Nucleic Acids Research, 1993, Vol. 21, No. 4 811-816

Characterization of a high-affinity binding site for a DNA-binding protein from sea urchin embryo mitochondria

Sohail A.Qureshi and Howard T.Jacobs*Robertson Institute of Biotechnology, Department of Genetics, University of Glasgow, Church St.,Glasgow G1 1 5JS, UK

Received January 5, 1993; Accepted January 19, 1993

ABSTRACT

Based on electrophoretic mobility shift assays, DNaseI footprinting and modification interference analyses wehave identified a sequence-specific DNA-bindingprotein in blastula stage mitochondria of the sea urchinStrongylocentrotus purpuratus, which interacts with abinding site around the major pause site for DNAreplication. This region straddles the boundary of thegenes for ATP synthase subunit 6 and cytochrome coxidase subunit 111, and contains also a prominent originof lagging-strand synthesis. The protein isthermostable, and its natural high-affinity binding sitecomprises the sequence 5'-AGCCT(N7)AGCAT-3'.Binding studies have demonstrated that two copies ofthe imperfect repeat, as well as the 7 bp spacingbetween them, are essential for tight binding. Basedon the location of its binding site, we tentativelydesignate the protein mitochondrial pause-regionbinding protein (mtPBP) 1.

INTRODUCTION

Sequence-specific DNA-binding proteins are known to beimportant regulators of mitochondrial transcription in mammals,and almost certainly play similar roles in plant and fungalmitochondria. These proteins include the enhancing factor fortranscriptional initiation, designated mtTF 1, which has recentlybeen shown to be a protein of the HMG-l family (1 -6), anda transcriptional termination factor, which interacts with theattenuator sequence downstream of the rRNA genes inmammalian mtDNAs (7, 8), and is believed to regulate therelative synthetic rates of different classes of RNA. In yeast, atleast one other protein, MTF1 (9, 10) seems to be involved inpromoter discrimination, although it requires the presence of thecore subunit of the mitochondrial RNA polymerase for sequence-specific DNA-binding (11).

Little is known, however, of the role of sequence-specificmtDNA-binding proteins in the regulation of DNA replicationor other processes in animal cell mitochondria, other than therequirement for mtTF 1 for replication primer synthesis at theheavy-strand origin (12). The known mechanics of metazoanmtDNA synthesis (13), involving D-loop formation, D-loop

expansion, second-strand priming at a specific site on thedisplaced strand, specific termination (presumably close to theleading-strand origin) and daughter molecule segregation, suggestthe likely involvement of other regulatory proteins, with specifictemplate-binding properties.We have become interested in the mechanism of mtDNA

replication in early sea urchin development, which isdevelopmentally regulated. Mitochondrial DNA is amplifiedduring oogenesis, but does not replicate in the mature oocyte,egg or early embryo (14), where it is maintained in the D-loopform (15). In earlier studies, we exploited the observation thatmtDNA replication may be precociously induced in enucleatedsea urchin eggs by ammonia activation (16, 17), to study thestructure of replication intermediates by gel electrophoretictechniques (18). This led us to identify replication pause sites,the most prominent of which maps close to a major site oflagging-strand initiation, near to the boundary of the genes forATP synthase subunit 6 (A6) and cytochrome c oxidase subunitIII (COIII).

In this paper, we report the identification and characterizationof a high-affinity binding site in the A6/COIII gene boundaryregion of sea urchin mitochondrial DNA, for a protein foundin blastula stage mitochondria. Because of the location of itsbinding site we suggest that the protein may play a role in mtDNAreplication, and tentatively designate it mitochondrial pause-region binding protein (mtPBP) 1.

