1
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
www.na u e.com/scien i ic epo s
Dii on cen e mu a ions in Ciona
in es inalis al e na i e oxidase
abolish enzyma ic ac i i y and
p e en escue o cy och ome
oxidase de iciency in lies
Ana Andjelko ić1, Ma cos T. Oli ei a1,2, Giuseppe Cannino1, Cag i Yalgin1,
P a een K. Dhandapani1,3, E ic Du ou 1, Pie e Rus in4, Ma en Szibo 1,3 & Howa d T. Jacobs1,3
The mi ochond ial al e na i e oxidase, AOX, ca ies ou he non p o on-mo i e e-oxida ion o
ubiquinol by oxygen in lowe euka yo es, plan s and some animals. He e we c ea ed a modi ied e sion
o AOX om Ciona ins es inalis, ca ying mu a ions a conse ed esidues p edic ed o be equi ed o
chela ion o he dii on p os he ic g oup. The modi ied p o ein was s ably exp essed in mammalian cells
o lies, bu lacked enzyma ic ac i i y and was unable o escue he pheno ypes o lies knocked down
o a subuni o cy och ome oxidase. The mu a ed AOX ansgene is hus a po en ially use ul ool in
s udies o he physiological e ec s o AOX exp ession.
The mi ochond ial al e na i e oxidase, AOX, ca ies ou he non p o on-mo i e e-oxida ion o ubiquinol by
molecula oxygen. Te minal elec on ans e by AOX cons i u es a pa allel sys em o ha p o ided by OXPHOS
complexes III and IV in plan s, ungi, p o is s and many animal phyla1. AOX is belie ed o become ac i a ed unde
s ess condi ions, when he OXPHOS cy och ome chain is o e loaded o una ailable.
In many o ganisms his is achie ed, a leas in pa , ia he egula ed exp ession o he AOX gene, which is
induced by a a ie y o s esses ele an o OXPHOS dys unc ion2,3. The enzyme is also inhe en ly esponsi e o
he me abolic signa u e o such s esses in di e en o ganisms. Fi s ly, i is ac i a ed by high le els o i s educed
subs a e, ubiquinol
4,5
, which is assumed o e lec a lowe a ini y o he subs a e han ha exhibi ed by OXPHOS
complex III, wi h which i compe es. Thus, unde no mal physiological condi ions, mos o he elec on low
om ubiquinol o oxygen is channelled h ough complexes III and IV, e en i AOX is physically p esen . Only i
ubiquinol le els inc ease, o example, i he enzyma ic capaci y o complexes III and IV becomes limi ing, will
AOX become unc ionally signi ican . In addi ion, AOX is allos e ically ac i a ed in many o ganisms by me abo-
li es whose le els inc ease unde condi ions o OXPHOS insu iciency, o example py u a e3, as well as by o he
me aboli es indica i e o cellula edox s a e.
Al hough he AOX gene has been los , du ing he cou se o e olu ion, in he lineages leading o he mos com-
plex and ad anced me azoan g oups, including mammals
1
, we easoned ha i s ein oduc ion by ansgenesis
should enable such animals o bu e many o he pa hological s esses esul ing om OXPHOS dys unc ion6.
Thus AOX could become a he apeu ic ool o ea ing mi ochond ial diseases and o he condi ions media ed by
OXPHOS dys unc ion7. P elimina y es s in model o ganisms, including cul u ed human cells8,9, D osophila10,11
and he mouse12, suppo his concep . In pa icula , he exp ession o AOX om he unica e Ciona in es inalis, was
shown o compensa e many o he pheno ypes esul ing om cy och ome oxidase (COX, complex IV) de iciency
in D osophila, including he knockdown o s uc u ally essen ial subuni s o he complex
11
. Howe e , i AOX is o
be o alue in e en ual he apy, he mechanism o his compensa ion needs o be es ablished. The hypo hesized
1BioMediTech and Tampe e Uni e si y Hospi al, Uni e si y o Tampe e, FI-33014, Finland. 2Depa amen o de
Tecnologia, Faculdade de Ciências Ag á ias e Ve e iná ias, Uni e sidade Es adual Paulis a “Júlio de Mesqui a Filho”,
14884-900 Jabo icabal, SP, B azil. 3Ins i u e o Bio echnology, Uni e si y o Helsinki, FI-00014, Finland. 4INSERM
UMR 1141 and Uni e si é Pa is 7, Facul é de Médecine Denis Dide o , Hôpi al Robe Deb é, 48, Boule a d Sé u ie ,
75019, Pa is, F ance. Co espondence and eques s o ma e ials should be add essed o H.T.J. (email: howa d.
[email p o ec ed])
Recei ed: 27 Augus 2015
Accep ed: 22 Oc obe 2015
Published: 17 Decembe 2015
OPEN
www.na u e.com/scien i ic epo s/
2
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
enzyma ic by-pass is only one o se e al possible such mechanisms. Exp ession o an ine ansgene, such as
GFP, in place o AOX, was unable o escue he pheno ypes p oduced by enginee ed de iciency o cy och ome
oxidase10,11. Howe e , his con ol canno be unambiguously in e p e ed, since he exp essed GFP was no a ge ed
o mi ochond ia, and e en i i we e, does no possess o he s uc u al ea u es o AOX ha enable i o inse in o
he inne mi ochond ial memb ane in a speci ic ashion and in e ac wi h o he componen s he eo .
