Between Protein Fold and Nucleophile Identity: Multiscale Modeling of the TEV Protease Enzyme–Substrate Complex

Abstract

The cysteine protease from the tobacco etch virus (TEVp) is a well-known and widely utilized enzyme. TEVp's chymotrypsin-like fold is generally associated with serine catalytic triads that differ in terms of a reaction mechanism from the most well-studied papain-like cysteine proteases. The question of what dominates the TEVp mechanism, nucleophile identity, or structural composition has never been previously addressed. Here, we use enhanced sampling multiscale modeling to uncover that TEVp combines the features of two worlds in such a way that potentially hampers its activity. We show that TEVp cysteine is strictly in the anionic form in a free enzyme similar to papain. Peptide binding shifts the equilibrium toward the nucleophile′s protonated form, characteristic of chymotrypsin-like proteases, although the cysteinyl anion form is still present and interconversion is rapid. This way cysteine protonation generates enzyme states that are a diversion from the most effective course of action, with only 13.2% of Michaelis complex sub-states able to initiate the reaction. As a result, we propose an updated view on the reaction mechanism catalyzed by TEVp. We also demonstrate that AlphaFold is able to construct protease−substrate complexes with high accuracy. We propose that our findings open a way for its industrious use in enzymological tasks. Unique features of TEVp discovered in this work open a discussion on the evolutionary history and trade-offs of optimizing serine triad-associated folds to cysteine as a nucleophile.

Challenges in Protein QM/MM Simulations with Intra-Backbone Link Atoms

“…It is generally accepted that serine deprotonation on histidine is coupled to the nucleophilic attack on the substrate carbon atom with immediate O–C bond formation. A stable deprotonated state may be expected from cysteine, but not from serine. The expected behavior is observed for the minimal sidechain-only system that is free of hyperpolarized link atoms.…”

“…59 Complex with VAHALAQTVP peptide was built with AlphaFoldmultimer-v1 60 with ColabFold v1.2.0 61 similarly to what was successfully done before for cysteine PA clan proteases. 62 Complex was then aligned to the 1R0R model in PyMol, 63 and peptide coordinates were transferred to it. The central ALA-LEU-ALA part was retained and capped with ACE and NME.…”

“…A Subtilisin Carlsberg enzyme–substrate system is built upon the 1R0R model . Complex with VAHALAQTVP peptide was built with AlphaFold-multimer-v1 with ColabFold v1.2.0 similarly to what was successfully done before for cysteine PA clan proteases . Complex was then aligned to the 1R0R model in PyMol, and peptide coordinates were transferred to it.…”

Modern non-polarizable force fields diverge in modeling the enzyme–substrate complex of a canonical serine protease

“…Its functional role is, however, unclear, since both its conformational variability and absence do not crucially affect the reaction. 41…”

“…Its functional role is, however, unclear, since both its conformational variability and absence do not crucially affect the reaction. 41 Representative members of each cluster are visualized in Fig. 4 (clusters 1, 2, 5-7).…”