A high resolution A-to-I editing map in the mouse identifies editing events controlled by pre-mRNA splicing

Pre-mRNA-splicing and adenosine to inosine (A-to-I) RNA-editing occur mostly cotranscriptionally. During A-to-I editing, a genomically encoded adenosine is deaminated to inosine by adenosine deaminases acting on RNA (ADARs). Editing-competent stems are frequently formed between exons and introns. Consistently, studies using reporter assays have shown that splicing efficiency can affect editing levels. Here, we use Nascent-seq and identify ∼90,000 novel A-to-I editing events in the mouse brain transcriptome. Most novel sites are located in intronic regions. Unlike previously assumed, we show that both ADAR (ADAR1) and ADARB1 (ADAR2) can edit repeat elements and regular transcripts to the same extent. We find that inhibition of splicing primarily increases editing levels at hundreds of sites, suggesting that reduced splicing efficiency extends the exposure of intronic and exonic sequences to ADAR enzymes. Lack of splicing factors NOVA1 or NOVA2 changes global editing levels, demonstrating that alternative splicing factors can modulate RNA editing. Finally, we show that intron retention rates correlate with editing levels across different brain tissues. We therefore demonstrate that splicing efficiency is a major factor controlling tissue-specific differences in editing levels.


Supplemental Figures
Supplemental Figure S1. The intronic coverage is higher for Nascent-seq than in poly(A) mRNA-seq.
The coverage for two different edited transcripts (A) Cog3, and (B) Grik1 is given for nascent-seq (yellow) and poly(A) mRNA-seq (green). The intron-exon structure of the gene is depicted below (blue); exons: bars; introns: thin lines. Editing sites are marked by arrows but can also be seen in the coverage profile (blue/red lines indicate mismatches).
Supplemental Figure S2. Sanger sequencing of cDNA (top rows) and corresponding gDNA (bottom rows) for all 22 editing sites tested to validate the pipeline to identify editing sites. Editing sites are marked by an arrow. Coordinates of the editing sites are given above the arrow (mm10).
Supplemental Figure S3. Venn diagram showing the overlap of identified editing sites between this study and previous studies (Danecek et al. 2012;Huntley et al. 2016;Liddicoat et al. 2016;Srivastava et al. 2017).
Supplemental Figure S4. The majority of editing sites identified using Nascent-seq in wildtype mice is edited below 20%. All editing sites have been binned according to their level of editing.
Supplemental Figure S5. (A, B) Comparison between bioinformatically predicted editing-complementary sequences (ECS) and randomized control sequences. (A) An empirical p-value was calculated by comparing the hybridization energy of the predicted ECS to the energies calculated from 1,000 dinucleotide shuffled input sequences. ECSs were accepted at a p-value level<=0.001. (B) To control the prediction, the same analysis was repeated with randomized editing site positions. Thereby, for each editing site an adenosine from the genome was randomly selected, but preserving its original genomic annotation and repeat status.
Supplemental Figure S6. Validation of bioinformatically predicted editing-complementary sequences (ECS). The genomic DNA encoding for edited transcripts and the corresponding predicted editing complementary sequence was cloned (+ ECS). Subsequently, the editing complementary region has been removed (the predicted secondary structure and the removed ECS (-ECS) regions is displayed in Supplemental Dataset S1). Following co-transfection with a plasmid coding for Flag-rADAR2 into HEK293 cells, RNA was isolated, reverse transcribed, amplified with target-specific primers, and submitted to Sanger sequencing. Genome coordinates: mm10.
Supplemental Figure S7. The persistence of the ECS increases editing levels. (A-C) Boxplots showing binned editing sites according to their log 10 ES/ECS coverage (red: log 10 ES/ECS coverage<0  ES saturated with ECS; blue: log 10 ES/ECS coverage>0  ES deprived of ECS) and the respective editing level (left side). The analysis was done separately for (A) exonic editing sites with an intronic ECS (exonicintronic), for (B) non-repeat associated sites, and for (C) repeat associated sites.
Supplemental Figure S8. Editing sites split according to their repeat status (compare to main figure 4).
The editing sites were split according to their repeat status (upper panels: non-repeat, lower panels: repeat). Only sites with a minimum read coverage of 10 are shown.
Boxplots showing the log 2 fold change (log 2 FC) for editing levels in untreated (DMSO) and treated (MEA) primary cells (left panels: bone marrow, right panels: primary neurons) separated into different genic locations as indicated below the panels (exonic, intronic, UTR, n.a.: not annotated/intergenic). Dots represent editing levels for individual sites that are changed significantly (red; p-value<0.05) or not significantly (grey; p-value>=0.05). Supplemental Figure S11. Deletion of NOVA1 or NOVA2 proteins does not perturb overall splicing efficiency. Re-analysis of publicly available RNA-seq data from cortices of 6 wildtype and 3 Nova1 -/and 3

Supplemental
Nova2 -/mice (please compare to main figure 5). The reads have been mapped to the mouse genome (mm10). The relative intron coverage for editing sites in wildtype, Nova1 -/-, and Nova2 -/mice is given.

Experimental validation of editing complementary site (ECS) predictions
A DNA sequence encompassing the randomly selected editing site and the predicted ECS was amplified from genomic mouse DNA and cloned into pcDNA3.1 (-) using primers fw 1 and rev 1 (Supplemental Table   S4). To remove the ECS primers fw 1/rev 2 and fw 2/rev 1 were used, respectively. Subsequently, both PCR products were combined and again amplified with primers fw 1 and rev 1. Restriction enzymes were used as indicated (Supplemental Table S4). To test editing, 3x10 5 Hek293 cells were plated per well using 6-well dishes. 24 h after plating, 0.5 µg of plasmid (plus or minus ECS) was co-transfected with 3.5 µg of a plasmid expressing Flag-rADAR2 using poly-ethyleneimine (Polysciences cat# 23966). 2 days after transfection RNA was isolated using Trifast™ (Peqlab cat# 30-2010) following the manufacturer's instructions. Subsequently, RNA was DNase I treated (New England Biolabs, cat# M0303). Following reverse transcription with random hexamers (Integrated DNA Technologies) and M-MuLV reverse transcriptase (New England Biolabs, cat# M0253) the editing site was PCR amplified using OneTaq Green PCR mix (New England Biolabs, cat# M0482) and primer pairs as indicated (Supplemental Table S5).
Following gel extraction using Monarch® DNA Gel Extraction Kit (New England Biolabs, cat# T1020) the PCR product was submitted to Sanger sequencing (Eurofins Genomics) using the primer as indicated (Supplemental Table S5). If not stated otherwise all custom primers were obtained from Microsynth Austria.

Supplemental Datasets
Supplemental Dataset S1: Structure predictions to validate bioinformatically predicted ECS. The structure has been predicted using RNAfold and visualized using forna (Lorenz et al. 2011;Kerpedjiev et al. 2015).
The editing site is shown in red; the editing complementary site is highlighted in blue. Light blue indicates the sequence that has been deleted in the deltaECS constructs.

Transcript: Armc9
Supplemental Dataset S2. Structure predictions of all edited transcripts used in Figure 6 were no experimental evidence for the location of the editing complementary site (ECS) is available. The structure has been predicted using RNAfold and visualized using forna. The exonic sequence is highlighted in yellow. The edited adenosine is marked in red. The intronic sequence is not colored.