Duplication and triplication with staggered breakpoints in human mitochondrial DNA

(1)

Duplication and triplication with staggered breakpoints in human

mitochondrial DNA

Celia H. Tengan

_´

a,c

_{, Carlos T. Moraes}

a,b,) a

Department of Neurology, UniÕersity of Miami School of Medicine, Miami, FL, USA

b

Department of Cell Biology and Anatomy, UniÕersity of Miami School of Medicine, Miami, FL, USA

c ₍ ₎

Disciplina de Neurologia Clınica, Escola Paulista de Medicina UNIFESP , Sao Paulo, Brazil´ ˜

Received 15 August 1997; accepted 5 November 1997

Abstract

Ž .

We identified a tandem duplication and triplication of a mitochondrial DNA mtDNA segment in the muscle of a 57-year-old man with no evidence of a neuromuscular disorder. A large triplication of a mtDNA coding region has not been

Ž

previously reported in humans. Furthermore, the rearrangements comprising 10–12% of the muscle mtDNA pool in the

. Ž

propositus were unique because the breakpoints were staggered at both ends between mtDNA positions 3263–3272 and

.

16 065–16 076 and contained no identifiable direct repeats. Both sides of the breakpoint were located approximately 35 bp

Ž .

downstream of regions that undergo frequent strand displacement by either transcription positions 3263–3272 or

Ž .

replication positions 16 065–16 076 , suggesting that topological changes generated by the movement of RNArDNA polymerases may be associated with the genesis of a subclass of mtDNA rearrangements. The presence of low levels of these rearrangements in other normal adults also suggest that these mutations are not rare. The characterization of these rearrangements shed light on potential alternative mechanisms for the genesis of mtDNA rearrangements.q_{1998 Elsevier}

Science B.V.

Keywords: mtDNA; Duplication; Superhelical stress; Recombination

1. Introduction

Two types of rearrangements have been described

Ž .

in human mitochondrial DNA mtDNA : deletions and duplications. MtDNA deletion is the most com-mon rearrangement found in patients with mitochon-drial disorders and are usually associated with

pro-Ž . w x

gressive external ophthalmoplegia PEO 1 ,

Ž .w x

Kearns–Sayre Syndrome KSS 2 and Pearson

Syn-)_{Corresponding author. Department of Neurology, University}

of Miami, 1501 NW 9th Ave., Miami, FL 33136, USA. Fax:

q1-305-243-4678; e-mail: cmoraes@mednet.med.miami.edu.

Ž . w x

drome PS 3 . In 1989, Poulton et al. described a large tandem duplication of mtDNA in two patients

w x

with KSS 4 . Since then, mtDNA tandem duplica-tions were reported in patients with different clinical manifestations such as KSS, PS, diabetes mellitus,

w x

renal tubulopathy and mitochondrial myopathy 5–9 . A 260-bp duplication of the D-loop region was

iden-w x

tified by Brockington et al. 10 . Brockington and colleagues suggested that this duplication could be a predisposing factor to mtDNA deletion generation

w x

but this was not confirmed by Manfredi et al. 11 . Small tandem duplications of the D-loop region were

w x

also detected in normal individuals 12,13 . Similarly,

Ž .

(2)

( ) C.H. Tengan, C.T. MoraesrBiochimica et Biophysica Acta 1406 1998 73–80

74

to deletions, many mtDNA duplications have direct

w x

repeats in the junction region 14,15 and when asso-ciated with deletions, they appear to have the same breakpoint. The origin of these rearrangements is still unknown, although the presence of direct repeats in the breakpoint region of some rearrangements sug-gests homologous recombination.

We identified low levels of a duplication and a triplication of a large mtDNA segment in an individ-ual who did not have a history of neuromuscular disorder. This case is unique because it is the first description of a large tandem triplication in humans and because both rearrangements involved staggered breakpoints with the absence of direct repeats. The conspicuous location of the breakpoints suggests that topological changes induced by the movement of polymerases on the DNA molecule may be involved in the generation of some types of mtDNA rearrange-ments.