MATERIALS AND METHODSPurification of mitochondria from blastulaeTwenty-four hour blastula embryos of the sea urchin S.purpuratuswere prepared by Dr F.J.Calzone (University of California,Irvine) according to Calzone et al. (19), stored at -70°C andlysed by thawing. After two low-speed spins at 900 gmax andone spin at 2500 gn,ax' mitochondria were harvested bycentrifugation of the supernatant at 10,000 gmax The crudemitochondrial pellet was resuspended in 0.25 M sucrose in DEVbuffer (30 mM Tris-HCI (pH 7.8), 0.25 M NaCl, 2 mM EDTA)and purified on a 0.6/1.2 M sucrose-DEV step gradient,centrifuged in a Beckman SW28. 1 rotor at 25,000 rpm for 30minutes at 4°C. Pure mitochondria were then resuspended in an

* To whom corrcspondence should be addressed

.'.) 1993 Oxfortl University Press

812 Nucleic Acids Research, 1993, Vol. 21, No. 4

equal volume of a buffer containing 10 mM Tlris HCI (pH 8.0).0.5 mM EDTA, 1 mM DTT. 20% glycerol. and frozen at-700C.

Heparin-sepharose chromatographyMitochondrial lysis and heparin-sephairose (Pharmacia)chromatography were performed as described (2-0). After washingthe column with 0.2 M KCI in buffeer A the protcilns WCer elutedeither with 1.0 M KCI or with successive steps otf 0.4 NI and1.0 M KCl. Bradford (Biorad) positive fractions Were pooled.concentrated by ultrafiltration (Millipore). dialvsed ag.2ainst abuffer containing 10 mM Tris-HCI (pH 8.0) 0. 1 NI KCl. 5mM MgCl, 1 mM DTT, 0.5 mM EDTA and50ct5 gflycerol.and stored at -20°C.

Recombinant clonesAM65, a genomic subclone of the A6/COIII gcne boundaryregion of S.purputratlus mtDNA (see Fig. 1)was provided by DrA .G. Mayhook.Oligonucleotides and DNA probesOligonucleotides were synthesized on an Applied Biosy stemlls 391DNA synthesizer, using phosphorimidite chemistry. andsubsequently purified on 20% denaturing polvacrylamide gels.Complementary single-stranded oligonucleotides 5'-AA.ATAG-GAGTAGCCTGCATCCAAGCATAT-3' (top) and 5'-AATA-CATATGCTTGGATGCAGGCTACTCC-3' (bottomn) wereannealed by incubation at 65°C with gradual lowering of thetemperature to 20°C, in a buffer containing 20 mM Tris-HCI(pH 7.6). 5 mM MgCI2, 0.1 mM DTT and 0. 01 mM EDTA.For modification interference experiments, the double-strandedoligonucleotide was cloned into the StinI site of pSP6'T7 (BRL)using standard techniques (22). Restriction fr-agimient and double-stranded oligonucleotide probes were labelled by filling in the3' recessed ends with appropriate radionucleotides (NEN. 3000Ci/mmol), using Klenow enzyme (Boehringer Mannheimn).

Construction of binding site variantsTo engineer mutations in the binding site a bottom-strandoligonucleotide containing degeneracies at three positions wassynthesized [5 '-AATACATA(T/G)GCTTGGAT(G,'C )CA(G/T)GCTACTCC-3'] and annealed to the top strand of theoligonucleotide described above. After filling in the 3' protrudingends the oligonucleotide mixture was cloned into the StinI siteof a modified plasmid pSP6-T7. containing an inactivated BawtiiHlsite. Recombinant plasmids were identified and sequencedl. FoIthe construction of spacing variants the Ba;uHI site in variantD (see Table 1) was cleaved, and partially or- completely fillediin using Klenow and various combinations of dNTPs. Wherenecessary, the remaining single-stranded DNA tails were digestedwith mung bean nuclease (Boehringer Maninheiimi) prior toligation. For use in electrophoretic mobility shift assays. theinserts of all variants were excised by digestion with EcoRI andXbaI, gel-purified, and radiolabelled as desci-ibed above.

