VIROLOGY 236, 18–29 (1997)ARTICLE NO. VY978739

P32/TAP, a Cellular Protein That Interacts with EBNA-1 of Epstein–Barr Virus

Yilong Wang,* Jon E. Finan,* Jaap M. Middeldorp,† and S. Diane Hayward*,‡,1

*Molecular Virology Laboratories, Department of Pharmacology and Molecular Sciences and ‡Department of Oncology, Johns HopkinsSchool of Medicine, Baltimore, Maryland 21205-2185; and †Organon-Teknika, 5281 RM Boxtel, The Netherlands

Received April 4, 1997; returned to author for revision June 3, 1997; accepted July 17, 1997

The Epstein – Barr virus (EBV) EBNA-1 protein has a central role in the maintenance of a latent EBV infection and is theonly virus-encoded protein expressed in all EBV-associated tumors. EBNA-1 is required for replication of the episomal formof the latent viral genome and transactivates the latency C and LMP-1 promoters. The mechanisms by which EBNA-1performs these functions are not known. Here we describe the cloning, expression, and characterization of a cellular protein,P32/TAP, which strongly interacts with EBNA-1. We show that P32/TAP is expressed at high levels in Raji cells and issynthesized as a proprotein of 282 amino acids (aa) that is posttranslationally processed by a two-step cleavage processto yield a mature protein of 209 aa. It has been previously reported that P32/TAP is expressed on the cell surface. Ourtransient expression assays detected full-length P32/TAP (1–282 aa) in the cytoplasm while mature P32/TAP proteinlocalized to the nucleus. Three lines of evidence support P32/TAP interaction with EBNA-1. First, in the yeast two-hybridsystem we mapped two interactive N-terminal regions of EBNA-1, aa 40–60 and aa 325–376, each of which containsarginine–glycine repeats. These regions interact with the C-terminal half of P32/TAP. Second, the full-length cytoplasmicP32/TAP protein can translocate nuclear EBNA-1 into the cytoplasm. Third, P32/TAP co-immunoprecipitated with EBNA-1.We have confirmed that a Gal4 fusion protein containing the C-terminal region of P32/TAP (aa 244–282) transactivatesexpression from a reporter containing upstream Gal4-binding sites. Deletion of the P32/TAP interactive regions of EBNA-1 severely diminished EBNA-1 transactivation of FRTKCAT in transient expression assays. Our data suggest that interactionwith P32/TAP may contribute to EBNA-1-mediated transactivation. q 1997 Academic Press

INTRODUCTION EBNA-1 is a DNA-binding protein which binds to a clusterof sites in EBV that together constitute the plasmid origin

Epstein–Barr virus (EBV) is a ubiquitous gamma her- of viral replication, oriP (Rawlins et al., 1985; Yates et al.,pes virus that is associated with several human dis- 1984), and to two sites immediately downstream of theeases, including infectious mononucleosis, Burkitt’s latency Qp (Nonkwelo et al., 1996; Rawlins et al., 1985;lymphoma, nasopharyngeal carcinoma, mixed cellularity

Schaefer et al., 1995; Smith and Griffin, 1992). The crystalHodgkins’ disease, and lymphoproliferative disease in

structure of the EBNA-1 DNA-binding domain bound toimmunocompromised hosts such as AIDS and transplant

DNA has been solved (Bochkarev et al., 1996). Purifiedpatients (Miller, 1990; Rickinson and Kieff, 1996). In

EBNA-1 lacks detectable enzymatic functions such as atransgenic mice, expression of EBNA-1 can induce tu-DNA-dependent ATPase or helicase (Frappier andmors (Wilson et al., 1996; Wilson and Levine, 1992). TheO’Donnell, 1991b; Middleton and Sugden, 1992), whichEBNA-1 protein has a centrol role in the maintenance ofare important activities for other viral origin-binding, repli-latent EBV infection, and it is the only virus-encoded pro-cation proteins such as UL9 of herpes simplex (Brucknertein required for replication of the episomal form of theet al., 1991), T antigen of simian virus 40 (Wold et al.,EBV genome (Yates et al., 1985). EBNA-1 is a phospho-1987), and E1 of bovine papillomavirus type 1 (Yang etprotein (Hearing and Levine, 1985) with an unusual struc-al., 1993). The exact mechanism whereby EBNA-1 contri-ture; it is separated into N- and C-terminal domainsbutes to oriP DNA replication is unclear. EBNA-1–EBNA-[amino acids (aa) 1–89 and 327–641] by a glycine – gly-1 interactions promote both intramolecular DNA loopingcine–alanine repeat sequence. The size of EBNA-1 var-(Frappier and O’Donnell, 1991a; Su et al., 1991) and inter-ies considerably in individual EBV-containing cell popula-molecular DNA interactions (Goldsmith et al., 1993;tions (from Ç69 to 94 kDa), depending on the number ofMackey et al., 1995) which may generate a DNA structurethese repeats within the protein (Heller et al., 1982).favorable for replication initiation. EBNA-1 may also beinvolved in the attachment of ori-P to the nuclear matrix

1 To whom correspondence and reprint requests should be (Jankelevich et al., 1992).addressed at Department of Pharmacology and Molecular Sciences, In cotransfection assays, EBNA-1 acts as a transcrip-Johns Hopkins School of Medicine, 725 N. Wolfe Street, Balti-

tional transactivator (Polvino-Bodnar and Schaffer, 1992;more, MD 21205-2185. Fax: (410) 955-8685. E-mail: diane_hayward@qmail.bs.jhu.edu. Sugden and Warren, 1989; Yates and Camiolo, 1988); it

180042-6822/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID VY 8739 / 6a46$$$$61 08-14-97 16:49:42 vira AP: VY

19P32/TAP INTERACTS WITH EBNA 1 PROTEIN

can transactivate heterologous promoters when multiple were cloned from pRA362 (Ambinder et al., 1990) andpWS61 (Shah et al., 1992).EBNA-1-binding sites are present within the target plas-

mid (Reisman and Sugden, 1986; Wysokenski and Yates, Plasmid pYW51, a modified SG5 vector (Stratagene,La Jolla, CA) expressing an amino-terminal Flag epitope1989) and contributes to down-regulation of the native

EBV Qp (Sample et al., 1992; Sung et al., 1994) and to (Eastman Kodak Co., New Haven, CT), was used to makethree Flag-tagged P32/TAP expression plasmids; P32/up-regulation of the latency Cp and LMP-1p (Gahn and

Sugden, 1995; Puglielli et al., 1996; Schlager et al., 1996; TAP(1-282)(pYW59) was constructed by cloning a BglIIfragment from a P32/TAP cDNA clone 64 into the BamHISugden and Warren, 1989). However, EBNA-1 lacks a

specific transcriptional activation domain. Fusion of dif- site of pYW51. Two P32/TAP mutations, P32/TAP(33–282) (pYW76) and P32/TAP(81–282) (pYW84), wereferent parts of EBNA-1 to the GAL4 DNA-binding domain

failed to generate a chimera capable of activating GAL4 cloned by digesting the P32/TAP cDNA with PstI andHindIII, blunt ending with Klenow or T4 DNA polymerase,reporters (M-R. Chen, unpublished results; Horner et al.,

