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:

Structures of fungal and plant acetohydroxyacid synthases

Abstract

Acetohydroxyacid synthase (AHAS), also known as acetolactate synthase, is a flavin adenine dinucleotide-, thiamine diphosphate- and magnesium-dependent enzyme that catalyses the first step in the biosynthesis of branched-chain amino acids1. It is the target for more than 50 commercial herbicides2. AHAS requires both catalytic and regulatory subunits for maximal activity and functionality. Here we describe structures of the hexadecameric AHAS complexes of Saccharomyces cerevisiae and dodecameric AHAS complexes of Arabidopsis thaliana. We found that the regulatory subunits of these AHAS complexes form a core to which the catalytic subunit dimers are attached, adopting the shape of a Maltese cross. The structures show how the catalytic and regulatory subunits communicate with each other to provide a pathway for activation and for feedback inhibition by branched-chain amino acids. We also show that the AHAS complex of Mycobacterium tuberculosis adopts a similar structure, thus demonstrating that the overall AHAS architecture is conserved across kingdoms.

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

Fig. 1: The structure of the ScAHAS and AtAHAS complexes.
Fig. 2: The role of ATP in assembling the ScAHAS complex.
Fig. 3: The ScRSU influences the conformation of the ScCSU.
Fig. 4: The effects of l-valine binding on the AtAHAS complex.

Similar content being viewed by others

Data availability

The structures of the ScAHAS complex, CnCSU–ScRSU–valine complex and AtAHAS in the presence and absence of l-valine have been deposited in the Protein Data Bank with accession numbers 6U9D and 6WO1, 6VZ8 and 6U9H, respectively. The cryo-EM map has been deposited in the Electron Microscopy Data Bank under accession numbers EMDB-21487 and EMDB-20700. Plasmids, strains and raw data are available from the corresponding author upon reasonable request.

References

  1. Duggleby, R. G. & Pang, S. S. Acetohydroxyacid synthase. J. Biochem. Mol. Biol. 33, 1–36 (2000).

    CAS  Google Scholar 

  2. Garcia, M. D., Nouwens, A., Lonhienne, T. G. & Guddat, L. W. Comprehensive understanding of acetohydroxyacid synthase inhibition by different herbicide families. Proc. Natl Acad. Sci. USA 114, E1091–E1100 (2017).

    Article  CAS  Google Scholar 

  3. Garcia, M. D. et al. Commercial AHAS-inhibiting herbicides are promising drug leads for the treatment of human fungal pathogenic infections. Proc. Natl Acad. Sci. USA 115, E9649–E9658 (2018).

    Article  CAS  Google Scholar 

  4. Nandula, V. K. et al. Herbicide metabolism: crop selectivity, bioactivation, weed resistance, and regulation. Weed Sci. 67, 149–175 (2019).

    Article  Google Scholar 

  5. Schloss, J. V., Ciskanik, L. M. & Van Dyk, D. E. Origin of the herbicide binding site of acetolactate synthase. Nature 331, 360–362 (1988).

    Article  CAS  ADS  Google Scholar 

  6. Lonhienne, T., Garcia, M. D. & Guddat, L. W. The role of a FAD cofactor in the regulation of acetohydroxyacid synthase by redox signaling molecules. J. Biol. Chem. 292, 5101–5109 (2017).

    Article  CAS  Google Scholar 

  7. Pang, S. S. & Duggleby, R. G. Regulation of yeast acetohydroxyacid synthase by valine and ATP. Biochem. J. 357, 749–757 (2001).

    Article  CAS  Google Scholar 

  8. Lee, Y. T. & Duggleby, R. G. Identification of the regulatory subunit of Arabidopsis thaliana acetohydroxyacid synthase and reconstitution with its catalytic subunit. Biochemistry 40, 6836–6844 (2001).

    Article  CAS  Google Scholar 

  9. Lang, E. J. M., Cross, P. J., Mittelstädt, G., Jameson, G. B. & Parker, E. J. Allosteric ACTion: the varied ACT domains regulating enzymes of amino-acid metabolism. Curr. Opin. Struct. Biol. 29, 102–111 (2014).