Electrophoretic mobility shift assays (EMSA) and DNase IfootprintingEMSAs were carried out in a total volume of 15 Ad. containineii1 ug of poly(dA-dT).poly(dA-dT) as nonspecific comiipetitorDNA, 0.2 ng of end-labelled DNA probe and 4 ,Lig of the 0.2 -0.4M KCI heparin-sepharose protein fiaction in a reaction consistineJof 20 mM HEPES-KOH (pH 7.9)1. 5 miM MLClh. 0.5> mM

EDTA, 50 maM KCI. I imiM DT'I and 7. S I cciol. The bindingreactions werc incubated at 20CC foI 30 imlinutes ancd then loadeddirectlN onto a 57% polvacrylamide (30 0.8) gel prepared in0.5xTBE (pH 8.3). which had becn pi-c-cooled at 4°C andprerun foi- 30 minutes at 200 'V. Follokv inm electrophoresis theeJels were diried and exposed to \-ra ilim (Kodak). For DNaseI footprinting, assays the complexes wcerc diLcsted with 25 n, ofDNase I (Worthington) in 35 jil of 5 mNM M(eCl an1d 5 mMCaClV solutioni for 15 seconds (at 20(CC. )Nase I was inactivatedby phenol. Rcactions \\ crc extrclted once \\ ith phenol-chlorotformll(5:I) acnd ethanol precipitated. The [)NA \Itas dissolved in 3 pAfoirma-imailde loladine buffer- anid clectroplioresed on an11 8 %sequencing gel. Maxam & Gilbeit chemistr\ (23) was perfoirmiedon Hvbond M&G paper. according to the instructions providedbx Ameicr-shaLml.

Protein- DNA contact analIsisThvi-nine- Cand guanine-specitic modifications were cai-ried out asdescribecd hv TIuSS C/ (a!. (24) and MNlaxam and Gilbert (23).respectivsely, with ImlinoI modifications. Both KMnO4 and DMSreactions were stopped bh chromatograph\ on Sephadex G-50 spuncolumiins equiliberated with TEN (10) imM Tris-HCI (pH 8.0).mM EDTA. 0. 1 M NaCl). The imioditficl DNA w\as purified

anid LseCd intl a five-fold scale-up of tEMSA as described ahove.

a -

N D . yt -

r~~~~~~,DrI

~(~f d r + L?3D

OCt11 07<(I(i(E';0 R c9010

A%1o6

Figure 1 (O)-ainlihationl th't pn7)-illt(a.t itotlhondliial eflonimC (a) S.hrclltiauephysical iiin p ot sCL. Li -chin nillDNA show inm thc t-RNNA 11(1 protcin coding efnesWxith thciu- if-nscriptional ociItifltitions indciC'act)dh\ h iltl ainows -I'hlC CpliC inltllOmttilns ate-c dcI0teCd LISinl" thte mammaliII1in 1oni.r t Itt-1v titc IclCIit' -sti-and ot icT-iiss y.() nd thic (I-otlincilt ot-Winiiiof i, it and i ntllt'iSis CtLtit-iCCI in C tliier'stidicic is ( ) w\ ith tilk clicction ot sindic sis ii iioth cisc s showi ll h! holditi. s (hb) NIip ot lonC ANN165i LiSCd in this stid(.% itil dieL "eCInCSe ilndiemitcciAh ATP ,iiItfMsC sUbuni1it h) mnld C011 sitO.lit it/H C\XICLisCsCUhClii 111

anild cst izCtiot sitcs inCi tcCI The.Cmilitlinls A,ttic puokli5ik i thlC pSP64 iCettrtIC shios LI, a htiikcn litlc rhc mapt co tdinati k)!tilC '-tOIoll ie .is showlnh.scI toii 1'i. 'I

Nucleic Acids Research, 1993, Vol. 21, No. 4 813

The bound and free DNAs were localized by autoradiography,excised, and eluted from the gel slices. Both KMnO4- and DMS-modified DNAs were cleaved with 10% piperdine at 90°C (23)and electrophoresed on a 20% sequencing gel.

