V-ATPase is a universal regulator of LC3-associated phagocytosis and non-canonical autophagy

Abstract

Non-canonical autophagy is a key cellular pathway in immunity, cancer, and neurodegeneration, characterized by conjugation of ATG8 to endolysosomal single membranes (CASM). CASM is activated by engulfment (endocytosis, phagocytosis), agonists (STING, TRPML1), and infection (influenza), dependent on K490 in the ATG16L1 WD40-domain. However, factors associated with non-canonical ATG16L1 recruitment and CASM induction remain unknown. Here, using pharmacological inhibitors, we investigate a role for V-ATPase during non-canonical autophagy. We report that increased V0-V1 engagement is associated with, and sufficient for, CASM activation. Upon V0-V1 binding, V-ATPase recruits ATG16L1, via K490, during LC3-associated phagocytosis (LAP), STING- and drug-induced CASM, indicating a common mechanism. Furthermore, during LAP, key molecular players, including NADPH oxidase/ROS, converge on V-ATPase. Finally, we show that LAP is sensitive to Salmonella SopF, which disrupts the V-ATPase-ATG16L1 axis and provide evidence that CASM contributes to the Salmonella host response. Together, these data identify V-ATPase as a universal regulator of CASM and indicate that SopF evolved in part to evade non-canonical autophagy.

Figure 1.

Rab7, LAMP1, and V-ATPase recruit prior to LC3 during LAP. (a and b) Representative time-lapse confocal microscopy images of OPZ phagocytosis in RAW264.7 cells expressing GFP-hLC3A and RFP-Rab7 (a) or LAMP1-RFP (b). Asterisks denote phagosomes and arrows mark acquisition of fluorescent markers. Scale bar, 5 μm, cropped box, 5 μm; time, min:s. (c) Quantification of fluorescent marker acquisition from 10 phagosomes in relation to GFP-hLC3A (time 0). Data represent mean ± SD. ****, P < 0.0001, unpaired t test. (d) Confocal images of OPZ containing phagosomes after 20 min, stained for GFP-hLC3A and ATP6V1A. Scale bar, 5 μm. (e) Quantification of phagosome markers from d. Data represent mean ± SEM from three independent experiments with >100 phagosomes analyzed per experiment. ****, P < 0.0001; **, P < 0.001, one-way ANOVA followed by Tukey multiple comparison test. (f) Representative time-lapse confocal microscopy images of LTR and GFP-hLC3A during OPZ phagocytosis. Asterisks denotes phagosomes and arrow marks acquisition of GFP-hLC3A. Scale bar, 2 μm; time, s. (g) Quantification of phagosome markers from f. Data represent mean ± SD from eight phagosomes.

Figure S1.

LAMP1 recruits to entotic vacuoles prior to LC3. (a) Representative time-lapse widefield microscopy images of entosis in MCF10A cells expressing GFP-hLC3A and LAMP1-RFP. Arrows mark acquisition of fluorescent markers. Scale bar, 15 μm; time, min:s. (b) Analysis of fluorescent marker intensity at the entotic vacuole membrane over time. (c) Quantification of LAMP1-RFP acquisition from 23 entotic vacuoles in relation to GFP-hLC3A (time 0). Data represent mean ± SD. ****, P < 0.0001, unpaired t-test.

Figure 2.

NADPH oxidase and V-ATPase are both required for LAP. (a) Confocal DIC and GFP-hLC3A images of OPZ phagocytosis in RAW264.7 cells treated with DPI (5 μM), GSK (10 μM), BafA1 (100 nM), or ConA (100 nM). Cropped insert showing pseudo-colored GFP fluorescence intensity. Asterisks denote phagosomes. Scale bar, 5 μm, cropped boxes, 7 μm. (b) Quantification of GFP-hLC3A intensity at phagosomes in cells treated as above. Data represent the mean of individual phagosome measurements from three independent experiments (red triangle, green square, blue circle); ****, P < 0.0001; ***, P < 0.0007, one-way ANOVA followed by Dunnett’s T3 multiple comparison test. (c) Representative confocal DIC images of NBT/formazan deposits in phagosomes from control, DPI, GSK, BafA1, and ConA treated RAW264.7 cells. Inserts are zoomed in phagosomes. Scale bar, 5 μm, cropped boxes 7 μm. (d) Luminol measurements of ROS during OPZ phagocytosis in RAW264.7 cells treated with DPI or BafA1. Data represent the mean ± SEM from three independent experiments. AU, arbitrary unit.

