NINJ1 mediates plasma membrane rupture by cutting and releasing membrane disks.

RESOURCE AVAILABILITY

Lead contact All information and requests for further resources and reagents should be directed to and will be fulfilled by the Lead Contact, Hao Wu, wu@crystal.harvard.edu .

Materials availability

All requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact. All reagents will be made available on request after completion of a Materials Transfer Agreement.

Data and code availability

The cryo-EM map has been deposited in the Electron Microscopy Data Resource (EMDB) under the accession number EMD-42301. The atomic coordinates have been deposited in the Protein Data Bank (PDB) under the accession number 8UIP. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines THP-1 (TIB-202, ATCC) cells were maintained in RPMI with GlutaMAX (Thermo Fisher Scientific, Cat. no: 11875093) supplemented with 10% FBS (Sigma), at 37 °C in 5% CO 2 and a humidified atmosphere. THP-1 NINJ1 KO (Clones 37, 38, 39) cells were generated using the CRISPR/cas9 genome editing procedure. THP-1 NINJ1 KO cells (Clone 39) were reconstituted with codon optimized NINJ1-GFP cloned into pLenti plasmid containing G418 selection.

Animals and cells Wild-type

C57BL/6 animals were purchased from Jackson Laboratories (strain #000664). All animal procedures were conducted under protocols approved by the Animal Care Committee at The Hospital for Sick Children and in accordance with animal care regulation and policies of the Canadian Council on Animal Care. Mice were housed in same-sex polycarbonate cages with ad libitum access to food and water. Housing rooms were temperature and humidity controlled with 14:10 h light:dark cycles. Primary bone marrow derived macrophages (BMDM) were harvested from the femurs of mixed-sex cohorts of wild-type mice. Bones were then washed with PBS under sterile conditions prior to flushing the marrow by cutting the ends and centrifuging them into sterile PBS. Following a wash in phosphate buffered saline (PBS), cells were plated in DMEM with 10 ng mL −1 M-CSF (315–02; Peprotech Inc, Cranbury, NJ). After 5 days of culture, BMDMs were detached from the dishes with TBS and 5 mM EDTA, resuspended in fresh DMEM and plated.

METHOD DETAILS Constructs and cloning

Full-length human NINJ1 and NINJ2 were cloned into the pDB-His-3C-MBP vector with an HRV 3C protease linker for E. coli expression. In addition, full-length human NINJ1 was cloned into pcDNA3-TEV-GFP-FLAG LIC 6D plasmid (Addgene #166835) for mammalian expression in Expi293 cells. NINJ1 and NINJ2 mutants were made both in the pDB-His-3C-MBP and the pcDNA3-TEV-GFP-FLAG LIC 6D plasmids. Protein expression and purification For E. coli expression, NINJ1 and NINJ2 constructs were transformed into BL21 (DE3) cells, grown to an OD600 of 0.6–0.8, cold shocked on ice water for 20 min, and induced overnight with 0.4 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) at 18 °C. Cells were harvested by centrifugation (4000 g , 20 min) and lysed by sonication in lysis buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, SIGMAFAST protease inhibitor). Cell lysate was then centrifuged (40,000 g , 1 h); the membrane fraction was resuspended in lysis buffer supplemented with 1 % DDM: 0.1% CHS (Anatrace) and incubated for 2 h at 4 °C for extracting NINJ1 from the membrane. Following another centrifugation (40,000 g , 1 h), the supernatant was incubated with amylose resin for 1 h at 4 °C. Bound resin was then washed by gravity flow with 50 column volume (CV) lysis buffer with 0.1 % DDM: 0.01% CHS and eluted with 3 CV elution buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, 0.1 % DDM: 0.01% CHS 50 mM maltose). Eluted proteins from amylose resin were concentrated and cleaved overnight with 3C protease at 4 °C. These cleaved protein samples were loaded onto a step-gradient of 25%, 30%, 35%, 40% sucrose in 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail (Sigma, Cat. no: S8830) and 0.002% LMNG (x2 LMNG CMC), and ultracentrifuged for 16 h at 40,000 rpm (MLS-50 swinging-bucket rotor, Beckman). Fractions of 300 μl were collected manually from the heavy fractions of the sucrose gradient, and NINJ1-containing fractions were merged and buffer-exchanged with Zeba ™ spin desalting columns (Fisher Scientific, Cat. no: PI87771) equilibrated with the buffer containing 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail (Sigma, Cat. no: S8830) and 0.002% LMNG. For NINJ1-GFP-FLAG expression in expi293F mammalian cells, one liter cells at 3×10 6 cells/ml were transfected with 1 mg plasmid using polyethylenimine (3 mg/l) as transfection reagent and addition of 5 mM glycine. Cells were harvested 24 h post-transfection. They were lysed by manual homogenization in a buffer containing 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail. The lysate was centrifuged at 42,000 RPM for 1 h (45 Ti fixed-angle rotor, Beckman), and the membrane fraction was solubilized in lysis buffer supplemented with 1 % DDM: 0.1% CHS (Anatrace) was incubated for 2 h at 4 °C, followed by another centrifugation step. The supernatant was incubated with FLAG beads over night at 4 °C. NINJ1-TEV-GFP-FLAG was eluted with 100 μg/ml 3xFLAG peptide (ApexBio, Cat. no: A6001). Eluted fractions were concentrated and incubated with TEV protease at room temperature for 30 min, which was sufficient for GFP-FLAG removal. The cleaved samples were loaded onto a step-gradient of 25%, 30%, 35%, 40% sucrose in 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail (Sigma, Cat. no: S8830) and 0.002% LMNG (x2 LMNG CMC), and ultracentrifuged for 16 h at 40,000 rpm. Untagged NINJ1 was isolated from the heavy fractions of the sucrose gradient. In vitro lipid blot assay Lipid binding assay with purified NINJ1 and NINJ2 was performed using PIP2 strip (Echelon Biosciences, Cat. no: P-6002) according to manufacturer instructions. The PIP2 strip membranes were blocked using 3% bovine serum albumin (BSA) in PNS with 0.1% Tween 20 (PBS-T) for 1 h following by incubation for 1 h with untagged NINJ1 and NINJ2 at 2 μg/ml diluted in 3% BSA in PBS-T. Next, the membranes were washed in PBS-T for 3 times (15 min total) and the bound proteins were visualized with anti-NINJ1 (R&D AF5105) and anti-NINJ2 (Abcam ab172627) antibodies (1:10000 in PBS-T for 1 h). All steps were performed at room temperature. All samples were analyzed at the same time under the same conditions. In vitro liposome assay Purified MBP-NINJ1 and MBP-NINJ2 proteins were mixed with liposomes containing 50% PC, 40% PA and 10% PI(4)P (from brain extract) at 1:3 protein: lipids ratio. The mixture was incubated for 12 h at 4 °C, together with 3C protease in order to allow MBP tag removal and NINJ1 and NINJ2 oligomerization into rings while being incorporated into liposomes. The samples were then subjected to ultracentrifugation at 40,000 rpm for 1 h for pelleting NINJ1 and NINJ2-containing fractions. The pellet fractions were visualized by negative staining EM. For cryo-EM, NINJ1 pellet fraction was further solubilized by 0.002% LMNG and subjected to sucrose gradient ultracentrifugation.

