Background:  The majority of new HIV infections worldwide result from the transmission of virus from male genital secretions (MGS). HIV exists in MGS in 2 forms:  cell-free HIV RNA (CF-RNA) in seminal plasma, and cell-associated HIV DNA (CA-DNA) in seminal lymphocytes. It is not known which of these reservoirs is the source of infection after sexual exposure.

Methods:  We studied 4 transmission pairs from a well-characterized cohort of recently HIV-infected men who have sex with men (MSM). Transmitting partnerships were verified by <1% genetic distance between bulk pol sequences of partners. Single genome amplification and sequencing of the C2V3 region of env was performed on RNA extracted from the blood of recipient partners. Clonal amplification and sequencing of the C2V3 region of env was also performed on the CF-RNA and CA-DNA extracted from MGS of source partners. At least 20 sequences from each sample were manually edited, aligned (MUSCLE), and used for phylogenetic analysis in a 1000-bootstrap-supported neighbor-joining HKY model (Geneious). For each source’s MGS, transmitting and non-transmitting amino acid sequences were compared using the Viral Epidemiology Signature Pattern Analysis program with a threshold set at 1.0, and then analyzed using the Shannon Entropy-Two program (LANL) with 1,000 randomizations to determine statistical confidence for the signature sequences identified.

Results:  Recipients’ blood HIV RNA was collected a mean of 59 days (range 21 to 85, SD 27) from the recipient’s estimated date of infection (REDI). Sources’ MGS were collected a mean of 72 days (range 19 to 133, SD 44) from REDI. Recipients’ blood HIV RNA sequences clustered completely and with 100% bootstrap support with their respective sources’ CF-RNA sequences and separately from CA-DNA sequences in all cases. Associated with transmission were 6 amino acid signatures (p <0.001 for each).

Conclusions:  In all studied cases of sexual transmission of HIV between MSM, the virus establishing infection in the recipient originated from the source’s seminal plasma. Virus in MGS that was transmitted could be distinguished from non-transmitted virus by an amino acid signature in the C2V3 region of env, which could be a target for development of a preventative vaccine. Further analyses are required to investigate the origin of seminal plasma virus in sources and to assess what residues correlate with the transmission signature among a larger number of transmission pairs.