CG3 (38) is a spontaneous EBV-carrying lymphoblastoid cell line that was derived from a Taiwanese patient with juvenile chronic myeloid leukemia (kindly provided by Dr Y.S. Chang). B95.8 (39) is a marmoset lymphoid cell line that was EBV-immortalized with a virus originating from an infectious mononucleosis patient, sampled in Africa. Raji (40), P3HR1 (41) and Daudi (42) are EBV-containing human lymphoid cell lines harboring viruses from Burkitt’s lymphoma patients. The cells were maintained in RPMI-1640 supplemented with 10% fetal calf serum (FCS), at 37°C in a humidified 5% CO₂ atmosphere. The 293 cell line (a human embryonal kidney cell line) (43) was grown in DMEM supplemented with 10% FCS. The 293-EBNA cells (i.e., 293 cells stably expressing B95.8 EBNA1; Invitrogen-Life Technologies, Carlsbad, CA, USA) were maintained in DMEM-10% FCS containing 250 μg/ml of Geneticin (G418; Gibco-Life Technologies, Carlsbad, CA, USA).

DNA was extracted from CG3 and 293-EBNA cells using a QIAamp DNA Mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions. The full-length EBNA1 gene and the sequence encoding its middle repeat region were PCR-amplified using the GC-RICH PCR System (Roche Molecular Biochemicals) according to the manufacturer’s instructions. The sequences of the utilized primers are shown in Table I, with the nucleotide numbering system used in this report being in accordance with that of Baer et al (44). For all PCR reactions discussed herein, the input DNAs were preheated at 95°C for 5 min prior to cycling, and the PCR products were subjected to a final extension for 10 min at 72°C. Briefly, 100 ng of DNA was combined with the provided reaction mixture (containing 0.5 M GC-RICH resolution solution) and subjected to PCR. To amplify the full-length EBNA1 gene, the F1/F2 primers were used along with the following reaction conditions: 30 cycles of 95°C for 1 min, 63°C for 2 min 30 sec and 72°C for 1 min. Nested PCR was required to successfully amplify the internal repeat sequences of EBNA1 (Fig. 1). Primer pairs GA1/GA2 and GA3/GA4 were used for the first and second round of PCR, respectively, under the aforementioned reaction conditions. The PCR products representing the full-length EBNA1 gene from CG3 and 293-EBNA cells were purified and cloned into a T-vector (pCRII-TOPO; Invitrogen) and three independent clones were sequenced (ABI377 DNA sequencing system; Perkin-Elmer/Applied Biosystems). From the clone, plasmids containing consensus sequences for pCRII-TOPO-CG3.1 and pCRII-TOPO-B95.8.1 were selected and used to generate EBNA1 eukaryotic expression vectors that expressed the CG3 and B95.8 EBNA1s, respectively (see below).

pCRII-TOPO-CG3.1 served as the template for PCR amplifying three fragments: the N fragment covering the N-terminus of EBNA1 (aa 1-90), and fragments FC and SC corresponding to two C-terminal regions downstream of the glycine-alanine repeat region (aa 354-641 and 459-641, respectively) (Fig. 1). The N, SC and FC fragments were amplified with primer pairs R1/R2, R3/R4 and R5/R4 (Table I and Fig. 1), respectively. For cloning purposes, restriction enzyme sites were included in the primers. The PCR products were then subjected to NdeI and BamHI double digestion followed by gel purification. The purified products were inserted in-frame at the C-terminus of the His-tag in pET-15b plasmids (Novagen) which had been pretreated with NdeI and BamHI. The resulting plasmids containing the N, FC and SC inserts were designated as pET_15b-hN, pET_15b-hFC and pET_15b-hSC, respectively, and their encoded fusion proteins were referred to as hN, hFC and hSC, respectively. These plasmids were used to transform BL21(DE3)pLysS (Novagen), and the fusion proteins were induced and purified as per the manufacturer’s instructions. Briefly, bacteria were grown to an OD₆₀₀ of 0.6, 1 mM IPTG was added, and the incubation was continued for 3 h via agitation. The bacterial pellet was collected by centrifugation at 2500 g for 15 min and the cells were lysed by freeze-thawing. The soluble recombinant EBNA1 proteins were purified with one-step metal chelation chromatography (Novagen). The purified hFC was then concentrated using Ultracel-3K (Millipore, Billerica, MA, USA). Equal amounts of the purified fractions were collected, resolved by SDS-PAGE and subjected to Coomassie blue staining. Western blot analysis, as previously described (45), was performed using an anti-His antibody (diluted 1:1,000; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA).

