Cryptic Pockets Repository through Pocket Dynamics Tracking and Metadynamics on Essential Dynamics Space: Applications to Mcl-1.

Detection of cryptic pockets (hidden protein pockets) is a hot topic in structure-based drug discovery, especially for drugging the yet undruggable proteome. The experimental detection of cryptic pockets is still considered an expensive endeavor. Thus, computational methods, such as atomistic simulations, are used instead. These simulation methods can provide a perspective on protein dynamics that overpasses the experimental X-ray structures' static and average view. Nonetheless, unbiased molecular dynamics (MD) simulations fall short to detect transient and cryptic pockets requiring the crossing of high-energy barriers. Enhanced sampling methods, such as Metadynamics, provide a solution to overcome the time-scale problem faced by unbiased MD simulations. However, these methods are still limited by the availability of collective variable space to capture the intricate parameters, leading to the opening of cryptic pockets. Unfortunately, the design of such collective variables requires a priori knowledge of the binding site, information that is by definition lacking for cryptic pockets. In this work, we evaluated the use of the Metadynamics biasing scheme on essential coordinates space as a general method for cryptic pocket detection. This approach was applied to an antiapoptotic protein: Mcl-1 as a test model. In addition to providing a broader characterization of Mcl-1's conformational space, we show the effectiveness of this method in drawing the full repository of Mcl-1's known and novel cryptic pockets in an unsupervised manner.