The m1A landscape on cytosolic and mitochondrial mRNA at single-base resolution

Abstract

Modifications on mRNA offer the potential of regulating mRNA fate post-transcriptionally. Recent studies suggested the widespread presence of N1-methyladenosine (m1A), which disrupts Watson–Crick base pairing, at internal sites of mRNAs1,2. These studies lacked the resolution of identifying individual modified bases, and did not identify specific sequence motifs undergoing the modification or an enzymatic machinery catalysing them, rendering it challenging to validate and functionally characterize putative sites. Here we develop an approach that allows the transcriptome-wide mapping of m1A at single-nucleotide resolution. Within the cytosol, m1A is present in a low number of mRNAs, typically at low stoichiometries, and almost invariably in tRNA T-loop-like structures, where it is introduced by the TRMT6/TRMT61A complex. We identify a single m1A site in the mitochondrial ND5 mRNA, catalysed by TRMT10C, with methylation levels that are highly tissue specific and tightly developmentally controlled. m1A leads to translational repression, probably through a mechanism involving ribosomal scanning or translation. Our findings suggest that m1A on mRNA, probably because of its disruptive impact on base pairing, leads to translational repression, and is generally avoided by cells, while revealing one case in mitochondria where tight spatiotemporal control over m1A levels was adopted as a potential means of post-transcriptional regulation.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Figure 1: Establishment of m1A-seq and characterization of 205 putative m1A-containing sites.
Figure 2: Cytosolic m1A sites share a sequence and structural motif and are strongly induced upon overexpression of TRMT6/TRMT61A.
Figure 3: Tissue- and development-specific methylation in ND5 catalysed via TRMT10C.
Figure 4: m1A-containing mRNAs are inefficiently translated.

Accession codes

Primary accessions

Gene Expression Omnibus

Referenced accessions

Gene Expression Omnibus

References

  1. 1

    Li, X. et al. Transcriptome-wide mapping reveals reversible and dynamic N1-methyladenosine methylome. Nat. Chem. Biol. 12, 311–316 (2016)

  2. 2

    Dominissini, D. et al. The dynamic N1-methyladenosine methylome in eukaryotic messenger RNA. Nature 530, 441–446 (2016)

  3. 3

    Hauenschild, R. et al. The reverse transcription signature of N-1-methyladenosine in RNA-seq is sequence dependent. Nucleic Acids Res. 43, 9950–9964 (2015)

  4. 4

    Jones, J. W. & Robins, R. K. Purine nucleosides. III. Methylation studies of certain naturally occurring purine nucleosides. J. Am. Chem. Soc. 85, 193–201 (1963)

  5. 5

    Mohr, S. et al. Thermostable group II intron reverse transcriptase fusion proteins and their use in cDNA synthesis and next-generation RNA sequencing. RNA 19, 958–970 (2013)

  6. 6

    Zubradt, M. et al. DMS-MaPseq for genome-wide or targeted RNA structure probing in vivo. Nat. Methods 14, 75–82 (2017)

  7. 7

    Wilusz, J. E., Freier, S. M. & Spector, D. L. 3′ end processing of a long nuclear-retained noncoding RNA yields a tRNA-like cytoplasmic RNA. Cell 135, 919–932 (2008)

  8. 8

    Hodgkinson, A. et al. High-resolution genomic analysis of human mitochondrial RNA sequence variation. Science 344, 413–415 (2014)

  9. 9

    Bar-Yaacov, D. et al. RNA-DNA differences in human mitochondria restore ancestral form of 16S ribosomal RNA. Genome Res. 23, 1789–1796 (2013)

  10. 10

    GTEx Consortium. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science 348, 648–660 (2015)

  11. 11

    Ozanick, S., Krecic, A., Andersland, J. & Anderson, J. T. The bipartite structure of the tRNA m1A58 methyltransferase from S. cerevisiae is conserved in humans. RNA 11, 1281–1290 (2005)

  12. 12

    Jühling, F. et al. tRNAdb 2009: compilation of tRNA sequences and tRNA genes. Nucleic Acids Res. 37, D159–D162 (2009)

