Ubiquilins regulate autophagic flux through mTOR signalling and lysosomal acidification

Abstract

Although the aetiology of amyotrophic lateral sclerosis (ALS) remains poorly understood, impaired proteostasis is a common feature of different forms of ALS. Mutations in genes encoding ubiquilins, UBQLN2 and UBQLN4, cause familial ALS. The role of ubiquilins in proteasomal degradation is well established, but their role in autophagy-lysosomal clearance is poorly defined. Here, we describe a crosstalk between endoplasmic reticulum stress, mTOR signalling and autophagic flux in Drosophila and mammalian cells lacking ubiquilins. We found that loss of ubiquilins leads to endoplasmic reticulum stress, impairs mTORC1 activity, promotes autophagy and causes the demise of neurons. We show that ubiquilin mutants display defective autophagic flux due to reduced lysosome acidification. Ubiquilins are required to maintain proper levels of the V0a/V100 subunit of the vacuolar H⁺-ATPase and lysosomal pH. Feeding flies acidic nanoparticles alleviates defective autophagic flux in ubiquilin mutants. Hence, our studies reveal a conserved role for ubiquilins as regulators of autophagy by controlling vacuolar H⁺-ATPase activity and mTOR signalling.

Fig. 1.. Ubiquilin Is Broadly Expressed and Required in the Developing and Adult Nervous System

(a) (a’) Schematic representation of the molecular lesion in ubqn gene, ubqn deletion (y^wing2+Δubqn), genomic rescue (GR) constructs and deficiency spanning the genomic region indicated (a”) Schematic representation of the nonsense mutation in Ubqn protein which contains an N-terminal UBiquitin-Like (UBL) domain, four Sti1 motifs, and a C-terminal UBiquitin-Associated (UBA) domain, capable of binding to proteasome, heat shock chaperones, and ubiquitinated proteins, respectively. (b) Immunofluorescence staining with GFP antibody of larval brain, ventral nerve cord (VNC), and adult brain of flies expressing GFP-tagged genomic Ubqn transgene (Ubqn-GFP construct). Scale bars, 40 μm. (c) qRT-PCR quantification showing ubqn transcript expression in ubqn¹ wandering third instar larvae compared to iso (y w FRT19A) larvae normalized against housekeeping gene (GAPDH). n= 3 independent biological samples, and 3 PCR replicates for each biological sample. Mean ± s.e.m. **p= 0.0028. (d) Western blot for Ubqn with protein lysates from third instar larvae of iso, ubqn¹; GR, and ubqn¹. Asterisks indicate non-specific bands. (e) H&E staining in trans- and frontal- head sections of control (iso) and ubqn¹ pre-pupae (P4 stage= 20h APF grown at room temperature). Neuropil is severely reduced in mutants (shown as light pink with H&E staining). Scale bars, 100 μm. (f) ERG traces from 15 and 45 day-old ey-FLP clones of iso (control), ubqn¹, and ubqn¹; GR raised in 12h light/12h dark cycle (LD) with quantification of ERG amplitudes. n= 5 (iso 15d and ubqn¹ 45d), n= 6 (iso 45d and ubqn¹; GR 45d) n= 7 (ubqn¹ 15d), n= 4 (ubqn¹; GR 15d) flies. Mean ± s.e.m. ns, not significant; **p= 0.0057, ****p< 0.0001. For all panels except 1e, three independent experiments were performed with similar results. For panel 1e, two independent experiments were performed with similar results. All statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 1c,f can be found in Table S9.