Mutagenesis analysis in vitro and in cells

NINJ1 and NINJ2 single mutants were made using Q5 site-directed Mutagenesis kit for both pDB-His-3C-MBP and the pcDNA3-TEV-GFP-FLAG LIC 6D plasmids. E. coli expressed MBP tagged mutants were purified and tested for oligomerization via negative staining EM and sucrose gradient profiles (for NINJ1). In addition, we performed in vitro liposome assay for the mutants to test the capability of the different types of mutants to rupture liposomes in comparison to WT. For GFP tagged mutants, they were studied in HEK 293T cells by performing cytotoxicity assay and visualizing cellular localization by confocal microscopy.

Negative-staining electron microscopy

For negative staining, 5 μl of NINJ1 and NINJ2 samples in detergents or from liposomes were placed on a copper grid (Electron Microscopy Sciences, cat. no: FCF400CU50), incubated for 1 min, washed twice with buffer containing 25 mM Tris-HCl, pH 7.5 and 150 mM NaCl, stained with 2% uranyl formate for 30 sec and air-dried. The images were collected at a Tecnai G2 Spirit BioTWIN or JEOL Transmission Electron Microscope (TEM) equipped with AMT 2k CCD camera at 49,000x magnification and 120 keV (HMS EM core facility). Cryo-EM data collection A 3 μl drop containing NINJ1 in LMNG (either reconstituted from liposomes as smaller rings, or detergents as mainly large rings) at 0.5 mg/ml was applied to a Lacey Carbon grid with ultrathin carbon support (Ted Pella, Cat. no: 01824G), blotted for 4 s, plunged into liquid ethane, and flash frozen using a FEI Vitrobot Mark IV at 100% humidity and 4 °C. Grid conditions were optimized during screening and small data set collection at the Pacific Northwest Center for Cryo-EM at Oregon Health & Science University (PNCC), the University of Massachusetts Cryo-EM Core (UMASS) and the Harvard Cryo-EM Center for Structural Biology (HMS) using FEI Talos Arctica (ThermoFisher) microscopes equipped with an autoloader (200 keV, Gatan K3 direct electron detector). Final datasets were collected at HMS using a Titan Krios microscope (ThermoFisher) operating at 300 keV and equipped with a BioQuantum Imaging Filter (Gatan) and K3 direct electron detector (Gatan). Automated data collection was performed using SerialEM software, 27 and the movies were obtained in counting mode at 105,000x magnification (0.825 Å/pix). For NINJ1 smaller rings purified from liposomes we collected 11,000 movies with 48 frames each, recorded at multiple defocus values from −1.2 to −2.4 μm and with multiple exposures per stage shift (5×4) introduced with image shift. Each movie was acquired at a dose rate 27.44 e/s per physical pixel and a total accumulated dose of 51.3 e/Å 2 over 1.31 s total exposure time. For NINJ1 sample with large rings purified in detergent, 15,296 movies with 47 frames each were recorded at multiple defocus values from −1 to −2.2 μm and with multiple exposures per stage shift (5×4) introduced with image shift. Each movie was acquired at a dose rate 12.588 e/s per physical pixel and a total accumulated dose of 52.08 e/Å 2 over 2.8 s total exposure time. Both data sets have 0.825 Å pixel size.

Cryo-EM data processing

Cryo-EM data processing software and support was provided by SBGrid consortium. 22 Raw movies were corrected by gain reference and beam-induced motion and combined into a motion-corrected micrograph using the MotionCorr2 algorithm. 28 The defocus value for each micrograph was determined with CTFFIND4. 29 Automated particle picking was performed in CryoSPARC: helical picker 23 for large ring segments, and Topaz training and automated picking for NINJ1 small ring data set. The dataset of NINJ1 smaller rings yielded 597,920 particles, which were extracted with no binning resulting in 0.825 Å pixel size. Several rounds of 2D classifications were applied, showing preferred top and bottom views of NINJ1 in rings of various sizes and shapes. A dataset of NINJ1 larger rings yielded 12,907,558 ring segments. 3 rounds of 2D classification resulted in a particle stack of 626,231 segments. 1 initial model was generated with ab initio reconstruction. Helical refinement mode, followed by the non-uniform 3D refinement using C1 symmetry resulted in final cryo-EM maps at 4.3 Å resolution. The described cryo-EM workflow for this dataset is presented in Figure S2 with gold-standard Fourier shell correlation (FSC). Post-processing of the map was performed with DeepEMhancer. 30 Model building and structure representation The atomic model was built using cryo-EM map obtained from the NINJ1 ring segments purified in the presence LMNG (4.3 Å resolution). Model building was performed using UCSF-Chimera 31 and Coot 32 The final model was further subjected to refinement in Phenix 26 , 33 with the starting model as a reference. Structure representations were generated using ChimeraX 24 and Pymol. 25 Generation of the HA-NINJ1 cells iBMDM NINJ1 knock-out cell line was generated as reported. 15 This cell line was transfected with a mouse HA-NINJ1 plasmid and stable clones isolated by G418 (0.4ug/ml) resistance. Transfection was done using FuGene HD transfection reagent (Promega) according to the manufacturer’s instructions. Multiple clones were isolated and validated. Cells are negative for mycoplasma.

Generation of NINJ1 KO and NINJ1-GFP cell lines

To generate NINJ1 KO and sgRNA control THP-1 cells, we transduced THP-1s (early passage, ATCC) with lentiCRISPR v2 (NINJ1 Genescript guide #6 GGCACATAGAAGGCGAAGCT, addgene #125836). For lentivirus preparation, 0.5×10 6 HEK293T cells in 6-well dishes were transfected (6 μL Fugene) with 1 μg of plasmid containing the construct of interest, 750 ng psPAX2 packaging plasmid and 250 ng pMD2.G envelope plasmid (Addgene plasmids #12260 and #12259). On the following day the medium was changed, and the virus-containing medium was collected an additional 24 h later. Virus-containing medium was concentrated with LentiX concentrator (Takara) and resuspended in RPMI (20X concentration). 1 million THP-1 cells seeded in 6-well plates were spinfected with 100 μL concentrated virus in 900 μL medium (2 h, 30 degrees, 1000 × g) containing 8 ug/mL polybrene. Following spinfection, THP-1 cells were recovered in fresh medium for 2 d prior to puromycin selection (3 μg/mL) for 3 days. Selected cells were single cell sorted into 96 well format with a BD FACSAria II cell sorter equipped with 100 μm nozzle. Cells were expanded before screening clones by western blot and subsequent genomic sequencing. We obtained 3 clones (37,38,39) in which NINJ1 was undetectable by western blot. Among the 3 clones, we used clone 39 for subsequent experiments given it had biallelic frameshift mutations (1 and 11 bp). For reconstituting NINJ1 KO cells with Doxycycline-inducible NINJ1-GFP (pInducer20), we produced virus and spinfected NINJ1 −/− THP-1 cells as above. Cells were selected with 250 μg/ml G418 for 2 weeks prior to use in further experiments. Pyroptosis and secondary necrosis induction In primary BMDMs and THP-1 cells pyroptosis was activated in cells primed with (0.5 μg/ml) LPS from E. coli serotype 055:B5 (Invivogen, Cat. no: tlrl-b5lps), which was first reconstituted at a stock concentration of 1 mg/ml. Cells were primed with LPS for a total of 4 hours, followed by induction with 10 μM or 20 μM nigericin (Sigma N7143; stock 10 mM in ethanol) for the final 15 min, 30 min or 60 min. Secondary necrosis was induced using ABT-199 in primary BMDMs (Tocris, 6960).

