ATG16L1 controls mammalian vacuolar proton ATPase

Abstract

The mechanisms governing mammalian proton pump V-ATPase function are of fundamental and medical interest. The assembly and disassembly of cytoplasmic V1 domain with the membrane-embedded V0 domain of V-ATPase is a key aspect of V-ATPase localization and function. Here, we show that the mammalian protein ATG16L1, primarily appreciated for its role in canonical autophagy and in noncanonical membrane atg8ylation processes, controls V-ATPase. ATG16L1 knockout elevated V-ATPase activity, increased V1 presence on endomembranes, and increased the number of acidified intracellular compartments. ATG16L1's ability to efficiently bind V-ATPase was required for its inhibitory role in endolysosomal acidification and for control of Mycobacterium tuberculosis infection in mice. These findings uncover a hitherto unappreciated role of ATG16L1 in regulating V-ATPase, a key pump governing acidification and functionality of the endolysosomal system along with its physiological roles.

Figure 1.. Inactivation of ATG16L1 hyperacidifies intracellular compartments in a V-ATPase-dependent manner.

A. (i) ATG16L1 immunoblot in Huh7^WT and Huh7^ATG16L1-KO cells. (ii)-(v) High Content Microscopy (HCM). (ii) Images from a bank of >15,000 primary objects/cells per experimental group) of Lysotracker Red in Huh7^WT and Huh7^ATG16L1-KO in full media. Masks: LTR (Lysotracker red), red; Hoechst 33342 nuclei, blue; cells/primary objects limits, white. Quantifications: (iii) LTR puncta/ cell, (iv) LTR area, and (v) LTR total intensity/cell. B. Schematic, FLAG-APEX2-ATG16L1 chromosomal insertion in Flp-In T-REx (Tet-ON) HEK293T cells used in proximity biotinylation LC-MS/MS proteomic analyses. FLAG-APEX2-ATG16L1 expression was induced with 10μg/mL of tetracyclin overnight, confirmed by Western blot (inset) and proximity biotinylation and subsequent LC-MS/MS and bioinformatics analyses carried out (described in methods). Volcano plot of ATG16L1 interactors (listed in Table S1) in cells incubated in EBSS or full medium (FM) for 90 min, highlighting V-ATPase components. C. Co-IP analysis of FLAG-ATG16L1 transfected into HEK293T cells in full medium or EBSS with endogenous V₁ components of V-ATPase. D. HCM quantification of (i) ATP6V1A puncta/cell and (ii) their representative images of HeLa^WT and HeLa^ATG16L1-KO in full medium. Masks: ATP6V1A puncta, green; other masks as in A. E. LTR HCM analysis in HeLa cells as in A: (i) LTR puncta/cell and (ii) HCM representative images of HeLa^WT and HeLa^ATG16L1-KO in EBBS with or without 100 nM Bafilomycin A1 (Baf A1) for 90 min stained with LTR (red mask) and Hoechst 33342 (nuclei, blue mask). F. HCM quantification of (i) ATP6V1A puncta/cell and (ii) their representative images of HeLa^WT and HeLa^ATG16L1-KO in EBSS for 90 min. Statistical significance was determined by two-way ANOVA (E) and paired t-test (A, D, F). Data, means ± SE, n ≥ 5 (biologically independent replicates, different plates); each HCM experiment: 1,000 valid primary objects/well, 5 to 6 wells technical replicates per plate. Scale bars: 5 μm.

Figure 2.. Localization analysis of hyperacidification markers by confocal microscopy.

A. Confocal microscopy images of LysoTracker (LTR; red), CD63 (AlexaFluor 488, green), and merged channels (including Hoechst 33342, blue, for nuclei) in HeLa^ATG16L1-KO. B. Confocal images of LTR, Lamp1 (AF488, green), and merged channels in HeLa^ATG16L1-KO. C. Confocal images of ATP6V1A (AF568, red), CD63 (AF488, green), and merged channels in HeLa^ATG16L1-KO. D. LTR (red), EEA1 (AF488, green), and merged channels in HeLa^ATG16L1-KO. E. Confocal images of LTR of LTR (red) and ER marker PDI (AF488, green) in HeLa^ATG16L1-KO. Tracings to the right, line profile intensities corresponding to dashed lines in the insets. Scale bars: 10 μm.

Figure 3.. Determination of luminal pH in endolysosomes of ATG16L1^KO cells.

