Protein-protein Interactions in the Allosteric Regulation of Protein Kinases
Patricia Pellicena and John Kuriyan
Protein–protein interactions involving the catalytic domain of protein kinases are likely to be generally important in the regulation of signal transduction pathways, but are rather sparsely represented in crystal structures. Recently determined structures of the kinase domains of the mitogen-activated protein kinase Fus3, the RNA-dependent kinase PKR, the epidermal growth factor receptor and Ca2+/calmodulin-dependent protein kinase II have revealed unexpected and distinct mechanisms by which interactions with the catalytic domain can modulate kinase activity.
Figure 1. Anatomy of a protein kinase catalytic domain. Ribbon diagram representing the main features of a prototypical kinase catalytic domain based on the structure of IRK (PDB code 1IR3) . The N-lobe is shown in green, the C-lobe is shown in purple and the activation loop is colored yellow. The substrate peptide, docked onto the activation loop, and an ATP analog are shown in black. Helix-C, which contains residues critical for catalysis, and helix-G, which is involved in docking protein substrates, are labeled.
Figure 2. Comparison of Fus3 in complex with a peptide derived from the scaffolding protein Ste5 and PKA. (a) The Ste5 peptide, shown in dark blue, docks onto Fus3 (PDB code 2F49), shown in white. (b) The Ste5 peptide follows a path similar to that of the C-terminal tail extension (dark blue) of PKA (PDB code 2CPK)
Figure 3. Protein kinases subjected to regulation via N-lobe interactions. In many protein kinases, regulation of activity occurs via a region that couples helix-C to regulatory domains, as in GRK2 (PDB code 2BCJ) and the Src family kinase Hck (PDB code 1QCF). In the protein kinases PknE (PDB code 2H34) and PKR (PDB code 2A1A), a similar back-to-back dimer is observed. In this arrangement, the kinase catalytic domains act as their own regulatory regions via the same N-lobe interface that is seen in GRK2 and the Src kinases.
Figure 4. Comparison of the mechanism of EGF receptor activation by dimerization with the Cdk2–cyclin A complex. In the EGF receptor asymmetric dimer, one kinase catalytic domain is activated by another kinase catalytic domain in a C-lobe to N-lobe interaction. This is analogous to the mechanism of activation of Cdk2 by cyclin A (PDB code 1FIN).
Figure 5. Structure of the CaMK-II kinase domain and regulatory region. (a) The kinase catalytic domains within the dimer are shown in blue and green. Dimerization occurs through the regulatory regions, which form a coiled-coil strut. The regulatory regions contain the CaM-binding site, shown in red. The location of the major regulatory autophosphorylation site, Thr286, is shown for one of the monomers (PDB code 2BDW). (b) Hypothetical reconstruction of the CaMK-II holoenzyme, based on the crystal structure and SAXS analysis, showing a possible location of the dimer unit within the dodecameric holoenzyme.