    Article  CAS  Google Scholar 

  10. Bansal, A., Karanth, N. M., Demeler, B., Schindelin, H. & Sarma, S. P. Crystallographic structures of IlvN·Val/Ile complexes: conformational selectivity for feedback inhibition of aceto hydroxyl acid synthases. Biochemistry 58, 1992–2008 (2019).

    Article  CAS  Google Scholar 

  11. Aravind, L. & Koonin, E. V. Gleaning non-trivial structural, functional and evolutionary information about proteins by iterative database searches. J. Mol. Biol. 287, 1023–1040 (1999).

    Article  CAS  Google Scholar 

  12. McCourt, J. A. & Duggleby, R. G. Acetohydroxyacid synthase and its role in the biosynthetic pathway for branched-chain amino acids. Amino Acids 31, 173–210 (2006).

    Article  CAS  Google Scholar 

  13. Lonhienne, T., Garcia, M. D., Fraser, J. A., Williams, C. M. & Guddat, L. W. The 2.0 Å X-ray structure for yeast acetohydroxyacid synthase provides new insights into its cofactor and quaternary structure requirements. PLoS ONE 12, e0171443 (2017).

    Article  Google Scholar 

  14. Kaplun, A. et al. Structure of the regulatory subunit of acetohydroxyacid synthase isozyme III from Escherichia coli. J. Mol. Biol. 357, 951–963 (2006).

    Article  CAS  ADS  Google Scholar 

  15. Petkowski, J. J. et al. Crystal structures of TM0549 and NE1324—two orthologs of E. coli AHAS isozyme III small regulatory subunit. Protein Sci. 16, 1360–1367 (2007).

    Article  CAS  Google Scholar 

  16. Xie, Y. et al. Interactions between the ACT domains and catalytic subunits of acetohydroxyacid synthases (AHASs) from different species. ChemBioChem 19, 2387–2394 (2018).

    Article  CAS  Google Scholar 

  17. Lonhienne, T. et al. Commercial herbicides can trigger the oxidative inactivation of acetohydroxyacid synthase. Angew. Chem. Int. Ed. 55, 4247–4251 (2016).

    Article  CAS  Google Scholar 

  18. Barak, Z. & Chipman, D. M. Allosteric regulation in acetohydroxyacid synthases (AHASs)—different structures and kinetic behavior in isozymes in the same organisms. Arch. Biochem. Biophys. 519, 167–174 (2012).

    Article  CAS  Google Scholar 

  19. Lonhienne, T. et al. Structural insights into the mechanism of inhibition of AHAS by herbicides. Proc. Natl Acad. Sci. USA 115, E1945–E1954 (2018).

    Article  CAS  Google Scholar 

  20. Lonhienne, T. et al. High resolution crystal structures of the acetohydroxyacid synthase-pyruvate complex provide new insights into its catalytic mechanism. ChemistrySelect 2, 11981–11988 (2017).

    Article  CAS  Google Scholar 

  21. Pang, S. S., Duggleby, R. G. & Guddat, L. W. Crystal structure of yeast acetohydroxyacid synthase: a target for herbicidal inhibitors. J. Mol. Biol. 317, 249–262 (2002).

    Article  CAS  Google Scholar 

  22. Belenky, I. et al. Many of the functional differences between acetohydroxyacid synthase (AHAS) isozyme I and other AHASs are a result of the rapid formation and breakdown of the covalent acetolactate–thiamin diphosphate adduct in AHAS I. FEBS J. 279, 1967–1979 (2012).

    Article  CAS  Google Scholar 

  23. Dai, S. et al. Low-barrier hydrogen bonds in enzyme cooperativity. Nature 573, 609–613 (2019).

    Article  CAS  ADS  Google Scholar 

  24. Lonhienne, T., Gerday, C. & Feller, G. Psychrophilic enzymes: revisiting the thermodynamic parameters of activation may explain local flexibility. Biochim. Biophys. Acta 1543, 1–10 (2000).

