A-to-I mRNA editing recodes hundreds of genes in dozens of species and produces endogenous protein isoforms in bacteria

Abstract

Adenosine-to-inosine (A-to-I) messenger RNA (mRNA) editing can affect the sequence and function of translated proteins and has been extensively investigated in eukaryotes. However, the prevalence of A-to-I mRNA editing in bacteria, its governing regulatory principles, and its biological significance are poorly understood. Here, we show that A-to-I mRNA editing occurs in hundreds of transcripts across dozens of gammaproteobacterial species, with most edits predicted to recode protein sequences. Furthermore, we reveal conserved regulatory determinants controlling editing across gammaproteobacterial species. Using Acinetobacter baylyi as a model, we show that mutating TadA, the mediating enzyme, reduces editing across all sites. Conversely, overexpressing TadA resulted in the editing of >300 transcripts, attesting to the editing potential of TadA. Notably, we show for the first time, at the protein level, that normal levels of A-to-I mRNA editing lead to wild-type bacteria expressing two protein isoforms from a single gene. Finally, we show that a TadA mutant with deficient editing activity does not grow at high temperatures, suggesting that RNA editing has a functional role in bacteria. Our work reveals that A-to-I mRNA editing in bacteria is widespread and has the potential to reshape the bacterial transcriptome and proteome.

Graphical Abstract

Figure 1.

Hundreds of mRNAs are edited and predicted to recode protein sequences across 64 gammaproteobacterial species. (A) Adenosine is deaminated to inosine, which is similar to guanosine in its base-pairing properties. Therefore, ribosomes and reverse transcriptases recognize inosine as guanosine. (B) TadA is the only known mRNA (and tRNA) A-to-I editing enzyme in bacteria. TadA requires a TACG (UACG) motif in the targeted RNA for its activity. (C) Four-base motif distribution around A>G mismatches (possible RNA editing sites) between RNA-seq data and the reference genome of 72 bacterial species. We found significant enrichment for the TACG motif (P-value <.0001; chi-square test for goodness of fit; marked in red). In black is the observed distribution, and in gray is the expected motif distribution of 1751 A>G sites when sampling randomly from the genome of the examined species. (D) Distribution and average editing level (percent of edited transcript) of the 381 mRNA editing events across 64 gammaproteobacterial species with detected editing. (E) Sanger sequencing of matched DNA and RNA samples of 11 genes/transcripts in three representative species. We observed a double peak of A and G (I) in 10/11 examined sites in mRNA (complementary DNA, cDNA) sequences but not the corresponding DNA (genome) sequences. Above each site is the gene’s name or locus tag. *Editing is present as a small peak. See Supplementary Fig. S4 for a closeup.

Figure 2.

Hundreds of mRNAs are edited and predicted to recode protein sequences across 64 gammaproteobacterial species. (A) Distribution of observed (red) and expected (black) positions of editing events within codons (P-value <.0001; chi-square test for goodness of fit; marked in red). (B) Distribution of observed (red) and expected (black) effects of editing events on protein sequences (P-value <.0001; chi-square test for goodness of fit; marked in red). (C) Distribution of observed effect of editing events on protein sequences. (D) The conservation of editing events. (E) Amino acid identity at the position recoded by conserved editing events across hundreds to thousands of gammaproteobacterial species.

Figure 3.

Bacterial mRNA editing occurs in a conserved seven-base motif embedded within a stem-loop structure. (A) WebLogo [48] of the 381 mRNA editing events detected in 64 gammaproteobacterial species. Position “0” is the edited site. (B) Distribution and editing levels of the 381 mRNA editing events across different seven-base motif combinations (data are provided in Supplementary Table S11). Mismatches to the conserved seven-base motif are marked in “x”. (C) Individually mutating different positions in the motif completely or nearly completely abolishes editing in two transcripts from different species. (D) Minimum free energy (MFE) secondary structure predicted by RNAfold [49] around the A-to-I editing site (red) for the 17 nucleotides composing the anticodon arm of tRNA^Arg2, and for the 17 nucleotides around the edited site in the transcripts of hokB and trpD. (E) MFE (mean and standard error) predictions for 17-nucleotide sliding windows centered on the positions indicated relative to all A-to-I mRNA editing events (n = 381) and control sites harboring all other YTACGAA motifs from all examined species (42 813 sites from 64 species). Statistical analysis on the 17 nucleotides surrounding A-to-I mRNA editing events and control sites (at position “0”) was conducted using Welch’s t-test (marked with a black arrow).

