei he he helix- loop-helix s uc u e, esponsible o he o ma ion o he ou - helix bundle wi h he adjacen pol g-b, no he h ee bshee s ound a he base o he ou -helix bundle a e clea ly obse ed o S . pu pu a us and T. adhae ens ( ig. 6B). As expec ed, he pol g-bmodels o D. melanogas e and Ci. in es inalis also ha e none o he s uc u al elemen s necessa y o he HLH-b3 olding. All hese p o eins a e, he e o e, mos likely unable o homodime ize using he same s uc u al ea u es adop ed by FIG.5.—Iden i ica ion o a new, pu a i e in e ac ing su ace be ween he ca aly ic and accesso y subuni s o e eb a e pol g.(A) Amino acid sequence alignmen s indica e he p esence o new mo i s in he pol g-aIP subdomain (boxed, le panel) and in he pol g-bmiddle domain (boxed, igh panel) o all species o Ve eb a a. Only he species o which bo h pol g-aand -bsequences we e e ie ed a e shown. The indica ed esidues in pol g-aa e diso de ed in he c ys al s uc u e o he human pol gholoenzyme (3IKM; Lee e al. 2009); likewise hose in pol g-ba e diso de ed in he c ys al s uc u e o he human pol g-bdime (2G4C; Fan e al. 2006). (B) Model o he human esidues indica ed in (A), sugges ing ha he p edic ed s uc u al elemen s a e in close p oximi y o each o he . The igh panel shows he possible in e ac ing egion in he absence o he newly iden i ied mo i s, as hey appea in he PDB iles 3IKM and 2G4C. P ime – empla e DNA binding o pol gwas modeled by Eu o e al. (2011). Colo s a e as in igu e 1. Oli ei a e al. GBE 950 Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om FIG.6.—Dime iza ion o e eb a e pol g-b h ough he o ma ion o he ou -helix bundle s uc u e. (A) Amino acid sequence alignmen indica es he p esence o he HLH-3bdomain (boxed) in all species o Ve eb a a and possibly in ew o he animal g oups. (B) Compa ison o he c ys al s uc u e o he human pol g-bdime and s uc u al models o pol g-bo T ichoplax adhae ens,S ongylocen o us pu pu a us,D osophila melanogas e , and Ciona in es inalis, showing ha only e eb a e pol g-bcan oldin oaHLH-3bs uc u e and he e o e o m he ou -helix bundle dime iza ion in e ace. The inse shows he h ee sho b-shee s a he base o he HLH-3bs uc u e. Fo mo e in o ma ion abou he s uc u al ea u es o he ou -helix bundle old, see Kam eka and Hech (1995) and Ca odeguas e al. (2001). E olu ion o Animal pol gGBE Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 951 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om Fan L, Sanschag in PC, Kaguni LS, Kuhn LA. 1999. The accesso y subuni o m DNA polyme ase sha es s uc u al homology wi h aminoacyl- RNA syn he ases: implica ions o a dual ole as a p ime ecogni ion ac o and p ocessi i y clamp. P oc Na l Acad Sci U S A. 96: 9527–9532. Fa num GA, Nu minen A, Kaguni LS. 2014. Mapping 136 pa hogenic mu a ions in o unc ional modules in human DNA polyme ase gamma es ablishes p edic i e geno ype-pheno ype co ela ions o he comple e spec um o POLG synd omes. Biochim Biophys Ac a. 1837:1113–1121. Fa CL, Wang Y, Kaguni LS. 1999. Func ional in e ac ions o mi ochon- d ial DNA polyme ase and single-s anded DNA-binding p o ein. Templa e–p ime DNA binding and ini ia ion and elonga ion o DNA s and syn hesis. J Biol