Identification of Novel Small Molecule Ligands for JAK2 Pseudokinase Domain

Abstract

Hyperactive mutation V617F in the JAK2 regulatory pseudokinase domain (JH2) is prevalent in patients with myeloproliferative neoplasms. Here, we identified novel small molecules that target JH2 of JAK2 V617F and characterized binding via biochemical and structural approaches. Screening of 107,600 small molecules resulted in identification of 55 binders to the ATP-binding pocket of recombinant JAK2 JH2 V617F protein at a low hit rate of 0.05%, which indicates unique structural characteristics of the JAK2 JH2 ATP-binding pocket. Selected hits and structural analogs were further assessed for binding to JH2 and JH1 (kinase) domains of JAK family members (JAK1-3, TYK2) and for effects on MPN model cell viability. Crystal structures were determined with JAK2 JH2 wild-type and V617F. The JH2-selective binders were identified in diaminotriazole, diaminotriazine, and phenylpyrazolo-pyrimidone chemical entities, but they showed low-affinity, and no inhibition of MPN cells was detected, while compounds binding to both JAK2 JH1 and JH2 domains inhibited MPN cell viability. X-ray crystal structures of protein-ligand complexes indicated generally similar binding modes between the ligands and V617F or wild-type JAK2. Ligands of JAK2 JH2 V617F are applicable as probes in JAK-STAT research, and SAR optimization combined with structural insights may yield higher-affinity inhibitors with biological activity.

Keywords: JAK inhibitor; JAK2 V617F; cytokine signaling; myeloproliferative neoplasm; pseudokinase.

Figure 1

Fluorescence polarization signal in screening. Data presented is decrease in FP [mP] for each 103,400 compounds screened. Verified 51 hits are marked with orange color.

Figure 2

Dose-response data and chemical structures of selected hits. (a) Normalized dose–response data JNJ-7706621 (1), reversine (7), CB_7644166 (4) from fluorescence polarization assay for recombinant JAK2 JH2 V617F. Binding of compounds in ATP pocket of proteins is observed as signal decrease. (b) Normalized dose–response data JNJ-7706621 (1), reversine (7), CB_7644166 (4) from fluorescence polarization assay for recombinant JAK2 JH1. (c) Chemical structure and IC50 value for binding against JAK2 JH2 V617F of confirmed hits. Hits were grouped based on chemical structure as diaminotriazole compounds, diaminotriazine compounds, aminopurine compounds, and other hits.

Figure 3

Chemical structures of selected analogs and binding to JAK2 JH2 and JH1 domains. (a) Diaminotriazole analogs, (b) diaminotriazine analogs, and (c) phenylpyrazolo-pyrimidone analogs. Subgroup R1 marked with yellow, R2 with green and R3 with blue background. Binding data are inhibition-% of tracer binding to recombinant JAK2 JH2 and JH1 domains at 100 µM analog concentration.

Figure 4

Crystal structures of selected hits and hits analogs with JAK2 JH2 wild-type and V617F mutant. The residues participating in hydrogen bonding and the gatekeeper Gln626 are shown as sticks. The ligand complexes of WT and V617F are superposed together. (a) Compound 1 (RMSD for aligned Cα atoms: 0.43). (b) Compound 36 (RMSD for aligned Cα atoms: 0.133). (c) Compound 16 (RMSD for aligned Cα atoms: 0.371). (d) Compound 7 (RMSD for aligned Cα atoms: 0.378). (e) Compound 24 (RMSD for aligned Cα atoms: 0.243). (f) Compound 6 (RMSD for aligned Cα atoms: 0.43). For JAK2 JH2 WT, the structure is shown in turquoise, the ligand and hydrogen bonds in orange. For JAK2 JH2 V617F, the structure is shown in magenta, the ligand and hydrogen bonds in yellow. JAK2 JH2—1 complex structure has been determined previously (pdb code 5WIN).

Figure 5

Differences induced by compound binding to the conformation of Phe594 and Phe595. (a) Comparison of JAK2 JH2 WT (PDB code 4FVQ, green) and JAK2 JH2 V617F (PDB code 4FVR, pink). (b) Comparison of JAK2 JH2 V617F-24 (magenta) and JAK2 JH2 V617F (pink). (c) Comparison of JAK2 JH2 V617F-6 (magenta) and JAK2 JH2 WT (green). (d) Comparison of JAK2 JH2 V617F-16 (magenta) and JAK2 JH2 WT (green).