Structural and Mechanistic Insights into Atypical Bacterial Topoisomerase Inhibitors

Abstract

Novel bacterial topoisomerase inhibitors (NBTIs) targeting DNA gyrase and topoisomerase IV constitute a new antibacterial class for deadly pathogens such as MRSA. While most NBTIs induce gyrase-mediated single-strand DNA breaks, a subset of amide NBTIs induces both single-strand and double-strand DNA breaks. Here, we report the X-ray crystal structures of two such amide NBTIs, 148 and 185, and demonstrate an unusual binding mode characterized by engagement of both GyrA D83 and R122. The synthesis of two isosteric triazole NBTIs is also described, one of which (342) affords only single-strand DNA breaks, while the other (276) also induces both single- and double-strand DNA breaks. A combination of docking and molecular dynamics simulations is employed to further investigate the potential structural underpinnings of differences in DNA cleavage.

Figure 1

(A) Chemical structures of NBTIs bearing a secondary amine (circled). Blue represents the DNA-binding moiety, black the linker, and red the gyrase-binding moiety. (B) Chemical structures of amide NBTIs with unusual binding mode.

Figure 2

Close-up view of interactions involving compounds 148 (A) and 185 (B). The full structure is shown in thin lines with carbon bonds in green, oxygen in red, nitrogen in blue, and sulfur in yellow. Compounds 148 and 185 are shown in thicker lines. Key gyrase side chains and DNA nucleotides are also shown as thick lines, with carbon bonds in cyan and yellow, respectively. Hydrogen bonds and other electrostatic interactions are shown as dashed lines with distances labeled. Notice that the amide groups of 148 and 185 have their dipole moments interacting favorably in a nearly linear arrangement with D83 on one side and R122 on the other. R122 also forms extensive interactions with the scissile phosphate and the adjacent phosphate in the 3′-direction. The red arrow indicates where the hydroxyl of Y123, if present (but mutated to F), would react with the phosphate that is cleaved. The Mn²⁺ atom is shown as a purple sphere.

Figure 3

Triazole-containing NBTIs as isosteres of amide-type NBTIs. Notably, 276 and 342 employ a piperidine linker moiety in place of the dioxane linker for ease of synthesis and a fluoronaphthyridine DNA-binding motif in place of a quinoline.

Figure 4

Compound 276 induction of DNA strand breaks in the presence of gyrase. (A) The ethidium-stained gel indicates positions of DNA after incubation with enzyme in the absence or presence of 276 (0.1 nM-10 μM). The various DNA forms are indicated as negatively supercoiled pBR322 DNA substrate ((−)SC), linearized DNA (Lin) representing double-strand breaks, and nicked open-circular DNA (Nick) representing single-strand breaks. Ciprofloxacin and gepotidacin were included as controls at the indicated concentrations. (B) Induction of single strand breaks (SSB) and double-strand breaks (DSB) induced by 276 (0.1 nM-100 μM). Percent DNA cleavage was calculated by assessment of the intensity of DNA bands relative to the intensity of the EcoR1 digested pBR322 band as described in Supporting Information. The percent DNA cleavage in enzyme controls was subtracted to yield the final results. Results shown are derived from 5 separate experiments run on separate days. Regression lines were generated using a 4-logistic curve fit in Sigmaplot 16 (Grafiti LLC, Palo Alto, CA).

Figure 5

Compound 342 induction of DNA strand breaks in the presence of gyrase. (A) The ethidium-stained gel indicates positions of DNA after incubation with enzyme in the absence or presence of 342 (0.1 nM-10 μM). Ciprofloxacin and gepotidacin were included as controls at the indicated concentrations. (B) Induction of single strand breaks (SSB) and double-strand breaks (DSB) by 342 (0.1 nM-100 μM) was quantified by the same method as that explained in Figure 4. Results shown are derived from 5 separate experiments run on separate days. Regression lines were generated using a 4-logistic curve fit in Sigmaplot 16 (Grafiti LLC, Palo Alto, CA).

Figure 6

Computational docking results in the S. aureus DNA gyrase crystal resolved with 148. (A) Compiled docked structures from all 14 compounds. Docked NBTIs are shown with a wire representation and binding site components shown as licorice. Symmetric binding site residues are labeled Near (magenta) and Far (yellow) based on their position relative to the bound NBTI. (B) Overlay of cocrystallized 148 (wire) with its docked pose (transparent). (C) Overlay of the aligned crystal of 185 (wire) with the docked pose (transparent). (D) Overlay of the previous docked pose of 88 in 2XCS (wire) and the newly docked pose (transparent). (E) Amide- and triazole-NBTI distance distributions to residue backbones (BB) and side chains (SC), grouped by SSB- (green) and DSB- (blue) subsets. Points show distances for individual NBTIs, color indicating an amide- or triazole-containing compound. (F) Docked poses of the two triazole NBTIs. 276 from the DSB-subset is in blue (top), and the SSB-subset 342 is in green (bottom). Statistical significance is given by an independent t test (* p < 0.05; ** p < 0.01; *** p < 0.001).

Figure 7

Molecular dynamics simulation analysis of NBTI functional group interactions with binding pocket residues. (A) Radius of gyration distributions split into SSB- and DSB-subsets. R_gyr denotes 3D radius of gyration, and ⊥R_gyr represents horizontal gyration, perpendicular to the longest axis of the NBTI. The points show trajectory-averaged values colored by NBTI linker motif. (B) Binding pocket map of the positions of the Near-side (magenta) and Far-side (yellow) residues relative to an NBTI (cyan). Dashed lines show distances measured throughout the simulations. Minimum distances between residue side chains and amide and triazole heavy-atoms were collected. (C) Aggregate distances measured throughout the simulation to the four interacting residues shown in part B. Average distances for each NBTI simulation replicate are shown as points colored by their motifs. (D) Per-compound distance distributions between R122_Near and the NBTIs shown in part C. Each distribution contains distances from three replicate trajectories. Time-series distances of 186 over replicates between (E) the amide or triazole to the side chain of R122_Near and (F) the guanidino group of R122_Near to the scissile phosphate. Curves were smoothed by rolling average with a window size of 1% total trajectory length. Gray regions indicate the equilibration period excluded from analyses. (G) Representative distances sampled from the trajectory denoted in E & F at 100 ns (top) and 150 ns (bottom). 186 is shown in blue, R122_Near in magenta, and the phosphate group oxygens in red. (H) Distances of the NBTI triazole and scissile phosphate to R122_Near, taken from the final frame of replicate 3 for 342 (top, green) and 276 (bottom, blue). (I) Per-replicate distance distributions between D83_Far and different NBTI motifs. (J) Representative interaction distances of D83_Far in yellow with SSB-inducing 89 in green (Top) and SSB- & DSB-inducing 148 in blue (Bottom). Statistical significance is given by the Mann–Whitney U-test for non-normal distributions (* p < 0.05; ** p < 0.01; *** p < 0.001).