Fig. 2.. Ubqn Is Required for Neuronal Function and Maintenance

(a) TEM images of the retinae of ey-FLP clones of ubqn¹; GR (control) and ubqn¹ raised in 12h light/12h dark cycle, at 1 day-old, 15 days-old, and 30 days-old. Scale bars, 2 μm. (b) Higher magnification TEM images of the retinae of ey-FLP clones of ubqn¹; GR (control) and ubqn¹ raised in 12h light/12h dark cycle, at 30 days old. R stands for rhabdomere. ubqn¹ retina displays degenerating photoreceptors, vacuolization (asterisks), rhabdomere loss (arrowheads), splitting rhabdomeres (arrow), glial death, and mitochondria accumulation. Glial cells are highlighted in purple and mitochondria are highlighted in magenta. Scale bars, left panels 1 μm, right panels 0.5 μm. (c) Quantification of photoreceptor number per ommatidium (n= 53 (control), n= 42 (ubqn¹) ommatidia from 3 biologically independent animals), vacuole number (n= 10 (control), n= 11 (ubqn¹) images (7 ommatidia/image) from 3 biologically independent animals), and mitochondria number (n= 10 (control), n= 11 (ubqn¹) ommatidia from 3 biologically independent animals) of TEM images from (b). Mean ± s.e.m. ****p< 0.0001. (d) TEM images of the ER in the retinae of 1 day-old ey-FLP clones of ubqn¹; GR (control) and ubqn¹ with quantification of normalized ER length (μm)/area (μm²). Scale bars, 0.5 μm. n= 12 (control), n= 11 (ubqn¹) images from 3 biologically independent animals. Mean ± s.e.m for vacuole and mitochondria number, and mean for quantification of % PR number/ommatidium. ****p< 0.0001. (e) Western blot for Phospho-eIF2α (S51) and total eIF2α with protein lysates from third instar larvae of ubqn¹; GR (control) and ubqn¹ with quantification of Phospho-eIF2α (S51) to total eIF2α ratio in ubqn¹; GR (control) and ubqn¹. n= 4 biologically independent samples. Mean ± s.e.m. ns, not significant; ***p= 0.0009. For all panels, three independent experiments were performed with similar results. All statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 2c-e can be found in Table S9.

Fig. 3.. Loss of Ubqn Affects mTOR Signaling and Promotes Autophagy Induction

(a) Western blot for mTOR-dependent phosphorylation of Akt, S6K, and 4E-BP with fat body protein lysates from third instar larvae of ubqn¹; GR (control) and ubqn¹. Quantification of relative P-Akt/Akt, P-S6K/S6K, and P-4EBP/Tubulin ratios. n= 3 biologically independent samples. Mean ± s.e.m. ns, not significant; *p= 0.0372, **p= 0.0035 (P-S6K/S6K), **p= 0.0047 (P-4EBP/Tubulin). (b) ubqn¹ clone in fat body of early third instar larvae expressing Atg1-GFP protein trap with quantification of normalized Atg1 punctae number/μm². Scale bar, 10 μm. n= 4 biologically independent samples. Mean ± s.e.m. **p= 0.007. (c) ubqn¹ clone in fat body of early third instar larvae expressing UAS-mCherry-Atg8a with Cg-GAL4 driver and quantification of normalized Atg8 punctae number/μm². Scale bar, 10 μm. n= 8 biologically independent samples. Mean ± s.e.m. ***p= 0.0003. (d) TEM images of the retinae of 15 day-old ey-FLP clones of ubqn¹; GR (control) and ubqn¹ and quantification of autophagic vesicle numbers. Scale bars, 0.5 μm. n= 4 flies. **p= 0.001, ***p= 0.0003 (autophagosome), ***p= 0.0004 (autolysosome and amphisome). (e) Western blot for p62 with fat body protein lysates from third instar larvae of ubqn¹; GR (control) and ubqn¹ and quantification of relative p62/actin levels. n= 5 biologically independent samples. Mean ± s.e.m. ***p = 0.0009. (f) Immunofluorescence staining of p62 and phalloidin (labeling rhabdomeres in photoreceptors) in whole eye clones of 2 day-old control (iso) and ubqn¹ and quantification of p62 fluorescence intensity. Scale bars, 1 μm. n= 5 (control), n= 7 (ubqn¹) biologically independent animals. Mean ± s.e.m. ***p = 0.0009. For all panels except 3b, three independent experiments were performed with similar results. For panel 3b, two independent experiments were performed with similar results. All statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 3 can be found in Table S9.

