Sequence requirements of the Epstein-Barr virus latent origin of DNA replication.

JOURNAL OF VIROLOGY, Mar. 1994, p. 1913-1925 Vol. 68, No. 3 0022-538X/94/$04.00+0

Copyright (C 1994, American Society for Microbiology

Sequence Requirements of the Epstein-Barr Virus

Latent

Origin

of

DNA

Replication

SARAH HARRISON, KIMBERLY FISENNE, AND JANETHEARING* Departmentof Microbiology, Health SciencesCenter, State UniversityofNew Yorkl

StonyBrook, New York 11794-5222 Received 14 October 1993/Accepted 8 December 1993

The Epstein-Barr virus (EBV) latent origin of DNA replication (oriP) is composed of two elements that containbinding sitesfor the sole viral gene product required for latent cycle replication, EBNA-1. One of these elements,regionI,functions as an EBNA-1-dependent enhancer for RNA polymeraseII-transcribedgenes, may play a rolein plasmid segregation, and is required for origin function in B cells latently infected with EBV. The second element,regionII,contains or isverynear the site ofinitiation of DNA replication. A genetic approach wastaken to determine the contribution of the EBNA-1 binding sites in oriP to origin function. Although region Iis required for the transient replication of plasmids bearing region II in EBV-infected B cells, a plasmid

lackingregion I but containing regionII,was observed toreplicate transiently in bothD98/RajiandHeLa cells

expressingEBNA-1. Thus,binding of EBNA-1 to region I is not absolutely required for the molecular events thatleadtoinitiation ofDNAreplication atregionII. Site-directed mutagenesis of the four EBNA-1-binding sites inregion II, individuallyandinvarious combinations,demonstrated that only twoEBNA-1-binding sites

arerequiredforregionIIfunction. The resultsobtainedwith these mutants, togetherwiththeanalysis of the replicativeabilityof plasmids containing insertions between EBNA-1-binding sites, suggested that the spatial relationship of the two sites is critical. Mutants that contain only two EBNA-1-binding sites separated by 26 to31bp in region II were not maintained as plasmids over many cell generations and were greatly reduced in their ability to replicate transiently inD98/Rajicells. TheEBNA-1-induced bending or untwistingof the DNA inEBNA-1-binding sites 1 and 4 in region II did not, however, demonstrate thisspatialconstraint. It may be concluded from these results that specific protein-protein interactions between EBNA-1 and/or between EBNA-1 and a cellular protein(s) are required for origin function.

In most Epstein-Barr virus (EBV)-infected B cells, viral

gene expression is limited and infectious virus is rarely

pro-duced(10). The EBVgenomesinsuch latently infected B cells

arepresent asmulticopyplasmids (38), and replication of these

EBV plasmids exhibits several important characteristics in

common with the replication of cellular chromosomes. EBV

chromosomes replicate once per cell division cycle (1), and replication occurs at a discrete time during the S phase (22, 23). Two additional features ofEBV latent-cycle replication make thisan attractive system to study the regulated replica-tion ofDNAineukaryotic cells. Onlyoneviralgeneproductis required for latent-cycle replication, the EBV nuclear antigen

1 (EBNA-1 [37, 53]). Thus, all other proteins required for replicationareprovided by the host cell.Additionally, replica-tion initiatesat awell-definedcis-acting element, oriP, and this

originofDNAreplicationdirects theregulated replication of recombinantplasmids in primate cells expressingEBNA-1(37, 52, 53). Investigation of the maintenance of EBV

chromo-somesin infected cellswill also further ourunderstanding of thevirus-host interactions underlying the lifelong human in-fections established by thisherpesvirus.

Previous studies have established that oriP is composed of

twoessential elements (37, 43). Region II, also referredto as

the dyad symmetry element, contains four binding sites for EBNA-1 (2, 41)and isat, or very near,thesite of initiation of latent cycle DNAreplication (17, 51). Region I, also termed the family of repeats, is located approximately 960 bp from

*Correspondingauthor. Mailingaddress:Department of

Microbi-ology, State Universityof New York, Life Sciences Building,Room

280, Stony Brook, NY 11794-5222. Phone:(516)632-8778. Fax:(516) 632-8891.

regionIIandcontains 20EBNA-1-binding sites (2,41). Region

Iisabsolutelyrequired for the stable maintenance ofplasmids bearingregionIIincellsexpressingEBNA-1 (37, 43),contains the termination site forreplication (17), functions as a

tran-scriptional enhancer element for RNA polymerase Il-tran-scribed genes (42), and allows for the prolonged retention of physically linkedDNA(33).Allof these activities attributedto

region IrequireEBNA-1 (11, 33, 37, 42, 43, 53).

Region I contains 20 EBNA-1-binding sites that are

sepa-rated by either 26or30bp (41).In contrast,regionIIcontains fourEBNA-1-binding sites. Sites 1 and 2areseparated by21

bp,asaresites 3 and 4; sites2and 3areseparated by 30 bp(41)

(Fig. 1). It is likely that differences in both the number and

spatial arrangement of EBNA-1 sites in regions I and II

contribute to the distinct biological properties of these

ele-ments.TwoorthreetandemcopiesofregionII canfunctionas anenhancer element foranRNApolymeraseIIpromoter,but the levelof enhancement is much lower than thatprovided by

asimilar numberof sites derived fromregionI(51). Similarly,

two and threetandem copies ofregion II activate replication

from regionII to alesserextentthan do fiveorseven repeats

from region I (51). Differences inthe interaction ofEBNA-1

withregionsIandIIhavebeen detectedbyusing KMnO4as a chemical probe for distorted DNA. EBNA-1 can induce the

bendingoruntwistingof theDNAduplexattwositesinregion

II (nucleotides [nt] 9046 and 9110 [Fig.

1])

but is unable to

distort the duplex in region I (16, 26). The distortion ofthe DNA may be important for the events that lead to

duplex

unwinding andreplicationinitiation (16,

26).

The presenceof

theKMnO4-reactive thymines in

EBNA-1-binding

sites

sepa-rated fromanadjacentsiteby21bp(Fig. 1)and theirabsence in EBNA-1-binding sites separated by26 or30 bp

suggested

1913

on November 9, 2019 by guest

http://jvi.asm.org/

1914 HARRISON ET AL.

4

3

V\/

2

1

\V

N

5'-CCCTA&!TCG&TAQCATATQCTTCCCGTGQGTA CAT GCTAT AATTAGG?TAVTGTAGTATATATACTACTACCCGGGATATGCTACCCGTTTADGTA-3'

3,-4;GATTAAGCTATCGAACCCACA¢TATACGATAACTTAATCCCAATCAGCCCATAT CCTTCGTATACGATGGGCAASCCCAA-5'

x7Z74KZZ7I1

21

bp

X

A

/

21

bp

30

bp

FIG. 1. Nucleotidesequenceof oriPregionII.The sequence of EBV nt 9024 to 9135(B95-8isolate)ispresented(4).Thepositionsof the four EBNA-1binding sites(numbered accordingtothe convention inreference41)are indicatedbythe numbersoverthe sequence andbythebars between the DNA strands. The number of basepairsseparating adjacentsites are indicatedatthe bottom of thefigure. Thebars above and beneath the DNA sequence represent nucleotides protected from DNaseIdigestion byEBNA-1(26).Breaksatthe ends of thebarsrepresentambiguities intheinterpretation ofthefootprintsbecause of the absence of nucleasecleavagesin naked DNAcontrols(26).The arrowheadsat nt9046(top strand) and nt 9110(bottom strand)indicate thepositions ofthymines oxidized by KMnO4in the presence of EBNA-1(16, 26).Thepositions of EBNA-1 dimers are representedby the ovals.

thatinteractions betweenEBNA-1 boundtosites 1 and 2 and between EBNA-1 bound to sites 3 and 4 were required for

EBNA-1 to bendor untwist the DNA.

To begin to understand how EBNA-1 interacts with the cellular replication machinery to direct replication of

oriP-bearingplasmids, agenetic approach has been usedto deter-mine the importance of EBNA-1-binding sites in oriP. The experiments presented here demonstrate that the molecular

eventsthatleadtothe initiation ofDNAreplicationatregion

II do not require, mechanistically speaking, the binding of

EBNA-1 to region I. These data provide additional evidence for a regulatory role for region I in oriP function (15). Experimentsarepresented that demonstrate that only twoof the fourEBNA-1-binding sitespresentinregionII, separated bytwo helical turns, are necessaryfor region II function and

suggest that specific interactions between EBNA-1 bound to sequencesinregion II arerequired forfunctional interactions with the cellular replication machinery.

MATERIALSAND METHODS

Cells and plasmids. D98 and D98/Raji cells (18) were

cultured in Dulbecco's modified Eagle's medium supple-mented with 10% fetal bovineserum (WhittakerBioproducts,

Walkersville, Md.). HeLa cells were grown in Dulbecco's

modified Eagle's mediumcontaining 10% iron-supplemented

calf serum (Whittaker Bioproducts). All cellswere grown as

monolayer cultures at 37°C in a humidified atmosphere

con-taining 5% CO2.

Plasmidsbearingintact oriP(pHEBo-1 andpHEBo-1.1)and deletion derivatives lackingeither region I or region II (pRII and pHEBo-ldlDS, respectively) have been described

previ-ously (26, 48). All recombinant plasmids used in replication assays were isolated from Escherichia coli DH1 cells (dam').

Transientexpressionof

EBNA-I

in HeLa cells was obtained by

transfection ofaplasmid, pAd/EBNA-1.3,in which synthesis of EBNA-1 is directed by the adenovirus major late promoter.