Figure 3.

LAP induction involves elevated ROS and raised pH in phagosomes. (a) Confocal microscopy measurement of phagosome LysoTracker intensity over time in RAW264.7 cells stimulated with OPZ ± DPI (5 μM) pretreatment. Data represent mean ± SEM of nine individual phagosomes across multiple independent experiments. (b) RAW264.7 cells were pretreated with DPI (5 μM) or DPI + CQ (100 μM) prior to stimulation with OPZ for 25 min. Representative confocal images of LysoTracker and GFP-hLC3A are shown. Scale bar, 5 μm. (c and d) Quantification of Lysotracker (c) and GFP-hLC3A (d) at phagosomes in cells treated as in b. Data represent the mean of individual phagosome measurements from three independent experiments (red triangle, green square, blue circle). ****, P < 0.0001, one-way ANOVA followed by Tukey multiple comparison test. (e) Luminol measurements of ROS during OPZ phagocytosis in primary murine BMDC and BMDM cells. Data represent the mean ± SEM from three independent experiments. (f) Representative confocal images of LysoTracker staining and DIC in BMDC and BMDM cells after stimulation with OPZ for 25 min. Arrows denote phagosomes in BMDMs and arrowheads denote phagosomes in BMDCs. Scale bar, 20 μm. (g) Quantification of phagosome Lysotracker intensity in cells treated as in f. Data represent the mean of 50 individual phagosomes. ****, P < 0.0001, unpaired t test. (h) Representative confocal images of BMDC and BMDM cells stimulated with OPZ for 25 min and stained for LC3. Arrows denote LC3 positive phagosomes. Scale bar, 10 μm. (i) Quantification of LC3 positive phagosomes from cells stimulated with OPZ for the indicated times. Data represent mean ± SEM from three independent experiments. ***, P < 0.0001; ***, P < 0.0003; **, P < 0.003, two-way ANOVA with multiple comparisons. AU, arbitrary unit.

Figure S2.

Confocal microscopy measurements of phagosome LysoTracker intensity over time in RAW264.7 cells pre-treated with either DPI (5 μM) or BafA1 (100 nM) prior to addition of OPZ. Data represent mean values normalized to time 0 ± SEM of nine individual phagosomes across multiple independent experiments.

Figure S3.

Effect of E64d treatment on LAP. Confocal images of GFP-hLC3A expressing RAW264.7 cells treated with DPI (5 mM), CQ (100 mM), or E64d + pepstatin (10 μg/ml) as indicated, followed by addition of OPZ for 25 min. Asterisks denote phagosomes and arrows mark GFP-LC3A positive phagosomes. Scale bar, 5 μm.

Figure 4.

Modulation of ROS regulates phagosomal V-ATPase recruitment during LAP. (a) Representative confocal images of ATP6V1A following 25 min OPZ stimulation of RAW264.7 cells pretreated with either DPI (5 μM) or BafA1 (100 nM). Asterisks denote phagosomes. Scale bar, 5 μm. (b) Quantification of ATP6V1A phagosome intensity from cells treated as in a. Data represent the mean of individual phagosome measurements from three independent experiments (red triangle, green square, blue circle). ****, P < 0.0001, one-way ANOVA followed by Tukey multiple comparison test. (c) Quantification of ATP6V1A phagosome intensity from RAW264.7 cells pretreated with DPI or DPI + CQ (100 μM). Data represent the mean of individual phagosome measurements from three independent experiments (red triangle, green square, blue circle). ****, P < 0.0001, one-way ANOVA followed by Tukey multiple comparison test. (d) Representative confocal images of ATP6V1A following 25 min OPZ stimulation of BMDC or BMDM cells. Asterisks denote phagosomes. Scale bar, 5 μm. (e) Quantification of ATP6V1A phagosome intensity from BMDC and BMDM cells pretreated ± DPI (5 μM). Data represent the mean of individual phagosome measurements from three independent experiments (red triangle, green square, blue circle). ****, P < 0.0001, one-way ANOVA followed by Tukey multiple comparison test.

Figure 5.

