Epstein-Barr switching - pnas.org · Proc. Natl. Acad. Sci. USA Vol. 88, pp. 3942-3946, May1991 Biochemistry Rolefor the Epstein-Barrvirus nuclearantigen2 in viral promoter switching

Proc. Natl. Acad. Sci. USAVol. 88, pp. 3942-3946, May 1991Biochemistry

Role for the Epstein-Barr virus nuclear antigen 2 in viral promoterswitching during initial stages of infection

(infection of primary B lymphocytes/Epstein-Barr virus nuclear antigen 2-dependent enhancer/Si nuclease protection/in vitrotranscription/DNase I footprinting)

MAXIMILIAN WOISETSCHLAEGER, XIAN W. JIN, CHANDRI N. YANDAVA, LISA A. FURMANSKI,JACK L. STROMINGER, AND SAMUEL H. SPECKDivision of Tumor Virology, Dana-Farber Cancer Institute and Department of Pathology, Harvard Medical School, 44 Binney Street, Boston, MA 02115

Contributed by Jack L. Strominger, January 2, 1991

ABSTRACT During latent Epstein-Barr virus (EBV) in-fection of human B lymphocytes, six viral nuclear antigens(EBNAs) are expressed from long primary transcripts bymeans of alternative splicing and alternative polyadenylylationsites. These transcripts initiate from one of two promoters, Cpor Wp, that function in a mutually exclusive fashion. Wp isexclusively utilized during the initial stages of infection ofprimary B lymphocytes, followed by a switch to Cp usage.These studies have been extended to show that (i) a mutantEBVstrain lacking the gene encoding EBNA 2 fails to switch fromWp to Cp usage in primary B lymphocytes, although the viruscontains a functional Cp; (ii) a region from -429 to -245 basepairs upstream of Cp is essential for Cp activity in B lympho-cytes, but only in the context of upstream and downstreamsequences; (iii) this region contains an EBNA 2-dependentenhancer; and (iv) DNase I protection employing nuclearextracts from B and T lymphocytes revealed a B-cell-specificfootprint in the region of the EBNA 2-dependent enhancer.These results support a model for viral promoter switchingduring the initial stages of infection in which Wp activity leadsto the expression of EBNA 2, followed by activation of Cpthrough the EBNA 2-dependent enhancer.

Epstein-Barr virus (EBV) is a human lymphotopic herpes-virus that predominantly establishes a latent infection in Blymphocytes, with a concomitant growth transformation ofthe infected cells. During viral latency, six EBV nuclearantigens (EBNAs) are known to be expressed, as well asthree viral membrane antigens. The viral genes encodingthese antigens are spread over nearly two-thirds of the172-kilobase (kb) viral genome. However, transcription ofthese genes initiates from promoters that are clustered nearthe fused terminal repeats on the viral episome (for review,see refs. 1-3).Two promoters have been identified (Cp and Wp) that are

involved in driving transcription of the six EBNA genes, andthey function in a mutually exclusive fashion (4). Infection ofperipheral B lymphocytes with EBV leads initially to exclu-sive Wp usage, followed by up-regulation of Cp (5). In thispaper the issue of the switch to Cp usage is addressed.

MATERIALS AND METHODSCells, Transfections, and Chloramphenicol Acetyltrans-

ferase (CAT) Assays. All cell lines were grown at 37°C inRPMI 1640 medium supplemented with 10% (vol/vol) fetalbovine serum. DNA transfections were carried out either byusing Lipofectin (BRL; ref. 4) or, in the case of the EBV-negative Burkitt lymphoma cell line DG75, by using DEAE-dextran as described (6). CAT assays were carried out as

described (7). The EBNA 2 expression vector employed inthe cotransfection experiments, pEAA6, has been described(8). SVpCAT contains an enhancerless simian virus 40(SV40) early promoter and was described earlier (9). EBVsequences from -429 (Avr II) to -245 (Sac I) upstream ofCpwere cloned in both orientations upstream of the SV40promoter ([-429/-245]SVpCAT and [-245/-429]SVp-CAT) or downstream of the CAT gene (SVpCAT[-429/-245]). In SVpCAT[-340/-245], a (shorter) 95-base-pair(bp) Kpn I/Sac I fragment upstream of Cp was cloned intoSVpCAT. The C1CAT plasmid series has been described (4).In Vitro Infection of Peripheral Blood Lymphocytes, RNA

