Abstract
The mechanism of conformationally dynamic proteins remains understudied because they are difficult to analyze structurally. For HIV-1 protease, the mechanism of cleavage by mature protease is well understood in large part because it forms a stable structure that is amenable to x-ray crystallography. However, self-cleavage or autoproteolysis of protease from the viral polyprotein involves transiently populated structures and is poorly understood. We probed autoproteolysis in HIV-1 using a yeast reporter and mutational scanning. We compared our results with mutational scanning of protease on viral fitness, which integrates both autocleavage and cutting by mature enzyme. We identified 220 mutations that were well tolerated for self-cleavage but not fitness. We analyzed three of these mutations (D30E, W42M, and P44L) using independent approaches. All three were capable of efficient self-cleavage in a bacterial assay of autoproteolysis but had strong defects in the mature form for cleavage of a peptide substrate. The separation of function mutations from the mutational scan clustered at hot-spot locations that do not impact autoproteolysis, likely because they are conformationally dynamic during self-cleavage. We used the mutational scanning results as guides for molecular simulations that provide models of autoproteolysis conformations. This work provides new perspectives on a structurally dynamic mechanism of autoproteolysis.IMPORTANCESelf-cleavage is a critical process for HIV-1 and many other viruses, but is poorly understood mechanistically because it involves a highly dynamic structure that is difficult to visualize. We used the structure-function information from a mutational scan of self-cleavage by HIV-1 protease to provide a new view of its mechanism. Models based on our mutational data indicate a core structure required for self-cleavage. The approaches used here are general and can be used to investigate self-cleavage in other viruses.