Immunoblotting of whole cell lysates

NINJ1-GFP cells were seeded at 0.7×10 6 cells/well on a 6-well tissue culture plate and treated with 25 ng/ml phorbol 12-myristate 13-acetate (PMA, P1585, Sigma) for 2 days and then recovered with complete medium (without PMA) for 24 h and supplemented with 1 μg/ml Doxycycline (Dox) or no Dox as a control. On the next day, cells were untreated or primed with 1 μg/ml LPS (Invivogen, Cat. no: tlrl-b5lps) for 4 h, followed by 1 h incubation with 10 μM nigericin (Sigma-Aldrich, Cat. no: N7143–5MG) for cell death stimulation. While collecting cells, we collected both pellet and supernatant fractions. The pellet fractions were collected by scraping cells with cell lifter with PBS and spinning at × 500 g for 5 min. The supernatant fractions were first spun down at 500 g for 5 min and then concentrated using 10K cutoff Centricon. SDS sample buffer was added to equal number of cells (by volume) and loaded to 4–20% non-reducing gel for western blot analysis. Moreover, in an additional analysis the concentrated supernatant fraction was ultracentrifuged for 1 hr at 136,000x g following western blot analysis for the pellet fraction Western blotting was performed using anti-NINJ1 (1:1000, R&D AF5105) and anti-sheep-HRP (1:5000, Cell Signaling, Cat. no: 7076S) secondary antibody. Blue-native PAGE of NINJ1 Primary mouse BMDMs were plated onto 10 cm dishes and treated as indicated in the figure legend. Following treatments, the supernatant (medium) was collected from each dish for centrifugation at 500 × g for 5 minutes, followed by 3000 × g for 10 min. Next, supernatant samples were concentrated using Centricon Plus-70 centrifugal filters (Millipore, UFC701008), after which the protein concentration of samples was determined using the BioRad DC protein concentration assay kit (5000112) as per the manufacturer’s instructions. 4 μg of sample was mixed with 4x NativePAGE sample buffer and Coomassie G-250 (ThermoFisher) and resolved using NativePAGE 3–12% Bis-Tris gels, transferred onto polyvinylidene difluoride (PDVF) membranes, and immunoblotted using rabbit monoclonal anti-mouse NINJ1 primary antibody (clone 25; kind gift from Dr. Kayagaki and Dr. Dixit at Genentech, Inc.) 3 . Immunofluorescence (IF) NINJ1-GFP THP-1 cells were plated on CELLview 4-compartment dishes (Greiner Bio-One), treated with 25 ng/ml phorbol 12-myristate 13-acetate (PMA, P1585, Sigma) for 2 days, and then incubated for 24 h with complete medium without PMA and supplemented with 1 μg/ml Doxycycline (Dox) or without Dox as a control. Cells were untreated or primed with 1 μg/ml LPS (Invivogen, Cat. no: tlrl-b5lps) for 4 h, followed by 1 h incubation with 10 μM nigericin (Sigma-Aldrich, Cat. no: N7143–5MG). Cells were fixed in 3% paraformaldehyde (PFA) for 30 min at 4 °C and permeabilized with 0.1% Triton X-100 for 10 min. Cells were incubated in PBS-Tween containing 3% BSA for 3 h, which minimized non-specific binding. After three washes with PBS-Tween, cells were incubated overnight at 4 °C with primary antibodies (human ASC 1:1000, Novus Biologicals, NBP1–78977 and human ASC-Phycoerythrin (PE) 1:1000, BioLegend 653903). NINJ1 was detected by the GFP signal, and no antibody was required in order to enhance the signal. After incubation, cells were washed and incubated with AlexaFluor647-conjugated anti-rabbit IgG (1:750, ThermoFisher, Cat. no: A27040 ) for 1 h at room temperature, washed with PBS (3 × 10 min) and then stained with Hoechst (1:500, Immunochemistry Technologies, Cat. no: 639). Cells were imaged using a Nikon Ti inverted microscope fitted with a Photometrics CoolSNAP HQ2 Peltier cooled CCD camera and Andor Zyla 4.2 sCMOS camera at x60 magnification using Plan Apo 60x/1.3 DIC objectives. Lumencor SpectraX LED illuminator was used. Chroma 49000 (DAPI), Chroma 29002 (green) and Chroma 49011 (far red) filter cubes were used. Image analysis was performed in Fiji 34 Live cell imaging NINJ1-GFP THP-1 cells were cultured in RPMI supplemented with GlutaMAX (Thermo Fisher Scientific, Cat. no: 11875093) supplemented with 10% FBS (Sigma) and neomycin (G418), at 37°C in 5% CO2 atmosphere. Cells treatment was carried out as for IF (above). For performing live cell imaging, we imaged LPS primed cells, induced with Dox for 24 h. Cells were imaged using a Nikon Ti inverted microscope fitted with a Photometrics CoolSNAP HQ2 Peltier cooled CCD camera and Andor Zyla 4.2 sCMOS camera at x60 magnification using Plan Apo 60x/1.3 DIC objective. Green and bright field channels were used. Time lapsed imaging was started upon addition of 10 μM nigericin. Images were recorded for over an hour with 2 min intervals. Image analysis was performed in Fiji 34 Lipid labeling of whole cell lysates Concentrated NINJ1-GFP THP-1 activated cells were labeled with BODIPY ™ 665/676 (Invitrogen B3932) at a final dye concentration of 10 μg/ml followed by 10%−40% sucrose ultracentrifugation. The pellet fraction was analyzed by spinning disk confocal imaging and negative staining EM. Confocal image analysis and plot profiles were performed in Fiji 34 LDH assay HEK 293T cells were seeded at 100,000 per well in 96-well plates, and NINJ1 constructs were transfected with lipofectamine 2000 with 0.25 μg DNA per well. 24 h following transfection, cell supernatants were analyzed for LDH activity using the CytoTox 96 ® Non-Radioactive Cytotoxicity Assay Kit (Promega, Cat. No: G1780). The samples’ absorbance at 490 nm was measured by a bioTEK plate reader. Minimal photon fluxes (MINFLUX) nanoscopy Primary cells from wild-type animals were harvested as described above and cultured on 1.5H glass coverslips. Cells were primed with LPS and stimulated to undergo pyroptosis as described above. Pyroptosis induction was confirmed by LDH release for each experimental replicate. Following treatments, cells were washed with PBS and fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M cacodylate buffer at room temperature for 15 min. The cells were washed 3 times in PBS, permeabilized with 0.1% Tween-20, then blocked in 10% donkey serum and 0.1% Tween-20 in PBS for 1 h in PBS. Rabbit monoclonal anti-mouse NINJ1 primary antibody (clone 25 3 at 10 μg/mL was then added overnight at 4 °C. The cells were washed three times with PBS supplemented with 0.1% Tween-20 before the addition of secondary antibody in PBS with 1% donkey serum for 1 h at room temperature. Alexa Fluor-647-conjugated donkey anti-rabbit secondary antibody (711-605-152) was used at a 1:1000 dilution. The samples were then washed with PBS and stabilized for MINFLUX imaging using 150 mm gold nanoparticles (EM.GC150, BBI Solutions) for 5–10 min, then mounted in imaging buffer, 50 mM Tris-HCl, 10 mM NaCl, 10% (w/v) glucose, and 64 μg/mL catalase, 0.4 mg/mL glucose oxidase, and 10 to 25 mM mercaptoethylamine, pH 8.0, as previously described, 35 , 36 A MINFLUX microscope (Abberior Instruments) equipped with a spatial light modulator-based beam shaping module and an electro-optical detector-based MINFLUX was used. Microscopy and MINFLUX measurements were performed using Imspector Software (v.16.3.15635-m2205-win64-MINFLUX; RRID:SCR_015249) using MINFLUX sequence templates seqIIF and DefaultIIF3D for 2D- and 3D-MINFLUX measurements as previously described. 37 Molecular precision was enabled through a reflection-based stabilization unit based on a 980-nm as previously described. 16 Before starting each MINFLUX measurement, the stabilization of the microscope was activated. Measurements were conducted with a stabilization precision of < 1 nm. A 642 nm (CW) excitation laser, a 405 nm activation laser, and detection windows in the range of 650–750 nm were used. The emitted fluorescence photons were counted using two avalanche photodiodes with the appropriate fluorescence filters. Molecular precision was enabled through a reflection-based stabilization unit based on a 980-nm laser as previously described. 16 For structure counting and ring-like structure diameter measurements, localizations were exported as 2D projections with a pixel size of 4 nm (2D) or 5–6 nm (3D) using Imspector Software. Individual structures were counted in ImageJ and classified according to shape: ring-like, punctate, or other. The diameter of each ring-like structure was considered as the largest measurable distance between points with zero localizations in the structure center and measured using the line tool in ImageJ-Fiji Software.

