the of vol. 266, no. 12, issue of april 25, pp. 7819-7826...

THE JOURNAL 0 1991 by The American Society for Biochemistry

OF BIOLOGICAL CHEMISTRY and Molecular Biology, Inc

Vol. 266, No. 12, Issue of April 25, pp. 7819-7826,1991 Printed in U. S. A.

Overproduction, Purification, and Characterization of EBNA1, the Origin Binding Protein of Epstein-Barr Virus*

(Received for publication, November 19, 1990)

Lori Frappier and Mike O’DonnellS From the Howard Huphes Medical Institute, Hearst Microbiology Department, Cornell University Medical College, New York, New York10021

The baculovirus expression system was used to overproduce the Epstein-Barr virus nuclear antigen, EBNA1, in insect cells. EBNAl overproduced via baculovirus expression (baculoEBNA1) was followed during purification to homogeneity using its ability to specifically retain the family of repeats of the latent origin of replication, oriP, onto nitrocellulose filters. A two-column procedure was developed which yields more than 1 mg of homogeneous baculoEBNA1 from 9 X 10’ insect cells (1.5 liters). Pure baculoEBNA1 had no detectable ATPase or helicase activity. Baculo- EBNAl was labeled with [32Plorthophosphate in uiuo, and analysis showed detectable levels of phosphoserine; no phosphothreonine or phosphotyrosine could be detected. The baculoEBNA1 appeared dimeric in solution, and a stoichiometry of 56 baculoEBNA1 monomers per 24 EBNAl binding sites in oriP suggests baculoEBNA1 binds its consensus site as a dimer. The binding of baculoEBNA1 to the dyad symmetry element of oriP ( K d - 2 nM) required more baculoEBNA1 and appeared less stable than the binding of baculo- EBNAl to the family of repeats in oriP (Ed - 0.2 nM).

Epstein-Barr virus (EBV)’ is the causative agent of infec- tious mononucleosis and is associated with some types of cancer (reviewed in Ref. 1). During the latent infection of human lymphocytes, multiple copies of the EBV genome are maintained as 172-kilobase pair extrachromosomal DNA plasmids (2). Only one viral protein, B V nuclear gntigen 1 (EBNAl), is essential for the replication of these EBV plasmids (3). EBNAl binds to the latent origin of replication, oriP, at multiple sites present in the two regions of oriP which were found to be necessary and sufficient for origin function (4, 5). One of these regions is composed of 20 tandem copies of a 30-bp sequence (family of repeats), each of which contains an EBNAl binding site. The other region includes four EBNAl binding sites (dyad symmetry element), two of which

*This research was supported by Grant ROI-GM38839 of the National Institute of General Medical Sciences, United States Public Health Service Grant ROI CA531-01 of the National Cancer Institute, Grant BRSG SO7 RR05396 of the Division of Research Resources, National Institutes of Health, and by a Medical Research Council of Canada fellowship (to L. F,). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

f To whom reprint requests should be addressed. ’ The abbreviations used are: EBV, Epstein-Barr virus; EBNA1,

Epstein-Barr virus nuclear antigen 1; bacuIoEBNA1, form of EBNAl overproduced using the baculovirus expression system; CIP, alkaline phosphatase from calf intestine; SDS, sodium dodecyl sulfate; PMSF, phenylmethylsulfonyl fluoride; bp, base pair(s); HEPES, 4-(2-hydrox- yethy])-l-piperazineethanesulfonic acid.

are located within a 65-bp region of dyad symmetry (5). The interaction of EBNAl with oriP occurs mainly through the carboxyl-terminal third of the protein (6). EBNAl activates oriP to function not only as an origin of replication but also as a plasmid maintenance element (7,8) and a transcriptional enhancer (9, 10).

Our interest in the latent phase of EBV replication is its extensive reliance on host factors, EBNAl being the only viral encoded protein required for latent phase replication from oriP. It is our hope that an in vitro EBV replication assay can be developed for use in isolating host factors that also participate in replication of human chromosomes. Since EBNAl is not abundant in EBV-infected cells, we have used the baculovirus expression system (11,12) to overproduce the EBNAl protein. This expression system was used because other proteins overproduced with the baculovirus system, including SV40 large T antigen (13) and the E2 protein of bovine papilloma virus,’ have been shown to be functionally active. We have purified baculoEBNA1 to homogeneity and characterized its phosphorylation state, aggregation state, and interaction with the EBV latent origin, oriP.

MATERIALS AND METHODS

Cells and Virus-The wild-type baculovirus, Autographa celifornica nuclear polyhedrosis virus (AcMNPV), and Spodoptera frugiperda (SF-9) cells used to propagate the baculoviruses, were kindly provided by Dr. Ora Rosen (Sloan-Kettering Cancer Center), with permission from Dr. Max D. Summers (Texas A & M University). SF-9 cells were grown as monolayer cultures in Grace’s medium (Gibco Labo- ratories) with 0.33% yeastolate and 0.33% lactalbumin hydrolysate (Difco) supplemented with 10% fetal bovine serum.

Plasmids-pVL941-SW was constructed from pVL941 (14) by Dr. Susan Wente in Dr. Ora Rosen’s laboratory, by insertion of an NcoI/ XbaIISpeI linker into the BamHI site of the polyhedrin gene in pVL941. Plasmid p205, containing the EBNAl gene with a 700-bp (t-20 bp) deletion in the glycine-alanine repeat region, was kindly provided by Dr. Bill Sugden (3). Plasmid pGEMoriP7 was constructed by ligatingRsaI/HindIII DNA linkers to the ends of the RsaI fragment of p220.2 (kindly provided by Bill Sugden (Ref. 15)) containing oriP and the EBNAl gene and inserting this DNA fragment into the Hind111 site of pGEMII (Promega Biotec, Madison, WI). pGEMoriP was constructed from pGEMoriP7 using AccI to excise 2 kilohase pairs of DNA containing the EBNAl gene followed by religation to give pGEMoriP, which contains the entire oriP sequence.

