There is great interest in developing selective protein kinase inhibitors by targeting allosteric sites, but these sites often involve protein-protein or protein-peptide interfaces that are very challenging to target with small molecules. Here we present a systematic approach to targeting a functionally conserved allosteric site on the protein kinase PDK1 called the PDK1-interacting fragment (PIF)tide-binding site, or PIF pocket. More than two dozen prosurvival and progrowth kinases dock a conserved peptide tail into this binding site, which recruits them to PDK1 to become activated. Using a site-directed chemical screen, we identified and chemically optimized ligand-efficient, selective, and cell-penetrant small molecules (molecular weight ∼ 380 Da) that compete with the peptide docking motif for binding to PDK1. We solved the first high-resolution structure of a peptide docking motif (PIFtide) bound to PDK1 and mapped binding energy hot spots using mutational analysis. We then solved structures of PDK1 bound to the allosteric small molecules, which revealed a binding mode that remarkably mimics three of five hot-spot residues in PIFtide. These allosteric small molecules are substrate-selective PDK1 inhibitors when used as single agents, but when combined with an ATP-competitive inhibitor, they completely suppress the activation of the downstream kinases. This work provides a promising new scaffold for the development of high-affinity PIF pocket ligands, which may be used to enhance the anticancer activity of existing PDK1 inhibitors. Moreover, our results provide further impetus for exploring the helix αC patches of other protein kinases as potential therapeutic targets even though they involve protein-protein interfaces.

Finding small molecules that target allosteric sites remains a grand challenge for ligand discovery. In the protein kinase field, only a handful of highly selective allosteric modulators have been found. Thus, more general methods are needed to discover allosteric modulators for additional kinases. Here, we use virtual screening against an ensemble of both crystal structures and comparative models to identify ligands for an allosteric peptide-binding site on the protein kinase PDK1 (the PIF pocket). We optimized these ligands through an analog-by-catalog search that yielded compound 4, which binds to PDK1 with 8 μM affinity. We confirmed the docking poses by determining a crystal structure of PDK1 in complex with 4. Because the PIF pocket appears to be a recurring structural feature of the kinase fold, known generally as the helix αC patch, this approach may enable the discovery of allosteric modulators for other kinases.

Allostery is an inherent feature of proteins and provides alternative routes to regulating function. Small-molecule allosteric inhibitors are often desirable; however, it remains challenging to identify surface sites in proteins which can bind small molecules and modulate function. We identified new allosteric sites in protein tyrosine phosphatase 1B (PTP1B) by combining multiple-temperature X-ray crystallography experiments and structure determination from hundreds of individual small-molecule fragment soaks. New modeling approaches reveal “hidden” low-occupancy conformational states for protein and ligands. Our results converge on a new allosteric site that is conformationally coupled to the active-site WPD loop, a hotspot for fragment binding, not conserved in the closest homolog, and distinct from other recently reported allosteric sites in PTP1B. Targeting this site with covalently tethered molecules allosterically inhibits enzyme activity. Overall, this work demonstrates how the ensemble nature of macromolecular structure revealed by multitemperature crystallography can be exploited for developing allosteric modulators.

Allostery is an inherent feature of proteins, but it remains challenging to reveal the mechanisms by which allosteric signals propagate. A clearer understanding of this intrinsic circuitry would afford new opportunities to modulate protein function. Here, we have identified allosteric sites in protein tyrosine phosphatase 1B (PTP1B) by combining multiple-temperature X-ray crystallography experiments and structure determination from hundreds of individual small-molecule fragment soaks. New modeling approaches reveal 'hidden' low-occupancy conformational states for protein and ligands. Our results converge on allosteric sites that are conformationally coupled to the active-site WPD loop and are hotspots for fragment binding. Targeting one of these sites with covalently tethered molecules or mutations allosterically inhibits enzyme activity. Overall, this work demonstrates how the ensemble nature of macromolecular structure, revealed here by multitemperature crystallography, can elucidate allosteric mechanisms and open new doors for long-range control of protein function.

Many proteins have small-molecule binding pockets that are not easily detectable in the ligand-free structures. These cryptic sites require a conformational change to become apparent; a cryptic site can therefore be defined as a site that forms a pocket in a holo structure, but not in the apo structure. Because many proteins appear to lack druggable pockets, understanding and accurately identifying cryptic sites could expand the set of drug targets. Previously, cryptic sites were identified experimentally by fragment-based ligand discovery and computationally by long molecular dynamics simulations and fragment docking. Here, we begin by constructing a set of structurally defined apo-holo pairs with cryptic sites. Next, we comprehensively characterize the cryptic sites in terms of their sequence, structure, and dynamics attributes. We find that cryptic sites tend to be as conserved in evolution as traditional binding pockets but are less hydrophobic and more flexible. Relying on this characterization, we use machine learning to predict cryptic sites with relatively high accuracy (for our benchmark, the true positive and false positive rates are 73% and 29%, respectively). We then predict cryptic sites in the entire structurally characterized human proteome (11,201 structures, covering 23% of all residues in the proteome). CryptoSite increases the size of the potentially "druggable" human proteome from ~40% to ~78% of disease-associated proteins. Finally, to demonstrate the utility of our approach in practice, we experimentally validate a cryptic site in protein tyrosine phosphatase 1B using a covalent ligand and NMR spectroscopy. The CryptoSite Web server is available at http://salilab.org/cryptosite.