In o de o p o ide a mo e applicable es o whe he he abili y o AOX o escue COX de iciency depends
on i s p ima y enzyma ic ac i i y, we sough o enginee he AOX in such a way as o des oy his ac i i y, whils
p oducing only a minimal e ec on he o e all s uc u e, s abili y and exp ession o he p o ein. To do his, we ook
ad an age o he ac ha AOX is well conse ed phylogene ically, ha he esidues con ibu ing o i s ac i e si e
ha e been cha ac e ized in a numbe o species, and ha he s uc u e o a ep esen a i e AOX, om he p o is an
pa asi e T ypanosoma b ucei, has ecen ly been published
13
. Using cu en ly a ailable bioin o ma ics ools, we
modelled he s uc u e o he Ciona in es inalis enzyme agains his empla e, p edic ed amino-acids equi ed o
binding he ca aly ically essen ial dii on moie y a he ac i e si e, and p oceeded ia alanine-subs i u ion mu agen-
esis o c ea e an exp essible e sion o he enzyme expec ed o lack enzyma ic ac i i y, despi e being p edic ed o
old o a simila o e all s uc u e. In se e al di e en con ex s (cul u ed human and D osophila cells, as well as
whole animals), we ound ha he mu a ed AOX was s ably exp essed bu de oid o de ec able enzyma ic ac i i y.
Fu he mo e, exp ession o he ansgene encoding he mu a ed AOX was unable o escue enginee ed COX de i-
ciency in he ly, con i ming ha his escue indeed depends on he enzyma ic ac i i y o AOX.
Ma e ials and Me hods
Sequence alignmen s and molecula modelling.
The sequences o AOX homologues ound by Blas P
sea ching we e aligned using he MUSCLE algo i hm buil in o he so wa e MEGA614, wi h de aul pa ame e s.
A homologous model o he s uc u e o one subuni o he C. in es inalis AOX was gene a ed using he so wa e
I-TASSER15, based on he c ys al s uc u e o he T ypanosoma b ucei AOX (PDB 3VV9:A)13 as empla e and he
mul iple sequence alignmen desc ibed abo e as inpu es ain . O he pa ame e s we e se as de aul . Selec ion
o he model was based upon he bes accu acy es ima ions p o ided by he C-sco es, es ima ed TM-sco es and
RMSD alues. Because he N- e minal egion (M1-K103) o he C. in es inalis AOX s uc u e could no be mod-
elled wi h high accu acy, his egion was elimina ed om he analysis. The dime ic model o C. in es inalis AOX
and he posi ioning o he wo dii on cen es (one pe subuni ) we e buil by o e lapping wo copies o he model
gene a ed by I-TASSER in o he c ys al s uc u e o he dime ic T. b ucei AOX using Pymol (www.pymol.o g).
Pymol was also used o analyze all s uc u e models and o p oduce he igu es.
Cloning p ocedu es and mu agenesis. Fo D osophila exp ession, he C. in es inalis AOX coding sequence,
including i s na u al s op codon, was ecloned om he pMT/V5-His B ec o (In i ogen), in which i had been
p e iously p opaga ed, in o he EcoRI si e o pUASTa B
16
. Based on he mul iple sequence alignmen shown in
Fig. S1, and he esul s o molecula modelling (see Resul s), PCR-based alanine subs i u ion mu agenesis and
ecloning we e ca ied ou acco ding o he scheme o Fig. S2. Mu a ions E239A, H242A, E344A and H347A we e
in oduced, using he plasmid-bo ne AOX cDNA as empla e, P u DNA polyme ase (S a agene) and oligonucleo-
ides (all shown 5′ o 3′ ) as ollows: GAAGCTGAAAATGcGAGAATGgcCTTAATGACTGCG and CGCAGTCA
TTAAGgcCATTCTCgCATTTTCAGCTTC o c ea e E239A/H242A, ollowed by ATCTGAGCTGAT
GcAGCACATgcCAGATCAGTCAAC and GTTGACTGATCTGgcATGTGCTgCATCAGCTCGGAT o c ea e
E344A/H347A (lowe case le e s indica e he si es o in oduced mu a ions). Fo exp ession in S2 cells, cons uc s
con aining he o iginal and mu a ed AOX cDNA inse s, again using he na u al s op codon, we e ecloned in o
he EcoRI si e o pAc5.1/V5-His B (In i ogen, USA) o c ea e pAC/AOX17 and pAC/mu AOX. Fo ansien
mammalian exp ession, he wild- ype and mu a ed AOX coding sequences we e ecloned, espec i ely, in o a
pBR322-de i ed kanR plasmid con aining he CAG p omo e 18 and bo ine g ow h ho mone poly(A) signal, oge he
wi h o he elemen s no ele an o he p esen s udy (copies o he e ope a o , loxP si es, insula o elemen s and
po ions o he po cine Gg a1 gene), o c ea e he exp ession cons uc s pCAG-AOX and pCAG-mu AOX. The
nucleo ide sequences o all clones we e con i med by Sange sequencing using he Big Dye Te mina o 3.1 ki
(Li e Technologies) and an ABI3130xl Gene ic Analyze , acco ding o he manu ac u e ’s speci ica ions.
D osophila s ocks and main enance.
Excep whe e s a ed, lies we e main ained and g own on s anda d
medium a 25°C, using a 12 h ligh /da k cycle, as p e iously10,19. Balance s, ecipien line w1118, he RNAi line o
CG9603 (Vienna D osophila RNAi Cen e line 106661), he ubiqui ous da-GAL4 d i e (Blooming on line 8641)
and he d i e line bea ing ela C155-GAL4 on ch omosome X and UAS-Dc 2 on ch omosome 2 (Blooming on line
25750), we e ob ained om s ock cen es. Φ C31 ecombinase-media ed-si e-di ec ed ansgenesis was used o
gene a e ansgenic ly lines (se ice p o ided by Bes Gene Inc, Chino Hills, CA), using ecipien lines wi h he
ollowing in eg a ion si es: a P18 (ch omosome X), a P40 (ch omosome 2) and a P2 (ch omosome 3), acco ding
o P ei e e al.
20
, employing he wild- ype and mu a ed AOX cons uc s cloned in pUASTa B and pUASTa B
i sel as emp y- ec o con ol. Following cha ac e iza ion, ansgenic lines we e main ained o e balance s app o-
p ia e o ch omosome X, 2 o 3, bea ing s anda d ma ke s (FM7, CyO, TM3Sb, espec i ely). T ansgenic lines
UAS-AOXF24 and UAS-AOXF6 we e desc ibed p e iously10.
