Riboswitches enable microbes to rapidly respond to changing levels of metabolites. A high-throughput platform reveals how RNA structural transitions kinetically compete during transcription in a new mechanism for riboswitch function.
To respond to an ever-changing environment, bacteria rely on RNA regulatory elements, which can change gene expression within seconds upon receiving a chemical signal. In this issue, Strobel et al.1 use a new method for visualizing RNA folding during transcription to show how a riboswitch responds to the alarmone ribonucleotide ZMP. Their high-throughput platform enables testing of many riboswitch variants and reveals the mechanism of transcriptional control.
Riboswitches, as their name suggests, function as molecular switches to turn a downstream gene ‘on’ or ‘off’ in response to ligand concentration2. The actual switch in gene expression is created by a change in the mRNA’s structure upon binding of a ligand that controls transcription termination or translation of the mRNA. The Clostridium beijerinckii ZMP riboswitch acts by forming an anti-termination pseudoknot that is stabilized by ZMP binding3. In the absence of ZMP, a terminator hairpin replaces the pseudoknot, stopping mRNA synthesis before the coding region (Fig. 1).
Riboswitches that regulate transcription operate on a tight time frame because they must respond to the ligand before RNA polymerase passes by the terminator downstream. Biophysical studies showed that the probability of transcription readthrough is controlled by the forward kinetics of ligand binding relative to the speed of RNA elongation4,5. This kinetically controlled switch depends on folding intermediates that poise the riboswitch to change conformation if the ligand binds. Therefore, how the RNA folds during transcription is crucial for riboswitch operation. Co-transcriptional folding has been challenging to study, however, owing to a lack of tools for measuring RNA structures within an ensemble of transcription intermediates.
Strobel et al.1 overcome this barrier to examine initial folding of the ZMP riboswitch using a clever co-transcriptional SHAPE-Seq6 method (Fig. 1). In this method, biotin–streptavidin roadblocks stall RNA polymerase at each position along the DNA template, generating transcripts of different lengths. The unpaired nucleotides are probed using SHAPE chemistry7 and read out using high-throughput sequencing, generating a matrix that illustrates the RNA structure at each step of transcription elongation. By comparing the folding pattern with and without ZMP, it is possible to pinpoint how much of the riboswitch must be transcribed for it to respond to ZMP binding.
A surprising result is that ZMP cannot stabilize the anti-termination pseudoknot until a 3′ hairpin (P3) is transcribed (Fig. 1). This suggests that the P3 hairpin itself, and not just the pseudoknot-forming nucleotides, plays an important role in the switch mechanism. To test this idea, they used high-throughput combinatorial mutagenesis to exhaustively search for mutations in the P3 stem and terminator hairpin that alter riboswitch function. The results showed that metastability of the P3 hairpin is critical for riboswitch function: too stable, transcription never terminates; too weak, and the anti-termination pseudoknot cannot form. Mutations in the terminator compensate mutations in P3, indicating that these two elements compete during the on/off response to ZMP.
The high-throughput pipeline for RNA-structure probing and combinatorial mutagenesis developed by Strobel et al.1 provides a fine-grained view of which RNA structures ‘win out’ during transcription and in response to ZMP. This powerful pipeline can be applied to many types of regulatory RNAs, revealing design principles for RNA switches. Roadblocks simplify the co-transcriptional folding pattern by allowing time for the RNA to adopt the most favored conformations at each step of elongation. A caveat to this approach is that it likely underestimates the number of transient structures present during RNA synthesis. In the future, extending RNA-structure probing to uninterrupted transcription may provide clearer timing of folding events.
Co-transcriptional folding is widely important, because base pairs form rapidly (microseconds to milliseconds) relative to the rate of RNA synthesis (seconds to minutes). As a result, local structures at the 5′ end form long before downstream sequences are transcribed8. By contrast, long-range tertiary interactions such as the ZMP anti-termination pseudoknot often form on timescales similar to those of transcription, creating an opportunity for regulation.
Interestingly, RNA structures formed early in transcription are often counterproductive and must be resolved, as was seen for vectorial folding of the twister ribozyme9. Indeed, Strobel et al.1 detected an initial non-native 5′ hairpin that later refolds into a native structure. In this sense, co-transcriptional folding may be viewed as an inevitable hazard that must be corrected through RNA-mediated conformational switching (such as strand invasion) or with the help of RNA chaperone proteins. Certain transient non-native structures may improve RNA function by facilitating later rearrangements.
The disparate timescales for RNA base pairing, tertiary folding and transcription means that transcripts frequently fold into more than one structure, leading to conformational exchange. Optimized switching mechanisms are likely widespread in non-coding RNAs that often include long-range interactions. Riboswitches demonstrate how these intrinsic features of RNA can be harnessed for rapid genetic control. High-throughput platforms such as the one used by Strobel et al.1 provide a new way of visualizing the regulatory opportunities created by kinetic coupling of synthesis and folding.
Strobel, E.J., Cheng, L., Berman, K.E., Carlson, P.D. & Lucks, J.B. Nat. Chem. Biol. https://doi.org/10.1038/s41589-019-0382-7 (2019).
Jones, C. P. & Ferré-D’Amaré, A. R. Annu. Rev. Biophys. 46, 1–27 (2016).
Kim, P. B., Nelson, J. W. & Breaker, R. R. Mol. Cell 57, 317–328 (2015).
Wickiser, K. J., Winkler, W. C., Breaker, R. R. & Crothers, D. M. Mol. Cell 18, 49–60 (2005).
Frieda, K. L. & Block, S. M. Science 338, 397–400 (2012).
Strobel, E. J., Watters, K. E., Nedialkov, Y., Artsimovitch, I. & Lucks, J. B. Nucleic Acids Res. 45, e109 (2017).
Wilkinson, K. A., Merino, E. J. & Weeks, K. M. Nat. Protoc. 1, 1610–1616 (2006).
Lai, D., Proctor, J. R. & Meyer, I. M. RNA 19, 1461–1473 (2013).
Hua, B., Panja, S., Wang, Y., Woodson, S. & Ha, T. J. Am. Chem. Soc. 140, 10067–10070 (2018).
The authors declare no competing interests.
About this article
Cite this article
Rodgers, M.L., Hao, Y. & Woodson, S.A. A newborn RNA switches its fate. Nat Chem Biol 15, 1031–1032 (2019). https://doi.org/10.1038/s41589-019-0391-6
PLOS Computational Biology (2020)
Homogeneous and Universal Detection of Various Targets with a Dual‐Step Transduced Toehold Switch Sensor