Epstein-Barr virus nuclear protein EBNA3C residues critical for maintaining lymphoblastoid cell growth

Abstract

Epstein-Barr virus (EBV) nuclear antigen 3C (EBNA3C) is essential for efficient conversion of primary human B lymphocytes to lymphoblastoid cell lines (LCLs) and for continued LCL growth. We used a transcomplementation assay in the context of LCLs transformed by an EBV with a conditional EBNA3C to identify the EBNA3C amino acids (aa) necessary for maintaining LCL growth. Surprisingly, we found that most EBNA3C aa were essential for continued LCL growth. Only EBNA3C mutants deleted for residues within aa 507–515, 516–620, 637–675, or 676–727 maintained full LCL growth, and EBNA3C mutants deleted for residues within aa 728–732 or 910–992 maintained slow LCL growth. In contrast, EBNA3C lacking aa 180–231, which mediate RBP-Jκ association and are necessary for EBNA3C abrogation of EBNA2-induced transcription through RBP-Jκ, could not support LCL growth. Furthermore, 2 EBNA3C alanine substitution mutants within aa 180–231, which were wild-type (wt) in abrogating EBNA2-mediated transcription through RBP-Jκ, maintained LCL growth, and 2 alanine substitution mutants within aa 180–231, which were null in abrogating EBNA2-mediated transcription through RBP-Jκ, did not maintain LCL growth. This indicates that EBNA3C regulation of transcription through RBP-Jκ is critical to maintaining LCL growth. Several other EBNA3C functions also are critical for LCL growth, because EBNA3C mutants deleted for residues within aa 130–159, 251–506, or 733–909 were wt in abrogating transcription through RBP-Jκ and expression level, but did not maintain LCL growth.

Epstein-Barr virus (EBV) is causally implicated in lymphocyte-proliferative diseases in immune-deficient persons, endemic Burkitt lymphomas, 50% of Hodgkin lymphomas, other B-cell and T-cell lymphomas, anaplastic nasopharyngeal carcinomas, and some gastric carcinomas (see ref. 1 for a review). In vitro, EBV infection efficiently converts human B-lymphocytes into lymphoblastoid cell lines (LCLs) (2, 3). In LCLs, EBV expresses 6 nuclear proteins (EBNA-1, -2, -3A, -3B, -3C, and -LP), 3 integral membrane proteins (LMP-1, -2A, and -2B), 2 small nonpolyadenylated RNAs (EBER1 and EBER2), BamHI A rightward transcripts, and micro RNAs (see ref. 4 for a review). Reverse genetic analyses have shown that EBNA-1, -2, -3A, -3C, -LP, and LMP1 are critical for LCL outgrowth; LMP2 and EBER2 also may be important (4–11). We undertook the experiments reported here to identify those EBNA3C residues necessary to maintaining LCL growth.

The EBNA3C, EBNA3B, and EBNA3A genes have similar structures, are located in tandem, encode similar proteins, and likely arose from a triplication of an Old World primate γ1-herpesvirus gene. The triplication may have occurred after continental separation, because a New World primate γ1-herpesvirus has only a single EBNA3-like gene (12–14). EBNA3C, EBNA3B, and EBNA3A are at least partially divergent in functions required for LCL growth. Recombinant EBV reverse genetic experiments indicate that EBNA3C and EBNA3A are critical for LCL outgrowth, whereas EBNA3B is not (10, 15–18). Furthermore, only EBNA3C can transcomplement a conditionally inactivated EBNA3C, and only EBNA3A can transcomplement a conditionally inactivated EBNA3A (15, 17).