Cell cul u e and ans ec ion.
HEK293T cells we e cul u ed as p e iously21. Pla es o 3 × 106 cells we e ans
-
ec ed wi h 24 μ g o he pCAG-AOX o pCAG-mu AOX plasmids o , as con ol, emp y ec o (pWPI, Addgene),
using 60 μ l Lipo ec amine® 2000 (In i ogen) unde manu ac u e ’s ecommended condi ions. D osophila S2 cells
we e g own and ans ec ed wi h pAc5.1/V5-His B o de i a i es as p e iously17.
www.na u e.com/scien i ic epo s/
3
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
Exp ession assays.
RNA ex ac ion and QRTPCR o measu e AOX ansc ip le els using RpL32 RNA as an
in e nal no maliza ion s anda d we e as p e iously desc ibed
10
, using RNA om 2 day-old adul male and emale
lies. P o ein ex ac ion om 2 day-old D osophila adul s and Wes e n blo s we e conduc ed essen ially as by
Fe nandez-Ayala e al.10, wi h he ollowing modi ica ions: o emales, 1% SDS was used o lysis ins ead o 1.5%
T i on X-100, lies we e p ocessed in ba ches o 30 ( emales) o 40 (males), SDS-PAGE used Any kD™ C i e ion™
TGX
™
18-well gels (Bio-Rad), P es ained P o ein Ladde (The mo-Scien i ic) and P oSie e
TM
EX Running and
T ans e Bu e s (Lonza), and memb anes we e ea ed in PBS-Tween® ins ead o TBS. P ima y an ibodies used
we e cus omized abbi an i-AOX10 (21s Cen u y Biochemicals, 1:10,000), abbi an i-α -ac ininin C-20-R (San a
C uz Bio echnology, 1:5,000) and mouse an i-ATP5A (Abcam, 1:50,000). Seconda y an ibodies we e Pe oxidase
Goa An i- abbi IgG and Ho se An i-mouse IgG (bo h om Vec o Labo a o ies, 1:10,000). Pos -nuclea ex ac s
(PN) om HEK293T cells we e p epa ed acco ding o Cannino e al.21. P o ein concen a ions we e measu ed
using he B ad o d assay.
Respi ome y. Oxygen consump ion o 5 × 106 human cells was measu ed 48 h a e ans ec ion, ollow-
ing pe meabiliza ion wi h 80 μ g/ml digi onin, in a Cla k- ype elec ode (Hansa ech Oxy he m sys em) using
espi a o y bu e A
22
a 37 °C. Complex II-d i en espi a ion was measu ed in he p esence o 10 mM ADP and
10 mM succina e. AOX-d i en (an imycin- esis an ) espi a ion was measu ed a e he u he addi ion o (60 ng/
ml) an imycin A, wi h sub ac ion o any esidual oxygen consump ion a e adding 100 μ M n-p opyl galla e.
Respi ome y on S2 cells was as desc ibed p e iously
17
and was also conduc ed on homogena es om 1–4 day-old
D osophila males. B ie ly, 25 males we e gen ly homogenized in 0.8 ml ice-cold isola ion bu e (250 mM suc ose,
5 mM T is-HCl, 2 mM EGTA, pH 7.4) and muslin- il e ed. Respi ome y was pe o med on 150 μ l aliquo s o his
homogena e, mixed wi h 500 μ l assay bu e (120 mM KCl, 5 mM KH
2
PO
4
, 3 mM HEPES-KOH, 1 mM EGTA,
1 mM MgCl2, 0.2% BSA, pH 7.2), subs a es (15 mM glyce ol-3-phospha e and 5 mM ADP) and inhibi o s as o
pe meabilized mammalian cells.
Beha iou al assays. Time o eclosion ollowing D osophila c osses was measu ed as p e iously23. Eggs om
pa en s c ossed wo days ea lie we e collec ed o e h ee consecu i e nigh s, and cul u ed a 25°C. Adul s less
han 24 h old we e collec ed and so ed on ice, a e which ba ches o 5 male lies we e placed in each emp y ial.
A e a 10 min wai ing pe iod, lies we e ipped down and hei subsequen beha iou eco ded using a DFK
21AF04 came a (The Imaging Sou ce, B emen, Ge many) and Media Reco de 2 so wa e (Noldus, Wageningen,
Ne he lands). The climbing index
11
o each ial was manually calcula ed om eco dings as he mean numbe
o lies which climbed 6 cm in 10 s in h ee ials. Climbing indices om di e en geno ypes we e compa ed by
one-way ANOVA wi h Bon e oni adjus men , using SPSS 12. The box plo was d awn wi h BoxPlo R (boxplo .
ye slab.com), wi h Tukey s yle whiske s ex ending o he da a poin ha is no mo e han 1.5 × IQR (in e qua ile
ange) om he edge o he box24.
Human subjec s. The wo k epo ed he e did no use human subjec s o any ma e ials de i ed om human
subjec s, o he han he eely a ailable cell-line HEK293T.
Resul s and Discussion
Modelling and c ea ion o mu a ed AOX ansgene.
Alignmen o he p edic ed Ciona in es inalis AOX
amino-acid sequence wi h he co esponding p o ein om o he axa, including T ypanosoma b ucei, e ealed
conse a ion o esidues implica ed in he o ganiza ion o he dii on cen e o he enzyme, as p e iously epo ed
by Shiba e al.