Fig. 4.. Ubiquilin Facilitates Lysosomal Acidification and Function

(a) Confocal images of fed early third instar larval fat body of y w (control) and ubqn¹ expressing UAS-GFP-mCherry-Atg8a driven by Cg-GAL4 (drives expression in the fat body, hemocytes, and lymph gland) and quantification of only mCherry expressing and GFP+mCherry expressing punctae/section (135μm²). Scale bars, 10 μm. n= 9 (control), n= 8 (ubqn¹) biologically independent samples. Mean ± s.e.m. **p= 0.006, ****p< 0.0001. (b) ubqn¹ clone in fat body of third instar larvae expressing UAS-LAMP1-GFP with Cg-GAL4 driver and higher magnification of LAMP1-GFP punctae to the right. Scale bar, 20 μm. (c) TEM images of lysosomes in 15 day-old retinae of ey-FLP clones of ubqn¹; GR (control) and ubqn¹. Scale bar, 0.5 μm. (d) Live imaging for LysoTracker (LT) dye in third instar larval fat body with ubqn¹ clones and quantification of normalized LT punctae number/μm² in ubqn¹ clones compared to the surrounding wild-type cells. Scale bar, 10 μm. n= 8 (control), n= 10 (ubqn¹) biologically independent samples. Mean ± s.e.m. ***p = 0.0004. (e) Live imaging for LysoTracker (LT) in third instar larval fat bodies of iso (control) and ubqn¹ expressing UAS-LAMP1-GFP with Cg-GAL4 driver and quantification of the normalized number of lysosomes (LAMP1-GFP punctae)/μm² and the number of lysosomes without LysoTracker/μm². Scale bars, 10 μm. n= 6 biologically independent samples. Mean ± s.e.m. ****p< 0.0001. (f) Live imaging for Magic Red dye that detects active Cathepsin B in third instar larval fat body with ubqn¹ clones and quantification of the number of punctae. Scale bar, 10 μm. n= 6 biologically independent samples. Mean ± s.e.m. **p = 0.0014. For all panels, three independent experiments were performed with similar results. All statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 4a, d, e, and f can be found in Table S9.

Fig. 5.. Ubqn Localizes to the Periphery of Lysosomes and Interacts with v-ATPase

(a) Schematic representation of v-ATPase with cytosolic V1 domain and integral membrane V0 domain. (b) Immunofluorescence staining with Ubqn antibody in fat body of third instar larvae expressing UAS-LAMP1-GFP with Cg-GAL4 driver. Arrowheads indicate Ubqn punctae at the periphery of lysosomes. Scale bars, 5 μm. (c) Western blot for Ubqn in homogenate (H) and lysosome (L) fractions from third instar larvae of iso (control) and ubqn¹ homogenate. Western blots for Vha55 (a V1 subunit), m-CTSL (mature Cathepsin L), Atg8-I, Lamin C, and Tubulin are shown to indicate lysosome enrichment upon subcellular fractionation. (d) Immunofluorescence staining with Ubqn antibody in fat body of third instar larvae expressing VhaSFD-GFP protein trap. Arrowheads indicate Ubqn punctae co-localizing with VhaSFD. Scale bars, 5 μm. (e) Manders’ colocalization coefficients, M1 and M2 for Ubqn and VhaSFD-GFP fluorescence from (e) compared to M1 and M2 for Ubqn and LAMP1-GFP fluorescence from (c). n= 10 (LAMP1), n= 13 (VhaSFD) biologically independent samples. Mean ± s.e.m. ****p< 0.0001. For all panels except 5c, three independent experiments were performed with similar results. For panel 5c, two independent experiments were performed with similar results. Statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 5e can be found in Table S9.

Fig. 6.. Ubiquilin Genetically Interacts with v-ATPase and Mediates Vesicle Acidification

(a) Western blot for V100 with protein lysates from third instar larvae of ubqn¹; GR (control) and ubqn¹ and quantification of full length (fl) and fragmented (fg) V100 protein levels. n= 5 biologically independent samples. Mean ± s.e.m. ns, not significant; **p = 0.0091. (b) Images of ubqn¹ pre-pupa, ubqn¹; v100³/+ and ubqn¹; vhaM8.9^EY03616/+ pharate adults. (c) Live imaging for LysoTracker dye in third instar larval fat body with ubqn¹ clones of v100³/+ and vhaM8.9^5-HA−1890/+ and quantitative analysis of normalized LysoTracker positive punctae number/μm². n= 16 (WT cell), n= 11 (ubqn¹ cell), n= 10 (ubqn¹; v100³/+ cell), and n= 12 (ubqn¹; vhaM8.9^5HA/+ cell) independent samples. Mean ± s.e.m. ns, not significant; **p= 0.0053 (ubqn¹ vs ubqn¹; v100³/+), **p= 0.0013 (ubqn¹ vs ubqn¹; vhaM8.9^5HA/+), ****p< 0.0001. Scale bars, 10 μm. (d) Western blot for p62 with fat body protein lysates from third instar larvae of ubqn¹; GR (control), ubqn¹, ubqn¹; v100³/+, and ubqn¹; vhaM8.9^EY03616/+ and quantification of normalized p62/actin levels. n= 5 biologically independent samples. Mean ± s.e.m. ns, not significant; **p= 0.001. (e) Live imaging of Nile red in fat body of third instar larvae expressing UAS-LAMP1-GFP driven by Cg-GAL4 driver after feeding the larvae with 1 mg/ml acidic nanoparticles (aNPs) loaded with Nile red for 3 hours. Scale bars, 10 μm. (f) Live imaging for LysoTracker in third instar larval fat body with ubqn¹ clones after feeding the larvae with 3 mg/ml aNPs for 3 hours. Scale bar, 10 μm. (g) Western blot for p62 with fat body protein lysates from third instar larvae of ubqn¹; GR (control) and ubqn¹ and quantification of normalized p62/actin levels. Larvae are fed with 0, 0.5, 1, or 3 mg/ml aNPs (3 hours) and transferred to standard fly food (3 hours) before tissue dissection. n= 5 biologically independent samples. Mean ± s.e.m. ns, not significant; *p= 0.0461, **p= 0.002, ***p= 0.0009. For all panels except 6e-g, three independent experiments were performed with similar results. For panels 6e-g, two independent experiments were performed with similar results. All statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 6a,c,d, and g can be found in Table S9.