RESULTSLocalization of the mtPBP1 binding siteIn order to investigate whether the major replication pause regionof sea urchin mtDNA contains binding-sites for sequence-specificDNA-binding proteins, the entire EcoRI-HindIll fragmentencompassing the A6/COIII junction region, from clone AM65(Fig. 1) was tested in EMSA, in the presence of a large excessof poly(dA-dT).poly(dA-dT), using the blastula stagemitochondrial lysate fractionated on heparin-sepharose. Thisresulted in the formation of a single complex whose specificitywas established by using increasing amounts of cold self-competitor, and monitoring the proportional loss in the intensityof the retarded band (data not shown). The binding site wasfurther investigated using shorter restriction fragments derivedfrom AM65. From the three fragments tested (i.e., EcoRl- AvaIl,

Figure 2. Determination of the protein binding site for mtPBPl (a) EMSAs usingthe 112 bp labelled AvalI-DraI fragment from clone AM65 (see Fig. 1). Variouseluates from the heparin-sepharose column were used as follows: no protein (lane

1), 4 yg 0.2-1.0 M KCI eluate (lane 2), 4 ytg 0.2- 0.4 M KCI eluate (lane 3)and 4 ,ug 0.4-1.0 M eluate (lane 4). (b) DNase I footprinting of the sense andantisense strands of the EcoRI- HindiJ fragment from AM65. Reactions contained:no protein (lane 1), 4 yg (lane 2) and 8 jtg (lane 3) of proteins from the 0.2-0.4M KCI eluate from heparin-sepharose. Footprinted regions are shown in brackets.

AvaII-DraI and Dral-HindIlI), only the AvaII-DraI fragmentproduced a band shift in the presence of a large quantity of thenonspecific competitor (Fig. 2a). Finer fractionation of theextracts on heparin-sepharose columns showed that this factorelutes in the 0.2-0.4 M KCl step.To localize the protein-binding site DNase-I footprinting was

carried out using the EcoRI-HindIHI fragment from AM 65,labelled on either strand. This experiment (Fig. 2b) gave afootprint on both strands, near to the 3' end of the A6 gene (nt9259 - 9284), which was resistant to competition from poly(dA-dT).poly(dA-dT).

EMSAs with the binding-site oligonucleotidesTo confirm the location of the protein binding site inferred fromDNase I footprinting, and to characterize further the protein(s)involved, a double-stranded oligonucleotide (oligo-PBP1)containing the sequence of the footprinted region in AM65 (nt9255 -9289) was synthesized (Fig. 3a). After radiolabelling,oligo-PBP1 was used in EMSAs in the presence of various proteinfractions. These experiments (Fig.3b, lanes 1-3) showed clearlythat a protein in the 0.2-0.4 M fraction binds to oligo-PBPIin the presence of a large excess of poly(dA-dT).poly(dA-dT).Although the binding reactions for EMSAs were carried out in50 mM KCI, increasing the salt concentration to 1.0 M did nothave a significant effect on the affinity of the protein for its site(data not shown).The thermostability of mtPBPI was also studied by treating

aliquots of the 0.2-0.4 M fraction at 20, 37, 50 and 100°C priorto using them in EMSA with oligo-PBPI. This experimentshowed mtPBPI to be completely resistant even to boiling, as

5'-AAATAGGAGT AGCCTGCATC CAAGCATATG TATT-3'3'-TTTATCCTCA TCGGACGTAG GTTCGTATAC ATAA-5'

Figure 3. EMSAs with mtPBP1 binding-site oligonucleotide. (a) Sequence ofoligonucleotide oligo-PBP1, derived from the DNase I footprint. (b) EMSAs usingoligo-PBP1 and various protein fractions. Lane 1: no protein; lane 2: 4 yg 0.2-0.4M KCI eluate; lane 3: 4 Ag 0.4-1.0 M KCI eluate; lanes 4-8: 4 /g heat-treatedprotein fractions. Proteins from the 0.2 -0.4 M KCI eluate were heated at theindicated temperatures for the indicated lengths of time, and immediately placedon ice for at least 10 minutes prior to incubation with oligo-PBPI. Lane 4:unheated; lane 5: 20°C for 30 min; lane 6: 37°C for 30 min; lane 7: 50°C for15 min; lane 8: 100°C for 5 min.