EBNA3C, EBNA3B, and EBNA3A likely regulate transcription. All 3 stably associate with RBP-Jκ/CBF1/CSL, the sequence-specific DNA-binding protein that mediates Notch receptor–induced transcription (19–23). RBP-Jκ plays a central role in EBV-induced cell growth, because EBNA2 associates with RBP-Jκ to activate the EBV Cp EBNA, LMP1, and LMP2 promoters and probably the cell c-Myc promoter as well (24–28). Transient transfection reporter assays have shown that EBNA3C, EBNA3B, or EBNA3A competes favorably with EBNA2 for interaction with RBP-Jκ and can down-regulate EBNA2 activation of the EBNA Cp promoter (20, 22, 23, 29, 30). EBNA3C also can coactivate the LMP1 promoter with EBNA2, a function not shared by EBNA3A or EBNA3B (31–35). Furthermore, EBNA3C interacts with other transcription factors and regulatory proteins, including PU.1, Spi-B, histone deacetylase 1 (HDAC1), CtBP, DP103, prothymosin-alpha, p300, Nm23-H1, SUMO1, and SUMO3 (33, 35–42). Moreover, comparison of RNAs in LCLs transformed by a recombinant cloned EBV with RNAs in LCLs transformed by mutants of the recombinant EBV identified an association between low EBNA3C expression and reduced TCL1A RNA and LCL growth (43). In addition to transcription regulation, EBNA3C may exert cell cycle regulatory effects through cyclin A, a cyclin A–dependent kinase (44, 45), an SCF^SKP2 ubiquitin ligase (46, 47), Chk2 (48), or p16^INK4a (15).

Our experiments used LCLs transformed by an EBV recombinant that expresses a conditionally active EBNA3C, wherein the EBNA3C C terminus is fused in frame to the N terminus of the 4-hydroxytamoxifen (4HT)-dependent mutant estrogen receptor (E3C-HT) (15). E3C-HT is inactive when the infected LCLs are incubated in medium without 4HT. Cell growth continues for several days, slows by day 6–10, and stops by day 20–30. Transduction with an EBNA3C expression vector before a shift to a medium without 4HT prevents growth arrest and maintains LCL growth, whereas EBNA3B or EBNA3A expression cannot maintain LCL growth (15). Because EBNA3C is needed to maintain E3C-HT–infected LCL growth in the absence of 4HT (15), the failure to transcomplement with a specifically mutated EBNA3C is likely to identify wild-type (wt) residues that are necessary for LCL growth.

Results

EBNA3C Deleted for Residues Within aa 1–506 or 733–909 Does Not Maintain LCL Growth, Whereas EBNA3C Deleted for Residues Within aa 507–620 or 637–727 Maintains wt LCL Growth and EBNA3C Deleted for Residues Within aa 728–732 or 910–992 Maintains Slow Growth.

Because LCLs established by infection with an EBV recombinant that expresses E3C-HT grow similarly to wt LCLs in media with 4HT, but experience growth arrest over 20–30 days in media without 4HT unless specifically transcomplemented by FLAG epitope–tagged wt EBNA3C (flagE3C) expression (15), we believed that this system was likely to be useful for identifying the EBNA3C residues necessary for maintaining LCL growth. We cloned FlagE3C deletion mutants (Fig. 1) into oriP plasmids and evaluated them for maintenance of E3C-HT–transformed LCL growth under nonpermissive conditions in media without 4HT (Fig. 2). As expected, E3C-HT LCLs transfected with control oriP plasmid grew in media with 4HT but stopped growing in the absence of 4HT, whereas E3C-HT LCLs transfected with an oriP plasmid expressing wt flagE3C continued to grow at the same rate in the presence or absence of 4HT (Fig. 2). E3C-HT LCLs that were transfected with an oriP plasmid expressing each EBNA3C deletion mutant also continued to grow in the medium with 4HT, indicating that expression of the EBNA3C deletion mutants does not prevent E3C-HT LCL growth (Fig. 2). However, in the medium without 4HT, some deletion mutants consistently maintained E3C-HT LCL growth as well as wt flagE3C, whereas others were less effective or ineffective in maintaining LCL growth (Fig. 2 and Table 1). Specifically, flagE3C lacking aa 507–515 (flagE3C Δ507–515), flagE3C Δ516–620, flagE3C Δ637–675, and flagE3C Δ676–727 maintained E3C-HT LCL growth in the absence of 4HT as well as wt EBNA3C (Fig. 2 and Table 1). In contrast, flagE3C Δ1–129, flagE3C Δ130–159, flagE3C Δ160–179, flagE3C Δ180–250, flagE3C Δ251–300, flagE3C Δ301–365, flagE3C Δ366–400, flagE3C Δ401–506, flagE3C Δ733–826, flagE3C Δ827–909, and control expression vector–transfected E3C-HT LCLs ceased growing at 20–30 days in medium without 4HT, and flagE3C Δ728–732 and flagE3C Δ910–992–transfected E3C-HT LCLs grew more slowly than flagE3C-transfected LCLs (Fig. 2A and Table 1). The transfected flagE3C deletion mutants were expressed at levels comparable to that of wt flagE3C on day 4 after E3C-HT LCL transfection, when cells were seeded into media with or without 4HT (Fig. 2B). These data indicate that EBNA3C residues within E3C aa 1–129, 130–159, 160–179, 180–250, 251–300, 301–365, 366–400, 401–506, 733–826, and 827–909 are indispensable for LCL growth and that residues within EBNA3C aa 728–732 and 910–992 are important for wt LCL growth, whereas residues within EBNA3C aa 507–620 and 637–727 are not required to maintain LCL growth (Fig. 2A and Table 1).