Mouse NINJ1 co-localization and imaging

For immunofluorescence labelling, iBMDMs stably expressing HA-NINJ1 cells were seeded onto glass coverslips. The next day the cells were treated or not with LPS (0.5 μg/mL) for a total of 3.5 h, during which nigericin (10μM) was added for the final 1 h. Cells were then washed with PBS and permeabilized using 0.1% Triton X-100 in PBS for 5 min and blocked with 1% BSA in PBS for 1 h. Cells were incubated with primary antibodies for 1 h, washed, then incubated with Alexa Fluor-488 or Cy3-conjugated secondary antibodies for 1 h. Prior to imaging, cells were stained with DAPI for 5 min. Antibodies used: anti-HA-tag (Cell Signaling, 3724), anti-KDEL (Enzo, ADI-SPA-827), anti-GM130 (BD Biosciences, 610823), anti-HSP60 (BD Biosciences, 611563). Cells were imaged as described in 38 by spinning disk confocal microscopy (Quorum) on a Zeiss Axiovert 200M microscope with a ×63 objective and an additional ×1.5 magnifying lens. Images were acquired by a CCD camera (Hamamatsu Photonics) driven by Volocity software.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical testing was calculated using Prism 9.0 (GraphPad Software Inc, La Jolla, CA; RRID:SCR_002798). Groups were compared using Student t test for two groups. Unless specified otherwise, all collected data was analyzed and a P value < 0.05 was considered statistically significant.

Materials availability

All requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact. All reagents will be made available on request after completion of a Materials Transfer Agreement.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines THP-1 (TIB-202, ATCC) cells were maintained in RPMI with GlutaMAX (Thermo Fisher Scientific, Cat. no: 11875093) supplemented with 10% FBS (Sigma), at 37 °C in 5% CO 2 and a humidified atmosphere. THP-1 NINJ1 KO (Clones 37, 38, 39) cells were generated using the CRISPR/cas9 genome editing procedure. THP-1 NINJ1 KO cells (Clone 39) were reconstituted with codon optimized NINJ1-GFP cloned into pLenti plasmid containing G418 selection.

Animals and cells Wild-type

C57BL/6 animals were purchased from Jackson Laboratories (strain #000664). All animal procedures were conducted under protocols approved by the Animal Care Committee at The Hospital for Sick Children and in accordance with animal care regulation and policies of the Canadian Council on Animal Care. Mice were housed in same-sex polycarbonate cages with ad libitum access to food and water. Housing rooms were temperature and humidity controlled with 14:10 h light:dark cycles. Primary bone marrow derived macrophages (BMDM) were harvested from the femurs of mixed-sex cohorts of wild-type mice. Bones were then washed with PBS under sterile conditions prior to flushing the marrow by cutting the ends and centrifuging them into sterile PBS. Following a wash in phosphate buffered saline (PBS), cells were plated in DMEM with 10 ng mL −1 M-CSF (315–02; Peprotech Inc, Cranbury, NJ). After 5 days of culture, BMDMs were detached from the dishes with TBS and 5 mM EDTA, resuspended in fresh DMEM and plated.

