Introduction

Macroautophagy (hereafter autophagy) is an intracellular degradation system delivering cytoplasmic components to the lysosome via autophagosomes (Mizushima and Komatsu, 2011). Autophagy contributes to intracellular quality control by degrading defective proteins and organelles, such as polyubiquitinated proteins and damaged mitochondria. Alternatively, autophagy is induced by starvation as an adaptation mechanism. These physiological roles of autophagy are associated with various diseases, including cancer and neurodegenerative disorders (Mizushima and Levine, 2020). Therefore, understanding and controlling autophagy in clinical contexts is essential.

Autophagosome formation is performed by the evolutionally conserved autophagy-related (ATG) proteins, which form several functional units to act directly on autophagosome formation (Nakatogawa, 2020). Among them, the Atg1/ULK complex is critical in initiating autophagosome formation by organizing the formation site where all the functional units, such as ATG9 vesicles and ATG8 conjugation system, colocalize (Hama, Kurikawa et al., 2023; Kannangara, Poole et al., 2021; Kishi-Itakura, Koyama-Honda et al., 2014; Ren, Nguyen et al., 2023; Suzuki, Kirisako et al., 2001; Yamamoto, Kakuta et al., 2012; Zhou, Liu et al., 2021). In mammals, the ULK complex includes ULK1/2, ATG13, ATG101, and RB1CC1/FIP200 (Mizushima, 2010; Wong, Puente et al., 2013). Structural and functional analysis has made considerable progress in the yeast counterpart, the Atg1 complex, whose core comprises Atg1 (ULK1/2 homolog), Atg13, and Atg17 (remote homolog of FIP200) (Fujioka, Suzuki et al., 2014; Li, Chung et al., 2014; Lin and Hurley, 2016; Noda, 2024; Ragusa, Stanley et al., 2012). Crystallographic analyses have unveiled the detailed interaction mode of Atg13 with Atg1 and Atg17 and the mechanism of the Atg1 complex organization (Fujioka et al., 2014). Moreover, structure-based in vitro and in vivo functional analyses have revealed that the Atg1 complex undergoes liquid-liquid phase separation to organize the autophagosome formation site on the vacuolar membrane to initiate autophagy (Fujioka, Alam et al., 2020). However, structural and functional analyses of the mammalian ULK complex lag behind that of the yeast Atg1 complex, and structural information on the interactions comprising the core of the ULK complex has long been limited to low-resolution electron microscopy (EM) data (Shi, Yokom et al., 2020). A recent preprint reported the cryo-EM structures of the human ULK1 core complex and the supercomplex between the ULK1 core complex and the class III phosphatidylinositol 3-kinase complex 1 at 4.2–6.84 Å resolution, which discovered the unprecedented interaction between these two critical complexes working at autophagy initiation (Chen, Ren et al., 2023); however, many of the interactions that construct the ULK1 core complex remained unresolved due to the insufficient quality of the density map. Recently, calcium transients on the endoplasmic reticulum surface were proposed to trigger the liquid-liquid phase separation of FIP200 to organize the autophagosome formation site, thereby initiating autophagy (Zheng, Chen et al., 2022). However, the lack of structural information makes elucidating the molecular mechanisms underlying these events challenging.

This study predicted the structure of the ULK1–ATG13–FIP200 complex using the AlphaFold2 multimer (Evans, O’Neill et al., 2021), which unveiled the detailed interactions between ATG13 and FIP200, ATG13 and ULK1, and ULK1 and FIP200. In vitro and in vivo mutational analysis confirmed all the predicted interactions, which were demonstrated to be important for autophagy. Furthermore, we found that the FIP200–ATG13 and ULK1–ATG13 interactions partially complement each other for the autophagy initiation. These findings establish the mechanism of ULK complex formation and provide a structural basis for understanding the ULK complex phase separation and regulation mechanism of autophagy in the mammalian context.