Fig. 1.

(A) A schematic diagram of EBNA3C showing known and putative functional domains, aa motifs, and binding sites for cell proteins. (B) Diagram of FLAG-tagged EBNA3C deletion mutants used in this study indicating EBNA3C residues that were deleted without affecting cell growth (dispensable), with an intermediate effect on cell growth (intermediate), or with marked effect on cell growth (critical).

Fig. 2.

Transcomplementation assays with EBNA3C or EBNA3C deletion mutants for maintenance of E3C-HT LCL growth under nonpermissive conditions. (A) E3C-HT LCLs were transfected with 30 μg of oriP plasmid expressing FLAG-tagged EBNA3C (flagE3C), indicated flagE3C mutants, or control oriP plasmid (Cont) and then cultured in medium containing 4HT for 4 days. The cells were then washed and resuspended at 5 × 10⁵ cells/5 mL of complete medium with (+) or without (-) 4HT in a 25-cm² culture flask (day 0). Every 6–8 days, cells were counted and cultures were fed with similar amounts of fresh medium. The numbers of viable cells derived from the initial cultures were calculated and plotted at each time point. The number “10E5” along the y-axis indicates that the cell number plotted should be multiplied by 100,000. The results are representative of 3 independent experiments. (B) Protein lysates made from these cells on day 0 (4 days after transfection) were immunoprecipitated with antibodies to FLAG and subjected to Western blot analysis with EBV-immune human serum to detect flagE3C and deletion mutants.

Table 1.

Summary of transcomplementation assays

Control	0/5	-*
flagE3C	5/5	wt†
Δ1–129	0/3	-, -, -
Δ130–159	0/3	-, -, -
Δ160–179	0/3	-, -, -
Δ180–250	0/3	-, -, -
Δ180–231	0/3	-, -, -
Δ251–300	0/3	-, -, -
Δ301–365	0/3	-, -, -
Δ366–400	0/3	-, -, -
Δ401–506	0/3	-, -, -
Δ507–515	3/3	wt, wt, wt
Δ516–620	3/3	wt, wt, wt
Δ637–675	3/3	wt, wt, wt
Δ676–727	3/3	wt, wt, wt
Δ728–732	3/3	sl,‡ sl, sl
Δ733–826	0/3	-, -, -
Δ827–909	0/3	-, -, -
Δ910–992	3/3	wt, sl, sl

*No growth maintenance.

^†Growth maintenance similar to that of flagE3C.

^‡Slower growth maintenance.

EBNA3C aa 180–231 Are Essential for RBP-Jκ Association and Maintainence of LCL Growth.

EBNA3C aa 180–240 are homologous to corresponding EBNA3A and EBNA3B residues, which also mediate RBP-Jκ association in yeast 2-hybrid and in vitro binding assays (23). We evaluated the potential effects of flagE3C deletion mutations on RBP-Jκ association in LCLs after the expression of wt flagE3C or deletion mutants in IB4 LCLs. Wt flagE3C or deletion mutants were immunoprecipitated with FLAG antibody–agarose beads, and the immunoprecipitates were subjected to Western blot analysis with EBNA3C immune human sera or RBP-Jκ antibody. FLAG antibody–agarose beads immunoprecipitated comparable levels of wt flagE3C and flagE3C deletion mutants (Fig. 3, EBV-immune HS). RBP-Jκ was significantly associated with wt flagE3C and all deletion mutants except flagE3C Δ180–250 and flagE3C Δ180–231 (Fig. 3). This indicates that EBNA3C residues within aa 180–231 and 180–250 are essential for RBP-Jκ association in LCLs, and that other residues are not required. Because EBNA3C aa 180–231 are essential for RBP-Jκ association and for maintenance of LCL growth (Table 1), association with RBP-Jκ likely is required for LCL growth.

