Skip to main content

Thank you for visiting 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:

Mutations in nuclear pore genes NUP93, NUP205 and XPO5 cause steroid-resistant nephrotic syndrome


Nucleoporins are essential components of the nuclear pore complex (NPC)1. Only a few diseases have been attributed to NPC dysfunction2,3,4. Steroid-resistant nephrotic syndrome (SRNS), a frequent cause of chronic kidney disease, is caused by dysfunction of glomerular podocytes5. Here we identify in eight families with SRNS mutations in NUP93, its interaction partner NUP205 or XPO5 (encoding exportin 5) as hitherto unrecognized monogenic causes of SRNS. NUP93 mutations caused disrupted NPC assembly. NUP93 knockdown reduced the presence of NUP205 in the NPC, and, reciprocally, a NUP205 alteration abrogated NUP93 interaction. We demonstrate that NUP93 and exportin 5 interact with the signaling protein SMAD4 and that NUP93 mutations abrogated interaction with SMAD4. Notably, NUP93 mutations interfered with BMP7-induced SMAD transcriptional reporter activity. We hereby demonstrate that mutations of NUP genes cause a distinct renal disease and identify aberrant SMAD signaling as a new disease mechanism of SRNS, opening a potential new avenue for treatment.

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

Access options

Buy this article

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

Figure 1: Homozygosity mapping and whole-exome sequencing identify recessive mutations in NUP93, NUP205 or XPO5 in eight families with SRNS.
Figure 2: Subcellular localization of NUP93 in podocytes and NUP93 knockdown resulting in reduced podocyte migration and proliferation and impaired resistance to oxidative stress.
Figure 3: NUP93 mutations interfere with NPC assembly, and a NUP205 mutation affects NUP93-NUP205 interaction.
Figure 4: NUP93, importin 7 and exportin 5 interact with SMAD proteins, and NUP93 mutations from individuals with SRNS abrogate the interaction of NUP93 with SMAD4 and importin 7 in coimmunoprecipitation in HEK293T cells.
Figure 5: Knockdown of NUP93 interferes with BMP7-induced activation of SMAD signaling.

Similar content being viewed by others

Accession codes


NCBI Reference Sequence

Protein Data Bank


  1. Grossman, E., Medalia, O. & Zwerger, M. Functional architecture of the nuclear pore complex. Annu. Rev. Biophys. 41, 557–584 (2012).

    Article  CAS  Google Scholar 

  2. Basel-Vanagaite, L. et al. Mutated nup62 causes autosomal recessive infantile bilateral striatal necrosis. Ann. Neurol. 60, 214–222 (2006).

    Article  CAS  Google Scholar 

  3. Tullio-Pelet, A. et al. Mutant WD-repeat protein in triple-A syndrome. Nat. Genet. 26, 332–335 (2000).

    Article  CAS  Google Scholar 

  4. Zhang, X. et al. Mutation in nuclear pore component NUP155 leads to atrial fibrillation and early sudden cardiac death. Cell 135, 1017–1027 (2008).

    Article  CAS  Google Scholar 

  5. Machuca, E., Benoit, G. & Antignac, C. Genetics of nephrotic syndrome: connecting molecular genetics to podocyte physiology. Hum. Mol. Genet. 18, R2, R185–R194 (2009).

    Article  Google Scholar 

  6. Smith, J.M., Stablein, D.M., Munoz, R., Hebert, D. & McDonald, R.A. Contributions of the Transplant Registry: The 2006 Annual Report of the North American Pediatric Renal Trials and Collaborative Studies (NAPRTCS). Pediatr. Transplant. 11, 366–373 (2007).

    Article  Google Scholar 

  7. Wiggins, R.C. The spectrum of podocytopathies: a unifying view of glomerular diseases. Kidney Int. 71, 1205–1214 (2007).

    Article  CAS  Google Scholar 

  8. Gee, H.Y. et al. ARHGDIA mutations cause nephrotic syndrome via defective RHO GTPase signaling. J. Clin. Invest. 123, 3243–3253 (2013).

    Article  CAS  Google Scholar 

  9. Gee, H.Y. et al. KANK deficiency leads to podocyte dysfunction and nephrotic syndrome. J. Clin. Invest. 125, 2375–2384 (2015).