13
. The ou in a ian glu ama e esidues and wo his idines co espond in Ciona AOX wi h E200,
E239, E290, E344, H242 and H347 (Fig. S1), numbe ed om he i s me hionine o he pu a i e p ep o ein. In he
T ypanosoma AOX s uc u e, he conse ed his idines pa icipa e in a hyd ogen bond ne wo k ha also includes
a conse ed y osine, Y297 in Ciona AOX (Fig. S1). S uc u al modelling (Fig.1) showed ha Ciona AOX can
old o an almos iden ical s uc u e as i s T ypanosoma coun e pa , igno ing he poo ly conse ed N- e minal
egion ( esidues 1–103 o he Ciona p o ein, Fig. S1). Fou alpha-helices enclose he dii on cen e o each p o ome
o he homodime ic p o ein, wi h he conse ed glu ama e and his idine esidues simila ly jux aposed as in he
T ypanosoma p o ein (Fig.1). Based on his s uc u e, we es ed he unc ional signi icance o he conse ed
esidues a he p edic ed dii on cen e, by mu a ing ou o hem o alanine (E239A, H242A, E344A, H347A), in
app op ia e ansgenic cons uc s o exp ession in mammalian cells and D osophila (Fig. S2). The mu a ions we e
p edic ed o des oy he binding o i on o he ac i e si e, whils only minimally dis u bing he o e all s uc u e
o each subuni .
Mu a ed AOX can be s ably exp essed in mammalian cells and lies. In o de o es i s unc ionali y,
he exp ession o he mu a ed AOX cons uc (mu AOX) was i s e i ied, ollowing ansien ans ec ion in o
cul u ed human cells. Based on Wes e n blo ing (Fig.2A), he mu AOX p o ein was he same size and compa-
ably exp essed as wild- ype AOX. Nex , he mu AOX ansgene, unde he con ol o he GAL4-dependen UAS
p omo e , was in oduced in o he D osophila genome by a ge ed inse ion a single si es on each ch omosome.
Pa allel con ol lines we e c ea ed, con aining wild- ype AOX and emp y ec o , inse ed a he same si es. Following
alida ion o he inse ions by PCR and sequencing, we measu ed ansgene exp ession di ec ed by he ubiqui ous
da-GAL4 d i e , a bo h RNA and p o ein le els, using QRTPCR (Fig.2B, C) and Wes e n blo ing (Fig.2D, E).
In bo h emales (Fig.2B) and males (Fig.2C), he exp ession o wild- ype and mu AOX we e simila a he
RNA le el, bu 3–4 old less han AOX in he p e iously c ea ed ansgenic lines, enginee ed by andom P-elemen
inse ion. A he p o ein le el, mu AOX showed sligh ly lowe exp ession han wild- ype AOX in bo h sexes, and
exp ession was again less han in he p e iously c ea ed lines (Fig.2D, E).
www.na u e.com/scien i ic epo s/
4
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
When exp essed ubiqui ously using he da-GAL4 d i e , he AOX and mu AOX ansgenes p oduced only
e y small changes in de elopmen al iming, mos o hem non-signi ican compa ed wi h he co esponding
ec o -only line (Fig.3).
Mu a ed AOX lacks de ec able enzyma ic ac i i y. The unc ionali y o he exp essed AOX a i-
an s was es ed by pola og aphy. Pe meabilized HEK293T cells, ollowing ansien ans ec ion wi h wild- ype
AOX, suppo ed app oxima ely 80% o he uninhibi ed oxygen consump ion, in he p esence o an imycin.
An imycin- esis an oxygen consump ion was unde ec able in pe meabilized cells ansien ly ans ec ed wi h he
mu AOX cons uc o emp y ec o (Fig.4A). A simila esul was ob ained a e ans ec ion o D osophila S2 cells.
A e ans ec ion wi h ei he o wo di e en AOX-exp essing cons uc s, whole-cell espi a ion in he p esence
o an imycin was 70–73% o he uninhibi ed a e, bu was unde ec able in con ol cells o cells ans ec ed wi h he
mu AOX cons uc (Table S1). Finally, in homogena es om male ansgenic lies ca ying a ge ed inse ions a he
same locus (on ch omosome 2), induced o exp ess he ansgene ubiqui ously using he da-GAL4 d i e , wild- ype
AOX suppo ed 14% o he uninhibi ed subs a e oxida ion a e in he p esence o an imycin (Fig.4B), whe eas
mi ochond ia om mu AOX- o emp y ec o - ansgenic lies showed no an imycin- esis an subs a e oxida ion.
In e e y pola og aphy expe imen , exp ession o he AOX ansgene was e i ied by Wes e n blo ing as pe Fig.2.
Mu a ed AOX is unable o escue COX knockdown in lies. The ac ha he mu a ed AOX is de oid
o de ec able enzyma ic ac i i y allowed us o use he newly c ea ed ansgenic lines o es whe he he p e iously
obse ed pheno ypic escue o lies knocked down o a subuni o cy och ome oxidase (Cox7a) was due o he
enzyma ic ac i i y o AOX o some o he p ope y con e ed by he AOX p o ein, when exp essed in D osophila.
Mo eo e , he ac ha he newly c ea ed ansgenic lines exp ess AOX a only abou 30% o he le el o he lines
p e iously s udied, allowed us o es whe he pheno ypic escue was quan i a i ely dependen on AOX exp es-
sion le el. Ubiqui ous knockdown o CG9603, he b oadly exp essed isogene o Cox7a, was p e iously shown o
p oduce pupal le hali y11, which was escued by high-le el exp ession o AOX.
To es he new ansgenic lines, we i s con i med ha he RNAi line used in he expe imen was de oid o
he addi ional inse ion p e iously epo ed o con e pupal le hali y un ela ed o speci ic a ge knockdown25
(Fig. S3). We hen combined he CG9603 RNAi line wi h AOX and con ol ansgenes, plus he da-GAL4 d i e
o induce simul aneous ansgene exp ession and Cox7a knockdown. Wild- ype AOX escued he le hali y, as
p e iously (Fig.5A), whe eas mu AOX o he emp y ec o we e unable o do so, con i ming ha AOX enzyma ic
ac i i y is equi ed o he escue.