Results

Structural prediction of the ULK1–ATG13–FIP200 complex identifies FIP200-interacting residues in ATG13

Previous studies reported that the C-terminal region of ULK1 and the C-terminal intrinsically disordered region of ATG13 bind to the N-terminal region of the FIP200 homodimer (Alers, Loffler et al., 2011; Ganley, Lam du et al., 2009; Hieke, Loffler et al., 2015; Hosokawa, Hara et al., 2009; Jung, Jun et al., 2009; Papinski and Kraft, 2016; Wallot-Hieke, Verma et al., 2018). However, structural details of the interactions forming the ULK complex remain elusive. To understand the mechanism of the ULK complex formation, we predicted the structure of the FIP200 [1–634 amino acids (aa)] dimer complexed with ATG13 (isoform c, 363–517 aa) and ULK1 (801–1,050 aa) by AlphaFold2 with the AlphaFold-Multimer option (Fig. 1A, Fig. S1). For clarity, flexible loop regions in FIP200 with low pLDDT scores were removed from Figure 1A. The overall structure is consistent with the low-resolution cryo-EM model reported previously (Shi et al., 2020), including the C-shape of the FIP200 dimer and the binding of ATG13 and ULK1 to the FIP200 dimer one by one, resulting in the 1:1:2 stoichiometry of the ULK1–ATG13–FIP200 complex. The two FIP200 molecules are named FIP200^A and FIP200^B, where FIP200^A is the one close to ULK1. Wide interactions were predicted between the four molecules, with a buried surface area of 3,658 Ų for FIP200^A–FIP200^B, 2,195 Ų for FIP200^A–ATG13, 1,966 Ų for FIP200^B–ATG13, 2,140 Ų for ATG13–ULK1, and 1,681 Ų for FIP200^A–ULK1. No interaction was observed between FIP200^B and ULK1.

Figure 1.

One ATG13 molecule binds to the FIP200 dimer through the interaction of residues 365–398 and 394–482 with FIP200^B and FIP200^A, respectively, resulting in a 1:2 stoichiometry of the ATG13–FIP200 complex. Previous studies using hydrogen-deuterium exchange coupled to mass spectrometry identified three regions (M1–M3) in ATG13 responsible for FIP200 binding (Shi et al., 2020). Among these sites, the side-chain of Phe377 in M1, Phe397 in M2, and Phe453 in M3, was deeply inserted into the hydrophobic pockets of FIP200 (Fig. 1B). Direct interaction between FIP200 (1–634) and Atg13 (363–517) was confirmed by isothermal titration calorimetry (ITC) with a Kd value of 6.6 μM, which was severely attenuated by alanine substitution at these three phenylalanine sites (F377A, F394A, F453A; FIP3A) (Fig. 1C, Fig. S2). Furthermore, the ATG13^FIP3A mutant expressed in atg13 KO cells displayed a significantly reduced coprecipitation rate of the endogenous FIP200 (Fig. 1D, E). These results confirm that ATG13 and FIP200 interact with each other via the residues predicted by AlphaFold2 in vitro and in vivo.

ULK1 and ATG13 interact via MIT-MIM interaction similar to yeast Atg1–Atg13

Next, we focused on the interaction between ULK1 and ATG13. We previously determined the crystal structure of the yeast Atg1–Atg13 complex, which revealed that the tandem two microtubule-interacting and transport (MIT) domains in Atg1 bind to the tandem two MIT-interacting motifs (MIMs) of ATG13 (Fig. 2A, right) (Fujioka et al., 2014). The predicted structure of the ULK1–ATG13 complex was quite similar to that of the yeast Atg1–Atg13 complex and was constructed by the ULK1^MIT1–ATG13^MIM(C) and ULK1^MIT2–ATG13^MIM(N) interactions (Fig. 2A, left). According to the predicted structure, Phe470 of ATG13^MIM(N) and Phe512 of ATG13^MIM(C) were inserted into the hydrophobic pocket of ULK1^MIT2 and ULK1^MIT1, respectively (Fig. 2B). Direct interaction between ULK1 (636–1,050) and Atg13 (363–517) was confirmed by ITC with a Kd value of 0.34 μM, which was severely attenuated by alanine substitution at these two phenylalanine sites (F470A, F512A; ULK2A) (Fig. 2C, Fig. S2). In atg13 KO cells, the ULK1 expression level was significantly reduced, suggesting that the lack of ATG13 severely destabilized ULK1. The exogenous expression of ATG13^WT, but not of the ATG13^ULK2A mutant, rescued the endogenous ULK1 expression (Fig. 2D, E). This result suggests that the ULK2A mutation in ATG13 attenuates the ATG13–ULK1 interaction, including in vivo. These results confirm that ATG13 and ULK1 interact via the residues predicted by AlphaFold2 in vitro and in vivo.

