RNA editing in Fern Plastomes Evolves Through Neutral Accumulation and Adaptive Co-option

Abstract

RNA editing is a post-transcriptional process that alters RNA sequences from their genomic templates, restoring functional codons in organellar transcripts. In plants, RNA editing is exclusive to the chloroplast and mitochondrial genomes but is mediated by large families of nuclear-encoded pentatricopeptide repeat (PPR) proteins. Despite its complexity and cost, plant RNA editing has long been viewed as a compensatory mechanism with no clear adaptive value, its origin and persistence attributed to constructive neutral evolution (CNE). However, the evolutionary forces that shape the gain, loss, and persistence of RNA editing sites across plant lineages remain poorly understood, especially in taxa retaining both C-to-U and U-to-C editing. Ferns are uniquely positioned for investigating these evolutionary processes: they exhibit a diversity of RNA editing frequencies, retain both editing types, and display dynamic organellar genome evolution. In Chapter 1, I characterized variation in plastid RNA editing within the fern genus Adiantum, providing the first infrageneric comparison of chloroplast RNA editing in ferns. I discovered striking variation in both the total number of editing sites and which specific sites are edited among closely related species, demonstrating that RNA editing evolves rapidly over shallow evolutionary timescales. Most of this turnover reflects gains and losses of nonsynonymous C-to-U edits, which vary extensively among species, whereas U-to-C edits are far more conserved. Notably, edits that restore start codons or remove internal stop codons showed the strongest phylogenetic conservation. This provided early evidence that different classes of RNA editing sites are shaped by distinct evolutionary pressures and that a functionally important subset may be maintained by selection. Chapter 2 expanded this phylogenetic perspective to deeper evolutionary timescales by comparing plastid editomes across two distantly related fern lineages, each exhibiting pronounced substitution rate heterogeneity: Vittarioideae (Pteridaceae) and Hymenophyllaceae. Using comparative phylogenetic methods, I demonstrated that nonsynonymous edits are progressively lost over evolutionary time, consistent with neutral evolution via CNE. In contrast, U-to-C edits that correct internal stop codons are selectively retained and even gained over time, occurring with lower, more variable efficiencies—patterns consistent with a regulatory role. These findings provide the first robust evidence that RNA editing in plants is not solely neutral: a subset of edits has been adaptively co-opted for post-transcriptional regulation of chloroplast gene expression. In Chapter 3, I investigated how nonsynonymous RNA editing sites evolve under relaxed selection in the Schizaeales, a fern order that includes lineages in the Schizaeaceae family with achlorophyllous, mycoheterotrophic gametophytes. Despite extensive plastome reduction, species in Schizaeaceae exhibited dramatic increases in C-to-U editing. Comparative analyses demonstrated that relaxed selection accelerates the accumulation of new editing sites via CNE, while expansion of the inverted repeat (IR) slows site loss by reducing backmutation rates. In addition, I identified a mechanistic basis that may facilitate rapid editome expansion: RNA editing factors in Schizaeales (and likely ferns more broadly) are modular, with PPR binding domains and catalytic editing domains encoded by separate proteins that assemble in trans. This modular architecture may permit rapid establishment and retention of new editing sites, without requiring the evolution of complete editing factors. In summary, this dissertation demonstrates that plastid RNA editing in ferns is shaped by both neutral and adaptive evolutionary forces. At shallow evolutionary scales (Chapter 1), RNA editing evolves rapidly, with substantial variation in both the number and identity of C-to-U editing sites among closely related species. In contrast, edits that modify start and internal stop codons are consistently conserved, pointing to functional constraint. At deeper timescales (Chapter 2), comparative phylogenetic analyses revealed that most RNA editing sites evolve through constructive neutral evolution (CNE), but a small subset—especially U-to-C edits that remove internal stop codons—shows clear signatures of adaptive co-option and likely regulate plastid gene expression. In lineages experiencing relaxed selection and plastome degradation (Chapter 3), nonsynonymous editing sites accumulate rapidly via CNE, and their persistence is further influenced by genome architecture, such as inverted repeat expansion, which slows their loss. Finally, I show that fern RNA editing factors are modular and act in trans, a feature that may facilitate rapid expansion of editomes. Overall, this work demonstrates that RNA editing evolves through both neutral accumulation and adaptive co-option: neutral processes generate molecular complexity that can be later co-opted by selection for regulatory roles.