Molecular determinants of the crosstalk between endosomal microautophagy and chaperone-mediated autophagy

Abstract

Chaperone-mediated autophagy (CMA) and endosomal microautophagy (eMI) are pathways for selective degradation of cytosolic proteins in lysosomes and late endosomes, respectively. These autophagic processes share as a first step the recognition of the same five-amino-acid motif in substrate proteins by the Hsc70 chaperone, raising the possibility of coordinated activity of both pathways. In this work, we show the existence of a compensatory relationship between CMA and eMI and identify a role for the chaperone protein Bag6 in triage and internalization of eMI substrates into late endosomes. Association and dynamics of Bag6 at the late endosome membrane change during starvation, a stressor that, contrary to other autophagic pathways, causes a decline in eMI activity. Collectively, these results show a coordinated function of eMI with CMA, identify the interchangeable subproteome degraded by these pathways, and start to elucidate the molecular mechanisms that facilitate the switch between them.

Keywords: Bag6; CP: Molecular biology; autophagy; chaperone; late endosome; lysosome; microautophagy; protein degradation; protein targeting; proteostasis; starvation.

Figure 1.. eMI and CMA share substrate proteins and display compensatory activity

(A–C) Immunoblot (A) of isolated rat liver lysosomes active for CMA (CMA⁺ Lys) and late endosomes (LE/MVBs) treated or not with protease inhibitors (PI) and then incubated with the indicated proteins. LAMP-1 and Rab7 are used as compartment markers. Quantification of the amount of substrate bound and internalized or degraded (Int/Deg) (B) and ratio of binding (top) and Int/Deg (bottom) of the indicated substrates by LE/MVBs relative to lysosomes (C); n = 4 independent experiments. (D) Representative images (left) and quantification (right) of eMI activity (as average number of fluorescent puncta per cell section) in mouse fibroblasts, control (Ctr) or knocked down for LAMP-2A (L2A(−)), stably expressing the KFERQ-Split Venus reporter to measure eMI. Nuclei are highlighted with Hoechst. Insets: higher-magnification images; n > 2,500 cells from three independent experiments. (E) Representative images (left) and quantification (right) of CMA activity (as average number of fluorescent puncta per cell section) in mouse fibroblasts, control (Ctr) or knocked down for Vps4A/B (Vps4(−)), stably expressing the KFERQ-Dendra reporter to measure CMA. Nuclei are highlighted with Hoechst. Insets: higher-magnification images; n > 2,500 cells from three independent experiments. (F) Proteolytic activity of intact (top) or detergent-disrupted (bottom) LE/MVBs isolated from wild-type (WT) and L2A-knockout (KO) mouse liver incubated with a pool of radiolabeled cytosolic proteins; n > 10 (intact) or 4 (broken) mice. (G and H) Immunoblot (G) and quantification (H) of binding and internalization/degradation of the indicated proteins in LE/MVBs isolated from WT or L2AKO mouse liver; n = 8 (GAPDH), 9 (Tau), and 4 (Cyclophilin A) mice per genotype. Data are the mean ± SEM and individual values. Two-way ANOVA with Bonferroni’s multiple comparisons post hoc test (B), one-way ANOVA with Tukey’s multiple comparison post hoc test (C), and unpaired t test (D–H) were used. Significant differences: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. See also Figures S1 and S2.

Figure 2.. A subset of the proteome degraded by CMA is rerouted to eMI upon CMA failure

(A) Left: schematic of comparative proteomics of LE/MVBs isolated from livers of wild-type (WT) and LAMP-2A knockout (L2AKO) mice injected or not with leupeptin (Leup) to block LE/MVB degradation. Right: log₂ fold change (log₂FC) in rates of protein degradation in LE/MVBs in WT compared with L2AKO. Black, proteins equally degraded in both groups; blue, reduced (light) or absent (dark) degradation in L2AKO; red, increased (light) or only (dark) degradation in L2AKO. (B) Voronoi flattened visualization using REACTOME pathway analysis of intracellular proteins undergoing degradation in LE/MVBs from WT mouse livers (top) and proteins no longer degraded or only degraded in LE/MVBs from L2AKO mice (bottom). (C) Percentage of LE/MVB proteins undergoing degradation only or at a higher rate in L2AKO (L2AKO > WT) mice is marked as the potentially rerouted proteome (red arrow). Roman numerals indicate the groups described in the text. (D) Analysis of KFERQ-like motifs in the protein groups in (C). (E–G) STRING analysis (E and F) and functional families with enrichment (G) of proteins displaying exclusive or higher degradation in LE/MVBs from L2AKO mice. All GO terms are statistically enriched, with p < 0.001. See also Figure S3.

