Applicability of a computational design approach for synthetic riboswitches

Improving Microbial Cell Factory Performance by Engineering SAM Availability

Methylated natural products are widely spread in nature. S-Adenosyl-l-methionine (SAM) is the secondary abundant cofactor and the primary methyl donor, which confer natural products with structural and functional diversification. The increasing demand for SAM-dependent natural products (SdNPs) has motivated the development of microbial cell factories (MCFs) for sustainable and efficient SdNP production. Insufficient and unsustainable SAM availability hinders the improvement of SdNP MCF performance. From the perspective of developing MCF, this review summarized recent understanding of de novo SAM biosynthesis and its regulatory mechanism. SAM is just the methyl mediator but not the original methyl source. Effective and sustainable methyl source supply is critical for efficient SdNP production. We compared and discussed the innate and relatively less explored alternative methyl sources and identified the one involving cheap one-carbon compound as more promising. The SAM biosynthesis is synergistically regulated on multilevels and is tightly connected with ATP and NAD(P)H pools. We also covered the recent advancement of metabolic engineering in improving intracellular SAM availability and SdNP production. Dynamic regulation is a promising strategy to achieve accurate and dynamic fine-tuning of intracellular SAM pool size. Finally, we discussed the design and engineering constraints underlying construction of SAM-responsive genetic circuits and envisioned their future applications in developing SdNP MCFs.

Mesoplasma florum: a near-minimal model organism for systems and synthetic biology

Mesoplasma florum is an emerging model organism for systems and synthetic biology due to its small genome (∼800 kb) and fast growth rate. While M. florum was isolated and first described almost 40 years ago, many important aspects of its biology have long remained uncharacterized due to technological limitations, the absence of dedicated molecular tools, and since this bacterial species has not been associated with any disease. However, the publication of the first M. florum genome in 2004 paved the way for a new era of research fueled by the rise of systems and synthetic biology. Some of the most important studies included the characterization and heterologous use of M. florum regulatory elements, the development of the first replicable plasmids, comparative genomics and transposon mutagenesis, whole-genome cloning in yeast, genome transplantation, in-depth characterization of the M. florum cell, as well as the development of a high-quality genome-scale metabolic model. The acquired data, knowledge, and tools will greatly facilitate future genome engineering efforts in M. florum, which could next be exploited to rationally design and create synthetic cells to advance fundamental knowledge or for specific applications.

Developing Complex RNA Design Applications in the Infrared Framework

Applications in biotechnology and bio-medical research call for effective strategies to design novel RNAs with very specific properties. Such advanced design tasks require support by computational tools but at the same time put high demands on their flexibility and expressivity to model the application-specific requirements. To address such demands, we present the computational framework Infrared. It supports developing advanced customized design tools, which generate RNA sequences with specific properties, often in a few lines of Python code. This text guides the reader in tutorial format through the development of complex design applications. Thanks to the declarative, compositional approach of Infrared, we can describe this development as a step-by-step extension of an elementary design task. Thus, we start with generating sequences that are compatible with a single RNA structure and go all the way to RNA design targeting complex positive and negative design objectives with respect to single or even multiple target structures. Finally, we present a "real-world" application of computational design to create an RNA device for biotechnology: we use Infrared to generate design candidates of an artificial "AND" riboswitch, which activates gene expression in the simultaneous presence of two different small metabolites. In these applications, we exploit that the system can generate, in an efficient (fixed-parameter tractable) way, multiple diverse designs that satisfy a number of constraints and have high quality w.r.t. to an objective (by sampling from a Boltzmann distribution).

