Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

m6A governs length-dependent enrichment of mRNAs in stress granules

Nature Structural & Molecular Biology volume 30, pages 1525–1535 (2023)Cite this article

Subjects

Abstract

Stress granules are biomolecular condensates composed of protein and mRNA. One feature of stress granule-enriched mRNAs is that they are often longer than average. Another feature of stress granule-enriched mRNAs is that they often contain multiple N6-methyladenosine (m6A) residues. m6A is bound by the YTHDF proteins, creating mRNA–protein complexes that partition into stress granules in mammalian cells. Here we show that length-dependent enrichment of mRNAs in stress granules is mediated by m6A. Long mRNAs often contain one or more long exons, which are preferential sites of m6A formation. In mammalian cells lacking m6A, long mRNAs no longer show preferential stress granule enrichment. Furthermore, we show that m6A abundance more strongly predicts which short or long mRNAs are enriched in stress granules, rather than length alone. Thus, mRNA length correlates with mRNA enrichment in stress granules owing to the high prevalence of m6A in long mRNAs.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on SpringerLink
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Knockout of Mettl3 and validation of stress response in M3KO MEFs.
Fig. 2: Quantification of stress differences in control and Mettl3-knockout MEFs.
Fig. 3: Effects of m6A depletion on total and stress granule mRNA abundance.
Fig. 4: m6A mediates length-dependent mRNA enrichment in stress granules.
Fig. 5: The effects of m6A on mRNA enrichment in stress granules is independent of length.

Similar content being viewed by others

Data availability

RNA-seq data and all associated analysis used for data visualization reported in this paper have been deposited in the Gene Expression Omnibus (GEO) under accession number GSE190121. Previously published datasets used in this study are available in GEO: GSE190121, GSE61995, GSM2300430, GSE92867, GSE78030, GSE177552, GSE177133, GSE99304 (U2OS), GSE90869 (NIH3T3). Source data are provided with this paper.

Code availability

All code used to perform analysis or create figures is available from the corresponding author upon reasonable request.

References

  1. Anderson, P. & Kedersha, N. Stress granules: the Tao of RNA triage. Trends Biochem. Sci. 33, 141–150 (2008).

    Article  CAS  PubMed  Google Scholar 

  2. Protter, D. S. W. & Parker, R. Principles and properties of stress granules. Trends Cell Biol. 26, 668–679 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Riggs, C. L., Kedersha, N., Ivanov, P. & Anderson, P. Mammalian stress granules and P bodies at a glance. J. Cell Sci. 133, jcs242487 (2020).

    Article  CAS  PubMed  Google Scholar 

  4. Ivanov, P., Kedersha, N. & Anderson, P. Stress granules and processing bodies in translational control. Cold Spring Harb. Perspect. Biol. 11, a032813 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Khong, A. et al. The stress granule transcriptome reveals principles of mRNA accumulation in stress granules. Mol. Cell 68, 808–820(2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Namkoong, S., Ho, A., Woo, Y. M., Kwak, H. & Lee, J. H. Systematic characterization of stress-induced RNA granulation. Mol. Cell 70, 175–187(2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Van Treeck, B. et al. RNA self-assembly contributes to stress granule formation and defining the stress granule transcriptome. Proc. Natl Acad. Sci. USA 115, 2734 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  8. Van Treeck, B. & Parker, R. Emerging roles for intermolecular RNA-RNA interactions in RNP assemblies. Cell 174, 791–802 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  9. Ries, R. J. et al. m6A enhances the phase separation potential of mRNA. Nature 571, 424–428 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Zaccara, S. & Jaffrey, S. R. A unified model for the function of YTHDF proteins in regulating m6A-modified mRNA. Cell 181, 1582–1595 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Zhu, T. et al. Crystal structure of the YTH domain of YTHDF2 reveals mechanism for recognition of N6-methyladenosine. Cell Res. 24, 1493–1496 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Patil, D. P., Pickering, B. F. & Jaffrey, S. R. Reading m6A in the transcriptome: m6A-binding proteins. Trends Cell Biol. 28, 113–127 (2017).

