Structural basis for GSDMB pore formation and its targeting by IpaH7.8 – Nature

Constructs and mutagenesis

The coding sequences of full-length GSDMs were cloned into a pET28-His-SUMO vector after the N-terminal His₆-SUMO tag. In regard to the GSDMB construct used for cryo-EM structural determination, a human rhinovirus 3C protease (3C) site (LEVLFQ/GP) was inserted after residue K239. The coding sequence of S. flexneri IpaH7.8 was cloned into a pET26b vector with a C-terminal 6XHis tag, and cloned into pET22b vector without an affinity tag for coexpression with GSDMB. Caspase-11 (96–373) was cloned into a pET22b vector to purify the active form of the p20–p10 complex. For the cellular experiments, GSDMs (full length) and GSDM-N were cloned into a pcDNA3.1 vector in which a FLAG-tag was fused at the C terminus while IpaH7.8 was inserted into a pCMV-HA vector, resulting in a fusion protein with an N-terminal HA tag. All mutations in this study were introduced using either the QuikChange Site-Directed Mutagenesis Kit (Stratagene) or Gibson Assembly Master Mix (New England BioLabs), and all plasmids were verified by sequencing.

Protein expression and purification

To obtain the GSDMB–IpaH7.8 complex, E. coli BL21 (DE3) cells harbouring the expression plasmids of pET28-His-SUMO-GSDMB and pET22b-IpaH7.8 were grown in lysogeny broth medium supplemented with 50 µg ml^–1 kanamycin and 100 µg ml^–1 ampicillin at 37 °C. Protein expression was induced by the addition of 0.5 mM isopropyl β-D-1-thiogalactopyranoside at 20 °C for 16 h when optical density (OD₆₀₀) reached 0.8. Cells were collected by centrifugation at 5,000g for 20 min. Harvested cells were lysed by sonication in a buffer containing 25 mM Tris-HCl pH 8.0, 150 mM NaCl, 2 mM β-mercaptoethanol and 25 mM imidazole. Lysates were centrifuged at 18,000g and 4 °C for 30 min to remove insoluble fractions. Supernatants containing recombinant proteins were purified using Ni-NTA agarose (Qiagen) according to the manufacturer’s instructions. His₆-SUMO tag removal was performed on a Ni-NTA column at 4 °C overnight with the addition of the recombinant Ulp1. Flowthrough non-tagged proteins were further purified using a Hitrap Q HP ion-exchange column (Cytiva), then a Superdex Increase 200 (10/300) size-exclusion column (Cytiva) in a buffer containing 25 mM HEPES pH 7.5 and 150 mM NaCl. All purified proteins were confirmed by Coomassie blue staining of SDS–PAGE.

Similar protocols were applied for the expression and purification of all individual GSDMs, IpaH7.8 and their mutants, except that the His₆-SUMO tag was retained for GSDMs used in the in vitro ubiquitination assay. All purified proteins were concentrated to approximately 5–10 mg ml^–1 before use.

The GZMA plasmid pET26b-GZMA was a kind gift from J. Lieberman³⁵. The plasmids of E1 (pET21d-hUbE1), E2 (pET15-hUbE2D2) and ubiquitin (pET15-Ub) were kind gifts from C. Wolberger, W. Harper and R. Klevit, respectively^36,37,38. Expression and purification followed previous protocols.

In vitro ubiquitination activity assay

In vitro ubiquitination reactions were performed in buffer A (25 mM HEPES pH 7.5, 50 mM NaCl, 0.1 mM DTT, 10 mM MgCl₂ and 5 mM ATP). Components were mixed as indicated at concentrations of 0.4 µM E1 (human UbE1), 2 µM E2 (human UbE2D2), 10 µM E3 (IpaH7.8^WT or IpaH7.8^C357A mutant), 200 µM ubiquitin and 10 µM GSDM (WT or indicated mutants). Reactions were incubated at 37 °C for 2 h and stopped by the addition of SDS–PAGE loading dye, followed by boiling for 5 min before electrophoresis. Ubiquitination was evaluated by Coomassie blue staining of SDS–PAGE.