Engineering U1-Based Tetracycline-Inducible Riboswitches to Control Gene Expression in Mammals

Synthetic riboswitches are promising regulatory devices due to their small size, lack of immunogenicity, and ability to fine-tune gene expression in the absence of exogenous trans-acting factors. Based on a gene inhibitory system developed at our lab, termed U1snRNP interference (U1i), we developed tetracycline (TC)-inducible riboswitches that modulate mRNA polyadenylation through selective U1 snRNP recruitment. First, we engineered different TC-U1i riboswitches, which repress gene expression unless TC is added, leading to inductions of gene expression of 3-to-4-fold. Second, we developed a technique called Systematic Evolution of Riboswitches by Exponential Enrichment (SEREX), to isolate riboswitches with enhanced U1 snRNP binding capacity and activity, achieving inducibilities of up to 8-fold. Interestingly, by multiplexing riboswitches we increased inductions up to 37-fold. Finally, we demonstrated that U1i-based riboswitches are dose-dependent and reversible and can regulate the expression of reporter and endogenous genes in culture cells and mouse models, resulting in attractive systems for gene therapy applications. Our work probes SEREX as a much-needed technology for the in vitro identification of riboswitches capable of regulating gene expression in vivo.

Elimination of editing plasmid mediated by theophylline riboswitch in Zymomonas mobilis

Zymomonas mobilis is regarded as a potential chassis for the production of platform chemicals. Genome editing using the CRISPR-Cas system could meet the need for gene modification in metabolic engineering. However, the low curing efficiency of CRISPR editing plasmid is a common bottleneck in Z. mobilis. In this study, we utilized a theophylline-dependent riboswitch to regulate the expression of the replicase gene of the editing plasmid, thereby promoting the elimination of exogeneous plasmid. The riboswitch D (RSD) with rigorous regulatory ability was identified as the optimal candidate by comparing the transformation efficiency of four theophylline riboswitch-based backbone editing plasmids, and the optimal theophylline concentration for inducing RSD was determined to be 2 mM. A highly effective method for eliminating the editing plasmid, cells with RSD-based editing plasmid which were cultured in liquid and solid RM media in alternating passages at 37 °C without shaking, was established by testing the curing efficiency of backbone editing plasmids pMini and pMini-RSD in RM medium with or without theophylline at 30 °C or 37 °C. Finally, the RSD-based editing plasmid was applied to genome editing, resulting in an increase of more than 10% in plasmid elimination efficiency compared to that of pMini-based editing plasmid. KEY POINTS: • An effective strategy for curing CRISPR editing plasmid has been established in Z. mobilis. • Elimination efficiency of the CRISPR editing plasmid was enhanced by 10% to 20% under the regulation of theophylline-dependent riboswitch RSD.

Riboswitches, from cognition to transformation

Riboswitches are functional RNA elements that regulate gene expression by directly detecting metabolites. Twenty years have passed since it was first discovered, researches on riboswitches are becoming increasingly standardized and refined, which could significantly promote people's cognition of RNA function as well. Here, we focus on some representative orphan riboswitches, enumerate the structural and functional transformation and artificial design of riboswitches including the coupling with ribozymes, hoping to attain a comprehensive understanding of riboswitch research.

Tapping the potential of synthetic riboswitches: reviewing the versatility of the tetracycline aptamer

Synthetic riboswitches are a versatile class of regulatory elements that are becoming increasingly established in synthetic biology applications. They are characterized by their compact size and independence from auxiliary protein factors. While naturally occurring riboswitches were mostly discovered in bacteria, synthetic riboswitches have been designed for all domains of life. Published design strategies far exceed the number of riboswitches found in nature. A core element of any riboswitch is a binding domain, called an aptamer, which is characterized by high specificity and affinity for its ligand. Aptamers can be selected de novo, allowing the design of synthetic riboswitches against a broad spectrum of targets. The tetracycline aptamer has proven to be well suited for riboswitch engineering. Since its selection, it has been used in a variety of applications and is considered to be well established and characterized. Using the tetracycline aptamer as an example, we aim to discuss a large variety of design approaches for synthetic riboswitch engineering and their application. We aim to demonstrate the versatility of riboswitches in general and the high potential of synthetic RNA devices for creating new solutions in both the scientific and medical fields.

Could artificial intelligence revolutionize the development of nanovectors for gene therapy and mRNA vaccines?