814 Nucleic Acids Research, 1993 Vol. 21, No. 4

Figure 4. Interterence with imtPBPl hinding, ht DMS aind KNInO4 ntaoditcation. (a) Icthvlation intcrt'ercnicc arnialysis ot' the scnsc and antisetse strands ol' the clonedEcoRI-BanolHI fraggment containing oligJo-PBPI and (b) oxidaltion interference anatlsis of the antisense strand of' the samic fraem-crten. Atcr modification with therespective reagents DNAs were used in a prepar-ative EMSA. Frec (F) and bound (B) DNAs were recvcred and, together with the input (I) DNA. were cleavedwith piperidine and analyzed on a 2O(' sequencing gel. Moditied residucs which interf'ere with mtPBPI binding are circled. (c) Summnary of the modification interferencedata: modified residues front oligo-PBPI which interftere with nmtPBPI hindine are 'asterisked. The imperf'ect repeats in which the contacted residucs lic are shownby bold lines.

no decrease in its activity was detected after heat treatment(Fig. 3b; lanes 4-8). In subsequent experiments this propertyof mtPBPl was exploited, and the 0.2-0.4 M fraction whichhad been boiled and centrifuged to remove precipitated proteinswas used at a concentration of 0.8 mg/mbl providing a further-5-fold purification of mtPBPI Further purification by Mono-Schromatography, and Southwestern blotting, revealed mtPBPIto comprise a single polypeptide species of appar-ent miiolecularweight 25 kD (data not shown).

Chemical modification and interference analsisTo localize the nucleotides in the binding site acting as contactpoints for the protein we decided to carry out modificationinterference experiments on thynmine and guanine residues. Forthis purpose oligo-PBP1 was cloned into the SmaI site ofpSP6/T7, and after excision the resulting 57 bp EcoRI-BacnfHIfragment was radiolabelled on either strand and used in theseexperiments. Methylation interference experiments were carriedout to probe the protein contact points with the various guanineresidues in the binding site. Methylation of the guanine r-esiduesat position 12 and 24 on the sense strand, and at positions 13.14, and 25 on the antisense strand reproducibly interfered withbinding (Fig. 4a).

Similarly, C-5/C-6 oxidation interference experimilents werecarried out to assay for protein contacts with the thymine residuesin the binding site. For this, thymine residues on denatured DNAwere modified with KMnO4 in the presence of 30 mMTris-HCI (pH 8.0), which according to Truss et al. (24) modifiesall thymines regardless of nearest neighbor sequence effects.Under these conditions we found the T at position 15 on the sensestrand to be hyporeactive to the oxidizinel a2lent. and, as a

Table 1. Sense-straidt DNA sequenccs 4)t the hindling-site and spacing variants

Variant DNNA Sequence

- AAATAGGAGTAGCCTGCATCCAAGCATATGTATT

A AAATAGGAGTAGCCTGCAICCcAGCATATGTATT

B AAATAGGAGTAGCtTGCATCCAAGCATATGTATT

C AAATAGGAGTAGCCTGgATCCAAGCcTATGTATT

D AAATAGGAGTAGCCTGgATCCAAGCATATGTATT

F AAATAGGAGTAGCCTGCATFCCAA-CATAT'CTATT

F AAATAGGAGTAGCaTGCATCCAAGCATATGTATT

i, AAA-AGGAGTAGCaTGgAT'CCAAGCATATGTATT

NI AAATAGGAGTAGCCTGAGCATATGTATT'

N2 AAATAGGAGTAGCCTCAAGCATATGTATT

N3 AAATAGGAGTAGCCTGCAAGCATATGT'AT¶I

N4 AAATAGGAGTAGCCTGGCAAGCATATGT'ATTI

N5 AAATAGGAGTAGCCTGGCCAAGCATATGTATT

N6 AAATAGGAGTAGCCTGGACCAAGCATATGT'ATT

N9 AAATAGGAGTAGCCTGGATATCCAAGCATATGTATT

NlI AAATAGGAGTAGCCTGGAT'CGA'TCCAAGCATATGTAI"

consequence, uninformative (data not shown). Other thymineresidues on the sense strand were mnodified normally but hadno effect on protein binding. On the antisense strand modificationof the thymine residues at positions 11 and 26 strongly interferedwith binding (Fig. 4b). The results of imodification interferenceanalysis are sUMmarized in Fiflure 4e. It is immediately apparent