TheEBNA-1-codingsequences(B95-8EBVisolate),together

with a polyomavirus late mRNA donor splice site, 1,037-bp

intervening sequence, and acceptor splice site, were removed from pMLPyA2K' (27) by digestion with BclI, followed by repairofthe ends with Klenowpolymerase anddigestionwith HindIII. This fragment was introduced into an adenovirus expressionvector, pNL3-C (kindly provided by R. Schneider,

New York University Medical Center), such that a cDNA containing the complete tripartite leader sequence present at the 5' end of all mature late mRNAs derived from the adenovirus major late promoter was positioned upstream of thepolyomavirus-derived 5'splice site.Arecombinant adeno-virus containing this hybrid transcription unit directed the

synthesisoffull-length EBNA-1 (28;unpublished data).

Mutagenesis of oriP.Double-point andinsertion mutations were introduced into region II by oligonucleotide-directed mutagenesis of this component of oriP subcloned in pBlue-script KS(Stratagene, LaJolla,Calif.) aspreviouslydescribed (26, 34). Sequences of themutagenicoligonucleotidesused for

creating the mutations are shown in Table 1. The mutated EBV sequences were substituted for wild-type sequences in

pHEBo-1.1, and the sequence of EBV nt 8992 to 9156 was determined for eachreconstruction(44)to ensurethat the only

mutationspresentin regionII werethose intended.

Nuclease protection and KMnO4 footprinting assays.

DNase I protection and KMnO4 footprinting assays were

performed as previously described with full-length EBNA-1 isolated from recombinant baculovirus-infected insect cells

(26). Immunoaffinity-purified EBNA-1 was used in some

ex-periments, whereas EBNA-1 further purified by

sequence-specificDNAaffinitychromatographywasusedin others (see figure legends). The quantity of EBNA-1 required to obtain

partial and complete protection of sites in region II in wild-type oriPwas determined for each preparation by DNase I

footprinting. Not all ofthe EBNA-1 present in the

immuno-affinity-purifiedmaterialboundtoDNA.Thus,more

immuno-affinity-purified EBNA-1 thanDNA affinity-purified EBNA-1 was required to achieve complete protection of the binding sites in oriP. All footprinting experiments involved plasmids bearing both regionsI and II (derivativesofpHEBo-1.1).

---x- Ix- an .

r

J.VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:2.612.66.552.71.246.2]

EBV LATENT ORIGIN 1915

TABLE 1. Nucleotide sequences ofoligonucleotidesused for site-directed mutagenesis of oriP region ll

Mutation Oligonucleotidesequence'

dpml 5'-CCTAAACGGGTCGCATATGCGTCCCGGGTA-3' (9111, 9120)

dpm2 5'-CCCGGGTAGTCGTATATACGATCCAGACTAACC-3' (9090, 9099)

dpm3 5'-CCAGACTAACCCTAATTCAATCGCATATGTGACCCAACGGGAAGCATATGC-3' (9057, 9066)

dpm4 5'-GCATATGTTACCCAACGGGACGCATATGCGATCGAATrAGGG-3' (9036,9045)

in1/2[5] 5'-GCATATGCTTCCCGGGATCCGTAGTAGTATATACTATCC-3' (9105)

in1/2[10] 5'-GCATATGCTTCCCGGAGTCTAGACTGTAGTAGTATATACTATCC-3' (9105)

in2/3[5] 5'-CCAGACTAACGGATCCCTAATTCAATAGC-3' (9079)

in213[10] 5'-CCAGACTAACAGTCTAGATACCTAATTCAATAGC-3' (9079)

in314[5] 5'-CCCTAATTCAATAGCATATGTTACCCATCTAGACGGGAAGCATATGC-3' (9052)

in3/4[10] 5'-CCCTAATTCAATAGCATATGTTACCCAAGTCTAGACTACGGGAAGCATATGCTATCG-3' (9052)

aThe nucleotide changes at the +51-5positions of the EBNA-1 recognition elements (2) and the nucleotides inserted betweenEBNA-1-binding sites in the mutagenicoligonucleotidesareunderlined. The numbers in parentheses indicate the positions in the EBV genome (numbered according to the B95-8 isolate [4]) that arealtered in the dpm mutants and the position of the insertions in the in mutants. Plasmids containing a 10-bp insertion between sites 2 and 3 also contain a G-to-T

changeat nt9078(boldface type)owing to an error in design of the oligonucleotide.

Replication assays. The ability of plasmids bearing wild-type and mutated oriPtoreplicate in vivo in D98 and D98/Raji cells wasdetermined byatransient-replication assay. Briefly, 10 ,ug ofplasmid DNA was introduced into cells by CaPO4 copre-cipitation (19), and low-molecular-weight DNA(29),digested with bothDpnI and Clal, was analyzed by Southern blotting with

32P-labeled

pHEBo-1.1 sequences generated by the ran-domprime method (13) aspreviouslydescribed (26).

The ability of plasmidsbearing wild-type and mutated oriP

toreplicate in HeLa cells was also determined by a transient-replication assay, as described above, with several modifica-tions. Increased quantities of either 1.1,

pHEBo-ldlDS, or pRII (20 ,ug) were coprecipitated with 20 ,ug of an EBNA-1expression vector, pAd/EBNA-1.3, or the expression vector lacking EBNA-1 coding sequences (pNL3-C).

Long-term maintenance of plasmids containing wild-type and mutated oriPwasassayed by introduction ofplasmidDNA

byCaPO4coprecipitation intoD98/Raji cells and selection of transformantsby culturing the cells in medium containing0.3

mgofhygromycinB(Sigma Chemical Co., St. Louis, Mo.) per ml. Low-molecular-weight DNA was isolated from pooled

transformants after 20or morecellgenerationsinthe presence ofhygromycin B, digested withClaI,andanalyzed by Southern blotting(26).

The results of transient-replication and long-term plasmid maintenance assays were quantitated by scanning the

mem-branes withanAmbisradioanalyticimagingsystemorbydirect scintillation counting of bands excised from the membranes.

RESULTS

RegionI-independenttransientreplicationof oriPregionII. Inadditiontobeing essential forthelong-term maintenance of

plasmidsinEBNA-1-expressingcells,oriPregionIisrequired

for the transientreplicationofplasmids containingregionIIin

the EBV-positive Burkitt's lymphoma cell line Raji (43, 51).

Not all of the EBNA-1-binding sites are required for this function ofregion I. SevenEBNA-1-bindingsites derivedfrom regionIfully restored transientreplication in anoriP-plasmid

lackingregionI,whereasplasmid bearingthree tandemcopies

ofregionIIinplaceofregionIreplicatedto70% ofaplasmid

containing intact oriP (51). We have observed that the

tran-sientreplicationofaplasmidwithadeletion ofregionI(pRII)

inRajicells isjust barely detectableoverbackground(datanot

shown). However,asshown inFig.2,regionI was notrequired

for transientreplication ofa plasmid containing region II in

D98/Raji cells. D98/Raji is a somatic cell hybrid formed between the D98 epithelial cell line and Raji cells

(18).

EBNA-1, encoded by the EBV genomes contributed by the Raji component of this cell hybrid, allows for the transient replication and long-term maintenance of plasmids bearing oriPinD98/Rajicells (37, 53).Intheexperimentwhose results

6 6

o o

aU

c c

=r

X X cn

-r' W W Cl)

m I I D

a ao a a CWo

a o0 X

U.J.a

Ul

1 2 34 56 7

FIG. 2. Transientreplication ofpRIIinD98/Rajicells. Aplasmid

bearing intact oriP(pHEBo-1; lane4) orderivatives lackingregionI

(pRII; lanes 1 and 2) or region II (pHEBo-1dlDS; lane 3) were

introduced into D98(lane 1)orD98/Raji(lanes2to4)cells. The level of plasmid replication was determined 93 h after transfection by digestion oflow-molecular-weight DNAisolated from the cells with Dpnl and Clal and analysis of the digestion products by Southern blotting with an EcoRI fragment of pHEBo-1, in common to all plasmidsused in theexperiment,astheprobe.Linearized

pRIl

(lane 5), pHEBo-ldiDS (lane 6),and pHEBo-1 (lane 7) (100 pg ofeach)

were includedasmarkers.The smallplasmidDNAfragmentspresent

atthe bottom of the blot representinputDNAthat didnotreplicate and therefore remained sensitive to Dpnl digestion as a result of methylation of its recognition sequence. Plasmids that replicated

becameresistant toDpnl digestion andwere linearizedbyClal.

VOL.68, 1994

on November 9, 2019 by guest

http://jvi.asm.org/

[image:3.612.58.562.84.196.2]

1916 HARRISON ET AL.

WT dIDS dIFR EBNA-1 - + - + - +

c] o o

U Lw

I_a I C:

w44

1 2 3 4 5 6 7 8 9

FIG. 3. Transient replication ofpRII in HeLa cells. HeLa cells

were cotransfected with plasmids bearing intact oriP (pHEBo-1.1;

lanes 1 and2),orlackingregionII (pHEBo-ldlDS; lanes3 and4)or

regionI(pRII;lanes 5and6,labeleddlFR)andanexpressionvector

containingthe EBNA-1openreadingframe(pAd/EBNA-1.3;lanes2, 4, and 6) or the vectoralone (pNL3-C; lanes 1, 3, and 5).Plasmid

replicationwasdeterminedasdescribed inthelegendtoFig. 2, except

that 32P-labeled pHEBo-1.1 was used as the probe. Lanes 7 to 9

contain 100 pg of linearized pHEBo-1.1, pHEBo-ldlDS, and pRII,

respectively. WT,wildtype.

are shown in Fig. 2, either a plasmid bearing intact oriP (pHEBo-1 [48]) or deletion derivatives that lack region I (pRII)orregionII(pHEBo-ldlDS)wereintroduced into D98

orD98/Rajicells andlow-molecular-weightDNAwasisolated

93 h later. The DNAwas restrictedwith bothDpnI andClaI,

and thedigestionproductswereanalyzed bySouthernblotting.

pRII replicated aswell aspHEBo-1 in this experiment (com-pare lanes2 and4). The inabilityofpRIItoreplicate inD98

cells is consistent with the previously described requirement

for EBNA-1 (lane 1) (37, 53). Furthermore, the inability of pHEBo-ldlDSto replicateinthis assay(lane 3)demonstrated

that the observed replication ofpRIIandpHEBo-1 required

the presence ofregion II.

