RNA Folding Research Articles

Gaussian Accelerated Molecular Dynamics in OpenMM.

Gaussian accelerated molecular dynamics (GaMD) is a computational technique that provides both unconstrained enhanced sampling and free energy calculations of biomolecules. Here, we present the implementation of GaMD in the OpenMM simulation package and validate it on model systems of alanine dipeptide and RNA folding. For alanine dipeptide, 30 ns GaMD production simulations reproduced free energy profiles of 1000 ns conventional molecular dynamics (cMD) simulations. In addition, GaMD simulations captured the folding pathways of three hyperstable RNA tetraloops (UUCG, GCAA, and CUUG) and binding of the rbt203 ligand to the HIV-1 Tar RNA, both of which involved critical electrostatic interactions such as hydrogen bonding and base stacking. Together with previous implementations, GaMD in OpenMM will allow for wider applications in simulations of proteins, RNA, and other biomolecules.

Caveats to Deep Learning Approaches to RNA Secondary Structure Prediction.

Machine learning (ML) and in particular deep learning techniques have gained popularity for predicting structures from biopolymer sequences. An interesting case is the prediction of RNA secondary structures, where well established biophysics based methods exist. The accuracy of these classical methods is limited due to lack of experimental parameters and certain simplifying assumptions and has seen little improvement over the last decade. This makes RNA folding an attractive target for machine learning and consequently several deep learning models have been proposed in recent years. However, for ML approaches to be competitive for de-novo structure prediction, the models must not just demonstrate good phenomenological fits, but be able to learn a (complex) biophysical model. In this contribution we discuss limitations of current approaches, in particular due to biases in the training data. Furthermore, we propose to study capabilities and limitations of ML models by first applying them on synthetic data (obtained from a simplified biophysical model) that can be generated in arbitrary amounts and where all biases can be controlled. We assume that a deep learning model that performs well on these synthetic, would also perform well on real data, and vice versa. We apply this idea by testing several ML models of varying complexity. Finally, we show that the best models are capable of capturing many, but not all, properties of RNA secondary structures. Most severely, the number of predicted base pairs scales quadratically with sequence length, even though a secondary structure can only accommodate a linear number of pairs.

RNAloops: a database of RNA multiloops.

MotivationKnowledge of the 3D structure of RNA supports discovering its functions and is crucial for designing drugs and modern therapeutic solutions. Thus, much attention is devoted to experimental determination and computational prediction targeting the global fold of RNA and its local substructures. The latter include multi-branched loops—functionally significant elements that highly affect the spatial shape of the entire molecule. Unfortunately, their computational modeling constitutes a weak point of structural bioinformatics. A remedy for this is in collecting these motifs and analyzing their features.ResultsRNAloops is a self-updating database that stores multi-branched loops identified in the PDB-deposited RNA structures. A description of each loop includes angular data—planar and Euler angles computed between pairs of adjacent helices to allow studying their mutual arrangement in space. The system enables search and analysis of multiloops, presents their structure details numerically and visually, and computes data statistics.Availability and implementationRNAloops is freely accessible at https://rnaloops.cs.put.poznan.pl.Supplementary information Supplementary data are available at Bioinformatics online.

Structural Basis for Fluorescence Activation by Pepper RNA.

Pepper is a fluorogenic RNA aptamer tag that binds to a variety of benzylidene-cyanophenyl (HBC) derivatives with tight affinity and activates their fluorescence. To investigate how Pepper RNA folds to create a binding site for HBC, we used antibody-assisted crystallography to determine the structures of Pepper bound to HBC530 and HBC599 to 2.3 and 2.7 Å resolutions, respectively. The structural data show that Pepper folds into an elongated structure and organizes nucleotides within an internal bulge to create the ligand binding site, assisted by an out-of-plane platform created by tertiary interactions with an adjacent bulge. As predicted from a lack of K+ dependence, Pepper does not use a G-quadruplex to form a binding pocket for HBC. Instead, Pepper uses a unique base-quadruple·base-triple stack to sandwich the ligand with a U·G wobble pair. Site-bound Mg2+ ions support ligand binding structurally and energetically. This research provides insight into the structural features that allow the Pepper aptamer to bind HBC and show how Pepper's function may expand to allow the in vivo detection of other small molecules and metals.

