A small RNA that cooperatively senses two stacked metabolites in one pocket for gene control

Abstract

Riboswitches are structured non-coding RNAs often located upstream of essential genes in bacterial messenger RNAs. Such RNAs regulate expression of downstream genes by recognizing a specific cellular effector. Although nearly 50 riboswitch classes are known, only a handful recognize multiple effectors. Here, we report the 2.60-Å resolution co-crystal structure of a class I type I preQ₁-sensing riboswitch that reveals two effectors stacked atop one another in a single binding pocket. These effectors bind with positive cooperativity in vitro and both molecules are necessary for gene regulation in bacterial cells. Stacked effector recognition appears to be a hallmark of the largest subgroup of preQ₁ riboswitches, including those from pathogens such as Neisseria gonorrhoeae. We postulate that binding to stacked effectors arose in the RNA World to closely position two substrates for RNA-mediated catalysis. These findings expand known effector recognition capabilities of riboswitches and have implications for antimicrobial development.

Fig. 1. Queuosine biosynthesis, the preQ₁ riboswitch consensus model and co-crystal structure of the Carnobacterium antarcticus (Can) riboswitch.

a The queuosine (Q) biosynthetic pathway proceeds through the 7-deazapurine metabolite preQ₁. b PreQ₁-I riboswitch subtypes shown as secondary structures based on covariation. Red, black and gray positions indicate 97, 90, and 75% sequence conservation. Asterisk indicates a specificity base predicted to recognize preQ₁. c Secondary structure of the Can riboswitch. Colors correspond to specific pseudoknot base pairing (P) and loop (L) sequences. PreQ₁ is depicted as “Q”. Noncanonical pairing is indicated by Leontis–Westhof symbols. The Shine-Dalgarno sequence (SDS) and anti-(a)SDS are highlighted in yellow and cyan. d Ribbon diagram of the global Can riboswitch fold. e Binding pocket floor overview wherein floor bases comprise the A28•G5-C18 base triple. Dashed lines depict hydrogen bonds here and elsewhere. f Overview of the pocket ceiling, which comprises the U32•A12•C8 base triple. The view highlights P2 bases in the aSDS and SDS.

Fig. 2. The Can preQ₁-I_I riboswitch pocket with two preQ₁ ligands and confirmation of ligand-to-RNA stoichiometry.

a Overview of fully occupied binding pocket. Interactions in the (b) α site and (c) β site. d Representative ITC experiment with titration of preQ₁ into WT Can RNA. The binding constant K_D, ligand-to-RNA stoichiometry N, and c value are shown.

Fig. 3. Riboswitch reporter assay and dose response in live bacteria.

a Schematic of the plasmid reporter. b Two-site binding model wherein preQ₁ can bind either site first. c Average GFPuv emission dependence on preQ₁; (inset) one-site binding by the Lrh preQ₁ riboswitch. d Bar graph showing fold repression of GFPuv emission for the Can, Lrh and mutant riboswitches with individual points shown. e Bar graph showing fold change in average EC₅₀ relative to Can riboswitch EC_50,2. Significance was determined by a two-tailed Student’s t test with Welch’s Correction (n = 3 biological replicates. *p ≤ 0.05). S.E.M. is shown in c and d; propagated errors are shown in e.