Clinical specimens and C. perfringens isolation
The study involving human clinical samples was approved by the Ethics Committee of Zhujiang Hospital, Southern Medical University (2025-KY-315-02). Informed consent was obtained from all participants or their guardians. Faecal samples were collected from patients diagnosed with NEC (Bell stage II or III). For bacterial isolation, faecal specimens (0.1 g) were homogenized in 1 ml PBS, and 100 μl of the homogenate was plated onto TSC (Hopebio, HB0253) supplemented with 500 μg ml−1 d-cycloserine (Proteintech, CM01307). Plates were incubated anaerobically (80% N2, 10% H2, 10% CO2, 37 °C, 18–24 h). Single black colonies were selected and cultured in brain–heart infusion (BHI) broth (Hopebio, HB8297-1) anaerobically for 4–6 h. C. perfringens was identified by PCR amplification of the 16S rRNA gene and genes encoding the major toxins (Alpha, Beta, Epsilon, Iota, CPE, NetB, Pfo and Delta toxin).
C. perfringens culture and supernatant toxicity
C. perfringens isolated from faeces of patients with NEC was recovered from glycerol stock, grown overnight in BHI at 37 °C anaerobically and subcultured for 48 h to an optical density at 600 nm (OD600) of around 2.5. The culture supernatant was concentrated around 20-fold. HeLa and HT29 cells and human intestinal organoids were treated with concentrated supernatant (10 µl per 500 µl medium) for 30 min, then imaged (Olympus microscope). LDH release was measured using a commercial kit (Thermo Fisher Scientific, C20301). 2% Triton X-100 was used as a 100% positive control.
16S rRNA sequencing of stool samples
Genomic DNA was extracted from faecal microbiota after overnight anaerobic culture using a magnetic bead-based kit (Tiangen, TGuide S96). The V3-V4 region of the 16S rRNA gene was amplified using primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′). Amplicons were indexed by a second PCR using Illumina-compatible primers, pooled, size-selected by gel electrophoresis and quantified. Paired-end sequencing (2 × 250 bp) was performed on the Illumina NovaSeq 6000 system. Reads were quality filtered using Trimmomatic and Cutadapt, followed by denoising, paired-end merging and chimera removal with DADA2 in R. ASVs were classified against SILVA (release 138) and UNITE (release 8.0) using BLASTn v.2.9.0 and QIIME2 Naïve Bayes (confidence threshold 0.7). Community composition and sample similarity were analysed using R packages. Sequencing data were submitted to the NCBI database (PRJNA1419495).
Genomic DNA extraction, sequencing and phylogenetic analysis
Bacterial DNA was extracted from pure C. perfringens isolates using the Bacterial/Fungal DNA extraction kit (magnetic beads) (Majorbio, T07-100). DNA was fragmented to about 400 bp using the Covaris M220 system, and libraries were prepared using the NEXTFLEX Rapid DNA-Seq Kit, and then processed for paired-end sequencing (2 × 150 bp) on the Illumina NovaSeq X Plus (Illumina). Raw reads were filtered with fastp (v.0.23.2) and assembled using SPAdes (v.3.15.5). The assembled genome was submitted to NCBI (PRJNA1419495). Seven genome assemblies (5 references from GenBank and 2 clinical isolates) were quality-controlled with QUAST (N50 ≥ 200 kb) and re-annotated with Prokka (v.1.14.6). Annotated sequences were clustered by OrthoSLC (v.0.2.4) for single-nucleotide polymorphism comparison. Core genes were aligned with Kalign (v.3.3.5) and concatenated into strain-specific core genomes, followed by RAxML (v.8.0.2) phylogenetic tree construction under GTRCAT with 1,000 bootstraps.
Mice
The Cd9f/f mice (T013111) were generated by GemPharmatech using CRISPR–Cas9 gene targeting technology. EllaiCre mice (T050183) and Vil1iCre mice (T004714) were purchased from GemPharmatech. Lyz2cre mice were gifted by P. Zhou. Sex- and age-matched C57BL/6 mice (both sexes; 4-week-old or neonatal) were purchased from Zhuhai BesTest Bio-Tech and randomly assigned to control or experimental groups (not blinded to investigators). Animals were bred and maintained (5 per cage) under SPF conditions with a 12 h–12 h light cycle at 22 ± 1 °C. Animal care and procedures were approved by the Ethics Committee of Guangzhou National Laboratory (GZLAB-AUCP-2024-08-A07) and followed the Guide for the Care and Use of Laboratory Animals. The number of mice per group was selected on the basis of previous publications.
