Int. J . Cancer: 44, 560-564 (1989) 0 1989 Alan R. Liss, Inc.

REPLICATION OF LATENT EPSTEIN-BARR VIRUS GENOMES IN NORMAL AND

Publication of the International Union Against Cancer Publication de I‘Union Internationale Contra le Cancer

MALIGNANT LYMPHOID CELLS Alice ADAMS, Tamara C. Pozos and Helen V. PURVEY Department of Medical Microbiology and Immunology, University of Minnesota, Duhth, MN 55812, USA.

DNA replication of 2 human lymphoid cell lines (U296 and Raji), latently infected with the Epstein-Barr virus, has been compared using a density transfer approach. Typical of non- malignant lymphoblastoid cells, U296 cells divided once in bromodeoxyuridine-supplemented medium to form hybrid but not heavy-density host DNA. Replication of the intracel- lular Epstein-Barr virus DNA was selectively inhibited in these cells with only 15% of the viral genomes duplicating once to form hybrid-density viral DNA. However, some heavy-density viral DNA was formed in the U296 cells and DNA synthesis can thus initiate again on newly duplicated viral genomes in cells that have traversed only a single S phase. These re$ulb contrast strongly with observations con- cerning the Burkitt-lymphoma-derived cell line. Lymphoma cells are not growth-inhibited and most of the latent Epstein- Barr virus genornes of the Raji line replicated once, and only once, in succes$ive S phases. While the majority of the 50 Epstein-Barr virus genomes of both the Raji and U296 cell lines are maintained as extra-chromosomal DNA plasmids, the control of their duplication is distinctly different in the respective malignant and non-malignant host cells.

The Epstein-Barr virus (EBV), one of 6 known human herpesviruses, can cause infectious mononucleosis (IM) and is implicated in the etiology of Burkitt’s lymphoma (BL) and nasopharyngeal carcinoma (NPC) (Klein, 1975). Both BL and NPC tumor cells contain multiple copies of the complete EBV genome, from which only defined regions are expressed in vivo (Dambaugh et al., 1979; Raab-Traub et al., 1983). EBV- DNA-positive lymphoid cell lines have been established from BL biopsy specimens. Lymphoid cell lines can also be isolated from IM patients as well as from healthy subjects latently infected with EBV (Nilsson, 1979). Like the in vivo BL tumor cells, most cell lines, irrespective of origin, harbor multiple copies of the EBV genome which, in the case of “non- producer” cell lines, are maintained as extrachromosomal, cir- cular DNA plasmids in the absence of productive virus repli- cation (Adams, 1979).

The BL-derived “lymphoma” lines differ from the so-called “lymphoblastoid” cell lines of non-malignant origin in a num- ber of properties, including morphology, chromosome abnor- malities, tumorigenicity in nude mice and growth response in various media (Nilsson and PontCn, 1975). In addition to the criteria used by Nilsson and PontCn to define the lymphoma and lymphoblastoid phenotypes, the 2 cell types differ in their growth response to 5’ bromodeoxyuridine (BUdR) (Henderson and Strauss, 1975). Thus, lymphoblastoid cells proceed through only a single cell doubling in medium that is 3 x lop5 M BUdR while lymphoma cells are not noticeably growth- inhibited. Replacement of thymidine residues by the bromi- nated base analogue increases DNA density, and density trans- fer can be used to monitor DNA replication. As expected from their growth response, only hybrid-density DNA, formed dur- ing a single cycle of semi-conservative DNA replication, was recovered from the lymphoblastoid cells, while heavy-density DNA, in which over 90% of the thymidine residues were re- placed by the brominated analogue, could be isolated from lymphoma cells.

The mode of latent EBV DNA replication has previously been investigated using Raji, a BL-derived line, as the exper- imental host cell. For Raji, the @mount of viral DNA doubles

early during the cellular S phase (Hampar et al. , 1974), with the plasmid forms replicating via theta structured intermediates (Gussander and Adams, 1984). When the density transfer ap- proach is used to follow EBV DNA replication, it appears that each of the 50 EBV DNA plasmids of Raji replicate only once in a given S phase (Adams, 1987).