    Article  Google Scholar 

  10. Ashraf, S. et al. ADCK4 mutations promote steroid-resistant nephrotic syndrome through CoQ10 biosynthesis disruption. J. Clin. Invest. 123, 5179–5189 (2013).

    Article  CAS  Google Scholar 

  11. Sadowski, C.E. et al. A single-gene cause in 29.5% of cases of steroid-resistant nephrotic syndrome. J. Am. Soc. Nephrol. 26, 1279–1289 (2015).

    Article  CAS  Google Scholar 

  12. Hildebrandt, F. et al. A systematic approach to mapping recessive disease genes in individuals from outbred populations. PLoS Genet. 5, e1000353 (2009).

    Article  Google Scholar 

  13. Otto, E.A. et al. Candidate exome capture identifies mutation of SDCCAG8 as the cause of a retinal-renal ciliopathy. Nat. Genet. 42, 840–850 (2010).

    Article  CAS  Google Scholar 

  14. Halbritter, J. et al. High-throughput mutation analysis in patients with a nephronophthisis-associated ciliopathy applying multiplexed barcoded array-based PCR amplification and next-generation sequencing. J. Med. Genet. 49, 756–767 (2012).

    Article  CAS  Google Scholar 

  15. Halbritter, J. et al. Identification of 99 novel mutations in a worldwide cohort of 1,056 patients with a nephronophthisis-related ciliopathy. Hum. Genet. 132, 865–884 (2013).

    Article  CAS  Google Scholar 

  16. Grandi, P., Doye, V. & Hurt, E.C. Purification of NSP1 reveals complex formation with 'GLFG' nucleoporins and a novel nuclear pore protein NIC96. EMBO J. 12, 3061–3071 (1993).

    Article  CAS  Google Scholar 

  17. Vollmer, B. & Antonin, W. The diverse roles of the Nup93/Nic96 complex proteins—structural scaffolds of the nuclear pore complex with additional cellular functions. Biol. Chem. 395, 515–528 (2014).

    Article  CAS  Google Scholar 

  18. Allende, M.L. et al. Insertional mutagenesis in zebrafish identifies two novel genes, pescadillo and dead eye, essential for embryonic development. Genes Dev. 10, 3141–3155 (1996).

    Article  CAS  Google Scholar 

  19. Hildebrandt, F. & Heeringa, S.F. Specific podocin mutations determine age of onset of nephrotic syndrome all the way into adult life. Kidney Int. 75, 669–671 (2009).

    Article  CAS  Google Scholar 

  20. Ruf, R.G. et al. Patients with mutations in NPHS2 (podocin) do not respond to standard steroid treatment of nephrotic syndrome. J. Am. Soc. Nephrol. 15, 722–732 (2004).

    Article  Google Scholar 

  21. Amlacher, S. et al. Insight into structure and assembly of the nuclear pore complex by utilizing the genome of a eukaryotic thermophile. Cell 146, 277–289 (2011).

    Article  CAS  Google Scholar 

  22. Brownawell, A.M. & Macara, I.G. Exportin-5, a novel karyopherin, mediates nuclear export of double-stranded RNA binding proteins. J. Cell Biol. 156, 53–64 (2002).

    Article  CAS  Google Scholar 

  23. Bohnsack, M.T. et al. Exp5 exports eEF1A via tRNA from nuclei and synergizes with other transport pathways to confine translation to the cytoplasm. EMBO J. 21, 6205–6215 (2002).

    Article  CAS  Google Scholar 

  24. Lund, E., Güttinger, S., Calado, A., Dahlberg, J.E. & Kutay, U. Nuclear export of microRNA precursors. Science 303, 95–98 (2004).

    Article  CAS  Google Scholar 

  25. Wente, S.R. & Rout, M.P. The nuclear pore complex and nuclear transport. Cold Spring Harb. Perspect. Biol. 2, a000562 (2010).

    Article  CAS  Google Scholar 

  26. Ebarasi, L. et al. Defects of CRB2 cause steroid-resistant nephrotic syndrome. Am. J. Hum. Genet. 96, 153–161 (2015).