Gene therapy enables the introduction of nucleic acids like DNA and RNA into host cells, and is expected to revolutionize the treatment of a wide range of diseases. This growth has been further accelerated by the discovery of CRISPR/Cas technology, which allows accurate genomic editing in a broad range of cells and organisms in vitro and in vivo. Despite many advances in gene delivery and the development of various viral and non-viral gene delivery vectors, the lack of highly efficient non-viral systems with low cellular toxicity remains a challenge. The application of cutting-edge technologies such as artificial intelligence (AI) has great potential to find new paradigms to solve this issue. Herein, we review AI and its major subfields including machine learning (ML), neural networks (NNs), expert systems, deep learning (DL), computer vision and robotics. We discuss the potential of AI-based models and algorithms in the design of targeted gene delivery vehicles capable of crossing extracellular and intracellular barriers by viral mimicry strategies. We finally discuss the role of AI in improving the function of CRISPR/Cas systems, developing novel nanobots, and mRNA vaccine carriers.

Electrochemical Aptasensors for Antibiotics Detection: Recent Achievements and Applications for Monitoring Food Safety

Antibiotics are often used in human and veterinary medicine for the treatment of bacterial diseases. However, extensive use of antibiotics in agriculture can result in the contamination of common food staples such as milk. Consumption of contaminated products can cause serious illness and a rise in antibiotic resistance. Conventional methods of antibiotics detection such are microbiological assays chromatographic and mass spectroscopy methods are sensitive; however, they require qualified personnel, expensive instruments, and sample pretreatment. Biosensor technology can overcome these drawbacks. This review is focused on the recent achievements in the electrochemical biosensors based on nucleic acid aptamers for antibiotic detection. A brief explanation of conventional methods of antibiotic detection is also provided. The methods of the aptamer selection are explained, together with the approach used for the improvement of aptamer affinity by post-SELEX modification and computer modeling. The substantial focus of this review is on the explanation of the principles of the electrochemical detection of antibiotics by aptasensors and on recent achievements in the development of electrochemical aptasensors. The current trends and problems in practical applications of aptasensors are also discussed.

Approaches in the photosynthetic production of sustainable fuels by cyanobacteria using tools of synthetic biology

Cyanobacteria, photosynthetic prokaryotic microorganisms having a simple genetic composition are the prospective photoautotrophic cell factories for the production of a wide range of biofuel molecules. The simple genetic composition of cyanobacteria allows effortless genetic manipulation which leads to increased research endeavors from the synthetic biology approach. Various unicellular model cyanobacterial strains like Synechocystis sp. PCC 6803 and Synechococcus elongatus PCC 7942 have been successfully engineered for biofuels generation. Improved development of synthetic biology tools, genetic modification methods and advancement in transformation techniques to construct a strain that can contain multiple foreign genes in a single operon have vastly expanded the functions that can be used for engineering photosynthetic cyanobacteria for the generation of various biofuel molecules. In this review, recent advancements and approaches in synthetic biology tools used for cyanobacterial genome editing have been discussed. Apart from this, cyanobacterial productions of various fuel molecules like isoprene, limonene, α-farnesene, squalene, alkanes, butanol, and fatty acids, which can be a substitute for petroleum and fossil fuels in the future, have been elaborated.

Cell-free riboswitches

The emerging community of cell-free synthetic biology aspires to build complex biochemical and genetic systems with functions that mimic or even exceed those in living cells. To achieve such functions, cell-free systems must be able to sense and respond to the complex chemical signals within and outside the system. Cell-free riboswitches can detect chemical signals via RNA-ligand interaction and respond by regulating protein synthesis in cell-free protein synthesis systems. In this article, we review synthetic cell-free riboswitches that function in both prokaryotic and eukaryotic cell-free systems reported to date to provide a current perspective on the state of cell-free riboswitch technologies and their limitations.