Nucleic Acids Research, 1993, Vol. 21, No. 4 815

a UNCHANGED A B C

b

Figure 5. EMSAs with variant binding-site probes. (a) Binding-site variants A-G(shown in Table 1) along with the fragment containing the wild-type sequence

(unchanged) were used in EMSAs, either with (lanes 1) no protein, or (lanes2-4) 0.8 tg of boiled 0.2-0.4 M KCI protein fraction, in the presence of 0,0.1 and 1 tzg respectively of poly(dA-dT).poly(dA-dT) as nonspecific competitorDNA. (b) EMSAs using the eight spacing variants shown in Table 1. Bindingconditions were the same as described above.

that the contacted residues lie within two copies of the imperfectlyrepeated sequence AGC(C/A)T, separated by 7 bp. These repeatslie approximately one turn of the DNA double helix apart.

EMSAs with binding site variantsIn order to understand the consequences for mtPBP 1 binding ofvarious point mutations in the binding site, as well as the effectof altered spacing between the two imperfect repeats, weconstructed two sets of binding-site variants, using themethodology outlined in Materials and Methods. We obtainedprogrammed as well as unprogrammed mutations in the bindingsite, as summarized in Table 1, where mutated bases are shownin lower case. The relative strength of mtPBP1 binding to thevariant sites was assessed by EMSA (Fig. 5), carried out in thepresence of increasing amounts of the non-specific competitorpoly(dA-dT).poly(dA-dT). Variants A and D behaved the sameway as the original binding site, indicating that the C to Gmutation at position 17, which creates a BamHI site between theimperfect repeats, is inert. This BamHI site was used in thecreation of the spacing variants. The binding of mtPBPl to variantC, in which both copies of the repeated sequence are AGCCT,appeared to be slightly stronger than to the original site, in whichthe second copy is AGCAT. All other site variants boundrelatively weakly, as shown by their inability to bind mtPBPlin the presence of 1 ,lg of the nonspecific competitor. EMSAswith the spacing variants indicated that none of them could interactwith mtPBP1 to form a complex in the presence of 1 jig of thenonspecific competitor (Fig. 5b). The 7 bp spacing between therepeats is therefore critical for tight binding. Oligonucleotidescontaining either of the two half-sites in isolation were also boundonly weakly by mtPBP1, even in the complete absence of anonspecific competitor (data not shown), indicating that thepresence of both half-sites is essential.

DISCUSSIONIn this study we have identified a protein which binds specificallyand strongly to a DNA sequence close to the A6/COIII geneboundary of the sea urchin mitochondrial genome. Since thisregion contains a strong pause site for leading-strand DNAreplication (18), we propose to refer to the protein asmitochondrial pause-region binding protein or mtPBP1.

Modification interference experiments have identified thebinding site for mtPBP1 in S. purpuratus mtDNA as the sequence5'- AGCCT(N7)AGCAT-3'. All five guanine residues (on bothstrands), as well as the two antisense-strand thymines which arerepresented in both copies of the imperfect repeat, are involvedin binding to mtPBP 1. No information was obtained from theseexperiments, concerning the two antisense-strand thymineresidues. Although the oxidation of the C-5/C-6 double bond ofthese residues does not interfere directly with binding, it ispossible that the protein may utilize these residues indirectly, forexample by making contacts with the complementary adenines.The fact that these two A-T pairs add to the symmetry of thebinding site supports the view that they are likely to be involvedin recognition. In addition, the 7 bp spacing between the imperfectrepeats, approximately equivalent to a helical turn, is crucial forstrong binding.

