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:

Visualization of a short-range Wnt gradient in the intestinal stem-cell niche

Abstract

Mammalian Wnt proteins are believed to act as short-range signals1,2,3,4, yet have not been previously visualized in vivo. Self-renewal, proliferation and differentiation are coordinated along a putative Wnt gradient in the intestinal crypt5. Wnt3 is produced specifically by Paneth cells6,7. Here we have generated an epitope-tagged, functional Wnt3 knock-in allele. Wnt3 covers basolateral membranes of neighbouring stem cells. In intestinal organoids, Wnt3-transfer involves direct contact between Paneth cells and stem cells. Plasma membrane localization requires surface expression of Frizzled receptors, which in turn is regulated by the transmembrane E3 ligases Rnf43/Znrf3 and their antagonists Lgr4-5/R-spondin. By manipulating Wnt3 secretion and by arresting stem-cell proliferation, we demonstrate that Wnt3 mainly travels away from its source in a cell-bound manner through cell division, and not through diffusion. We conclude that stem-cell membranes constitute a reservoir for Wnt proteins, while Frizzled receptor turnover and ‘plasma membrane dilution’ through cell division shape the epithelial Wnt3 gradient.

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: Endogenous Wnt3 protein is localized to basolateral plasma membranes in the intestinal crypt.
Figure 2: Wnt3 transfer requires direct cell contact and has a limited range.
Figure 3: Frizzled receptors act as membrane tether of Wnt3.
Figure 4: Cell proliferation influences Wnt3 surface level and signal range.

Similar content being viewed by others

References

  1. van den Heuvel, M., Nusse, R., Johnston, P. & Lawrence, P. A. Distribution of the wingless gene product in Drosophila embryos: a protein involved in cell-cell communication. Cell 59, 739–749 (1989)

    Article  CAS  PubMed  Google Scholar 

  2. Thorpe, C. J., Schlesinger, A., Carter, J. C. & Bowerman, B. Wnt signaling polarizes an early C. elegans blastomere to distinguish endoderm from mesoderm. Cell 90, 695–705 (1997)

    Article  CAS  PubMed  Google Scholar 

  3. Goldstein, B., Takeshita, H., Mizumoto, K. & Sawa, H. Wnt signals can function as positional cues in establishing cell polarity. Dev. Cell 10, 391–396 (2006)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Alexandre, C., Baena-Lopez, A. & Vincent, J.-P. Patterning and growth control by membrane-tethered Wingless. Nature 505, 180–185 (2014)

    Article  CAS  ADS  PubMed  Google Scholar 

  5. Clevers, H. The intestinal crypt, a prototype stem cell compartment. Cell 154, 274–284 (2013)

    Article  CAS  PubMed  Google Scholar 

  6. Gregorieff, A. et al. Expression pattern of Wnt signaling components in the adult intestine. Gastroenterology 129, 626–638 (2005)

    Article  CAS  PubMed  Google Scholar 

  7. Sato, T. et al. Paneth cells constitute the niche for Lgr5 stem cells in intestinal crypts. Nature 469, 415–418 (2011)

    Article  CAS  ADS  PubMed  Google Scholar 

  8. Coudreuse, D. Y. M., Roël, G., Betist, M. C., Destrée, O. & Korswagen, H. C. Wnt gradient formation requires retromer function in Wnt-producing cells. Science 312, 921–924 (2006)

    Article  CAS  ADS  PubMed  Google Scholar 

  9. Kiecker, C. & Niehrs, C. A morphogen gradient of Wnt/β-catenin signalling regulates anteroposterior neural patterning in Xenopus. Development 128, 4189–4201 (2001)

    CAS  PubMed  Google Scholar 

  10. Zecca, M., Basler, K. & Struhl, G. Direct and long-range action of a wingless morphogen gradient. Cell 87, 833–844 (1996)

    Article  CAS  PubMed  Google Scholar 

  11. Neumann, C. J. & Cohen, S. M. Long-range action of Wingless organizes the dorsal-ventral axis of the Drosophila wing. Development 124, 871–880 (1997)