A. RaVit (ratiometric HCM in vitro test) in live cells transfected with RpHlourin2 (a ratiometric GFP) fused to a compartment-specific protein (e.g. LAMP1). RaVit is an HCM process using automated imaging system (high content microscope CX7 Cellomics). Cells in 96 well lpates are transfected with RpHlourin2 fusions and imaged live. Cells are identified by cell mask (Plasma membrane CellMask Deep Red), for setting regions of interest (ROI A, channel 1, Ex:650/Em:702) and RpHlourin2 (e.g, LAMP1 fusion) for target I (Ex:386/Em:521nm; channel 2) and target II (Ex:485nm/Em:521nm; channel 3). Once the images are collected, the total fluorescence emission intensity of target I and target II is divided to derive a ratio. The experimental value (ratio of intensities at 521 nm upon sequential illumination at 386 nm and 485 nm) is converted into pH value based on a calibration curve of ratios generated with permeabilized transfected cells incubated/equilibrated in pH buffers ranging from 3.5 to 7.5. B. Calibration curve for Lamp1-RpHLuorin2 in cells permeabilized and equilibrated with external buffers of indicated pH.. D. Sequential rendering of representative images: Masks, cells with overlayed algorithm-imposed masks (primary objects, cells) and targets (Lamp1-RpHLuorin2 puncta), followed by corresponding raw images acquired under indicated wavelengths, raw ratio of the channels (excitation at 386nm and 485nm, emission at 521nm), and corresponding heatmap; Image calculator and LUT (lookup table), ImageJ; red, lower pH. Data, means ± SE, two-way ANOVA, n ≥ 5 (biologically independent replicates, different plates; 5 technical replicates/wells (>400 Lamp1-RpHLourin2 transfected cells per well) per plate per experimental group.

Figure 4.. ATG16L1 regulates V-ATPase activity.

A. Scheme, FITC fluorescence quenching by V-ATPase-dependent acidification in FITC-dextran loaded compartments. B. In vitro fluorescence quenching of FITC- dextran (FITC Ex: 485/20, Em: 528/20) in endolysosomal compartments (cell extracts) from HeLa^WT and HeLa^ATG16L1-KO after 90s of ATP addition. C. R.F.U. (relative fluorescence units) difference between HeLa^WT and HeLa^ATG16L1-KO cell extract at 75sec reaction time point. Data, means ± SE (n=4 independent experiments) statistical significance was determined by nonlinear regression (B) and paired t-test (C). D. Scheme, in vitro ATPase assay with magnetically isolated lysosomes; phosphate released from ATP is quantified colorimetrically (650 nm). E. (i) Normalized lysosomal ATPase activity measure (released phosphate) in assays with magnetically isolated lysosomes from HeLa^WT and HeLa^ATG16L1-KO, (ii) As in (i) with or without the V-ATPase inhibitor concanamycin A (1μM), a control for the specificity of the ATPase activity releasing the phosphate. Data, means ± SE (n=4 independent experiments); statistical significance was determined by paired t-test. F. AMP, ADP and ATP levels (i) and calculated Energy charge (ii) in HeLa^WT and HeLa^ATG16L1-KO. Data, means ± SE (n=5 independent experiments); statistical significance was determined by paired t-test.

Figure 5.. ATG16L1 inhibits V-ATPase activity in vitro.

A. Scheme, ACMA assay. ACMA freely exchanges between the buffer and the lumen of magnetically purified lysosomal organelles and is trapped in the lumen upon protonation which also cases quenching of ACMA fluorescence. B. Fluorimetric quantification of ACMA quenching in magnetically isolated lysosomes from HeLa^WT and HeLa^ATG16L1-KO (grown in full medium) after ATP addition (Ex: 360/40, Em: 460/40). C. Quantification of ACMA quenching after subtraction of ConA (Concanamycin A) from non-treated with ConA curves. D. Purified ATG16L1, Coomassie blue stain. E. Fluorimetric quantification of ACMA quenching in magnetically isolated lysosomes from HeLa^WT and HeLa^ATG16L1-KO with or without added purified ATG16L1 protein. F. Quantification of ACMA quenching as in C. Data, means ± SE (n≥3 independent experiments); statistical significance was determined by paired t-test G. Summary of findings: ATG16L1 inhibits proton pumping and ATPase activities of V-ATPase in endolysosomal compartments.

Figure 6.. ATG16L1 regulates V-ATPase assembly.

A, B. ATP6V1A immunoblot (A) and band intensity (ATG16L1) quantification (B) in HeLa^WT and HeLa^ATG16L1-KO. C-E. ATP6V1A and ATP6V0D1 immunoblot (C) and quantification (D,E) without or upon treatment of cells with Bafilomycin A1 (lysosomal degradation inhibitor) or MG132 (proteasomal inhibitor); cell lysates from HeLa^WT and HeLa^ATG16L1-KO. F. ATP6V1A and ATP6V0D1 (immunoblots) after fractionation (whole cell lysates, membrane, and cytosol) of HeLa^WT and HeLa^ATG16L1-KO cell extracts. G. Ratio of V₁/ V₀ in membrane preparations from HeLa^WT and HeLa^ATG16L1-KO cells. H. Summary of findings: V-ATPase assembly is regulated by ATG16L1. Statistical significance was determined by paired t-test (B, G) and by two-way ANOVA (D, E). Data, means ± SE, n ≥ 3 (biologically independent experiments).

