ZEBRA and a Fos-GCN4 chimeric protein differ in their DNA-binding specificities for sites in the Epstein-Barr virus BZLF1 promoter.

0022-538X/91/084033-09$02.00/0

Copyright© 1991, AmericanSocietyfor Microbiology

ZEBRA

and

a

Fos-GCN4

Chimeric Protein

DNA-Binding

Specificities for Sites in the

Virus

BZLF1 Promoter

Differ in Their

Epstein-Barr

NAOMITAYLOR,' ERIK FLEMINGTON,2 JOHN L. KOLMAN,1RAYMOND P. BAUMANN,' SAMUEL H. SPECK,2 ANDGEORGE

MILLER'3

4*

Departments of Molecular Biophysics and Biochemistry,'Epidemiology and Public Health,3 andPediatrics,4 Yale University School of Medicine, New Haven, Connecticut 06510, andDivision of Tumor Virology,

Dana-Farber CancerInstitute, andDepartmentof Pathology, Harvard Medical School, Boston, Massachusetts 021152

Received12October1990/Accepted16April 1991

Epstein-Barr virus (EBV) encodesa protein, ZEBRA, which enables the virustoswitchfrom alatenttoa

lytic lifecycle. Thebasicdomain of ZEBRA is homologoustotheFos/Junoncogenefamily, and both proteins

bindthe canonical AP-1site(TGAGTCA). However, ZEBRAdoesnotcontainaleucine zipper dimerization domain which has beenshown tobe necessaryforDNAbinding of Fos/Jun proteins. Additionally, ZEBRA bindstosites which deviatefromthe AP-1consensus sequence.Thus,itwasofinteresttodefinethedomain of theZEBRAprotein required forDNAbinding.We have determined bymutagenesisthatZEBRA residues 172 to227,representing the basic domainandaputative dimerization domain,arerequiredfor specific bindingto AP-1 and divergent sites. Mutagenesisofthe basic aminoacids 178 to 180 or187 to 189 abrogates ZEBRA binding to all DNA target sequences. These residues are conserved in Fos and arealso necessary for Fos DNA-binding activity.Wehave foundthataFos-GCN4 chimera and ZEBRA have differentcognatebinding specificities. The autoregulated BZLF1 promoter contains three divergent AP-1 sequences, ZIHA (TGAGCCA),ZIIIB(TTAGCAA),andZ-AP-1-octamer(TGACATCA).ZEBRAbindswithhigh specificityto ZIIIAand ZUIBbutweaklytothe Z-AP-1octamer. Conversely, theFos-GCN4 chimera recognizesonly the Z-AP-1 octamer. ZEBRA binds the ZIIIA and ZIIIB sites together in a noncooperative fashion, while Fos-GCN4 bindsthesesitesas a higher-ordercomplex. Additionally, we havefound thatflankingsequences influencebinding of Fos-GCN4toadegenerate AP-1site(TGAGCAA). The characteristic binding specificities of ZEBRA andcellular AP-1 proteinssuggest thatthey differentiallyaffect viraland cellular transcription. Epstein-Barr virus (EBV) encodes a protein, ZEBRA,

which enablesthe virustoswitch fromalatenttoalytic life cycle (8, 9, 41). EBV is latent in lymphoid cells butcan be activated invitroby phorbolesters, butyrate,orexpression of the ZEBRAgene(8, 9, 15, 29, 41, 46). The mechanism by which ZEBRA initiates the cascade of viral replication has been thefocus of much study. Numerousgroupshavefound that ZEBRA directly transactivates several promoters of lytic cyclegenes, includingaregion integraltotheEBVlytic originofreplication (7, 8, 13, 15-17, 20, 28, 41, 43).

Exon II of ZEBRA ishomologoustothe basic domain of the Fos/Junproto-oncogene protein family (11). In Fos/Jun complexes, thisdomain has been showntointeractdirectly witha DNA heptamer known asthe tetradecanoyl phorbol

acetate (TPA) response element or AP-1 site (TGAGTCA) (3, 26, 34). Itwasdetermined that ZEBRA shares theability of theFos/JuncomplextobindanAP-1consensussite(11). AnAP-1 sequence (MS-AP-1) is locatedin thepromoterof anEBVearlygene, BSLF2-BMLF1(MS-EA) (11). ZEBRA can transactivatetranscription throughthis site, although it is not yet clear whether AP-1 sequences are required for transactivation ofEBV genes in infected lymphocytes (18, 43).

ZEBRA also binds sites on the EBV genome which deviate from the AP-1 consensus sequence (11, 13, 27, 28,

37, 43). TheautoregulatedBZLF1 (ZEBRA)promoter

con-*Corresponding author.

tains three divergent AP-1 elements (13). Two sites, TGAGCCA (ZIIIA)andTTAGCAA(ZIIIB), 129 and 116bp upstream of the transcriptional start site are arranged in a tail-to-tail fashion andare separated by six nucleotides. An additional AP-1 sequence, TGACATCA (Z-AP-1 octamer),

islocatedfurther downstreamat-67.Thissequence,which wasfirst identified in thejun promoterby Angelet al. (2), mediates activationbyTPA andJun/AP-1. Recent work has shown thatalthoughthissite is TPAinducible in the ZEBRA promoter,it doesnotappeartobetransactivatedbyZEBRA (12, 13).

Sinceboth ZEBRA and Fos/Junareinvolved in transcrip-tional regulation of gene expression, it was important to

compare the DNA-binding activities of these proteins. Therefore,westudiedthe relativeDNA-bindingaffinitiesof ZEBRA and a Fos-GCN4 chimera for the divergent AP-1 sitesfound in the autoregulatedBZLF1 promoter.

MATERIALS ANDMETHODS

Plasmids. Thefull-length BZLF1cDNA(aminoacids[aa] 1to245) (30)wasclonedas aNaeI-BamHIfragmentintothe SmaI-BamHI sites of the trpE bacterial expression vector

pATH 11. TheRajiBZLF1 cDNA andtheexpressionvector

were the gifts of A. Sergeant and T. J. Koerner,

respec-tively. BZLF1 deletionmutantswereconstructedasfollows:

Z141-245

wasclonedasaNheI-BamHIfragment,Z172-245was

cloned as aBsmI-BamHI fragment, Zl-198 wascloned as a NaeI-PstI fragment, and

Zl-227

was cloned as a

BamHI-4033

on November 10, 2019 by guest

http://jvi.asm.org/

4034 TAYLOR ET AL.

HinclI

fragment.

All

fragments

were cloned into the

appro-priate

sites ofpATH vectors.

ZA111-159

was constructed

by

bidirectional Bal 31 exonuclease

digestion

of

Zl-245

atthe

unique

NheI site(aa 141). Exon Iofthe BZLF1 open

reading

frame, WZ1_167,

was cloned into the

pATH

vector from

WZhet (42), and exon II of

BZLF1, Z167-202,

was cloned

from

genomic

BamHI Zas aPvuII-HincII

fragment.

TwoadditionalmutantZEBRA

proteins

wereconstructed

by

site-directed

mutagenesis

of the B95-8 BZLF1 cDNA cloned in SP64. This cDNA was the

gift

ofP. Farrell

(11).

Amino acids 178to 180 and 187 to 189 were mutated from

KRY to EELand RKCto EES,

respectively.

Site-directed

mutagenesis

was

performed

as described

by

the

manufac-turer

(Bio-Rad).

