CmiCASP1 cleaves CmiGSDMA/B to execute pyroptosis in C. milii.

(A) Timeline of GSDM gene family evolution showing phylogenetic branching in million years ago (MYA), the origin of ancestral GSDM, gene duplications giving rise to GSDME and PJVK, and lineage-specific expansions of GSDMA, GSDMB, GSDMC, and GSDMD in vertebrates. (B) Structural comparison of CmiGSDMA/B and HsaGSDMD. The predicted structure of CmiGSDMA/B (purple) was superimposed with the structure of HsaGSDMD (PDB: 6N9O) (yellow) using PyMOL. The N- and C-terminal domains are indicated and separated by a dashed line. (C) Phylogenetic analysis of caspases. Accession numbers are listed in Supplementary file 1. (D) RMSD and TM-score for structural comparisons of CmiCASP1 with HsaCASP1 and HsaCASP3. RMSD was calculated using PyMOL, and TM-score was calculated using Foldseek. (E) The distinct substrate-recognition pockets are shown in different colors, with the residues forming each pocket colored accordingly. (F) Hydrophobicity analysis of the S4 pocket. The region outlined by the blue dashed line corresponds to the S4 pocket. Cyan indicates hydrophilic regions, whereas yellow indicates hydrophobic regions. The magenta sticks indicate bound peptide ligands: FLTD in HsaCASP1 and DEVD in HsaCASP3. Residues forming the S4 pocket are labeled in black. The yellow label indicates the maximum distance accommodated by the S4 pocket. (G) Bright-field images of HeLagsdmd/e DKO cells transfected with CmiGSDMA/B, CmiCASP1 or both, with or without Z-VAD-FMK treatment. Red arrows mark swollen, bubbling pyroptotic cells. (H) Microscopic analysis of cell death phenotypes using Annexin V-FITC/PI staining. Confocal microscopy images of HeLagsdmd/e DKO cells transfected as indicated. (I) Flow cytometry analysis of Annexin V-FITC and PI double-stained cells. HEK293T cells were transfected and stimulated as indicated, followed by double staining with Annexin V-FITC and PI. (J) LDH release assay. LDH release was measured in the supernatants of HEK293T cells transfected with the indicated constructs. (K) ATP-based cell viability assay of HEK293T cells under different transfection and treatment conditions, as indicated. (L) Western blot analysis showing CmiCASP1-mediated cleavage of CmiGSDMA/B in HEK293T cells in the presence or absence of Z-VAD-FMK. Data are presented as mean ± SD from three independent experiments. The p values were calculated using Student’s t-test.

CmiGSDMA/B is cleaved by CmiCASP1 at D241 and D288.

(A) Cartoon diagram showing the domain architecture of GSDMs and inflammatory caspase-mediated cleavage. (B) Western blot analysis of CmiCASP1-mediated cleavage of CmiGSDMA/B and its mutants in HEK293T cells. (C) LDH release assay in HEK293T cells co-expressing CmiCASP1 with CmiGSDMA/B or its mutants. (D) Flow cytometric analysis of Annexin V-FITC and PI double stained HEK293T cells transfected with the indicated constructs. (E) Cell viability assay of HEK293T cells transfected with the indicated expression vectors. (F) Bright-field images of HeLagsdmd/e DKO cells co-transfected with CmiCASP1 and CmiGSDMA/B or its mutants. Red arrows indicate cells with pyroptotic morphology. (G) Microscopic analysis of cell death phenotypes by Annexin V-FITC/PI staining in HeLagsdmd/e DKO cells under the indicated transfection conditions. (H) Western blot analysis of CmiCASP1-mediated cleavage of CmiGSDMA/B and its mutants in HEK293T cells treated with Z-VAD-FMK. (I-K) LDH release, flow cytometric analysis of Annexin V-FITC/PI staining, and cell viability assay of Z-VAD-FMK treated HEK293T cells transfected with the indicated expression vectors. (L) Bright-field images of HeLagsdmd/e DKO cells under different transfection and treatment conditions, as indicated. Red arrows indicate cells with pyroptotic morphology. Data are shown as mean ± SD from three independent experiments. The p values were calculated using Student’s t-test.