Figure 2.

Direct interaction between ULK1 and FIP200 is essential for autophagy

Although AlphaFold2 predicted the direct interaction between ULK1 and FIP200 over a reasonably large contact area (1681 Ų) (Fig. 3A), a previous study reported that ULK1 required ATG13–ATG101 for FIP200 binding in vitro (Shi et al., 2020). In our predicted structure, ULK1^MIT1 and ULK1^MIT2 interact directly with FIP200, with the Leu967 of ULK1^MIT1 and Phe997 of ULK1^MIT2 inserted on the hydrophobic groove of FIP200 (Fig. 3A, B). Although we failed to measure the affinity between ULK1 and FIP200 by ITC due to aggregation upon their mixing, an in vitro pull-down assay indicated that ULK1 directly interacted with FIP200, which was significantly attenuated by the FIP2A (L967A, F997A) mutation (Fig. 3C, D). Consistently, the ULK1^FIP2A mutant lost the interaction with FIP200 while partially retaining the interaction with ATG13 in cells (Fig. 3E, F). These results confirm the direct ULK1–FIP200 interaction in vitro and in vivo. Our results establish the mode of the three independent interactions, ATG13–FIP200, ATG13–ULK1, and ULK1–FIP200, in the ULK complex.

Figure 3.

To investigate the physiological importance of the ULK1–FIP200 interaction, we assessed autophagy flux using the Halo-LC3 processing assay (Rudinskiy, Bergmann et al., 2022; Yim, Yamamoto et al., 2022). Ulk1,2 DKO mouse embryonic fibroblasts (MEFs) expressing FLAG-tagged wild-type ULK1 produce a stronger band of cleaved Halo than cells which do not express ULK1. The intensity of this cleaved Halo band corresponds to the amount of Halo-LC3 delivered to lysosomes in an autophagy-dependent manner. Ulk1,2 DKO MEFs expressing the ULK1^FIP2A mutant had partially but significantly reduced processed Halo bands upon starvation compared to ULK1^WT-expressing cells, confirming that the direct ULK1–FIP200 interaction is crucial, although partially, for autophagy activity (Fig. 3G, H). This partial phenotype would be due to an indirect interaction between ULK1 and FIP200 via ATG13, which was supported by the observation that the ULK1^FIP2A mutant colocalized with FIP200 in cells, although less efficiently than ULK1^WT (Fig. 3I).

Triadic ULK1–ATG13–FIP200 interactions are critical for autophagic flux

We compared the autophagic flux of impaired ATG13–ULK1 and ATG13–FIP200 interactions to confirm the hypothesis that indirect triadic complexes are functional for autophagy. We generated knock-in (KI) cell lines that included ATG13^FIP3A, ATG13^ULK2A, and ATG13^FU5A(F377A, F394A, F453A, F470A, F510A; FU5A) followed by a 3xFLAG-tag in the genomic ATG13 locus of HeLa cells (Fig. 4A, B). Since overexpressed ATG13-FLAG is more than 20-fold higher than its endogenous level, the mutation significance can be evaluated at more physiological expression levels using the KI cell lines (Fig. S3). The ATG13^FIP3A and ATG13^ULK2A mutants displayed partial colocalization with FIP200, whereas the ATG13^FU5A mutant displayed scarce colocalization with FIP200 (Fig. 4C). Next, we expressed Halo-LC3 in these KI cells and performed a Halo processing assay. Cleaved Halo was partially reduced in ATG13^FIP3A and ATG13^ULK2A cells compared to ATG13^WT cells. In contrast, cleaved Halo in ATG13^FU5A cells was almost equivalent to that in the knock-out cells (Fig. 4D, E). These results suggest that ATG13–ULK1 and ATG13–FIP200 bindings complement each other in autophagy function and that ULK1, ATG13, and FIP200 directly bind to each other to organize the robust ULK complex.

