RNA Translocation through Protein Nanopores: Interlude of the Molten RNA Globule

Direct translocation of RNA with secondary structures using single-molecule electrophoresis through protein nanopores shows significant fluctuations in the measured ionic current, in contrast to unstructured single-stranded RNA or DNA. We developed a multiscale model combining the oxRNA model for RNA with the 3-dimensional Poisson-Nernst-Planck formalism for electric fields within protein pores, aiming to map RNA conformations to ionic currents as RNA translocates through three protein nanopores: α-hemolysin, CsgG, and MspA. Our findings reveal three distinct stages of translocation (pseudoknot, melting, and molten globule) based on contact maps and current values. Two translocation modes emerge: fast and slow. In the fast mode, the speed is determined by the electric field, independent of pore geometry. In the slow mode, the molten globule stage is the rate-determining factor in slowing the translocation, instead of the previous paradigm of melting of the base pairs. Using these insights, we propose a neural network framework to identify and reconstruct RNA secondary structures from ionic current windows. We find that the electric field distribution, not the nanopore geometry, drives the molten globule stage. Our results explain the large current fluctuations. These results provide a fundamental understanding of the role of secondary and tertiary structures in the translocation of RNA in direct RNA translocation platforms based on single-molecule electrophoresis. This work offers design rules for new protein pores and real-time imaging of the secondary structures of RNA.