5732–5742 Nucleic Acids Resea ch, 2016, Vol. 44, No. 12 Published online 25 Ap il 2016
doi: 10.1093/na /gkw302
Mi ochond ial ansc ip ion e mina ion ac o 1
di ec s pola eplica ion o k pausing
Yonghong Shi1,2,†, Vik o Posse1,†, Xue eng Zhu1,2, Anne K. Hy ¨
a inen3,Howa d
T. Jacobs3,4, Ma ia Falkenbe g1,* and Claes M. Gus a sson1,*
1Ins i u e o Biomedicine, Uni e si y o Go henbu g, P.O. Box 440, SE-405 30 Go henbu g, Sweden, 2Cen e o
Molecula Medicine, Na ional Hea Lung and Blood Ins i u e, NIH, Be hesda, MD 20892, USA, 3BioMediTech and
Tampe e Uni e si y Hospi al, FI-33014, Uni e si y o Tampe e, Finland and 4Ins i u e o Bio echnology, FI-00014,
Uni e si y o Helsinki, Finland
Recei ed Feb ua y 17, 2016; Re ised Ap il 11, 2016; Accep ed Ap il 12, 2016
ABSTRACT
Du ing eplica ion o nuclea ibosomal DNA ( DNA),
clashes wi h he ansc ip ion appa a us can cause
eplica ion o k collapse and genomic ins abili y. To
a oid his p oblem, a eplica ion o k ba ie p o-
ein is si ua ed downs eam o DNA, he e p e en -
ing eplica ion in he di ec ion opposi e DNA an-
sc ip ion. A po en ial candida e o a simila unc-
ion in mi ochond ia is he mi ochond ial ansc ip-
ion e mina ion ac o 1 (MTERF1, also deno ed
mTERF), which binds o a sequence jus downs eam
o he ibosomal ansc ip ion uni . P e ious s udies
ha e shown ha MTERF1 p e en s an isense an-
sc ip ion o e he ibosomal RNA genes, a p ocess
which we he e show o be independen o he an-
sc ip ion elonga ion ac o TEFM. Impo an ly, we
now demons a e ha MTERF1 a es s mi ochond ial
DNA (m DNA) eplica ion wi h dis inc pola i y. The
e ec is explained by he abili y o MTERF1 o ac as a
di ec ional con ahelicase, blocking m DNA unwind-
ing by he mi ochond ial helicase TWINKLE. This
conclusion is also suppo ed by
in i o
e idence ha
MTERF1 s imula es TWINKLE pausing. We conclude
ha MTERF1 can di ec pola eplica ion o k a es
in mammalian mi ochond ia.
INTRODUCTION
Human mi ochond ial DNA (m DNA) is a ci cula ,
double-s anded molecule. The wo s ands di e in hei
buoyan densi y in alkaline cesium chlo ide g adien s and
hey a e he e o e called he ligh s and (L-s and) and he
hea y s and (H-s and), espec i ely (1). Human m DNA
encodes 13 componen s o he espi a o y chain. The
double-s anded genome also p oduces 22 RNAs and 2
RNAs needed o mi ochond ial ansla ion. T ansc ip-
ion is ini ia ed om he hea y- and ligh -s and p omo -
e s (HSP and LSP), which gene a e polycis onic, nea ull
genome-sized ansc ip s co e ing he wo s ands. The p i-
ma y ansc ip s a e hen p ocessed o p oduce indi idual
RNA molecules (2).
The molecula machine ies esponsible o ansc ip ion
and eplica ion o mammalian m DNA a e dis inc om
hose ound in he nucleus. Fo ins ance, DNA syn he-
sis is pe o med by he he e o ime ic DNA polyme ase
␥(POL␥), which consis s o a ca aly ic subuni (POL␥A)
and wo iden ical accesso y subuni s (POL␥B) (3,4). The
eplica i e m DNA helicase TWINKLE is a homohexam-
e ic p o ein complex wi h simila i y o he bac e iophage
T7 p imase/helicase gene 4 p o ein (T7 gp4), bu wi h-
ou he p imase ac i i y p esen in he phage p o ein (5–
8). A mi ochond ial single-s anded DNA binding p o ein
(m SSB) p o ec s he single-s anded pa en al s and du -
ing m DNA eplica ion and s imula es bo h POL␥and
TWINKLE ac i i ies (6,9–12). Mi ochond ia also con-
ain a specialized ansc ip ion machine y, which include a
monome ic RNA polyme ase (POLRMT) and wo acces-
so y ac o s, mi ochond ial ansc ip ion ac o s A and B2
(13–16). The e is also a mi ochond ial ansc ip ion elon-
ga ion ac o (TEFM), which helps POLRMT o ansc ibe
longe s e ches o RNA, and o bypass egions gene a ing
highly s uc u ed RNA (17–19). In addi ion o p oducing
mRNA, RNA and RNA, he mi ochond ial ansc ip ion
machine y also gene a es p ime s o ini ia ion o m DNA
eplica ion a he majo o igins o he wo s ands (20).
Fo a long ime, i was belie ed ha mi ochond ial an-
sc ip ion is egula ed a he le el o e mina ion. This
idea was based on he iden i ica ion o he MTERF1 ( o -
me ly deno ed as mTERF), a p o ein ha binds sequence-
speci ically o a 28-bp egion immedia ely downs eam o
*To whom co espondence should be add essed. Tel: +46 31 7863826; Fax: +46 31 416108; Email: claes[email p o ec ed]
Co espondence may also be add essed o Ma ia Falkenbe g. Tel: +46 31 7863444; Fax: +46 31 416108; Email: ma ia. alkenbe [email protected]
†These au ho s con ibu ed equally o he wo k as i s au ho s.
C
The Au ho (s) 2016. Published by Ox o d Uni e si y P ess on behal o Nucleic Acids Resea ch.
