A genomic and bioinformatics analysis of the integration of HIV in peripheral blood mononuclear cells.

The mechanistic basis of the target-site preference of lentivirus DNA integration is not well understood. In the present in silico study, we describe the integrational profile of simultaneous HIV-1 and HIV-2 infection. A total of 352 genomic DNA sequences from human peripheral blood mononuclear cells (PBMCs) obtained from GenBank and possessing the 5' LTR of HIV were used to characterize the structure and composition of local chromatin associated with high frequency integration sites. These sequences were aligned with the draft human genome (hg18) using BLAST (NCBI) and BLAT (UCSC) in order to derive information about chromosome localization, functional aspects of coding protein genes, CpG island number, and repetitive elements flanking integration sites. No significant differences in the integrational profile between HIV-1 and HIV-2 were found. However, we observed a tendency in both lentiviruses to integrate in the vicinity of protein coding genes. Multiple regression analysis showed a strong correlation between the number of genes and the number of CpG islands in regions with high integration frequency, mainly in chromosome 17 (R = 0.95, p < 0.05). Our results provide strong evidence that HIV-1 and HIV-2 have common genomic environments in the local chromatin regions with high gene density and CpG islands. The understanding of local genomic environments with a high frequency of integration would be the starting point to develop novel antiviral strategies for lentiviral infection.