Many biomolecules undergo liquid–liquid phase separation to form liquid-like condensates that mediate diverse cellular functions1,2. Autophagy is able to degrade such condensates using autophagosomes—double-membrane structures that are synthesized de novo at the pre-autophagosomal structure (PAS) in yeast3,4,5. Whereas Atg proteins that associate with the PAS have been characterized, the physicochemical and functional properties of the PAS remain unclear owing to its small size and fragility. Here we show that the PAS is in fact a liquid-like condensate of Atg proteins. The autophagy-initiating Atg1 complex undergoes phase separation to form liquid droplets in vitro, and point mutations or phosphorylation that inhibit phase separation impair PAS formation in vivo. In vitro experiments show that Atg1-complex droplets can be tethered to membranes via specific protein–protein interactions, explaining the vacuolar membrane localization of the PAS in vivo. We propose that phase separation has a critical, active role in autophagy, whereby it organizes the autophagy machinery at the PAS.
Subscribe to Journal
Get full journal access for 1 year
only $3.83 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
All relevant data are available from the authors. Source data for gels and blots are provided as Supplementary Information. Source Data for graphs are provided with the paper.
Banani, S. F., Lee, H. O., Hyman, A. A. & Rosen, M. K. Biomolecular condensates: organizers of cellular biochemistry. Nat. Rev. Mol. Cell Biol. 18, 285–298 (2017).
Shin, Y. & Brangwynne, C. P. Liquid phase condensation in cell physiology and disease. Science 357, eaaf4382 (2017).
Suzuki, K. et al. The pre-autophagosomal structure organized by concerted functions of APG genes is essential for autophagosome formation. EMBO J. 20, 5971–5981 (2001).
Kim, J., Huang, W. P., Stromhaug, P. E. & Klionsky, D. J. Convergence of multiple autophagy and cytoplasm to vacuole targeting components to a perivacuolar membrane compartment prior to de novo vesicle formation. J. Biol. Chem. 277, 763–773 (2002).
Wang, Z. & Zhang, H. Phase separation, transition, and autophagic degradation of proteins in development and pathogenesis. Trends Cell Biol. 29, 417–427 (2019).
Yamamoto, H. et al. The intrinsically disordered protein Atg13 mediates supramolecular assembly of autophagy initiation complexes. Dev. Cell 38, 86–99 (2016).
Suzuki, K., Kubota, Y., Sekito, T. & Ohsumi, Y. Hierarchy of Atg proteins in pre-autophagosomal structure organization. Genes Cells 12, 209–218 (2007).
Yamamoto, H. et al. Atg9 vesicles are an important membrane source during early steps of autophagosome formation. J. Cell Biol. 198, 219–233 (2012).
Kawamata, T., Kamada, Y., Kabeya, Y., Sekito, T. & Ohsumi, Y. Organization of the pre-autophagosomal structure responsible for autophagosome formation. Mol. Biol. Cell 19, 2039–2050 (2008).
Cheong, H., Nair, U., Geng, J. & Klionsky, D. J. The Atg1 kinase complex is involved in the regulation of protein recruitment to initiate sequestering vesicle formation for nonspecific autophagy in Saccharomyces cerevisiae. Mol. Biol. Cell 19, 668–681 (2008).
Courchaine, E. M., Lu, A. & Neugebauer, K. M. Droplet organelles? EMBO J. 35, 1603–1612 (2016).
Wei, M. T. et al. Phase behaviour of disordered proteins underlying low density and high permeability of liquid organelles. Nat. Chem. 9, 1118–1125 (2017).
Mahen, R., Jeyasekharan, A. D., Barry, N. P. & Venkitaraman, A. R. Continuous polo-like kinase 1 activity regulates diffusion to maintain centrosome self-organization during mitosis. Proc. Natl Acad. Sci. USA 108, 9310–9315 (2011).
Kroschwald, S., Maharana, S. & Simon, A. Hexanediol: a chemical probe to investigate the material properties of membrane-less compartments. Matters 1–7, https://doi.org/10.19185/matters.201702000010 (2017).
Rabouille, C. & Alberti, S. Cell adaptation upon stress: the emerging role of membrane-less compartments. Curr. Opin. Cell Biol. 47, 34–42 (2017).
