Conformational dynamics of RNA G4C2 and G2C4 repeat expansions causing ALS/FTD using NMR and molecular dynamics studies.

G4C2 and G2C4 repeat expansions in chromosome 9 open reading frame 72 (C9orf72) are the most common cause of genetically defined amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), or c9ALS/FTD. The gene is bidirectionally transcribed, producing G4C2 repeats [r(G4C2)exp] and G2C4 repeats [r(G2C4)exp]. The c9ALS/FTD repeat expansions are highly structured, and structural studies showed that r(G4C2)exp predominantly folds into a hairpin with a periodic array of 1 × 1 G/G internal loops and a G-quadruplex. A small molecule probe revealed that r(G4C2)exp also adopts a hairpin structure with 2 × 2 GG/GG internal loops. We studied the conformational dynamics adopted by 2 × 2 GG/GG loops using temperature replica exchange molecular dynamics (T-REMD) and further characterized the structure and underlying dynamics using traditional 2D NMR techniques. These studies showed that the loop's closing base pairs influence both structure and dynamics, particularly the configuration adopted around the glycosidic bond. Interestingly, r(G2C4) repeats, which fold into an array of 2 × 2 CC/CC internal loops, are not as dynamic. Collectively, these studies emphasize the unique sensitivity of r(G4C2)exp to small changes in stacking interactions, which is not observed in r(G2C4)exp, providing important considerations for further principles in structure-based drug design.

Study of an RNA-Focused DNA-Encoded Library Informs Design of a Degrader of a r(CUG) Repeat Expansion.

A solid-phase DNA-encoded library (DEL) was studied for binding the RNA repeat expansion r(CUG)exp, the causative agent of the most common form of adult-onset muscular dystrophy, myotonic dystrophy type 1 (DM1). A variety of uncharged and novel RNA binders were identified to selectively bind r(CUG)exp by using a two-color flow cytometry screen. The cellular activity of one binder was augmented by attaching it with a module that directly cleaves r(CUG)exp. In DM1 patient-derived muscle cells, the compound specifically bound r(CUG)exp and allele-specifically eliminated r(CUG)exp, improving disease-associated defects. The approaches herein can be used to identify and optimize ligands and bind RNA that can be further augmented for functionality including degradation.

Low-Molecular Weight Small Molecules Can Potently Bind RNA and Affect Oncogenic Pathways in Cells.

RNA is challenging to target with bioactive small molecules, particularly those of low molecular weight that bind with sufficient affinity and specificity. In this report, we developed a platform to address this challenge, affording a novel bioactive interaction. An RNA-focused small-molecule fragment collection (n = 2500) was constructed by analyzing features in all publicly reported compounds that bind RNA, the largest collection of RNA-focused fragments to date. The RNA-binding landscape for each fragment was studied by using a library-versus-library selection with an RNA library displaying a discrete structural element, probing over 12.8 million interactions, the greatest number of interactions between fragments and biomolecules probed experimentally. Mining of this dataset across the human transcriptome defined a drug-like fragment that potently and specifically targeted the microRNA-372 hairpin precursor, inhibiting its processing into the mature, functional microRNA and alleviating invasive and proliferative oncogenic phenotypes in gastric cancer cells. Importantly, this fragment has favorable properties, including an affinity for the RNA target of 300 ± 130 nM, a molecular weight of 273 Da, and quantitative estimate of drug-likeness (QED) score of 0.8. (For comparison, the mean QED of oral medicines is 0.6 ± 0.2). Thus, these studies demonstrate that a low-molecular weight, fragment-like compound can specifically and potently modulate RNA targets.

Quantum machine learning corrects classical forcefields: Stretching DNA base pairs in explicit solvent.

In order to improve the accuracy of molecular dynamics simulations, classical forcefields are supplemented with a kernel-based machine learning method trained on quantum-mechanical fragment energies. As an example application, a potential-energy surface is generalized for a small DNA duplex, taking into account explicit solvation and long-range electron exchange-correlation effects. A long-standing problem in molecular science is that experimental studies of the structural and thermodynamic behavior of DNA under tension are not well confirmed by simulation; study of the potential energy vs extension taking into account a novel correction shows that leading classical DNA models have excessive stiffness with respect to stretching. This discrepancy is found to be common across multiple forcefields. The quantum correction is in qualitative agreement with the experimental thermodynamics for larger DNA double helices, providing a candidate explanation for the general and long-standing discrepancy between single molecule stretching experiments and classical calculations of DNA stretching. The new dataset of quantum calculations should facilitate multiple types of nucleic acid simulation, and the associated Kernel Modified Molecular Dynamics method (KMMD) is applicable to biomolecular simulations in general. KMMD is made available as part of the AMBER22 simulation software.