2. Materials and methods

2.1. DNA samples

We analyzed muscle DNA obtained from autopsy material of the propositus, a 57-year-old man. Ac-cording to the autopsy report, the cause of death was pneumonia and no history or pathological signs of neuromuscular disorder were mentioned. Genomic DNA was obtained by organic extraction as

previ-w x

ously described 2 . Autopsy muscle from six

addi-Ž .

tional individuals ages 3, 37, 49, 57, 67, 91 years without evidence of neuromuscular disorders were

also analyzed for the presence of the rearrangements observed in the propositus.

2.2. DNA analysis

Genomic DNA was independently digested with three different restriction enzymes: PÕuII, Pst I and

DraI. Digested DNA was separated by

electrophore-sis through a 0.6% agarose gel, transferred to a nylon membrane which was hybridized to two mtDNA

32 _{Ž .}

specific P-labeled probes: A corresponding to Ž . mtDNA positions 11 680 to 12 406 and B corre-sponding to mtDNA positions 526 to 1768. The probes were obtained by PCR amplification of the specific regions of mtDNA, purified from agarose gel ŽQuiex II, Qiagen and labeled by the random primer.

Ž .

labeling kit Boehringer Mannheim as recommended by the manufacturer. The hybridized membranes were exposed to X-ray films. The bands were quantified by scanning and analyzing the autoradiograms with the

Ž

aid of an image analysis software IMAGE V 1.61, .

NIH freeware .

The breakpoint region was amplified by PCR us-ing about 500 ng of muscle DNA, 50 pmol of

Ž .

forward primer mtDNA positions 3116–3134 and

Ž .

reverse primer mtDNA position 16 496–16 465 , 20

Ž .

nmol of each deoxynucleotide triphosphate dNTP , 10X PCR buffer, 2.5 units of Taq DNA polymerase ŽBoehringer . Amplification conditions were as fol-. lows: 30 cycles consisting of 94₈C for 1 min, 55₈C for 1 min, 728C for 1 min followed by a last exten-sion step at 72₈C for 10 min. PCR products were separated by electrophoresis through an ethidium

bro-Ž .

mide stained agarose gel 1% and then visualized under a UV source. Direct sequencing of gel-purified

Ž .

Fig. 1. Characterization of the duplicationrtriplication of a mtDNA segment. Total muscle DNA from the subject SUBJ and a normal

Ž . Ž . Ž . Ž .

control CTRL were digested with Pst I panel A , DraI panel B , and PÕuII panel D , electrophoresed through a 0.6% agarose gel,

Ž . Ž

transferred to a nylon membrane and hybridized to probe A mtDNA positions 11 680 to 12 406 or probe B mtDNA positions 526 to

.

1768 . Migration of molecular weight markers are indicated on each panel. D, digested DNA; U, undigested DNA. Panel C shows the plot obtained from densitometric analysis of a autoradiography from the Pst I digestion seen in panel A. The relative intensities of the

Ž . Ž

peaks calculated from a lesser-exposed autoradiography showed different relative intensities of the normal and abnormal bands trip and

. Ž .

dup with probe A and probe B. Panel E shows in schematic form the structure of the wild-type mtDNA WT , the mtDNA with a

Ž . Ž . Ž .

duplication DUP , with a triplication TRIP and the putative molecule with a large deletion DEL, not observed in the blots . The approximate sites for PÕuII, Pst I, and DraI cleavage, and the location of probes A and B used in the Southern blot analyses are shown.

Ž

Panel E also shows the location of the PCR primers used to identify rearranged molecules. The diagnostic 580 bp PCR shown only in the

.

(3)

(4)

76

PCR products was performed with the fmol DNA

Ž .

sequencing kit Promega as suggested by the manu-facturer. PCR products were also cloned into pCRII

Ž .

plasmid vector TA cloning, Invitrogen and se-quenced by an ABI automated sequencer.

3. Results

During an unrelated study, we analyzed the muscle mtDNA of several individuals with no history of a mitochondrial disorder. One of these individuals, a 57 year old man, showed an abnormal mtDNA hybridiz-ing band on a Southern blot after digestion with

PÕuII. The blot showed the normal 16.5-kb linearized

Ž mtDNA band and a smaller one of about 4 kb Fig.