Figure 4.

Discussion

The interaction mechanism of ULK1–ATG13–FIP200, the core of the ULK complex, has been extensively studied (Alers et al., 2011; Ganley et al., 2009; Hieke et al., 2015; Hosokawa et al., 2009; Jung et al., 2009; Papinski and Kraft, 2016; Shi et al., 2020). However, most of these studies were not based on structural information, and the knowledge of the interaction mechanism remained limited. Notably, the evidence for ULK1–FIP200 interaction has been limited to in vitro studies (Shi et al., 2020), and its physiological significance has not been characterized. In this study, AlphaFold2-based point mutational analysis provides solid evidence for the direct ULK1–ATG13, ATG13–FIP200, and ULK1–FIP200 bindings in vitro and in vivo. Additionally, the triadic interaction is complementary in the cell.

One major structural difference between the mammalian ULK complex we predicted here and the yeast Atg1 complex is that one ATG13 binds to two FIP200s within the same FIP200 dimer in mammals, whereas one Atg13 binds to two Atg17s in the distinct Atg17 dimers in yeast (Fig. 4F). The latter binding mode of Atg13 bridges Atg17 dimers to each other to form a higher-order assemblage (Yamamoto, Fujioka et al., 2016), which is considered to be the Atg1 complex’s phase separation mechanism (Fujioka et al., 2020). Conversely, ATG13 cannot bridge FIP200 dimers to each other and thus cannot induce a higher-order assemblage of the ULK complex, which necessitates another mechanism for phase separation. A recent study reported that calcium transients on the endoplasmic reticulum surface trigger FIP200 phase separation (Zheng et al., 2022); however, the detailed mechanisms are unresolved. Further studies are required to elucidate the mechanisms of phase separation in organizing the autophagosome formation sites in mammals.

In addition to Atg17, budding and fission yeasts have Atg11 as a closer homolog of FIP200 (Kim, Kamada et al., 2001; Li et al., 2014; Pan, Shao et al., 2020; Sun, Li et al., 2013; Yorimitsu and Klionsky, 2005). Atg11 and Atg17 interact with Atg1 and Atg13, respectively, whereas Atg11–Atg13 and Atg17–Atg1 interactions have not been reported. Arabidopsis thaliana, which belongs to the Archaeplastida group, a supergroup distant from Opisthokonta (including yeasts and mammals), also has Atg1, Atg13, and Atg11 (Burki, Roger et al., 2020; Li et al., 2014). The A. thaliana Atg11 interacts with Atg13 but not with Atg1 (Li et al., 2014). The triadic interaction of Atg1/ULK1, ATG13, and Atg11/Atg17/FIP200 has not been reported for any species other than mammals. This triadic interaction is likely beneficial for cells that form the complicated cellular communities that comprise the metazoans. The most likely physiological significance is the fine-tuning of tissue-specific autophagic activity. Consistent with this idea, several splicing variants of ATG13 are present in vertebrates (Alers, Wesselborg et al., 2014; Hieke et al., 2015; Jung et al., 2009). In humans, at least five splicing variants are known, including the isoform 3 lacking the C-terminal ULK1-binding site (MIMs, Fig. 2A) (Alers et al., 2014; Jung et al., 2009). The avian Atg13 has seven splicing variants, some lacking parts of the FIP200 interaction site (Alers et al., 2014; Hieke et al., 2015). This study found that a single disruption of the FIP200–ATG13 or ULK1–ATG13 interaction results in a partial reduction of autophagy activity (Fig. 4D). In the individual context, the tissue-and organ-specific fine-tuning of the autophagy activity may occur by regulating the expression amount and ratio of splicing variants that lack some of the triad interactions. Generating KI mice with the point mutations identified in this study might help investigate the significance of the triadic binding at the individual level.

