Discovery of holoenzyme-disrupting chemicals as substrate-selective CK2 inhibitors

Abstract

CK2 is a constitutively active protein kinase overexpressed in numerous malignancies. Interaction between CK2α and CK2β subunits is essential for substrate selectivity. The CK2α/CK2β interface has been previously targeted by peptides to achieve functional effects; however, no small molecules modulators were identified due to pocket flexibility and open shape. Here we generated numerous plausible conformations of the interface using the fumigation modeling protocol, and virtually screened a compound library to discover compound 1 that suppressed CK2α/CK2β interaction in vitro and inhibited CK2 in a substrate-selective manner. Orthogonal SPR, crystallography, and NMR experiments demonstrated that 4 and 6, improved analogs of 1, bind to CK2α as predicted. Both inhibitors alter CK2 activity in cells through inhibition of CK2 holoenzyme formation. Treatment with 6 suppressed MDA-MB231 triple negative breast cancer cell growth and induced apoptosis. Altogether, our findings exemplify an innovative computational-experimental approach and identify novel non-peptidic inhibitors of CK2 subunit interface disclosing substrate-selective functional effects.

Figure 1

Computational identification of inhibitors of CK2α/CK2β interaction. (a) Ribbon diagrams of the four models of the CK2α/CK2β interface used in this study. The models differ in the position of the V101-P109 loop and demonstrate varying degree of openness of the binding site at the backbone level. The most open conformation, closely resembling the CK2β bound state of the CK2α, appears too flat to produce any appreciable small-molecule binding pockets; (b) The four structures of the CK2α/CK2β interface were subjected to fumigation and evaluated for druggability using ICM Pocket Finder algorithm. Fumigation resulted in larger and more drug-like pocket envelopes (white wire meshes) as compared to the original crystal structures. Four best models (framed) were selected and used for virtual ligand screening. The protein is represented by its solvent accessible surface and colored by molecular interaction properties: green – aliphatic, white – aromatic, blue – hydrogen bond (HB) donor, red – HB acceptor; (c) Distribution of compound binding scores predicted by ICM for the four selected fumigated CK2α/CK2β interface models. The models based on PDB 1m2r and 3bw5 (formerly 1ymi) appeared the most productive. The 1na7-based model featured the narrowest, and the 1om1-based model the widest binding pocket, both leading to the decreased number of low-scoring hits.

Figure 2

Characterization of CK2 inhibition by compound 1. (a) Chemical structure of compound 1. (b) Compound 1 inhibits the phosphorylation of a CK2β-dependent peptide substrate (⦁) and only shows a weak effect on the phosphorylation of a CK2β-independent substrate (⚬). (c) CK2α (40 nM) was incubated without (⦁,○) or with 50 μM compound 1 (▴,▵) in the presence of increasing concentrations of CK2β and assayed for phosphorylation of CK2β-independent (⦁,▴) or CK2β-dependent peptide substrates (⚬,▵). (d) Inhibition of CK2β-dependent phosphorylation activity by 100 μM compound 1 is non-competitive towards the CK2β-dependent peptide substrate. (e) Lineweaver-Burk double reciprocal plots are consistent with a competitive inhibition toward CK2β. Compound 1 concentration: 100, 50, 25, 12.5 and 0 μM. (f) Lineweaver-Burk double reciprocal plots are consistent with a mixed-type inhibition toward ATP. Compound 1 concentration: 100, 50, 37.5 and 0 μM. (g,h) Plate-bound MBP-CK2β was incubated with [³⁵S]methionine-labeled CK2α in the absence or presence of unlabeled CK2α, Pc peptide, compound 1 (g) or increasing concentrations of compound 1 (h). As a positive control, a 10-fold molar excess of untagged CK2α, was used (defined as 100% competition), and the value for 0% competition was obtained in the absence of any competitor. (i) GST or GST-CK2α were immobilized on biosensor surfaces and incubated in the absence or presence of the indicated concentrations of compound 1; MBP-CK2β binding was analyzed across these surfaces using SPR technique. SPR signals are expressed as percentages of MBP-CK2β binding to GST-CK2α in the absence of compound 1. Error bars represent the SEM of two biological replicates derived from technical triplicates.

Figure 3

The crystal structures of 4 and 6 bound to CK2α. (a) The structure of 4 (green, pdb:6FVF) bound to the interface site of CK2α. CK2α is shown as the surface representation and two of the important hydrogen bond interactions with the pocket are highlighted. All distances are in Angstroms. The 2Fo-Fc map contoured at 1σ is displayed in green. (b) The Interactions of 6 (blue, pdb:6FVG) with the interface site of CK2α (green). All of the hydrogen bonding and salt bridge interactions are highlighted. (c) The superimposed structures of 4 (green), 6 (blue) and CK2β (purple, pdb:4NH1) binding in the interface site of CK2α. (d) The structure of 6 binding to the interface site of CK2α shown as the surface representation.

Figure 4

Uptake and cellular effects of compound 6. Anti-CK2β immunoprecipitates were prepared from MCF10A cells incubated with DMSO (0.5%) or 75 μM 6 for 4 h. The corresponding immunoprecipitates were analyzed for the presence of CK2 subunits by Western blot (a) and quantified in (b) or for CK2 activity with CK2β-independent and CK2β-dependent peptide substrate (c). Data are representative of three biological replicates, uncropped blots are shown in Supplementary Fig. S16 and statistical analysis using Wilcoxon signed rank test showed a p-value of 0.03 in both cases (*). (d) In situ proximity ligation images of MCF10A cells incubated with 50 μM 6 for 3 h. (e) Number of fluorescent dots per cell was quantified using the BlobFinder software. n = 120 cells from two independent experiments. Mann Whitney test was used and results were found significant (*p-value = 0.0003 and **p-value = 0.044).

Figure 5

Cellular effects of compounds 3, 4 and 6. MCF10A cells were incubated for 24 h with either DMSO or 30 μM TAT or 30 μM TAT-Pc or 25 μM 6 (a) or with increasing concentrations of 6 (b) or 3 (c). MCF10A cells were incubated for 48 h with either DMSO or with increasing concentrations of 3 or 6 (d). Cells were then lysed and analyzed by western blot with the indicated antibodies. Blot images are representative of at least three independent experiments. Uncropped blots are shown in Supplementary Fig. S16. (e) MDA-MB231 cells were treated with 40 μM 6 (⚬), 4 (▴) or 3 (X). (f) MDA-MB231 cells were treated for 30 h with increasing concentrations of 6 (⚬) or 4 (▴). Insert: cells were treated for 30 h with 80 μM 6 or 10 μM CX-4945. (g) Cells were incubated with 80 μM 6 or 4 in the absence or presence of 20 μM Z-VAD. Statistical analysis using a two-way ANOVA with multiple comparison, Uncorrected Fisher’s LSD, showed that cell death was not significantly affected by Z-VAD for 4 (p = 0.1907, *) but was significantly affected for 6 (p = 0.0006, **). (h) MDA-MB231 cells (⚬) or MCF10A cells (⦁) were incubated for 24 h with 80 μM 6. Cell death was automatically quantified from images captured every 3 h for the duration of the experiments using an Essen IncuCyte Zoom live-cell microscopy incubator. Error bars represent the SEM of two biological replicates derived from technical triplicates.