Heteromeric clusters of ubiquitinated ER-shaping proteins drive ER-phagy – Nature

To study the function of ARL6IP1 and the consequences of its disruption, we investigated Arl6ip1 KO mice and different cell culture models, including fibroblasts from a patient carrying the homozygous mutation AR6IP1^K193Ffs, from the patient’s father (unaffected and carrying the heterozygous allele) and from a healthy individual. Mouse experiments were performed on a C57BL/6 background after backcrossing for more than four generations. Mice were maintained in groups of up to 3 mice per cage at 21 ± 2 °C, air humidity of ≥45%, 15-fold air exchange, 14–10 h day–night cycle and maximum 500 lx. Mice had free access to standard mouse chow and water. Littermates of the same sex were randomly assigned to experimental groups. Experiments were conducted in a blinded manner with regard to cell, mouse and human genotypes. Figure legends include details of replicate experiments used to generate datasets. All animal experiments were approved by the Thüringer Landesamt für Lebensmittelsicherheit und Verbraucherschutz (registration numbers 02-055/14 and UKJ-17-006). Studies using human fibroblasts were approved by the local ethics committee.

Plasmids are presented in Supplementary Table 1. cDNAs were cloned into the pDONR223 vector using a BP Clonase Reaction kit (Invitrogen, 11789100) and further recombined into the Gateway destination vectors pcDNA5-FRT/TO-N-mCherry-EGFP, pcDNA3.1-N-HA, pHAGE-GFP, pcDNA3.1-N-Flag, pcDNA3.1-C-Flag, pcDNA3.1-N-SBP-Flag, pGEX6-GST, and the biomolecular complementation affinity purification system vectors pDEST-V1-ORF, pDEST-V2-ORF, pDEST-ORF-V1 and pDEST-ORF-V2 using a LR Clonase Reaction kit (Invitrogen). Plasmids encoding untagged ARL6IP1, Trx–His-tagged ARL6IP1 and Trx–His–ARL6IP1-7KR were generated by subcloning ARL6IP1 from pGEX6P1 into pPal7 and pET32a, respectively. ARL6IP1 in pGEX6P1 was cloned by PCR using pClneo-ARL6IP1 WT-GFP as template. The three sgRNA guides of ARL6IP1 were cloned into the pLentiCRISPR v2 vector.

Primers are presented in Supplementary Table 2, primary antibodies in Supplementary Table 3 and secondary antibodies are presented in Supplementary Table 4.

Generation of Arl6ip1 KO mice

The EUCOMM embryonic stem cell clone HEPD0752_7_A11 (Source Bioscience) was injected into C57BL/6J donor blastocysts. Next, 15–30-week-old F₁ female offspring from C57BL/6J and CBA/J matings served as foster mice. Resulting chimeras were mated with C57BL/6J. For all experiments, littermates were used, which had been backcrossed for at least four generations. Genotyping was performed by PCR with three primers (Arl6ip1-forward: 5′-GTAATATTCTGAGCACTGCCT-3′, Arl6ip1-KO-reverse: 5′-TGCCATAATGACCTAATACTGTTGTG-3′, Arl6ip1-WT-reverse: 5′-CTAAGCACAGGCTATGAACC-3), which produced a WT band of 537 bp or the KO band of 350 bp.

Generation of ARL6IP1 CRISPR–Cas9 KO cell lines

The ARL6IP1 knockout U2OS cell line was generated using a lentiviral CRISPR–Cas9 system. sgRNAs are reported in Supplementary Table 5 (design at https://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-design). The lentiviral plasmids were generated as previously reported²⁸. The forward and reverse oligonucleotides were annealed and phosphorylated using T4 polynucleotide kinase (BioLabs). The oligonucleotides were ligated into the Cas9 vector pLenti-Puro-v2 and pLenti-Puro-EGFP using the BsmBI site. The lentiviral plasmids were co-transfected into HEK293T cells together with the packaging vectors pPAX2 and pDM2.G for lentivirus production. After 48 h, the medium containing lentiviral particles was collected, centrifuged to remove dead cells and stored at –80 °C. To generate the ARL6IP1 KO cell line, fresh U2OS or HeLa cells were infected with lentiviral particles with the three different sgRNAs for 48 h and then selected using 5 μg ml^–1 of puromycin. The surviving cells were maintained in DMEM supplemented with 2 μg ml^–1 puromycin. When using pLenti-Puro-EGFP as backbone of the sgRNA, cells were also FACS-sorted for GFP expression (SONY SH800S Cell Sorter, version 2.1.6). KO was verified by western blotting.

