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:

Multisite phosphorylation code of CDK

Nature Structural & Molecular Biology volume 26, pages 649–658 (2019)Cite this article

Subjects

Abstract

The quantitative model of cyclin-dependent kinase (CDK) function states that cyclins temporally order cell cycle events at different CDK activity levels, or thresholds. The model lacks a mechanistic explanation, as it is not understood how different thresholds are encoded into substrates. We show that a multisite phosphorylation code governs the phosphorylation of CDK targets and that phosphorylation clusters act as timing tags that trigger specific events at different CDK thresholds. Using phospho-degradable CDK threshold sensors with rationally encoded phosphorylation patterns, we were able to predictably program thresholds over the entire range of the Saccharomyces cerevisiae cell cycle. We defined three levels of CDK multisite phosphorylation encoding: (i) serine−threonine swapping in phosphorylation sites, (ii) patterning of phosphorylation sites, and (iii) cyclin-specific docking combined with modulation of CDK activity. Thus, CDK can signal via hundreds of differentially encoded targets at precise times to provide a temporally ordered phosphorylation pattern required for cell division.

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

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

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

Fig. 1: CDK thresholds can be encoded into substrates.
Fig. 2: Binary coding of CDK thresholds using S-T swapping.
Fig. 3: Cyclin-specific encoding of CDK thresholds.
Fig. 4: Cyclin rearrangements change the order of the thresholds.
Fig. 5: Non-consensus CDK sites as specificity filters: separating the CDK oscillator from the other kinase activities.
Fig. 6: Helper networks in output site phosphorylation and the number of phosphorylation sites.

Similar content being viewed by others

Data availability

Source data for Fig. 1g,i–k, 2a,b,d–h, 3b–h, 4a,c,d,h,i, 5a–c, 6b,c and Supplementary Figs. 2a,c−f, 3b,c, 4a−q, 5a−c,e,h, 6a–f are available online. All other data are available upon request.

References

  1. Holt, L. J. et al. Global analysis of Cdk1 substrate phosphorylation sites provides insights into evolution. Science 325, 1682–1686 (2009).

    Article  CAS  Google Scholar 

  2. Fisher, D. et al. Phosphorylation network dynamics in the control of cell cycle transitions. J. Cell Sci. 125, 4703–4711 (2012).

    Article  CAS  Google Scholar 

  3. Holt, L. J. Regulatory modules: Coupling protein stability to phopshoregulation during cell division. FEBS Lett. 586, 2773–2777 (2012).

    Article  CAS  Google Scholar 

  4. Kõivomägi, M. et al. Multisite phosphorylation networks as signal processors for Cdk1. Nat. Struct. Mol. Biol. 20, 1415–1424 (2013).

    Article  Google Scholar 

  5. Valk, E. et al. Multistep phosphorylation systems: tunable components of biological signaling circuits. Mol. Biol. Cell 25, 3456–3460 (2014).

    Article  Google Scholar 

  6. Stern, B. & Nurse, P. A quantitative model for the cdc2 control of S phase and mitosis in fission yeast. Trends Genet. 12, 345–350 (1996).

    Article  CAS  Google Scholar 

  7. Coudreuse, D. & Nurse, P. Driving the cell cycle with a minimal CDK control network. Nature 468, 1074–1079 (2010).

    Article  CAS  Google Scholar 

  8. Swaffer, M. P., Jones, A. W., Flynn, H. R., Snijders, A. P. & Nurse, P. CDK Substrate phosphorylation and ordering the cell cycle. Cell 167, 1750–1761.e16 (2016).

    Article  CAS  Google Scholar 

  9. Oikonomou, C. & Cross, F. R. Rising cyclin-CDK levels order cell cycle events. PLoS One 6, e20788 (2011).

    Article  CAS  Google Scholar 

  10. Loog, M. & Morgan, D. O. Cyclin specificity in the phosphorylation of cyclin-dependent kinase substrates. Nature 434, 104–108 (2005).

    Article  CAS  Google Scholar 

  11. Örd, M. & Loog, M. How the cell cycle clock ticks. Mol. Biol. Cell 30, 169–172 (2019).