    Article  CAS  Google Scholar 

Download references

Acknowledgements

Crystallization was performed at the University of Queensland Remote-Operation Crystallization and X-ray diffraction facility (UQROCX). Data collection was carried out at the Australian Synchrotron, part of ANSTO and made use of the Australian Cancer Research Foundation (ACRF) detector. This work was supported by funds from the National Health and Medical Research Council (grant numbers 1087713 and 1147297), and the Australian Government Department of Agriculture and Water Resources (CT-06). We thank the Bio-Electron Microscopy facility (iHuman Institute) at ShanghaiTech University and the Cryo-EM Facility Center of Southern University of Science and Technology (Shenzhen) for providing support and acknowledge the facilities and scientific and technical assistance of the Australian Microscopy & Microanalysis Research Facility at the Centre for Microscopy and Microanalysis, The University of Queensland. T.C. is supported by a Wellcome Trust grant (209407/Z/17/Z), awarded to R. Read. We thank R. Duggleby for initiating the early functional and structural studies on AHASs.

Author information

Authors and Affiliations

Authors

Contributions

L.W.G., T.L., G.S., Z.R., C.M.W., R.P.M., J.A.F. and N.P.W. conceived the project. T.L. and Y.S.L. established the protein-purification protocols, generated mutants and performed kinetic experiments. M.D.G. performed the initial purification and cryo-EM experiments with AtAHAS. Y.S.L., Y.G., Q.W. and M.J.L. collected and processed the cryo-EM data. T.L. crystallized and solved the structures of the fungal AHASs. T.L. and L.W.G. collected the X-ray data for the fungal AHASs. T.L., Y.S.L., L.W.G., G.S. and M.J.L. interpreted the data. L.B. assisted with cryo-EM and ReFyn analysis. T.C. performed the final refinement steps for the cryo-EM structure of AtAHAS in complex with l-valine. T.L. and Y.S.L. prepared the figures. T.L., Y.S.L., L.W.G. and G.S. wrote the paper and T.L., Y.S.L., L.W.G., G.S., C.M.W., M.J.L., Z.R., N.P.W., J.A.F., R.P.M. and T.C. edited the paper.

Corresponding authors

Correspondence to Thierry Lonhienne or Luke W. Guddat.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature thanks Robert Sammons, Kai Tittmann and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

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 The assembly of the ScAHAS complex.

a, General view of the interactions that stabilize the ScAHAS complex. The 16 chains in the complex are signified by the letters A–P. Each RSU has N-terminal and C-terminal extensions that span through the core of the structure. As an example, two regions of the N-terminal extension of the RSU (chain L) form interactions with four neighbouring subunits (chains D (top inset), O, P (middle inset) and M (bottom inset)). b, The assembly of the ScRSU octamer monitored by mass photometry (ReFeyn, see Supplementary Methods). For this experiment, the extended form of the RSU was used (RSUE; see Supplementary Methods). In the absence of ATP (left), the majority of the RSUs (97%) have an average molecular mass of 63 kDa, corresponding to the RSUE (31.4 kDa) dimer. In the presence of 1 mM ATP (right), there are three peaks that correspond to the dimer (34%), the tetramer (13%) and the octamer (49%). Partial assembly of the octamer was observed because the RSU could only be used at a maximum of 130 nM in this experiment due to saturation of the detector (Supplementary Methods). c, d, Stereo views of the binding of ATP to the ScAHAS complex. c, The coordination of the six phosphates by Mg2+ and the stabilization of their negative charges by six arginine and two lysine residues from two different RSUs. d, Stabilization of the adenine and ribosyl moieties of one of the ATPs by three RSUs. Dashed black lines represent electrostatic interactions. Thick dashed green lines represent π–π interactions.

Extended Data Fig. 2 Structure of a catalytic centre in the CSU dimer of the ScAHAS complex.