    CAS  PubMed  Google Scholar 

  12. Strigini, M. & Cohen, S. M. Wingless gradient formation in the Drosophila wing. Curr. Biol. 10, 293–300 (2000)

    Article  CAS  PubMed  Google Scholar 

  13. Galli, L. M., Barnes, T. L., Secrest, S. S., Kadowaki, T. & Burrus, L. W. Porcupine-mediated lipid-modification regulates the activity and distribution of Wnt proteins in the chick neural tube. Development 134, 3339–3348 (2007)

    Article  CAS  PubMed  Google Scholar 

  14. Muñoz, J. et al. The Lgr5 intestinal stem cell signature: robust expression of proposed quiescent ‘+4’ cell markers. EMBO J. 31, 3079–3091 (2012)

    Article  PubMed  PubMed Central  Google Scholar 

  15. Barker, N. et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature 449, 1003–1007 (2007)

    Article  CAS  ADS  PubMed  Google Scholar 

  16. Farin, H. F., Van Es, J. H. & Clevers, H. Redundant sources of Wnt regulate intestinal stem cells and promote formation of Paneth cells. Gastroenterology 143, 1518–1529.e7 (2012)

    Article  CAS  PubMed  Google Scholar 

  17. Kabiri, Z. et al. Stroma provides an intestinal stem cell niche in the absence of epithelial Wnts. Development 141, 2206–2215 (2014)

    Article  CAS  PubMed  Google Scholar 

  18. Sato, T. et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 459, 262–265 (2009)

    Article  CAS  ADS  PubMed  Google Scholar 

  19. Janda, C. Y., Waghray, D., Levin, A. M., Thomas, C. & Garcia, K. C. Structural basis of Wnt recognition by Frizzled. Science 337, 59–64 (2012)

    Article  CAS  ADS  PubMed  PubMed Central  Google Scholar 

  20. Liu, P. et al. Requirement for Wnt3 in vertebrate axis formation. Nature Genet. 22, 361–365 (1999)

    Article  CAS  PubMed  Google Scholar 

  21. Yin, X. et al. Niche-independent high-purity cultures of Lgr5+ intestinal stem cells and their progeny. Nature Methods 11, 106–112 (2014)

    Article  CAS  PubMed  Google Scholar 

  22. Hao, H.-X. et al. ZNRF3 promotes Wnt receptor turnover in an R-spondin-sensitive manner. Nature 485, 195–200 (2012)

    Article  CAS  ADS  PubMed  Google Scholar 

  23. Koo, B.-K. et al. Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 488, 665–669 (2012)

    Article  CAS  ADS  PubMed  Google Scholar 

  24. Chen, P.-H., Chen, X., Lin, Z., Fang, D. & He, X. The structural basis of R-spondin recognition by LGR5 and RNF43. Genes Dev. 27, 1345–1350 (2013)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Schwank, G. et al. Functional repair of CFTR by CRISPR/Cas9 in intestinal stem cell organoids of cystic fibrosis patients. Cell Stem Cell 13, 653–658 (2013)

    Article  CAS  PubMed  Google Scholar 

  26. Lecourtois, M., Alexandre, C., Dubois, L. & Vincent, J.-P. Wingless capture by Frizzled and Frizzled2 in Drosophila embryos. Dev. Biol. 235, 467–475 (2001)

    Article  CAS  PubMed  Google Scholar 

  27. Gurney, A. et al. Wnt pathway inhibition via the targeting of Frizzled receptors results in decreased growth and tumorigenicity of human tumors. Proc. Natl Acad. Sci. USA 109, 11717–11722 (2012)

    Article  CAS  ADS  PubMed  PubMed Central  Google Scholar 

  28. Chen, B. et al. Small molecule-mediated disruption of Wnt-dependent signaling in tissue regeneration and cancer. Nature Chem. Biol. 5, 100–107 (2009)

    Article  CAS  ADS  Google Scholar 

  29. Ritsma, L. et al. Intestinal crypt homeostasis revealed at single-stem-cell level by in vivo live imaging. Nature 507, 362–365 (2014)

