Large-Scale Structure-Based Prediction and Identification of Novel Protease Substrates Using Computational Protein Design

Abstract

Characterizing the substrate specificity of protease enzymes is critical for illuminating the molecular basis of their diverse and complex roles in a wide array of biological processes. Rapid and accurate prediction of their extended substrate specificity would also aid in the design of custom proteases capable of selectively and controllably cleaving biotechnologically or therapeutically relevant targets. However, current in silico approaches for protease specificity prediction, rely on, and are therefore limited by, machine learning of sequence patterns in known experimental data. Here, we describe a general approach for predicting peptidase substrates de novo using protein structure modeling and biophysical evaluation of enzyme-substrate complexes. We construct atomic resolution models of thousands of candidate substrate-enzyme complexes for each of five model proteases belonging to the four major protease mechanistic classes-serine, cysteine, aspartyl, and metallo-proteases-and develop a discriminatory scoring function using enzyme design modules from Rosetta and AMBER's MMPBSA. We rank putative substrates based on calculated interaction energy with a modeled near-attack conformation of the enzyme active site. We show that the energetic patterns obtained from these simulations can be used to robustly rank and classify known cleaved and uncleaved peptides and that these structural-energetic patterns have greater discriminatory power compared to purely sequence-based statistical inference. Combining sequence and energetic patterns using machine-learning algorithms further improves classification performance, and analysis of structural models provides physical insight into the structural basis for the observed specificities. We further tested the predictive capability of the model by designing and experimentally characterizing the cleavage of four novel substrate motifs for the hepatitis C virus NS3/4 protease using an in vivo assay. The presented structure-based approach is generalizable to other protease enzymes with known or modeled structures, and complements existing experimental methods for specificity determination.

MFPred: Rapid and accurate prediction of protein-peptide recognition multispecificity using self-consistent mean field theory

“…Additionally, substrate binding involves flexibility on the protease side, with two loops (“flaps”) that are mobile and close over the binding pocket. Incorporation of greater backbone flexibility on both the receptor and peptide parts of the HIVPR1-peptide interface may help improve predictions, as previously observed by us and others [31–33]. …”

“…We have previously found that modeling a near-attack conformation for the acylation step was successful in discriminating between known cleaved and uncleaved peptides [31]. Therefore, starting from structures of protease-substrate complexes in a near-attack conformation, we performed MFPred-based specificity prediction.…”

“…Second, we generated peptide backbone ensembles by threading on a varying number of known substrate (cleaved) peptides using three different Rosetta-based backbone sampling protocols (FastRelax [34], FlexPepDock [35], and Backrub [36]) separately to further diversify the peptide backbone ensemble. In each case, geometric constraints [31] were used to limit the scissile peptide bond to a near-attack conformation and the catalytic residues to an active conformation. The MFPred simulations were then performed on all backbone ensembles and their results were compared to each other (Fig 2).…”

“…For further details on the curation of the protease datasets, please see our recent study [31]. To generate a specificity profile for each protease, we first removed duplicates from the set of cleaved peptides and then calculated the frequency of each amino acid at each position.…”

Electrostatic recognition in substrate binding to serine proteases

“…To predict the specificity of proteases, Pethe et al used a structure‐based approach that ranks possible substrates according to interaction energies and reorganization penalties. Their scheme outperforms conventional methods that focus solely on knowledge‐based prediction of substrate preferences.…”