Protein kinase B (PKB)-selective inhibitors were designed.. synthesized, and cocrystallized using the AGC kinase family protein kinase A (PKA, often called cAMP-dependent protein kinase),PKA has been used as a surrogate for other members of this family and indeed for protein kinases in general. The high homology between PKA and PKB includes very Similar ATP binding sites and hence similar binding pockets for inhibitors, with only few amino acids that differ between the two kinases. A series of these sites were mutated in PKA in order to improve the surrogate model for a design of PKB-selective inhibitors. Namely, the PKA to PKB exchanges F187L and Q84E enable the design of the selective inhibitors described herein which mimic ATP but extend further into a site not occupied by ATP. In this pocket, selectivity over PKA can be achieved by the introduction of bulkier substituents. Analysis of the cocrystal structures. and binding studies were performed to rationalize the selectivity and improve the design.

Protein–protein interactions serve as the molecular engine that drives the formation and disassembly of intracellular signaling pathways. Antagonists of these interactions could play key roles as both biological reagents and therapeutic compounds. However, much of the early work in this area with peptides revealed that these species, in general, bind with modest affinity to their protein targets. In addition, when these studies first commenced nearly 20 years ago, the technology for the intracellular delivery of peptides and modified analogs thereof was rudimentary. In the intervening years, not only has this technology dramatically improved, but the global role that protein–protein interactions play in transducing intracellular signals has become simply too obvious to ignore. With the introduction of combinatorial library methods, it is now a simple matter to identify consensus sequences recognized by protein interaction domains. An array of strategies has now been developed to transform these otherwise modest binding consensus sequences into high-affinity ligands. These strategies include the design of high-affinity replacements for key amino acid residues in consensus peptides, the construction of HEP (2005) 167:11--44 Springer-Verlag 2005

Protein kinases comprise the major enzyme family critically involved in signal transduction pathways; posttranslational modifications affect their regulation and determine signaling states. The prototype protein kinase A (PKA) possesses an N-terminal alpha-helix (Helix A) that is atypical for kinases and is thus a major distinguishing feature of PKA. Its physiological function may involve myristoylation at the N-terminus and modulation via phosphorylation at serine 10. Here we describe an unusual structure of an unmyristoylated PKA, unphosphorylated at serine 10, with a completely ordered N-terminus. Using standard conditions (e.g., PKI 5-24, ATP site ligand, MEGA-8), a novel 2-fold phosphorylated PKA variant showed the ordered N-terminus in a new crystal packing arrangement. Thus, the critical factor for structuring the N-terminus is apparently the absence of phosphorylation of Ser10. The flexibility of the N-terminus, its myristoylation, and the conformational dependence on the phosphorylation state are consistent with a functional role for myristoylation.

As the key mediators of eukaryotic signal transduction, the protein kinases often cause disease, and in particular cancer, when disregulated. Appropriately selective protein kinase inhibitors are sought after as research tools and as therapeutic drugs; several have already proven valuable in clinical use. The AGC subfamily protein kinase C (PKC) was identified early as a cause of cancer, leading to the discovery of a variety of PKC inhibitors. Despite its importance and early discovery, no crystal structure for PKC has yet been reported. Therefore, we have co-crystallized PKC inhibitor bisindolyl maleimide 2 (BIM2) with PKA variants to study its binding interactions. BIM2 co-crystallized as an asymmetric pair of kinase-inhibitor complexes. In this asymmetric unit, the two kinase domains have different lobe configurations, and two different inhibitor conformers bind in different orientations. One kinase molecule (A) is partially open with respect to the catalytic conformation, the other (B) represents the most open conformation of PKA reported so far. In monomer A, the BIM2 inhibitor binds tightly via an induced fit in the ATP pocket. The indole moieties are rotated out of the plane with respect to the chemically related but planar inhibitor staurosporine. In molecule B a different conformer of BIM2 binds in a reversed orientation relative to the equivalent maleimide atoms in molecule A. Also, a critical active site salt bridge is disrupted, usually indicating the induction of an inactive conformation. Molecular modeling of the clinical phase III PKC inhibitor LY333531 into the electron density of BIM2 reveals the probable binding mechanism and explains selectivity properties of the inhibitor.