Comprehensive sequence-to-function mapping of cofactor-dependent RNA catalysis in the glmS ribozyme

Abstract

Massively parallel, quantitative measurements of biomolecular activity across sequence space can greatly expand our understanding of RNA sequence-function relationships. We report the development of an RNA-array assay to perform such measurements and its application to a model RNA: the core glmS ribozyme riboswitch, which performs a ligand-dependent self-cleavage reaction. We measure the cleavage rates for all possible single and double mutants of this ribozyme across a series of ligand concentrations, determining k_cat and K_M values for active variants. These systematic measurements suggest that evolutionary conservation in the consensus sequence is driven by maintenance of the cleavage rate. Analysis of double-mutant rates and associated mutational interactions produces a structural and functional mapping of the ribozyme sequence, revealing the catalytic consequences of specific tertiary interactions, and allowing us to infer structural rearrangements that permit certain sequence variants to maintain activity.

Fig. 1. RNA array assay for ribozyme self-cleavage.

a Crystal structure of the T. tengcongensis glmS ribozyme, PDB 2Z75, with duplex elements P1, P2, P2.1, and P2.2 indicated (core domain: dark colors; non-core domain, pale cyan). b Secondary structure of the 66-nt core construct studied; duplex color scheme same as in a. Flanking sequences and the region targeted for random mutagenesis by doped synthesis are also indicated. The sequence of the core construct is from ref. ; the secondary structure is derived from crystallographic data (refs. ^, ). c Schematic of sequential steps for the RNA array assay. Variants were transcribed in situ on a sequencing chip. Hybridized Cy3- and Cy5-dye labeled DNA oligomers scored the presence of the ribozyme. The Cy3 TIRF signal is lost upon cleavage, which releases the 5′ product. d Representative cleavage records for the consensus ribozyme and the 13 variants are indicated (color-coded), displayed on both short and long time scales (data points are median per-tile intensity values for the given variant). e, Representative Michaelis–Menten curve fits to cleavage rates, derived from RNA array data (as in d), as a function of GlcN6P concentration (error bars, std. err.). Same color code as in panel d. Source data are available in the Source Data file.

Fig. 2. Mutation frequency is tightly correlated with effects on the cleavage rate.

Structural heatmaps of k_cat (a) and apparent K_M (b) for all single mutants and basepaired-partner double mutants. Magenta arrowheads indicate consensus ribozyme values. c Structural heatmap of mutation frequency for point mutants, and previously reported basepair covariation, from ref. (Methods). d–f Scatterplots of mutation frequency vs. cleavage rate (d), K_M (e), and k_cat/K_M (f) for point mutants of the ribozyme core. Dotted magenta lines indicate consensus ribozyme values. g Scatterplot of K_M vs. k_cat for all variants for which both parameters were determined. Source data are available in the Source Data file.

Fig. 3. Catalytic consequences of insertions and deletions.

Heatmap of the catalytic rates for double-incorporation insertion and point deletion mutants of the core glmS ribozyme. Source data are available in the Source Data file.

Fig. 4. Effects of mutations on cleavage rate at 10 mM GlcN6P.

a Heatmap of rates for all measured single and double mutants, with the first constituent point mutation indicated on the ordinate and the second on the abscissa. Sequence regions participating in duplex structures are labeled and color-coded as in Fig. 1. Heatmap of cleavage rates of all single mutants and basepaired-partner double mutants superimposed on the ribozyme tertiary (b) and secondary (c) structure, with secondary elements indicated in the same color scheme as a. The structure in b is based on PDB 2Z75. Source data are available in the Source Data file.

Fig. 5. Double-mutant rescue interactions (for 10 mM GlcN6P).

a Heatmap of double-mutant rescue, as defined in the main text. Sequence regions participating in duplex structures are labeled and color-coded, as in Fig. 4. b–f Examples of rescue interactions, illustrating different candidate mechanisms: secondary-structure rearrangements (b, e), tertiary contacts seen in crystal structures (d, f), and novel tertiary interactions (c). The crystal structure in c, d, and f is PDB 2Z75. Source data are available in the Source Data file.