Preparation, and S1 Nuclease Protection. Virus was preparedfrom Akata cells (an immortalizing strain of EBV) (10) byinduction with anti-IgG as described (5). Virus was preparedfrom the clone-13 cell line (an EBV-positive, EBNA 2-neg-ative Burkitt lymphoma cell line) after induction with phorbol12-myristate 13-acetate at 20 ng/ml for 10 days. Infectionswere carried out as described (5). Total cellular RNA wasprepared by the method of Auffray and Rougeon (11). Fiftymicrograms of RNA was used in S1 nuclease protectionanalyses, employing specific oligonucleotides diagnostic forCp or Wp activity (4).Cloning Cp. EcoRI/Xho I restriction endonuclease-

digested genomic DNA from the Cp-using lymphoblastoidcell line JY and from clone-13 was cloned into the Uni-ZAPXR vector (Stratagene), and positive clones were recoveredby screening the library with an oriP-containing probe. DNAsequencing revealed a new Xho I site in the clone fromclone-13 at position 13,143 (with respect to the B95.8 map),just upstream of the BamHI C/Wjunction at 13,215, resultingin a 5.8-kb insert containing only BamHI C sequences. TheXho I site in the clone-13 clone was converted to a BamHIsite, and the entire insert was used to replace the equivalentEcoRI/BamHI fragment in CWWlglobin (see Fig. 3), whichresulted in the deletion of 72 bp from the BamHI C fragment.

Preparation of Extracts, in Vitro Transcription, and DNaseI Footprinting. Nuclear extracts were prepared as described(12) with slight modifications (9). In some cases whole-cellextracts were prepared by omitting the crude nuclei prepa-ration step. DNase I footprinting and in vitro transcriptionwere carried out as described (11, 13, 14).

RESULTSAn EBNA 2 Mutant EBV Strain Fails to Switch from Wp to

Cp Usage in Primary B Lymphocytes. Based on the kinetics ofthe switch from Wp to Cp and the appearance of the EBNAs,it seems likely that one or more of these viral antigens areeither directly or indirectly involved in promoter switching(5, 15). EBNA 2 and EBNA 4 appear to be the first viral

Abbreviations: EBV, Epstein-Barr virus; EBNA, EBV nuclearantigen; CAT, chloramphenicol acetyltransferase; SV40, simianvirus 40.

3942

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 88 (1991) 3943

antigens expressed upon infection of primary B lymphocytes(15). To begin to address their possible roles in promoterswitching, an infection time course was carried out byemploying the nonimmortalizing clone-13 strain of EBV (asubclone of the mutant P3HR-1 strain) (16). This viruscontains a deletion that removes the entire EBNA 2 gene andthe exons encoding the carboxyl terminus of EBNA 4 (17).

Infection of adult peripheral blood lymphocytes withclone-13 virus led to detectable Wp activity within 24 hrpostinfection, which remained constant over the first 6 days(Fig. 1B). However, there was no detectable Cp activity overthe same time range. In contrast, infection ofadult peripheralblood lymphocytes with the immortalizing Akata strain ofEBV led to detectable Cp activity by 48 hr postinfection, anda pronounced down-regulation of Wp activity was observedby 96 hr postinfection with a concomitant rise in Cp activity(Fig. 1A).Clone-13 Virus Contains a Functional Cp. Since the