AFos-GCN4 chimera which containsthe Fosbasic

DNA-binding domain,

aa 126to 162, fused inframetothe GCN4

leucine

zipper domain,

aa251to281,was thekind

gift

ofT.

Kouzarides and E. Ziff(24). A Sau3AI-XbaI fragment

con-taining

theFos-GCN4chimerawasclonedinto the

BamHI-XbaI sites ofpATH1.

Protein expression. Fusion

proteins

were expressed as described

previously

(22, 42). Escherichia coliAG1

contain-ing

thevariousTrpE-BZLF1deletionmutants wasgrownin 5 ml ofM9 medium

containing

1% Casamino Acids

(Difco)

andinduced for4hin the presenceof

indoleacrylic

acid(20

,ug/ml).

Proteins from0.5 ml of cells were

electrophoresed

on a sodium

dodecyl

sulfate

(SDS)-polyacrylamide gel

and stained with Coomassiebrilliant blue. The level of induced

protein

in each

preparation

was estimated

visually,

and

volumes ofcells containing equivalent amounts ofprotein

(3.5

to 5

ml)

were

pelleted

and

resuspended

in 1 ml of6 M

urea

(1).

Cellsweresonicatedtwice for20seachtime. The

urea-solubilized

protein

fractionwas

dialyzed

threetimes for

45mineachtimeat4°C

against

15 mM

N-2-hydroxyethylpi-perazine-N'-2-ethanesulfonic

acid

(HEPES; pH 7.9)-75

mM

KCl-0.2 mM EDTA-1 mM dithiothreitol-7.5%

glycerol.

Constructs in

pBluescript

KS were transcribed in vitro

(31)

and thentranslatedinawheat germextract asdescribed

by

the manufacturer

(Promega).

Synthetic oligonucleotides.

The

synthetic

double-stranded

oligonucleotides

used in

binding

studies were as follows

(sequence

reflects top strand in a 5'-to-3'

direction;

core

sequences are underlined and mutated bases are noted in

bold

type).

MS-AP-1: TCTTCATGAGTCAGTGCTTC MS-AP-1: TCTTCATTAGTCAGTGCTTC ZIIIA: CTAGCTATGCATGAGCCACAGATC

ZIIIA*: CTAGCTATGCATGAGCAACAGATC

ZIIIAm: CTAGCTATGCAGAATTCACAGATC

ZIIIB: CTAGCAGGCAITGCTAATGTACCGATC

ZIIIB*: CTAGCAGGCATTGCTCATGTACCGATC

ZIIIBm: CTAGCAGGATCCGCTAATGTACCGATC

Double:

GACTATGCATGAGCCACAGGCADGCJAATGTACCGA

(containsZIIIA +ZIIIB)

Z-AP-1octamer:CTAGAAACCATGACATCACAGAGGATC Controloligonucleotide: TGGCATGCTGCTGACATCTGGC

All

oligonucleotides

had XbaI sites at the 5' end and

BamHI sites at the 3'

end,

with the exceptionof the AP-1 and Double

oligonucleotides,

which had CT and AG

over-hangs

at the 5' and 3' ends, respectively. Radiolabeled

oligonucleotides

were prepared by using the Klenow

frag-ment ofE. coliDNApolymerase and 32P-labeled

deoxynu-cleotides.

Gel retardationassays. DNA

binding

reactions were

per-formedasdescribed

previously

(11).Reaction mixtureswere incubated atroomtemperature(22°C)or4°C for30min.A 4 x

10-13

M concentrationof the appropriate

32P-end-labeled

double-strandedoligonucleotideand 2 ,ul ofprotein

(approx-imately 100ng)were incubated ina25-,ul volume of 15 mM HEPES(pH

7.9)-75

mMKCl-0.2 mM EDTA-4 mM dithio-threitol-7.5% glycerol-1 pigofpoly(dI-dC). Gel retardation

competitionassayswereconducted with5-, 20-,and 50-fold molar excesses of cold competitor over the 32P-labeled MS-AP-1 probe. The Double oligonucleotide was used in

2.5-, 10-, and 25-fold molarexcesses since it contains two

potential binding sites (ZIIIA and ZIIIB). Samples were

electrophoresedthrough 4%polyacrylamide-0.5x TBEgels at room temperature or 4°C. Gels were then dried and

autoradiographed. Films fromtwooligonucleotide competi-tionexperiments were quantified by densitometer scanning on a Kodak

Biolmage

densitometer with Visage 2000 soft-ware. Disruption of complex formationwas determinedby comparing the ratios of shifted complex in lanes with and withoutcompetitor.

RESULTS

Determination of the ZEBRA DNA-binding domain by deletionalmutagenesis. A BZLF1 cDNA obtained fromRaji

cells was cloned in frame with the trpE coding sequences contained inthepATH 11

expression

vector

(Fig.

1B). The 75-kDafusion

protein

bounda

20-bp oligonucleotide

which encompasses an AP-1 consensus sequence (MS-AP-1) de-rived from the promoter of the EBV early gene, MS-EA

(Fig.

1Cand

D,

Z1-245).

Severaldeletionmutantsdefined the

domain of the protein which bound the AP-1 site (Fig. 1C and D). All of the mutants exhibited some temperature-sensitiveDNA-binding activity.Mutant

Z141-245

encoded the

smallest

protein,

which behaved like the

wild-type

ZEBRA

(Zl-245)

in bindingDNAatroomtemperatureand 4°C (Fig. lA).However,eventhismutantdidnotbindDNAaswellat

4°C

as did the

wild-type

Zl-245-

The

region

of ZEBRA encoded by exon II, which shares homology with the Fos basic domain (11), was insufficient to elicit DNA-binding

activity (Fig.

1CandD,

Z168202).

Although

ZA111-159

bound

DNA

unstably

(smearin

Fig. 1C)

at roomtemperature(22°C)

and

Z172-245

didnotbindDNAunder theseconditions

(Fig.

1C), both mutant proteins bound DNA at 4°C (Fig. 1D). Deletion ofthe18carboxyl amino acids ofZEBRAresulted in a mutant

(Zl-227)

which bound the AP-1 site at room temperaturebutnotat4°C

(Fig.

1CandD).Furthermore,all

ofthe ZEBRAdeletionmutantsshowed thesamepatternof

binding

tothe

divergent

AP-1

sites,

ZIIIAandZIIIB,present

in the BZLF1 promoter (data not shown). Thus, ZEBRA amino acids 172 to 227 were

required

for DNA

binding,

althoughneighboringsequences mayhaveinfluenced

protein

folding

or

stability

under sometemperatureconditions.

Site-directedmutagenesisofthe ZEBRAbasicdomain. The

region

ofZEBRAencoded

by

exonII ofBZLF1 (aa 168to

202)ispartially homologoustothebasicregionof Fos(Fig. 2A).It hasrecentlybeen shown thataa144to146and 154to 156 are required for the DNA-binding ability of Fos (35).

These basic amino acids are conserved at

positions

178 to 180 and 188 to 190 of ZEBRA. Therefore, we determined

whether thesamebasicamino acidsinthis

region

contribute

to the DNAbinding of ZEBRA. These residues were mu-tatedfrom Lys-Arg-Tyrto Glu-Glu-Leu and from

Arg-Lys-Cys

to

Glu-Glu-Ser, respectively (Fig.

