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.

  • Letter
  • Published:

Directional tissue migration through a self-generated chemokine gradient

Abstract

The directed migration of cell collectives is a driving force of embryogenesis1,2,3. The predominant view in the field is that cells in embryos navigate along pre-patterned chemoattractant gradients2. One hypothetical way to free migrating collectives from the requirement of long-range gradients would be through the self-generation of local gradients that travel with them4,5, a strategy that potentially allows self-determined directionality. However, a lack of tools for the visualization of endogenous guidance cues has prevented the demonstration of such self-generated gradients in vivo. Here we define the in vivo dynamics of one key guidance molecule, the chemokine Cxcl12a, by applying a fluorescent timer approach to measure ligand-triggered receptor turnover in living animals. Using the zebrafish lateral line primordium as a model, we show that migrating cell collectives can self-generate gradients of chemokine activity across their length via polarized receptor-mediated internalization. Finally, by engineering an external source of the atypical receptor Cxcr7 that moves with the primordium, we show that a self-generated gradient mechanism is sufficient to direct robust collective migration. This study thus provides, to our knowledge, the first in vivo proof for self-directed tissue migration through local shaping of an extracellular cue and provides a framework for investigating self-directed migration in many other contexts including cancer invasion6.

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

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

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

Figure 1: Tandem fluorescent protein timers (tFTs) allow quantification of chemokine-dependent receptor turnover in vivo.
Figure 2: Ligand-regulated chemokine receptor lifetime gradient indicates self-generated migration.
Figure 3: Role of Cxcr4b and Cxcr7 sink activities in chemokine gradient formation.
Figure 4: Chemokine gradient formation by Cxcr7 drives collective migration.

Similar content being viewed by others

References

  1. Friedl, P. & Gilmour, D. Collective cell migration in morphogenesis, regeneration and cancer. Nature Rev. Mol. Cell Biol. 10, 445–457 (2009)

    Article  CAS  Google Scholar 

  2. Rørth, P. Whence directionality: guidance mechanisms in solitary and collective cell migration. Dev. Cell 20, 9–18 (2011)

    Article  Google Scholar 

  3. Montell, D. J. Morphogenetic cell movements: diversity from modular mechanical properties. Science 322, 1502–1505 (2008)

    Article  ADS  CAS  Google Scholar 

  4. Dambly-Chaudière, C., Cubedo, N. & Ghysen, A. Control of cell migration in the development of the posterior lateral line: antagonistic interactions between the chemokine receptors CXCR4 and CXCR7/RDC1. BMC Dev. Biol. 7, 23 (2007)

    Article  Google Scholar 

  5. Scherber, C. et al. Epithelial cell guidance by self-generated EGF gradients. Integr. Biol. 4, 259–269 (2012)

    Article  Google Scholar 

  6. Friedl, P., Locker, J., Sahai, E. & Segall, J. E. Classifying collective cancer cell invasion. Nature Cell Biol. 14, 777–783 (2012)

    Article  Google Scholar 

  7. Ghysen, A. & Dambly-Chaudière, C. The lateral line microcosmos. Genes Dev. 21, 2118–2130 (2007)

    Article  CAS  Google Scholar 

  8. David, N. B. et al. Molecular basis of cell migration in the fish lateral line: role of the chemokine receptor CXCR4 and of its ligand, SDF1. Proc. Natl Acad. Sci. USA 99, 16297–16302 (2002)

    Article  ADS  CAS  Google Scholar 

  9. Haas, P. & Gilmour, D. Chemokine signaling mediates self-organizing tissue migration in the zebrafish lateral line. Dev. Cell 10, 673–680 (2006)

    Article  CAS  Google Scholar 

  10. Streichan, S. J., Valentin, G., Gilmour, D. & Hufnagel, L. Collective cell migration guided by dynamically maintained gradients. Phys. Biol. 8, 045004 (2011)