    Article  Google Scholar 

  12. Songyang, Z. et al. Use of an oriented peptide library to determine the optimal substrates of protein kinases. Curr. Biol. 4, 973–982 (1994).

    Article  CAS  Google Scholar 

  13. Suzuki, K. et al. Identification of non-Ser/Thr-Pro consensus motifs for Cdk1 and their roles in mitotic regulation of C2H2 zinc finger proteins and Ect2. Sci. Rep. 5, 7929 (2015).

    Article  CAS  Google Scholar 

  14. Schulman, B. A., Lindstrom, D. L. & Harlow, E. D. Substrate recruitment to cyclin-dependent kinase 2 by a multipurpose docking site on cyclin A. Biochemistry 95, 10453−10458.(1998).

  15. Wilmes, G. M. et al. Interaction of the S-phase cyclin Clb5 with an RXL docking sequence in the initiator protein Orc6 provides an origin-localized replication control switch. Genes Dev. 18, 981–991 (2004).

    Article  CAS  Google Scholar 

  16. Bhaduri, S. & Pryciak, P. M. Cyclin-specific docking motifs promote phosphorylation of yeast signaling proteins by G1/S Cdk complexes. Curr. Biol. 21, 1615–1623 (2011).

    Article  CAS  Google Scholar 

  17. Kõivomägi, M. et al. Dynamics of Cdk1 substrate specificity during the cell cycle. Mol. Cell 42, 610–623 (2011).

    Article  Google Scholar 

  18. Archambault, V., Buchler, N. E., Wilmes, G. M., Jacobson, M. D. & Cross, F. R. Two-faced cyclins with eyes on the targets. Cell Cycle 4, 125–130 (2005).

    Article  CAS  Google Scholar 

  19. Kõivomägi, M. et al. Cascades of multisite phosphorylation control Sic1 destruction at the onset of S phase. Nature 480, 128–131 (2011).

    Article  Google Scholar 

  20. McGrath, D. A. et al. Cks confers specificity to phosphorylation-dependent CDK signaling pathways. Nat. Struct. Mol. Biol. 20, 1407–1414 (2013).

    Article  CAS  Google Scholar 

  21. Kõivomägi, M. & Skotheim, J. M. Docking interactions: cell-cycle regulation and beyond. Curr. Biol. 24, R647–R649 (2014).

    Article  Google Scholar 

  22. Hodge, A. & Mendenhall, M. The cyclin-dependent kinase inhibitory domain of the yeast Sic1 protein is contained within the C-terminal 70 amino acids. Mol. Gen. Genet. 262, 55–64 (1999).

    Article  CAS  Google Scholar 

  23. Doncic, A., Falleur-Fettig, M. & Skotheim, J. M. Distinct interactions select and maintain a specific cell fate. Mol. Cell 43, 528–539 (2011).

    Article  CAS  Google Scholar 

  24. King, R. W., Deshaies, R. J., Peters, J. M. & Kirschner, M. W. How proteolysis drives the cell cycle. Science 274, 1652–1659 (1996).

    Article  CAS  Google Scholar 

  25. Zhou, P. & Howley, P. M. Ubiquitination and degradation of the substrate recognition subunits of SCF ubiquitin–protein ligases. Mol. Cell 2, 571–580 (1998).

    Article  CAS  Google Scholar 

  26. Hao, B., Oehlmann, S., Sowa, M. E., Harper, J. W. & Pavletich, N. P. Structure of a Fbw7-Skp1-Cyclin E complex: multisite-phosphorylated substrate recognition by SCF ubiquitin ligases. Mol. Cell 26, 131–143 (2007).

    Article  CAS  Google Scholar 

  27. Lyons, N. A. et al. Sequential primed kinases create a damage-responsive phosphodegron on Eco1. Nat. Struct. Mol. Biol. 20, 194–201 (2013).