  13. Fu, Y. & Zhuang, X. m6A-binding YTHDF proteins promote stress granule formation. Nat. Chem. Biol. 16, 955–963 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Gao, Y. et al. Multivalent m6A motifs promote phase separation of YTHDF proteins. Cell Res. 29, 767–769 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

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

    Article  CAS  PubMed  Google Scholar 

  16. Batista, P. J. et al. m6A RNA modification controls cell fate transition in mammalian embryonic stem cells. Cell Stem Cell 15, 707–719 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Ke, S. et al. m6A mRNA modifications are deposited in nascent pre-mRNA and are not required for splicing but do specify cytoplasmic turnover. Genes Dev. 31, 990–1006 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Huang, H. et al. Histone H3 trimethylation at lysine 36 guides m6A RNA modification co-transcriptionally. Nature 567, 414–419 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Wang, Z. et al. m6A deposition is regulated by PRMT1-mediated arginine methylation of METTL14 in its disordered C-terminal region. EMBO J. 40, e106309 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Yue, Y. et al. VIRMA mediates preferential m6A mRNA methylation in 3' UTR and near stop codon and associates with alternative polyadenylation. Cell Discov. 4, 10 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  21. Geula, S. et al. Stem cells. m6A mRNA methylation facilitates resolution of naive pluripotency toward differentiation. Science 347, 1002–1006 (2015).

    Article  CAS  PubMed  Google Scholar 

  22. Poh, H. X., Mirza, A. H., Pickering, B. F. & Jaffrey, S. R. Alternative splicing of METTL3 explains apparently METTL3-independent m6A modifications in mRNA. PLoS Biol. 20, e3001683 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. De Leeuw, F. et al. The cold-inducible RNA-binding protein migrates from the nucleus to cytoplasmic stress granules by a methylation-dependent mechanism and acts as a translational repressor. Exp. Cell. Res. 313, 4130–4144 (2007).

    Article  PubMed  Google Scholar 

  24. Kimball, S. R., Horetsky, R. L., Ron, D., Jefferson, L. S. & Harding, H. P. Mammalian stress granules represent sites of accumulation of stalled translation initiation complexes. Am. J. Physiol. Cell Physiol. 284, C273–C284 (2003).

    Article  CAS  PubMed  Google Scholar 

  25. Zhang, X., Shu, X. E. & Qian, S.-B. O-GlcNAc modification of eIF4GI acts as a translational switch in heat shock response. Nat. Chem. Biol. 14, 909–916 (2018).

    Article  CAS  PubMed  Google Scholar 

  26. Cheng, Y. et al. m6A RNA methylation maintains hematopoietic stem cell identity and symmetric commitment. Cell Rep. 28, 1703–1716 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Kedersha, N. L., Gupta, M., Li, W., Miller, I. & Anderson, P. RNA-binding proteins Tia-1 and Tiar link the phosphorylation of Eif-2α to the assembly of mammalian stress granules. J. Cell Biol. 147, 1431–1442 (1999).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Jolly, C., Usson, Y. & Morimoto, R. I. Rapid and reversible relocalization of heat shock factor 1 within seconds to nuclear stress granules. Proc. Natl Acad. Sci. USA 96, 6769–6774 (1999).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Dewey Colleen, M. et al. TDP-43 is directed to stress granules by sorbitol, a novel physiological osmotic and oxidative stressor. Mol. Cell. Biol. 31, 1098–1108 (2011).

    Article  CAS  PubMed  Google Scholar 

  30. Patil, D. P. et al. m6A RNA methylation promotes XIST-mediated transcriptional repression. Nature 537, 369–373 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Aulas, A. et al. Stress-specific differences in assembly and composition of stress granules and related foci. J. Cell Sci. 130, 927–937 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  32. van Leeuwen, W. & Rabouille, C. Cellular stress leads to the formation of membraneless stress assemblies in eukaryotic cells. Traffic 20, 623–638 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  33. Sommer, S., Lavi, U. & Darnell, J. E. Jr. The absolute frequency of labeled N-6-methyladenosine in HeLa cell messenger RNA decreases with label time. J. Mol. Biol. 124, 487–499 (1978).

    Article  CAS  PubMed  Google Scholar 

  34. Zaccara, S., Ries, R. J. & Jaffrey, S. R. Reading, writing and erasing mRNA methylation. Nat. Rev. Mol. Cell Biol. 20, 608–624 (2019).

    Article  CAS  PubMed  Google Scholar 

  35. Du, H. et al. YTHDF2 destabilizes m6A-containing RNA through direct recruitment of the CCR4–NOT deadenylase complex. Nat. Commun. 7, 12626 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Olarerin-George, A. O. & Jaffrey, S. R. MetaPlotR: a Perl/R pipeline for plotting metagenes of nucleotide modifications and other transcriptomic sites. Bioinformatics 33, 1563–1564 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Van Nostrand, E. L. et al. A large-scale binding and functional map of human RNA-binding proteins. Nature 583, 711–719 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  38. Yang, P. et al. G3BP1 is a tunable switch that triggers phase separation to assemble stress granules. Cell 181, 325–345(2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Matheny, T., Van Treeck, B., Huynh, T. N. & Parker, R. RNA partitioning into stress granules is based on the summation of multiple interactions. RNA 27, 174–189 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Xia, Z. et al. Dynamic analyses of alternative polyadenylation from RNA-seq reveal a 3′-UTR landscape across seven tumour types. Nat. Commun. 5, 5274 (2014).