    Article  ADS  Google Scholar 

  11. Weijer, C. J. Collective cell migration in development. J. Cell Sci. 122, 3215–3223 (2009)

    Article  CAS  Google Scholar 

  12. Boldajipour, B. et al. Control of chemokine-guided cell migration by ligand sequestration. Cell 132, 463–473 (2008)

    Article  CAS  Google Scholar 

  13. Naumann, U. et al. CXCR7 functions as a scavenger for CXCL12 and CXCL11. PLoS ONE 5, e9175 (2010)

    Article  ADS  Google Scholar 

  14. Lee, E. et al. CXCR7 mediates SDF1-induced melanocyte migration. Pigment Cell Melanoma. Res. 26, 58–66 (2013)

    Article  CAS  Google Scholar 

  15. Rajagopal, S. et al. β-arrestin- but not G protein-mediated signaling by the ‘decoy’ receptor CXCR7. Proc. Natl Acad. Sci. USA 107, 628–632 (2010)

    Article  ADS  Google Scholar 

  16. Minina, S., Reichman-Fried, M. & Raz, E. Control of receptor internalization, signaling level, and precise arrival at the target in guided cell migration. Curr. Biol. 17, 1164–1172 (2007)

    Article  CAS  Google Scholar 

  17. Khmelinskii, A. et al. Tandem fluorescent protein timers for in vivo analysis of protein dynamics. Nature Biotechnol. 30, 708–714 (2012)

    Article  CAS  Google Scholar 

  18. Knaut, H., Werz, C. & Geisler, R. The Tübingen 2000 Screen Consortium & Nüsslein-Volhard, C. A zebrafish homologue of the chemokine receptor Cxcr4 is a germ-cell guidance receptor. Nature 421, 279–282 (2003)

    Article  ADS  CAS  Google Scholar 

  19. Kettleborough, R. N. W. et al. A systematic genome-wide analysis of zebrafish protein-coding gene function. Nature 496, 494–497 (2013)

    Article  ADS  CAS  Google Scholar 

  20. Valentin, G., Haas, P. & Gilmour, D. The chemokine SDF1a coordinates tissue migration through the spatially restricted activation of Cxcr7 and Cxcr4b. Curr. Biol. 17, 1026–1031 (2007)

    Article  CAS  Google Scholar 

  21. Distel, M., Wullimann, M. F. & Köster, R. W. Optimized Gal4 genetics for permanent gene expression mapping in zebrafish. Proc. Natl Acad. Sci. USA 106, 13365–13370 (2009)

    Article  ADS  CAS  Google Scholar 

  22. Gilmour, D., Knaut, H., Maischein, H.-M. & Nüsslein-Volhard, C. Towing of sensory axons by their migrating target cells in vivo. Nature Neurosci. 7, 491–492 (2004)

    Article  CAS  Google Scholar 

  23. Peri, F. & Nüsslein-Volhard, C. Live imaging of neuronal degradation by microglia reveals a role for v0-ATPase a1 in phagosomal fusion in vivo. Cell 133, 916–927 (2008)

    Article  CAS  Google Scholar 

  24. Levoye, A., Balabanian, K., Baleux, F., Bachelerie, F. & Lagane, B. CXCR7 heterodimerizes with CXCR4 and regulates CXCL12-mediated G protein signaling. Blood 113, 6085–6093 (2009)

    Article  CAS  Google Scholar 

  25. Nakayama, M. et al. Spatial regulation of VEGF receptor endocytosis in angiogenesis. Nature Cell Biol. 15, 249–260 (2013)

    Article  CAS  Google Scholar 

  26. Müller, A. et al. Involvement of chemokine receptors in breast cancer metastasis. Nature 410, 50–56 (2001)

    Article  ADS  Google Scholar 

  27. Hernandez, L., Magalhaes, M. A. O., Coniglio, S. J., Condeelis, J. S. & Segall, J. E. Opposing roles of CXCR4 and CXCR7 in breast cancer metastasis. Breast Cancer Res. 13, R128 (2011)

    Article  CAS  Google Scholar 

  28. Luker, K. E. et al. Scavenging of CXCL12 by CXCR7 promotes tumor growth and metastasis of CXCR4-positive breast cancer cells. Oncogene 31, 4750–4758 (2012)

    Article  CAS  Google Scholar 

  29. Zhang, Y., Muyrers, J. P., Testa, G. & Stewart, A. F. DNA cloning by homologous recombination in Escherichia coli. Nature Biotechnol. 18, 1314–1317 (2000)