Liposome leakage assay

The liposome leakage assay was performed following an established protocol⁹. Briefly, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine and PS or CL (Avanti Polar Lipids) were mixed at the indicated ratio in a glass tube. The solvent chloroform was evaporated under a stream of nitrogen gas for 30 min. The dry lipid film was then rehydrated with buffer B (25 mM HEPES pH 7.5, 150 mM NaCl) supplemented with 50 mM 6-FAM (Tokyo Chemical Industry). 6-FAM-loaded liposomes were then extruded through a 1 μm membrane (Whatman Nuclepore) using a mini-extruder (Avanti Polar Lipids). To remove unencapsulated 6-FAM, extruded liposomes were subjected to a PD-10 desalting column (Cytiva) equilibrated with buffer B. For the liposome leakage assay, liposomes were incubated with proteins of GSDMB/D and/or IpaH7.8 with or without activating enzymes (GZMA for GSDMB and caspase-11^p10/p20 for GSDMD). Reactions were performed on a 384-well plate, with release of 6-FAM dye monitored by fluorescence at 517 nm using a SpectraMax M5 plate reader (Molecular Devices) with excitation at 495 nm for 60 min at 1 min intervals.

Liposome pulldown assay

Liposomes were prepared as described above, except that fluorescent dye was not used. Liposomes were incubated with GSDMB/D proteins in/without the presence of IpaH7.8 at various molar ratios with or without activating enzymes. Mixtures were incubated for 30 min at 4 °C before sedimentation at 20,000g for 30 min at 4 °C. Supernatants were transferred immediately to new tubes and pellets were washed twice with buffer B, then resuspended in an equal volume of buffer. Proteins in both pellets and supernatant were then analysed by Coomassie blue staining of SDS–PAGE.

ITC assay

Protein concentrations of non-tagged GSDMs and IpaH7.8 were measured in triplicate using a NanoDrop One Microvolume UV-Vis Spectrophotometer (Thermo Fisher Scientific) based on their extinction coefficients. Isothermal titration calorimetry measurements were performed at 20 °C using a VP-ITC microcalorimeter (MicroCal). Experiments were performed by the injection of 250 μl of IpaH7.8 solution (200 μM) into a sample cell containing 2 ml of GSDMB (10 μM) in 25 mM Tris-HCl pH 8.0 and 150 mM NaCl. In total, 25 injections were administered at 300 s intervals. In regard to human GSDMD (hGSDMD) and mouse GSDMD (mGSDMD), 250 μl of IpaH7.8 solution (625 μM) was titrated into a sample cell containing 2 ml of either hGSDMD (40 μM) or mGSDMD (40 μM). All ITC data were analysed using Origin Software provided by the manufacturer and fitted to a one-site binding model.

Cell culture and transfection

The 293T cells were obtained from the American Type Culture Collection and were frequently checked in regard to their morphological features and functionalities. Cells were grown in DMEM (Gibco) supplemented with 10% (v/v) fetal bovine serum (Gibco) and 2 mM l-glutamine at 37 °C in a 5% CO₂ incubator. Transient transfection in 293T cells was performed using Lipofectamine 3000 (Thermo Fisher Scientific) following the manufacturers’ instructions.

Immunoprecipitation assays

For detection of GSDMB ubiquitination in cells, pcDNA-FLAG-GSDMB (isoform 1) was cotransfected with pCMV-HA-IpaH7.8 (WT or indicated mutants) into HEK293T cells in a 10 cm tissue culture dish. Eight hours after transfection, a final concentration of 10 μM bortezomib (Sigma Aldrich) was added to the cell culture to reduce proteasome-mediated protein degradation. After a further 8 h, cells were collected and lysed in 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5% NP-40 and 1× protease inhibitor cocktail (Sigma Aldrich). Lysate was added to 25 μl of Anti-FLAG M2 Magnetic Beads (Sigma Aldrich, no. M8823) and incubated at 4 °C for 3 h with gentle rotation. Beads were washed three times with PBS buffer then eluted with 50 μl of PBS buffer containing 100 μg ml^–1 FLAG peptide (Sigma Aldrich, no. F3290). Eluted samples were boiled with an equal volume of 2× SDS Loading buffer (Bio-Rad) then processed for immunoblotting with one of the following antibodies: anti-FLAG (Sigma Aldrich, no. F1804, 1:1,000), anti-actin (Cell Signaling Technology, no. 3700S, 1:1,000), anti-HA (Cell Signaling Technology, no. 3724S, 1:1,000) or anti-ubiquitin (Thermo Fisher Scientific, no. PA3-16717, 1:1,000).

