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.

eIF5B and eIF1A reorient initiator tRNA to allow ribosomal subunit joining

Abstract

Translation initiation defines the identity and quantity of a synthesized protein. The process is dysregulated in many human diseases1,2. A key commitment step is when the ribosomal subunits join at a translation start site on a messenger RNA to form a functional ribosome. Here, we combined single-molecule spectroscopy and structural methods using an in vitro reconstituted system to examine how the human ribosomal subunits join. Single-molecule fluorescence revealed when the universally conserved eukaryotic initiation factors eIF1A and eIF5B associate with and depart from initiation complexes. Guided by single-molecule dynamics, we visualized initiation complexes that contained both eIF1A and eIF5B using single-particle cryo-electron microscopy. The resulting structure revealed how eukaryote-specific contacts between the two proteins remodel the initiation complex to orient the initiator aminoacyl-tRNA in a conformation compatible with ribosomal subunit joining. Collectively, our findings provide a quantitative and architectural framework for the molecular choreography orchestrated by eIF1A and eIF5B during translation initiation in humans.

This is a preview of subscription content

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Fig. 1: Real-time analysis of eIF5B-mediated ribosomal subunit joining in humans.
Fig. 2: eIF1A resides on initiation complexes until the 60S subunit joins.
Fig. 3: Cryo-EM structure of an eIF1A and eIF5B-bound initiation complex.
Fig. 4: eIF5B contacts with eIF1A and \({{\bf{Met-tRNA}}}_{{\bf{i}}}^{{\bf{M}}{\bf{e}}{\bf{t}}}\) mediate 60S subunit joining.

Data availability

Raw files are available upon request to J.D.P. The cryo-EM map and final model of the late 40S initiation complex have been deposited in the Electron Microscopy Data Bank under accession EMD-26067 and Protein Data Bank under accession 7TQL, respectively.

Code availability

All custom codes used in this study to process and analyse single-molecule data are available publicly under an open-source license at: https://github.com/puglisilab/Lapointe-2022-Nature.

References

  1. Tahmasebi, S., Khoutorsky, A., Mathews, M. B. & Sonenberg, N. Translation deregulation in human disease. Nat. Rev. Mol. Cell Biol. 19, 791–807 (2018).

    CAS  PubMed  Google Scholar 

  2. Robichaud, N., Sonenberg, N., Ruggero, D. & Schneider, R. J. Translational control in cancer. Cold Spring Harb. Perspect. Biol. 11, a032896 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  3. Pain, V. M. Initiation of protein synthesis in eukaryotic cells. Eur. J. Biochem. 236, 747–771 (1996).

    CAS  PubMed  Google Scholar 

  4. Merrick, W. C. & Pavitt, G. D. Protein synthesis initiation in eukaryotic cells. Cold Spring Harb. Perspect. Biol. 10, a033092 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. Hershey, J. W. B., Sonenberg, N. & Mathews, M. B. Principles of translational control. Cold Spring Harb. Perspect. Biol. 11, a032607 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  6. Aitken, C. E. & Lorsch, J. R. A mechanistic overview of translation initiation in eukaryotes. Nat. Struct. Mol. Biol. 19, 568–576 (2012).

    CAS  PubMed  Google Scholar 

  7. Sokabe, M. & Fraser, C. S. Toward a kinetic understanding of eukaryotic translation. Cold Spring Harb. Perspect. Biol. 11, a032706 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  8. Hinnebusch, A. G. The scanning mechanism of eukaryotic translation initiation. Annu. Rev. Biochem. 83, 779–812 (2014).

    CAS  PubMed  Google Scholar 

  9. Hinnebusch, A. G. Structural Insights into the mechanism of scanning and start codon recognition in eukaryotic translation initiation. Trends Biochem. Sci. 42, 589–611 (2017).

    CAS  PubMed  Google Scholar 

  10. Hashem, Y. & Frank, J. The jigsaw puzzle of mRNA translation initiation in eukaryotes: a decade of structures unraveling the mechanics of the process. Annu. Rev. Biophys. 47, 125–151 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  11. Pestova, T. V. et al. The joining of ribosomal subunits in eukaryotes requires eIF5B. Nature 403, 332–335 (2000).

    ADS  CAS  PubMed  Google Scholar 

  12. Lee, J. H. et al. Initiation factor eIF5B catalyzes second GTP-dependent step in eukaryotic translation initiation. Proc. Natl Acad. Sci. USA 99, 16689–16694 (2002).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  13. Shin, B. S. et al. Uncoupling of initiation factor eIF5B/IF2 GTPase and translational activities by mutations that lower ribosome affinity. Cell 111, 1015–1025 (2002).

    CAS  PubMed  Google Scholar 

  14. Wang, J. et al. eIF5B gates the transition from translation initiation to elongation. Nature 573, 605–608 (2019).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  15. Fernández, I. S. et al. Molecular architecture of a eukaryotic translational initiation complex. Science 342, 1240585 (2013).

    PubMed  Google Scholar 

  16. Wang, J. et al. Structural basis for the transition from translation initiation to elongation by an 80S–eIF5B complex. Nat. Commun. 11, 5003 (2020).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  17. Yamamoto, H. et al. Structure of the mammalian 80S initiation complex with initiation factor 5B on HCV-IRES RNA. Nat. Struct. Mol. Biol. 21, 721–727 (2014).

    CAS  PubMed  Google Scholar 

  18. Choi, S. K. et al. Physical and functional interaction between the eukaryotic orthologs of prokaryotic translation initiation factors IF1 and IF2. Mol. Cell. Biol. 20, 7183–7191 (2000).

    CAS  PubMed  PubMed Central  Google Scholar 

  19. Olsen, D. S. Domains of eIF1A that mediate binding to eIF2, eIF3 and eIF5B and promote ternary complex recruitment in vivo. EMBO J. 22, 193–204 (2003).