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-nc/4.0/), which
pe mi s non-comme cial e-use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. Fo comme cial e-use, please con ac
jou nals.pe [email protected]
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
Nucleic Acids Resea ch, 2016, Vol. 44, No. 12 5733
he 16S RNA gene (21,22). Acco ding o he belie a he
ime, MTERF1 egula ed he le els o hea y-s and an-
sc ip ion ha p oceeded beyond i s binding si e, a model
ha was pu o wa d o explain why RNA ansc ip s we e
almos 50 imes mo e abundan han he mRNA ansc ip s
p oduced downs eam o he MTERF1 binding si e (23).
Mo e ecen ly, in i o analysis o cul u ed human cells (24)
and he M e 1−/−knockou mouse e u ed his idea (25).
Loss o MTERF1 does no a ec he ela i e s eady-s a e
le els o RNAs and mRNA encoded on he H-s and. In-
s ead, hese ela i e RNA le els a e mos likely explained
by di e ences in s abili y, wi h RNA ha ing a much longe
hal -li e. S udies o he M e 1−/−knockou mouse model
ins ead e ealed ha M e 1 p e en s an isense ansc ip-
ion om en e ing he RNA gene egion. In he absence
o M e 1 an isense ansc ip ion p og esses beyond he
RNA genes and may also in e e e wi h ligh -s and p o-
mo e (LSP) unc ion. These in i o indings ag ee nicely
wi h he e ec s o MTERF1 on ansc ip ion e mina ion
in i o, which ound ha he p o ein comple ely blocks
L-s and ansc ip ion, bu only has a mino e ec on H-
s and ansc ip ion (26,27).
Acco ding o he s and-asymme ic model o m DNA
eplica ion, DNA syn hesis is con inuous on bo h s ands
and each s and con ains a dis inc , majo o igin o
eplica ion, O iH and O iL (1,28). Replica ion is s and-
asynch onous, i.e. syn hesis o he H-s and is i s ini-
ia ed a O iH and he eac ion p oceeds app oxima ely
wo hi ds o he way a ound he ci cula genome, displac-
ing he non- empla e H-s and, which becomes ansien ly
complexed wi h m SSB (29). Two-dimensional aga ose gel-
elec opho esis has also iden i ied he p esence o RNA
annealed o he displaced, non- empla e H-s and DNA
(30,31). I has been p oposed ha hese RNA species can
eplace m SSB and s abilize ssDNA du ing eplica ion, bu
his emains a deba ed issue (29,32). L-s and DNA syn he-
sis is ini ia ed a O iL by a p ocess ha is belie ed o in ol e
ac i a ion o he o igin when i becomes single-s anded.
The main ea u es o his s and-asymme ic eplica ion
mode a e suppo ed by a numbe o di e en obse a ions,
including elec on mic oscopy analysis (28), biochemical
analysis o eplica ion in e media es in bo h wild- ype cells
and issues, and whe e exp ession o key eplica ion ac o s
has been manipula ed (33–35), in i o sa u a ion mu agen-
esis (36), he in i o occupancy pa e n o m SSB (29), and
he possibili y o econs i u e speci ic s eps o his p ocess
in i o (12,20,37).
How DNA eplica ion and ansc ip ion a e coo dina ed
in human mi ochond ia is no known. Un egula ed clashes
be ween eplica ion and ansc ip ion may a ec gene an-
sc ip ion and/o impai DNA eplica ion (38). To a oid
such p oblems in nuclea DNA eplica ion, he mo ing
eplica ion o k can be a es ed close o he 3-end o highly
ansc ibed ibosomal DNA ( DNA). A eplica ion o k
ba ie p e en s DNA eplica ion om p og essing in a di-
ec ion opposi e ha o DNA ansc ip ion, which sup-
posedly p e en s head-on collisions be ween eplica ion and
ansc ip ion machine ies (38). In mouse, he mTTF-1 p o-
ein binds o Sal boxes si ua ed jus downs eam o he i-
bosomal ansc ip ion uni s and block eplica ion wi h op-
posi e pola i y o ha o ibosomal ansc ip ion (39). In
budding yeas , he Fob1 p o ein ensu es ha DNA epli-
ca ion only occu s in he di ec ion o ansc ip ion and ha
he eplica ion machine y does no dis u b ansc ip ion
o DNA (40). I a simila eplica ion o k ba ie p o ein
also exis s in mi ochond ia is no known, bu he bind-
ing o MTERF1 jus downs eam o he 3-end o DNA
makes i an in e es ing candida e o such a unc ion. In
suppo o his idea, a p e ious epo demons a ed ha
o e exp ession o MTERF1 s imula es eplica ion pausing
a MTERF1 binding si es in i o. The exp ession le els o
MTERF1 co ela ed wi h he s eng h o he pausing and
i was hypo hesized ha MTERF1 ac s o coo dina e he
passage o eplica ion and ansc ip ion complexes (41). In
his epo , we use in i o biochemis y o demons a e
ha MTERF1 ac s as a eplica ion o k ba ie and de-
lays eplica ion o k p og ession. The e ec is explained by
MTERF1’s abili y o wo k as a con ahelicase wi h dis inc
pola i y.
MATERIALS AND METHODS
Recombinan p o eins
TWINKLE, m SSB, POL␥A, POL␥B, and MTERF1 we e
exp essed and pu i ied as desc ibed p e iously (6,12,27,42).
POLRMT, TFAM, TFB2M and TEFM we e exp essed and
pu i ied as desc ibed in (19).
In i o ansc ip ion expe imen s
The HSP and LSP ansc ip ion empla es we e cloned as
desc ibed in (13) and linea ized wi h Ps I and EcoRI e-
spec i ely. The ansc ip ion eac ion olumes we e 25 ␮l
and con ained 25 mM T is-HCl pH 8.0, 10 mM MgCl2,40
mM NaCl, 100 ␮g/mL BSA, 10 mM DTT, 400 ␮MATP,
150 ␮M GTP, 150 ␮MCTP10␮M UTP, 0,02 ␮M␣-32P
UTP (3000 Ci/mmol), 4 U RNase inhibi o Mu ine (New
England Biolabs), 4 nM o indica ed plasmid empla e. The
eac ions con ained POLRMT (20 nM), TFAM (200 nM),
and TFB2M (60 nM). TEFM (40 nM) and MTERF1 (2,
8 and 32 nM) we e added as indica ed. The eac ions we e
s opped and analyzed on 4% dena u a ing polyac ylamide
gels as desc ibed p e iously (13).
