A key target in the treatment of HIV-1/AIDS has been the viral protease. Here we first studied in silico the evolution of protease resistance. Primary active site resistance mutations were found to weaken interactions between protease and both inhibitor and substrate P4-P4' residues. We next studied the effects of secondary resistance mutations， often distant from the active site， on protease binding to inhibitors and substrates. Those secondary mutations contributed to the rise of multi-drug resistance while also enhancing viral replicative capacity. Here many secondary resistance mutations were found in the HIV-1 protease substrate-grooves， one on each face of the symmetrical protease dimer. The protease active site binds substrate P4-P4' residues， while the substrate-groove allows the protease to bind residues P12-P5/P5'-P12'， for a total of twenty-four residues. The substrate-groove secondary resistance mutations were found to compensate for the loss of interactions between the inhibitor resistant protease active site and substrate P4-P4' residues， due to primary resistance mutations， by increasing interactions with substrate P12-P5/P5'-P12' residues. In vitro experiments demonstrated that a multi-drug resistant protease with substrate-groove resistance mutations was slower than wild-type protease in cleaving a peptide substrate， which did not allow for substrate-groove interactions， while it had similar activity as wild-type protease when using a Gag polyprotein in which cleavage-site P12-P5/P5'-P12' residues could be bound by the protease substrate-grooves. When the Gag MA/CA cleavage site P12-P5/P5'-P12' residues were mutated the multi-drug resistant protease cleaved the mutant Gag significantly slower， indicating the importance of the protease S-grooves in binding to substrate.