STAT3 associates with vacuolar H+-ATPase and regulates cytosolic and lysosomal pH

Abstract

Dysregulated intracellular pH is emerging as a hallmark of cancer. In spite of their acidic environment and increased acid production, cancer cells maintain alkaline intracellular pH that promotes cancer progression by inhibiting apoptosis and increasing glycolysis, cell growth, migration, and invasion. Here we identify signal transducer and activator of transcription-3 (STAT3) as a key factor in the preservation of alkaline cytosol. STAT3 associates with the vacuolar H⁺-ATPase in a coiled-coil domain-dependent manner and increases its activity in living cells and in vitro. Accordingly, STAT3 depletion disrupts intracellular proton equilibrium by decreasing cytosolic pH and increasing lysosomal pH, respectively. This dysregulation can be reverted by reconstitution with wild-type STAT3 or STAT3 mutants unable to activate target genes (Tyr705Phe and DNA-binding mutant) or to regulate mitochondrial respiration (Ser727Ala). Upon cytosolic acidification, STAT3 is transcriptionally inactivated and further recruited to lysosomal membranes to reestablish intracellular proton equilibrium. These data reveal STAT3 as a regulator of intracellular pH and, vice versa, intracellular pH as a regulator of STAT3 localization and activity.

Fig. 1

STAT3 localizes to the lysosomal membrane. a Representative images of A549-RFP-STAT3 cells labeled with the indicated organelle markers (blue). Values, mean percentage of RFP-STAT3 puncta colocalizing with the indicated organelle marker ± SD of three independent experiments with ≥ 10 cells/sample analyzed in each. Colocalization analysis was not applicable (NA) in KDEL-BFP-labeled cells due to the diffuse staining. The areas marked with white squares are magnified in upper right corners. Images of live cells were taken with 60× magnification using Zeiss LSM700 confocal microscope. See Supplementary information, Fig. S1 for colocalization of STAT3 and lysosomes in other cells. b Representative immunoblots of STAT3 and the indicated organelle markers in total cell lysates or the indicated flow throughs (FT) and immunoprecipitates (eluate) from HeLa cells. n = 3. c Representative immunoblots of the indicated proteins in total cell lysates or the indicated fractions of HeLa cells (left) and quantification of P-Y705-STAT3 and P-S727-STAT3 levels relative to total STAT3. d Representative immunoblots (top) and quantification (bottom) of the indicated proteins in lysates of HeLa cell lysosomes purified by iron-dextran method and left untreated or treated with 10 µg/ml proteinase K and 100 µg/ml digitonin for 10 min at 25 °C when indicated. CTSD, cathepsin D. e Representative images (left) and quantification of cytosolic RFP-STAT3 puncta (right) in A549-RFP-STAT3 cells left untreated or treated with 100 ng/mL IL6 for 30 min and stained with Hoechst. Error bars, SD of three independent experiments, with ≥ 10 cells analyzed/sample. P- values were calculated by one-way ANOVA combined with Dunnett’s multiple comparisons test (c) or two-tailed, homoscedastic Student’s t-test (e). The optimal slice thickness (∼350 nm) of confocal images (a, e) was defined by the Zeiss zen software. Scale bar, 10 µm

Fig. 2

STAT3 interacts with V-ATPase on the lysosomal membrane. a V-ATPase subunits detected by high-accuracy mass spectrometry in anti-Flag immunoprecipitates of lysosomal lysates from HeLa cells transiently transfected with pBCMV-STAT3-Flag-puro. Protein amount in relation to STAT3 was estimated from summed peptide intensities (iBAQ algorithm in Maxquant software). The criteria for false discovery rate (FDR) of peptide spectrum match was set to 0.05 and protein FDR to 0.01. Numbers of peptides with identification P-values < 0.01 are indicated. Experiment was repeated once with similar results. b Representative immunoblots of endogenous STAT3 and ATP6V1A (V1A) in anti-STAT3 immunoprecipitates (IP) and lysates of crude lysosome fractions of HeLa cells. n = 3. c Representative immunoblots of the indicated proteins in anti-HA immunoprecipitates (IP) and lysates of crude lysosome fractions of HeLa cells transiently transfected with pCDNA3.1-HA-ATP6V1A. n = 3. d A representative SR-SIM image of a HeLa cell stained with anti-STAT3 and anti-ATP6V0D1 antibodies. The areas marked with white squares are magnified in both sides. Scale bars, 5 µm (middle) and 200 nm (left and right). e Representative images of the indicated SR-SIM projections (left) and colocalization analysis (right) of a punctum staining with anti-STAT3 and anti-ATP6V0D1 (V0D1) antibodies in HeLa cells. Scale bars, 145 nm (yellow) and 200 nm (white). f Representative images (left) and quantification (right) of PLA puncta with antibodies against STAT3 and ATP6V1A in HeLa CRISPR control cell clone (C-4) and STAT3-KO clone (STAT3-KO-11). Lysosomes were visualized by loading with cascade blue dextran. See Fig. 4a for the immunoblot verifying STAT3 depletion. g Representative images of PLAs using antibodies against STAT3 and ATP6V1A in CRISPR control and STAT3-KO human pancreatic duct epithelial cells (H6C7) and human mammary fibroblasts (HMF3) counterstained with DAPI. See Supplementary information, Fig. S2c for the immunoblot verifying STAT3 depletion. h Representative images (left) and quantification (right) of PLA puncta with antibodies against STAT3 and LAMP1 or CD63 in HeLa-C-4 clone. Nuclei were counterstained with DAPI. Cell images in f–h were taken with 60×  magnification using Zeiss LSM700 confocal microscope. The optimal slice thickness (∼350 nm) was defined by the Zeiss zen software. The images of H6C7 cells in g are integrated stacks. Scale bar, 10 µm. Error bars, SD of three independent experiments with ≥ 10 cells analyzed/sample. P-values were calculated by two-tailed, homoscedastic Student’s t-test (f) or one-way ANOVA combined with Dunnett’s multiple comparisons test (h)