    Article  CAS  PubMed  Google Scholar 

  41. Ke, S. et al. A majority of m6A residues are in the last exons, allowing the potential for 3′ UTR regulation. Genes Dev. 29, 2037–2053 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Jain, A. & Vale, R. D. RNA phase transitions in repeat expansion disorders. Nature 546, 243–247 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Khong, A., Matheny, T., Huynh, T. N., Babl, V. & Parker, R. Limited effects of m6A modification on mRNA partitioning into stress granules. Nat. Commun. 13, 3735 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Jain, S. et al. ATPase-modulated stress granules contain a diverse proteome and substructure. Cell 164, 487–498 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Irgen-Gioro, S., Yoshida, S., Walling, V. & Chong, S. Fixation can change the appearance of phase separation in living cells. eLife 11, e79903 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  46. Zhong, S. et al. MTA is an Arabidopsis messenger RNA adenosine methylase and interacts with a homolog of a sex-specific splicing factor. Plant Cell 20, 1278–1288 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Andrews, S. FastQC: a Quality Control Tool for High Throughput Sequence Data (Babraham Bioinformatics, 2010).

  48. Dodt, M., Roehr, J. T., Ahmed, R. & Dieterich, C. FLEXBAR-flexible barcode and adapter processing for next-generation sequencing platforms. Biology 1, 895–905 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  49. Howe, K. L. et al. Ensembl 2021. Nucleic Acids Res. 49, D884–D891 (2021).

    Article  CAS  PubMed  Google Scholar 

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

    Article  CAS  PubMed  Google Scholar 

  51. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  52. Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer, 2016).

  53. Xiang, Y. et al. RNA m6A methylation regulates the ultraviolet-induced DNA damage response. Nature 543, 573–576 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods 25, 402–408 (2001).

    Article  CAS  PubMed  Google Scholar 

  55. Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Kent, W. J., Zweig, A. S., Barber, G., Hinrichs, A. S. & Karolchik, D. BigWig and BigBed: enabling browsing of large distributed datasets. Bioinformatics 26, 2204–2207 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank all members of the Jaffrey laboratory for comments and suggestions, and the lab of M. Kharas for generously providing Mettl3flox/flox mice that we used to make the MEF cell lines. This work was supported by the National Institutes of Health grants R35NS111631, R01CA186702, S10 OD030335 (S.R.J.), T32CA062948, F32CA22104 (B.F.P.), and F31CA254763 (R.J.R.), the Agency for Science, Technology, and Research (A*STAR) (H.X.P.), and National Research Foundation of Korea grant NRF-2020R1C1C1009253 (S.N.).

Author information

Authors and Affiliations

  1. Department of Pharmacology, Weill Cornell Medicine, Cornell University, New York, NY, USA

    Ryan J. Ries, Brian F. Pickering, Hui Xian Poh & Samie R. Jaffrey

  2. Department of Biochemistry, College of Natural Sciences, Kangwon National University, Chuncheon, Republic of Korea

    Sim Namkoong

Authors
  1. Ryan J. Ries
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Brian F. Pickering
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Hui Xian Poh
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Sim Namkoong
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Samie R. Jaffrey
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

R.J.R. and S.R.J. conceived the project and designed the experiments. R.J.R. performed all experiments unless stated otherwise. B.F.P. isolated MEFs and established cell lines used in the study. H.X.P. performed thin-layer chromatography of m6A in mRNA. S.N. purified the stress granules. R.J.R. designed and prepared the figures. R.J.R. and S.R.J. wrote the paper. The paper was read and approved by all authors.

Corresponding author

Correspondence to Samie R. Jaffrey.

Ethics declarations

Competing interests

S.R.J. is scientific founder of, advisor to, and/or owns equity in Chimerna Therapeutics, Lucerna Technologies, and 858 Therapeutics. The other authors declare no competing interests.

Peer review

Peer review information

Nature Structural & Molecular Biology thanks Benjamin Wolozin, Gene Yeo and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Sara Osman and Dimitris Typas, in collaboration with the Nature Structural & Molecular Biology team. Peer reviewer reports are available.