1995). Therefore the mechanism of EBNA-1-mediated adding BamHI linkers, and ligating into pYW51. A full-length P32/TAP BamHI/BglII fragment was also removedtransactivation is also unclear. The inability to detect an

intrinsic transcriptional activation function prompted us from cDNA clone 64 and ligated into the BamHI site ofpSG5(YW164). Using similar PCR methods, we con-to explore the possibility that a cellular protein(s) inter-

acting with EBNA-1 might contribute to its transactivation structed EBNA-1 variants by inserting EBNA-1 frag-ments into another modified SG5 plasmid, pMRC69ability. A yeast two-hybrid screen for interacting proteins

identified P32/TAP, a protein that has been previously (Chen et al., 1994), to generate pMRC72[E1(D102-325)],pYW47[E1(D102-325, D40-60)], pYW48[E1(D102-381)],described by several different investigators and that has

been implicated in splicing (Honore et al., 1993; Krainer and pYW102[E1(376-641)].The target plasmids used for CAT assays, 51Gal4-et al., 1990), transactivation (Yu et al., 1995a,b), and re-

ceptor functions (Eggleton et al., 1994, 1995; Ghebrehi- TKCAT and FRTKCAT, have been described previously(Hsieh and Hayward, 1995; Reisman and Sugden, 1986).wet et al., 1994, 1995). To address the biological rele-

vance of the EBNA-1 interaction we examined the pro- The P32/TAP Gal4 fusion plasmid pYW61 contains a full-length P32/TAP, and pYW62 contains a domain from aacessing and cellular localization of P32/TAP and mapped

the interaction to the two arginine–glycine-rich domains 244 to aa 282, which was generated by PCR amplificationwith primers LGH1637 and LGH2179.within EBNA-1. Deletion of these regions abrogated the

transcriptional activity of EBNA-1, implicating P32/TAPas a possible mediator of EBNA-1-dependent transacti- Yeast two-hybrid screenvation.

Saccharomyces cerevisiae Y190 and Y187 and yeastplasmids pAS1-CYH2 and pACTII were obtained from

MATERIALS AND METHODSStephen Elledge (Durfee et al., 1993). A human lympho-cyte cDNA library in which the cDNAs are joined to thePlasmid constructionsGal4 activation domain (aa 768–881) in pACT was pur-chased from Clontech (Palo Alto, CA). Y190 plasmidsThe plasmids used in the yeast two-hybrid assay were

cloned into the vector pAS1-CHY2 (Durfee et al., 1993). were transformed to Trp prototrophy by the method ofSchiestl and Geitz (1989) with pAS-EBNA-1, which con-The different EBNA-1 segments were generated by PCR

amplification and, after restriction enzyme digestion, the tained EBNA-1 fused to the Gal4 DNA-binding domain(1–147) in pAS1-CYH2. A single colony was grown in SCproducts were fused in-frame to the Gal4 DNA-binding

domain. Expression of the fusion proteins was confirmed Trp0 medium and transformed with library DNA usingsingle-strand DNA as carrier. Aliquots were taken fromby Western blotting with anti-Gal4 antibody (Upstate Bio-

technology). Gal4-E1(D102-325) (pYW18), which contains each transformation mix and plated on SC medium lack-ing tryptophan and leucine to determine the transforma-a full-length EBNA-1 with a deletion from aa 102 to 325,

was PCR-amplified from p367 (Yates and Camiolo, 1988) tion efficiency. The transformation mix was then platedon 15-cm petri dishes containing SC medium lackingwith primers LGH734 and LGH312; pWS41[Gal4-E1

(D102-381)] was amplified from p385(d145) (a gift from tryptophan, leucine, and histidine but including 40 mM3-AT (Sigma) and incubated at 307 for 5–7 days. His/John Yates) with primers LGH312 and LGH734; pYW44

[Gal4-E1(1-89; D40-60)] and pYW95[Gal4-E1(D102-325, colonies were then screened for b-galactosidase activityusing a filter lift assay (Breeden and Nasmyth, 1985).D40-60)] were amplified from p378d1GA (a gift from John

Yates) with primers LGH2147, LGH901, and LGH2148; Colonies were transferred onto nitrocellulose filters, per-meabilized by freezing in liquid nitrogen for 5 sec, andpYW96 [Gal4 - E1(1-89 ) ] , pYW97 [Gal4 - E1( 85-641 ) ], and

pYW98[Gal4-E1(376-641)] were amplified from p367 with thawed at room temperature. Filters were then overlaidon Whatman 3MM papers saturated with Z-buffer/X-galprimers LGH2147, LGH2149, LGH2150, and LGH2148;

pYW11[Gal4-E1(459-641)] and pYW42[Gal4-E1(408-641)] solution (60 mM Na2HPO4r7H2O, 40 mM NaH2PO4rH2O,

AID VY 8739 / 6a46$$$$62 08-14-97 16:49:42 vira AP: VY

20 WANG ET AL.

10 mM KCl, 1 mM MgSO4r7H2O, 35 mM b-mercaptoetha- with PBS. The cells were incubated in cold PBS con-taining 0.2% Triton 100 for 15 min on ice and then rinsednol, 0.4 mg/ml X-gal), and incubated at 307. The time

required for color development was 4–6 hr. The positive twice with PBS. Primary antibodies were diluted in PBSand were typically reacted for 1 hr at 377. To removecDNAs were recovered from yeast using SC medium

selection and transformed into DH5a via electroporation excess antibody, the slides were washed four times inPBS. Fluorochrome-conjugated secondary antibodiesusing a Bio-Rad GenePulser. The positive clones were

sequenced using Sequenase 2.0 (U.S. Biochemicals) ac- were reacted in the same manner. Excess secondaryantibodies were removed by four washes in PBS. Thecording to the manufacturer’s specifications. Sequence

analysis and homology searches were performed using processed slides were then mounted with mounting me-dium (Sigma). The prepared slides were examined byBLAST (NIH) and GeneWorks (IntelliGenetics, Inc.). Pro-

tein interaction was quantified by determining b-galac- confocal optics and an Olympus inverted microscope.tosidase activity: Cultures (2.5 ml) were grown in the

Co-immunoprecipitationappropriate selection medium to an OD600 of 1.0. Cellswere then prepared and permeabilized as described by Cos-1 cells were transfected with 8 mg each of Flag-the library supplier (Clontech). Cell pellets were resus- tagged P32/TAP and the different EBNA-1 constructions.pended and o-nitrophenyl-b-D-galactoside substrate was Forty-eight hours after transfection the cells were lysedadded. The amount of liberated O-nitrophenol was deter- in 500 ml of isotonic lysis buffer (142.5 mM KCl, 5 mMmined at OD420 . MgCl2 , 10 mM HEPES, pH 7.2, 1 mM EGTA, pH 8.0, and

0.2% NP-40). One hundred microliters of the extract wasTransient expression and Western immunoblot

kept for direct analysis of the proteins present. Four hun-analysis

dred microliters of the extract was incubated with anti-Flag rabbit antibody (Santa Cruz, Inc.) overnight at 47Vero, Cos-1, 293, and CV-1 cells were maintained infollowed by the addition of 100 ml of Protein A–Sepha-Dulbecco’s modified Eagle’s medium plus 10% fetal calfrose (Pharmacia, Sweden). After incubation for 3 hr atserum at 377 in 5% CO2 and were transfected using cal-47, these samples, along with the samples for direct anal-cium phosphate-BES [N*N-bis (2-hydroxyethyl)-2-ami-ysis, were subjected to 10% SDS–PAGE and transferrednoethanesulfonic acid]-buffered saline (Chen and Oka-to a nitrocellulose membrane. The individual proteinsyama, 1987). Cells were lysed 60 hr after transfection inwere detected with monoclonal antibody EBNA.OT1x and21 sample buffer [50 mM Tris, 4% sodium dodecyl sulfateECL (Amersham, England). Co-immunoprecipitation of(SDS), 20% glycerol, 0.04% bromophenol blue, 200 mMendogenous P32/TAP was performed similarly usingdithiothreitol] at 2 1 104 cells/ml and were sonicated293 cells transfected with 8 mg of wild-type EBNA-1briefly. The lysate was electrophoresed through a 10%DNA. The cell extract was immunoprecipitated withpolyacrylamide gel. After blocking in Tris-buffered salineEBNA.OT1x monoclonal antibody and after immunoblot-(TBS), 5% nonfat dry milk, 0.1% Tween 20 for 1 hr andting, proteins were detected with anti-P32 antibody (Ghe-washing with TBS, 0.1% Tween 20, the filter was thenbrehiwet et al., 1994).incubated with Flag M2 monoclonal antibody (Kodak