Templa es p epa a ion
We cloned a DNA agmen co esponding o n 3056–
3559 o he mi ochond ial human genome con aining he
MTERF1 binding si e (n 3231–3253) be ween he HindIII
and EcoRI si es in he pBluesc ip SK(+) and pBluesc ip
SK(−) ec o s (Agilen Technologies; La Jolla, CA, USA).
The pBluesc ip SK(−) MTERF1 cons uc was used as
a empla e o si e-di ec ed PCR mu agenesis eac ions o
gene a e a mu an a ian o he MTERF1 binding si e con-
aining a 4-bp dele ion (n 3241–3244). All cons uc s we e
con i med by sequencing and used o isola e ssDNA ol-
lowing he manu ac u e ’s p o ocol (S a agene). To p o-
duce he olling-ci cle DNA eplica ion empla es, we an-
nealed a 70-me oligonucleo ide (5-40[T]- AGT TAC CAA
TGC TTA ATC AGT GAG GCA CCT-3) o he pBlue-
sc ip SK(+) and (5-40[T]-TCG CTG AGA TAG GTG
CCT CAC TGA TTA AGC-3’) o he pBluesc ip SK(−)
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
5734 Nucleic Acids Resea ch, 2016, Vol. 44, No. 12
ssDNA (7.5 pmol) and syn hesized he second s and as de-
sc ibed p e iously (20).
To p epa e he helicase subs a es indica ed in he igu e
legends, h ee di e en oligonucleo ides; Helicase (WT) o -
wa d; TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT
TTT TTT TTT AAG ATG GCA GAG CCC GGT AAT,
Helicase (WT) e e se; TTT TTT TTT TTT TTT TTT TTT
TTT TTT TTT TTT TTT TTT TTT ATT ACC GGG CTC
TGC CAT CTT A, o Helicase (del) e e se; TTT TTT TTT
TTT TTT TTT TTT TTT TTT TTT TTT TTT TTT TCG
ATT ACC GGG GCC ATC TTA AC we e labeled wi h
32P a he 5- e mini using T4 polynucleo ide kinase (S a a-
gene) and annealed o ssDNA (desc ibed abo e) con aining
he co esponding DNA sequence c ea ing he helicase sub-
s a es indica ed in he igu e legends. A e annealing, he
samples named SK+Te 1, SK-Te 1, and SK- TERF1 del
we e un h ough Cen icon 100 columns (Millipo e Com-
pany) o emo e unannealed oligonucleo ide as p e iously
desc ibed (6).
Rolling-ci cle m DNA eplica ion
The eac ion mix u es (20 ␮l) con ained 10 mol o he in-
dica ed dsDNA empla e and we e pe o med as desc ibed
p e iously bu in he p esence o MTERF1 when indica ed
in he igu e legends. The eac ions we e incuba ed a 37◦C
and s opped a e 40 min (o a he imes indica ed) by
adding 200 ␮l o ansc ip ion s op bu e , 0.5% SDS and
2␮l o p o einase K. A e ano he 45 min incuba ion a
42◦C, he samples we e p ecipi a ed and analyzed as de-
sc ibed be o e (12).
Helicase assays
DNA empla es o TWINKLE helicase assays we e p e-
pa ed as desc ibed abo e. The eac ion mix u e (25 ␮l) con-
ained 20 mol o he indica ed DNA subs a e, 20 mM T is–
Cl (pH 7.6), 4.5 mM MgCl2, 10 mM DTT, 3 mM UTP, 0.1
mg/ml BSA, 40 mM NaCl, 600 mol TWINKLE o 100
mol T7 gp4 p o ein and 500 mol o MTERF1 i no hing
else is indica ed in he igu e legends. The eac ions we e in-
cuba ed a 32◦C o 20 min (o as indica ed) and s opped by
he addi ion o 2 ␮l s op bu e (90 mM EDTA, 6% SDS,
30% glyce ol, and 0.25% b omphenol blue), and analyzed
on a 6% non-dena u ing polyac ylamide gel.
