Establishment of persistent Epstein-Barr virus (EBV) infection requires transition from a program of full viral latency gene expression (latency III) to one that is highly restricted (latency I and 0) within memory B lymphocytes. It is well established that DNA methylation plays a critical role in EBV gene silencing, and recently the chromatin boundary protein CTCF has been implicated as a pivotal regulator of latency via its binding to several loci within the EBV genome. One notable site is upstream of the common EBNA gene promoter Cp, at which CTCF may act as an enhancer-blocking factor to initiate and maintain silencing of EBNA gene transcription. It was previously suggested that increased expression of CTCF may underlie its potential to promote restricted latency, and here we also noted elevated levels of DNA methyltransferase 1 (DNMT1) and DNMT3B associated with latency I. Within B-cell lines that maintain latency I, however, stable knockdown of CTCF, DNMT1, or DNMT3B or of DNMT1 and DNMT3B in combination did not result in activation of latency III protein expression or EBNA gene transcription, nor did knockdown of DNMTs significantly alter CpG methylation within Cp. Thus, differential expression of CTCF and DNMT1 and -3B is not critical for maintenance of restricted latency. Finally, mutant EBV lacking the Cp CTCF binding site exhibited sustained Cp activity relative to wild-type EBV in a recently developed B-cell superinfection model but ultimately was able to transition to latency I, suggesting that CTCF contributes to but is not necessarily essential for the establishment of restricted latency. E pstein-Barr virus (EBV) establishes a lifelong, largely quiescent (latent) infection within B lymphocytes of its human host.This requires the concerted actions of the viral latency-associated genes, several of which are believed to facilitate a germinal center (GC)-like reaction to promote differentiation of infected B cells into ones phenotypically defined as memory B cells and which serve as the primary reservoir of EBV within persistently infected individuals (reviewed in reference 59). During the establishment of latency in vivo, infected B cells must transition through several programs of EBV latency gene transcription, beginning with expression of the full complement of latency proteins (the latency III program), i.e., six nuclear antigens (EBNAs) and three integral plasma membrane proteins (LMPs), that is associated with a rapid EBV-induced expansion of infected cells. Thereafter, expression proceeds through a more restricted program limited to EBNA1, LMP1, and LMP2 (latency II) and ultimately to a complete restriction of EBV protein expression in the memory B cell (latency 0 [alternatively, the latency program]) (reviewed in reference 44). During subsequent periods of limited cell division, reactivation of expression of the EBV genome-maintenance protein EBNA1 alone (latency I) occurs to ensure against loss of the episomal viral genome (12).With the exception of latency 0, each of the viral latency prog...

An ordered silencing of Epstein-Barr virus (EBV) latency gene transcription is critical for establishment of persistent infection within B lymphocytes, yet the mechanisms responsible and the role that the virus itself may play are unclear. Here we describe two B-cell superinfection models with which to address these problems. In the first, Burkitt lymphoma (BL) cells that maintain latency I, when superinfected, initially supported transcription from the common EBNA promoters Wp and Cp (latency III) but ultimately transitioned to latency I (Cp/Wp silent), an essential requirement for establishment of EBV latency in vivo. We used this model to test whether the early lytic-cycle gene BHLF1, implicated in silencing of the Cp/Wp locus, is required to establish latency I. Upon superinfection with EBV deleted for the BHLF1 locus, however, we have demonstrated that BHLF1 is not essential for this aspect of EBV latency. In the second model, BL cells that maintain Wp-restricted latency, a variant program in which Cp is silent but Wp remains active, sustained the latency III program of transcription from the superinfecting-virus genomes, failing to transition to latency I. Importantly, there was substantial reduction in Wp-mediated protein expression from endogenous EBV genomes, in the absence of Cp reactivation, that could occur independent of a parallel decrease in mRNA. Thus, our data provide evidence of a novel, potentially posttranscriptional mechanism for trans-repression of Wp-dependent gene expression. We suggest that this may ensure against overexpression of the EBV nuclear antigens (EBNAs) prior to the transcriptional repression of Wp in cis that occurs upon activation of Cp.