2A).Thesemutations

abrogated

DNA

binding

totheconsensus(MS-AP-1)aswell

as

divergent

AP-1sites(ZIIIAandZIIIB)

(Fig.

2B).

Equiv-alentamountsof the in vitro-translatedwild-typeandmutant

proteins

wereusedin eachreaction(datanotshown). These

results demonstrate that the samebasic residues in ZEBRA J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

ZF RBRA

A CDNA

pATI' dcrivativcs

zI,414.

~i1J1-154

A:

1I- 4

C

4~~~~~

I~~~~~

+(4C)

;I. ,

II

22CIncubations

D

is-Z7..

f (22C)

B 7

r X -Z

Kd

4CIncubations

84~ ~ ~ 4 "O.

~~~w'*

....~~ ~ ~ ~ ~ ~ ~ ~ ..

and Fos are required for binding to AP-1 and related se-quences. However, deletional mutagenesis showed that

thesebasic residuesarenecessarybutnotsufficient for DNA binding (Fig. 1).

Binding ofZEBRAandFos-GCN4tooligonucleotides bear-ing sbear-ingle AP-1 and AP-1-like sequences. It was previously shown that ZEBRA footprinted sites ZIIIA at -129 (TGAGCCA) and ZIIIB at -116 (TTAGCAA) but not the Z-AP-1 octamersequence (TGACATCA) 67 bpupstreamof the BZLF1 transcriptional start site (13; Fig. 3A). A gel retardation assay wasusedtocomparetherelative affinities of ZEBRA and Fos-GCN4 for these sites in the BZLF1 promoter. TheFos basic domain has been showntobind the AP-1 consensus site as a homodimer when fused to the GCN4 leucine zipper (18, 24, 40). A Fos-GCN4 chimera, expressed as a TrpE fusion protein, bound the MS-AP-1 consensussite butnotZIIIBorZIIIA. ZIIIBdiffers from the consensusAP-1 siteat3bpandZIIIAcontainsasinglebase pair change from the consensus heptanucleotide AP-1 se-quence (Fig. 3B). Fos-GCN4 bound the Z-AP-1 octamer with high affinity, whereas ZEBRA interacted only weakly with this sequence. Thus, Fos-GCN4 binds the AP-1

oc-tamer and heptanucleotide sequences while ZEBRA binds the MS-AP-1, ZIIIA, and ZIIIB sites with relatively high affinity (Fig. 3B). The TrpE-ZEBRA and TrpE-Fos-GCN4 fusionproteins werederived from theurea-solubilized frac-tion oftotal E. coliproteinextracts.ZEBRAandFos-GCN4, expressed inthis system,bindDNAashomodimers (23). A control extractof bacterial protein did not exhibit specific bindingto anyofthese sites(datanotshown).

FIG. 1. DNA-binding activities of ZEBRAmutants. (A)

Struc-turesof ZEBRA deletionmutants. The ZEBRAcDNA is shownat

the top (30), and clones are designated accordingto the ZEBRA amino acidspresentintheconstruct.

Z,,111-159

isnamedaccordingto

the amino acids that have been deleted. Zl-167 was derivedfrom

genomic DNA (42). Translated sequences that are derived from

ZEBRA intron sequences are indicated by a hatched bar, and

sequencesfrom theexpressionplasmid polylinkerareindicatedbya

stippled bar. (B) Expression of TrpE-ZEBRA fusion proteins in bacteria. Total protein from bacteria that carried various TrpE-ZEBRA deletionproteinswasresolvedby SDS-polyacrylamide gel electrophoresis and stained with Coomassie brilliant blue. The

fusionprotein is indicated with dots in each lane. (C) Gel retardation

assayshowing binding ofTrpE-ZEBRA deletionmutants toa20-bp

32P-labeled oligonucleotide (4 x 10-' M) containing the AP-1

consensus sequence (11). Extract from bacteria transformed with

the trpE plasmid pATH 11 wasused as anegative control (vector lane). The various proteinswereeach incubated with the

oligonu-cleotideprobe and 1 p.gofpoly(dI-dC) in binding buffer for 30 min priortoelectrophoresis. Binding reactions and electrophoresiswere

performed at room temperature (22°C). Complexes were

electro-phoresedon a4%acrylamide gel in 0.5x TBE. (D) Gel retardation

assayof thesameTrpE-ZEBRA deletionmutantsperformedasfor panel C except that binding reactions and electrophoresis were carriedout at4°C.

Binding of ZEBRA and Fos-GCN4 to an oligonucleotide containing twoZEBRAresponseelements, ZIIIAand ZIIIB. Since the ZIIIA and ZIIIB sites are separated by only six nucleotides in the ZEBRA promoter, westudied the ability of ZEBRA and Fos-GCN4tobind both sitessimultaneously. AnoligonucleotidetermedDouble,whichincludes both sites intheorientation found in the BZLF1promoter,wasusedin gelband shiftassays. ZEBRAwasabletobindtwo siteson thisoligonucleotide simultaneously, asdemonstrated bythe

appearance of two band shift products in a mobility shift

assay(Fig. 3B). Surprisingly,the Fos-GCN4chimera,which didnotbindasingle ZIIIAorZIIIBsite,wasable to bindthe

Double oligonucleotide (ZIIIA plus ZIIIB), generating a moreslowly migrating complex (Fig. 3B).

C1 1 I

_1 - . - = -

-> N N N N N N NS N

Om S

< -C < ig

1- <z

zS

Cm

I

z Fl.-.,4,.

7 -4.

I

;I

" . =7

;.I

I .

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.612.57.291.79.407.2]

4036 TAYLOR ET AL.

A

132-LSPEEEEKRRIRRERNKMAAAXCRNRRRELTDTLQA-167

i1l

II

I1: II

167-SLEECDSELEIKRYKNRVASRKCRAKFKQLLQHYREV-203

---EEL

---EES

---B Z:A

,-% a',

OD r1.

Ns N rsJ

c-FOs

A

ZEBRA Z-M178-80 Z-M187-89

2 1,E tvIS At^

c~W. c c: OD K- OD cc

Ns N N .N rs N

ZEBRA (BZLFI) Promoter

ZIIIA ZIIIB

-140 GAAAc'l'A'I'GCATGAGCCACAGGCA'rl1'O;C'T'AATGTACCTCATAGACACACC

CTTTrGAC'TACG'T'ACT' CGGr G'G CCCG AACGATT.ACATIGGAGTATCTGTGTGG

Z-AP-1 OCTAMER

-90 T'AAA'I"I"I'A0CACG'G'CCCAAACCATGACATCACAGAGGAGGCTGGTGCCTT

A'ITTr I'AAATECGTlGCAGGGrGrT'I"ICGGTACTrGTrAGTGTCTCCTCCGACCACGGAA

-40 GGCTITT'AAAG0GGGAG0ATGTTCAGACAGGTAACTCACTAAACATTGCACCTT CCGAAA'I"IT'CCCCI'CTACAATCTGTCCATTGAGTGATTTGTAACGTGGAA

MS-EA (BSLF2-BMLF1) Promoter

ACGGTCACCTTCATGAGTc GTGcrTCGCCGGTCGCTGTGGGGCCAATCA

BZ_Z_P-1

[image:4.612.276.566.76.474.2]

B Octamer MS-AP-1 ZIIIA ZIIIB Double

FIG. 2. Effects of point mutations in the basic domain ofZEBRA.

(A) Amino acid sequencesof human c-FosexonIII(aa 132to167) (44) and BZLF1exonII(aa 168to202) (5). The conserved residues

in this highly basic domain are indicated by lines (11). ZEBRA

mutantswereobtainedby site-directed mutagenesisofthe BZLF1 cDNAcloned inSP64. Amino acids 178to180 and 187to189were

changedasshown for Z-M178-80andZ-M187-89,respectively. (B) Gel retardation assay demonstrating theinability ofZEBRA con-structs mutated at aa 178 to 180 (Z-M178-80) or 187 to 189 (Z-M187-89) to bind MS-AP-1, ZIIIA, or ZIIIB. The wild-type (Z-WT) and mutated ZEBRAconstructs(Z-M178-180 and Z-M187-189)weretranscribed in vitrowithSP6 polymeraseandtranslatedin awheatgermextract.Equivalentamountsof theresultingtranslated

proteinswere each incubated withthefollowing 32P-labeled

oligo-nucleotide probes: MS-AP-1,ZIIIA, and ZIIIB(seeMaterials and

Methods for sequences). These complexes were then resolvedby gel electrophoresison a4%acrylamide gelin0.5x TBE.

Wenextstudied the order ofbindingtositesonthe Double oligonucleotide. Serial dilutions of ZEBRA and Fos-GCN4 wereincubated with 4x 10-13M 32P-labeledDoubleprobe. Atlowproteinconcentrations(0.25 pl),ZEBRAboundonly a single site on the oligonucleotide (one shifted complex),