    Article  CAS  Google Scholar 

  30. Pau, G., Fuchs, F., Sklyar, O., Boutros, M. & Huber, W. EBImage—an R package for image processing with applications to cellular phenotypes. Bioinformatics 26, 979–981 (2010)

    Article  CAS  Google Scholar 

  31. Westerfield, M. The Zebrafish Book: A Guide for the Laboratory use of Zebrafish (Brachydanio rerio) 2nd edn (University of Oregon Press, 1993)

    Google Scholar 

  32. Siekmann, A. F., Standley, C., Fogarty, K. E., Wolfe, S. A. & Lawson, N. D. Chemokine signaling guides regional patterning of the first embryonic artery. Genes Dev. 23, 2272–2277 (2009)

    Article  CAS  Google Scholar 

  33. Kwan, K. M. et al. The Tol2kit: a multisite gateway-based construction kit for Tol2 transposon transgenesis constructs. Dev. Dyn. 236, 3088–3099 (2007)

    Article  CAS  Google Scholar 

  34. Canals, M. et al. Ubiquitination of CXCR7 controls receptor trafficking. PLoS ONE 7, e34192 (2012)

    Article  ADS  CAS  Google Scholar 

  35. Preibisch, S., Saalfeld, S. & Tomancak, P. Globally optimal stitching of tiled 3D microscopic image acquisitions. Bioinformatics 25, 1463–1465 (2009)

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We are grateful to J. Ellenberg, F. Peri and S. De Renzis for comments on the manuscript, S. Streichan and the Gilmour laboratory for suggestions. We thank R. Koester and M. Distel for the lateral line Gal4-driver, E. Busch-Nentwich and D. Stemple for the cxcr7 mutant identified by the Zebrafish Mutation Project, the European Molecular Biology Laboratory (EMBL) Monoclonal Antibody Facility for Cxcr7 and Cxcr4b antibodies, the EMBL Advanced Light Microscopy Facility for imaging assistance, B. Klaus from the EMBL Centre for Statistical Data Analysis for advice and A. Gruia for fish care. We acknowledge funding from the European Commission’s FP7 Network of Excellence ‘Systems Microscopy’ (W.H. and J.D.B.), Marie Curie FP6 (A.F.-M.), European Molecular Biology Organization (A.Kh) and Deutsche Forschungsgemeinschaft SFB 488 (D.G.).

Author information

Authors and Affiliations

Authors

Contributions

D.G. and E.D. designed study with input from M.K. and A.Kh. E.D. performed all experiments, C.Q. helped perform the nerve rescue experiments. J.D.B. and W.H. developed the data analysis methods. G.V. developed the BAC-complementation of Cxcr4b. A.Ku generated the cxcr4b:NLS-tdTomato line, S.D. generated the UAS:cxcr7-GFP line, L.R.N. designed and cloned the hsp70:cxcl12a-GFP_T2A_mKate2-CAAX construct, A.F.-M. designed and validated the Cxcr7 and Cxcr4b monoclonal antibodies. D.G. and E.D. wrote the paper with input from all authors.

Corresponding author

Correspondence to Darren Gilmour.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Extended data figures and tables

Extended Data Figure 1 cxcr4b:cxcr4b-tFT BAC rescues primordium migration defect in cxcr4b mutants.

a, Image of cxcr4b mutant embryos, where primordium migration (white arrowheads) is made visible by cxcr4b:NLS-tdTomato (red). Fish carrying cxcr4b:cxcr4b-tFT transgenes are identifiable by an independent transgenic marker that labels the lens blue (cry:CFP, asterisks). b, In all cxcr4b:cxcr4b-tFT transgenic embryos the primordium migrated normally, whereas non-transgenics showed migration defects typical of cxcr4b mutants20. Dot-plot, median and interquartile range of the indicated number of samples are shown (Welch’s t-test, ***P < 0.0001).