Riboswitches for Controlled Expression of Therapeutic Transgenes Delivered by Adeno-Associated Viral Vectors

Vectors developed from adeno-associated virus (AAV) are powerful tools for in vivo transgene delivery in both humans and animal models, and several AAV-delivered gene therapies are currently approved for clinical use. However, AAV-mediated gene therapy still faces several challenges, including limited vector packaging capacity and the need for a safe, effective method for controlling transgene expression during and after delivery. Riboswitches, RNA elements which control gene expression in response to ligand binding, are attractive candidates for regulating expression of AAV-delivered transgene therapeutics because of their small genomic footprints and non-immunogenicity compared to protein-based expression control systems. In addition, the ligand-sensing aptamer domains of many riboswitches can be exchanged in a modular fashion to allow regulation by a variety of small molecules, proteins, and oligonucleotides. Riboswitches have been used to regulate AAV-delivered transgene therapeutics in animal models, and recently developed screening and selection methods allow rapid isolation of riboswitches with novel ligands and improved performance in mammalian cells. This review discusses the advantages of riboswitches in the context of AAV-delivered gene therapy, the subsets of riboswitch mechanisms which have been shown to function in human cells and animal models, recent progress in riboswitch isolation and optimization, and several examples of AAV-delivered therapeutic systems which might be improved by riboswitch regulation.

A convolutional neural network for the prediction and forward design of ribozyme-based gene-control elements

Ribozyme switches are a class of RNA-encoded genetic switch that support conditional regulation of gene expression across diverse organisms. An improved elucidation of the relationships between sequence, structure, and activity can improve our capacity for de novo rational design of ribozyme switches. Here, we generated data on the activity of hundreds of thousands of ribozyme sequences. Using automated structural analysis and machine learning, we leveraged these large data sets to develop predictive models that estimate the in vivo gene-regulatory activity of a ribozyme sequence. These models supported the de novo design of ribozyme libraries with low mean basal gene-regulatory activities and new ribozyme switches that exhibit changes in gene-regulatory activity in the presence of a target ligand, producing functional switches for four out of five aptamers. Our work examines how biases in the model and the data set that affect prediction accuracy can arise and demonstrates that machine learning can be applied to RNA sequences to predict gene-regulatory activity, providing the basis for design tools for functional RNAs.

Beyond Plug and Pray: Context Sensitivity and in silico Design of Artificial Neomycin Riboswitches

Gene regulation in prokaryotes often depends on RNA elements such as riboswitches or RNA thermometers located in the 5' untranslated region of mRNA. Rearrangements of the RNA structure in response, e.g., to the binding of small molecules or ions control translational initiation or premature termination of transcription and thus mRNA expression. Such structural responses are amenable to computational modelling, making it possible to rationally design synthetic riboswitches for a given aptamer. Starting from an artificial aptamer, we construct the first synthetic transcriptional riboswitches that respond to the antibiotic neomycin. We show that the switching behaviour in vivo critically depends not only on the sequence of the riboswitch itself, but also on its sequence context. We therefore developed in silico methods to predict the impact of the context, making it possible to adapt the design and to rescue non-functional riboswitches. We furthermore analyse the influence of 5' hairpins with varying stability on neomycin riboswitch activity. Our data highlight the limitations of a simple plug-and-play approach in the design of complex genetic circuits and demonstrate that detailed computational models significantly simplify, improve, and automate the design of transcriptional circuits. Our design software is available under a free licence on GitHub (https://github.com/xileF1337/riboswitch_design).

Artificial Intelligence in Aptamer-Target Binding Prediction

Aptamers are short single-stranded DNA, RNA, or synthetic Xeno nucleic acids (XNA) molecules that can interact with corresponding targets with high affinity. Owing to their unique features, including low cost of production, easy chemical modification, high thermal stability, reproducibility, as well as low levels of immunogenicity and toxicity, aptamers can be used as an alternative to antibodies in diagnostics and therapeutics. Systematic evolution of ligands by exponential enrichment (SELEX), an experimental approach for aptamer screening, allows the selection and identification of in vitro aptamers with high affinity and specificity. However, the SELEX process is time consuming and characterization of the representative aptamer candidates from SELEX is rather laborious. Artificial intelligence (AI) could help to rapidly identify the potential aptamer candidates from a vast number of sequences. This review discusses the advancements of AI pipelines/methods, including structure-based and machine/deep learning-based methods, for predicting the binding ability of aptamers to targets. Structure-based methods are the most used in computer-aided drug design. For this part, we review the secondary and tertiary structure prediction methods for aptamers, molecular docking, as well as molecular dynamic simulation methods for aptamer-target binding. We also performed analysis to compare the accuracy of different secondary and tertiary structure prediction methods for aptamers. On the other hand, advanced machine-/deep-learning models have witnessed successes in predicting the binding abilities between targets and ligands in drug discovery and thus potentially offer a robust and accurate approach to predict the binding between aptamers and targets. The research utilizing machine-/deep-learning techniques for prediction of aptamer-target binding is limited currently. Therefore, perspectives for models, algorithms, and implementation strategies of machine/deep learning-based methods are discussed. This review could facilitate the development and application of high-throughput and less laborious in silico methods in aptamer selection and characterization.