Additional information

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

Extended data

Extended Data Fig. 1 Validation of 4-OHT-mediated loss of Mettl3 and additional stress responses.

a, Mettl3 is depleted with high efficiency in Mettl3f/f MEFs cells treated with 4-OHT for 48 h. First lane shows Mettl3 in cells untreated with 4-OHT; Second lane shows Mettl3 in the same cells 7 days after treatment with 4-OHT (500 nM). The middle lane is blank. The lasts two lanes show protein collected from cells passaged for two weeks after 4-OHT treatment. All subsequent experiments were performed between one and two weeks after 4-OHT treatment. b, m6A was measured using thin-layer chromatography of twice-polyA-purified RNA extracted from MEFs on 8 days after treatment with 4-OHT. Control mRNA, left; mRNA from Mettl3 knockout cells (M3KO), right. Dotted circle shows the position at which m6A migrates, demonstrating a near-complete loss of m6A and METTL3 protein. c, 4-OHT treatment results in a substantial reduction in Mettl3 mRNA expression relative to control. Reads per kilobase per million reads (RPKM) was calculated for Mettl3 in samples from control and Mettl3 KO cells. Total RNA-seq reads mapping to Mettl3 mRNA are reduced substantially. Two biological replicates per condition. d, Average number of unique reads per base was calculated for Mettl3 exon 4 (GRCm38.p6; ENSMUSE00001262239; Chr14:52,298,640−52,298,815) which is deleted after Cre activation. Reads mapping to exon 4 were reduced by approximately 95%. Exon 4 codes for a zinc-finger domain required for methylation activity23. Two biological replicates per condition. e, Control (blue) shows a substantially higher total number of mapped reads relative to Mettl3 KO (red). The dotted lines delineate the boundaries of exon 4 within Mettl3. Scissors signify Cre-mediated removal of this region after treatment with 4-OHT in Mettl3 KO cells. We find a substantial loss of Mettl3 reads in Mettl3 KO cells and the loss of the deleted region in exon 4. Two biological replicates per condition. f, Response to sorbitol (400 mM, 2 h, 37 °C) and thapsigargin (1 µM, 1 h, 37 °C) in control MEFs. Ythdf2, red; Tiar, green, DAPI, blue. G, Response to sorbitol and thapsigargin stress in Mettl3 KO cells. Conditions are as described in f.

Source data

Extended Data Fig. 2 Ythdf2 and Tiar partition efficiency into stress granules.

a, Schematic depicting alternative method of collecting background fluorescence for partition coefficients. Since the region used for calculating the cytoplasmic intensities in Fig. 2a,d could have placement bias, we integrated total cytoplasmic fluorescence using average pixel intensity. The sampled region (dotted line, bottom), nuclear region (shaded gray and outlined, based on DAPI), stress granule (shaded gray, dotted outline, based on Tiar), and all other stress granules targeted for analysis are excluded from the cytoplasm. Stress granule intensity averages were measured relative to the integrated cytoplasmic average for each cell to obtain partition coefficients (see b). b, Alternative partition coefficient analysis, corresponding to schematic in a. Partition coefficient analysis was performed using the identical set of granules analyzed in the ‘arsenite’ condition in Fig. 2c. Boxplots depict the median (heavy line), upper and lower quartile medians (box bounds), and data maxima and minima (whiskers). Open circles depict individual measurements for stress granules. Control: n = 83, Mettl3 KO: n = 70. Two-sided student’s t-test, p < 0.05, with no adjustments. c, Comparison methods depicted in Fig. 2a and a. Ythdf2 partitioning measured in 150+ total stress granules (control: n = 83, Mettl3 KO: n = 70) using both methods. The ratio between partition coefficients in the ‘localized’ background measurements (Fig. 2a) and the partition coefficients in the ‘cytoplasmic’ background measurements (a) are shown on the y-axis. ‘1’ indicates background intensity values of both methods were equal, >1 indicates the localized method resulted in a higher partition coefficient estimate, <1 indicates the localized method underestimated the partition coefficient relative to the cytoplasmic method. Boxplots: median (heavy line), upper and lower quartile medians (box bounds), and data maxima and minima (whiskers). Generally, the localized method slightly underestimates partition coefficients, and this effect is stronger in control cells. d-f, Changes in fluorescence intensity values in stress granules after heat shock, thapsigargin, and sorbitol. The results are similar to measured average partition coefficients for entire granules in Fig. 2c. Tiar (green) demonstrates bounds of analyzed granules. Error bars, s.e.m. d, heat shock (control, n = 12 granules; Mettl3 KO, n = 11); e, thapsigargin (control, n = 10; Mettl3 KO, n = 11); f, sorbitol (control, n = 13; Mettl3 KO, n = 14).