while athigh concentrations (5 ,ul), itbegantointeract with both sequences on a single oligonucleotide molecule (two shifted complexes) (Fig. 4A). However, Fos-GCN4

gener-ated a single complex of slower mobility with the Double probe, regardlessofprotein concentration(Fig. 4B).

Nucleotide specificity of ZEBRA and Fos-GCN4 binding sites. Point mutations were introduced into the MS-AP-1, ZIIIA,and ZIIIBoligonucleotides inanattempt toestablish whichbase pairs mediate specific DNAbinding by ZEBRA and Fos-GCN4. The oligonucleotides created were MS-AP-1* (TTAGTCA), ZIIIA* (TGAGCAA), and ZIIIB*

(T-CAGCAA).

ZIIIA*and ZIIIB* contain an identical 7-bp

core recognition site (TGAGCAA) but differ in their sur-roundingsequences(see Materials and Methods). Addition-ally, the ZIIIAand ZIIIB oligonucleotides were mutatedat fourpositions togenerate the ZIIIAm and ZIIIBm oligonu-cleotides.

ZEBRA bound with relatively high affinity to the

MS-z z

O G , O < C)

0 = L 0N(Do

> N > N LL >

z z z

= 0 (

U m c

N LL > N LL N U.

FIG. 3. BindingspecificitiesofZEBRAandFos-GCN4 forAP-1 and BZLF1 promoter sequences. (A) Sequences of the ZEBRA (BZLF1)and MS(BSLF2-BMLF1)promoters.AP-1-likesequences are underlined (11, 13), and the transcriptional start site ofthe BZLF1 gene is marked. (B) Gel retardation assay showing the specificities of TrpE (vector), TrpE-ZEBRA(Zl-245),and TrpE-Fos-GCN4 (Fos-GCN4) proteins for oligonucleotides containing the AP-1and BZLF1 promoter sequences: Z-AP-1 octamer,ZIIIA,and ZIIIB (13). The Double oligonucleotide contains sites ZIIIA and ZIIIBin theorientation foundinthe BZLF1promoter. Equivalent

amounts ofbacterially expressedfusion proteins were incubated with eacholigonucleotide(4x 10-13M) for30minpriortoanalysis

oftheretardationcomplexes by gelelectrophoresis.

AP-1, ZIIIA, and ZIIIB sequences

containing

a

single point

mutation (MS-AP-1*, ZIIIA*, andZIIIB*) (Fig.

5A).

How-ever,itwasunabletobind either the ZIIAmorZIIIBm

site,

eachof which contains4-bp changes (Fig.5A). The lackof

binding to the ZIIIAm or ZIIIBm site confirms the

impor-tance of the identified core element. Fos-GCN4 was more fastidious in itsbindingsiterequirements; it

recognized

the ZIIIB*(TGAGCAA)sequencebutdidnotinteract with any of the other mutated

oligonucleotides

(Fig.

5B).

ZIIIA* and ZIIIB* differ from the AP-1 consensus at two

nucleotides,

while ZIIIB differs at three positions and ZIIIA differs at J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.612.330.568.230.469.2]

ZEBRA DNA-BINDING SPECIFICITY 4037 A

0 - N G O

- a 0 0 - L

-_1W D:

*-.P

MS Double(ZIIIA*71113)

API

B N

L °~

- 0 o o

-MS Double(711

AP-1

FIG. 4. Binding of ZEBRA (A) and Fos-GCN4 (B) oligonucleotide. ZEBRA and Fos-GCN4 proteins we] gel retardation assay with 4 x 10-13 M 32P-labelec Double (ZIIIA plus ZIIIB) probe. Various volum

extract(indicated in microliters) wereincubated with

or Double probe in the presence of poly(dI-dC). Fos-GCN4 both bind to the MS-AP-1 site as a sii indicatedasS.Mobility shift complexes withaslowe

indicatedasD.

only one. ZIIIA* and ZIIIB* contain the same

nition sequence, but their surrounding sequ These results suggest that ZEBRA is considi tolerant of changes within the 7-bp recogniti4 than isFos-GCN4.

Competition analysis to determine the affinitic and Fos-GCN4 for AP-1 and ZEBRA promote

The relative affinities ofZEBRA and Fos-GCN4 for AP-1-° °, related sequences were assayed by competition analysis (Fig. 6). Complex formation between ZEBRAorFos-GCN4 and theMS-AP-1

probe

was

assayed

in thepresenceof three

M4-. D concentrations of cold competitor oligonucleotides. The

relativereduction ofAP-1complex formationisproportional

S tothe affinity of ZEBRAorFos-GCN4 for the

oligonucleo-tidecompetitor. Disruptionwasquantified by comparingthe integrated intensity of each shifted complexasafunction of the concentration of oligonucleotidecompetitor (Fig.6C and D).TheZEBRAcomplex formedwith theMS-AP-1 sitewas

most efficiently competed for with MS-AP-1, ZIIIA*, and ZIIIB* (Fig. 6A and C). A 5-fold molar excess of these

oligonucleotides

eliminated ZEBRA

binding

to 32P-labeled

MS-AP-1,whereas a50-fold molarexcess ofZIIIA,ZIIIB, orDoubleoligonucleotidewasrequiredtoachieveasimilar A4ZIlIB) level ofinhibition (Fig. 6A and C). The

ZEBRA/MS-AP-1

complexwasminimally reducedbytheaddition ofa50-fold ItotheDouble molarexcessof Z-AP-1octamerandwasnotaffected

by

the

retitrated ina addition ofunlabeledZIIIAm, ZIIIBm, or a control

oligonu-I MS-AP-1 or cleotide with no AP-1-like sequences (Fig. 6A and C and

tes of protein datanotshown). Thus,ZEBRA binds withhighest affinityto

itheMS-AP-1 TGAGTCA (MS-AP-1) and TGAGCAA(ZIIIA* andZIIIB*)

ZEBRA and sites and with a lower affinity to TTAGCAA (ZIIIB),

ngle complex, TGAGCCA(ZIIIA), TTAGTCA (MS-AP-1*), and the

Dou-,rmobilityare bleoligonucleotide (Fig. 6C and 7).

The Fos-GCN4/MS-AP-1 complex was efficiently

com-peted for with a 5-fold molar excess of Z-AP-1 octamer,

20-fold molarexcessof MS-AP-1,or50-fold molarexcessof

7-bp recog- ZIIIB* (Fig. 6B and D). Surprisingly, the Double oligonu-ences vary. cleotide,whichwasboundbyFos-GCN4 inagelretardation erably more assay, was unable to disruptthe formation ofa complexat

on sequence the MS-AP-1 site (Fig. 4; Fig. 6B and D). None of the remaining divergentAP-1 sitescompetedwith the formation esofZEBRA ofaFos-GCN4/MS-AP-1 ata50-fold molarexcess(datanot

!r sequences. shown). A secondcompetition experimentwasalso

quanti-_D

0E < <E

L0 0 o0

*E

wt =

N v - A s <

III

FIG. 5. Gel retardationassayillustrating the binding specificities of ZEBRA and Fos-GCN4 formutantrecognitionsequences.MS-AP-1, ZIIIA, andZIIIBoligonucleotidesweremutatedat asingle position (noted by*); the basesubstitution is underlined.Twooligonucleotides mutated at multiple positions are designated ZIIIAm and ZIIIBm. The control oligonucleotide contains no AP-1-related sequences. Bacterially expressed TrpE-ZEBRA (A) orTrpE-Fos-GCN4 (B) was incubated with these32P-labeledoligonucleotides, and binding was

analyzed byamobility shiftassay. VOL.65, 1991

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.612.57.290.72.248.2]

4038 TAYLOR ET AL.

A

Z AP-1 Octamer MS-AP-1 MS-AP-1' ZIIIA 0 52050 0 5 20 50 0 5 20 50 0 5 20 50

ZEBRA w " w asz

ZlIlABZIIIB Zll Double

0 520 50 0 5 20 50 0 5 20 50 0 5 20 50

ZEBRA

B

Competitor: Z-AP-1Octamer MS-AP-1 ZIIIB' Double

0 52055 05205 0520 50 0 5 20 50

Eaawt.r 2Ifg.au.&add.

MS-AP-1 BSLF2 TCTrCAfGGTGCTTC

TGrrcA313TGCTTC

z

mO

N

U-1 2

2

-Z-AP-1 BZLF1 CTAGAMCCAG~wAGGATC 3 1

(Octamer)

ZIIIA BZLF1 CTAGCTATGCA1AGATC

ZIIIA* - CTAGCTATGCAGATC

2

-1

-Fos-GCN4 _

C

a 0

._

.0

Clcn 0

u CL

H

n

~~Ll

!

5Foldexcess!

E 20Fold excess

