Guiding Peptide Kinetics via Collective-Variable Tuning of Free-Energy Barriers

TL;DR

The results show that kinetic effects of point mutations can be inferred from minimal local sampling, providing a practical route to the rational engineering of conformational transition rates without exhaustive simulations or large training datasets.

Abstract

While recent advances in AI have transformed protein structure prediction, protein function is also often strongly influenced by the thermodynamic and kinetic features encoded in its underlying free-energy surface. Here, we propose a framework to rationally modify these surfaces in order to control conformational transition rates, built on the Collective Variables for Free Energy Surface Tailoring (CV-FEST) framework, and validate it on point mutations of the miniprotein Chignolin. The framework relies on Harmonic Linear Discriminant Analysis (HLDA) based collective variables (CVs) constructed from short molecular dynamics trajectories restricted to the metastable states, requiring only limited sampling within each state. Notably, the HLDA CV derived solely from the wild-type system already provides residue-level scores that predict whether mutations at specific positions are likely to accelerate or slow conformational transitions. Furthermore, we find that the leading HLDA eigenvalue associated with the derived CV, a quantitative measure of the one-dimensional statistical separation between folded and unfolded ensembles, correlates strongly with conformational transition rates across mutations. Together, these results show that kinetic effects of point mutations can be inferred from minimal local sampling, providing a practical route to the rational engineering of conformational transition rates without exhaustive simulations or large training datasets.

Figures (10)

• Figure 1: (a) Illustration of representative backbone-distance descriptors used as input features for HLDA construction. (b) Aggregated per-residue HLDA weights derived from the leading eigenvector of the WT projection, obtained using Eq. \ref{['eq:residue-importance']}.
• Figure 2: Correlation between WT residue importance and unfolding kinetics. (a) Aggregated WT residue importance $I_r$ versus mean MFPT change due to mutations at residue $r$ (Spearman $\rho = 1.00$; Pearson $r = 0.96$, $p = 5.4\times10^{-4}$). (b) Spearman correlation between mean MFPT and aggregated WT residue importance $I_r$ as a function of the RMSD threshold used to define first-passage times.
• Figure 3: Relationship between HLDA eigenvalues $\lambda$ and unfolding kinetics. (a) Logarithm of the MFPT ratio for each mutant plotted against the corresponding HLDA eigenvalue. A Pearson coefficient of $r=-0.68$ is obtained, with a corresponding $p$-value of $7.8\times10^{-6}$, and a Spearman rank coefficient of $\rho=-0.66$, indicating a monotonic and linear association between state separation and unfolding kinetics. (b) Spearman coefficient $\rho$ between MFPT and HLDA eigenvalue as a function of the RMSD threshold used to define first-passage times, demonstrating robustness of the observed relationship across threshold choices.
• Figure 4: Schematic illustration of the relationship between state separation as captured by the HLDA CV and the peptide's kinetic barrier. The folded and unfolded states are represented as parabolic free-energy basins projected onto the one-dimensional HLDA CV. (a) A smaller separation between the basin minima results in a lower intersection energy and a reduced free-energy barrier. (b) A larger separation shifts the intersection to a higher energy, increasing the effective barrier height.
• Figure 5: Statistical validation of the exponential first-passage-time model. (a) Empirical cumulative distribution function (ECDF) of WT FPTs compared with the theoretical exponential CDF estimated from the fastest subset of trajectories. (b) Fraction of mutants whose first-passage-time distributions pass the Kolmogorov--Smirnov and Lilliefors tests as a function of the RMSD threshold used to define unfolding.