The backs of New Zealand White rabbits were subcutaneously immunized with 300 μg of hN or hSC in 0.5 ml complete Freund’s adjuvant (v/v 1:1) on day 0, and with the same amount of the corresponding protein in incomplete Freund’s adjuvant (v/v 1:1) (Sigma-Aldrich) on day 20 and were bled on day 27. The obtained antisera were designated Ab-N and Ab-SC, respectively. The animal experimental protocol was reviewed and approved by the Chang Gung University Animal Ethics Committee. The antibody responses were examined by western blot analysis using purified recombinant proteins and extracts prepared from EBV-containing cell lines and EBNA1-transfected 293 cells.

pCRII-TOPO-CG3.1 and pCRII-TOPO-B95.8.1 were digested with BamHI, filled in with Klenow (New England Biolabs, Beverly, MA, USA), and cleaved with XbaI to release the inserts of interest. Plasmid pCMV-EBNA (Clontech), which encoded EBNA1 of the B95.8 strain, was treated with PstI, filled in with Klenow, and digested with XbaI to remove the B95.8 EBNA1 gene, which served as the vector. The insert and vector were gel purified and ligated to generate the eukaryotic expression plasmids, pCMV-CG3EBNA1 and pCMV-B95.8EBNA1. These plasmids were transfected into 293 cells. Briefly, 5×10⁵ cells/wel) were seeded in 6-well (35-mm diameter) plates, grown for 18 h (~60–70% confluency), and then transfected with 4 μg DNA and 10 μl Lipofectamine (Invitrogen), according to the manufacturer’s instructions. At 48 h post-transfection, the cell extracts were collected, resolved by 10% SDS-PAGE, and subjected to western blot analysis (45) with rabbit Ab-SC antibodies. Proteins extracted from 293-EBNA (Invitrogen) and CG3 cells were used as positive controls. To generate 293 cells stably expressing EBNA1 proteins of B95.8 or CG3 origin, or the drug-resistant gene alone (as a control), EBNA1 expression vector- and empty vector-transfected cells were selected in the presence of 500 μg/ml Geneticin (G418; Gibco) and then maintained in medium containing 250 μg/ml Geneticin.

Stably transfected 293 cells (2×10⁵ cells/well) were seeded in 6-well plates. At 16 h post-seeding, the cells were washed twice with serum-free medium and then cultured in the absence and presence of FCS. Cell viability was determined at various time points using trypan blue exclusion. Proteins were also extracted at different time points and subjected to western blot analysis using antibodies against poly(ADP-ribose) polymerase (PARP) and actin (diluted 1:600 and 1:5,000, respectively; both from Millipore).

Size polymorphisms and sequence variations have been reported in EBNA1 proteins from different EBV isolates. CG3 is an EBV-containing lymphoblastoid cell line established from a Taiwanese patient with chronic myeloid leukemia (38). To the best of our knowledge, no previous study has reported the size and sequence of its EBNA1 protein. Consequently, PCR was amplified with the full-length EBNA1 gene from CG3 cells. As shown in Fig. 1, PCR reactions using primers F1/F2 and nested primers GA1/GA2 and GA3/GA4 were performed to amplify the full-length EBNA1 gene and its middle domain, respectively. DNA extracted from B95.8 cells was amplified as a control. The full-length gene and its middle region from B95.8 cells were previously predicted to be 2392 and 1177 bp in length, respectively (44). We obtained PCR products consistent with those predicted sizes (Fig. 2A and 2B, lane 2). The PCR products representing the full-length gene and the middle region of EBNA1 from CG3 cells (Fig. 2A and B, lane 1) appeared to be ~150 bp longer than those from B95.8 cells (Fig. 2A and B, lane 2). These data are consistent with previous studies that the molecular mass polymorphisms of EBNA1 across different EBV isolates reflect different numbers of internal glycine-alanine repeats (24–26).

Sequence analysis revealed that the N- and C-terminal regions of the CG3 EBNA1 gene did not contain any insertion or deletion, but harbored a series of nucleotide changes relative to the B95.8 clone. The deduced amino acid sequences of the CG3 clone relative to the B95.8 EBNA1 sequence are shown in Fig. 3. Based on the amino acid at position 487 and the characterized sequence variations at other positions, we determined that the CG3 EBNA1 belongs to the V-val subtype. We identified 4 (aa 16, 18, 20 and 85) and 12 (aa 364, 410, 411, 418, 439, 487, 499, 502, 524, 528, 533 and 594) amino acid substitutions relative to the B95.8 sequence in the N- and C-terminal regions, respectively, of the CG3 EBNA1. By contrast, only three amino acid substitutions (aa 85, 364 and 410) were identified between the CG3 EBNA1 and the EBNA1 isolated from NPC patients of Hong Kong (28), which was the first reported V-val EBNA1.