clone-13 Burkitt lymphoma cell line (from which clone-13virus was induced) utilizes Wp exclusively (4), the failure ofclone-13 virus to switch to Cp usage upon infection ofprimary B lymphocytes may be due to the presence of anonfunctional Cp in this strain ofEBV. To directly determinewhether Cp is functional in the clone-13 genome, a 5.8-kbregion of the genome containing Cp was cloned (from theEcoRI site at 7316 bp to the Xho I site at 13,143 bp). As acontrol, the same region was also cloned from the Cp-usinglymphoblastoid cell line JY (18) (see Materials and Methods).Both the clone-13 and JY fragments were cloned into a

reporter construct (CWWlglobin; see Fig. 3 for structure) inplace of the cloned fragment isolated from the B95.8 cell line(5, 19). These reporter plasmids were introduced back intothe clone-13 cell line and the EBV-negative Burkitt lym-phoma cell line DG75 (20), and Cp activity was assessed (Fig.1C). In both cell lines, no difference in Cp activity wasobserved with the B95.8, JY, and clone-13 reporter plasmids.Thus, Cp from any of these viral strains introduced into theCWWlglobin reporter construct functions normally in cy-cling B lymphocytes (i.e., in both clone-13 and DG75 cells).The lack of Cp activity from the endogenous virus in theclone-13 cell line is therefore not the result of a nonfunctional

Cp. These results suggest that the failure to switch from Wpto Cp usage in primary cells infected with clone-13 virus isdue to the absence of one or more of the EBNAs.A Region Essential for Cp Activity Functions as an Enhancer

in EBNA 2-Positive Cell Lines but Not in the EBNA 2-NegativeClone-13 Cell Line or in EBV-Negative Cell Lines. To initiallymap cis-acting elements involved in regulating Cp activity,several deletions in the region upstream of Cp were gener-ated. These deletion constructs were assayed in the EBV-positive Burkitt lymphoma cell lines Jijoye (which expressesEBNA 2 and is the parent of P3HR-1 and clone-13) (21) andclone-13 and in the EBV-negative Burkitt lymphoma cell lineDG75 (Fig. 2). Deletion of a 1.2-kb region (-2199 to -1021bp upstream of Cp) had no effect on Cp activity in any of thecell lines assayed (see construct dl, Fig. 2). However,deletion of a 776-bp region from - 1021 to -245 bp abrogatedCp activity almost entirely (see construct d2, Fig. 2). Asmaller deletion from -429 to -245 bp also had the sameeffect, mapping essential sequences to a 184-bp region (seeconstruct d6, Fig. 2). Deletion of the sequences from -1021to -429 bp (see construct d5, Fig. 2; data not shown) did notaffect Cp activity, indicating that the essential cis-actingelements map entirely within the region from -429 to -245bp.To determine whether this region functions as an enhancer,

it was cloned in both orientations upstream of the enhancer-less SV40 early promoter and transfected into various celllines (Table 1). In the EBNA 2-positive Jijoye cell line, theregion from -429 to -245 bp exhibited enhancer activity inboth orientations (see experiments 1-3, Table 1). However,in the EBNA 2-negative cell line clone-13, this region did notfunction as an enhancer, but rather it significantly suppressedthe activity of the SV40 early promoter. Furthermore, in theEBV-negative cell line DG75, this region also did not exhibitenhancer activity.EBNA 2 Is Required for Enhancer Activity. To determine

whether EBNA 2 expression is required for enhancer func-tion, several cell lines were cotransfected with the appropri-ate reporter construct and either a control plasmid or anEBNA 2 expression plasmid. In DG75 cells the region from-429 to -245, cloned either upstream or downstream of a

