Submission Details

Design Rationale:

We have been taking a unique approach to this problem. Rather than targeting the Sars Cov2 protease in the native state, we are attempting to stabilize folding intermediates by designing compounds that are predicted to bind non-native cavities within these intermediates, thus stabilizing them. Our approach is akin to allosteric drug design, except that we aim to stabilize less structured states which may be prone to proteolytic degradation and/or aggregation in the cell. We expect it may be harder for the virus to evolve resistance to such compounds, as nonnative cavities often involve residues that are buried in the native state--thus, mutation of these residues may be detrimental to native stability. We model folding intermediates of a single subunit of the protease's N-terminal domain, on the assumptions that the domains fold independently, and dimerization precedes subunit folding. To predict these intermediates, we apply the algorithm in (1) (see Notes), which can rapidly simulate folding of complex proteins at the atomistic level while accounting for non-native states. We then grow compounds that are predicted to bind these intermediates at specific cavities using the OpenGrowth algorithm detailed in (2), keeping only compounds that predicted to bind with sub-micromolar affinity whose synthetic accessibility is less than 4.0, and which are predicted to bind a large number of representative snapshots assigned to a given folding intermediate with a similar binding profile, as computed using the AlphaSpace method (3). Three such compounds, each bound to a predicted folding intermediate, are included as separate submissions.

Other Notes:

Citations: 1. Bitran et. al. PNAS (2020), 117 (3), 1485-1495. 2. Chérot et. al. (2016) J. Med. Chem. 59, 9, 4171-4188. 3. Katigbak et. al. (2020) J. Chem. Inf. Model. 60, 3, 1494-1508.

Inspired By:

Download PDB File