    Article  CAS  Google Scholar 

  27. Gee, H.Y. et al. Mutations in EMP2 cause childhood-onset nephrotic syndrome. Am. J. Hum. Genet. 94, 884–890 (2014).

    Article  CAS  Google Scholar 

  28. Gant, T.M. & Wilson, K.L. Nuclear assembly. Annu. Rev. Cell Dev. Biol. 13, 669–695 (1997).

    Article  CAS  Google Scholar 

  29. Sachdev, R., Sieverding, C., Flötenmeyer, M. & Antonin, W. The C-terminal domain of Nup93 is essential for assembly of the structural backbone of nuclear pore complexes. Mol. Biol. Cell 23, 740–749 (2012).

    Article  CAS  Google Scholar 

  30. Grandi, P. et al. Nup93, a vertebrate homologue of yeast Nic96p, forms a complex with a novel 205-kDa protein and is required for correct nuclear pore assembly. Mol. Biol. Cell 8, 2017–2038 (1997).

    Article  CAS  Google Scholar 

  31. Theerthagiri, G., Eisenhardt, N., Schwarz, H. & Antonin, W. The nucleoporin Nup188 controls passage of membrane proteins across the nuclear pore complex. J. Cell Biol. 189, 1129–1142 (2010).

    Article  CAS  Google Scholar 

  32. Chen, X. & Xu, L. Specific nucleoporin requirement for Smad nuclear translocation. Mol. Cell. Biol. 30, 4022–4034 (2010).

    Article  CAS  Google Scholar 

  33. Yao, X., Chen, X., Cottonham, C. & Xu, L. Preferential utilization of Imp7/8 in nuclear import of Smads. J. Biol. Chem. 283, 22867–22874 (2008).

    Article  CAS  Google Scholar 

  34. Dudley, A.T., Lyons, K.M. & Robertson, E.J. A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye. Genes Dev. 9, 2795–2807 (1995).

    Article  CAS  Google Scholar 

  35. Nakamura, J. & Yanagita, M. Bmp modulators in kidney disease. Discov. Med. 13, 57–63 (2012).

    PubMed  Google Scholar 

  36. Meng, X.M. et al. Disruption of Smad4 impairs TGF-β/Smad3 and Smad7 transcriptional regulation during renal inflammation and fibrosis in vivo and in vitro. Kidney Int. 81, 266–279 (2012).

    Article  CAS  Google Scholar 

  37. Mitu, G.M., Wang, S. & Hirschberg, R. BMP7 is a podocyte survival factor and rescues podocytes from diabetic injury. Am. J. Physiol. Renal Physiol. 293, F1641–F1648 (2007).

    Article  CAS  Google Scholar 

  38. Leeuwis, J.W. et al. Direct visualization of Smad1/5/8-mediated transcriptional activity identifies podocytes and collecting ducts as major targets of BMP signalling in healthy and diseased kidneys. J. Pathol. 224, 121–132 (2011).

    Article  CAS  Google Scholar 

  39. Klaassen, I. et al. Response to cyclosporine in steroid-resistant nephrotic syndrome: discontinuation is possible. Pediatr. Nephrol. 30, 1477–1483 (2015).

    Article  Google Scholar 

  40. Kruglyak, L., Daly, M.J., Reeve-Daly, M.P. & Lander, E.S. Parametric and nonparametric linkage analysis: a unified multipoint approach. Am. J. Hum. Genet. 58, 1347–1363 (1996).

    CAS  PubMed  PubMed Central  Google Scholar 

  41. Strauch, K. et al. Parametric and nonparametric multipoint linkage analysis with imprinting and two-locus-trait models: application to mite sensitization. Am. J. Hum. Genet. 66, 1945–1957 (2000).

    Article  CAS  Google Scholar 

  42. Gudbjartsson, D.F., Jonasson, K., Frigge, M.L. & Kong, A. Allegro, a new computer program for multipoint linkage analysis. Nat. Genet. 25, 12–13 (2000).

    Article  CAS  Google Scholar 

  43. Sayer, J.A. et al. The centrosomal protein nephrocystin-6 is mutated in Joubert syndrome and activates transcription factor ATF4. Nat. Genet. 38, 674–681 (2006).