Orchestrating ribosomal RNA folding during ribosome assembly.

Construction of the eukaryotic ribosome is a complex process in which a nascent ribosomal RNA (rRNA) emerging from RNA Polymerase I hierarchically folds into a native three-dimensional structure. Modular assembly of individual RNA domains through interactions with ribosomal proteins and a myriad of assembly factors permit efficient disentanglement of the error-prone RNA folding process. Following these dynamic events, long-range tertiary interactions are orchestrated to compact rRNA. A combination of genetic, biochemical, and structural studies is now providing clues into how a nascent rRNA is transformed into a functional ribosome with high precision. With this essay, we aim to draw attention to the poorly understood process of establishing correct RNA tertiary contacts during ribosome formation.

3D Tiled Code Generation for Nussinov’s Algorithm

Current state-of-the-art parallel codes used to calculate the maximum number of pairs for a given RNA sequence by means of Nussinov’s algorithm do not allow for achieving speedup close up to the number of the processors used for execution of those codes on multi-core computers. This is due to the fact that known codes do not make full use of and derive benefit from cache memory of such computers. There is a need to develop new approaches allowing for increasing cache exploitation in multi-core computers. One of such possibilities is increasing the dimension of tiles in generated target tiled code and assuring a similar size of generated tiles. The article presents an approach allowing us to produce 3D parallel code with tiling calculating Nussinov’s RNA folding, i.e., code with the maximal tile dimension possible for the loop nest, executing Nussinov’s algorithm. The approach guarantees that generated tiles are of a similar size. The code generated with the presented approach is characterized by increased code locality and outperforms all closely related ones examined by us. This allows us to considerably reduce execution time required for computing the maximum number of pairs of any nested structure for larger RNA sequences by means of Nussinov’s algorithm.

RNAi-Based Therapeutics and Novel RNA Bioengineering Technologies.

RNA interference (RNAi) provides researchers with a versatile means to modulate target gene expression. The major forms of RNAi molecules, genome-derived microRNAs (miRNAs) and exogenous small interfering RNAs (siRNAs), converge into RNA-induced silencing complexes to achieve posttranscriptional gene regulation. RNAi has proven to be an adaptable and powerful therapeutic strategy where advancements in chemistry and pharmaceutics continue to bring RNAi-based drugs into the clinic. With four siRNA medications already approved by the US Food and Drug Administration (FDA), several RNAi-based therapeutics continue to advance to clinical trials with functions that closely resemble their endogenous counterparts. Although intended to enhance stability and improve efficacy, chemical modifications may increase risk of off-target effects by altering RNA structure, folding, and biologic activity away from their natural equivalents. Novel technologies in development today seek to use intact cells to yield true biologic RNAi agents that better represent the structures, stabilities, activities, and safety profiles of natural RNA molecules. In this review, we provide an examination of the mechanisms of action of endogenous miRNAs and exogenous siRNAs, the physiologic and pharmacokinetic barriers to therapeutic RNA delivery, and a summary of the chemical modifications and delivery platforms in use. We overview the pharmacology of the four FDA-approved siRNA medications (patisiran, givosiran, lumasiran, and inclisiran) as well as five siRNAs and several miRNA-based therapeutics currently in clinical trials. Furthermore, we discuss the direct expression and stable carrier-based, in vivo production of novel biologic RNAi agents for research and development. SIGNIFICANCE STATEMENT: In our review, we summarize the major concepts of RNA interference (RNAi), molecular mechanisms, and current state and challenges of RNAi drug development. We focus our discussion on the pharmacology of US Food and Drug Administration-approved RNAi medications and those siRNAs and miRNA-based therapeutics that entered the clinical investigations. Novel approaches to producing new true biological RNAi molecules for research and development are highlighted.