Motor performance

For the beam walk test, mice were placed on an elevated beam of 1 m in length and 4 cm in width, with the home cage at the end. After habituation on three consecutive days, the mouse was videotaped from behind during its movement on the beam. The foot base angle of the hind limb was measured at the moment when the toe was lifted.

For the grip strength analysis, mice were lifted at the tail base, brought to a trapeze-shaped handle connected with a force sensor (Grip Strength Meter, Ugo Basile). When the mouse spontaneously grabbed the handle, the mouse was gradually pulled away from the handle until it was released.

For the electrophysiological analysis of peripheral nerves, anaesthetized mice (100 mg kg^–1 ketamine and 16 mg kg^–1 xylazine) were placed on a heating pad. One pair of needle electrodes with a tip distance of 5 mm (WE30030.1H10, Science Products) was inserted near the base of the tail and a second pair 30 mm distal to the stimulation site close to the tip of the tail. For the analysis of motor fibres, the stimulus was applied through the proximal electrodes and the response recorded using the distal electrodes. Compound muscle action potentials and sensory nerve action potentials were evoked with increasing intensity (0–15 V, increment of 1 V, 50 µs duration, interstimulus interval of 20 s). Sum action potentials were filtered (high-pass filter 3 Hz, low-pass filter 1.3 kHz) and digitized with a sampling frequency of 10 kHz. Amplitudes were determined from peak to peak.

Histology, neuron count and TEM of mouse tissues

Mice were deeply anaesthetized and transcardially perfused with PBS (pH 7.4) followed by 4% paraformaldehyde (PFA) in PBS for 10 min. After dissection, tissues were post-fixed in 4% PFA in PBS for at least 1 h. Tissues were incubated in sucrose (10% sucrose for 4 h and in 30% sucrose overnight at 4 °C), frozen on dry ice and cut with a sliding microtome (Leica SM 2000R) in 30-µm-thick free-floating sections and stored in PBS supplemented with sodium azide at 4 °C until further use. For NeuN staining, free-floating sections were permeabilized with 0.25% Triton X-100, blocked in 5% normal goat serum (NGS) for 1 h and incubated with mouse anti-NeuN (Millipore) 1:500 at 4 °C overnight. After washing, sections were incubated with the appropriate secondary antibodies (Invitrogen) at 1:1,000 and Hoechst 33342 (Thermo Scientific). Sections were mounted with Fluoromount-G (Southern Biotech). Images were acquired using Cellobserver Z1 (Zeiss) with the tile-scan module and further analysed using ImageJ.

For paraffin embedding, the samples were dehydrated overnight in a series of ethanol and xylol baths (Leica TP20 Tissue Processor), embedded with paraffin (Leica HistoCore Arcadia) and cut into 5 µm sections with a microtome (ThermoScientific Microm HM 355S). For histological analyses, sections were either stained with haematoxylin and eosin (Sigma-Aldrich) or cresyl-violet (Sigma-Aldrich). Images were acquired using a Zeiss AxioLab A1 microscope and further analysed using ImageJ.

For TEM of tissue sections, animals were perfused with 2.5% glutaraldehyde in PBS unless indicated otherwise. For the analysis of DRGs, mice were perfused with 4% PFA and 2% glutaraldehyde in PBS. After dissection, tissues were post-fixed overnight. Tissues were contrasted with 1% osmium tetroxide, dehydrated and infiltrated with epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate, mounted on copper grids and viewed with a Philips CM10 or Zeiss EM 900 digital (DRGs) transmission electron microscope.