C

Akata virusc

24hr 48hr 96hr0

0 c3'0 3. 0F5 3 05

B

Cl-13 virus ,

IDwow1 Pl

mco

z

a,

a c= ~-cr = 4

!M%, 3

DG75 cells

Cl- 3 cells 6 '(Cl probe) m

a Qa0 0

< uS 3

> On -

i

0

protectedfragments

-_

protected WI4 4protected fragments

fragments

FIG. 1. Infection time course of adult peripheral blood lymphocytes infected with either Akata virus (A) or clone-13 (CI-13) virus (B). Totalcellular RNA was prepared from cells harvested at the indicated times postinfection, hybridized to C1 exon-specific and WO/W1 splice-junction-specific oligonucleotides, followed by digestion with S1 nuclease to determine Cp and Wp activities, respectively. (C) Analysis of Cpactivity ofexogenous constructs transfected into either the clone-13 or DG75 cell line. RNA was harvested 72 hr posttransfection and hybridizedto C1 exon-specific and WO/W1 splice-junction-specific oligonucleotides (5), followed by digestion with S1 nuclease. The resulting products wereresolved on a 10%o urea/acrylamide gel. Cp(B95.8), CWWlglobin reporter construct containing viral BamHI C sequences cloned from the B95.8marmoset cell line; Cp(JY), CWWlglobin reporter construct containing viral BamHI C sequences cloned from the JY cell line; Cp(CI-13),CWWlglobin reporter construct containing viral BamHI C sequences cloned from the clone-13 cell line. The structure of the CWWlglobinreporter construct is shown in Fig. 3.

A

inputprobe

Biochemistry: Woisetschlaeger et al.

3944 Biochemistry: Woisetschlaeger et al.

JIJOYE

pinput _ _probe

protected s.

fragments

C

oriP

Exons-4018

-2199

wCp E Wp

co mC1 C2

+1Fel

WO W2Wi

y

Y1_ Y3Y2

--Globin

--noi

n[loi

-1021

-1021 -245I

-429 -245

-1021 -429i~~ ~ ~~~

FIG. 2. Analysis ofCp activityin several deletion constructs. TheJijoye, clone-13, and DG75 celllines were transfected with the in-dicated reporter constructs; RNAwas prepared 72 hr posttransfec-tion and analyzed by S1 nucleaseprotection employing a C1 exon-

wt specific probe. The resulting prod-d1 ucts were resolved on a 10%o urea/acrylamide gel. To distinguish be-d2 tween endogenous and exogenous

Cp activity in the Jijoye cell line,d6 the C1 exon of the plasmid was

d5 tagged by site-directed mutagene-sis as described (5). wt, wild type.

CAT reporter construct containing the enhancerless SV40early promoter, exhibited EBNA 2-dependent enhancer ac-tivity (see experiments 6-8, Table 1). Deletion of the se-quences from -429 to -340 bp eliminated virtually all of theEBNA 2-dependent enhancer activity (see experiments 6 and9, Table 1). EBNA 2-dependent enhancer activity was furthersubstantiated by employing reporter constructs in which theregion from -429 to -245 bp was in its physiological settingupstream of Cp and the CAT gene was linked to the C1 exon(C1CAT constructs). In all cell lines examined, cotransfec-tion ofthe EBNA 2 expression vector with the -820/ClCATand -559/ClCAT constructs resulted in increased activity.However, no apparent EBNA 2-dependent enhancer activitywas observed with the -245/ClCAT plasmid.The EBNA 2 Enhancer Region Is Essential for Cp Activity in

EBNA 2-Negative Burkitt Lymphoma Cell Lines in the Pres-ence of Upstream and Downstream Sequences but in ThisContext Is Not Responsive to EBNA 2. As already shown in

Table 1. Relative activities of various CAT reporter constructsin the presence and absence of EBNA 2

Expression Relative activityExp. Reporter construct plasmid DG75 Clone-13 Jijoye

1 SVpCAT Control (1.0) (1.0) (1.0)2 [-429/-245]SVpCAT Control 0.8 0.4 5.53 [-245/-429]SVpCAT Control 0.7 0.2 4.34 SVpCAT[-429/-245] Control 0.65 SVpCAT[-340/-245] Control 1.26 SVpCAT EBNA 2 1.77 [-429/-245]SVpCAT EBNA 2 13.4 -8 SVpCAT[-429/-245] EBNA 2 6.49 SVpCAT[-340/-245] EBNA 2 2.110 -820/ClCAT Control (1.0) (1.0) (1.0)11 -559/ClCAT Control 1.0 2.2 0.712 -245/ClCAT Control 0.7 5.4 0.113 CWY2CAT Control 0.614 -820/ClCAT EBNA 2 5.8 3.6 2.515 -559/ClCAT EBNA 2 9.6 5.1 1.416 -245/ClCAT EBNA 2 0.7 3.2 0.117 CWY2CAT EBNA 2 1.0

-, not determined.

