Bafilomycin A₁ (Baf), puromycin, blasticidin S and Hanks' balanced salt solution (HBSS) were purchased from FUJIFILM Wako Pure Chemical Corporation. Rapamycin was obtained from LC Laboratories. Hydroxychloroquine (HCQ) was obtained from Cayman Chemical Company. The GAK inhibitor (GAKi; cat. no. 538770), Y-27632 (cat. no. 688000) and cycloheximide were purchased from Calbiochem (Merck KGaA). DAPI was obtained from Sigma-Aldrich (Merck KGaA). Protease inhibitor cocktail and phosphatase inhibitor cocktail were obtained from Nacalai Tesque, Inc. Baf and rapamycin were dissolved in DMSO and used at a concentration of 10 nM, and the same amount of DMSO was used for negative control. HCQ was dissolved in water and used at a concentration of 50 µM. GAKi was dissolved in DMSO and used at the indicated concentrations (3, 10 or 30 µM), and the same amount of DMSO was used for negative control. Y-27632 was dissolved in water and used at a concentration of 10 µM. Cycloheximide was dissolved in DMSO and used at a concentration of 30 µM. Cells treated with Baf (8 or 24 h), rapamycin (24 h), HCQ (8 or 24 h), GAKi (8 or 12 h), Y-27632 (8 h) or cycloheximide (3, 9 or 24 h) at 37°C with 5% CO₂ for an indicated period of time.

The following primary antibodies were used in the present study: Anti-GAK (cat. no. M057-3; Medical & Biological Laboratories Co., Ltd.), anti-phosphorylated (p)-mTOR (cat. no. 5536), anti-mTOR (cat. no. 2983), anti-p-ribosomal protein S6 kinase (S6K; cat. no. 9234), anti-S6K (cat. no. 2708), anti-p-S6 (cat. no. 4858), anti-S6 (cat. no. 2217;), anti-p-AMPK (cat. no. 2535), anti-AMPK (cat. no. 5832), anti-LC3B for immunocytochemistry (cat. no. 2775), anti-p62/sequestosome 1 (SQSTM1; cat. no. 7695), anti-CTSB (cat. no. 31718), and anti-CTSD (cat. no. 2284), anti-ROCK1 (cat. no. 4035; all from Cell Signaling Technology, Inc.), anti-lysosomal associated membrane protein (LAMP)1 (cat. no. sc-20011), anti-LAMP2 (cat. no. sc-18822), anti-β-actin (cat. no. sc-47778), anti-GAPDH (cat. no. sc-32233; all from Santa Cruz Biotechnology, Inc.), anti-LC3B for immunoblotting (cat. no. NB600-1384; Novus Biologicals, LLC), anti-FLAG (cat. no. F1804; Sigma-Aldrich; Merck KGaA) and anti-V5 (R960-25; Invitrogen; Thermo Fisher Scientific, Inc.). Horseradish peroxidase-conjugated secondary antibodies against mouse IgG (cat. no. 115-035-003) and rabbit IgG (cat. no. 711-035-152) were purchased from Jackson ImmunoResearch Laboratories, Inc. The Alexa Fluor 555-conjugated secondary antibody against mouse IgG (cat. no. A21424) and Alexa Fluor 488-conjugated secondary anti-body against rabbit IgG (cat. no. A11034) were obtained from Invitrogen (Thermo Fisher Scientific, Inc.).

The human NSCLC cell lines (A549 and H596), human pancreatic cancer cell line (PANC-1) and human liver cancer cell line (Hep-G2) were purchased from American Type Culture Collection. Cells were cultured in RPMI-1640 (Sigma-Aldrich; Merck KGaA) supplemented with 10% heat-inactivated FBS (Gibco; Thermo Fisher Scientific, Inc.) at 37°C with 5% CO₂. Stable A549/GFP-LC3-mCherry-LC3ΔG cells were cultured under the same conditions as A549 cells with the addition of puromycin (2 µg/ml). RPMI medium without serum and glutamine (RPMI-Gln) was prepared by dissolving RPMI-1640 without glutamine and sodium bicarbonate (cat. no. 05918; Nissui Pharmaceutical Co., Ltd.) in water, followed by autoclaving and then adding sterilized sodium bicarbonate. 293T cells were cultured in DMEM (Sigma-Aldrich; Merck KGaA) supplemented with 10% FBS at 37°C with 5% CO₂.