Inc.), anti-P32/TAP antibody (gift of B. Ghebrehiwet (Ghe-CAT assaybrehiwet et al., 1994)), or EBNA-1 monoclonal antibody

EBNA.OT1x (Chen et al., 1993) in TBS, 0.1% Tween 20 at CV-1 cells were plated in six-well cluster dishes at 2room temperature for 1 hr. The filter was washed three 1 105 cells per well the day before transfection. Cellstimes for 15 min in TBS–Tween and then incubated in a were harvested 40 hr after transfection and assays for1:5000 dilution of alkaline phosphatase-conjugated goat CAT were performed as previously described (Ling et al.,anti-mouse antibody or anti-rabbit antibody. The filter 1993).was washed three times and developed with an ECL kit(Amersham, England) followed by exposure to Kodak X- RESULTSray film.

EBNA-1 interacts with P32/TAP in a yeast two-hybridImmunofluorescence assay assay

Evidence that EBNA-1 interacts with P32/TAP emergedVero cells were plated at approximately 0.8 1 105 cellsper well onto glass tissue culture chamber slides (Nunc). from analyses of a yeast two-hybrid screen. Yeast ex-

pressing a Gal4 DNA-binding domain fusion of the full-DNA transfections were performed as described above,with a total of 2 mg of DNA transfected per well. At 24 hr length EBNA-1 in which the glycine–alanine repeat

region was deleted were transformed with a human lym-posttransfection, cells were washed twice in phosphate-buffered saline (PBS) and then incubated with PBS con- phocyte cDNA library. Approximately 1.5 million trans-

formants were placed under selection. Of the 63 cDNAtaining 3% paraformaldehyde (pH 8.0) for 10 min at roomtemperature. The fixative was removed by two washes plasmids whose products interacted specifically with

AID VY 8739 / 6a46$$$$62 08-14-97 16:49:42 vira AP: VY

21P32/TAP INTERACTS WITH EBNA 1 PROTEIN

TABLE 1 between amino acids 325 and 376 of EBNA-1. Each ofthese regions contains arrays of arginine and glycineStructure of P32/TAP cDNA Clones Selected in a Yeastresidues.Two-Hybrid Screen for Interaction with EBNA-1a

P32/TAP EBNA-1 EBNA-1 P32/TAP (aa 1–282) is cleaved twice within the N-(aa)b (1–641D102 – 325) (459–641) SNF-1 P53 terminus to form the mature protein

1–282 / 0 0 0 Krainer et al. (1991) viewed P32/TAP as a full-length19–282 / 0 0 0 protein of 209 amino acids with a CTG (Leu) initiator22–282 / 0 0 0

codon. However, Honore et al. (1993) observed that P32/25–282 / 0 0 0TAP is produced as a proprotein of 282 amino acids with35–282 / 0 0 0a conventional ATG start codon. These authors further43–282 / 0 0 0

46–282 / 0 0 0 demonstrated that the 209-amino-acid protein is the ma-53–282 / 0 0 0 ture form that is produced by posttranslational pro-55–282 / 0 0 0

cessing of the 282-amino-acid proprotein.60–282 / 0 0 0To examine P32/TAP processing, a series of P32/TAP64–282 / 0 0 0

variants tagged with the Flag epitope at the N-terminus71–282 / 0 0 0103–282 / 0 0 0 was constructed in a pSG5 background. Extracts of trans-

fected Cos-1 cells were analyzed by Western blot witha The specificity of interaction was tested by cotransforming yeast

anti-Flag monoclonal antibody and anti-P32/TAP mono-Y190 with these P32 cDNAs and EBNA-1 (1–641D102–325), EBNA-1clonal antibody (a gift from Berhane Ghebrehiwet, SUNY).(459–641), pAS/SNF1, or pAS/P53. b-Galactosidase activity was mea-As shown in the blot probed with anti-P32/TAP antibodysured by a filter lift assay, which resulted in blue color (/), indicating

interaction or white color (0), indicating no interaction. in Fig. 3A, the P32/TAP (aa 1–282) construct expressedb The 25 interactive clones identified represented 13 separate mainly the full-length P32/TAP which is about 39 kDa, as

cDNAs.well as two smaller forms of 35 and 32 kDa. P32/TAP(aa 33–282) expressed two polypeptides, the smaller ofwhich was the same size as the 32-kDa product gener-EBNA-1, 31 were partially sequenced and 25 of theseated from P32/TAP (aa 1–282) and the same size aswere found to match the published sequence of P32/endogenous P32/TAP from Raji cells. These smaller poly-TAP (Table 1). The different 5* termini of the P32/TAPpeptides were not observed in the blot probed with anti-cDNA clones indicate that the EBNA-1 interaction do-Flag antibody (shown in Fig. 3B), indicating that themain is located after amino acid 103 of P32/TAP. P32smaller polypeptides were formed by removal of the N-was initially identified as a protein in a HeLa cell extractterminus of P32/TAP. P32/TAP (aa 81–282) expressedthat copurified with the SF2/ASF splicing factor (Honoreonly one band which was similar in size to endogenouset al., 1993; Krainer et al., 1990, 1991) while TAP wasP32/TAP from Raji cells. These observations are inidentified by affinity chromatography as a protein inter-agreement with those of Honore et al. (1993) in sug-acting with the HIV Tat transactivator and cloned from agesting that the mature form of the P32/TAP protein isMolt 3 T-cell library (Yu et al., 1995b). P32/TAP is highlyproduced by posttranslational processing of the 282-conserved and the mouse homolog YL2 (Luo et al., 1994)amino-acid proprotein. Furthermore, we demonstratehas 92% identity with human P32/TAP (Fig. 1). P32/TAPthat there are two separate cleavage sites in the N-termi-transcripts are expressed in most human tissues (Yu etnal region of P32/TAP located between aa 1 and 33 andal., 1995b) and in a variety of different cell lines (Deb andbetween aa 33 and 71. The band detected in the vectorDatta, 1996; Luo et al., 1994; Yu et al., 1995b).lane may represent a more rapidly migrating form of theendogenous P32/TAP in Cos-1 cells. P32/TAP expres-Mapping of P32/TAP interaction domains in EBNA-1sion has been described in both lymphoid and non-

The unexpectedly large number of P32/TAP cloneslymphoid cell lines. In agreement with these reports, we

whose products interacted with EBNA-1 prompted furtherdetected P32/TAP by immunoblot analysis in a variety of

investigation into the nature of the association betweenB-cell lines including CA46, DG75, Akata, IB4, P3HR-1,

P32/TAP and EBNA-1. To map the EBNA-1 domain inter-and B95-8 (data not shown).