Construction of the EBNAI Recombinant Bueubvirus (AcMNPV- EBNA1,J-The EBNAl gene was excised from p205 using RsaI and Ban, which remove the first seven codons. The initiating methionine was regenerated upon ligation into the baculovirus transfer vector pVL941-SW to yield pVL941/EBNA1 (Fig. 1). pVL941/EBNA1 and AcMNPV DNA were cotransfected into SF-9 cells by the calcium phosphate precipitation method as described in Summers and Smith (16). Five days post-transfection, serial dilutions of the medium from the transfected cells were plated with SF-9 cells in 96-well plates. After amplification of the virus for 4 days, the cells were screened for

~

* M. Lusky, personal communication.

7819

7820 Epstein-Barr Virus Nuclear Antigen 1

the presence of virus containing the EBNAl gene by dot blot analysis. The medium from a positive well was then used in a plaque assay (16), and a recombinant (nonoccluded) plaque was picked, analyzed for the presence of the EBNAl gene by dot blots, and subjected to one more round of plaque purification. Virus from one of the resulting recombinant plaques was amplified in SF-9 cells. Total DNA was prepared from these cells, digested with restriction enzymes, and analyzed by Southern blot hybridizations to verify the presence of the complete EBNAl RsaI-BalI fragment in the recombinant virus.

DNA Oligonucleotide Affinity Column-The complementary DNA oligonucleotides,5' AGATTAGGATAGCATATGCTACCCAGATAT 3' and 5' TCTATATCTGGGTAGCATATGCTATCCTAA 3', were synthesized on an Applied Biosystems 381A DNA synthesizer and purified by polyacrylamide gel electrophoresis. Hybridization of the two DNA oligomers yields the consensus 30-bp repeated sequence of oriP (with directional 5' overhangs) to which EBNAl binds (6). An equal weight of each oligomer (220 pg each) was mixed, hybridized, phosphorylated, and ligated as described (17). The ligated oligomers were coupled to 10 ml of Sepharose CL-2B (Pharmacia LKB Bio- technology Inc.) freshly activated with cyanogen bromide (17).

Nitrocellulose Filter Binding Assays-During purification, baculoEBNA1 was followed and quantitated by its ability to specifically retain a 900-bp fragment of oriP containing 20 copies of the 30- bp repeated sequence (family of repeats fragment) onto nitrocellulose filters. The family of repeats fragment was excised from pGEMoriP with EcoRI and NcoI, purified by agarose gel electrophoresis followed by electroelution, and quantitated by measuring the absorbance at 260 nm. The oriP repeat fragment was end-labeled by filling in the 3"recessed ends using the Klenow fragment of DNA polymerase I with four dNTPs and [a-32P]TTP. Assays for baculoEBNAl were performed by incubating an aliquot (20-200 ng of protein) of each fraction with 10-100 fmol of the end-labeled family of repeats fragment for 10 min at 23 "C, in 25 pl of 50 mM HEPES (pH 7.5), 5 mM MgCIz, and 300 mM NaCl containing 2.5-5 pg of calf thymus DNA. Reaction mixtures were then diluted with 900 p1 of 50 mM HEPES (pH 7.5), 5 mM MgCI, and immediately filtered through 0.45-pm HA filters (Millipore). The filters were dried and counted by liquid scintillation.

In nitrocellulose filter binding assays of baculoEBNA1 with the dyad symmetry element of oriP, a 300-bp fragment of oriP containing the dyad and its four associated EBNAl binding sites was incubated with baculoEBNA1 as described for the family of repeats. This dyad symmetry element fragment was excised from pGEMoriP with HindIII and EcoRV, gel-purified, quantitated by absorbance at 260 nm, and end-labeled as described for the family of repeats fragment.

In assays in which bacuIoEBNA1 was incubated with the complete oriP sequence, a 2-kilobase pair DNA fragment containing oriP was prepared from pGEMoriP and end-labeled near the dyad. This fragment was prepared by linearizing pGEMoriP with HindIII, filling in the 3"recessed ends with [a-3ZP]TTP using the Klenow fragment of DNA polymerase I, then digesting with BamHI. The HindIII to BamHI fragment containing the complete oriP sequence was gel- purified and incubated with baculoEBNAl as described for the family of repeats fragment.

Protein Determinations-Protein concentration was determined by the method of Bradford (18) using bovine serum albumin as a standard. The concentration of pure baculoEBNAl was determined bv amino acid analvsis (Sloan-Kettering Institute, Microchemistry Laboratory).

PhosDhate and Methionine LabelinE of BaculoEBNAl-SF-9 cells (2.8 X i? cells, 10 X 150-cmZ flasks) were infected with recombinant EBNAl baculovirus as described above. Twenty-four hours post- infection, the media was replaced with phosphate-free or methionine- free Grace's media (Gibco) supplemented with 0.33% lactalbumin hydrolysate and 1 mCi of [32P]orthophosphate or [35S]methionine (Du Pont-New England Nuclear). Cells were labeled for 18 h before nuclei were prepared. Labeled baculoEBNA1 was purified as described in the legend to Table I.

Phosphoamino Acid Analysis of BaculoEBNAI-Acid hydrolysis of baculoEBNA1 and resolution of phosphoamino acids was performed by the method of Cooper et al. (19). Four-microgram samples of pure ["P]baculoEBNAl (4 pl) were added to 50 pl of 6 N constant boiling HCl (Pierce Chemical Co.) and heated to 110 "C in a screw-cap 1.5- ml Eppendorf tube for 1, 2, or 4 h. The samples were lyophilized and resuspended in 2 pl of distilled water containing 4 pg each of phosphoserine, phosphothreonine, and phosphotyrosine markers. One mi- croliter of each sample (2 p g of hydrolyzed baculoEBNA1) was spotted onto a strip of Whatman No. 3MM paper and subjected to electro-

phoresis in 0.5% pyridine, 5% acetic acid for 10 min at 2000 V. The paper was then dried and stained with ninhydrin to visualize the phosphoamino acid markers. 32P-Labeled amino acids of baculo- EBNAl were identified by autoradiography.

Phosphatase Treatment-Complete dephosphorylation of 32P-labeled baculoEBNA1 was achieved by treating 0.25 pg of [3zPpl baculoEBNA1 with 9 units of CIP (Sigma) in 20 p1 of 10 mM Tris- HCI (pH 8.01, 1 mM MgClz, 0.1 mM ZnCI?, 300 mM NaCl for 1 h at 25 "C.