Figure 3.. Changes in CMA activity are associated with discrete remodeling of the LE/MVB proteome

(A and B) Immunoblot (A) and quantification (B) for the indicated proteins of LE/MVBs isolated from livers of wild-type (WT) and LAMP-2A knockout (L2AKO) mice; n = 6 mice. The residual LAMP-2A band in L2AKO mice is due to cross-reactivity of the antibody with other LAMP-2 isoforms; these mice have no LAMP-2A mRNA. (C) Volcano plot showing log₂ fold change (log₂FC) of levels of non-substrate LE/MVB proteins between WT and L2AKO mice against their negative log₁₀-transformed p values. Names of proteins with significantly higher levels in L2AKO LE/MVBs are shown. (D) Immunoblot for chaperones and co-chaperones in LE/MVBs and CMA⁺ lysosomes isolated from livers of fed (F) or 48-h-starved (S) rats. Top right: markers of endolysosomes. Bottom right: heatmap of the abundance of the indicated proteins from quantification of immunoblots on the left; n = 3 rats per group. (E) Immunoblot of LE/MVBs from livers of WT and L2AKO mice for the indicated chaperones. (F–I) Co-immunoprecipitation of proteins with Hsc70 from LE/MVBs, CMA⁺ lysosomes, and cytosol. Experimental schematic (F), Venn diagram of proteins recovered in the immunoprecipitation of each compartment (G), percentage of KFERQ-like motifs in those proteins (H), and table of identified chaperones (I). Ponceau staining is shown as a loading control in (A) and (E). Data in (B) are the mean ± SEM and individual values. Multiple t test did not reveal statistical differences. See also Figures S4 and S5 and Table S1.

Figure 4.. Bag6 is required for eMI degradation of a subset of the proteome

(A–C) Representative images (A) and quantification (B and C) of eMI activity in NIH3T3 mouse fibroblasts stably expressing the KFERQ-Split Venus reporter control (Ctr) or knocked down (−) for Bag6 using shRNA (A and B) or siRNA (C). Where indicated, cells were treated with ammonium chloride and leupeptin (+N/L) to block degradation. Nuclei are highlighted with DAPI. Insets: higher magnification images; n = 16 fields from four independent experiments for (B) and n > 2,500 cells from four independent experiments (C). (D–F) Representative images (D) and quantification (E and F) of CMA activity in NIH3T3 mouse fibroblasts stably expressing the KFERQ-Dendra reporter Ctr or knocked down (−) for Bag6 using shRNA (D and E) or siRNA (D). Nuclei are highlighted with DAPI. Insets: higher magnification images; n > 30 cells from four independent experiments (E) and n > 2,500 cells from four independent experiments (F). (G and H) Degradation of long-lived proteins in Ctr and Bag6(−) NIH3T3 fibroblasts. Rates of total (G) and endolysosomal (H) protein degradation are shown; n = 3 independent experiments. (I–L) Quantification of eMI activity (I and J) or CMA activity (K and L) using NIH3T3 mouse fibroblasts stably expressing the KFERQ-Split Venus reporter or KFERQ-Dendra reporter, respectively, upon the indicated knockdowns (KD). Data in (I) and (K) are relative to values in control cells. Data in (J) and (L) are relative to the indicated single KD; n > 2,500 cells from four (I and J) or eight or nine (K and L) independent experiments. (M) Representative immunoblots of Bag6 (left), Hsc70 (middle), and Alix (right) after blue native electrophoresis of WT and LAMP-2A KO LE/MVBs. Samples were run in the same gel, and dotted lines indicate where the membrane was cut to blot for each protein separately. Quantification of the high-molecular-weight form of Bag6 is shown on the right; n = 3 independent experiments. (N–R) Comparative proteomics of LE/MVBs isolated from Ctr and Bag6(−) cells. Venn diagram of proteins in both groups (N), presence of KFERQ-like motifs (O), fraction of the LE/MVB proteome degraded in a Bag6-dependent or -independent manner (P), functional families with protein enrichment in the group of proteins detected only in LE/MVBs in the presence of Bag6 (Q), and STRING analysis of the same proteins (R). Data in (B), (C), (E), (F), and (G)–(M) are the mean ± SEM and individual values. Two-way ANOVA with Bonferroni’s (B, C, and G) or Tukey’s (H) multiple comparisons post hoc test, unpaired t test (E, F, J, L, and M), and one-way ANOVA with Bonferroni’s multiple comparisons post hoc test (I and K) were used. Significant differences: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. All GO terms in (Q) are statistically enriched, with p < 0.05. See also Figure S5 and Tables S1 and S2.