Target sequences for CRISPR interference (42) were designed using CRISPR direct (http://crispr.dbcls.jp/) provided by the Database Center for Life Science. The target sequences for human GAK were GCA GTC GGC GCT CGA CTT CT (exon 1; N-terminus) or GAG GCC CCC TTT CGT GCG ACA (exon 5; kinase domain). Subsequently, two complementary oligonucleotides with BpiI restriction sites for guide RNAs (gRNAs) were synthesized at Eurofins Genomics and cloned into the pSpCas9(BB)-2A-Puro (PX459) vector (cat. no. 48139; Addgene, Inc.), deposited by the Feng Zhang Laboratory (43). A549 cells (2×10⁶ cells/100 mm dish) were transfected with pX459-gRNA (10 µg) using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Following culture at 37°C with 5% CO₂ overnight, cells were treated with 2 µg/ml puromycin for 2 days. Surviving cells were diluted in growth medium (RPMI-1640 supplemented with 10% heat-inactivated FBS) to prepare a cell suspension (5 cells/ml). Then, 100 µl cell suspension was added to each well of a 96-well plate. The expression of GAK in the expanded colonies was detected by immunoblotting, described in the immunoblotting section, using an anti-GAK antibody to select GAK-depleted colonies. The genome sequences of the edited locus in selected colonies were confirmed by Sanger DNA sequencing performed at Eurofins Genomics, which demonstrated that the expected mutation and frameshifting were present in each exon of GAK (Fig. S1). After cloning and expansion, A549/GAK-knockout (KO) cells were cultured under the same conditions as A549 cells and used for subsequent experiments.

The stable cell line was generated by transfecting A549/GAK-KO cells (2×10⁶ cells/100 mm dish) with plasmid DNA (pMRX-IP-GFP-LC3-mCherry-LC3ΔG; 10 µg) using Lipofectamine 2000 according to the manufacturer's instructions and cultured at 37°C with 5% CO₂ overnight. As previously described (44), the pMRX-IP-GFP-LC3-mCherry-LC3ΔG plasmid used in the present study was recombined with the pMRX-IP-GFP-LC3-RFP-LC3ΔG plasmid, which was a kind gift from Dr N. Mizushima (University of Tokyo, Japan) (45). After selecting the transfected cells using puromycin (1 µg/ml) from the day after transfection, single clones of the cells were isolated and the expression of transfected genes was confirmed by immunoblotting. After cloning and expansion, A549/GAK-KO/GFP-LC3-mCherry-LC3ΔG cells were cultured under the same conditions as A549 cells with the addition of 1 µg/ml puromycin. Since A549/GAK-KO cells are more sensitive to puromycin than A549 cells, puromycin was used at a concentration of 1 µg/ml.