SaliP induces non-canonical autophagy and V-ATPase V0–V1 engagement. (a) Cartoon showing the differential effects of BafA1 and SaliP on V0–V1 association. (b) Confocal images of LTR staining in WT MCF10A cells treated with BafA1 (100 nM) or SaliP (2.5 μM). Scale bar, 15 μm. Quantification of LTR intensity. Data represent mean ± SD from eight fields of view. ****, P < 0.0001. (c) WT and ATG13^−/− MCF10A cells were treated with BafA1 (100 nM) or SaliP (2.5 μM) for 1 h. Western blotting was performed to probe for LC3 (I and II forms marked) and GAPDH. Quantification of LC3II/I levels. Data represent mean ± SEM from three independent experiments. ***, P < 0.0001; *, P < 0.02, unpaired t test. (d) Representative confocal images of ATG13^−/− MCF10A cells stained for LAMP1 and GFP-hLC3A following treatment with SaliP (2.5 μM) for 1 h. (e) Confocal images of entotic corpse vacuoles in WT MCF10A cells treated with monensin (100 μM) for 1 h and stained for LAMP1 and GFP-hLC3A. Scale bar, 10 μm. (f) Quantification of GFP-hLC3A recruitment to LAMP1 positive entotic corpse vacuoles (ECVs) following treatment with monensin (Mon, 100 μM), SaliP (2.5 μM), or BafA1 (100 nM) for 1 h. Data represent mean ± SEM from three independent experiments. ****, P < 0.0001; **, P < 0.008, unpaired t test. (g) Quantification of GFP-hLC3A recruitment to entotic vacuoles during inner cell death in WT MCF10A cells treated ± BafA1 (100 nM) or SaliP (2.5 μM). Data represent mean ± SEM from three independent experiments. ****, P < 0.0001, unpaired t test. (h) Quantification of duration of GFP-hLC3A recruitment to entotic vacuoles during inner cell death in cells treated ± SaliP (2.5 μM). Data represent mean ± SD from five examples. (i) Widefield GFP-hLC3 and DIC time-lapse images of entotic cell-in-cell structures in MCF10A cells treated ± SaliP (2.5 μM). Asterisks denote inner cells and arrows denote GFP-hLC3A recruitment to entotic vacuoles. Scale bar, 10 μm; time, h:min. Source data are available for this figure: SourceData F5.

Figure 6.

Increased V0–V1 engagement is associated with activation of non-canonical autophagy. (a) WT MCF10A cells were treated with BafA1 (100 nM), SaliP (2.5 μM), monensin (100 μM), or CQ (100 μM) for 1 h. Following fractionation, input, membrane, and cytosol fractions were probed for ATP6V1A, ATP6V0D1, and LC3 by Western blotting. LC3II/LC3I ratios are shown below. (b) Quantification of V0–V1 association from experiments as shown in a. Data represent mean ± SD from four to seven independent experiments. ***, P < 0.003; **, P < 0.01; *, P < 0.04, unpaired t test. (c) HeLa cells were stimulated with either mTOR inhibitor AZD8055 (1 μM) or TRPML1 agonist C8 (2 μM) for 90 min. Following fractionation, membrane and cytosol fractions were probed for ATP6V1B2, ATP6V0D1 by Western blotting. (d) Quantification of V0–V1 association from experiments as shown in c. Data represent mean ± SD from two independent experiments. Source data are available for this figure: SourceData F6.

Figure 7.

ATG16L1 interacts with V-ATPase during non-canonical autophagy. (a and b) ATG16L1^−/− RAW264.7 cells and those re-expressing Flag-tagged WT and K490A ATG16L1, were treated with OPZ for 25 min (a) or STING agonist DMXAA (50 μg/ml) for 1 h (b). Input lysates and Flag immunoprecipitations were probed for ATP6V1A, ATG16L1, and ATG5 by Western blotting. (c) ATG16L1^−/− HCT116 cells and those re-expressing FLAG-tagged WT and K490A ATG16L1 were treated with TRPML1 agonist C8 (2 μΜ) for 30 min. Input lysates and FLAG immunoprecipitations were probed by Western blotting as above. (d) ATG16L1^−/− RAW264.7 cells and those re-expressing FLAG-tagged WT and K490A ATG16L1 were treated with mTOR inhibitor PP242 (1 μM) for 1 h. Input lysates and Flag immunoprecipitations were probed by Western blotting as above. (e) Confocal images of RAW264.7 cells expressing WT or K490A ATG16L1 treated with OPZ (25 min), DMXAA (1 h), or PP242 (1 h) and stained for LC3. Asterisks denote phagosomes. Scale bar, 10 μm. (f) Confocal images of GFP-rLC3B HCT116 cells expressing WT or K490A ATG16L1 treated with C8 (30 min). Scale bar, 15 μm. (g) ATG16L1^−/− HCT116 cells and those re-expressing FLAG-tagged WT and K490A ATG16L1, were treated with BafA1 (100 nM) or SaliP (2.5 μM) for 1 h. Input lysates and FLAG immunoprecipitations were probed by Western blotting as above. Source data are available for this figure: SourceData F7.