Source data

Extended Data Fig. 3 Effects of m6A loss on total and stress granule mRNA relative abundance.

a, Fig. 3a presented as boxplots, median (heavy line), upper and lower quartile medians (box), interquartile range (whiskers, IQR = 1.5). Outliers plotted separately. n = 15110 transcripts; 0 m6A sites, n = 9832; 1 m6A site, n = 2807; 2 m6A sites, n = 1326; 3 m6A sites, n = 570; 4 + m6A sites, n = 575. Note, plots are not normalized to increased RNA abundance in Mettl3 KO. b, Fig. 3b presented as boxplots as in a. n = 14613 transcripts; 0 m6A sites, n = 9335; 1 m6A site, n = 2806; 2 m6A sites, n = 1326; 3 m6A sites, n = 571; 4 + m6A sites, n = 575. non-m6A-containing mRNAs are relatively unchanged between control and Mettl3 KO, while m6A-containing transcripts are substantially reduced in control due to the destabilizing effects of m6A. c, Boxplot of log2-fold change for mRNAs in SG relative to total mRNAs in control cells. mRNA levels in SG are normalized, which considers the expression of each mRNA in control cells. Figure 3c presented as boxplots as in a. n = 14638 transcripts; 0 m6A sites, n = 9362; 1 m6A site, n = 2806; 2 m6A sites, n = 1326; 3 m6A sites, n = 571; 4 + m6A sites, n = 573. m6A-containing transcripts are enriched in arsenite-induced SG. d, Boxplot of log2-fold change for mRNAs in SG normalized to total mRNAs in Mettl3 knockout. Figure 3d presented as boxplots as in a. n = 14877 transcripts; 0 m6A sites, n = 9594; 1 m6A site, n = 2809; 2 m6A sites, n = 1328; 3 m6A sites, n = 570; 4 + m6A sites, n = 576. When normalized to transcript expression levels, m6A-mRNAs no longer exhibit enrichment in arsenite-induced SG in KO. e, Cumulative distribution plot of log2-fold change for transcript abundance in arsenite-induced SG. Far left, expression is not normalized to cellular transcript expression. After adjustment for increased m6A-mRNA expression in Mettl3 knockout (second plot), the final two plots show the m6A-mRNA enrichment in third plot is lost in Mettl3 knockout. Annotation for m6A was GSM230043010 compared to GSE6199524 for Fig. 3a-d. Either dataset showed similar results. f, Same as e, except m6A annotation from transcripts with m6A sites in GSM2300430 that intersect with m6A peaks in GSE6199510,24. Again, similar results with e and Fig. 3a-d. g, Heatmap of m6A transcript proportions in mRNAs enriched and depleted from SGs. Far left is from ref. 5, others are from ref. 6 using their definitions of enrichment/depletion. Ref. 6 described groups with increasing enrichment. h, qPCR of transcripts in SG. Expression values were normalized to the ratio of difference in expression from WT and KO SG relative to that in WT and control total mRNA (see Methods). The negative effect on stress granule enrichment in Mettl3 KO is increasingly pronounced in transcripts with more m6A. Horizontal line, mean log2-fold change. Whiskers, SEM. n = 3 biological replicates.

Extended Data Fig. 4 YTHDF proteins CLIP to stress granule-enriched transcripts.

a, G3BP1 and TIAR show weak correlation between mRNA binding and the level of mRNA enrichment in stress granules. The x-axis indicates the density of eCLIP reads for either G3BP1 or TIAR for each mRNA. The y-axis shows the log2-fold change for mRNA abundance in stress granule mRNAs relative to total mRNAs generated in a previous stress granule transcriptome in U2OS cells5. Both G3BP1 and TIAR show that mRNAs exhibit a relatively small correlation between CLIP read density and stress granule enrichment. n = 7878 transcripts. b, YTHDF proteins show stronger correlations between mRNA binding and the level of mRNA enrichment in stress granules. YTHDF binding and the correlation with stress granule mRNA enrichment was calculated as in c. All the YTHDF proteins show a stronger correlation between stress granule enrichment and CLIP read density compared to G3BP1 and TIAR, with the strongest correlation for YTHDF2. n = 7878 transcripts.

