EBNA-3B- and EBNA-3C-regulated cellular genes in Epstein-Barr virus-immortalized lymphoblastoid cell lines

Abstract

The cellular pathways that Epstein-Barr virus (EBV) manipulates in order to effect its lifelong persistence within hosts and facilitate its transmission between hosts are not well understood. The EBV nuclear antigen 3 (EBNA-3) family of latent infection proteins consists of transcriptional regulators that influence viral and cellular gene expression in EBV-infected cells. To identify EBNA-3B- and EBNA-3C-regulated cellular genes potentially important for virus infection in vivo, we studied a lymphoblastoid cell line (LCL) infected with an unusual EBV mutant, where a genetic manipulation to delete EBNA-3B also resulted in a significant decrease in EBNA-3C expression and slower than normal growth (3B(-)/3C(low)). Transcriptional profiling was performed on the 3B(-)/3C(low) LCLs, and comparison of mutant and wild-type LCL profiles resulted in a group of 21 probe sets representing 16 individual genes showing statistically significant differences in expression. Further quantitative reverse transcription-PCR analyses comparing 3B(-)/3C(low) LCLs to a previously described EBNA-3B mutant (3B(-)) where EBNA-3C expression was normal revealed three potential EBNA-3B-repressed genes, three potential EBNA-3C-repressed genes, and two potential EBNA-3C-activated genes. The most highly EBNA-3C-repressed gene was Jagged1, a cell surface ligand and inducer of the Notch receptor signaling pathway that is usurped by EBV genes essential for B-cell immortalization. 3B(-)/3C(low) LCLs expressed increased levels of Jagged1 protein and were able to more efficiently induce functional Notch signaling, and this signaling was dependent on Notch cleavage by gamma-secretase. However, inhibiting gamma-secretase-mediated Notch cleavage did not rescue 3B(-)/3C(low) LCL growth, suggesting that EBNA-3C-mediated repression of this signaling pathway did not contribute to LCL growth in tissue culture. Similarly, expression of the chemokine receptor CXCR4 was reproducibly upregulated in EBNA-3B-null LCLs. Since deletion of EBNA-3B has no significant impact on B-cell immortalization in tissue culture, this finding suggested that EBNA-3B-mediated regulation of CXCR4 may be an important viral strategy for alteration of B-cell homing in the infected host. These studies identify two cellular genes that do not contribute to EBV-induced B-cell growth but whose expression levels are strongly EBNA-3 regulated in EBV-infected primary B cells. These EBV-manipulated cellular pathways may be important for virus survival or transmission in humans, and their independence from EBV-induced B-cell growth makes them potential targets for testing in vivo with the rhesus lymphocryptovirus animal model for EBV infection.

FIG. 1.

Schematic of EBNA-3B mutations made in the EBV BAC. F plasmid sequences for prokaryotic replication were inserted into the EBV internal repeat 1 (IR1) region. The wild-type BAC contains the full complement of EBNA-3A, -3B, and -3C genes. EBNA-3B was knocked out in the 3B⁻/3C^low BAC by replacement of the entire EBNA-3B open reading frame with a kanamycin resistance cassette (Kan^r) flanked by FRT sites. The 3B⁻ BAC has been described previously and contains a single FRT site in place of the EBNA-3B open reading frame (2).

FIG. 2.

3B⁻/3C^low LCLs are impaired for cell growth and express low levels of EBNA-3C protein. (A) Growth curves for wild-type and 3B⁻/3C^low LCLs. Wild-type and two independently derived 3B⁻/3C^low LCLs (3B⁻/3C^low A1 and A2) were seeded at an initial density of 4 × 10⁵ cells per ml in a 24-well plate at day 0, and viable cell counts were taken over a period of 7 days. Data are means ± standard deviations for three replicates. (B) Expression of EBV latent proteins in 3B⁻/3C^low LCLs. Whole-cell lysates were analyzed for expression of EBV latent proteins by Western blotting. EBV-immune human serum was used to detect EBNA-1, -2, -3A, -3B, and -3C; EBNA-3C alone was detected using the A10 monoclonal antibody (30); EBNA-LP was detected using the JF186 monoclonal antibody (9); and LMP-1 was detected using the S12 monoclonal antibody (27). Controls included an EBV-negative Burkitt lymphoma line (BJAB), an LCL infected with the B95-8 virus (B95-8), and an LCL derived from B95-8-immortalized umbilical-cord B cells (IB4) (21) which lacks full-length EBNA-3B protein expression (38) but is normal for expression of all the other latent proteins. An LCL infected with the wild-type BAC-derived virus (wild-type) and two independently derived 3B⁻/3C^low LCLs are shown.

