Integrated IP-MS and Optoproteomics Reveal Dynamic Remodeling of Protein Interaction Networks During NLRP3 Inflammasome Activation Authors: Yuehui Zhang, Kairong Zhang, Chaozhi Jin, Limin Shang, Shanshan Wu, Mingyang Zhang, Yanzhi Yuan, Yuchen Liu, Jian Wang Published: 2026 ABSTRACT AbstractNOD-like receptor family pyrin domain-containing 3 (NLRP3) is a key intracellular sensor that assembles inflammasome complexes to activate inflammatory caspases, yet the molecular composition and stimulus-dependent organization of the NLRP3 inflammasome remain incompletely defined. Here, we systematically characterized dynamic remodeling of the NLRP3 inflammasome interactome using endogenous immunoprecipitation coupled with mass spectrometry (IP-MS) targeting both NLRP3 and its adaptor protein apoptosis-associated speck-like protein containing a CARD (ASC) in bone marrow-derived macrophages. Quantitative interactome profiling across NLRP 3 inflammasome priming and activation states identified 1496 stimulus-associated protein interactions (1032 non-redundant), revealing extensive and stage-specific reorganization of NLRP3- and ASC-centered interaction networks. Integration of these datasets with microscopy-guided in situ protein photo-labeling (optoproteomics) of ASC specks provided spatial context to inflammasome assembly and identified inner membrane mitochondrial protein (IMMT) as an activation-associated binding partner of both NLRP3 and ASC. Functional analyses demonstrated that Immt depletion enhanced caspase-1 and gasdermin D cleavage, interleukin-1β secretion, lactate dehydrogenase release, and ASC oligomerization in response to NLRP3 agonist, indicating a negative regulatory role for IMMT in inflammasome activation. Together, this study establishes a stimulus-resolved proteomic framework for NLRP3 inflammasome interaction networks and reveals mitochondrial IMMT as a regulator of inflammasome assembly and inflammatory signaling. Innate immune responses to infection and tissue damage rely on the ability of host cells to sense a broad spectrum of pathogenic and endogenous danger signals (1). This surveillance is mediated by pattern recognition receptors (PRRs), which initiate inflammatory signaling cascades upon detection of pathogen-associated or danger-associated molecular patterns. Among these receptors, NLRP3 has emerged as a central intracellular sensor that assembles a multiprotein inflammasome complex to drive inflammatory caspase activation and pyroptotic cell death (2). Assembly and activation of the NLRP3 inflammasome occur through a tightly regulated two-step process consisting of priming and activation (3). Priming signals, delivered through receptors such as Toll-like receptors, TNF (tumor necrosis factor) receptors, or IL-1 (interleukin-1) receptors, induce NF-κB-dependent transcriptional upregulation of NLRP3 and pro-inflammatory cytokines, including pro-IL-1β and pro-IL-18 (4). Subsequent exposure to diverse secondary stimuli, including extracellular ATP, nigericin, monosodium urate crystals, and a wide range of microbial pathogens, triggers inflammasome assembly, leading to caspase-1 activation, processing of IL-1β and IL-18, cleavage of gasdermin D, and execution of pyroptosis (5, 6, 7). At the molecular level, NLRP3 inflammasome assembly is orchestrated through coordinated protein-protein interactions involving NLRP3, ASC, NEK7 (NIMA-related kinase 7), and caspase-1. While genetic and biochemical studies have established essential roles for these core components, the broader interaction landscape that governs inflammasome assembly, regulation, and signal propagation remains incompletely defined. In particular, how NLRP3- and ASC-centered interaction networks are dynamically reorganized across distinct activation stages and inflammatory stimuli remains a major unresolved question in the field. Mass spectrometry-based protein interaction mapping has provided valuable insights into inflammasome-associated proteins (8, 9). However, most existing studies have been constrained by experimental design. Many interaction datasets rely on overexpression systems, engineered affinity tags, or non-inflammatory cell lines, limiting their relevance to physiological inflammasome activation. Proximity labeling strategies, such as APEX- or BioID-based approaches, have further expanded the capacity to capture transient and spatially restricted interactions, yet their application to NLRP3 inflammasome biology has largely been confined to heterologous systems or single activation conditions (10, 11). Previously, we applied thermal proteome profiling in combination with IP-MS to probe NLRP3 inflammasome regulation (12). While these studies demonstrated the value of unbiased proteomics for identifying inflammasome regulators, a comprehensive and stimulus-resolved characterization of dynamic NLRP3 and ASC interaction networks in primary inflammatory cells remains lacking. Here, we address these limitations by integrating endogenous IP-MS and microscopy-guided optoproteomics to define the dynamic and spatially resolved interaction landscape of the NLRP3 inflammasome in bone marrow-derived macrophages (BMDMs). By profiling both NLRP3- and ASC-associated protein interactions under priming and multiple canonical activation conditions, and spatially mapping proteins proximal to assembled ASC specks, we uncover stimulus-dependent remodeling of inflammasome interaction networks and identify previously unrecognized regulatory components that modulate inflammasome activation. Experimental Procedures Cell Culture Primary BMDMs were isolated from femurs of C57BL/6J mice and differentiated in RPMI 1640 medium (Gibco, #C11875500BT) supplemented with 10% fetal bovine serum (FBS; Gibco, #A5256701), 50 ng/ml macrophage colony-stimulating factor (M-CSF; Novoprotein, #CB34), and 1% penicillin-streptomycin (Cell World, #C0160-611) for 7 days. Approximately 1.5 × 107 BMDMs can be obtained from each mouse following isolation and induction. Immortalized bone marrow-derived macrophages (iBMDMs) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, #11995500BT) supplemented with 10% FBS and 1% penicillin-streptomycin. All cells were maintained at 37 °C in a humidified incubator with 5% CO2. Antibodies The following primary antibodies were used: anti-NLRP3 (AdipoGen, #AG-20B-0014), anti-ASC (AdipoGen, #AG-25B-0006), anti-caspase-1 (p20) (AdipoGen, #AG-20B-0042), anti-IL-1β (Abcam, #ab9722), anti-gasdermin D (GSDMD; Abcam, #ab209845), anti-ASC (Selleck, #F0468), anti-IMMT (Proteintech, #68226-1-Ig), anti-GAPDH (HRP-conjugated; Proteintech, #HRP-60004), normal rabbit IgG (Proteintech, #B900610), and normal mouse IgG (Beyotime, #A7028). Secondary antibodies included goat anti-rabbit IgG (heavy chain; EASYBIO, #BE0106), HRP-conjugated