Nex , we in es iga ed he e ec s o CG9603 knockdown and i s po en ial escue by AOX, using he
neu on-speci ic d i e ela
C155
-GAL4. P e iously, i was shown ha his p oduces a locomo o de ec in newly
Figu e 1. S uc u al modelling and mu agenesis o ac i e si e o Ciona in es inalis AOX. (A) Model o
he ac i e si e o he Ciona (Ci) enzyme, g een, compa ed wi h he s uc u e o he T ypanosoma b ucei (Tb)
AOX, blue. In bo h cases, he dii on si e (i on moie ies in o ange, hyd oxyl in pink) is bu ied in a ou alpha-
helix bundle. Fo cla i y, only one p o ome is shown. (B) Conse ed esidues binding he dii on cen e show
an iden ical a angemen in he Ci model (g een) as in he Tb s uc u e (blue). (C) The esidues selec ed o
alanine-subs i u ion mu agenesis in he Ci enzyme (he e shown in blue), alongside he esul ing modelled
s uc u e.
www.na u e.com/scien i ic epo s/
5
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
Figu e 2. Exp ession o AOX ansgenes in mammalian cells and D osophila. (A) Wes e n blo o p o ein
ex ac s om HEK293T cells ans ec ed wi h wild- ype and mu a ed AOX cons uc s (w , mu ) o emp y
ec o (V), p obed o AOX and o ATP syn hase subuni α as loading con ol. (B,C) Rela i e AOX exp ession
a RNA le el, based on QRTPCR, in (B) emales and (C) males o di e en D osophila lines ansgenic o
wild- ype o mu a ed AOX, o emp y ec o , inse ed on ch omosomes X, 2 and 3, as shown, in combina ion
wi h he ubiqui ous da-GAL4 d i e . New w (wild- ype) and mu AOX lines we e hose c ea ed by si e-
speci ic in eg a ion a de ined ch omosomal si es using he Φ C31 sys em; old w AOX lines we e UAS-AOXF6
(ch omosome 2) and UAS-AOXF24 (ch omosome 3). Fo males, all alues we e signi ican ly di e en om
emp y- ec o lines; old w AOX lines we e signi ican ly di e en om new w AOX lines (p < 0.001, ANOVA
ollowed by pos -hoc Bon e oni-co ec ed es ), bu mu AOX and new w AOX lines we e no signi ican ly
di e en om each o he . S a is ical analysis o emales ga e simila esul s, al hough g ea e sample- o-sample
a ia ion o old w AOX lines yielded only p < 0.05 compa ing hem wi h new w o mu AOX lines. (D,E)
Wes e n blo o p o ein ex ac s om he same lies (amoun s as shown), p obed o AOX o , as loading con ol,
ei he ATP syn hase subuni α o α -ac inin, as indica ed.
Figu e 3. De elopmen al ime o eclosion o AOX ansgenic lies. Eclosion day (mean +SD) o emales and
males o di e en D osophila lines ansgenic o wild- ype (w ) o mu a ed (mu ) AOX, o emp y ec o ( ),
inse ed on ch omosomes X, 2 and 3, as shown, in combina ion wi h he ubiqui ous da-GAL4 d i e . *deno es
signi ican di e ence om lies o he same sex om he emp y ec o line on he same ch omosome, p < 0.05
(S uden ’s es ).

www.na u e.com/scien i ic epo s/
6
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
eclosed lies
11
. To po en ia e he pheno ype, we included UAS-Dc 2 in he backg ound, so as o inc ease he pene-
ance o RNAi. Wi hou concomi an AOX escue, he esul ing lies showed a se e e locomo o de ec as measu ed
by hei inabili y o climb he walls o he ial, in a s anda d nega i e geo axis assay (Fig.5B). High-le el exp ession
o AOX p oduced, as be o e, a clea escue, whils lowe -le el exp ession using he newly c ea ed ansgenic lines
p oduced only a modes pheno ypic imp o emen (wild- ype AOX), o no imp o emen a all (mu AOX, Fig.5B).
S uc u al conclusions.
Al e na i e oxidases a e membe s o a supe amily o me alloenzymes, cha ac e ized
by a common ca aly ic unc ion o ac i a ion o molecula oxygen, and by common s uc u al elemen s de ining
he ca aly ic dii on cen e, including he ou -helix bundle old and a mo i comp ising wo his idine esidues,
ou ca boxyla e g oups, and a b idging ca boxyla e g oup ac oss he dii on cen e26–28. The c ys al s uc u e o
he ypanosomal enzyme indica es ha i is a homodime wi h each monome comp ising six long and ou
sho α -helices13. The subuni s in e ac wi h each o he ia α -helices 2, 3 and 4, whe eas he hyd ophobic egion
o med by α -helices 1, 2, 4 and 5 is p oposed o ancho he p o ein o he inne su ace o he mi ochond ial inne
memb ane. A se ies o conse ed a ginine esidues, capable o in e ac ing wi h phospholipid head-g oups, may
Figu e 4. Respi ome y o AOX- ans ec ed cells and lies. Oxygen consump ion (% esis an o an imycin,
as de ined in Ma e ials and Me hods) o (A) pe meabilized, ansien ly ans ec ed cells, and (B) homogena es
om male ansgenic lies induced o exp ession using da-GAL4 d i e , exp essing wild- ype (w ) o mu a ed
(mu ) AOX o emp y ec o ( ). The lies had ansgenic inse ions on ch omosome 2. *deno es signi ican
di e ence om ec o -only lies.
Figu e 5. AOX escue o Cox7a de iciency. (A) Su i al (%) om egg o eclosion o lies o he indica ed
geno ypes, all bea ing he da-GAL4 d i e and he CG-9603 knockdown (RNAi) cons uc . Lines es ed
con ained ei he no addi ional ansgene (–), ec o only ( ), wild- ype (w ) o mu a ed AOX (mu ), in each case
on ch omosome 3. (B) Boxplo o climbing index o lies o he indica ed geno ypes. All lies ca ied he ela C155
-GAL4 d i e on ch omosome X plus UAS-Dc 2 wi h o wi hou he CG9603 knockdown (RNAi) cons uc on
ch omosome 2, and he indica ed AOX ansgene on ch omosome 3 (AOX7.1 is he Φ C31- a ge ed inse ion).