acting with P32/TAP, a series of EBNA-1 deletions wasmade. Each EBNA-1 variant was cloned as a GAL4 DNA- Intact and processed forms of P32/TAP differ in theirbinding domain fusion and transformed into yeast strain intracellular localizationY190 containing the plasmid expressing P32/TAP as aGal4 activation domain fusion. The EBNA-1–P32/TAP in- P32/TAP has been reported to be a complement C1q

subunit-binding protein that is expressed on the plasmateraction was quantitated using a lacZ assay (Fig. 2).The results revealed two separate P32/TAP interaction membrane (Ghebrehiwet et al., 1994), a protein that co-

purifies with the SF2 splicing factor (Honore et al., 1993;domains located between amino acids 40 and 60 and

AID VY 8739 / 6a46$$$$63 08-14-97 16:49:42 vira AP: VY

22 WANG ET AL.

FIG. 1. Amino acid sequence comparison of P32 homologs. YL2 is a mouse isolate and is 208 aa in length (Luo et al., 1994) and P32/gC1qR/TAP proteins are human isolates (Ghebrehiwet et al., 1994; Honore et al., 1993; Krainer et al., 1991; Yu et al., 1995b) and are 282 (P32/C1qR) and279 (TAP) amino acids long. Identical amino acids are boxed.

FIG. 2. Mapping P32/TAP interactive domains in EBNA-1. Diagrammatic representation of the different Gal4–EBNA-1 fusion constructions tested.Yeast Y190 plasmids were cotransformed with each of the Gal4–EBNA-1 constructions and Gal4-P32/TAP [Gal4 activation domain II (aa 768–881)–P32/TAP (aa 1–282)]. The black boxes indicate interaction domains for P32/TAP. The results of liquid culture assays to quantitate b-galactosidaseactivity are summarized on the right.

AID VY 8739 / 6a46$$8739 08-14-97 16:49:42 vira AP: VY

23P32/TAP INTERACTS WITH EBNA 1 PROTEIN

with transfected full-length EBNA-1, as demonstrated bythe immunoprecipitation assay shown in Fig. 5B.

P32/TAP (aa 1–282) relocates EBNA-1 into thecytoplasm

Additional evidence for a physical interaction betweenP32/TAP and EBNA-1 came from a visualization of theintracellular distribution of EBNA-1 in P32/TAP cotrans-fected cells. Anti-EBNA-1 monoclonal antibody was usedto detect EBNA-1, and anti-Flag rabbit antibody was usedto detect P32/TAP (Fig. 6). Figure 6D shows single stain-ing of individually transfected Vero cells expressingFIG. 3. Expression of endogenous and Flag-tagged transfected P32/EBNA-1, mature P32/TAP (aa 81–282), and full-lengthTAP. Cos-1 cells were transfected with the indicated Flag-tagged P32/

TAP constructions or the vector pYW51 (a pSG5-based vector express- P32/TAP (aa 1–282). Figures 6A and 6B show that P32/ing the Flag epitope). The positive control was Raji cell extract. Cell TAP and EBNA-1 cotransfected Vero cells stained withextracts were electrophoresed through SDS–polyacrylamide gels and anti-P32/TAP antibody (red) and anti-EBNA-1 antibodytransferred to nitrocellulose filters, and protein expression was visual-

(green). Diffuse nuclear P32/TAP (aa 81–282) and EBNA-ized using anti-P32/TAP monoclonal antibody or Flag M2 monoclonal1 colocalized in the nucleus. The cytoplasmic P32/TAPantibody and the ECL detection system. The positions of molecular

weight size markers are indicated. (aa 1–282) translocated most of the nuclear EBNA-1 pro-tein to the cytoplasm and perfectly colocalized withEBNA-1 in spots surrounding the nucleus. P32/TAP does

Krainer et al., 1991), an HIV Rev interactive protein (Luo not interact indiscriminantly with nuclear proteins sinceet al., 1994), and an HIV TAT interactive protein (Yu et P32/TAP (aa 1–282) did not relocate the human cytomeg-al., 1995b). In order to investigate the intracellular distri- alovirus nuclear regulatory protein, IE1 (Fig. 6C).bution of P32/TAP, we transfected Vero cells with Flag-

The P32/TAP interaction domains of EBNA-1 appeartagged P32/TAP (aa 1–282) and two Flag-tagged deletionto be critical for EBNA-1-dependent transactivationvariants expressing aa 33–282 and aa 81–282. The

transfected cells were immunostained with both anti- EBNA-1 can activate transcription from heterologousFlag and anti-P32/TAP antibodies. Surprisingly, three dif- promoters when at least seven EBNA-1-binding sitesferent staining patterns were observed (Fig. 4). P32/TAP(aa 1–282) remained in the cytoplasm and formed con-densed globules that were concentrated around the nu-cleus. P32/TAP (aa 33–282) showed diffuse staining overboth the cytoplasm and the nucleus. P32/TAP (aa 81 –282) localized predominantly to the nucleus and had adiffuse staining pattern. In order to eliminate the possibil-ity that N-terminal tagging might affect processing of theprotein, we cloned the full-length P32/TAP in pSG5 with-out the Flag epitope and found similar results when thecells were stained with anti-P32/TAP antibody (data notshown). It has been reported that the mature form of P32/TAP comprises aa 71–282 (Honore et al., 1993). P32/TAP (aa 81–282) is only 10 aa smaller than this maturepolypeptide and gives predominantly nuclear staining,thus strongly suggesting that mature P32/TAP is a nu-clear form.

P32/TAP co-immunoprecipitates with EBNA-1

P32/TAP was isolated because of its strong bindingto EBNA-1 in a yeast two-hybrid assay. To determine

FIG. 4. Cellular localization of P32/TAP. Immunofluorescence photo-whether P32/TAP can interact with EBNA-1 in mamma-micrographs of Vero cells transfected with Flag-tagged P32/TAP ex-lian cells, co-immunoprecipitation assays were per-pression constructions. Transfected Vero cells were fixed and incu-formed (Fig. 5). Both cytoplasmic P32/TAP (aa 1 – 282)bated with Flag M2 monoclonal antibody followed by FITC-conjugated

and nuclear P32/TAP (aa 81–282) co-immunoprecipi- goat anti-mouse immunoglobulin G antiserum. (A) Flag-tagged P32/tated with full-length EBNA-1 in cotransfected cells (Fig. TAP (aa 1–283). (B) Flag-tagged P32/TAP (aa 33–283). (C) Flag-tagged

P32/TAP (aa 81–283).5A, lanes 2 and 3). Endogenous P32/TAP also interacted

AID VY 8739 / 6a46$$$$63 08-14-97 16:49:42 vira AP: VY

24 WANG ET AL.

complete amino-terminus containing both domains abro-gated the ability of EBNA-1 to activate FRTKCAT.

