Human CMV transcripts: an overview.

The human CMV (HCMV) genome consists of an approximately 230-kb dsDNA and is predicted to contain over 165 open reading frames. Although the entire sequence of the laboratory-adapted AD169 strain of HCMV was first available in 1991, the precise number and nature of viral genes and gene products are still unclear. Fewer than 100 predicted genes have been convincingly elucidated with respect to their expression patterns, transcript structure and transcription characteristics. The high gene number of HCMV creates a crowded genome with many overlapping transcriptional units. 3´- or 5´-coterminal overlapping polycistronic transcripts could use a common promoter element or a poly-A signal. 3´-coterminal monocistronic transcripts could encode 'nested' open reading frames, which possess different initiation but the same termination sites. As a virus with eukaryotic cells as the host, HCMV has the capacity to splice out introns during transcription. Major alternately spliced mRNA species of HCMV originate primarily, but not exclusively, from the immediate early gene regions. Alternate splicing patterns of the mRNAs could encode a number of gene products with different sizes. In recent years, some antisense and noncoding transcripts of HCMV have been reported. These RNAs probably have functions in genomic replication or the regulation of gene expression.