Coding relationship links RNA G-quadruplexes and protein RGG motifs in RNA-binding protein autoregulation

Abstract

RNA G-quadruplexes (rG4s), the four-stranded structures formed by guanine-rich RNA sequences, are recognized by regions in RNA-binding proteins (RBPs) that are enriched in arginine-glycine repeats (RGG motifs). Importantly, arginine and glycine are encoded by guanine-rich codons, suggesting that some RGG motifs may both be encoded by and interact with rG4s in autogenous messenger RNAs (mRNAs). By analyzing transcriptome-wide rG4 datasets, we show that hundreds of RGG motifs in humans are at least partly encoded by rG4s, with an increased incidence for longer RGG motifs (~10 or more residues). Using randomized genetic codes, we demonstrate that the rG4/RGG coding relationship derives from the universal genetic code’s structure. Moreover, we show that proteins, which contain RGG motifs encoded by experimentally detected rG4s, are significantly enriched in RNA binding relative to all RGG-containing proteins. Finally, using enhanced crosslinking and immunoprecipitation (eCLIP) data, we identify several prominent RBPs, including FUS, FMRP, and G3BP1, which interact with autogenous mRNAs in regions where RGG motifs are encoded by rG4s. Our results define a physically realistic mechanism behind autogenous mRNA/protein interactions that is hardwired in the genetic code structure and may contribute to the establishment of autoregulatory feedback loops in the cell.