To construct the lentiviral expression vector, pLentiN-GAK, GAK cDNA was amplified by PCR. Total RNA was isolated from A549 cells using the NucleoSpin RNA kit (cat. no. U0955C; Takara Bio, Inc.) according to the manufacturer's protocol. cDNA was synthesized from the total RNA using the SuperScript III First-Strand Synthesis system (cat. no. 18080-051; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. cDNA for GAK was amplified using PrimeSTAR HS DNA Polymerase (cat. no. R010A; Takara Bio, Inc.) according to the manufacturer's protocol. The primers used for amplification were as follows: forward, 5′-CAC CGG TAC CCA CCA TGT CGC TGC TGC AGT CGG CGC TC-3′ and reverse, 5′-AGC TGA ATT CTC ACT TGT CGT CGT CAT CCT TGT AGT CGA AGA GGG GCC GGG AGC CCT G-3′. The thermocycling conditions were as follows: 98°C for 1 min; followed by 30 cycles of 98°C for 10 sec, 60°C for 5 sec and 72°C for 4 min; and 72°C for 5 min. During PCR amplification, FLAG-tag was fused at the C-terminus of GAK. The amplified GAK cDNA was subcloned into the pcDNA6 mammalian expression vector (Invitrogen; Thermo Fisher Scientific, Inc.). The cloned sequences were verified by Sanger sequencing at Eurofins Genomics (Fig. S2). The DNA inserts encoding GAK were then subcloned into the pLentiN lentiviral expression vector, which was a kind gift from Dr Karl Munger (cat. no. 37444; Addgene, Inc.) using restriction enzymes, SpeI and EcoRI (46). To construct the lentiviral short hairpin RNA (shRNA) vectors, pLKO.1-Rho-associated protein kinase (ROCK)1, oligonucleotides for shRNAs targeting ROCK1 on appropriate sites and a non-targeting (NT) control oligonucleotide were synthesized (Table SI). The oligonucleotides were phosphorylated, annealed and inserted into the pLKO.1 puro lentiviral shRNA vector, which was a kind gift from Dr Bob Weinberg (cat. no. 8453; Addgene, Inc.) (47). The production and transduction of lentiviral particles were performed as previously described (48). Briefly, 293T cells (1×10⁷ cells/100 mm dish) were cotransfected with the constructed plasmid (pLentiN-GAK, pLentiN-empty, pLKO.1-ROCK1 or pLKO.1-NT; all 4 µg), pMD2.G (a gift from Dr Didier Trono; cat. no. 12259; Addgene, Inc.; 2 µg) and psPAX2 (a gift from Dr Didier Trono; cat. no. 12260; Addgene, Inc.; 2 µg) using PEI-MAX (cat. no. 24765-1, Polysciences, Inc.; 24 µg) and cultured at 37°C with 5% CO₂. At 2 days post-transfection, the cultured supernatant containing lentivirus was harvested and filtrated. The titer of lentiviral particles was not determined. A549 and A549/GAK-KO cells were plated (2×10⁵ cells/well) in 6-well plate for 24 h before transduction and infected with the prepared lentivirus (150 µl) at 37°C with 5% CO₂ overnight, and selected using blasticidin S (10 µg/ml for A549 cells; 5 µg/ml for A549/GAK-KO cells) or puromycin (2 µg/ml for A549 cells; 1 µg/ml for A549/GAK-KO cells) from the day after infection. In the present study, the following cell lines were established: A549/GAK-KO/GAK-rescued, A549/GAK-KO/sham-rescued, A549/GAK-overexpression (OE), A549/sham-OE, A549/GAK-KO/shROCK1#1, A549/GAK-KO/shROCK1#2 and A549/GAK-KO/shNT cells. A549/GAK-KO/sham-rescued and A549/sham-OE cells were established using pLentiN-empty. The established cell lines were cultured under the same conditions as A549 cells, except for the supplementation of blasticidin S for GAK-rescued cells or puromycin for ROCK1-knockdown cells.

Total cellular proteins were extracted from A549 and A549/GAK-KO cells under the following conditions at 37°C with 5% CO₂: i) Basal-fed condition (RPMI-1640 supplemented with 10% FBS); ii) starved condition using RPMI-Gln (2, 4, 6 or 8 h); iii) basal-fed condition with 30 µM cycloheximide (3, 9 or 24 h); iv) starved condition using HBSS (2 h); v) refed condition (starvation using HBSS or RPMI-Gln for 1 h, followed by fed conditions for 1 h) using lysis buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1.0% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, and protease and phosphatase inhibitor cocktails. Each sample was sonicated on ice for 20 pulses (0.5 sec on/off) to solubilize the aggregated proteins using a Branson 450D Sonifier (Emerson Electric Co.). Protein concentrations were measured using the BCA Protein Assay kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Equal amounts (10 µg) of proteins were resolved by SDS-PAGE (10% gel or 5-20% gradient gel) and transferred onto Immobilon-P PVDF membranes (Merck KGaA). Following blocking with 5% non-fat milk for 1 h at room temperature, the membranes were probed with primary antibodies at 4°C overnight. Primary antibodies used for immunoblotting were as follows: Anti-LC3B (1:2,000), anti-GAK (1:500), anti-LAMP1 (1:1,000), anti-LAMP2 (1:1,000), anti-p-mTOR (1:1,000), anti-mTOR (1:1,000), anti-p-AMPK (1:1,000), anti-AMPK (1:1,000), anti-p-S6K (1:1,000), anti-S6K (1:1,000), anti-p-S6 (1:1,000), anti-S6 (1:1,000), anti-CTSB (1:1,000), anti-CTSD (1:1,000), anti-SQSTM1 (1:1,000), anti-ROCK1 (1:1,000), anti-FLAG (1:1,000), anti-V5 (1:5,000), anti-GAPDH (1:1,000), and anti-β-actin (1:1,000). These membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (anti-mouse IgG or anti-rabbit IgG; both 1:2,500) for 1 h at room temperature. Immunoreactive proteins were detected using Immobilon Western HRP substrate detection reagents (Merck KGaA). Densitometry analysis was performed using a WSE-6300 Luminograph III molecular imager (ATTO Corporation) and CSAnalyzer4 software version 2.3.1 (ATTO Corporation).

