RNArchitecture: a database and a classification system of RNA families, with a focus on structural information

RNA3DB: A structurally-dissimilar dataset split for training and benchmarking deep learning models for RNA structure prediction

With advances in protein structure prediction thanks to deep learning models like AlphaFold, RNA structure prediction has recently received increased attention from deep learning researchers. RNAs introduce substantial challenges due to the sparser availability and lower structural diversity of the experimentally resolved RNA structures in comparison to protein structures. These challenges are often poorly addressed by the existing literature, many of which report inflated performance due to using training and testing sets with significant structural overlap. Further, the most recent Critical Assessment of Structure Prediction (CASP15) has shown that deep learning models for RNA structure are currently outperformed by traditional methods. In this paper we present RNA3DB, a dataset of structured RNAs, derived from the Protein Data Bank (PDB), that is designed for training and benchmarking deep learning models. The RNA3DB method arranges the RNA 3D chains into distinct groups (Components) that are non-redundant both with regard to sequence as well as structure, providing a robust way of dividing training, validation, and testing sets. Any split of these structurally-dissimilar Components are guaranteed to produce test and validations sets that are distinct by sequence and structure from those in the training set. We provide the RNA3DB dataset, a particular train/test split of the RNA3DB Components (in an approximate 70/30 ratio) that will be updated periodically. We also provide the RNA3DB methodology along with the source-code, with the goal of creating a reproducible and customizable tool for producing structurally-dissimilar dataset splits for structural RNAs.

A potential biological signature of 7-methylguanosine-related lncRNA to predict the immunotherapy effects in bladder cancer

Background

Bladder urothelial carcinoma (BLCA) is the second prevalent genitourinary carcinoma globally. N7-methylguanosine (m7G) is important for tumorigenesis and progression. This study aimed to build a predictive model for m7G-related long non-coding RNAs (lncRNAs), elucidate their role in the tumor immune microenvironment (TIME), and predict immunotherapy response in BLCA.

Methods

We first used univariate Cox regression and coexpression analyses to identify m7G-related lncRNAs. Next, the prognostic model was built by utilizing LASSO regression analysis. Then, the prognostic significance of the model was examined utilizing Kaplan-Meier survival analysis, receiver operating characteristic (ROC) curves, nomogram, and univariate, multivariate Cox regression. We also analyzed Gene set enrichment analyses (GSEA), immune analysis and principal component analysis (PCA) in risk groups. To further predict immunotherapy effectiveness, we evaluated the predictive ability for immunotherapy in 2 risk groups and clusters using tumor immune dysfunction and exclusion (TIDE) score and Immunophenoscore (IPS).

Results

Seven lncRNAs related to m7G were used to create a model. The calibration plots for the model suggested a strong fit with the prediction of overall survival (OS). The area under the curve (AUC) for first, second, and third years was respectively, 0.722, 0.711, and 0.686. In addition, the risk score had strong correlation with TIME features and genes linked to immune checkpoint blockade (ICB). TIDE scores were dramatically different between two risk groups (p < 0.05), and IPS scores were markedly different between two clusters (p < 0.05).

Conclusion

Our research constructed a novel m7G-related lncRNAs that could be used to predict patient outcomes and the effectiveness of immunotherapy in BLCA. Immunotherapy may be more effective for the low-risk group and cluster 2.

Expression pattern, tumor immune landscape, and prognostic value of N7‑methylguanosine regulators in bladder urothelial carcinoma