    Article  CAS  Google Scholar 

  28. Al-Zain, A., Schroeder, L., Sheglov, A. & Ikui, A. E. Cdc6 degradation requires phosphodegron created by GSK-3 and Cdk1 for SCFCdc4 recognition in Saccharomyces cerevisiae . Mol. Biol. Cell 26, 2609–2619 (2015).

    Article  CAS  Google Scholar 

  29. Godfrey, M. et al. PP2ACdc55 Phosphatase imposes ordered cell-cycle phosphorylation by opposing threonine phosphorylation. Mol. Cell 65, 393–402.e3 (2017).

    Article  CAS  Google Scholar 

  30. Örd, M., Venta, R., Möll, K., Valk, E. & Loog, M. Cyclin-specific docking mechanisms reveal the complexity of M-CDK function in the cell cycle. Mol. Cell https://doi.org/10.1016/j.molcel.2019.04.026 (2019).

  31. Bhaduri, S. et al. A docking interface in the cyclin Cln2 promotes multi-site phosphorylation of substrates and timely cell-cycle entry. Curr. Biol. 25, 316–325 (2015).

    Article  CAS  Google Scholar 

  32. Welcker, M. et al. Fbw7 dimerization determines the specificity and robustness of substrate degradation. Genes Dev. 27, 2531–2536 (2013).

    Article  CAS  Google Scholar 

  33. Nash, P. et al. Multisite phosphorylation of a CDK inhibitor sets a threshold for the onset of DNA replication. Nature 414, 514–521 (2001).

    Article  CAS  Google Scholar 

  34. Yang, X., Lau, K.-Y., Sevim, V. & Tang, C. Design principles of the yeast G1/S switch. PLoS Biol. 11, e1001673 (2013).

    Article  CAS  Google Scholar 

  35. Cross, F. R., Archambault, V., Miller, M. & Klovstad, M. Testing a mathematical model of the yeast cell cycle. Mol. Biol. Cell 13, 52–70 (2002).

    Article  CAS  Google Scholar 

  36. Cross, F. R., Yuste-Rojas, M., Gray, S. & Jacobson, M. D. Specialization and targeting of B-Type cyclins. Mol. Cell 4, 11–19 (1999).

    Article  CAS  Google Scholar 

  37. Hu, F. & Aparicio, O. M. Swe1 regulation and transcriptional control restrict the activity of mitotic cyclins toward replication proteins in Saccharomyces cerevisiae . Proc. Natl Acad. Sci. USA 102, 8910 (2005).

    Article  CAS  Google Scholar 

  38. Keaton, M. A. et al. Differential susceptibility of yeast S and M phase CDK complexes to inhibitory tyrosine phosphorylation. Curr. Biol. 17, 1181–1189 (2007).

    Article  CAS  Google Scholar 

  39. Gunawardena, J. Multisite protein phosphorylation makes a good threshold but can be a poor switch. Proc. Natl Acad. Sci. USA 102, 14617–14622 (2005).

    Article  CAS  Google Scholar 

  40. Miller, C. J. & Turk, B. E. Homing in: Mechanisms of substrate targeting by protein kinases. Trends Biochem Sci. 43, 380–384 (2018).

    Article  CAS  Google Scholar 

  41. Chang, F. & Herskowitz, I. Identification of a gene necessary for cell cycle arrest by a negative growth factor of yeast: FAR1 is an inhibitor of a G1 cyclin, CLN2. Cell 63, 999–1011 (1990).

    Article  CAS  Google Scholar 

  42. Gartner, A., Nasmyth, K. & Ammerer, G. Signal transduction in Saccharomyces cerevisiae requires tyrosine and threonine phosphorylation of FUS3 and KSS1. Genes Dev. 6, 1280–1292 (1992).

    Article  CAS  Google Scholar 

  43. Ferrell, J. E. Tripping the switch fantastic: how a protein kinase cascade can convert graded inputs into switch-like outputs. Trends Biochem. Sci. 21, 460–466 (1996).

    Article  CAS  Google Scholar 

  44. Ferrell, J. E., Ha, S. H. & Ha, S. H. Ultrasensitivity part II: multisite phosphorylation, stoichiometric inhibitors, and positive feedback. Trends Biochem. Sci. 39, 556–569 (2014).

