Using in-cell SHAPE-Seq and simulations to probe structure–function design principles of RNA transcriptional regulators

Abstract

Antisense RNA-mediated transcriptional regulators are powerful tools for controlling gene expression and creating synthetic gene networks. RNA transcriptional repressors derived from natural mechanisms called attenuators are particularly versatile, though their mechanistic complexity has made them difficult to engineer. Here we identify a new structure-function design principle for attenuators that enables the forward engineering of new RNA transcriptional repressors. Using in-cell SHAPE-Seq to characterize the structures of attenuator variants within Escherichia coli, we show that attenuator hairpins that facilitate interaction with antisense RNAs require interior loops for proper function. Molecular dynamics simulations of these attenuator variants suggest these interior loops impart structural flexibility. We further observe hairpin flexibility in the cellular structures of natural RNA mechanisms that use antisense RNA interactions to repress translation, confirming earlier results from in vitro studies. Finally, we design new transcriptional attenuators in silico using an interior loop as a structural requirement and show that they function as desired in vivo. This work establishes interior loops as an important structural element for designing synthetic RNA gene regulators. We anticipate that the coupling of experimental measurement of cellular RNA structure and function with computational modeling will enable rapid discovery of structure-function design principles for a diverse array of natural and synthetic RNA regulators.

Comparison of SHAPE reagents for mapping RNA structures inside living cells

“…However, recent reports have suggested that 1M7, an isatoic anhydride reagent related to NMIA, was able to robustly modify both ribosomal RNA as well as mRNAs inside both mammalian and bacterial cells (McGinnis et al 2015;Smola et al 2015;Takahashi et al 2016;Watters et al 2016). These apparently contradictory results stimulated us to further investigate the differences between the two types of SHAPE reagents.…”

“…Published work has reported in vivo modification of RNA in E. coli cells; however, no cDNA truncation blots were shown (McGinnis et al 2015;Smola et al 2015;Takahashi et al 2016). To address the possibility that our lack of in vivo signal using 1M7 is due to the type of cells used, we repeated the SHAPE experiment in HST08 E. coli cells following precisely the published protocols for cell culture, RNA modification, and RNA extraction (McGinnis et al 2015).…”

Achieving large dynamic range control of gene expression with a compact RNA transcription-translation regulator

“…To test this hypothesis, we truncated the antisense RNA sequence to the elements necessary for initial RNA-RNA kissinghairpin interactions that were shown to be essential for the transcriptional regulatory decision (11). Specifically, hairpin 2 of the pT181 antisense makes contact with the first hairpin of the sense target region of the attenuator that contains the anti-terminator.…”

“…Antisense-mediated RNA transcriptional regulators are particularly versatile because they regulate RNA synthesis as a function of an RNA input and thus can be used to create RNA-only genetic circuitry (2,10). RNA genetic circuits have many potential advantages over proteinbased circuits including the possibility of leveraging advances in RNA folding algorithms and design rules for part design (11,12) and their natural fast circuit dynamics (10). However, RNA transcriptional regulators still suffer from low dynamic range in comparison to protein-based regulators.…”

Exploring RNA structure and dynamics through enhanced sampling simulations

“…The AMBER force field is by far the most adopted empirical force field. In some works, alternative modifications were tested, including modified non-bonded parameters [17,18,22,25,38,39] or dihedral reparametrizations [24,26,39,53]. A limited number of applications used the CHARMM force field, either for very short simulations [35] or for simulations where RNA backbone was constrained [55].…”