. Ž

1D . The smaller band 12% of the total mtDNA .

pool could correspond to a large 12.5 kb deletion or a 4.0 kb duplication of mtDNA. To identify the type of rearrangement we performed a digestion with Pst I,

Ž .

using two probes probe A and B . The normal mtDNA has two sites for Pst I, at mtDNA positions

Ž

6910 and 9020, giving rise to two fragments 14.5 kb .

and 2 kb . An mtDNA molecule with a duplication including the PÕuII site would give rise to a fragment

larger than 14.5 kb. On the other hand, a mtDNA molecule with a 12.5-kb deletion sparing the PÕuII

site, would lack sites for Pst I, and would run as undigested mtDNA during electrophoresis. Probe A Žsee panel E would detect the fragments from the. normal and duplicated mtDNA molecules, but not from the deleted molecules and probe B would detect deleted, duplicated and normal mtDNA molecules. Probe A revealed the presence of the wild-type

14.5-Ž

kb band and two additional larger bands approxi-.

mately 18.5 kb and 22.5 kb . Probe B detected the

Ž .

same bands Fig. 1A , which did not migrate as undigested DNA. The two larger bands could repre-sent a duplication and a triplication of a 4-kb segment of mtDNA. To provide further evidence that the largest fragment had a third repeated segment of the duplicated molecule we digested total DNA with

DraI which would have one site in the triplicated and

duplicated segment, giving rise to only one probe B-detectable abnormal fragment if the segments in-serted were the same. A Southern blot of the DraI digest with probe B showed only one extra band with 3.8 kb, corresponding to approximately 10% of the

Ž .

total mtDNA Fig. 1B . The abnormal bands ob-served in the propositus DNA were not obob-served in the control sample even after overexposure of the

Ž .

different blots to X-ray films not shown . The DraI digest result was compatible with the coexistence of a duplicated and triplicated molecule which contained the same breakpoint region.

Comparing the intensity of the abnormal bands with the two probes showed that the larger band Žcorresponding to the triplicated mtDNA was weaker.

Ž .

with probe A Fig. 1C . Probe A signal reflects the real ratio of duplicated to triplicated mtDNA, as the segment hybridized to probe A is not duplicated or triplicated. On the other hand, probe B hybridized to a segment that is duplicated and triplicated, explain-ing the stronger signal for the triplicated mtDNA with this probe. The normalized ratio of the different

Ž

bands i.e., signal intensity divided by the number of

. Ž .

repeats for probe B was 1:0.10:0.05 wt:dup:trip , a ratio that was similar to the one observed with probe

Ž . Ž .

A 1:0.13:0.04 Fig. 1C . The ratio of mutated to Ž

wild-type mtDNAs determined to be 10–12% in the .

PÕuII and DraI blots was probably overestimated in

the Pst I X-rays because we had to overexpose the blots to observe the triplicated mtDNA. Taking into account all three Southern blot experiments, we con-cluded that the mutated mtDNA fraction corresponds to 10–12%, with 6.7–8% duplicated and 3.3–4% triplicated. Although panel E of Fig. 1 shows the structure of a putative deleted mtDNA, such molecule was not observed in our blots.

Fig. 2. DNA sequence of the breakpoint regions from the rear-ranged mtDNA. The figure shows DNA sequences from seven

Ž .

unique plasmid clones from 12 sequenced clones with an inserted PCR segment containing the breakpoint region of the mtDNA rearrangement observed in the propositus. The exact breakpoint position is shown under each sequence. Regions in brackets were not present in the abnormal junction. In bold is an

Ž .

(5)

The breakpoint region was amplified by PCR us-ing oligonucleotide primers as described in Section 2. We could amplify an approximately 580-bp fragment using a forward primer corresponding to mtDNA position 3116–3134 and a reverse primer correspond-ing to mtDNA position 16 496–16 465. We directly sequenced this PCR fragment but the sequence was

Ž

unreadable from mtDNA position 3263 using the . forward PCR primer as the sequencing primer . The PCR fragment was cloned into a pCRII vector and 12 independent clones sequenced. Sequence results of the 12 clones showed seven different breakpoints which varied from positions 3263 to 3272 and 16 065 to 16 076. No direct repeats were identified in the breakpoint region and we detected an insertion of a G

Ž .

at position 16 064 in one clone Fig. 2 .