Chem. 274:14779–14785. Fou y F. 1989. Cloning and sequencing o he nuclea gene MIP1 encoding he ca aly ic subuni o he yeas mi ochond ial DNA polyme ase. J Biol Chem. 264:20552–20560. Fukuoh A, e al. 2014. Sc een o mi ochond ial DNA copy numbe main- enance genes e eals essen ial ole o ATP syn hase. Mol Sys Biol. 10:734. Fus e JM, e al. 2010. Mi ochond ial RNA polyme ase is needed o ac i- a ion o he o igin o ligh -s and DNA eplica ion. Mol Cell. 37: 67–78. Ga cia-Gomez S, e al. 2013. P imPol, an a chaic p imase/polyme ase op- e a ing in human cells. Mol Cell. 52:541–553. Gissi C, Iannelli F, Pesole G. 2008. E olu ion o he mi ochond ial genome o Me azoa as exempli ied by compa ison o congene ic species. He edi y (Edinb). 101:301–320. Hance N, Eks and MI, T i uno ic A. 2005. Mi ochond ial DNA polyme ase gamma is essen ial o mammalian emb yogenesis. Hum Mol Gene . 14:1775–1783. Hol IJ, Lo ime HE, Jacobs HT. 2000. Coupled leading- and lagging-s and syn hesis o mammalian mi ochond ial DNA. Cell 100:515–524. Hy a inen AK, e al. 2007. The mi ochond ial ansc ip ion e mina ion ac o mTERF modula es eplica ion pausing in human mi ochond ial DNA. Nucleic Acids Res. 35:6458–6474. Hy a inen AK, Pohjoismaki JL, Hol IJ, Jacobs HT. 2011. O e exp ession o MTERFD1 o MTERFD3 impai s he comple ion o mi ochond ial DNA eplica ion. Mol Biol Rep. 38:1321–1328. Iyenga B, Luo N, Fa CL, Kaguni LS, Campos AR. 2002. The accesso y subuni o DNA polyme ase gamma is essen ial o mi ochond ial DNA main enance and de elopmen in D osophila melanogas e .P ocNa l Acad Sci U S A. 99:4483–4488. Iyenga B, Roo e J, Campos AR. 1999. The amas gene, iden i ied as a mu a ion ha dis up s la al beha io in D osophila melanogas e , codes o he mi ochond ial DNA polyme ase ca aly ic subuni (DNApol-gamma125). Gene ics 153:1809–1824. Jiang ZJ, e al. 2007. Compa a i e mi ochond ial genomics o snakes: ex ao dina y subs i u ion a e dynamics and unc ionali y o he duplica e con ol egion. BMC E ol Biol. 7:123. Joe s P, e al. 2013. Mi ochond ial ansc ip ion e mina o amily mem- be s mTTF and mTe 5 ha e opposing oles in coo dina ion o m DNA syn hesis. PLoS Gene . 9:e1003800. Joe s P, Jacobs HT. 2013. Analysis o eplica ion in e media es indica es ha D osophila melanogas e mi ochond ial DNA eplica es by a s and-coupled he a mechanism. PLoS One 8:e53249. Kaguni LS. 2004. DNA polyme ase gamma, he mi ochond ial eplicase. Annu Re Biochem. 73:293–320. Kam eka S, Hech MH. 1995. P o ein Mo i s. 7. The ou -helix bundle: wha de e mines a old? FASEB J. 9:1013–1022. Ka oh K, Toh H. 2008. Recen de elopmen s in he MAFFT mul iple se- quence alignmen p og am. B ie Bioin o m. 9:286–298. Ko honen JA, Gaspa i M, Falkenbe g M. 2003. TWINKLE Has 50!30 DNA helicase ac i i y and is speci ically s imula ed by mi ochond ial single-s anded DNA-binding p o ein. J Biol Chem. 278: 48627–48632. Ko honen JA, Pham XH, Pelleg ini M, Falkenbe g M. 2004. Recons i u ion o a minimal m DNA eplisome in i o. EMBO J. 23:2423–2429. La o DV, Boo e JL, B own WM. 2000. The comple e mi ochond ial DNA sequence o he ho seshoe c ab Limulus polyphemus. Mol Biol E ol. 17:813–824. Lec enie N, Van De B uggen P, Fou y F. 