How does precursor RNA structure influence RNA processing and gene expression?

Abstract

RNA is a fundamental biomolecule that has many purposes within cells. Due to its single-stranded and flexible nature, RNA naturally folds into complex and dynamic structures. Recent technological and computational advances have produced an explosion of RNA structural data. Many RNA structures have regulatory and functional properties. Studying the structure of nascent RNAs is particularly challenging due to their low abundance and long length, but their structures are important because they can influence RNA processing. Precursor RNA processing is a nexus of pathways that determines mature isoform composition and that controls gene expression. In this review, we examine what is known about human nascent RNA structure and the influence of RNA structure on processing of precursor RNAs. These known structures provide examples of how other nascent RNAs may be structured and show how novel RNA structures may influence RNA processing including splicing and polyadenylation. RNA structures can be targeted therapeutically to treat disease.

Keywords: RNA structure; alternative splicing; polyadenylation; precursor RNA; splicing; transcription.

Figure 1. RNA structures influence precursor RNA processing

(A) Hairpin elements can block 5′ splice site recognition by interfering with U1 snRNP binding. MAPT1 RNA exon 10 alternative splicing is controlled by hairpin structure at the 5′ splice site. (B) RNA structure can bring distal elements in close proximity. The global fold of the HBB RNA is mediated by SRSF1 binding and orients the 5′ and 3′ splice sites for U1 snRNP interaction and efficient splicing. (C) Recognition of RNA elements are control processing. MBNL1 protein binds to its own RNA. MBNL1 binding causes remodeling of the RNA structure around the branchpoint and represses exon 5 inclusion. (D) Transcriptome-wide structural analysis of nascent RNA found clear structural ‘steps’ in proximity to efficiently spliced exons (top). The structure ‘steps’ around frequently skipped exon are less evident (bottom). Cartoon depiction based on [12].

Figure 2. Identifying potential functional structures in MCL1 RNA using nucleotide conservation and protein binding sites

(A) Schematic of human SMN2 exon 7 surrounded by 150 intronic nucleotides on both sides. Known RNA structures in SMN2 are annotated (red). (B and C) Nucleotide conservation in SMN2. Higher values indicate more conservation. A region of high conservation extends into the 5′ splice site of exon 7 and overlaps with TSL2 and ISTL1 (boxed). (D) Protein binding sites across SMN2 from ENCODE (gray) and published studies (green) are mapped on to the schematic. Binding sites for the RBP TIA1 overlap with TSL3 and ISTL2 structures (boxed). (E) Schematic of human MCL1 exon 2 surrounded by 150 intronic nucleotides on both sides. (F,G) Nucleotide conservation in SMN2. Higher values indicate more conservation. A region of high conservation extends into the 3′ splice site of exon 2 and overlaps with the branchpoint region (boxed). (H) Protein binding sites across MCL1 from ENCODE (gray) and published studies (green). Binding sites for regulatory RBPs, SRSF1 and hnRNPF/H are indicated (boxes). For both RNAs, schematics were visualized with Geneious Prime v2022.1.1 [203]. Branchpoints were annotated based on experimental data [80]. Conservation data were retrieved from UCSC table browser (Cons 100 Verts, phastCon, phyloP100way) [79]. ENCODE eCLIP data were retrieved as bigBed narrowPeak annotations filtered by a P-value of < 0.05, and annotations collapsed for overlapping peaks of the same protein [85]. All data reference hg38.

Figure 3. Identifying potential functional structures in MCL1 RNA using genomic variation and RNA structure models

(A) Schematic of human SMN2 exon 7 surrounded by 150 intronic nucleotides on both sides. Known RNA structures in SMN2 are annotated (red). (B) Genomic variants in SMN2 from a variety of sources including variants that are disease-associated (blue), somatic (orange), predicted to change RNA structure (riboSNitches, purple) and inherited (green). The ClinVar disease-associated variant in SMN2 is a riboSNitch (arrow). Likewise, a somatic mutation in Hairpin Element 2 is a riboSNitch (arrow). (C) Arc diagram of the SMN2 RNA structure generated with published SHAPE-MaP data [204] showing highly probable base pairing (>80%, green) and moderately probable base pairing (30–80%, blue). Predicted structures overlap with published structural elements (boxes). (D) Schematic of human MCL1 exon 2 surrounded by 150 intronic nucleotides on both sides. (E) Genomic variation in the MCL1 exon 2 region, colored as indicated above. RiboSNitches in proximity to the ClinVar disease-associated variant overlap with somatic mutations (arrow). We also highlight a riboSNitch that is a somatic mutation (arrow). (F) MCL1 precursor RNA structure base-pairing probabilities generated from in vitro 5NIA SHAPE-MaP data and analyzed with shapemapper2 [205]. For both RNAs variants were retrieved from gnomAD [99], COSMIC [98], RNAsnp screening [102], and ClinVar [93]. RNA structures were generated with the RNAStructure package functions (partition and ProbabilityPlot [206]) and visualized in IGV [207].

Figure 4. RNA interactions with RBPs

(A) The RNA-binding domain of U1A snRNP protein (gray) binds the U1 snRNA stem loop II (blue). Base stacking of A11 and C12 between amino acids Phe56 and Asp92 (maroon). Amino acids Ser46, Ser48, Leu49, and Arg52 (light orange) lock the protein into the hole defined by the RNA structure and interact with bases C11-G16. (PDB 1URN [115]). (B) U1A amino acids Lys20 and Lys22 contribute electrostatic interactions that stabilize the phosphodiester backbone of the RNA. Lys23 interacts with the U1A protein loop located in the open RNA (pink) (PDB 1URN [115]). (C) U1A RMM dimer binds PIE RNA structure. Amino acids are highlighted in the same colors as outlined above (PDB 1DZ5, [121]). (D) DROSHA dsRBD (PDB 6V5B [128]) in complex with pri-miR-16-2. Amino acids Ser1293, His1294, and Arg1296 interact with ribose in the minor groove and Tyr1298 interacts with the minor groove phosphate backbone. Gln1318 electrostatically interacts with the phosphate backbone of the major groove [127].

Figure 5. RNA structure impacts precursor RNA processing and influences gene expression

Schematic showing RNA processing in the nucleus (left, blue) and the impact of processing on gene expression in the cytoplasm (right, pink). (A) Alternative splicing can result in either exon skipping or exon inclusion (left, top). In MAPT RNA, exon skipping is promoted by hairpin formation at the 5′ splice site of exon 10. Exon skipping produces 3R and 4R transcripts and their corresponding protein isoforms. The 3R and 4R MAPT proteins have different biological functions. (B) MBNL1 binding to MBNL1 RNA at the 3′ splice site of intron 4 promotes exon skipping. Exon skipping produces a transcript missing a bipartite nuclear localization motif and a cell-wide protein isoform. Exon inclusion produces a transcript that is translated into a nuclear MBNL1 protein isoform. (C) In SMN2 RNA, exon skipping is promoted by hairpin formation at the 5′ splice site of exon 7. Exon skipping results in a protein isoform of SMN that is less stable than the full length SMN containing exon 7. Levels of SMN are associated with the severity of spinal muscular atrophy. (D) Most RNAs are polyadenylated at the 3′UTR (bottom). In SMN2 processing the 3′ end of the transcript contains a PIE structural element bound by U1A that inhibits polyadenylation. Little or no polyadenylation leads to transcript instability and a decrease in RNA levels.