EBNA-1canphysicallylink thetwoessentialcomponentsof

oriP (40) and form a DNA loop when regions I and II are

presentonthesamemolecule(15, 47).Theaffinityof EBNA-1

forregion IIisapproximately10-foldlower than itsaffinityfor

region I (14),and EBNA-1-region II complexesare sensitive

to the presence of nonspecific competitor DNA (47). The stability of EBNA-1-region II complexes, however, is

in-creasedwhenregion Iispresent,presumablyas aresult of the

formation of the DNAloop (47). The consequences of loop

formationinvivoarenot known,but loopingmayplayarole

inthe regulationoforiPfunction(15). BecauseD98/Raji cells carryEBVDNA(18),weconsideredthepossibilitythatregion

I,presentonendogenousEBVDNA,andregion II,presenton

thetransfected plasmid, facilitatedthe transientreplication of

pRIIin trans via interactions with EBNA-1. The experiment

whoseresultsareshowninFig.3demonstratesthatthis isnot

the case. In this experiment, pRII, pHEBo-ldIDS, and pHEBo-1.1 were cotransfected with either anexpression

plas-mid carrying EBNA-1 coding sequences under the control of the adenovirus majorlate promoter (pAd/EBNA-1.3; lanes

2,

4, and 6) or the vector lacking EBNA-1 (pNL3-C; lanes 1, 3,

and 5) into human cells that lack EBV DNA (HeLa). Both pHEBo-1.1 andpRIIreplicated transiently in HeLacellsto a

similar extent in an EBNA-1-dependent manner, whereas replication of pHEBo-ldlDS was undetectable. This

experi-ment further demonstrated that replication ofplasmids bear-ing region II can occur in the absence of regionI. Although the difference(s) between Raji versusD98/Raji and HeLa cells that accounts for the observed difference in replicative ability of pRII is notknown, theseexperiments show thatregionIisnot

absolutely required for the molecular events that leadtoorigin

activation and suggest that region I provides a regulatory function.

Cooperative binding ofEBNA-1 to sites in region II. Two-dimensional agarose gel analysis of replicative intermediates (17) and mutational analysis of oriP (51) have shown that the site of initiation of DNA replication is within, or close to, region II. The data presented in Fig. 2and 3 underscore the central role of region II in the EBV latent origin of DNA

replication. Region IIcontains four EBNA-1-binding sites (2, 41); however, all four sites are not required for originfunction (9, 52). Because previous genetic analysis of region II relied upon large deletions that removed EBNA-1 sites, as well as

flanking sequences that might be important (9, 52), and because internal deletions that removed sites 2 and 3 without affecting sites 1 and 4 were not analyzed, we introduced double pointmutations into each of the four EBNA-1-binding sites (dpml,dpm2, dpm3, anddpm4 [Table 1; Fig. 1]) to determine the contribution of each site to origin function. Each dpm

consisted ofbasepairtransversions at the +5/- 5 positionsof the EBNA-1 recognition element that greatly reduced the

binding of EBNA-1 to synthetic oligonucleotides carrying these mutations (2). Insertions of 5 and 10 bp were also introduced betweenadjacent EBNA-1 sites (in1!2[5], inl/2[10], in2/3[5], in2/3[10], in3/4[5], andin3/4[10] [Table 1; Fig. 1]) to

investigate the spatial arrangement of these sites. Chemical footprinting experiments have shown that thecontacts made by

EBNA-1 lie on one face of the DNA helix (2, 16, 26, 31).

Insertion of 5 bp between adjacent sites would cause the

contacts made by EBNA-1 bound to these sites to lie on

opposite faces of the helix, whereas insertion of 10 bp would

make only a small change in the orientation of the EBNA-1

dimersbut wouldincrease the distance between them. The ability of EBNA-1 to bind to sites in region II in plasmidscontaining dpml through dpm4 was determined by a DNase I protection assay (Fig. 4) (20, 26). The sequences

protected by EBNA-1 in a plasmid containing wild-type oriP

are shown in lanes 6 and 7. Mutation of site 1 abolished protection of site 1 aswell as the adjacent site 2 but did not

affect binding of EBNA-1 to sites 3 and 4 (lanes 9 and 10). Similarly, mutation of site4eliminated protectionofsite4 as

well asthe adjacent site 3 without affecting binding to sites 1 and 2 (lanes 18 and 19). Mutation of site 2 eliminated the binding of EBNA-1 to site 2 but did notaffect binding to the three nonmutated sites (lanes 12 and 13). Mutation of site 3 abolished EBNA-1 binding to that site but did not affect binding to site 4 (lanes 15 and 16). These results revealedthat thebinding of EBNA-1 to the outer paired binding sites (sites 1 and 2, and sites 3 and 4) iscooperative.

Although the binding of EBNA-1 to the outer paired

bindingsites is cooperative, theinteractions between EBNA-1 dimers bound to these sites that lead to cooperativebinding

J. VIROL.

.ow

ow

IV,.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.96.258.74.315.2]

EBV LATENT ORIGIN 1917

wr dpml dpm2 dpm3 dpm4

G A T

<_9'ffot

C 0 1 2 0 1 20o 1 2 0 1 2 0 1 2 o L9EBNA-1

7*;g= S

3

I3

4bS~~~~~4

2

-*<~

!

! 5

*-- 0

"b40~~~~~

_ffi-

4

r .

1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 1 6 1 7 1 8 1 9 20

FIG. 4. Protection of sequences in region II by EBNA-1 in

plas-mids containing individual EBNA-1 site mutations. Supercoiled pHEBo-1.1(lanes 5to7)orits derivativescontaining dpm1(lanes8to 10), dpm2 (lanes 11 to13), dpm3 (lanes14to 16),ordpm4 (lanes17to

20) (500ngofeach)wereincubated in the absence of EBNA-1(lanes

5, 8, 11, 14, 17,and20),orwith I ,ug (lanes 6, 9, 12, 15,and 18) or2 jig(lanes 7, 10, 13, 16,and19)of EBNA-1 purifiedfrom recombinant baculovirus-infected Sf9 cellsby immunoaffinity chromatography (26).

Thebinding-reactionmixturesweretreated with DNaseI,and

nucle-ase cleavage sites on the top strand ofregion II were detectedby a

primerextensionassay(20)witharadiolabeled oligonucleotide

com-plementary to EBV nt 9161 to 9175 (26). Lanes I to 4 contain

sequencingreactionsgeneratedwith thesamelabeledoligonucleotide primer (44).Thepositionsof the four EBNA-1 recognition elements

(2, 41)are indicatedbythe barsonthe left.WT,wildtype.

are tolerant ofchanges in the relative orientation and spatial

arrangement of the binding sites (Fig. 5). Analysis of the

sequences in region II protected by EBNA-1 in plasmids

containing insertion mutations revealed that introduction of either 5 bp (Fig. SA) or 10 bp (Fig. SB) between adjacent EBNA-1 sites did not affect the cooperative binding of EBNA-1 tosites 1 and 2 and tosites 3 and 4.

Two EBNA-1-binding sites suffice for region II function in

D98/Raji cells. The plasmids carrying EBNA-1-binding-site and insertion mutations inregion II also contain regionI and the bacterialgeneconferringresistancetohygromycin B under the control of the herpes simplex virus thymidine kinase promoter (48). We introduced these plasmids into D98/Raji cells, selected transformants thatwere resistant tohygromycin B, and examined low-molecular-weight DNA prepared from pooledtransformants after 20ormorecell generations under drug selection (long-term replication assay). Approximately equal (within twofold) numbers of transformants were

ob-tained with thewild-typeandmutatedplasmids,and each pool

analyzed representedtheprogeny ofapproximately 104

origi-nal transformants (transformation efficiencies, approximately 1%). The results presented in Fig. 6 show that none of the

single-binding-site or insertion mutations introduced into re-gion IIabolished the ability ofplasmidsbearingthese muta-tions to be maintained as extrachromosomal elements. The copynumber of transformantsestablished withdpm1and allof the in mutants was similar to that of a plasmid

bearing

wild-type region II (average copy number of 15 plasmidsper

cell transformed withpHEBo-1.1 [comparelanes 2and7with

lanes 3 and 8 to 13]). The plasmid copy number in cells transformed with dpm2, dpm3, and dpm4 was reduced (one

copy percell fordpm2anddpm4 transformants,two

copies

per

cell fordpm3 transformants

[lanes

4 to

6]).

Previous

experi-ments have shown that the plasmid copy number of

oriP-bearingrecombinant

plasmids

is determined

by

the amountof

plasmid DNA introduced into cells

during

the transfection

procedure and is not due to

amplification

of the inputDNA

(54).Becauseplasmids

containing dpm2, dpm3,

and

dpm4

are maintainedat one or more copies percell, we conclude that theseplasmidscontain functional

origins.