A549, A549/GAK-KO and A549/GAK-rescued cells were seeded (4×10⁴ cells/well) onto glass coverslips in a 24-well culture plate and incubated for 48 h. After washing with PBS, cells were fixed with 100% methanol at −30°C for 15 min. After washing with TBST (0.05% Tween-20), cells were blocked with 10% newborn calf serum (Gibco; Thermo Fisher Scientific, Inc.) in TBST at room temperature for 1 h, followed by incubation with primary antibodies in TBST containing 1.5% newborn calf serum and 0.1% BSA (Nacalai Tesque, Inc.) at 4°C overnight. Primary antibodies used for immunocytochemistry were follows: Anti-LC3B (1:100) and anti-LAMP2 (1:100). After washing with TBST, cells were incubated with Alexa Fluor 488-conjugated anti-rabbit IgG and/or Alexa Fluor 555-conjugated anti-mouse IgG (both 1:1,000) along with DAPI at 37°C for 1 h. Cells were washed again with TBST then mounted in ProLong Diamond Antifade Mountant (Invitrogen; Thermo Fisher Scientific, Inc.). Stained cells were visualized using a Zeiss LSM700 confocal laser scanning fluorescence microscope (Zeiss GmbH) equipped with PlanApochromat 63x/1.4 or 100x/1.4 oil DIC (Zeiss GmbH). All images were acquired and processed using ZEN 2012 software (version 8.1.0.484; Zeiss GmbH). Object-based fluorescence intensity was measured using ImageJ software (version 1.50i; National Institutes of Health). The size of lysosomes was measured using ZEN 2012 software. The number of autophagosomes was automatically recognized based on size and intensity thresholds using the Spot function in Imaris ×64 software (version 9.1.0; Oxford Instruments Bitplane) (49).

A549/GFP-LC3-mCherry-LC3ΔG and A549/GAK-KO/GFP-LC3-mCherry-LC3ΔG cells were plated (8×10³ cells/well) in a 96-well plate for 24 h before treatment. Fluorescence intensities derived from GFP-LC3 and mCherry-LC3ΔG were monitored during the 24 h-culture at 37°C with 5% CO₂ using an IncuCyte ZOOM cell imaging system (Essen BioScience) under the following conditions: i) Basal-fed condition (RPMI-1640 supplemented with 10% FBS); ii) basal-fed condition with 10 nM Baf; iii) basal-fed condition with 10 nM rapamycin; iv) basal-fed condition with DMSO; v) starved condition using HBSS; vi) starved condition using RPMI-Gln; vii) starved condition using HBSS with GAKi (3, 10 or 30 µM); and viii) starved condition using RPMI-Gln with GAKi (3, 10 or 30 µM). Autophagic flux was presented as alterations in the relative intensities of GFP/mCherry calculated using Microsoft Excel version 16.51 (Microsoft Corporation) from the measured fluorescence intensities of GFP and mCherry.

A549, A549/GAK-KO and A549/GAK-rescued cells were seeded (4×10⁴ cells/well) into glass-bottom cell culture dishes (Greiner Bio-One International GmbH). After a 48-h culture, cells were incubated under basal-fed (RPMI-1640 supplemented with 10% FBS) or starvation conditions (RPMI-Gln) for 8 h at 37°C with 5% CO₂ and then incubated for 1 h with 50 nM LysoTracker Red DND-99 (cat. no. L7528; Invitrogen; Thermo Fisher Scientific, Inc.) or 1 µM LysoSensor Green DND-189 (cat. no. L7535; Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C with 5% CO₂. Stained cells were visualized using an LSM700 confocal laser scanning fluorescence microscope equipped with PlanApochromat 63x/1.4 oil DIC. All images were acquired and processed using ZEN 2012 software. The object-based fluorescence intensity was measured using ImageJ software.