Materials and Methods

Cell culture

HeLa cells and MEFs were cultured in Dulbecco’s Modified Eagle medium (DMEM) (043-30085; Fujifilm Wako Pure Chemical Corporation) supplemented with 10% fetal bovine serum (FBS) (172012; Sigma-Aldrich) in a 5% CO₂ incubator at 37°C. Cells were washed twice with phosphate-buffered saline (PBS) and incubated in amino acid-free DMEM (048-33575; Fujifilm Wako Pure Chemical Corporation) without FBS for starvation. atg13 KO HeLa cells (Hama et al., 2023) and Ulk1 Ulk2 double knockout MEFs (Cheong, Lindsten et al., 2011) were described previously.

Plasmids for mammalian cells

Plasmids for stable expression were generated as follows: DNA fragments encoding human ULK1 (NM_003565.4) and ATG13 (Hosokawa et al., 2009) were inserted into the retroviral plasmid pMRX-IP (puromycin-resistant marker) (Kitamura, Koshino et al., 2003; Saitoh, Nakayama et al., 2003) with a 3 × FLAG epitope tag. Site-directed mutagenesis was used to introduce ATG13 FIP3A, ATG13 ULK2A, ATG13 FU5A, and ULK1 FIP2A mutations. pMRXIB-ratLC3B was described previously (Yim et al., 2022). Plasmids for ATG13 KI were generated as follows: sgRNAs targeting the exon 14 of ATG13 (5′-GGCCTCCCCTCACGATGTCT-3′) were inserted into the BpiI site of PX458 [pSpCas9(BB)-2A-GFP; Addgene #48138]. Donor plasmids for ATG13 KI were generated as follows: 500 base-pair (bp) and 682 bp homology arms were amplified from the genomic ATG13 locus around exon 14 and inserted to flank the KI cassettes (described in Fig. 4A) in the donor plasmid (pKnockIn). The coding sequence of 3xFLAG-tagged ATG13 with or without FIP3A, ULK2A, or FU5A mutations was amplified from stable expression plasmids (described above) and inserted just behind the 5’ homology arm.

Stable expression in HeLa cells by retroviral infection

HEK293T cells were transfected with the retroviral plasmid with pCG-gag-pol and pCG-VSV-G (a gift from Dr. T. Yasui, National Institutes of Biomedical Innovation, Health and Nutrition) using FuGENE HD (E2311; Promega) for 4 h in Opti-MEM (31985-070; Gibco). After cell cultivation for 2 days in DMEM, the retrovirus-containing medium was harvested, filtered through a 0.45-µm filter unit (Ultrafree-MC; Millipore), and added to HeLa cells or MEFs with 8 μg/ml polybrene (H9268; Sigma-Aldrich). After 24 h, 2 μg/ml puromycin (P8833; Sigma-Aldrich) or 2.5 μg/ml blasticidin (022-18713; Fujifilm Wako Pure Chemical Corporation) were used to select the stable transformants.

Plasmids for protein preparation

To construct the expression plasmid encoding an N-terminal maltose binding protein (MBP) followed by an HRV3C protease site, the genes were amplified by PCR and cloned into the pET15b vector. The genes were amplified by PCR and cloned into the pET15b-MBP vector for the plasmids encoding N-terminal MBP-tagged ULK1, ATG13, and FIP200. Similarly, for the plasmids encoding N-terminal glutathione S-transferase (GST)-tagged ULK1, the genes were amplified by PCR and cloned into the pGEX6p-1 vector. The NEBuilder HiFi DNA Assembly Master Mix (New England BioLabs) was used to assemble the PCR fragments. PCR-mediated site-directed mutagenesis was used to introduce the mutations that led to the specified amino acid substitutions. All constructs were sequenced to confirm their identities.

