(134d) Roles of Conserved Tryptophans in Trimerization of HIV-1 Membrane-Proximal External Regions:  Implications for Virucidal Design Via Alchemical Free-Energy Molecular Simulations

Type-1 Human Immunodeficiency Virus (HIV-1), the viral agent responsible for AIDS, leads to the death of over two million people each year and represents a global health epidemic. Despite the success of highly active antiretroviral therapies (HAART) in extending lifespans of infected individuals, prevention of transmission to uninfected individuals is still problematic. We recently developed a novel microbicidal recombinant fusion protein comprising the lectin cyanovirin-N (CVN) to the amphipathic peptide membrane-proximal external region (MPER) of gp41, attached by a flexible polypeptide linker of controllable length (Contarino et al. 2013 Antimicrob. Agents Ch. 57:4743-4750). This class of molecules, termed “dual-acting virucidal entry inhibitors” (DAVEI’s), induce poration of the virus evidently by triggering the viral fusion machinery of the envelope spike complexes to which they bind. Recent work on optimizing DAVEI’s has focused on understanding where on virus the DAVEI MPER binds, and we recently showed that the binding site is likely to involve endogenous MPER segments, and, in particular, that of the many tryptophan residues on MPER, the only one used by DAVEI appears to be the most N-terminal one (Parajuli et al. 2016; Biochemistry 55:6100-6114). In this current work, we sought to use molecular modeling and simulation to structurally rationalize this observation. Our approach consists of first building a model MPER trimer and then performing double-decoupling alchemical free-energy perturbation calculations to assess the thermodynamic importance of each tryptophan in the three sets of four tryptophans per MPER segment. We used the asymmetric MPER triple helix structure displayed by the 3G9R PDB entry (Liu et al. 2009; Biochemistry 48:2915-2923). Using MD-equilibrated explicit-water systems of both the trimer and each protomer, we use FEP to compute the reversible work associated with mutating each tryptophan separately to alanine. The difference between this work in the context of the trimer and the single protomer is the ∆∆G associated with that mutation on the the binding energy of the protomer to the trimer. We determined that the most N-terminal tryptophan (position 5) was significant on all three chains of the trimer and the the third tryptophan (residue 11) was significant on all but the one chain of the trimer (designated chain C by the original submitters), where significance is assigned for |∆∆G| > 1 kcal/mol. This single chain also happens to have several important differences in side-chain rotomers and thus intersubunit contacts relative to the other two protomers in the 3G9R structure. Observing the interaction energy between the mutant chain and the interacting wild type chains, we observed that each of the mutations resulted in a decrease in interaction energy between the mutated chain and its two bonding partners. We further observed that the mutations resulting in a significant |∆∆G|, had a larger decrease in interaction energy (< -4 kcal/mol) while non-significant mutations had a smaller interaction energy change. In light of the experimental results, we therefore posit that the binding conformation of chain C against the other two is possibly a good model for the binding of a DAVEI MPER to a (likely transiently exposed) endogenous MPER dimer on a viral envelope spike. Future work will be directed a further tests of this model using mutagenesis on viral spikes as well as assessing roles of other potentially important residues in the MPER sequence.