We attempted to detect the same rearrangements in muscle DNA of normal individuals of different ages. PCR amplification was performed using the same

Ž

pair of primers 13 380 bp apart in the wild-type .

sequence which gave a 580-bp product from the subject’s muscle DNA. We could detect the 580-bp

Ž

fragment in three additional individuals 49, 57 and .

91 years old, Fig. 3 . Additional bands, probably

Fig. 3. PCR detection of the rearranged mtDNA in normal adults.

Ž

PCR-amplified fragments spanning the breakpoint region of the

.

subject from muscle DNA of different controls were

elec-Ž .

trophoresed through a 1% agarose gel panel A . Sbj, amplifica-tion from subject’s muscle DNA. MW, 100-bp molecular weight ladder. ‘y’, amplification without template DNA. The specific 580-bp band, containing the breakpoint region of the duplication,

Ž

was observed in amplifications from three normal controls ages

. Ž .

49, 57, 91 and the subject panel A . To confirm the identity of the 580-bp fragment, the bands were purified from the gel and digested with DdeI and MspI which are known to digest the

Ž w

fragment spanning the breakpoint position of sites and the

x .

breakpoint region are depicted in panel B . The digestion pattern of the 580-bp band was determined by agarose gel

electrophore-Ž .

sis panel C .

originated from different rearrangements, were also

Ž .

observed in some samples Fig. 3A . To assure that the 580-bp bands corresponded to the same break-point region observed in the propositus, we excised and purified the band from the gel and subjected the

Ž .

DNA to digestion with DdeI and MspI Fig. 3B,C . The restriction pattern of the bands amplified from the additional ‘normal’ controls was indistinguishable from the one observed with the amplification from

Ž .

the propositus Fig. 3C . Although the PCR approach does not allow us to distinguish between duplication, triplication or deletion, it does implicate the same breakpoint region in large-scale rearrangements.

4. Discussion

w x

Although it was described in humans in 1989 4 , the genesis and pathophysiology of mtDNA duplica-tions is still unclear. Because of the co-existence of mtDNA duplications and deletions in many individu-als, it was suggested that mtDNA deletions and dupli-cations were originated from the same recombination

w x w x

event 7 . Poulton et al. 7 re-analyzed 14 patients with KSS and mtDNA deletions and found duplica-tions of mtDNA in all of them, suggesting that duplicated mtDNA molecules could be intermediate forms that would evolve to a deleted mtDNA. This idea was supported by the finding that the level of duplicated mtDNA decreased with time while deleted

w x

mtDNA increased 16 . The presence of direct repeats in many of the duplication or deletion junction re-gions suggests that homologous recombination is as-sociated with the recombination mechanism. How-ever, a significant group of patients with mtDNA rearrangements do not have direct repeats in the

w x

breakpoint region 14 . These rearrangements, termed

w x

class II by Mita et al. 14 , may arise by a different recombination process.

In this report, we describe an individual with a large tandem duplication and triplication in his mus-cle mtDNA. The breakpoint region of these rear-rangements was staggered at both ends, and no iden-tifiable direct repeat was observed. These findings suggest a nonhomologous recombination mechanism for these rearrangements. The breakpoint region was

Ž .

(6)

78

Fig. 4. Location of rearrangements’ breakpoints. Schematic repre-sentation of the double-stranded human mtDNA. Aborted

replica-Ž .

tion events of the heavy strand D-loop formation and transcrip-tion of the rRNA region are represented in the diagram. The rearrangements’ breakpoints are 30–43 bp downstream of the two mtDNA regions which undergo frequent strand displacement by transcription or replication, potentially generating regions of superhelical tension or other topologic alterations. OH, origin of the heavy strand replication. HSP, heavy strand promoter region.

Ž16 065–16 076 . One of these regions. Ž16 065– .

16 076 was previously found to be associated with

w x

staggered-breakpoint deletions 17,18 . These dele-tions were observed in different families with autoso-mal dominant PEO and were part of a larger group of

w x

different mtDNA deletions 17,18 . Zeviani et al. suggested that the 16 065–16 076 region could be prone to recombination because it forms hairpin structures which may interact with trans-acting

fac-w x

tors involved with replication 17 .