Most EBV gene functions, including the virus-encoded DNA polymerase, are not expressed in latently infected cells (Ooka et al., 1986) and the host cell probably provides most of the functions needed for plasmid duplication. Observations made on EBV DNA replication in Raji may, therefore, not be representative of what occurs in non-malignant cells. In this report the effect of BUdR on latent EBV DNA replication of Raji and U296, a normal lymphoblastoid cell line, is com- pared.

MATERIAL AND METHODS

Origin and growth of cell lines Raji, established from a BL tumor (Pulvertaft, 1965) was the

reference lymphoma line. The U296 line, derived from periph- eral blood of a patient without malignant disease (Nilsson, 197 l), was selected as the representative lymphoblastoid line. Both lines exhibit the phenotypic properties associated with their respective malignant and non-malignant origins. The U296 line was chosen for these DNA replication studies be- cause, like Raji, it maintains about 50 extrachromosomal EBV genomes in the total absence of any evidence of productive virus replication.

Both prototype lines were propagated at 37°C in a 5% CO, atmosphere as stationary suspension cultures in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 units penicillidml and 100 pg streptomycidml. Two weeks prior to their use in a density transfer experiment, the cultures were maintained in log phase by periodic dilution in complete me- dium to which 50 pg tylosidml had been added as a preventive measure to preclude aberrant BUdR incorporation known to result from Mycoplasma contamination (Henderson and Strauss, 1975). Density labelling was initiated by the addition of 1/100 vol. of 3 m~ BUdR and all subsequent operations were performed under subdued yellow light to retard formation of undesirable photo products (Wacker et al., 1962).

Cultures were monitored for induction of the lytic EBV cy- cle by direct immunofluorescence (Klein et al., 1971). Fewer than 0.02% early antigen (EA)-positive cells were detected in any of the cultures when cells were collected for DNA analy- sis.

Isolation and analysis of DNA At various times after the addition of BUdR, cells were col-

lected, washed with phosphate-buffered saline (PBS) and con- centrated cell suspensions in PBS were lysed with detergent. DNA was extracted by one of 2 methods. The standard method included enzymatic digestions to remove both RNA and pro- teins, followed by phenol extraction and ethanol precipitation

Received: May 24, 1989

REPLICATION OF LATENT EB VIRUS GENOMES 56 1

E : 12- 0 d

'D 8 -

n I 6- a a 4 -

a 2 -

10 ~

a, N ._ ._ L

>

2

0 - N m Y

0 7

of the DNA (Adams el al., 1973). Precipitated DNA was dis- solved in T.E. (0.01 M Tris HCI, 0.001 M EDTA) PH 8 prior to analysis. With the alternative method, DNA was isolated in the continued presence of 1~ salt without RNase treatment or exposure to organic solvents (Gussander and Adams, 1984). Solid CsCl was added to give a final density of 1.73 g/cm3 and the material was passed 6 times through a 1-inch long, 26- gauge syringe needle to shear the DNA to a mean molecular weight of lo7. Samples containing 10-20 pg DNA/ml were centrifuged for 60 hr at 20°C in the Beckman Ti 50.1 or Ti 50.2 rotor at a maximum force of 140,000 g. About 40 fractions of equal volume were collected from the bottom of the tube and diluted to 600 pl prior to A,, measurements.

