The reduced efficiency with which herpes simplex virus type 1 (HSV-1) mutants establish latent infections in vivo has been a fundamental obstacle in efforts to determine the roles of individual viral genes in HSV-1 reactivation. For example, in the absence of the "nonessential" viral immediate-early protein, ICP0, HSV-1 is severely impaired in its ability to (i) replicate at the site of inoculation and (ii) establish latency in neurons of the peripheral nervous system. The mouse ocular model of HSV latency was used in the present study to determine if the conditions of infection can be manipulated such that replication-impaired, ICP0-null mutants establish wild-type levels of latency, as measured by viral genome loads in latently infected trigeminal ganglia (TG). To this end, the effects of inoculum size and transient immunosuppression on the levels of acute replication in mouse eyes and of viral DNA in latently infected TG were examined. Clinical interest in herpes simplex virus type 1 (HSV-1) and HSV-2 centers on their ability to reactivate from latency and cause recurrent herpetic diseases such as herpes labialis, stromal keratitis, genital herpes, and opportunistic infections of immunosuppressed individuals (37, 58). Despite long-standing interest in the problem (3, 13) and significant advances in our understanding of the molecular events in HSV-1 replication (38), the events that lead from latency to reactivation remain poorly understood. Two factors that have impeded our understanding of latency and reactivation are (i) the lack of definitive in vitro models of HSV latency and (ii) the fact that animalbased models of latency are not amenable to the analysis of HSV reactivation at the molecular level. Regarding the first point, although quiescent infections can be established in several different cell types (2,9,39,47,59), the relevance of existing "in vitro latency models" to HSV latency in vivo is unclear. Regarding the second point, although the establishment of latency in animal models closely parallels the natural history of HSV infection in humans (22,41,48), the effect of eliminating a viral gene product on reactivation is difficult to study because many HSV mutants replicate poorly in animals.Comparison of the reactivation efficiencies of null mutant viruses to that of wild-type virus is a potentially powerful approach to identifying viral genes involved in HSV reactivation. The effect of a mutation in a given gene on reactivation efficiency can be measured accurately, however, only when equal numbers of mutant and wild-type viral genomes are present in latently infected ganglia at the time of reactivation. Given that the efficient establishment of latency is dependent on viral replication at the site of inoculation (27,42,43), a fundamental obstacle to the use of viral mutants to study reactivation is that mutations in many "nonessential" viral genes impair the ability of HSV-1 to replicate in animals. For example, attempts to define the roles of ICP0, ICP22, and the virion host shutoff protein in ...