Ch oma in immunop ecipi a ion analysis
MTERF1-MycHis exp essing Flp-InTM T-RexTM-293
cells (41) we e cul u ed in DMEM medium wi h 5% e al
bo ine se um, hyg omycin B(50␮g/ml) and blas icidin-HCl
(15 ␮g/ml) in 10 ×75 cm2 issue cul u e lasks o 80% con-
luence. O hese, 5 ×75 cm2 lasks cells we e induced o ex-
p ess he Myc- agged mTERF ansgene by addi ion o 100
ng/ml doxycycline, whe eas he o he 5 ×75 cm2 lasks we e
kep as a non-induced con ol. A e 24 h induc ion, he
cells we e ha es ed and washed once wi h ice-cold phos-
pha e bu e ed saline (PBS). Isola ion o mi ochond ia was
ca ied ou ollowing a p o ocol desc ibed in (43). The mi o-
chond ia we e washed once wi h ice-cold PBS, and nex in-
cuba ed wi h 1% o maldehyde in PBS o 10 min a . The
c osslinking eac ion was quenched by addi ion o glycine
( inal conc. 125 mM) and hen incuba ed o i e addi ional
minu es. A e washing wice in ice-cold PBS, mi ochond ia
we e lysed in 25 mM HEPES-KOH (pH 7.6), 10% glyce ol,
5mMMgCl
2, 0.5 mM EDTA, 0.5% ween-20, 0.15 M KCl,
1 mM phenylme hlsul onyl luo ide, 2 mM peps a in A, 0.6
mM leupep in, and 2 mM benzamidine. The mi ochond ial
lysa es we e sonica ed in a Bio up o UCD 200TM (Di-
agenode) o 10 min a high ou pu , wi h in e als o 30 s
on and 30 s o , and hen cen i uged o 5 min a 14 000 ×
g. Aliquo s o 100 ␮l w and induced supe na an we e in-
cuba ed wi h 1.5 ␮l o human TWINKLE polyclonal an i-
body (Ag ise a, Sweden) o 1.5 ␮l abbi IgG (ab37415, Ab-
cam) o e nigh in a o a o a 4◦C. 50 ␮l o p o ein A beads
(GE heal hca e) we e added o supe na an s and he indi-
ca ed an ibody o 1 h a 4◦C. A e wash and elu ion, sam-
ples we e incuba ed o e nigh a 65◦C o e e se c osslink-
ing. RNA con amina ions we e emo ed by incuba ion wi h
100 ng/ml RNaseA o 15 min a 37◦C, and p o eins we e
emo ed by addi ion o 20 ␮g p o einase K and incuba ion
o 2 h a 56◦C. DNA was pu i ied by phenol/chlo o o m
ex ac ion, ollowed by e hanol p ecipi a ion. The pu i ied
DNA was used o eal- ime PCR analysis (Bio-Rad) wi h
he p ime pai s lis ed in Supplemen a y Table S1. Quan-
i ica ions we e pe o med using eal ime PCR So wa e
(Bio-Rad) and Excel (Mic oso ); a ios o IP/inpu a e de-
pic ed in he igu es a e sub ac ing a ios ob ained om
he abbi IgG con ol. Th ee independen biological epli-
ca es we e ca ied ou o ChiP analysis. A e ages and s an-
da d de ia ions we e calcula ed and plo ed in Mic oso
Excel.
RESULTS
MTERF1 e ec i ely e mina es ansc ip ion in he p esence
o TEFM
The ansc ip ion elonga ion ac o TEFM was jus e-
cen ly iden i ied (17–19). Be o e p oceeding o s udies o
MTERF1 and DNA eplica ion, we decided o in es iga e i
he p esence o TEFM a ec ed MTERF1 ac i i ies in i o
and conclusions d awn om hese p e ious s udies (27).
In mi ochond ia, ansc ip ion can app oach he MTERF1
binding si e om wo di e en di ec ions. To mimic each o
hese si ua ions, we used linea ized DNA empla es wi h ei-
he he HSP o LSP p omo e , ollowed by he MTERF1
binding si e o ien ed in he same di ec ion ela i e he p o-
mo e as ha obse ed in i o (Figu e 1A). In ou analysis,
we added inc easing concen a ions o MTERF1 o an-
sc ip ion eac ions con aining POLRMT, TFAM, TFB2M,
and he DNA empla e. Using he LSP empla e in he ab-
sence o MTERF1 we obse ed syn hesis o he expec ed
ull-leng h un-o ansc ip s (Figu e 1B, lane 1, LSP RO)
and also a p e iously desc ibed p e- e mina ed ansc ip
a CSBII (CSBII PT). When we added inc easing amoun s
o MTERF1, we obse ed he appea ance o an addi ional
ansc ip wi h he expec ed size o an RNA ansc ip p e-
ma u ely e mina ed a he MTERF1-binding si e (Figu e
1B, lanes 2–4, MTERF1 PT). In ag eemen wi h p e ious
epo s, addi ion o TEFM s imula ed ansc ip ion and
abolished ansc ip ion e mina ion a CSBII. Howe e , i
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
Nucleic Acids Resea ch, 2016, Vol. 44, No. 12 5735
MTERF1 0
-TEFM
000
+TEFM -TEFM +TEFM
CSBII PT
1 2 3 4 5 6 7 8 9 10111213141516
LSP HSP
HSP RO
MTERF1 PT
LSP RO
MTERF1 PT
MTERF1
LSP MTERF1
Runo (RO)
P e e m CSBII (CSBII PT)
Runo (RO)
P e e m MTERF1 (MTERF1 PT)
HSP
P e e m MTERF1 (MTERF1 PT)
A
B
Figu e 1. MTERF1 e ec s on ansc ip ion. (A) A schema ic desc ip ion o he empla es used. The a ows indica es he di ec ion o he MTERF1
binding si e ela i e he ansc ip ion machine y. (B)In i o ansc ip ion wi h he ecombinan ansc ip ion machine y (POLRMT, TFAM and TFB2M)
on MTERF1 binding si e-con aining empla es, wi h ansc ip ion app oaching MTERF1 ei he in he LSP di ec ion (lanes 1–8) o he HSP di ec ion
(lanes 9–16). The espec i e p omo e has been used in hese expe imen s (LSP o lanes 1–8 and HSP o lanes 9–16). The bands co esponding he uno
p oduc s (LSP RO and HSP RO), as well as he p ema u e ansc ip ion e mina ions a CSBII (CSBII PT) and he MTERF1 binding si e (MTERF1 PT)
a e indica ed. TFEM was added o eac ions in lanes 5–8 o LSP and lanes 13–16 o HSP.
did no a ec MTERF1-dependen ansc ip ion e mina-
ion (Figu e 1B, lanes 5–8).
We nex analyzed e ec s on he HSP empla e. In he ab-
sence o MTERF1 we obse ed syn hesis o he expec ed
ull-leng h un-o ansc ip s (Figu e 1B, lane 9, HSP RO).
Addi ion o inc easing concen a ions o MTERF1 was
much less e ec i e in e mina ing ansc ip ion in his o i-
en a ion and we only obse ed he appea ance o a e y
weak band a he expec ed size o an RNA ansc ip e -
mina ed a he MTERF1-binding si e (Figu e 1B, lanes
10–12, MTERF1 PT). Addi ion o TEFM abolished his
weak band (Figu e 1B, lanes 13–16). F om ou obse a-
ions, we conclude ha MTERF1 e mina es ansc ip ion
ini ia ed om LSP also in he p esence o TEFM. In con-
as , TEFM does dec ease he al eady e y weak MTERF1
e mina ion ac i i y on ansc ip ion ini ia ed om HSP.