FIG. 3.

Spontaneous restoration of normal EBNA-3C expression in a 3B⁻/3C^low LCL after several months in culture. (A) Expression of the EBNA-3 proteins in a 3B⁻/3C^low LCL after prolonged culture. Whole-cell lysates were analyzed for expression of EBNA-3 proteins by Western blotting. EBV-immune serum was used to detect all the EBNA-3s, and the A10 monoclonal antibody was used to detect EBNA-3C alone. (B) Schematic of the mutation in the EBNA-3C restored 3B⁻/3C^low LCL. The DNA region between the EBNA-3A and EBNA-3C open reading frames was PCR amplified and sequenced. Kanamycin^R/Kan^R represents the kanamycin resistance cassette inserted in the place of the EBNA-3B ORF; FRT represents the FLP recombinase target sites flanking the Kan^r cassette; a and b represent nonviral, unique flanking sequences on either side of the FRT sites. The numbers of EBV nucleotides present between the nonviral insert and the EBNA-3A stop codon or EBNA-3C start codon are indicated by the numbered base pairs (bp).

FIG. 4.

Altered cellular gene expression in 3B⁻/3C^low BAC-infected LCLs. Shown are genes displaying significant changes in expression levels in 3B⁻/3C^low LCLs relative to expression levels in all wild-type LCLs grouped together (B95-8, P3HR1, and wild-type BAC-derived LCLs). Vertical columns represent data obtained for each individual cell line: for two LCLs immortalized with B95-8 virus (B95-8 clones 1 and 2), three LCLs derived by homologous recombination of EBNA-LP and EBNA-2 in P3HR1 EBV (P3HR1 clones 1, 2, and 3), two LCLs derived from the wild-type EBV BAC (wild-type A1 and A2), and three LCLs derived from the 3B⁻/3C^low BAC (3B⁻/3C^low A1, A2, and B1). Horizontal rows represent data obtained for each individual gene across all cell lines. The color of each data point represents the normalized signal-to-control ratio for a given gene in a specific cell line. Identities of individual genes are as follows: JAG1, jagged1; NCALD, neurocalcin D; AKR1C1, aldoketoreductase family, member C1; UST, uronyl-2-sulfotransferase; CHST4, carbohydrate sulfotransferase 4; CXCR4, chemokine (C-X-C motif) receptor 4; ENTH, enthoprotin; AIM1, absent in melanoma 1; P101-PI3K, phosphoinositide-3-kinase, regulatory subunit, polypeptide p101; TTF2, transcription termination factor, RNA polymerase II; JWA, vitamin A responsive; FLNA, filamin A, alpha; PDXK, pyridoxal kinase; ITGA4, integrin α4; EVI2A, ecotropic viral integration site 2A; TCL1A, T-cell leukemia/lymphoma 1A. GenBank accession numbers for the two unknown genes are shown.

FIG. 5.

Quantitative RT-PCR analysis of cellular gene expression levels in 3B⁻/3C^low and 3B⁻ BAC-derived LCLs. Quantitative RT-PCR was performed on RNA derived from wild-type, 3B⁻/3C^low, and 3B⁻ LCLs. Average C_T values for each gene were normalized to GAPDH levels, and data displayed are changes (n-fold) in expression levels of 3B⁻/3C^low and 3B⁻ LCLs relative to expression levels in wild-type LCLs. Data are means ± standard deviations for two independent experiments performed in triplicate on the same RNA preparation.

FIG. 6.

