Identification of Thiourea-Based Inhibitors of the B-Cell Lymphoma 6 BTB Domain via NMR-Based Fragment Screening and Computer-Aided Drug Design

Abstract

Protein-protein interactions (PPI) between the transcriptional repressor B-cell lymphoma 6 (BCL6) BTB domain (BCL6^BTB) and its corepressors have emerged as a promising target for anticancer therapeutics. However, identification of potent, drug-like inhibitors of BCL6^BTB has remained challenging. Using NMR-based screening of a library of fragment-like small molecules, we have identified a thiourea compound (7CC5) that binds to BCL6^BTB. From this hit, the application of computer-aided drug design (CADD), medicinal chemistry, NMR spectroscopy, and X-ray crystallography has yielded an inhibitor, 15f, that demonstrated over 100-fold improved potency for BCL6^BTB. This gain in potency was achieved by a unique binding mode that mimics the binding mode of the corepressor SMRT in the aromatic and the HDCH sites. The structure-activity relationship based on these new inhibitors will have a significant impact on the rational design of novel BCL6 inhibitors, facilitating the identification of therapeutics for the treatment of BCL6-dependent tumors.

Figure 1.

Chemical structures of previously reported inhibitors for the BCL6^BTB.

Figure 2.

(A) Chemical structure of hit 7CC5; (B) NMR titration experiment to determine K_D value for binding of 7CC5 and BCL6^BTB; (C) Superposition of the ¹H-¹⁵N HSQC spectra of 150 μM BCL6^BTB with 5% DMSO (black) and 500 μM 7CC5 (red); (D) Crystal structure of BCL6^BTB-7CC5 complex (PDB 6C3N). The mFo-DFc electron density map contoured at 3.0 sigma is shown for 7CC5 and selected protein residues involved in the interactions with 7CC5 are shown as sticks.

Figure 3.

The crystal binding mode of 7CC5 overlaid with the SILCS apolar (green), H-bond donor (blue), H-bond acceptor (red), negative (orange) and positive (cyan) FragMaps. FragMaps are shown at contour levels of −1.0 kcal/mol for the generic apolar H-bond donor and H-bond acceptor maps and at −1.5 kcal/mol for the aromatic, aliphatic, negative and positive maps.

Figure 4.. Characterization of the binding of 15a and 15f to BCL6^BTB.

(A) Chemical structures of 15a and 15f; (B) NMR titration experiment to determine K_D values for binding of 15a and 15f to BCL6^BTB; (C) Binding affinity of 15f to BCL6^BTB determined using isothermal titration calorimetry; (D) Crystal structure of BCL6^BTB-15a complex (PDB 6CQ1). The 2mFo-DFc electron density map contoured at 1.0 sigma is shown for 15a and selected protein residues involved in the interactions with 15a are shown as sticks; (E) Crystal structure of BCL6^BTB-15f complex (PDB 6C3L). The 2mFo-DFc electron density map contoured at 1.0 sigma is shown for 15f and selected protein residues involved in the interactions with 15f are shown as sticks; (F) Superposition of compounds 7CC5 (magenta), 15a (yellow), and 15f (cyan) from BCL6^BTB-complex crystal structures. HDCH site is labeled.

Figure 5.

SILCS predicted (magenta) and X-ray crystal (yellow) binding orientations of compound 15f overlaid on the SILCS FragMaps and the BCL6^BTB crystal structure. Apolar (green), H-bond donor (blue) and acceptor (red), and positively charged (cyan) SILCS FragMaps are shown. H-bond donor and H-bond acceptor FragMaps are set to a cutoff of −0.6 kcal/mol, while Apolar and positively charged FragMaps are set to a cutoff of −1.5 kcal/mol.

Figure 6.

Correlation plots for (A) Lg(α%) with LGFE scores and (B) NMR perturbation data with LGFE scores. Correlation coefficient R² and predictive index (PI) value are shown. Lg(α%) represents the common logarithm (log10) value of remaining AlphaLisa signal (in percentage) in the presence of an inhibitor (50 µM).

Scheme 1.

Synthesis of compounds 7a-i.^a

Scheme 2.

Synthesis of inhibitors 13, 15a-h.^a