A549 and A549/GAK-KO cells were plated (2×10⁵ cells/well) in 6-well plate for 24 h before treatment and further cultured for 8 h at 37°C with 5% CO₂ under basal-fed (RPMI-1640 supplemented with 10% FBS), starved condition (RPMI-Gln) or basal-fed with 50 µM HCQ conditions. Then, the cells were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.3) for 1 h at 4°C. The samples were further fixed in 1% osmium tetroxide for 1 h at 4°C, dehydrated in a graded ethanol series (30-100%) and embedded in Quetol 812 epoxy resin (Nisshin EM) for 2 days at 60°C. Ultrathin sections (70 nm) were cut using an Ultracut J microtome. The sections were stained with uranium acetate (Merck KGaA) for 15 min at room temperature in a dark box. After washing with water, the sections were further stained with lead nitrate for 10 min at room temperature and subjected to electron microscopy using a transmission electron microscope (JEM-1400Flash; JEOL, Ltd.).

Total RNA was extracted from A549 and A549/GAK-KO cells using the NucleoSpin RNA kit (Takara Bio, Inc.). Total RNA was reverse transcribed into cDNA using the PrimeScript RT Master Mix kit (Takara Bio, Inc.) according to the manufacturer's instructions. Subsequently, qPCR was performed using the SYBR Premix Ex Taq II Tli RNase H Plus kit (Takara Bio, Inc.). The sequences of primers used for qPCR are listed in Table SI. qPCR was performed using a Thermal Cycler Dice Real-Time System TP800 (Takara Bio, Inc.) and the following thermocycling conditions: Initial cDNA denaturation at 95°C for 30 sec; followed by 45 cycles of denaturation at 95°C for 5 sec; and simultaneous annealing and extension at 60°C for 30 sec. The data were analyzed using Thermal Cycler Dice Real-Time System Software (Takara Bio, Inc.). To determine relative gene expression, the 2^−ΔΔCq method was used (50). mRNA expression levels were normalized to the internal reference gene GAPDH.

The cDNA of ROCK1 was amplified in the same manner as GAK and cloned into the pcDNA6 vector. The primers used for amplification were as follows: forward, 5′-CAC CGC GGC CGC CAC CAT GTC GAC TGG GGA CAG TTT TG-3′ and reverse, 5′-AGC TGG CGC GCC CGC TAG GTT TGT TTG GGG CAA GC-3′. The expression vectors (4 µg each) for GAK-FLAG and ROCK1-V5 were transfected into 293T cells (6×10⁶ cells/dish) using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) and cultured at 37°C with 5% CO₂. At 48 h post-transfection, cells were lysed in 500 µl lysis buffer [50 mM Tris-HCl (pH 8.0), containing 150 mM NaCl and 1.0% NP-40] containing a protease inhibitor cocktail. Immunoprecipitation was performed on transfected cell lysates (100 µl each) using anti-FLAG M2 magnetic beads (cat. no. M8823; Sigma-Aldrich; Merck KGaA; 10 µl). The beads were washed three times with lysis buffer using magnetic separation. Immunoprecipitates were eluted with lysis buffer containing 100 µg/ml 3xFLAG peptide (cat. no. F4799; Sigma-Aldrich; Merck KGaA) and recovered using magnetic separation. Immunoprecipitates or total cell lysates were analyzed by immunoblotting using anti-FLAG (1:1,000) and anti-V5 (1:5,000) primary antibodies according to the aforementioned protocol.

All data are presented as the mean ± standard deviation from at least three independent experiments. The unpaired two-tailed Student's t-test or one-way ANOVA followed by Dunnett's or Bonferroni's post hoc test were used to analyze the data using Microsoft Excel version 16.51 (Microsoft Corporation) with the add-in software, Excel Statistical Program File ystat2008 (Igakutosho-shuppan, Ltd.). One-way ANOVA followed by Tukey's post hoc test and two-way ANOVA followed by Bonferroni's post hoc test were used to analyze the data using SPSS version 27 software (IBM Corp.). P<0.05 was considered to indicate a statistically significant difference.