Induction of experimental NEC and antibody neutralization
Experimental NEC was induced in 7-day-old C57BL/6 pups (body weight: 3–3.5 g) over 4 days. Pups were gavaged every 4 h with isotonic infant formula (Similac Advance, 40 μl g−1) containing 1 × 108 CFU of C. perfringens twice daily. Age-matched controls received formula alone. Intestines were collected at P11 for histology. For antibody neutralization, neonatal mice (P0) were housed with dams until P6, then randomly assigned to five groups: (1) negative control; (2) Delta toxin+ C. perfringens (clinical 03) + isotype IgG; (3) Delta toxin+ C. perfringens (clinical 03) + anti-Delta toxin monoclonal antibody; (4) NCTC3181 + isotype IgG; (5) NCTC3181 + anti-Delta toxin monoclonal antibody. From P7 to P11, pups received a daily i.p. injection of 10 µg of the respective antibody in 50 µl sterile PBS using a 34-gauge needle. Throughout, all pups were maintained in thermoneutral cages, fed isotonic infant formula and gavaged twice daily with 1 × 108 CFU C. perfringens. At P11, pups were euthanized for tissue collection and histopathological analysis (GZLAB-AUCP-2025-05-A03). The experimental design schematic was created using BioRender (https://www.biorender.com/).
Cell lines
The HeLa (CCL-2), A549 (CRM-CCL-185), 5637 (HTB-9A), HEK293T (CRL-321), HT29 (HTB-38), RAW264.7 (TIB-71) and HUVEC (PCS-100-013) cell lines were obtained from ATCC. The BV2 (BGT-CMP-219) cell line was purchased from Amyjet Scientific. iBMDMs were provided by F. Shao (National Institute of Biological Sciences, Beijing)51. CD9-KO HeLa cells were obtained from T. Gallagher52. All cell lines were cultured in DMEM or RPMI-1640 medium (Gibco) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin at 37 °C with 5% CO2. All lines were confirmed to be mycoplasma free and authenticated.
Primary macrophage isolation and stimulation
Peritoneal macrophages were collected from mice by lavage with cold PBS containing 2% FBS and 1 mM EDTA. The lavage fluid was centrifuged, and the supernatant was aliquoted and stored at −80 °C for subsequent analysis. The cell pellet was resuspended in FACS buffer or complete DMEM (10% FBS, 1% penicillin–streptomycin). For in vitro stimulation, all purified toxins were endotoxin-depleted before incubation with macrophages.
Human intestinal organoids
De-identified human intestinal organoids were derived from paediatric endoscopic biopsies at Boston Children’s Hospital and provided by the Harvard Digestive Disease Center Organoid Core Facility (IRB-P00000529). Biopsies were digested with 2 mg ml−1 collagenase type I (Life Technologies, 17018029) at 37 °C for 40 min with agitation then centrifuged (500g, 4 °C, 5 min). Crypts were resuspended in Matrigel (Corning, 356231) and polymerized at 37 °C, then cultured in Matrigel with growth medium (50% L-WRN conditioned medium, 45% DMEM/F12, 1% GlutaMAX, 1% N-2/B-27, 10 mM HEPES, 100 μg ml−1 primocin/normocin, 500 nM A83-01, 500 μM N-acetyl-cysteine, 50 ng ml−1 EGF, 10 nM [Leu15]gastrin I, 10 mM nicotinamide, 10 μM SB202190; refreshed every 2 days). After 6–8 days, organoids were passaged with cell recovery solution (Corning, 354253) at 4 °C for 1 h, centrifuged, resuspended in fresh Matrigel, replated and cultured for 2 days, followed by overnight toxin treatment. Cell viability was measured using a MTT assay.
Constructs
Full-length toxin genes were synthesized by Tsingke. Constructs for expressing Delta toxin, GST–Delta toxin or mutated Delta toxin (V17E/T18E) were generated by subcloning Delta toxin (residues 29–318) into the pET-22b (Addgene, 69744-3) vector with a C-terminal His6 tag or the pGEX-4T-1 (Addgene, 27458001) vector with an N-terminal GST tag.