    CAS  PubMed  PubMed Central  Google Scholar 

  20. Acker, M. G., Shin, B.-S., Dever, T. E. & Lorsch, J. R. Interaction between eukaryotic initiation factors 1A and 5B is required for efficient ribosomal subunit joining. J. Biol. Chem. 281, 8469–8475 (2006).

    CAS  PubMed  Google Scholar 

  21. Fringer, J. M., Acker, M. G., Fekete, C. A., Lorsch, J. R. & Dever, T. E. Coupled release of eukaryotic translation initiation factors 5B and 1A from 80S ribosomes following subunit joining. Mol. Cell. Biol. 27, 2384–2397 (2007).

    CAS  PubMed  PubMed Central  Google Scholar 

  22. Pisareva, V. P. & Pisarev, A. V. eIF5 and eIF5B together stimulate 48S initiation complex formation during ribosomal scanning. Nucleic Acids Res. 42, 12052–12069 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  23. Chen, J. et al. High-throughput platform for real-time monitoring of biological processes by multicolor single-molecule fluorescence. Proc. Natl Acad. Sci. USA 111, 664–669 (2014).

    ADS  CAS  PubMed  Google Scholar 

  24. Lapointe, C. P. et al. Dynamic competition between SARS-CoV-2 NSP1 and mRNA on the human ribosome inhibits translation initiation. Proc. Natl Acad. Sci. USA 118, e2017715118 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

  25. Sokabe, M. & Fraser, C. S. Human eukaryotic initiation factor 2 (eIF2)–GTP–Met–tRNAi ternary complex and eIF3 stabilize the 43 S preinitiation complex. J. Biol. Chem. 289, 31827–31836 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  26. Koch, A., Aguilera, L., Morisaki, T., Munsky, B. & Stasevich, T. J. Quantifying the dynamics of IRES and cap translation with single-molecule resolution in live cells. Nat. Struct. Mol. Biol. 27, 1095–1104 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  27. Acker, M. G. et al. Kinetic analysis of late steps of eukaryotic translation initiation. J. Mol. Biol. 385, 491–506 (2009).

    CAS  PubMed  Google Scholar 

  28. Huang, B. Y. & Fernández, I. S. Long-range interdomain communications in eIF5B regulate GTP hydrolysis and translation initiation. Proc. Natl Acad. Sci. USA 117, 1429–1437 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  29. Passmore, L. A. et al. The eukaryotic translation initiation factors eIF1 and eIF1A induce an open conformation of the 40S ribosome. Mol. Cell 26, 41–50 (2007).

    CAS  PubMed  Google Scholar 

  30. Hashem, Y. et al. Structure of the mammalian ribosomal 43S preinitiation complex bound to the scanning factor DHX29. Cell 153, 1108–1119 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  31. Lomakin, I. B. & Steitz, T. A. The initiation of mammalian protein synthesis and mRNA scanning mechanism. Nature 500, 307–311 (2013).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  32. Weisser, M., Voigts-Hoffmann, F., Rabl, J., Leibundgut, M. & Ban, N. The crystal structure of the eukaryotic 40S ribosomal subunit in complex with eIF1 and eIF1A. Nat. Struct. Mol. Biol. 20, 1015–1017 (2013).

    CAS  PubMed  Google Scholar 

  33. Hussain, T. et al. Structural changes enable start codon recognition by the eukaryotic translation initiation complex. Cell 159, 597–607 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  34. Des Georges, A. et al. Structure of mammalian eIF3 in the context of the 43S preinitiation complex. Nature 525, 491–495 (2015).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  35. Simonetti, A. et al. eIF3 peripheral subunits rearrangement after mRNA binding and start-codon recognition. Mol. Cell 63, 206–217 (2016).

    CAS  PubMed  Google Scholar 

  36. Llácer, J. L. et al. Translational initiation factor eIF5 replaces eIF1 on the 40S ribosomal subunit to promote start-codon recognition. eLife 7, e39273 (2018).

    PubMed  PubMed Central  Google Scholar 

  37. Brito Querido, J. et al. Structure of a human 48S translational initiation complex. Science 369, 1220–1227 (2020).

    ADS  CAS  PubMed  Google Scholar 

  38. Simonetti, A., Guca, E., Bochler, A., Kuhn, L. & Hashem, Y. Structural insights into the mammalian late-stage initiation complexes. Cell Rep. 31, 107497–107497 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  39. Kratzat, H. et al. A structural inventory of native ribosomal ABCE1–43S pre-initiation complexes. EMBO J. 40, e105179 (2021).

    CAS  PubMed  Google Scholar 

  40. Nag, N. et al. eIF1A/eIF5B interaction network and its functions in translation initiation complex assembly and remodeling. Nucleic Acids Res. 44, 7441–7456 (2016) .

    CAS  PubMed  PubMed Central  Google Scholar 

  41. Marintchev, A., Kolupaeva, V. G., Pestova, T. V. & Wagner, G. Mapping the binding interface between human eukaryotic initiation factors 1A and 5B: A new interaction between old partners. Proc. Natl Acad. Sci. USA 100, 1535–1540 (2003).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  42. Shao, S. et al. Decoding mammalian ribosome-mrna states by translational GTPase complexes. Cell 167, 1229–1240.e1215 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  43. Pisarev, A. V. et al. Specific functional interactions of nucleotides at key −3 and +4 positions flanking the initiation codon with components of the mammalian 48S translation initiation complex. Genes Dev. 20, 624–636 (2006).

    CAS  PubMed  PubMed Central  Google Scholar 

  44. Suresh, S. et al. eIF5B drives integrated stress response-dependent translation of PD-L1 in lung cancer. Nat. Cancer 1, 533–545 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  45. Roll-Mecak, A., Cao, C., Dever, T. E. & Burley, S. K. X-ray structures of the universal translation initiation factor IF2/eIF5B. Cell 103, 781–792 (2000).