Fig. 2, the region from -429 to -245 bp upstream of Cp isessential for Cp activity in both EBNA 2-positive and EBNA2-negative Burkitt lymphoma cell lines in the context ofupstream and downstream sequences. However, whenC1CAT constructs containing limited upstream and down-stream sequences were assayed, deletion ofthis region (in theabsence of the EBNA 2 expression vector) reduced activityonly in the EBNA 2-positive cell line Jijoye (see experiments11 and 12, Table 1). Indeed, deletion of these sequencesresulted in increased activity in the clone-13 cell line, con-sistent with the observation that these sequences act as asuppressor in clone-13 cells in the absence of EBNA 2(compare experiments 1 and 2 and also 11 and 12, Table 1).

It is possible, however, that with the short C1CAT con-structs the observed activity did not originate from Cp butrather from promoters associated with the plasmid. To di-rectly address this issue, transcription initiating from Cp wasassessed by employing an in vitro transcription system withclone-13 and DG75 whole-cell extracts. With both extracts,the initial observation that the region from -429 to -245 bpis essential for Cp activity, in the context of upstream anddownstream sequences (CWWlglobin and CWWld6globin),was reproduced in the in vitro transcription system (Fig. 3).However, comparison of Cp activity exhibited by either-820/ClCAT or -559/ClCAT and -245/ClCAT in the invitro system revealed that the -559 to -245 bp sequence isnot required for Cp activity in these constructs.

Extending the upstream region to include oriP did notsignificantly alter the dependence of Cp activity on the d6(-429 to -245 bp) region (see oriP/C1 vs. oriP/Cld6, Fig. 3).However, deletion of the sequences from -534 to -245 bpdid appreciably diminish Cp activity in the context of exten-sive downstream sequences (see -534/CWW1 vs. -245/CWW1, Fig. 3), although this effect was small compared tothe effect of deleting these sequences in the context of bothupstream and downstream sequences (see CWW1 vs.CWWld6, Fig. 3).

Since the EBNA 2-dependent enhancer region is essentialfor Cp activity in the context of extensive upstream anddownstream sequences, even in the absence of EBNA 2, itwas important to determine whether EBNA 2 expressioninfluenced Cp activity in this setting. Cotransfection ofDG75

CLONE-1 31~V- N (0

0

DG75

._

Proc. Natl. Acad. Sci. USA 88 (1991)

Proc. Natl. Acad. Sci. USA 88 (1991) 3945

Clone-13 DG75 DG75 Clone-13

G id

Hi_ @__ En @

as 0C~y

F

Ai.

coriP Cp

,, f.Exons CIC2

-4018 +1

-429 -245

-534

E wpw

WO W2

WI

-245

CAT orlP/C

820 { onP/Cld6

5? -820/Cl

.X{Z-559/Cl

-245/Cl

E wp

WIwWI

kL~lin CWW1-4Z CWWld6

= -534/CWWI= -245/CWW1

FIG. 3. Influence of upstream and downstream sequences on thedependence of Cp on the EBNA 2 enhancer region. The indicatedreporter constructs were assayed by employing an in vitro transcrip-tion system with either DG75 or clone-13 whole-cell extracts. RNAwas recovered from the reactions, hybridized with a C1 exon-specificoligonucleotide probe, and digested with S1 nuclease. The resultingproducts were resolved on a 10%o urea/acrylamide gel. The filledarrowhead indicates total protection of the input probe, presumablyresulting from transcription initiating upstream of Cp. The openarrowhead indicates specific protection of transcripts initiating fromCp.