Full-length CD9 genes from human, mouse, bovine, pig, zebrafish, chicken, Drosophila and C. elegans were synthesized by Twist Bioscience and cloned into the pcDNA3.1 vector (Invitrogen, V80020) via Gibson Assembly (NEB, E2621). Human CD9 and CD82 chimeric proteins were constructed as follows: CD9EC1CD82 (CD9(1–33)–CD82(33-53)–CD9(56–221)), CD9EC2CD82 (CD9(1–111)–CD82(111–228)–CD9(196–221)), CD9EC1EC2CD82 (CD9(1–33)–CD82(33–53)–CD9(56–111)–CD82(111–228)–CD9(196–221)), CD82EC1CD9 (CD82(1–32)–CD9(34–55)–CD82(54–250)), CD82EC2CD9 (CD82(1–110)–CD9(112–195)–CD82(229–250)), CD82EC1EC2CD9 (CD82(1–32)–CD9(34–55)–CD82(54–110)–CD9(112–195)–CD82(229–250)), with a GSGSEF linker and a triple-Flag tag at their N termini. These constructs were cloned into the pcDNA3.1 vector by Gibson Assembly. The constructs for expressing Beta toxin, Pfo or α-Hla were generated into the pET-28a vector.
Protein purification
Recombinant proteins (Delta toxin, Beta toxin, Pfo, α-Hla) with either a His6 or GST tag were expressed in E. coli BL21 (DE3, Thermo Fisher Scientific, EC0114) using autoinduction medium at 20 °C for 16 h. Bacterial pellets were lysed by sonication in purification buffer (200 mM NaCl, 20 mM Tris-HCl pH 7.5, 10% glycerol) and clarified by centrifugation. GST-tagged proteins were purified on GSTrap columns (Cytiva, 7513101), eluted with buffer containing 10 mM reduced glutathione (Sigma-Aldrich, PHR1359), pH 7.5. His-tagged proteins were purified on HisTrap nickel columns (Cytiva, 17528601), washed with PP buffer containing 20 mM imidazole, and eluted with a linear imidazole gradient (20–500 mM). Proteins were further purified through HiTrap Q (Cytiva, 17515601) ion-exchange and HiLoad 16/60 Superdex 75 (Cytiva, 28989333) gel-filtration columns using buffer containing 200 mM NaCl, 20 mM Tris-HCl pH 7.5. Purified proteins were concentrated and treated with an endotoxin removal kit (Beyotime, C0268). Protein concentrations were determined from measurements of absorbance at 280 nm (BioDrop spectrophotometer) using extinction coefficients calculated using the ProtParam tool (https://web.expasy.org/protparam/).
BBB permeability assay
Mice received i.p. injections of Delta toxin at different doses and timepoints: 20 µg (7 days before dye injection), 100 µg (6 h before) or 1 mg LPS (6 h before). For NLRP3 inhibition, mice received a single i.p. injection of 10 mg per kg MCC950 (TargetMol, T3701) before acute Delta toxin challenge. For the chronic model, mice were i.p. administered MCC950 1 day before toxin exposure, followed by repeated administration of MCC950 every other day for 7 days. BBB permeability was assessed by retro-orbital injection of 2% Evans blue. After 2 h of circulation, mice were anaesthetized with Avertin (250 mg per kg) and transcardially perfused with PBS; the brains were dissected and photographed. For quantification, the right hemisphere was weighed, minced and incubated in 50% (w/v) trichloroacetic acid. Homogenates were centrifuged (15,000g, 30 min, 4 °C). Evans blue fluorescence was measured at 620/680 nm using the TECAN SPARK reader. The Evans blue concentration was calculated from a standard curve and normalized to brain weight. The experimental design schematic was created using BioRender (https://www.biorender.com/).
In vivo brain MRI
Mice received i.p. injections of 20 μg Delta toxin (7 days before imaging), 100 μg Delta toxin (4 h before imaging) or vehicle. Brain MRI was performed on a Bruker BioSpec 94/30 USR 9.4 T preclinical scanner (Bruker BioSpin MRI). T2-weighted anatomical images were acquired using a rapid acquisition with relaxation enhancement (RARE) sequence. Images were reconstructed and analysed using MicroDicom DICOM Viewer software.