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
5736 Nucleic Acids Resea ch, 2016, Vol. 44, No. 12
MTERF1 pauses m DNA eplica ion in an o ien a ion-
dependen manne in i o
We nex add essed he e ec s o MTERF1 on m DNA
eplica ion. The pu i ied POL␥, m SSB, and TWINKLE
p o eins oge he o m a minimal mi ochond ial eplisome,
which in olling ci cle eplica ion can suppo syn hesis o
single-s anded DNA s e ches o mo e han 20 000 n (12).
We cons uc ed a DNA empla e ha con ained a epli-
ca ion o k on which he eplica ion machine y could be
loaded. The empla e also con ained an app oxima ely 3.5
kb long dsDNA egion and a ee 3-end ha can p ime
leading-s and DNA syn hesis. On his ype o ci cula em-
pla e, leading-s and DNA syn hesis coupled o con inu-
ous unwinding o he double-s anded empla e can, in
p inciple, p og ess inde ini ely. To moni o he e ec s o
MTERF1, we in oduced a 38-bp sequence co e ing he
MTERF1 egion, in wo di e en o ien a ions (Figu e 2A).
Fi s , we incuba ed he ci cula empla es wi h he com-
ponen s o he m DNA eplica ion machine y and moni-
o ed DNA syn hesis using adiolabeled dNTPs in a olling-
ci cle expe imen (Figu e 2B). To es i MTERF1 has
a eplica ion o k-ba ie ac i i y, we added inc easing
amoun s o MTERF1 and moni o ed e ec s on DNA syn-
hesis. When he MTERF1 si e was posi ioned in he o -
wa d o ien a ion ( eplica ion p og essing in he same o i-
en a ion as RNA ansc ip ion ini ia ed a HSP), we could
no de ec a signi ican e ec on olling-ci cle eplica ion
(Figu e 2B, lanes 6–10). In con as , when he MTERF1
si e was in he e e se o ien a ion ( eplica ion p og essing
in he opposi e o ien a ion o RNA ansc ip ion, i.e. an-
isense ansc ip ion ini ia ed a LSP), we obse ed a d a-
ma ic dec ease in DNA syn hesis (Figu e 2B, lanes 1–5).
These esul s indica ed ha MTERF1 blocks olling-ci cle
DNA eplica ion in a di ec ion-dependen manne .
Addi ion o MTERF1 no only led o inhibi ion o DNA
eplica ion, bu also o he o ma ion o eplica ion p od-
uc s o de ined leng hs (e.g. Figu e 2B, lane 5). To u he
elucida e his e ec , we pe o med a ime-cou se expe i-
men . We used he olling-ci cle empla e wi h he MTERF1
si e o ien ed in he e e se di ec ion. We incuba ed his em-
pla e wi h ou econs i u ed in i o DNA eplica ion sys em
in he p esence o cons an amoun s o MTERF1 (Figu e
2C). In he absence o MTERF1, he eplica ion p oduc s
o med a con inuous smea (Figu e 2C, lanes 1–4). In he
p esence o MTERF1 p o ein, we obse ed he o ma ion
o speci ic-sized eplica ion p oduc s (Figu e 2C, lanes 6–9)
The size in e als be ween neighbou ing eplica ion p od-
uc s we e cons an , abou 3.5 kb, which co esponded o
he size o he DNA empla e molecule, sugges ing ha he
eplica ion o k a es ed in he same egion each ime he
ci cle was a e sed. In addi ion, he i s band was a abou
5.5 kb, co esponding o he size expec ed o eplica ion
pausing a he MTERF1 binding si e du ing he i s such
cycle. We nex epea ed he expe imen wi h an iden ical
empla e, bu wi h he MTERF1 binding si e o ien ed in he
o wa d di ec ion. Wi h his empla e, we ailed o obse e
any e ec s o he MTERF1 p o ein on DNA syn hesis (Fig-
u e 2D). Ou esul s suppo he conclusion ha MTERF1
a es s he m DNA eplica ion machine y when app oach-
ing he 3-end o he RNA ansc ip ion uni , wi h he same
di ec ionali y as ansc ip ion om LSP. The ime-cou se
expe imen also demons a ed ha m DNA eplica ion was
no de ini i ely e mina ed a he MTERF1 binding si e, bu
ha MTERF1 s imula ed pausing a his si e, since sho e
eplica ion p oduc s we e chased in o longe p oduc s o e
ime (Figu e 2C, compa e 5.5 kb p oduc in lanes 7–9).
To e i y he ele ance o MTERF1 binding o he
obse ed e ec s on m DNA eplica ion, we p epa ed a
olling ci cle eplica ion empla e wi h 4-bp dele ion in he
MTERF1 binding si e (Figu e 3and Supplemen a y Fig-
u e S1). The e ec s o MTERF1 on m DNA eplica ion
we e moni o ed in a ime-cou se expe imen . The binding
si e mu a ion comple ely abolished he o k-pausing e ec
o MTERF1.