N7-Methylguanosine (m7G) modification is important in post-transcriptional regulation. dysregulation of m7G RNA modification has been reported to be markedly associated with cancer. However, its importance in bladder urothelial carcinoma (BLCA) remains poorly characterized. The present study systematically analyzed mRNA gene expression data and clinical information from The Cancer Genome Atlas and further constructed robust risk signatures for the four regulators of m7G RNA modification (nudix hydrolase 11, gem nuclear organelle-associated protein 5, eukaryotic translation initiation factor 3 subunit D and cytoplasmic FMR1 interacting protein 1). The differential expression and cell function of m7G-related genes in bladder cancer cells were verified by reverse transcription-quantitative PCR, Cell Counting Kit-8 and colony formation assays. The four-gene-based model could accurately predict the prognosis of BLCA. Nomogram-based clinical decisions had a higher net benefit compared with that of individual predictors. Through immune infiltration analysis, it was found that immune cell infiltration affected the prognosis of patients with BLCA. Finally, the present study identified potential therapeutics that differ between high and low-risk groups based on four genes. In summary, the current findings revealed an essential role for m7G RNA modification regulators in BLCA, and developed risk signatures as promising prognostic markers in patients with BLCA.

Proteostasis Deregulation in Neurodegeneration and Its Link with Stress Granules: Focus on the Scaffold and Ribosomal Protein RACK1

The role of protein misfolding, deposition, and clearance has been the dominant topic in the last decades of investigation in the field of neurodegeneration. The impairment of protein synthesis, along with RNA metabolism and RNA granules, however, are significantly emerging as novel potential targets for the comprehension of the molecular events leading to neuronal deficits. Indeed, defects in ribosome activity, ribosome stalling, and PQC—all ribosome-related processes required for proteostasis regulation—can contribute to triggering stress conditions and promoting the formation of stress granules (SGs) that could evolve in the formation of pathological granules, usually occurring during neurodegenerating effects. In this review, the interplay between proteostasis, mRNA metabolism, and SGs has been explored in a neurodegenerative context with a focus on Alzheimer’s disease (AD), although some defects in these same mechanisms can also be found in frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS), which are discussed here. Finally, we highlight the role of the receptor for activated C kinase 1 (RACK1) in these pathologies and note that, besides its well characterized function as a scaffold protein, it has an important role in translation and can associate to stress granules (SGs) determining cell fate in response to diverse stress stimuli.

Comprehensive analysis of 7-methylguanosine and immune microenvironment characteristics in clear cell renal cell carcinomas

Clear cell renal cell carcinoma (ccRCC) is one of the most common tumors in the urinary system. ccRCC has obvious immunological characteristics, and the infiltration of immune cells is related to the prognosis of ccRCC. The effect of immune checkpoint therapy is related to the dynamic changes of the tumor immune microenvironment (TIM). The 7-methylguanosine (m7G) is an additional mRNA modification ability besides m6A, which is closely related to the TIM and affects the occurrence and development of tumors. At present, the correlations between m7G and the immune microenvironment, treatment, and prognosis of ccRCC are not clear. As far as we know, there was no study on the relationship between m7G and the immune microenvironment and survival of clear cell renal cell carcinomas. A comprehensive analysis of the correlations between them and the construction of a prognosis model are helpful to improve the treatment strategy. Two different molecular subtypes were identified in 539 ccRCC samples by describing the differences of 29 m7G-related genes. It was found that the clinical features, TIM, and prognosis of ccRCC patients were correlated with the m7G-related genes. We found that there were significant differences in the expression of PD-1, CTLA4, and PD-L1 between high- and low-risk groups. To sum up, m7G-related genes play a potential role in the TIM, treatment, and prognosis of ccRCC. Our results provide new findings for ccRCC and help to improve the immunotherapy strategies and prognosis of patients.

Towards SINEUP-based therapeutics: Design of an in vitro synthesized SINEUP RNA

SINEUPs are a novel class of natural and synthetic non-coding antisense RNA molecules able to increase the translation of a target mRNA. They present a modular organization comprising an unstructured antisense target-specific domain, which sets the specificity of each individual SINEUP, and a structured effector domain, which is responsible for the translation enhancement. In order to design a fully functional in vitro transcribed SINEUP for therapeutics applications, SINEUP RNAs were synthesized in vitro with a variety of chemical modifications and screened for their activity on endogenous target mRNA upon transfection. Three combinations of modified ribonucleotides-2'O methyl-ATP (Am), N6 methyl-ATP (m6A), and pseudo-UTP (ψ)-conferred SINEUP activity to naked RNA. The best combination tested in this study was fully modified with m6A and ψ. Aside from functionality, this combination conferred improved stability upon transfection and higher thermal stability. Common structural determinants of activity were identified by circular dichroisms, defining a core functional structure that is achieved with different combinations of modifications.