    Article  CAS  ADS  PubMed  PubMed Central  Google Scholar 

  30. Habib, S. J. et al. A localized Wnt signal orients asymmetric stem cell division in vitro. Science 339, 1445–1448 (2013)

    Article  CAS  ADS  PubMed  PubMed Central  Google Scholar 

  31. Farley, F. W., Soriano, P., Steffen, L. S. & Dymecki, S. M. Widespread recombinase expression using FLPeR (flipper) mice. Genesis 28, 106–110 (2000)

    Article  CAS  PubMed  Google Scholar 

  32. Lallemand, Y., Luria, V., Haffner-Krausz, R. & Lonai, P. Maternally expressed PGK-Cre transgene as a tool for early and uniform activation of the Cre site-specific recombinase. Transgenic Res. 7, 105–112 (1998)

    Article  CAS  PubMed  Google Scholar 

  33. Barrow, J. R. et al. Ectodermal Wnt3/β-catenin signaling is required for the establishment and maintenance of the apical ectodermal ridge. Genes Dev. 17, 394–409 (2003)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. El Marjou, F. et al. Tissue-specific and inducible Cre-mediated recombination in the gut epithelium. Genesis 39, 186–193 (2004)

    Article  CAS  PubMed  Google Scholar 

  35. Koo, B.-K. et al. Controlled gene expression in primary Lgr5 organoid cultures. Nature Methods 9, 81–83 (2012)

    Article  CAS  Google Scholar 

  36. Lamprecht, M. R., Sabatini, D. M. & Carpenter, A. E. CellProfiler: free, versatile software for automated biological image analysis. Biotechniques 42, 71–75 (2007)

    Article  CAS  PubMed  Google Scholar 

  37. Jmol: an open-source Java viewer for chemical structures in 3D. http://www.jmol.org/

  38. Morgenstern, B. DIALIGN: multiple DNA and protein sequence alignment at BiBiServ. Nucleic Acids Res. 32, W33–W36 (2004)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Schepers, A. G., Vries, R., van den Born, M., van de Wetering, M. & Clevers, H. Lgr5 intestinal stem cells have high telomerase activity and randomly segregate their chromosomes. EMBO J. 30, 1104–1109 (2011)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank M. van den Born, C. Kroon-Veenboer, J. van Es, T. Darvishi and S. van den Brink and the Hubrecht imaging facility for technical support. We thank A. Gurney for the pan-Fzd antibody and B.-K. Koo for discussions. M.M.M. is supported by the European Research Council (ERC StG 242958), the European Union (FP7 Marie-Curie ITN 608180 ‘WntsApp’) and the Netherlands Scientific Organization (NWO-ECHO 711.013.012 and NWO-VICI 91815604).

Author information

Authors and Affiliations

Authors

Contributions

H.F.F. conceived the study, performed the experiments and wrote the manuscript. I.J. and O.B. performed experiments and provided intellectual input. M.H.M. and D.V.F.T. performed experiments. J.K. performed ES blastocyst injection. K.P., S.A. and P.J.P. provided microscopic support and reagents. M.M.M. and H.C. conceived the study and wrote the manuscript.

Corresponding authors

Correspondence to Henner F. Farin or Hans Clevers.

Ethics declarations

Competing interests

H.C. is an inventor on several patents involving the organoid culture system.

Extended data figures and tables

Extended Data Figure 1 Generation of an epitope-tagged Wnt3 mouse allele.