Native Molecular Weight Determination-The sedimentation coefficient of haculoEBNA1 was measured by layering 40 pg of baculo- EBNAl either alone or along with 60pg of molecular weight standards (apoferritin, IgG, bovine serum albumin, ovalbumin, and myoglobin) in 200 p1 of 25 mM Tris-HCI (pH 7.5), 300 mM NaCl, 0.5 mM EDTA, 10% glycerol onto 12-ml10-30% glycerol gradients containing 25 mM Tris-HC1 (pH 7.5), 300 mM NaC1, 0.5 mM EDTA. Gradients were spun for 40 h at 270,000 X g at 5 "C in a TH-641 rotor. After centrifugation, fractions of 160 pl were collected from the bottom of each tube.

The Stokes radius of baculoEBNA1 was determined by injecting 40 pg of baculoEBNA1 along with 60 pg of molecular weight standards in 200 pl of 25 mM Tris-HCI (pH 7.5), 300 mM NaC1,0.5 mM EDTA, 10% glycerol onto a 30-ml fast protein liquid chromatography Super- ose 12 gel filtration column. The column was developed in the same buffer. Fractions of 160 pl were collected. Two microliters of each fraction from the glycerol gradients and gel filtration columns were assayed for the presence of baculoEBNA1 using the nitrocellulose filter binding assay described above. BaculoEBNAl and the molecular weight standards were visualized after SDS-polyacrylamide gel electrophoresis (20) analysis by staining with Coomassie Blue.

Stoichiometry of BaculoEBNAl on OriP-Thirty-five micrograms (7.7 pmol as plasmid circles) of pGEMoriP7 were incubated with excess 35S-labeled baculoEBNA1 (56 pg, 1.1 nmol as monomer) for 10 min at 37 "C in 200 pl of 20 mM HEPES (pH 7.5), 5 mM MgCIZ, 300 mM NaC1, 40% glycerol. The reaction was gel-filtered over a 5- ml Bio-Gel A-5m column at 4 "C in the same buffer, and 140-pl fractions were collected. [36S]BaculoEBNAl in each fraction was quantitated by counting 30 pl in a scintillation counter. The molar quantity of DNA in each fraction was measured upon diluting 100 p1 of column fraction with 400 pl of column buffer and measuring the absorbance at 260 nm (assuming 1 absorbance unit equals 50 pg/ml DNA). Approximately 90% of the radioactivity and absorbance at 260 nm was recovered after gel filtration.

AvaI Endonuclease Protection Assay-The 300-bp HindIII to EcoRV fragment of pGEMoriP containing the dyad symmetry element (see "Nitrocellulose Filter Binding Assays" above) was end- labeled using the Klenow fragment of DNA polymerase I .and TTP to fill in the HindIII end of the fragment. BaculoEBNAl was incubated with 10 fmol of the 32P-labeled dyad fragment in a 20-4 reaction containing 50 mM HEPES (pH 7.5), 300 mM NaCI, 5 mM MgCI, for 10 min at room temperature. The reactions were then diluted to 50 mM NaCl and incubated with 30 units of AvaI at 37 "C for 3 min. Digestions were stopped by the addition of SDS to 1%. Half of each reaction was then subjected to electrophoresis on a 6% polyacrylamide gel, which was dried prior to autoradiography.

RESULTS

Expression of EBNAl in Baculouirus-The EBNAl gene was excised from plasmid p205 (3) and inserted into the pVL941-SW baculovirus transfer vector as described in Fig. 1. The resulting plasmid, pVL941-EBNA1, contained the EBNAl gene, which translates into a BO-kDa protein lacking six amino-terminal amino acids and approximately 230 con- tiguous glycine and alanine residues of the glycine-alanine repeat region. Neither of these regions was essential for EBNAl-dependent replication in vivo when tested separately (mutants CAT-EBNA1 and DL7 in Ref. 21). Recombination of pVL941-EBNA1 with AcMNPV wild-type baculovirus DNA resulted in a recombinant baculovirus (AcMNPV- EBNA1) containing the EBNAl gene controlled by the strong polyhedrin gene promoter. The EBNAl produced by AC- MNPV-EBNA1 will be referred to here as baculoEBNAl. BaculoEBNAl is not a fusion protein, as the EBNAl gene was placed directly adjacent to the only ATG sequence present

Epstein-Barr Virus Nuclear Antigen 1 7821 q6 +l

Pallhed""

PVL941-SW

CATGG C C GGTAC 4 Phosphatase

4 PolI(Klenow)

CAT- GTACC

anp p m c c m GGTAC

FIG. 1. Construction of the EBNAl baculovirus transfer vector pVL941-EBNA1. The sequence of the oligonucleotide linkers inserted in the polyhedrin gene of the baculovirus transfer vector pVL941-SW is shown above the plasmid. The underlined ATG is the only ATG sequence in the 5' region of the polyhedrin gene and was used as the start codon for translation of the EBNAl gene. After linearization of pVL941-SW with NcoI and digestion with bacterial alkaline phosphatase, the 3"recessed ends were extended with the Klenow fragment of DNA polymerase I. The EBNAl gene was excised from p205 with RsaI and BalI enzymes, which remove the first seven codons of the gene, and ligated into pVL941-SW to form pVL941- EBNA1. Hygromycin phosphotransferase (hph) and p-lactamase (amp) genes are also shown.

in the 5' region of the polyhedrin gene in pVL941-SW (Fig. 1).

In initial experiments, SF-9 monolayers were infected with AcMNPV-EBNA1 and harvested a t 24-h intervals to deter- mine the time course of baculoEBNA1 expression. Baculo- EBNAl protein levels peaked approximately 48 h post-infection as determined by the ability of whole cell extracts to specifically retain the oriP repeat fragment on a nitrocellulose filter. The level of oriP binding activity correlated with the appearance on Coomassie-stained SDS-polyacrylamide gels of a 50-kDa protein that was not present in SF-9 cells infected with wild-type baculovirus (data not shown).