Distinct functional activities of N241 and N288.

(A) Representative bright-field images of HeLagsdmd/e DKO cells transfected with vectors expressing CmiGSDMA/B, N241 or N288. Red arrows indicate pyroptotic cells. (B) Microscopic analysis of cell death phenotypes by Annexin V-FITC/PI staining in HeLagsdmd/e DKO cells transfected with expression vectors, as indicated. (C-E) Flow cytometry analysis of Annexin V-FITC and PI double-stained, LDH release assay, and ATP-based cell viability assay in HEK293T cells transfected with the indicated expression vectors. (F) The structure of CmiGSDMA/B predicted using AlphaFold. The N- and C-terminal domains are indicated and separated by a dashed line. The green region represents the additional segment present in N288 relative to N241 and the red triangle indicates the β1-β2 loop. (G) Subcellular localization of GFP-tagged N241 and N288 in HeLagsdmd/e DKO cells. (H) Western blot analysis showing CmiCASP1-mediated cleavage of N288 in HEK293T cells and its inhibition by Z-VAD-FMK treatment. (I-K) Flow cytometry analysis of Annexin V-FITC/PI staining, LDH release assay, and cell viability assay in HEK293T cells co-expressing CmiCASP1 and N288 in the presence or absence of Z-VAD-FMK. (L) Representative bright-field images of HeLagsdmd/e DKO cells expressing N241 alone or together with N288. Red arrows indicate cells with pyroptotic morphology. (M-O) Flow cytometry analysis of Annexin V-FITC/PI staining, LDH release assay, and cell viability assay in HEK293T cells expressing N241 alone or together with N288, as indicated. (P) Co-immunoprecipitation analysis of the interaction between N241 and N288 in HEK293T cells. (Q) Subcellular localization of GFP-tagged N241 in the presence or absence of mCherry-tagged N288 in HeLagsdmd/e DKO cells. Quantitative data are presented as mean ± SD from three independent experiments. The p values were calculated using Student’s t-test.

CmiCASP1 directly binds LPS to trigger activation.

(A) Sequence alignment of caspase CARD domains generated using MAFFT. The conserved amino acid residues are highlighted in blue. Residues important for LPS binding to caspase-4/11 are highlighted in red. (B-D) Hydropathy plot of the CARD domains of mammalian caspase-1/4/5/11, CmiCASP1 and CmiCASP1-mutant generated using ProtScale. In the CmiCASP1-mutant, conserved positively charged residues within the predicted LPS-binding surface were replaced by alanine. Dashed box indicates the distinctive hydrophobic grooves within CARD domains of caspase-1 and caspase-4 homologs and CmiCASP1. (E) Streptavidin-based pull-down assay to detect biotin-conjugated LPS binding to Flag-tagged catalytically dead CmiCASP1-C279A, CmiCASP1-C279A-ΔCARD, the isolated CARD domain, and the CmiCASP1-C279A-mutant in transfected HEK293T cell lysates. (F) Streptavidin-based pull-down assay to detect the binding of biotin-labeled LPS or Pam3CSK4 to Flag-tagged CmiCASP1-C279A and CmiCASP1-C279A-mutant, with the indicated unlabeled ligands for competition experiments. (G) LPS induced activation of purified CmiCASP1. Caspase activity was determined by measuring the fluorescence intensity of free AMC hydrolyzed from Z-VAD-AMC. (H) Purified CmiGSDMA/B was cleaved in vitro by CmiCASP1, which was activated by LPS.

N241 exhibits antibacterial activity against Gram-negative bacteria.