goat anti-mouse IgG (R&D Systems, #HAF007), Alexa Fluor 488-conjugated goat anti-mouse IgG (Jackson ImmunoResearch, #115-545-003), and Cy3-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch, #711-165-152). Inflammasome Priming and Activation in Macrophages Primary BMDMs were seeded in 12-well plates at a density of 5 × 105 cells per well in RPMI 1640 medium supplemented with 10% FBS. iBMDMs were seeded in 12-well plates at a density of 2.5 × 105 cells per well in DMEM supplemented with 10% FBS. Cells were cultured overnight at 37 °C in a humidified incubator with 5% CO2. On the following day, the culture medium was replaced with Opti-MEM (Cell World, #C2656-862) prior to inflammasome stimulation. For canonical NLRP3 inflammasome activation, BMDMs were primed with LPS (400 ng/ml; Sigma-Aldrich, #L4391) for 4 h, followed by stimulation with ATP (5 mM; Roche, #10127523001) for 30 min, nigericin (10 μM; InvivoGen, #tlrl-nig) for 30 min, or monosodium urate (MSU) crystals (200 μg/ml; InvivoGen, #tlrl-msu) for 5 h. For iBMDMs, cells were primed with LPS (1 μg/ml) for 5 h and subsequently stimulated with nigericin (20 μM) for 1 h. Following stimulation, culture supernatants and cell lysates were collected separately and subjected to immunoblotting analysis or IP-MS. Affinity Purification of NLRP3 and ASC Complexes for Mass Spectrometry To identify proteins associated with the NLRP3 or ASC proteins, affinity purification was performed using lysates from BMDMs subjected to inflammasome priming and activation as described above. Briefly, cells were lysed by sonication in ice-cold IP lysis buffer, and cell debris was removed by centrifugation at 12,000 rpm for 10 min at 4 °C. A small aliquot of the supernatant was mixed with 2 × SDS loading buffer and boiled for input analysis. The remaining cleared lysates were incubated overnight at 4 °C with protein A/G magnetic beads (MedChemExpress, #HY-K0202) pre-bound to anti-NLRP3 or anit-ASC antibodies. Following incubation, the beads were washed twice with IP buffer containing 100 mM NaCl and once with IP buffer containing 500 mM NaCl to reduce nonspecific binding. Immunoprecipitated protein complexes were eluted by resuspending the beads in 100 mM NaCl IP buffer supplemented with an equal volume of 2 × SDS loading buffer, followed by boiling at 100 °C for 10 min under denaturing conditions. Eluted proteins were resolved by SDS-PAGE and visualized by gel staining. Gel bands were excised based on staining intensity, pooled into fractions and subjected to overnight digestion with sequencing-grade trypsin at 37 °C. The resulting peptides were extracted, desalted, and prepared for LC-MS/MS analysis. Three replicates were performed per group, with each replicate requiring 1.5 × 107 BMDMs for a total of 5.4 × 108 cells derived from 34 mice. LC-MS/MS Analysis of Immunoprecipitated Samples LC-MS/MS analysis was performed on a Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific) coupled online to an EASY-nLC 1000 nanoflow liquid chromatography system (Thermo Fisher Scientific). Peptides were initially loaded onto a 1 cm self-packed trap column (150 μm inner diameter, 1.9 μm ReproSil-Pur C18-AQ resin; Dr Maisch GmbH) in solvent A (0.1% formic acid in HPLC-grade water). After loading and washing, peptides were separated on a 15 cm analytical column (150 μm inner diameter, 1.9 μm ReproSil-Pur C18-AQ resin; Dr Maisch GmbH) using a 78 min nonlinear gradient from 5% to 95% solvent B (0.1% formic acid in acetonitrile) at a flow rate of 600 nl/min. Peptides were ionized by electrospray ionization at a spray voltage of 2.0 kV, with the capillary temperature set to 320 °C. Data were acquired in data-dependent acquisition mode using Orbitrap detection for both MS1 and MS2 scans (OT-OT mode). Full MS scans were acquired over an m/z range of 300-1400 at a resolution of 70,000, with an automatic gain control (AGC) target of 3 × 106 and a maximum injection time of 60 ms. The top 20 most intense precursor ions were selected for higher-energy collisional dissociation (HCD) fragmentation with a normalized collision energy of 27%. Dynamic exclusion was set to 18 s. MS2 spectra were acquired in the Orbitrap at a resolution of 17,500, with an AGC target of 5 × 104 and a maximum injection time of 80 ms. IP-MS Data Analysis Raw MS data were searched against the UniProt mouse (Mus musculus) protein database (release 20260128, containing 25,710 protein sequences including isoforms) using Mascot software (version 2.6.0). Trypsin was specified as a proteolytic enzyme, with up to three missed cleavages allowed. Precursor ion charges were restricted to 2+, 3+, and 4+. The precursor mass tolerance was set to 20 ppm, and the fragment ion mass tolerance was set to 0.05 Da. Monoisotopic mass calculation and MS/MS ion search were enabled, using Mascot generic format and ESI_TRAP instrument settings. No fixed modification was specified because no reduction and alkylation treatment was performed. Variable modifications included methionine oxidation, protein N-terminal acetylation, and cysteine destreak. Peptide-spectrum matches were filtered to a false discovery rate (FDR) of 1% using Pepdistiller (13), and protein-level FDR was controlled at 1% using PANDA (14). An ion score cutoff of 10 was applied for peptide identification. Proteins identified with fewer than two unique peptides across all combined biological replicates were excluded from subsequent analyses. To identify high-confidence protein-protein interactions, spectral count data were analyzed using SAINTexpress (15). Interactions with a SAINT AvgP score ≥0.9 were considered high confidence and retained for downstream analyses. Differential interaction scores (DISs) were calculated from SAINTexpress results as previously described (16). For each bait-prey pair, a condition-specific interaction probability (S) was derived from the SAINT AvgP value under each condition. To quantify differential enrichment between two conditions A and B, a raw DIS was calculated as: raw DISAtoB = SA × (1 − SB). This metric captures interactions that are strongly enriched under condition A but absent or weak under condition B. Because interaction changes may occur in either direction, a signed DIS was determined by comparing the two opposing raw scores (raw DISAtoB vs raw DISBtoA). The larger of the two values was retained, and its sign was used to indicate the favored condition: positive values indicate enrichment in the comparison condition (e.g., Nigericin-, ATP-, or MSU-stimulated), whereas negative values indicate enrichment in the reference condition (LPS-primed). A cutoff of |DIS| > 0.5 was applied to define condition-specific interactions. Protein interaction networks were visualized using Cytoscape (version 