Fig. 3.

EBNA3C aa 180–231 are required for RBP-Jκ association. IB4 LCLs were transfected with 30 μg of oriP plasmid expressing FLAG-tagged EBNA3C (flagE3C), indicated flagE3C deletion mutants, or control oriP plasmid (Cont). At 48 h after transfection, the flagE3C protein complexes were immunoprecipitated with antibodies to FLAG and subjected to Western blot analysis with RBP-Jκ–specific antibody or EBV-immune human serum (HS).

EBNA3C Maintenance of LCL Growth Also Requires Residues Necessary for Full Inhibition of EBNA2-Mediated Transcription Activation Through RBP-Jκ and Residues Necessary for Other Activities.

In transient transfection assays in non–EBV-infected BJAB B-lymphoma cells, EBNA2 transactivated the EBNA Cp promoter through RBP-Jκ, and coexpression of wt EBNA3C or EBNA3A fully repressed EBNA2 transactivation, whereas EBNA3C deleted for aa 180–231 and EBNA3A deleted for aa 170–240 (which are required for RBP-Jκ association) did not abrogate EBNA2 transactivation or maintain LCL growth, consistent with RBP-Jκ association being essential for transcription regulation and continued LCL growth (Fig. 1) (20, 22, 23, 29, 30). Notably, EBNA3A deleted for aa 300–386 still associated with RBP-Jκ, but was deficient in abrogating EBNA2- and RBP-Jκ–mediated transcription and could not maintain LCL growth, indicating that abrogation of EBNA2- and RBP–Jκ mediated transcription can be indicative of another EBNA3A element and activity essential for continued LCL growth, a regulator of RBP–Jκ interaction or a repressive component of EBNA3A (49). To evaluate the existence of a similar component in EBNA3C abrogation of EBNA2 transcriptional regulation through RBP-Jκ that might be required for LCL growth, we examined flagE3C deletion mutants for their activity in repression of EBNA2- and RBP-Jκ–mediated transactivation of EBNA Cp promoter activity (Fig. 4).

Fig. 4.

Effects of EBNA3C or EBNA3C deletions on repression of EBNA2 activation of a multimerized Cp promoter. BJAB cells were transfected with the pLuc-Cp reporter construct containing 8 copies of the RBP-Jκ binding site along with EBNA2 (E2+) or with the indicated flagE3C construct. Firefly luciferase activity was normalized to the renilla luciferase activity from cotransfected pRL-TK. The results are an average of triplicate samples and are representative of 3 experiments with similar results. Bars indicate standard errors.

As expected, EBNA2 transactivated the Cp promoter and flagE3C associated with RBP-Jκ, fully repressed EBNA2- and RBP-Jκ–mediated Cp promoter transactivation, and maintained LCL growth, whereas flagE3C Δ180–250, which is null in RBP-Jκ association, failed to repress Cp promoter transactivation and did not maintain LCL growth (Figs. 2–4). Interestingly, flagE3C deletion mutants Δ1–129 and Δ160–179, associated with RBP-Jκ, were significantly less repressive of EBNA2 transactivation than wt flagE3C, and did not maintain LCL growth (Figs. 2–4). This indicates that EBNA3C aa 1–129 and 160–179 are important for abrogating EBNA2- and RBP-Jκ–mediated promoter transactivation and maintaining LCL growth, even though they are not required for RBP-Jκ association (Figs. 3 and 4). Furthermore, all flagE3C deletion mutants that completely or partially support LCL growth (Δ507–515, Δ516–620, Δ637–675, Δ676–727, Δ728–732, and Δ910–992) associated with RBP-Jκ and abrogated EBNA2 and RBP-Jκ transcription (Table 1 and Figs. 2–4). However, flagE3C Δ130–159, Δ251–300, Δ301–365, Δ366–400, Δ401–506, Δ728–732, Δ733–826, Δ827–909, and Δ910–992 associated fully with RBP-Jκ, strongly repressed EBNA2- and RBP-Jκ–mediated Cp transactivation, and were fully or partially impaired in maintaining LCL growth (Figs. 2–4). This indicates that transcription regulation through RBP-Jκ is necessary but not sufficient for the EBNA3C effects in maintaining LCL growth. EBNA3C also requires many other residues to maintain LCL growth. These residues are likely to mediate other functions necessary for maintaining LCL growth.

