The amino acid arginine can recognize RNA with striking diversity and a level of specificity that forms the basis of numerous essential biological interactions. Nearly 30 years ago, Frankel and co-workers postulated a theoretical model for Tat-mediated recognition of HIV-1 TAR in which a specific arginine residue forms four hydrogen bonds with two nearby phosphates to yield a two-pronged readout interaction.1 This ‘arginine fork’ model was posited to exist widely in RNA-protein recognition but languished without a substantial structural database. Fortunately, a collection of high-resolution RNA-protein structures is now available, providing a structurally rich landscape to systematically evaluate the fork hypothesis. indeed, a variation of the arginine fork was observed recently that featured: (i) a single arginine forming hydrogen bonds to the phosphate backbone and the Hoogsteen edge of a guanine nucleobase;and (ii) simultaneous engagement of the arginine guanidinium group in cation-π stacks with flanking nucleobases. Using this experimentally based framework for arginine-mediated base-specific readout, we conducted a structural bioinformatic search of an RNA-protein library of crystal structures retrieved from the protein data bank. Filtering was applied to satisfy the two arginine-fork criteria above. The results revealed four distinct classes of arginine-forks. We report arginine rotamer and RNA backbone conformation (suite) analysis of the arginine-fork interfaces. Additionally, we also present a practical application of the arginine-fork, which was used to improve NMR modeling of the HIV-1 Tat-TAR structure, and to correct atypical arginine-guanine interactions. Overall, our results can improve modeling of RNA-protein interfaces and provide insight into arginine-mediated recognition of specific RNA sequences, which is relevant to drug discovery. 1. Calnan, Barbara J., et al. Science (1991): 1167-1171. 2. Belashov, Ivan A., et al. Nucleic acids research 46.13 (2018): 6401-6415.
Read full abstract