Figure 4.

TadA mediates A-to-I mRNA editing in A. baylyi and has the potential to reshape the transcriptome. (A) RNA editing level (%) of mRNAs and tRNA^Arg2 in the WT and TadA^D54E mutant strains of A. baylyi determined from RNA-seq data. The means and standard errors of four biological replicates conducted on different days are shown (N= 4). (B) Relative change in editing levels in the TadA^D54E strain compared to the WT strain as measured by RNA-seq. Each dot represents the reduction in average editing level in each edited RNA. See calculations in Supplementary Table S13. (C) Sanger sequencing of the mRNAs and tRNA^Arg2 with the highest level of editing shown in panel (A). (D) RNA editing level (%) of mRNAs and tRNA^Arg2 in the TadA^D54E mutant supplemented with TadA^WT or mRFP from the pBTK402 plasmid determined from RNA-seq data. The means and standard errors of four biological replicates conducted on different days are shown (N = 4). (E) Average and standard error of mRNA editing events detected in the TadA^D54E strain overexpressing TadA^WT (433 sites) or mRFP (6 sites) from the pBTK402 plasmid. Each dot represents the average editing level of a given site as detected in all four biological replicates. (F) Sanger sequencing of the mRNAs and tRNA^Arg2 with the highest level of editing shown in panel (C). Statistical analysis in panel (A), (B), and (D) was conducted using Student’s t-test followed by Benjamini–Hochberg false discovery rate (FDR) correction (in panels A and D), and in panel (E) using Welch’s t-test: P-value ≤.01 (**), ≤.001 (***), and ≤.0001 (****).

Figure 5.

A-to-I mRNA editing introduces protein isoforms in bacteria. Left: Representative MS/MS spectrum of edited (C38; top) and non-edited (Y38; bottom) RpsU peptides, and their normalized frequencies. Black arrows mark identified peptides and their mass in the MS/MS spectra that show a shift in mass corresponding to the presence of a tyrosine or cysteine at the edited site and can be compared between the two MS/MS spectra (different mass). The gray arrow marks an example of a peptide and its mass in the MS/MS spectra that does not include the edited site (same mass). All peptides were discovered with FDR ≤ 0.01. The peaks’ weight, font size, and axis were adjusted from the original figure for better visualization and only the peaks of identified peptides that correspond to the edited or non-edited peptides are shown. A comprehensive mass distribution and the original MS/MS spectra can be found in Supplementary Fig. S6 and Supplementary Table S18. Right: Relative peptide frequencies in the TadA^D54E mutant and the WT strain of A. baylyi. The mean and standard error of three biological replicates conducted on different days (N = 3) are shown. Statistical analysis was conducted using Welch’s t-tests: *P-value ≤.05.

Figure 6.

TadA mutant with deficient editing activity does not grow at high temperatures. (A) Growth assays with the WT and TadA^D54E mutant strain of A. baylyi in LB. The mean and standard error of five biological replicates conducted on different days (N = 5), each with 42 technical replicates, are shown. The dashed gray line represents the end of the log phase in the WT strain and was used for statistical analysis using Student’s t-tests: ****P-value ≤.0001. (B) A-to-I RNA editing level (%) of mRNAs and tRNA^Arg2 in WT A. baylyi determined from RNA-seq data of samples that grew at 42°C to mid-logarithmic phase. The means and standard errors of three biological replicates conducted on different days are shown (N = 3).