Riboswitches are structural RNA elements that control gene expression. These naturally occurring RNA sensors are of continued interest as antibiotic targets, molecular sensors, and functional elements of synthetic circuits. Here, we describe affinity-based profiling of the flavin mononucleotide (FMN) riboswitch to characterize ligand binding and structural folding. We designed and synthesized photoreactive ligands and used them for photoaffinity labeling. We showed selective labeling of the FMN riboswitch and used this covalent interaction to quantitatively measure ligand binding, which we demonstrate with the naturally occurring antibiotic roseoflavin. We measured conditional riboswitch folding as a function of temperature and cation concentration. Furthermore, combining photoaffinity labeling with reverse transcription revealed ligand binding sites within the aptamer domain with single-nucleotide resolution. The photoaffinity probe was applied to cellular extracts of Bacillus subtilis to demonstrate conditional folding of the endogenous low-abundant ribD FMN riboswitch in biologically derived samples using quantitative PCR. Lastly, binding of the riboswitch-targeting antibiotic roseoflavin to the FMN riboswitch was measured in live bacteria using the photoaffinity probe.

No abstract

Introduction Memory use For the investigation of the performance of a mathematical filter for variance reduction, a previously described [1-I computer program for peak recognition and data-processing for a continuous-flow system was adapted to accomodate subroutines for base-line drift compensation and between-run variance reduction. The subroutine for between-run variance reduction was based on the ideas and algorithms reported by Jansen and Bonants [2]. Whereas for most components determined with continuousflow analysis a within-run coefficient of variation (CV) of about 1-1.5% can be obtained, which compares favourably with newer analytical techniques like centrifugal analysis, the between-run CV of 2"5-3.5% was judged to be too large. From the work of Jansen and Bonants it can be concluded that between-run variance reductions between 40 and 70% are obtainable by applying a digital filtering technique to the information content of control sera placed and analysed in every run.