    CAS  PubMed  Google Scholar 

  46. Kuhle, B. & Ficner, R. eIF5B employs a novel domain release mechanism to catalyze ribosomal subunit joining. EMBO J. 33, 1177–1191 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  47. Zheng, A. et al. X-ray structures of eIF5B and the eIF5B–eIF1A complex: the conformational flexibility of eIF5B is restricted on the ribosome by interaction with eIF1A. Acta Crystallogr. D 70, 3090–3098 (2014).

    CAS  PubMed  Google Scholar 

  48. Kaledhonkar, S. et al. Late steps in bacterial translation initiation visualized using time-resolved cryo-EM. Nature 570, 400–404 (2019).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  49. Marshall, R. A., Aitken, C. E. & Puglisi, J. D. GTP hydrolysis by IF2 guides progression of the ribosome into elongation. Mol. Cell 35, 37–47 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

  50. Ling, C. & Ermolenko, D. N. Initiation factor 2 stabilizes the ribosome in a semirotated conformation. Proc. Natl Acad. Sci. USA 112, 15874–15879 (2015).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  51. Sprink, T. et al. Structures of ribosome-bound initiation factor 2 reveal the mechanism of subunit association. Sci. Adv. 2, e1501502 (2016).

    ADS  PubMed  PubMed Central  Google Scholar 

  52. Johnson, A. G. et al. RACK1 on and off the ribosome. RNA 25, 881–895 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  53. Yin, J., Lin, A. J., Golan, D. E. & Walsh, C. T. Site-specific protein labeling by Sfp phosphopantetheinyl transferase. Nat. Protoc. 1, 280–285 (2006).

    CAS  PubMed  Google Scholar 

  54. Fraser, C. S., Berry, K. E., Hershey, J. W. B. & Doudna, J. A. eIF3j is located in the decoding center of the human 40S ribosomal subunit. Mol. Cell 26, 811–819 (2007).

    CAS  PubMed  Google Scholar 

  55. Özeş, A. R., Feoktistova, K., Avanzino, B. C. & Fraser, C. S. Duplex unwinding and ATPase activities of the DEAD-box helicase eIF4A are coupled by eIF4G and eIF4B. J. Mol. Biol. 412, 674–687 (2011).

    PubMed  PubMed Central  Google Scholar 

  56. Feoktistova, K., Tuvshintogs, E., Do, A. & Fraser, C. S. Human eIF4E promotes mRNA restructuring by stimulating eIF4A helicase activity. Proc. Natl Acad. Sci. USA 110, 13339–13344 (2013).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  57. Sokabe, M., Fraser, C. S. & Hershey, J. W. The human translation initiation multi-factor complex promotes methionyl-tRNAi binding to the 40S ribosomal subunit. Nucleic Acids Res. 40, 905–913 (2012).

    CAS  PubMed  Google Scholar 

  58. Damoc, E. et al. Structural characterization of the human eukaryotic initiation factor 3 protein complex by mass spectrometry. Mol. Cell. Proteomics 6, 1135–1146 (2007).

    CAS  PubMed  Google Scholar 

  59. Korlach, J. et al. Selective aluminum passivation for targeted immobilization of single DNA polymerase molecules in zero-mode waveguide nanostructures. Proc. Natl Acad. Sci. USA 105, 1176–1181 (2008).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  60. Aitken, C. E., Marshall, R. A. & Puglisi, J. D. An oxygen scavenging system for improvement of dye stability in single-molecule fluorescence experiments. Biophys. J. 94, 1826–1835 (2008).

    CAS  PubMed  PubMed Central  Google Scholar 

  61. Erickson, F. L. & Hannig, E. M. Ligand interactions with eukaryotic translation initiation factor 2: role of the gamma-subunit. EMBO J. 15, 6311–6320 (1996).

    CAS  PubMed  PubMed Central  Google Scholar 

  62. Kapp, L. D. & Lorsch, J. R. GTP-dependent recognition of the methionine moiety on initiator tRNA by translation factor eIF2. J. Mol. Biol. 335, 923–936 (2004).

    CAS  PubMed  Google Scholar 

  63. Juette, M. F. et al. Single-molecule imaging of non-equilibrium molecular ensembles on the millisecond timescale. Nat. Methods 13, 341–344 (2016).

    PubMed  PubMed Central  Google Scholar 

  64. Bronson, J. E., Fei, J., Hofman, J. M., Gonzalez, R. L. & Wiggins, C. H. Learning rates and states from biophysical time series: a Bayesian approach to model selection and single-molecule FRET data. Biophys. J. 97, 3196–3205 (2009).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  65. Lawson, M. R. et al. Mechanisms that ensure speed and fidelity in eukaryotic translation termination. Science 373, 876–882 (2021).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  66. Zivanov, J., Nakane, T. & Scheres, S. H. W. Estimation of high-order aberrations and anisotropic magnification from cryo-EM data sets in. IUCrJ 7, 253–267 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  67. Rohou, A. & Grigorieff, N. CTFFIND4: fast and accurate defocus estimation from electron micrographs. J. Struct. Biol. 192, 216–221 (2015).

    PubMed  PubMed Central  Google Scholar 

  68. Casañal, A., Lohkamp, B. & Emsley, P. Current developments in Coot for macromolecular model building of electron cryo-microscopy and crystallographic data. Protein Sci. 29, 1069–1078 (2020).

    PubMed  PubMed Central  Google Scholar 

  69. Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  70. Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. D 74, 531–544 (2018).

    CAS  Google Scholar 

  71. Nicholls, R. A., Long, F. & Murshudov, G. N. Low-resolution refinement tools in REFMAC5. Acta Crystallogr. D 68, 404–417 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

  72. Williams, C. J. et al. MolProbity: more and better reference data for improved all-atom structure validation. Protein Sci. 27, 293–315 (2018).

    CAS  PubMed  Google Scholar 

  73. Goddard, T. D. et al. UCSF ChimeraX: meeting modern challenges in visualization and analysis. Protein Sci. 27, 14–25 (2018).

    CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We thank M. Lawson, J. Carette and other members of the Puglisi and Carette laboratories for helpful guidance, discussions and feedback; and P. Sarnow and the Sarnow laboratory for sharing cell culture equipment. Some of this work was performed at the Stanford-SLAC Cryo-EM Facilities, supported by Stanford University, SLAC and the National Institutes of Health S10 Instrumentation Programs. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. C.P.L. was a Damon Runyon Fellow supported by the Damon Runyon Cancer Research Foundation (DRG-#2321-18); C.A. was supported by a Stanford Bio-X Fellowship; and J.W. was supported by a postdoctoral scholarship from the Knut and Alice Wallenberg Foundation (KAW 2015.0406). This work was funded, in part, by the National Institutes of Health (GM011378 and AG064690 to J.D.P.; GM092927 to C.S.F.; and K99GM144678 to C.P.L.) and by the Intramural Research Program of the National Institutes of Health (to T.E.D.).

Author information

Authors and Affiliations

Authors

Contributions

Conceptualization: C.P.L., I.S.F. and J.D.P. Methodology: C.P.L., R.G., M.S., C.A., J.W., N.V., B.S.-S., E.M., T.E.D., C.S.F., I.S.F. and J.D.P. Resources: C.P.L., R.G., M.S., C.A., N.V. and C.S.F. Investigation: C.P.L., R.G., E.M. and I.S.F. Visualization: C.P.L., I.S.F. and J.D.P. Funding acquisition: T.E.D., C.S.F. and J.D.P. Project administration: T.E.D., C.S.F., I.S.F. and J.D.P. Supervision: T.E.D., C.S.F., I.S.F. and J.D.P. Writing, original draft: C.P.L., I.S.F. and J.D.P. Writing, review and editing: C.P.L., R.G., M.S., C.A., J.W., T.E.D., C.S.F., I.S.F. and J.D.P.

Corresponding authors

Correspondence to Israel S. Fernández or Joseph D. Puglisi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature thanks Katrin Karbstein, Assen Marintchev and Meni Wanunu for their contribution to the peer review of this work. Peer review 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 figures and tables

Extended Data Fig. 1 Real-time monitoring of human ribosomal subunit recruitment to β-globin mRNA.

a. Human 40S and 60S ribosomal subunits were labeled with fluorescent dyes on the N-terminus of uS19 or C-terminus of uL18, respectively. These labeling positions yield an inter-subunit FRET signal upon formation of translation-competent 80S ribosomes. The structural model was obtained from PDB: 4UG0. Of note, the first nine N-terminal amino acids of uS19 and last two C-terminal acids of uL18 were unresolved in the structure, which also does not account for the fused ybbR tags (11 amino acids). Thus, the reported distance between the 40S and 60S labeling sites is only an approximation. b. A real-time single molecule fluorescence assay using zero-mode waveguides (ZMWs) on a custom PacBio RSII DNA sequencer. Components of interest are tethered to the imaging surface within individual ZMWs, and reaction components are added directly to the surface. Time-resolved, real-time 4-color fluorescence emission was monitored across ~150,000 ZMWs after excitation with a 532 nm laser. The order and time elapsed between relevant fluorescence signals (e.g., green & orange) were determined in ~100–200 ZMWs with the desired signals. Rates of association and dissociation were determined using probability based statistical models; cumulative distribution functions of the observed times were calculated and subsequently fit to exponential functions to yield association or dissociation rates, as appropriate. c. Schematic of a real-time single-molecule assay to monitor ribosomal subunit joining on tethered β-globin mRNA (the ‘tethered mRNA’ setup). β-globin mRNA with a 5' m7G cap, poly(A)30 tail, and 3'-terminal biotin moiety was tethered to the neutravidin-coated ZMW imaging surface. Immediately after tethering, eIFs 4A, 4B, 4G, and 4E were added. After start of data acquisition and excitation via the 532 nm laser at 30 °C, the 43S PIC (5 nM via the labeled 40S-Cy3), unlabeled eIF5B (1 µM), 60S-Cy5 subunits (100 nM), and 1 mM of ATP and GTP were added. Unlabeled 43S PIC components were present at \(\ge \) 1.5-fold excess during each step of the experiment. Fluorescence data were acquired for 600 s. d. Cartoon schematic of theoretical single-molecule fluorescence data where 40S and 60S subunits were recruited to an mRNA to form 80S ribosomes. The 40S subunit association time (t40S) was defined as the time elapsed from addition of the 43S PIC until appearance of the Cy3 signal. The 60S subunit association time (t60S) was defined as the time elapsed from 40S subunit association until appearance of the 40S(Cy3,donor)-60S(Cy5,acceptor) FRET signal. e. Example single-molecule fluorescence trace with FRET efficiency (EFRET) plot. Recruitment of the 40S subunit (as the 43S PIC) was indicated by a burst of Cy3 (green) fluorescence intensity. 60S subunit joining was indicated by appearance of the 40S-60S FRET signal. f. Density maps of normalized 40S-Cy3 and 60S-Cy5 fluorescence intensities post-synced to 60S subunit joining (n = 128). As in the individual trace shown in panel D, joining of the 60S subunit led to an anti-correlated decrease in Cy3 and increase in Cy5 signals, indicative of intersubunit FRET when the 80S ribosome formed. g. Histogram and single gaussian function fit (line) of the observed EFRET in the real-time initiation assay. The mean EFRET (µ) was 0.65 ± 0.1 and the standard deviation (σ) was 0.17 ± 0.1 (n = 168), consistent with structural predictions for translation-competent 80S ribosomes. h. Table of 40S loading and 60S joining efficiencies (left) and a plot of observed 40S association times (right) in the indicated conditions. 40S subunit loading efficiency (80% labeling efficiency in all experiments) was defined as the fraction of 1,000 analyzed ZMWs with at least one stable (> 10 s) 40S subunit association event. The concentration of 40S subunits (5 nM) was optimized to yield a single recruitment event per ZMW, which is most probable when < 30% of ZMWs contain a recruitment event, as predicted by Poisson distribution statistics. 60S subunit joining efficiency (with errors propagated from 95% CIs) was defined the fraction of recruited 40S subunits (300 analyzed) with a 60S subunit joining event (indicated by Cy3-Cy5 FRET), which was normalized to account for relative 60S subunit labeling efficiencies (35 or 80% labeled). 40S association time (t40S) was defined as above (n = 135, 128, 131, 155, and 137, from left to right), and the median observed times are represented by the orange lines. i. Cumulative probability plot of 40S and 60S subunit association times and the fits to single-exponential functions, which yielded apparent association rates of kon,40S ≈ 0.049 ± 0.002 s−1 and kon,60S ≈ 0.033 ± 0.001 s−1 when added at 5 nM and 100 nM (final concentration), respectively (n = 128).

