How was membrane permeability produced in an RNA world?

Darwinian evolution in an RNA World required that catalysts be encapsulated in membranes since this would allow superior catalysts to benefit from the products of their own reactions. However, typical membranes are relatively impermeable to polar and complex molecules and, thus, even primitive cells had to have RNA-based mechanisms for the uptake of external nutrients and the excretion of waste products. Nucleic acids form weak non-specific contacts with the surface of the lipid membrane in the presence of divalent cations, and strongly binding species can be obtained in the course of SELEX experiments. The only currently suggested mechanism for the production of permeability is through formation of supramolecular RNA complexes capable of destabilizing and transiently opening lipid membranes by action from one side, but neither natural nor selected examples of RNA channels or transmembrane shuttles are known so far. The necessity to evolve proteins could be strongly driven by the need to build defined hydrophobic structures that when integrated into membranes could provide selective permeability.