TEM of cultured cells

MEFs and human fibroblasts were fixed by adding an equal volume of double strength fixative (4% PFA, 5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4) to the culture medium and incubated for 20 min at room temperature. The fixative mixture was then replaced with one volume of single strength fixative (2% PFA and 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4) for another 2 h at room temperature. After 5 washes with 0.1 M sodium cacodylate buffer (pH 7.4), cells were processed for dehydration and embedding in Epon resin²⁹. To preserve their original morphology, the monolayer culture of fibroblasts of the patient and of his father were embedded in their original position in their culture flasks. By contrast, WT and Arl6ip1 KO MEFs were scraped into 2% low-melting-point agarose before the dehydration process and the embedding in Epon resin, as previously described²⁹. Subsequently, 70 nm ultrathin sections were cut using a Leica EM UC7 ultra microtome (Leica Microsystems) and stained with uranyl acetate and lead citrate²⁹. Cell sections were analysed using an 80 kV transmission electron microscope CM100bio TEM (FEI).

Skeletal muscle fibre bundle staining

Muscles freshly dissected from 2-month-old mice were fixed in 2% PFA for 15 min and subsequently washed with PBS. Fibre bundles were prepared and used for further analyses. After overnight permeabilization with 0.2% Triton X-100 in PBS, samples were blocked with 5% NGS for 1 h followed by an incubation with α-bungarotoxin-Alexa 555 (Invitrogen) 1:500 and mouse anti-NF200 overnight at 4 °C. After washing with PBS, single myofibre bundles were incubated with the corresponding secondary antibodies (Invitrogen) in a dilution of 1:1,000 for 1 h at room temperature. Nuclei were stained with Hoechst 33258 (Invitrogen). Myofibres were washed with PBS and mounted using Fluoromount-G (Southern Biotech). Images were acquired using a Zeiss LSM 880 confocal microscope with Airyscan using the z-stack module. Z-projections with maximum intensities processed using ImageJ are shown.

Cell culture

HEK293T, U2OS and HeLa cells were obtained from the American Type Culture Collection. Their identities were authenticated by STR analysis. U2OS TRex cells were provided by S. Blacklow (Brigham and Women’s Hospital and Harvard Medical School), HeLa TRex were provided by S. Taylor (Manchester University). WT, Fam134b KO and Arl6ip1 KO MEFs were isolated from embryos and immortalized using SV40 large T antigen. All cell lines were regularly tested for mycoplasma contamination using a LookOut Mycoplasma PCR Detection kit (Sigma). Cells were maintained at 37 °C with 5% CO₂ in DMEM medium (Gibco) supplemented with 10% FBS (Gibco) and 100 U ml^–1 penicillin and streptomycin (Gibco).

Inducible cell lines were induced with 1 µg ml^–1 doxycycline (Sigma-Aldrich).

Bafilomycin A1 (LC-Laboratories) was used at a concentration of 200 ng ml^–1, Torin1 (LC-Laboratories) at 250 nM. EBSS medium was obtained from Gibco. For each treatment, cells were plated the day before to perform the experiments when cells had a confluence of 50–60%. For transient expression, DNA plasmids were transfected using GeneJuice (Merck-Millipore), Turbofect (Thermo Scientific) or Lipofectamine 2000 (Invitrogen).

U2OS TRex cell lines were used to generate stable cell lines using the Flp-IN TRex system (Invitrogen) or the lentiviral vector p-Lenti N-HA. For ARL6IP1 knockdown experiments, the respective cells were transfected with either 30 pmol siNT (non-targeting sequence, Qiagen) or with 30 pmol double-stranded ARL6IP1 siRNA (Integrated DNA Technologies; hs.RiARL6IP1.13.2) using Lipofectamine RNAiMAX transfection reagent (Invitrogen, 13778075). For ARL6IP1 and FAM134B double-knockdown experiments, HeLa cells were transfected with ARL6IP1 siRNA and FAM134B siRNA (siRNA RETREG1 18 J-016936-18-0002 and siRNA RETREG1 21 J-016936-21-0002, respectively). Experiments were performed 72 h after transfection.

Protein isolation from cells and tissue lysates

Cells were collected and lysed in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% (v/v) NP-40, 1% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, 1 mM EDTA and complete protease inhibitor (Roche)). Tissue lysates were prepared using a Ultra-Turrax T8 tissue homogenizer (IKA-WERKE) in RIPA buffer. After sonication, homogenates were spun down at 16,900g to remove nuclei and insoluble debris. The supernatant was stored at –80 °C until further use.