Cellular degradation assay

One each of plasmids pcDNA-FLAG-GSDMB and pcDNA-FLAG-GSDMD (250 ng) (WT or indicated mutants) was cotransfected with 500 ng of the pCMV-HA-IpaH7.8 plasmid into HEK293T cells seeded in a 12-well plate at 1.5 × 10⁵ cells per well. After 40 h cells were lysed in RIPA buffer (Thermo Fisher Scientific), added to an equal volume of 2× SDS Loading buffer (Bio-Rad) and processed for immunoblotting.

Cytotoxicity assay

Cell death was determined by Hoechst/PI double-staining assay: 150 ng of the indicated pcDNA-FLAG-GSDMB construct or 75 ng of pcDNA-FLAG-GSDMD plasmid (FL, NT or indicated mutants) was transfected into HEK293T cells seeded in a 96-well plate at 2 × 10⁴ cells per well. For IpaH7.8 inhibition, pcDNA-FLAG-GSDMB-NT (isoform 4) was cotransfected with 200 ng of the pCMV-HA-IpaH7.8 plasmid (WT or C357A). Transfected cells were then cultured for up to 40 h. At the start of the assay, cells were stained with 30 μM PI (Sigma Aldrich) for 10 min followed by 15 μM Hoechst 33342 (Thermo Fisher Scientific) for 15 min at 37 °C in the dark. Afterwards, cells were visualized using a ZOE Fluorescent Cell Imager (Bio-Rad). Cell death was quantified and expressed as the percentage of PI-positive cells among total cells (Hoechst-stained cells).

Bacterial growth inhibition assay

Escherichia coli DH5α was grown overnight in BHI medium, then diluted the following day at 1:100 in BHI and grown for a further 2 h at 37 °C until exponential phase. Next, 1 ml of the bacterial culture was collected with centrifugation at 5,000g for 2 min and resuspended in buffer B to a final bacterial cell density of 5 × 10⁸ ml^–1. For the killing assay, 5 μl of bacteria was added to a 15 μl reaction containing 10 μM full-length GSDMB in the absence or presence of GZMA. Reactions were performed at 37 °C for 2 h. After incubation, 5 μl of treated bacteria was seeded into 200 μl of BHI in flat-bottomed, 96-well plates. Bacterial growth was monitored by reading absorbance at 600 nm over 6 h using a SpectraMax M5 plate reader (Molecular Devices). Numbers of recovered colony-forming units (CFUs) were calculated by normalization of the OD₆₀₀ of treated bacteria (with GZMA in reactions) to untreated bacteria (buffer or without GZMA).

GSDMB pore reconstitution and purification

Purified GSDMB isoform 1 was added to the prepared liposomes, followed by the addition of 3C protease to initiate pore formation. The reaction proceeded on ice for 3 h. Liposomes loaded with GSDMB pore were solubilized by 2% C12E8 (Anatrace) to extract pores. To remove poorly behaving particles and GSDMB-C, samples were further purified using a Superose 6 (10/300) Increase size-exclusion column (Cytiva) equilibrated with buffer B (25 mM HEPES pH 7.5, 150 mM NaCl and 0.006% C12E8).

Negative-staining electron microscopy

For negative staining, 10 μl of the GSDMB–IpaH7.8 complex or GSDMB pore was applied to a glow-discharged, carbon-coated copper grid (Electron Microscopy Sciences). The sample was incubated on the grid for 1 min, stained with 1% uranyl acetate for 1 min and blotted dry. Grids were imaged on a Hitachi H-7650 transmission electron microscope equipped with a 2k CCD camera (Advanced Microscopy Techniques) at the UCONN Health Electron Microscopy Facility.

Cryo-EM grid preparation and data acquisition

For the GSDMB–IpaH7.8 complex, 3.5 μl of freshly purified sample at 0.5 mg ml^–1 was applied to plasma glow-discharged, Quantifoil holey copper grids (R 1.2/1.3, 400 mesh, Electron Microscopy Sciences) using a Vitrobot Mark IV (Thermo Fisher Scientific) set at blotting force 4, blotting time 5.5 s, 100% humidity and 4 °C. Blotted grids were immediately plunged into liquid ethane and transferred to liquid nitrogen for storage. One cryo-EM dataset was collected at the Case Western Reserve University cryo-EM facility on a Titan Krios electron microscope (Thermo Fisher Scientific) equipped with a K3 Summit direct electron detector (Gatan) and a post-column energy filter (Gatan) in counting mode using serialEM. A total of 3,128 movies were recorded at defocus values ranging from −0.8 to −2.5 μm at magnification ×105,000 and pixel size 0.414 Å. For each movie, 58 frames were acquired over 5.25 s at an approximate total dose of 66.95 e⁻ Å^–2.