Nucleic acid hybridization The DNA in an appropriate aliquot of selected gradient frac-

tions was fixed either to individual 13-mm nitrocellulose filter discs or to a single slot of a 24-well slot-blot filter. Viral DNA sequences were quantitated by hybridization with 32P-labelled RNA probes (Lindahl et al., 1976). RNA complementary to the entire EBV genome was prepared by transcription of B95- 8 virion DNA with E . coli RNA polymerase (holoenzyme, Boehringer, La Jolla, CA). RNA probes to defined regions of the EBV genome were prepared using plasmids consisting of B a d 1 and EcoRI fragments of B95-8 DNA cloned in either the pUC18 or pUC19 vector as template. The recombinant DNA plasmids were first linearized with an appropriate restric- tion endonuclease, and run-off transcripts were initiated from either the T7 or T3 promoters, located adjacent to opposite ends of the EBV DNA inserts, with the corresponding, bacte- riophage, RNA polymerase according to the protocols supplied with the enzymes (BRL, Gaithersburg, MD).

RESULTS

The growth response of U296 to 3 X M BUdR was identical to that reported for 10 other lymphoblastoid cell lines by Henderson and Strauss (1975). Thus, the cell density of treated cultures did not increase more than 2-fold during a 96-hr incubation period. In the absence of BUdR, U296 cells have a generation time of 29 hr with the cell density of control cultures increasing 8-fold in 4 days. Preliminary DNA analysis conf i ied that the cells do not traverse more than a single S phase in the presence of BUdR. It was concluded from these results that U296 is representative of the lymphoblastoid cell type and a suitable prototype for characterization of latent EBV DNA synthesis of a non-malignant cell.

Figure 1 shows the density distribution of DNA isolated from Raji (a) and U296 (b), grown in 3 X lop5 M BUdR medium. Because the generation time of Raji is 5-6 hr shorter than that of U296, the 36-hr Raji and 48-hr U296 samples are compared. At these times, roughly equivalent numbers of cells of the respective control cultures had traversed more than one complete S phase. If DNA synthesis had not been affected by BUdR, similar amounts of hybrid- and heavy-density DNA should have been formed. Comparison of the density gradient profiles reveals clear differences in both host and viral DNA synthesis of the representative lymphoma and lymphoblastoid cell lines.

The A,, curve of Raji shows that both hybrid- and heavy- density cellular DNAs were formed during the 36-hr incubation period (Fig. la). Moreover, the amount of EBV DNA associ- ated with the heavy (fractions 5-10), hybrid (fractions 11-18), and light (fractions 19-25) density peaks parallels that of total DNA. The cell and viral DNAs of the various density forms reach a peak in different fractions due to differences in base composition. Thus, the intrinsic density of EBV DNA (58% GC) is greater than that of host DNA (43% GC). However, more thymidine residues are available for BUdR replacement

141 , I 4

5 10 15 20 25 30

Fraction N u m b e r

n

5 10 15 20 25 30

Fraction Number

1 2

- 1 0

-0 8

- 0 6

0 4

0 2

0 0

0 W

P

0 4

0 3

0

0 2 P

0 1

0 0

FIGURE 1 - CsCl density gradients of DNA from cells grown in the presence of 3 x lo-' M BUdR. Sheared DNAs, isolated by standard methods employing proteinase K and RNase treatments followed by phenol extraction, were mixed with solid CsCl to give an R.I. = 1.4020 and centrifuged for 60 hr. In (a) the DNA from 48 X lo6 Raji cells incubated in BUdR medium for 36 hr was cen- trifuged in the Beckman Ti 50.2 rotor. In (b) the DNA of 12 X lo6 U296 cells, grown for 48 hr, was fractionated using the Beckman Ti 50.1 rotor. For Raji, 200-kl aliquots of each fraction were alkaline- denatured and the DNA was fixed to individual, 13-mm nitrocellulose discs, then, following hybridization, the amount of EBV DNA was quantitated by scintillation counting. Only 40 ~1 of each U296 fraction was used per well of the 24-well slot-blot filter and after exposure of the probed filter to Kodak X-OMAT AR, X-ray film, the relative amount of hybridization was determined from densitometer tracings. The profile of total DNA (C-.), was determined by AZm measure- ments and (A--A) indicates the distribution of EBV DNA sequences as measured by hybridization with 3ZP-EBV cRNA.

in cellular DNA and the separation of the 2 DNAs decreases in the hybrid density region. When fully replaced by BUdR, cel- lular DNA is more dense than the corresponding form of EBV DNA. It is estimated that 53% of the EBV DNA plasmids and 54% of the total DNA, present in the Raji cells at the time of BUdR addition, duplicated to form the respective hybrid den- sity species at 36 hr.