~~~~~~~~~~~~~~~~~~ ..

§< §< - 8

Oligonucleotide Competitor

r% U 100

90

a 80

0

m 70 U- 60 50

a

0 40

30

20

10

nI

5 Fold excess

L_ 20 Fold excess

<

Oligonucleotide Competitor

FIG. 6. Competition experiment to evaluate the affinities of

ZEBRA and Fos-GCN4 for AP-1 and ZEBRA promotersites. A

32P-labeled MS-AP-1 oligonucleotide (4 x 10-13M)wasincubated

with bacterially expressed TrpE-ZEBRA (A)or TrpE-Fos-GCN4

(B)inthepresenceof three concentrations of cold competitorDNA.

Complex formationwasresolved by gelelectrophoresis. Only the shiftedcomplexesareshown, and thenatureofthe competitor and itsmolarexcess overthe MS-AP-1 probeareindicated. (C)Results

quantified by scanning of the relative intensity of each shifted complex.Thepercentdisruption of theZEBRA/MS-AP-1complex isplotted for each oligonucleotide competitor.(D) Resultsquantified asforpanel C fordisruption of the Fos-GCN4/MS-AP-1 complex.

ZIIIB BZLF1 CTAGCAGGCAtd1GTACCGATC 2

-ZIIIB* - CTAGCAGGCAf1GTACCGATC 1 3

FIG. 7. Evaluation ofZEBRAandFos-GCN4 binding site affin-ities. Affinities were determined from the competition analysis in Fig. 6. Nucleotide substitutions in the recognition sequences are

noted by bold type. Arelative affinity of1 representsthe highest affinity; - indicates that the protein did not bind the sequence specifically.

fied;therelative affinities of ZEBRA and Fosforthe various

oligonucleotides were found to be the same in the two experiments (data not shown).

DISCUSSION

Comparison ofZEBRA and FosDNA-binding domains. The

construction andexpressionof ZEBRA deletion mutantsin E. coli enabled us to determine the domains within the ZEBRAproteinwhicharerequired forDNAbinding. Unlike

allotheridentifiedtranscriptionfactorswhichbind the AP-1 consensus sequence, ZEBRA does not contain a leucine zipperdirectly adjacent to the basic region (25, 45). How-ever,these studies indicate that theregion3' tothe ZEBRA basic domain (aa 199 to 227) is required for DNA binding.

The C-terminal 18 amino acids (228 to 245) may also con-tribute todimerformationorprotein stability,sincemutants

lacking this region bind DNA in a temperature-sensitive fashion(Fig. 1CandD).Thisregion,exonIII(aa 199to245),

is nothomologous toanyknown dimerizationdomain (45). However, it has been hypothesized thatthis domain forms an alpha helix and that dimerization involves a coiled-coil interaction(14, 23,27, 36). These resultsareconsistent with those of Packham et al., who found that mutants which deleted the same regions ofexonI or III bound DNA in a

salt-dependentmanner,whereas thewild-typeprotein bound

DNA under allconditions (37).

Site-directedmutagenesisofc-fos previouslyidentified the amino acids necessary for DNA recognition of theFos/Jun complex (35). Mutagenesis of Fos amino acids 144 to146or 154 to 156 completely abrogated DNA binding (35). The

corresponding amino acids are conserved in the basic do-main of ZEBRA exon

II,

and proteins mutated in these

positions wereunableto bind anyof theDNA-binding sites

(MS-AP-1, ZIIIA, or ZIIIB) (Fig. 2). These results suggest that while the DNA-binding specificities of Fos-GCN4 and ZEBRA differ, the same two clusters of basic amino acids arenecessaryfor theDNA-binding abilities of bothproteins.

Furtherwork is neededtodefine the amino acids in the basic

region which confer theunique binding specificities of Fos-J. VIROL.

I

Li

#0

was -

1#

4 '. 4p

0*0

aw

'O

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.612.340.532.71.248.2] [image:6.612.64.300.73.612.2]

GCN4 and ZEBRA. We have conducted dimerization do-mainswapexperimentstoshow that DNAtargetspecificity, invitro, is contained exclusively within the basic region and is not influenced by the nature of the ZEBRA or GCN4

dimerization domain (23).