DNA partitions into triplets under tension in the presence of organic cations, with sequence evolutionary age predicting the stability of the triplet phase.

Using atomistic simulations, we show the formation of stable triplet structure when particular GC-rich DNA duplexes are extended in solution over a timescale of hundreds of nanoseconds, in the presence of organic salt. We present planar-stacked triplet disproportionated DNA (Σ DNA) as a possible solution phase of the double helix under tension, subject to sequence and the presence of stabilising co-factors. Considering the partitioning of the duplexes into triplets of base pairs as the first step of operation of recombinase enzymes like RecA, we emphasise the structure-function relationship in Σ DNA. We supplement atomistic calculations with thermodynamic arguments to show that codons for 'phase 1' amino acids (those appearing early in evolution) are more likely than a lower entropy GC-rich sequence to form triplets under tension. We further observe that the four amino acids supposed (in the 'GADV world' hypothesis) to constitute the minimal set to produce functional globular proteins have the strongest triplet-forming propensity within the phase 1 set, showing a series of decreasing triplet propensity with evolutionary newness. The weak form of our observation provides a physical mechanism to minimise read frame and recombination alignment errors in the early evolution of the genetic code.

NMR structures of small molecules bound to a model of an RNA CUG repeat expansion.

Trinucleotide repeat expansions fold into long, stable hairpins and cause a variety of incurable RNA gain-of-function diseases such as Huntington's disease, the myotonic dystrophies, and spinocerebellar ataxias. One approach for treating these diseases is to bind small molecules to the structured RNAs. Both Huntington's disease-like 2 (HDL2) and myotonic dystrophy type 1 (DM1) are caused by a r(CUG) repeat expansion, or r(CUG) exp . The RNA folds into a hairpin structure with a periodic array of 1×1 nucleotide UU loops (5'C U G/3'G U C; where the underlined nucleotides indicate the Us in the internal loop) that sequester various RNA-binding proteins (RBP) and hence the source of its gain-of-function. Here, we report NMR-refined structures of single 5'C U G/3'G U C motifs in complex with three different small molecules, a diguandinobenzoate ( 1 ), a derivative of 1 where the guanidino groups have been exchanged for imidazole ( 2 ), and a quinoline with improved drug-like properties ( 3 ). These structures were determined using nuclear magnetic resonance (NMR) spectroscopy and simulated annealing with restrained molecular dynamics (MD). Compounds 1 , 2 , and 3 formed stacking and hydrogen bonding interactions with the 5'C U G/3'G U C motif. Compound 3 also formed van der Waals interactions with the internal loop. The global structure of each RNA-small molecule complexes retains an A-form conformation, while the internal loops are still dynamic but to a lesser extent compared to the unbound form. These results aid our understanding of ligand-RNA interactions and enable structure-based design of small molecules with improved binding affinity for and biological activity against r(CUG) exp . As the first ever reported structures of RNA r(CUG) repeats bound to ligands, these structures can enable virtual screening campaigns combined with machine learning assisted de novo design.

Fragment-Based Approaches to Identify RNA Binders.

Although fragment-based drug discovery (FBDD) has been successfully implemented and well-explored for protein targets, its feasibility for RNA targets is emerging. Despite the challenges associated with the selective targeting of RNA, efforts to integrate known methods of RNA binder discovery with fragment-based approaches have been fruitful, as a few bioactive ligands have been identified. Here, we review various fragment-based approaches implemented for RNA targets and provide insights into experimental design and outcomes to guide future work in the area. Indeed, investigations surrounding the molecular recognition of RNA by fragments address rather important questions such as the limits of molecular weight that confer selective binding and the physicochemical properties favorable for RNA binding and bioactivity.

Optimization of a Protein-Targeted Medicine into an RNA-Specific Small Molecule.

