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.