Extended Data Fig. 5 Effects of m6A loss on stress granule enrichment and model of exon structure in stress granule-enriched mRNAs.

a, Boxplot of log2-fold change for mRNA abundance in SG from Fig. 4c. Median (heavy line), upper/lower quartile medians (box), interquartile range (whiskers, IQR = 1.5). Outliers plotted separately. n = 13312 total transcripts; 0-1 kb, n = 848; 1-1.5 kb, n = 1280; 1.5-2 kb, n = 1601; 2-2.5 kb, n = 1512; 2.5-3 kb, n = 1526; 3-4 kb, n = 2391; 4-5 kb, n = 1696; 5-6 kb, n = 903; 6+ kb, n = 1554. b, as in a, for data from Fig. 4d. n = 13512 total transcripts; 0-1 kb, n = 839; 1-1.5 kb, n = 1296; 1.5-2 kb, n = 1620; 2-2.5 kb, n = 1537; 2.5-3 kb, n = 1556; 3-4 kb, n = 2448; 4-5 kb, n = 1720; 5-6 kb, n = 914; 6+ kb, n = 1582. c, Correlation of length and mRNA stress granule enrichment in U2OS cells after arsenite. Transcript length and log2-fold change data from a ref. 5. r measured using the correlation test. d, As in c, except NIH3T3 cells after arsenite treatment from ref. 6. e,f, As in c, except control MEFs (e) and Mettl3 KO MEFs (f). This shows that the length effect is dependent on m6A. g, Mean internal exon lengths in SG-enriched (1384 transcripts, >0.5 log2-fold change) and SG-depleted transcripts (1495 transcripts, <-0.5log2-fold change) used to calculate Fig. 4f. Transcripts were ordered by length and split into equal-sized bins. The difference is plotted in black or red in h and Fig. 4f. SD and maximum ranges in Fig. 4f based on sampling internal exons randomly 1000 times (see Methods). h, As in Fig. 4h but zoomed in on the y-axis. None of the shorter exon bins reached a high threshold for significance according to the two-tailed Z-test (p < 0.0001). i, Z-scores for differences in mean internal exon lengths for SG-enriched and SG-depleted transcripts. Z-scores were calculated for each bin relative to the theoretical standard deviation of randomly distributed internal exons sampled from the transcriptome as described in g and in Methods. The difference in mean exon lengths in the two bins containing the largest exons (far left, red dots) resulted in a substantially higher Z-score than all other bins and had a high statistical significance (p < 0.0001, asterisk) relative to all other bins. This indicates that long internal exons in SG-enriched exons are significantly longer than SG-depleted exons.

Extended Data Fig. 6 m6A loss leads to slight preference for proximal poly(A) site selection, but does not substantially shorten mRNAs.

a, m6A correlates with a slight preference for distal polyadenylation site (PAS) usage. DaPars40 provides a percentage that predicts proximal (negative values) or distal (positive values) shifts in PAS usage relative to another condition. Each dot represents a unique transcript isoform. Best-fit line (blue) and standard deviation (light gray). P-values: Fisher’s exact test comparing control to Mettl3 KO transcripts. Benjamini-Hochberg adjustment (false discovery rate=5%) applied to all p-values before log transformation. Transcripts are assigned to the ‘proximal’ (6388 transcripts) and ‘distal’ (8425 transcripts) group if the change in site usage (that is, the relative use of proximal vs. distal PAS for all the transcript copies from a gene) is >5% in control relative to Mettl3 KO. Transcripts with <5% change are not assigned (1780 transcripts). b,c, We wanted to determine how overall length of transcripts could be affected by selection of distal or proximal PAS. Boxplot (b) depicts the length change between proximal and distal PAS usage in all transcripts (n = 16593) depicted in a. Boxplot depicts the median (heavy line), upper and lower quartile medians (box), and interquartile range (whiskers, IQR = 1.5). Outliers plotted separately. The median length change is ~272 nt. ~70% of alternative PAS will cause a length change of <0.5 kb. ~16% have an alternative PAS that will lead to a change in transcript length >1 kb. d, Proximal or distal PAS preferences do not substantially alter the length-dependent effect of mRNA enrichment in control stress granules. Longest transcript isoform is plotted on the y-axis. The effect is quantified for both ‘proximal APA’ transcripts (1960 transcripts) and ‘distal APA’ transcripts (2612 transcripts) as determined in a. Length-dependent enrichment is seen irrespective of changes in APA sites. e, Proximal or distal polyadenylation site preferences do not substantially alter the length-dependent effect of mRNA enrichment in Mettl3 KO SG. Analysis is performed as in d, except lengths are plotted against the enrichment of mRNAs in Mettl3 KO stress granules. In contrast to d, no clear relationship can be observed between SG enrichment for transcripts that show either distal or proximal APA site shifts in control cells. The fit line was measured using the correlation test (proximal, r = 0.14, p < 4.8×10-10; distal, r = -0.07, p < 7.7×10-).