Alanine Point Mutants Within EBNA3C aa 180–231 That Are wt in Abrogating EBNA2-Induced Transcription Through RBP-Jκ Maintain LCL Growth, Whereas Point Mutants That Do Not Abrogate EBNA2-Induced Transcription Through RBP-Jκ Do Not Maintain LCL Growth.

To further exlore the relationship between EBNA3C regulation of transcription through RBP-Jκ and LCL growth maintenance, we introduced alanine substitution mutations within EBNA3C aa 180–231, which are required for association with RBP-Jκ (Fig. 3), abrogation of EBNA2- and RBP-Jκ–mediated activation of the Cp promoter (Fig. 4), and maintenance of LCL growth (Fig. 5D). FlagE3C with ₂₀₉TFGC₂₁₂ substituted to ₂₀₉AAAA₂₁₂ was expressed at wt levels (Fig. 5B), was deficient in RBP-Jκ association and in abrogation of EBNA2 activation (ref. 23 and Fig. 5 B and C), and did not maintain LCL growth (Fig. 5D). FlagE3C ₁₈₂AAA_{184, 195}AAA₁₉₇, and ₂₁₈AAA₂₂₀ also were expressed at wt levels (Fig. 5B) and associated with RBP-Jκ at wt levels (Fig. 5B). But FlagE3C mutants ₁₉₅AAA and ₂₁₈AAA repressed EBNA2- and RBP-Jκ–dependent transcription as well as wt flagE3C and maintained E3C-HT LCL growth under nonpermissive conditions as well as wt flagE3C (Fig. 5 C and D), whereas ₁₈₂AAA did not repress EBNA2- and RBP–Jκ dependent transcription and did not maintain E3C-HT LCL growth under nonpermissive conditions (Fig. 5 C and D). Thus, alanine point mutant EBNA3C abrogation of EBNA2 transcription through RBP-Jκ correlated with alanine point mutant support of LCL growth. These data further indicate that EBNA3C regulation of EBNA2 transcription through RBP-Jκ is critical to maintaining LCL growth.

Fig. 5.

EBNA3C alanine point mutant repressive effects on EBNA2- and RBP-Jκ–dependent transactivation correlate with the ability to maintain E3C-HT LCL growth under nonpermissive conditions. (A) The sequence of the EBNA3C region required for association with RBP-Jκ (E3C aa 180–231). Site-specific mutagenesis was used to make 4 alanine substitution mutants (₁₈₂AAA, ₁₉₅AAA, ₂₀₉AAAA, and ₂₁₈AAA). (B) IB4 LCLs were transfected with 30 μg of oriP plasmid expressing FLAG-tagged EBNA3C (flagE3C), flagE3C mutants, or control oriP plasmid (Cont). At 48 h after transfection, protein complexes were immunoprecipitated with FLAG antibody and subjected to Western blot analysis with RBP-Jκ–specific antibody or EBV-immune human serum. (C) BJAB cells were transfected with the pLuc-Cp reporter plasmid containing 8 copies of an RBP-Jκ binding site, with an EBNA2 (E2+) expression plasmid, and with the indicated flagE3C or flagE3C mutant plasmid. Relative firefly luciferase activity was normalized to the renilla luciferase activity from cotransfected pRL-TK. The results are the average of triplicate samples and are representative of 2 experiments with similar results. Bars indicate standard errors. (D) E3C-HT LCLs were transfected with 30 μg of oriP plasmid expressing flagE3C, flagE3C mutants, or control oriP plasmid (Cont) and cultured in medium with 4HT for 5 days. The cells were then washed and resuspended at 1 × 10⁶ cells/10 mL of complete medium with (+) or without (-) 4HT in a 25-cm² culture flask (day 0). Every 5–6 days, cells were counted and cultures were fed with similar amounts of fresh medium. Total numbers of viable cells derived from the initial cultures were calculated and plotted at each time point. The number “10E5” along the y-axis indicates that the cell number plotted should be multiplied by 100,000. The results are representative of 3 independent experiments with similar results. (E) Protein lysates made from the cells shown in D on day 0 (5 days after transfection) were subjected to Western blot analysis with DYKDDDDK Tag–specific antibody.