Fig. 7.. Ubiquilin’s Role in Lysosomal Acidification Is Conserved in Human Neuronal Cells

(a) Western blot for Phospho-S6K (T389), S6K, Phospho-Ulk1(S757), and Ulk1 with Daoy whole cell lysates of siCtrl and UBQLN1+2+4 knockdown (siUBQLNs) cells and quantification of phosphorylated/total protein ratios. n= 3 biologically independent samples. Mean ± s.e.m. **p= 0.0017 (P-S6K/S6K), **p= 0.0099 (P-Ulk1/Ulk1). (b) Western blot for LC3-I and LC3-II with Daoy whole cell lysates of siCtrl and siUBQLNs cells with or without 4 hour Bafilomycin A1 treatment and quantification of LC3-II/LC3-I ratio. n= 5 biologically independent samples. Mean ± s.e.m. ***p= 0.0008, ****p< 0.0001. (c) Immunofluorescence staining with p62 and DAPI in siCtrl and siUBQLNs Daoy cells with or without 4 hour Bafilomycin A1 treatment and quantification of normalized mean p62 fluorescence intensity per cell. n= 5 images (~40 cells). Mean ± s.e.m. **p= 0.0015, ****p< 0.0001. Scale bars, 10 μm. (d) Live imaging for LysoTracker (LT) in siCtrl and siUBQLNs Daoy cells and quantification of normalized mean LT fluorescence intensity per cell. n= 10 (NC) and n= 13 (siUBQLNs) images (~100 cells). Mean ± s.e.m. ****p< 0.0001. Scale bars, 10 μm. (e) Western blot for LAMP1 and v-ATPase subunits: V0a1, V1B2, M8.9, V1C1, and V1H with Daoy whole cell lysates of siCtrl and siUBQLNs cells. fl: full length, fg: fragmented. For all panels, three independent experiments were performed with similar results. All statistics were determined by two-sided Student’s t-test. Statistics source data for Fig. 7a-d can be found in Table S9.

Fig. 8.. Expression of ALS-variant containing UBQLN2 (UBQLN2^P497H) impairs lysosomal degradation

(a-d) Western blots for UBQLN2, dUbqn, Phospho-eIF2α (S51), total eIF2α, p62, Atg8, and V100 with fat body protein lysates from third instar larvae of w; Cg-GAL4/UAS-Luciferase, w; Cg-GAL4/UAS-FLAG-dUbqn, w; Cg-GAL4/UAS-UBQLN2^WT-HA, and w; Cg-GAL4/UAS-UBQLN2^P497H-HA. Two independent experiments were performed with similar results. (e) Live imaging for LysoTracker (LT) in third instar fat body of w; Cg-GAL4/UAS-Luciferase, w; Cg-GAL4/UAS-FLAG-dUbqn, w; Cg-GAL4/UAS-UBQLN2^WT-HA, and w; Cg-GAL4/UAS-UBQLN2^P497H-HA. Two independent experiments were performed by n=10 biologically independent samples with similar results. Scale bars, 20 μm. (f) Ubiquilins regulate autophagic flux through mTOR signaling and lysosomal acidification