For GSDMB pores, detergent-solubilized GSDMB pores were concentrated to 0.6 mg ml^–1 then frozen onto Quantifoil holey copper grids coated with ultrathin carbon film (R 1.2/1.3, 400 mesh, Electron Microscopy Sciences). Briefly, a 3 μl drop of GSDMB pore sample was applied to a plasma glow-discharged lacey carbon grid mounted on a Vitrobot. The grid was then blotted with filter paper for 6 s at blotting force 10 after a waiting time of 2 s. Humidity and temperature in the Vitrobot were set to 100% and 4 °C, respectively, throughout the operation. The blotted grid was then plunged into liquid ethane and transferred to liquid nitrogen for storage. The cryo-EM dataset was collected at the cryo-EM facility at the University of Massachusetts Chan Medical School on a Titan Krios electron microscope (Thermo Fisher Scientific) equipped with a K3 Summit direct electron detector (Gatan) and a post-column energy filter (Gatan). A total of 6,376 movies were collected in counting mode, each containing 50 frames and a total exposure dose of 50 e^–1 Å^–2. Magnification was set to 105,000, pixel size was 0.83 Å and defocus range −1.0 to −2.0 μm.

Cryo-EM image processing

Raw movies were corrected by gain reference and for beam-induced motion and summed into motion-corrected images using MotionCor2 (ref. ³⁹). CTF parameters were determined using CTFFind4 (ref. ⁴⁰) and refined later in cryoSPARC⁴¹.

For the GSDMB–IpaH7.8 complex, after particle picking using the general model in crYOLO⁴² the coordinates (1,522,742 particles in total) were transferred to cryoSPARC for subsequent processing. Several rounds of 2D classification were performed to eliminate ice, carbon edges and false-positive particles containing noise. Frequently featured classes containing 307,276 particles were selected and subjected to ab initio 3D reconstruction followed by heterogeneous refinement. The optimal class, containing 113,959 particles, was selected for homogenous and non-uniform refinement⁴³. Resolution of the final electron density map was estimated at 3.8 Å, based on the gold-standard Fourier shell correlation (FSC) criterion of 0.143 (ref. ⁴⁴). The local resolution distribution of the map was determined by ResMap⁴⁵. The density map sharpened in cryoSPARC was used to produce figures.

For GSDMB pores, a total of 692,212 particles were initially extracted by both manual and automated particle picking in cryoSPARC. Two-dimensional classification was performed in cryoSPARC to eliminate ice, carbon edges and false-positive particles containing noise. After 2D classification, 156,037 particles were imported into Relion-4.0 for 3D classification with an initial model generated de novo in cryoSPARC using the same particle set. C1 symmetry was used for the first round of 3D classification; 3D classes with relatively clear features of C24 symmetry were selected for an extra round of 3D classification with C24 symmetry to discard bad particles. Next, 41,799 particles from the 3D class with optimal resolution were imported back to cryoSPARC for non-uniform refinement. With C24 symmetry, the resolution of the GSDMB pore map was 4.96 Å as measured by gold-standard FSC of 0.143. Focus refinement with a mask excluding the β-barrel region improved local resolution of the GSDMB globular domain to 4.48 Å.

Model building and structure analysis

Atomic models of both the IpaH7.8–GSDMB complex and GSDMB pore were built and refined into cryo-EM density using Coot⁴⁶ and PHENIX⁴⁷. For the IpaH7.8–GSDMB complex, AlphaFold2-predicted structures of IpaH7.8 and GSDMB were used as starting models²⁷. Models of IpaH7.8 and GSDMB were docked into EM density as a rigid body in UCSF Chimera⁴⁸ then manually adjusted in Coot. The structural model of the complex was further refined using ‘phenix.real_space_refine’, with secondary structure restraints and Coot iteratively. The quality of the atomic model was evaluated by Molprobity⁴⁹. For GSDMB pores, the structure of GSDMB in the IpaH7.8–GSDMB complex was used as a starting model. A similar procedure was then performed for further adjustment and refinement. Figures were prepared using PyMOL (Schrödinger) and UCSF Chimera.

Reporting summary

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