Following acute infection, herpes simplex virus (HSV) establishes latency in sensory neurons, from which it can reactivate and cause recurrent disease. Available antiviral therapies do not affect latent viral genomes; therefore, they do not prevent reactivation following therapy cessation. One possible curative approach involves the introduction of DNA double strand breaks in latent HSV genomes by rare-cutting endonucleases, leading to mutagenesis of essential viral genes. We tested this approach in an in vitro HSV latency model using the engineered homing endonuclease (HE) HSV1m5, which recognizes a sequence in the HSV-1 gene U_L19, encoding the virion protein VP5. Coexpression of the 3′-exonuclease Trex2 with HEs increased HE-mediated mutagenesis frequencies up to sixfold. Following HSV1m5/Trex2 delivery with adeno-associated viral (AAV) vectors, the target site was mutated in latent HSV genomes with no detectable cell toxicity. Importantly, HSV production by latently infected cells after reactivation was decreased after HSV1m5/Trex2 exposure. Exposure to histone deacetylase inhibitors prior to HSV1m5/Trex2 treatment increased mutagenesis frequencies of latent HSV genomes another two- to fivefold, suggesting that chromatin modification may be a useful adjunct to gene-targeting approaches. These results support the continuing development of HEs and other nucleases (ZFNs, TALENs, CRISPRs) for cure of chronic viral infections.