Latently infected resting CD4+ T cells provide a long-term reservoir for human immunodeficiency virus type 1 (HIV-1) and are likely to represent the major barrier to virus eradication in patients on combination antiretroviral therapy. The mechanisms by which viruses enter the latent reservoir and the nature of the chemokine receptors involved have not been determined. To evaluate the phenotype of the virus in this compartment with respect to chemokine receptor utilization, full-length HIV-1 env genes were cloned from latently infected cells and assayed functionally. We demonstrate that the majority of the viruses in the latent reservoir utilize CCR5 during entry, although utilization of several other receptors, including CXCR4, was observed. No alternative coreceptors were shown to be involved in a systematic fashion. Although R5 viruses are present in the latent reservoir, CCR5 was not expressed at high levels on resting CD4+ T cells. To understand the mechanism by which R5 viruses enter latent reservoir, the ability of an R5 virus, HIV-1 Ba-L, to infect highly purified resting CD4+ T lymphocytes from uninfected donors was evaluated. Entry of Ba-L could be observed when virus was applied at a multiplicity approaching 1. However, infection was limited to a subset of cells expressing low levels of CCR5 and markers of immunologic memory. Naive cells could not be infected by an R5 virus even when challenged with a large inoculum. Direct cell fractionation studies showed that latent virus is present predominantly in resting memory cells but also at lower levels in resting naive cells. Taken together, these findings provide support for the hypothesis that the direct infection of naive T cells is not the major mechanism by which the latent infection of resting T cells is established.