Ba s indica e medians, boxes show he i s and hi d qua iles pe cen iles, whiske s a e plo ed acco ding o
he Tukey scheme (K zywinski and Al man, 2014). Signi ican di e ences based on ANOVA a e indica ed by
ho izon al lines (black, ed) deno ing p < 0.05 and 0.001, espec i ely. A single ou lie poin is indica ed by an
open ci cle.
www.na u e.com/scien i ic epo s/
7
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
assis inne memb ane ancho age13. Ou s uc u e modelling o he C. in es inalis AOX sugges s ha he same
s uc u al elemen s a e conse ed in animal AOXs, and ha he enzyme is also a homodime inse ed in o he
mi ochond ial inne memb ane.
In addi ion, he model p edic s ha he ac i e si e, and he e o e he mechanism o oxygen ac i a ion, a e also
conse ed in animal AOXs. The ou -helix bundle, which ac s as a s uc u al pla o m o he binding o he wo
i on a oms, bu ies he ac i e si e deep in a hyd ophobic en i onmen . In T. b ucei AOX, glu ama e esidues 123, 162,
213 and 266, in addi ion o a hyd oxo-b idge, a e esponsible o di ec ly coo dina ing he dii on cen e. The cen e
is u he s abilized by a edox-ac i e y osine esidue29,30, Y220, and wo his idine esidues (H165 and H269),
which a e wi hin hyd ogen-bond dis ances o E123, E169 and E213. The C. in es inalis AOX model indica es ha
he homologous esidues E200, E239, E290, E344, Y297, H242 and H347 o ganize he ac i e si e in he same way.
Func ional conclusions. In heo y, he mu agenesis o a single glu ama e esidue should be enough o des -
abilize he dii on cen e
31
. Howe e , aking ad an age o he p oximi y in he DNA sequence o he codons o
E239 and H242 and o hose o E344 and H347, we we e able o c ea e alanine subs i u ions o ou impo an
ac i e si e esidues simul aneously. Acco ding o ou model, hese mu a ions should dis up i on binding, hus
gene a ing a mu an de oid o ca aly ic ac i i y, wi hou any majo dis u bance o he o e all p o ein s uc u e.
These p edic ions a e suppo ed by he ac ha he mu an and wild- ype p o eins we e exp essed a compa able
le els in mammalian cells and in lies, bu ha no enzyma ic ac i i y could be de ec ed.
Impo an ly, he mu a ed enzyme was unable o escue he o ganismal pheno ypes a ising om enginee ed
cy och ome oxidase de iciency. In heo y, he ac ion o a o eign p o ein in a enua ing such pheno ypes could be
due o any o se e al di e en mechanisms, o which he p o ision o an enzyma ic by-pass o ubiquinol oxida ion
is only one. In p e ious wo k we ound ha Ciona AOX, when exp essed in D osophila mi ochond ia, dec eased
he ne p oduc ion o mi ochond ial ROS e en unde non-inhibi ed condi ions10,32. The mechanism o his emains
unknown, bu one possibili y is ha AOX is able o ac di ec ly o indi ec ly as an an ioxidan , e.g. by binding and
quenching quinone adicals ia some o he mechanism. S udies in a ious o ganisms ha e suppo ed he idea
ha a hyd ophobic pocke , loca ed be ween α -helices 2 and 3, binds and channels ubiquinone o he ac i e si e33,
which migh be in ol ed in such an ac i i y.
A second possibili y would be a ho me ic esponse o dis up ion o he inne mi ochond ial memb ane o
i s p o ein complexes by he o eign p o ein. The induc ion o a a ie y o de ence pa hways o p o ec cells om
inc eased ROS, dis u bed p o ein, lipid o edox homeos asis, o al e ed mi ochond ial u no e o dynamics,
migh equip he o ganism o cope wi h he addi ional bu ela ed s esses o espi a o y insu iciency. Many s ud-
ies in model o ganisms suppo his concep o ‘mi oho mesis’
34
. Whils we canno ule ou ha such e ec s a e
ma e ial in o he con ex s, ou indings do exclude hem in ega d o he de elopmen al le hali y p oduced by
global cy och ome oxidase knockdown, o he locomo o dys unc ion esul ing om i s knockdown speci ically in
neu ons11. Based on ou indings, ha mu AOX canno compensa e hese pheno ypes, we in e ha he escue o
hese e ec s o cy och ome oxidase de iciency by AOX is almos ce ainly due o i s enzyma ic ac i i y as a quinol
oxidase, hough o mally we canno exclude o he , unknown e ec s o i on binding. A equi emen o enzyma ic
ac i i y migh no be ue o e e y pheno ypic ea u e con e ed by AOX in model o ganisms. Ou indings indica e
a obus way o es his in ega d o all po en ial such pheno ypes, allowing he mechanisms by which AOX ac s
o be p obed, con olled o e i ied.
Se e al quan i a i e issues a e also add essed by ou indings. The i s is ha he ex en o pheno ypic escue
depends in some ins ances on he AOX exp ession le el, bu in o he cases, such as he escue o he de elopmen al
le hali y caused by ubiqui ous COX knockdown, is an all-o -none phenomenon. We sugges ha his e lec s a
h eshold e ec whe ein e en he h ee- old lowe exp ession le el o AOX, when in eg a ed a speci ic si es by
Φ C31-media ed ecombina ion (in compa ison wi h P elemen -media ed in eg an s c ea ed p e iously), exceeds
a h eshold alue equi ed o main ain me abolic homeos asis and comple e de elopmen . In con as , he lowe
exp ession le el o he a ge ed in eg an s ga e a clea ly weake escue o locomo o dys unc ion, when COX was
knocked down only in neu ons, oughly in p opo ion o he dec eased exp ession le el.
