Inhibition of a viral enzyme by a small-molecule dimer disruptor

Abstract

We identified small-molecule dimer disruptors that inhibit an essential dimeric protease of human Kaposi's sarcoma-associated herpesvirus (KSHV) by screening an alpha-helical mimetic library. Next, we synthesized a second generation of low-micromolar inhibitors with improved potency and solubility. Complementary methods including size exclusion chromatography and 1H-13C HSQC titration using selectively labeled 13C-Met samples revealed that monomeric protease is enriched in the presence of inhibitor. 1H-15N HSQC titration studies mapped the inhibitor binding site to the dimer interface, and mutagenesis studies targeting this region were consistent with a mechanism where inhibitor binding prevents dimerization through the conformational selection of a dynamic intermediate. These results validate the interface of herpesvirus proteases and other similar oligomeric interactions as suitable targets for the development of small-molecule inhibitors.

Figure 1

KSHV Pr dimer interface and helical mimetic inhibitors of KSHV Pr activity. (a) The interface of monomer A (gray) and the α-helix 5 of monomer B (blue) is shown from two viewpoints. The α-helix 5 (dark gray) and other interface residues (orange) on monomer A form key contacts with α-helix 5 of monomer B. Interfacial residues Met197 (gray and blue), Trp109 (orange), and Ile201 (green) are highlighted throughout the paper. The active site (red) is located 15Å away from the dimer interface. (b) Chemical structures of KSHV Pr activity inhibitors identified by screening an α-helical mimetic library. Compounds 2-8 are structural analogues of initial hit compound 1.

Figure 2

DD2 disrupts the KSHV Pr dimer. (a) Wt KSHV Pr (5 μM) was incubated with either DMSO (dotted) or 30 μM DD2 (solid). Mixtures were loaded onto a size exclusion chromatography column, pre-equilibrated with running buffer containing either DMSO or 30 μM DD2 respectively. (b) The elution profile of KSHV Pr variant W109A (5 μM) in buffer alone indicated the absence of the dimeric species. (c) The overlaid ¹H-¹³C HSQC spectra of selectively labeled ¹³C methionine wt KSHV Pr (23 μM) in the presence of 0 μM (red), 6 μM (yellow), 12 μM (green) and, 24 μM (blue) DD2. The spectra of apo protease (red) exhibited individual resonances for the N-terminus Met1 (M1), as well as the interfacial Met197 in the dimeric and monomeric states.

Figure 3

Kinetic studies with DD2 show evidence of mixed-type inhibition. (a) A standard enzyme assay was performed using 1 μM KSHV Pr with final substrate (P6) concentration range of 80 - 2 μM, across inhibitor concentrations of 10 (orange), 7 (brown), 5 (purple), 3 (green), 1 (blue), and 0 (red) μM. Double-reciprocal plots of initial reaction velocities are shown. (b) IC₅₀ of inhibition by DD2 is higher for the tighter-K_D KSHV Pr variant. Activities of wt and M197L KSHV protease at 1 μM were monitored in the presence of increasing concentrations of DD2 (~0.14μM – 100μM). Data is plotted as the ratio of experimental end-point fluorescence (F), and un-inhibited end-point fluorescence (F0), as a function of increasing inhibitor concentration. The IC₅₀ of inhibition were 2.6 ± 0.4 μM and 7.3 ± 0.8 μM for wt (dotted) and M197L (solid) KSHV Pr respectively. Data represents mean values ± s.d. (n=3).

Figure 4

¹H-¹⁵N-HSQC titration data map the DD2 binding site to the dimer interface. (a) The overlaid ¹H-¹⁵N-HSQC spectra of the monomeric variant, KSHV Pr M197D (170 μM) in the absence (red) and presence (blue) of 170 μM DD2 is shown. The zoomed region highlights the Trp109 side-chain indole NH (boxed). The full ¹H-¹⁵N-HSQC spectra is included in Supplementary Fig. 3 online. (b) Perturbed resonances from the titration study are mapped onto a monomeric unit of the KSHV Pr crystal structure. Residues are color coded based on the extent of peak intensity loss observed at the 1 molar equivalence titration point with DD2.

Figure 5

Monomer Trap model of inhibition by compound DD2. Inhibitor binding to a monomer of KSHV Pr directly competes with the process of dimerization, which is a required step in enzyme activation. When bound to interface residues, DD2 shifts the equilibrium towards a pre-existing folding intermediate, hence trapping the protease in an inactive monomeric state.