cells with CWY2CAT (see wt in Fig. 2 for structure) andeither a control plasmid or an EBNA 2 expression vectorrevealed that there was little transactivation by EBNA 2(experiments 13 and 17, Table 1). This observation wasconfirmed by S1 nuclease protection monitoring transcrip-tion initiating from Cp (data not shown). These data suggestthat elements within the EBNA 2 enhancer interact withdistal elements and that these interactions preclude furtherenhancement by EBNA 2.B Lymphocyte-Specific Transcription Factor Binding in the

Region Containing the EBNA 2-Dependent Enhancer. Toidentify regions binding transcription factors that may beimportant for EBNA 2-dependent enhancer activity, DNaseI protection was carried out by employing nuclear extractsprepared from several lymphocyte cell lines (Fig. 4A). A largefootprint extending from approximately -380 to -330 bp wasobserved with all B-lymphocyte extracts examined. Protec-tion ofthis region by extract prepared from the EBV-negativeBurkitt lymphoma cell line Ramos indicates that it is notdependent on the presence of EBV antigens. Notably, ex-tracts prepared from the T-cell line Jurkat only protected afraction of the extensive B-lymphocyte footprint (Fig. 4B). Inaddition, extracts prepared from the clone-13 cell line spe-cifically protected the region from about -310 to -330 bp.The latter result may be of some interest since in clone-13cells the EBNA 2 enhancer acts as a suppressor in theabsence of EBNA 2, an effect not observed in the EBV-negative Burkitt lymphoma cell line DG75.

DISCUSSIONIn this paper we have shown that infection of peripheral Blymphocytes with an EBNA 2-negative EBV mutant did not

result in a switch from Wp to Cp usage. This failure to switchfrom Wp to Cp usage could not be attributed to a mutant Cpin the viral genome. Furthermore, characterization of theregion upstream of Cp identified a region that is essential forCp activity in the context of upstream and downstreamsequences and that contains an EBNA 2-dependent en-hancer. However, EBNA 2-responsive enhancer functioncould only be demonstrated in the absence of extensiveupstream and downstream sequences.

Several important conclusions can be inferred from thisdata: (i) since clone-13 virus fails to switch from Wp to Cpusage in primary B lymphocytes, expression of some of theEBNAs is required in these cells for Wp to Cp switching; (ii)the preferential utilization of Cp over Wp in all Burkittlymphoma cell lines examined (both EBNA 2-positive andEBNA 2-negative) indicates that in cycling B lymphocytesEBNA 2 is not required for Cp activity; (iii) the inability ofEBNA 2 to enhance Cp activity in the context of upstreamand downstream sequences indicates that elements within theenhancer are involved in interactions with distal elements andthat these interactions preclude further enhancement byEBNA 2.When the DNA sequence of the EBNA-dependent en-

hancer was compared to the EBV BamHI C and W se-quences, two regions of striking homology were identified.The B-cell-specific portion of the DNase I footprint containsthe sequence 5'-GAGTGCTATCTTTGGA-3', which is ho-mologous to a region in oriP at -3927 to -3943 bp upstreamof Cp (5'-GAGTGCTATCcTTtGc-3') and to another regionlocated downstream of Cp in the BamHI W fragment atposition +2829 to +2844 bp (5'-tgcTGCTATCTTraGA-3').Given the striking homology common to these three se-quences it is likely that they bind the same factor(s), whichmay be involved in a three-component interaction betweenthe EBNA 2-dependent enhancer and distal elements, thusforming a tripartite regulatory structure that is essential forthe activity of Cp.The observation that Cp cannot be transactivated by

