A quantitative ribonuclease protection assay (RPA) was developed in order to rapidly and accurately measure the levels and timing of latency-associated transcript (LAT) expression in ganglia latently infected with wild-type and mutant herpes simplex virus (HSV). Use of this assay in parallel with measurement of viral titers in murine trigeminal ganglia demonstrated that the peak of viral replication precedes the peak and subsequent plateau of LAT expression. This plateau of LAT expression was unaltered from Day 7 through the end of the experimental period on Day 28, suggesting that LAT does not further accumulate during latency of wild-type virus. RPA analyses of trigeminal ganglia latently infected with HSV-1 mutants containing specific alterations in the LAT TATA box, cyclic AMP-response element (CRE), and both TATA and CRE were performed. Mutation of the upstream TATA box reduced LAT expression to 25% of wild-type or marker-rescued virus levels, whereas mutation of the CRE did not significantly affect LAT expression in vivo whether in the presence or absence of the TATA box. These experiments demonstrate a specific requirement for the upstream promoter TATA box for wild-type LAT expression. Further examination of the role of the CRE and the TATA box by transient expression assays suggests that the CRE is important for inducible activity and that its interaction with the TATA box requires stereospecific alignment.