Fig 2: FIP200 interacts with TBK1. Flp-In T-REx 293 cells inducibly expressing GFP or GFP-FIP200 were treated with 0.1 µg/ml doxycycline for 16 h to induce expression. Afterwards, the GFP proteins were purified overnight using GFP-trap beads. Cleared cellular lysates (input) and purified proteins were immunoblotted for the indicated proteins. The full-length blots are presented in Figure S5.

Fig 3: Loss of FIP200 further enhances TBK1 accumulation and activation caused by defective autophagy. (A) Atg3 WT and atg3 KO MEFs were transfected with either 20 nM negative control siRNA or 20 nM Fip200 siRNA for 72 h, and the cleared cellular lysates were immunoblotted for the indicated proteins. The asterisk indicates a nonspecific band. The full-length blots are presented in Figure S9. The densities of the bands on immunoblots from three independent experiments were quantified and normalized to those of Actin. For each protein, the normalized densities of atg3 KO samples were normalized to that of the respective untransfected sample. The boxes represent the highest and the lowest value, while the centerline shows the mean. Two-way ANOVA with Tukey's multiple comparisons test was used to determine differences between treatments. n.s. = not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. (B + C) atg3 KO MEFs were transfected with either 20 nM negative control siRNA or 20 nM Fip200 siRNA for 72 h and immunostained for TBK1 pS172 (B; fixed in 4% PFA) or TBK1 in combination with either FIP200 or SQSTM1/p62 pS403 (C; fixed in 100% MeOH). For (B), the number, area, and intensity of TBK1 pS172-positive structures of at least 232 cells per cell line were quantified using ImageJ 1.53c and normalized to those of untransfected atg3 KO cells. The boxes represent the highest and the lowest value, while the centerline shows the mean from four independent experiments. For IF imaging, DAPI was used to stain nuclei.

Fig 4: Mutation of residues I131, R133, V134 and Y138 in ATG13 is sufficient to inhibit its interaction with ATG101. (A) Computational alanine scanning of the ATG13-ATG101 interface was performed using the structure of the human ATG13-ATG101 HORMA heterodimer (PDB ID: 5C50, ref. 24) and the DrugScorePPI webserver (ref. 37) (upper panel). ??G denotes binding free energy differences for wild-type residue-to-Ala mutations; residues yielding ??G > 1 kcal mol-1 are considered binding hot spots. In the middle panel, per-residue effective binding energies (?Gbinding) computed by the MM-GB/SA approach (ref. 39, 40) are shown. Residues considered hot spots according to both methods are colored in red. In the lower panel, the localization of these residues in the ATG13 interface is shown. (B) atg13 KO MEFs stably expressing VenusC-ATG101 and VenusN-ATG13 (wild-type or the indicated mutants) were trypsinized and analyzed for Venus fluorescence using a flow cytometer. The median of fluorescence intensity for each sample was normalized to control cells lacking VenusN-ATG13 expression (“none”) and was plotted in a bar diagram representing mean ± SEM. Samples without significant difference display identical letters (Student t test, 2-sample assuming unequal variances). Representative data are plotted in a histogram. Cell lysates were examined for expression of the indicated proteins by immunoblotting. (C) atg13 KO MEFs retrovirally transfected with empty vector or cDNA encoding either HA-ATG13 or the indicated variants were lysed and cleared cellular lysates were subjected to immunopurification with anti-ATG101 antibodies and a protein A/G-Sepharose mix. Purified proteins were subjected to SDS-PAGE and analyzed by immunoblotting for RB1CC1, ULK1, or HA. (D) atg13 KO MEFs stably expressing the indicated HA-tagged ATG13 variants were grown on glass cover slips one day prior to incubation with starvation medium (EBSS) for 2 h, fixation and permeabilization. Immunofluorescence for HA (covance MMS-101P) and RB1CC1 was performed. An inverse confocal laser scanning microscope was used for imaging. Please note that we detected a high number of HA-positive puncta varying in size and intensity, of which only a minor portion colocalized with RB1CC1 puncta. This might be due to the exogenous expression of HA-ATG13 variants. Puncta and colocalization per cell quantification was done using fiji software. A minimum of 89 cells per stimulation was analyzed. Data represent mean + SEM. Statistical analysis using the Student t test, 2-sample assuming unequal variances was performed comparing EBSS to DMEM for each individual cell line. **P < 0.01, ***P < 0.001. Scale bar: 20 µm.

Fig 5: Non-conserved GABARAP residues are key to selective LC3-interacting region (LIR) binding. a, c, e Quantification of endogenous pericentriolar material 1 (PCM1) (a), ULK1 (c), and FIP200 (e) binding to indicated GFP-GABARAP constructs expressed and immunoprecipitated by GFP-TRAP from HEK293A cells. Representative immunoblot of GFP-TRAP immunoprecipitation (IP) experiments is shown in Supplementary Fig. 6e. b, d, f Quantification of endogenous PCM1 (b), ULK1 (d), and FIP200 (f) binding to indicated GFP-LC3B constructs expressed and immunoprecipitated by GFP-TRAP from HEK293A cells. Representative immunoblot of GFP-TRAP IP experiments is shown in Supplementary Fig. 6f. Statistical analysis using one-way analysis of variance (ANOVA) test; mean ± s.d.; data from at least three independent experiments. ****p ≤ 0.0001; ***p ≤ 0.001; **p ≤ 0.01; *p ≤ 0.05. g Affinities (Kd values) of LIR peptides to indicated ATG8 proteins measured by bio-layer interferometry (BLI). Color code represents fold changes relative to wild-type (WT) GABARAP and LC3B, respectively (data are mean ± s.d., n = 2–5)