DISCUSSION

EBNA-1 acts as the origin-binding protein for oriP repli-cation during latency but lacks detectable enzymatic ac-tivities such as DNA-dependent ATPase or helicase,which are intrinsic to other viral origin-binding replicationproteins such as E1 of papillomavirus, UL9 of herpessimplex virus, or T antigen of simian virus 40. Further-more, EBNA-1 acts as a transactivator of viral latencypromoters but lacks a detectable activation domain. DNA

FIG. 5. Co-immunoprecipitation of EBNA-1 and P32/TAP. (A) Cos-1 binding is essential for both replication and transactiva-cells were transfected with 8 mg of each of the Flag-tagged P32/TAP tion functions (Ambinder et al., 1991; Kirchmaier and Sug-and EBNA-1 constructions. 48 hr after transfection the cells were lysed den, 1997; Polvino-Bodnar and Schaffer, 1992). DNA bind-and the extract was incubated with anti-Flag M2 monoclonal antibody

ing by EBNA-1 produces significant structural alterationsfollowed by the addition of 100 ml of Protein A–Sepharose. The immu-in the DNA (Frappier and O’Donnell, 1991a, 1992; Gold-noprecipitates were subjected to 10% SDS–PAGE and transferred to

a nitrocellulose membrane. The individual proteins were detected with smith et al., 1993; Hearing et al., 1992; Hsieh et al., 1993;the monoclonal antibody EBNA.OT1x and ECL. Lane 1, EBNA-1; lane Mackey et al., 1995; Su et al., 1991) which are anticipated2, EBNA-1 plus P32/TAP (aa 1 – 282); lane 3, EBNA-1 plus P32/TAP (aa to contribute to both replication and transactivation activi-81–282); lane 4, P32/TAP (aa 1–282); lane 5, P32/TAP (aa 81–282).

ties and EBNA-1 also binds to RNA (Snudden et al., 1994).(B) Co-immunoprecipitation of endogenous P32/TAP from 293 cellsHowever, it seems likely that there are additional contri-transfected with 8 mg of wild-type EBNA-1.

butions made by EBNA-1 and the most likely scenariois that these activities would be mediated by EBNA-1interacting proteins. The cellular protein P32/TAP was(Wysokenski and Yates, 1989) are added and can acti-

vate transcription from the latency C and LMP-1 promot- isolated in a two-hybrid screen for such interacting part-ners.ers which are positioned approximately 10 kb from the

oriP EBNA-1-binding sites in the EBV genome. However, P32/TAP is a previously described protein whose pub-lished properties are somewhat difficult to reconcile.EBNA-1 lacks a detectable transactivation domain. It has

been reported that P32/TAP contains a transcriptional P32/TAP was originally isolated as a component of abiochemically purified splicing complex that containedactivation domain that interacts with the general tran-

scription factor TFIIB (Yu et al., 1995a). To determine p32 and p33 polypeptides (Krainer et al., 1991). Subse-quent work showed that the p33 protein SF2/ATF had allwhether P32/TAP contributes to EBNA-1 transactivation,

we first confirmed that P32/TAP did indeed contain an the properties of a functional splicing factor and P32/TAP was not required for splicing activity (Mayeda et al.,activator domain. In agreement with the published report

we found that a Gal4 fusion with intact P32/TAP (aa 1 – 1992), leaving the relevance of the original associationan open question. Other associations between P32/TAP282) showed no appreciable activation. However, an im-

munofluorescence assay showed that this fusion protein and splicing have been made but in each case theseare indirect. P32/TAP was detected in a yeast two-hybridwas retained in the cytoplasm despite the presence of

the nuclear localization signal within Gal4 (1–147). On screen for proteins that interacted with the HIV-1 Revprotein (Luo et al., 1994). Rev binds to the Rev responsethe other hand, a Gal4 fusion that localized to the nucleus

and contained P32/TAP (aa 244–282) gave 40- to 50- element in HIV RNAs and is believed to influence RNAsplicing. In the presence of Rev, singly spliced and un-fold activation of a cotransfected reporter containing five

copies of the Gal4-binding site upstream of the promoter spliced genomic viral RNAs are efficiently transported tothe cytoplasm. In transient expression assays, cotrans-(Fig. 7). We then generated three EBNA-1 variants, each

of which had the gly-gly-ala repeat region (aa 102 – 325) fection of P32/TAP with Rev led to increased activity ofa CAT reporter in which the CAT gene was present withindeleted in addition to specific deletions of: (i) the first

EBNA-1 P32/TAP interaction domain (aa 40–60); (ii) the an intron and could be expressed only from an unsplicedtranscript. However, these results would potentially alsosecond P32/TAP interaction domain (aa 325–381); and

(iii) the entire amino-terminus, including both P32/TAP be compatible with transcriptional activation by Rev-teth-ered P32/TAP. Also detected in a yeast two-hybrid screeninteraction domains (Fig. 8). In cotransfection assays,

parental EBNA-1(D102-325) transactivated the CAT ex- was an association with another viral protein, the productof ORF P, an HSV ORF transcribed in latent infectionpression from the FRTKCAT reporter containing the

EBNA-1-binding sites from the ori-P family of repeats (Bruni and Roizman, 1996). In IF assays ORF P colocal-ized with a splicing factor, SC35, but the concurrent pres-approximately 10- to 30-fold (Fig. 8). Deletion of either

of the P32/TAP interaction domains or deletion of the ence of P32/TAP was not examined.

AID VY 8739 / 6a46$$$$63 08-14-97 16:49:42 vira AP: VY

25P32/TAP INTERACTS WITH EBNA 1 PROTEIN

FIG. 6. Confocal photomicrographs showing EBNA-1 colocalization with P32/TAP in transfected cells. (A) Vero cells cotransfected with EBNA-1(Daa 102–325) and Flag-tagged P32/TAP (aa 81–282). EBNA-1 was detected by EBNA.OT1x mAb and FITC-conjugated sheep anti-mouse antibody(green fluorescence on the left). Flag-tagged P32/TAP was detected using anti-Flag rabbit antibody followed by Texas red-conjugated horse anti-rabbit second antibody (red fluorescence in the middle). The yellow staining in the confocal overlay on the right indicates colocalization of the twoproteins. (B) Vero cells cotransfected with EBNA-1 (Daa 102–325) and Flag-tagged P32/TAP (aa 1–282). EBNA-1 (green fluorescence on the left)and Flag-tagged P32/TAP (red fluorescence in the middle) colocalize (yellow) in the cytoplasm in the confocal overlay on the right. (C) Vero cellscotransfected with HCMV IE1 and Flag-tagged P32/TAP (aa 1–282). IE1 was detected by anti-IE1 mAb and FITC-conjugated sheep anti-mouseantibody (green fluorescence on the left). Flag-tagged P32/TAP is shown in the middle (red fluorescence). The confocal overlay on the right showsno evidence for colocalization. (D) Vero cells transfected with (left) EBNA-1 (Daa 102–325); (middle) Flag-tagged P32/TAP (aa 81–282), and (right)Flag-tagged P32/TAP (aa 1 – 282).

AID VY 8739 / 6a46$$873908-14-97 16:49:42 vira AP: VY

26 WANG ET AL.

P32/TAP had been reported to be processed from thefull-length 282-amino-acid protein to a mature form thatlacks the amino-terminal 73 amino acids. We demon-strated that this processing results from two independentcleavage events and used immunofluorescence assaysto examine the intracellular localization of the differentforms of P32/TAP. Interestingly, while the mature form ofthe protein localized to the nucleus, the full-length proteinappeared in the cytoplasm. Thus it is plausible that P32/TAP could function as a plasma membrane receptor and,after processing, relocate to the nucleus to serve a sec-ond gene regulatory function. There is precedent for sucha mechanism. The Notch receptor is a large transmem-brane protein that participates in intercellular signalingevents that determine cell fate decisions (Artavanis-Tsa-konas et al., 1995). The cleaved intracellular domain ofNotch, NotchIC, localizes to the nucleus and activates

