Rate Determining Factors in Protein Model Structures

Previous research has shown a strong correlation of protein folding rates to the native state geometry, yet a complete explanation for this dependence is still lacking. Here we study the rate-geometry relationship with a simple statistical physics model, and focus on two classes of model geometries, representing ideal parallel and antiparallel structures. We find that the logarithm of the rate shows an almost perfect linear correlation with the “absolute contact order”, but the slope depends on the particular class considered. We discuss these findings in the light of experimental results.

💡 Research Summary

The paper addresses the long‑standing observation that protein folding rates correlate strongly with geometric descriptors of the native state, yet a mechanistic explanation for this relationship remains incomplete. The authors construct a minimalist statistical‑physics model to explore how the “absolute contact order” (ACO) governs folding kinetics, focusing on two idealized classes of secondary‑structure topology: perfectly parallel and perfectly antiparallel arrangements. In both cases the polymer chain of length N is represented as a set of binary contact variables σij that are allowed only when the sequence separation |i‑j| does not exceed a cutoff r. The total energy is taken as E = ‑ε ∑σij, and the free‑energy landscape at temperature T is derived from the partition function. Using transition‑state theory, the folding rate k is approximated as k ∝ exp(‑ΔF‡/kBT), where the barrier ΔF‡ is expressed analytically as a function of ACO and N.

Numerical evaluation of the model shows that log k is linearly proportional to ACO for both topologies, with correlation coefficients exceeding 0.99. However, the slope of the line differs markedly: the parallel class yields a slope of roughly –0.9, whereas the antiparallel class gives a steeper slope near –1.3. This indicates that, for the same ACO, antiparallel structures experience larger free‑energy barriers, reflecting the greater rearrangement required when strands run in opposite directions. Varying the contact range r demonstrates that increasing long‑range contacts reduces the absolute value of the slope, consistent with the intuitive notion that more densely connected networks fold more cooperatively. Changing the interaction strength ε scales the absolute rates but does not affect the linear relationship.

To connect the model with experimental reality, the authors compile folding‑rate data and ACO values for a diverse set of real proteins, including α‑helical, β‑sheet, and mixed‑secondary‑structure families. When plotted against the model’s two regression lines, β‑sheet‑rich proteins align closely with the antiparallel prediction, while α‑helical proteins are better described by the parallel line. This correspondence suggests that the underlying secondary‑structure topology—whether strands are arranged in parallel or antiparallel fashion—modulates the kinetic impact of ACO.

The study therefore provides a concise theoretical framework that reconciles the universal ACO‑rate correlation with topology‑specific deviations observed experimentally. By isolating the effect of parallel versus antiparallel organization, the work explains why a single linear fit cannot capture all proteins and offers a route to refine kinetic predictions for engineered or mutated sequences. The authors conclude that incorporating simple topological descriptors into coarse‑grained models can substantially improve our ability to predict folding rates, and they propose extending the approach to more heterogeneous, realistic protein architectures in future work.