Extended Data Fig. 2 Real-time analysis of eIF5B-mediated ribosomal subunit joining in humans.

a. Fluorescence scans of SDS-PAGE gels that analyzed purified and fluorescently-labeled eIF5B proteins. ‘WT’ corresponds to the wild-type protein with the N-terminal domain truncated (residues 1-586 removed) and an N-terminal ybbR peptide tag (11 amino acids). ‘H706E’ corresponds to the GTPase deficient version of the protein. Each protein was analyzed by SDS-PAGE twice. The full scans of the cropped gel are available in Supplementary Figure 1. b. Plot of the relative light units (RLU) from nanoLuciferase in vitro translation assays where the indicated eIF5B proteins were supplemented in the extract (840 nM final concentration). Three replicates (n = 3) were completed and are overlayed on the bar plots. Error bars represent standard error of the mean. c. Structural model of the yeast 80S initiation complex bound by eIF5B (PDB: 6WOO16) used to predict the potential distance between the human 40S subunit (ybbR tag fused to N-term of uS19) and eIF5B (ybbR tag fused to residue 587) labeling sites. The relatively long predicted distance (>130 Å) precluded detection using FRET. eIF5B therefore was monitored via direct excitation and emission in all experiments. d. Theoretical real-time single-molecule fluorescence data from a ZMW with sequential association of the 40S subunit (green), eIF5B (orange), and 60S subunit (red), which was followed by departure of eIF5B from the newly-formed 80S initiation complex. The dwell times were defined as follows: between appearance of 40S and eIF5B signals, as the eIF5B association time (t1); between appearance of eIF5B and 60S signal, indicated by 40S-60S FRET, as 60S joining time (t2); and between appearance of 60S until loss of eIF5B, as eIF5B departure time (t3). e. Cumulative probability plots of the observed times at the indicated concentrations for eIF5B and 60S subunits at 30 °C. In all experiments, the 43S PIC was present at 10 nM (final concentration, via 40S-Cy3). WT and H706E indicate whether wild-type or catalytically-inactive eIF5B-Cy3.5 were present, respectively. Lines represent fits of observed data to single- or double-exponential functions, which yielded the indicated association or dissociation rates. All errors represent 95% confidence intervals (C.I.). The number of molecules analyzed were: 136 (black), 148 (purple), 134 (blue), 146 (red), and 150 (gray). f. Cumulative probability (top) and Eyring (bottom) plots of observed 60S joining and eIF5B departure times at 20, 25, 30, and 35 °C. These data were obtained using tethered, pre-equilibrated 48S initiation complexes (see, Extended Figure 3a for more details on the experiment setup) to provide sharper focus on eIF5B dynamics and 60S subunit joining. eIF5B-Cy3.5 and 60S-Cy5 subunits were present at 20 nM and 100 nM, respectively. Lines on the cumulative probability plots represent fits of the observed data to single or double-exponential functions, which yielded the indicated rates (errors represent 95% C.I.). Eyring plots (bottom) were modeled via linear regression analyses, which yielded the indicated enthalpies and entropies of activation. The number of molecules analyzed were: 116 (20 °C), 109 (25 °C), 120 (30 °C), and 44 (35 °C). g. Example single-molecule fluorescence data (top) and density plots (bottom) from the tethered mRNA setup when catalytically-inactive eIF5B (H706E) was examined at 30 °C. 40S-Cy3, eIF5B(H706E)-Cy3.5, and 60S-Cy5 subunits were present at 10 nM, 20 nM, and 100 nM, respectively. eIF5B(H706E) was unable to depart the 80S initiation complex in the presence of 1 mM GTP.

Extended Data Fig. 3 eIF5B associated more rapidly with pre-equilibrated 48S initiation complexes.