Figure 7.. Effects of canonical autophagy and membrane atg8ylation systems on the control of V-ATPase.

A, B. Absence of effects of canonical autophagy factors (FIP200 and ATG13). HCM quantification of LTR puncta/cell in HeLa^WT, HeLa ^ATG16L1-KO, HeLa ^FIP200-KO and HeLa^ATG13-KO cells. Images in C are from a large machine-acquired/processed bank of HCM data. C,D. LTR puncta/cell in HeLa^WT, HeLa ^ATG16L1-KO, HeLa ^ATG3-KO and HeLa^ATG5-KO cells; C, HCM images from the HCM bank. White masks, outlines of cells/primary objects; LTR, red mask (puncta quantified by HCM); Hoechst 33342 (nuclei) blue circle mask. Scale bars: 5 μm. E,F. ATG16L1 immunoblots (E) and band intensity quantification (F) in cell lysates from HeLa^WT and HeLa^ATG5-KO (second lane in each set, MG132 treatment; note no effect). G. LTR puncta/cell in HeLa^WT, HeLa^ATG5-KO, and HeLa^ATG16L1-KO cells transfected with GFP or GFP-ATG16L1⁴ (NM_001190266.1) in Full Media. Statistical significance was determined by two-way ANOVA (A,C), paired t-test (F), and one-way ANOVA (G). Data, means ± SE, n ≥ 5 (biologically independent replicates, different plates); each HCM experiment: 1,000 valid primary objects/well, 5–6 wells technical replicates per plate. Scale bars: 5 μm.

Figure 8.. Effects of ATG16L1^E230 truncation and mutations in ATG16L1 V₁H-interaction sites.

A. Schematic, ATG16L1 full length and ATG16L1^E230 indicating regions of interaction with known binding partners based on existing crystal structures, when applicable. B,C. Co-IP (B) and quantification (C) of ATP6V1A with FLAG, FLAG-ATG16L1^FL and FLAG-ATG16L1^E230. Data, means ± SE, n=3 (biologically independent replicates). D. LTR puncta/cell in HeLa^WT and HeLa^ATG16L1-KO cells transfected with FLAG, FLAG-ATG16L1^FL or FLAG-ATG16L1^E230. Data, means ± SE, n=4 (biologically independent replicates). E. Overlay of AlphaFold 3 predicted interactions between ATG16L1-CC dimer (region 78–230; CC, predicted coiled-coil) and V1H imposed on the known structure of human V-ATPase (PDB ID: 6WM2). ATG16L1A and ATG16L1B, A and B chains of ATG16L1 in its dimer. F. Residues participating in putative interactions predicted by AF3 Multimer and Chimera X analyses. Site detail was flipped and rotated to enable viewing of a putative ATG16L1-ATP6V1H binding pocket; the mutated amino acids in ATG16L1 are indicated (CC-chain A in yellow, and CC-chain B in orange). G. LTR puncta/cell in HeLa^WT and HeLa^ATG16L1-KO cells transfected with GFP, GFP-ATG16L1^FL, GFP-ATG16L1^*V1H in EBSS. Data, means ± SE, n=3 (biologically independent replicates). H, I. Co-IP (H) and quantification (I) of ATP6V1A with GFP, GFP-ATG16L1^FL and GFP-ATG16L1^*V1H. Data, means ± SE, n=4 (biologically independent replicates). HCM experiments in D and G: 1,000 primary objects/well, 4–6 wells technical replicates per plate. Statistical significance was determined by one-way ANOVA.

Figure 9.. Susceptibility of Atg16l1^E230 mice to M. tuberculosis infection.

A, B Kaplan-Meier survival rate of mice infected with M. tuberculosis Erdman. Mouse groups (littermates): Atg16L1^E230/E230, Atg16L1^WT/WT (homozygous control) and Atg16L1 ^E230/WT (heterozygous control). Aerosol infection: low dose (A); intermediate dose (B). Gehan-Breslow-Wilcoxon and Mantel-Cox (log-rank) tests.

Figure 10.. Model of ATG16L1 as a regulator of mammalian V-ATPase.

Top, ATG16L1 AlphaFold predicted structure and its WD and CCD regions, the latter containing the section of ATG16L1 used in modeling its interactions with V₁H. Bottom left: Under homeostatic conditions (e.g. cells with intact lysosomes) ATG16L1 inhibits V-ATPase to maintain optimal normo-acidic pH of endolysosomal organelles and prevent futile ATP hydrolysis. In the absence of ATG16L1 this process is perturbed, and V-ATPase is excessively activated causing hyperacidification of endolysosomal compartments. Right, under stress conditions such as starvation or endomembrane injury, ATG16L1 is mobilized as atg8ylation E3 ligase to support canonical autophagy or noncanonical processes known as CASM or VAIL.