The FDA-approved tests for diagnosis of HIV exposure depend on detection of specific antibody in serum. HIV infection is missed in some individuals because they score seronegative by the standard clinical EIA and Western blot assays. This apparent immunological "silent" period following infection may last for months and has been reported to be as long as 3 years in rare cases. Is there truly a lack of an immune response or is there a more subtle, narrowly focused antibody response in these HIV-infected individuals which is not detected by the current tests? Using a nondenaturing serological assay (immunofluorescence of live infected T-cells), we found that each of four infected individuals "seronegative" by the standard tests did possess antibody against native HIV proteins expressed on infected cells. These antibodies reacting with native HIV antigenic epitopes were of the IgG isotype, they cross-reacted with many, but not all, of seven random HIV-1 isolates, and one of the sera immunoprecipitated HIV gp160 from NP-40-solubilized infected cells. These results show that seronegative, high-risk, infected individuals can actually be seropositive and that different types of assays using native antigenic epitopes may be required for screening. Implementation of these findings thus may decrease HIV transmission. These results also highlight the importance of protein conformation for many natural viral antigenic epitopes.