a. Schematic of an alternative experimental setup (‘tethered 48S’) used to examine eIF5B association with pre-equilibrated 48S initiation complexes. In this setup, a pre-formed 48S initiation complex at equilibrium on the β-globin mRNA was tethered to the ZMW imaging surface in the presence of 1 mM ATP and GTP. After removal of untethered components, data acquisition began via excitation with a 532 nm laser, and (final concentrations) 20 nM eIF5B-Cy3.5, 100 nM 60S-Cy5 subunits, 1 µM eIF1A, and 1 mM ATP and GTP were added. ZMWs with tethered 48S PICs were identified by the initial presence of 40S-Cy3 fluorescence signal (green). b. Example single-molecule fluorescence data from a tethered 48S PIC experiment where eIF5B association led to 60S subunit joining and rapid departure of eIF5B from the 80S initiation complex at 30 °C. The dwells were defined as in the ‘tethered mRNA’ experimental setup (see, Extended Data Fig. 2d). c. Comparison of observed eIF5B association times (t1) in tethered mRNA versus tethered 48S PIC experimental setups at 30 °C. The pre-equilibration of the 48S PIC prior to tethering allowed the slow upstream step to proceed, which enabled rapid and concentration-dependent eIF5B association. Fits of the observed association times are represented by the lines, which yielded the indicated rates (n = 172 and 111 for 20 & 40 nM, respectively). In one experiment (magenta, n = 113), the 48S PIC was formed and pre-equilibrated in the absence of eIF5, and then 2.5 nM eIF5 was added simultaneously with labeled eIF5B and 60S subunits to the tethered 48S PIC. The ‘tethered mRNA’ data were replotted from Extended Data Fig. 2e to facilitate comparisons. All errors represent 95% C.I. d. Example single-molecule fluorescence data from the tethered 48S complex experimental setup where wild-type eIF5B-Cy3.5 (20 nM) was added to pre-equilibrated 48S initiation complxes either in the absence of 60S subunits (top) or when 60S subunits (100 nM) were present but GTP was replaced with 1 mM GDPNP in free solution (bottom). e. Plot of the total lifetime of wild-type eIF5B-Cy3.5 on tethered 48S complexes in the indicated conditions. In the presence of 60S subunits and 1 mM GTP, eIF5B remained on initiation complexes for a total of ~15 s (t2 + t3). The total lifetime of eIF5B was lengthened approximately 34-fold to ~510 s when GTP was replaced with 1 mM of the non-hydrolyzable analog, GDPNP. eIF5B total lifetime also was lengthened dramatically (to ~ 130 s) when 60S subunits were omitted and 1 mM GTP was present. f. Table of 60S subunit joining and eIF5B association efficiencies when eIF2–\({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\)–GTP or eIF2–\({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\)–GDPNP were present at 30 °C. 60S subunit joining efficiency was quantified in the tethered mRNA experimental setup with 1 µM unlabeled eIF5B present. The 95% CI of the observed efficiencies are indicated, with 200 or 204 40S loading events or 48S complexes analyzed in each experiment, respectively. Observed efficiencies were corrected to account for the 40S and 60S subunit labeling efficiencies, which were 80% and 38%, respectively, with errors propagated from the observed 95% CIs. eIF5B association efficiency (45% labeled) was quantified in the tethered 48S PIC setup (to facilitate analyses) in the absence of 60S subunits.

Extended Data Fig. 4 eIF1A resides on initiation complexes until the 60S subunit joins.

a. Structural model that depicts the proximity of eIF1A and 40S subunit labeling sites. eIF1A was fluorescently labeled via an S102C substitution followed by reaction with Cy5-maleimide dye, which yields a fully-active protein in translation assays, as described previously25. 40S subunits were labeled with Cy3 on the N-terminus of uS19, via a fused ybbR tag (11 amino acid), as above. The relative proximity of the two labeling sites suggested the presence of eIF1A on the 40S subunit could be monitored via FRET. Of note, the first three N-terminal amino acids of uS19 were unresolved in the structure, which also does not account for the fused ybbR tags (11 amino acids). Thus, the reported distance between the 40S and eIF1A labeling sites is only an approximation. b, c. Theoretical (top) and example (middle, bottom) real-time single-molecule fluorescence data from a ZMW where doubly labeled 43S PICs (10 nM via 40S-Cy3; eIF1A-Cy5), 1 µM unlabeled eIF5B, and 200 nM 60S-Cy5.5 subunits were added to tethered eIF4ABGE-mRNA complexes at 30 °C. During imaging, eIF1A-Cy5 was present at 4.5-fold molar excess relative to 40S subunits. Key association and dissociation events are labeled on the theoretical plot. The lifetime of eIF1A on the 48S PIC (t1A,48S) was defined as the duration of individual 40S(Cy3,donor)-eIF1A(Cy5,acceptor) FRET events. In these experiments, 60S subunit joining was indicated by appearance of 40S(Cy3,donor)-60S(Cy5.5,acceptor) FRET. eIF1A lifetime on the 80S complex (t1A,80S) was defined as the time elapsed from 60S subunit joining until eIF1A departure (loss of 40S-eIF1A FRET). Panel B depicts an mRNA where eIF1A was co-recruited with the 40S subunit (signal begins in 40S-eIF1A FRET state) and that initial, co-recruited eIF1A protein was present until the 60S subunit joined. Panel C depicts a complex where multiple eIF1A binding events occurred prior to 60S subunit joining. In both cases, eIF1A departed either concomitantly (left) with or within a few hundred milliseconds (right) after 60S subunit joining. d. Plot of FRET efficiency (EFRET) distributions for 40S(Cy3,donor)-eIF1A(Cy5,acceptor) FRET. ‘All’ represents all observed 40S-eIF1A FRET events (n = 433), ‘60S joined’ represents events that overlapped with 60S subunit joining (n = 175), and ‘No 60S joining’ represents events that did not overlap with 60S subunit joining (n = 258). The lines represent fits of the observed data to single gaussian functions, which yielded the indicated means (µ) and standard deviations (σ). e. Cumulative probability plot of observed eIF1A reassociation times with the 48S initiation complex prior to 60S subunit joining at 30 °C (n = 258). eIF1A-Cy5 was present at ~45 nM in free solution. The line represents a fit of the observed data to a double-exponential function, which yielded the indicated rates. All errors represent 95% C.I. f. Cumulative probability plot of observed eIF1A lifetimes on either the 48S or 80S initiation complexes when either eIF5B-GTP or eIF5B-GDPNP was present at 30 °C. The line represents a fit of the observed data to either single- or double-exponential functions, which yielded the indicated rates. All errors represent 95% C.I. g. Table of 60S subunit joining efficiency when eIF1A-Cy5 was present in or eIF1A was absent from initiation reactions. The 95% CI of the observed efficiencies are indicated, with 500 40S loading events analyzed in each experiment. Observed efficiencies were corrected to account for the 40S and 60S subunit labeling efficiencies, which were 80% and 80%, respectively, with errors propagated from the observed 95% CIs. h. Cumulative probability plot of observed 60S subunit joining times at 30 °C when either 40 nM eIF5B-Cy3.5 (gold, n = 134) or 45 nM eIF1A-Cy5 (slate, n = 175) was present. These findings further confirmed that fluorescently-labeled eIF1A and eIF5B are fully active, as the proteins yielded identical rates of 60S subunit joining as to when the unlabeled version is present (unlabeled eIF1A was present in the eIF5B-Cy3.5 experiment, and vice versa). Moreover, the rare instances where the 60S subunit joined when eIF1A was absent from the reaction (reported in panel H) occurred 4-fold slower (grey, n = 100). The lines represent fits of the observed data to single-exponential functions, which yielded the indicated rates. All errors represent 95% C.I.