  13. 13

    Safra, M., Nir, R., Farouq, D. & Schwartz, S. TRUB1 is the predominant pseudouridine synthase acting on mammalian mRNA via a predictable and conserved code. Genome Res. http://dx.doi.org/10.1101/gr.207613.116 (2017)

  14. 14

    Petropoulos, S. et al. Single-cell RNA-seq reveals lineage and X chromosome dynamics in human preimplantation embryos. Cell 165, 1012–1026 (2016)

  15. 15

    Bar-Yaacov, D. et al. Mitochondrial 16S rRNA is methylated by tRNA methyltransferase TRMT61B in all vertebrates. PLoS Biol. 14, e1002557 (2016)

  16. 16

    Yan, L. et al. Single-cell RNA-seq profiling of human preimplantation embryos and embryonic stem cells. Nat. Struct. Mol. Biol. 20, 1131–1139 (2013)

  17. 17

    Pikó, L. & Taylor, K. D. Amounts of mitochondrial DNA and abundance of some mitochondrial gene transcripts in early mouse embryos. Dev. Biol. 123, 364–374 (1987)

  18. 18

    Vilardo, E. et al. A subcomplex of human mitochondrial RNase P is a bifunctional methyltransferase—extensive moonlighting in mitochondrial tRNA biogenesis. Nucleic Acids Res. 40, 11583–11593 (2012)

  19. 19

    Hudson, G. et al. Clinical expression of Leber hereditary optic neuropathy is affected by the mitochondrial DNA-haplogroup background. Am. J. Hum. Genet. 81, 228–233 (2007)

  20. 20

    Torroni, A. et al. Haplotype and phylogenetic analyses suggest that one European-specific mtDNA background plays a role in the expression of Leber hereditary optic neuropathy by increasing the penetrance of the primary mutations 11778 and 14484. Am. J. Hum. Genet. 60, 1107–1121 (1997)

  21. 21

    Carelli, V. et al. Haplogroup effects and recombination of mitochondrial DNA: novel clues from the analysis of Leber hereditary optic neuropathy pedigrees. Am. J. Hum. Genet. 78, 564–574 (2006)

  22. 22

    Brown, M. D., Sun, F. & Wallace, D. C. Clustering of Caucasian Leber hereditary optic neuropathy patients containing the 11778 or 14484 mutations on an mtDNA lineage. Am. J. Hum. Genet. 60, 381–387 (1997)

  23. 23

    Schwartz, S. et al. Transcriptome-wide mapping reveals widespread dynamic-regulated pseudouridylation of ncRNA and mRNA. Cell 159, 148–162 (2014)

  24. 24

    Carlile, T. M. et al. Pseudouridine profiling reveals regulated mRNA pseudouridylation in yeast and human cells. Nature 515, 143–146 (2014)

  25. 25

    Lovejoy, A. F., Riordan, D. P. & Brown, P. O. Transcriptome-wide mapping of pseudouridines: pseudouridine synthases modify specific mRNAs in S. cerevisiae. PLoS ONE 9, e110799 (2014)

  26. 26

    Edelheit, S., Schwartz, S., Mumbach, M. R., Wurtzel, O. & Sorek, R. Transcriptome-wide mapping of 5-methylcytidine RNA modifications in bacteria, archaea, and yeast reveals m5C within archaeal mRNAs. PLoS Genet. 9, e1003602 (2013)

  27. 27

    Mills, R. M. & Brinster, R. L. Oxygen consumption of preimplantation mouse embryos. Exp. Cell Res. 47, 337–344 (1967)

  28. 28

    Ginsberg, L. & Hillman, N. ATP metabolism in cleavage-staged mouse embryos. J. Embryol. Exp. Morphol. 30, 267–282 (1973)

  29. 29

    Biggers, J. D. & Borland, R. M. Physiological aspects of growth and development of the preimplantation mammalian embryo. Annu. Rev. Physiol. 38, 95–119 (1976)

  30. 30

    Wassarman, P. M. & Josefowicz, W. J. Oocyte development in the mouse: an ultrastructural comparison of oocytes isolated at various stages of growth and meiotic competence. J. Morphol. 156, 209–235 (1978)

  31. 31

    Schwartz, S. et al. High-resolution mapping reveals a conserved, widespread, dynamic mRNA methylation program in yeast meiosis. Cell 155, 1409–1421 (2013)

