Second-generation CK2α inhibitors targeting the αD pocket

Abstract

CK2 is a critical cell cycle regulator that also promotes various anti-apoptotic mechanisms. Development of ATP-non-competitive inhibitors of CK2 is a very attractive strategy considering that the ATP binding site is highly conserved among other kinases. We have previously utilised a pocket outside the active site to develop a novel CK2 inhibitor, CAM4066. Whilst CAM4066 bound to this new pocket it was also interacting with the ATP site: herein, we describe an example of a CK2α inhibitor that binds completely outside the active site. This second generation αD-site binding inhibitor, compound CAM4712 (IC₅₀ = 7 μM, GI₅₀ = 10.0 ± 3.6 μM), has numerous advantages over the previously reported CAM4066, including a reduction in the number of rotatable bonds, the absence of amide groups susceptible to the action of proteases and improved cellular permeability. Unlike with CAM4066, there was no need to facilitate cellular uptake by making a prodrug. Moreover, CAM4712 displayed no drop off between its ability to inhibit the kinase in vitro (IC₅₀) and the ability to inhibit cell proliferation (GI₅₀).

Fig. 1. Structure of CAM4066 and pro-CAM4066. Zwitterionic elements are coloured in dark green, amide bonds in blue and the difference between CAM4066 and its prodrug highlighted in purple. The interaction between CAM4066 and the highly conserved Lys68 is shown as red dashed line. The flexible linker is circled in orange. The αD pocket and ATP binding site are reported as black curves.

Fig. 2. Optimisation strategy adopted in this project using 1 as the fragment starting point.

Fig. 3. (a) A cross section of the αD pocket of 1 (green) bound to CK2α. The first two water molecules in the water channel are shown (PDB: ; 5CSH). (b) The structure of the αD loop (in grey) in the closed conformation of the apo protein (PDB: ; 5CSP). Tyr125 fills the mouth of the water channel. The binding mode of 1 (in green) superimposed on the apo structure is shown, with 1 occupying the same position as Phe121.

Fig. 4. Cross sections of the αD and ATP pockets: first and second rows are lateral views and last row is a view from the top. The water channel is highlighted in blue and the hydrophobic pocket in green for complexes that are not making the most of the space available. (a) Co-crystal structure of 1 with CK2α (PDB: ; 5CSH); (b) co-crystal structure of 2 with CK2α (PDB: ; 5ORH); (c) co-crystal structure of 3 with CK2α (PDB: ; 5ORJ); (d) co-crystal structure of 14 with CK2α (PDB: ; 5OTR); (e) co-crystal structure of 15 with CK2α (PDB: ; 5OTZ).

Fig. 5. (a) Co-crystal structure of 1 (in pink) and ATP bound to CK2α (PDB: ; 5CSH) and (b) surface representation of CK2α with 1 bound. (c) Co-crystal structure of 21 (in green) bound to CK2α. The binding mode of ATP (grey) in the ATP site when 1 is bound is superimposed upon the structure. (d) Co-crystal structure of 21 (in green) bound to the surface representation of CK2α. (e) Co-crystal structure of 22 (in blue) bound to CK2α. The binding mode of ATP (grey) in the ATP site when 1 is bound is superimposed upon the structure. (f) Co-crystal structure of 22 (in blue) bound to the surface representation of CK2α.

Fig. 6. Compounds 35, 36 and CAM4712 with respective data.

Fig. 7. (a) Western blot analysis showing the specific CK2 phosphorylation target: AKT1 serine 129. HCT116 cells were treated with 2 × GI₅₀ of CX4945 (20 μM), CAM4712 (20 μM) or pro-CAM4066 (20 μM) for 72 hours (b) dose response curve for the inhibition of the growth of HCT116 cells by CAM4712, pro-CAM4066 and CX4945. All graphs show the mean ± SEM of not less than three independent experiments with each in triplicate.