SINEUPs: a novel toolbox for RNA therapeutics

RNA molecules have emerged as a new class of promising therapeutics to expand the range of druggable targets in the genome. In addition to ‘canonical’ protein-coding mRNAs, the emerging richness of sense and antisense long non-coding RNAs (lncRNAs) provides a new reservoir of molecular tools for RNA-based drugs. LncRNAs are composed of modular structural domains with specific activities involving the recruitment of protein cofactors or directly interacting with nucleic acids. A single therapeutic RNA transcript can then be assembled combining domains with defined secondary structures and functions, and antisense sequences specific for the RNA/DNA target of interest.

As the first representative molecules of this new pharmacology, we have identified SINEUPs, a new functional class of natural antisense lncRNAs that increase the translation of partially overlapping mRNAs. Their activity is based on the combination of two domains: an embedded mouse inverted SINEB2 element that enhances mRNA translation (effector domain) and an overlapping antisense region that provides specificity for the target sense transcript (binding domain). By genetic engineering, synthetic SINEUPs can potentially target any mRNA of interest increasing translation and therefore the endogenous level of the encoded protein.

In this review, we describe the state-of-the-art knowledge of SINEUPs and discuss recent publications showing their potential application in diseases where a physiological increase of endogenous protein expression can be therapeutic.

Systematics for types and effects of RNA variations

Systematics is described for annotation of variations in RNA molecules. The conceptual framework is part of Variation Ontology (VariO) and facilitates depiction of types of variations, their functional and structural effects and other consequences in any RNA molecule in any organism. There are more than 150 RNA related VariO terms in seven levels, which can be further combined to generate even more complicated and detailed annotations. The terms are described together with examples, usually for variations and effects in human and in diseases. RNA variation type has two subcategories: variation classification and origin with subterms. Altogether six terms are available for function description. Several terms are available for affected RNA properties. The ontology contains also terms for structural description for affected RNA type, post-transcriptional RNA modifications, secondary and tertiary structure effects and RNA sugar variations. Together with the DNA and protein concepts and annotations, RNA terms allow comprehensive description of variations of genetic and non-genetic origin at all possible levels. The VariO annotations are readable both for humans and computer programs for advanced data integration and mining.

Discovery of new group I-D introns leads to creation of subtypes and link to an adaptive response of the mitochondrial genome in fungi

Group I catalytic introns are widespread in bacterial, archaeal, viral, organellar, and some eukaryotic genomes, where they are reported to provide regulatory functions. The group I introns are currently divided into five types (A-E), which are themselves distributed into several subtypes, with the exception of group I type D intron (GI-D). GI-D introns belong to the rarest group with only 17 described to date, including only one with a putative role reported in fungi, where it would interfere with an adaptive response in the cytochrome b (COB) gene to quinone outside inhibitor (QoI) fungicide resistance. Using homology search methods taking into account both conserved sequences and RNA secondary structures, we analysed the mitochondrial genomes or COB genes of 169 fungal species, including some frequently under QoI selection pressure. These analyses have led to the identification of 216 novel GI-D introns, and the definition of three distinct subtypes, one of which being linked with a functional activity. We have further uncovered a homing site for this GI-D intron type, which helps refine the accepted model of quinone outside inhibitor resistance, whereby mobility of the intron across fungal mitochondrial genomes, would influence a fungus ability to develop resistance to QoIs.