Ligand-dependent tRNA processing by a rationally designed RNase P riboswitch

We describe a synthetic riboswitch element that implements a regulatory principle which directly addresses an essential tRNA maturation step. Constructed using a rational in silico design approach, this riboswitch regulates RNase P-catalyzed tRNA 5′-processing by either sequestering or exposing the single-stranded 5′-leader region of the tRNA precursor in response to a ligand. A single base pair in the 5′-leader defines the regulatory potential of the riboswitch both in vitro and in vivo. Our data provide proof for prior postulates on the importance of the structure of the leader region for tRNA maturation. We demonstrate that computational predictions of ligand-dependent structural rearrangements can address individual maturation steps of stable non-coding RNAs, thus making them amenable as promising target for regulatory devices that can be used as functional building blocks in synthetic biology.

Methods and Applications of In Silico Aptamer Design and Modeling

Aptamers are nucleic acid analogues of antibodies with high affinity to different targets, such as cells, viruses, proteins, inorganic materials, and coenzymes. Empirical approaches allow the design of in vitro aptamers that bind particularly to a target molecule with high affinity and selectivity. Theoretical methods allow significant expansion of the possibilities of aptamer design. In this study, we review theoretical and joint theoretical-experimental studies dedicated to aptamer design and modeling. We consider aptamers with different targets, such as proteins, antibiotics, organophosphates, nucleobases, amino acids, and drugs. During nucleic acid modeling and in silico design, a full set of in silico methods can be applied, such as docking, molecular dynamics (MD), and statistical analysis. The typical modeling workflow starts with structure prediction. Then, docking of target and aptamer is performed. Next, MD simulations are performed, which allows for an evaluation of the stability of aptamer/ligand complexes and determination of the binding energies with higher accuracy. Then, aptamer/ligand interactions are analyzed, and mutations of studied aptamers made. Subsequently, the whole procedure of molecular modeling can be reiterated. Thus, the interactions between aptamers and their ligands are complex and difficult to understand using only experimental approaches. Docking and MD are irreplaceable when aptamers are studied in silico.

Evaluating Terminator Strength Based on Differentiating Effects on Transcription and Translation

Transcription terminators play a role in terminating the progress of gene transcription, and are thus essential elements in the gene circuit. Terminators have two main functions: terminating gene transcription and improving the stability of gene transcripts during translation. We therefore considered the detailed characteristics of terminators in relation to their different roles in gene transcription and translation, including transcription shut-down degree (α) and upstream mRNA protection capacity (β), and apparent termination efficiency (η) reflecting the overall regulatory effect of the terminator. Based on a dual-reporter gene system, we constructed three terminator-probe plasmids to investigate each characteristic in Escherichia coli. According to multiple regression analysis, the transcription shut-down degree and the upstream mRNA protection capacity contributed almost equally to the apparent termination efficiency. Sequence analysis of 12 terminators demonstrated that the terminator sequence was dominated by GC bases, and that a high ratio of GC bases in the stem structure of terminators might be associated with a high degree of transcription shut-down. This comprehensive characterization of terminators furthers our understanding of the role of terminators in gene expression and provides a guide for synthetic terminator design.

Riboflow: Using Deep Learning to Classify Riboswitches With ∼99% Accuracy

Riboswitches are cis-regulatory genetic elements that use an aptamer to control gene expression. Specificity to cognate ligand and diversity of such ligands have expanded the functional repertoire of riboswitches to mediate mounting apt responses to sudden metabolic demands and signal changes in environmental conditions. Given their critical role in microbial life, riboswitch characterisation remains a challenging computational problem. Here we have addressed the issue with advanced deep learning frameworks, namely convolutional neural networks (CNN), and bidirectional recurrent neural networks (RNN) with Long Short-Term Memory (LSTM). Using a comprehensive dataset of 32 ligand classes and a stratified train-validate-test approach, we demonstrated the accurate performance of both the deep learning models (CNN and RNN) relative to conventional hyperparameter-optimized machine learning classifiers on all key performance metrics, including the ROC curve analysis. In particular, the bidirectional LSTM RNN emerged as the best-performing learning method for identifying the ligand-specificity of riboswitches with an accuracy >0.99 and macro-averaged F-score of 0.96. An additional attraction is that the deep learning models do not require prior feature engineering. A dynamic update functionality is built into the models to factor for the constant discovery of new riboswitches, and extend the predictive modeling to new classes. Our work would enable the design of genetic circuits with custom-tuned riboswitch aptamers that would effect precise translational control in synthetic biology. The associated software is available as an open-source Python package and standalone resource for use in genome annotation, synthetic biology, and biotechnology workflows.

Small but large enough: structural properties of armless mitochondrial tRNAs from the nematode Romanomermis culicivorax

As adapter molecules to convert the nucleic acid information into the amino acid sequence, tRNAs play a central role in protein synthesis. To fulfill this function in a reliable way, tRNAs exhibit highly conserved structural features common in all organisms and in all cellular compartments active in translation. However, in mitochondria of metazoans, certain dramatic deviations from the consensus tRNA structure are described, where some tRNAs lack the D- or T-arm without losing their function. In Enoplea, this miniaturization comes to an extreme, and functional mitochondrial tRNAs can lack both arms, leading to a considerable size reduction. Here, we investigate the secondary and tertiary structure of two such armless tRNAs from Romanomermis culicivorax. Despite their high AU content, the transcripts fold into a single and surprisingly stable hairpin structure, deviating from standard tRNAs. The three-dimensional form is boomerang-like and diverges from the standard L-shape. These results indicate that such unconventional miniaturized tRNAs can still fold into a tRNA-like shape, although their length and secondary structure are very unusual. They highlight the remarkable flexibility of the protein synthesis apparatus and suggest that the translational machinery of Enoplea mitochondria may show compensatory adaptations to accommodate these armless tRNAs for efficient translation.

Molecular mechanisms for dynamic regulation of N1 riboswitch by aminoglycosides

A synthetic riboswitch N1, inserted into the 5'-untranslated mRNA region of yeast, regulates gene expression upon binding ribostamycin and neomycin. Interestingly, a similar aminoglycoside, paromomycin, differing from neomycin by only one substituent (amino versus hydroxyl), also binds to the N1 riboswitch, but without affecting gene expression, despite NMR evidence that the N1 riboswitch binds all aminoglycosides in a similar way. Here, to explore the details of structural dynamics of the aminoglycoside-N1 riboswitch complexes, we applied all-atom molecular dynamics (MD) and temperature replica-exchange MD simulations in explicit solvent. Indeed, we found that ribostamycin and neomycin affect riboswitch dynamics similarly but paromomycin allows for more flexibility because its complex lacks the contact between the distinctive 6' hydroxyl group and the G9 phosphate. Instead, a transient hydrogen bond of 6'-OH with A17 is formed, which partially diminishes interactions between the bulge and apical loop of the riboswitch, likely contributing to riboswitch inactivity. In many ways, the paromomycin complex mimics the conformations, interactions, and Na+ distribution of the free riboswitch. The MD-derived interaction network helps understand why riboswitch activity depends on aminoglycoside type, whereas for another aminoglycoside-binding site, aminoacyl-tRNA site in 16S rRNA, activity is not discriminatory.