MTERF1 is a con ahelicase
POL␥is s ic ly dependen on TWINKLE o eplica-
ion using a dsDNA empla e. We decided o moni-
o i MTERF1 could unc ion by blocking TWINKLE-
dependen DNA unwinding. To his end, we c ea ed
DNA helicase subs a es by annealing oligonucleo ides 32P-
labeled in he 5-end ( o sequences, see Ma e ials and Me h-
ods) o he complemen a y egion o pBluesc ip II SK
single-s anded DNA (SK+ o SK−) o o m helicase sub-
s a es wi h a sho double-s anded egion and a 40-n 5-
single-s anded ail. The double-s anded egion con ained
he MTERF1-binding si e in he o wa d o e e se o ien-
a ion ela i e o he 40-n 5-single-s anded ail. In he ab-
sence o MTERF1, he TWINKLE helicase was able o un-
wind he DNA subs a es. Addi ion o inc easing amoun s
o MTERF1 had no e ec on DNA unwinding o he em-
pla e wi h he MTERF1 binding si e in he ‘ o wa d’ o i-
en a ion (Figu e 4A, lanes 1–6). In con as , inc easing
amoun s o MTERF1 dose-dependen ly blocked DNA un-
winding o he empla e wi h he MTERF1 binding si e in
he e e sed o ien a ion (Figu e 4A, lanes 7–12). We con-
cluded ha he abili y o MTERF1 o block he TWIN-
KLE DNA helicase is dependen o he ela i e o ien a ion
o he MTERF-binding si e, since we obse ed a con ahe-
licase ac i i y when unwinding p og essed in an o ien a ion
opposi e ha o RNA ansc ip ion, bu when we e e sed
he o ien a ion o he binding si e, he con ahelicase ac i -
i y was los .
We nex pe o med a ime-cou se expe imen wi h con-
s an amoun s o MTERF1. We used he o k-like heli-
case subs a es wi h he double-s anded egion con aining
a binding si e o MTERF1 in he e e se di ec ion (Figu e
4B, lanes 1–10). We also included a subs a e in which he
MTERF1 binding si e had been mu a ed by dele ing 4 bp,
AGAG, in he middle o he MTERF1 binding si e (Fig-
u e 4B, lanes 11–20 and Supplemen a y Figu e S1). As ex-
pec ed, MTERF1 was able o block TWINKLE DNA un-
winding on he wild- ype binding si e empla e (Figu e 4B,
lanes 7–10). The e ec was abolished when he MTERF1
binding si e was mu a ed (Figu e 4B, lanes 17–20). To a-
cili a e u he analysis and o in es iga e i he MTERF1
con ahelicase ac i i y was speci ic o TWINKLE, we also
in es iga ed MTERF1 e ec s on a ela ed DNA helicase,
he bac e iophage T7 gene 4 p o ein (gp4) (Figu e 4C).
The MTERF1 and gp4 p o ein concen a ions we e kep
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om

Nucleic Acids Resea ch, 2016, Vol. 44, No. 12 5737
10 kb
8 kb
6 kb
3 kb
5 kb
1 2 3 4 5 6 7 8 9 10
MTERF1
- - - - + + + +
5 10 20 40 M 5 10 20 40 M
BA
1 2 3 4 5 6 7 8 9 10
C
Re e se Fo wa d
MTERF1 (pmol)
0 0.1 0.2 0.5 1.0 2.0 0 0.1 0.2 0.5 1.0 2.0
Re e se
Templa e
MTERF1
- - - - + + + +
5 10 20 40 5 10 20 40
Fo wa d
1 2 3 4 5 6 7 8 9
D
Templa e
MTERF1
3´
5´
Time (min) Time (min)
RF
Figu e 2. E ec s o human ecombinan MTERF1 on m DNA eplica ion in i o. The p epa a ion o he olling ci cle empla es and he eplica ion
eac ions we e pe o med as desc ibed in ‘Ma e ials and Me hods’. (A) Schema ic p esen a ions o he DNA empla es; he a ows show he di ec ion o
MTERF1 binding si es. (B) Pola pausing ac i i y o ecombinan human MTERF1 on m DNA eplica ion. Inc easing amoun s o MTERF1 we e added
as indica ed and pausing was only obse ed when eplica ion app oached MTERF1 in he di ec ion e e se o ha o expec ed DNA ansc ip ion. (C)
Time cu e o eplica ion wi h MTERF1 binding si e-con aining empla e in e e se di ec ion. Lanes 1–4 we e eac ions done in he absence o MTERF1
and lanes 6–9 in he p esence o MTERF1 (750 mol). Lanes 5 and 10 con ains a labeled 1 kb DNA ladde wi h sizes indica ed on he igh . (D) Time
cu e o eplica ion wi h MTERF1 binding si e-con aining empla e in o wa d di ec ion. Lanes 1–4 a e eac ions in he absence o MTERF1 and lanes
6–9 in he p esence o MTERF1 (750 mol).
MTERF1
- + + + + - - - - - + + + +
Time (min)
40 5 10 20 40 40 5 10 20 40 5 10 20 40
1 2 3 4 5 6 7 8 9 10 11 12 13 14
TWINKLE
- + + + + - + + + + + + + +
Re e se Re e se (DEL)
Figu e 3. MTERF1 dependen pausing o m DNA eplica ion in i o is
dependen on he MTERF1 binding si e. In i o eplica ion wi h olling
ci cle empla es con aining a w MTERF1 binding si e in he e e se di-
ec ion (lanes 1–5) o a MTERF1 binding si e wi h a 4-bp dele ion in a
e e se di ec ion (lanes 6–14).
cons an and we again moni o ed DNA unwinding o he
empla e in a ime cou se expe imen . In he absence o
MTERF1, he gp4 helicase could unwind bo h empla es
(Figu e 4C, lanes 3–6 and 13–16). Addi ion o MTERF1
blocked gp4-dependen unwinding o he subs a e con-
aining he MTERF1 binding si e in he e e se o ien a-
ion (Figu e 4C, lanes 7–10). MTERF1 did no a ec he
gp4-d i en unwinding eac ion, when he binding si e was
mu a ed (Figu e 4B, lanes 17–20). MTERF1 can hus e i-
cien ly block DNA unwinding by gp4.