RNA-Puzzles toolkit: a computational resource of RNA 3D structure benchmark datasets, structure manipulation, and evaluation tools

Significant improvements have been made in the efficiency and accuracy of RNA 3D structure prediction methods during the succeeding challenges of RNA-Puzzles, a community-wide effort on the assessment of blind prediction of RNA tertiary structures. The RNA-Puzzles contest has shown, among others, that the development and validation of computational methods for RNA fold prediction strongly depend on the benchmark datasets and the structure comparison algorithms. Yet, there has been no systematic benchmark set or decoy structures available for the 3D structure prediction of RNA, hindering the standardization of comparative tests in the modeling of RNA structure. Furthermore, there has not been a unified set of tools that allows deep and complete RNA structure analysis, and at the same time, that is easy to use. Here, we present RNA-Puzzles toolkit, a computational resource including (i) decoy sets generated by different RNA 3D structure prediction methods (raw, for-evaluation and standardized datasets), (ii) 3D structure normalization, analysis, manipulation, visualization tools (RNA_format, RNA_normalizer, rna-tools) and (iii) 3D structure comparison metric tools (RNAQUA, MCQ4Structures). This resource provides a full list of computational tools as well as a standard RNA 3D structure prediction assessment protocol for the community.

RNA 3D structure prediction guided by independent folding of homologous sequences

BACKGROUND

The understanding of the importance of RNA has dramatically changed over recent years. As in the case of proteins, the function of an RNA molecule is encoded in its tertiary structure, which in turn is determined by the molecule's sequence. The prediction of tertiary structures of complex RNAs is still a challenging task.

RESULTS

Using the observation that RNA sequences from the same RNA family fold into conserved structure, we test herein whether parallel modeling of RNA homologs can improve ab initio RNA structure prediction. EvoClustRNA is a multi-step modeling process, in which homologous sequences for the target sequence are selected using the Rfam database. Subsequently, independent folding simulations using Rosetta FARFAR and SimRNA are carried out. The model of the target sequence is selected based on the most common structural arrangement of the common helical fragments. As a test, on two blind RNA-Puzzles challenges, EvoClustRNA predictions ranked as the first of all submissions for the L-glutamine riboswitch and as the second for the ZMP riboswitch. Moreover, through a benchmark of known structures, we discovered several cases in which particular homologs were unusually amenable to structure recovery in folding simulations compared to the single original target sequence.

CONCLUSION

This work, for the first time to our knowledge, demonstrates the importance of the selection of the target sequence from an alignment of an RNA family for the success of RNA 3D structure prediction. These observations prompt investigations into a new direction of research for checking 3D structure "foldability" or "predictability" of related RNA sequences to obtain accurate predictions. To support new research in this area, we provide all relevant scripts in a documented and ready-to-use form. By exploring new ideas and identifying limitations of the current RNA 3D structure prediction methods, this work is bringing us closer to the near-native computational RNA 3D models.

Evolutionary Trends in RNA Base Selectivity Within the RNase A Superfamily

There is a growing interest in the pharmaceutical industry to design novel tailored drugs for RNA targeting. The vertebrate-specific RNase A superfamily is nowadays one of the best characterized family of enzymes and comprises proteins involved in host defense with specific cytotoxic and immune-modulatory properties. We observe within the family a structural variability at the substrate-binding site associated to a diversification of biological properties. In this work, we have analyzed the enzyme specificity at the secondary base binding site. Towards this end, we have performed a kinetic characterization of the canonical RNase types together with a molecular dynamic simulation of selected representative family members. The RNases' catalytic activity and binding interactions have been compared using UpA, UpG and UpI dinucleotides. Our results highlight an evolutionary trend from lower to higher order vertebrates towards an enhanced discrimination power of selectivity for adenine respect to guanine at the secondary base binding site (B2). Interestingly, the shift from guanine to adenine preference is achieved in all the studied family members by equivalent residues through distinct interaction modes. We can identify specific polar and charged side chains that selectively interact with donor or acceptor purine groups. Overall, we observe selective bidentate polar and electrostatic interactions: Asn to N1/N6 and N6/N7 adenine groups in mammals versus Glu/Asp and Arg to N1/N2, N1/O6 and O6/N7 guanine groups in non-mammals. In addition, kinetic and molecular dynamics comparative results on UpG versus UpI emphasize the main contribution of Glu/Asp interactions to N1/N2 group for guanine selectivity in lower order vertebrates. A close inspection at the B2 binding pocket also highlights the principal contribution of the protein ß6 and L4 loop regions. Significant differences in the orientation and extension of the L4 loop could explain how the same residues can participate in alternative binding modes. The analysis suggests that within the RNase A superfamily an evolution pressure has taken place at the B2 secondary binding site to provide novel substrate-recognition patterns. We are confident that a better knowledge of the enzymes' nucleotide recognition pattern would contribute to identify their physiological substrate and eventually design applied therapies to modulate their biological functions.