Extended Data Fig. 5 eIF1A and eIF5B reside simultaneously on initiation complexes when the 60S subunit joins.

a. Schematic of single-molecule experiments that examined eIF1A and eIF5B dynamics on the 48S initiation complex at equilibrium. Pre-formed 48S initiation complexes on β-globin mRNA were tethered at equilibrium within ZMWs in the presence of 1 mM ATP and GTP. After removal of untethered components, data acquisition began via excitation with a 532 nm laser, and an imaging mix that contained (final concentrations) 10 nM eIF1A-Cy5 (red) and 20 nM eIF5B-Cy3.5 (orange) was present at 30 °C. ZMWs with tethered 48S PICs were identified by the initial presence of 40S-Cy3 fluorescence signal (green). b. Rudimentary structural comparison where a low-resolution model of human eIF5B (PDB: 4UJC) was docked onto a high-resolution model of a mammalian 48S PIC post-recognition of the start codon (PDB: 6YAL). From these crude analyses, eIF1A was predicted to be within FRET distance (< 80 Å) of the N-terminus of truncated eIF5B, consistent with the eIF5B(Cy3.5, donor)-eIF1A(Cy5,acceptor) FRET signal observed in our single-molecule assays. c. Example single-molecule fluorescence data that depicts either: top, a complex with 40S(Cy3,donor)-eIF1A(Cy5,acceptor) FRET in the absence of eIF5B signal (‘eIF5B absent’); bottom, a complex with both 40S(Cy3,donor)-eIF1A(Cy5,acceptor) and eIF5B(Cy3.5,donor)-eIF1A(Cy5,acceptor) FRET, (‘eIF5B present’). Focused analyses were conducted on both forms of the 48S PIC to derive kinetic parameters. d. Cumulative probability plots of observed eIF1A reassociation times with (left) or eIF1A lifetimes on (right) the 48S initiation complex at 30 °C at equilibrium. eIF1A-Cy5 was present at 10 nM. The lines represent fits of the observed data to double-exponential functions, which yielded the indicated rates. All errors represent 95% C.I. 831 and 589 eIF1A binding events were analyzed when eIF5B was present or absent, respectively. e. Schematic of the four-color single-molecule experiment. The doubly labeled 43S PIC (10 nM by 40S-Cy3; eIF1A-Cy5), 40 nM eIF5B-Cy3.5, and 200 nM 60S-Cy5.5 subunit were added to b-globin mRNA tethered within ZMWs in the presence of saturating concentrations of eIFs 4A, 4B, 4G, and 4E, and 1 mM ATP and GTP at 30 °C. During imaging, eIF1A-Cy5 was present at 4.5-fold molar excess relative to the 40S subunit. Fluorescence data were acquired for 600 s with excitation via the 532 nm laser. The potential FRET signals are indicated in the box. f, g. Theoretical (panel F) and example (panel G) four-color single-molecule fluorescence data from a ZMW where a loaded doubly labeled 43S PIC (40S-Cy3, green; eIF1A-Cy5, red) was bound by eIF5B-Cy3.5 (orange), which was followed by 60S-Cy5.5 (purple) subunit joining. The 43S PIC was recruited in a 40S(Cy3, donor)-eIF1A(Cy5, acceptor) FRET state. Once eIF5B bound, the eIF1A-Cy5 signal increased due to eIF5B(Cy3.5, donor)-eIF1A(Cy5, acceptor) FRET. Joining of the 60S subunit was indicated by appearance of Cy5.5 fluorescence signal due to 40S(Cy3, donor)-60S(Cy5.5, acceptor) FRET. Departure of eIF1A-Cy5 and eIF5B-Cy3.5 was indicated by loss of Cy5 and Cy3.5 fluorescence signals, respectively. In panel G, the full experimental window is depicted on the left, and the middle and right panels represent zoomed views of the indicated time windows. Given bleed through across the four fluorescent channels, the fluorescence signals in each channel were made transparent before relevant events for presentation here. h. Quantification of eIF1A and eIF5B occupancy upon 60S subunit joining in the four-color single-molecule experiment. The potentially convoluted FRET signals and fluorescence bleed through among channels and factors precluded rigorous kinetic analyses, as exact frames for association and dissociation were extremely challenging to assign. However, the experiment did allow the presence of eIF1A and eIF5B upon 60S subunit joining to be quantified unmistakably. On a large majority of 48S initiation complexes (183/271), eIF1A-Cy5 signal was present when the 60S subunit joined, which corresponded to an estimated eIF1A occupancy of about 85 ± 5 % after correction for eIF1A labeling efficiency (~80 %). Of those eIF1A-bound 48S complexes, nearly half (88/183) also contained eIF5B-Cy3.5 signal when the 60S subunit joined, which indicated that about 90 ± 10 % of the 48S PICs contained both eIF1A and eIF5B, after correction for eIF5B labeling efficiency (~45%). eIF1A preceded eIF5B association on nearly all (79/88; 90 ± 20 %) 48S complexes that contained both labeled proteins when the 60S subunit joined.