Extended Data Fig. 7 Analysis of mRNA enrichment using short isoforms.

a-b, Cumulative distribution plots of differential expression using shortest isoforms. Data analysis and interpretation is essentially the same as Fig. 4c,d. c-d, Box plots of differential expression using shortest isoforms. Data analysis and interpretation is essentially the same as Extended Data Fig. 5a,b. Boxplots depict the median (heavy line), upper and lower quartile medians (box), and interquartile range (whiskers, IQR = 1.5). Outliers are plotted separately. For c, control: n = 13311 transcripts; 0-0.5 kb, n = 1450; 0-1 kb, n = 2889; 1-1.5 kb, n = 1404; 1.5-2 kb, n = 1475; 2-2.5 kb, n = 1246; 2.5-3 kb, n = 1121; 3-4 kb, n = 1593; 4-5 kb, n = 933; 5-6 kb, n = 477; 6+ kb, n = 723) and for d, Mettl3 KO (n = 13511 transcripts; 0-0.5 kb, n = 1449; 0-1 kb, n = 2923; 1-1.5 kb, n = 1422; 1.5-2 kb, n = 1500; 2-2.5 kb, n = 1270; 2.5-3 kb, n = 1148; 3-4 kb, n = 1627; 4-5 kb, n = 949; 5-6 kb, n = 485; 6+ kb, n = 739).

Extended Data Fig. 8 Re-analysis of the CDS and 3’UTR effect on stress granule enrichment in U2OS cells.

a, Comparison of CDS lengths used in Khong et al. 2017 with re-annotated CDS lengths from the longest transcript isoforms in GRCh37.p13. Left, CDS lengths used in the original annotation provided by Khong et al. 2017. Right, CDS lengths retrieved for the longest mRNA isoforms from GRCh37.p13. Boxplots depict the median (heavy line), upper and lower quartile medians (box), and interquartile range (whiskers, IQR = 1.5). Outliers plotted separately. SG-enriched mRNAs defined as >1.0 log2-fold change or greater (Khong et al. 2017, n = 1627; longest isoform, n = 1616). SG-depleted mRNAs defined as <-1.0 log2-fold change (Khong et al. 2017, n = 1780; longest isoform, n = 1762). mRNAs between -1.0 and 1.0 log2-fold change are grouped as ‘neither’ (Khong et al. 2017, n = 7789; longest isoform, n = 7733). Re-analysis showed discrepancies of the region transcript length sums (5’ UTR + CDS + 3’ UTR) relative to the total lengths listed in the supplemental data. Since the dataset in Khong et al. 2017 does not contain transcript identifiers, the source of the discrepancy in lengths could not be determined. However, the re-annotation has a limited effect on CDS transcript lengths. b, Since the effects seen in a are relatively small, there was very little effect on the correlation between CDS and mRNA enrichment in stress granules in the original annotation (top panel, r = 0.63, p < 2.2 ×10-16) and the re-annotation (bottom panel, r = 0.60, p < 2.2 ×10-16) according to the correlation test. c, Comparison of 3’UTR lengths used in Khong et al. 2017 with re-annotated 3’UTR lengths from the longest transcript isoforms in GRCh37.p13. Boxplots prepared as in a. Outliers plotted separately. Left, 3’UTR lengths used in the original annotation provided by Khong et al. 2017 (depleted, n = 1780; neither, n = 7789; enriched, n = 1627). Right, 3’UTR lengths retrieved for the longest mRNA isoforms from GRCh37.p13 (depleted, n = 1762; neither, n = 7733; enriched, n = 1616). This resulted in a substantial increase in the length of 3’UTRs in each category. d, The general increase in the 3’UTR length had a substantial effect on the correlation between 3’UTR length and mRNA enrichment in stress granules when comparing the original annotation (top panel, r = 0.22, p < 2.2 ×10-16) and the re-annotation (bottom panel, r = 0.38, p < 2.2 ×10-16) according to the correlation test.