Both regions involved in the rearrangements in the present report have the common feature of being located immediately downstream of mtDNA regions that undergo frequent strand displacement by either

Ž . Ž

transcription region 3263–3272 or replication re-.

gion 16 065–16 076, Fig. 4 . Transcription of the heavy-strand usually stops at positions 3237–3249 within the tRNALeuŽUUR. gene. Only 4% of the tran-scriptional events starting at the heavy-strand pro-moter region proceed beyond this region. This strong transcription termination is mediated by a factor ŽmTERF that footprints positions 3229–3256 of the.

w x Ž

mtDNA 19 . One of the breakpoint regions mtDNA .

positions 3263–3272 is 34–43 bp downstream from

X

w x

the 3 end of this frequently transcribed region 20 . Ž Coincidentally, the mtDNA region 16 065–16 076 the

.

other breakpoint is located 30–41 bp downstream X

Ž

from the 3 end of the displacement loop D-loop, .

Fig. 4 . The D-loop is a conformational structure

w x

caused by frequently aborted replication events 21 . Transcription and replication have been associated with the generation of superhelical tension in

w x

double-stranded circular DNA 22 . The movement of a DNA unwinding protein along a double-stranded DNA generates positive superhelical tension in the DNA in front of it and a negative tension behind it

w23–26 . The D-loop unwinding of the human mtDNAx

could transfer a positive superhelical tension capable

Ž .

of causing 60 supercoils 10 ntrsupercoil . However, the previously described association of the mtDNA with the inner mitochondrial membrane via protein

w x

complexes 27–29 may make difficult the formation of supercoils in vivo, causing increased superhelical tension immediately downstream of the DNA un-winding. In solution, D-loop-containing mtDNAs from Xenopus oocytes display increased superhelic-ity when compared to mtDNAs lacking a D-loop, indicating that the displacement loop does induce

w x

superhelical tension to the mtDNA 30 .

Trans-acting factors interacting with regions of superhelical tension play a key role in controlling DNA topology and integrity. A mitochondrial topoi-somerase I-like activity identified in human and

w x

bovine mitochondria 31,32 showed a preferential supercoil relieving activity at the transcription

termi-w x

nation region at the end of the 16S rRNA gene 33 , one of the breakpoint regions described in this report. Both factors, namely DNA unwinding and topoiso-merase activity, may play a role in recombination mechanisms, either by inducing single-strand breaks caused by increased superhelical tension, or causing illegitimate rejoining of topoisomerase I-induced nicks. In any case, superhelical stress may favor the

w x

formation of recombinogenic free DNA ends 34 . It is important to emphasize that the model described above is based on indirect evidence, and at this moment, it is purely hypothetical. The crucial direct demonstration of the existence of localized superheli-cal tension in mtDNA is difficult to obtain because the association of the mtDNA with the mitochondrial

w x

(7)

the molecule, would be lost upon purification of the mtDNA. Alternatively, transcription and replication can stimulate recombination not by inducing superhe-lical stress but simply by changing DNA topology. In either case, the association between transcription, replication and recombination, which has been

de-Ž w x.

scribed in different systems reviewed in Ref. 35 , may also play a role in mtDNA recombination.

Southern blot-detectable levels of large tandem duplications of mtDNA have not been previously reported in ‘normal’ individuals. We were unable to obtain enough clinical data to ensure that this individ-ual did not have any kind of clinical manifestation, as the muscle specimen was obtained from an autopsy from the medical examiner’s office. However, this is an unlikely scenario because of the relatively low

Ž .

levels of the mutated mtDNA species 10–12% , and the lack of direct evidence that mtDNA duplications are pathogenic. Likewise, a large triplication of a mtDNA coding region has not been previously re-ported in humans. Triplication of a 260-bp

non-cod-w x

ing region was reported by Manfredi et al. 11 but it could only be detected by PCR. A triplication of a 9-bp non-coding region between COX II and the tRNALys genes has also been reported as a neutral

w x

polymorphism 36 . Although we hypothesize a mechanism involving topological alterations for the genesis of the duplication, the triplication may have arisen by slip mispairing or other form of homolo-gous recombination from a duplicated molecule