To analyze the role of GAK in the autophagy-lysosome system, GAK-KO A549-derived cell lines were established (Fig. 1A). GAK was completely knocked out in clones 1-1 and 2-1, but not completely in clone 1-2. The expression of LC3B-II, an autophagosome marker protein, was significantly enhanced in all three clones compared with that in wild-type cells (Fig. 1A and B), whereas the expression of LC3B mRNA in GAK-KO cells was similar to that in wild-type cells (Fig. 1C). Therefore, in the following experiments, clone 1-1 was primarily used for the GAK-KO cells, and clones 1-2 and 2-1 were used as needed. The significant enhancement of LC3B-puncta formation in GAK-KO cells compared with that in wild-type cells was detected by immunofluorescence analysis (Fig. 1D-F).

To analyze autophagic flux, a GFP-LC3-mCherry-LC3ΔG-based system was used (44,45). Brief ly, GFP-LC3-mCherry-LC3ΔG is cleaved into equimolar amounts of GFP-LC3 and mCherry-LC3ΔG. GFP-LC3 is degraded by autophagy, whereas mCherry-LC3ΔG remains in the cytosol and serves as an internal control. Therefore, autophagic flux can be estimated by calculating the GFP/mCherry signal ratio. In the present study, GAK-KO cells showed significantly higher GFP/mCherry ratios under basal-fed conditions compared with those in wild-type cells, suggesting that autophagic flux levels were lower in GAK-KO cells compared with those in wild-type cells (Fig. S3A and B). Baf suppressed autophagic flux, as indicated by an increased GFP/mCherry ratio, and cytoplasmic accumulation of GFP-LC3 was more prominent in wild-type cells compared with that in GAK-KO cells (Figs. 1G and S3C). This result suggested that the additional inhibitory effect of the autophagy inhibitor was decreased in GAK-KO cells because the autophagic flux was already inhibited in the steady-state. By contrast, as indicated by a reduced GFP/mCherry ratio, rapamycin induced autophagic flux, and the elimination rates of GFP-LC3 were more prominent in wild-type cells compared with those in GAK-KO cells (Figs. 1H and S3D). Similarly, starvation conditions (HBSS or RPMI-Gln medium) induced autophagic flux, and the elimination rates of GFP-LC3 were more prominent in wild-type cells compared with those in GAK-KO cells (Figs. 1I, J, S3E and S3F). Collectively, these results suggested that GAK disruption decelerated autophagic flux under basal-fed conditions and reduced the acceleration of autophagic flux under autophagy-inducing conditions, such as starvation.

To analyze the autophagy-related signaling pathways, starvation-refeeding experiments were performed. Both in wild-type and GAK-KO cells, starvation significantly decreased the phosphorylation of mTOR (Ser2448), S6K (Thr389) and S6 (Ser235/236) and induced the phosphorylation of AMPK (Thr172) compared with that in the fed group, indicating that the mTOR signaling pathway was repressed by starvation. Conversely, refeeding significantly induced the phosphorylation of mTOR, S6K and S6 and decreased the phosphorylation of AMPK compared with that in the starved group, indicating that the mTOR signaling pathway was reactivated by refeeding (Fig. S4). Therefore, GAK disruption did not impair starvation and refeeding-induced fluctuations in mTOR signaling. Collectively, these results suggested that the loss of GAK stagnated autophagic flux, but displayed limited effects on the signals that induce autophagosome formation, as well as the formation of autophagosomes.

Transmission electron microscopy (TEM) revealed a notable accumulation of cytoplasmic vesicles in GAK-KO cells (Fig. 2A). To characterize the cytoplasmic vesicles, immunofluorescence analysis was performed. The size distribution of LAMP2-positive vesicles, including lysosomes and autolysosomes, in GAK-KO cells was larger compared with that in wild-type cells (Fig. 2B and C). The protein expression levels of LAMP1 and LAMP2 in GAK-KO cells were significantly higher compared with those in wild-type cells (Fig. S5A and B), whereas the expression levels of LAMP1 and LAMP2 mRNA in GAK-KO cells were similar to those in wild-type cells (Fig. S5C). These data indicated that the de novo synthesis of lysosome-related proteins was not upregulated; however, lysosome-related proteins accumulated in GAK-KO cells under basal-fed conditions. The enlargement of LAMP2-positive vesicles in GAK-KO cells was more prominent under starvation conditions (Fig. 2D-F). Transduction of GAK-KO cells with a GAK-encoding lentivirus canceled the GAK-KO cell-specific phenotype, enlarged LAMP2-positive vesicles (Fig. 2D-F). Additionally, it was shown that the number of LAMP2-positive vesicles in GAK-KO cells was significantly lower compared with that in wild-type cells, and that transduction of GAK-KO cells with a GAK-encoding lentivirus restored the number of LAMP2-positive vesicles (Fig. 2D, E and G). Although the expression of lysosome-related genes was not decreased (Fig. S5), the number of lysosomes decreased, suggesting that lysosomal reformation was suppressed in GAK-KO cells.