Since the two half-sites are almost symmetrical, and bindingis enhanced by converting both copies to AGCCT, it is plausiblethat mtPBP1 interacts with its site as a dimer. This type ofinteraction is common among DNA-binding proteins, for example

816 Nucleic Acids Research, 1993, Vol. 21, No. 4

the catabolite activator protein (CAP) of E. coli, which binds asa dimer to two symmetrical half-sites separated by 6 bp, andyields a single retarded complex on EMSA (25). Like many suchproteins, the binding of mtPBP I is dramatically weakened if thespacing is perturbed. A binding site comprising short directrepeats with a constrained spacing resembles that for the phagelambda clI repressor (26), and also recalls the architecture ofthe binding sites for the nuclear receptor family (27).The binding site for mtPBP 1 is highly conserved in the related

sea urchin species Paracentrotus lividus (28). diverged from S.purpuratus by at least 40 Myr (29). suggesting that it is offunctional significance. More strikingly, the sequenceAGC(C/A)T(N)7AGC(C/A)T occurs once, and only once, in alldeuterostome mitochondrial genomes we have examined, thoughnot always in the same location. In mouse, as in sea urchins.it is found in the 3' end of A6, where amino acid sequenceconstraints arguably favour its occurrence. However, in humanand Xenopus mtDNAs it occurs in two unrelated regions of ND4.and in bovine mtDNA in the D-loop. We are unable to evaluatethe possible significance of this observation, other than to notethat the given sequence would be expected to occur by chanceonly about once every 250,000 base-pairs.The relationship between mtPBPI and DNA-binding proteins

studied in other systems remains to be determined. Itsthermostability recalls that of mtTF 1, although it does not formthe characteristic laddered pattern of complexes which mtTF Iexhibits on EMSA (2), and its binding-site preference appearsto be much higher. There is also no reason to associate it withthe 40 kD mitochondrial protein from P.lividus, studied byRoberti et al. (30), as there is no similarity in the binding sitesfor the two proteins.The location of the binding-site for mtPBPl does not suggest

any obvious role for this protein in transcription. Although wecannot exclude the possibility that it acts as a termination factor.we have been unable to detect any 3' ends of transcripts of theA6 gene, other than precisely at the gene boundary (31).A possible function for mtPBPI in DNA replication is

suggested by the proximity of its binding site to the major pausesite for leading strand replication, and a prominent lagging-strandorigin (18). We suggest, by analogy with the tus protein of E. coli.which functions in the termination of chromosomal and plasmidreplication (32), that mtPBPI, when bound, might stall anadvancing replication complex by antagonizing the action of itsassociated DNA helicase. This could facilitate the formation onthe displaced strand of the priming structure for lagging-strandinitiation, for example, by impeding the access of single-strandbinding protein, or by preventing interference from upstreamsequences. Alternatively, the local presence of DNA polymeraseand other components of the replication machinery might allowthe efficient initiation of lagging-strand synthesis. To test theseideas we are attempting to purify mtPBP1 to homogeneity. inorder to assay its activities ini vitro.

REFERENCES1. Fisher, R. P Topper.J. N. and Clayton.D A. 1987) Ce 50, 247 -250.2. Fisher,R.P. and Clavton. D.A. (1988) MWo!. Cc/I. Biol. 8, 3496-3509.3. Fisher,R.P.. Parisi,M.A. aind (la1vton.D. A. (1989) Gencs Dei, 3,

220212 17.4. ParisiMN1.A. and Clayton,[).A. i199 1) Science 252, 965 -969.5. Difflev.J.F.X. aInd Stillman.B. (1991) Pr-oc. k,itl. Icatd. Sti. USA 88,

7864 -7868.6. Fisher.R.P.. Lisowskv.T.. Breen,GA.NI. antcd CIaIvton,DA. (1991) 1. BRi!.

C/wein. 266, 9153-9160.7. Christianson.T.W. and Clavton,D..A. 11988) 0M1. Ce/i. Biol. 8, 4502-- 4509.8. KrusecB., Narasinmhan,N. and Attardi.G. (1989) Cell 58, 391 -3979. Jang.S.H. and Jachning.J.A. (1991) J. Bio(,l. Chem. 266, 22671 -22677.