EBNA 2 in reporter constructs containing extensive up-stream and downstream sequences raises the question of thesignificance of the EBNA 2-dependent enhancer and the roleof EBNA 2, if any, in promoter switching. This, in turn,relates to the more fundamental question of the respectiveroles of Wp and Cp. There are three obvious differencesbetween Wp and Cp. First, there are multiple copies of Wp,which may be important for generating sufficient transcrip-tion of EBNA 4 and EBNA 2 during the initial phase ofinfection. Second, the 5' ends of Cp- and Wp-initiatedtranscripts are distinct, and this may influence the splicing ofthe long primary transcripts. Finally, the 5' regulatory re-gions of Cp and Wp are different, and this may reflect a needto run these promoters in different activation and differenti-ation states of the infected cells.

While it appears certain that EBNAs 2 and 4 are expressedfrom Wp-initiated transcripts during the initial phase ofinfection,it is not known whether the other EBNAs are also expressedfrom Wp transcripts or whether switching to Cp is requiredbefore they can be expressed. There is some evidence in supportof a model in which expression of EBNAs 1, 3a, 3b, and 3crequires switching to Cp usage. It has been shown that infectionofprimary B lymphocytes with clone-16 virus (both clone-13 andclone-16 are subclones of the P3HR-1 cell line) leads to expres-sion of EBNA 4 but not EBNAs 1, 3a, 3b, or 3c (19). Further-more, no EBNA 1 or 3 cDNA clones have been characterizedthat initiate from Wp in Wp-using cell lines, leaving unresolvedthe question of their origin. In addition, a recent report demon-strated successful immortalization ofhuman PBLs with clone-13virus that had regained a functional EBNA 2 gene (22). Analysisof promoter usage in these cell cultures revealed Cp usage in

Biochemistry: Woisetschlaeger et al.

3946 Biochemistry: Woisetschlaeger et al.

+ LU coAzC r

-334b -

-382bp-

SF

Cp(+)I

+ WL coa Z 0 9*

CI-13

.R

v~V:

so ..'

I

lkIt-

M* -__

B L C)

<X <00 a: - Z

I

327bp~

--362bpA U- ~ ~ ~ -8b

-327 bp ^

FIG. 4. DNase I footprintinganalysis of the region containingthe EBNA 2-dependent enhancer.Footprinting was performed withcrude nuclear extracts preparedfrom the indicated cell lines[Ramos, an EBV-negative Burkittlymphoma cell line; Ramos/B95.8, EBV-converted Ramos; Ji-joye, an EBV-positive Burkittlymphoma cell line; clone-13 (Cl-13), an EBV-positive, EBNA2-negative Burkitt lymphoma cellline; Jurkat, a T-cell line]. Pro-tected regions are bracketed, andthe relative positions of the pro-tected regions with respect totranscription initiation from Cpare indicated. NE, no extract.Cp(-) and Cp(+) indicate the non-coding and coding strands,respectively.

addition to some Wp activity (M.W. and S.H.S., unpublisheddata).

If full expression of all the EBNAs requires switching fromWpto Cp usage, then the role ofEBNA 2 and theEBNA 2-dependentenhancer may be during a discrete time period in the initial stagesof infection prior to full transformation of the cells. One possiblepoint ofaction forEBNA 2 may be in the first G1 phase ofthe cellcycle. It has been shown that infection of resting B lymphocyteswith clone-13 virus activates the infected cells to progress intoearly G1 phase ofthe cell cycle (23). The cells arrest at that point,however, and do not complete a full round of the cycle. Thus,during the early stages of G1, EBNA 2 expression may directlyor indirectly up-regulate Cp via the Cp-associated EBNA 2-de-pendent enhancer. This in turn may allow expression ofthe otherEBNAs, some of which may be required to complete the cellcycle. In contrast, in cycling B-cell lines, EBNA 2 is obviouslynot required for the activity ofthis enhancer but rather functionsin cooperation with upstream and downstream elements.

Note Added in Proof. The EBNA 2-dependent enhancer reported herehas also been recently identified by Sung et al. (24).

