Efficient detection of cryptic pockets on challenging drug targets using MD-free co-vibrational modes

Allosteric and cryptic binding sites present valuable prospects in drug discovery, enabling selectivity, diminishing toxicity, and unlocking difficult therapeutic targets. However, cryptic pockets are notoriously elusive, challenging experimental methods to capture them and often escaping detection. Computational techniques for detecting flexible cryptic sites typically involve resource-intensive, unbiased molecular dynamics simulations or more efficient, albeit still computationally demanding, biased MD techniques that exploit low-frequency protein motions.

In this study, researchers employed an innovative variant of the recently developed and significantly faster MD-free COVIB algorithm to efficiently identify cryptic pockets in two complex targets: the KRAS oncoprotein and the N-terminal domain of the STAT5B transcription factor. Utilizing both experimental and AI-modelling structures, the researchers uncovered numerous cryptic sites, some of which remained undocumented.

These sites were prioritized and validated using a pharmacophore-optimized photoaffinity-tagged fragment library through mass spectrometry-based peptide mapping and nuclear magnetic resonance spectroscopy. Beyond unveiling new, targetable regions on the surface of these crucial oncotargets, the team has devised a protocol merging COVIB-based cryptic pocket detection with photoaffinity-based experimental peptide mapping.

This robust approach promises a systematic and effective strategy for exploring unique and accessible cryptic pockets in therapeutic targets.