Protein expression and purification

E. coli BL21 (DE3) cells were used to express all recombinant proteins. After cell lysis, MBP-FIP200 (1–634) was purified by affinity chromatography using Amylose Resin High Flow (New England Biolabs). Next, MBP was cleaved with the human rhinovirus 3C protease. The eluates were then desalted with 20 mM Tris-HCl, pH 8.0, and 150 mM NaCl utilizing a Bio-Scale Mini Bio-Gel P-6 desalting column (Bio-Rad Laboratories). Subsequently, the cleaved MBP was removed by reappliying FIP200 (1–634) to an Amylose Resin High Flow column. For the pull-down assay, MBP-FIP200 (1–634) eluted from the amylose resin continued to be purified on a HiLoad 26/60 Superdex 200 PG column eluted with 20 mM Tris-HCl, pH 8.0, and 150 mM NaCl.

Affinity chromatography with Amylose Resin High Flow (New England Biolabs) was used to purify the MBP-Atg13 (363–517) and MBP-ULK1 (636–1,050) proteins. They were further purified on a HiLoad 26/60 Superdex 200 PG column and eluted with a buffer containing 20 mM Tris-HCl, pH 8.0, and 150 mM NaCl. Similarly, GST-ULK1 (636–1050) underwent initial purification using a GST-accept resin (Nacalai Tesque, 09277-14). This step was followed by further purification on a HiLoad 26/60 Superdex 200 PG column using the same elution buffer as above.

In vitro pull-down assay

Purified proteins were incubated with GST-accept beads (Nacalai Tesque) at 4°C for 30 min. The beads were washed three times with PBS, and proteins were eluted with 10 mM glutathione in 50 mM Tris-HCl (pH 8.0). SDS-PAGE was used to separate the samples, and protein bands were detected by One Step CBB (BIO CRAFT).

Isothermal titration calorimetry

The binding of FIP200 (1–634) to MBP-ATG13 (363–517) and MBP-ATG13 (363–517) to MBP-ULK1 (636–1,050) was measured by ITC, with a MICROCAL PEAQ-ITC calorimeter (Malvern) and stirring at 750 rpm at 25°C. All proteins were prepared in a solution of 20 mM Tris-HCl, pH 8.0, and 150 mM NaCl. The titration of MBP-ATG13 (363–517) with FIP200 (1– 634) involved 18 injections of 2 μl of the MBP-ATG13 (363–517) solution (229 μM) at 150 s intervals into a sample cell containing 300 μl of FIP200 (1–634) (4 μM). The titration of MBP-ULK1 (636–1,050) with MBP-ATG13 (363–517) involved 18 injections of 2 μl of the MBP-ULK1 (636–1,050) solution (232 μM) at 150 s intervals into a sample cell containing 300 μl of MBP-ATG13 (363–517) (22.9 μM). The isotherm was integrated and fitted with the one-side-binding model of the Malvern MicroCal PEAQ-ITC analysis software. The error of each parameter indicates the fitting error.

Generation of KI cell lines

HeLa cells were transfected with the PX458-based plasmid expressing sgRNA for the ATG13 gene and donor plasmids using FuGENE HD (E2311; Promega). One day after transfection, the KI cells were selected with 1 mg/mL G418 for 14 days. After single clones were isolated, clones positive for the FLAG tag and negative for wild-type ATG13 were screened by immunoblotting.

Preparation of whole cell lysates and immunoblotting

HeLa cells were lysed with 0.2% n-octyl-β-D-dodecyl maltoside in lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, EDTA-free protease inhibitor cocktail (04080; Nacalai Tesque)] for 15 min on ice. After centrifugation at 20,000 × g for 15 min at 4°C, the supernatants were mixed with 20% volumes of 6 × SDS-PAGE sample buffer. The samples were subjected to SDS-PAGE and transferred to Immobilon-P PVDF membranes (IPVH00010; EMD Millipore). The primary antibodies used for immunoblotting were rabbit polyclonal antibodies against FIP200 (17250-1-AP; ProteinTech), ATG13 (13273; Cell Signaling), and ULK1 (8054; Cell Signaling) and mouse monoclonal antibody against β-actin (A2228; Sigma-Aldrich). Horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (111-035-144; Jackson ImmunoResearch) was used as a secondary antibody. The HRP-conjugated mouse monoclonal antibody against DDDDK-tag (M185-7; MBL) was used for FLAG tag detection. Immobilon Western Chemiluminescent HRP Substrate (P90715; EMD Millipore) was used to visualize the signals detected by an image analyzer (ImageQuant LAS 4000; Cytiva). Fiji software (ImageJ; National Institutes of Health) was used to adjust contrast and brightness and quantify the signals (Schindelin, Arganda-Carreras et al., 2012).