The results for the U296 line are shown in Figure lb. In contrast to Raji, no trace of A,,, absorbing material is seen in fractions 5 and 6, where heavy-density chromosomal DNA would be expected to band. Though somewhat less U296 DNA has replicated to form hybrid-density material, this should not have precluded the recovery of at least some heavy-density host DNA. Thus, BUdR-treated U296 cells are inhibited from traversing more than a single S phase.

562 ADAMS ET AL

Comparison of the hybridization profiles of Figure 1 indi- cates significant differences in the mode of latent EBV DNA replication in the 2 cell lines. EBV DNA replication is selec- tively inhibited in the BUdR-treated U296 culture with only about 9% of the EBV DNA initially present having been du- plicated at a time when 43% of the host chromosomal DNA was transferred to the hybrid density. Moreover, a shoulder of more dense EBV DNA sequences is seen in fractions 9-12 of the U296 gradient (Fig. lb), indicating that DNA synthesis can again initiate on the newly duplicated, hybrid-density EBV DNA sequences. Evidence of some heavy-density EBV DNA was first noted in the U296 DNA sample isolated at 36 hr, with the amount of this material increasing from a level correspond- ing to half that of the hybrid EBV DNA peak at 48 hr to nearly equal amounts of the 2 forms at 96 hr. Assuming the dense material to be equivalent to fully BUdR-substituted duplex DNA, it appears that viral DNA synthesis preferentially ini- tiates several times on only a few of the intracellular EBV genomes of the U296 line and that, in contrast to Raji, this may occur within a single S phase.

The results obtained with DNA samples isolated at other times are summarized in Figure 2. From hybridization and A,, profiles similar to those presented in Figure 1, the areas of the hybrid- and light-density peaks of both viral and total DNA were estimated and the percentage of DNA replicating at the various times calculated. It is noted that, once formed, the absolute amount of hybrid-density DNA will not decrease even if, as is the case with Raji, some of these sequences replicate further to generate heavy-density DNA.

About 20% of the total DNA present in Raji cells at the time of BUdR addition did not replicate prior to the culture entering stationary phase and some cells apparently did not cycle in the BUdR medium, even though 9&95% of the cells of the initial inocula appeared viable by exclusion of Trypan blue. How-

20 40 60 80 100

Time (hours)

FIGURE 2 - Percentage of total and EBV DNA present at the time of BUdR addition that replicated at various times. The absolute amounts of hybrid and light (unreplicated) density DNAs were esti- mated and the percentage of initial DNA replicated was calculated as:

YZ amount hybrid DNA

amount light DNA + Y2 amount hybrid DNA % = x loo.

For Raji (0, 0) the 12- and 66-hr samples were from one batch of BUdR-treated cells and the 24-, 36-, 48- and 96-hr samples were from a second batch of cells. Likewise, with U296 (A, A) the 4-, 9- and 64-hr points were from one treated culture and the lo-, 24-, 35-, 48- and 96-hr samples were from another. The closed symbols and solid lines indicate results for cellular DNA and the open symbols and dashed lines represent EBV DNA.

ever, the amount of host and EBV DNA replicating at various times is roughly equivalent. Somewhat fewer EBV DNA mol- ecules appear to have replicated, but the difference seen in Figure 2 may, in part, be due to an under-estimation of the amount of hybrid EBV DNA. The separation between the light-, hybrid- and heavy-density viral DNA species is not as well defined as that for host DNA (Fig. la) and minor errors in estimating the relative concentrations of EBV DNA could eas- ily result in the lower plateau value shown. Still, it cannot be excluded that a few of the intracellular EBV genomes of Raji may escape duplication during a given S phase. However, no heavy-density DNA was detected prior to a second S phase (Adams, 1987) which is inconsistent with a predominantly random mode of template selection.