1-Deazaguanosine-Modified RNA: The Missing Piece for Functional RNA Atomic Mutagenesis.

Atomic mutagenesis is the key to advance our understanding of RNA recognition and RNA catalysis. To this end, deazanucleosides are utilized to evaluate the participation of specific atoms in these processes. One of the remaining challenges is access to RNA-containing 1-deazaguanosine (c1G). Here, we present the synthesis of this nucleoside and its phosphoramidite, allowing first time access to c1G-modified RNA. Thermodynamic analyses revealed the base pairing parameters for c1G-modified RNA. Furthermore, by NMR spectroscopy, a c1G-triggered switch of Watson-Crick into Hoogsteen pairing in HIV-2 TAR RNA was identified. Additionally, using X-ray structure analysis, a guanine–phosphate backbone interaction affecting RNA fold stability was characterized, and finally, the critical impact of an active-site guanine in twister ribozyme on the phosphodiester cleavage was revealed. Taken together, our study lays the synthetic basis for c1G-modified RNA and demonstrates the power of the completed deazanucleoside toolbox for RNA atomic mutagenesis needed to achieve in-depth understanding of RNA recognition and catalysis.

Web-based platform for analysis of RNA folding from high throughput chemical probing data.

RNA structures play critical roles in regulating gene expression across all domains of life and viruses. Chemical probing methods coupled with massively parallel sequencing have revolutionized the RNA structure field by enabling the assessment of many structures in their native, physiological context. Previously, we developed Dimethyl-Sulfate-based Mutational Profiling and Sequencing (DMS-MaPseq), which uses DMS to label the Watson-Crick face of open and accessible adenine and cytosine bases in the RNA. We used this approach to determine the genome-wide structures of HIV-1 and SARS-CoV-2 in infected cells, which permitted uncovering new biology and identifying therapeutic targets. Due to the simplicity and ease of the experimental procedure, DMS-MaPseq has been adopted by labs worldwide. However, bioinformatic analysis remains a substantial hurdle for labs that often lack the necessary infrastructure and computational expertise. Here we present a modern web-based interface that automates the analysis of chemical probing profiles from raw sequencing files (http://rnadreem.org). The availability of a simple web-based platform for DMS-MaPseq analysis will dramatically expand studies of RNA structure and aid in the design of structure-based therapeutics.

The Broad Institute Scores Another Victory in Its Battle with the University of California over the Patenting of CRIPSR

The Broad Institute Scores Another Victory in Its Battle with the University of California over the Patenting of CRIPSR

Machine learning predicts translation initiation sites in neurologic diseases with nucleotide repeat expansions.

A number of neurologic diseases associated with expanded nucleotide repeats, including an inherited form of amyotrophic lateral sclerosis, have an unconventional form of translation called repeat-associated non-AUG (RAN) translation. It has been speculated that the repeat regions in the RNA fold into secondary structures in a length-dependent manner, promoting RAN translation. Repeat protein products are translated, accumulate, and may contribute to disease pathogenesis. Nucleotides that flank the repeat region, especially ones closest to the initiation site, are believed to enhance translation initiation. A machine learning model has been published to help identify ATG and near-cognate translation initiation sites; however, this model has diminished predictive power due to its extensive feature selection and limited training data. Here, we overcome this limitation and increase prediction accuracy by the following: a) capture the effect of nucleotides most critical for translation initiation via feature reduction, b) implement an alternative machine learning algorithm better suited for limited data, c) build comprehensive and balanced training data (via sampling without replacement) that includes previously unavailable sequences, and d) split ATG and near-cognate translation initiation codon data to train two separate models. We also design a supplementary scoring system to provide an additional prognostic assessment of model predictions. The resultant models have high performance, with ~85-88% accuracy, exceeding that of the previously published model by >18%. The models presented here are used to identify translation initiation sites in genes associated with a number of neurologic repeat expansion disorders. The results confirm a number of sites of translation initiation upstream of the expanded repeats that have been found experimentally, and predict sites that are not yet established.

Artificial intelligence-assisted cryoEM structure of Bfr2-Lcp5 complex observed in the yeast small subunit processome

Eukaryotic ribosome is maturated through an elaborate process that includes modification, processing and folding of pre-ribosomal RNA (pre-rRNAs) by a series of ribosome assembly intermediates. More than 70 factors participate in the dynamic assembly and disassembly of the small subunit processome (90S) inside nucleolus, leading to the early maturation of small subunit. The 5’ domain of the 18S rRNA is the last to be incorporated into the stable 90S prior to the cleavage of pre-rRNA at the A1 site. This step is facilitated by the Kre33-Enp2-Bfr2-Lcp5 protein module with the participation of the DEAD-box protein Dbp4. Though structures of Kre33 and Enp2 have been modeled in previously observed 90S structures, that of Bfr2-Lcp5 complex remains unavailable. Here, we report an AlphaFold-assisted structure determination of the Bfr2-Lcp5 complex captured in a 3.99 Å − 7.24 Å cryoEM structure of 90S isolated from yeast cells depleted of Pih1, a chaperone protein of the 90S core assembly. The structure model is consistent with the protein-protein interaction results and the secondary structures of recombinant Bfr2 and Bfr2-Lcp5 complex obtained by Circular Dichroism. The Bfr2-Lcp5 complex interaction mimics that of exosome factors Rrp6-Rrp47 and acts to regulate 90S transitions.

Consistent Picture of Phosphate–Divalent Cation Binding from Models with Implicit and Explicit Electronic Polarization

The binding of divalent cations to the ubiquitous phosphate group is essential for a number of key biological processes, such as DNA compaction, RNA folding, or interactions of some proteins with membranes. Yet, probing their binding sites, modes, and associated binding free energy is a challenge for both experiments and simulations. In simulations, standard force fields strongly overestimate the interaction between phosphate groups and divalent cations. Here, we examine how different strategies to include electronic polarization effects in force fields─implicitly, through the use of scaled charges or pair-specific Lennard-Jones parameters, or explicitly, with the polarizable force fields Drude and AMOEBA─capture the interactions of a model phosphate compound, dimethyl phosphate, with calcium and magnesium divalent cations. We show that both implicit and explicit approaches, when carefully parameterized, are successful in capturing the overall binding free energy and that common trends emerge from the comparison of different simulation approaches. Overall, the binding is very moderate, slightly weaker for Ca2+ than Mg2+, and the solvent-shared ion pair is slightly more stable than the contact monodentate ion pair. The bidentate ion pair is higher in energy (or even fully unstable for Mg2+). Our results thus suggest practical ways to capture the divalent cations with biomolecular phosphate groups in complex biochemical systems. In particular, the computational efficiency of implicit models makes them ideally suited for large-scale simulations of biological assemblies, with improved accuracy compared to state-of-the-art fixed-charge force fields.

Fuzzy RNA recognition by the Trypanosoma brucei editosome.

The assembly of high molecular mass ribonucleoprotein complexes typically relies on the binary interaction of defined RNA sequences or precisely folded RNA motifs with dedicated RNA-binding domains on the protein side. Here we describe a new molecular recognition principle of RNA molecules by a high molecular mass protein complex. By chemically probing the solvent accessibility of mitochondrial pre-mRNAs when bound to the Trypanosoma brucei editosome, we identified multiple similar but non-identical RNA motifs as editosome contact sites. However, by treating the different motifs as mathematical graph objects we demonstrate that they fit a consensus 2D-graph consisting of 4 vertices (V) and 3 edges (E) with a Laplacian eigenvalue of 0.5477 (λ2). We establish that synthetic 4V(3E)-RNAs are sufficient to compete for the editosomal pre-mRNA binding site and that they inhibit RNA editing in vitro. Furthermore, we demonstrate that only two topological indices are necessary to predict the binding of any RNA motif to the editosome with a high level of confidence. Our analysis corroborates that the editosome has adapted to the structural multiplicity of the mitochondrial mRNA folding space by recognizing a fuzzy continuum of RNA folds that fit a consensus graph descriptor.