We obse ed ha he MTERF1-dependen block o
DNA unwinding was no comple e, since he e was a weak,
bu no iceable inc ease in DNA unwinding a la e ime
poin s, e en in he p esence o MTERF1 (Figu e 4B, com-
pa e lanes 6 and 10). To in es iga e he possibili y ha
MTERF1 s imula es helicase pausing a he han e mi-
na ion, we pe o med a mo e de ailed ime-cou se expe i-
men using he DNA helicase subs a e wi h he MTERF1
binding si e in bo h o wa d (Figu e 5A) and e e se di-
ec ion (Figu e 5B). In he o wa d di ec ion, he addi ion
o MTERF1 did no a ec he DNA unwinding a e (Fig-
u e 5A and C). In he e e se di ec ion, i.e. ha co e-
sponding o TWINKLE app oaching MTERF1 in a di ec-
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
5738 Nucleic Acids Resea ch, 2016, Vol. 44, No. 12
MTERF1 - - - - -
A
B
MTERF1 - - - - - - + + + + - - - - - - + + + +
- - -
Time (min)
TWINKLE
T7 gp4
- - - -
*
MTERF1
F
*
MTERF1
R
*
MTERF1
R
*
MTERF1
R
DEL
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 2 3 4 5 6 7 8 9 10 11 12
- - - - - - + + + + - - - - - - + + + +
Time (min) - - - -
*
MTERF1
R
*
MTERF1
R
DEL
MTERF1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
C
S
P
S
P
S
P
Figu e 4. MTERF1 has con ahelicase ac i i y. MTERF1 blocks TWINKLE dependen DNA unwinding in one di ec ion, bu no he o he . (A) Helicase
assays we e pe o med as desc ibed in ‘Ma e ials and Me hods’. The a ows show he di ec ionali y o he MTERF1 binding si e; he s a indica es he
adioac i i y-labeling si e o he oligonucleo ide. Lanes 1 and 7, subs a e hea ed o 100◦C be o e loading; lanes 2 and 8, un ea ed subs a e. Inc easing
amoun s o MTERF1 (100, 250 and 500 mol) we e added as indica ed. (B) Time-cou se expe imen using a helicase subs a e wi h MTERF1 binding si e
in e e se di ec ion o wi h a mu a ed MTERF1 binding si e (4-bp dele ion). Lanes 1 and 11, un ea ed subs a e; lanes 2 and 12, subs a e hea ed o 100◦C
be o e loading. TWINKLE (600 mol) is added o lanes 3–10 and 13–20. (C) As in B, bu wi h T7 gp4 ins ead o TWINKLE. Lanes 1 and 11, subs a e
hea ed o 100◦C be o e loading; lane 2 and 12, un ea ed subs a e. MTERF1 (500 mol) was added o lanes 7–10 and lanes 17–20. T7 gp4 p o ein (100
mol) was added o lanes 3–10 and 13–20. S, double-s anded subs a e; P, single-s anded p oduc .
ion opposi e ha o DNA ansc ip ion, MTERF1 inhib-
i ed DNA unwinding. In e es ingly, MTERF1 did no com-
ple ely block TWINKLE-dependen unwinding, ins ead
slowed i down signi ican ly (Figu e 5B and D). We epea ed
he expe imen wi h he gp4 helicase and ob ained simila
esul s (da a no shown). Based on hese obse a ions we
conclude ha MTERF1 is a po en con ahelicase ha de-
lays TWINKLE-dependen DNA unwinding in a di ec ion-
dependen way.
MTERF1 s imula es TWINKLE pausing in i o
A p e ious s udy has demons a ed ha o e exp ession o
MTERF1 can inc ease eplica ion pausing a MTERF1
binding si es in i o (41). To in es iga e i he ob-
se ed e ec co ela es wi h inc eased TWINKLE paus-
ing in i o, we pe o med ch oma in immunop ecipi a-
ion (ChIP) using se en di e en p ime pai s ha co -
e ed he MTERF1 binding si e and su ounding egions.
In cells o e -exp essing a C- e minally Myc epi ope- agged
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
Nucleic Acids Resea ch, 2016, Vol. 44, No. 12 5739
Figu e 5. MTERF1 s imula es DNA helicase pausing, no e mina ion. (Aand B) DNA unwinding was measu ed in a ime-cou se expe imen in he
absence o p esence o MTERF1 (500 mol), using wo di e en subs a es wi h he MTERF1 binding si e in o wa d (A) o e e se o ien a ion (B). (C
and D) DNA unwinding e iciency was quan i ied om phospho image images o he subs a e wi h he MTERF1 binding in he o wa d (C) o e e se
(D) o ien a ion. All esul s we e no malized o he esul s o 60 min unwinding in he absence o MTERF1 o he espec i e empla e.
12345678
WT
OE
Pe cen age o Inpu (%)
123
456 78
MTERF1 00140082
2.6
2.2
1.8
1.4
1.0
TWINKLE occupancy
Figu e 6. ChIP analysis o TWINKLE occupancy. ChIP analysis o
TWINKLE occupancy in non-induced cells (da k g ey) o cells o e ex-
p essing MTERF1 (ligh g ey). The loca ions o PCR agmen s used o
moni o TWINKLE occupancy a e indica ed, he ed box co esponds o
he MTERF1 binding si e.
MTERF1, we no ed a peak o TWINKLE p o ein adja-
cen o he MTERF1 binding si e (Figu e 6). The peak was
ba ely no iceable in he non-induced con ol cells. The e-
sul s con i m ha MTERF1 can inc ease TWINKLE occu-
pancy le els a he MTERF1 binding si e, hence suppo ing
he idea ha MTERF1 ac s as a con ahelicase, inducing
TWINKLE pausing also in i o.