Fixed-parameter tractable sampling for RNA design with multiple target structures

Background

The design of multi-stable RNA molecules has important applications in biology, medicine, and biotechnology. Synthetic design approaches profit strongly from effective in-silico methods, which substantially reduce the need for costly wet-lab experiments.

Results

We devise a novel approach to a central ingredient of most in-silico design methods: the generation of sequences that fold well into multiple target structures. Based on constraint networks, our approach supports generic Boltzmann-weighted sampling, which enables the positive design of RNA sequences with specific free energies (for each of multiple, possibly pseudoknotted, target structures) and GC-content. Moreover, we study general properties of our approach empirically and generate biologically relevant multi-target Boltzmann-weighted designs for an established design benchmark. Our results demonstrate the efficacy and feasibility of the method in practice as well as the benefits of Boltzmann sampling over the previously best multi-target sampling strategy—even for the case of negative design of multi-stable RNAs. Besides empirically studies, we finally justify the algorithmic details due to a fundamental theoretic result about multi-stable RNA design, namely the #P-hardness of the counting of designs.

Conclusion

introduces a novel, flexible, and effective approach to multi-target RNA design, which promises broad applicability and extensibility.

Our free software is available at: https://github.com/yannponty/RNARedPrint Supplementary data are available online.

Electronic supplementary material

The online version of this article (10.1186/s12859-019-2784-7) contains supplementary material, which is available to authorized users.

Tuning the Performance of Synthetic Riboswitches using Machine Learning

Riboswitch development for clinical, technological, and synthetic biology applications constantly seeks to optimize regulatory behavior. Here, we present a machine learning approach to improve the regulation of a tetracycline (tc)-dependent riboswitch device composed of two individual tc aptamers. We developed a bioinformatics model that combines random forest analysis with a convolutional neural network to predict the switching behavior of such tandem riboswitches. We found that both biophysical parameters and the hydrogen bond pattern influence regulation. Our new design pipeline led to significant improvement of the tc riboswitch device with a dynamic range extension from 8.5 to 40-fold. We are confident that our novel method not only results in an excellent tc-dependent riboswitch device but further holds great promise and potential for the optimization of other riboswitches.

Synthetic Biology of Small RNAs and Riboswitches

In bacteria and archaea, small RNAs (sRNAs) regulate complex networks through antisense interactions with target mRNAs in trans, and riboswitches regulate gene expression in cis based on the ability to bind small-molecule ligands. Although our understanding and characterization of these two important regulatory RNA classes is far from complete, these RNA-based mechanisms have proven useful for a wide variety of synthetic biology applications. Besides classic and contemporary applications in the realm of metabolic engineering and orthogonal gene control, this review also covers newer applications of regulatory RNAs as biosensors, logic gates, and tools to determine RNA-RNA interactions. A separate section focuses on critical insights gained and challenges posed by fundamental studies of sRNAs and riboswitches that should aid future development of synthetic regulatory RNAs.

RNA Switches for Synthetic Biology

In addition to coding for protein sequences, RNA molecules encode a diverse set of gene-regulatory elements. RNA switches are one class of gene-regulatory elements that control protein expression in a manner that is dependent on the concentration of specific ligand molecules. These allosteric gene-regulatory elements have been shown as useful tools in engineering diverse cell types to display novel function. In particular, RNA switches have been used as genetically encoded biosensors and conditional controllers to direct cellular decisions based on the system's changing environment. A significant focus in the field has been the generation of novel RNA switches that are tailored for different biological systems. We review approaches that have been used to generate RNA switches, which leverage the unique physical properties of RNA and the myriad ways in which RNA can modulate gene expression.

Solving the RNA design problem with reinforcement learning

We use reinforcement learning to train an agent for computational RNA design: given a target secondary structure, design a sequence that folds to that structure in silico. Our agent uses a novel graph convolutional architecture allowing a single model to be applied to arbitrary target structures of any length. After training it on randomly generated targets, we test it on the Eterna100 benchmark and find it outperforms all previous algorithms. Analysis of its solutions shows it has successfully learned some advanced strategies identified by players of the game Eterna, allowing it to solve some very difficult structures. On the other hand, it has failed to learn other strategies, possibly because they were not required for the targets in the training set. This suggests the possibility that future improvements to the training protocol may yield further gains in performance.

