Research ArticleGENETICS

Analysis and minimization of cellular RNA editing by DNA adenine base editors

See allHide authors and affiliations

Science Advances  08 May 2019:
Vol. 5, no. 5, eaax5717
DOI: 10.1126/sciadv.aax5717
  • Fig. 1 RNA and DNA editing activity of each TadA monomer in ABEmax.

    (A) ABEmax (shown as a schematic model) comprises three proteins fused in a single chain: TadA-TadA*-Cas9(D10A). (B) The two TadA monomers (shown as a schematic model) in ABEmax. The schematic models in (A) and (B) are generated from independently solved Cas9 [Protein Data Bank (PDB) id: 4un3] and E. coli TadA (PDB id: 1z3a) structures, as the structure of ABE has not yet been solved. (C) Average A-to-I conversion frequency in three mRNA transcripts from each treatment analyzed by high-throughput sequencing (HTS). (D) The number of adenosines within a 220- to 240-nt region of the indicated mRNA that are converted to inosine [read as a G after cDNA synthesis and DNA sequencing] at a detectable level (≥0.1%). Cas9(D10A) controls show the number of adenosines that are edited by endogenous cellular adenosine deaminases. The amplified regions of RSL1D1, CTNNB1, and IP90 mRNA have 46, 59, and 77 sequenced adenosines, respectively. (E) DNA base editing at seven genomic loci from ABEmax or by ABEmax with mutations at catalytic Glu59 in TadA or TadA*. The protospacer position of the target A and the sequence context of the A are shown. (F) RNA editing frequencies at various adenosines within the RSL1D1 amplicon after treatment with the indicated base editors. The adenosine homologous to TadA’s native substrate is at position 152 within the amplicon. (G) On-target DNA base editing with the low-density lipoprotein receptor (LDLR) sgRNA leads to a U-to-C (red to blue) edit in the LDLR mRNA in the transcriptome-wide RNA sequencing (RNA-seq) data. Alignments were visualized in the Integrated Genomics Viewer (IGV) and aligned to hg38. (H) Transcriptome-wide RNA-seq analysis showing the number of high-confidence (Phred quality score, ≥20; see Materials and Methods) A-to-I variant calls after treatment with the indicated base editors. The line represents the number of A-to-I conversions in the transcriptome from endogenous deaminase activity as measured in the Cas9(D10A) control samples. (I) The average frequency (%) of A-to-I RNA editing across all transcripts. For (A) to (F), data are shown as individual data points and means ± SD for n = 3 independent biological replicates performed on different days. For (H) and (I), data are shown as means ± SEM. The alignment was generated by combining reads from three independent biological replicates performed on different days.

  • Fig. 2 Design and testing of ABEmax mutants with reduced RNA editing activity.

    Views of the structure of S. aureus TadA bound to a minimized version of its native substrate (tRNAArg2) (PDB id: 2B3J) (23), showing the residues homologous to Arg47 (A), Asp108 (B), and Ala106 (C) in E. coli TadA. Asp108 is mutated to Asn108 in the evolved TadA*, while Ala106 is mutated to Val106 in TadA* (1). (D) DNA base editing at seven genomic loci from ABEmax or ABEmax mutants. (E) The number of adenosines converted to inosine at a detectable level (>0.1%) within a 220- to 240-nt region of the indicated mRNA by ABEmax or ABEmax mutants. The amplified regions of RSL1D1, CTNNB1, and IP90 mRNA have 46, 59, and 77 sequenced adenosines, respectively. The Cas9(D10A) controls show the number of adenosines that are edited because of endogenous A-to-I editing activity. (F) Average A-to-I RNA editing frequencies by ABEmax or ABEmax mutants among 46 adenosines in RSL1D1, 59 in CTNNB1, and 77 in IP90 mRNA transcripts. (G) On-target DNA base editing with the LDLR sgRNA leads to a U-to-C edit in the LDLR mRNA in the transcriptome-wide RNA-seq data. Alignments were visualized in the IGV and aligned to hg38. (H) Transcriptome-wide RNA-seq analysis showing the number of high-confidence (Phred quality score, ≥20; see Materials and Methods) A-to-I variant calls after treatment with the indicated base editors. The line represents the number of A-to-I conversions in the transcriptome from endogenous deaminase activity as measured in the Cas9(D10A) control samples. (I) The average frequency (%) of A-to-I RNA editing across all transcripts. For (C) to (F), data are shown as individual data points and means ± SD for n = 3 independent biological replicates performed on different days. For (H) and (I), data are shown as means ± SEM. The alignment was generated by combining reads from three independent biological replicates performed on different days.