3.10.1) (17). Microscopy-Guided Photo-Labeling of ASC Specks (Optoproteomics) For optoproteomic labeling of proteins proximal to ASC specks, approximately 1.5 × 106 BMDMs were seeded per well in one-well chamber slides (Cellvis, #C1-1.5H-N), totaling 4.5 × 107 cells derived from 3 mice. For each biological replicate, five chamber slides were subjected to photo-labeling (PL), and an additional five slides were processed in parallel as unlabeled controls (UL). Three biological replicates were performed per group. Following overnight culture, the medium was replaced with OPTI-MEM (Cell World, #C2656-862), and cells were primed with LPS (400 ng/ml) for 4 h, followed by stimulation with nigericin (10 μM) for 30 min to induce ASC speck formation. Cells were washed with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 15 min at room temperature. After two PBS washes, cells were permeabilized with PBS containing 0.2% Triton X-100 for 15 min at 4 °C. Cells were then blocked in PBS supplemented with 0.1% Tween-20 and 3% bovine serum albumin (BSA) for 1 h at room temperature. Microscopy-guided photo-labeling was performed using the Synlight-Rich kit (SynCell, #SYN-PU0106) according to the manufacturer’s instructions. Briefly, cells were sequentially incubated with Block 1 and Block 2 reagents for 30 min and 15 min, respectively, followed by incubation with rabbit anti-ASC primary antibody at 4 °C overnight. After three washes with PBS, cells were incubated with Cy3-conjugated donkey anti-rabbit secondary antibody for 1 h at room temperature. ASC specks were identified using the image processing function of the Microscoop system, which automatically generated spatial foci corresponding to ASC specks and guided localized photo-labeling within these regions across the entire chamber slide. Following completion of photo-labeling, reactions were quenched using the quencher solution (three times, 5 min each), and cells were thoroughly washed with PBS containing 0.1% Triton X-100 to remove unreacted probes prior to downstream affinity purification. Unlabeled control samples were processed in parallel under identical conditions. To assess biotinylation efficiency, labeled cells were stained with the “Verify” reagent included in the Synlight-Rich kit for 1 h at room temperature and imaged by fluorescence microscopy (Nikon TE2000). For confocal imaging, nuclei were counterstained with DAPI (Sigma-Aldrich, D9542). Streptavidin Enrichment and On-Bead Digestion of Photolabeled Proteins Photolabeled and unlabeled control cells were scraped and collected into separate tubes using scraping buffer (Solution A, Synpull reagent kit, SynCell). Samples were centrifuged at 5000×g for 3 min at 4 °C, after which the supernatant was removed, leaving approximately 100 μl residual volume with the cell pellet. Cells were lysed in SDT buffer (50 mM Tris-HCl, 2% SDS, 50 mM DTT, pH 7.8) followed by sonication and heating at 95 °C for 10 min to ensure complete protein denaturation. To reduce the SDS concentration to below 0.5%, lysates were diluted with RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, pH 7.5). Insoluble material was removed by centrifugation at 4 °C for 10 min, and the clarified supernatants were transferred to fresh tubes. Streptavidin beads were added to each sample, and biotinylated proteins were enriched by incubation at 4 °C for 2 h with gentle rotation. Beads were washed sequentially three times with RIPA buffer, once with RIPA buffer containing 1% SDS, once with RIPA buffer containing 1 M NaCl, and finally twice with 50 mM ammonium bicarbonate (NH4HCO3). Reduction and alkylation were performed directly on beads by incubation with 5 mM DTT for 30 min at room temperature, followed by 10 mM chloroacetamide (CAA) for 30 min in the dark. After two additional washes with ammonium bicarbonate, proteins were digested on-bead with sequencing-grade trypsin at 37 °C for 2 h. The supernatant containing released peptides was transferred to a new tube. The beads were rinsed with half the initial trypsin volume, and the rinse was combined with the first digest. The pooled peptides were further digested overnight at 37 °C with continuous agitation. Digestion was terminated by adding trifluoroacetic acid (TFA) to a final concentration of 0.3% (v/v). Peptides were desalted using C18 StageTips and dried by vacuum centrifugation prior to LC–MS/MS analysis. DIA-MS and Data Analysis of ASC-Speck Proximal Proteins Peptides were reconstituted in 0.1% formic acid and analyzed using a nanoflow liquid chromatography system (UltiMate 3000, Thermo Fisher Scientific) coupled online to a trapped ion mobility spectrometry, time-of-flight mass spectrometer (timsTOF Pro 2, Bruker Daltonics). For each run, peptides were loaded onto a reversed-phase C18 analytical column (25 cm × 75 μm inner diameter, 1.9 μm particle size) maintained at 60 °C. Peptide separation was performed at a flow rate of 300 nl/min using a 60-min linear gradient, with mobile phase A consisting of water containing 0.1% formic acid (v/v) and mobile phase B consisting of 80% acetonitrile with 0.1% formic acid (v/v). In the 60 min gradient, peptides were eluted from 6% to 30% B over 48 min, followed by an increase to 40% B over 12 min and a final ramp to 99% B over 3 min. Mass spectrometric data were acquired in data-independent acquisition mode using parallel accumulation-serial fragmentation (DIA-PASEF). The LC system was coupled online with a timsTOF Pro 2 mass spectrometer operated under the following conditions: capillary voltage, 1400 V; dry gas flow, 3 L/min; dry gas temperature, 180 °C. TIMS separation was performed over an ion mobility range of 1/K0 = 0.6 to 1.6 V s/cm2. MS1 spectra were acquired over a mass range of m/z 100–1700. For DIA-PASEF acquisition, a total of 32 MS/MS windows were defined per cycle, each with a width of 25 Da and a 1 Da overlap, corresponding to 32 mass steps per cycle and a total cycle time of approximately 1.8 s. Collision energy was ramped linearly as a function of ion mobility, from 20 eV at 1/K0 = 0.6 V s/cm2 to 59 eV at 1/K0 = 1.6 V s/cm2. Specifically, ions with 1/K0 < 0.6 V s/cm2 were fragmented at 20 eV, ions within 0.6 to 1.6 V s/cm2 were subjected to a linear energy ramp from 20 to 59 eV, and ions with 1/K0 > 1.6 V s/cm2 were fragmented at 59 eV. Raw DIA files from both photolabeled and unlabeled control samples were analyzed together using Spectronaut (version 18.0.230605.50606) in directDIA mode. Protein identification was performed using the Pulsar search engine against a UniProt mouse (M. musculus) database (release 20260128, containing 25,710 protein sequences including isoforms). The enzymes used were trypsin and LysC (trypsin/P), with cleavage specificity and a maximum of two missed cleavages allowed. DIA data were