Binding sitespecificitiesof ZEBRA and Fos-GCN4.ZEBRA and Fos bind similar DNA sequences, and both ZEBRA expression and TPA treatment result in activation ofEBV lytic viral DNAreplication (8,9, 15, 41, 46). An underlying biological question is whether Fos, or other cellular AP-1 proteins, mediates TPA induction of the EBV lytic cycle. Whilewehavenotaddressedthisquestion directly,wehave begun by assessing the binding affinities of ZEBRA and Fos-GCN4 for several sites in the BZLF1 and MS-EA promoters. We chosetocompareZEBRA bindingtothat of Fos-GCN4, since the first indication that ZEBRA was a DNA-binding transactivatorcamefromarecognition of the homology between the basic domains of ZEBRA and c-Fos (11). While Fos generally binds DNA as a Fos/Jun het-erodimer, we used an artificial Fos-GCN4 construct, con-taining the Fos basic domain and the GCN4 leucine zipper, to allow Fos to bind DNA as ahomodimer. Manygroups have previously shown that the target site specificity ofa protein isnotaffected by thenatureof its leucine zipper (24, 32, 34, 40).

Several groups have recently examined ZEBRA binding activity in EBV promoters and the origin of EBV lytic replication (orilyt) (11, 13, 17, 21, 27, 28, 37, 43), but there hadbeennostudiesdirectly comparing the ability of Fos and ZEBRA to bind the same sites. Since our work was com-pleted, Chang etal. (6) reported onthebindingspecificities ofZEBRAand Fos/Junin ori lyt. Ourworkon the binding specificities ofZEBRAand Fosfor sites in the BZLF1 and MS-EApromotersprovides both complementary andnovel observations. Chang etal. demonstrated that ZEBRA binds an AP-1 consensus site with higher affinity than it binds divergent heptamer sites inori lyt, whereas we have found that it binds equivalently to an AP-1 site and a divergent

sequence (ZIIIA* or ZIIIB*) with high affinity (6). ZIIIA* and ZIIIB* contain the same coreheptamersequencefound

in ZRE2 inori lyt (6). Their datasuggestthatZEBRAbinds ZRE2 with high affinity (6). Both studies show that Fos/Jun and Fos-GCN4 are restrictive in binding divergent AP-1 heptamersites. However, wehavefound novel binding ofa Fos-GCN4 homodimer to a divergent heptamer sequence flankedbydyadsymmetry (ZIIIB*).

ZEBRAbounda muchbroaderrange of DNAsequences thandidFos-GCN4(Fig. 4). ZEBRAbound thepalindromic AP-1site,but italsoboundtheZIIIAandZIIIB sites,which do not contain perfect dyads. Unexpectedly, oligonucleo-tides withbase pair substitutions which disrupt the partial dyadnatureofZIIIAorZIIIB(TGAGCCAtoTGAGCAAor TTAGCAA to TGAGCAA) were bound by ZEBRA with increased affinity (Fig. 5 and 7). This departs from the proposed scissors-grip model ofDNA binding by dimeric proteins,inwhicheach proteinmonomerishypothesizedto bind anidentical DNA half-site (45). Although it is known that ZEBRA(expressedin vitroor asaTrpE fusion protein) binds DNA as a homodimer (6, 14, 23, 27), it is unclear whether both monomers contact or recognize symmetrical units on DNA. There are two possibilities which could explainZEBRAbindingto asymmetricsites: first, onlyone subunit ofthe ZEBRAhomodimer contacts DNA; or sec-ond, each subunit recognizes a wide variety of half-site

sequences,therebyrecognizingasymmetricsequences. Our preliminary data suggest that the second hypothesis

ac-countsforZEBRA's

asymmetrical

recognition ofDNA(23).

It has also beenfound that the homodimeric glutocorticoid

receptorbinds DNAasymmetricallyandis more sensitiveto mutations in the right half of its binding site (reviewed in reference 4).

Fos-GCN4bindsas ahomodimertothepalindromicAP-1 site (19, 24, 40; Fig. 3). Both Fos/Jun heterodimers and artificial Foshomodimershavebeen showntorecognizethe AP-1consensussequenceasymmetrically (39).This suggests that the Fos basic domains in a homodimer interact with heptamer target sequences in a manner equivalent to a

Fos/Jun heterodimer. However, a Fos-GCN4 homodimer

did not bind any of the divergent AP-1 sites with high affinity. Itwasunabletorecognize ZIIIA,ZIIIB,orZIIIA*,

allof whichwererecognized byZEBRAhomodimers(Fig.4 and7). Fos-GCN4 didbind ZIIIB*(TGAGCAA), although it

didnotbindZIIIA*; botholigonucleotides containthesame

7-bp core recognition sequence. This result suggests that

flanking nucleotides influence binding ability. ZIIIB*, but not ZIIIA*, contains the bases CA andTG adjacent to the

recognition sequence increasing the dyad symmetry ofthe

site, CATGAGCAATG. The importance of flanking se-quences in the binding affinity of Fos was noted by Nakabeppuand Nathans (32). Theyfound that substitutions ofcytosineand

guanosine

foradenine andthyminebasesat

flanking positions 1 and 10 of the cyclic AMP response element (CRE) resulted in greater than a 90% decrease of homodimericFosorFos/Junbinding activity (32).Our novel result indicates that the binding site specificity of Fos is

broader than originally proposed; Fos-GCN4 can bind an

asymmetric heptamer sequence with two mismatches from

the AP-1 consensus in an appropriate flanking

oligonucleo-tide environment.

While ZEBRA hasabroaderspecificitythanFos-GCN4 in

recognizingvariousheptamersequencesrelatedtothe AP-1 consensus, only Fos-GCN4 binds the Z-AP-1 octamer se-quence

(TGACATCA)

located in the BZLF1 promoter with

high affinity.This is interesting sincethis sitewasoriginally defined as the Jun/AP-1 response element in thejun pro-moterand is TPA

responsive

(2). Thus,ZEBRAisunableto binda subset of the knownAP-1/TPA-responsiveelements.

Fos/JunaswellasartificialFoshomodimersareknownto bind the octamer CRE

(TGACGTCA),

which differs from theZ-AP-1 octamer atposition5 (GtoA) (6, 12, 32, 33, 38, 39). Although the sites are similar, Deutsch et al. have shown thattheyhavedifferent

signal-responsive

properties;

that

is,

CREs are not

responsive

to TPA (10). Chang et al.

have shown that ZEBRAdoesnotbindaCRE(6).

Although

ZEBRAdoes notbind the

TPA-responsive

Z-AP-1octamer

site,bindingtothis site may beimportantin theinductionof the EBVlyticcascade. Forexample,Fosoranother cellular AP-1proteinmaybind this site in vivo(2, 3, 10, 26),

leading

to low levels of transcription from the BZLF1 promoter. ZEBRAcan then bindto the ZIIIA and ZIIIB sites in the BZLF1 promoter,

resulting

in autoactivation of this gene

(13).

Binding of ZEBRA and Fos-GCN4 to an oligonucleotide with twobinding sites. We studied ZEBRA and Fos-GCN4 bindingto

multiple

sitesastheyare

arranged

in the BZLF1 promoter.The arrangements of ZEBRA bindingsites in its ownpromoterand in the

lytic-origin divergent

promoterare very similar (28). ZRE1 (TGTGTAA) and ZRE2

(TGAG

CAA)

(identical

totheheptamercoresequenceintheZIIIA*

andZIIIB*

oligonucleotides)

in the

lytic

origin

are

arranged

in a tail-to-tailfashion

separated

by

10

bp

(28).

Sites ZIIIA (TGAGCCA) and ZIIIB (TTAGCAA) are

arranged

in the

on November 10, 2019 by guest

http://jvi.asm.org/

4040 TAYLOR ET AL.

same manner in the BZLF1 promoter, separated by 6 bp. This arrangement may be a common mode of targeting ZEBRAtranscriptional controlinthe EBV genome.