Fujioka, Y. et al. Structural basis of starvation-induced assembly of the autophagy initiation complex. Nat. Struct. Mol. Biol. 21, 513–521 (2014).
Kamada, Y. et al. Tor-mediated induction of autophagy via an Apg1 protein kinase complex. J. Cell Biol. 150, 1507–1513 (2000).
Yeh, Y. Y., Wrasman, K. & Herman, P. K. Autophosphorylation within the Atg1 activation loop is required for both kinase activity and the induction of autophagy in Saccharomyces cerevisiae. Genetics 185, 871–882 (2010).
Kamber, R. A., Shoemaker, C. J. & Denic, V. Receptor-bound targets of selective autophagy use a scaffold protein to activate the Atg1 kinase. Mol. Cell 59, 372–381 (2015).
Torggler, R. et al. Two independent pathways within selective autophagy converge to activate Atg1 kinase at the vacuole. Mol. Cell 64, 221–235 (2016).
Patel, A. et al. ATP as a biological hydrotrope. Science 356, 753–756 (2017).
Memisoglu, G., Eapen, V. V., Yang, Y., Klionsky, D. J. & Haber, J. E. PP2C phosphatases promote autophagy by dephosphorylation of the Atg1 complex. Proc. Natl Acad. Sci. USA 116, 1613–1620 (2019).
Ragusa, M. J., Stanley, R. E. & Hurley, J. H. Architecture of the Atg17 complex as a scaffold for autophagosome biogenesis. Cell 151, 1501–1512 (2012).
Woodruff, J. B., Hyman, A. A. & Boke, E. Organization and function of non-dynamic biomolecular condensates. Trends Biochem. Sci. 43, 81–94 (2018).
Scott, S. V. et al. Apg13p and Vac8p are part of a complex of phosphoproteins that are required for cytoplasm to vacuole targeting. J. Biol. Chem. 275, 25840–25849 (2000).
Jeong, H. et al. Mechanistic insight into the nucleus-vacuole junction based on the Vac8p–Nvj1p crystal structure. Proc. Natl Acad. Sci. USA 114, E4539–E4548 (2017).
Milovanovic, D., Wu, Y., Bian, X. & De Camilli, P. A liquid phase of synapsin and lipid vesicles. Science 361, 604–607 (2018).
Spitzer, M., Wildenhain, J., Rappsilber, J. & Tyers, M. BoxPlotR: a web tool for generation of box plots. Nat. Methods 11, 121–122 (2014).
Ward, J. J., Sodhi, J. S., McGuffin, L. J., Buxton, B. F. & Jones, D. T. Prediction and functional analysis of native disorder in proteins from the three kingdoms of life. J. Mol. Biol. 337, 635–645 (2004).
Adams, A., Gottschling, D. E., Kaiser, C. A. & Stearns, T. Methods in Yeast Genetics (Cold Spring Harbor Laboratory Press, 1998).
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
Ries, J. & Schwille, P. Studying slow membrane dynamics with continuous wave scanning fluorescence correlation spectroscopy. Biophys. J. 91, 1915–1924 (2006).
Ries, J., Chiantia, S. & Schwille, P. Accurate determination of membrane dynamics with line-scan FCS. Biophys. J. 96, 1999–2008 (2009).
Kolaczyk, E. D. & Dixon, D. D. Non-parametric estimation of intensity maps using Haar wavelets and Poisson noise characteristics. Astrophys. J. 534, 490–505 (2000).
Kolaczyk, E. D. Non-parametric estimation of gamma-ray burst intensities using Haar wavelets. Astrophys. J. 483, 340–349 (1997).
Rossum, G. V. Python tutorial. Technical Report CS-R9526 (Centrum voor Wiskunde en Informatica, 1995).
Magatti, D. & Ferri, F. Fast multi-tau real-time software correlator for dynamic light scattering. Appl. Opt. 40, 4011–4021 (2001).
Fukuda, S. et al. High-speed atomic force microscope combined with single-molecule fluorescence microscope. Rev. Sci. Instrum. 84, 073706 (2013).
Imamura, M. et al. Probing structural dynamics of an artificial protein cage using high-speed atomic force microscopy. Nano Lett. 15, 1331–1335 (2015).
