ReaxFF-Guided Optimization of VIRIP-Based HIV-1 Entry Inhibitors

Abstract

Peptides hold great promise for safe and effective treatment of viral infections. However, their use is often constrained by limited efficacy and high production costs, especially for long or complex peptide chains. Here, we used ReaxFF molecular dynamics (MD) simulations to optimize the size and activity of VIRIP (Virus Inhibitory Peptide), a naturally occurring 20-residue fragment of α1-antitrypsin that binds the HIV-1 GP41 fusion peptide (FP), thereby blocking viral fusion and entry into host cells. Specifically, we used the NMR structure of the complex between an optimized VIRIP derivative (VIR-165) and the HIV-1 gp41 FP for ReaxFF-guided in silico analysis, evaluating the contribution of each amino acid in the interaction of the inhibitor with its viral target. This approach allowed us to reduce the size of the HIV-1 FP inhibitor from 20 to 10 amino acids (2.28-1.11 kDa). HIV-1 infection assays showed that the size-optimized VIRIP derivative (soVIRIP) retains its broad-spectrum anti-HIV-1 capability and is nontoxic in the vertebrate zebrafish model. Compared to the original VIRIP, soVIRIP displayed more than 100-fold higher antiviral activity (IC₅₀ of ∼120 nM). Thus, it is more potent than a dimeric 20-residue VIRIP derivative (VIR-576) that was proven safe and effective in a phase I/II clinical trial. Our results show that ReaxFF-based MD simulations represent a suitable approach for the optimization of therapeutic peptides.

Figure 1

C- and N-terminal interactions mediate binding of VIR-148 to the HIV-1 gp41 FP. (A) Schematic presentation of HIV-1 attachment and membrane anchoring (upper panel). HIV-1 binds to CD4 and a coreceptor (CXCR4 or CCR5) on CD4+ T cells. The viral gp41 FP (red) is blocked by VIRIP. The lower panel shows VIR-148 (modified from PDB 2JNR(26)) bound to the HIV-1 gp41 FP. (B) Predicted contribution of each amino acid in VIR-148 to the interaction energy with gp41 FP (left panel). ReaxFF MD simulations for selected alanine mutations: E2A, P10A, and F17A (right panel).

Figure 2

Short N- and C-terminal fragments linked by a cysteine-bridge show efficient anti-HIV-1 activity. (A) TZM-bl reporter cells were incubated with increasing concentrations of VIR-148 or the indicated individual or mixed N- and C-terminal fragments (left). Cells were subsequently infected with HIV-1 NL4-3, and beta-galactosidase activity was determined 3 days postinfection. ReaxFF-based molecular dynamics simulation of the interaction between the indicated C–C-linked short VIR-148 fragments and the HIV-1 gp41 FP (middle). The structure after 150 ps (left) and ReaxFF MD simulation of the total interaction energy between cysteine-bridge-coupled split VIR-148 and the HIV-1 gp41 FP (right). (B) Mass spectrometry analysis of an equimolar mix of the two cysteine-containing VIR-148 fragments in cell culture medium.

Figure 3

Size optimization of VIR-148. (A) TZM-bl reporter cells were incubated with increasing concentrations of VIR-148, VIR-102, or the indicated internally truncated itVIRIP derivatives. Cells were subsequently infected with HIV-1 NL4-3, and beta-galactosidase activity was determined 3 days after infection. (B) Hemolytic effect of indicated compounds on human erythrocytes. Each tile represents the area under the curve (AUC) for hemolytic activity. Full hemolysis is represented by a Triton-X control. (C) Amino acid alignment of the FP regions (red) of HIV-1 CH042, CH077, SIVmac239 and HIV-2 ROD10. (D) TZM-bl cells were incubated with increasing concentrations of VIR-148 or soVIRIP and subsequently infected with the HIV-1 CH042 or CH077, SIVmac239 or HIV-2 ROD10. (E) Exemplary ReaxFF structure of VIR-148 (blue) and soVIRIP (orange) binding to the HIV-1 gp41 FP (surface, red) after 150 ps of simulation time. The contribution of each amino acid to the binding interaction is indicated by color saturation. (F) Time-resolved ReaxFF-based molecular dynamics simulation of the interaction between soVIRIP and the HIV-1 FP. The trajectories are shown for VIR-148 (blue) and soVIRIP (orange). (G) Correlation between the free energy contribution of each amino acid of soVIRIP with the corresponding amino acid in VIR-148, based on ReaxFF simulations.

Figure 4

soVIRIP is not toxic to embryonic zebrafish. Twenty-four hours post fertilization, dechorionated zebrafish embryos were exposed to DMSO, NRC-03 (cytotoxic control), abamectin (neurotoxic control), or increasing amounts of the indicated peptide for 24 h. Data shown are derived from 60 embryos per group, sampled in two independent experiments.