Logo image
Back
Study of HIV-1 viral engagement of a novel virucidal molecule using molecular dynamics
Dissertation   Open access

Study of HIV-1 viral engagement of a novel virucidal molecule using molecular dynamics

Steven T. Gossert
Doctor of Philosophy (Ph.D.), Drexel University
Jun 2020
DOI:
https://doi.org/10.17918/00000051
pdf
Gossert_Steven_20202.91 MBDownloadView
PDF Open Access Open Access (License Unspecified)

Abstract

Drugs--Design HIV infections Molecular dynamics--Simulation methods Free energy perturbation
The Dual-Acting Virolytic Entry Inhibitors, or DAVEI's and the Dual-engaging Lytic Inhibitors or DLI's are both classes of recombinant chimera developed as inhibitors of HIV-1 through the targeting the the Envelope glycoprotein. DAVEI consists of a glycan-binding lectin, a flexible, polypeptide repeat unit linker and a fragment of the membrane-proximal internal region (MPER) of HIV-1 Env gp41, while DLI consists of a small molecule CD4 mimetic compound or smCD4mc, a flexible PEO repeat unit linker, and a fragment of the MPER region. The unique feature of DAVEI and DLI is the proposed mechanism, dual-engagement interaction with Env, simultaneously engaging the gp120 and gp41 regions to trigger virolysis of HIV-1 virions. DAVEI molecules proposed engagement sites are surface glycans on gp120 for the lectin and the endogenous gp41 MPER for the MPERDAVEI. DLI molecules proposed engagement sites are the CD4 binding pocket of gp120 for the smCD4mc and endogenous gp41 MPER for the DLI-MPER. For both molecules the linker generates space between the active heads of the molecule to allow for dual-engagement and is a necessary component as the active heads in mixture do not cause lysis of HIV-1 virions. In this thesis, we detail the use of molecular dynamics (MD) simulations to understand the mechanisms by which DAVEI and DLI molecules function including the MPERDAVEI · MPEREnv the use of various linker lengths to generate space between active heads compared to measured distances on HIV-1 models, the generation of fitted models of DLI molecules to determine the dual-engagement capabilities of new DLI molecules given the proposed binding regions, and to understand the interaction of the DLI smCD4mc with a lipid bilayer representative of the HIV-1 virion. First, in order to assess the MPERDAVEI · MPEREnv interaction, we considered what form this interaction would take, knowing MPEREnv forms a trimeric structure as part of HIV-1 cell entry, we hypothesized MPERDAVEI would form a mixed trimer complex with two Env MPER. We then sought to explore the stability of this structure using four amino acid point mutations via alchemical free-energy perturbation on multiple tryptophan in the structure changing them to alanine. These mutations were performed on the chain representing MPERDAVEI in the complex and the results compared to previous mutation studies of DAVEI. We found that on two of the protomers of the trimer two tryptophans (HXBc2 numbers: W666 and W672) were essential in the stability of the trimer while on the third protomer only one tryptophan (W666) was essential to the stability of the trimer. We show that this is a result of a unique placement of the W666 for this protomer which rotates the backbone of this protomer. These findings show that the unique protomer is consistent with previous DAVEI tryptophan mutation lysis studies suggesting that this model could be representative of MPERDAVEI interactions with MPEREnv trimer. Next, we used structural modelling to rationalize the dependence of DAVEI potency on the molecular length of the linker connecting the two active heads. We used temperature accelerated molecular dynamics and on-the-fly parameterization to generate free energy vs end-to-end distance for two different linker lengths, a length of zero (His6) and a length of two ([Gly4Ser]2His6). Previous studies show that DAVEI-L0 is non-lytic while DAVEI-L2 is lytic meaning that the minimum distance between the two engagement sites of DAVEI lies somewhere between DAVEI-L0 and DAVEI-L2. In addition, we generation a envelope model based on a cryo-electron microscopy-derived structure of a cleaved, soluble Env construct, with glycans added as the putative lectin engagement site and MPER added as the putative engagement site for MPERDAVEI. We used MD simulations to measure distances between the lectin C-terminus, where the linker begins, to the Env gp41, where MPERDAVEI would begin, and compared this to free energy vs end-to-end distance of the linkers. We determined that none of the glycans were close enough to MPEREnv to allow for dual-engagement of DAVEI-L0 while several would allow for DAVEI-L2 to dual engagement. These findings are consistent with previous linker length dependence studies of DAVEI and supports the proposed engagement model of DAVEI. Next, we examined the dual-engagement possibilities of DLI molecules using a model of Env and steered MD to the proposed engagement sites. We then released constraints on the system to allow the complex to relax from the engineered position to determine if dual-engagement would be maintained. We tested three different linker L1, L3 and L7 each of which were determined to be lytic. We found that all three linker lengths tested were able to maintain dual-engagement for at least one DLI on the complex although L1 required extended steered equilibration to maintain this dual-engagement which showed a high degree of flexibility in the Env structure when smaller linker lengths were used. With this model, we were able to propose that DLIs are functioning with a similar mechanism to DAVEI molecules. Finally, we examined the interaction of DLI molecules with membrane which resulted in an undesired, off-target effect. We used Umbrella Sampling and WHAM to generate a free energy vs z-distance (representative of membrane depth) for BNM-III-170 and two hydrophilic mutations of BNM-III-170 (hydroxyl-BNM-III-170 and cyano-BNM-III-170). We determined that BNM-III-170 had a minimum free energy when inserted vertically into the membrane with the hydrophobic ring embedded, while hydroxyl-BNM-III-170 still embedded in the membrane it was not to the same depth as BNM-III-170 and cyano-BNM-III-170 sat horizontally on the surface of the membrane. This supports the hypothesis that BNM-III-170 hydrophobicity is leading to an interaction with the membrane causing DLI to have an off-target effect and proposes that a hydrophilic mutation to BNM-III-170 could be used to reduce this membrane interaction and eliminate the off-target effect of DLI. Further studies should be performed on new hydrophilic BNM-III-170 structures to further decrease membrane interaction and ensure CD4 binding pocket interaction is maintained through these mutation.

Metrics

59 File views/ downloads
27 Record Views

Details

Logo image