    Article  CAS  Google Scholar 

  44. Boyden, L.M. et al. Mutations in kelch-like 3 and cullin 3 cause hypertension and electrolyte abnormalities. Nature 482, 98–102 (2012).

    Article  CAS  Google Scholar 

  45. Watanabe, M., Masuyama, N., Fukuda, M. & Nishida, E. Regulation of intracellular dynamics of Smad4 by its leucine-rich nuclear export signal. EMBO Rep. 1, 176–182 (2000).

    Article  CAS  Google Scholar 

  46. Saleem, M.A. et al. A conditionally immortalized human podocyte cell line demonstrating nephrin and podocin expression. J. Am. Soc. Nephrol. 13, 630–638 (2002).

    CAS  Google Scholar 

  47. Zariwala, M.A. et al. ZMYND10 is mutated in primary ciliary dyskinesia and interacts with LRRC6. Am. J. Hum. Genet. 93, 336–345 (2013).

    Article  CAS  Google Scholar 

  48. Eisenhardt, N., Schooley, A. & Antonin, W. Xenopus in vitro assays to analyze the function of transmembrane nucleoporins and targeting of inner nuclear membrane proteins. Methods Cell Biol. 122, 193–218 (2014).

    Article  CAS  Google Scholar 

  49. Franz, C. et al. MEL-28/ELYS is required for the recruitment of nucleoporins to chromatin and postmitotic nuclear pore complex assembly. EMBO Rep. 8, 165–172 (2007).

    Article  CAS  Google Scholar 

Download references


We are grateful to the families and study individuals for their contributions. We thank the Yale Center for Mendelian Genomics for whole-exome sequencing analysis and M. Mihatsch and H. Hopfer (Basel, Switzerland) for providing histology data. This research was supported by a grant from the US National Institutes of Health to F.H. (DK076683) and by core funding of the Max Planck Society to W.A. H.Y.G. is supported by the National Research Foundation of Korea, Ministry of Science, ICT and Future planning (2015R1D1A1A01056685) and faculty seat money from the Yonsei University College of Medicine (2015-32-0047). W.T. is supported by the ASN Foundation for Kidney Research. F.O. is supported by the European Community's Seventh Framework Programme (FP7/2007-2013) (EURenOmics; grant 2012-305608). The Nephrogenetics Laboratory at Hacettepe University was established by the Hacettepe University Infrastructure Project (grant 06A101008). F.H. was also supported by the Howard Hughes Medical Institute, the Doris Duke Charitable Foundation and a Warren E. Grupe Professorship.

Author information

Authors and Affiliations



D.A.B., C.E.S., S.K., S.L., W.L.P., S.A., H.Y.G., S.S., M.A., W.T., J.R., D.S., W.-I.C., T.H., R.P.L. and F.H. generated whole-genome linkage data, performed exome capture with massively parallel sequencing, and performed whole-exome evaluation and mutation analysis. D.A.B. performed luciferase reporter gene assays, podocyte in vitro experiments and coimmunoprecipitation experiments. D.A.B., S.K. and J.A.L. performed immunofluorescence and subcellular localization studies in tissues and cell lines by confocal microscopy. W.A. and S.A.A. performed depletion and add-back assays and coimmunoprecipitation experiments analyzing Nup93-Nup205 interaction. M.J.K., M.P., F.O., M.K., R. Bogdanovic, R. Büscher, E.S., U.H., D.A.B., C.E.S., S.K., S.L., W.L.P., S.A., H.Y.G., S.S. and F.H. recruited patients and gathered detailed clinical information for the study. All authors critically reviewed the manuscript. F.H. conceived and directed the project and wrote the manuscript.

Corresponding author

Correspondence to Friedhelm Hildebrandt.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Homozygosity mapping and WES identify NUP93 mutations in two additional families with steroid-resistant nephrotic syndrome.