a, Stereo view of the superimposition of the catalytic centres of a CSU dimer in the CSU–RSU complex (light brown) and the CSU dimer in the absence of the RSU (light blue) (PDB code 5FEM). The conformation of the active site and the location of ThDP, Mg2+ and FAD are unaltered, as well as the binding mode of the herbicide, BSM. FAD, ThDP and BSM are represented as magenta and orange ball-and-stick models in the structures of the ScAHAS CSU–RSU complex and of the ScAHAS CSU alone, respectively. b. KM for pyruvate for the only CSU of ScAHAS (left; assayed using 0.28 μM CSU) and for the ScAHAS complex (right; assayed using 0.14 μM CSU, 1 mM ATP and 8.5 μM RSU). These results appear to contradict previous studies that show that the KM of pyruvate increases in presence of the ScRSU (2.1 mM for CSU alone and 13.3 mM for the complex)7. However, this discrepancy can be explained by the different buffers used in the assay reported here and those performed previously7 (Supplementary Methods). Data are mean ± s.e.m. of technical triplicates. Data were fitted to the Michaelis–Menten equation to obtain the KM values. c, Stereo view of the superimposition of the α–β linker and nearby regions in the structures of the ScAHAS complex (light brown) and CSU alone (light blue). The dashed line indicates residues that are disordered (not visible) in the structure of CSU alone.

Extended Data Fig. 3 Interactions between a ScCSU dimer and neighbouring RSUs in the ScAHAS complex.

a, Stereo view of a CSU dimer and interacting RSUs. The CSU dimer is represented by chains A (dark grey) and B (light brown), and includes the cofactors FAD (yellow transparent surface) and ThDP (orange transparent surface) and the herbicide BSM (blue transparent surface). The RSU dimer is represented by chains C (green) and D (magenta). The two bound ATP pairs are shown as a red transparent surfaces. Chain G (cyan) and chain P (violet) are RSUs that interact with chain B and chain A, respectively, through their N-terminal extensions. b. Stereo view and magnification of the α-helixes of the ACT domain of RSU (chain C). The colours of the chains and cofactors are as in a. BSM has been omitted for clarity. Dashed black lines represent electrostatic bonds. The first α-helix (α1) of the ACT domain in the RSU interacts with the Q-loop of the CSU (chain B) (that is, R101 interacts with Q202) whereas the second α-helix (α2) interacts with the β-domain of the CSU (chain A) (that is, N134 and R137 interact with I413 and P410, respectively). c, Stereo view and magnification of the α–β linker of a CSU. The chains are coloured as in a. Dashed black lines represent electrostatic bonds and white thick dashed lines represent hydrophobic interactions. The α–β linker of CSU (chain B) interacts with the ‘contact loop’ of the RSU (chain C), two β-strands of the ACT domain of the RSU (chain D) and the N-terminal extension of the RSU (chain G).

Extended Data Fig. 4 Regulatory mechanism of ScAHAS.