Discussion

Transcomplementation of a conditional EBNA3C under nonpermissive conditions with wt or specifically mutated EBNA3C has enabled the identification of EBNA3C residues that are critical to maintaining LCL growth and those that are not. Despite similar wt and mutant EBNA3C expression levels, only wt EBNA3C and 4 deletion mutants maintained wt LCL growth and 2 deletions maintained slow LCL growth, whereas 10 deletion mutants failed to maintain LCL growth. These data indicate that EBNA3C has multiple domains that are necessary for maintaining LCL growth.

The discovery that residues throughout most of EBNA3C, including residues within aa 1–129, 130–159, 160–179, 180–250, 251–300, 301–365, 366–400, 401–506, 733–826, and 827–909, are critical to maintaining LCL growth is surprising, especially because these residues are not required for wt protein expression levels. Indeed, only EBNA3C residues within aa 507–620 and 637–727 are not necessary for LCL growth, and residues within aa 728–732 and 910–992 are important for full wt LCL growth. This broad requirement of EBNA3C residues for maintaining LCL growth contrasts with the more restricted set of EBNA3A residues necessary for LCL growth under nonpermissive conditions in EBNA3A-HT–infected LCLs (49). As reported previously, EBNA3A residues within aa 2–124, 410–612, and 620–820 are not required for LCL growth, and deletion of residues within aa 240–300 and 827–944 results in slower LCL growth; only residues within aa 170–240, 300–386, and 386–410 are necessary for maintaining LCL growth (49). Nevertheless, we found that the number and extent of EBNA3C deletions that failed to abrogate EBNA2 activation of transcription through RBP-Jκ and failed to maintain LCL growth were similar to those seen with EBNA3A. EBNA3C and EBNA3A deletions of the RBP-Jκ association domains failed to abrogate and failed to maintain growth, as did EBNA3C N-terminal deletions of aa 1–124 and 160–179 or EBNA3A deletion of aa 300–386. These latter EBNA3C and EBNA3A deletions failed to block EBNA2 activation through RBP-Jκ and were functionally deficient in an aspect of RBP-Jκ interaction or repression that is essential for EBNA3C or EBNA3A maintenance of LCL growth but does not affect RBP-Jκ association. In contrast, the larger number and cumulatively more extensive EBNA3C deletions that cannot maintain LCL growth were wt in abrogation of EBNA2 activation through RBP-Jκ, and thus are unlikely to be misfolded but likely to identify residues that interact with other cell proteins critical to continued LCL growth.

In the EBNA3C experiments described here and in previous EBNA3A experiments (49), deletion of EBNA3C residues within aa 180–231 and EBNA3A residues within aa 170–240, which are essential and sufficient for RBP-Jκ association (23, 29, 50), also abrogated EBNA2 transcription regulation through RBP-Jκ and were null mutants for LCL growth, indicating that transcription regulation through association with RBP-Jκ is essential for the unique EBNA3C or EBNA3A roles in maintaining LCL growth. Furthermore, EBNA3C or EBNA3A (49) alanine substitution mutants within the RBP-Jκ association domains that retained association with RBP-Jκ but failed to abrogate EBNA2- and RBP-Jκ–mediated transcription were unable to maintain LCL growth, whereas wt EBNA3C and 2 EBNA3C alanine substitution mutants, which abrogated EBNA2- and RBP-Jκ–mediated transcription, maintained LCL growth. Moreover, EBNAC deleted for aa 1–129 and 160–179 and EBNA3A deleted for aa 300–386 associated with RBP-Jκ but were deficient in abrogating EBNA2- and RBP-Jκ–mediated transcription and in maintaining LCL growth. This indicates that EBNA3C or EBNA3A inhibition of EBNA2- and RBP-Jκ–mediated transcription activation detects a second EBNA3C- or EBNA3A-RBP-Jκ– dependent activity that is essential for LCL growth. The essential activity of these EBNA3C or EBNA3A residues in LCL growth could be in regulating RBP–Jκ interaction, mediating repression through RBP-Jκ, or regulating EBNA2 or another RBP-Jκ–associated transcription factor. EBNA3C, EBNA3A, or EBNA3B can independently repress EBNA2 activation of Cp, but this is not the precise EBNA3C activity detected in the E3C-HT LCL growth maintenance assay. The uniform correlation of EBNA3C and EBNA3A residues important for transcription regulation through RBP-Jκ with importance for LCL growth provides strong evidence that EBNA3C and EBNA3A transcription regulation through RBP-Jκ are complex, unique, and essential EBNA3C and EBNA3A functions for LCL growth.