METHOD DETAILS Constructs and cloning

Full-length human NINJ1 and NINJ2 were cloned into the pDB-His-3C-MBP vector with an HRV 3C protease linker for E. coli expression. In addition, full-length human NINJ1 was cloned into pcDNA3-TEV-GFP-FLAG LIC 6D plasmid (Addgene #166835) for mammalian expression in Expi293 cells. NINJ1 and NINJ2 mutants were made both in the pDB-His-3C-MBP and the pcDNA3-TEV-GFP-FLAG LIC 6D plasmids. Protein expression and purification For E. coli expression, NINJ1 and NINJ2 constructs were transformed into BL21 (DE3) cells, grown to an OD600 of 0.6–0.8, cold shocked on ice water for 20 min, and induced overnight with 0.4 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) at 18 °C. Cells were harvested by centrifugation (4000 g , 20 min) and lysed by sonication in lysis buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, SIGMAFAST protease inhibitor). Cell lysate was then centrifuged (40,000 g , 1 h); the membrane fraction was resuspended in lysis buffer supplemented with 1 % DDM: 0.1% CHS (Anatrace) and incubated for 2 h at 4 °C for extracting NINJ1 from the membrane. Following another centrifugation (40,000 g , 1 h), the supernatant was incubated with amylose resin for 1 h at 4 °C. Bound resin was then washed by gravity flow with 50 column volume (CV) lysis buffer with 0.1 % DDM: 0.01% CHS and eluted with 3 CV elution buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, 0.1 % DDM: 0.01% CHS 50 mM maltose). Eluted proteins from amylose resin were concentrated and cleaved overnight with 3C protease at 4 °C. These cleaved protein samples were loaded onto a step-gradient of 25%, 30%, 35%, 40% sucrose in 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail (Sigma, Cat. no: S8830) and 0.002% LMNG (x2 LMNG CMC), and ultracentrifuged for 16 h at 40,000 rpm (MLS-50 swinging-bucket rotor, Beckman). Fractions of 300 μl were collected manually from the heavy fractions of the sucrose gradient, and NINJ1-containing fractions were merged and buffer-exchanged with Zeba ™ spin desalting columns (Fisher Scientific, Cat. no: PI87771) equilibrated with the buffer containing 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail (Sigma, Cat. no: S8830) and 0.002% LMNG. For NINJ1-GFP-FLAG expression in expi293F mammalian cells, one liter cells at 3×10 6 cells/ml were transfected with 1 mg plasmid using polyethylenimine (3 mg/l) as transfection reagent and addition of 5 mM glycine. Cells were harvested 24 h post-transfection. They were lysed by manual homogenization in a buffer containing 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail. The lysate was centrifuged at 42,000 RPM for 1 h (45 Ti fixed-angle rotor, Beckman), and the membrane fraction was solubilized in lysis buffer supplemented with 1 % DDM: 0.1% CHS (Anatrace) was incubated for 2 h at 4 °C, followed by another centrifugation step. The supernatant was incubated with FLAG beads over night at 4 °C. NINJ1-TEV-GFP-FLAG was eluted with 100 μg/ml 3xFLAG peptide (ApexBio, Cat. no: A6001). Eluted fractions were concentrated and incubated with TEV protease at room temperature for 30 min, which was sufficient for GFP-FLAG removal. The cleaved samples were loaded onto a step-gradient of 25%, 30%, 35%, 40% sucrose in 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM TCEP, supplemented with protease inhibitor cocktail (Sigma, Cat. no: S8830) and 0.002% LMNG (x2 LMNG CMC), and ultracentrifuged for 16 h at 40,000 rpm. Untagged NINJ1 was isolated from the heavy fractions of the sucrose gradient. In vitro lipid blot assay Lipid binding assay with purified NINJ1 and NINJ2 was performed using PIP2 strip (Echelon Biosciences, Cat. no: P-6002) according to manufacturer instructions. The PIP2 strip membranes were blocked using 3% bovine serum albumin (BSA) in PNS with 0.1% Tween 20 (PBS-T) for 1 h following by incubation for 1 h with untagged NINJ1 and NINJ2 at 2 μg/ml diluted in 3% BSA in PBS-T. Next, the membranes were washed in PBS-T for 3 times (15 min total) and the bound proteins were visualized with anti-NINJ1 (R&D AF5105) and anti-NINJ2 (Abcam ab172627) antibodies (1:10000 in PBS-T for 1 h). All steps were performed at room temperature. All samples were analyzed at the same time under the same conditions. In vitro liposome assay Purified MBP-NINJ1 and MBP-NINJ2 proteins were mixed with liposomes containing 50% PC, 40% PA and 10% PI(4)P (from brain extract) at 1:3 protein: lipids ratio. The mixture was incubated for 12 h at 4 °C, together with 3C protease in order to allow MBP tag removal and NINJ1 and NINJ2 oligomerization into rings while being incorporated into liposomes. The samples were then subjected to ultracentrifugation at 40,000 rpm for 1 h for pelleting NINJ1 and NINJ2-containing fractions. The pellet fractions were visualized by negative staining EM. For cryo-EM, NINJ1 pellet fraction was further solubilized by 0.002% LMNG and subjected to sucrose gradient ultracentrifugation.

Mutagenesis analysis in vitro and in cells

NINJ1 and NINJ2 single mutants were made using Q5 site-directed Mutagenesis kit for both pDB-His-3C-MBP and the pcDNA3-TEV-GFP-FLAG LIC 6D plasmids. E. coli expressed MBP tagged mutants were purified and tested for oligomerization via negative staining EM and sucrose gradient profiles (for NINJ1). In addition, we performed in vitro liposome assay for the mutants to test the capability of the different types of mutants to rupture liposomes in comparison to WT. For GFP tagged mutants, they were studied in HEK 293T cells by performing cytotoxicity assay and visualizing cellular localization by confocal microscopy.

Negative-staining electron microscopy

For negative staining, 5 μl of NINJ1 and NINJ2 samples in detergents or from liposomes were placed on a copper grid (Electron Microscopy Sciences, cat. no: FCF400CU50), incubated for 1 min, washed twice with buffer containing 25 mM Tris-HCl, pH 7.5 and 150 mM NaCl, stained with 2% uranyl formate for 30 sec and air-dried. The images were collected at a Tecnai G2 Spirit BioTWIN or JEOL Transmission Electron Microscope (TEM) equipped with AMT 2k CCD camera at 49,000x magnification and 120 keV (HMS EM core facility). Cryo-EM data collection A 3 μl drop containing NINJ1 in LMNG (either reconstituted from liposomes as smaller rings, or detergents as mainly large rings) at 0.5 mg/ml was applied to a Lacey Carbon grid with ultrathin carbon support (Ted Pella, Cat. no: 01824G), blotted for 4 s, plunged into liquid ethane, and flash frozen using a FEI Vitrobot Mark IV at 100% humidity and 4 °C. Grid conditions were optimized during screening and small data set collection at the Pacific Northwest Center for Cryo-EM at Oregon Health & Science University (PNCC), the University of Massachusetts Cryo-EM Core (UMASS) and the Harvard Cryo-EM Center for Structural Biology (HMS) using FEI Talos Arctica (ThermoFisher) microscopes equipped with an autoloader (200 keV, Gatan K3 direct electron detector). Final datasets were collected at HMS using a Titan Krios microscope (ThermoFisher) operating at 300 keV and equipped with a BioQuantum Imaging Filter (Gatan) and K3 direct electron detector (Gatan). Automated data collection was performed using SerialEM software, 27 and the movies were obtained in counting mode at 105,000x magnification (0.825 Å/pix). For NINJ1 smaller rings purified from liposomes we collected 11,000 movies with 48 frames each, recorded at multiple defocus values from −1.2 to −2.4 μm and with multiple exposures per stage shift (5×4) introduced with image shift. Each movie was acquired at a dose rate 27.44 e/s per physical pixel and a total accumulated dose of 51.3 e/Å 2 over 1.31 s total exposure time. For NINJ1 sample with large rings purified in detergent, 15,296 movies with 47 frames each were recorded at multiple defocus values from −1 to −2.2 μm and with multiple exposures per stage shift (5×4) introduced with image shift. Each movie was acquired at a dose rate 12.588 e/s per physical pixel and a total accumulated dose of 52.08 e/Å 2 over 2.8 s total exposure time. Both data sets have 0.825 Å pixel size.

Cryo-EM data processing

Cryo-EM data processing software and support was provided by SBGrid consortium. 22 Raw movies were corrected by gain reference and beam-induced motion and combined into a motion-corrected micrograph using the MotionCorr2 algorithm. 28 The defocus value for each micrograph was determined with CTFFIND4. 29 Automated particle picking was performed in CryoSPARC: helical picker 23 for large ring segments, and Topaz training and automated picking for NINJ1 small ring data set. The dataset of NINJ1 smaller rings yielded 597,920 particles, which were extracted with no binning resulting in 0.825 Å pixel size. Several rounds of 2D classifications were applied, showing preferred top and bottom views of NINJ1 in rings of various sizes and shapes. A dataset of NINJ1 larger rings yielded 12,907,558 ring segments. 3 rounds of 2D classification resulted in a particle stack of 626,231 segments. 1 initial model was generated with ab initio reconstruction. Helical refinement mode, followed by the non-uniform 3D refinement using C1 symmetry resulted in final cryo-EM maps at 4.3 Å resolution. The described cryo-EM workflow for this dataset is presented in Figure S2 with gold-standard Fourier shell correlation (FSC). Post-processing of the map was performed with DeepEMhancer. 30 Model building and structure representation The atomic model was built using cryo-EM map obtained from the NINJ1 ring segments purified in the presence LMNG (4.3 Å resolution). Model building was performed using UCSF-Chimera 31 and Coot 32 The final model was further subjected to refinement in Phenix 26 , 33 with the starting model as a reference. Structure representations were generated using ChimeraX 24 and Pymol. 25 Generation of the HA-NINJ1 cells iBMDM NINJ1 knock-out cell line was generated as reported. 15 This cell line was transfected with a mouse HA-NINJ1 plasmid and stable clones isolated by G418 (0.4ug/ml) resistance. Transfection was done using FuGene HD transfection reagent (Promega) according to the manufacturer’s instructions. Multiple clones were isolated and validated. Cells are negative for mycoplasma.