Alam, J. M., Kobayashi, T. & Yamazaki, M. The single-giant unilamellar vesicle method reveals lysenin-induced pore formation in lipid membranes containing sphingomyelin. Biochemistry 51, 5160–5172 (2012).
Alam, J. M. & Yamazaki, M. Spontaneous insertion of lipopolysaccharide into lipid membranes from aqueous solution. Chem. Phys. Lipids 164, 166–174 (2011).
Karal, M. A., Alam, J. M., Takahashi, T., Levadny, V. & Yamazaki, M. Stretch-activated pore of the antimicrobial peptide, magainin 2. Langmuir 31, 3391–3401 (2015).
Yamazaki, M. The single GUV method to reveal elementary processes of leakage of internal contents from liposomes induced by antimicrobial substances. Adv. Planar Lipid Bilayers Liposomes 7, 121–142 (2008).
Noda, T., Matsuura, A., Wada, Y. & Ohsumi, Y. Novel system for monitoring autophagy in the yeast Saccharomyces cerevisiae. Biochem. Biophys. Res. Commun. 210, 126–132 (1995).
We thank H. Yamamoto, T. Kawamata, Y. Kamada and D. S. Goldfarb for providing plasmids and strains for yeast experiments, Y. Ishii for assistance with protein preparation and H. Tochio for critical advice. This work was supported in part by JSPS KAKENHI grant number 25111004, 18H03989, 19H05707 (to N.N.N.), 15H01651, 17H05894, 17K07319 (to Y.F.), 19K16344 (to D.N.), 16H06375 (to Y. Ohsumi), 26119003, 17H06121 (to T.A.), 16H06280, 18H04853 (to K.S.), 19H05795, 19H03394, 16H06280 (to Y. Okada), 18H04751, 19H22520 (to K.M.), JST CREST grant number JPMJCR13M7 (to N.N.N.), JPMJCR15G2 (to Y. Okada) and grants from RIKEN (pioneering project ‘Dynamic Structural Biology’ to Y. Okada), from the Takeda Science Foundation (to N.N.N. and Y.F.), from Mochida Memorial Foundation for Medical and Pharmaceutical Research (to N.N.N.), from Tokyo Biochemical Research Foundation (to N.N.N. and J.M.A.), and from the Naito Foundation (to N.N.N. and Y.F.).
The authors declare no competing interests.
Peer review information Nature thanks Peter Hinterdorfer, Fulvio Reggiori, Jeffrey Woodruff and Li Yu 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
a, Measurement of alkaline phophatase activity of yeast cells overexpressing GFP–Atg13 performed based on previous reports44. Data are mean ± s.d. (n = 3 independent experiments). Bar colour indicates hours after onset of nitrogen starvation. b, Dissolution of Atg13–GFP puncta after addition of nitrogen source. Experiments were repeated independently twice with similar results. c, Two additional examples of GFP–Atg13 fluorescence recovery, related to Fig. 1a. Experiments were repeated independently three times with similar results, which are shown here and in Fig. 1a. d, Fluorescence of endogenously expressed Atg1–GFP, Atg13–GFP and Atg17–GFP puncta rapidly recovers after photobleaching. Rapa and −N indicate rapamycin treatment and nitrogen starvation, respectively. Experiments were repeated independently twice with similar results. e, Kymograph of FCS data shown in Fig. 1b. f, Two additional examples of partial fluorescence recovery of giant GFP–Atg13 droplets, related to Fig. 1d. DIC of Fig. 1d experiment is also shown. Experiments were repeated independently three times with similar results, which are shown here and in Fig. 1d. g, Coalescence of GFP–Atg13 puncta observed after removing 1,6-hexanediol. The images are the sum of five z-slices of GFP fluorescent images. Experiments were repeated independently three times with similar results. h, Two additional examples of the coalescence of two PAS precursors, related to Fig. 1g. Experiments were performed three times with similar results, which are shown here and in Fig. 1g. i, An additional example of Ostwald ripening of the PAS, related to Fig. 1h, i. The images are the sum of four z-slices of GFP fluorescent images. The bottom graph shows the line profile of fluorescence intensity in the top image. Experiments were repeated independently twice with similar results, which are shown here and in Fig. 1i. Scale bars, 2 μm, except in f, 1 μm.