The enlargement of lysosomes/autolysosomes is accompanied by acidification defects that promote the accumulation of undegradable cargo in these organelles (14). To assess lysosomal function in GAK-KO cells, GAK-KO cells were stained with LysoTracker dye, which is usually trapped in acidified compartments, such as lysosomes. Decreased staining with LysoTracker dye was detected in GAK-KO cells compared with that in wild-type cells under basal-fed conditions, indicating the elevation of lysosomal pH in GAK-KO cells (Fig. 3A and B). However, under starvation conditions, the intensity of LysoTracker staining in GAK-KO cells was at the same level as that in wild-type cells (Fig. 3A and B). Staining with the LysoSensor dye displayed similar results to the LysoTracker dye (Fig. S6A and B). The maturation rates of CTSB in GAK-KO cells were significantly lower compared with those in wild-type cells under both basal-fed and starvation conditions (Fig. 3C and D). However, mature CTSD expression levels were similar between wild-type and GAK-KO cells (Fig. 3C and D). In addition, the degradation rates of p62/SQSTM1, an autophagy substrate, were also similar between wild-type and GAK-KO cells (Fig. 3E and F). The TEM images revealed a significant accumulation of cytoplasmic vesicles in GAK-KO cells, but the vesicles contained little undigested cargo (Fig. 2A). Collectively, these results indicated that the loss of GAK altered the intra-lysosomal conditions, but essentially maintained the hydrolytic capacity in lysosomes and/or autolysosomes, suggesting that the alteration of intra-lysosomal conditions may not be a major cause of the stagnation of the autophagic flux.

Subsequently, the formation of autophagosomes and autolysosomes under starvation conditions was assessed. The accumulation of enlarged LC3B-positive puncta was detected in GAK-KO cells at 4-24 h after starvation (Figs. 4A, 4B, S7A and S7B). In addition, enlarged LC3B-positive puncta and enlarged LAMP2-positive vesicles were observed side by side in GAK-KO cells (Fig. 4A). Transduction of GAK-KO cells with a GAK-encoding lentivirus canceled the accumulation of enlarged LC3B-positive puncta and the enlargement of LAMP2-positive vesicles (Fig. 4A and B). In GAK-OE cells, similar to wild-type cells, a small number of enlarged LC3B-positive puncta were detected at 24 h after starvation, but these were not observed at 10 h after starvation (Fig. S7C-E). The TEM results verified the accumulation of enlarged autophagosomes and autolysosomes in GAK-KO cells under starvation conditions (Fig. 4C). These observations indicated that the enlarged LAMP2-positive vesicles observed under starvation conditions (Figs. 2D and 4A) were primarily autolysosomes.

Chemical inhibition of GAK kinase activity also induced the accumulation of enlarged LC3B-positive puncta and LAMP2-positive vesicles in a dose-dependent manner under starvation conditions (Fig. 5A-C). GAK-KO cells showed an increase in lysosomal size even under basal-fed conditions (Fig. 2D and F). By contrast, wild-type cells in the presence of the GAKi showed an increase in lysosomal size under starvation conditions, but no increase under basal-fed conditions (Fig. 5A and C), suggesting that the increase in lysosomal size involves more than just the kinase activity of GAK, at least under basal-fed conditions. In addition to A549 cells, H596, PANC-1, and Hep-G2 cells also showed enlarged LC3B-positive puncta in the presence of the GAKi under starvation conditions (Fig. S8). As observed in GAK-KO cells, the GAKi reduced autophagic flux, which was demonstrated by increased elimination rates of GFP-LC3 in a GFP-LC3-mCherry-LC3ΔG-based system, in wild-type cells under starvation conditions (Figs. 5D, 5E, S9A and S9B). Wild-type cells in the presence of 30 µM GAKi showed a more considerable decrease in autophagic flux compared with that observed in GAK-KO cells (Figs. 1I, 1J, 5D and 5E), suggesting that high concentrations of the GAKi may affect proteins other than GAK.