Generation of NINJ1 KO and NINJ1-GFP cell lines

To generate NINJ1 KO and sgRNA control THP-1 cells, we transduced THP-1s (early passage, ATCC) with lentiCRISPR v2 (NINJ1 Genescript guide #6 GGCACATAGAAGGCGAAGCT, addgene #125836). For lentivirus preparation, 0.5×10 6 HEK293T cells in 6-well dishes were transfected (6 μL Fugene) with 1 μg of plasmid containing the construct of interest, 750 ng psPAX2 packaging plasmid and 250 ng pMD2.G envelope plasmid (Addgene plasmids #12260 and #12259). On the following day the medium was changed, and the virus-containing medium was collected an additional 24 h later. Virus-containing medium was concentrated with LentiX concentrator (Takara) and resuspended in RPMI (20X concentration). 1 million THP-1 cells seeded in 6-well plates were spinfected with 100 μL concentrated virus in 900 μL medium (2 h, 30 degrees, 1000 × g) containing 8 ug/mL polybrene. Following spinfection, THP-1 cells were recovered in fresh medium for 2 d prior to puromycin selection (3 μg/mL) for 3 days. Selected cells were single cell sorted into 96 well format with a BD FACSAria II cell sorter equipped with 100 μm nozzle. Cells were expanded before screening clones by western blot and subsequent genomic sequencing. We obtained 3 clones (37,38,39) in which NINJ1 was undetectable by western blot. Among the 3 clones, we used clone 39 for subsequent experiments given it had biallelic frameshift mutations (1 and 11 bp). For reconstituting NINJ1 KO cells with Doxycycline-inducible NINJ1-GFP (pInducer20), we produced virus and spinfected NINJ1 −/− THP-1 cells as above. Cells were selected with 250 μg/ml G418 for 2 weeks prior to use in further experiments. Pyroptosis and secondary necrosis induction In primary BMDMs and THP-1 cells pyroptosis was activated in cells primed with (0.5 μg/ml) LPS from E. coli serotype 055:B5 (Invivogen, Cat. no: tlrl-b5lps), which was first reconstituted at a stock concentration of 1 mg/ml. Cells were primed with LPS for a total of 4 hours, followed by induction with 10 μM or 20 μM nigericin (Sigma N7143; stock 10 mM in ethanol) for the final 15 min, 30 min or 60 min. Secondary necrosis was induced using ABT-199 in primary BMDMs (Tocris, 6960).

Immunoblotting of whole cell lysates

NINJ1-GFP cells were seeded at 0.7×10 6 cells/well on a 6-well tissue culture plate and treated with 25 ng/ml phorbol 12-myristate 13-acetate (PMA, P1585, Sigma) for 2 days and then recovered with complete medium (without PMA) for 24 h and supplemented with 1 μg/ml Doxycycline (Dox) or no Dox as a control. On the next day, cells were untreated or primed with 1 μg/ml LPS (Invivogen, Cat. no: tlrl-b5lps) for 4 h, followed by 1 h incubation with 10 μM nigericin (Sigma-Aldrich, Cat. no: N7143–5MG) for cell death stimulation. While collecting cells, we collected both pellet and supernatant fractions. The pellet fractions were collected by scraping cells with cell lifter with PBS and spinning at × 500 g for 5 min. The supernatant fractions were first spun down at 500 g for 5 min and then concentrated using 10K cutoff Centricon. SDS sample buffer was added to equal number of cells (by volume) and loaded to 4–20% non-reducing gel for western blot analysis. Moreover, in an additional analysis the concentrated supernatant fraction was ultracentrifuged for 1 hr at 136,000x g following western blot analysis for the pellet fraction Western blotting was performed using anti-NINJ1 (1:1000, R&D AF5105) and anti-sheep-HRP (1:5000, Cell Signaling, Cat. no: 7076S) secondary antibody. Blue-native PAGE of NINJ1 Primary mouse BMDMs were plated onto 10 cm dishes and treated as indicated in the figure legend. Following treatments, the supernatant (medium) was collected from each dish for centrifugation at 500 × g for 5 minutes, followed by 3000 × g for 10 min. Next, supernatant samples were concentrated using Centricon Plus-70 centrifugal filters (Millipore, UFC701008), after which the protein concentration of samples was determined using the BioRad DC protein concentration assay kit (5000112) as per the manufacturer’s instructions. 4 μg of sample was mixed with 4x NativePAGE sample buffer and Coomassie G-250 (ThermoFisher) and resolved using NativePAGE 3–12% Bis-Tris gels, transferred onto polyvinylidene difluoride (PDVF) membranes, and immunoblotted using rabbit monoclonal anti-mouse NINJ1 primary antibody (clone 25; kind gift from Dr. Kayagaki and Dr. Dixit at Genentech, Inc.) 3 . Immunofluorescence (IF) NINJ1-GFP THP-1 cells were plated on CELLview 4-compartment dishes (Greiner Bio-One), treated with 25 ng/ml phorbol 12-myristate 13-acetate (PMA, P1585, Sigma) for 2 days, and then incubated for 24 h with complete medium without PMA and supplemented with 1 μg/ml Doxycycline (Dox) or without Dox as a control. Cells were untreated or primed with 1 μg/ml LPS (Invivogen, Cat. no: tlrl-b5lps) for 4 h, followed by 1 h incubation with 10 μM nigericin (Sigma-Aldrich, Cat. no: N7143–5MG). Cells were fixed in 3% paraformaldehyde (PFA) for 30 min at 4 °C and permeabilized with 0.1% Triton X-100 for 10 min. Cells were incubated in PBS-Tween containing 3% BSA for 3 h, which minimized non-specific binding. After three washes with PBS-Tween, cells were incubated overnight at 4 °C with primary antibodies (human ASC 1:1000, Novus Biologicals, NBP1–78977 and human