Figure S4.

Quantification of FLAG-ATG16L1 co-immunoprecipitation of V-ATPase subunits. (a) ATG16L1^−/− HCT116 cells and those re-expressing FLAG-tagged WT and K490A ATG16L1 were treated with monensin (100 μM) + IN-1 (1 μM) for 45 min. Input lysates and FLAG immunoprecipitations were probed for ATP6V1A, ATG16L1, and ATG5 by Western blotting. (b) Confocal images of GFP-rLC3B HCT116 cells expressing WT or K490A ATG16L1 treated as in a. Scale bar, 15 μm. (c and d) Quantification of ATP6V1A pulldown in Flag-ATG16L1 immunoprecipitations as shown in Fig. 7, a–d, and Fig. S4 a. Data represent means normalized to unstimulated controls from two to three independent experiments. (e) ATG16L1^−/− RAW264.7 cells and those re-expressing Flag-tagged WT and K490A ATG16L1 were treated with OPZ for 25 min. Input lysates and Flag immunoprecipitations were probed for ATP6V1B2 and ATG16L1 by Western blotting. Source data are available for this figure: SourceData FS4.

Figure 8.

The Salmonella effector protein SopF blocks LAP and non-canonical autophagy. (a) Confocal images of GFP-hLC3A expressing RAW264.7 cells transfected with mCherry-SopF or empty vector and stimulated with OPZ for 25 min. Asterisks denote phagosomes and arrows mark GFP-LC3A positive phagosomes and dashed line marks outline of the cell. Scale bar, 5 μm. (b) Quantification of the percentage of phagocytosing mCherry-SopF or empty vector expressing cells that contain GFP-hLC3A positive phagosomes following OPZ stimulation. Data represent the mean ± SEM from three independent experiments. ***, P < 0.0003, unpaired t test. (c) Representative confocal DIC images of NBT/formazan deposits in phagosomes from empty vector and mCherry-SopF expressing RAW264.7 cells. Scale bar, 5 μm. (d) WT MCF10A cells expressing mCherry-SopF or empty vector were treated with SaliP (2.5 μM) or monensin (100 μM) for 1 h. Following fractionation, input, membrane, and cytosol fractions were probed for ATP6V1A, ATP6V0d1, and mCherry by Western blotting. Data representative of two repeats. Source data are available for this figure: SourceData F8.

Figure 9.

Non-canonical autophagy contributes to the Salmonella host response. (a and b) ATG16L1^−/− HCT116 cells and those re-expressing FLAG-tagged WT and K490A ATG16L1, were infected with either WT (a) or ΔsopF Salmonella (b) for 1 h and imaged by confocal microscopy. Scale bar, 10 μm. (c) Quantification of GFP-rLC3B positive Salmonella in cells treated as in a and b. At least 100 bacteria were counted for each condition. Data represent mean ± SD from three independent experiments. **, P < 0.006; ****, P < 0.0001, one-way ANOVA followed by Tukey multiple comparison test.

Figure 10.

Model of CASM activation. (1) Stimuli that induce perturbations in endolysosomal ion and pH balance drive (2) V0–V1 engagement that (3) promotes the recruitment of ATG16L1 through its WD40 CTD and results in CASM. SopF modifies V-ATPase to inhibit ATG16L1 interaction while BafA1and ConA interfere with V-ATPase activity, both resulting in inhibition of CASM. During LAP, NOX2-dependent ROS production consumes phagosomal H⁺ protons in a DPI and GSK sensitive manner, which alters phagosome pH and drives the SopF-sensitive interaction between V-ATPase and ATG16L1, which then directs ATG8 lipidation to phagosomes.