a, Domain organization of Atg1-complex components. Grey regions indicate IDRs consisting of ten or more residues predicted to be disordered by DISOPRED29. Bar length is approximately proportional to the number of residues. b, SDS–PAGE of purified SNAP-tagged proteins used for in vitro analyses. Experiments were repeated independently twice with similar results. For gel source data, see Supplementary Fig. 1. c, An additional example of coalescence of Atg1-complex droplets observed in vitro, related to Fig. 2c. The right panel shows the change of the aspect ratio during coalescence. Experiments were repeated independently twice with similar results, which are shown here and in Fig. 2c. d, Formation of liquid droplets of the scaffold complex and their dissociation by 1,6-hexanediol treatment. Experiments were repeated independently six times with similar results. e, Quantification of the residual droplet area in d. Data are mean ± s.d. (n = 6 independent experiments). ****P = 3.9 × 10−6, two-sided t-test. f, The effect of pH on the formation of scaffold droplets. The concentrations of NaCl and Atg13–Atg17–Atg29–Atg31 are 500 mM and 4 μM, respectively. The experiment was repeated independently three times with similar results. g, Phase diagram of the formation of scaffold droplets at indicated NaCl concentrations and pH values. The protein concentration is 4 μM. Experiment was performed once. h, Time-course analysis of droplet area in Fig. 2g. Data are mean ± s.d. (n = 3 independent experiments). i, Phase diagram of droplet formation upon mixing of Atg13 and Atg17–Atg29–Atg31 at indicated protein concentrations. Representative images at a, b and c in the diagram are shown above the diagram. Experiment was performed once.
a, Purification of TORC1 from yeast. The experiment was repeated independently twice with similar results. b, Confirmation of the specificity of the anti-T226-P antibodies. Experiment was performed once. c, Quantification of the results in Fig. 3c. Data are mean ± s.d. (n = 3 independent experiments). d, e, Phosphorylation-mediated band shifts of Atg1, Atg13 and Atg29 upon incubation of the Atg1 complex with ATP analysed by conventional (d) and Phos-tag SDS–PAGE (e). Experiment was repeated independently three times with similar results. f, Effect of Atg1-mediated phosphorylation on Atg1-complex droplets. Bottom graph shows time-course analysis of droplet area. Data are mean ± s.d. (n = 3 independent experiments). g, SDS–PAGE of the recombinant Ptc2 used in this study. Experiments were repeated independently twice with similar results. h, Mn2+-dependent dephosphorylation of Atg1 and Atg13 by Ptc2. Experiment was performed once. For gel source data, see Supplementary Fig. 1 (a, b, d, e, g, h).
Photobleaching experiments were performed for the scaffold droplets in vitro. White and yellow arrow heads indicate photobleached and non-bleached droplets, respectively. Photobleaching was performed between 0 and 3 s. Non-bleached droplets coalesce, whereas bleached droplets do not. Experiments were repeated independently twice with similar results. Scale bars, 5 μm.
Extended Data Fig. 5 Droplets tethered to GUVs via specific Atg13–Vac8 interaction retain their liquid-like nature.
a, b, Deletion of VAC8 results in mislocalization of Atg5–GFP puncta away from the vacuole (a) and GFP–Atg8 puncta (b). Atg5–GFP and GFP–Atg8 were observed following 3 or 5 h rapamycin incubation, respectively. Data are mean ± s.d. (n = 3 independent experiments). *P = 0.0119, **P = 0.0044, two-sided t-test. c, Lack of tethering of Atg1-complex droplets to Vac8-free GUVs. Experiments were repeated independently 20 times with similar results. d, Impaired interaction of the Vac8 mutant with Atg13 demonstrated by in vitro pull-down assay. Experiments were repeated independently three times with similar results. For gel source data, see Supplementary Fig. 1. e, Near complete lack of tethering of scaffold droplets to Vac8 mutant-anchored GUVs. Experiments were repeated independently 30 times with similar results. f, Additional examples of time-dependent change in the number and size of scaffold droplets on Vac8-GUVs, related to Fig. 5g. The number and average area of droplets ± s.d. (n = droplet numbers) are shown. Experiments were repeated independently five times with similar results, which are shown here and in Fig. 5g. g, FRAP experiments of Atg1–SNAP in Atg1-complex droplets attached to Vac8-anchored multilamellar vesicles. Multilamellar vesicles were used instead of GUVs to reduce the movement of droplets on vesicles. The bottom graph indicates the ratio of fluorescence intensity at each time point in comparison to the initial intensity. Data from seven independent experiments are shown.