The

ability

of

plas-mids containing certain combinations of these

binding-site

mutations tobe maintained

extrachromosomally

in

EBNA-1-expressingcells

(data presented

below)

further supports this conclusion.

In addition to

unit-length molecules,

cells transformed

by

the

dpm

1 and

dpm3

mutant

plasmids

contain

plasmids

that

migrate between pHEBo-1.1 linear

species

and the more slowly

migrating

Clal

species

derived from

endogenous

EBV DNA in

D98/Raji

cells

(Fig.

6, lanes 3 and

5).

We have not

determined the structure of these

plasmids,

but

they

are

reminiscent of the plasmids detected in

D98/Raji

cells

trans-fected with

plasmids bearing

deletions

extending

into the

dyad

symmetryelement in

region

11

(9).

Because the individualmutation of sites 1 and4 eliminated

EBNA-1bindingatboth themutatedsite andthe

adjacent

site

(sites

2 and 3,

respectively

[Fig. 4])

in

vitro,

these results

suggestedthat onlytwoof the four

EBNA-1-binding

sites are

necessary for region II function. Additional support for this conclusionwas provided by the results of

replication

experi-ments performed with

plasmids containing

various combina-tions of the EBNA-1 site mutations

(dpm 1+2, dpm3+4,

dpml+4,

and

dpm2+3).

The sequences in oriP

region

II in

plasmidscontainingthesemutations thatwere

protected

from

DNase I digestion by EBNA-1 are shown in

Fig.

7. In this

experiment, the

protection

of site 1 at a lower EBNA-1

concentration and the

subsequent

protection

of site 2 at a

higher EBNA-1 concentration in

wild-type region

II are

evi-dent (lanes 6 and 7) and are consistent with the observed

cooperative

binding

of EBNA-1 to these sites

(Fig.

4).

The results with

dpml+2

and

dpm3+4

are identical to those

obtained with either

dpml

or

dpm4 singly (Fig.

4);

EBNA-1 failedtoprotectthe mutatedsites fromnuclease

digestion

but

completely protected

the nonmutated sites

(Fig. 7,

lanes

9, 10,

12,and

13).

Although

one

might

have

predicted

EBNA-1 tobe unable tobindtosites 2and3 in

dpml

+4on the basisofthe

results

presented

in

Fig. 4,

protection

of sites 2 and 3 was

observed

(lane

16).

The

affinity

of EBNA-1 for sites 2and

3,

however,wasreduced

by

themutations in sites 1 and

4,

since

larger amounts of EBNA-1 were

required

for

protection.

Togetherwiththe data

presented

in

Fig. 4,

this result indicates thatinteractionsoccurbetweenEBNA-1 dimers boundtosites

2 and 3 that promote stable DNA

binding

in the absence of EBNA-1 bound to sites 1 and 4. Evidence that EBNA-1-EBNA-1 interactions can occurbetween

recognition

elements

VOL.68, 1994

on November 9, 2019 by guest

http://jvi.asm.org/

[image:5.612.63.299.71.383.2]

1918 HARRISON ET AL.

A WT in3/4[5] in2/3[5] inl/2[5]

G A T C, ° o o 00o ° °0 C0C,°0 0 ngEBNA-1

ol co C\M It CO N It co CMJ 1t co

4 3

ww

-~~

o _ X

57 6it 8$diS

B WT in3/4[10] in2/3[10] inl/2[10]

G A T C, oa a a a°, ° ,a° ° ° ngEBNA-1

C\j st a:> C\ t co CM I co CM Ie co

0s--_d.--__hA tit

-I

a.w_ L__ : ._

R3 ww_

4,7.

iUIMiUi

' '0

1 2 3 4 5 6 7 8 9 10 1 1 121314 1516 1718 19 20

1 2 3 4 5 6 7 8 9 10 1112 13 14 1516 1718 19 20

FIG. 5. Protection of sequences in regionII byEBNA-1 inplasmidscontaining insertions between EBNA-1 sites. The binding ofEBNA-1 to

region II in plasmids containing 5-bp (A) or 10-bp (B) insertions between adjacent EBNA-1 sites was determined by DNase I footprint analysis

as described in the legend to Fig. 4, except that the EBNA-1 was further purified by sequence-specific DNA affinity chromatography (26). pHEBo-1.1 (lanes 5 to8) oritsderivatives containing insertions between sites 3 and4(lanes 9 to 12), sites2and3(lanes 13 to 16), and sites 1 and2(lanes 17to20) (500 ng ofeach) were incubated without EBNA-1 (lanes 5, 9, 13, and 17) or with 200 ng (lanes 6, 10, 14, and 18),400ng (lanes7,11, 15, and 19), or 800 ng(lanes 8, 12, 16, and 20) of EBNA-1. Sequencing ladders generated with the same radiolabeled oligonucleotide

are present inlanes 1 to 4, and thepositions of the EBNA-1 recognition elements are indicated on the left. WT, wild type.

separated by threehelical turnswasprovidedby the

coopera-tivebinding ofEBNA-1 to sites 1 and 2 and to sites3 and 4 withinl/2[10] andin3/4[10],respectively (Fig. 5B).

The ability of plasmids bearing these combination dpmsto

replicatetransientlyandbe maintainedovermany cell

gener-ations in D98/Raji cells was determined (Fig. 8). As for the plasmids containing single EBNA-1 site mutations and inser-tions between EBNA-1 sites, similar number of hygromycin-resistantcolonies were derived with pHEBo-1.1 and plasmids containing thedpml+2anddpm3+4 mutations(ranging from

4.1 x 104 for dpm3+4 to 5.9 x 104 for pHEBo-1.1). In

contrast, 5- to 10-fold fewer stable transformants were

ob-tained with plasmids containing the dpml+4 and dpm2+3 mutations. These reduced numbers were not due to less

efficient transfections with these plasmids, because similar

numbersofhygromycin-resistantcolonies wereobservedatday 7 following transfection. At this time, however, many of the

colonies began dyingandonly10 to20% continuedtoincrease in size. Such transient drug resistance has been previously

reported with plasmids that lack regionIIbut contain region I and maybeexplained bythe ability of region I to prolong the

retention of the plasmid in the presence of EBNA-1 (43). Stable transformants established with plasmids bearing

dpml +2 and dpm3+4 contained autonomously replicating

plasmid DNA at a copy number similar to that found in cells transformed with aplasmidbearing wild-type oriP(pHEBo-1.l

[Fig. 8A, lanes 3 to 5]). These results are consistent with the observed inability of EBNA-1 to bind to site 2 or site 3 in plasmids with mutations in site 1 or site4,respectively,and the ability of these plasmids to replicateautonomouslyin

EBNA-1-positive cells (Fig. 4 and6).

Stable transformants established with plasmids containing dpml+4anddpm2+3, in contrast, did not contain extrachro-mosomal plasmid DNA(Fig. 8A, lanes 5 and6). This defect could be explained byeither a failure of DNA replication to initiate at region II or a defect in the segregation of the

plasmids during expansion and passage ofthe transformants. Examination of the ability of plasmids bearing dpml+4 and dpm2+3 to replicate transiently in D98/Raji cells (Fig.

8B) revealed that these mutants were greatly reduced in their ability to replicate over a 93-h period compared with pHEBo-1.1 and plasmids containing dpml+2 and dpm3+4. This experiment demonstrated that the primary

defect in these mutants is in the initiation of DNA

replica-tion.Together, these resultsdemonstrate thatonlytwoof the four EBNA-1 sites present in region II are required for this

J VlIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.74.542.69.396.2]

EBV LATENT ORIGIN 1919

c) co co co

EiE rAi r'j4 j I~

_ m mmm m m m m m m m

co w w w w w w ww w w w

0.0.0.0.0

t * Y _ w _ _ _

wT dpml+2 dpm3+4 dpml+4 dpm2+3

G A T C 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0

4

3 4. ;

4i AJUt

>__-i_S _ w~~4

1 2 3 4 5 6 7 8 9 10 1 1 12 13

FIG. 6. Replication of plasmids containing point mutations in

individual EBNA-1 sites and insertions between EBNA-1 sites. The

presence ofplasmid DNAcarried by pooled hygromycin B-resistant

D98/Raji cells establishedbytransfectionwith pHEBo-1.1 (lanes 2 and

7) orits derivatives containing dpml (lane 3),dpm2 (lane 4), dpm3

(lane 5), dpm4 (lane 6), inl/2[5] (lane 8),inl/2[10] (lane 9),in2/3[5] (lane 10),in2/3[10](lane 11), in3/4[5] (lane 12),orin3/4[10](lane 13)

was determined after 20 or more cell generations.

Low-molecular-weightDNAfrom2 x 106cellswasdigested withClal andanalyzed

bySouthern blotting with radiolabeled pHEBo-1.1 asthe probe. The

most slowlymigrating species detected by the probe in lanes 2 to 6 represents the Clal fragment containing oriP from the endogenous

EBV genomes in D98/Raji cells as evidenced by its presence in

nontransfected D98/Raji cells (lane 1).

element to function and suggested that the spatial

arrange-ment of EBNA- 1 at these sites was important for origin

function.