a, The specific activity of ScCSU measured at 30 °C as a function of the concentration of ScCSU, in the presence (blue data, right y axis) of saturating concentrations of ScRSU (8.5 μM) and ATP (1 mM) or in their absence (black data, left y axis). Data (technical triplicates) were fitted using Supplementary equation (3) (Supplementary Methods), yielding a Kd of 0.033 ± 0.005 μM for the CSU in the presence of the RSU and ATP, and Kd = 0.63 ± 0.05 μM for the CSU alone. Data are mean ± s.e.m. b, Measurement of the energy of activation (Ea) of ScAHAS. Arrhenius plots were generated for the CSU alone (black data, technical triplicates) and for CSU in the presence of saturating concentrations of RSU and ATP (blue data, technical triplicates) (Supplementary Methods). Data points between 24 and 36 °C (7 temperatures) were used to perform a linear regression calculation to obtain an Ea value (the absolute value of the slope multiplied by the gas constant value24 (R = 8.314 kJ/mol)) of 51.7 ± 1.5 kJ/mol for the CSU in presence of the RSU and ATP, and Ea = 51.5 ± 2.1 kJ/mol for the CSU alone. Data are mean ± s.e.m. (s.e.m. of the slope value multiplied by R). c, Measurement of the thermodynamic parameters of the activation at 30 °C. Left, the kcat of ScCSU at 30 °C was measured in the presence (blue data, right y axis) of saturating amounts of ScRSU (33 μM) and ATP (1 mM) or in their absence (black data, left y axis). Data points are technical replicates (n = 5). Data are mean ± s.e.m. of the five replicates. Right, values for the thermodynamic parameters of the activation at 30 °C. The values of the change in Gibbs free energy (ΔG#), enthalpy (ΔH#), and temperature and entropy (TΔS#) were calculated using the equations derived from the transition state theory24. The comparison of ΔH# (49.2 versus 48.9 kJ/mol) and TΔS# (−16.23 versus −20 .3 kJ/mol) values (of CSU + RSU versus CSU, respectively) that account for the value of ΔG#G# = ΔH# − TΔS#) revealed that the decrease in ΔG# (which is directly associated with the increase in kcat) is largely entropy-driven. Data are mean ± s.e.m. of the kcat (left) for ΔG# and of the Ea for ΔH# and TΔS#, according to the literature24. d, e, Effect of the R101A mutation on the activity of ScRSU. d, Activation of ScAHAS by ScRSUE(WT) or ScRSUE(R101A). The specific activity of ScAHAS (130 nM in the assay) was measured at different RSU concentrations with a saturating concentration of ATP (1 mM). The data (technical triplicates) were fitted to the Hill equation, showing that the mutation increases the Kd between the CSU and the RSU from 0.27 ± 0.02 (ScRSUE(WT)) to 6 ± 0.3 μM (ScRSUE(R101A)). Data are mean ± s.e.m. e, The specific activity of the ScCSUs as a function of their concentration, alone (black data) and in the presence of 1 mM ATP and 2.6 μM ScRSUE(WT) (blue data, saturating concentration) or 20 μM ScRSUE(R101A) (red data, saturating concentration). The data (technical triplicates) were fitted to Supplementary equation (3) (Supplementary Methods), yielding a Kd of 0.04 ± 0.005 and 0.036 ± 0.007 μM for the CSU in the presence of RSUE(WT) and RSUE(R101A), respectively. For the CSU alone, the Kd is 0.62 ± 0.04 μM confirming the value shown in a. The values of kcat—which are also derived from Supplementary equation (3)—are 11.1 ± 0.2, 40.9 ± 1.1 and 19.5 ± 0.8 s−1 for the CSU alone, the CSU with RSUE(WT) and the CSU with RSUE(R101A), respectively, showing that the R101A mutation strongly affects catalysis. Data are mean ± s.e.m.

Extended Data Fig. 5 Model for the regulation of the ScAHAS reaction.

A hybrid image showing the catalytic centres taken from an asymmetric ScCSU dimer in the presence of pyruvate (grey and brown) (PDB code 5INU) and the α1-helices (pink and green) of the RSU from the complex. The different conformations of FAD and the Q-loops, and the He-ThDP intermediate or ThDP are shown. The Q-loops and their N-terminal and C-terminal extensions have been drawn in brown and dark grey for catalytic centre 1 and 2, respectively. The model shows how the conformational change of the active sites during the catalytic cycle20 triggers a movement of the Q-loops that is complementary between the catalytic centres (red arrows). This feature facilitates the synchronization of the reactions that occur in the two catalytic centres of a CSU dimer13. In the ScCSU dimer, direct contact between the Q-loops is made through their N- and C-terminal extensions, which transmit the movements between catalytic centres (red arrows). When the complex is formed, the RSU works at two levels. (1) The RSU creates a ‘rigid scaffold’ (magenta) that is formed by interactions of the ACT regions (that is, the α2-helix and two β-strands of the ACT domain) with the α- and β-domains and the α–β linker of the CSUs (Extended Data Fig. 3). This restrict movements of the CSU that is not involved in the communication between active sites. (ii) The α1-helices of the RSUs create a bridge (α1-bridge) between Q-loops (curve line in magenta) that increases the efficiency of the mechanical transmission between catalytic centres. Both aspects are involved in decreasing the entropy of the enzyme–substrate complex before activation, leading to an increase in kcat.

