Optimized Preparation of Segmentally Labeled RNAs for NMR Structure Determination.

RNA structures are significantly underrepresented in public repositories (∼ 100-fold compared to proteins) despite their importance for mechanistic understanding and for development of structure prediction/validation tools. A substantial portion of deposited RNA structures have been determined by NMR (∼30%), but most comprise fewer than 60 nucleotides due to complications associated with NMR signal overlap. A promising approach for applying NMR to larger RNAs involves use of a mutated DNA polymerase (TGK) that can extend "primer" RNA strands generated independently by synthetic or enzymatic methods [Haslecker et al., Nature Commun. 2023]. In attempts to employ this technology, we uncovered sequence- and enzyme-dependent complications for most constructs examined that prohibited preparation of homogeneous samples. By using TGK extension efficiency and NMR as guides, we identified non-templated run-on by wild-type T7-RNA polymerase (RNAP(WT)) as the primary source of product heterogeneity. Use of 2'-O-methylated DNA templates did not prevent RNAP(WT) run-on for most constructs examined. However, primer RNAs with appropriate 3' end homogeneity were obtained in high yield using a recently described T7 RNAP mutant designed for improved immunogenic behavior. Minor spectral heterogeneity sometimes observed for 3' residues, caused by partial premature TGK termination, could be moved to sites downstream of the RNA region of interest by employing extended template DNAs that encode additional non-interacting 3' nucleotides. We additionally present an approach for large-scale synthesis of homogeneous template DNA required for TGK extension. With these modifications, segmentally labeled RNAs appropriate for high resolution structural studies are now routinely obtainable.