Riboswitches are motifs in the untranslated regions (UTRs) of RNA transcripts that sense metabolite levels and modulate the expression of the corresponding genes for metabolite import, export, synthesis, or degradation. All riboswitches contain an aptamer: an RNA structure that, upon binding ligand, folds to expose or sequester regulatory elements in the adjacent sequence through alternative nucleotide pairing. The coupling between ligand binding and aptamer folding is central to the regulatory mechanisms of thiamine pyrophosphate (TPP) riboswitches and has not been fully characterized. Here, we show that TPP aptamer folding can be decomposed into ligand-independent and -dependent steps that correspond to the formation of secondary and tertiary structures, respectively. We reconstructed the full energy landscape for folding of the wild-type (WT) aptamer and measured perturbations of this landscape arising from mutations or ligand binding. We show that TPP binding proceeds in two steps, from a weakly to a strongly bound state. Our data imply a hierarchical folding sequence, and provide a framework for understanding molecular mechanism throughout the TPP riboswitch family. Riboswitches that sense the essential coenzyme thiamine pyrophosphate (TPP) are found in all kingdoms of life, and regulate thiamine synthesis at the level of transcription, translation, or splicing (1-3). Members of the TPP riboswitch family share sequence elements, architectures, and modes of ligand binding (4-9). The TPP-binding aptamer in the 3′ UTR of the thiC gene from Arabidopsis thaliana possesses a "tuning-fork" architecture (10) comprising two sensor helix arms (P2∕3 and P4∕5) and a switch helix (P1), all stemming from a central junction (J2∕4) (Fig. 1A). The aptamer is thought to bind its ligand as the sensor helix arms are brought together, and bulges (J2∕3 and J4∕5) in the arms join to form a bipartite binding pocket. Purine riboswitches also resemble tuning forks (11, 12), but in contrast have a single binding pocket comprised largely of nucleotides from the central junction.It has been proposed that riboswitches may be sorted into two functional types (13): Type I and Type II, of which the purine and TPP riboswitches are prototypic examples, respectively [as more riboswitch sequences and structures have been determined, additional classification schemes have been discussed (1, 14)]. The two types are distinguished by binding pocket architecture and the scale of the structural rearrangement accompanying ligand binding, with Type I and Type II undergoing local and longdistance rearrangements, respectively. An earlier single-molecule study (15) of the Type I pbuE aptamer from Bacillus subtilis, which binds adenine, revealed that secondary and tertiary structure formation were interleaved during folding, in that a competent binding site (constituting a tertiary element) was formed prior to the closure of the base of the switch helix, P1 (a secondary element). Here, we have extended the single-molecule approaches used previously to in...
The folding dynamics of riboswitches are central to their ability to modulate gene expression in response to environmental cues. In most cases, a structural competition between the formation of a ligand-binding aptamer and an expression platform (or some other competing off-state) determines the regulatory outcome. Here, we review single-molecule studies of riboswitch folding and function, predominantly carried out using single-molecule FRET or optical trapping approaches. Recent results have supplied new insights into riboswitch folding energy landscapes, the mechanisms of ligand binding, the roles played by divalent ions, the applicability of hierarchical folding models, and kinetic vs. thermodynamic control schemes. We anticipate that future work, based on improved data sets and potentially combining multiple experimental techniques, will enable the development of more complete models for complex RNA folding processes.
The thiamine pyrophosphate (TPP) riboswitch is a cis-regulatory element in mRNA that modifies gene expression in response to TPP concentration. Its specificity is dependent upon conformational changes that take place within its aptamer domain. Here, the role of tertiary interactions in ligand binding was studied at the single-molecule level by combined force spectroscopy and Förster resonance energy transfer (smFRET), using an optical trap equipped for simultaneous smFRET. The ‘Force-FRET’ approach directly probes secondary and tertiary structural changes during folding, including events associated with binding. Concurrent transitions observed in smFRET signals and RNA extension revealed differences in helix-arm orientation between two previously-identified ligand-binding states that had been undetectable by spectroscopy alone. Our results show that the weaker binding state is able to bind to TPP, but is unable to form a tertiary docking interaction that completes the binding process. Long-range tertiary interactions stabilize global riboswitch structure and confer increased ligand specificity.DOI: http://dx.doi.org/10.7554/eLife.12362.001
There is broad interest in creating RL agents that can solve many (related) tasks and adapt to new tasks and environments after initial training. Model-based RL leverages learned surrogate models that describe dynamics and rewards of individual tasks, such that planning in a good surrogate can lead to good control of the true system. Rather than solving each task individually from scratch, hierarchical models can exploit the fact that tasks are often related by (unobserved) causal factors of variation in order to achieve efficient generalization, as in learning how the mass of an item affects the force required to lift it can generalize to previously unobserved masses. We propose Generalized Hidden Parameter MDPs (GHP-MDPs) that describe a family of MDPs where both dynamics and reward can change as a function of hidden parameters that vary across tasks. The GHP-MDP augments model-based RL with latent variables that capture these hidden parameters, facilitating transfer across tasks. We also explore a variant of the model that incorporates explicit latent structure mirroring the causal factors of variation across tasks (for instance: agent properties, environmental factors, and goals). We experimentally demonstrate state-of-the-art performance and sample-efficiency on a new challenging MuJoCo task using reward and dynamics latent spaces, while beating a previous state-of-the-art baseline with > 10× less data. Using test-time inference of the latent variables, our approach generalizes in a single episode to novel combinations of dynamics and reward, and to novel rewards.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.