Key concepts:

Figure 1. ATP competitive inhibition of catalytic cores of protein kinases. Key tenets of structure-based design of inhibitors utilizing inactive and active conformation of catalytic cores. (a) Schematic representation of the conserved catalytic core of the protein kinase family: a, N-terminal domain; b, C-terminal domain; c, inhibitor bounding at the cleft between the two domains potentially involving (f, e) two hydrophobic pockets and d, hydrogen bonds (one, two or three) that anchor the inhibitor to the linker between the N- and C-terminal lobes. Protein kinases exist in inactive-unphosphorylated form (g, green) and active phosphorylated form (g, yellow). (b) On phoshorylation the activation loop (g) undergoes conformational changes and creates a template for the protein substrate. Inhibitor binding interaction differs for the unphosphorylated (inactive, green) and phosphorylated (active, yellow) states of protein kinase. (c) Ribbon diagram of the crystal structures of active insulin receptor (IR) kinase domain in the ‘open’ apoprotein conformation (1IRK.pdb; Hubbard et al., 1994) with the activation loop (green) in the inactive unphosphorylated state. (d) IR in the ‘closed’ ATP-bound conformation with the activation loop (yellow) in the active triphosphorylated state (phosphotyrosines – stick) (1GAG.pdb; Parang et al., 2001). (e) Ribbon diagram of the crystal structure of active protein kinase A (PKA) kinase domain in the ‘open’ apoprotein (1CTP.pdb; Karlsson et al., 1993) with the activation loop (yellow) in the active state (phosphothreonine – stick) and (f) the ‘closed’ ATP-bound conformations of PKA with a bound ATP analogue (stick) (1ATP.pdb; Zheng et al., 1993) with the activation loop (yellow) in the active state (phosphothreonine – stick). The ‘open’ and ‘closed’ conformations relative to the ATP-binding pocket of fully active protein kinase result from the rotation and translation of the N-terminal lobe in respect to the C-terminal lobe. The axis of rotation passes through the region linking the N- and C-terminal lobes. The red residue marks the ‘gatekeeper’ residue M1076 in IR and M120 in PKA. Substrate peptides, or peptide inhibitors, are represented as magenta ribbons.

Figure 2. Fighting off patients' resistance through reversible inhibition by going after both inactive and active conformations of catalytic core of c-Abl kinase. Structure-based design of reversible inhibitors aiming at CML patients' resistance to imatinib. (a) Co-crystal structure of imatinib bound to the inactive unphosphorylated catalytic core of c-Abl (2HYY.pdb; Cowan-Jacob et al., 2007). (b) Co-crystal structure of nilotinib bound to the inactive unphosphorylated catalytic core of c-Abl (3CS9.pdb; Weisberg et al., 2005). (c) Co-crystal structure of desatinib bound to the active phosphorylated catalytic core of c-Abl (2GQG.pdb; Tokarski et al., 2006). Chronic myeloid leukaemia (CML) patient drug resistant mutations (O'Hare et al., 2007) for each drug are mapped on the co-crystals structures to illustrate the general location within the catalytic core (side chains displayed are the native structure and not the mutation). Green side-chains (sticks) are of intermediate resistance, whereas red side-chain positions (sticks) are for strong resistance. A common, strong drug resistant mutant T315I (red) in CML cancer patients lies in the conserved linker region next to the axis of rotation of the N-terminal lobe. Drug space filling models and chemical structures with ligand hydrogen bond interactions with the enzyme are shown on the left side of panels a, b and c.

Figure 3. Fighting off resistance through irreversible inhibition. Structure-based design of reversible and irreversible inhibitors of active conformation of EGFR kinase catalytic core. (a) Gefinitib bound to the catalytic core of EGFR containing the wild-type T790 (red) with the activation loop in the active conformation, but unphosphorylated (2ITY.pdb; Yun et al., 2007). (b) Irreversible inhibitor HKI-272 bound to the catalytic core of EGFR-containing mutant T790M (red) with the activation loop in the active conformation, but unphosphorylated (2JIV.pdb; Yun et al., 2008). Drug chemical structures and space filling models along with hydrogen-bond network interactions with the enzyme are on the left sides of panels a and b.

Figure 4. Inhibiting downstream kinases. Structure-based design of MEK inhibitor with the first non-ATP competitive inhibitor of kinase and the structure-based design of PI3 kinase inhibitors. (a) Co-crystal structure of MEK kinase domain with inhibitor clinical candidate PD 0325901 and ADP (3EQG.pdb; Fischmann et al., 2009). (i) Chemical structures of ligands and hydrogen bond network interactions between ligands and the enzyme along with a (ii) space-filling models. (iii) A ribbon drawing of MEK kinase in complex with ADP and PD 0325901. The activation region (yellow) forms a helix. Inhibitors of MEK are unique to the vast majority of protein kinase inhibitors, the mode of binding at a site adjacent to ATP, so that the inhibitor is noncompetitive with ATP. The analogous gatekeeper residue, M143, is positioned in contact with both the ATP and PD 0325901 inhibitor. (b) PI3 kinase gamma with inhibitor GDC0941 (3DBS.pdb; Folkes et al., 2008). (i) Chemical structure of inhibitor GDC0941 and the hydrogen bond interactions with the enzyme along with a (ii) space-filling model of GDC0941. (iii) A ribbon drawing of PI3K gamma with the extended N-terminal region (magenta), the kinase N-terminal domain (blue), the kinase C-termial domain (light blue), the activation loop (yellow), the gatekeeper residue I879 (red), and GDC0941 (stick).

Figure 5. Going ‘downstream’  Cancer-activated kinases in RAS/B-RAF/MEK and PI3KA/mTOR/AKT phosphate signalling pathways. Receptor tyrosine kinase (RTK) and downstream kinases with cancer-activating mutations are outlined in dark blue. Activating mutations have been identified in chronic myeloid leukaemia, gastrointestinal stromal cancer, nonsmall cell lung cancer, skin melanoma cancer and in colon cancer.