Abstract
Beta- and gammaherpesviruses late transcription factors (LTFs) target viral promoters containing a TATT sequence to drive transcription after viral DNA replication has begun. Human cytomegalovirus (HCMV), a betaherpesvirus, uses the UL87 LTF to bind both TATT and host RNA polymerase II (Pol II), whereas the UL79 LTF has been suggested to drive productive elongation. Here we apply integrated functional genomics (dTag system, PRO-Seq, ChIP-Seq, and promoter function assays) to uncover the contribution of diversity in LTF target sequences in determining degree and scope to which LTFs drive viral transcription. We characterize the DNA sequence patterns in LTF-responsive and -unresponsive promoter populations, determine where and when Pol II initiates transcription, identify sites of LTF binding genome-wide, and quantify change in nascent transcripts from individual promoters in relation to core promoter sequences, LTF loss, stage of infection, and viral DNA replication. We find that HCMV UL79 and UL87 LTFs function concordantly to initiate transcription from over half of all active viral promoters in late infection, while not appreciably affecting host transcription. Both LTFs act on and bind to viral early-late and late kinetic-class promoters. Over one-third of these core promoters lack the TATT and instead have a TATAT, TGTT, or YRYT. The TATT and non-TATT motifs are part of a sequence block with a sequence code that correlates with promoter transcription level. LTF occupancy of a TATATA palindrome shared by back-to-back promoters is linked to bidirectional transcription. We conclude that diversity in LTF target sequences shapes the LTF-transformative program that drives the viral early-to-late transcription switch.