DISCUSSION
DNA eplica ion and ansc ip ion a e undamen al p o-
cesses ha mus be pe o med wi h high ideli y. The en-
zymes esponsible, DNA polyme ase and RNA polyme ase,
ac in coo dina ion wi h a numbe o addi ional ac o s
and a el o long dis ances on he DNA empla e. To he
bes o ou knowledge, DNA eplica ion and ansc ip ion
in mi ochond ia a e no sepa a ed in ime, which means
ha hese enzyma ic machine ies may use he same DNA
empla e and he e o e ha e he po en ial o in e e e wi h
each o he . S udies in many di e en sys ems ha e demon-
s a ed ha head-on collision be ween ansc ip ion and
eplica ion can cause dec eased eplica ion o k s abili y,
leading o eplica ion s ess (38). This helps o explain why
highly ansc ibed genes a e associa ed wi h inc eased le -
els o DNA mu a ions ( ansc ip ion associa ed mu a ions,
TAM) and ecombina ion ( ansc ip ion associa ed ecom-
bina ion, TAR) (44–47). Head-on collision may also a ec
he many p o eins associa ed wi h co- ansc ip ional RNA
p ocessing, which in mi ochond ia include he p ecise en-
donucleoly ic clea age o he polycis onic ansc ip s. Bac-
e ial genomes seem o ha e e ol ed o minimize he p ob-
lem o eplica ion - ansc ip ion collision. Fo ins ance,
in Esche ichia coli, he se en highly ansc ibed ibosomal
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om
5740 Nucleic Acids Resea ch, 2016, Vol. 44, No. 12
Figu e 7. A schema ic igu e o he po en ial ou come o a collision be ween ansc ip ion and DNA eplica ion and he MTERF1 binding si e. S ep
1. The mi ochond ial eplisome pauses a he MTERF1 binding si e. T ansc ip ion ini ia ed om HSP app oaches he MTERF1 binding si e du ing
ansc ip ion o he DNA egion. Al e na i e S ep 2A. T ansc ip ion displaces MTERF1 om he empla e, bu is e mina ed immedia ely a e when
POLRMT collides wi h TWINKLE enci cling he empla e s and. Al e na i e S ep 2B. T ansc ip ion displaces MTERF1 and TWINKLE om he
empla e s and. POLRMT con inues o ansc ibe o ano he 25–75 n be o e i e mina es due o lack o p ocessi i y on ssDNA. TWINKLE is eloaded
be o e H-s and DNA syn hesis can be esumed.
genes a e mainly p esen in egions nea he o igin o epli-
ca ion (48), o ien ed in a way so ha eplica ion and an-
sc ip ion ake place in he same di ec ion (49). Ano he
means by which collisions be ween eplica ion and an-
sc ip ion a e p e en ed is ia speci ic o k-p og ession ba -
ie s ound a highly ansc ibed DNA in euka yo ic nuclei
(40).
Simila o o k-ba ie p o eins in many o he sys ems,
MTERF1 binds immedia ely downs eam o he DNA
ansc ip ion uni (21). P e ious epo s ha e demon-
s a ed ha MTERF1 blocks an isense ansc ip ion ac oss
he DNA egion in bo h human and mouse m DNA
(24,25,27). As demons a ed he e, MTERF1 can also a es
he mo ing eplica ion wi h he same di ec ionali y as ha
obse ed o ansc ip ion. MTERF1 hus ac s as a epli-
ca ion o k ba ie and causes a delay in he p og ession o
DNA eplica ion in he di ec ion opposi e ha o DNA
ansc ip ion. As a consequence, MTERF1 can ac o a oid
o manage head-on collisions be ween eplica ion and an-
sc ip ion machine ies in he DNA egion, he eby ensu -
ing ha POLRMT has comple ed syn hesis o highly s uc-
u ed RNA molecules be o e collision wi h he m DNA
syn hesis machine y. Ou indings also p o ide a biochem-
ical explana ion o he obse a ion ha o e exp ession o
MTERF1 causes eplica ion pausing in i o (41). The e ec
is explained by he abili y o MTERF1 o unc ion as a con-
ahelicase wi h dis inc pola i y. In suppo o his no ion,
we employed ‘ch oma in’ immunop ecipi a ion o demon-
s a e ha MTERF1 s imula es TWINKLE pausing in i o.
I is emp ing o specula e ha MTERF1 suppo s a p o-
cess whe eby POLRMT comple es RNA ansc ip ion ini-
ia ed om HSP be o e H-s and DNA eplica ion con-
inues o e he DNA egion. Acco ding o his model,
he mi ochond ial eplisome pauses a he MTERF1 bind-
ing si e (Figu e 7, s ep 1). Once POLRMT has passed he
DNA egion, ansc ip ion would acili a e he displace-
men MTERF1 and he block o m DNA eplica ion in he
opposi e di ec ion is emo ed (Figu e 7, s ep 2). This model
p edic s ha POLRMT encoun e s he paused TWINKLE
helicase, which enci cles he ansc ibed s and. The mos
plausible ou come o his collision is ha POLRMT e mi-
na es ansc ip ion and ha he H-s and DNA syn hesis
is esumed (Figu e 7, s ep 2A). This inding would be in
line wi h indings in many o he sys ems, whe e ansc ip-
ion no mally yields o a mo ing eplisome (38). A simi-
la concep has been p oposed o sea u chin (50,51)and
D osophila (52) m DNA. In ac , he sea u chin m DBP
p o ein unc ions as a con ahelicase, inasmuch as i blocks
he helicase ac i i y o he he e ologous simian i us 40
la ge T an igen (50).
Po en ially, he in e ac ion wi h MTERF1 weakens ha
o POLRMT and he empla e, which allows TWINKLE o
displace POLRMT om he empla e s and. Te mina ion
o ansc ip ion in his manne would be a a e e en , un-
likely o signi ican ly a ec he ela i e le els o RNA and
mRNA molecules p oduced downs eam o he MTERF1
binding si e. The al e na i e ou come (Figu e 7, s ep 2B) ap-
pea s less likely. I POLRMT displaces TWINKLE i would
en ail eplisome eloading a e passage o he ansc ip ion
machine y, a p ocess ha could lead o misp iming and in-
c eased le els o mu a ions. In addi ion, POLRMT has e y
low p ocessi i y on ssDNA and, i i managed o displace
a Tampe e Uni e si y Lib a y. Depa men o Heal h Sciences on Sep embe 28, 2016h p://na .ox o djou nals.o g/Downloaded om