Computational modeling of cotranscriptional RNA folding

Abstract

An RNA folds as it is transcribed. RNA folding during transcription differs fundamentally from thermodynamic folding. While thermodynamic folding reaches an equilibrium ensemble of structures, cotranscriptional folding is a kinetic process where the RNA structure evolves as the chain elongates during transcription. This dynamic folding pathway causes cotranscriptional structures often to deviate from thermodynamic predictions, as the system rarely reaches equilibrium. Since these cotranscriptional effects can persist in the mature RNA's structure, understanding this kinetic process is crucial. While experimental studies of cotranscriptional folding have been successful, they remain resource-intensive. Computational modeling has emerged as an increasingly practical and powerful approach for investigating these dynamics. This short review examines current computational methods and tools for simulating cotranscriptional folding, with the goal of advancing our understanding of RNA folding mechanisms.

Keywords: Computational modeling; Cotranscriptional folding; Kinetic pathway; RNA folding; Structure prediction.

Graphical abstract

Fig. 1

An illustration of RNA folding during transcription in the cell. As the RNA chain elongates, three kinetic intermediates form sequentially, each containing metastable helices shown in color. Some intermediates, such as a three-way junction, may be biologically functional prior to the RNA refolding into its thermodynamically stable structure, such as the extended stem-loop.

Fig. 2

A comparative scheme of the basepair-based (upper route) and helix-based (lower route) kinetic path approaches to RNA cotranscriptional folding simulations. The simulated kinetic transition network is updated from left to right with RNA chain elongation by one nucleotide.

Fig. 3

A practical example of reconstructing an RNA cotranscriptional folding pathway through the combined use of cotranscriptional chemical probing and computational modeling is shown here. The candidate structures and kinetic folding pathway were predicted using Vfold2D and the landscape-zooming model . Folding pathway population kinetics were inferred from cotranscriptional SHAPE experiments conducted by Watters et al. . Experimentally identified unpaired nucleotides were highlighted in yellow on the predicted 2D structures, showing strong agreement with the computationally predicted secondary structures.