Fig. 3

STAT3 interacts with V-ATPase via its coiled-coil domain. a Domain structure of STAT3 with mutations used in b indicated below. SH2, Src homology 2 domain; TAD, transactivation domain. b Quantification of PLA (anti-Flag and anti-ATP6V1A) puncta in HeLa cells expressing the indicated Flag-tagged STAT3 constructs (top). Error bars, SD of three independent experiments with ≥ 10 cells analyzed/sample. P-values were calculated by one-way ANOVA combined with Dunnett’s multiple comparisons test. Rough estimates of the relative expression levels of Flag-tagged STAT3 constructs are shown below the histogram as percentages of the expression of the wild-type STAT3. See Supplementary information, Fig. S3b for representative immunoblots. Bottom, representative images of PLAs taken with 60×  magnification using Zeiss LSM700 confocal microscope. The optimal slice thickness (∼350 nm) was defined by the Zeiss zen software. Lysosomes were visualized with cascade blue dextran. Scale bar, 10 µm. c Representative immunoblots of the indicated proteins in total cell lysates or the indicated fractions of HeLa-STAT3-KO cells reconstituted with STAT3-Y705F (20 µg protein/lane). n = 3

Fig. 4

STAT3 regulates lysosomal pH and activity. a Lysosomal pH determined by FITC/TMR ratio in a HeLa CRISPR control cell clone (C-4), STAT3-KO clones (KO-1 and −11), and KO-11 clone reconstituted with wild-type (WT) or mutated (Y705F, DBM, S727A) STAT3 constructs. Representative immunoblots show STAT3 and ACTB (loading control) protein levels in the clones. Standard curve for pH measurements is shown in Supplementary information, Fig. S4a. b Lysosomal pH determined as in a in CRISPR control and STAT3-KO HMF3 and H6C7 cells. Standard curve for pH measurements is shown in Supplementary information, Fig. S4a. c Volume of acidic compartment (VAC) in HeLa cell clones described in a analyzed by flow cytometer after 5 min staining with 75 nM Lysotracker Green. Relative fluorescence intensities are shown on the left. A representative flow cytometry profile is shown on the right. For other flow cytometry profiles and gating of the cells, see Supplementary information, Fig. S4c and d. d Representative immunoblots of LAMP1, STAT3, and cathepsin B (CTSB) in lysosomal lysates of the indicated HeLa cell clones. The histogram shows ratios between the active (25 kDa) and inactive (31 kDa) CTSB as percentages of the value in C4 control clone. e AlexaFluor 488-Dextran degradation in the indicated HeLa clones loaded with 0.4 mg/ml AlexaFluor 488-dextran for 20 min, washed, and fixed with or without a 4 h chase period. Representative images taken with 60× magnification using Zeiss LSM700 confocal microscope are shown on the right. Error bars, SD of ≥ 3 independent experiments. A minimum of 10 cells/sample were analyzed in a, b, and e. P-values were calculated by one-way ANOVA combined with Dunnett’s multiple comparisons test (a, c) or two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (b, d) for multiple comparisons, or by two-tailed, homoscedastic Student’s t-test (e)