a, Reported three-dimensional protein structure of Xwnt8 (in turquoise) bound to Fzd8-CRD (in blue) (Protein Data Bank accession number 4F0A)19. Dotted line depicts the location of the amino (N) terminus, where the HA epitope tag was introduced. Note that the position of the tag does not interfere with receptor binding. The image was created with the jmol-viewer37. Glycosyl and palmitoyl groups are shown in green and yellow, respectively. b, Amino-acid alignment of the N-terminal region of Wnt3 and other Wnt-family members. Predicted signal peptides are indicated in red. The HA tag (YPYDVPDYASL) was inserted after position Q41 that is labelled in blue. The first residue of Xwnt8 that is resolved in the crystal structure in a (T32) is shaded in green. Small letters indicate non-conserved residues that could not be aligned by the software38. The degree of local similarity is marked by asterisks. A score of five asterisks represents maximal similarity, which is not reached using the default settings of the algorithm. c, HA-knock-in strategy: a targeting vector that comprised the normal exon 2, a Frt-site flanked Pgk-neomycin resistance-polyA cassette and an inverted HA-inserted exon 2 (placed in intron 1 of the gene) was used. LoxP and LoxM sides were introduced in a configuration that allowed excision of the wild-type exon 2 and inversion of the HA-tagged exon 2 by Cre-mediated recombination (orientation of LoxP and LoxM sites is indicated with arrowheads). Transgenic mice were crossed to Pgk-Cre mice for germline recombination of the allele. This was done because the non-recombined allele was apparently a null allele as no homozygous offspring could be obtained. We suspect that the antisense configuration of the HA-modified exon 2 could result in RNA duplex formation and masking of the regular splice acceptor site of exon 2. Consistently, we observed a shorter transcript lacking wild-type exon 2 from the non-recombined allele by RT–PCR analysis. d, Southern blot analysis of ES cell clones to confirm correct targeting. Genomic BamHI digest was performed (see scheme in c).

Extended Data Figure 2 A permissive internal location for introduction of epitope tags in Wnt proteins.

a, Amino-acid sequences of tagged Wnt versions used in this study. Signal peptides are labelled in red. Protein alignment (Clustal Omega program, Uniprot); asterisk, colon and dot symbols indicate full conservation, and groups of strongly and weakly similar properties, respectively. b, Lentiviral rescue experiment in Wnt-dependent (Wnt3Δ/Δ) small intestinal organoids. Top: average organoid number (±s.d. in n = 3 independent wells) in the absence of exogenous Wnt-conditioned medium (Wnt-CM). Introduction of an internal HA- or Flag-tag does not interfere with rescue activity of Wnt3 lentivirus. Bottom: cell morphology after 5 days (passage 0) in the absence or presence of Wnt-CM. Scale bar, 50 μm. c, Introduction of an internal Flag-tag in mouse Wnt3a results in functional protein secretion in L-cells. Top: TOPFlash assay using conditioned media from control L-cells and stable lines expressing Wnt3a and Wnt3a–Flag. Bottom: immunoblots of the conditioned media measured above using anti-Wnt3a and anti-Flag antibodies. Fully scanned western blots are shown.

Extended Data Figure 3 Wnt signalling status in Wnt3-HA knock-in crypts and organoids.

Quantitative RT–PCR analysis of freshly isolated crypts (a) or established organoids (b, c) from the small intestine. a, b, Mean normalized expression levels in Wnt3HA/HA relative to Wnt3HA/+-samples (n = 6 mice, per genotype). c, Relative expression to Wnt3+/+ organoids (n = 4 independent wells). Error bars, s.d., no significant changes were found (P > 0.05 as determined by Student’s t-test). Normal expression of stem-cell markers (Lgr5, Olfm4), Wnt pathway activity (Axin2) and Paneth cell markers (Lyz1, Defa6 and Wnt3) indicates that introduction of the HA-tag does not interfere with Wnt signalling.

Extended Data Figure 4 Immunodetection of endogenous Wnt3-HA protein expression.