In vivo mouse brain photoacoustic imaging
Mice received i.p. injections of 20 μg Delta toxin (7 days before imaging), 100 μg Delta toxin (4 h before imaging) or vehicle. Raster scanning (10 mm × 10 mm × 2.4 mm) was performed using a small animal photoacoustic imaging system with a 532 nm pulsed laser (pulse width ≤ 8 ns, repetition rate 10 kHz, GAni, Guangzhou G-Cell Technology). Deionized water was used as the acoustic coupling medium. Photoacoustic signals were detected by a PVDF ultrasonic transducer (30 MHz centre frequency), amplified, digitized and reconstructed in real time using PAM 2.0 software (Guangzhou G-Cell Technology), with system synchronization controlled by FPGA.
Immunofluorescence staining
Mice were anaesthetized and perfused with ice-cold PBS, followed by 4% paraformaldehyde (PFA). Brains were dissected and post-fixed in 4% PFA (24 h, 4 °C), then dehydrated in 30% sucrose (in PBS) at 4 °C until saturated. Tissue was embedded in Tissue-Tek OCT compound (14020108926, Leica) and frozen on dry ice. Coronal sections (8 µm for vascular and tight junction analysis; 30 µm free-floating for microglial and astrocytic analysis) were cut using a cryostat (CM3050-S, Leica). The sections were permeabilized with 0.5% Triton X-100, blocked with 5% BSA, then incubated overnight at 4 °C with the following primary antibodies: ZO-1 (1:500, Invitrogen, 40-2200), IBA1 (1:1,000, Wako, 019-19741), GFAP (1:1,000, Servicebio, GB11096-50). After washing, the sections were incubated with fluorescent secondary antibodies (1 h at room temperature; Alexa Fluor 594, 1:1,000, Invitrogen, A110012; Alexa Fluor 488, 1:1,000, Invitrogen, A11008). The sections were mounted with antifade mounting medium with DAPI (Vector Labs, H-1200-10) and coverslipped. Images were acquired using the LSM800 confocal microscope (ZEISS) or a slide imaging scanner (SLIDEVIEW VS200, Evident Scientific). The ZO-1 MFI was quantified using ImageJ and normalized to the controls. IBA1 and GFAP morphologies were analysed by Imaris-based 3D reconstruction and skeletonization (Bitplane). Microglia and astrocytes were counted per field of view and normalized to cells per mm².
Flow cytometry and antibodies
Single-cell suspensions from peritoneal lavage fluid, intestine and brain were filtered (70 μm) and centrifuged (400g, 5 min, 4 °C). Cells were blocked with anti-CD16/32 antibody (2.4G2, Elabscience, E-AB-F0997A) in FACS buffer on ice for 15 min. For surface marker staining, cells were incubated with fluorochrome-conjugated antibodies in FACS buffer (30 min, 4 °C). Viability was assessed using Live/Dead Fixable Violet Dead Cell Stain (Invitrogen, L34959). The following fluorescence conjugated anti-mouse CD45 (30-F11, 103112), CD4 (RM4-5, 100550), CD8 (53-6.7, 100752), CD11b (M1/70, 101210), F4/80 (BM8, 123110), Ly6G (1A8, 127606) and CD9 (MZ3,124815) antibodies were purchased from BioLegend. CD19 (1D3, 565965) was purchased from BD Biosciences. Data were collected using the BD LSRFortessa cytometer (BD Biosciences) or the SONY ID7000 Spectral Cell Analyzer, and analysed using FlowJo v.10.8.1.