(A) Effect of CmiGSDMA/B and its fragments on bacterial viability. HEK293T cell lysates containing full-length CmiGSDMA/B, N241 or N288 fragments were incubated with Gram-negative bacteria (E. coli, E. tarda strain EIB202, and V. parahaemolyticus) and Gram-positive bacteria (S. aureus, L. monocytogenes, and B. megaterium). Bacterial viability was assessed by CFU assays. Only plates containing 30-300 colonies were used for CFU calculation. (B) Representative serial dilution plating of E. coli treated with the indicated GSDM-containing lysates. E. coli suspended in PBS was incubated with the indicated lysates at 37°C for 40 minutes, serially diluted 10-fold and plated on LB agar for CFU determination. (C) Representative fluorescence microscopy images showing the viability of E. coli assessed by DMAO/PI staining. Bacteria were incubated for 40 minutes with the indicated concentrated lysates collected from transfected HEK293T cells. DMAO permeates both live and dead bacteria, while PI stains only dead bacteria. (D, E) Flow cytometry analysis of DMAO and PI double-stained E. coli (D) or S. aureus (E) after incubation with the indicated concentrated lysates. (F, G) Structural analysis of N241 highlights putative membrane contact and insertion surfaces enriched in basic residues including three basic patches (BPs). (H) Lipid strip binding assay compared the differences in lipid binding between purified wild-type N241, N288, and N241-m. (I) CFU assay measuring the viability of E. coli after incubation with lysates containing N241 and N241-m. (J) Representative fluorescence microscopy images showing the viability of S. aureus assessed by DMAO/PI staining. (K) LDH release assay of HEK293T cells transfected with the indicated vectors. Quantitative data are presented as mean ± SD from three independent experiments. The p values were calculated using Student’s t-test.

Working model of LPS-responsive CmiCASP1-CmiGSDMA/B pathway.

Schematic model of the CmiCASP1-mediated non-canonical pyroptosis pathway in C. milii, highlighting its potential contributions to antibacterial immune responses.

Characteristics of CmiGSDMA/B and CmiCASP1 in Callorhinchus milii.

(A) Phylogenetic analysis of CmiGSDMA/B. GSDM members identified in C. milii are labeled in red. The yellow stars indicate the GSDMA/B branches of cartilaginous fish. (B) The domain architecture of C. milii caspases predicted by InterProScan, showing the presence and arrangement of N-terminal interaction modules, such as CARD and DED domains, and the conserved caspase protease core comprising the p20 and p10 domains. (C) A heatmap showing sequence similarity among full-length caspases. Higher sequence similarity is indicated by darker blue. CmiCASP1 is highlighted in red. (D) Multiple sequence alignment shows the key amino acid residues in HsCASP1 that form the substrate-binding pockets and catalytic site, as well as their corresponding positions in CmiCASP1. Sequence and structural analyses predict that the S1 pocket residues of CmiCASP1 (R173, Q277, S332) correspond to those of HsaCASP1 (R179, Q283, S339), whereas the S2, S3, and S4 pocket residues V331/F333, R334, and D335/I341/K374 correspond to V338/W340, R341, and H342/V348/R383 in HsaCASP1, respectively. The conserved catalytic motif QACRG is also indicated, with C279 in CmiCASP1 corresponding to the catalytic cysteine C285 in HsaCASP1.

Homologs of inflammasome components in cartilaginous fish.

(A) Schematic representation of key inflammasome components and domains architectures in C. milii. (B, C) Multiple sequence alignment of IL-1β and IL-18 homologs, respectively. The alignments were generated using the MAFFT algorithm. Conserved amino acid residues are highlighted in blue. The red box represents the predicted or previously identified caspase-1 cleavage sites. (D) Schematic representation of the domain architectures of inflammasome-related homologs in C. plagiosum. (E) A heatmap of inflammasome-related gene expression across eleven tissues in C. plagiosum. Color intensity corresponds to expression level, with darker shades indicating higher expression.

Effect of N241 on the activity of S. aureus.

(A) Representative serial dilution plating of S. aureus treated with the indicated GSDM lysates. S. aureus suspended in PBS was incubated with the indicated lysates at 37°C for 40 min, serially diluted 10-fold and plated on LB agar for CFU determination. (B) Representative fluorescence microscopy images of bacterial viability assessed by DMAO/PI staining. Bacteria were incubated for 40 min with the indicated concentrated lysates collected from transfected HEK293T cells. (C) Sequence alignment showing basic residues within the predicted basic patches (BPs) of GSDM proteins. Basic residues within the BPs are highlighted in red, and dashes indicate alignment gaps. Red labels below the alignment indicate the corresponding basic residues in CmiGSDMA/B.