Only 50% of the cellular DNA sequences replicated in BUdR-treated U296 cultures where the cell density increased nearly 2-fold. Much of the light-density DNA must, therefore, come from cells that were in S or G, phase at the time of BUdR addition and were able to complete mitosis but unable to ini- tiate a new S phase. Hybridization indicates that only 15% of the EBV DNA of U296 cells replicates once to form hybrid- density material.

The hybridization data presented in Figures 1 and 2 were obtained with 32P EBV cRNA, a probe specific for the entire EBV genome. The U296 data do not, therefore, distinguish between whether only a few of the approximately 50 EBV genomes replicate or if only a limited region of each genome is duplicated. A number of identical slot blots were prepared and these were hybridized with different probes. Analysis of one DNA sample with probes recognizing sequences mapping to 3 scattered regions of the EBV genome (4-7, 66-69 and 138- 142 kbp) is shown in Figure 3. No particular sequence had preferentially replicated to form the limited amount of hybrid- density viral DNA. The small shifts in the peak position of the light-density DNAs are sequence-dependent and analogous dis- placements were seen on analysis of both Raji and U296 con- trol DNA gradients. Results with 3 additional probes, recog- nizing sequences between 1 3 4 7 , 155-160 and 160-167 kbp, were similar to those shown in Figure 3. Thus, hybrid DNA indicative of replication of 13 f 1% of the DNA initially present was found with all 6 probes. The cRNA probe indi- cated that 14% of the EBV DNA had replicated and, while the 69-138 kbp region of the EBV genome was not analyzed with individual probes, it is unlikely that any significant differences would have been observed. Analysis of other U296 gradients with EBV probes that recognize sequences within this region (84-88, 92-93, 102-104, 125-129 and 143-145 kbp) did not reveal any sequence(s) to be selectively duplicated. All probes tested hybridized with some heavier than hybrid density EBV DNA sequences but the values are low and the data are not sufficiently accurate to conclude whether particular sequences may accumulate as fully BUdR-substituted DNA.

The DNA preparation analyzed in Figure 3 was isolated by the high-salt method and the high background of A,, material is due to impurities which are removed by the phenol extraction and ethanol precipitation steps of the alternative purification procedure used in Figure 1. Moreover, the DNA of Figure 3 was not treated with RNase and it is likely that the small peak of A,,,labsorbing material seen in fractions 7 and 8 is RNA. No equivalent heavy-density material was detected with an RNase treated DNA sample isolated at 96 hr. The hybridization profiles presented in Figures 3 and l b are similar and it is unlikely that the reduced level of EBV DNA replication seen is due to selective loss or degradation of viral DNA during the isolation procedure.

DISCUSSION

Latent EBV DNA replication of a typical lymphoblastoid

563 REPLICATION OF LATENT EB VIRUS GENOMES

10 20 30 40

Fraction Number

1 .o

0 .8

0 . 8 0 (0

2 0 .4

0.2

0.0

FIGURE 3 - Analysis of CsCl density gradient fractionated U296 DNA with different hybridization probes. DNA, from cells grown for 64 hr in BUdR medium, was isolated by the high-salt method and fractionated on an 18 ml CsCl gradient in the Beckman Ti 50.2 rotor. A,, (0-0) was determined and 30-p.1 aliquots of individual fractions were fixed to individual slots of several filters which were then hybridized with different probes. The probes used consisted of the EcoRI J (A--A) fragment of B95-8 DNA cloned in the pUC18 vector and the BamHI U (V---V) and Bam HI b (G-0) fragments cloned in pUC19 vectors. Hybridization data were quantitated from densitometer tracing of autoradiographs of the slot-blot filters.

cell line is more impaired by BUdR than host DNA synthesis. All EBV genomes are located in the nucleus where the sub- strate pools are presumably the same as for replication of host DNA. Moreover, few viral genes are expressed and most of the enzymatic functions needed for EBV DNA duplication must be shared. The mechanism by which viral DNA synthesis is se- lectively repressed probably involves altered template recog- nition by only one or a few proteins. Likely candidates for such control substances would be one or more of the half-dozen EBV-encoded nuclear antigens or EBNA proteins expressed by latently infected cells (Reedman and Klein, 1973).