Jagged1 protein expression is increased in 3B⁻/3C^low BAC-derived LCLs. (A) Whole-cell lysates from wild-type, 3B⁻/3C^low, and 3B⁻ BAC-derived LCLs were analyzed for EBNA-3 and Jagged1 expression by Western blotting. EBV-immune human serum was used to detect the EBNA-3 proteins; Jagged1 protein was detected using the H-114 polyclonal antibody. (B) Downregulation of Jagged1 expression in 3B⁻/3C^low LCLs that are restored for EBNA-3C expression. Whole-cell lysates from the 3B⁻/3C^low LCL restored for normal EBNA-3C expression were analyzed for Jagged1 expression by Western blotting. The bottom two panels are reproduced from Fig. 3A, for reference and loading control purposes.

FIG. 7.

3B⁻/3C^low LCLs are able to induce Notch signaling. (A) Notch-induced luciferase activity in U2OS cells stimulated with wild-type or 3B⁻/3C^low LCLs. U2OS cells were transfected with a Notch luciferase reporter construct containing four RBP-Jκ binding sites in its promoter (17). Twenty-four hours posttransfection, cells were overlaid with either wild-type or 3B⁻/3C^low LCLs at different concentrations. Forty-eight hours posttransfection, cells were harvested and lysates assayed for luciferase activity. Data shown are stimulation levels (n-fold) relative to those in unstimulated controls and represent the means ± standard deviations for three replicates. (B) Inhibition of 3B⁻/3C^low LCL-stimulated, Notch-induced luciferase activity by the γ-secretase inhibitor compound E. U2OS cells were transfected with luciferase reporter constructs and treated with 1 μM compound E. Twenty-four hours posttransfection, cells were overlaid with either wild-type or 3B⁻/3C^low LCLs at a density of 1 × 10⁶ cells/well. Forty-eight hours posttransfection, cells were harvested and lysates assayed for luciferase activity. Data shown are stimulation levels (n-fold) relative to those in untreated, unstimulated controls and represent the means ± standard deviations for three replicates.

FIG. 8.

Notch signaling induced by 3B⁻/3C^low BAC-derived LCLs is not responsible for their slow growth. Wild-type and 3B⁻/3C^low LCLs (3B⁻/3C^low A1 and 3B⁻/3C^low B1) were treated with either dimethyl sulfoxide or 1 μM compound E (+E) at day 0. Cells were seeded at an initial density of 4 × 10⁵ cells per ml in a 24-well plate at day 0, and cell counts were taken over a period of 6 days. Data shown are means ± standard deviations for three replicates.

FIG. 9.

RNA and cell surface protein expression of CXCR4 in 3B⁻ LCLs from multiple donors. (A) Average expression levels of CXCR4 mRNA in 3B⁻ LCLs, relative to expression levels in wild-type LCLs. Multiple independent wild-type and 3B⁻ LCLs were generated from four different donors, and quantitative RT-PCR for CXCR4 was performed on RNA isolated from these cells. Average C_T values for CXCR4 were normalized to GAPDH levels, and data displayed are changes (n-fold) in expression levels in 3B⁻ LCLs relative to expression levels in wild-type LCLs. Data are means ± standard deviations for independent experiments performed in triplicate on independently derived 3B⁻ LCLs. (B) Cell surface expression of CXCR4 protein in wild-type (wt) and 3B⁻ LCLs from different donors. Wild-type and 3B⁻ LCLs were stained for CXCR4 expression by using a PE-conjugated antibody directed against CXCR4 (clone 1D9). A PE-conjugated, isotype-matched antibody was used as a control. The shaded area shows CXCR4 staining of wild-type LCLs, and the bold line represents staining of 3B⁻ LCLs. The mean fluorescence intensities (MFI) of each sample are indicated on their respective graphs.

FIG. 10.

Increased CXCR4 expression on 3B⁻ LCLs restores their ability to migrate in response to ligand. Wild-type and 3B⁻ LCLs from two different donors (D9 and D13) were loaded into the upper chamber of a Transwell insert, and their chemotactic responses to 100 ng/ml CXCL12 (SDF-1α) were compared after 4 h. Percent migration was calculated as the fraction of total cells migrating across the Transwell filter, after subtracting the number of cells migrating in response to medium alone.