FIG. 7. P32/TAP contains a transcriptional activation domain. Tran- expression of genes containing CBF1-binding sites insient expression assay in Vero cells which received 1 mg of the 51 their regulatory sequences by binding to the repressionGal4TKCAT reporter plasmid and 2 mg of the indicated Gal4-P32/TAP

domain of DNA-bound CBF1 (Hsieh et al., 1996, 1997;effector DNAs. The results shown are an average of two experimentsJarriault et al., 1995). The EBV EBNA-2 protein also inter-with the standard deviation indicated.acts with CBF1 and functions, in part, by mimickingNotchIC (Hsieh and Hayward, 1995). Another example isthe sterol regulatory element-binding proteins (SREBPs),

P32/TAP has been identified as a transcriptional acti- which attach to the endoplasmic reticulum. This tran-vator (Yu et al., 1995b). P32/TAP was isolated by affinity scription factor is released from the membrane in re-chromatography as a protein interacting with HIV-1 Tat. sponse to sterols by two sequential cleavages and theGal4-P32/TAP fusion constructions were created and the liberated domain translocates to the nucleus and acti-presence of a transcriptional activation domain was dem- vates genes controlling cholesterol synthesis (Sakai etonstrated. This domain was negatively charged and mu- al., 1996).tation of two hydrophobic residues (leucines) within the While a case can be made for a gene regulation func-domain abolished the activation function, indicating that tion for P32/TAP, the question remains as to the biologi-the activation observed was not merely a nonspecificcharge effect. The potential for P32/TAP to function asan activator was strengthened by the demonstration thatP32/TAP interacted in vitro with the basal transcriptionfactor TFIIB (Yu et al., 1995a).

While connections between P32/TAP and either splic-ing or transactivation would be consistent with a generegulation function in the nucleus, the identification ofP32/TAP as a plasma membrane receptor (Eggleton etal., 1995; Ghebrehiwet et al., 1994; Peerschke et al., 1994)initially appears contradictory. Ghebrehiwet et al. (1994)isolated from Raji cells a protein, gC1qR, that binds tothe globular heads of gC1q molecules and inhibits com-plement-mediated lysis of sheep red blood cells. Theyused the amino acid sequence of the purified protein todesign PCR primers and cloned the P32/TAP cDNA froma B-cell library. The recombinant protein had the proper-ties of the purified protein in terms of size, sequence,and ability to inhibit lysis of sheep red blood cells. P32/

FIG. 8. Deletion of the P32/TAP interaction domains abolishes EBNA-TAP does not contain a recognizable transmembrane1-mediated transactivation. Transient expression assay in CV-1 cellsdomain but it has also been found to bind to hyaluronicwhich were transfected with the reporter FRTKCAT (2 mg) and the

acid (Deb and Datta, 1996), a high-molecular-weight non- EBNA-1 constructions diagrammed in the lower panel. Deletion of ei-sulfated glycosaminoglycan, and this may be relevant to ther of the P32/TAP interaction domains (aa 40–60 or aa 325–381) is

associated with loss of transactivation.its apparent surface expression.

AID VY 8739 / 6a46$$$$63 08-14-97 16:49:42 vira AP: VY

27P32/TAP INTERACTS WITH EBNA 1 PROTEIN

Bruckner, R. C., Crute, J. J., Dodson, M. S., and Lehman, I. R. (1991). Thecal relevance of the interaction between P32/TAP andherpes simplex virus 1 origin binding protein: A DNA helicase. J.EBNA-1. Relevance can usually be determined by corre-Biol. Chem. 266, 2669–2674.

lating loss of interaction with loss of function but these Bruni, R., and Roizman, B. (1996). Open reading frame P—A herpesexperiments are not completely decisive in the case of simplex virus gene repressed during productive infection encodes

a protein that binds a splicing factor and reduces synthesis of viralEBNA-1. The interaction mapped to the two arginine –proteins made from spliced mRNA. Proc. Natl. Acad. Sci. USA 93,glycine-regions of EBNA-1 and deletion of either of these10423–10427.regions severely impaired EBNA-1 transactivation, a re-

Chen, C., and Okayama, H. (1987). High-efficiency transformation ofsult which is consistent with previous reports (Puglielli mammalian cells by plasmid DNA. Mol. Cell. Biol. 7, 2745–2752.et al., 1996; Yates and Camiolo, 1988). The aa 325 – 376 Chen, M.-R., Middeldorp, J. M., and Hayward, S. D. (1993). Separation

of the complex DNA binding domain of EBNA-1 into DNA recognitionarginine–glycine domain is required for RNA binding byand dimerization subdomains of novel structure. J. Virol. 67, 4875–EBNA-1 (Snudden et al., 1994) for both intrastrand DNA4885.looping and interstrand DNA linking as well as for EBNA-

Chen, M.-R., Zong, J.-C., and Hayward, S. D. (1994). Delineation of a1 oligomerization (Goldsmith et al., 1993; Laine and Frap- sixteen amino acid sequence that forms a core DNA recognitionpier, 1995; Mackey et al., 1995), and hence loss of trans- motif in the Epstein–Barr virus EBNA-1 protein. Virology 205, 486–

495.activation function upon loss of this domain may be oc-Deb, T. B., and Datta, K. (1996). Molecular cloning of human fibroblastcurring for a variety of reasons. The aa 40–60 region of

hyaluronic acid-binding protein confirms its identity with P-32, a pro-EBNA-1 has been associated with DNA linking but nottein co-purified with splicing factor SF2. J. Biol. Chem. 271, 2206–

with any other function, and P32/TAP binding to this re- 2212.gion is of interest. The loss of transactivation that occurs Durfee, T., Becherer, K., Chen, P.-L., Yeh, S.-H., Yang, Y., Kilburn, A. E.,

Lee, W.-H., and Elledge, S. J. (1993). The retinoblastoma protein asso-on removal of this region is provocative but cannot un-ciates with the protein phosphatase type 1 catalytic subunit. Genesequivocally be linked to loss of P32/TAP binding. TheDev. 7, 555–569.two arginine–glycine domains of EBNA-1 interacted in-

Eggleton, P., Ghebrehiwet, B., Coburn, J. P., Sastry, K. N., Zaner, K. S.,dependently with P32/TAP in the yeast two-hybrid assay and Tauber, A. I. (1994). Characterization of the human neutrophiland the EBNA-1 deleted for the aa 40–60 region retains C1q receptor and functional effects of free ligand on activated neutro-

phils. Blood 84(5), 1640–1649.the second aa 325–376 region. However, it is possibleEggleton, P., Ghebrehiwet, B., Sastry, K. N., Coburn, J. P., Zaner, K. S.,that while each EBNA-1 domain is sufficient for protein –

Reid, K. B., and Tauber, A. I. (1995). Identification of a gC1q-bindingprotein tethering to P32/TAP in the yeast assay system,protein (gC1q-R) on the surface of human neutrophils. Subcellular

the presence of both regions is necessary for a functional localization and binding properties in comparison with the cC1q-R.interaction that is capable of mediating transactivation J. Clin. Invest. 95(4), 1569–1578.

Frappier, L., and O’Donnell, M. (1991a). Epstein–Barr virus nuclearas would be required in the cotransfection CAT assays.antigen 1 mediates a DNA loop within the latent replication originP32/TAP is a protein whose processing and intracellularof Epstein–Barr virus. Proc. Natl. Acad. Sci. USA 88, 10875–10879.localization are suggestive of a regulatory function and

Frappier, L., and O’Donnell, M. (1991b). Overproduction, purification,the possibility that it may be contributing to EBNA-1- and characterization of EBNA-1, the origin binding protein of EBV. J.mediated transactivation is intruiging. Biol. Chem. 266, 7819–7826.