analyzed using Spectronaut with dynamic mass tolerance settings enabled for both precursor and fragment ion extraction. Mass tolerances were automatically optimized by Spectronaut based on instrument mass accuracy and nonlinear retention time calibration. Carbamidomethylation of cysteine residues was specified as a fixed modification, while protein N-terminal acetylation, deamidation of asparagine or glutamine, and oxidation of methionine were included as variable modifications. A False Discovery Rate (FDR) threshold of 0.01 (1%) was applied at the PSM (peptide spectrum match), peptide, and protein group levels. Protein groups identified with fewer than two unique peptides across all combined biological replicates were excluded from subsequent analyses. Protein group-level intensity-based absolute quantification (iBAQ) values exported from Spectronaut were used for downstream quantitative analysis. Quantitative values were log2-transformed prior to statistical analysis. Differential protein abundance between experimental groups was assessed using the limma package in R. Linear models were fitted for each protein, and empirical Bayes moderation was applied to stabilize variance estimates. A design matrix reflecting experimental group assignments was constructed, and contrasts were defined to compare photolabeled versus unlabeled samples. Resulting p-values were adjusted for multiple testing using the Benjamini-Hochberg method. Proteins with an absolute fold change >1.3 and an adjusted p-value <0.05 were considered significantly enriched proximal proteins. Co-immunoprecipitation Assay iBMDMs were lysed in ice-cold lysis buffer containing 50 mM Tris–HCl (pH 7.5), 150 mM NaCl, 1.5 mM MgCl2, 0.2% (v/v) NP-40, 5% (v/v) glycerol, and a protease inhibitor cocktail (Selleck, #B14002). Lysates were sonicated on ice (30 W, 4 s on/6 s off cycles for a total of 1 min) and clarified by centrifugation at 20,000×g for 10 min at 4 °C. The resulting supernatants were pre-cleared by incubation with protein A/G agarose beads (MedChemExpress, #HY-K0202) and normal IgG at 4 °C for 2 h. After removal of the beads, pre-cleared lysates were incubated with fresh protein A/G beads and the indicated antibodies overnight at 4 °C. Immunoprecipitated proteins were washed and subsequently analyzed by Western blotting. siRNA-Mediated Gene Silencing Small-interfering RNAs (siRNAs) targeting mouse Immt (si-Immt-1: 5′-GCACACUCCAACAUACUGAAATT-3′; si-Immt-2: 5′-CCCGUUUCUAUGCUGUUCAAATT-3′) and a non-targeting negative control siRNA were synthesized by GenePharma. For siRNA transfection, iBMDMs were transfected with 100 nM siRNA using CALNP RNAi in vitro transfection reagent (Beijing D-Nano Therapeutic, #DN001-10) according to the manufacturer’s instructions. Cells were incubated for 42 h post-transfection prior to subsequent inflammasome activation or downstream analyses. Cytokine Quantification and LDH Release Assays Cell culture supernatants were collected following NLRP3 inflammasome activation and analyzed for mouse IL-1β (Solarbio, #SEKM-0002), TNF-α (Solarbio, #SEKM-0034), and IL-6 (Solarbio, #SEKM-0007) using enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturers’ instructions. Lactate dehydrogenase (LDH) release was measured in culture supernatants immediately after stimulation using the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega, #G1780). LDH release was calculated relative to a maximum lysis control, with background absorbance values from blank culture medium subtracted. ASC Oligomerization Assay For the ASC oligomerization assay, iBMDMs cultured in 6-well plates were primed with LPS (1 μg/ml) for 5 h, followed by stimulation with nigericin (20 μM) for 1 h. Cells were lysed in Tris-buffered saline (TBS) containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5% Triton X-100, and EDTA-free protease inhibitor cocktail. Cell lysates were centrifuged at 8000×g for 15 min at 4 °C. The resulting pellets were washed twice with PBS and resuspended in 60 μl PBS containing 4 mM disuccinimidyl sulfoxide (DSS; Thermo Fisher Scientific, #A33545). Crosslinking was carried out at 37 °C for 30 min and terminated by centrifugation at 8000×g for 15 min. The pellets were dissolved in SDS sample buffer and subjected to immunoblot analysis. Immunofluorescence Microscopy Cells were washed with PBS and fixed with 4% paraformaldehyde for 15 min at room temperature. After fixation, cells were washed with PBS and permeabilized with 0.3% Triton X-100 for 15 min on ice. Non-specific binding was blocked by incubation with 3% BSA in PBS for 1 h at room temperature. Cells were then incubated with the indicated primary antibodies overnight at 4 °C, followed by incubation with appropriate fluorophore-conjugated secondary antibodies for 1 h at room temperature. Nuclei were counterstained with DAPI. Fluorescence images were acquired using an LSM 880 laser scanning confocal microscope (Zeiss). Structure Prediction Using AlphaFold3 Previously reported PPIs identified by IP-MS were selected for structural modeling. Amino acid sequences of the interacting proteins were retrieved from UniProt. Structural predictions of protein complexes were performed using AlphaFold3 (AF3) in multimer mode with default parameters, which include automated multiple sequence alignment generation and template searching. For each protein complex, multiple structural models were generated by AF3. The top-ranked model was selected based on the predicted TM-score (pTM) and interface predicted TM-score (ipTM), which were subsequently used for downstream structural inspection and comparative analyses. All structure predictions were conducted using the public AlphaFold3 server (https://alphafoldserver.com/). Experimental Design and Statistical Rationale All experiments were designed with appropriate biological replicates and controls. For IP-MS and optoproteomics profiling, three independent biological replicates were performed for each experimental condition (e.g., LPS priming, ATP, nigericin, and MSU stimulation). Unlabeled controls were included in optoproteomics experiments to distinguish true proximal labeling from background biotinylation. Quantitative proteomic comparisons were conducted using normalized intensity data (iBAQ for DIA; spectral counts and SAINT AvgP values for IP-MS). DIS were computed to assess condition-specific protein–protein interactions, and significance thresholds were applied as described in the corresponding Methods subsections. For functional enrichment analysis, differentially interacting proteins were subjected to Gene Ontology (GO) and pathway enrichment using the clusterProfiler package in R (18). Multiple testing corrections for enrichment analyses were performed using the Benjamini-Hochberg method, and terms with