Western blotting

Proteins were denatured at 90 °C for 5 min in Laemmli buffer, resolved by SDS–PAGE and transferred to methanol-activated PVDF membranes (Amersham Hybond P 0.45 µm). Membranes were blocked for 1 h in 10% skim milk in TBS-T (Tris-buffered saline with Tween, 20 mM Tris, 150 mM NaCl, 0.1% Tween 20, pH 7.5) and incubated overnight at 4 °C with the specific primary antibody followed by 1 h incubation with the respective secondary antibody at room temperature. Detection was carried out using Clarity Western ECL substrate (Bio-Rad) and a LAS 4000 automated detection system (GE Healthcare). Bands were quantified using ImageJ.

Real-time qPCR

RNA was isolated by TRIzol–chloroform extraction. RNA was reverse-transcribed using a GoScript reverse transcription kit (Promega). qPCR was performed with a final amount of 20 ng of cDNA and EvaGreen Mix (Bio-Rad) with primer pairs for either mouse Gapdh (forward, GCTCATGACCACAGTCCAT; reverse, GTCATCATACTTGGCAGGTTT), mouse Arl6ip1 (forward, GCTCTAATAAATGGACCACTG; reverse, GCACAAATGTCACAATCAGGT), human GAPDH (forward, GAAGGCTGGGGCTCATTT; reverse, GGACTGTGGTCATGAGTC) or human ARL6IP1 (forward, GCTCCAATAAATGGACCACTGA; reverse, GGAAGTCACTATCAGGTAGGT) on a CFX96 Touch Real-Time PCR detection system (Bio-Rad).

Immunofluorescence and autophagic flux analysis

For immunostainings, cells were fixed with 4% PFA at room temperature or ice-cold methanol, washed with PBS and permeabilized with 0.25% Triton X-100 in PBS at room temperature for 10  min, or with 0.1% saponin in PBS at room temperature for 1 min followed by blocking with 5% NGS in PBS for 1  h at room temperature. Incubation with primary antibodies diluted in 5% NGS and 0.25% Triton X-100 in PBS was carried out overnight at 4 °C or at room temperature for 1 h. After three consecutive washes with PBS, secondary antibodies were incubated for 1 h at room temperature. After three washes, cells were incubated at room temperature with Hoechst 33258 (Invitrogen) for 10 min. After a final wash with PBS, the coverslips were mounted with Fluoromount-G solution (ThermoFisher, 00-4958-02). Images were acquired using a confocal microscope (Zeiss LSM 880 with Airyscan).

For the analysis of ER-phagy, cells were transiently transfected with the mCherry–GFP–FAM134B reporter construct. Cells were fixed with ice-cold methanol for 10 min 24 h after transfection. Then cells were permeabilized with 0.25% Triton X-100 in PBS and blocked with 4% NGS for 1 h and stained for LC3B. Images were taken with a LSM 880 and analysed using the ComDet (v.0.5.5) plugin for ImageJ (https://github.com/ekatrukha/ComDet; settings: particle size = 10 pixels, co-localization distance = 7, intensity threshold = 20). The cell border was selected and the cell area determined. Only signals within the cell border were analysed. The intensity threshold was set at 1,000 for all channels, except for human fibroblasts, for which the threshold for mCherry was set at 200. Pearson’s coefficients were calculated using the JaCOP plugin in ImageJ and normalized to the ER area.

To assess whether ARL6IP1 is involved in bulk autophagy, we knocked down ARL6IP1 in U2OS cells stably expressing the mCherry–GFP–LC3 reporter. Autophagy was triggered by 6 h of EBSS exposure or 6 h of 250 nM Torin1 exposure. Images were acquired with a high-content microscope–Yokogawa CQ1 confocal imaging cytometer. To assess whether ARL6IP1 is involved in mitophagy, ARL6IP1, CRISPR–Cas9 KO HeLa cells were transfected with the mitophagy reporter mCherry–GFP–FIS1. Autophagic flux was triggered with 40 μM CCCP for 4 h. Pexophagy was assessed after siRNA-mediated knockdown of ARL6IP1 in U2OS cells after induction of the doxycycline-inducible reporter mCherry–GFP–PMP34 (ref. ³⁰) at baseline or after starvation with EBSS for 20 h. Images were acquired with a confocal microscope (Zeiss LSM 880 with Airyscan). The red and yellow puncta were manually counted.