w37,38 .x

The peculiar features of the mtDNA rearrangement in our subject led us to hypothesize that DNArRNA polymerases may cause conformational changes or superhelical tension which may play a role in the genesis of a subgroup of mtDNA rearrangements which are not associated with homologous recom-bination of direct repeats. Homologous recomrecom-bination may also act together with localized topological alter-ations in the mtDNA as small direct repeats have been associated with staggered breakpoints at the end

w x

of the D-loop region 17,18 . Although further studies will be needed to validate this concept, this hypothe-sis also raises the possibility that enzymes associated with the control of mtDNA topology could be poten-tial candidates for mutations in mendelian-inherited syndromes presenting with mtDNA rearrangements

w39,40 .x

Acknowledgements

This work was supported by the NIH grant EY10804, the Muscular Dystrophy Association and by the PEW Charitable Trust. Dr. C.H. Tengan was

Ž .

supported by CNPq Brazilian Research Council . We thank Dr. David A. Clayton for useful discus-sions on mtDNA superhelicity.

References

w x1 C.T. Moraes, S. DiMauro, M. Zeviani, A. Lombes, S. Shanske, A.F. Miranda, H. Nakase, E. Bonilla, L.C. Wernec, S. Servidei, I. Nonaka, Y. Koga, A. Spiro, K.W. Brownell, B. Schmidt, D.L. Schotland, M.D. Zupanc, D.C. DeVivo,

Ž .

E.A. Schon, L.P. Rowland, New Engl. J. Med. 320 1989 1293–1299.

w x2 M. Zeviani, C.T. Moraes, S. DiMauro, H. Nakase, E. Bonilla,

Ž .

E.A. Schon, L.P. Rowland, Neurology 38 1988 1339–1346.

w x3 A. Rotig, V. Cormier, S. Blanche, J.-P. Bonnefont, F.¨

Ledeist, N. Romero, J. Schmitz, P. Rustin, A. Fischer, J.-M.

Ž .

Saudubray, A. Munnich, J. Clin. Invest. 86 1990 1601– 1608.

w x4 J. Poulton, M.E. Deadman, R.M. Gardiner, Nucleic Acids

Ž .

Res. 17 1989 10223–10229.

w x5 A. Rotig, J.-L. Bessis, N. Romero, V. Cormier, J.-M. Saudubray, P. Narcy, G. Lenoir, P. Rustin, A. Munnich,

Ž .

Am. J. Hum. Genet. 50 1992 364–370.

w x6 D.R. Dunbar, P.A. Moonie, R.J. Swingler, D. Davidson, R.

Ž .

Roberts, I.J. Holt, Hum. Mol. Genet. 2 1993 1619–1624.

w x7 J. Poulton, M.E. Deadman, L. Bindoff, K. Morten, J. Land,

Ž .

G. Brown, Hum. Mol. Genet. 2 1993 23–30.

w x8 S.W. Ballinger, J.M. Shoffner, S. Gebhart, D.A. Koontz,

Ž .

D.C. Wallace, Nat. Genet. 7 1994 458–459.

w x9 A. Superti-Furga, W. Schoenle, P. Tuchschmid, R. Caduff,

Ž .

V. Sabato, D. DeMattia, Eur. J. Pediatr. 152 1993 44–50.

w10 M. Brockington, M.G. Sweeney, S.R. Hammans, J.A. Mor-x

Ž .

gan-Hughes, A.E. Harding, Nat. Genet. 4 1993 67–71.

w11 G. Manfredi, S. Servidei, E. Bonilla, S. Shanske, E.A.x

Ž .

Schon, S. DiMauro, C.T. Moraes, Neurology 45 1995 762–768.

w12 A. Torroni, M.T. Lott, M.F. Cabell, Y.-S. Chen, L. Lavergne,x

Ž .

D.C. Wallace, Am. J. Hum. Genet. 55 1994 760–776.

w13 H.-C. Lee, C.-Y. Pang, H.-S. Hsu, Y.-H. Wei, FEBS Lett.x

Ž .

354 1994 79–83.

w14 S. Mita, R. Rizzuto, C.T. Moraes, S. Shanske, E. Arnaudo,x

G. Fabrizi, Y. Koga, S. DiMauro, E.A. Schon, Nucleic

Ž .