adjusted p values < 0.05 were considered significant. To contextualize NLRP3 and ASC interaction data within existing knowledge, publicly curated protein interaction datasets were retrieved from BioGRID (19), STRING (20), and IntAct (21). Quantitative data from cytokine ELISAs and LDH release assays are presented as mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism 9. Unless otherwise specified, comparisons between more than two groups were assessed by one-way analysis of variance (ANOVA). A p value < 0.05 was considered statistically significant. Results Stimulus-Resolved Remodeling of the NLRP3 and ASC Interactomes During Inflammasome Activation To systematically define how inflammasome activation reshapes protein–protein interaction (PPI) networks, we performed IP-MS to profile NLRP3- and ASC-associated proteins in BMDMs under defined activation conditions. Cells were primed with LPS and subsequently exposed to canonical NLRP3-activating signals, including nigericin, ATP, or MSU, enabling stimulus-resolved interrogation of inflammasome-associated interactomes. Mass spectrometry data quality was stringently assessed prior to downstream analysis. A single MSU-treated replicate with low protein and peptide identifications was excluded from further analysis. Peptide and protein identifications were consistent across other biological replicates, and replicate reproducibility was confirmed by high Pearson correlation coefficients (Supplemental Fig. S1 and Supplemental Table S1A). Using antibodies against NLRP3 or ASC, we generated six condition-specific interactomes corresponding to distinct combinations of target protein and activation signal (Fig. 1A). Interaction confidence was evaluated using SAINTexpress (15), and interactions with AvgP ≥0.9 were retained as high-confidence PPIs (Supplemental Table S1B). Systematic mapping of NLRP3-ASC interactions across inflammatory stimuli.A, workflow of the interaction-screening pipeline for NLRP3 and ASC under distinct stimulation conditions, including sample preparation, affinity purification, mass spectrometry, and computational scoring. B, comparison of the number of NLRP3 and ASC interacting proteins identified in this study versus previously reported interactors in major public databases. “Recall” indicates previously reported interactions recovered in our dataset. C, recovered NLRP3 interactions across different stimulation conditions. D, UpSet plot showing overlap of PPIs identified under each stimulation condition. E, distribution of Jaccard indices illustrating pairwise similarity of interactomes across stimulation conditions. To place the experimentally derived interactomes in the context of existing knowledge, we compared our dataset with curated NLRP3 and ASC interaction data from BioGRID, STRING, and IntAct. Across all conditions, a total of 1496 PPIs were identified, corresponding to 1032 non-redundant interactions, in contrast to 81 previously reported interactors (Fig. 1B; Supplemental Fig. S2A). Notably, most known PPIs were originally identified in unstimulated cells or heterologous expression systems, which likely explains the limited overlap with our stimulus-resolved macrophage dataset. Totally 16 of the 81 known interactions (19.75%) were recalled (Fig. 1C; Supplemental Fig. S2B, Supplemental Table S1C) by this study. Most of the PPIs identified in this study have not been previously reported, suggesting that inflammasome-associated interaction landscapes may undergo substantial remodeling in an activation-dependent manner. Quantitative comparison across conditions revealed pronounced stimulus-specific differences in interactome size and composition (Supplemental Fig. S2C). Among NLRP3 interactomes, MSU stimulation yielded 969 PPIs, followed by ATP stimulation (368 PPIs), whereas LPS priming alone and nigericin stimulation obtained 80 and 27 PPIs, respectively. A total of 33 and 19 PPIs were detected for ASC-associated interactomes under LPS and nigericin stimulation, respectively. These differences indicate that the distinct activating signals strongly influence interaction network complexity. Overlap analysis further revealed that NLRP3 interactomes induced by MSU and ATP shared a substantial number of 294 common interactors (Fig. 1D). In contrast, overlap between other NLRP3 conditions was minimal (≤30 shared interactors), and the nigericin-induced NLRP3 interactome exhibited particularly limited similarity to those induced by MSU or ATP, suggesting engagement of a more restricted or distinct interaction program. Despite their obligatory co-assembly within the inflammasome, the NLRP3 and ASC interactomes exhibited minimal overlap under the same stimulation conditions, with Jaccard similarity indices consistently below 0.03 (e.g., NLRP3/LPS vs. ASC/LPS = 0.02). By comparison, higher similarity was observed between different activation states of the same bait protein (e.g., NLRP3/MSU vs. NLRP3/ATP = 0.35; Fig. 1E). Given the increasing application of AlphaFold-based approaches for PPI prediction, we assessed whether predicted structural compatibilities aligned with experimentally observed NLRP3 and ASC interactions. All previously reported interactions, including those reproduced in our dataset, were assigned uniformly low confidence scores by AlphaFold (Supplemental Fig. S2D; Supplemental Table S1C). This discrepancy likely reflects fundamental limitations of current structure prediction frameworks, which are optimized for stable, well-defined binary complexes. In contrast, inflammasome assembly involves highly dynamic, signal-dependent, and often transient interactions that may be mediated indirectly or stabilized only within higher-order assemblies. Consistent with this result, several interactions previously validated under inflammasome-activating conditions were detected in our dataset. For example, HERC6 (HECT and RLD domain containing E3 ubiquitin protein ligase family member 6) and TRIM28 (tripartite motif containing 28) have been reported to associate with NLRP3 in LPS-primed macrophages, with interactions persisting following ATP stimulation (22, 23). Similarly, the tyrosine kinase Syk has been shown to interact with both NLRP3 and ASC and to promote inflammasome assembly in response to canonical stimuli (24), including in pathogen-induced contexts (25). The reproducible detection of these interactions under matched stimulation conditions underscores the ability of our IP-MS strategy to capture biologically relevant, context-dependent protein assemblies that are currently inaccessible to purely computational prediction approaches. Stimulus-Specific Perturbation Profiling Reveals Distinct NLRP3 Interaction Programs To quantitatively characterize how distinct activation signals remodel the NLRP3 and ASC