Extended Data Figure 2 Cxcl12a-induced internalization targets Cxcr4b to lysosomal degradation.

ac, Cxcr4b is sent for lysosomal degradation in wild-type embryos. a, cxcr4b:cxcr4b-GFP embryo stained with the lysosomal marker LysoTracker Red DND-99. Left, maximum intensity projection of a primordium confocal z-stack (scale bar, 20 μm) showing Cxcr4b–GFP (green) and LysoTracker (red) localization. Right, magnification of single confocal plane showing LysoTracker and Cxcr4b–GFP co-localization (arrowheads; scale bar, 5 μm). b, Impaired lysosomal acidification leads to accumulation of endogenous Cxcr4b in intracellular vesicles. Anti-Cxcr4b (maximum intensity projections) of Bafilomycin A1-treated embryos (bottom) and controls (top; DMSO (dimethylsulphoxide)). c, Bafilomycin A1 treatment results in an increased number of Cxcr4b–tFT vesicles. Images show maximum intensity projections of z-stacks (scale bar, 20 μm). df, High Cxcl12a levels cause increased degradation of endogenous Cxcr4b. d, Anti-Cxcr4b immunohistochemistry 3 h after heat-shock induction (1 h at 38 °C) of hsp70:cxcl12a (sum projection of confocal z-stacks). e, Quantification of Cxcr4b levels within the primordium (box–and–whiskers plot, one-way ANOVA test with Bonferroni’s multiple comparison adjustment, ***P < 0.001) showing significantly reduced levels in hsp70:cxcl12a transgenics compared to non-transgenic controls. This effect is suppressed by Bafilomycin A1 treatment, resulting in increased accumulation of Cxcr4b-positive intracellular vesicles (f). Insets show single plane magnification of the squared area. Scale bars, 20 μm. NS, not significant.

Extended Data Figure 3 Cxcr4b–tFT lifetime ratio is stable over several hours.

ac, Spatially resolved profiles of Cxcr4b–tFT-expressing primordia at 32 h.p.f. (t = 0) and 35 h.p.f. (t = 3 h) show that although sfGFP (a) and TagRFP (b) median fluorescence intensities increase through time, their ratio instead stays constant and maintains a gradient across the front–rear axis of the tissue (c). Median and MAD of 7 samples acquired on the same day with identical imaging conditions are shown. d, Ratiometric images of one representative sample through time. Lookup table = red/green fluorescence intensity ratio. Scale bar, 20 μm. All lifetime measurements were performed during this time window, whenever possible.

Extended Data Figure 4 Increased turnover in primordium leading domain results in graded distribution of endogenous Cxcr4b receptor.

Anti-Cxcr4b immunohistochemistry reveals a gradient of endogenous receptor along the primordium. a, Cxcr4b antibody (red) and a homogeneous membrane counter-label given by cldnb:lynGFP transgenic line (green) displayed in segmented (top, sum projection of segmented stack) and ratiometric form (bottom, calibration bar = Cxcr4b antibody/Lyn–GFP fluorescence intensity ratio). b, Cxcr4b antibody/Lyn–GFP fluorescence intensity ratio profiles. Median and MAD after normalization of individual profiles by the front-most 15 μm are plotted.

Extended Data Figure 5 Cxcr4b–tFT lifetime ratio gradient results from protein turnover not from a transcriptional gradient.

a, Schematic representation of the modified BACs used to create Cxcr4b–tFT and mem-tFT reporter lines. b, Predicted behaviour of Cxcr4b–tFT and mem-tFT in response to Cxcl12a. c, d, Ratiometric images of 36 h.p.f. Cxcr4b–tFT (c) and mem-tFT (d) of whole-mount transgenic embryos. Segmentation was performed using intensity-based thresholding on sfGFP channel. Scale bars, 200 μm. Calibration bar = normalized red/green ratio. e, f, Raw primordium images (merge of green and red channels) corresponding to ratiometric images shown in Fig. 2a, b. Scale bars, 20 μm. g, Ratiometric images of cxcr4b:mem-tFT before (top) and 10 h after (bottom) heat-shock-induced pulse of Cxcl12a. Scale bars, 20 μm. h, Spatially resolved lifetime ratio analysis of the sample in g. Although overall the ratio increases, during the long time lag, no front to rear gradient could be detected, in contrast to the behaviour of Cxcr4b–tFT (compare to Fig. 2e, f).