    Article  CAS  Google Scholar 

  45. Hein, J. B., Hertz, E. P. T., Garvanska, D. H., Kruse, T. & Nilsson, J. Distinct kinetics of serine and threonine dephosphorylation are essential for mitosis. Nat. Cell Biol. 19, 1433–1440 (2017).

    Article  CAS  Google Scholar 

  46. Huis In’t Veld, P. J. et al. Molecular basis of outer kinetochore assembly on CENP-T. eLife 5, e21007 (2016).

  47. Gould, K. L. & Nurse, P. Tyrosine phosphorylation of the fission yeast cdc2+protein kinase regulates entry into mitosis. Nature 342, 39–45 (1989).

    Article  CAS  Google Scholar 

  48. Morla, A. O., Draetta, G., Beach, D. & Wang, J. Y. J. Reversible tyrosine phosphorylation of cdc2: Dephosphorylation accompanies activation during entry into mitosis. Cell 58, 193–203 (1989).

    Article  CAS  Google Scholar 

  49. McNulty, J. J. & Lew, D. J. Swe1p responds to cytoskeletal perturbation, not bud Size, in S. cerevisiae. Curr. Biol. 15, 2190–2198 (2005).

    Article  CAS  Google Scholar 

  50. Longtine, M. S. et al. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast 14, 953–961 (1998).

    Article  CAS  Google Scholar 

  51. Janke, C. et al. A versatile toolbox for PCR-based tagging of yeast genes: new fluorescent proteins, more markers and promoter substitution cassettes. Yeast 21, 947–962 (2004).

    Article  CAS  Google Scholar 

  52. Puig, O. et al. The tandem affinity purification (TAP) Method: A general procedure of protein complex purification. Methods 24, 218–229 (2001).

    Article  CAS  Google Scholar 

  53. Ubersax, J. A. et al. Targets of the cyclin-dependent kinase Cdk1. Nature 425, 859–864 (2003).

    Article  CAS  Google Scholar 

  54. McCusker, D. et al. Cdk1 coordinates cell-surface growth with the cell cycle. Nat. Cell Biol. 9, 506–515 (2007).

    Article  CAS  Google Scholar 

  55. Reynard, G. J., Reynolds, W., Verma, R. & Deshaies, R. J. Cks1 is required for G(1) cyclin-cyclin-dependent kinase activity in budding yeast. Mol. Cell. Biol. 20, 5858–5864 (2000).

    Article  CAS  Google Scholar 

  56. Doncic, A., Eser, U., Atay, O. & Skotheim, J. M. An algorithm to automate yeast segmentation and tracking. PLoS One 8, e57970 (2013).

    Article  CAS  Google Scholar 

  57. Jehl, P., Manguy, J., Shields, D. C., Higgins, D. G. & Davey, N. E. ProViz—a web-based visualization tool to investigate the functional and evolutionary features of protein sequences. Nucleic Acids Res. 44, W11–W15 (2016).

    Article  CAS  Google Scholar 

  58. Waterhouse, A. M., Procter, J. B., Martin, D. M. A., Clamp, M. & Barton, G. J. Jalview Version 2–a multiple sequence alignment editor and analysis workbench. Bioinformatics 25, 1189–1191 (2009).

    Article  CAS  Google Scholar 

  59. Krystkowiak, I. & Davey, N. E. SLiMSearch: a framework for proteome-wide discovery and annotation of functional modules in intrinsically disordered regions. Nucleic Acids Res. 45, W464–W469 (2017).

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We would like to thank D. Morgan, P. Pryciak, D. Kellogg, M. Kõivomägi, and J. Skotheim for valuable comments on the manuscript. We thank M. Peter (ETH Zürich) for providing NES-Cdc4 construct and J. Mihhejev for technical assistance. The work was supported by ERC Consolidator Grant 649124 and Estonian Science Agency grants Nr. IUT2–21 and PRG550 to M.L.