interaction landscapes upon LPS priming, we applied a DIS framework to compare condition-specific interaction strengths across activation states (Fig. 2A; Supplemental Table S2A). In this framework, positive DIS values indicate interactions preferentially enriched upon canonical NLRP3 stimulation, whereas negative values indicate interactions enriched during LPS priming alone. Interactions with |DIS| > 0.5 were considered conditionally preferential and retained for downstream analyses. Quantitative scoring and perturbation profiling of inflammasome-dependent PPIs.A, Workflow for computing DIS to identify stimulus-specific interactions. B, Scatterplots of SAINTexpress AvgP scores for all PPIs under LPS priming and activation stimulation of NLRP3 inflammasome. C, Density distribution of DIS values, illustrating global interaction perturbation patterns across stimuli. D, Scatterplots highlighting condition-dependent interaction changes for PPIs under different inflammasome stimuli. E, Principal component analysis of DIS values for the PPIs obtained under LPS priming and NLRP3 activation stages. F, Enriched representative biological processes of LPS-preferential PPIs. G, Endosomal transport-related proteins and their perturbation patterns across stimulation conditions. H, Representative biological processes of activation signal-preferential PPIs. I, The proteins associated with Cellular stress-associated processes and their interaction patterns. We first compared SAINTexpress interaction probabilities (AvgP) across LPS priming and activation conditions. Under nigericin stimulation, most interactions remained clustered near the unchanged region, indicating limited redistribution of interaction strengths relative to LPS priming (Fig. 2B). In contrast, ATP and MSU stimulation resulted in pronounced shifts toward activation signal-enriched interactions, reflecting the emergence of NLRP3-associated interactions that were weak or undetectable during LPS priming alone. Consistent with this observation, DIS distributions for ATP and MSU conditions exhibited clear right-skewed tails, with MSU stimulation producing the strongest activation signal-preferential interaction profile (Fig. 2C). To further resolve interactions exhibiting the pronounced condition-specific preference, PPIs were ranked according to their DIS values (Fig. 2D). Principal component analysis based on DIS profiles across all prey proteins revealed robust separation of LPS priming and NLRP3 activation conditions, indicating that distinct activation signals produce divergent perturbation patterns on the NLRP3 interaction network (Fig. 2E). We next sought to determine whether these perturbation patterns reflected underlying biological processes associated with the priming stage of NLRP3 activation. Gene Ontology analysis of interactions preferentially detected during LPS priming revealed enrichment of processes linked to acute response, metabolic regulation, endosomal transport, and redox stress (Fig. 2F and Supplemental Table S2B). These pathways have been associated with the establishment of the primed inflammasome state, in which macrophages upregulate inflammatory signaling components while simultaneously maintaining organelle integrity and metabolic balance. Notably, enrichment of endosomal transport-related terms, including acidic amino acid transport, intracellular cholesterol transport, and phagosome maturation, suggests that LPS priming induces a lysosomal-endosomal reprogramming program (Fig. 2G). This program optimizes organelle homeostasis and vesicular trafficking, thereby establishing cellular readiness for subsequent NLRP3 activation signals that require intact phagosome-lysosome function. In contrast to the interaction patterns preferentially observed during LPS priming, interactions preferentially toward activation-signal stimulation were enriched for processes related to cellular stress responses, energy metabolism, immune response, and organelle dysfunction (Fig. 2H, Supplemental Table S2C). Notably, activation signal-preferential interactions showed prominent enrichment for mitochondrial stress-associated processes, including response to oxidative stress, mitochondrial calcium ion homeostasis, and inner mitochondrial membrane organization (Fig. 2I). Additional enriched categories included innate immune response-activating signaling and type I interferon production, as well as pathways involved in vesicle cargo loading, and responses to endoplasmic reticulum stress. These enrichment patterns indicate that activation-signal stimulation is associated with preferential engagement of interaction modules linked to mitochondrial metabolic activity and stress-responsive cellular pathways during inflammasome activation. Optoproteomic Profiling Identifies the Proximal Proteome of ASC Specks During Inflammasome Activation To investigate the spatial proteome during inflammasome activation, we applied a microscopy-guided optoproteomics strategy to selectively label proteins around ASC specks upon nigericin stimulation (Fig. 3A). ASC specks were first visualized by immunofluorescence staining, followed by targeted photoactivation of biotin-conjugated probes within ASC-positive foci. Successful proximity labeling was confirmed by a marked quench of ASC immunofluorescence signal (red) with strong enrichment of biotin signal (green) at ASC speck sites, indicating efficient and spatially restricted photo-biotinylation (Fig. 3B). Optoproteomics identifies ASC speck-proximal proteins.A, workflow of ASC speck optoproteomics, including microscopy-guided photolabeling, enrichment, and mass spectrometry-based proteomic profiling. B, immunofluorescence showing colocalization of ASC and streptavidin before and after photolabeling, confirming spatially restricted biotinylation of ASC speck-associated proteins. C, differential proteomics analysis of optoproteomics-enriched proteins (fold change >1.3, adjusted p < 0.05). D, Venn diagram showing overlap between optoproteomics-identified proteins and the proteins detected by IP-MS under nigericin stimulation. E, overlap of proteins shared between the datasets shown in panel D, identified using NLRP3 or ASC as bait under nigericin stimulation. F, gene ontology enrichment analysis of the common proteins in panel D for biological process and cellular component. G, interaction network of the common proteins in panel D. Solid edges denote experimentally identified interactions; dashed edges represent STRING-supported interactions (confidence >0.4). Edge color indicates the corresponding DIS score. Optoproteomics datasets were of high quality, with consistent peptide and protein identification depth, and high reproducibility among biological replicates (Supplemental Fig. S3, A–D). Mass spectrometric