Figure 5.. LE/MVB-associated Bag6 is required for eMI

(A and B) Organelle stability of Bag6. Representative immunoblots (A) of LE/MVBs and CMA-active (CMA⁺) lysosomes incubated at 37°C for the indicated times with or without protease inhibitors (PI) with examples of constitutive (GBA) and substrate proteins (GAPDH). The percentage of Bag6 remaining in each compartment at the indicated times is shown in (B); n = 3 independent experiments. (C and D) Organelle topology of Bag6. Representative immunoblots (C) and quantification (D) of LE/MVBs and CMA⁺ lysosomes incubated with increasing concentrations of trypsin only or with Triton X-100 (TX) to facilitate access to luminal proteins. GBA and mTOR are shown as examples of luminal and membrane-associated proteins, respectively. The percentage of Bag6 relative to the untreated samples is shown in (D); n = 3 independent experiments. (E and F) Representative immunoblots (E) and quantification (F) of LE/MVBs preincubated or not with PI and/or an antibody against Bag6 (anti-Bag6) and then incubated with the indicated substrates. Ponceau staining is shown as a loading control. Data are presented relative to the total amount of substrate protein detected in samples incubated with PI but not the antibody; n = 5 (GAPDH, α-syn), 3 (Tau), and 4 (Cyclophilin A) independent experiments. (G–I) Representative Tau-STED microscopy images of isolated LE/MVBs incubated with Tau and immunolabeled for the indicated proteins (G and H, left); single-channel images (top) and 2.5D density plots (bottom) (G and H). Examples of fluorescence coincidence signals and intensity plotting of each signal along the perimeter of the organelles (G and H, right). Schemes on the right show proposed topology for the indicated proteins. The linear correlations between luminal content of Hsc70 (top) and Bag6 (bottom) with Tau internalization are shown in (I); n = 3 animals, ≥30 LE/MVBs per condition. (J and K) Representative Tau-STED microscopy images of isolated LE/MVBs incubated with or without Tau and immunolabeled for the indicated proteins; single-channel images (J, top) and 2.5D density plots (J, middle). Examples of fluorescence coincidence signals and intensity plotting of each signal along the perimeter of the organelles are shown in (J), bottom. Percentage of Hsc70 (K, top) and Bag6 (K, bottom) in the lumen of LE/MVBs incubated in the presence (+) or absence (−) of Tau; n = 5 animals, ≥50 LE/MVBs per condition. Data are the mean ± SEM and individual values. Two-way ANOVA with Bonferroni’s multiple comparisons post hoc test (B, C, D, and F) and unpaired t test (K) were used. Simple linear regression was used in (I). Significant differences: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. See also Figure S6.

Figure 6.. eMI activity is inhibited by the absence of nutrients

(A and B) Immunoblot of LEs/MVBs isolated from livers of fed or 24-h-starved (Stv) rats treated or not with protease inhibitors (PI) and then incubated with the indicated proteins (A). Quantification of the amount of substrate protein bound (B, left) and internalized/degraded (B, right) relative to fed conditions; n = 10 (Tau) and 6 (Cyclophilin A) independent experiments. (C and D) Immunoblot (B) and quantification (C) of LE/MVBs isolated from livers of rats fed or starved for the indicated periods of time and incubated with Tau (left) or GAPDH (right) as in (A). Values are expressed relative to fed rats; n = 4 independent experiments. (E and F) Representative immunoblot for the indicated proteins (E) and quantification (F) of LE/MVBs isolated from livers of fed and 24-h-starved (Stv) rats injected or not with leupeptin to block LE/MVB degradation. Data are expressed relative to degradation in LE/MVBs from fed mice; n = 4 (GAPDH), 3 (LRRK2, aldolase, Cyclophilin A), or 2 (hexokinase) rats. (G) Left: representative confocal images of LE/MVBs isolated from fed or 24-h-starved rats and immunostained for the indicated proteins after incubating with Tau. Inset: overlapping mask (in white) of higher-magnification image. Right: quantification of the percentage of co-localization; n ≥ 25 fields from three independent experiments. (H) Proteolytic activity of detergent-disrupted LE/MVBs isolated from livers of fed or 24-h-starved rats incubated with a pool of radiolabeled cytosolic proteins. Values are expressed relative to those in fed rats; n = 16 independent experiments. Ponceau staining is shown in (A), (C), and (E) as a loading control. Data are the mean ± SEM and individual values. Two-way ANOVA (D) and unpaired t test (B, F, and H) were used. Significant differences: *p < 0.05 and **p < 0.01; ns, not significant. See also Figure S7.

Figure 7.. Starvation changes levels and topology of LE/MVB-associated Bag6

(A and B) Representative immunoblot for the indicated proteins (A) and quantification (B) of LE/MVBs isolated from livers of fed or 24-h-starved rats. Ponceau staining is shown as loading control; n = 6 (Fed) and 5 (Stv) rats. (C and D) Representative immunoblot for the indicated chaperones (C) and quantification (D) in homogenates (Hom), cytosol (Cyt), and LE/MVBs (LE) isolated from livers of fed (F) or 24-h-starved (S) rats. Values in (D) are expressed relative to average fed levels; n = 3–9 rats. (E–H) Representative STED microscopy images of LE/MVBs isolated from fed or 24-h-starved mouse livers incubated or not with Tau and immunolabeled for the indicated proteins. Single-channel images (left) and 2.5D density plots for each one (right) are shown (E). Quantification of the percentage of each protein present in the LE/MVB lumen (F) and correlation between luminal Hsc70 and Bag6 levels per LE/MVB from fed and starved mice (G) and overlapped correlations between both chaperones in the same LE/MVB samples as in (E) to illustrate the effect of starvation (left) or of incubation with Tau (right) (H); n = 3 animals, ≥30 LE/MVBs per condition. Data are the mean ± SEM and individual values. Unpaired (B) and one-sample multiple (D) t test and two-way ANOVA with Tukey’s multiple comparisons post hoc test (F) were used. Simple linear correlations were used in (G) and (H). Significant difference: *p < 0.05; ns, not significant. See also Figure S7.