Purification of BaculoEBNAl-The purification of baculoEBNA1, from approximately 9 X 10' SF-9 cells infected with AcMNPV-EBNA1 baculovirus, is summarized in Table I. The cells were harvested 46 h post-infection and separated into nuclear and cytoplasmic fractions. Origin binding activity and the 50-kDa protein produced upon infection were detected only in the nuclear fraction (Fig. 2). Dialysis against low salt buffers was avoided at all stages of purification, since baculoEBNA1 had a tendency to precipitate at under 250 mM NaC1. Hence the nuclear extract was diluted to 350 mM NaCl and loaded onto a column of heparin-agarose. The heparin- agarose column was washed with 500 mM NaCl and the origin binding activity was eluted from the column with 1 M NaCl (Fig. 2). At this point the 50-kDa protein appeared to be a t

TABLE I Purification of baculoEBNA1 from insect cells

SF-9 cells were seeded into 30 150-cm2 culture flasks (3 X lo7 cells/ flask) (Corning), allowed to attach, then infected with AcMNPV- EBNAl at a multiplicity of infection of three. The cells were harvested 46 h post-infection, washed in 500 ml of ice-cold phosphate- buffered saline, and resuspended on ice in 110 ml of hypotonic buffer (20 mM HEPES (pH 7.5), 1 mM MgCl,, 1 mM PMSF) using a Dounce homogenizer with pestle B. Nuclei were collected upon centrifugation a t 1000 X g for 10 min a t 5 "C, washed in 80 ml of cold hypotonic buffer, and resuspended with the Dounce homogenizer and pestle B in 30 ml of 20 mM HEPES (pH 7.5). 1 M NaCl, 1% Nonidet P-40, 10% glycerol, 1 mM MgCIz, 1 mM PMSF, followed by incubation for 1 h on ice. The nuclear extract was clarified by centrifugation a t 39,000 X g for 80 min a t 5 'C. All column chromatography procedures to follow were a t 4 "C. The clarified nuclear extract was diluted with buffer A (20 mM HEPES (pH 7.51, 0.5 mM EDTA, 2 mM dithiothreitol, 1 mM PMSF, 20% glycerol) to a final conductivity equivalent to 350 mM NaCl before loading onto a 30-ml heparin-agarose column (Bio-Rad). The heparin-agarose column was washed with 60 ml of buffer A containing 500 mM NaCl at 0.27 ml/min; then baculoEBNA1 was eluted with buffer A containing 1 M NaCI. Fractions of the 1 M NaCl eluate containing oriP binding activity were pooled (26 ml), diluted to 350 mM NaCl with buffer A, then loaded onto 9 ml of the DNA oligonucleotide affinity column in three batches. The DNA affinity column was washed with 18 ml of 350 mM NaCI, and then baculoEBNAl was eluted using buffer A containing 2 M NaCI. T o concentrate baculoEBNA1, the 2 M NaCl eluate containing 33% of the oriP binding activity (50 ml) was dialyzed against 500 mM NaC1, diluted with buffer A to a conductivity equivalent to 260 mM NaCl (105 ml), and loaded onto a 1-ml Mono Q column. BaculoEBNAl was eluted with buffer A containing 500 mM NaC1. Aliquots of active fractions (20 ulltube) were stored a t -70 "C.

~~

Fraction Protein Activity 22:; Purification Yield

mg units" unitslmg -fold %

Cytoplasm 264 0 0 Nuclear extract 84 6600 79 4 100 Heparin-agarose 9.8 4000 408 21 61 Oligonucleotide affinity 1.4 2200 1571 80 33

a One unit is defined as the amount of baculoEBNA1 required to retain 1 pmol of the oriP family of repeats onto a nitrocellulose filter (see "Materials and Methods").

(kdi

92

6e

31

22

FIG. 2. Purification of baculoEBNA1. Cytoplasmic and nuclear fractions of SF-9 cells infected with AcMNPV-EBNAl baculovirus were prepared as described in the legend to Table I. Baculo- EBNAl was purified to homogeneity from the nuclear extract by a combination of heparin-agarose column chromatography (heparin), and oriP DNA affinity column chromatography ( D N A affinity). The homogeneous baculoEBNAl was concentrated on a Mono Q column (Mono Q). Molecular mass standards are shown (STDS). Samples were subjected to electrophoresis on a 10% SDS-polyacrylamide gel and were visualized by staining with Coomassie Blue.

7822 Epstein-Barr Virus

least 30% pure. The 1 M NaCl eluate from the heparin-agarose column was diluted to 350 mM NaCl and loaded onto a DNA oligonucleotide affinity column containing the oriP 30-bp repeated sequence. The oriP affinity column was eluted using 2 M NaCl to yield 1.4 mg of homogeneous 50-kDa baculo- EBNAl protein (Fig. 2). The baculoEBNA1 protein was then concentrated on a Mono Q column to 0.5 mg/ml, then frozen in 20-pl aliquots at -70 "C (see legend to Table I).

Biochemical Assays of BaculoEBNAl-Homogeneous baculoEBNA1 was assayed for the ability to hydrolyze ATP, GTP, CTP, UTP, dATP, dGTP, dCTP, and TTP in 1,3, and 10 mM MgC12, in the absence of DNA and in the presence of either oriP-containing duplex DNA or single-stranded DNA? The 7 subunit of Escherichia coli DNA polymerase I11 holo- enzyme was used as a positive control for ATP hydrolysis (22). No hydrolysis of any nucleoside triphosphate by baculoEBNA1 was detected (data not shown).

Although all known helicases are ATPases, we tested baculoEBNA1 in the standard oligonucleotide displacement type of helicase assay (23). BaculoEBNAl was examined for an ability to displace, from single-stranded circular bacteriophage 4x174 DNA, a "P-end-labeled flush DNA 30-mer, a 5'-tailed DNA 30-mer, and a 3'-tailed DNA 46-mer: The SV40 large T antigen was used as a positive control (24). Although the SV40 T antigen displaced each of these DNA oligonucleotides, no helicase activity was detected for baculoEBNA1, consistent with its lack of ATPase activity. We have also tested baculoEBNA1 for DNA polymerase, DNA ligase, endonuclease, exonuclease, and topoisomerase activities without positive results (not shown).

Phosphoamino Acid Analysis of BaculoEBNAl-Baculo- EBNAl was labeled in vivo with ["P]orthophosphate and purified to homogeneity. BaculoEBNAl was the major 32P- labeled protein in the nuclear extract and was not detected in the cytoplasm (Fig. 3). Treatment of pure [32P]baculoEBNAl with CIP resulted in loss of all detectable radioactive phosphate from baculoEBNA1 (Fig. 3). Since CIP has previously been shown to dephosphorylate serine residues only (25, 26), baculoEBNA1 is presumably phosphorylated only on serine.