Frappier, L., and O’Donnell, M. (1992). EBNA-1 distorts oriP, the Ep-stein–Barr virus latent replication origin. J. Virol. 66, 1786–1790.ACKNOWLEDGMENTS

Gahn, T. A., and Sugden, B. (1995). An EBNA-1-dependent enhancerWe thank Michael Delannoy for assistance with the confocal micros- acts from a distance of 10 kilobase pairs to increase expression of

copy. Berhane Ghebrehiwet for anti-P32/TAP antiserum, and Feng the Epstein–Barr virus LMP gene. J. Virol. 69, 2633–2636.Chang for manuscript preparation. This work was supported by the Ghebrehiwet, B., Kew, R. R., Gruber, B. L., Marchese, M. J., Peerschke,National Institutes of Health (RO1 CA30356). S.D.H. is an American E. I., and Reid, K. B. (1995). Murine mast cells express two types ofCancer Society awardee (FRA429). C1q receptors that are involved in the induction of chemotaxis and

chemokinesis. J. Immunol. 155(5), 2614–2619.Ghebrehiwet, B., Lim, B. L., Peerschke, E. I., Willis, A. C., and Reid, K. B.REFERENCES

(1994). Isolation, cDNA cloning, and overexpression of a 33-kD cellsurface glycoprotein that binds to the globular ‘‘heads’’ of C1q. J. Exp.Ambinder, R. F., Mullen, M., Chang, Y.-N., Hayward, G. S., and Hayward,Med. 179(6), 1809–1821.S. D. (1991). Functional domains of Epstein–Barr virus nuclear anti-

Goldsmith, K., Bendell, L., and Frappier, L. (1993). Identification ofgen EBNA-1. J. Virol. 65, 1466–1478.EBNA1 amino acid sequences required for the interaction of theAmbinder, R. F., Shah, W. A., Rawlins, D. R., Hayward, G. S., and Hay-functional elements of the Epstein–Barr virus latent origin of DNAward, S. D. (1990). Definition of the sequence requirements for bind-replication. J. Virol. 67, 3418–3426.ing of the EBNA-1 protein to its palindromic target sites in Epstein–

Hearing, J., Mulhaupt, Y., and Harper, S. (1992). Interaction of Epstein–Barr virus DNA. J. Virol. 64, 2369–2379.Barr virus nuclear antigen 1 with the viral latent origin of replication.Artavanis-Tsakonas, S., Matsuno, K., and Fortini, M. E. (1995). NotchJ. Virol. 66, 694–705.signaling. Science 268, 225 – 232.

Hearing, J. C., and Levine, A. J. (1985). The Epstein–Barr virus nuclearBochkarev, A., Barwell, J. A., Pfuetzner, R. A., Bochkareva, E., and Frap-antigen (BamHI K antigen) is a single-stranded DNA binding phos-pier, L. (1996). Crystal structure of the DNA-binding domain of thephoprotein. Virology 145, 105–116.Epstein–Barr virus origin-binding protein, EBNA1, bound to DNA.

Heller, M., van Santen, V., and Kieff, E. (1982). Simple repeat sequenceCell 84, 791–800.in Epstein–Barr virus DNA is transcribed in latent and productiveBreeden, L., and Nasmyth, K. (1985). Regulation of the yeast HO gene.

Cold Spring Harbor Symp. Quant. Biol. 50, 643–650. infections. J. Virol. 44, 311–320.

AID VY 8739 / 6a46$$$$64 08-14-97 16:49:42 vira AP: VY

28 WANG ET AL.

Honore, B., Madsen, P., Rasmussen, H. H., Vandekerckhove, J., Celis, Polvino-Bodnar, M., and Schaffer, P. A. (1992). DNA binding activity isrequired for EBNA 1-dependent transcriptional activation and DNAJ. E., and Leffers, H. (1993). Cloning and expression of a cDNA cov-

ering the complete coding region of the P32 subunit of human pre- replication. Virology 187, 591–603.mRNA splicing factor SF2. Gene 134(2), 283–287. Puglielli, M. T., Woisetschlaeger, M., and Speck, S. H. (1996). oriP is

Horner, D., Lewis, M., and Farrell, P. J. (1995). Novel hypotheses for essential for EBNA gene promoter activity in Epstein–Barr virus-the roles of EBNA-1 and BHRF1 in EBV-related cancers. Intervirology immortalized lymphoblastoid cell lines. J. Virol. 70, 5758–5768.38, 195–205. Rawlins, D. R., Milman, G., Hayward, S. D., and Hayward, G. S. (1985).

Hsieh, D.-J., Camiolo, S. M., and Yates, J. L. (1993). Constitutive binding Sequence specific DNA-binding of EBNA-1 to clustered sites in theof EBNA-1 protein to the EBV replication origin, oriP, with distortion plasmid maintenance region. Cell 42, 859–868.of DNA structure during latent infection. EMBO J. 12, 4933–4944. Reisman, D., and Sugden, B. (1986). Transactivation of an Epstein–

Hsieh, J. J.-D., and Hayward, S. D. (1995). Masking of the CBF1/RBPJk Barr viral transcriptional enhancer by the Epstein–Barr viral nucleartranscriptional repression domain by Epstein–Barr virus EBNA2. Sci- antigen 1. Mol. Cell. Biol. 6, 3838–3846.ence 268, 560–563. Rickinson, A. B., and Kieff, E. (1996). Epstein–Barr virus. In ‘‘Field’s

Hsieh, J. J.-D., Henkel, T., Salmon, P., Robey, E., Peterson, M. G., and Virology’’ (B. N. Field, D. M. Knipe, and P. M. Howley, Eds.), 3rd ed.,Hayward, S. D. (1996). Truncated mammalian Notch1 activates CBF1/ Vol. 2, pp. 2397–2446. Raven Press, New York.RBPJk-repressed genes by a mechanism resembling that of Epstein– Sakai, J., Duncan, E. A., Rawson, R. B., Hua, X., Brown, M. S., andBarr virus EBNA2. Mol. Cell. Biol. 16, 952–959. Goldstein, J. L. (1996). Sterol-regulated release of SREBP-2 from cell

Hsieh, J. J.-D., Nofziger, D. E., Weinmaster, G., and Hayward, S. D. membranes requires two sequential cleavages, one within a trans-(1997). EBV immortalization: Notch2 interacts with CBF1. J. Virol. 71, membrane segment. Cell 85, 1037–1046.1938–1945. Sample, J., Henson, E. B. D., and Sample, C. (1992). The Epstein–Barr

Jankelevich, S., Kolman, J. L., Bodnar, J. W., and Miller, G. (1992). A virus nuclear protein 1 promoter active in type 1 latency is autoregu-nuclear matrix attachment region organizes the Epstein–Barr viral lated. J. Virol. 66, 4654–4661.plasmid in Raji cells into a single DNA domain. EMBO J. 11, 1165– Schaefer, B. C., Strominger, J. L., and Speck, S. H. (1995). Redefining1176. the Epstein–Barr virus-encoded nuclear antigen EBNA-1 gene pro-

Jarriault, S., Brou, C., Logeat, F., Schroeter, E. H., Kopan, R., and Israel, moter and transcription initiation site in group I Burkitt lymphomaA. (1995). Signalling downstream of activated mammalian Notch. cell lines. Proc. Natl. Acad. Sci. USA 92, 10565–10569.Nature 377, 355–358. Schiestl, R. H., and Gietz, R. D. (1989). High efficiency transformation

Kirchmaier, A. L., and Sugden, B. (1997). Dominant-negative inhibitors of intact yeast cells using single stranded nucleic acids as a carrier.of EBNA-1 of Epstein – Barr virus. J. Virol. 71, 1766–1775. Curr. Genet. 16, 339–346.