Extended Data Figure 6 Cxcr7-mediated internalization of Cxcl12a–GFP at the tissue rear.

a, Film-strip montages of z-stacks showing mosaic expression of Cxcl12a–GFP resulting in preferential internalization at the rear of the primordium (outlined by white line, corresponding to Supplementary Video 3). Stacks were acquired 1–2 h after heat-shock induction (1 h at 38 °C) of hsp70:cxcl12a-GFP_T2A_mKate2-CAAX-injected embryos. Accumulation of Cxcl12a–GFP-positive vesicles in the primordium (yellow arrows) and Cxcl12a–GFP-expressing cells within the primordium are indicated (white arrowheads). Scale bar, 20 μm. b, Anti-GFP and anti-Cxcr7 immunohistochemistry showing co-localization of Cxcl12a–GFP (green) and Cxcr7 (red) in vesicles at the rear of the primordium. Images show maximum intensity projection. Scale bars, 10 μm.

Extended Data Figure 7 Validation of internalization-defective chemokine receptor forms.

a, Cartoon illustrating the Gal4/UAS strategy used to induce mosaic expression of chemokine receptors in the primordium. b, Schematic of internalization-defective Cxcr4b form (Cxcr4*), as described in ref. 16. Mutated putative phosphorylation sites (S to A) are shown in orange. c, Examples of Gal4/UAS-driven expression of Cxcr4b–GFP (green or greyscale, top) or Cxcr4b*–GFP (green or greyscale, bottom) in cxcr4b:NLS-tdTomato (red) transgenic embryos. Both receptor forms localize at the plasma membrane (arrowheads). Higher magnification of the squared area shows internalized Cxcr4b (arrows), but not Cxcr4b*, under unstimulated conditions, presumably in response to endogenous Cxcl12a. Scale bars, 10 μm. d, e, Heat-shock-induced Cxcl12a expression causes complete internalization of Cxcr4b–GFP (d), whereas Cxcr4b*–GFP remains at the plasma membrane (e). Thus, Cxcr4b* does not internalize Cxcl12a. f, An internalization-defective form of Cxcr7 (Cxcr7*) was created by deleting the 40 C-terminal amino acids of the receptor. Putative ubiquitination and phosphorylation target residues, conserved in humans, are shown in red and orange, respectively. g, GFP fusion of full-length Cxcr7 localizes in vesicles (arrows in the inset), consistent with antibody stainings of endogenous Cxcr7 (Fig. 3a). h, C terminus deletion results in clear receptor re-localization to the plasma membrane (arrowheads). Cxcr7* is only occasionally found in vesicles (arrows). Expression of both receptor forms was driven by Gal4/UAS. Scale bars, 50 μm (10 μm in the insets).

Extended Data Figure 8 Cxcr7–GFP misexpression alters migrating primordium morphology.

ac, Primordia at 36 h.p.f. (marked by NLS-dtTomato) in wild type (WT) conditions (a) or misexpressing Cxcr4b–GFP (b) or Cxcr7–GFP (c) under the control of the Gal4/UAS system. d, cxcr4b−/− primordium misexpressing Cxcr7–GFP (green). Scale bars, 10 μm. e, Overlay of outlines of several primordia in the same conditions as in ac. Primordia are aligned by the leading edge (right). Ectopic Cxcr7–GFP, but not Cxcr4b–GFP, expression results in primordium elongation. This phenotype is Cxcr4b dependent, as elongated morphology is never observed in Cxcr4b mutants misexpressing Cxcr7–GFP (d).

Extended Data Figure 9 Cxcr4a is not required for primordium migration.

a, In situ hybridization reveals cxcr4a expression in the primordium (36 h.p.f. embryo). Inset shows higher magnification of the squared area comprising the primordium. b, Images show primordium migration (red = cxcr4b:NLS-tdTomato) at 48 h.p.f. in the presence (top) or absence (bottom) of Cxcr4a. Asterisk marks heart oedema, characteristic of cxcr4a−/−(corresponding to cxcr4aum20/um20)32. c, Quantification of primordium migration in cxcr4a−/− embryos. The absence of Cxcr4a does not affect migration when Cxcr4b is present. Dot-plot, median, interquartile range and number of analysed samples are shown. Unpaired t-test.