Further identification of phosphorylated residues in

' Nucleotide hydrolysis assays were performed by incubating 200 ng of baculoEBNAl with 50 p~ [(U-'~P]- or [y-"'P]nucleoside triphosphate and deoxynucleoside triphosphate in 10 pl of 20 mM Tris-HCI (pH 7.5) and 1, 3, or 10 mM MgCIZ for 30 min at 37 "C. Additional assays for nucleoside triphosphate and deoxynucleoside triphosphate hydrolysis were performed in the presence of 50 ng of bacteriophage M13 single-stranded DNA at the three MgC12 concentrations, as well as in the presence of 75 ng of pGEMoriP at the three MgCI2 concentrations. ATPase activity was also tested in the presence of 2 and 8 mM sodium acetate. Samples (0.5 pl) of reaction mixtures were spotted on polyethyleneimine cellulose thin layer chromatography plates and developed in 0.8 M acetic acid, 0.8 M LiCl (when cy-:'*P- labeling was used) or 1 M ammonium formate, 0.5 M LiCl (when y- ""P-labeling was used). Reaction products were identified by autoradiography. ' In separate experiments three different synthetic DNA oligonu-

cleotides were hybridized to bacteriophage 6x174 single-stranded DNA to give either 1) flush (30-mer), 2) 5"tailed (30-mer with 20 nucleotides annealed), or 3) 3"tailed (46-mer with 30 nucleotides annealed) helicase substrates. The annealed oligonucleotides were 3'- end-labeled using either [ ( u - ~ ~ P I ~ C T P and the Klenow fragment of DNA polymerase I (flush and 5"tailed substrates) or using terminal transferase (3"tailed substrate). Each helicase substrate was then purified from unhvbridized olieonucleotide bv eel filtrat,ion on Rio- Gel A-1.5m. Helicase assays were performed by incubating 400 ng of baculoEBNA1 with 9 fmol of DNA substrate in 30 mM HEPES (pH 7.5), 4 mM ATP, 7 mM MgCIz, 1 mM dithiothreitol for 30 min at 37 "C. Positive control reactions contained 400 ng of SV40 large T antigen. Reaction products were analyzed for oligonucleotide displacement on a 15% polyacrylamide gel.

Nuclear Antigen 1

(kd4 CYt-

92- rL

* n-

22-

CIP: - +

r) '+EBNAl

FIG. 3. Phosphate labeling and phosphatase digestion of baculoEBNA1. SF-9 cells were infected with the AcMNPV-EBNA1 baculovirus and labeled with [32P]orthophosphate as described under "Materials and Methods." Labeled cells were separated into cytoplasmic (cyt) and nuclear ( n u ) fractions, and baculoEBNAl was purified to homogeneity from the nuclear extract. Pure ["PI baculoEBNAl was incubated at 25 "C for 1 h either with (+) or without (-) CIP. Samples were subjected to electrophoresis on 12% SDS-polyacrylamide gels and "P-labeled proteins were detected upon autoradiography of wet gels.

. . . : Thr-P . . . . . . . .... . .

: : : : Tyr-p ... .. . . ... ..

8

origin- 0 o 1 2 (hrsl

FIG. 4. Phosphoamino acid analysis of baculoEBNAl. Pure ["P]baculoEBNAl was hydrolyzed in 6 N HCI (constant boiling) at 110 "C for 1 or 2 h, then combined with unlabeled phosphoserine (Ser-P), phosphothreonine (Thr-P), and phosphotyrosine (Tyr-P) markers. The mixture of hydrolyzed amino acids and phosphoamino acid standards was separated by high voltage paper electrophoresis. Positions of phosphoamino acid markers were visualized by ninhydrin staining and are indicated by dotted circles. Positions of 32P-labeled amino acids and nonhydrolyzed baculoEBNA1 (0-h data lane) were identified by autoradiography.

baculoEBNA1 was performed by acid hydrolysis of ["PI

high voltage paper electrophoresis (Fig. 4). Samples of ["PI baculoEBNA1 hydrolyzed for 1, 2, and 4 h were analyzed to ensure identification of any [32P]phosphothreonine, which requires longer hydrolysis times, or [32P]phosphotyrosine, which is less stable to acid hydrolysis (19). Upon electropho-

baculoEBNA1 and separation of the phosphoamino acids by

Epstein-Barr Virus Nuclear Antigen 1 7823

resis, all the radioactive phosphate in baculoEBNA1 migrated in the position of phosphoserine; no radioactivity was detected at positions of phosphothreonine or phosphotyrosine (Fig. 4). After 4 h of acid hydrolysis most of the radioactive phosphate was detected as free phosphate (data not shown).

Native Molecular Mass-BaculoEBNAl was analyzed by glycerol gradient sedimentation; an s value of 4.6 was obtained by comparison with protein markers with known s values (Fig. 5A). A Stokes radius of 50 A for baculoEBNA1 was determined by gel filtration analysis and comparison with protein standards of known Stokes radius (Fig. 5B). In both glycerol gradient and gel filtration analyses, oriP binding activity coeluted with the baculoEBNA1 protein visualized in SDS- polyacrylamide gel analysis of the column fractions (data not shown). The s value and Stokes radius were combined in the equation of Siege1 and Monty (27) to calculate a native molecular mass of 94 kDa for baculoEBNA1. The amino acid sequence of EBNAl deduced from the DNA sequence of the EBNAl gene predicts a molecular mass of 50 kDa for a baculoEBNAl monomer (28). Hence, the native molecular mass of baculoEBNA1 indicates that baculoEBNA1 is a dimer.