Author information

Authors and Affiliations

  1. Institute of Technology, University of Tartu, Tartu, Estonia

    Mihkel Örd, Kaidi Möll, Alissa Agerova, Rait Kivi, Ilona Faustova, Rainis Venta, Ervin Valk & Mart Loog

Authors
  1. Mihkel Örd
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Kaidi Möll
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Alissa Agerova
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Rait Kivi
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Ilona Faustova
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Rainis Venta
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Ervin Valk
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Mart Loog
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

M.Ö. and M.L. directed the study. M.Ö., K.M., A.A., R.V., and I.F. cloned the constructs and made the strains. M.Ö., R.K. and E.V. purified the proteins. M.Ö. performed the microscopy and in vitro kinase experiments, M.Ö., K.M., and A.A. performed the western blotting. M.Ö. and M.L. wrote the manuscript.

Corresponding author

Correspondence to Mart Loog.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information: Anke Sparmann was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

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

Integrated supplementary information

Supplementary Figure 1 Quantification of sensor-GFP dynamics and analysis of sensor phosphorylation by western blotting.

(a, b) Plots showing examples of degradation timing and duration quantification for an early sensor (WT RxL in ‘a’) and a late sensor (T33+4 in ‘b’). To determine the sensor degradation timing and duration values for individual cells, the nuclear fluorescence levels of the sensor-GFP were plotted over time and fitted with a Gaussian function. Timing denotes the time from 50% nuclear exit of Whi5-mCherry in late G1 to sensor levels dropping to 50% of peak level. Degradation duration marks the time of sensor level decreasing from 80% to 20%. (c) Western blot images that show multisite phosphorylation and degradation of the threshold sensors (described in Fig. 1h) in the cell cycle. Cells were arrested in G1 using pheromone and released to cell cycle. The lysates were separated using Phos-tag SDS-PAGE. Source Data.

Source Data

Supplementary Figure 2 Cdk1 thresholds based on serine-threonine swapping in phosphorylation sites.

(a) Degradation duration values for the indicated sensors in single cells. The numbers above the plot show the mean values for each sensor, the error bars are 95% CI of the mean. (b) Multisite phosphorylation and degradation dynamics of the 5SP-3TP and 5TP-3SP sensors were analyzed by western blotting of synchronized cultures. The cells were arrested in G1 using pheromone arrest and released to cell cycle. The lysates were separated using Phos-tag SDS-PAGE, the western blot images are shown. (c) Degradation duration values for individual cells expressing the indicated sensors. The numbers above the plot show the mean values for each sensor, the error bars are 95% CI of the mean. (d) Plots showing the mean nuclear levels of the indicated threshold sensors in wild-type and Cdc55 non-expressing strain backgrounds. (e, f) Timing and degradation duration values for individual cells of the wild-type and cdc55 strain backgrounds. The numbers above the plot show the mean values for each sensor, the error bars are 95% CI. Source Data.

Source Data

Supplementary Figure 3 Localization and quantitative expression profiles of cyclins.

The localization and levels of cyclin-Citrine fusion proteins were followed in unperturbed cell cycles of single cells using time-lapse microscopy. (a) Images from fluorescent channels fused with phase-contrast images show the nuclear-cytoplasmic shuttling of Whi5-mCherry and the expression of indicated cyclin-Citrine fusion proteins. Images were taken every 3 minutes. The export of 50% of Whi5-mCherry at Start takes place at time 0 in the cell marked with an asterisk. (b) The times of cyclin-Citrine nuclear levels reaching 50% of their peak levels after Start in single cells. (c) Peak nuclear fluorescence intensities of cyclin-Citrine fusion proteins in the cell cycle in individual cells. The error bars in ‘b’ and ‘c’ show 95% CI of the mean. Source Data.

Source Data

Supplementary Figure 4 Cyclin docking motifs and cyclin hydrophobic patch control the degradation timing of the CDK threshold sensors.

