Background:  APOBEC3G and F are primarily expressed in the lymphoid and myeloid cell lineage and a variety of human tissues. APOBEC3G is located in the cell cytoplasm of non-permissive cells and deaminates the C residues to U residues on the minus strand of DNA, resulting in either the degradation of the HIV DNA or HIV hypermutation in HIV-1 gag, env, nef, and LTR. Here we studied the relevance of G-A hypermutation in relation to expression levels of APOBEC3G levels in vivo, in the blood of an HIV-infected non-progressor, in the context of HIV dementia and non-progression of HIV disease Although nothing is known regarding the association of G-A hypermutation with HIV dementia, an inverse correlation between APOBEC3G mRNA levels and HIV viral load has been seen in HIV patients.

Methods:  Full-length viral sequences (>9.4 kb) were amplified by polymerase chain reaction (PCR) and sequences were obtained over time from blood and cerebrospinal fluid (CSF) of an HIV⁺ non-progressor, who has been infected with a non-evolving/attenuated strain in the blood and a normally replicative HIV strain in the CSF for >20 years. The HYPERMUT Program Package was used to determine hypermutation. APOBEC3G expression levels were measured by Western blot analysis of whole peripheral blood mononuclear cells (PBMC) extracts. Band intensities were quantitated by densitometry (Imagequant program).

Results:  Over time, full genome sequencing showed distinct patterns of gag gene G-A hypermutation between blood- and CSF-derived virus. This is the first demonstration of this phenomenon. While the gag-pol gene was interrupted by 5 major stop codons in blood-derived virus as a consequence of extensive G-A hypermutation, showing replication incompetence, the CSF virus was devoid of G-A hypermutation, showing normal replicative capacity. Although the absence of hypermutation in the CSF-derived HIV correlated with normal replicative ability of the HIV and development of dementia in the study patient, the measurement of APOBEC3G levels in peripheral blood cells from this long-term non-progressors and control progressor failed to reveal significant difference in the levels of APOBEC3 protein in vivo and a direct relationship with G-A hypermutation.

Conclusions:  Our findings support the hypothesis that the hypermutated HIV genomes evolve slowly or as a result of G-A hypermutation and can induce replication incompetence in HIV, which may be related to differential compartmental evolution of HIV within the same individual. It may have immense significance in long-term concealment of HIV in vivo and its reactivation.