Spatial requirements for EBNA-1 sites in oriP regionII.To

test the hypothesis that the spatial arrangement ofEBNA-1 dimersbound tosites in region IIis critical for origin function,

weintroduced 5- and 10-bp insertions between sites 1 and 2in

a dpm3+4 background (dpm3+4;inl12[5] and dpm3+4;inl/

2[10]) and between sites 3 and 4 in a dpml +2 background

(dpml+2;in3/4[5] and dpml +2;in3/4[10]). Aswasobserved for

plasmids carrying insertion mutations between adjacent EBNA-1 sites in a wild-type region II background (Fig. 5),

EBNA-1bound tothe sitesseparatedby the insertions inthese mutated plasmids (Fig.9). However, all four of these mutants

weregreatly reduced in their abilityto replicate transiently in D98/Raji cells (Fig. 10, lanes 4 to 7). Analysis of stable transformants established with these mutants revealed the complete absence of extrachromosomal plasmid DNA in

pHEBo-1.1(dpm1+2;in3/4[10]), pHEBo-l.1(dpm3+4;in1/2[5]),

andpHEBo-1.l(dpm3+4;in1/2[10])andapproximately 0.2copy per cell forpHEBo-1.1(dpml+2;in3/4[5]) transformants (data

not shown). These results indicate that only twobinding sites for EBNA-1, separated bytwohelical turns, are necessaryfor

region II function. Because the insertions altered the DNA

sequence between EBNA-1 sitesaswell asthe spacing of the

binding sites, it is possible that the binding ofacellularprotein

tothe minorgroovebetweenadjacent EBNA-1 sites (2, 16,31),

*__

hl

_~~~~~ - x

_~~~~~~~~~wi_ *

1 2 3 4 5 6 7 8 9 10 11 12 13 1415 16 17 18 19 20

FIG. 7. DNase I footprint analysis of pHEBo-1.1 derivatives

con-taining combinations of EBNA-1 site mutations. The sequences in regionIIcontaining doublepoint mutations in both sites 1and 2(lanes 8to 10), sites 3 and 4 (lanes 11 to13), sites I and4 (lanes 14to16),

orsites 2 and 3(lanes 17to20) protectedfrom nucleasedigestionby EBNA-1 weredeterminedasdescribed in thelegendtoFig.4.Lanes

I to4 contain sequencing ladders generated with the same

radiola-beled oligonucleotideasusedfor the footprintingassay,and lanes5to

7 contain footprinting reactions generated with wild-type (WT) pHEBo-1.1.

or its interactionwith EBNA-1, was altered by the insertions

and that thisprotein is required for origin function.

Compari-sonof thenucleotides between sites1 and 2with the sequence

between sites 3 and 4 (Fig. 1) does not reveal any striking

sequence similarities. Thus,ifa cellularprotein interacts with the EBNA-1-DNA complex at these sites, it may do so by

makingcontactswith phosphate residues(35,46).

Sequence requirements for the EBNA-1-induced distortion of region II. It has been shown previously that EBNA-1

induces the distortion of the DNA helix in oriP regionIIat two

sites (Fig. 1). These distortions, detected by oxidation of non-base-paired thymines by KMnO4, represent the bending

or untwisting of the duplex by EBNA-1 as opposed to helix unwinding (16, 26). The requirement for EBNA-1-induced distortion of region II for origin function is not known, but

distortion oforigin DNA appears tobe a common activity of

DNAreplication initiator proteins (6, 8, 32,45). Because of the location of these helical distortions between EBNA-l-binding sitesseparatedbytwohelicalturnsand the absence ofsimilar distortions in region I and between region II sites 2 and 3, it

has beenproposed that interactions between EBNA-1 dimers bound to sites 1 and 2 and to sites 3 and 4 are required for EBNA-1 to bend or untwist the DNA (26). The mutated origins described in this studywereexamined for the

EBNA-M9

EBNA-2

VOL. 68, 199)4

on November 9, 2019 by guest

http://jvi.asm.org/

[image:7.612.323.556.72.370.2] [image:7.612.91.266.74.299.2]

