Hotter isn't faster for a melting RNA hairpin

Abstract

We investigate the denaturation dynamics of nucleic acids through extensive molecular dynamics simulations of a coarse-grained RNA hairpin model. In apparent contradiction with Arrhenius' law, our findings reveal that the denaturation time of RNA hairpins is a non-monotonous function of temperature for molecules longer than few persistence lengths, with an optimal temperature above the melting point, $T_m$, at which denaturation occurs fastest. This anomaly arises from the existence of two distinct pathways: ``unidirectional'' unzipping, progressing from one end to the other and favored near $T_m$, and ``bidirectional'' denaturation, where competing unzipping events initiate from both ends at higher temperatures. The two regimes manifest distinct scaling laws for the melting time \textit{vs.} length, $L$, and are separated by a crossover temperature $T_\times$, with $(T_\times-T_m) \sim L^{-1}$. The results highlight the significant role of the helical structure in the out-of-equilibrium dynamics of RNA/DNA denaturation and unveil multiple surprises in a decades-old problem.