  32. 32

    Shishkin, A. A. et al. Simultaneous generation of many RNA-seq libraries in a single reaction. Nat. Methods 12, 323–325 (2015)

  33. 33

    Engreitz, J. M. et al. The Xist lncRNA exploits three-dimensional genome architecture to spread across the X chromosome. Science 341, 1237973 (2013)

  34. 34

    Machnicka, M. A. et al. MODOMICS: a database of RNA modification pathways--2013 update. Nucleic Acids Res. 41, D262–D267 (2013)

  35. 35

    Dobin, A . et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013)

  36. 36

    Schwartz, S. et al. Perturbation of m6A writers reveals two distinct classes of mRNA methylation at internal and 5′ sites. Cell Reports 8, 284–296 (2014)

  37. 37

    Li, B. & Dewey, C. N. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12, 323 (2011)

  38. 38

    Robinson, M. D. & Oshlack, A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol. 11, R25 (2010)

  39. 39

    Robinson, M. D., McCarthy, D. J. & Smyth, G. K . edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010)

  40. 40

    Gandin, V. et al. Polysome fractionation and analysis of mammalian translatomes on a genome-wide scale. J. Vis. Exp. 87, http://dx.doi.org/10.3791/51455 (2014)

  41. 41

    Dominissini, D. et al. Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq. Nature 485, 201–206 (2012)

Download references

Acknowledgements

This project was supported by the Israel Science Foundation (543165), the European Research Council under the European Union’s Horizon 2020 research and innovation programme (grant agreement number 714023), by the Abisch-Frenkel-Stiftung, by research grants from The Abramson Family Center for Young Scientists, the David and Fela Shapell Family Foundation INCPM Fund for Preclinical Studies, the Estate of David Turner, and the Berlin Family Foundation New Scientist Fund.

Author information

Affiliations

Authors

Contributions

M.S., A.S.-C., R.N., D.B., M.E., W.R., N.S.-G., and S.S. designed the experiments. M.S. performed the m1A-seq optimizations and experiments, the luciferase assays, and the genetic perturbations. M.S. and A.S.-C. performed the targeted sequencing experiments and the mitochondrial RNA half-life experiments. M.S. and R.W. performed the polysomal fractionations. R.N. performed the oligoarray experiment. A.N. assisted in the GTEx analysis. S.S. and M.S. performed the computational analysis. S.S. wrote the manuscript with input from all authors.

Corresponding author

Correspondence to Schraga Schwartz.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Additional information

Reviewer Information Nature thanks Y. Motorin and the other anonymous reviewer(s) for their contribution to the peer review of this work.

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended Data Figure 1 Analytical pipeline for identification of m1A sites.

a, Overview of the analytical pipeline, which was applied to each of the four data sets. The pipeline utilizes two key statistical tests to identify ‘A’ containing sites that either harbour misincorporations at higher levels following immunoprecipitation with an anti-m1A antibody compared with input samples or compared with Dimroth-treated samples. In addition, the pipeline requires that minimum levels of at least two distinct types of misincorporation be observed for a given site. bd, Truncation and misincorporation rates are reproducible between replicates, both with SuperScript and with TGIRT enzymes. Shown are scatterplots based on 205 putative m1A sites, indicating the correlation between replicates of truncation rates using SuperScript (b) misincorporation rates using SuperScript (c), and misincorporation rates using TGIRT (d). e, Comparison of truncation rates using TGIRT (y axis) and SuperScript (x axis), on the basis of the 205 detected sites. Sites are colour-coded on the basis of the indicated classes of RNA. f, Comparison of misincorporation rates using TGIRT and SuperScript, plotted as in e. g, Comparison of misincorporation rates in IP samples compared with IP + Dimroth, in both cases using TGIRT. h, Comparison of misincorporation rates in IP samples compared with input, using TGIRT. i, Serial dilutions of RNA extracted from cells overexpressing TRMT6/TRMT61A within RNA extracted from cells depleted of TRMT6/TRMT61A. The x axis denotes the percentage of the RNA originating from cells overexpressing TRMT6/TRMT61A, and the y axis captures the misincorporation rate, as measured via targeted sequencing of the three presented sites.