I may also be no ed ha he amoun o an imycin- esis ance con e ed upon espi a ion in homogena es om
he a ge ed in eg an s was s ill app oxima ely 14%, compa ed wi h app oxima ely 20% o he P elemen -media ed
in eg an s, e en hough hey a e exp essed a a much highe le el. The le el o espi a o y an imycin- esis ance in
he ly may a y be ween issues, and his 20% maximum may e lec only he p ope ies o he p edominan class
o mi ochond ia. Mos o he espi a o y capaci y in adul lies is es ed in he ligh muscles, whe e mi ochond ia
make up almos one- hi d o he o al issue mass
35
. The appa en uppe limi o how much elec on low can be
di e ed h ough AOX p obably e lec s speci ic ea u es o his issue and i s ene ge ic needs. The limi could be
dic a ed by he cons ain s o memb ane a chi ec u e, o example, i much o he ubiquinone pool is channelled
di ec ly om complex I o complex III ia espi a o y supe complexes, such ha i equilib a es only slowly wi h
ee ubiquinones a ailable o AOX36. Mos o he espi a o y ac i i y in adul D osophila indeed esides in supe -
complexes37. Such a phenomenon may accoun o he in e ed h eshold e ec on he escue o de elopmen al
le hali y. Con e sely, he o ganiza ion o he espi a o y chain may di e in o he issues, such as in neu ons, whe e
a mo e g aded esponse o he AOX exp ession le el is e iden .
In conclusion, mu AOX o e s a use ul ool o u u e s udies o he mechanism(s) whe eby exp ession o
Ciona AOX modi ies he pheno ypes o model o ganisms, po en ially con ibu ing he e en ual de elopmen o
AOX-based he apies.
Re e ences
1.
McDonald, A. E., Vanle be ghe, G. C. & S aples, J. F. Al e na i e oxidase in animals: unique cha ac e is ics and axonomic dis ibu ion.
J. Exp. Biol. 212, 2627–2634 (2009).
2.
Feng, H. e al. Exp ession and signal egula ion o he al e na i e oxidase genes unde abio ic s esses. Ac a. Biochim. Biophys. Sin.
45, 985–994 (2013).
www.na u e.com/scien i ic epo s/
8
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
3.
Vanle be ghe, G. C. Al e na i e oxidase: a mi ochond ial espi a o y pa hway o main ain me abolic and signaling homeos asis
du ing abio ic and bio ic s ess in plan s. In . J. Mol. Sci. 14, 6805–6847 (2013).
4.
Hoe nagel, M. H. & Wiskich, J. T. Ac i a ion o he plan al e na i e oxidase by high educ ion le els o he Q-pool and py u a e.
A ch. Biochem. Biophys. 355, 262–270 (1998).
5.
Cas o-Gue e o, N. A., K ab, K. & Mo eno-Sánchez, R. The al e na i e espi a o y pa hway o euglena mi ochond ia. J Bioene g.
Biomemb . 36, 459–469 (2004).
6.
Rus in, P. & Jacobs, H. T. Respi a o y chain al e na i e enzymes as ools o be e unde s and and coun e ac espi a o y chain
de iciencies in human cells and animals. Physiol. Plan 137, 362–370 (2009).
7.
El-Khou y, R. e al. Enginee ing he al e na i e oxidase gene o be e unde s and and coun e ac mi ochond ial de ec s: s a e o he
a and pe spec i es. B . J. Pha macol. 171, 2243–2249 (2014).
8.
Hakkaa , A., Dassa, E. P., Jacobs, H. T. & Rus in, P. Allo opic exp ession o a mi ochond ial al e na i e oxidase con e s cyanide
esis ance o human cell espi a ion. EMBO Rep. 7, 341–345 (2006).
9.
Dassa, E. P. e al. Exp ession o he al e na i e oxidase complemen s cy och ome c oxidase de iciency in human cells. EMBO Mol.
Med. 1, 30–36 (2009).
10.
Fe nandez-Ayala, D. J. e al. Exp ession o he Ciona in es inalis al e na i e oxidase (AOX) in D osophila complemen s de ec s in
mi ochond ial oxida i e phospho yla ion. Cell Me ab. 9, 449–460 (2009).
11.
Kemppainen, K. K. e al. Exp ession o al e na i e oxidase in D osophila amelio a es di e se pheno ypes due o cy och ome oxidase
de iciency. Hum. Mol. Gene . 23, 2078–2093 (2014).
12.
El-Khou y, R. e al. Al e na i e oxidase exp ession in he mouse enables bypassing cy och ome c oxidase blockade and limi s
mi ochond ial ROS o e p oduc ion. PLoS Gene . 9, e1003182 (2013).
13.
Shiba, T. e al. S uc u e o he ypanosome cyanide-insensi i e al e na i e oxidase. P oc. Na l. Acad. Sci. USA 110, 4580–4585 (2013).
14. Tamu a, K., S eche , G., Pe e son, D., Filipski, A. & Kuma , S. MEGA6: Molecula E olu iona y Gene ics Analysis e sion 6.0. Mol.
Biol. E ol. 30, 2725–2729 (2013).
15.
Bazzoli, A., Te amanzi, A. G. & Zhang, Y. Compu a ional p o ein design and la ge-scale assessmen by I-TASSER s uc u e assembly
simula ions. J Mol. Biol. 407, 764–776 (2011).
16.
Bischo , J., Maeda, R. K., Hedige , M., Ka ch, F. & Besle , K. An op imized ansgenesis sys em o D osophila using ge m-line-speci ic
ϕ C31 in eg ases. P oc. Na l. Acad. Sci. USA 104, 3312–3317 (2007).
17. Fukuoh, A. e al. 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 (2014).
18.
Niwa, H., Yamamu a, K. & Miyazaki, J. E icien selec ion o high-exp ession ans ec an s wi h a no el euka yo ic ec o . Gene 108,
193–199 (1991).
19.
Sanz, A. e al. Exp ession o he yeas NADH dehyd ogenase Ndi1 in D osophila con e s inc eased li espan independen ly o die a y
es ic ion. P oc. Na l. Acad. Sci. USA 107, 9105–9110 (2010).
20. P ei e , B. D. e al. Re inemen o ools o a ge ed gene exp ession in D osophila. Gene ics 186, 735–755 (2010).
21.
