Viruses do not encode translation machinery and thus operate under the constraints of host protein synthesis. However, the compact nature of viral genomes has resulted in the evolution of specialized strategies to maximize their coding capacity. The KSHV genome, for example, contains a scarcity of poly(A) sites, often with a single signal allocated for several consecutive open reading frames (ORFs). In many cases alternative splicing or gene-specific promoters allow production of nested sets of transcripts, enabling most KSHV ORFs to be positioned as the 5’ gene for normal cap-dependent translation. However, there exist some genes that appear to be translated as downstream genes from polycistronic mRNAs. We are currently exploring noncanonical mechanisms that enable such genes to be efficiently recognized by the host translation machinery.