Self-Assembling RNA Nanoparticle for Gene Expression Regulation in a Model System

In the search for enzymatically processed RNA fragments, we found the novel three-way junction motif. The structure prediction suggested the arrangement of helices at acute angle approx. 60°. This allows the design of a trimeric RNA nanoparticle that can be functionalized with multiple regulatory fragments. Such RNA nano-object of equilateral triangular shape was applied for gene expression regulation studies in two independent cellular systems. Biochemical and functional studies confirmed the predicted shape and structure of the nanoparticle. The regulatory siRNA fragments incorporated into the nanoparticle were effectively released and triggered gene silencing. The regulatory effect was prolonged when induced with structuralized RNA compared to unstructured siRNAs. In these studies, the enzymatic processing of the motif was utilized for function release from the nanoparticle, enabling simultaneous delivery of different regulatory functions. This methodology of sequence search, RNA structural prediction, and application for rational design opens a new way for creating enzymatically processed RNA nanoparticles.

The State of Long Non-Coding RNA Biology

Transcriptomic studies have demonstrated that the vast majority of the genomes of mammals and other complex organisms is expressed in highly dynamic and cell-specific patterns to produce large numbers of intergenic, antisense and intronic long non-protein-coding RNAs (lncRNAs). Despite well characterized examples, their scaling with developmental complexity, and many demonstrations of their association with cellular processes, development and diseases, lncRNAs are still to be widely accepted as major players in gene regulation. This may reflect an underappreciation of the extent and precision of the epigenetic control of differentiation and development, where lncRNAs appear to have a central role, likely as organizational and guide molecules: most lncRNAs are nuclear-localized and chromatin-associated, with some involved in the formation of specialized subcellular domains. I suggest that a reassessment of the conceptual framework of genetic information and gene expression in the 4-dimensional ontogeny of spatially organized multicellular organisms is required. Together with this and further studies on their biology, the key challenges now are to determine the structure–function relationships of lncRNAs, which may be aided by emerging evidence of their modular structure, the role of RNA editing and modification in enabling epigenetic plasticity, and the role of RNA signaling in transgenerational inheritance of experience.

The 2018 Nucleic Acids Research database issue and the online molecular biology database collection

The 2018 Nucleic Acids Research Database Issue contains 181 papers spanning molecular biology. Among them, 82 are new and 84 are updates describing resources that appeared in the Issue previously. The remaining 15 cover databases most recently published elsewhere. Databases in the area of nucleic acids include 3DIV for visualisation of data on genome 3D structure and RNArchitecture, a hierarchical classification of RNA families. Protein databases include the established SMART, ELM and MEROPS while GPCRdb and the newcomer STCRDab cover families of biomedical interest. In the area of metabolism, HMDB and Reactome both report new features while PULDB appears in NAR for the first time. This issue also contains reports on genomics resources including Ensembl, the UCSC Genome Browser and ENCODE. Update papers from the IUPHAR/BPS Guide to Pharmacology and DrugBank are highlights of the drug and drug target section while a number of proteomics databases including proteomicsDB are also covered. The entire Database Issue is freely available online on the Nucleic Acids Research website (https://academic.oup.com/nar). The NAR online Molecular Biology Database Collection has been updated, reviewing 138 entries, adding 88 new resources and eliminating 47 discontinued URLs, bringing the current total to 1737 databases. It is available at http://www.oxfordjournals.org/nar/database/c/.