Western blot analysis using cell lysates from wild-type controls or Wnt3HA/HA organoids after normal culture or directed differentiation to Paneth cell s (DAPT/CHIR-99021 treatment for 6 days (ref. 21)). a, In microscopic images, Paneth cell differentiation is evident by presence of a dark granular structure. b, Anti-HA western blot shows a specific signal corresponding to the Wnt3-HA protein (arrow; expected molecular mass of the mature protein chain is 39 kDa). As positive control, lysates from HEK293T cells transfected with an expression plasmid encoding the Wnt3-HA cDNA were used. Full-scan western blot is shown. c, Wnt3-HA staining (green signal) on small intestinal cryosections of wild-type control and Wnt3HA/HA mice. Counterstaining of secretory granules (wheat germ agglutinin; WGA) and cell membranes (anti-Epcam). Data from Fig. 1e are shown in bright-field and single confocal channels (×20 objective) with magnified crypt region (right). Note that the HA-signal in Paneth cells is mutually exclusive with WGA-positive apical granula. d, High-magnification confocal image of crypt (×63 objective) merged with bright-field channel or nuclear staining (DAPI). Paneth cells were identified by their granular morphology (asterisks); arrows label crypt membranes adjacent to Paneth cells. Single confocal image is shown (for the entire z-stack see Supplementary Video 1). e, Whole-mount staining of wild-type control and Wnt3HA/HA small intestinal organoids. Confocal images or z-projections of co-stainings as in c (×20 objective). f, High-magnification confocal image (×63 objective) of a crypt region in Wnt3HA/HA organoid. Scale bars, 50 μm (a, c, e) and 10 μm (d, f).

Extended Data Figure 5 Diffusible Wnt activity in organoids is neither sufficient nor necessary to support growth.

a, Culture after co-embedding of dsRED-labelled wild-type organoids with GFP-labelled wild-type or Wnt3Δ/Δ organoids in Matrigel. Organoid fragments were seeded in a 1:1 ratio. Wnt3Δ/Δ cells cannot be propagated alone or in the presence of wild-type cells. Images after seeding (P0; day 3) and after passage 1 (P1; day 10). b, Co-culture at higher seeding density using dsRED-labelled wild-type organoids or in vitro differentiated Paneth cells21 with GFP-labelled Wnt3Δ/Δ organoids in a 3:1 ratio. Images of the same wells are shown at 1 and 5 days after seeding. c, d, Quantification of results shown in a and b. Mean relative number of organoids (±s.d.) from n = 3 independent experiments. e, Anti-HA immunodetection following co-culture of Wnt3HA/HA organoids or in vitro differentiated Paneth cells with dsRED-labelled wild-type recipient organoids (3:1 ratio). Inserts show growth in Matrigel. Confocal z-projected images are shown. Note that dsRED-positive cells remain negative for Wnt3-HA. f, Bead depletion experiment. Wild-type, Wnt3HA/HA or Wnt3Δ/Δ organoids were either embedded Matrigel alone or together with anti-HA affinity beads (blue asterisks) to sequester diffusible HA-tagged Wnt. Bright-field images after 6 days of culture; note that Wnt3HA/HA organoids display unaffected morphology in the presence of beads. Wnt3Δ/Δ organoids do not grow in the presence of beads and L-cell derived Wnt3a-HA CM (black arrows) demonstrating efficient depletion. Scale bars, 500 μm (a, b) and 100 μm (e, f).

Extended Data Figure 6 Wnt3 protein localization and messenger RNA expression depend on the R-spondin (Rspo) concentration.

a, Wnt3-HA immunostaining (Wnt3HA/HA organoids) after culture for 6 days in 1%, 5% or 25% Rspo-conditioned medium. Counterstaining of plasma membranes (Epcam) and secretory granules (WGA). Scale bar, 25 μm. b, RT–PCR analysis of stem-cell and differentiation markers in small intestinal organoids following 6 days of culture in variable concentrations of Rspo. Shown are mean normalized expression levels (±s.d.) in n = 3 independent wells relative to organoids cultured in the normal concentration of 5% Rspo. Markers of stem cells (Lgr5, Olfm4), Wnt activity (Axin2), Paneth cells (Lyz1, Defa6, Wnt3), enterocytes (Alpi), endocrine cells (Chga) and goblet cells (Muc2). Note that the messenger RNA expression of Wnt3 is sensitive to reduced Rspo concentration.

Extended Data Figure 7 CRISPR/Cas9 induced mutations in organoids.

Sequence analysis of indel mutations in targeted regions of mouse APC (a), Rnf43 (b) and Znrf3 (c) genomic loci. Two independent clonal organoid lines were analysed for each genotype. Genomic loci were PCR amplified, fragments were subcloned in E. coli and five or six colonies were sequenced per clonal line. For APC (clone#1 and #2) and Znrf3 (clone#1) mutant lines, hemizygosity was found, suggesting a larger genomic deletion on the other chromosome. sgRNA target sequences are shown in blue with the PAM sequence marked in bold. Arrows, Cas9 cleavage sites. Inserted nucleotides are shown in red.

