Characterization of the CBF2 binding site within the Epstein-Barr virus latency C promoter and its role in modulating EBNA2-mediated transactivation

Abstract

The Epstein-Barr virus (EBV) EBNA2 protein is a transcriptional activator that regulates viral and cellular gene expression and is also essential for EBV-driven immortalization of B lymphocytes. The EBNA2-responsive enhancer in the viral latency C promoter (Cp) binds two cellular factors, CBF1 and CBF2. The precise role of the CBF2 protein for Cp enhancer function is presently unclear. CBF2 does not appear to interact with EBNA2 and binds just downstream of CBF1 between positions -339 and -368 in the Cp EBNA2 enhancer. Within this region an 8-bp sequence, CAGTGCGT, can be found, and a similar sequence is also located downstream of CBF1 binding sites in other EBNA2-responsive promoters. Previous studies have indicated that mutations and methylation in this sequence affect EBNA2 responsiveness. To investigate the requirements for CBF2 binding, we synthesized a series of oligonucleotides carrying double transversion mutations spanning both the conserved core sequence and outside flanking sequences. Surprisingly, mutations outside of the conserved core sequence in 4 bases immediately flanking the 5' end, GGTT, had the most deleterious effect on CBF2 binding. Mutations in the conserved core had a gradient effect, with those near the 5' end having the most deleterious effects on CBF2 binding. In addition, the affinities of CBF2 for binding to the LMP-1, LMP-2, and CD23 promoters were also measured. These promoters contain the conserved core but lack the 5' flanking GGTT motif and bound CBF2 weakly or not at all. Using Cp reporter plasmids containing CBF2 mutant binding sites, we were also able to show that at lower doses of EBNA2, Cp transactivation required a functional CBF2 binding site but that higher doses of EBNA2 transactivated CBF2 mutant promoters to 40% of wild-type levels. These assays indicate that CBF2 is important for EBNA2-mediated transactivation of the viral latency Cp. In addition, CBF2 activity was found to be associated with two polypeptides of 27 and 33 kDa.

FIG. 1

Structure of Cp reporter constructions. A schematic illustration of the EBV genomic region containing the Cp and upstream sequences is shown at the top. Indicated are the episomal origin of replication (ori P), glucocorticoid response element (GRE), EBNA2 response element (E2RE), and the KpnI and Sau3A restriction sites. The arrow indicates the site of transcription initiation. Below are shown the Cp sequences cloned into reporter vectors (Cp −1021 to +3) and a multimerized version of the EBNA2 response element corresponding to positions −330 to −430 also cloned into a reporter vector (8X Cp E2RE).

FIG. 2

CBF1 and CBF2 bind distinct sequences in the Cp EBNA2 enhancer which are conserved in other EBNA2-responsive promoters. (A) Nuclear extracts from CA46 cells were incubated with a radiolabeled Cp sequence from positions −330 to −430 in the presence or absence of competitor oligonucleotides and analyzed by EMSA. Lanes: 1, probe only; 2, CA46 extract; 3, CA46 extract and cold 30-mer oligonucleotide competitor from positions −359 to −388 (CBF1 binding element); 4, CA46 extract and cold 30-mer oligonucleotide competitor from positions −339 to −368 (CBF2 binding element). (B) Comparison of the nucleotide sequences of EBNA2-responsive promoters and locations of their putative conserved CBF2 binding sequences. Shown also is the location of the CBF1 binding sites previously characterized in these promoters. The LMP-1 promoter (LMP-1p) contains in its natural context the consensus CBF1 binding site in the antisense orientation and is shown on top as the reverse complement strand to aid in the comparison of similarities. LMP-2Ap, LMP-2A promoter; CD23p, CD23 promoter.

FIG. 3

Mutagenesis analysis of the conserved 8-bp sequence in the Cp and identification of residues crucial for CBF2 binding. Nuclear extracts were mixed with 0-, 2.5-, 5-, 25-, and 125-fold molar excesses of the different unlabeled mutant oligonucleotides. A fixed amount of ³²P-wild-type (wt) oligonucleotide was then added to the shift reaction mixture. The reaction mixtures were separated on 4.5% nondenaturing polyacrylamide gels, dried, and autoradiographed. Amounts of CBF2 complex were quantified with a Betascope 603 blot analyzer. (A) The EMSA gel shows binding of nuclear extract containing CBF2 to a 30-mer oligonucleotide probe from positions −339 to −368 alone or with increasing (triangle) amounts of cold competitor oligonucleotide of the same sequence (the wild type [wt]) or mutant 11 (mut. 11) or mutant 12 (mut. 12) oligonucleotide. The asterisk denotes a nonspecific band that does not compete with the CBF2-specific oligonucleotide probes and appears in lanes containing probe only. (B) Summary of competition results. The sequences shown represent the central 30 bp of the 36-mer oligonucleotide. The 8-nucleotide conserved sequence is shaded. The mutations consist of double transversion mutations, and they are shown in boldface type and underlined. The numbers in column a are concentrations of unlabeled mutant oligonucleotides required for 50% competition. The numbers in column b are percentages representing the ability of each oligonucleotide to compete for CBF2 relative to that of the wt element, which was set at 100%. Results are averages from three independent experiments.

FIG. 4

The LMP-1, LMP-2A, and CD23 promoters fail to bind CBF2. Competitions were carried out as described for Fig. 3. The sequences shown represent the central 30 bp of the 36-mer oligonucleotide. The 8-bp putative CBF2 binding sites are shaded. The natural changes between the sequences of the LMP-1, LMP-2A, and CD23 promoters and Cp are shown in boldface type and underlined. The numbers in column a are concentrations of unlabeled mutant oligonucleotides required for 50% competition. The numbers in column b are percentages representing the ability of each oligonucleotide to compete for CBF2 relative to that of the wild-type element, which was set at 100%. Results are averages from three independent experiments.

FIG. 5

Analysis of the EBNA2 responsiveness of wild-type (wt) and CBF2 Cp mutants. (A) Results of single-dose transfections to compare levels of EBNA2 transactivation of reporter plasmids containing eight tandem copies of the EBNA2 enhancer. DG75 cells were cotransfected with 2 μg of the target plasmid and 0.5 μg of the effector EBNA2-expressing plasmid. Results are averages from two independent experiments. mut., mutant. (B) Dose responses for comparison of levels of EBNA2 transactivation of reporter plasmids containing the Cp −1021-to-+3 sequence. Cells were transfected with 2 μg of target plasmid and 0.2, 0.4, 0.8, and 1.6 μg of effector EBNA2-expressing plasmid. Results are averages from five independent experiments. Standard errors of the means are indicated with T bars.

FIG. 6

UV-cross-linking analysis demonstrating that the CBF2 binding site specifically interacts with two polypeptides of 33 and 25 kDa. (A) Competition with different amounts of unlabeled competitors. Reaction mixtures containing UV-cross-linked proteins were resolved on sodium dodecyl sulfate–12.5% polyacrylamide gels, dried, and autoradiographed. To some cross-linking reaction mixtures cold competitor oligonucleotides were added as indicated. Lanes 1 to 3, competition with 75-, 50-, and 25-fold molar excesses, respectively, of mutant 3 (mut.3); lanes 4 to 6, competition with 75-, 50-, and 25-fold molar excesses, respectively, of the wild-type (wt) sequence; lane 7, no competition. (B) Competition with 50-fold molar excesses of wild-type, mutant 10, mutant 2, and mutant 3 oligonucleotides. Asterisks indicate specific bands of 33 and 25 kDa interacting with the CBF2 probe. Lane 1 contains no competitor.