N-terminal to the RBP-Jκ binding domain, EBNA3C aa 1–129, 130–159, and 160–179 were essential for LCL outgrowth, whereas EBNA3A did not require aa 2–124, and the significance of EBNA3A aa 125–169 has not yet been investigated. EBNA3C aa 130–190 can bind Skp2 and thereby affect c-Myc or cyclin A/cdk2 stability (47, 51). Furthermore, EBNA3C aa 1–129 and 160–179 may recruit other factors besides RBP-Jκ to the EBNA3C N terminus, affect EBNA3C folding into a functional RBP-Jκ–binding or –repressive domain, or effect Skp2 interaction.

Some deletion mutants that repressed transcription through RBP-Jκ (e.g., EBNA3C Δ130–159, Δ251–300, Δ301–365, Δ366–400, Δ401–506, Δ733–826, and Δ827–909) did not maintain LCL growth. This indicates that EBNA3C regulation of transcription through RBP-Jκ is required but is not sufficient for LCL growth maintenance. The protein interactions that mediate these essential effects on cell growth remain to be identified.

One important EBNA3C transcription effect through RBP-Jκ is coactivation of the LMP1 promoter with EBNA2 (27, 31, 32). Chromatin immunoprecipitation with antibody to EBNA3C localizes EBNA3C to the LMP1 promoter proximal RBP-Jκ site (32). EBNA3C aa 181–365 also can interact with Spi-1/PU.1, another important transcription factor in EBNA2 and EBNA3C activation of the LMP1 promoter (27, 35). EBNA3C residues within aa 251–300 and 301–365 are essential for EBNA3C maintenance of LCL growth. Furthermore, prothymosin α can interact with EBNA3C aa 366–400 (36), and these residues also are critical for LCL growth. More precise mapping of the EBNA3C–Spi-1 and –prothymosin α interaction sites is needed to enable further genetic testing of their importance in maintaining LCL growth.

Although analysis of EBNA3C effects in LMP1 promoter activation in transient promoter and reporter assays has identified EBNA3C aa 507–513 (33, 34) within an aa 343–545 repressive domain (52) as important for EBNA3C coactivation, EBNA3C aa 507–515 were found to be not important for maintaining LCL growth. However, LMP1 expression did not change under nonpermissive conditions for E3C-HT expression (15). In addition, EBNA3C aa 637–675 can bind Nm23-H1 (41), but aa 637–675 are not essential for LCL growth, indicating that Nm23-H1 interaction with this domain is less important for LCL growth.

More C-terminal to the RBP-Jκ site, deletion of the EBNA3C aa 728–732 CtBP repressor binding site or of the EBNA3A aa 827–944 CtBP repressor binding sites (42, 53, 54) resulted in slower but persistent LCL growth. This indicates that these EBNA3C and EBNA3A CtBP binding sites are not essential for EBNA3C- and EBNA3A-mediated LCL growth. The striking correlation between CtBP binding and EBNA3C or EBNA3A cooperation with (Ha-) Ras in the immortalization of rat fibroblasts (42, 53, 54) may be due to enforced EBNA3C or EBNA3A expression–induced high-level recruitment of CtBP. CtBP can bind CtIP, which has an LXCXE motif that can induce pRb release of E2F and enable cell cycle progression (55). But in LCL growth maintenance, CtBP binding is less important for EBNA3C than transcription regulation through RBP-Jκ. In contrast, EBNA3C aa 733–909 are essential for LCL growth and overlap with aa 724–826, a glutamine- and proline-rich transactivation domain (29, 56), which may bind p300 or HDAC1 (36, 38–40) and may be important in transcription regulation.

LCLs with low EBNA3C expression underexpress TCL1A (43), which also may be important in EBNA3C maintenance of LCL growth (57). Studies are underway to explore TCL1A regulation by EBNA3C, identify residues that may regulate TCL1A expression, and identify other EBNA3C-regulated RNAs by profiling RNA abundances in E3C-HT LCLs under permissive and nonpermissive conditions.

Materials and Methods

Cell Lines.

BJAB is an EBV-negative B-lymphoma cell line. IB4 is an LCL transformed with B95–8 strain EBV. E3C-HT EBV-infected LCLs (15) were maintained in RPMI medium 1640 supplemented with 15% FBS, L-glutamine, streptomycin, penicillin, and 400 nM 4HT (Sigma). All other cell lines were maintained in RPMI medium 1640 supplemented with 10% FBS, L-glutamine, streptomycin, and penicillin.