A Test and Refinement of Folding Free Energy Nearest Neighbor Parameters for RNA Including N6-Methyladenosine

RNA folding free energy change parameters are widely used to predict RNA secondary structure and to design RNA sequences. These parameters include terms for the folding free energies of helices and loops. Although the full set of parameters has only been traditionally available for the four common bases and backbone, it is well known that covalent modifications of nucleotides are widespread in natural RNAs. Covalent modifications are also widely used in engineered sequences. We recently derived a full set of nearest neighbor terms for RNA that includes N6-methyladenosine (m6A). In this work, we test the model using 98 optical melting experiments, matching duplexes with or without N6-methylation of A. Most experiments place RRACH, the consensus site of N6-methylation, in a variety of contexts, including helices, bulge loops, internal loops, dangling ends, and terminal mismatches. For matched sets of experiments that include either A or m6A in the same context, we find that the parameters for m6A are as accurate as those for A. Across all experiments, the root mean squared deviation between estimated and experimental free energy changes is 0.67 kcal/mol. We used the new experimental data to refine the set of nearest neighbor parameter terms for m6A. These parameters enable prediction of RNA secondary structures including m6A, which can be used to model how N6-methylation of A affects RNA structure.

Isomorphic Fluorescent Nucleosides Facilitate Real-Time Monitoring of RNA Depurination by Ribosome Inactivating Proteins.

Ribosome-inactivating proteins, a family of highly cytotoxic proteins, interfere with protein synthesis by depurinating a specific adenosine residue within the conserved α-sarcin/ricin loop of eukaryotic ribosomal RNA. Besides being biological warfare agents, certain RIPs have been promoted as potential therapeutic tools. Monitoring their deglycosylation activity and their inhibition in real time have remained, however, elusive. Herein, we describe the enzymatic preparation and utility of consensus RIP hairpin substrates in which specific G residues, next to the depurination site, are surgically replaced with tz G and th G, fluorescent G analogs. By strategically modifying key positions with responsive fluorescent surrogate nucleotides, RIP-mediated depurination can be monitored in real time by steady-state fluorescence spectroscopy. Subtle differences observed in preferential depurination sites provide insight into the RNA folding as well as RIPs' substrate recognition features.

Hyperpolarized water as universal sensitivity booster in biomolecular NMR.

NMR spectroscopy is the only method to access the structural dynamics of biomolecules at high (atomistic) resolution in their native solution state. However, this method's low sensitivity has two important consequences: (i) typically experiments have to be performed at high concentrations that increase sensitivity but are not physiological, and (ii) signals have to be accumulated over long periods, complicating the determination of interaction kinetics on the order of seconds and impeding studies of unstable systems. Both limitations are of equal, fundamental relevance: non-native conditions are of limited pharmacological relevance, and the function of proteins, enzymes and nucleic acids often relies on their interaction kinetics. To overcome these limitations, we have developed applications that involve 'hyperpolarized water' to boost signal intensities in NMR of proteins and nucleic acids. The technique includes four stages: (i) preparation of the biomolecule in partially deuterated buffers, (ii) preparation of 'hyperpolarized' water featuring enhanced 1H NMR signals via cryogenic dynamic nuclear polarization, (iii) sudden melting of the cryogenic pellet and dissolution of the protein or nucleic acid in the hyperpolarized water (enabling spontaneous exchanges of protons between water and target) and (iv) recording signal-amplified NMR spectra targeting either labile 1H or neighboring 15N/13C nuclei in the biomolecule. Water in the ensuing experiments is used as a universal 'hyperpolarization' agent, rendering the approach versatile and applicable to any biomolecule possessing labile hydrogens. Thus, questions can be addressed, ranging from protein and RNA folding problems to resolving structure-function relationships of intrinsically disordered proteins to investigating membrane interactions.