The EBNA-1 protein is essential for plasmid maintenance (Yates et al., 1985) and altered EBNA-1 binding to ori P , the proposed origin of latent EBV DNA replication, could explain the U296 results. The ori P locus consists of 2 AT-rich se- quences mapping to the BamHI C fragment of the EBV ge- nome. When in the cis configuration, these 2 short (600 and 114 bp) sequences are sufficient for maintenance of recombi- nant DNA plasmids in EBNA-1 positive cells (Reisman et al., 1985). The EBNA-1 protein binds preferentially to the 30-bp repeat units of the longer, 33% GC, region of ori P (Rawlins et al., 1985). Only when the 20 30-bp repeat units, arranged in tandem, are saturated will EBNA-1 bind to a region of dyad symmetry within the second, 43% GC, 114-bp sequence of ori P and initiate DNA synthesis. This model for the control of EBV plasmid copy number predicts random initiation of DNA synthesis on individual EBV DNA molecules. Had purely ran- dom template selection occurred in the BUdR-treated cells, the amounts of heavy- and light-density EBV DNA should, after one cell doubling, have been equal to each other, but less than that of hybrid-density material. This was not observed for ei- ther cell in Figure 1. However, incorporation of BUdR into one strand of DNA can affect protein binding (Lin and Riggs, 1972; Rutter et al., 1973) and the U296 data could be ex- plained if a limiting amount of EBNA-1 were to preferentially bind to the BUdR substituted ori P locus of early replicating

genomes. Further initiation of DNA synthesis on unreplicated molecules would be suppressed because of insufficient EBNA, while newly duplicated, hybrid-density genomes, with bound EBNA-1, would tend to replicate again. BUdR substitution is known to increase the thermal stability of DNA (Lapeyre and Bekhor, 1974) which might reduce the frequency at which DNA synthesis initiates on hybrid density genomes saturated with EBNA-1. The net result would be both a reduced level of EBV DNA replication and a preference for DNA synthesis to initiate on newly duplicated molecules.

Normal human B lymphocytes do not grow in continuous cell culture unless latently infected with EBV (Pope, 1979) and viral functions are thought to be essential for growth transfor- mation of these cells. Another EBNA protein, EBNA-2, is implicated as one of the proteins needed for B-cell immortal- ization (Dambaugh et al., 1986). If EBV gene expression is required for the cells to cycle, then inhibition of EBV DNA replication, with a resulting decrease in genes coding for viral functions, could cause the growth inhibition observed with lymphoblastoid cell lines.

The intracellular EBV genomes of several BL-derived lym- phoma lines are deleted for the EBNA-2 region. Moreover, rare, EBV-negative, BL lymphoma lines have been isolated (Klein et al., 1974) and EBV genetic information may not be essential to the immortality of malignant cells. All Burkitt cells have specific chromosome translocations which invariably in- volve the c-myc locus (Lenoir, 1986) and altered c-myc ex- pression may supersede the need for EBV-encoded immortal- ization functions.

The replication of latent EBV genomes of Raji was non- random and unaffected by BUdR substitution, suggesting that the control of viral DNA synthesis in lymphoma cells may be influenced by host factors. The c-myc gene codes for a nuclear protein with DNA binding properties (Hann and Eisenman, 1984) which could augment the interaction of EBNA-1 with

564 ADAMS ET AL.

the ori-P locus. Human c-myc may substitute for T antigen in the initiation of SV40 DNA. Thus, SV40 origin-containing plasmids replicate in the absence of viral T antigen in both Raji and HL-60 cells (Iguchi-Ariga et al., 1988). HL-60 is an EBV- negative, promyelocytic leukemia-derived line which ex- presses high levels of the myc oncogene. Large (250 kbp) extra-chromosomal DNA plasmids present in a subclone of HL-60 cells replicate in a similar fashion to the EBV genomes of Raji (Von Hoff et al., 1988). Thus, after one generation in BUdR medium, only hybrid-density plasmid DNA forms were recovered and, as with Raji, replication of the 250-kbp plas-

REFERE

ADAMS, A, , The state of the virus genome in transformed cells and its relationship to host cell DNA. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Burr virus, pp. 155-183, Springer, Berlin (1979). ADAMS, A,, Replication of latent Epstein-Barr virus genomes in Raji cells. J. Virol., 61, 1743-1746 (1987). ADAMS, A,, LINDAHL, T. and KLEIN, G., Linear association between cellular DNA and Epstein-Barr virus DNA in a human lymphoblastoid cell line. Proc. nat. Acad. Sci. (Wash.), 70, 1888-1892 (1973). DAMBAUGH, T., HENNESSY, K., FENNEWALD, S. and KIEFF, E., The virus genome and its expression in latent infection. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Burr virus, recent advances, pp. 1345, J. Wiley, New York (1986). DAMBAUGH, T., NKRUMAH, F.K., BIGGAR, R.J. and KIEFF, E., Epstein- Ban virus RNA in Burkitt tumour tissue. Cell, 16, 313-322 (1979). GUSSANDER, E. and ADAMS, A,, Electron microscopic evidence for rep- lication of circular Epstein-Barr virus genomes in latently infected Raji cells. J. Virol., 52, 54F556 (1984). HAMPAR, B., TANAKA, A,, NONOYAMA, M. and DERGE, J., Replication of the resident repressed Epstein-Barr virus genomes during the early S phase (S-I period) of non-producer Raji cells. Proc. nat. Acad. Sci. (Wash.), 71, 631-633 (1974). HANN, S.R. and EISENMAN, R.N., Proteins encoded by the human c-myc oncogene: differential expression in neoplastic cells. Molec. Cell. Biol., 4, 24862497 (1984). HENDERSON, E.E. and STRAUSS, B., Differences in the incorporation of bromodeoxyuridine by human lymphoblastoid cell lines. Cell, 5 , 38 1-387 (1975). IGUCHI-AIUGA, S.M.M., ITANI, T., YAMAGUCHI, M. and ARIGA, H., c-myc protein can be substituted for SV40 T antigen in SV40 DNA repli- cation. Nucl. Acid Res., 15, 4889-4899 (1988). KLEIN, G., The Epstein-Barr virus and human malignancies. N. Engl. J. Med., 293, 1353-1357 (1975). KLEIN, G., GERGELY, L. and GOLDSTEIN, G., Two-colour immunofluo- rescence studies on EBV-determined antigens. Clin. exp. Immunol., 8, 593402 (1971). KLEIN, G., LINDAHL, T., JONDAL, M., LEIBOLD, W., MENEZES, J., NILS- SON, K. and SUNDSTROM, C., Continuous lymphoid cell lines with char- acteristics of B cells (bone-marrow-derived), lacking the EBV genome and derived from three human lymphomas. Proc. nat. Acad. Sci. (Wash.), 71, 3283-3286 (1974). LAPEYRE, J.-N. and BEKHOR, I., Effects of 5-bromo-2’-deoxyundine and dimethyl sulfoxide on properties of structure of chromatin. J . mol. Biol., 89, 137-162 (1974). LENOIR, G.M., Role of the virus, chromosomal translocations and cellular oncogenes in the etiology of Burkitt’s lymphoma. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Barr virus, recent advances, pp. 183- 205, J. Wiley, New York (1986). LIN, S.-Y. and B c c s , A.D., lac operator analogues: bromodeoxyuridine

mids paralleled host (HL-60) DNA synthesis during subse- quent S phases. Malignant lymphoid cells appear to control the semi-conservative replication of extrachromosomal plasmids in such a way that each molecule divides once, and only once, during the cell cycle.