analysis identified 1980 and 1982 proteins in photolabeled and unlabeled control samples, respectively (Supplemental Table S3A). Differential abundance analysis revealed 555 proteins significantly enriched (Supplemental Table S3B) in the proximity-labeled group (fold change >1.3; adjusted p < 0.05) (Fig. 3C), including core inflammasome component NLRP3, supporting the specificity and biological relevance of the ASC speck-proximal proteome. To identify proteins tightly associated with ASC speck assembly, we integrated the optoproteomics dataset with NLRP3- and ASC-centered interactomes obtained by IP-MS under nigericin stimulation. This comparison identified 11 proteins shared between the ASC speck-proximal proteome and the IP-MS-sourced interactions (Fig. 3D). Among these proteins, IMMT was the only protein detected in proximity to ASC specks and interacted with both NLRP3 and ASC (Fig. 3E). Gene Ontology enrichment analysis of the 11 shared proteins revealed biological processes closely aligned with canonical NLRP3 inflammasome activation. Enriched terms included mitochondrial calcium ion homeostasis, cellular respiration, and endolysosomal TLR signaling pathway (Fig. 3F, Supplemental Table S3C). These processes have been reported in upstream or permissive events required for NLRP3 inflammasome activation. For example, mitochondrial Ca2+ overload promotes ROS production and NLRP3 activation (26), while enhanced cellular respiration provides the metabolic priming required for inflammasome assembly (27). Additionally, endolysosomal TLR signaling cooperates with mitochondrial stress to license NLRP3 activation (28). Consistent with these functional annotations, enriched cellular component terms included the ATPase-dependent transmembrane transport complex, mitochondrial matrix, and inner mitochondrial membrane protein complexes. The ATPase-dependent transmembrane transport complex is involved in maintaining ion homeostasis, and its enrichment may reflect the ionic fluxes critical for NLRP3 inflammasome activation (29). Moreover, ASC specks represent supramolecular signaling platforms in which ASC oligomerization recruits’ proteins to enable downstream inflammasome signaling (30). Collectively, these results indicate that optoproteomics captures proteins residing within the microenvironment supporting ASC speck assembly. Finally, we examined whether ASC speck-proximal proteins preferentially participate in activation signal-dependent interaction remodeling. Integration of DIS revealed that the two preferential PPIs upon nigericin stimulation (DISnig > 0.5) were NLRP3-IMMT and ASC-IMMT (Fig. 3G). Integration of STRING-supported interactions (confidence score >0.4) provided additional evidence for potential associations among NLRP3, ASC, and IMMT (Fig. 3G). These findings suggest that IMMT may participate in nigericin-induced inflammasome activation. IMMT Associates with NLRP3 and ASC in an Activation-Dependent Manner Based on the previous evidence, we next determine whether IMMT associates with NLRP3 and ASC under inflammasome-activating conditions. To examine endogenous interactions, co-immunoprecipitation assays (Co-IP) were performed in LPS-primed iBMDMs in the presence or absence of nigericin stimulation. The results revealed a detectable association between IMMT and NLRP3 at the priming stage, which was markedly enhanced following nigericin stimulation (Fig. 4A). This activation-dependent increase is consistent with the stimulus-resolved interaction remodeling observed in the proteomic analyses and supports the notion that IMMT engagement is strengthened upon inflammasome activation rather than during priming alone. IMMT interacts with NLRP3 and ASC.A, Co-IP analysis of IMMT and NLRP3 in LPS-primed iBMDMs with or without nigericin stimulation. Cell lysates were immunoprecipitated (IP) and immunoblotted (IB) with indicated antibodies. B, colocalization of IMMT and NLRP3 by immunofluorescence in iBMDMs under LPS priming with or without nigericin. C, Co-IP analysis of IMMT and ASC under LPS plus nigericin stimulation. D, Confocal imaging showing colocalization of IMMT and ASC in iBMDMs upon NLRP3 inflammasome activation. To further assess the spatial relationship between IMMT and NLRP3, we performed confocal microscopy in iBMDMs. After nigericin stimulation, NLRP3 redistributed from a diffuse cytoplasmic pattern into discrete punctate structures characteristic of inflammasome assembly. Under nigericin treatment, IMMT exhibited clear colocalization with NLRP3-positive puncta, whereas such colocalization was not observed in unstimulated cells (Fig. 4B). These observations indicate that IMMT is spatially recruited to sites of NLRP3 aggregation during inflammasome activation. We next investigated whether IMMT is also associated with the adaptor protein ASC. Co-IP experiments demonstrated that IMMT efficiently pulled down ASC in iBMDMs following nigericin stimulation (Fig. 4C). Consistent with these biochemical findings, confocal imaging revealed pronounced colocalization between IMMT and ASC under inflammasome-activating conditions (Fig. 4D), further supporting the involvement of IMMT within ASC-containing inflammasome assemblies. Taken together, these data provide orthogonal support for the proteomics-based identification of IMMT as a component linked to inflammasome assembly. IMMT Negatively Regulates NLRP3 Inflammasome Activation To investigate the functional role of IMMT in NLRP3 inflammasome regulation, we examined the effects of Immt knockdown on multiple hallmarks of inflammasome activation, including caspase-1 processing, cytokine secretion, pyroptotic cell death, and ASC oligomerization. Immunoblot analysis revealed that silencing Immt in iBMDMs greatly enhanced LPS/nigericin-induced cleavage of caspase-1 (p20) and secretion of IL-1β (p17), while the levels of pro-caspase-1 and pro-IL-1β remained unchanged (Fig. 5A). Consistently, ELISA measurements demonstrated that IL-1β secretion was significantly increased in Immt knockdown cells upon nigericin stimulation compared with control cells (Fig. 5B). In line with the elevated caspase-1 activity, Immt knockdown significantly increased cleavage of gasdermin D (GSDMD) and augmented lactate dehydrogenase (LDH) release following LPS/nigericin stimulation (Fig. 5, C and D), indicating enhanced pyroptosis. Finally, to evaluate the effect of IMMT on inflammasome assembly, we examined ASC oligomerization. Crosslinking of ASC with disuccinimidyl suberate (DSSO) revealed higher levels of ASC oligomers in Immt knockdown cells compared with controls under nigericin treatment (Fig. 5E), suggesting that IMMT may restrict ASC polymerization during inflammasome activation. These results indicate that IMMT acts as a negative regulator of NLRP3 