Protein-targeted small molecule medicines often bind RNAs and affect RNA-mediated pathways in cells. Historically, small molecule engagement and modulation of RNA have not been considered in medicine development; however, RNA should be considered both a potential on- and off-target. Kinase inhibitors have emecrged as common RNA binders with dovitinib, a classic receptor tyrosine kinase (RTK) inhibitor, inhibiting RTKs and the biogenesis of oncogenic microRNA-21 through direct engagement. In this study, we use knowledge of the molecular recognition of both protein and RNA targets by dovitinib to design molecules that specifically inhibit the RNA target but lack activity against canonical protein targets in cells. As it is now becoming apparent that RNA can be both an on- and off-target for small molecule medicines, this study lays a foundation to use design principles to maximize desired compound activity while minimizing off-target effects.

Methods for the study of ribonuclease targeting chimeras (RiboTACs).

Recently, a class of heterobifunctional small molecules called ribonuclease targeting chimeras (RiboTACs) have been developed that selectively induce degradation of RNAs in cells. These molecules function by recruiting latent ribonuclease (RNase L), an endoribonuclease involved in the innate immune response, to targeted RNA structures. The RiboTACs must activate RNase L in proximity to the RNA, resulting in cleavage of the RNA and downstream degradation. To develop and validate a new RiboTAC, several steps must be taken. First, small molecule activators that bind to RNase L must be identified. Next, since RNase L is only catalytically active upon ligand-induced homodimerization, the capability of identified small molecules to activate RNase L must be assessed. RNase L-activating small molecules should then be coupled to validated RNA-binding small molecules to construct the active RiboTAC. This RiboTAC can finally be assessed in cells for RNase L-dependent degradation of target RNAs. This chapter will provide several methods that are helpful to develop and assess RiboTACs throughout this process, including recombinant RNase L expression, methods to assess RNase L engagement in vitro such as saturation transfer difference nuclear magnetic resonance (STD NMR), an in vitro assay to assess activation of RNase L, and cellular methods to demonstrate RNase L-dependent cleavage.

Transcriptome-Wide Studies of RNA-Targeted Small Molecules Provide a Simple and Selective r(CUG)exp Degrader in Myotonic Dystrophy.

Myotonic dystrophy type 1 (DM1) is caused by a highly structured RNA repeat expansion, r(CUG)exp, harbored in the 3' untranslated region (3' UTR) of dystrophia myotonica protein kinase (DMPK) mRNA and drives disease through a gain-of-function mechanism. A panel of low-molecular-weight fragments capable of reacting with RNA upon UV irradiation was studied for cross-linking to r(CUG)expin vitro, affording perimidin-2-amine diazirine (1) that bound to r(CUG)exp. The interactions between the small molecule and RNA were further studied by nuclear magnetic resonance (NMR) spectroscopy and molecular modeling. Binding of 1 in DM1 myotubes was profiled transcriptome-wide, identifying 12 transcripts including DMPK that were bound by 1. Augmenting the functionality of 1 with cleaving capability created a chimeric degrader that specifically targets r(CUG)exp for elimination. The degrader broadly improved DM1-associated defects as assessed by RNA-seq, while having limited effects on healthy myotubes. This study (i) provides a platform to investigate molecular recognition of ligands directly in disease-affected cells; (ii) illustrates that RNA degraders can be more specific than the binders from which they are derived; and (iii) suggests that repeating transcripts can be selectively degraded due to the presence of multiple ligand binding sites.

Computational Investigation of Bending Properties of RNA AUUCU, CCUG, CAG, and CUG Repeat Expansions Associated With Neuromuscular Disorders.