Rescue experiment

Arl6ip1 WT and KO MEFs were seeded in 24-well-plates at 40,000 cells per well. After 24 h, cells were transfected with the mCherry–GFP–FAM134B plasmid in combination with either the ARL6IP–HA or the ARL6IP1-7KR–HA plasmid. After 48 h, cells were fixed with ice-cold methanol for 10 min, permeabilized with 0.25% Triton X-100 in PBS and blocked with 4% NGS for 1 h and stained for LC3B and HA and further processed as described above. Images were taken with a LSM 880 and analysed using the ComDet (v.0.5.5) plugin for ImageJ (https://github.com/ekatrukha/ComDet; settings: particle size = 10 pixels, co-localization distance = 10, intensity threshold = 200/200/15). Only signals within the cell border were analysed.

ER stress induction and cell viability count

MEFs or human fibroblasts were seeded in 6-well-plates and cultured to 70–80% confluency. Cells were washed with PBS and incubated with new medium with 1.5 µM thapsigargin (Sigma, T9033-5MG) or 5 µg ml^–1 tunicamycin (Santa Cruz) without or in combination with 1 µM MG132 (Calbiochem, 474787-10MG) to inhibit the proteasome. After 24 h for MEFs and 48 h for human fibroblasts, the culture medium was removed, cells washed with PBS and trypsinized. All cells were pooled, centrifuged at 800 r.p.m. for 5 min (Heraeus Sepatech Megafuge 2.0R) and resuspended in fresh medium. Cell viability was measured by trypan blue exclusion with an automatic counting device (Bio-Rad TC20 automatic cell counter).

Interactome analysis and sample preparation for MS

For the LC–MS interactome analysis, HA–FAM134B expression was induced in U2OS cells with doxycycline. HEK293T cells were transiently transfected with the constructs V1-ARL6IP1, V2-ARL6IP1, V1-FAM134B and V2-FAM134B. For in vivo ubiquitination, HEK293T cells were transiently transfected with the plasmids V1-ARL6IP1, V2-ARL6IP1, AMFR-V2 and AMFR-V2-C356G-H361A. After 24 h, cells were lysed with 1% Triton X-100 IP buffer (50 mM Tris-HCL, 150 mM NaCl and 0.5 mM EDTA). Lysis buffer without detergents was added to protein lysates to dilute Triton X-100 to 0.3%. Then samples were incubated with HA-agarose beads (Sigma-Aldrich, A2095) or GFP-Trap beads (Chromotek, gta-20) overnight at 4 °C on a rotating platform. Protein-bound beads were washed three times with lysis buffer supplemented with 0.1% Triton X-100 and once with lysis buffer without detergents. HA IP samples were incubated with 40 µl denaturing buffer (2% sodium deoxycholate, 1 mM tris(2-carboxyethyl)phosphine, 4 mM chloroacetamide and 50 mM Tris-HCl pH 8.5) and heated at 95 °C for 10 min. Samples were mixed 1:1 with 500 ng LysC (Promega), incubated for 3 h at 37 °C and digested with 500 ng of trypsin in 50 mM Tris-HCl, pH 8.5, overnight at 37 °C. GFP IP samples were denatured with 25 µl denaturing buffer at 60 °C for 30 min. After cooling down, the samples were digested with 25 µl 50 mM Tris with 1 µl of trypsin (500 ng) at 37 °C overnight. Reactions were stopped by addition of 150 µl of isopropanol containing 1% trifluoroacetic acid (TFA). Peptides were cleaned up by loading them onto SDB-RPS stage tips (Sigma). After one wash with 1% TFA in isopropanol and one wash with 0.2% TFA in water, peptides were eluted using 80% acetonitrile and 1.25% ammonia. Eluted peptides were dried, tandem mass tagged labelled and processed for LC–MS measurements.