Extended Data Figure 2 Misincorporation levels at putative sites on the basis of GTEx data set.

Shown are misincorporation levels across eight lncRNA + mRNA and two rRNA sites on the basis of a random sampling of 4,114 GTEx RNA-seq samples. Boxplots represent median, interquartile range, distribution, and outliers, as indicated in Fig. 2h.

Extended Data Figure 3 m1A detection as a function of gene expression, on the basis of analysis of sites identified following TRMT6/TRMT61A overexpression.

a, All genes were divided into five bins, on the basis of the indicated levels of expression. The number of m1A harbouring sites in each bin is plotted. b, As in a, but normalized by number of genes within each bin. Error bars, binomial confidence intervals.

Extended Data Figure 4 Peak detection following m1A-seq reveals a strong enrichment towards the 5′ terminus.

a, Representative meta-gene profiles of read coverage in m1A-IP (top) or input (bottom). Shown are the first 500 nt immediately following the transcription start site (left) and immediately preceding the annotated 3′ terminus of the gene (right). Genes are divided into five bins, on the basis of expression levels. b, Distribution of expression levels (on the basis of input samples) of all genes harbouring high-confidence peaks. On this distribution are overlaid, in red, the genes in which putative m1A sites were detected. c, Consistently identified peaks are highly enriched towards the 5′ terminus of the gene. Each peak was classified into one of five segments, as in ref. 41, in the following order: transcription start site (TSS) if the peak is present in the first 200 nt of the gene; 5′ UTR if in the 5′ UTR region but outside the transcription start site region; stop codon region, comprising 200 nt on both sides of the stop window; CDS region for peaks within the CDS; and 3′ UTR for remaining 3′ UTR peaks. Each peak is scored on the basis of the number of experiments in which it was detected, whereby more robustly identified peaks should be considered as ones of higher confidence. The stacked bar plots summarize the relative proportion of peaks in each segment, and the right-most bar plots the relative amount of ‘gene architecture’ taken up by each of these segments. d, Analysis of transcription start site peaks. The number of reads beginning (rather than overlapping) at each of the first 50 annotated transcribed bases was calculated across IP and input samples (Methods), and the log(fold change) between the two was derived. All fold changes were binned into six bins (as plotted) and the fraction of positions harbouring an ‘A’ are plotted as a function of this binned fold change, revealing that positions at the transcription start site that are enriched in IP samples over input samples are biased towards beginning with an ‘A’. Error bars, binomial error.

Extended Data Figure 5 Characterization of sequence and structure at m1A sites, and requirement of the TRMT6/TRMT61A machinery for its catalysis in the cytosol.

a, Predicted secondary structures of sequence environment surrounding putative m1A-containing sites in cytoplasmic mRNA and lncRNAs (see also Fig. 2a). b, Predicted secondary structure at the vicinity of mitochondrial m1A sites. A common secondary structure surrounding mitochondrial sites is depicted, featuring a ‘UAAA’ motif in the loop, stabilized by a stem structure. c, Misincorporation levels at the indicated sites, measured by targeted sequencing, in cells depleted of TRMT6/TRMT61A via siRNAs, compared with mock-treated controls. Error bars, binomial confidence intervals. d, e, Distributions of loop length (d) and stem length (e) among the 384 putative m1A sites in mRNA + lncRNAs, compared with randomly shuffled controls. These values are derived from RNAfold.

Extended Data Figure 6 Validation of putative m1A sites, via m6A-seq of Dimroth-converted RNA extracted from TRMT6/TRMT61A mRNA.

a, Coverage plot along the cytosolic 28S rRNA across the four indicated samples. The known m1A and m6A sites are indicated. b, Coverage plots as in a for the mitochondrial 16S rRNA. The normalized coverage levels are indicated to the left of the track (note the orders-of-magnitude higher coverage at the m1A sites upon m6A-IP in Dimroth-treated samples, compared with the controls). c, Quantification of percentage of coverage in a 100 nt centred region around putative m1A sites out of overall coverage of the gene in two biological replicates of mRNA extracted from cells overexpressing TRMT6/TRMT61A; this mRNA was subjected to Dimroth treatment followed by m6A-seq. Quantifications were obtained for 40 sites exhibiting misincorporation levels greater than 10% upon TRMT6/TRMT61A overexpression in addition to the 2 sites on the 16S and 28S rRNA molecules. d, Comparison of ‘% Coverage in peak’ (as in c) between RNA subjected to Dimroth treatment and m6A-IP compared with RNA only subjected to m6A-IP. e, Misincorporation levels across the indicated conditions across 8 sites (of the 42 tested) in which a significant P value (P < 0.05) was obtained when comparing Dimroth + m6A-IP samples with their corresponding input (significance indicated by ‘*’), and/or when comparing Dimroth + m6A-IP to no-Dimroth + m6A-IP (significance indicated by ‘#’). Error bars, binomial error after pooling of two biological replicates.