Extended Data Figure 10 Statistical analysis of tissue-scale lifetime ratios.

a, Summary of tissue-scale lifetime ratio distributions at 32 h.p.f. (data corresponding to Figs 2d and 3c). b, Quantile-quantile (Q-Q) plots of the analysed data sets. c, Statistical tests results.

Supplementary information

Supplementary Information

This file contains a Supplementary Discussion. (PDF 101 kb)

Increased levels of Cxcl12a induce Cxcr4b-tFT internalisation and reduce lifetime ratio.

Cxcr4b-tFT response to a pulse of Cxcl12a induced by heat-shock (hsp70:Cxcl12a, below, corresponding to Fig. 1e) and control (above, heat-shocked, non-transgenic control). Increased levels of Cxcl12a induce endocytosis of Cxcr4b-tFT, which is sent for lysosomal degradation. Fluorescent signal recovers at the plasma-membrane through new synthesis. Scale bar = 50 µm. Time indicates minutes after the heat-shock pulse. Time interval = 12 min. (MOV 3425 kb)

Reestablishment of Cxcr4b-tFT activity gradient.

Front to rear differences in Cxcr4b-tFT lifetime ratios are re-established after erasure by a heat-shock pulse of Cxcl12a (corresponding to Fig. 2e). Scale bar = 50 µm. Time indicates minutes after the heat-shock pulse. Time interval = 62 min. (MOV 473 kb)

Cxcl12a-GFP internalisation is higher at the rear of the primordium

Labelled Cxcl12a (green), expressed outside the primordium (expressing cells are marked by red mKate2-CAAX) results in ligand accumulation in the tissue rear, in deposited rosettes and interneuromast cells (white outline). Four examples are shown. Stacks, shown as film-strip montages in Extended Data Fig. 6, were acquired 1 to 2 hours after heat-shock induction (1 hour at 38 °C) of embryos injected with hsp70:Cxcl12a-GFP_T2A_mKate2-CAAX. First row corresponds to Fig. 3g. Scale bar = 20 µm. (MOV 6594 kb)

The posterior lateral line nerve follows the migrating primordium.

The pLLN (nerve:mem-GFP, green) co-migrates with the primordium (red = NLS-tdTomato) and never overtakes its leading edge. Scale bar = 50 µm. Time interval = 12 min. (MOV 5396 kb)

Nerve:Cxcr7-GFP impairs primordium migration.

The presence of Cxcr7-GFP (green, below) in the pLLN decreases migration speed in an otherwise wild-type primordium (red = NLS-tdTomato). Above, a wild-type control is shown. Scale bar = 50 µm. Time interval = 10 min. (MOV 6773 kb)

Impaired migration in a Cxcr7 mutant embryo

This video shows strongly impaired directional primordium migration in a cxcr7-/- mutant embryo (red = NLS-tdTomato, corresponding to Fig. 4f). Scale bar = 50 µm. Time interval = 23 min. (MOV 4108 kb)

Primordium migration rescue by nerve:Cxcr7 is position dependent.

Cxcr7-GFP (green) expression in the pLLN rescues primordium migration in a cxcr7-/- mutant embryo (red = NLS-tdTomato, corresponding to Fig. 4g). Notice that primordium migration stalls when the nerve overtakes the leading edge of the tissue, but resumes as soon as it falls back to its ‘follower’ position. Scale bar = 50 µm. Time interval = 23 min. (MOV 5734 kb)

Primordium migration rescued by nerve:Cxcr7.

Example of cxcr7-/- mutant primordium rescued by nerve:Cxcr7-GFP. The pLLN never overtakes and the primordium migrates at constant speed. Scale bar = 50 µm. Time interval = 23 min. (MOV 928 kb)

PowerPoint slides

Rights and permissions

Reprints and permissions

About this article

Cite this article

Donà, E., Barry, J., Valentin, G. et al. Directional tissue migration through a self-generated chemokine gradient. Nature 503, 285–289 (2013). https://doi.org/10.1038/nature12635

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nature12635

This article is cited by

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