Stoichiometry of BaculoEBNAl Binding to OriP--35S-La- beled baculoEBNA1 protein was prepared in uiuo by metabolic labeling using [35S]methionine followed by purification to

u) I 10 20 30 40 50 60

Fraction

B kda: 440 158 66 45 17 + + + + + A

u) L, Y I 20 30 40 50 60

Fraction

FIG. 5. Native aggregation state of baculoEBNA1. BaculoEBNAl was combined with the protein standards apoferritin (upo; 440 kDa), IgG (158 kDa), bovine serum albumin (BSA; 66 kDa), ovalbumin ( o m ; 45 kDa) and myoglobin (myo; 17 kDa), then analyzed by glycerol gradient sedimentation (A) or gel filtration on Superose 12 ( B ) as described under “Materials and Methods.” BaculoEBNAl was identified in column fractions by the nitrocellulose filter binding assay. The sedimentation coefficient (s) and Stokes radius of baculoEBNA1 were determined by comparison to the positions of protein standards of which the s values and Stokes radii are known.

homogeneity. The [35S]baculoEBNAl was used to measure the number of baculoEBNA1 molecules bound to oriP under conditions of saturating baculoEBNA1. A plasmid containing the complete oriP sequence was incubated with increasing amounts of [35S]baculoEBNAl then gel-filtered to separate [35S]baculoEBNAl bound to DNA in the excluded fractions from the unbound [35S]baculoEBNAl in the included fractions. Upon saturation of oriP with baculoEBNA1, indicated by the appearance of baculoEBNA1 in the included fractions, the molar ratio of baculoEBNA1 monomers per oriP DNA which comigrated in the excluded fractions was 56 to 1 (Fig. 6). Since there are 24 EBNAl binding sites in oriP, the stoichiometry of 2.3 baculoEBNA1 monomers per EBNAl binding site indicates that baculoEBNA1 bound its site as a dimer, consistent with the native molecular weight of baculoEBNA1 and the palindromic structure of the consensus EBNAl binding site.

Effect of Salt on Binding of BaculoEBNAl to the Family of Repeats and Dyad Symmetry Element-The effect of NaCl concentration on baculoEBNA1 binding to the family of repeats or the dyad symmetry element of oriP was studied using the nitrocellulose filter binding assay. BaculoEBNAl (50 ng) was incubated with 40 fmol of 32P-labeled dyad fragment or 32P-labeled repeat fragment in various concentrations of NaCl and in the presence of excess (2.5 pg) calf thymus DNA (Fig. 7). The binding profile indicates that the specific interaction of baculoEBNA1 with the dyad symmetry element was maximum at 250-300 mM NaCl and dropped off sharply at higher NaCl concentrations. Binding of baculoEBNA1 to the family of repeats, however, remained stable up to 500 mM NaC1. Hence, the relative binding strength of baculoEBNA1 for the family of repeats uersus the dyad symmetry element depended on the salt concentration. The apparent requirement of high salt for binding baculoEBNA1 to labeled DNA in these experiments may be attributed to efficient competi- tion by nonspecific calf thymus DNA at low NaCl concentration.

BaculoEBNAl Binding to the Dyad. Symmetry Element- The interaction of baculoEBNA1 with the family of repeats and dyad symmetry element of oriP was also assessed by examining the amount of baculoEBNA1 required to retain each element on nitrocellulose filters. Increasing amounts of baculoEBNA1 were incubated with 10 fmol of 32P-end-labeled

0 - 5 m w 0 - E,

10 20 30 40

Fraction

FIG. 6. Stoichiometry of [36S]baculoEBNAl bound to oriP DNA. [35S]BaculoEBNAl was incubated with pGEMoriP7, then gel- filtered to separate [35S]baculoEBNAl bound to pGEMoriP7 in the excluded fractions from unbound baculoEBNA1 in the included fractions as described under “Materials and Methods.” Fractions were analyzed for DNA and [35S]baculoEBNAl as described under “Ma- terials and Methods.”


I I

NaCl concentration (rnM)

FIG. 7. Salt dependence of baculoEBNA1 binding to the family of repeats and the dyad symmetry element. Baculo- EBNAl (50 ng) was incubated with 40 fmol of "P-end-labeled DNA containing either the dyad symmetry element (closed circles) or the family of repeats (open circles) in the presence of 2.5 pg of calf thymus DNA and various concentrations of NaCI. After 10 min at 23 "C, the reaction mixture was filtered through nitrocellulose, and the DNA retained on the filters was quantitated by liquid scintillation.

100 200 300 400 500 600

w EcoAV

Dyac

fmol EBNAI (as dimer)

FIG. 8. Effect of the family of repeats on binding of baculoEBNA1 to the dyad symmetry element. Top, diagram of oriP showing the disposition of EBNAl binding sites (boxes). Bottom, 10 fmol of :'2P-labeled DNA fragment containing either the family of repeats (open circles), the dyad symmetry element (closed circles), or the complete oriP (closed triangles) were incubated with various amounts of baculoEBNAl (shown as fmol dimers) in 50 mM HEPES (pH 7.5), 300 mM NaC1,5 mM M&12 for 10 min a t 23 "C. Reactions containing the family of repeats or dyad symmetry element were then filtered through nitrocellulose. Reactions containing the complete oriP (closed triangles) were treated with 50 units of EcoRV for 3 min at 37 "C to separate the family of repeats from the end-labeled dyad symmetry element (see scheme, top) prior to filtration through nitrocellulose.

repeat or dyad DNA fragment in 20 pl of buffer containing 300 mM NaCl and no calf thymus DNA. Retention of the dyad symmetry element onto nitrocellulose appeared to have a threshold where significant retention was not observed below 20 baculoEBNA1 dimers per dyad fragment (200 ng in Fig. 8, closed circles), but full retention was achieved a t 50 baculoEBNA1 dimers per dyad (500 ng in Fig. 8). It would seem from this behavior that baculoEBNA1 must reach a critical concentration before it binds the dyad symmetry element. The apparent K d for baculoEBNA1 binding to the dyad symmetry element calculated from these data is 2 nM (assuming four baculoEBNA1 dimers were bound per dyad symmetry element). The family of repeats was retained onto nitrocellulose at lower levels of baculoEBNA1 than required

Densitometric Analysis

l7-7 I

~ w ~ w ~ M ) u x ) M o ~ ~ ~ M ~ tmol E m 1 (U dlmr)

FIG. 9. Protection of the AuaI site in the dyad symmetry element by baculoEBNA1. A, the 300-bp DNA fragment containing the dyad symmetry element, "2P-end-labeled at one end only (*), was incubated with various amounts of baculoEBNA1 (shown as fmol dimers) prior to digestion with AuaI and electrophoresis on a 6% polyacrylamide gel as described under "Materials and Methods." The DNA was visualized by autoradiography of dried gels. Scheme of DNA fragment ( top) shows EBNAl consensus binding sites (boxes). B, the AuaI-protected bands in the autoradiograph in A were quantitated by a laser densitometer (LKB Bromma Ultroscan XL).

for binding the dyad symmetry element (Fig. 8, open circles). An apparent K d for baculoEBNA1 binding to the family of repeats was calculated to be 0.2 nM (assuming four baculo- EBNAl dimers were bound per family of repeats).