We found that two ZEBRA homodimers can bind ZIIIA and ZIIIB simultaneously on the Double oligonucleotide

(Fig. 3). We have not determined which site isfilledinitially,

although footprinting data have demonstrated that ZEBRA binds ZIIIB withhigheraffinitythan ZIIIA (13, 27).Gel shift

experiments with limitingdilutions ofZEBRAproteinshow

that botharebound only afterasignificant portion of single sites arefilled. Furthermore, competition experiments

indi-cate that ZEBRA does not have a greater affinity for the

Double oligonucleotide than for either single site (Fig. 6A andC).Thus, ZEBRAbindsnoncooperativelytothese sites. However, binding of ZEBRA to these two adjacent sites

may allow forsynergistictransactivation.

Fos-GCN4, which did not bind oligonucleotides bearing single ZIIIA orZIIIB sites, bound the Double oligonucleo-tide(containing ZIIIAandZIIIB)as ahigh-molecular-weight complex (Fig.3and4). Theassociation ofthisslow-mobility complex can be disrupted by addition ofZEBRA protein,

suggesting that the two proteins compete for a similar sequence ontheDoubleprobe(unpublished data). However,

the Doubleoligonucleotide didnotinterferewithFos-GCN4/

MS-AP-1 complex formation(Fig. 6B andD). One explana-tion for this phenomenon is that these associations are noncompetitive because the twoprobes bind different

Fos-GCN4complexes. The MS-AP-1 probeisbound by a Fos-GCN4homodimer,whereas theDoubleprobemaybindonly

a Fos-GCN4 complex with more than two monomers and

perhaps with otherproteins.

The interaction ofZEBRA with cellular AP-1 sequences mayresultin thedisplacement ofFos orothercellular AP-1

factors, thereby precluding theiractivity. Alternatively, by binding AP-1 sites, ZEBRA may transactivate (orpossibly

transrepress) cellular genes in a manner analogous to that with cellular AP-1 transcription factors. AP-1 proteins are

probably not able to activate transcription of EBV genes

through ZEBRA-responsive elements because atleast one

AP-1-binding protein, Fos-GCN4, is unable to bind those sites. The divergent nature oftheZEBRA-responsive sites makes it unlikely that cellular AP-1 proteins can have an

ongoingrolepromotingviraltranscriptionandactivatingthe EBVlytic cycle. Additionally, theabilityof Fostobind the TPA-responsive Z-AP-1 octamerand CRE sites,which are notboundby ZEBRA,indicatesthat ZEBRA cannotsimply

substitute for Fos. The ability ofthese related proteins to

bindsimilaraswellasdistinct sites adds yet another level of

complexity to the relationship of these transactivators in

regulating cellular and viral geneexpression.

ACKNOWLEDGMENTS

We aregratefulto E. Manet, A. Sergeant, and P. Farrell for the gifts ofBZLF1cDNAs, T.Kouzarides and E.Zifffor the Fos-GCN4 construct, and T. J.Koernerfor the pATHexpression plasmids. We also thank D. Crothers for his support and advice. N.T. would like toacknowledgethe support of E. & E. Lichtenstein.

This work was supported by Public Health Service grants CA12055, CA16038, and A122959 (to G.M.); RO1CA143 and R01CA52004 (to S.H.S.); IF32CA08482-01 (to E.F.); and CA09159-15 (toR.B.) from the National Institutes of Health and by grant NP 669K (toG.M.)from the American CancerSociety.N.T. is supported by the Medical Scientist Training Program of the NationalInstitutes of Health.

REFERENCES

1. Angel, P.,E. A.Allegretto,S.Okino,K.Hattori,W.J.Boyle,T. Hunter,and M.Karin.1988.Oncogene junencodesa sequence-specifictrans-activator similartoAP-1. Nature (London)332: 166-171.

2. Angel, P., K. Hattori,T. Smeal,and M. Karin. 1988. Thejun proto-oncogeneispositively autoregulated byitsproduct,Jun/ AP-1. Cell55:875-885.

3. Angel,P., M. Imagawa, R.Chiu, B. Stein, R.J. Imbra,H.J. Rahmsdorf,C.Jonat,P.Herrlich,and M.Karin.1987.Phorbol ester-induciblegenescontainacommoncis elementrecognized byaTPA-modulatedtrans-actingfactor. Cell 49:729-739. 4. Beato, M. 1989. Gene regulation by steroid hormones. Cell

56:335-344.

5. Biggin, M.,M. Bodescot,M.Perricaudet, and P.Farrell.1987. Epstein-Barr virus gene expression in P3HR1-superinfected Rajicells. J. Virol. 61:3120-3132.

6. Chang, Y.-N.,D.L.-Y.Dong, G. S.Hayward, andS. D. Hay-ward. 1990. TheEpstein-Barrvirus Zta transactivator:a

mem-ber of thebZIPfamilywithuniqueDNA-binding specificityand

a dimerization domain that lacks the characteristic heptad leucinezippermotif.J.Virol.64:3358-3369.

7. Chavrier, P.,H.Gruffat,A.Chevallier-Greco,M. Buisson,and A.Sergeant.1989. TheEpstein-Barrvirus(EBV) earlypromoter DRcontainsacis-actingelementresponsivetothe EBV

trans-activatorEB1andanenhancer with constitutive and inducible activities.J. Virol.63:607-614.

8. Chevallier-Greco, A., E. Manet, P. Chavrier, C. Mosnier, J. Daillie, and A.Sergeant. 1986. BothEpstein-Barrvirus(EBV) encoded trans-acting factors, EB1 and EB2, are required to

activatetranscriptionfroman EBVearlypromoter. EMBO J. 5:3243-3249.

9. Countryman, J.,andG.Miller. 1985. Activationofexpression

of latent Epstein-Barr herpesvirus after gene transfer with a small clonedsubfragment ofheterogeneous viralDNA. Proc. Natl.Acad.Sci. USA81:7632-7636.

10. Deutsch,P.J.,J.P.Hoeffier,J.L.Jameson,andJ.F. Habener. 1988. Cyclic AMP and phorbol ester-stimulated transcription

mediatedbysimilarDNAelements thatbinddistinctproteins.

Proc.Natl. Acad.Sci. USA 85:7922-7926.

11. Farrell,P.J., D. T. Rowe, C. M. Rooney, and T.Kouzarides. 1989. Epstein-Barr virus BZLF1 trans-activator specifically

bindsto a consensusAP-1site and is relatedtoc-fos. EMBOJ. 8:127-132.

12. Flemington, E., andS. Speck. 1990. Identification of phorbol esterresponseelements in thepromoter ofEpstein-Barrvirus putativelyticswitchgene BZLF1. J. Virol. 64:1217-1226. 13. Flemington, E.,and S.Speck. 1990. Autoregulationof

Epstein-Barr virusputative lyticswitch gene BZLF1. J.Virol. 64:1227-1232.

14. Flemington, E.,andS. H.Speck.1990. Evidenceforcoiled-coil dimerformation by an Epstein-Barr virus transactivator that lacksaheptadrepeatof leucine residues.Proc.Natl. Acad. Sci. USA87:9459-9463.