Extended Data Fig. 9 Alternative RPKM comparison of stress granule enrichment of mRNAs based on length.

a, Comparison of the percentage of reads per kilobase per million (RPKM) between stress granule and total mRNA samples grouped by transcript length. Average RPKM values for mRNAs containing m6A sites were calculated and compared across samples. The percentages are approximated to the last digit given. b, Percentage change between stress granule and total mRNA RPKM in control and Mettl3 KO grouped by mapped m6A sites. The data shows the percentage differences depicted between each of the categories presented in a. The percentages are approximated to the hundredth. In control MEFs, a length-dependent enrichment effect is observed for all mRNAs longer than 1.5 kb. In Mettl3 KO MEFs, only mRNAs that are very short ( < 1.5 kb) or very long ( > 5 kb) are enriched in stress granules. Notably, shorter mRNAs in the 0-1.5 kb group comprise 50-60% of total overall reads, indicating the abundance of these transcripts. These reads are greatly de-enriched in control stress granules. In contrast, these abundant, short transcripts are enriched in Mettl3 KO stress granules. c, Estimating the effects of m6A on the stress granule transcriptome. Blue arrowed lines depict observations in control MEFs, while red arrowed lines represent observations during in Mettl3 KO. In the left panel, the m6A-dependent effect on stress granule enrichment is shown. In cells with Mettl3, m6A-DF interactions exert a strong effect on mRNA recruitment. The m6A effect is absent in cells lacking Mettl3. In the right panel, the length effect on stress granule enrichment is shown. Under standard conditions, long RNAs interact with one another through RNA-RNA interactions and multivalent DF-m6A effects. In the absence of m6A, shorter mRNAs enter stress granules. Some longer mRNAs are affected to a lesser degree by loss of m6A owing to compensatory RNA-RNA or other RNA-protein interactions.

Extended Data Fig. 10 mRNA enrichment as a function of transcript length and m6A in NIH3T3 cells.

a, Boxplots demonstrating the correlation between m6A levels and stress granule enrichment in NIH3T3 cells. The y-axis shows the level of enrichment of transcripts in stress granules relative to the cytoplasm. The x-axis shows the number of m6A sites per transcript. Boxplots depict the median (heavy line), upper and lower quartile medians (box bounds), and interquartile range (whiskers, IQR = 1.5) Each dot represents a unique individual transcript. n = 8301 transcripts; 0 m6A sites, n = 5607; 1 m6A site, n = 1015; 2 m6A sites, n = 1023; 3 m6A sites, n = 300; 4 + m6A sites, n = 356. At the top of each boxplot the range of transcript lengths contained in the plot below is shown. Transcript size increases from left to right. Generally, regardless of transcript length, m6A has a cumulative effect on the likelihood of enrichment in stress granules. b, Scatter plot demonstrating the correlation between m6A levels and stress granule enrichment in NIH3T3 cells. The y-axis shows the relative level of enrichment of transcripts in stress granules relative to the cytoplasm. The x-axis shows the length of individual transcripts on a logarithmic scale. The number of m6A sites in each individual transcript is color-coded as shown (yellow=0, light green=1, blue-green=2, blue=3, violet=4 + ). Dotted lines are shown connected to the boxplots in panel a that summarize the scatter data per m6A site. As can be seen, the highly methylated mRNAs of any mRNA length exhibit higher stress granule enrichment.

Supplementary information

Reporting Summary

Peer Review File

Supplementary Tables 1–3

Supplementary Table 1: Differential expression analysis of RNA-seq data from control and Mettl3-KO cells. Analysis was performed using the DESeq2 package in R52. RPKM values provided for each condition are the average of two biological replicates. Short and long transcript isoform lengths were obtained from Ensembl GRCm38.75 (mm10)50. m6A site counts were obtained from m6A-seq data for E13.5 MEFs (GSE61995)22 and mESCs (GSM2300430)18. Supplementary Table 2: APA site usage analysis of RNA-seq data from control and Mettl3-KO cells. Analysis was performed using the DaPars algorithm in Python41. Raw RNA-seq counts from two biological replicates were used as input. Supplementary Table 3: The m6A-dependent ΔFCE of transcripts in Mettl3- and m6A-depleted cells relative to control cells. A positive ΔFCE indicates lower dependence on m6A for stress granule enrichment; a negative ΔFCE indicates higher dependence on m6A for stress granule enrichment.

Source data

Source Data Fig. 2

Statistical source data for Fig. 2b,c,e.

Source Data Extended Data Fig. 1

Unprocessed western blot

Source Data Extended Data Fig. 2

Statistical source data

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ries, R.J., Pickering, B.F., Poh, H.X. et al. m6A governs length-dependent enrichment of mRNAs in stress granules. Nat Struct Mol Biol 30, 1525–1535 (2023). https://doi.org/10.1038/s41594-023-01089-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41594-023-01089-2

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing