Polyubiquitin chain-induced p62 phase separation drives autophagic cargo segregation

Abstract

Misfolded proteins can be degraded by selective autophagy. The prevailing view is that ubiquitin-tagged misfolded proteins are assembled into aggregates by the scaffold protein p62, and the aggregates are then engulfed and degraded by autophagosomes. Here we report that p62 forms droplets in vivo which have liquid-like properties such as high sphericity, the ability to undergo fusion, and recovery after photobleaching. Recombinant p62 does not undergo phase separation in vitro; however, adding a K63 polyubiquitin chain to p62 induces p62 phase separation, which results in enrichment of high-molecular weight ubiquitin signals in p62 droplets. Mixing recombinant p62 with cytosol from p62^-/- cells also results in p62 phase separation in a polyubiquitination-dependent manner. Mechanistically, p62 phase separation is dependent on p62 polymerization, the interaction between p62 and ubiquitin, and the valence of the polyubiquitin chain. Moreover, p62 phase separation can be regulated by post-translational modifications such as phosphorylation. Finally, we demonstrate that disease-associated mutations in p62 can affect phase separation. We propose that polyubiquitin chain-induced p62 phase separation drives autophagic cargo concentration and segregation.

Fig. 1

p62 forms liquid droplets in vivo. a Correlative light-electron microscopy (CL-EM) of NRK cells transiently transfected with GFP-p62 and Lamp1-mCherry constructs. Scale bar, 2 µm. The insert in the left panel shows a p62 body. The inserts in the right panel show p62 bodies contained within an autophagosome (upper) and an autolysosome (lower). b CL-EM of Atg12^−/− cells transiently transfected with GFP-p62 and Lamp1-mCherry constructs. Scale bar, 2 µm. The insert shows a p62 body. c Western blot analysis of wild-type and Atg12^−/− cells with the indicated antibodies. d GFP-labeled p62 forms p62 bodies in Atg12^−/− NRK cells. An enlargement of the boxed p62 body is shown in the insert. Scale bar, 5 µm. e Rendered 3D shapes of a p62 body. Cells were fixed with 4% PFA. The panels show the XY, XZ, and YZ planes. Scale bar, 1 µm. f A plot showing the sphericity of p62 bodies (n = 44). Error bar represents SD. g Fusion of p62 bodies. Scale bar, 1 µm. h Left panels: fluorescence intensity recovery of a p62 body after photobleaching. Scale bar, 2 µm. Right panel: quantification of fluorescence intensity recovery of a photobleached p62 body

Fig. 2

Polyubiquitin chains induce p62 phase separation in vitro. a The purity of purified mCherry-p62 was analyzed by Coomassie blue staining. b The indicated concentrations of mCherry-p62 were incubated in phase separation assay buffer and visualized by confocal microscopy. Scale bar, 5 µm. c Schematic diagram of the reaction to synthesize K63 polyubiquitin chains in vitro. Ub, monoubiquitin; E1, Ubiquitin-activating enzyme E1; Ube2V2/Ube2N, an E2 complex. d The in vitro K63 polyubiquitin chain synthesis reaction with (+) and without (-) ATP was analyzed by western blot with an anti-K63 polyubiquitin chain antibody. e mCherry-p62 was mixed with the in vitro-synthesized K63 polyubiquitin chain from d and visualized by confocal microscopy. mCherry-p62, 20 µM; monoubiquitin, 80 µM. Scale bar, 5 µm. f Schematic diagram of the sedimentation assay to separate the condensed liquid droplets and the supernatant. g The phase separation reaction from e was separated by centrifugation as shown in f and the pellet (p62 droplets) and supernatant were analyzed by western blot using an anti-K63 polyubiquitin chain antibody. S, supernatant; P, pellet. h Fusion of p62 droplets formed during the in vitro phase separation process in e. Scale bar, 1 µm. i Left panels: fluorescence intensity recovery of a p62 droplet formed in vitro in the presence of K63 polyubiquitin chains after half-bleaching. Scale bar, 2 µm. Right panel: quantification of fluorescence intensity recovery in the bleached region of p62 droplets (n = 3). j Western blot analysis of wild type and p62^−/− cells with the indicated antibodies. k The S150 cytosolic fraction from p62^−/− cells was pretreated with and without the deubiquitinating enzyme USP5 (0.025 mg/mL) for 2 h and analyzed by western blotting using an anti-ubiquitin antibody. l mCherry-p62 was mixed with the S150 cytosolic fraction pretreated with or without the deubiquitinating enzyme USP5 from p62^−/− cells and the reaction was visualized by confocal microscopy. mCherry-p62, 40 µM; cytosol, 90 mg/mL. Scale bar, 5 µm. m p62 droplets and supernatant from l were separated by centrifugation and analyzed by western blot using antibodies against ubiquitin, K63 polyubiquitin chains and K48 polyubiquitin chains. S supernatant, P pellet

Fig. 3

p62 phase separation is dependent on the valence of the polyubiquitin chains, and polyubiquitin chains can diffuse freely in p62 droplets. a Schematic diagram of linear ubiquitin chains containing 6 (Ubx6) or 8 (Ubx8) ubiquitin molecules. b In vitro phase separation assay of mCherry-p62 with the linear Ubx8 and Ubx6. mCherry-p62, 15 µM; Ubiquitin, 80 µM. Scale bar, 5 µm. c Phase diagrams of mCherry-p62 and linear polyubiquitin chains. The concentrations are in terms of the modules. The number on the x and y axes means the dilution ratio of p62 protein (30uM) and Ub (160uM).The red circles indicate phase separation, and the black circles indicate no phase separation. The boxed red circles show the concentrations used in b. d Schematic diagram of GFP-labeled linear octa-ubiquitin (GFP-Ubx8) e In vitro phase separation assay of mCherry-p62 with GFP-Ubx8. mCherry-p62, 20 µM; GFP-Ubx8, 5 µM. Scale bar, 5 µm. f Top panels: fluorescence intensity recovery of an in vitro-formed mCherry-p62 droplet in the presence of GFP-Ubx8 after photobleaching. Scale bar, 2 µm. Bottom left panel: quantification of fluorescence intensity recovery of mCherry-p62 and GFP-Ubx8 in the bleached p62 droplet. Bottom right panel: enlargement of the boxed area in the left panel (n=3).

Fig. 4

LC3 is recruited into p62 bodies. a Atg12^−/− NRK cells were transfected with mCherry-LC3b and GFP-p62 and observed by confocal microscopy. Scale bar, 5 µm. b Upper panels: fluorescence intensity recovery of GFP-p62 and mCherry-LC3b in a bleached p62 body. Scale bar, 2 µm. Bottom left panel: quantification of fluorescence intensity recovery of GFP-p62 and mCherry-LC3b in the bleached p62 body. Bottom right panel: enlargement of the boxed area in the left panel (n=5). c In vitro phase separation assay of mCherry-p62 with linear Ubx8. GFP-LC3b was added after the phase separation. mCherry-p62, 20 µM; Ubx8: 5 µM; GFP-LC3b, 10 µM. Scale bar, 10 µm. d Upper panels: fluorescence intensity recovery of GFP-LC3b and mCherry-p62 in an in vitro-formed mCherry-p62 droplet from c after photobleaching. Scale bar, 2 µm. Bottom left panel: quantification of fluorescence intensity recovery of GFP-LC3b and mCherry-p62 in the bleached p62 droplet. Bottom right panel: enlargement of the boxed area in the left panel (n=3)

Fig. 5

The PB1 domain and UBA domain of p62 are required for polyubiquitin chain-induced phase separation and autophagic degradation of p62. a Schematic diagram of the p62 protein domain structure. The location of p62 mutations is shown underneath. b The indicated GFP-p62 constructs were expressed in Atg12^−/− cells and observed by confocal microscopy. Scale bar, 5 µm. c. Cells from b were quantified for the number of p62 bodies. >45 cells from b were assessed blind and quantified. Error bars indicate SD (n = 3). d NRK cells transiently expressing the indicated GFP-p62 constructs were starved for the indicated time. Cell lysates were analyzed by western blot with the indicated antibodies. e Quantification of the indicated western blots in d. f The indicated mutated recombinant mCherry-p62 proteins were mixed with linear Ubx8 and the reaction was visualized by confocal microscopy. mCherry-p62, 5 µM; linear Ubx8, 2.5 µM. Scale bar, 5 µm

Fig. 6

Phosphorylation of S403 promotes polyubiquitin chain-induced phase separation, p62 body formation and autophagic degradation. a The indicated GFP-p62 constructs were expressed in Atg12^−/− cells and observed by confocal microscopy. Scale bar, 5 µm. b Cells from a were quantified for the number of p62 bodies. In all, >45 cells from a were assessed blind and quantified. Error bars indicate SD (n = 3). c Cells from a were quantified for the area of p62 bodies. In total, >38 cells from a were assessed blind and quantified. Error bars indicate SD (n = 3). d NRK cells transiently expressing the indicated GFP-p62 constructs were starved for the indicated time. Cell lysates were analyzed by western blot with the indicated antibodies. e Quantification of the indicated western blots in d. f The indicated mutated recombinant mCherry-p62 proteins were mixed with linear Ubx8 and the reaction was visualized by confocal microscopy. mCherry-p62, 5 µM; linear Ubx8, 2.5 µM. Scale bar, 5 µm

Fig. 7

Two p62 mutations identified in Paget’s disease of bone (PDB) affect p62 body formation, autophagic degradation of p62, and p62 phase separation. a The indicated GFP-p62 constructs were expressed in Atg12^−/− cells and observed by confocal microscopy. Scale bar, 5 µm. b Cells from a were quantified for the number of p62 bodies. In all, >45 cells from a were assessed blind and quantified. Error bars indicate SD (n = 3). c NRK cells transiently expressing the indicated GFP-p62 constructs were starved for the indicated time. Cell lysates were analyzed by western blot with the indicated antibodies. d Quantification of the indicated western blots in c. e The indicated mutated recombinant mCherry-p62 proteins were mixed with linear Ubx8 and the reaction was visualized by confocal microscopy. mCherry-p62, 5 µM; linear Ubx8, 2.5 µM. Scale bar, 5 µm

Fig. 8

Model for the formation of p62 bodies and their role in selective autophagy