Anchor extension: a structure-guided approach to design cyclic peptides targeting enzyme active sites

Abstract

Despite recent success in computational design of structured cyclic peptides, de novo design of cyclic peptides that bind to any protein functional site remains difficult. To address this challenge, we develop a computational “anchor extension” methodology for targeting protein interfaces by extending a peptide chain around a non-canonical amino acid residue anchor. To test our approach using a well characterized model system, we design cyclic peptides that inhibit histone deacetylases 2 and 6 (HDAC2 and HDAC6) with enhanced potency compared to the original anchor (IC50 values of 9.1 and 4.4 nM for the best binders compared to 5.4 and 0.6 µM for the anchor, respectively). The HDAC6 inhibitor is among the most potent reported so far. These results highlight the potential for de novo design of high-affinity protein-peptide interfaces, as well as the challenges that remain.

Structure-based prediction of HDAC6 substrates validated by enzymatic assay reveals determinants of promiscuity and detects new potential substrates

“…First, the HDAC6 substrate with the highest k cat / K M value (EGK Ac FVR, derived from prelamin A, see Table 1 ) was docked into the binding pocket of a solved DD2 HDAC6 structure (Protein Data Bank 29 , PDB ID: 6WSJ 30 ) using Rosetta FlexPepDock refinement 31 . The top 5 best scoring models were selected as templates.…”

“…To assess if performance was affected by the specific crystal structure selected we repeated the analysis using different structures. The DD2 domain of HDAC6 was solved several times, bound to different ligands: (1) a cyclic peptide (PDB ID: 6WSJ 30 ) (2) a tripeptide substrate attached to coumarin (PDB ID: 5EFN 4 ), and (3) an inhibitor (PDB ID: 7JOM 33 ). Best performance was achieved with the cyclic peptide (6WSJ), while the structure with the inhibitory small molecule performed worst (Table 2 B and Supplementary Fig.…”

“…We explored structural differences between these proteins that could explain this finding. Comparison of HDAC6 DD2 (PDB ID: 6WSJ 30 ) and HDAC8 (PDB ID: 2V5W 47 ) structures (Fig. 5 ) highlight as main difference a loop involved in forming the binding pocket that could lead to differences in binding selectivity.…”

“…Running FlexPepBind requires the creation of a starting structure to generate a template (or a set of templates, as in the present study) for threading peptides. We used the structure of the Danio rerio HDAC6 catalytic domain 2 (DD2) which was crystallized in a complex with a cyclic peptide substrate (PDB ID: 6WSJ 30 ). To enforce a catalysis-competent binding conformation, we defined constraints that characterize substrate binding as revealed by the solved structures of HDAC6 bound to ligands.…”

Computational Site Saturation Mutagenesis of Canonical and Non-Canonical Amino Acids to Probe Protein-Peptide Interactions

“…Application of these stringent filters to our dataset of 420 point mutations on CP2 removed 183 designs; 65 of which were not in agreement with the in vitro dataset. Despite the benefit of these additional filters for assessing macrocycle designs, predictive modeling of side chains that stabilize a binding interface remains challenging and not always accurate, as reported elsewhere ( Nguyen et al, 2019 ; Mulligan et al, 2020 ; Hosseinzadeh et al, 2021 ). This is in part because macrocycles are dynamic molecules that are sensitive to small perturbations which may alter backbone and side chain conformations.…”