1997. Mi ochond ial DNA poly- me ases om yeas o man: a new amily o polyme ases. Gene 185: 147–152. Lee YS, e al. 2010. Each monome o he dime ic accesso y p o ein o human mi ochond ial DNA polyme ase has a dis inc ole in con e ing p ocessi i y. J Biol Chem. 285:1490–1499. Lee YS, Kennedy WD, Yin YW. 2009. S uc u al insigh in o p ocessi e human mi ochond ial DNA syn hesis and disease- ela ed polyme ase mu a ions. Cell 139:312–324. Lewis DL, Fa CL, Kaguni LS. 1995. D osophila melanogas e mi ochon- d ial DNA: comple ion o he nucleo ide sequence and e olu iona y compa isons. Insec Mol Biol. 4:263–278. Lewis DL, Fa CL, Wang Y, Lagina AT 3 d, Kaguni LS. 1996. Ca aly ic subuni o mi ochond ial DNA polyme ase om D osophila emb yos. Cloning, bac e ial o e exp ession, and biochemical cha ac e iza ion. J Biol Chem. 271:23389–23394. Lewis SC, Joe s P, Wilcox S, G i i h JD, Jacobs HT, Hyman BC. 2015. A olling ci lce eplica ion mechanism p oduces mul ime ic la ia s o mi ochondi al DNA in Caeno habdi is elegans. PLoS Gene . 11: e1004985. Li W-H. 1997. Molecula e olu ion. Sunde land (MA): Sinaue Associa es. Lim SE, Longley MJ, Copeland WC. 1999. The mi ochond ial p55 accesso y subuni o human DNA polyme ase gamma enhances DNA binding, p omo es p ocessi e DNA syn hesis, and con e s N-e hylmaleimide esis ance. J Biol Chem. 274:38197–38203. Logan DT, Mazau ic MH, Ke n D, Mo as D. 1995. C ys al s uc u e o glycyl- RNA syn he ase om The mus he mophilus.EMBOJ.14: 4156–4167. Longley MJ, P asad R, S i as a a DK, Wilson SH, Copeland WC. 1998. Iden i ica ion o 50-deoxy ibose phospha e lyase ac i i y in human DNA polyme ase gamma and i s ole in mi ochond ial base excision epai in i o. P oc Na l Acad Sci U S A. 95: 12244–12248. Mi e a M, Blax e ML, Bi d DM, McCa e JP. 2005. Compa a i e geno- mics o nema odes. T ends Gene . 21:573–581. Mo aes CT, Bacman SR, Williams SL. 2014. Manipula ing mi ochond ial genomes in he clinic: playing by di e en ules. T ends Cell Biol. 24: 209–211. Oli ei a MT, Ga esse R, Kaguni LS. 2010. Animal models o mi ochond ial DNA ansac ions in disease and ageing. Exp Ge on ol. 45:489–502. Oli ei a MT, Kaguni LS. 2010. Func ional oles o he N- and C- e minal egions o he human mi ochond ial single-s anded DNA-binding p o- ein. PLoS One 5:e15379. Oli ei a MT, Kaguni LS. 2011. Reduced s imula ion o ecombinan DNA polyme ase gamma and mi ochond ial DNA (m DNA) helicase by a - ian s o mi ochond ial single-s anded DNA-binding p o ein (m SSB) co ela es wi h de ec s in m DNA eplica ion in animal cells. J Biol Chem. 286:40649–40658. Olson MW, Wang Y, Elde RH, Kaguni LS. 1995. Subuni s uc u e o mi ochond ial DNA polyme ase om D osophila emb yos. Physical and immunological s udies. J Biol Chem. 270:28932–28937. Philippe H, e al. 2009. Phylogenomics e i es adi ional iews on deep animal ela ionships. Cu Biol. 19:706–712. Pohjoismaki JL, e al. 2010. Mammalian mi ochond ial DNA eplica ion in e media es a e essen ially duplex bu con ain ex ensi e ac s o RNA/DNA hyb id. J Mol Biol. 397:1144–1155. Oli ei a e al. GBE 958 Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om Reyes A, e al. 2013. Mi ochond ial DNA eplica ion p oceeds ia