Extended Data Fig. 6 Binding site interactions, cryo-EM map for l-valine in AtRSU and consequences of l-valine binding.

a, Cryo-EM densities of the two BCAA-binding sites (site 1 and 2) occupied by l-valine. l-Valine is shown as a green stick model. The cryo-EM map (5.5σ) is overlaid. b, Stereo view of the interaction of l-valine in site 1 with the α1-helices of the ACT domains. Dashed lines represent hydrogen bonds (black), van der Waals interactions (blue) and hydrophobic interactions (yellow). c, Superimposition of the free AtAHAS complex and the l-valine-bound complex, representing one of the four CSU RSU pairs that interact with each other. The CSU dimers are shown in green and light blue for structures of the complex in the absence and presence of l-valine, respectively. The RSUs are shown in dark green and magenta for the repeats in the structure of the complex in the absence of l-valine, and cyan and coral for the repeats in the structure of the complex with l-valine bound. The cofactors FAD (yellow) and ThDP (orange) are shown in the structure of the l-valine free complex only. l-Valine (ball and stick representation) has a light-yellow transparent surface around the outside. The comparison of the structures shows that the CSU dimer expands when l-valine is bound to the ACT domains of the RSUs.

Extended Data Fig. 7 Structural changes in the ACT domains of AtRSU that are triggered by the binding of l-valine.

a, Stereo view of the translocation of R110 induced by the binding of l-valine in site 1 of the ACT domain. When l-valine is present the backbone of the α1-helix of repeat 1 is shifted (black arrows), causing a marked translocation of R110 and the breaking of the R110–D204 salt bridge. b, Stereo view showing the conformational change in α2 caused by the bending of α1. The conformational change of α1 induced by l-valine triggers new interactions to occur between α1 and α2 (that is, the interaction of V138 (α2) with R110 (α1) and L139 (α2) with Y112 (α1)) leads to the bending of α2). The colour scheme is as described in Extended Data Fig. 6c.

Extended Data Fig. 8 The structure of the fungal hybrid complex formed by the CnCSU and ScRSU in the presence of l-valine and comparisons with the AtAHAS complexes.

a, The overall structure of the hybrid complex resembles the ScAHAS and AtAHAS complexes, possessing the characteristic Maltese cross shape. b, The addition of ATP leads to the activation of the fungal hybrid complex, whereas the addition of l-valine inhibits its activity. This experiment was performed once. c, Left, the cryo-EM map at 3.4 Å resolution (8σ level) in the AtAHAS complex with l-valine overlaid and optimally fitted. Middle, the hybrid CnScAHAS complex with the Fo − Fc electron density (3.5σ level) with l-valine overlaid and optimally fitted. Right, the binding mode of l-valine to E. coli RSU (structure determined in the absence of the CSU at 2.3 Å resolution; PDB code 5YPW). In all three complexes, the carboxylate group of l-valine forms three hydrogen bonds to backbone amides of the RSUs (I114, V332, L333 in AtAHAS; I105, V90, L91 in CnScAHAS; V38, V23, M24 in the E. coli RSU alone). The amide group of l-valine forms either two or three hydrogen bonds to the RSU (N113 side chain, I114 carbonyl oxygen, D328 side chain in AtAHAS; N104 side chain, I105 carbonyl oxygen in CnScAHAS; N37 side chain, V38 carbonyl oxygen, Q19 side chain in E. coli RSU alone). The side chain of l-valine is stabilized by hydrophobic and/or van der Waals interactions with nearby hydrophobic side chains. The similarity of the network of interactions involving the binding of l-valine that is observed in all three complexes, and the fact that these three sites align when the sequences are compared (Supplementary Fig. 2) shows that the binding mode of l-valine is conserved in bacterial, fungal and plant AHASs. d, Comparison of the conformational changes that occur in the AtAHAS and in the CnScAHAS complexes on l-valine binding. These changes are highly conserved in the two complexes as emphasized by the arrows.

Extended Data Table 1 Data collection and refinement statistics for ScAHAS complexes
Extended Data Table 2 Cryo-EM data collection, refinement and validation statistics for the AtAHAS complexes

Supplementary information

Supplementary Information

This file contains the Supplementary Materials and Methods used for the experiments conducted in this Article. It also contains Supplementary References, Supplementary Table 1 and Supplementary Figures 1-18.

Reporting Summary

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lonhienne, T., Low, Y.S., Garcia, M.D. et al. Structures of fungal and plant acetohydroxyacid synthases. Nature 586, 317–321 (2020). https://doi.org/10.1038/s41586-020-2514-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41586-020-2514-3

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