Structure-activity analysis of vinylogous urea inhibitors of human immunodeficiency virus-encoded ribonuclease H

Abstract

Vinylogous ureas 2-amino-5,6,7,8-tetrahydro-4H-cyclohepta[b]thiophene-3-carboxamide and N-[3-(aminocarbonyl)-4,5-dimethyl-2-thienyl]-2-furancarboxamide (compounds 1 and 2, respectively) were recently identified to be modestly potent inhibitors of the RNase H activity of HIV-1 and HIV-2 reverse transcriptase (RT). Both compounds shared a 3-CONH(2)-substituted thiophene ring but were otherwise structurally unrelated, which prevented a precise definition of the pharmacophore. We have therefore examined a larger series of vinylogous ureas carrying amide, amine, and cycloalkane modifications of the thiophene ring of compound 1. While cycloheptane- and cyclohexane-substituted derivatives retained potency, cyclopentane and cyclooctane substitutions eliminated activity. In the presence of a cycloheptane ring, modifying the 2-NH(2) or 3-CONH(2) functions decreased the potency. With respect to compound 2, vinylogous ureas whose dimethylthiophene ring contained modifications of the 2-NH(2) and 3-CONH(2) functions were investigated. 2-NH(2)-modified analogs displayed potency equivalent to or enhanced over that of compound 2, the most active of which, compound 16, reflected intramolecular cyclization of the 2-NH(2) and 3-CONH(2) groups. Molecular modeling was used to define an inhibitor binding site in the p51 thumb subdomain, suggesting that an interaction with the catalytically conserved His539 of the p66 RNase H domain could underlie inhibition of RNase H activity. Collectively, our data indicate that multiple functional groups of vinylogous ureas contribute to their potencies as RNase H inhibitors. Finally, single-molecule spectroscopy indicates that vinylogous ureas have the property of altering the reverse transcriptase orientation on a model RNA-DNA hybrid mimicking initiation plus-strand DNA synthesis.

FIG. 1.

(A) Structures of the HIV-1 RNase H inhibitors 2-amino-5,6,7,8-tetrahydro-4H-cyclohepta[b]thiophene-3-carboxamide (compound 1) and N-[3-(aminocarbonyl)-4,5-dimethyl-2-thienyl]-2-furancarboxamide (compound 2). (B and C) Effects of order-of-addition on RNase H activity of compound 1. (B) No inhibitor. ○, preincubation of RT with Mg²⁺, hydrolysis initiated with RNA-DNA hybrid; □, RT alone, hydrolysis initiated with DNA-RNA hybrid and Mg²⁺; ▴, RT preincubated with DNA-RNA hybrid, hydrolysis initiated by adding Mg²⁺. (C) With inhibitor. ○, preincubation of RT with Mg²⁺, hydrolysis initiated by adding RNA-DNA hybrid; □, preincubation of RT with inhibitor, hydrolysis initiated by adding RNA-DNA hybrid and Mg²⁺; ▴, preincubation of RT with DNA-RNA hybrid, hydrolysis initiated by adding Mg²⁺ and inhibitor; ⋄, preincubation of RT with DNA-RNA hybrid and inhibitor, hydrolysis initiated by adding Mg²⁺.

FIG. 2.

Inhibition of HIV-1 RNase H activity by vinylogous ureas containing cycloalkyl-substituted 2-amino-thiophene-3-carboxamides. IC₅₀ values are reported only for HIV-1 RT and are the averages of triplicate assays.

FIG. 3.

Inhibition of HIV-1 RNase H activity by 2-amino- and 3-carboxamide-substituted cyclohepta[b]thiophenes. IC₅₀s are reported only for HIV-1 RT and are the averages of triplicate assays.

FIG. 4.

Inhibition of HIV-1 RNase H activity by 2-amino-substituted 4,5-dimethylthiophene-3-carboxamides. IC₅₀s are reported only for HIV-1 RT and are the averages of triplicate assays.

FIG. 5.

(A) Eliminating the 4-NO₂ of compound 16 (compound 23) or replacing it with an —OCH₃ function (compound 24) eliminates its inhibitory potency. (B) Compound 16 specifically inhibits the RNase H activity of HIV-1 RT. DNA-dependent DNA polymerase activity was determined at inhibitor concentrations of 5 μM (lane 3) and 50 μM (lane 4). Lane 1, no enzyme; lane 2, no inhibitor. The migration positions of the unextended and fully extended primer are indicated.

FIG. 6.

(A) Effect of temperature change upon affinity of compound 1 (•) and β-thujaplicinol (○) for p66/p51 HIV-1 RT. (B) ΔG, ΔH, and ΔS estimates for inhibitor binding and thermodynamic profiles for inhibition of HIV-1 RNase H by compound 1 and β-thujaplicinol. ΔG, ΔH, and ΔS are represented by open, gray, and black rectangles, respectively.

FIG. 7.

DNA polymerase and RNase H inhibitors can influence HIV-1 RT orientation on the PPT-containing RNA-DNA hybrid. (A) Cartoon depicting the high and low FRET states assumed by RT on the nucleic acid duplex. RNA and DNA strands are depicted in pink and black, respectively. 5′ termini are indicated by filled circles, and 3′ termini are indicated by arrows. RNase H-deficient HIV-1 RT was derivatized with Cy3 (green) on the p66 RNase H C terminus and bound to a surface-immobilized hybrid labeled with Cy5 near the DNA template 3′ terminus (red). The fingers, thumb, and RNase H domains of RT are indicated F, T, and H, respectively. (Upper cartoon) Enzyme adopting a polymerizing orientation with the fingers domain of the polymerase subunit over the PPT 3′ terminus brings Cy3 and Cy5 into proximity, yielding a high FRET signal. (Lower cartoon) separation of the fluorophores when the RNase H domain is in the vicinity of the primer 3′ terminus produces a low FRET signal (1, 20). (B) FRET histograms depicting enzyme orientation in the absence of inhibitor (black trace), in the presence of nevirapine (gold trace), or in the presence of RNase H inhibitor compound 1 (green trace). (C and D) FRET histograms depicting RT orientation in the presence of compounds 14 and 16, respectively.

FIG. 8.

Lowest-energy docking conformers of select vinylogous ureas on the surface of HIV-1 RT. AutoDock 4.2 positions compound 1 (A and B) and compound 16 (C and D) in the conformations shown within a small binding pocket located at the junction between the p51 thumb subdomain (green) and the RNase H domain (magenta). A surface representation of this pocket is depicted in panels A and C, while panels B and D highlight residues in close proximity to or forming hydrogen bonds with the docked conformers of compounds 1 and 16, respectively.