RNA-based dynamic genetic controllers: development strategies and applications

•

Unique properties of RNA lead to the development of RNA-based dynamic genetic controllers.

•

Natural riboswitches are re-engineered to detect new molecules.

•

RNA-based regulatory mechanisms are exploited to construct novel dynamic RNA controllers.

•

Computational methods and in vitro–in vivo selection enable de novo design of dynamic RNA controllers.

•

Dynamic RNA controllers are utilized for metabolic engineering and synthetic biology.

Dynamic regulation of gene expression in response to various molecules is crucial for both basic science and practical applications. RNA is considered an attractive material for creating dynamic genetic controllers because of its specific binding to ligands, structural flexibility, programmability, and small size. Here, we review recent advances in strategies for developing RNA-based dynamic controllers and applications. First, we describe studies that re-engineered natural riboswitches to generate new dynamic controllers. Next, we summarize RNA-based regulatory mechanisms that have been exploited to build novel artificial dynamic controllers. We also discuss computational methods and high-throughput selection approaches for de novo design of dynamic RNA controllers. Finally, we explain applications of dynamic RNA controllers for metabolic engineering and synthetic biology.

Computational design of small transcription activating RNAs for versatile and dynamic gene regulation

A longstanding goal of synthetic biology has been the programmable control of cellular functions. Central to this is the creation of versatile regulatory toolsets that allow for programmable control of gene expression. Of the many regulatory molecules available, RNA regulators offer the intriguing possibility of de novo design-allowing for the bottom-up molecular-level design of genetic control systems. Here we present a computational design approach for the creation of a bacterial regulator called Small Transcription Activating RNAs (STARs) and create a library of high-performing and orthogonal STARs that achieve up to ~ 9000-fold gene activation. We demonstrate the versatility of these STARs-from acting synergistically with existing constitutive and inducible regulators, to reprogramming cellular phenotypes and controlling multigene metabolic pathway expression. Finally, we combine these new STARs with themselves and CRISPRi transcriptional repressors to deliver new types of RNA-based genetic circuitry that allow for sophisticated and temporal control of gene expression.

Design of Artificial Riboswitches as Biosensors

RNA aptamers readily recognize small organic molecules, polypeptides, as well as other nucleic acids in a highly specific manner. Many such aptamers have evolved as parts of regulatory systems in nature. Experimental selection techniques such as SELEX have been very successful in finding artificial aptamers for a wide variety of natural and synthetic ligands. Changes in structure and/or stability of aptamers upon ligand binding can propagate through larger RNA constructs and cause specific structural changes at distal positions. In turn, these may affect transcription, translation, splicing, or binding events. The RNA secondary structure model realistically describes both thermodynamic and kinetic aspects of RNA structure formation and refolding at a single, consistent level of modelling. Thus, this framework allows studying the function of natural riboswitches in silico. Moreover, it enables rationally designing artificial switches, combining essentially arbitrary sensors with a broad choice of read-out systems. Eventually, this approach sets the stage for constructing versatile biosensors.

Exploring the Mutational Robustness of Nucleic Acids by Searching Genotype Neighborhoods in Sequence Space

To assess the mutational robustness of nucleic acids, many genome- and protein-level studies have been performed, where nucleic acids are treated as genetic information carriers and transferrers. However, the molecular mechanisms through which mutations alter the structural, dynamic, and functional properties of nucleic acids are poorly understood. Here we performed a SELEX in silico study to investigate the fitness distribution of the l-Arm-binding aptamer genotype neighborhoods. Two novel functional genotype neighborhoods were isolated and experimentally verified to have comparable fitness as the wild-type. The experimental aptamer fitness landscape suggests the mutational robustness is strongly influenced by the local base environment and ligand-binding mode, whereas bases distant from the binding pocket provide potential evolutionary pathways to approach the global fitness maximum. Our work provides an example of successful application of SELEX in silico to optimize an aptamer and demonstrates the strong sensitivity of mutational robustness to the site of genetic variation.