ACKNOWLEDGEMENTS

This work was supported by the University of Minnesota Graduate School and grant MV-37 1 from the American Cancer Society.

.NCES

substitution in the lac operator affects the rate of dissociation of the lac repressor. Proc. nat. Acad. Sci. (Wash.), 69, 2574-2576 (1972). LINDAHL, T., ADAMS, A., BJURSELL, G., BORNKAMM, G.W., KASCHKA- DIERICH, C. and JEHN, U., Covalently closed circular duplex DNA of Epstein-Ban virus in a human lymphoid cell line. J. mol. Biol., 102, 511-530 (1976). NILSSON, K., High-frequency establishment of human immunoglobulin- producing lymphoblastoid lines from normal and malignant lymphoid tis- sue and peripheral blood. Int. J . Cancer, 8, 4 3 2 4 2 (1971). NILSSON, K., The nature of lymphoid cell lines and their relationship to virus. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Burr virus, pp. 225-281, Springer, Berlin (1979). NILSSON, K. and PONTBN, J., Classification and biological nature of es- tablished human hematopoietic cell lines. Int. J . Cancer, 15, 321-341 (1975). OOKA, T., LENOIR, G.M., DECAUSSIN, G., BORNKAMM, G.W. and DAILLIE, J., Epstein-Ban virus-specific DNA polymerase in virus non- producer Raji cells. J. Virol., 58, 671-675 (1986). POPE, J.H., Transformation by the virus in virro. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Burr virus, pp. 205-223, Springer, Ber- lin (1979). FLJLVERTAFT, R.J.V., A study of malignant tumors in Nigeria by short term tissue culture. J. clin. Path., 18, 261-273 (1965). RAAB-TRAUB, N., HOOK, R., YANG, C.-S., HENRY 11, B. and PAGANO, J.S., Epstein-Barr virus transcription in nasopharyngeal carcinoma. J. Virol., 48, 58@590 (1983). RAWLINS, D.R., MILMAN, G., HAYWARD, S.D. and HAYWARD, G.S., Sequence-specific DNA binding of the Epstein-Barr virus nuclear antigen (EBNA-I) to clustered sites in the plasmid maintenance region. Cell, 42, 859-868 (1985). REEDMAN, B.M. and KLEIN, G., Cellular localization of an Epstein-Barr virus-associated complement-fixing antigen in producer and non-producer lymphoblastoid cell lines. Int. J. Cancer, 11, 499-520 (1973). REISMAN, D., YATES, J. and SUGDEN, B., A putative origin of replication of plasmids derived from Epstein-Ban virus is composed of two cis-acting components. Mol. cell. Biol., 5 , 1822-1832 (1985). RUTTER, W.J., PICTET, R.L. and MORRIS, P.W., Toward molecular mechanisms of developmental processes. Ann. Rev. Biochem., 42, 601- 646 (1973). VON HOW, D.D., NEEDHAM-VANDEVANTER, D.R., JUCEL, J., WINDLE, B.E. and WAHL, G.M., Amplified human myc oncogenes localized to replicating submicroscopic circular DNA molecules. Proc. mt. Acad. Sci. (Wash.), 85, 48044808 (1988). WACKER, A,, MENNINGMANN, H.D. and SZYBALSKI, W., Effects of “visible” light on 5-bromouracil-labelled DNA. Nature (Lond.), 196, 685-686 (1962). YATES, J.L., WARREN, N. and SUGDEN, B., Stable replication of plasmids derived from Epstein-Barr virus in various mammalian cells. Nature (Lond.), 313, 812-815 (1985).