ASC-Phycoerythrin (PE) 1:1000, BioLegend 653903). NINJ1 was detected by the GFP signal, and no antibody was required in order to enhance the signal. After incubation, cells were washed and incubated with AlexaFluor647-conjugated anti-rabbit IgG (1:750, ThermoFisher, Cat. no: A27040 ) for 1 h at room temperature, washed with PBS (3 × 10 min) and then stained with Hoechst (1:500, Immunochemistry Technologies, Cat. no: 639). Cells were imaged using a Nikon Ti inverted microscope fitted with a Photometrics CoolSNAP HQ2 Peltier cooled CCD camera and Andor Zyla 4.2 sCMOS camera at x60 magnification using Plan Apo 60x/1.3 DIC objectives. Lumencor SpectraX LED illuminator was used. Chroma 49000 (DAPI), Chroma 29002 (green) and Chroma 49011 (far red) filter cubes were used. Image analysis was performed in Fiji 34 Live cell imaging NINJ1-GFP THP-1 cells were cultured in RPMI supplemented with GlutaMAX (Thermo Fisher Scientific, Cat. no: 11875093) supplemented with 10% FBS (Sigma) and neomycin (G418), at 37°C in 5% CO2 atmosphere. Cells treatment was carried out as for IF (above). For performing live cell imaging, we imaged LPS primed cells, induced with Dox for 24 h. Cells were imaged using a Nikon Ti inverted microscope fitted with a Photometrics CoolSNAP HQ2 Peltier cooled CCD camera and Andor Zyla 4.2 sCMOS camera at x60 magnification using Plan Apo 60x/1.3 DIC objective. Green and bright field channels were used. Time lapsed imaging was started upon addition of 10 μM nigericin. Images were recorded for over an hour with 2 min intervals. Image analysis was performed in Fiji 34 Lipid labeling of whole cell lysates Concentrated NINJ1-GFP THP-1 activated cells were labeled with BODIPY ™ 665/676 (Invitrogen B3932) at a final dye concentration of 10 μg/ml followed by 10%−40% sucrose ultracentrifugation. The pellet fraction was analyzed by spinning disk confocal imaging and negative staining EM. Confocal image analysis and plot profiles were performed in Fiji 34 LDH assay HEK 293T cells were seeded at 100,000 per well in 96-well plates, and NINJ1 constructs were transfected with lipofectamine 2000 with 0.25 μg DNA per well. 24 h following transfection, cell supernatants were analyzed for LDH activity using the CytoTox 96 ® Non-Radioactive Cytotoxicity Assay Kit (Promega, Cat. No: G1780). The samples’ absorbance at 490 nm was measured by a bioTEK plate reader. Minimal photon fluxes (MINFLUX) nanoscopy Primary cells from wild-type animals were harvested as described above and cultured on 1.5H glass coverslips. Cells were primed with LPS and stimulated to undergo pyroptosis as described above. Pyroptosis induction was confirmed by LDH release for each experimental replicate. Following treatments, cells were washed with PBS and fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M cacodylate buffer at room temperature for 15 min. The cells were washed 3 times in PBS, permeabilized with 0.1% Tween-20, then blocked in 10% donkey serum and 0.1% Tween-20 in PBS for 1 h in PBS. Rabbit monoclonal anti-mouse NINJ1 primary antibody (clone 25 3 at 10 μg/mL was then added overnight at 4 °C. The cells were washed three times with PBS supplemented with 0.1% Tween-20 before the addition of secondary antibody in PBS with 1% donkey serum for 1 h at room temperature. Alexa Fluor-647-conjugated donkey anti-rabbit secondary antibody (711-605-152) was used at a 1:1000 dilution. The samples were then washed with PBS and stabilized for MINFLUX imaging using 150 mm gold nanoparticles (EM.GC150, BBI Solutions) for 5–10 min, then mounted in imaging buffer, 50 mM Tris-HCl, 10 mM NaCl, 10% (w/v) glucose, and 64 μg/mL catalase, 0.4 mg/mL glucose oxidase, and 10 to 25 mM mercaptoethylamine, pH 8.0, as previously described, 35 , 36 A MINFLUX microscope (Abberior Instruments) equipped with a spatial light modulator-based beam shaping module and an electro-optical detector-based MINFLUX was used. Microscopy and MINFLUX measurements were performed using Imspector Software (v.16.3.15635-m2205-win64-MINFLUX; RRID:SCR_015249) using MINFLUX sequence templates seqIIF and DefaultIIF3D for 2D- and 3D-MINFLUX measurements as previously described. 37 Molecular precision was enabled through a reflection-based stabilization unit based on a 980-nm as previously described. 16 Before starting each MINFLUX measurement, the stabilization of the microscope was activated. Measurements were conducted with a stabilization precision of < 1 nm. A 642 nm (CW) excitation laser, a 405 nm activation laser, and detection windows in the range of 650–750 nm were used. The emitted fluorescence photons were counted using two avalanche photodiodes with the appropriate fluorescence filters. Molecular precision was enabled through a reflection-based stabilization unit based on a 980-nm laser as previously described. 16 For structure counting and ring-like structure diameter measurements, localizations were exported as 2D projections with a pixel size of 4 nm (2D) or 5–6 nm (3D) using Imspector Software. Individual structures were counted in ImageJ and classified according to shape: ring-like, punctate, or other. The diameter of each ring-like structure was considered as the largest measurable distance between points with zero localizations in the structure center and measured using the line tool in ImageJ-Fiji Software.