(a) Homozygosity mapping identifies recessive candidate loci. In two families with nephrotic syndrome, A1626 and A2241, non-parametric LOD scores (NPL) were calculated and plotted across the human genome. The x axis shows Affymetrix 250K StyI array SNP positions on human chromosomes concatenated from the p terminus (left) to the q terminus (right). Genetic distance is given in cM. Maximum NPL peaks (red circles) indicate candidate regions of homozygosity by descent. The NUP93 locus (arrowhead) is positioned within one of the maximum NPL peaks on chromosome 16q. (b) Primer location for RT-PCR analysis of an individual with a splice-site mutation in NUP93 (red and blue arrows, referring to c and d, respectively). (c,d) RT-PCR analysis of the NUP93 transcript in patient A1394-21 harboring the c.1537+1G>A mutation in intron 13 and a control individual using primers located in exons 11 and 14 (c) and exons 12 and 14 (d). In addition to the wild-type bands of 518 bp (c) or 405 bp (d) corresponding to normal splicing, bands of 326 bp (c) and 213 bp (d), respectively, were detected in the cDNA of individual A1394-21. (e) Sanger sequencing of the RT-PCR product from A1394-21 showed that the c.1537+1G>A mutation leads to aberrant splicing, with in-frame skipping of exon 13 (192 bp).

Supplementary Figure 2 Sequencing traces.

Sanger sequencing traces are shown for ten individuals from eight families with mutations in NUP93 (a), NUP205 (b) and XPO5 (c). Arrowheads denote altered nucleotides. p, paternal; m, maternal; WT, wild type.

Supplementary Figure 3 Renal histologies of individuals A2403, A3256 and A1394 with recessive mutations in NUP93.

(a) H&E staining at high magnification (63×) shows diffuse mesangial sclerosis in A2403-21 and focal-segmental glomerulosclerosis in A3256-21 and A1394-21. (b) H&E staining at low magnification (40×) shows tubular atrophy, tubular dilation with protein casts and cysts with interstitial infiltrations. (c) Transmission electron microscopy images show partial foot process effacement (arrowheads) (15,000× magnification). N/D, no data. Scale bars: 30 μm (a,b) and 2 μm (c).

Supplementary Figure 4 Interaction of nuclear pore complex (NPC) proteins NUP93 and NUP205 with SMAD, importin 7 (IPO7) and exportin 5 (XPO5) showing the positions of the mutations that abrogate interactions.

(a) Subcomplexes of the NPC are depicted and labeled in different colors. (b) Inset from a. NUP93 and NUP155 are tethered to the nuclear membrane by NUP53. The three-dimensional structure of NUP93 is shown in relation to the likely positions of interaction with NUP205, IPO7 or SMADs. The positions of the NUP93 and NUP205 mutations that abrogate the interaction and were found in individuals with SRNS are marked by yellow lightning signs. (c) BMP7 binding to its receptor causes phosphorylation of SMAD1/5/8, resulting in dissociation from the receptor and binding to SMAD4. The SMAD4/1/5/8 complex interacts with importin 7, enabling the SMAD complex to enter the nucleus through the NPC via NUP93 interaction, bind to cognate DNA sites and initiate transcription of SMAD downstream targets. (d) XPO5 binds to SMAD4 for nuclear export. Deletion of the nuclear export signal of SMAD4 abrogates interaction with XPO5.

Supplementary Figure 5 NUP93, NUP205 and XPO5 localize to developing podocytes.

(ac) Neonatal rat coronal kidney sections at embryonic day 16.5 p.c. were stained with antibodies against NUP93, NUP205 or XPO5 (green) and the podocyte nuclear marker WT1 (red). Note that NUP93 (a) and NUP205 (b) mark nuclear envelopes with a nuclear rim pattern in podocyte precursor cells (arrowheads) at the capillary loop stage and in other cell types in developing kidney. Anti-XPO5 antibody (c) prominently marks the nuclear content of cells in developing kidney with a dot-like pattern (arrow). Scale bars, 25 μm.

Supplementary Figure 6 Knockdown of NUP93 and its interactor NUP188 but not NUP153 impairs the migratory phenotype of human podocytes in vitro.

As compared to basal migration rate (dotted black line), the migration rate of human podocytes is increased after administration of serum as a chemoattractant (solid black line). Knockdown of NUP93 impairs the podocyte migration rate in vitro (blue line). Knockdown of NUP188, encoding a direct interaction partner of NUP93, causes a similar but less pronounced reduction in the cell migration rate (yellow line). In contrast, knockdown of the gene encoding the nuclear basket protein NUP153 does not affect cell migration (purple line).