Extended Data Figure 8 Differential efficiency of Wnt3 transfer to APC- and Rnf43/Znrf3-deficient cells.

a, Re-associated epithelia to test Wnt3 decoration on receiving cells (dsRED positive). The z-projected confocal images show depleted HA-signal on APC mutant cells (arrowheads) and enriched staining on Rnf43/Znrf3 mutant cells (asterisks). Wnt3 range: receiving cells were colour-coded depending on the success of transfer and the distance to the closest Wnt3-HA neighbour. b, Average percentages of each distance fraction in n = 10 re-associated organoids (±s.d.). Data were compared by t-test to re-association experiment using wild-type receiving cells (shown in Fig. 2f). ***P < 10−3; **P < 10−2; NS, non-significant. Scale bars, 10 μm.

Extended Data Figure 9 Pharmacological inhibition of proliferation in organoids.

Cell cycle status of wild-type mouse small intestinal organoids was determined by flow cytometry. EDU incorporation assay in controls and 24 h after administration of EGFR-inhibitor (Gefinitib), MEK-inhibitor (PD-0325901) or CDK4/6-inhibitor (Palbociclib) to the regular culture medium. EDU was added 1 h before collection and dissociation of cells for FACS analysis. a, Original FACS data. Single cells were gated using FSC/SSC characteristics. EDU signals were plotted against 7-AAD signals (DNA content). b, Inhibitor titration. Relative percentage of EDU-positive cells is shown for two independent experiments (grey line shows average). Experiments in Fig. 4 were performed at concentrations indicated by arrows. Note that Lgr5+ intestinal stem cells double every 19–20 h in vitro21, comparable to the cell cycle length in vivo39 (21.5 h).

Supplementary information

Wnt3-HA localization in small intestinal crypt of Wnt3HA/HAi mice. z-stack of confocal images (63× objective)

HA signal alone (left) and merged images with nuclear staining (DAPI, middle) and bright field channel (right) are shown. Paneth cells can be identified by their granular morphology. Note the prominent perinuclear signal in Paneth cells, but a lack of overlap with secretory granules in these cells. Plasma membranes of the stem cells, which are adjacent to Paneth cells are positive for Wnt3-HA. Images are 30 μm in width. (MOV 1243 kb)

Wnt3-HA localizes to plasma membranes of the stem cell compartment in organoids.

3-D projections of confocal image stacks of single 'crypts' of Wnt3HA/HA organoid after of HA staining with (left) or without (right) permeabilization step. Below light microscopic images are shown. Image width is 82 μm. (MOV 1974 kb)

Wnt3-HA remains highly restricted to plasma membranes close to producing cells.

Confocal 3-D projection of a re-associated organoid. Single bud is shown with HA-producing Paneth cells (Wnt3HA/HA, labeled by dsRED expression) and non-labeled Wnt3HA/HA cells cultured in normal growth medium. HA staining (green signal) shows local transfer of Wnt3-HA protein to adjacent cells. Image width is 100 μm. (MOV 887 kb)

Crypt budding after IWP-2 washout.

Time-lapse videos during 36 hours of culture, bright field images are shown. Left panel shows growth in continued presence of IWP-2/Wnt-CM. Organoids remain spherical and non-polarized. Middle panel: Following washout of IWP-2/Wnt-CM and change to normal culture medium (ENR) formation of crypt protrusions is observed. Right panel: Addition of EGFRi (right) to normal culture medium after IWP-2/Wnt-CM washout blocks crypt formation. Only rudimentary protrusions are formed, demonstrating that EGFR signaling (and cell proliferation) is crucial for this morphogenetic program. (MOV 3441 kb)

PowerPoint slides

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Farin, H., Jordens, I., Mosa, M. et al. Visualization of a short-range Wnt gradient in the intestinal stem-cell niche. Nature 530, 340–343 (2016). https://doi.org/10.1038/nature16937

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nature16937

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