Optimization of RNA Display Using GC‐Clamp Modifications to Improve Genetic Detection of RNA‐Protein Interactions

Non‐coding small regulatory RNAs (sRNAs) have an important role in bacterial stress responses. The binding of sRNAs to their target mRNAs is facilitated by RNA‐binding proteins such as Hfq or ProQ. The Berry Lab has developed a bacterial three‐hybrid (B3H) assay to detect the binding of RNA with both of these RNA chaperones in vivo, by connecting the strength of an RNA‐protein interaction to the expression of a reporter gene. The interaction between a “prey” protein fused to the α‐subunit of RNA polymerase (RNAP) and a “bait” RNA tethered upstream of a test promoter stabilizes the binding of RNAP and increases transcription of the reporter gene lacZ. Despite the promise of the B3H system and its success in detecting many high‐affinity interactions, low signal‐to‐noise for other RNA‐protein interactions currently limits the broader utility of the assay. Computational structure predictions suggested that certain RNAs of interest could misfold when hybridized with other components of the hybrid RNA construct, e.g. an MS2 hairpin (MS2hp) or an exogenous intrinsic terminator. Such misfolding would likely disrupt a bait RNA’s interaction with the prey protein. To avoid this limitation, this study aims to optimize the hybrid RNA construct to improve the breadth of detectable interactions in the B3H assay. To this end, we designed new hybrid RNA constructs with the addition of a GC‐clamp – a short insert of guanines (G) and cytosines (C) flanking a region of interest – to promote proper folding and optimal display of RNA. Several sRNAs and 5’UTRs were cloned into GC‐clamp designs and their in vivo interactions with Hfq were tested in the B3H assay. Preliminary results demonstrate the promise of a short GC‐clamp in improving the B3H signal for many sRNA‐Hfq interactions, and we are currently working to test these GC constructs with additional RNAs. Increasing the sensitivity and generalizability of the B3H assay to study bacterial RNA‐protein interactions will help shed light on the molecular mechanisms of RNA‐chaperone proteins and the important processes in bacteria they regulate, such as adaptation to stress, biofilm formation and virulence.

Sub-3-Å cryo-EM structure of RNA enabled by engineered homomeric self-assembly.

High-resolution structural studies are essential for understanding the folding and function of diverse RNAs. Herein, we present a nanoarchitectural engineering strategy for efficient structural determination of RNA-only structures using single-particle cryogenic electron microscopy (cryo-EM). This strategy-ROCK (RNA oligomerization-enabled cryo-EM via installing kissing loops)-involves installing kissing-loop sequences onto the functionally nonessential stems of RNAs for homomeric self-assembly into closed rings with multiplied molecular weights and mitigated structural flexibility. ROCK enables cryo-EM reconstruction of the Tetrahymena group I intron at 2.98-Å resolution overall (2.85 Å for the core), allowing de novo model building of the complete RNA, including the previously unknown peripheral domains. ROCK is further applied to two smaller RNAs-the Azoarcus group I intron and the FMN riboswitch, revealing the conformational change of the former and the bound ligand in the latter. ROCK holds promise to greatly facilitate the use of cryo-EM in RNA structural studies.

3’ Terminal end‐labeling of structured RNA by Klenow DNA polymerase

RNA labeling is a key step in studying RNA localization, folding, and structural dynamics using fluorescence based assays. Current methods of fluorophore attachment to long (>100 nucleotides) and structured RNAs are hampered by low labeling yields and limitations in attachment position. These challenges have led to an interest in developing alternative methods that would enable the site‐specific labeling of long and structured RNAs. To this end, we have developed a site‐specific, chemo‐enzymatic method of fluorophore incorporation to RNAs. This novel method uses Klenow enzyme, a DNA polymerase, to add modified nucleotides to RNAs 3’ end in a templated manner. To facilitate this modified nucleotide addition in structured RNAs, we designed helper DNA oligonucleotides to disrupt any pre‐existing RNA intramolecular structure. Using this method, we have achieved efficient 3’ terminal end‐labeling of chemically synthesized and in‐vitro transcribed RNAs of different lengths with cyanine dyes.This method will be routinely used in fluorescence‐based assays that investigate the structure‐function relationships of RNAs involved in cancer and HIV progression.