RNA editing, the correction of genomic errors at the level of messenger RNA, has evolved multiple times independently, but it remains unclear exactly how this process has evolved. A model for the emergence of RNA editing proposed that RNA editing activity pre-exists but there is no substrate for editing activity to act upon. Subsequently, mutation results in appearance of editable nucleotide sites, which may be fixed by genetic drift. Fixation results in RNA editing becoming indispensable for protein production from affected genes1. We sought to test this model by asking whether slippage-type editing can evolve under experimental conditions designed to maximize the impact of genetic drift.
We previously showed that, in the bacterial endosymbiont Buchnera, RNA polymerase slips at long poly(A/T) tracts, leading to stochastic incorporation or removal of As or Us in the nascent messenger RNA. This results in a heterogeneous population of mRNAs. RNA polymerase slippage was thus shown to correct natural frameshift mutations in mRNA, but the stochasticity of correction suggested there was reduced expression efficiency, as only some mRNAs carried corrected reading frames2.
In an evolution experiment using Escherichia coli, we subjected 10 lines to daily single-cell bottlenecks. Following 50 days of bottlenecking, one line showed an observable reduction in growth rate. Genome sequencing revealed the emergence of 38 frameshift mutations that appear to require slippage-type editing for gene expression. We present data showing that slippage-type editing indeed rescues frameshift mutations and that protein production is reduced, consistent with this type of mutation being slightly deleterious. Our results support the hypothesis that, under conditions favouring genetic drift, editing readily emerges. To our knowledge, this is the first experimental demonstration of the evolutionary drivers for the emergence of RNA editing.