(a–c) Plots showing the degradation duration of indicated sensors in individual cells. The numbers above the plots show mean degradation durations, the error bars show 95% CI. (d) The dynamics of the mean ±SEM intensities of threshold sensors based on T33(+4) with or without RxL motif in wild-type or clb5(hpm) clb3(hpm) cells from Start. (e, f) Similar plots as in ‘d’ showing the effect of PxF motif in the threshold sensors based on T33(+4) in wild-type or clb3(hpm) cells, and the LxF motif in wild-type or clb2(hpm) cells. (g-i) Plots showing the 50% degradation timing values for individual cells of the strains presented on the graphs ‘d-f’. The error bars show 95% CI of the mean. (j-l) Plots showing the degradation duration values for individual cells of the strains presented on the graphs ‘d-f’. (m-q) Quantified phosphorylation signals from the in vitro kinase assays presented in Figure 3 ‘i’. The error bars show standard deviation. Source Data.

Source Data

Supplementary Figure 5 Comparison of phosphorylation and degradation of threshold sensors in the nucleus and cytoplasm.

(a) Cytoplasmic fluorescence intensities of cyclins fused to Citrine averaged over a population of cells synchronized at the time of 50% of Whi5 nuclear export in late G1. See also Supplementary Fig 3a. The graph shows mean ±SEM. (b) The time of cyclin-Citrine cytoplasmic levels reaching 50% of peak levels after Start in single cells. The error bars show 95% CI of the mean. (c) Peak cytoplasmic fluorescence intensities of cyclin-Citrine fusion proteins in the cell cycle of individual cells. The error bars show 95% CI of the mean. (d) Diagrams of the threshold sensors used in ‘e’ and ‘h’. (e) Graph showing the dynamics of the threshold sensors with NES and different cyclin-specific docking motifs. Plot shows mean ±SEM. (f, g) Cells expressing the wild-type threshold sensor without cyclin docking motifs under the ADH1 promoter from pRS315 vector were synchronized in G1 using pheromone arrest and released to cell cycle. Western blot images that show the multisite phosphorylation and degradation of the sensor with either NLS or NES. The proteins were separated using Phos-tag SDS-PAGE. (h) Graphs showing the mean intensities of the indicated cytoplasmic threshold sensors. The error bars are ±SEM. Source Data.

Source Data

Supplementary Figure 6 Multisite phosphorylation code in ordering of CDK substrate phosphorylation.

(a, b) Graphs showing the dynamics of an intermediate (T33(+4)) and late (T5S T33S) threshold sensors with and without S-CDK- (RxL) and M-CDK-specific (LxF) docking motifs. The plot shows mean nuclear fluorescence intensities of the sensors, the error bars are ±SEM. Due to the poor intrinsic activity of S-CDK compared to M-CDK, the compensatory effect of docking motifs of the intermediate and late sensors is different. (c) The degradation timing from Start of indicated sensors containing different number of phosphorylation sites in single cells. (d) The degradation duration of the sensors in individual cells. (e) The timing of degradation of 50% of the indicated sensors from Start in single cells. In ‘c-e’, the numbers above the plot show mean values for each sensor, the error bars are 95% CI of the mean. (f) Full set of differentially encoded CDK threshold sensors used in this study covering the entire span of the cell cycle are plotted relatively to Start as t=0 (50% Whi5-mCherry nuclear exit). The plot shows mean nuclear levels of the sensors. (g) The barcodes showing predicted phosphorylation sites and potential cyclin docking motifs in disordered regions of a selected set of Cdk1 targets with different degrees of docking connectivity. The arrows below each substrate are the predicted Cks1-mediated connections and the arrows above each substrate show cyclin docking mediated phosphorylation of indicated sites. Source Data.

Source Data

Supplementary information

Supplementary Information

Supplementary Figures 1–6, Supplementary Notes 1–8, Supplementary Data Set 1

Reporting Summary

Supplementary Table 1

Supplementary Table 2

Supplementary Table 3

Source data

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Örd, M., Möll, K., Agerova, A. et al. Multisite phosphorylation code of CDK. Nat Struct Mol Biol 26, 649–658 (2019). https://doi.org/10.1038/s41594-019-0256-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41594-019-0256-4

This article is cited by

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