LC–MS analysis

Dried peptides were reconstituted in 2% acetonitrile, 0.1% TFA and analysed on a Q Exactive HF mass spectrometer coupled to an easy nLC 1200 (ThermoFisher Scientific) using a 35-cm-long, 75 µm inner diameter fused-silica column packed in-house with 1.9 µm C18 particles (Reprosil pur, Dr. Maisch) and kept at 50 °C using an integrated column oven (Sonation). Peptides were eluted using a nonlinear gradient from 4 to 28% acetonitrile over 45 min and directly sprayed into the mass spectrometer equipped with a nanoFlex ion source (ThermoFisher Scientific). Full scan MS spectra (300–1,650 m/z) were acquired in profile mode at a resolution of 60,000 at m/z 200, a maximum injection time of 20 ms and an automatic gain control target value of 3 × 10⁶ charges. Up to 15 most intense peptides per full scan were isolated using a 1.4 Th window and fragmented using higher energy collisional dissociation (normalized collision energy of 28). MS/MS spectra were acquired in centroid mode with a resolution of 30,000, a maximum injection time of 45 ms and an automatic gain control target value of 1 × 10⁵. Single charged ions, ions with a charge state above 4 and ions with unassigned charge states were not considered for fragmentation, and dynamic exclusion was set to 20 s to minimize the acquisition of fragment spectra of already acquired precursors.

Proteomics data processing

MS raw data were processed using MaxQuant (v.1.6.10.43) applying default parameters. Acquired spectra were searched against the human ‘one sequence per gene’ database (taxonomy identifier 9606) downloaded from UniProt (accessed 3 March 2020; 20,531 sequences) and a collection of 244 common contaminants (“contaminants.fasta” provided with MaxQuant) using the Andromeda search engine integrated in MaxQuant^31,32. Identifications were filtered to obtain false discovery rates (FDRs) below 1% for both peptide spectrum matches (minimum length of 7 amino acids) and proteins using a target–decoy strategy³³. Protein quantification and data normalization relied on the MaxLFQ algorithm implemented in MaxQuant³⁴. The MS proteomics data have been deposited to the ProteomeXchange Consortium³⁵ through the PRIDE partner repository³⁶ with the dataset identifiers PXD032718, PXD032720 and PXD039184. All acquired raw files were processed using MaxQuant (v.1.6.10.43) and the implemented Andromeda search engine. For protein assignment, spectra were correlated with the UniProt human database (v.2019) including a list of common contaminants. Searches were performed with tryptic specifications and default settings for mass tolerances for MS and MS/MS spectra. Carbamidomethyl at cysteine residues, oxidations at methionine, acetylation at the N terminus were defined as a fixed modification. The minimal peptide length was set to 7 amino acids and the FDR for proteins and peptide-spectrum matches to 1%. The match-between-run feature with a time window of 1 min was used. For further analysis, Perseus software (v.1.6.6.0) was used and first filtered for contaminants and reverse entries as well as proteins that were only identified by a modified peptide.

Proximity ligation assays

Proximity ligation assays (PLAs) were performed using a Duolink in situ red starter kit mouse/rabbit (DUO92101, Sigma-Aldrich) according to the manufacturer’s instructions with rabbit anti-FAM134B (ref. ¹) and rabbit anti-ARL6IP1 (PRS3305, Sigma-Aldrich) antibodies using the Minus (DUO92010) and the Plus probe (DUO92009). Rabbit anti-FAM134B-Plus was used in a 1:5 dilution and rabbit anti-ARL6IP1-Minus in a 1:10 dilution on WT, Arl6ip1 and Fam134b KO MEFs after PFA (4%) fixation and permeabilization with 0.25% (v/v) Triton X-100 in PBS.