a “boo - lace” mechanism in ol ing he inco po a ion o p ocessed ansc ip s. Nucleic Acids Res. 41:5837–5850. Reyes A, Yang MY, Bowmake M, Hol IJ. 2005. Bidi ec ional eplica ion ini ia es a si es h oughou he mi ochond ial genome o bi ds. J Biol Chem. 280:3242–3250. Ronquis F, e al. 2012. M Bayes 3.2: e icien Bayesian phylogene ic in- e ence and model choice ac oss a la ge model space. Sys Biol. 61: 539–542. Sali A, Blundell TL. 1993. Compa a i e p o ein modelling by sa is ac ion o spa ial es ain s. J Mol Biol. 234:779–815. Spelb ink JN, e al. 2000. In i o unc ional analysis o he human mi o- chond ial DNA polyme ase POLG exp essed in cul u ed human cells. J Biol Chem. 275:24818–24828. S iban J, Fa num GA, Ho de SL, Kaguni LS. 2014. The N- e minal do- main o he D osophila mi ochond ial eplica i e DNA helicase con- ains an i on-sul u clus e and binds DNA. J Biol Chem. 289: 24032–24042. Suyama M, To en s D, Bo k P. 2006. PAL2NAL: obus con e sion o p o ein sequence alignmen s in o he co esponding codon align- men s. Nucleic Acids Res. 34:W609–W612. Szczepanowska K, Fou y F. 2010. A clus e o pa hogenic mu a ions in he 30-50exonuclease domain o DNA polyme ase gamma de ines a no el module coupling DNA syn hesis and deg ada ion. Hum Mol Gene . 19:3516–3529. Wang Y, Fa CL, Kaguni LS. 1997. Accesso y subuni o mi ochond ial DNA polyme ase om D osophila emb yos. Cloning, molecula anal- ysis, and associa ion in he na i e enzyme. J Biol Chem. 272: 13640–13646. Wang Y, Kaguni LS. 1999. Baculo i us exp ession econs i u es D osophila mi ochond ial DNA polyme ase. J Biol Chem. 274: 28972–28977. Wan ooij S, e al. 2008. Human mi ochond ial RNA polyme ase p imes lagging-s and DNA syn hesis in i o. P oc Na l Acad Sci U S A. 105: 11122–11127. We ne e CM, Kaguni LS. 1986. A mi ochond ial DNA polyme ase om emb yos o D osophila melanogas e . Pu i ica ion, subuni s uc u e, and pa ial cha ac e iza ion. J Biol Chem. 261:14764–14770. Wol YI, Koonin EV. 2001. O igin o an animal mi ochond ial DNA poly- me ase subuni ia lineage-speci ic acquisi ion o a glycyl- RNA syn he- ase om bac e ia o he The mus-Deinococcus g oup. T ends Gene . 17:431–433. Yakubo skaya E, Chen Z, Ca odeguas JA, Kiske C, Bogenhagen DF. 2006. Func ional human mi ochond ial DNA polyme ase gamma o ms a he e o ime . J Biol Chem. 281:374–382. Yang MY, e al. 2002. Biased inco po a ion o ibonucleo ides on he mi ochond ial L-s and accoun s o appa en s and-asymme ic DNA eplica ion. Cell 111:495–505. Yasukawa T, e al. 2006. Replica ion o e eb a e mi ochond ial DNA en ails ansien ibonucleo ide inco po a ion h oughou he lagging s and. EMBO J. 25:5358–5371. Ye F, Samuels DC, Cla k T, Guo Y. 2014. High- h oughpu sequencing in mi ochond ial DNA esea ch. Mi ochond ion 17C:157–163. Young MJ, e al. 2011. Biochemical analysis o human POLG2 a ian s associa ed wi h mi ochond ial disease. Hum Mol Gene . 20: 3052–3066. Associa e edi o : Sa ah Schaack E olu ion o Animal pol gGBE Genome Biol. E ol. 7(4):943–959. doi:10.1093/gbe/e 042 Ad ance Access publica ion Ma ch 3, 2015 959 a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on No embe 15, 2016h p://gbe.ox o djou nals.o g/Downloaded om