To facilitate the analysis of autophagosome-lysosomal fusion, HCQ, which blocks the late phase of autophagy by increasing the lysosomal pH (51-53), was used. After 8 h of treatment with HCQ, most of the LC3B-positive puncta were enclosed within the LAMP2-positive vesicles in wild-type and GAK-rescued cells (Fig. 6A, B and E). By contrast, in GAK-KO cells, most of the LC3B-positive puncta were in contact with the LAMP2-positive vesicles but were not enclosed within them (Fig. 6A, B, and E). However, after 24 h of treatment with HCQ, most of the LC3B-positive puncta were enclosed within the LAMP2-positive vesicles in GAK-KO and wild-type cells (Fig. 6C-E). After treating GAK-KO cells with HCQ for 8 h, the TEM results also indicated the accumulation of incompletely fused autolysosomes, which presented as snowman-like structures in which spheres were connected (Fig. 6F). Taken together, these results indicated that autophagosome-lysosome fusion was delayed in GAK-KO cells and that GAK kinase activity was required to regulate these steps.

Subsequently, ALR in GAK-KO cells was analyzed. mTOR, which is a key negative regulator of autophagy, is reactivated upon prolonged starvation, triggering ALR (19). Phosphorylation of S6K (Thr389), an indicator of mTOR activity, was reduced by starvation, but induced after 8 h of starvation in wild-type cells. By contrast, S6K phosphorylation remained decreased even after 8 h of starvation in GAK-KO cells (Fig. 7A and B). Therefore, mTOR was reactivated in wild-type cells but not in GAK-KO cells.

Prolonged starvation has been reported to induce the formation of reformation tubules from the autolysosomes, which is a characteristic feature of ALR, and the impairment of ALR results in the enlargement of autolysosomes (19,54-56). The accumulation of enlarged LAMP2-positive vesicles in GAK-KO cells under starvation conditions was observed (Figs. 2D, 4A and 7C). Although significant tubule formation, as reported in previous studies (19,54,56), was not confirmed in A549 cells, the deformation of LAMP2-positive vesicles was observed in wild-type and GAK-rescued cells, but not in GAK-KO cells (Fig. 7C and D). These results suggested that ALR was delayed or impaired in GAK-KO cells, which was consistent with the observation that there were fewer LAMP2-positive vesicles in GAK-KO cells compared with in wild-type cells (Fig. 2G).

Actin cytoskeleton and myosin motor proteins work together to serve essential roles in each stage of autophagy, such as the formation and transport of vesicular cargoes, the fusion between autophagosomes and lysosomes, and the formation of reformation tubules (20,21). GAK disruption stagnated autophagosome-lysosome fusion and impaired ALR, suggesting the dysregulation of actomyosin in GAK-KO cells. Therefore, the ROCK inhibitor Y-27632 was used to suppress the function of ROCK (ROCK1 and ROCK2) (57), which mediates actin-myosin contractility and regulates actin cytoskeleton dynamics (58). Treatment with Y-27632 attenuated the accumulation of enlarged LC3B-positive puncta detected in GAK-KO cells under starvation conditions (Figs. 8A, 8B, S10A and S10B). ROCK1 knockdown also attenuated the accumulation of enlarged LC3B-positive puncta detected in GAK-KO cells under starvation conditions (Fig. 9A-C). Furthermore, the addition of Baf caused enlargement of the autophagosomes in ROCK1-knockdown GAK-KO cells, which indicated that ROCK1 knockdown in GAK-KO cells restored the fusion between autophagosomes and lysosomes to normal rather than impairing autophagosome formation (Fig. S11). However, no interaction was detected between GAK and ROCK1 (Fig. S12). In addition, treatment with Y-27632 (Fig. 8A and C) or ROCK1 knockdown (Fig. 9B and D) reversed the enlargement of LAMP2-positive vesicles in GAK-KO cells under starvation conditions. Collectively, these results indicated that the disturbance of lysosomal dynamics and the stagnation of autophagic flux due to GAK loss involved a pathway for ROCK-mediated cytoskeletal control (Fig. S13).