Fluorescence protease protection assay

We followed a fluorescence protease protection assay protocol as previously described³⁷. In brief, 75,000 COS-7 cells per well were seeded on 18 mm coverslips coated with 0.1 mg ml^–1 PLL. The next day, cells were transfected with the respective constructs using Lipofectamine 2000. After two more days, cells were washed with pre-warmed intracellular buffer (50 mM HEPES, pH 7.5, 23  mM NaCl, 3 mM MgCl₂, 100 nM CaCl₂, 1 mM EGTA and freshly added 107 mM potassium glutamate, 1 mM ATP and 2  mM dithiothreitol) and transferred to a heated perfusion chamber filled with the same buffer. Live cell imaging for both GFP and RFP was initiated on a Zeiss Cell Observer Z1 with a frame every 20 s starting with a pre-permeabilization image followed by manual administration of 18 µM digitonin. After 120 s, the buffer was replaced by intracellular buffer containing 6 mM freshly added trypsin. Analysis was carried out using ImageJ by drawing the outline of the selected cell and measuring the mean fluorescence intensity of the surrounded area subtracted by the background intensity taken from a cell free spot of the same frame. For further analysis, the area under the curve was calculated between 160 and 720 s.

Purification of recombinant proteins

Trx–His fusion proteins were purified from Escherichia coli treated with 0.5 mM IPTG (overnight, 18 °C). Eluted proteins were concentrated using Amicon Ultra-4-10k centrifugal filter units (Millipore) and then either dialysed at 4 °C against HN buffer (20 mM HEPES/KOH pH 7.4, 150 mM NaCl and 2.5 mM dithiothreitol) (material used for Fig. 4 and Extended Data Fig. 5 analyses) or against liposome buffer (25 mM HEPES-KOH, pH 7.2, 25 mM KCl, 2.5 mM magnesium acetate and 100 mM potassium glutamate) (protein used for studies presented in Fig. 2).

Liposome preparation, liposome incubations, freeze-fracturing and TEM

Liposomes were prepared using Folch-fraction type I lipids (Sigma-Aldrich) according to previously described procedures^10,38. Liposome co-floatation assays were performed as previously reported¹⁰. In brief, liposomes and purified recombinant protein were incubated for 15 min at 37 °C in 0.3 M sucrose in liposome buffer, mixed with 75% sucrose in liposome buffer, overlaid with 200  μl 35% sucrose and 200 μl liposome buffer and then centrifuged at 200,000g for 30 min at 28 °C. Six fractions were collected from top to bottom and analysed by SDS–PAGE and fluorescence-based western blotting using a LICOR Odyssey system (LICOR Bioscience).

For shaping assays presented in Fig. 2, 1 mg of liposomes was incubated with 5 μM protein in liposome buffer containing 0.3 M sucrose for 15 min at 37 °C. For shaping assays presented in Fig. 4 and Extended Data Fig. 5, 1 mg of liposomes was incubated at 37 °C with 2.5 μM protein in HN buffer containing 0.3 M sucrose for 15 min. Thereafter, 20 μg proteinase K was added to avoid liposomal aggregates. The reaction was performed for 40 min at 45 °C (ref. ³⁹). Small aliquots of the liposome suspension were then used for freeze-fracturing. The grids with the samples were systematically explored using an EM 900 electron microscope (Zeiss) operated at 80 kV. Images were acquired with a Wide-angle Dual Speed 2K (Tröndle) CCD camera. The diameters of liposomes were determined using ImageJ.

In vitro ubiquitination assay of ARL6IP1 with recombinant AMFR and sample preparation for MS

For the ubiquitination assay with AMFR, 1 µM purified Trx-His–ARL6IP1, 10 µM ubiquitin (in-house), 10 mM ATP and 10 mM MgCl₂ were incubated with 0.8 µM AMFR (provided by B. Schulman, Max Planck Institute of Biochemistry), 100 nM E1 UBA1 (in-house) and 0.8 µM E2 UBE2G2 (Biotechne) in 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, at 37 °C for 2 h. The reaction mixture was analysed by SDS–PAGE and Coomassie staining or immunoblot analysis for GST (Cell Signaling Technology), His (Cell Signaling Technology) and ubiquitin (Cell Signaling Technology) or by mapping of ubiquitinated lysines by MS.