10. Lisowskv,T. and Michaelis.G. (1988) MWo!. Gen. Genet. 214, 218 22311. Schinkel.A.H.. Groot-Koerkamp.M.J.A and Ta'EhaIk.HF. (1988) EMBO. .1J

7, 3255 3262.12. Chang.D.D., Fisher.R.P. and Clasyton.D. A (1987) Bioch/i,. Bioph.v.s. Aclr

909, 85-91.13. Clavton,D.A. (1982). Cell 28, 693 7()z14. Piko.L. (1969) Amt. Zoo. 9, 1118.15. Jacobs,H.T., HerhertE.R. and Rankinc.J. (1989) Nucleic A(cu/Ls PN 17,

8949 8965.16. Rinaldi.A.-M.. Carra,E. Salcher-Cillari.l. and Oli0va.A.O. (1983) Ce!l Bi,ol.

heit. Rep. 7, 211 218.17. Rinaldi.A.-M. and Giudice.G. (1985) In Ietz.C.B3. aInd Monro .A. (cds.).

The Biolo2v ot Fertilization. Acadermic Pr-ess. New York.vol. III. pp.367- 377.

18. Mayhook.A.G.. RinaldiA-M. and JacohK.HH.T. (1992) Pr-oc. ROY. Soc. I,otl/.B. 248, 85-94.

19. Calzone.F.J., Hoog.C.. Teplow.D.B.. Cuttinv.A.E.. Zellelr.R.WA..Britten.R.J. and Davidson.E.H. (1991) Development 112, 335-350).

20. Shuev,D.J. and Attardi.G. (1985) J. Bio!. Chietm. 260, 1952--- 195821. Jacobs.H.T.. Elliott.D.J.. M/lath.V.13. 'and Farquharson,A. (1988) J. o()/.

Biol. 202,185 - 217.22. Maniatis,T.. Fritsch,E.F. and Sambrook.J. (1982) MoleculaIr Cloning. A

Laboratorv Manual. Cold Spring Harbor Unliversity Pr-css. Cold SpriilngHarbor.

23. Maxam,A. and GilbertoWN. (1980) Mfethodi En-,vmol/. 65, 499-56(024. Truss,M.. Chalepakis.G. .and Beato.lI. (1980) Prfoc. Nrw. A4cd. Sc!. US,4

87, 7180-7184.25. Gartenberg.M.R. and Crothers.D.M. (1988) NVtwre 333, 824 82926. Ho.Y. and Rosenherg.M. (1982) Ann. MWicr-obiol. 133A. 215-238.27. Kliewer. S.A., UmnesonoK. Hevman. R.A.. Mlangelsdorl.D.J.. Dycki.

and Evans,R.M. (1992) Proc. Ntrl. Aca.el Sci. USA 89, 1448 1452.28. Cantatore,P.. Roberti,M.. RainaldiCG., CadalCta,NI.N. and Saccone.C. (1989)

J. Biol. Chem. 264, 10965- 10975.29. Smith,A.B. (1988) Mol. Biol. Evol. 5, 345- 365.30. RobertiNi.. Mustich.A.. GadaletaI,M.N and CaIntatore P. (1991) N/cetl(

Acids Res. 19, 6249-6254.31. Mayhook A.G. (1992) Ph. D. Thesis. Universits of Glaseoo.32. Lee.E.H.. Komber-A.. HidakaN.. Kohav"ashiT. aInd Horiuchi,T. 1989)

PIroc. Natl. .Actd. Sci. USIA 86. 9104 91(18.

ACKNOWLEDGEMENTSWe are indebted to Frank Calzone (University of California.Irvine) for the provision of frozen lysates of S.purpuratusblastulae, and of lab facilities for preparing mitochondria fromthese lysates. We also thank Andrew Mayhook, for supplyingthe clone AM65, which he constructed. We are also grateful forfinancial support from the Royal Society, Medical ResearchCouncil, the European Community and NATO.