1920 HARRISON ET AL.

A. B.

$ +

cc 0 0

co co c c

co ww w w

Ell El

CDo M a: a:

~~~.

_.CO_.E E

El

| ;E

c

I : - 6 6 6

m m m m co

co

a IJ I I Iu I

C]t: X X L L m

_ .

W"

,'.

dpm3+4;dpm3+4; dpml+2; dpml+2;

'N inl/2[5]inl/2[10] in3/4[5] n3/4[10]

0120120120120120 gg EBNA-1

4

3

2

[image:8.612.67.278.70.336.2]

1 2 34 5 6 1 2 34 5 67

FIG. 8. Replication of plasmids containing combinations of EBNA-1sitemutationsinD98/Rajicells.(A)Theamountofplasmid DNAinhygromycin B-resistant D98cellsestablishedby transfection withpHEBo-1.1(lane 2)ormutantformscontainingdpml+2 (lane 3), dpm3+4 (lane 4), dpml+4(lane5),ordpm2+3 (lane 6)was deter-mined as described in the legend to Fig. 5. Low-molecular-weight DNA from D98/Raji cells was analyzed in lane 1. (B) Transient replication ofplasmidscontainingdpml+2(lane 4), dpm3+4 (lane5),

dpml+4 (lane 6), and dpm2+3 (lane 7) in D98/Raji cells was

comparedwith thereplicativeability of wild-type pHEBo-1.1 in D98 (lane 1)andD98/Raji(lane 3)cellsasdescribed in thelegendtoFig. 2.The absence oflow-molecular-weightDNAhybridizingtotheprobe inmock-transfectedD98/Rajicellsis shownin lane 2.

1-induced distortions at nt 9046 (Fig. 11A and 12A) and nt

9110(Figs. llB and 12B) totestthishypothesis.The

require-ments for both EBNA-1 andKMnO4 to detect these distor-tions, and the location of the KMnO4-reactive thymines in

region II,areshown in Fig. 11, lanes7 and 8.

The results obtainedwith the fourEBNA-1sitemutantsand six insertion mutants are presented in Fig. 11, lanes 9 to 18. Theonly mutation that eliminated the ability ofEBNA-1 to

bendoruntwist theduplexat nt9110wasdpml (Fig. liB,lane

9). Similarly, theonly mutationthat eliminated the EBNA-1-induced distortion at nt 9046 was dpm4 (Fig. 1lA, lane 9).

Because EBNA-1 wasabletodistort theDNA at nt9110 in the absence ofbindingtosite2(dpm2; Fig. 11A, lane 10) andatnt

9046in the absence ofbinding tosite 3(dpm3; Fig. 1 B,lane 11), it may be concluded that protein-protein interactions betweenEBNA-1 boundto sites1 and2 orbetweenEBNA-1

bound to sites 3 and 4 are not required for these helical distortions. The results obtained with the insertionmutants are

generally consistent with the effects of the single-site mutations onthe distortion ofregionIIbyEBNA-1.Thedistortion of the

DNA at nt9110was notaffected by the insertion of either 5or

10bp between sites1and2(Fig. 11, lanes 13and16).EBNA-1 was able to bend or untwist the DNA in site 4 in mutants

containing insertions between sites 3 and 4, but the thymine

1 2 3 4 5 6 7 8 9 10 11 12 13 1415 16

FIG. 9. DNaseIfootprintanalysisofplasmidscontainingin

muta-tions in combination dpm backgrounds. The ability of EBNA-1 to

protectregion IIsequences inpHEBo-1.1 (lanes 1 to3)andmutant

derivatives containing dpm3+4;inlI2[5] (lanes 4to 6), dpm3+4;inll 2[10] (lanes7to9),

dpml+2;in3/4[5]

(lanes10to 12),anddpml+2;

in314[10](lanes13to16)wasdeterminedasdescribed in thelegendto

Fig.4with0

pLg

(lanes 1, 4, 6, 10, 13,and 16),1 ,ug(lanes2,5, 8, 11, and 14), or2 p.g (lanes 3, 6, 9, 12,and 15) of DNAaffinity-purified EBNA-1 (26).WT,wild type.

oxidizedby KMnO4wasnt9045asopposedto nt9046(Fig. 11,

lanes 15 and18).Thechange in the KMnO4-reactivethymine

in mutants containinginsertionsbetween sites 3 and 4 might

suggestthat interactions betweenEBNA-1boundtosites 3 and

4affect thecontactsmadebetween EBNA-1andsequencesin site 4. This explanation seems unlikely because insertions betweensites1and 2didnotresult inoxidationof thethymine

at nt9111 andmutation of sites 2and 3hadno effectonthe

bending or untwisting of DNA in region II by EBNA-1. A more likely explanation is that the sequences created bythe insertions between sites3 and4 directly altered theability of

EBNA-1 to distort the DNAinsite4.

Although regionIis alsopresent ontheplasmidDNAsused

in the experimentwhose resultsare presented inFig. 11 and

EBNA-1 canmediate the formation ofa DNA loopbetween

oriPregionsIand11(15, 47), previous experimentshaveshown that thedistortion ofsequences inregion IIbyEBNA-1 does

notrequire the participation ofEBNA-1boundtoregionI,nor doesEBNA-1-region I complex formation inhibit the distor-tion of theDNA(26).The results of theexperimentshown in

Fig. 11 do not, however, rule out potential interactions be-J.VIROL.

1

on November 9, 2019 by guest

http://jvi.asm.org/

[image:8.612.326.548.74.408.2]

EBV LATENT ORIGIN 1921

cM

c om cl .._l *_l *.E_l

+

El El El E

8 _.__-7 7

I :, o o 0 0

a: tm m m m m

0)X LU LU LU

a) a) I I I I I c oac a. Q a

1 2 3 4 5 6 7

FIG. 10. Replicative abilityofplasmids containingin mutations in

combinationdpm backgrounds in D98/Rajicells. The transient

repli-cation ofpHEBo-1LI (lane3)andderivatives containing dpm3+4;inlI/

2[51 (lane 4), dpm3+4;inl/2[10] (lane 5), dpml+2;in314[51 (lane 6),

and dpml+2;in3/4[10] (lane 7) in D98/Raji cells was determined as

described in the legend to Fig. 2. Low-molecular-weight DNA from

mock-transfected D98/Raji cells (lane 2) and D98 cells transfected

withpHEBo-1.1 (lane 1)wasalsoanalyzed.

tween EBNA-1 bound to site 4 with dimers bound to sites 1

and/or 2 or between EBNA-1 bound to site 1 with dimers boundtosites 3 and/or4. Theshort distance separatingthese EBNA-1 sites might appear to preclude such interactions in the absence ofsharp bendingof theintervening DNA,and the absence ofKMnO4-reactive thyminesin thesequence separat-ing these sites when bound to EBNA-1 (Fig. 11) (16, 26) suggests that sharp bendingof the DNA does not occur (5).

However, such long-range interactions between EBNA-1 dimers inregion II have beensuggested bytheappearance of

a novel DNase I-hypersensitive site in EBNA-1 site 4 in mutantscontaining pointmutations in site 2(26). This obser-vation implied that a protein-DNA complex involving more

than four EBNA-1 dimers formsat region II. Examinationof the ability of EBNA-1 toinduceKMnO4 reactivity at nt9046

indpml+2;in3/4[5] and dpml+2;in3/4[10] (Fig. 12A, lanes 7 and 8) and at nt 9110 in dpm3+4;inl/2[5] and dpm3+4;inll 2[10] (Fig. 12B, lanes 7 and 8), however, revealed that these

long-range interactions are not required for the

EBNA-1-induceddistortions ofnt9046 and 9110. We concludefrom the

results of KMnO4 footprinting and replication experiments with these mutants that the EBNA-1-induced distortion of

region II is notsufficient for originfunction.

DISCUSSION

The EBV latent origin ofDNA replication isan important

modelforinvestigatingthe

regulated replication

of

eukaryotic

chromosomes. Bidirectional

replication (17)

initiates from a

defined DNA sequence

(17, 43, 52)

during early

S

phase

(22,

23) and requires only one viral gene

product,

the initiator

protein EBNA-1

(37, 53).

StudiesonEBNA-1 and its interac-tions with oriP have indicated that EBNA-1

performs only

a

subset of the functions

provided by

the most

highly

character-ized replication initiator

protein

that functions in

eukaryotic

cells, simian virus 40

large

T

antigen

(7).

In the presence of

ATP,Tantigenformstwohexamersattheviral

origin

ofDNA

replication andinducestwostructuralchanges

(melting

ofone arm of an inverted repeat sequence and distortion of an

A+T-rich

sequence).

Anintrinsic helicase

activity

ofT

antigen

extends the melted

region

to form a

larger

single-stranded

region,

and T

antigen

further interacts with DNA

polymerase

ot

primase

(12)

andRP-A

(39)

tofacilitate the

synthesis

of the first RNA

primer

andOkazaki

fragment.

T

antigen

alsoacts as a helicase at the

replication

forks

during

elongation (7).

PurifiedEBNA-1doesnotunwind theDNA

duplex

inoriP

(16,

26), and intrinsicATPase and helicaseactivitieshavenotbeen

found associated withEBNA-1

(3, 14, 40). Thus,

itseems

likely

thatone or more cellular

proteins

cooperatewith EBNA-1 to

carryouttheinitialstages inthe initiationof DNA

replication

atoriP. These

proteins

maybe novel ormaybe among those

requiredforsimian virus40 DNA

replication

(49).

Ifthe latter

is true, studies of the interaction of EBNA-1 with these

proteins

may

identify previously

unknown activities

required

for theinitiation ofDNA

replication

atcellular

origins

thatare

not essential

(or

are

masked)

in the simian virus 40 in vitro

replication

system. An

important

step toward

investigation

of

the cellular

proteins

required

for the initiation of DNA

replication

from oriPis the characterization ofthe molecular interactionsbetween EBNA-1 and the

origin

anddefinition of the sequence

requirements

for afunctional

origin.

The

exper-iments

presented

in this report

provide important

information

on both fronts.

The

ability

of

plasmids

lacking region

I to

replicate

tran-siently

in two different human cell lines revealed that this

component oforiP is not

required

for the initiation of DNA

replication

at

region

II in all cell types.

Previously published

experiments,

as well as

unpublished experiments

from our

laboratory,

have determined that

region

I is

required

for transient

replication

in

Raji cells,

an

EBV-positive

Burkitt's

lymphoma

cell line

(43, 51). Thus,

it appears that differences between

Raji

and

D98/Raji

or HeLa cellsare

responsible

for

the

varying

requirement

for

region

I in theinitiation ofDNA

replication.

One

possible

candidatethat

might

differ,

either in

concentration and/or in

posttranslational

modifications

be-tweenthesecelltypes, is EBNA-1. EBNA-1 is

phosphorylated

on serine residuesin

logarithmic

cultures of EBV-infectedB

cells

(25),

but

nothing

is known about the

effect(s)

of

phos-phorylation

on EBNA-1 function or whether cell division

cycle-dependent

changes

in EBNA-1

phosphorylation

occur.

Investigation

of

potential

differences in the levelor

posttrans-lational modification of EBNA-1 between

Raji

and

D98/Raji

cells may

yield

clues about the

regulated

replication

of

oriP-bearing

plasmids.

Although region

I is nonessential for efficient transient

replication

in

D98/Raji cells,

it is