Extended Data Figure 7 ND5 methylation levels are genetically determined, are in part controlled by a SNP two bases upstream, and are increased in stable transcripts.

a, Snapshot of randomly sampled reads aligned to the ND5 locus. Reads originating from the heavy strand are depicted in red, reads from the light strand in purple. Top: IP sample; bottom: input sample. Misincorporations are apparent only in reads originating from the heavy strand. b, Misincorporation levels at the ND5 locus in DNA and RNA samples of five individuals. For RNA, the distribution of misincorporation reads are shown across all tissues; for DNA, the measurement consists of a single measurement available in GTEx. c, Misincorporation levels at ND5:1374 measured at the indicated time-points following ethidium-bromide-mediated transcriptional arrest (n = 3); Points, mean; error bars, s.e.m. d, Correlations between misincorporation levels at ND5:1374 in skin versus brain samples, from the same individuals. e, Histogram of all pairwise correlation coefficients between tissues (but from the same individuals); note that values are centred around 0.5, rather than around 0 if they were independent of each other. f, Misincorporation rates at ND5:1374 obtained via strand-specific targeted sequencing of the ND5 locus across six lymphoblastoid cell lines, two harbouring a G13708A SNP and four wild-type samples.

Extended Data Figure 8 m1A represses translation.

a, Representative sucrose gradient, indicating the division into fractions on the basis of the number of polysomes associated with them. b, Scheme of experimental design. The wild-type m1A-containing stretch from the PRUNE gene was cloned either in-frame and upstream of firefly luciferase (CDS construct) or as a 3′ UTR element (3′ UTR construct). Control or TRMT6/TRMT61A-overexpressing cells were co-transfected with each of these plasmids along with a plasmid expressing Renilla luciferase. c, d, Renilla-normalized firefly luciferase levels in TRMT6/TRMT61A-overexpressing cells, standardized by this value in the non-overexpressing (control) cells (n = 3). Note that for the CDS construct the presented data are identical to those in Fig. 4 and are re-plotted for convenience.

Extended Data Figure 9 Cytosolic m1A does not affect steady-state levels of mRNA or mRNA stability.

a, Comparison of expression levels of all genes acquiring robust levels of m1A (defined as misincorporation levels greater than 10%) upon overexpression of TRMT6/TRMT61A, compared with wild-type (non-overexpressing) counterparts. b, Misincorporation levels in the four indicated genes, in cells overexpressing TRMT6/TRMT61A, in a 6 h time course following cycloheximide treatment (n = 3). Error bars, s.e.m.

Supplementary information

Supplementary Information

This file contains Supplementary Notes 1-10 and Supplementary References.

Life Sciences Reporting Summary

Supplementary Table 1

Alignment statistics for all four analysed datasets, including total number of reads and the percentage of aligned reads.

Supplementary Table 2

Dataset of 277 putative m1A sites, obtained after intersection of the four analysed datasets – see Supplementary Information document for full description.

Supplementary Table 3

Dataset of 690 consistently identified peaks, used in the context of analyses presented in Extended Data Fig. 4 revealing strong bias towards 5’ termini of genes. The dataset is in bed format.

Supplementary Table 4

Dataset of 495 putative m1A sites, obtained following overexpression of TRMT6/TRMT61A.

Supplementary Table 5

A list of primer sequences used in this study.

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Safra, M., Sas-Chen, A., Nir, R. et al. The m1A landscape on cytosolic and mitochondrial mRNA at single-base resolution. Nature 551, 251–255 (2017). https://doi.org/10.1038/nature24456

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.