Recombinant CG3 EBNA1 proteins covering aa 1-90, 354-641 and 459-641 were expressed as His-tagged N-terminal fusion proteins in E. coli, and were designated as hN-, hFC- and hSC-EBNA1, respectively (Fig. 1). SDS-PAGE and Coomassie blue staining were used to examine the EBNA1 fusion proteins before (Fig. 4A–C, lane C) and after (lanes 1–4) their purification by metal chelation chromatography. The sizes of the EBNA1 fusions were as expected, except that doublet bands were observed for hN (Fig. 4A, lanes 2 and 3). The molecular nature of these doublets remains unknown. As shown in Fig. 4D, fraction 2 of hN (lane 1), fraction 3 of hSC (lane 2), and concentrated pooled hFC (fractions 2 and 3, lane 3) were stained with Coomassie blue. These purified EBNA1 fusions were then analyzed by western blotting using an anti-His antibody (Fig. 4E).

The purified hN and hSC fusion proteins were used to raise antibodies in rabbits, and the resulting antisera were designated as Ab-N and Ab-SC, respectively. As expected, our western blot analysis showed that Ab-N reacted with purified hN (although a significant background was observed; Fig. 5A), while Ab-SC detected hSC and hFC but not hN (Fig. 5B). We determiend the ability of Ab-SC to detect EBNA1 from cell lines carrying various EBNA1 subtypes. As reported in the present and other studies (46,47), the EBNA1 proteins in the B95.8, P3HR1, Daudi and CG3 cell lines correspond to the P-ala, V-leu, P-thr and V-val subtypes, respectively. The EBNA1 protein in Raji cells corresponds to the P-ala subtype, but has amino acid substitutions with respect to B95.8 EBNA1 (47). As shown in Fig. 5C, the Ab-SC antibody reacted with EBNA1 in all of the tested EBNA1-positive cells. As reported previously (26), the molecular weights of the EBNA1 proteins in Raji, B95.8, Daudi, and P3HR1 cells were 69, 79, 78, and 75 kDa, respectively. We obtained results consistent with those prior findings (Fig. 5C), and found that the CG3 EBNA1 was larger than the B95.8 EBNA1 in our western blot analysis (similar to the gene-based findings described earlier and shown in Fig. 2). Collectively, these results showed that our rabbit polyclonal anti-EBNA1 antibodies detected EBNA1 fusion proteins expressed in E. coli, and EBNA1 proteins of various subtypes in different EBV-associated cell lines.

Since most of the previous functional studies involved EBNA1 cloned from the B95.8 prototype (which belongs to the P-ala subtype), we subcloned the full-length V-val CG3 EBNA1 gene into a mammalian expression vector. The resulting plasmid, pCMV-CG3EBNA1, was transfected into 293 cells, and EBNA1 expression was examined by western blot analysis using our rabbit Ab-SC antibody (Fig. 5D). The electrophoretic mobility of the expressed CG3 EBNA1 was retarded as compared to that of B95.8 EBNA1, and the plasmid-expressed CG3 and B95.8 EBNA1 proteins were indistinguishable from those of the corresponding EBV-positive cells. These findings indicated that the full-length EBNA1 gene was successfully amplified from EBV-positive samples and expressed in the cultured cells, which in turn should facilitate studies on the functional differences among the different EBNA1 subtypes.

Growth factor-independent proliferation is considered to be a hallmark of cancer. To study the effect of CG3 EBNA1 on the cell proliferation and apoptosis of serum-starved cells, we generated 293 cells stably expressing CG3 EBNA1 (V-val subtype), B95.8 EBNA1 (P-ala subtype), or the drug-resistance gene alone (empty vector, control). Expression of CG3 EBNA1 (from two independent clones) and B95.8 EBNA1 was confirmed by western blot analysis using our Ab-SC antibody (Fig. 6A). The cells were then cultured in the absence of FCS, and cell viability was determined by trypan blue exclusion. As expected, 293 cell-expressing empty vector demonstrated loss of viability within three days of serum deprivation (Fig. 6B). Cells expressing B95.8 EBNA1 grew exponentially only for one day after serum withdrawal, and then plateaued through to the end of the experiment (day 5). As shown in Fig. 6B, however, the 293 cells expressing two independent clones of CG3 EBNA1 were able to proliferate under serum starvation throughout the experimental period. We also used western blotting to examine the status of PARP, a substrate cleaved by caspases during apoptosis, in 293 cells expressing empty vector, B95.8 EBNA1, and CG3 EBNA1 following serum withdrawal. As shown in Fig. 6C, cells expressing empty vector showed marginal induction of PARP cleavage after 63 h and significant PARP cleavage after 66 h of serum starvation. PARP cleavage was observed after 69 h of serum starvation in cells expressing B95.8 EBNA1. Of note, 293 cells expressing CG3 EBNA1 under serum starvation inhibited PARP cleavage at all of the tested time points. Taken together, these results suggested that CG3 EBNA1 blocked serum withdrawal-induced apoptosis and promoted cell proliferation.