A targeted covalent small molecule inhibitor of HIV-1 fusion

Abstract

We describe a low molecular weight covalent inhibitor targeting a conserved lysine residue within the hydrophobic pocket of HIV-1 glycoprotein-41. The inhibitor bound selectively to the hydrophobic pocket and exhibited an order of magnitude enhancement of anti-fusion activity against HIV-1 compared to its non-covalent counterpart. The findings represent a significant advance in the quest to obtain non-peptide fusion inhibitors.

Figure 1.

Docked pose of 1 in the hydrophobic pocket of gp41 (PDB 2xra), produced by induced fitting (Schrodinger Inc.) allowing movement of side chains of Gln 575 and Trp 571. A salt bridge from one carbonyl oxygen on the ligand to Lys574-εNH₂ is shown as an orange dotted line, and a hydrogen bond is predicted from the second carbonyl oxygen to the lysine εNH₂.

Figure 2.

Aromatic ¹H NMR spectra of 1 and 2. In the STP ester, downfield chemical shifts occurred for one of the indole rings, while the other was unchanged. The tentative assignment shown for well-resolved resonances was based on the predicted 3D structure of 2 (Supplementary Information).

Figure 3.

A. Protein construct C26(L4)N50 used to evaluate pocket binding and adduct formation. NHR and CHR domains are swapped relative to their order in gp41, shown directly below the sequence. Boundaries of the different domains of gp41 are given (HXB2 numbering) as well as the length of the NHR (50 residues) and CHR (39 residues). The CHR is truncated in the construct and the two lysine residues are bolded. B. Schematic figure of folded construct, revealing an exposed HP (residues 565–581). C. Maldi-ESI-MS spectra of reaction mixtures of 2 with C26(L4)N50 at the indicated stoichiometries., shown over the range 7 – 15 kDa. At stoichiometry 0, DMSO was used in the reaction.

Figure 4.

A, B. Dose response curves for 1 (black) and 2 (green) in an HXB2 - pseudotyped virus infection assay (A) and HXB2-Env mediated cell – cell fusion assay (B). IC₅₀’s were obtained by fitting luminescence data (solid symbols, solid lines) and cell viability was measured by fluorescence using resazurin reagent (open symbols, dashed lines). Experiments were repeated in sextuplicate and error bars shown are the standard deviation. C, D. Dose response curves for 1 (C) and 2 (CD) in infection assays using HXB2-Env (black, solid symbols, solid lines), JRFL-Env (black, open symbols, dashed lines) or A-MLV (blue, solid symbols, solid lines) pseudotyped virus. Corresponding cell viability is shown (blue, open symbols, dashed lines).