Ubiquitination assays in cells, co-immunoprecipitations and TUBE2 pull-down

The ubiquitination of ARL6IP1 was assessed in HEK293T cells transfected with Myc–ubiquitin, HA–ARL6IP1 and either AMFR–Flag or its catalytically inactive RING mutant. Cells were lysed in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1% Triton X-100, 10 mM N-ethylmaleimide and protease inhibitors (Roche Diagnostics, 5892791001)). The lysates were then incubated on ice for 15 min and centrifuged at 12,000g at 4 °C for 30 min. Next, 40 µl of the supernatant was collected, mixed with Laemmli sample buffer, boiled for 5 min at 95 °C and stored at –20 °C as input control. Myc-tagged ubiquitinated proteins were immunoprecipitated from lysates cleared with Myc-Trap Agarose (Chromotek, yta-10). Beads were washed three times with lysis buffer, heated at 95 °C for 5 min, subjected to SDS–PAGE and analysed by immunoblotting to detect the HA-Tag. For other immunoprecipitation assays, cleared lysates were incubated with GFP-Trap (Chromotek, gta-20), HA-agarose beads (Sigma-Aldrich, A2095) or TUBE2 agarose beads (Life Sensors, UM402) and incubated at 4 °C overnight. The next day, tubes were centrifuged (800g, 4 °C, 5 min), the supernatants removed and the beads washed with ice-cold lysis buffer. Input and co-precipitated fractions were analysed by SDS–PAGE and immunoblotting. For co-immunoprecipitation of endogenous ARL6IP1 and FAM134B, a confluent 15 cm dish of Arl6ip1 KO or WT MEFs was collected. Lysates were cleared by centrifugation at 12,000g for 10 min and incubated with the ARL6IP1 primary antibody at 4 °C overnight. Protein A agarose beads (Roche, 11719408001) were added and incubated at 4 °C for 4 h. Beads were then washed three times with lysis buffer, re-suspended in Laemmli buffer and boiled. Supernatants were analysed by SDS–PAGE and immunoblotting.

Modelling and simulations of ARL6IP1

The atomic model of human ARL6IP1 was built using the AI-based AlphaFold (v.2) program⁴⁰. Five models were constructed, and the top-ranked model was chosen as it had maximal overlap with predicted secondary structures and consensus transmembrane topology, a higher pLDDT score and a relatively lower predicted alignment error (AF confidence measure).

Coarse-grained (CG) molecular dynamics simulations were performed using the MARTINI model (v.2.2)^41,42. CG models of ubiquitinated and non-ubiquitinated versions were built by using martinize.py⁴³. DSSP assignments were used to generate backbone restraints to preserve local secondary structure^44,45. In the ubiquitinated protein (ARL6IP1-K96-Ub), the iso-peptide bond between K96 and the terminal glycine (G76) of ubiquitin (Protein Data Bank identifier 1UBQ) was modelled by modifying the side chain lysine bead (SC2/+1) into a neutral backbone bead (BB/0) and restraining the distance between the terminal bead of ubiquitin and the lysine side chain to 0.35 nm with a force constant of k = 1,250 kJ (mol nm²)^–1. CG protein models were embedded into POPC (16:0-18:1 PC) bilayers spanning the xy plane of a periodic simulation box (20 × 20 × 20 nm³) solvated with CG-water containing 150 mM NaCl using the insane.py script⁴³. All systems were first energy minimized and then equilibrated using the Berendsen thermostat⁴⁶ and barostat⁴⁷ along with position restraints on protein backbone beads followed by production runs with a 20 fs time step for a total of 10 μs. The system temperature and pressure were maintained at 310 K and 1 atm with the velocity rescaling thermostat⁴⁸ and the semi-isotropic Parrinello–Rahman barostat⁴⁹, respectively. All simulations were performed using gromacs (v.2019.3)^50,51.

Statistical analysis

All experiments were performed in at least three independent biological replicates unless indicated otherwise. Data are presented as the mean ± s.e.m. unless indicated otherwise. Data analyses were performed using GraphPad Prism 9. For statistical analysis, raw data were analysed for normal distribution with the Kolmogorov–Smirnov test or with graphical analysis using Q-Q-plot. If appropriate, we used one-way analysis of variance (with Bonferroni post-hoc test unless indicated otherwise), repeated-measures two-way analysis of variance, Kruskal–Wallis H-test, Student’s t-test (two-sided unless indicated otherwise) or Mann–Whitney U-test. P values less than 0.05 were considered significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.