Extended Data Fig. 6 Fourier Shell Correlation curves, local resolution, particle distribution plot and representative cryo-EM density examples.

a. Fourier Shell Correlation (FSC) curves computed for independently refined half-maps of the final subgroup of particles before masking (blue) and after masking (black). The red curve corresponds to FSC curve of phase-randomized structure factors beyond 7Å. b. Model versus map FSC (black). Red and blue curves correspond to a model-map overfitting validation test performed for the final model. A per-atom, random distortion of 0.5Å was introduced in the model which was subsequentially refined against half-map 1 only. FSC between the distorted and refined model against half-map 1 (blue) and half-map 2 (red, not included in the refinement) are nearly identical. c. Unsharpened map colored according to local resolution calculations with the resolution scale indicated on the left and a zoomed view for \({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\)/eIF1A/eIF5B on the right. d. Eulerian angular distribution for the final refined set of particles in a mollweide spherical projection. e. Several views of the final post-processed cryo-EM density used for model building and refinement. The areas of the map to which the different views belong are indicated.

Extended Data Fig. 7 eIF5B, through contacts with eIF1A, reorients \({{\bf{Met-tRNA}}}_{{\bf{i}}}^{{\bf{M}}{\bf{e}}{\bf{t}}}\) conformation in initiation complexes.

a. Post-processed cryo-EM density with eIF1A colored red, \({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\) green, and eIF5B blue. While nearly the entire flexible C-terminal tail of eIF1A was disordered, extra density proximal to W1207 of eIF5B was observed, which likely corresponds to I141 in the eIF1A C-terminal tail. b. Position of eIF1A and \({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\) in scanning, post-scanning, and eIF5B-bound initiation complexes. eIF1A remains bound in the ribosomal A site throughout initiation, until the 60S subunit joins. c. Structural models of the 40S subunit (with ribosomal proteins omitted) that indicate movement of the 18S rRNA in the indicated states relative to the final position in the elongation-competent 80S ribosome. In both the post-scanning and eIF5B-bound states, the 40S head region resembles the conformation in the 80S ribosome. d. Comparison of \({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\) conformations in scanning-competent (orange), post-scanning (grey), and eIF5B-bound 48S (green) complexes. eIF5B reorients the elbow and acceptor stem region of \({\text{Met-tRNA}}_{{\rm{i}}}^{{\rm{Met}}}\) to avoid steric clashes (black arrowheads) with the incoming 60S subunit.

Extended Data Fig. 8 eIF5B contacts with eIF1A and \({{\bf{Met-tRNA}}}_{{\bf{i}}}^{{\bf{M}}{\bf{e}}{\bf{t}}}\) mediate 60S subunit joining.

a. Alignments of human and yeast eIF5B protein sequences at the indicated regions. b. Fluorescence scans of a SDS-PAGE gel that analyzed the indicated purified and fluorescently-labeled eIF5B proteins. All proteins had the N-terminal domain removed (residues 1-586) and contained an N-terminal ybbR peptide tag (11 amino acids). Each protein was analyzed by SDS-PAGE twice. The full scans of the cropped gel are available in Supplementary Figure 1. c. Schematic of the single-molecule experiment to assess the role of the indicated residues in eIF5B. The 43S PIC (10 nM via 40S-Cy3 subunits, green), wild-type or mutant eIF5B-Cy3.5 (orange, at 20 or 40 nM), and 100 nM 60S-Cy5 (red) subunit were added to β-globin mRNA tethered within ZMWs in the presence of saturating concentrations of eIF4ABGE and 1 mM ATP and GTP at 30 °C. Fluorescence data were acquired for 600 s with excitation via the 532 nm laser. d. Theoretical real-time single-molecule fluorescence data from a ZMW with sequential association of the 40S subunit (green), eIF5B (orange), and 60S subunit (red), which was followed by departure of eIF5B from the newly-formed 80S initiation complex. In this experiment, the lifetimes of all eIF5B binding events that began prior to 60S subunit joining were analyzed. Events that concluded prior to 60S subunit joining were classified as ‘samples’, and events that concluded after the 60S subunit joined were classified as ‘productive’. To facilitate cross-comparisons in these experiments, the 60S subunit joining time was defined as the time elapsed from 40S subunit loading (appearance of green fluorescence) until 60S subunit joining, as indicated by 40S(Cy3, donor)-to-60S(Cy5, acceptor) FRET. e. Example single-molecule fluorescence data obtained when the indicated eIF5B proteins were present. The left plots represent the fluorescence intensities throughout the entire 600 s data collection window, and the right are zoomed views of events of interest. f. Fraction of observed eIF5B binding events on the 48S complex that were samples or productive, defined as in panel D. g. Cumulative probability plots of observed eIF5B lifetimes on the 48S complex when the indicated eIF5B protein was present at the indicated concentration (20 or 40 nM). The lines represent fits of the observed data to double-exponential functions, which yielded the indicated rates. All errors represent 95% C.I. The number of binding events analyzed (n) are indicated in the figure. h. Table of the 60S subunit joining efficiency in ZMWs that contained at least one 40S subunit loading event when the indicated eIF5B proteins were present at 40 nM. The 95% CI of the observed efficiencies are indicated, with 1,000 40S loading events analyzed in each experiment. Relative efficiencies were determined relative to the wild-type protein, with errors propagated from the 95% CIs. i. Cumulative probability plots of observed 60S joining times on the 48S complex when the indicated eIF5B protein was present at the indicated concentration (20 or 40 nM). The lines represent fits of the observed data to exponential functions, which yielded the indicated rates. All errors represent 95% C.I. The number of binding events analyzed (n) are indicated in the figure.

Extended Data Table 1 Cryo-EM data collection, model refinement, and validation statistics

Supplementary information

Supplementary Information

This file contains Supplementary Fig. 1 and legends for Supplementary Videos 1 and 2.

Reporting Summary

Peer Review File

Supplementary Video 1

Supplementary Video 2

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lapointe, C.P., Grosely, R., Sokabe, M. et al. eIF5B and eIF1A reorient initiator tRNA to allow ribosomal subunit joining. Nature 607, 185–190 (2022). https://doi.org/10.1038/s41586-022-04858-z

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41586-022-04858-z

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.

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