Immunoprecipitation

Cells were lysed with 1% 3-[(3-cholamidopropyl)-dimethylammonium]-1-propanesulfonate (CHAPS) in lysis buffer and centrifuged at 20,000 g for 15 min. The supernatants were incubated with anti-FLAG M2 affinity gel (A2220; Sigma-Aldrich) for 1 h at 4°C with gentle rotation. The beads were washed three times in washing buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, and 0.5% CHAPS), and the proteins were eluted with the SDS-PAGE sample buffer.

Immunostaining and fluorescence microscopy

Cells were grown on segmented cover glass (SCC-002; Matsunami Glass IND.), washed with PBS, and fixed with 4% paraformaldehyde (163-20145; Fujifilm Wako Pure Chemical Corporation) in PBS for 10 min. The cells were permeabilized with 50 µg/ml digitonin in PBS for 5 min and blocked with 3% BSA in PBS for 30 min. Primary antibodies in PBS and 3% BSA were added, and samples were incubated for 1 h at room temperature. Cells were washed five times with PBS, incubated with secondary antibodies in PBS with 3% BSA for 1 h at room temperature, and washed five times with PBS. Rabbit polyclonal antibodies against FIP200 (17250-1-AP; ProteinTech) and mouse monoclonal antibodies against FLAG (F1804; Sigma-Aldrich) were used as primary antibodies for immunostaining. Alexa Fluor 488-conjugated anti-mouse IgG (also cross-adsorbed to rabbit IgG and rat IgG) (A-11029; Thermo Fisher Scientific) and Alexa Fluor 555-conjugated anti-rabbit IgG (also cross-adsorbed to mouse IgG) (A-31572; Thermo Fisher Scientific) were used as secondary antibodies. Cells were observed under a confocal microscope (ECLIPSE Ti2; Nikon).

Halo-LC3 processing assay

Cells were treated with 100 nM tetramethylrhodamine (TMR)-conjugated HaloTag ligand (G8251, Promega) for 15 min. After washing twice with PBS, cells were cultured in the starvation medium for 1 h and lysed as described above. Proteins were separated by SDS-PAGE, and in-gel TMR fluorescence was detected using a fluorescent imaging system (Odyssey M; Li-COR).

Statistical analysis

GraphPad Prism 9 software (GraphPad Software) was used for statistical analyses. The statistical methods are described in each figure legend.

Data availability

All the data underlying this study are available in the published article and its online supplemental material.

Acknowledgements

We would like to thank Craig B. Thompson for providing the Ulk1/2 DKO MEFs; Shoji Yamaoka for the pMRXIP plasmid; Teruhito Yasui for the pCG-gag-pol and pCG-VSV-G plasmids; Sekiko Kurazono and Elijah William Caldwell for assistance with protein preparation. This work was supported in part by JSPS KAKENHI Grant Number JP19H05707, JP23K20044, JP23K06667, JP24H00060 (to NNN), JP23K14140, JP22KJ0044 (to Y.H.), JP21H05731, JP23H02429, JP23K27122, JP23H04923 (to Y.F.), JP21H05256 (to H.Y.), JP22H04919 (to NM), CREST, Japan Science and Technology Agency Grant number JPMJCR20E3 (to NNN), ERATO, Japan Science and Technology Agency Grant number JPMJER1702 (to NM), PRIME, Japan Agency for Medical Research and Development Grant number JP20gm6410009 (to YF), and grants from the Takeda Science Foundation (to NNN, YF).

Disclosures

The authors declare no competing interests exist.