Isolation of mouse intestine lamina propria cells and brain immune cells
Brain and intestinal tissues were dissected from anaesthetized mice and rinsed with cold PBS. For brain, single-cell suspensions were prepared using a Dounce homogenizer in RPMI 1640. For the intestine, segments were placed in Ca2+/Mg2+-free HBSS, and the mucosa was separated and incubated in a pre-digestion solution (containing 1 mM DTT and 2 mM EDTA in Ca2+/Mg2+-free HBSS) with shaking (200 rpm, 30 min, 37 °C). The intestines were washed twice with RPMI 1640 and then minced into small pieces. The digested tissues were then incubated in a digestion solution supplemented with 2 mg ml−1 collagenase IV (Yeasen, 40510ES76) and 5 U ml−1 DNase I (Sigma-Aldrich, D4263) with shaking (200 rpm, 1 h, 37 °C). Digested tissues were filtered through a 70-μm strainer (Absin, abs7306) and centrifuged (500g, 10 min, 18 °C). Immune cells were isolated by Percoll gradient centrifugation. In brief, 3 ml 30% stock isotonic Percoll (SIP; prepared as 9:1 (v/v) Percoll (Cytiva, 17089101) to 10× Ca2+/Mg2+-free HBSS) was added to the cell suspensions, layered onto 2 ml 70% SIP and centrifuged (500g, 30 min, 18 °C) with minimal/no braking. The 70–30% interphase was collected, diluted with 8 ml HBSS, centrifuged (500g, 7 min, 18 °C) and resuspended in 1 ml FACS buffer on ice.
Cell viability assay (MTT assay)
Cells were plated in 96-well microplates overnight (around 70% confluence) and exposed to twofold serial dilutions of toxins overnight at 37 °C. MTT (0.5 mg ml−1, Research Products International, M92050) was added and incubated for 4 h at 37 °C. A total of 100 μl solubilization solution (10% SDS in 0.01 M HCl) was then added to each well and incubated overnight at room temperature. Absorbance was measured at 580 nm using a microplate reader (BMG Labtech, FLUOstar Omega). Vehicle controls were included. Viability curves were analysed using Origin software (v.8.5).
Histology analysis and scoring
Tissue was fixed in 10% neutral-buffered formalin (24 h at room temperature), transferred to PBS (4 °C, 24 h), dehydrated in graded ethanol, cleared in xylene and embedded in paraffin. The sections (6 µm) were cut and stained with H&E. For intestine tissues, a well-established NEC histopathological scoring system (0 to 4) was applied: 0, no damage; 1, slight submucosal/lamina propria separation and villus tip epithelial sloughing; 2, moderate submucosal/lamina propria separation and mild villous necrosis; 3, severe submucosal/lamina propria separation, submucosal/muscular oedema and extensive villous sloughing; 4, complete villous loss and full thickness necrosis. For brain tissues, a 0–4 scale histopathological grading was performed based on the severity of neuronal damage and tissue lesions: 0, normal morphology; 1, minimal (sporadic neuronal shrinkage without structural disruption); 2, mild (focal neuronal loss in limited brain regions); 3, moderate (extensive neuronal damage with potential progression); 4, severe (diffuse neuronal necrosis involving the entire brain region examined).
RNA extraction and qPCR with reverse transcription
Total RNA from cells was extracted with TRIzol reagent (Magen, R4801). The RNA quantity and quality were measured by NanoDrop. cDNA was synthesized from 1 μg RNA using the StarScript III One-Step RT-PCR Kit for qPCR (Genstar, A235-01). Quantitative PCR (qPCR) was performed in a 384-well format on a Real-Time PCR Detection System (Bio-Rad) using 2×RealStar Power SYBR qPCR Mix (Genstar, A311-01). Relative quantitation was performed using the \({2}^{(-\varDelta \varDelta {C}_{{\rm{T}}})}\) method using Actb (for mouse cells) for normalization. A list of the specific RT–qPCR primers is provided in Supplementary Table 3.
ELISA
For serum antibody detection, 96-well plates were coated with Delta toxin (1 μg ml−1) in coating buffer (Solarbio, C1055) overnight at 4 °C. After washing in PBST, the plates were blocked with 5% BSA in PBS (1 h, room temperature) and then incubated with Delta toxin monoclonal antibody standards (1 μg ml−1, threefold serial dilution) or serum samples (starting at 1:100, threefold serial dilution) for 2 h at room temperature. After washing, HRP-conjugated goat anti-human IgG (H+L) (Yeasen, 33501ES60, 1:10,000) was added for 1 h at room temperature, followed by TMB (Solarbio, SEKF104) development and stop solution (Solarbio, SEKF101). The absorbance was measured at 450 nm (630 nm reference) using a TECAN SPARK reader. Antibody titres were determined from dilution curves.
