Functional identification of nucleotides conferring substrate specificity to retroviral integrase reactions.

The long terminal repeats (LTRs) that flank the retroviral DNA genome play a distinct role in the integration process by acting as specific substrates for the integrase (IN). The role of LTR sequences in providing substrate recognition and specificity to integration reactions was investigated for INs from human immunodeficiency virus type 1 (HIV-1), Moloney murine leukemia virus (M-MuLV), human T-cell leukemia virus type 1 (HTLV-1), and human T-cell leukemia virus type 2 (HTLV-2). Overall, these INs required specific LTR sequences for optimal catalysis of 3'-processing reactions, as opposed to strand transfer and disintegration reactions. It is of particular note that in strand transfer reactions the sites of integration were similar among the four INs. In the 3'-processing reaction, sequence specificity for each IN was traced to the three nucleotides proximal to the conserved CA. Reactions catalyzed by M-MuLV IN were additionally influenced by upstream regions. The nucleotide requirements for optimal catalysis differed for each IN. HIV-1 IN showed a broad range of substrate specificities, while HTLV-1 IN and HTLV-2 IN had more defined sequence requirements. M-MuLV IN exhibited greater activity with the heterologous LTR substrates than with its own wild-type substrate. This finding was further substantiated by the high levels of activity catalyzed by the IN on modified M-MuLV LTRs. This work suggests that unlike the other INs examined, M-MuLV IN has evolved with an IN-LTR interaction that is suboptimal.