Krainer, A. R., Conway, G. C., and Kozak, D. (1990). Purification and Schlager, S., Speck, S. H., and Woisetschlager, M. (1996). Transcriptioncharacterization of pre-mRNA splicing factor SF2 from HeLa cells. of the Epstein–Barr virus nuclear antigen 1 (EBNA1) gene occursGenes Dev. 4, 1158 – 1171. before induction of the BCR2 (Cp) EBNA gene promoter during the

Krainer, A. R., Mayeda, A., Kozak, D., and Binns, G. (1991). Functional initial stages of infection in B cells. J. Virol. 70, 3561–3570.expression of cloned human splicing factor SF2: Homology to RNA-

Shah, W. A., Ambinder, R. F., Hayward, G. S., and Hayward, S. D. (1992).binding proteins, U1 70K, and Drosophila splicing regulators. Cell

Binding of EBNA-1 to DNA creates a protease-resistant domain that66, 383–394.

encompasses the DNA recognition and dimerization functions. J.Laine, A., and Frappier, L. (1995). Identification of Epstein–Barr virus

Virol. 66, 3355–3362.nuclear antigen 1 protein domains that direct interactions at a dis-

Smith, P. R., and Griffin, B. E. (1992). Transcription of the Epstein–Barrtance between DNA-bound proteins. J. Biol. Chem. 270, 30914–virus gene EBNA-1 from different promoters in nasopharyngeal carci-30918.noma and B-lymphoblastoid cells. J. Virol. 66, 706–714.Ling, P. D., Ryon, J. J., and Hayward, S. D. (1993). EBNA-2 of herpesvirus

Snudden, D. K., Hearing, J., Smith, P. R., Grasser, F. A., and Griffin, B. E.papio diverges significantly from the type A and type B EBNA-2(1994). EBNA-1, the major nuclear antigen of Epstein–Barr virus,proteins of Epstein– Barr virus but retains an efficient transactivationresembles ‘‘RGG’’ RNA binding proteins. EMBO J. 1354, 4840–4847.domain with a conserved hydrophobic motif. J. Virol. 67, 2990–3003.

Su, W., Middleton, T., Sugden, B., and Echols, H. (1991). DNA loopingLuo, Y., Yu, H., and Peterlin, B. M. (1994). Cellular protein modulatesbetween the origin of replication of Epstein–Barr virus and its en-effects of human immunodeficiency virus type 1 Rev. J. Virol. 68,hancer site: Stabilization of an origin complex with Epstein–Barr3850–3856.nuclear antigen 1. Proc. Natl. Acad. Sci. USA 88, 10870–10874.Mackey, D., Middleton, T., and Sugden, B. (1995). Multiple regions

Sugden, B., and Warren, N. (1989). A promoter of Epstein–Barr viruswithin EBNA1 can link DNAs. J. Virol. 69, 6199–6208.that can function during latent infection can be transactivated byMayeda, A., Zahler, A. M., Krainer, A. R., and Roth, M. B. (1992). TwoEBNA-1, a viral protein required for viral DNA replication during latentmembers of a conserved family of nuclear phosphoproteins are in-infection. J. Virol. 63, 2644–2649.volved in pre-mRNA splicing. Proc. Natl. Acad. Sci. USA 89, 1301–

Sung, N. S., Wilson, J., Davenport, M., Sista, N. D., and Pagano, J. S.1304.(1994). Reciprocal regulation of the Epstein–Barr virus BamHI-F pro-Middleton, T., and Sugden, B. (1992). EBNA-1 can link the enhancermoter by EBNA-1 and an E2F transcription factor. Mol. Cell Biol. 14,element to the initiator element of the Epstein–Barr virus plasmid7144–7152.origin of DNA replication. J. Virol. 66, 489–495.

Wilson, J. B., Bell, J. L., and Levine, A. J. (1996). Expression of Epstein–Miller, G. (1990). Epstein – Barr virus. In ‘‘Virology’’ (B. N. Fields, D. M.Barr virus nuclear antigen-1 induces B cell neoplasia in transgenicKnipe, R. M. Chanock, M. S. Hirsch, J. L. Melnick, T. P. Monath, andmice. EMBO J. 15, 3117–3126.B. Roizman, Eds.), 2nd ed., pp. 1921–1958. Raven Press, New York.

Wilson, J. B., and Levine, A. J. (1992). The oncogenic potential of Ep-Nonkwelo, C., Skinner, J., Bell, A., Rickinson, A., and Sample, J. (1996).stein–Barr virus nuclear antigen 1 in transgenic mice. Curr. Top.Transcription start sites downstream of the Epstein–Barr virus (EBV)Microbiol. Immunol. 182, 375–384.Fp promoter in early-passage Burkitt lymphoma cells define a fourth

Wold, M. S., Li, J. J., and Kelly, T. J. (1987). Initiation of simian virus 40promoter for expression of the EBV EBNA-1 protein. J. Virol. 70, 623–DNA replication in vitro: Large-tumor-antigen-and origin-dependent627.unwinding of the template. Proc. Natl. Acad. Sci. USA 84, 3643–3647.Peerschke, E. I., Reid, K. B., and Ghebrehiwet, B. (1994). Identification

of a novel 33-kDa C1q-binding site on human blood platelets. J. Wysokenski, D. A., and Yates, J. L. (1989). Multiple EBNA-1 binding sitesare required to form an EBNA-1 dependent enhancer and to activateImmunol. 152(12), 5896 – 5901.

AID VY 8739 / 6a46$$$$64 08-14-97 16:49:42 vira AP: VY

29P32/TAP INTERACTS WITH EBNA 1 PROTEIN

a minimal replicative origin within oriP of Epstein–Barr virus. J. Virol. Yates, J. L., Warren, N., and Sugden, B. (1985). Stable replication ofplasmids derived from Epstein–Barr virus in a variety of mammalian63, 2657–2666.

Yang, L., Mohr, I., Fouts, E., Lim, D. A., Nohaile, M., and Botchan, M. cells. Nature (London) 313, 812–815.Yu, L., Loewenstein, P. M., Zhang, Z., and Green, M. (1995a). In vitro(1993). The E1 protein of the papillomavirus BPV-1 is an ATP depen-

dent DNA helicase. Proc. Natl. Acad. Sci. USA 90, 5086–5090. interaction of the human immunodeficiency virus type 1 Tat transacti-vator and the general transcription factor TFIIB with the cellular pro-Yates, J., Warren, N., Reisman, D., and Sugden, B. (1984). A cis acting

element from the Epstein – Barr viral genome that permits stable repli- tein TAP. J. Virol. 69(5), 3017–3023.Yu, L., Zhang, Z., Loewenstein, P. M., Desai, K., Tang, Q., Mao, D.,cation of recombinant plasmids in latently infected cells. Proc. Natl.

Acad. Sci. USA 81, 3806 – 3810. Symington, J. S., and Green, M. (1995b). Molecular cloning and char-acterization of a cellular protein that interacts with the human immu-Yates, J. L., and Camiolo, S. M. (1988). Dissection of DNA replication

and enhancer activation functions of Epstein–Barr virus nuclear anti- nodeficiency virus type 1 Tat transactivator and encodes a strongtranscriptional activation domain. J. Virol. 69(5), 3007–3016.gen 1. Cancer Cells 6, 197 – 205.

AID VY 8739 / 6a46$$$$64 08-14-97 16:49:42 vira AP: VY