Abstract
Small-molecule CD4 mimics (SMCM's) bind to the gp120 subunit of the HIV-1 envelope glycoprotein (Env) and have been optimized to block cell infection in vitro. The lack of the V1/2 and V3 loops and the presence of the β2/3 and β20/21 strands (bridging sheet) in the available structures of the monomeric gp120 core may limit its applicability as a target for further synthetic optimization of SMCM potency and/or breadth. Here, we employ a combination of binding-site search, docking, estimation of protein-ligand interaction energy, all-atom molecular dynamics, and ELISA-based CD4-binding competition assays to create, characterize, and rationalize models of first- and second-generation of SMCM's bound to the distinct, trimeric BG505 SOSIP.664 structures 4NCO and 4TVP containing V1/2 and V3 loops with no bridging sheet. We demonstrate that the in silico neutralization of the highly conserved D368 is necessary to obtain the correct orientation of SMCM in their binding site when docking against the monomeric gp120 core. The computational results correlate with IC(50)'s measured in CD4 binding competition ELISA and with K(D)'s measured on gp120 core monomer. This supports the hypothesis that the 4NCO trimeric structure represents a viable target for further SMCM's optimization with advantages over both the 4TVP trimer and gp120 core monomer. Finally, the docking protocol has been optimized to screen compounds that can clearly interact with the highly conserved residue D368, increasing the likelihood of future optimizations to arrive at SMCM's with a broader spectrum of activity.