Plasmids.

OriP plasmids expressing FLAG-EBNA3C (flagE3C) under the control of SV40 promoter have been described previously (15). Deletion mutants of flagE3C (Δ1–129, Δ130–159, Δ160–179, Δ180–231, Δ180–250, Δ251–300, Δ301–365, Δ366–400, Δ401–506, Δ507–515, Δ516–620, Δ637–675, Δ676–727, Δ728–732, Δ733–826, Δ827–909, and Δ910–992), and flagE3C alanine substitution mutants ₁₈₂AAA (₁₈₂AAA substituted for ₁₈₂MGY), ₁₉₅AAA (₁₉₅AAA substituted for ₁₉₅VPN), ₂₀₉AAAA (₂₀₉AAAA substituted for ₂₀₉TFGC), and ₂₁₈AAA (₂₁₈AAA substituted for ₂₁₈TLN) were constructed from the oriP plasmid containing flagE3C using PCR mutagenesis. The deletion and alanine substitution mutations were verified by sequencing.

Transcomplementation Assay.

First, 5 × 10⁶ E3C-HT–infected LCLs were transfected with 30 μg of oriP plasmid expressing flagE3C, flagE3C mutants, or control oriP plasmid through electroporation. For electroporation, LCLs were resuspended in 400 μL of complete medium with DNA in a cuvette. After a 10-min incubation at 25 °C, the culture was pulsed with 220 V at 950 μF using a Bio-Rad Gene Pulser II. The LCL transfection efficiency was 20%–40%, as estimated from the EGFP-positive cell fraction on day 4 after transfection with an oriP plasmid expressing EGFP (data not shown). Transfected LCLs were cultured in complete medium with 4HT for 4–5 days. The cells were then washed and resuspended at 5 × 10⁵ cells/5 mL or 1 × 10⁶ cells/10 mL of complete medium with or without 4HT in a 25-cm² culture flask. Every 5–8 days, viable cell numbers were determined by hemocytometry based on trypan blue exclusion. Cultures were then split, and total viable cell numbers were calculated relative to the initial culture.

Western Blot Analysis.

Total cell lysates or immunoprecipitated proteins were separated by SDS-PAGE, blotted onto nitrocellulose membrane, and reacted with EBV-immune human sera, rabbit polyclonal antiserum to RBP-Jκ (20), or DYKDDDDK Tag antibody (Cell Signaling). Membranes were reacted with HRP-conjugated species-specific secondary antibodies (GE Healthcare) and developed with a chemiluminescent reagent (GE Healthcare).

Immunoprecipitation.

IB4 cells (5 × 10⁶) were transfected with 30 μg of the oriP plasmid expressing flagE3C, flagE3C mutants, or a control oriP plasmid through electroporation (Bio-Rad Gene Pulser II; 230 V, 950 μF). After 48 h, the cells were lysed in immunoprecipitation (IP) buffer (150 mM NaCl, 1% Nonidet P-40, 50 mM Tris [pH 7.4], 2 mM EDTA) containing protease inhibitors (10 μg of aprotinin per mL, 0.5 μM phenylmethylsulfonylfluoride) for 1 h, and then centrifuged to remove insoluble debris. The supernatant was incubated overnight with M2-conjugated agarose beads (Sigma) at 4 °C. The beads were washed 4 times with IP buffer. Proteins were eluted with SDS sample buffer and subjected to Western blot analysis with EBV-immune human serum, DYKDDDDK Tag–specific, or RBP-Jκ–specific antibodies.

Reporter Assay.

BJAB cells (4 × 10⁶) in log-phase growth were electroporated with 0.05 μg of pRL-TK, 5 μg of pLuc-Cp reporter construct, and 1 μg of pSG5-EBNA2 alone or in combination with 15 μg of oriP plasmid expressing flagE3C wt or mutants in a Bio-Rad Gene Pulser II at 220 V and 950 μF. Two days after transfection, the cells were lysed in passive lysis buffer (Promega), and luciferase activities were measured with the PromegaDual-Luciferase Reporter Assay system. Firefly luciferase activities were corrected for transfection efficiency based on renilla luciferase activity.