We thank Dr. K. Takata for providing the Akata cell line and Dr.L. Rymo for providing both the DG75 cell line and the EBNA 2expression vector. We would also like to thank Dr. E. Flemington forhelp with the in vitro transcription analysis. This research wassupported by Grants CA43143 (S.H.S.) and CA7554 (J.L.S.) from theNational Institutes of Health and by Postdoctoral Fellowships toM.W. from the Fonds zur Foerderung der Wissenschaftlichen For-schung, Vienna, and to X.W.J. from the Lady Tata Memorial Trust.

1. Klein, G. (1989) Adv. Viral Oncol. 8.

2. Epstein, M. A. & Achong, B. G. (1986) The Epstein Barr Virus: RecentAdvances (Heinemann, London).

3. Kieff, E. & Liebovitz, D. (1990) in Virology, eds. Fields, B., Knipe, D.,Chanock, R., Hirsh, M., Melnick, J., Monath, T. & Roizman, B. (Raven,New York), pp. 1889-1920.

4. Woisetschlaeger, M., Strominger, J. L. & Speck, S. H. (1989) Proc.Nati. Acad. Sci. USA 86, 6498-6502.

5. Woisetschlaeger, M., Yandava, C. N., Furmanski, L. A., Strominger,J. L. & Speck, S. H. (1990) Proc. Natl. Acad. Sci. USA 87, 1725-1729.

6. Ricksten, A., Olsson, A., Andersson, T. & Rymo, L. (1988) NucleicAcids Res. 16, 8391-8410.

7. Gorman, C. M., Moffat, L. F. & Howard, B. H. (1982) Mol. Cell. Biol.2, 1044-1051.

8. Ricksten, A., Svensson, C., Welinder, C. & Rymo, L. (1987) J. Gen.Virol. 68, 2407-2418.

9. Flemington, E. & Speck, S. H. (1990) J. Virol. 64, 1217-1226.10. Takada, K. & Ono, Y. (1989) J. Virol. 63, 445-449.11. Auffray, C. & Rougeon, T. (1980) Ear. J. Biochem. 107, 303-314.12. Dignam, J. D., Lebovitz, R. M. & Roeder, R. G. (1983) Nucleic Acids

Res. 11, 1475-1489.13. Lee, W., Mitchell, P. & Tjian, R. (1987) Cell 49, 741-752.14. Hai, T., Horikoshi, M., Roeder, R. G. & Green, M. R. (1988) Cell 54,

1043-1051.15. Rooney, C., Howe, J. G., Speck, S. H. & Miller, G. (1989) J. Virol. 63,

1531-1539.16. Rabson, M., Gradoville, L., Heston, L. & Miller, G. (1982) J. Virol. 44,

834-844.17. Jeang, K. & Hayward, D. S. (1983) J. Virol. 48, 135-148.18. Terhorst, C., Parkham, P., Mann, D. L. & Strominger, J. L. (1976) Proc.

Natl. Acad. Sci. USA 73, 910-914.19. Skare, J. & Strominger, J. L. (1980) Proc. Natl. Acad. Sci. USA 77,

3860-3864.20. Ben-Bassat, H., Goldblum, N., Mitrani, S., Goldblum, T., Yoffey, J. M.,

Cohen, M. M., Bentwich, Z., Ramot, B., Klein, E. & Klein, G. (1977)Int. J. Cancer 19, 27-33.

21. Brandsman, J. & Miller, G. (1980) Proc. Natl. Acad. Sci. USA 77,6851-6855.

22. Hammerschmidt, W. & Sugden, B. (1989) Nature (London) 340,393-397.23. Gordon, J., Walker, L., Guy, G., Brown, G., Rowe, M. & Rickinson, A.

(1986) Immunology 58, 591-598.24. Sung, N. S., Kenney, S., Gutsch, D. & Pagano, J. S. (1991) J. Virol. 65,

2164-2169.

Proc. Natl. Acad. Sci. USA 88 (1991)

Ab

0