Supplementary Figure 7 NUP93 interacts with NUP205.

GFP-tagged NUP205 precipitates Myc-tagged NUP93 upon co-overexpression in HEK293 cells. Mutant clones reflecting the alleles of NUP93 identified in individuals with SRNS do not abrogate the interaction.

Supplementary Figure 8 Myc-tagged NUP93 interacts with GFP-tagged SMAD4 upon co-overexpression in HEK293 cells.

Clones reflecting the C-terminal mutations identified in individuals with steroid-resistant nephrotic syndrome (Lys442Asn_fs*14, Gly591Val, Tyr629Cys) abrogate the interaction.

Supplementary Figure 9 Knockdown of NUP93 interferes with BMP7-induced activation of SMAD signaling.

(a) Upon knockdown of NUP93 using two different shRNAs in HEK293 cells, BMP7 fails to induce efficient translocation of SMAD4 from the cytoplasm to the nucleus as compared to negative control (scrambled shRNA). (b) Whereas transfection with wild-type Nup93 rescues the cytoplasm to nucleus translocation of SMAD4 (white arrows), all five mutants detected in individuals with SRNS fail to rescue SMAD4 translocation in HEK293 cells. Scale bars: 5 μm (a) and 25 μm (b).

Supplementary Figure 10 Knockdown of NUP93 but not NUP188 decreases SMAD reporter activity upon BMP7 stimulation.

HEK293 cells were transfected with luciferase reporter constructs under the control of a SMAD-responsive promoter allowing measurement of SMAD activity in response to BMP7 stimulation (rc-BMP7). As compared to scrambled control siRNA, knockdown of NUP93 significantly reduces SMAD reporter activity (lane 2 versus 3). Knockdown of NUP153, encoding a nuclear basket protein, induces less pronounced but significant reduction of SMAD reporter activity. Knockdown of NUP188, encoding a direct interactor of NUP93, does not reduce SMAD reporter activity.

Supplementary Figure 11 SMAD1/5 phosphorylation upon BMP7 treatment and NUP93 knockdown in HEK293 cells.

Treatment of HEK293 cells with recombinant BMP7 induces phosphorylation of SMAD1/5 after 60 min (top). Transfection of HEK293 cells with shRNA against human NUP93 results in efficient knockdown of NUP93 as compared to scrambled control (middle). The β-actin control (bottom) confirms equal loading.

Supplementary Figure 12 Characterization of the antibodies against NUP93 and NUP205 by immunoblotting.

(ac) Characterization of the anti-NUP93 antibody (mouse, Santa Cruz, sc-374400). (a) The antibody detects N-terminally Myc-tagged full-length human NUP93 (N-Myc-hs_NUP93_FL), as well as N-terminally GFP-tagged full-length mouse Nup93 (N-GFP-mm_NUP93_FL) upon overexpression in HEK293 cells. (b) Detection of endogenous NUP93 in the nuclear fraction of human podocytes at the expected relative mobility of 93 kDa. (c) Knockdown of NUP93 in human podocytes using three different siRNAs demonstrates the specificity of the detected band at ~93 kDa. (df) Characterization of the anti-NUP205 antibody (rabbit polyclonal, Sigma, HPA024574). (d) The antibody detects N-terminally GFP-tagged full-length human NUP205 (N-GFP-hs_NUP205_FL), as well as N-terminally Myc-tagged full-length human NUP205 (N-Myc-hs_NUP205_FL) upon overexpression in HEK293T cells. (e) Detection of endogenous NUP205 in the nuclear fraction of human podocytes at the expected relative mobility of 205 kDa. (f) Knockdown of NUP205 in human podocytes using four different siRNAs demonstrates the specificity of the detected band at ~205 kDa. The β-actin and SP1 controls confirm equal loading.

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–12 and Supplementary Table 1. (PDF 1484 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Braun, D., Sadowski, C., Kohl, S. et al. Mutations in nuclear pore genes NUP93, NUP205 and XPO5 cause steroid-resistant nephrotic syndrome. Nat Genet 48, 457–465 (2016).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


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