absolutely

required

for the

long-term maintenanceof

oriP-bearing

plasmids

in these cells

(37,

43).

In addition to

fulfilling

a

regulatory

role,

three

observations suggest a role for EBNA-1 and

region

I in

plasmid

segregation during

mitosisand may

explain

the

abso-VOL.68, 1994

on November 9, 2019 by guest

http://jvi.asm.org/

[image:9.612.124.226.72.395.2]

1922 HARRISON ET AL.J.Vo.

EBNA-1 + +

A KMnO4 - +- U

G AT C cli

~f

C;~ C't) q

C') C\j CY)

ClI Cl .~C lI

B EBNA-1 - - + + ..

KMnO4

-G A T C .vi.j _{cl cl cl-lclIc:}Cmi co CMJ rV

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

FIG. 11. Distortion of sequencesinwild-typeand mutated region IIDNAsbyEBNA-1. Thebendingoruntwisting of theDNAduplexatnt

9046(A)ornt91 10(B)ofregion IIwasdeterminedbytheabilityofEBNA-1Itoinduce theoxidationofthymines byKMnO4(16, 26).Oxidized

thymines prevent DNA elongation by Klenow polymerase and are evident as termination sites in primer extension experiments (5). Immunoaffinity-purifiedEBNA-1(2

jig)

wasincubated with 0.5

jig

ofsupercoiled plasmidDNAcontaining wild-typeormutatedoriP, the reactions

were treated with KMnO4, and the products were analyzed by primer extension with end-labeled oligonucleotides (those in panel A were

complementarytoEBVnt9161 to9175;those inpanelBcorrespondedtoEBVnt8989to9005)aspreviouslydescribed(26).Both EBNA-1 and

KMnO4wererequiredfor detection of the distortionsatEBV nt9046 and 9110(lanes 8);omittingeitherEBNA-1 (lanes 6),KMnO4 (lanes7),

orbothEBNA-lI andKMnO4(lanes5) resultedinthe absence of chain termination atthesesites. TheabilityofEBNA-1 todistortthe DNA in the

pHEBo-l.1

derivativecontainingdpml (lanes 9), dpm2 (lanes 10),dpm3

(lanies

11),

dprn4

(lanes 12),

inl/2[5]

(lanes 13),

in2/3[51

(lanes 14),

in3!4[51

(lanes 15),

inl/2[10]

(lanes 16),

in2/3[10]

(lanes 17), or

in3/4[10]

(lanes 18) was determined. Sequencingladders containing the same

end-labeledoligonucleotides are shown in lanes I to4.

lute

requirement

for

region

Iin the

long-term

maintenance of

oriP-bearing plasmids. Region

I

provides

for the

prolonged

retentionof

physically

linked sequences in an

EBNA-1-depen-dent manner

(33). Additionally,

both EBNA-1 and EBV genomesareassociated with

metaphase

chromosomes

(21,

24).

Mutational

analysis

of

region II,

summarized in Table 2,

showed that not all of the EBNA-1 sites in this element are

required

for

origin

function. Two ofthe four sitesaresufficient,

provided

that

they

aretwohelicalturnsapart.Our

experiments

examining

the

replicative ability

of mutants

containing only

two EBNA-1 sites,

separated

by

21

bp,

showed that these

plasmids

wereableto

replicate

overmany cell

generations

but didnotaddress their

stability.

Additional

experiments

measur-ing plasmid

loss rates in the absence of selection for the

hygromycin

resistance gene are

required

to examine this

possibility.

No apparent differences between the

replicative

ability

of

plasmids bearing

mutations in sites 1 and 2orinsites 3 and 4 were observed,

suggesting

thatthese

paired

sites are

functionally equivalent

and, when

assayed

in

D98/Raji

cells,

redundant. The

spatial

requirement

for these sites does not

reflect an

inability

of EBNA-1 to bind in vitro to sites

separated by

either 26 or31

bp

orthe

inability

of EBNA-1 to bend or untwist the DNA

duplex

within site 1 or4. Rather,

these data suggest that

protein-protein

interactions between EBNA-1 dimers bound tothese

adjacent

sites and/orbetween EBNA-1 andacellular

protein(s)

necessary for

origin

function

require

a

specific spatial

arrangement of EBNA- 1. These

results,

together

with the

availability

of mutants that are

defective for

replication

because this

spatial requirement

isnot met, will be useful for

investigation

of the cellular

protein(s)

that interacts with EBNA-1 atthe

origin.

Previous

analyses

ofthe sequences in

region

II

required

for

plasmid

maintenance reliedondeletion mutations

(9,

52).

The

phenotypes

of

plasmids containing

mutations in EBNA-1 sites 4 and 3

(see above)

are

generally

consistent with the

pheno-types of mutants

containing

deletions from EBV nt 8997 to 9065

(9):

deletion of site 4 and most of site 3

(together

with sequences

flanking

site

4)

didnotabolish

plasmid

maintenance but did result in the appearance ofsome

plasmids

with altered structures. However, deletions

extending

from nt 9518 into J.VIROL.

Ao

vow

AW

mm*

on November 9, 2019 by guest

http://jvi.asm.org/

[image:10.612.107.495.74.401.2]

EBV LATENT ORIGIN 1923

A B.

-o

EBNA-1 + + lc EBNA-i +

KMnO4 -+ C\+ KMnO4 -+ 1

G A T

;

C G A T C

- -S;

---ss - *

- -* e -31t ''

- *h

--

^-

-s

-

--

-- --

-

-

^

--E *

--

--

--

--

--

--

-

-

-

-

p--

O.-_

1-40

do

a

qatoa

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

FIG. 12. EBNA-1-induced distortion of region IIin plasmids

con-tainingin mutations in combination dpm backgrounds. The distortion ofthe DNA at EBV nt 9046 (A) or9110 (B) was determined by KMnO4 footprinting (as described in the legend to Fig. 11) (25). Plasmid DNA (0.5 p.g)containing wild-type oriP (lanes 5 and 6) or

dpml+2;in3/4[5] (panel A, lane 7),dpml+2;in3/4[10] (panel A, lane 8), dpm3+4;inl/2[5] (panel B, lane 7),ordpm3+4;inl/2[5] (panel B,

lane 8) was incubated with 2 ,ug of DNA affinity-purified EBNA-1. Sequencing ladders generated with thesame oligonucleotide primers

areshown in lanes 1 to4.

regionII resulted in less thanoneplasmidpercell (deletion to nt9106removing EBNA-1 site 1andflankingsequences)or no

detectableplasmid maintenance (deletiontont9095 removing site 1 and half of site 2 [9]). These data may suggest that

sequences flanking the EBNA-1 sites that were affected by

these deletions are important for origin function. It is also possible that the deletions moved inhibitoryvector sequences

closer to the EBNA-1 sites and that this juxtaposition was

responsible for the loss of replicative ability. One indication that thelatterpossibility iscorrectwasthe observed ability of

adifferentoriP-bearing plasmidtotolerate adeletion

extend-ing fromnt 9518 to9106 (52).

Whyaretherearefour EBNA-1 sites inregionIIifonlytwo

are required for origin function? Additional sites may be required for plasmid replication, long-term plasmid

mainte-nance, orregulation of origin activity in EBV-infected B cells. The nucleotidesequence of the latent origin of DNA replica-tion of the related baboon virus, herpesvirus papio (HVP), resembles EBV oriP in a number of ways (36). Sequences similarto the EBNA-1 recognition element are present as 10

TABLE 2. Summaryof the effects of EBNA-I-binding site and insertion mutations upon the EBNA-1-induced distortion

andreplicative function of oriP regionII

Residue(s) Transient Plasmid Mutation withKMnO4

replicationh

maintenance"

reactivity'

dpn1 9046 ND"d +

dpmn2 9046, 9110 ND +

dpm3 9046,9110 ND +

dpmn4 9110 ND +

inl1/2[5] 9046, 9110 ND +

inl1/2[10] 9046,9110 ND +

in2/3[5] 9046, 9110 ND +

in2!3[10] 9046,9110 ND +

in3/4[5] 9045,9110 ND +

in3/4[10]

9045, 9110 ND +

dpml+2 ND + +

dpm3+4 ND + +

dpml+4 ND -

-dpm2+3 ND

-dpm3+4;in1/2[51 9110

-dprm33+4;inl/2[1I0] 9110

-dpmI+2;in3/4[5] 9045

-dpmI+2;in3/4[I0J 9045

-"The thymine residue(s) in pHEBo-1.1-derivatives containing the indicated mutationsthat are oxidized by KMnO4 in the presence of EBNA-1 are given. Numbering isaccording to reference 4.

"The ahilityof plasmids bearing the indicated mutations to replicate tran-siently inD98/Raji cells to a level similar to wild-type is indicated by a plus sign. Mutants thatreplicatedat 10%lorless ofwild-typearerepresented by a minus

sign.

'Theability of mutated plasmids to be maintained at one or more copies per cell inD98/Rajicellsis indicatedby a plus sign. Mutated plasmids that are not present asextrachromosomal elements inD98/Raji cells are indicated by a minus sign.

" ND, not done.

tandemcopiesin one region of theorigin,and these can serve as a transcriptional enhancerin HVP-infected cells (36). Five related repeatsare located 764bpfrom theenhancerelement,

and these may make up the oriP region II equivalent of the HVP latentorigin of DNAreplication. We are limited in our

ability to make comparisons between this sequence and oriP

region II because of lack of information on the sequence

requirements of theHVP EBNA-1counterpart. Itis, however, interesting that,asfor oriP,morethantworepeatsarepresent and the spacingof fourrepeatsin theHVPsequenceis similar

tothatseenin oriP: two paired sites, each separated by21 bp.

The distance between the innertwosites is22bp, asopposed

tothe 30bpseparating sites2and 3 in oriPregion11(2, 36, 41).

Thesimilarities betweenthe latentorigins ofDNAreplication

of EBV and HVP suggest that morethantwoinitiator protein-binding sites are required for an origin function not revealed

byourassays.

oriP region II is also referred to as the dyad symmetry

element becauseof the presence ofa65-bp dyad sequencethat

spans EBNA-1 sites 3and 4 (41, 43,52). Computer modeling

of this sequence predicted the formation of a number of alternativestem-loopstructures (30,41,50), and evidence for oneofthese structures wasobtainedby probing nakedplasmid

DNA containing region II with a single-strand-specific

endo-nuclease (50). Stem-loopstructures are not, however,formed inregionIIinthe presence ofEBNA-1 (16,27).Analysisof the

thermodynamic stability of the dyad symmetry element also

predicted that this sequence has reduced helical stability

relative to flanking sequences, and the presence of similar easily unwound sequences in other replication origins

sug-VOL.68, 1994

on November 9, 2019 by guest

http://jvi.asm.org/

[image:11.612.98.258.74.406.2]

1924 HARRISON ET AL.

gestedthat this featuremightbeimportantfororiginfunction (50).Althoughthesignificanceoftheseinvitroobservationsis not clear, the ability of plasmids containing mutations in EBNA-1 sites 3 and4 to replicate autonomously showedthat binding of EBNA-1 to sites in the dyad symmetryelement is not requiredfor activation of theorigin.

