1-(-D-Ribofuranosyl)-2,5,6-trichlorobenzimidazole (TCRB) and its 2-bromo analog, BDCRB, are potent and selective inhibitors of human cytomegalovirus (HCMV) DNA processing and packaging. Since they are readily metabolized in vivo, analogs were synthesized to improve biostability. One of these, 1-(-L-ribofuranosyl)-2-isopropylamino-5,6-dichlorobenzimidazole (1263W94; maribavir), inhibits viral DNA synthesis and nuclear egress. Resistance to maribavir was mapped to UL97, and this viral kinase was shown to be a direct target of maribavir. In the present study, an HCMV strain resistant to TCRB and BDCRB was passaged in increasing concentrations of maribavir, and resistant virus was isolated. This strain (G2) grew at the same rate as the wild-type virus and was resistant to both BDCRB and maribavir. Resistance to BDCRB was expected, because the parent strain from which G2 was isolated was resistant due to known mutations in UL56 and UL89. However, no mutations were found in UL97 or other relevant open reading frames that could explain resistance to maribavir. Because sequencing of selected HCMV genes did not identify the resistance mutation, a cosmid library was made from G2, and a series of recombinant G2 wild-type viruses were constructed. Testing the recombinants for sensitivity to maribavir narrowed the locus of resistance to genes UL26 to UL32. Sequencing identified a single coding mutation in ORF UL27 (Leu335Pro) as the one responsible for resistance to maribavir. These results establish that UL27 is either directly or indirectly involved in the mechanism of action of maribavir. They also suggest that UL27 could play a role in HCMV DNA synthesis or egress of HCMV particles from the nucleus.