The binding of baculoEBNA1 to the dyad symmetry element was further examined by an AuaI endonuclease protection assay. An AuaI site was present at the junction of two of the four EBNAl binding sites in the dyad symmetry element (Fig. 9). Increasing amounts of baculoEBNA1 were incubated with 10 fmol of the dyad symmetry element, end-labeled with 32P a t one end only. The reaction was then treated with sufficient AuaI to completely digest the DNA within 3 min at 37 "C. Digestions were stopped with SDS and subjected to polyacrylamide gel electrophoresis to separate DNA fragments cut by AuaI from uncut (Ad-protected) DNA (Fig. 9). As in the nitrocellulose binding assay, the AuaI protection analysis showed that a 20-fold molar excess of baculoEBNA1 dimers (200 ng in Fig. 9) was required over the dyad fragment to detect protection of the AuaI site, followed by a very sharp increase in protection against AuaI at levels above 20 baculoEBNA1 dimers per dyad symmetry element.5

The small difference between the AuaI protection assay (Fig. 9) and the nitrocellulose filter binding assay (Fig. 8) showed approximately 1.5 times more baculoEBNA1 was needed to bind the dyad symmetry element onto a nitrocellulose filter relative to the amount of baculoEBNAl needed to protect the AuaI site. This may be due to the requirement for baculoEBNAl to bind to only one particular site

Epstein-Barr Virus Nuclear Antigen 1 7825

I n vivo the dyad symmetry element is accompanied by the family of repeats within oriP which may affect the interaction of EBNAl with the dyad symmetry element. Therefore we examined the interaction of baculoEBNA1 with the dyad symmetry element in the complete oriP sequence. Baculo- EBNAl was incubated with oriP labeled with 32P at the end near the dyad. Just prior to filtration through nitrocellulose, the family of repeats was separated from the dyad symmetry element by digestion with EcoRV (Fig. 8). For each assay an aliquot was removed prior to filtration, quenched with SDS, and analyzed in an agarose gel to confirm that EcoRV had completely separated the dyad from the oriP DNA. The results showed significant amounts of dyad symmetry element were retained onto nitrocellulose at lower levels of baculo- EBNAl (200 fmol and less) in the presence of the family of repeats than in their absence (Fig. 8, closed triangles). How- ever, further along in the titration, more baculoEBNA1 was required to bind the dyad on oriP than to bind the dyad alone. Complete retention onto nitrocellulose of the isolated family of repeats and dyad symmetry fragments required 300 and 500 fmol of baculoEBNA1, respectively. Hence, it seems a paradox that even 800 fmol of baculoEBNA1 was not sufficient to retain onto nitrocellulose more than half of the dyad fragment when it was within the context of oriP. Possible explanations include the following. The presence of the family of repeats may destabilize the interaction of baculoEBNA1 with the dyad. A less stable complex of baculoEBNA1 with the dyad may assemble in the presence of the family of repeats. The nonessential region of oriP between the family of repeats and dyad symmetry element may influence the nitrocellulose binding assay, or the presence of the dyad may cause more cooperative binding of baculoEBNA1 to the family of repeats, effectively decreasing the availability of baculo- EBNAl for binding the dyad.

DISCUSSION

In this report we describe the overproduction of EBNA1, the viral encoded protein which binds the latent phase origin (oriP) of EBV, in the baculovirus system and its purification to homogeneity. Like EBNAl from latently infected B cell lines (4, 29, 33), the baculoEBNA1 bound tightly to oriP, arrested replication forks within or near the oriP family of repeats,6 and was phosphorylated on serine residues. Since phosphorylation can modulate protein function (30-32), it seems likely that initiation of replication from oriP will be regulated by phosphorylation of EBNA1.

The palindromic nature of each EBNAl consensus site suggests that EBNAl binds its DNA site as a dimer. Indeed the baculoEBNA1 appeared to be a dimer in solution and the stoichiometry of 56 baculoEBNA1 molecules per 24 EBNAl binding sites in the oriP sequence was consistent with EBNAl binding its site as a dimer as predicted (34).

Increasing evidence suggests replication initiates within the dyad symmetry element of oriP (33,35). Replication initiation in the dyad is greatly stimulated by the family of repeats (35). One mechanism by which the repeats might activate the dyad is by altering the interaction of EBNAl with the dyad symmetry element. The nitrocellulose filter binding assay sug- gested that the family of repeats reduced the concentration of baculoEBNA1 required to initiate binding to the dyad symmetry element. Provided the interaction of EBNAl with

in the dyad symmetry element to protect it from AuaI, whereas retention of the dyad onto nitrocellulose may require baculoEBNA1 bound to another site or multiple baculoEBNA1 molecules bound to multiple sites in the dyad symmetry element.

V. Dhar and C. L. Schildkraut, personal communication.

the dyad symmetry element is important for the initiation of replication from oriP, then the stimulation of dyad binding by the family of repeats at low EBNAl concentration may be one mechanism by which the repeats enhance replication from oriP.

EBNAl is essential for latent EBV replication, yet the precise biochemical function of EBNAl remains elusive. The baculoEBNA1 protein should prove useful in biochemical assays to analyze the mechanism by which EBNAl activates oriP to function as an origin of replication, a plasmid maintenance element, and a transcriptional enhancer (21). We find no ATPase (or other nucleoside triphosphatase), helicase, ligase, topoisomerase, DNA polymerase, exonuclease, or endonuclease activities associated with baculoEBNA1. The absence of ATPase and helicase activity suggests EBNAl plays a different role in replication than the large T antigen of SV40. It is always possible, however, that the true activity of EBNAl will only be revealed upon binding other proteins or by modification at a specific site(s). Furthermore, we cannot exclude the possibility that, although the six amino-terminal amino acids and glycine-alanine repeat region of EBNA1, lacking in baculoEBNAl, are nonessential for EBNAl function in vivo (21), they may affect the biochemical activity of EBNAl in uitro.