  • Fig. 3 Analysis of A-to-I RNA edits found in transcriptome-wide RNA-seq.

    (A) Classification of the position in which an A-to-I RNA edit was found. “Five-kilobase downstream” refers to mutations that occur within 5-kb downstream of a coding gene, and “5-kb upstream” refers to mutations that occur within the region 5-kb upstream of a coding gene. (B) For edits in protein coding regions of mRNAs, edits were classified into synonymous or nonsynonymous mutations. (C) For nonsynonymous A-to-I edits in protein-coding regions of RNA, SIFT was used to predict the effect on protein function for these edits. High- or low-confidence calls (indicated in parentheses in the figure) were made according to the standard parameters of the prediction software (see Materials and Methods).

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/5/eaax5717/DC1

    Fig. S1. Indel frequencies associated with ABEmax and engineered ABEmax mutants.

    Fig. S2. DNA base editing and indel formation in HeLa cells from ABEmax and ABEmax mutants.

    Fig. S3. DNA base editing, indel formation, and RNA editing in U2OS and K562 cells harvested 48 hours after nucleofection with ABEmax, ABEmax mutants, or Cas9(D10A).

    Fig. S4. DNA base editing, indel formation, and RNA editing in HEK293T cells harvested 5 days after transfection with ABEmax or ABEmax mutants.

    Fig. S5. Off-target DNA base editing associated with the HEK site 2 locus by ABEmax and ABEmax mutants.

    Fig. S6. Off-target DNA base editing associated with the HEK site 3 locus by ABEmax and ABEmax mutants.

    Fig. S7. Off-target DNA base editing associated with the HEK site 4 locus by ABEmax and ABEmax mutants.

    Fig. S8. DNA base editing, indel formation, and RNA editing in HEK293T cells harvested 48 hours after transfection with ABEmax, ABEmaxAW, ABEmaxQW or ABEmax(TadA* A106V).

    Fig. S9. A-to-I RNA editing across the transcriptome for ABEmax, ABEmaxAW, ABEmax(TadA E59A), and Cas9(D10A).

    Fig. S10. Depiction of plasmid maps used in this study.

    Table S1. Guide RNA sequences.

    Table S2. Primers used for amplification of genomic DNA or cDNA for HTS.

    Table S3. List of amplicon sequences used for alignment and analysis of HTS reads.

    Table S4. List of primers used to amplify genomic off-target loci.

    Table S5. List of interrogated off-target genomic loci (28), with guide RNA sequences and amplicons used for alignment.

    Table S6. List of plasmid accession numbers from Addgene.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Indel frequencies associated with ABEmax and engineered ABEmax mutants.
    • Fig. S2. DNA base editing and indel formation in HeLa cells from ABEmax and ABEmax mutants.
    • Fig. S3. DNA base editing, indel formation, and RNA editing in U2OS and K562 cells harvested 48 hours after nucleofection with ABEmax, ABEmax mutants, or Cas9(D10A).
    • Fig. S4. DNA base editing, indel formation, and RNA editing in HEK293T cells harvested 5 days after transfection with ABEmax or ABEmax mutants.
    • Fig. S5. Off-target DNA base editing associated with the HEK site 2 locus by ABEmax and ABEmax mutants.
    • Fig. S6. Off-target DNA base editing associated with the HEK site 3 locus by ABEmax and ABEmax mutants.
    • Fig. S7. Off-target DNA base editing associated with the HEK site 4 locus by ABEmax and ABEmax mutants.
    • Fig. S8. DNA base editing, indel formation, and RNA editing in HEK293T cells harvested 48 hours after transfection with ABEmax, ABEmaxAW, ABEmaxQW or ABEmax(TadA* A106V).
    • Fig. S9. A-to-I RNA editing across the transcriptome for ABEmax, ABEmaxAW, ABEmax(TadA E59A), and Cas9(D10A).
    • Fig. S10. Depiction of plasmid maps used in this study.
    • Table S1. Guide RNA sequences.
    • Table S2. Primers used for amplification of genomic DNA or cDNA for HTS.
    • Table S3. List of amplicon sequences used for alignment and analysis of HTS reads.
    • Table S4. List of primers used to amplify genomic off-target loci.
    • Table S5. List of interrogated off-target genomic loci (28), with guide RNA sequences and amplicons used for alignment.
    • Table S6. List of plasmid accession numbers from Addgene.

    Download PDF

    Files in this Data Supplement:

Navigate This Article