For tissue Delta toxin quantification and distribution, tissues were weighed, homogenized in PMSF-supplemented PBS (10 μl mg−1) and subjected to three freeze–thaw cycles. After centrifugation (15,000g, 15 min, 4 °C), the supernatants were collected. The plates were coated with anti-Delta toxin mouse monoclonal antibody (1 μg ml−1) overnight at 4 °C and blocked. Serially diluted Delta toxin standards and tissue supernatants were added for 2 h at room temperature. HRP-conjugated rabbit anti-mouse IgG (H+L) (Yeasen, 771475ES60; 1: 10,000) was added for 1 h, followed by the above ELISA procedure. Tissue Delta toxin levels were normalized to uninfected controls.
Mouse IL-1β was measured using a high-sensitivity ELISA kit (ABclonal, RK04878) according to the manufacturer’s instructions.
Western blotting
Cells were lysed in RIPA buffer (Beyotime, P0013D) containing phosphatase (Roche, 4906845001) and protease inhibitors (Roche, 11836153001). Protein concentration was determined (Bio-Rad, 5000111) and normalized across samples with 20 μg protein per sample loaded. After transfer, the membranes were blocked with 5% non-fat milk in TBST (Tris-buffered saline with 0.5% Tween-20) and incubated overnight at 4 °C with primary antibodies. After washing, the membranes were incubated with HRP-conjugated secondary antibodies: anti-rabbit IgG (H+L) (Proteintech, SA00001-2) and anti-mouse IgG (H+L) (Proteintech, SA00001-1). Protein bands were visualized using Clarity Western ECL substrate (SPARK, TECAN) and imaged. The following primary antibodies were used: mouse monoclonal antibodies against GST (8-326, Thermo Fisher Scientific, MA4-004), GAPDH (1E6D9, Proteintech, 60004-1-Ig); rabbit monoclonal antibody against CD9 (SA35-08, HUABIO, HA750027), tissue factor (SN20-16, HUABIO, ET1611-95), GSDMD (E9S1X, CST, 39754), NLRP3 (PSH19-43, HUABIO, HA724075), caspase 1 (1O18, Proteintech, 81482-1-RR), IL-1β (EPR16805-15, Abcam, ab234437), phospho-NF-κB p65 (Ser536) (93H1, CST, 3033), phospho-AKT (Ser473) (D9E, CST,4060), phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) (D13.14.4E, CST, 4370), BAX (D2E11, CST, 5023), cleaved caspase 3 (Asp175) (5A1E, CST, 9664) and β-actin (13E5, CST, 4970); and cleaved N-terminal GSDMD polyclonal antibody (Wanleibio, WL05411), cleaved-caspase 1 polyclonal antibody (Zenbio, 341030) and β-tubulin polyclonal antibody (Proteintech, 10094-1-AP).
Genome-wide CRISPR–Cas9 screen and generation of KO cells
HeLa-Cas9 cells were generated using lentiviruses (LentiCas9-Blast, Addgene, 52962) and selected using 10 µg ml−1 blasticidin S (RPI, B12150.01). The GeCKO-V2 sgRNA library was obtained from Addgene (1000000049). Sublibraries A and B were independently packed into lentiviruses. HeLa-Cas9 Cells were transduced with sgRNA lentiviral libraries at an multiplicity of infection of 0.3. Infected cells were selected with 5 µg ml−1 puromycin (Thermo Fisher Scientific, A1113830) for 1 week. Then, 3.3 × 107 cells for sublibrary A or 2.9 × 107 cells for sublibrary B were plated onto 15 cm culture dishes to ensure enough sgRNA coverage. These cells were either saved as an initial library control or exposed to 0.01 µg ml−1 Delta toxin for 24 h. The surviving cells were washed and reseeded within toxin-free medium until about 70% confluence, followed by the next round of selection with 0.01 µg ml−1 Delta toxin for 24 h. The genomic DNA of surviving cells was extracted using a commercial kit (Qiagen, 13323). The DNA fragments containing the sgRNA sequences were amplified by PCR using primers lentiGP-1_F (5′-AATGGACTATCATATGCTTACCGTAACTTGAAAGTATTTCG-3′) and lentiGP-3_R (5′-ATGAATACTGCCATTTGTCTCAAGATCTAGTTACGC-3′). Next-generation sequencing was performed by a commercial vendor (Genewiz, Illumina HiSeq). The selected sgRNA sequences (LRP6: TGGTTCTGACCGTGTAGTAT; CLEC12B: CATTGCCACATCTTAGGACA) were cloned into the LentiGuide-Puro vector (Addgene, 52963). The KO cells were generated by lentiviral transduction of sgRNAs into HeLa-Cas9 cells. Mixed populations of transduced cells were selected with puromycin (5 µg ml−1). A list of CRISPR–Cas9 screening results is provided in Supplementary Table 4.