Fig. 5

STAT3 enhances V-ATPase activity. a Representative immunoblots (left) and quantification (right) of the indicated V-ATPase subunits from lysates of HeLa CRISPR control clone (C-4) and STAT3-KO clones (KO-1 and −11). TUBA1A served as a loading control. b Numbers of lysosomal genes whose expression analyzed by RNA-Seq was decreased or increased over ≥ 1.5-fold (P ≤ 0.05) in HeLa-STAT3-KO cells as compared to HeLa-C4 control cells. Lysosomal genes were defined as genes whose protein products localize to lysosomes according to either Gene Ontology or Kyoto Encyclopedia of Genes and Genomes databases. See Supplementary information, Fig. S5b for the list of altered genes. c Representative immunoblots of the indicted proteins from lysates of HeLa cells transfected with the indicated siRNAs 72 h earlier. n = 3. d Representative images (left) and quantification (right) of PLA puncta with antibodies against ATP6V1A (V1A) and ATP6V0D1 (V0D1) in HeLa CRISPR control (C-4) and STAT3-KO (STAT3-KO-11) cells, as well as in HeLa cells transfected with the indicated siRNAs 72 h earlier. DNA was stained with DAPI. Images were taken with 60× magnification using Zeiss LSM700 confocal microscope. The optimal slice thickness (∼350 nm) was defined by the Zeiss zen software. Scale bar, 10 µm. e Quantification of PLA puncta with antibodies against STAT3 and V1A in HeLa cells (left) or Flag and V0D1 in HeLa-STAT3-Flag cells (right). Cells were transfected with the indicated siRNAs 72 h earlier. f Activity of V-ATPase in the presence of 30 µg/mL superfolder-GFP (sfGFP; control) or ΔN-STAT3-sfGFP. V-ATPase was immunoprecipitated with anti-HA magnetic beads from lysosomal lysates of HeLa cells transiently transfected with pCDNA3.1-HA-ATP6V1A. When indicated, the samples were treated with 100 nM bafilomycin A1. Right, standard curve for the measurement of the free phosphate ion used to estimate the ATP consumption. Protein blot for recombinant proteins is shown in Supplementary information, Fig. S5c. Error bars, SD of ≥ 3 independent experiments. A minimum of ten cells/sample were analyzed in d, e. P-values were calculated by one-way ANOVA combined with Dunnett’s multiple comparisons test (a, d), DEseq2 (b), or by two-tailed, homoscedastic Student’s t-test (e, f)

Fig. 6

STAT3 regulates cytosolic pH. a Intensity of lysosomal RFP-STAT3 in A549-RFP-STAT3 cells loaded with 0.4 mg/ml cascade blue dextran for 1 h, chased for 5 h, and treated with EBSS for 4 h, 0.1 µM bafilomycin A1 or 10 µM niclosamide for 1 h, 25 µM EIPA in the absence of NaHCO₃ or the presence of 50 mM propionate for 0.5 h, or with 1 mM LLOMe for 15 min. Histograms show mean lysosomal RFP intensities/cell (top) and distribution of lysosomal RFP intensities in a cell population (bottom). Representative images of live cells taken with 60× magnification using Zeiss LSM700 confocal microscope are shown on the right. Scale bar, 10 µm. b Representative immunoblots of LAMP1 and STAT3 in lysosomal lysates of HeLa cells left untreated or treated as in a, except for LLOMe treatment that was for 1 h. The histogram shows relative ratios of STAT3/LAMP1. Cytosolic acidification caused by these treatments is shown in Supplementary information, Fig. S6a. c Representative immunoblots of the indicated proteins in lysates of HeLa cells treated as in b. The histogram shows relative ratios of P-Y705-STAT3 and P-S727-STAT3. d Representative immunoblots of the indicated proteins in lysates of HeLa cells treated with 25 µM EIPA + 50 mM propionate for the indicated times. CCND1, cyclin D1; BIRC5, survivin. n = 3. e Quantitative PCR analysis of CCND1 mRNA levels in HeLa cells left untreated or treated with 25 µM EIPA + 50 mM propionate for 4 h. ACTA1 mRNA served as an internal control. f Quantification of STAT3/LAMP2 ratios in immunoblots of proteins from lysosomes immunopurified with anti-LAMP1. See Supplementary information, Fig. S6d for a representative blot. g Quantification of PLA (anti-STAT3 and anti-ATP6V1A) puncta in MCF7-vector and MCF7-p95DNErbB2 cells. Error bars, SD of ≥ 3 independent experiments. P-values were calculated by one-way ANOVA combined with Dunnett’s multiple comparisons test (b, c, d, f and g) or by two-tailed, homoscedastic Student’s t-test (e) in comparison with the untreated cells

Fig. 7

Lysosomal STAT3 protects cells against cytosolic acidification. a Cytosolic pH in the indicated HeLa clones determined by image analysis of cells loaded with pHrodo™ Green AM. Standard curve is displayed on the right. b, c Cell death of the indicated HeLa cell clones (b) and HeLa-STAT3-KO-11 cells transiently transfected with vector or STAT3-Y705F (c) left untreated, starved in EBSS for 6 h, or treated with 25 µM EIPA + 50 mM propionate for 24 h was determined by propidium iodine uptake. Error bars, SD of ≥ 3 independent experiments. P-values were calculated by one-way ANOVA combined with Dunnett’s multiple comparisons test (a, b) or by two-tailed, homoscedastic Student’s t-test (c)