Substrate recognition in HIV-1 protease: a computational study.

HIV-1 protease is a crucial enzyme for the life cycle of the human immunodeficiency virus, the retrovirus that triggers AIDS. It is well documented that HIV-1 protease mediates the cleavage of Gag, Gag-Pol, and Nef precursor polyproteins and is highly selective concerning the set of 12 different amino acid sequences that cleaves. However, the governing principles and physical parameters, which determine substrate recognition and specificity, remain poorly understood despite the many speculative proposals that abound in the literature. In fact, it has been difficult so far to circumvent the fact that protease's substrates share little sequence identity and lack an obvious consensus binding motif. We have used microsecond time scale MD simulations to quantitatively show that some sequences of the polyprotein Gag-Pol that are not cleaved (nonsubstrates) have in fact a higher affinity to the active site of HIV-1 protease than a substrate; i.e., recognition is not governed by affinity to the active site. On the basis of a detailed analysis of the results and experimental data, we propose that the recognition is based on the geometric specificity of PR:Gag and PR:Gag-Pol multiprotein complex, that selects which residues lie in the specific position that makes them accessible to the active site for cleavage.