Partial purification and substrate analysis of bacterially expressed HIV protease by means of monoclonal antibody.

Retroviruses code for a specific protease which is essential for polyprotein precursor processing and viral infectivity. The HIV-specific protease has been predicted to be an aspartic protease which is located at the amino terminus of the pol gene. We have prepared several constructs for bacterial expression of the protease. Two of them span the whole protease region and result in its autocatalytic activation. Analysis of the dynamics of this activation indicates a two-step process which starts at the carboxy terminus and ends at the amino terminus of the protease. The activated protease is a molecule of 9 kd as evidenced by monoclonal antibody in immunoblot analysis. A construct in which the carboxy terminus of the protease is deleted results in a stable, enzymatically inactive 27-kd protein which proved useful as substrate since it contains one of the predicted cleavage sites. The stability of this protein indicates that the carboxy-terminal sequences of the protease are essential for its activity and its autocatalytic activation. The protease which is very hydrophobic was solubilized by acetone treatment and passaged over ultrogel and propylagarose columns for partial purification. It elutes as a dimer and tends to aggregate. It is inhibited by pepstatin A in agreement with its expected active site and its theoretical classification as aspartic protease. Cleavage of the gag precursor results in the mature capsid protein, p17. The protease does not, however, cleave the denatured 27-kd substrate or the denatured gag precursor. Therefore its specificity appears to be not solely sequence- but also conformation-dependent. This property needs to be taken into account for the development of protease inhibitors for therapy of AIDS.