Under growing conditions, hyperphosphorylation of Atg13 by TORC1 inhibits Atg1-complex formation and phase separation. Upon starvation, TORC1 activity is inhibited and Atg13 is dephosphorylated by PP2C phosphatases, which leads to Atg1-complex formation. The Atg1 complex then undergoes phase separation to form a liquid droplet (early PAS), which is tethered to the vacuolar membrane through Atg13–Vac8 interaction. The early PAS activates Atg1 kinase by accelerating autophosphorylation and at the same time recruits downstream Atg factors, thereby transforming into the mature PAS from which isolation membrane is generated. Continuous phosphorylation and dephosphorylation of Atg13 at the PAS would contribute to maintaining the liquid property of the PAS.
Supplementary Figure 1: Uncropped gel images.
Supplementary Table 1: Yeast strains used in this study.
Supplementary Table 2: Plasmids used for yeast experiments in this study.
FRAP of overexpressed Atg13-GFP puncta. Rapid recovery of fluorescence of overexpressed GFP-Atg13 puncta after photobleaching. This video was used to create the images presented in Fig. 1a. Photobleaching was performed at frame 6 in this video. 1 frame, 0.174 s; Bar, 2 μm.
Partial FRAP of a giant PAS. Rapid recovery (within 0.7 s) in a partially quenched region at the centre of the giant PAS after photobleaching. This video was used to create the images presented in Fig. 1d. Partial photobleaching was performed at frame 6 in the video. 1 frame, 0.174 s; Bar, 1 μm.
Effect of 1,6-hexanediol treatment on the PAS. Original video for data presented in Fig. 1e assessing the effect of 1,6-hexanediol treatment on Atg13-GFP puncta formation. 1 frame, 6 s; Bar, 2 μm.
Formation process of the PAS. Multiple Atg13-GFP puncta form in response to rapamycin treatment (10 min) before coalescing into a larger droplet. This video was used to generate the images shown in Fig. 1f. 1 frame, 11.4 s; Bar, 2 μm.
Coalescence of two PAS precursors. Close-up of two Atg13-GFP puncta coalescing to form a single punctum. This video was used to generate the images and aspect ratio data presented in Fig. 1g. 1 frame, 0.09 s; Bar, 2 μm.
Ostwald ripening of PAS precursors. Video of Ostwald ripening of Atg13-GFP puncta shown in Fig. 1h. 1 frame, 10.7 s; Bar, 2 μm.
HS-AFM observation of Atg17 molecules in droplets. HS-AFM videos of Atg17-SNAP in scaffold droplets on non-coated coverslip before (top) and after FFT bandpass filter (bottom).
HS-AFM observation of Atg17 molecules in mature droplets. HS-AFM videos of Atg17-SNAP in scaffold droplets on a 3-aminopropyltriethoxysilane-coated coverslip before (top) and after FFT bandpass filter (bottom).
Early PAS droplets on Vac8-GUVs coalesce to form fewer but larger droplets. Video of scaffold droplets attached to Vac8-GUVs. Three different fluorescence signals (blue, mKalama1 from Vac8; red, SNAP-Surface 549 from Atg13; purple, SNAP-Surface Alexa Fluor 647 from Atg17) were merged to create this video. 1 frame, 2.2 s; Bar, 20 μm.
About this article
Cite this article
Fujioka, Y., Alam, J.M., Noshiro, D. et al. Phase separation organizes the site of autophagosome formation. Nature 578, 301–305 (2020). https://doi.org/10.1038/s41586-020-1977-6
Biomolecular condensates in cell biology and virology: Phase-separated membraneless organelles (MLOs)
Analytical Biochemistry (2020)