15. Gradoville, L., E. Grogan, N. Taylor, and G. Miller. 1990. Differences in the extent of activation of Epstein-Barr virus

replicativegeneexpressionamongfournon-producercell lines

stablytransformedbyoriP/BZLF1plasmids. Virology 178:345-354.

16. Hammerschmidt, W., and B. Sugden. 1988. Identificationand characterization oforilyt,alyticoriginofreplicationof

Epstein-Barrvirus.Cell55:427-433.

17. Hardwick,J.M.,P. M.Lieberman,andS. D.Hayward.1988.A

newEpstein-Barrvirustransactivator, R,inducesexpressionof

acytoplasmicearlyantigen.J. Virol.62:2274-2284.

18. Holley-Guthrie, E. A., E. B. Quinlivan, E.-C. Mar, and S. Kenney.1990.TheEpstein-Barrvirus(EBV)BMRF1promoter forearlyantigen(EA-D)isregulated bythe EBV trans-activa-tors, BRLF1 andBZLF1, in acell-specific manner. J. Virol. 64:3753-3759.

19. Hope,I.A.,and K.Struhl.1987.GCN4,aeukaryotic

transcrip-tional activatorprotein,bindsas adimertotargetDNA.EMBO J. 6:2781-2784.

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

20. Kenney, S.,J.Kamine, E. Holley-Guthrie,J.C. Lin, E. C. Mar, andJ. Pagano.1989. The Epstein-Barr virus (EBV)geneBZLF1

immediate-earlygeneproductdifferentially affects latentversus

productive EBVpromoters.J.Virol. 63:1729-1736.

21. Kenney, S., J. Kamine,D. Markovitz, R. Fenrick, and J. Pagano.

1988. An Epstein-Barr virus immediate-early gene product

trans-activates gene expression from the human immunodefi-ciency long terminal repeat. Proc. Natl. Acad. Sci. USA 85: 1652-1656.

22. Kleid, D. G., D. Yansura, B. Small, D. Dowbenko, D.M. Moore,

M.J. Grubman, P. D. McKercher, D. 0.Morgan, B.H.

Rob-ertson,andH. L.Bachrach. 1981. Cloned viral protein vaccine for foot and mouth disease responses in cattle and swine. Science214:3834-3838.

23. Kolman, J., N. Taylor,and G. Miller.Unpublished data. 24. Kouzarides, T., and E. Ziff. 1989. Leucine zippers of fos, jun

andGCN4 dictatedimerization specificity and thereby control DNAbinding. Nature (London)340:568-571.

25. Landschulz, W. H., P.F. Johnson, and S.L. McKnight. 1988. The DNA-binding domain of theratliver nuclearprotein C/EBP isbipartite. Science 243:1681-1688.

26. Lee, W., P.Mitchell, and R.Tjian. 1987. Purified transcription factor AP-1 interacts with TPA-inducible enhancer elements. Cell49:741-752.

27. Lieberman, P. M., andA. J. Berk. 1990. In vitro transcriptional activation, dimerization, and DNA-binding specificity of the Epstein-Barr virus Zta protein. J. Virol. 64:2560-2568. 28. Lieberman, P. M., M. J. Hardwick, J. Sample, G. S. Hayward,

and S. D. Hayward. 1990. The Zta transactivator involved in inductionof lytic cycle geneexpression in Epstein-Barr virus-infectedlymphocytes bindstoboth AP-1 and ZRE sites intarget promoterand enhancer regions. J. Virol. 64:1143-1155. 29. Luka, J., B. Kallin, and G. Klein. 1979. Induction of the

Epstein-Barr virus (EBV) cycle in latently infected cells by n-butyrate. Virology 94:228-231.

30. Manet, E., H. Gruffat, M. C. Trescol-Biemont, N. Moreno, P. Chambard, J.F.Giot, and A. Sergeant. 1989.Epstein-Barr virus bicistronic mRNAs generated by facultative splicing code for

twotranscriptional trans-activators. EMBO J. 8:1819-1826. 31. Melton, D. A., P. A. Krieg, M. R. Rebagliati, T. Maniatis, K.

Zinn, and M. R. Green. 1984. Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res. 12:7035-7056.

32. Nakabeppu, Y., and D. Nathans. 1989. The basic region of Fos mediatesspecific DNA binding. EMBO J.8:3833-3841.

33. Nakabeppu,Y., K. Ryder, and D. Nathans. 1988. DNA binding

activitiesofthreemurineJunproteins: stimulation byFos. Cell 55:907-915.

34. Neuberg, M., J. Adamkiewicz, J. B. Hunter, and R. Muller. 1989. A Fos protein containing the leucine zipper forms a homodimerwhichbinds totheAP1binding site.Nature (Lon-don)341:243-245.

35. Neuberg, M., M. Schuermann, J. B. Hunter, and R. Muller. 1989. Twofunctionallydifferent regions inFos arerequired for the sequence-specific DNA interaction of the Fos/Jun protein complex.Nature(London)338:589-590.

36. O'Shea, E. K., R. Rutkowski, and P. S. Kims. 1989. Evidence that theleucine zipperisacoiled-coil. Science 243:538-543. 37. Packham, G., A. Economou, C. M. Rooney, D. T. Rowe, and

P.J. Farrell. 1990. Structureandfunction of the Epstein-Barr virus BZLF1protein. J.Virol. 64:2110-2116.

38. Rauscher, F. J., P. J. Boulalas, B. R. Franza, Jr., and T. Curran. 1988. Fos and Jun bindcooperatively tothe AP-1site: recon-struction invitro.GenesDev.2:1687-1699.

39. Risse, G., K. Jooss, N.Manfred, H. J. Bruller,and R. Muller. 1989.Asymmetricalrecognition of the palindromicAP1binding site (TRE)byFosprotein complexes. EMBOJ. 8:3825-3832. 40. Sellers,J.W., and K. Struhl.1989.ChangingFosoncoproteinto

aJun-independent DNA-binding protein with GCN4 dimeriza-tionspecificity by swapping 'leucine zippers.'Nature(London) 341:74-76.

41. Takada, K., N.Shimizu, S. Sakuma,andY. Ono. 1986. trans-activation of the latentEpstein-Barrvirus(EBV)genomeafter transfection of theEBV DNAfragment.J.Virol.57:1016-1022. 42. Taylor,N., J. Countryman, C. Rooney, D. Katz, and G. Miller. 1989. Expression ofthe BZLF1latency-disriptinggenediffers in standard and defective Epstein-Barr viruses. J. Virol. 63: 1721-1728.

43. Urier, G.,M.Buisson, P. Chambard, and A.Sergeant.1989. The Epstein-Barr virus early protein EB1 activates transcription from differentresponsive elements includingAP-1binding sites. EMBOJ.8:1447-1453.

44. van Straaten, F., R. Muller, T. Curran, C. van Beveren, and I.M. Verma.1983.Completesequenceofahumanc-oncgene: deduced amino-acidsequenceof the human c-fosgeneprotein. Proc. Natl. Acad. Sci. USA 80:3183-3187.

45. Vinson, C. R., P. B.Sigler,andS. L. McKnight. 1989. Scissors-gripmodel forDNArecognition by afamily of leucine zipper proteins. Science 246:911-916.

46. zur Hausen,H.,F.J.O'Neill,U. K.Freese, and E. Hecker.1978. Persisting oncogenic herpesvirus induced by the tumor pro-moter TPA.Nature(London)272:373-375.

on November 10, 2019 by guest

http://jvi.asm.org/