ACKNOWLEDGMENTS

Wethank R. Glaser forprovidingthe D98 and D98/Rajicelllines, N. Mausfor construction ofpRII,R.Schneider forprovidingpNL3-C, and B. Sugden for providing pHEBo-1. We thank P. Hearing for critical review of themanuscriptandmanystimulatingdiscussions.

This work was supported by grant VM-13A from the American CancerSociety.

REFERENCES

1. Adams,A.1987.Replicationof latentEpstein-Barrvirusgenomes

inRajicells. J. Virol. 61:1743-1746.

2. Ambinder, R. F.,M. Mullen,Y.-N. Chang, G. S. Hayward,and

D. S.Hayward. 1991. Functional domains ofEpstein-Barrvirus nuclearantigenEBNA-1. J.Virol. 65:1466-1478.

3. Ambinder,R.F.,W.A. Shah,D.R.Rawlins,G.S.Hayward,and

S. D.Hayward.1990.Definitionof thesequencerequirementsfor binding of the EBNA-1 protein toitspalindromictargetsites in Epstein-Barrvirus DNA. J. Virol. 64:2369-2379.

4. Baer, R.,A. T.Bankier,M.D.Biggin,P. L.Deininger,P.J. Farrell, T.J.Gibson,G.Hatfull,G. S.Hudson,S. C.Satchwell,C.Sequin,

P. S. Tuffnell, and B. G. Barrell. 1984. DNA sequence and

expressionof the B95-8Epstein-Barrvirusgenome.Nature (Lon-don)310:207-211.

5. Borowiec, J. A.,F. B.Dean, P. A.Bullock,andJ. Hurwitz. 1990. Bindingandunwinding-howTantigenengagestheSV40origin of DNA replication.Cell 60:181-184.

6. Borowiec, J. A., and J. Hurwitz. 1988. Localized melting and structural changes in the SV40 origin ofreplication induced by

T-antigen. EMBO J. 7:3149-3158.

7. Borowiec, J. A.,L.Zhang,S.Sasse-Dwight,andJ.D.Gralla. 1987.

DNAsupercoilingpromotesformation ofabentrepressionloop inlac DNA. J. Mol. Biol. 196:101-111.

8. Bramhill, D., and A. Kornberg. 1988. Duplexopening by dnaA proteinatnovelsequencesininitiationofreplicationattheorigin ofthe E. colichromosome. Cell 52:743-755.

9. Chittenden,T.,S.Lupton,and A.J.Levine.1989.Functional limits oforiP, the Epstein-Barrvirus plasmid origin of replication. J. Virol. 63:3016-3025.

10. Dambaugh, T.,K.Hennessy,S.Fennewald,andE.Kieff.1986. The virus genome anditsexpression in latent infection, p. 13-45. In M. A.Epstein and B. G. Achong (ed.), TheEpstein-Barrvirus:

recentadvances. JohnWiley&Sons,Inc.,NewYork.

11. Dhar, V., and C. L. Schildkraut. 1991. Role of EBNA-1 in arresting replicationforksattheEpstein-BarrvirusoriPfamilyof tandem repeats. Mol. Cell. Biol. 11:6268-6278.

12. Dornreiter,I.,A.Hoss,A. K.Arthur,and E.Fanning.1990. SV40

T antigenbinds directly to the large subunit of purified DNA polymerase alpha.EMBO J.9:3329-3336.

13. Feinberg,A.P.,andB.Vogelstein.1983.Atechniquefor radiola-beling DNArestriction endonuclease fragments to high specific

activity. Anal. Biochem. 132:6-13.

14. Frappier, L., andM. O'Donnell. 1991. Overproduction,

purifica-tion,andcharacterization ofEBNA1,theoriginbindingproteinof Epstein-Barrvirus. J. Biol. Chem. 266:7819-7826.

15. Frappier, L., and M. O'Donnell. 1991. Epstein-Barr nuclear antigen1mediatesaDNAloopwithinthe latentreplication origin ofEpstein-Barrvirus. Proc. Natl. Acad. Sci. USA 88:10875-10879. 16. Frappier, L.,andM.O'Donnell. 1992. EBNA-1distortsoriP,the Epstein-Barrvirus latentreplication origin.J.Virol. 66:1786-1790.

17. Gahn,T.A.,andC. L.Schildkraut. 1989. TheEpstein-Barrvirus originofplasmid replication,oriP,contains both the initiation and termination sites ofDNAreplication.Cell 58:527-535.

18. Glaser, R., and M. Nonoyama. 1974. Host cell regulation of induction ofEpstein-Barrvirus. J. Virol. 14:174-176.

19. Graham, F. L., and A. J.vanderEb.1973. Anewtechniquefor the assayofinfectivityofhuman adenovirus S DNA.Virology 5:456-467.

20. Gralla, J. D. 1985. Rapid "footprinting" on supercoiled DNA. Proc.Natl. Acad. Sci. USA 82:3078-3081.

21. Grogan, E. A., W. P. Summers, S. Dowling, D. Shedd, L. Gradoville, and G. Miller. 1983. TwoEpstein-Barr viral nuclear neoantigens distinguishedby gene transfer, serology, and

chromo-somebinding.Proc. Natl. Acad. Sci.USA 80:7650-7653. 22. Gussander, E., and A. Adams. 1984. Electron microscopic

evi-dence for replication of circular Epstein-Barrvirus genomes in latently infected Raji cells. J. Virol. 52:549-556.

23. Hampar, B.,A. Tanaka,M. Nonoyama, and J. G.Derge. 1974. Replication of theresident repressed Epstein-Barr virus genome during the early Sphase(S-1 period) of nonproducerRajicells. Proc.Natl. Acad. Sci. USA 71:631-633.

24. Harris, A., B. D. Young, and B. E. Griffin. 1985. Random association of Epstein-Barr virus genomes with host cell

met-aphase chromosomes in Burkitt's lymphoma-derived cell lines. J. Virol.56:328-332.

25. Hearing, J. C., and A. J. Levine. 1985. The Epstein-Barr virus nuclear antigen (BamHI K antigen) is a single-stranded DNA binding phosphoprotein. Virology145:105-116.

26. Hearing, J., Y. Mulhaupt, and S. Harper. 1992. Interaction of Epstein-Barr virusnuclear antigen 1 with the viral latent origin of replication. J. Virol. 66:694-705.

27. Hearing, J. C., J.-C.Nicolas, and A. J. Levine. 1984. Identification of Epstein-Barr virus sequences that encode a nuclearantigen expressed in latently infected lymphocytes. Proc. Natl. Acad. Sci. USA81:4373-4377.

28. Hennessy, K., and E. Kieff. 1983. One of twoEpstein-Barr virus nuclear antigens contains a glycine-alanine copolymer domain. Proc. Natl. Acad. Sci. USA 80:5665-5669.

29. Hirt, B. 1967. Selectiveextraction of polyoma DNA from infected mousecell cultures. J. Mol. Biol. 26:365-369.

30. Karlin, S., and B. E. Blaisdell. 1987. A model for thedevelopment of the tandem repeat units in the EBV ori-P region and a

discussion oftheirpossible function. J. Mol. Evol. 25:215-229. 31. Kimball, A. S., G. Milman, and T. D. Tullius. 1989.

High-resolutionfootprints of the DNA-binding domain of Epstein-Barr virusnuclearantigen1. Mol. Cell. Biol.9:2738-2742.

32. Koff, A., J. R. Schwedes, and P. Tegtmeyer. 1991. Herpes simplex virus origin-binding protein (UL9) loops and distorts the viral replication origin. J. Virol. 65:3284-3292.

33. Krysan, P.J., S. B.Haase, and M. P. Calos. 1989. Isolation of human sequences that replicate autonomously in human cells. Mol. Cell.Biol. 9:1026-1033.

34. Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesis withoutphenotypicselection. Proc. Natl. Acad. Sci. USA 82:488-492.

35. Lee,D.K.,M.Horikoshi,and R. G. Roeder.1991. Interaction of TFIIDin theminor groove of the TATA element. Cell 67:1241-1250.

36. Loeb, D. D., N. Schwab Sung, R. L. Pesano, C. J. Sexton, C.

HutchisonIII,andJ.S.Pagano. 1990. Plasmid originof replica-tion ofherpesvirus papio: DNA sequence and enhancer function. J. Virol.64:2876-2883.

37. Lupton, S., andA.J. Levine. 1985. Mappinggenetic elements of Epstein-Barrvirusthat facilitate extrachromosomal persistence of Epstein-Barr virus-derived plasmids in human cells. Mol. Cell. Biol.5:2533-2542.

38. Mecsas,J.,andB.Sugden.1987. Replication of plasmids derived frombovinepapillomavirus type1and Epstein-Barr virus in cells inculture. Annu. Rev. Cell Biol. 3:87-108.

39. Melendy, T., and B. Stillman. 1993. An interaction between replication protein Aand SV40 Tantigen appears essential for primosome assembly during SV40 DNA -replication. J. Biol. Chem.268:3389-3395.

40. Middleton,T.,and B.Sugden. 1992. EBNA1 can link the enhancer elementtotheinitiator element of theEpstein-Barrvirusplasmid originof DNAreplication.J. Virol.66:489-495.

41. Rawlins,D.R., G. Milman, S.D.Hayward, and G. S. Hayward. 1985. Sequence-specific DNAbinding of the Epstein-Barrvirus J. VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

EBV LATENT ORIGIN 1925 nuclear antigen (EBNA-1) to clustered sites in the plasmid

maintenanceregion. Cell 42:859-868.

42. Reisman, D., and B.Sugden. 1986.transactivation ofan Epstein-Barr viral transcriptional enhancer by the Epstein-Barr viral nuclearantigen 1.Mol.Cell. Biol. 6:3838-3846.

43. Reisman, D., J.Yates,and B.Sugden. 1985.Aputative origin of replication of plasmidsderived from Epstein-Barr virus is

com-posed oftwocis-actingcomponents. Mol.Cell.Biol. 5:1822-1832.

44. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467.

45. Schnos, M., K. Zahn, R B. Inman, and F. R. Blattner. 1988. Initiation protein induced helix destabilization at the lambda origin: apreprimingstepinDNAreplication. Cell 52:385-395. 46. Starr, D. B., and D. K.Hawley. 1991. TFIID binds in the minor

grooveofthe TATA box. Cell 67:1231-1240.

47. Su, W., T. Middleton, B. Sugden, and H. Echols. 1991. DNA

looping between the origin of replication ofEpstein-Barr virus and its enhancersite: stabilization ofanorigin complex with

Epstein-Barr nuclear antigen 1. Proc. Natl. Acad. Sci. USA 88:10870-10874.

48. Sugden,B.,K.Marsh, and J. Yates. 1985. Avectorthatreplicates

as a plasmid and can be efficiently selected in B lymphoblasts

transformedby Epstein-Barr virus. Mol. Cell. Biol.5:410-413. 49. Tsurimoto, T., T. Melendy, and B. Stillman. 1990. Sequential

initiation of lagging and leading strand synthesis bytwodifferent polymerase complexesat theSV40 DNA replication origin.

Na-ture (London) 346:534-539.

50. Williams, D. L., and D. Kowalski. 1993. Easily unwound DNA

sequences andhairpinstructuresin theEpstein-Barr virus origin of plasmid replication. J. Virol. 67:2707-2715.

51. Wysokenski, D. A., and J. L. Yates.1989.MultipleEBNA1-binding sitesarerequiredtoformanEBNA1-dependent enhancerandto

activate a minimal replicative origin within oriP ofEpstein-Barr virus. J. Virol. 63:2657-2666.

52. Yates, J., N. Warren, D. Reisman, and B. Sugden. 1984. A

cis-acting element from the Epstein-Barrvirusgenome that

per-mits stable replication of recombinant plasmidsinlatently infected cells. Proc. Natl. Acad. Sci. USA81:3806-3810.

53. Yates, J. B., N. Warren, and B.Sugden. 1985.Stablereplication of plasmids derived from Epstein-Barrvirus in various mammalian

cells.Nature(London)313:812-815.

54. Yates, J. L., and N. Guan. 1991. Epstein-Barr virus-derived

plasmids replicate onlyoncepercellcycleandarenotamplified afterentryinto cells. J. Virol. 65:483-488.

VOL.68, 1994

on November 9, 2019 by guest

http://jvi.asm.org/