BACKGROUND: Many participants in microbicide trials remain uninfected despite ongoing exposure to HIV-1. Determining the emergence and nature of mucosal HIV-specific immune responses in such women is important, since these responses may contribute to protection and could provide insight for the rational design of HIV-1 vaccines. METHODS AND FINDINGS: We first conducted a pilot study to compare three sampling devices (Dacron swabs, flocked nylon swabs and Merocel sponges) for detection of HIV-1-specific IgG and IgA antibodies in vaginal secretions. IgG antibodies from HIV-1-positive women reacted broadly across the full panel of eight HIV-1 envelope (Env) antigens tested, whereas IgA antibodies only reacted to the gp41 subunit. No Env-reactive antibodies were detected in the HIV-negative women. The three sampling devices yielded equal HIV-1-specific antibody titers, as well as total IgG and IgA concentrations. We then tested vaginal Dacron swabs archived from 57 HIV seronegative women who participated in a microbicide efficacy trial in Southern Africa (HPTN 035). We detected vaginal IgA antibodies directed at HIV-1 Env gp120/gp140 in six of these women, and at gp41 in another three women, but did not detect Env-specific IgG antibodies in any women. CONCLUSION: Vaginal secretions of HIV-1 infected women contained IgG reactivity to a broad range of Env antigens and IgA reactivity to gp41. In contrast, Env-binding antibodies in the vaginal secretions of HIV-1 uninfected women participating in the microbicide trial were restricted to the IgA subtype and were mostly directed at HIV-1 gp120/gp140.