Elucidation of the precise role of EBNAl in replication and the mechanism(s) of replication control at oriP would be greatly facilitated by development of an in vitro system ca- pable of initiating replication from oriP. We hope that the availability of EBNAl produced in baculovirus will enable the development of such a cell-free replication ~ y s t e m , ~ in a manner analogous to that by which supplementation of cell extracts with SV40 large T antigen enabled the development of an in vitro system for replication from the SV40 origin (36- 38).

Acknowledgments-We would like to thank Drs. Ora Rosen, Susan Wente, David Russell, and Max Summers for the gift of pVL941-SW and for their expert advice on the baculovirus expression system. We are also grateful to Bill Sugden for providing the p205 plasmid and Carol Prives for the methionine-free Graces media. We would also like to thank Dr. Monika Lusky, Dr. Paula Traktman, Rachel Rem- pel, Maija Skangalis, Rene Onrust, Todd Stukenberg, and Patricia Studwell for help and advice throughout the course of this work.

REFERENCES

1. Zur Hausen, H. (1981) DNA Tumor Viruses (J. Tooze, ed) pp. 747-795, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

2. Lindahl, T., Adams, A., Bjursell, G., Bornkamm, G. W., Kashaka- Dierich, C., and Jehn, U. (1976) J. Mol. Biol. 102, 511-530

3. Yates, J. L., Warren, N., and Sugden, B. (1985) Nature 313,

4. Jones, C. H., Hayward, S. D., and Rawlins, D. R. (1989) J. Virol.

5. Reisman, D., Yates, J., and Sugden, B. (1985) Mol. Cell. Biol. 5 ,

6. Rawlins, D. R., Milman, G., Hayward, S. D., and Hayward, G. S.

7. Krysan, P. J., Haase, S. B., and Calos, M. P. (1989) Mol. Cell.

8. Chittendon, T., Lupton, S., and Levine, A. (1989) J. Virol. 63,

9. Reisman, D., and Sugden, B. (1986) Mol. Cell. Biol. 6, 3838-3846

812-815

63, 101-110

1822-1832

(1985) Cell 42,859-868

Biol. 9, 1026-1033

3016-3025

10. Sugden, B., and Warren, N. (1989) J Virol. 63, 2644-2649 11. Miller, L. K. (1988) Annu. Reu. Microbiol. 42, 177-199 12. Luckow, V. A., and Summers, M. D. (1988) Biotechnology 6,47-

13. Murphy, C. I., Weiner, B., Bikel, I., Piwnica-Worms, H., Bradley,

The EBNAl recombinant baculovirus (AcMNPV-EBNA1) will

55

be provided upon request.


M. K., and Livingston, D. (1988) J. Virol. 62,2951-2959 14. Luckow, V. A., and Summers, M. D. (1989) Virology 170,31-39 15. DuBridge, R. B., Tang, P., Hsia, H. C., Leong, P.-M., Miller, J.

16. Summers, M. D., and Smith, G. E. (1987) Tex. Agric. Exp. Stn.

17. Kadonaga, J. T., and Tjian, R. (1986) Proc. Natl. Acad. Sci. U. S.

18. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 19. Cooper, J. A., Sefton, B. M., and Hunter, T. (1983) Methods

20. Laemmli, U. K. (1970) Nature 227,680-685 21. Yates, J. L., and Camiolo, S. M. (1988) Cancer Cells 6, 197-205 22. Tsuchihashi, Z., and Kornberg, A. (1989) J. Biol. Chem. 264,

23. Matson, S. W. (1986) J. Biol. Chem. 261, 10169-10175 24. Goetz, G. S., Dean, F. B., Hurwitz, J., and Matson, S. W. (1988)

25. Shaw, S. B., and Tegtmeyer, P. (1981) Virology 116,88-96 26. Klausing, K., Scheidtmann, K.-H., Baumann, E. A., and Knip-

H., and Calos, M. P. (1987) Mol. Cell. Bwl. 7, 379-387

Bull. 1665, 27-31

A. 83,5889-5893

Enzymol. 99,387-402

17790-17795

J. Biol. Chem. 263, 383-392

pers, R. (1988) J. Virol. 62,1258-1265

27.

28.

29. 30.

31. 32.

33. 34.

35.

36.

37.

38.

Siegel, L. M., and Monty, K. J. (1966) Biochim. Biophys. Acta

Baer, R., Bankier, A. T., Biggin, M. D., Deininger, P. L., Farrell, P. J., Gibson, T. J., Hatfull, G., Hudson, G. S., Satchwell, S. C., Siguin, C., Tuffnell, P. S., and Barrell, B. G. (1984) Nature

112,346-362

310,207-211 Hearing, J. C., and Levine, A. J. (1985) Virology 145, 105-116 Donaldson, R. W., and Gerner, E. W. (1987) Proc. Natl. Acad.

Gould, K. L., and Nurse, P. (1989) Nature 342,39-45 McVey, D., Brizuela, L., Mohr, I., Marshak, D. R., Gluzman, Y.,

Gahn, T. A., and Schildkraut, C. L. (1989) Cell 58,527-535 Ambinder, R. F., Shah, W. A., Rawlins, D. R., Hayward, G. S.,

Wysokinski, D. A., and Yates, J. L. (1989) J. Virol. 63, 2657-

Li, J. J., and Kelly, T. J. (1984) Proc. Natl. Acad. Sci. U. S. A.

Stillman, B. W., and Gluzman, Y. (1985) Mol. Cell. Biol. 6, 2051-

Wobbe, C. R., Dean, F., Weissbach, L., and Hurwitz, J. (1985)

Sci. U. S. A . 84,759-763

and Beach, D. (1989) Nature 341,503-507

and Hayward, S. D. (1990) J. Virol. 64,2369-2379

2666

81,6973-6977

2060

Proc. Natl. Acad. Sci. U. S. A. 82,5710-5714

the of vol. 266, no. 12, issue of april 25, pp. 7819-7826...

Documents