Expansions of RNA AUUCU, CCUG, CAG, and CUG repeats cause spinocerebellar ataxia type 10, myotonic dystrophy type 2, Huntington's disease, and myotonic dystrophy type 1, respectively. By performing extensive molecular dynamic simulations, we investigated the bending propensities and conformational landscapes adopted by 3×3, 2×2, and 1×1 internal loops observed in RNA AUUCU, CCUG, CAG, and CUG repeat expansions using model systems having biologically relevant repeat sizes. We show that the conformational variability experienced by these loops is more complex than previous reports where a variety of unconventional hydrogen bonds are formed. At the global scale, strong bending propensity was observed in r(AUUCU)10, r(CCUG)15, r(CAG)20, and r(CUG)20, and, to a lesser extent, in r(AUUCU)4, r(CCUG)10, r(CAG)10, and r(CUG)10. Furthermore, RNA CAG repeats exhibit a tendency toward bent states with more than 50% of observed conformations having bending angles greater than 50°, while RNA CUG repeats display relatively linear-like conformations with extremely bent conformations accounting for less than 25% of the observed structures. Conformations experienced by RNA AUUCU repeats are a combination of strongly bent and kinked structures. The bent states in RNA CCUG repeats mostly fall into the moderately bent category with a marginal ensemble experiencing extreme bending. The general pattern observed in all the bent structures indicates the collapse of the major groove width as the mechanical trigger for bending, which is caused by alteration of base pair step parameters at multiple locations along the RNA due to local distortions at the loop sites. Overextension is also observed in all the RNA repeats that is attributed to widening of the major groove width as well as undertwisting phenomenon. This information and the rich structural repository could be applied for structure based small molecule design targeting disease-causing RNAs. The bending propensities of these constructs, at the global level, could also have implications on how expanded RNA repeats interact with proteins.

Evaluating Geometric Definitions of Stacking for RNA Dinucleoside Monophosphates Using Molecular Mechanics Calculations.

RNA modulation via small molecules is a novel approach in pharmacotherapies, where the determination of the structural properties of RNA motifs is considered a promising way to develop drugs capable of targeting RNA structures to control diseases. However, due to the complexity and dynamic nature of RNA molecules, the determination of RNA structures using experimental approaches is not always feasible, and computational models employing force fields can provide important insight. The quality of the force field will determine how well the predictions are compared to experimental observables. Stacking in nucleic acids is one such structural property, originating mainly from London dispersion forces, which are quantum mechanical and are included in molecular mechanics force fields through nonbonded interactions. Geometric descriptions are utilized to decide if two residues are stacked and hence to calculate the stacking free energies for RNA dinucleoside monophosphates (DNMPs) through statistical mechanics for comparison with experimental thermodynamics data. Here, we benchmark four different stacking definitions using molecular dynamics (MD) trajectories for 16 RNA DNMPs produced by two different force fields (RNA-IL and ff99OL3) and show that our stacking definition better correlates with the experimental thermodynamics data. While predictions within an accuracy of 0.2 kcal/mol at 300 K were observed in RNA CC, CU, UC, AG, GA, and GG, stacked states of purine-pyrimidine and pyrimidine-purine DNMPs, respectively, were typically underpredicted and overpredicted. Additionally, population distributions of RNA UU DNMPs were poorly predicted by both force fields, implying a requirement for further force field revisions. We further discuss the differences predicted by each RNA force field. Finally, we show that discrete path sampling (DPS) calculations can provide valuable information and complement the MD simulations. We propose the use of experimental thermodynamics data for RNA DNMPs as benchmarks for testing RNA force fields.

A Small Molecule Exploits Hidden Structural Features within the RNA Repeat Expansion That Causes c9ALS/FTD and Rescues Pathological Hallmarks.

The hexanucleotide repeat expansion GGGGCC [r(G4C2)exp] within intron 1 of C9orf72 causes genetically defined amyotrophic lateral sclerosis and frontotemporal dementia, collectively named c9ALS/FTD. , the repeat expansion causes neurodegeneration via deleterious phenotypes stemming from r(G4C2)exp RNA gain- and loss-of-function mechanisms. The r(G4C2)exp RNA folds into both a hairpin structure with repeating 1 × 1 nucleotide GG internal loops and a G-quadruplex structure. Here, we report the identification of a small molecule (CB253) that selectively binds the hairpin form of r(G4C2)exp. Interestingly, the small molecule binds to a previously unobserved conformation in which the RNA forms 2 × 2 nucleotide GG internal loops, as revealed by a series of binding and structural studies. NMR and molecular dynamics simulations suggest that the r(G4C2)exp hairpin interconverts between 1 × 1 and 2 × 2 internal loops through the process of strand slippage. We provide experimental evidence that CB253 binding indeed shifts the equilibrium toward the 2 × 2 GG internal loop conformation, inhibiting mechanisms that drive c9ALS/FTD pathobiology, such as repeat-associated non-ATG translation formation of stress granules and defective nucleocytoplasmic transport in various cellular models of c9ALS/FTD.