Diversity-Oriented Synthesis as a Strategy for Fragment Evolution against GSK3β

Abstract

Traditional fragment-based drug discovery (FBDD) relies heavily on structural analysis of the hits bound to their targets. Herein, we present a complementary approach based on diversity-oriented synthesis (DOS). A DOS-based fragment collection was able to produce initial hit compounds against the target GSK3β, allow the systematic synthesis of related fragment analogues to explore fragment-level structure-activity relationship, and finally lead to the synthesis of a more potent compound.

Keywords: Diversity oriented synthesis; GSK3β; fragment growing; fragment-based drug discovery.

Figure 1

Logical chemical synthesis strategy related to the fragments would impact optimization and modification of fragment hits. Fragments constructed using highly modular syntheses would enable rapid substitutions around a fragment core. Therefore, the generation of fragment-SAR using modular fragment syntheses would aid the fragment evolution process as a complementary approach to structure-guided approaches.

Figure 2

(a) Three different build/couple/pair pathways leading to a library of skeletally and stereochemically diverse fragments. (b) Fragment hits from a DSF screening at 2.5 mM against GSK3β. Compounds displaying a thermal shift ≥0.5 °C are shown.

Figure 3

(a) Synthesis of 1S as a representation of the B/C/P pathway. (b) DSF and biochemical assay results of analogues of 1S. (c) Chemical modifications on 1R to identify key binding interactions and probe potential pocket for fragment growing. DSF assays were performed at 2.5 mM unless otherwise indicated; biochemical assays were performed at 1 mM.

Figure 4

(a) STD NMR spectrum of 1R against GSK3β. (b) WaterLOGSY NMR spectrum of 1R against GSK3β. Both NMR techniques further validated binding of fragment 1R against GSK3β. (c) Using ITC, 1R was determined to bind against GSK3β with K_d = 610 μM (LE = 0.37).

Figure 5

Small set of compounds were synthesized, extending from a growth vector identified from previous SAR. For ranking purposes, DSF assays were performed at 1.25 mM for comparison, and point inhibition biochemical assays were performed at 1 mM.

Figure 6

(a) ITC K_d determination of 15R against GSK3β K_d = 9.1 μM. (b) Biochemical inhibition measurements of 15R (IC₅₀ = 18 μM) and 15S (IC₅₀ = 322 μM) against GSK3β. (c) IC₅₀ measurement of 15R and 15S against CDK.

Figure 7

X-ray crystal structure of 1R and 15R against GSK3β. The 2F_o-F_c (1σ) and F_o-F_c maps (3σ) around 15R are shown as blue and orange meshes, respectively. Polar interactions are displayed as red dashed lines. PDB ID for 1R, 4J71; for 15R, 4J1R.