Acids Res. 18 1990 561–567.

w15 M. Zeviani, C. Gellera, M. Pannacci, G. Uziel, A. Prelle, S.x

Ž .

Servidei, S. DiDonato, Ann. Neurol. 28 1990 94–97.

w16 J. Poulton, K.J. Morten, D. Marchington, K. Weber, G.K.x

Ž

Brown, A. Rotig, L. Bindoff, Muscle and Nerve Supple-¨

. Ž .

ment 3 1995 S154–S158.

w17 M. Zeviani, S. Servidei, C. Gellera, E. Bertini, S. DiMauro,x

Ž .

(8)

80

w18 A.R. Moslemi, A. Melberg, E. Holme, A. Oldfors, Ann.x

Ž .

Neurol. 40 1996 707–713.

w19 B. Kruse, N. Narasimhan, G. Attardi, Cell 58 1989 391–x Ž .

397.

w20 J. Montoya, G.L. Gaines, G. Attardi, Cell 34 1983 151–x Ž .

159.

w21 D.A. Clayton, Cell 28 1982 693–705.x Ž .

w22 L.F. Liu, J.C. Wang, Proc. Natl. Acad. Sci. U.S.A. 84x

Ž1987 7024–7027..

w23 D.C. Eichler, T.S. Wang, D.A. Clayton, D. Korn, J. Biol.x

Ž .

Chem. 252 1977 7888–7893.

w24 J.S. Minden, K.J. Marians, J. Biol. Chem. 260x Ž1985.

9316–9325.

w25 M. Dodson, F.B. Dean, P. Bullock, H. Echols, J. Hurwitz,x

Ž .

Science 238 1987 964–967.

w26 J.C. Wang, A.S. Lynch, Curr. Opin. Genet. Dev. 3 1993x Ž .

764–768.

w27 T.B. Kazakova, M.P. Melinikova, E.D. Sverdlov, Mol. Cell.x

Ž .

Biochem. 14 1977 37–45.

w28 D.A. Potter, J.M. Fostel, M. Berninger, M.L. Pardue, T.R.x

Ž .

Cech, Proc. Natl. Acad. Sci. U.S.A. 77 1980 4118–4122.

w29 M. Barat, D. Rickwood, C. Dufresne, J.C. Mounolou, Exp.x

Ž .

Cell Res. 157 1985 207–217.

w30 J.C. Callen, M. Tourte, N. Dennebouy, J.C. Mounolou, Exp.x

Ž .

Cell Res. 143 1983 115–125.

w31 M.J. Kosovsky, G. Soslau, Biochim. Biophys. Acta 1164x

Ž1993 101–107..

w32 J.-H. Lin, F.J. Castora, Arch. Biochem. Biophys. 324 1995x Ž .

293–299.

w33 Z. Topcu, F.J. Castora, Biochim. Byophys. Acta 1264 1995x Ž .

377–387.

w34 E.C. Friedberg, G.C. Walker, W. Siede, DNA Repair andx

Mutagenesis, ASM Press, Washington, DC, 1995, pp. 22–24.

w35 S. Gangloff, M.R. Lieber, R. Rothstein, Experientia 50x

Ž1994 261–269..

w36 G. Passarino, O. Semino, G. Modiano, A.S. Santachiara-Be-x

Ž .

nerecetti, Am. J. Hum. Genet. 53 1993 609–618.

w37 E.A. Schon, R. Rizzuto, C.T. Moraes, H. Nakase, M. Ze-x

Ž .

viani, S. DiMauro, Science 244 1989 346–349.

w38 J.M. Shoffner, M.T. Lott, A.S. Voljavec, S.A. Soueidan,x

D.A. Costigan, D.C. Wallace, Proc. Natl. Acad. Sci. U.S.A.

Ž .

86 1989 7952–7956.

w39 A. Suomalainen, J. Kaukonen, P. Amati, R. Timonen, M.x

Haltia, J. Weissembach, M. Zeviani, H. Somer, L. Peltonen,

Ž .

Nat. Genet. 9 1995 146–151.

w40 J.A. Kaukonen, P. Amati, A. Suomalainen, A. Rotig, M.G.x ¨

Piscaglia, F. Salvi, J. Weissenbach, G. Fratta, G. Comi, L.

Ž .