Mouse NINJ1 co-localization and imaging

For immunofluorescence labelling, iBMDMs stably expressing HA-NINJ1 cells were seeded onto glass coverslips. The next day the cells were treated or not with LPS (0.5 μg/mL) for a total of 3.5 h, during which nigericin (10μM) was added for the final 1 h. Cells were then washed with PBS and permeabilized using 0.1% Triton X-100 in PBS for 5 min and blocked with 1% BSA in PBS for 1 h. Cells were incubated with primary antibodies for 1 h, washed, then incubated with Alexa Fluor-488 or Cy3-conjugated secondary antibodies for 1 h. Prior to imaging, cells were stained with DAPI for 5 min. Antibodies used: anti-HA-tag (Cell Signaling, 3724), anti-KDEL (Enzo, ADI-SPA-827), anti-GM130 (BD Biosciences, 610823), anti-HSP60 (BD Biosciences, 611563). Cells were imaged as described in 38 by spinning disk confocal microscopy (Quorum) on a Zeiss Axiovert 200M microscope with a ×63 objective and an additional ×1.5 magnifying lens. Images were acquired by a CCD camera (Hamamatsu Photonics) driven by Volocity software.

Supplementary Material 1 Figure S1. NINJ1 and NINJ2 purification, related to Figure 1 (A) SDS-PAGE of MBP-NINJ1 and MBP-NINJ2 purification from E. coli expressed. (B) Negative staining EM image of MBP-NINJ1. NINJ1 large rings were hardly detected before MBP removal. (C) SDS-PAGE of NINJ1-GFP-FLAG purification and detection by in-gel GFP fluorescence at the expected molecular weight. (D) SDS-PAGE (top) and negative staining EM image (bottom) of NINJ1-GFP-FLAG after GFP-FLAG removal by TEV. (E) Liposome breakdown by NINJ1, shown by negative staining EM images before and after NINJ1 incorporation into liposomes. Liposome membrane breakdown by NINJ1 is clearly shown. All data are representative of three or more independent experiments. 2 Figure S2. Cryo-EM flow charts of NINJ1 data processing, related to Figure 2 (A) Flow chart for cryo-EM data processing of NINJ1 rings reconstituted from liposomes. (B) A raw cryo-EM micrograph of NINJ1 large rings in detergent. (C) Chart-flow for cryo-EM data processing of NINJ1 large ring segments purified from detergent. (D-E) NINJ1 subunits fitted in the cryo-EM map viewed for individual helices (D), the hydrophobic, concave side of the ring (E, left) and the hydrophilic, convex side of the ring (E, right). 3 Figure S3. Structural representation on NINJ1 ring segments, related to Figure 2 and 3 (A) Model of a longer NINJ1 ring segment, showing the high-density face labeled in deep purple and low-density face labeled in orange based on the 2D classification shown in (B). (B) 2D classification of NINJ1 ring segments, showing that all classes possess a high-density concave face and a low-density convex face. (C) Western blots of anti-Cas9, anti-NINJ1 and anti-GAPDH (loading control) for WT, control sgRNA cells and single NINJ1 knockout clones. The data are representative of three or more independent experiments. (D) Genomic sequencing for NINJ1 KO clones (clones 37, 38 and 39). 4 Figure S4. Immunofluorescence imaging of NINJ1 and organelle localization, related to Figures 3 (A) Confocal imaging of LPS-primed, Dox-induced NINJ1-GFP cells (green) stained with the mitochondria marker MitoTracker (magenta), and DNA (Hoechst dye, blue). Scale bar: 10 μM. (B) IF confocal imaging of LPS-primed, or LPS/Nigericin-treated, Dox-induced NINJ1-GFP cells (green) for the Golgi marker GM130 (magenta), and DNA (Hoechst dye, blue). Scale bar: 10 μM. (C) IF confocal imaging of NINJ1 KO iBMDMs reconstituted with HA-NINJ1 (green or magenta) cells co-localized with the following organelle markers: MemBrite (green, for plasma membrane), KDEL (magenta, for ER), GM130 (magenta, for Golgi), HSP60 (magenta, for mitochondria) and EEA1 (magenta, for early endosome). Scale bar: 26 μM. (D) Quantitative analysis of IF experiments of ASC speck formation upon LPS/Nigericin activation of NINJ1 KO THP-1 cells reconstituted with DOX-inducible NINJ1-GFP, fixed and imaged 24 h after Dox introduction or DMSO as a control. Data shown are mean ± SEM based on n=3 fields of view for each condition. Statistical significance was assessed by 2-tailed t-test. *** denotes p < 0.0001. (E) IF confocal imaging of Dox-induced NINJ1-GFP THP-1 cells in the presence of NLRP3 inhibitor MCC950, in the presence of LPS or LPS + Nigericin, visualized for NINJ1 (green), ASC (magenta), and DNA (Hoechst dye, blue). Scale bar: 10 μM. (F) Confocal imaging of transiently co-expressed NINJ1-GFP and GSDMD-NT-mCherry, imaged 24 h posted after DNA transfection. visualized by NINJ1 (green), GSDMD (magenta), and DNA (Hoechst dye, blue). Scale bar: 10 μM Data in (A-B) and (E-F) are representative of three or more independent experiments. 5 Figure S5. EM imaging of purified NINJ1 mutants in detergents and in reconstituted liposomes, related to Figure 5 (A-E) Negative staining EM images to examine NINJ1-containing structures in detergents and with liposomes, for WT NINJ1 and MBP control (A), mutants at the TM helix kinks (B), mutants at the hydrophobic face (C), mutants at the hydrophilic face (D), and mutants at a potential PI(4)P-binding surface. The scores for oligomerization and for liposome rupture relative to WT NINJ1 are shown at the top left corners of each images. For oligomerization, + = many rings as for WT NINJ1; +/− = some filaments and rings in a background of small dots; - = a few filaments in a background of small dots; - - = barely any rings or filaments with only a background of small dots. For liposome rupture: + = dissolved small rings as for WT NINJ1; +/− = small rings plus a few undissolved liposomes; - = lots of undissolved liposomes; - - = largely undissolved liposomes. All Data are representative of three or more independent experiments. 6 Figure S6. Oligomerization of NINJ1 mutants by sucrose gradient and cellular localization in HEK 293T cells, related to Figure 5 (A) Sucrose gradient profiles of purified NINJ1 mutants in comparison to NINJ1 WT. (B) Confocal imaging of GFP tagged NINJ1 WT and mutants (green) co-stained with the plasma membrane dye, CellMask (magenta), and DNA (Hoechst dye, blue). Cells were fixed 24 h after DNA transfection and imaged. (C) SDS-PAGE of liposome pelleting assay for NINJ1 mutants at the potential PI(4)P-binding surface. Sup: supernatant. All Data are representative of three or more independent experiments. 7 Figure S7.

Analysis of NINJ2 mutants and NINJ1

MINFLUX microscopy during secondary necrosis, related to Figure 5 and 6 (A) A table summarizing the mutational analysis in vitro and in cells for NINJ2 mutants. The mutants were scored based on the following assays: 1) oligomerization by EM imaging, 2) liposome rupture, 3) cell death (LDH release), and 4) plasma membrane localization by confocal imaging. (B) Negative staining EM imaging of purified NINJ2 mutants in detergents and in reconstituted liposomes. (C) Confocal imaging of GFP tagged NINJ2 WT and mutants (green) co-stained with the plasma membrane dye, CellMask (magenta) and DNA (Hoechst dye, blue). Cells were fixed 24 h after DNA transfection and imaged. Scale bar: 10 μM. (D) Cytotoxicity assay of NINJ2 and mutants in HEK293T cells measured based on % of LDH release, at 24 h post transfection. Data shown are mean ± SEM of three or more independent experiments. Statistical significance was assessed by 2-tailed t-test. ** and * denote p < 0.001 and 0.01, respectively. (E) Blue-native PAGE gel and western blot of endogenous NINJ1 in the supernatant of primary BMDMs upon pyroptosis activation in comparison to untreated cells. (F) 2D-MINFLUX images (confocal inset) of primary mouse BMDMs without (control; left) or with secondary necrosis stimulation (ABT-199; right). The BMDMs were stimulated by treatment with ABT-199 or DMSO (vehicle). Cells were fixed in 4% paraformaldehyde and 0.1% glutaraldehyde, followed by immunolabelling with anti-NINJ1 antibody (rabbit monoclonal clone 25, Genentech, Inc) and Alexa Fluor 647-conjugated whole IgG secondary antibody. top ; Samples were imaged by MINFLUX. Small dashed, white boxes indicate images shown in (bottom). Scale bars: 500 nm. bottom ; Representative structures of each classification category for ABT-199 treated cells, ring-like (i), discrete puncta (ii), or other (iii). Scale bar: 40 nm. (G) Pie charts for NINJ1 structures counted and classified as ring-like, discrete puncta, or other for control BMDMs (DMSO) and those stimulated to undergo secondary necrosis (ABT-199). (H) A graph of the largest measured inner diameters of the identified ring-like structures in the 2D plane, shown with a median of 23.9 nm (interquartile range 21.0, 65.8 nm). Data in (B) and (C) are representative three independent experiments, and data in (F-H) are representative of two independent experiments. 8 Movie S1. NINJ1 mediated membrane rupture in THP-1 cells, related to Figure 4 Live cell imaging of a THP-1 cell expressing NINJ1-GFP showing cell death leading cell rupture mediated by NINJ1. Imaging started after 4 h of LPS priming and upon Nigericin addition. Cells were imaged for 66 min at 2 min intervals using a 60x objective on a spinning disk confocal microscope. 9 Movie S2. 3D MINFLUX nanoscopy of endogenous NINJ1 control (LPS-primed) primary mouse macrophages, related to Figure 6 The white cube shown has a length of 100 nm along each axis. The rainbow-colored scale indicates distance in the Z-plane in units of meters (e.g., 3.5e-07 = 350 nm). 10 Movie S3. 3D MINFLUX nanoscopy of structures formed by endogenous NINJ1 in LPS-primed mouse macrophages stimulated to undergo pyroptosis, related to Figure 6 The white cube shown has a length of 100 nm along each axis. The rainbow-colored scale indicates distance in the Z-plane in units of meters (e.g., 4e-07 = 400 nm). 11