inflammasome activation. IMMT negatively regulates NLRP3 inflammasome activation.A, immunoblot analysis of caspase-1 (p20) and IL-1β (p17) in supernatants (Sup) or cell lysates of iBMDMs transfected with control or Immt siRNAs, primed with LPS, followed by nigericin treatment. B, ELISA quantification of IL-1β secretion under the same conditions as (A). C, immunoblot of GSDMD cleavage in iBMDMs following Immt knockdown and NLRP3 inflammasome activation. D, LDH release assay showing pyroptotic cell death in Immt-silenced iBMDMs upon NLRP3 activation. E, ASC oligomerization assessed by DSSO crosslinking and immunoblot in Immt knockdown iBMDMs following LPS priming and nigericin stimulation. Statistical analyses in (B) and (D) were performed using two-way ANOVA with Bonferroni post hoc test; ∗∗∗∗p < 0.0001. Discussion The NLRP3 inflammasome is a dynamic signaling platform whose activation relies on tightly regulated PPIs. Although multiple regulatory factors have been reported, a systematic, stimulus-resolved view of how NLRP3 and its adaptor ASC engage protein networks during physiological inflammasome activation remains limited. Here, we combined endogenous IP-MS of NLRP3 and ASC with optoproteomics of ASC specks to characterize stimulus-associated interaction landscapes in primary BMDMs under canonical inflammasome activation conditions. This integrated approach enabled comparative profiling of protein associations detected by complementary strategies, revealing that the composition of NLRP3- and ASC-associated protein networks varied across stimulation conditions. Different activation signals were associated with partially distinct interaction patterns, consistent with context-dependent modulation of inflammasome-associated protein assemblies. Moreover, the limited overlap between NLRP3 and ASC interactomes observed under the same stimulation conditions indicates that interaction profiles are influenced by the bait protein analyzed, underscoring the methodological and spatial specificity of inflammasome-associated interactions. A major finding of our study is the identification of IMMT as a context-dependent regulator of the NLRP3 inflammasome. IMMT associated with both NLRP3 and ASC upon inflammasome activation, and knockdown of Immt enhanced caspase-1 and GSDMD cleavage, IL-1β secretion, LDH release, and ASC oligomerization. These results support a model in which IMMT acts as a negative regulator, potentially fine-tuning inflammasome assembly and downstream inflammatory responses. This observation complements previous reports linking mitochondrial dynamics to inflammasome regulation and suggests that IMMT might contribute to mitochondrial-mediated control of NLRP3 signaling. Compared with previous interaction studies of NLRP3 inflammasome that relied on overexpression systems, single stimuli, or non-physiological cell lines, our dataset captures endogenous protein complexes under multiple physiologically relevant conditions. The combination of IP-MS and optoproteomics enables simultaneous detection of core and spatially restricted interactions, providing a multidimensional view of inflammasome assembly that may not be accessible with conventional approaches. In conclusion, this work provides a comprehensive, stimulus-resolved interactome of the NLRP3 inflammasome and identifies IMMT as a novel negative regulator. These findings offer mechanistic insights into inflammasome regulation and provide a proteomic resource that can guide future studies on stimulus-specific inflammasome signaling and potential therapeutic targets for NLRP3-driven inflammatory diseases. Limitations of the Study This study has several limitations. First, while endogenous IP-MS and ASC speck optoproteomics provide a comprehensive view of NLRP3- and ASC-associated protein networks, these approaches cannot unambiguously distinguish direct physical interactions from indirect or spatially proximal associations within higher-order inflammasome assemblies. Other methods, such as cross-linking mass spectrometry (XL-MS), represent a promising approach to resolving direct physical interactions. Second, our analyses were performed in BMDMs under canonical activation conditions, and the interactomes identified here may differ in other cell types or in more complex inflammatory contexts. Finally, although functional experiments indicate that IMMT negatively regulates NLRP3 inflammasome activation, the precise molecular mechanisms by which IMMT influences inflammasome assembly and signaling remain to be determined. Data Availability The mass spectrometry proteomics data generated in this study have been deposited to the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX (31, 32) partner repository with the dataset identifier PXD079555. In addition, NLRP3 interactome data obtained from BMDMs under LPS and nigericin stimulation have been previously reported and are publicly available via the ProteomeXchange Consortium through the iProX partner repository under the dataset identifier PXD063269. Supplemental data This article contains supplemental data. Conflict of Interests The authors declare no competing interests. Acknowledgments The authors wish to acknowledge the State Key Laboratory of Medical Proteomics, the National Center for Protein Science (Beijing) (NCPSB) for performing mass spectrometry (Mingchao Wang, Bin Fu, Yuping Xie, Shuhui Ji and Wenzhe Chen) and imaging analysis (Shaoqiong Yi, Chunhua Zhang and Yecheng Jin). Author Contributions Y. Z., Y. L., and J. W. writing–original draft; Y.Z., Y. L., and J. W. writing–review & editing; Y. Z., C. J., K. Z., L. S., S. W., and M. Z., investigation; Y. L., and J. W. data curation; Y. L. and J. W. visualization; Y. L. and J. W. formal analysis; J. W. funding acquisition; J. W. conceptualization; J. W. supervision. Funding and Additional Information This work was supported by the National Key R&D Program of China to J. W. (2025YFA1309201). Contributor Information Yuchen Liu, Email: 57994893@qq.com. Jian Wang, Email: wangjian@bmi.ac.cn. Supplemental Data Supplemental Table S1 Supplemental Table S2 Supplemental Table S3 Supplementary Figures References Associated Data Supplementary Materials Supplemental Table S1 Supplemental Table S2 Supplemental Table S3 Supplementary Figures Data Availability Statement The mass spectrometry proteomics data generated in this study have been deposited to the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX (31, 32) partner repository with the dataset identifier PXD079555. In addition, NLRP3 interactome data obtained from BMDMs under LPS and nigericin stimulation have been previously reported and are publicly available via the ProteomeXchange Consortium through the iProX partner repository under the dataset identifier PXD063269.