Cell surface binding and immunostaining
Cells were seeded onto glass coverslips (Hampton, HR3-239). The constructs were transduced into cells for 24 h and exposed to GST-Delta toxin (0.4 μg ml−1) on ice for 60 min. After washing with cold PBS, cells were fixed with 4% PFA for 20 min, blocked with 10% goat serum for 40 min and stained with primary antibodies against GST and Flag (1:500) for 1 h, followed by fluorescence-labelled secondary antibodies for 1 h. Slides were mounted with DAPI-containing medium (SouthernBiotech, 0100-20). Images were captured using the Olympus DSU-IX81 spinning-disk confocal system, pseudocoloured and analysed using ImageJ (v.1.52o).
Pull-down and immunoblot analysis
Cell lysates were collected in 1 ml lysis buffer (PBS, 1% Triton X-100, 0.1% SDS, plus a protease inhibitor cocktail (Sigma-Aldrich), 1 ml per 10 cm dish) and incubated with 20 μg GST-tagged Delta toxin proteins for 4 h at 4 °C. Pull-down experiments were performed using 15 µl glutathione agarose beads. Beads were washed, pelleted and boiled. The samples were subjected to SDS–PAGE and western blotting. Blot images were taken using the Fuji LAS3000 imaging system and analysed using ImageJ (v.1.52o).
BLI assays
The binding between Delta toxin and human GST–CD9 was measured using a BLI assay with the BLItz system (ForteBio). In brief, 10 μg ml−1 biotin-labelled human GST–CD9 (Abcam, ab152262) was immobilized onto capture biosensors (GST Biosensors, ForteBio) and washed with Dulbecco’s PBS (DPBS) containing 0.5% BSA. Empty biosensors served as a control. These biosensors were then exposed to variable concentrations of Delta toxin in solution, followed by dissociation in DPBS (0.5% BSA). Binding affinities (KD) were estimated using the BLItz software (ForteBio).
Preparation of polyclonal antibodies against Delta toxin
Mice (BALB/c, aged 6 weeks) were immunized subcutaneously with 100 µg of purified, non-toxic mutant Delta toxin (V17E/T18E) emulsified in complete Freund’s adjuvant (CFA, Sigma-Aldrich, AR001), followed by two booster doses (50 µg in Incomplete Freund’s Adjuvant, Sigma-Aldrich, F5506) at 2-week intervals. Polyclonal antibodies were purified using protein G affinity chromatography. The final antibody titre exceeded 1:100,000 as determined by ELISA.
Preparation of monoclonal antibodies against Delta toxin
BALB/c mice (aged 6 weeks) were immunized subcutaneously with 100 µg of non-toxic Delta toxin mutant (V17E/T18E) in complete Freund’s adjuvant, followed by four boosters in incomplete Freund’s adjuvant. Serum titres were monitored by ELISA. Splenocytes from the highest-titre mouse were fused with SP2/0 myeloma cells. The fused cells were seeded into plates and selected in RPMI 1640-HAT medium (Sigma-Aldrich, H0262) supplemented with 10%FBS, 1% penicillin–streptomycin and gentamicin for 2 weeks. The supernatants were screened by ELISA and cytotoxicity neutralization assays to identify functional antibody-secreting clones. Positive hybridomas were subcloned by limiting dilution and amplified in RPMI 1640-HT medium (Sigma-Aldrich, H0137) supplemented with 10% FBS and 1% penicillin–streptomycin. Stable clones were characterized by ELISA and isotyped, with one lead clone chosen for small-scale antibody purification through protein A affinity chromatography, followed by large-scale production in ascitic fluid, yielding 1 mg of purified antibody. All hybridoma lines were cryopreserved at 1 × 106 cells per vial for long-term storage.
Quantification and statistical analysis
All experiments were performed at least three times. Data are presented by descriptive statistics such as mean ± s.d. Statistical analyses were performed using the GraphPad Prism v.9.0. Statistical details of the experiments are presented in the relevant figure legends and Methods sections for each specific technique.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