Cannino, G. e al. Glucose modula es espi a o y complex I ac i i y in esponse o acu e mi ochond ial dys unc ion. J Biol. Chem.
287, 38729–38740 (2012).
22. Ch e ien, D. e al. Re e ence cha s o espi a o y chain ac i i ies in human issues. Clin. Chim. Ac a. 228, 53–70 (1994).
23. Toi onen, J. M. e al. echnical knockou , a D osophila model o mi ochond ial dea ness. Gene ics 159, 241–254 (2001).
24. K zywinski, M. & Al man, N. Poin s o Signi icance: Visualizing samples wi h box plo s. Na u e Me hods 11, 119–120 (2014).
25. G een, E. W., Fedele, G., Gio gini, F. & Ky iacou, C. P. A D osophila RNAi collec ion is subjec o dominan pheno ypic e ec s. Na .
Me hods 11, 222–223 (2014).
26.
Be hold, D. A., Voe odskaya, N., S enma k, P., G äslund, A. & No dlund, P. EPR s udies o he mi ochond ial al e na i e oxidase.
E idence o a dii on ca boxyla e cen e . J Biol. Chem. 277, 43608–43614 (2002).
27. Be hold, D. A. & S enma k, P. Memb ane-bound di-i on ca boxyla e p o eins. Annu. Re . Plan Biol. 54, 497–517 (2003).
28.
Simone, F., Reisne , E. & Lippa d, S. J. Cu en challenges o modeling dii on enzyme ac i e si es o dioxygen ac i a ion by biomime ic
syn he ic complexes. Chem. Soc. Re . 39, 2768–2779 (2010).
29.
Albu y, M. S., A ou i , C., C ich on, P. G. & Moo e, A. L. S uc u e o he plan al e na i e oxidase. Si e-di ec ed mu agenesis p o ides
new in o ma ion on he ac i e si e and memb ane opology. J Biol. Chem. 277, 1190–1194 (2002).
30.
A ou i , C., Albu y, M. S., C ich on, P. G. & Moo e, A. L. Explo ing he molecula na u e o al e na i e oxidase egula ion and
ca alysis. FEBS Le . 510, 121–126 (2002).
31. Ajayi, W. U., Chaudhu i, M. & Hill, G. C. Si e-di ec ed mu agenesis e eals he essen iali y o he conse ed esidues in he pu a i e
dii on ac i e si e o he ypanosome al e na i e oxidase. J Biol. Chem. 277, 8187–8193 (2002).
32.
Sanz, A., Fe nández-Ayala, D. J., S e ana os, R. K. & Jacobs, H. T. Mi ochond ial ROS p oduc ion co ela es wi h, bu does no di ec ly
egula e li espan in D osophila. Aging 2, 200–223 (2010).
33.
Albu y, M. S., Ellio , C. & Moo e, A. L. Towa ds a s uc u al elucida ion o he al e na i e oxidase in plan s. Physiol. Plan 137,
316–327 (2009).
34. Yun, J. & Finkel, T. Mi oho mesis. Cell Me ab. 19, 757–766 (2014).
35.
Le enbook, L. & Williams, C. M. Mi ochond ia in he ligh muscles o insec s III. Mi ochond ial cy och ome c in ela ion o he
aging and wing bea equency o lies. J. Gen. Physiol. 39, 497–512 (1956).
36.
Geno a, M. L. & Lenaz, G. Func ional ole o mi ochond ial espi a o y supe complexes. Biochim. Biophys. Ac a. 1837, 427–443
(2014).
37.
Celo o, A. M., Chiu, W. K., Van Voo hies, W. & Palladino, M. J. Modes o me abolic compensa ion du ing mi ochond ial disease
using he D osophila model o ATP6 dys unc ion. PLoS One 6, e25823 (2011).
Acknowledgemen s
We hank Tony Moo e o use ul discussions, Filippo Scialo o he cons uc ion o he o iginal AOX plasmid o
exp ession in S2 cells, Dmi o Gospoda yo o c i ical eading o he manusc ip and Samuli Ha ikainen, E eliina
Kaulio, Tea Tuomela, Essi Ki i an a, Ou i Ku onen, Me ja Jokela and Maa i Myöhänen o echnical assis ance.
Funding was p o ided by Academy o Finland (CoE g an 272376), he Eu opean Resea ch Council (ad anced g an
232738 o HTJ), he EU (Ma ie Cu ie In e na ional Incoming Fellowship 328988 o MTO), Tampe e Uni e si y
Hospi al Medical Resea ch Fund, and he Sig id Juselius Founda ion.
Au ho Con ibu ions
A.A., M.T.O., H.T.J. and P.R. concei ed and planned he p ojec . A.A., M.T.O., G.C., C.Y. and P.K.D. conduc ed
he labo a o y wo k and analysis. H.T.J., M.S. and E.D. supe ised he labo a o y wo k and con ibu ed analysis
and insigh s. H.T.J. and M.T.O. compiled he igu es and d a ed he manusc ip .
Addi ional In o ma ion
Supplemen a y in o ma ion accompanies his pape a h p://www.na u e.com/s ep
www.na u e.com/scien i ic epo s/
9
Scien i ic RepoR s | 5:18295 | DOI: 10.1038/s ep18295
Compe ing inancial in e es s: The au ho s decla e no compe ing inancial in e es s.
How o ci e his a icle: Andjelko ić, A. e al. Dii on cen e mu a ions in Ciona in es inalis al e na i e oxidase
abolish enzyma ic ac i i y and p e en escue o cy och ome oxidase de iciency in lies. Sci. Rep. 5, 18295; doi:
10.1038/s ep18295 (2015).
This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. The images
o o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons license,
unless indica ed o he wise in he c edi line; i he ma e ial is no included unde he C ea i e Commons license,
use s will need o ob ain pe mission om he license holde o ep oduce he ma e ial. To iew a copy o his
license, isi h p://c ea i ecommons.o g/licenses/by/4.0/