The leucine-rich repeat (LRR) domain of NLRP3 is required for NLRP3 inflammasome activation in macrophages

The NLRP3 inflammasome is a critical component of innate immunity that defends the host from microbial infections. However, its aberrant activation contributes to the pathogenesis of several inflammatory diseases. Activation of the NLRP3 inflammasome induces the secretion of proinflammatory cytokines IL-1β and IL-18 and pyroptotic cell death. NLRP3 contains a leucine-rich repeat (LRR) domain at its C terminus. Although posttranslational modifications in this LRR domain have been shown to regulate NLRP3 inflammasome activation, the role of the entire LRR domain in NLRP3 inflammasome activation remains controversial. Here, we generated mouse macrophages that express an endogenous NLRP3 mutant lacking the LRR domain. Deletion of the LRR domain diminished NLRP3 inflammasome activation in macrophages. Furthermore, using NLRP3-deficient macrophages that are reconstituted with NLRP3 mutants lacking the LRR domain, we found that deletion of the LRR domain inhibited NLRP3 inflammasome activation. Mechanistically, deletion of the LRR domain inhibited NLRP3 self-association, oligomerization, and interaction with the essential regulator NEK7. Our results demonstrate a critical role for the LRR domain in NLRP3 inflammasome activation.

The NLRP3 inflammasome is a critical component of innate immunity that defends the host from microbial infections. However, its aberrant activation contributes to the pathogenesis of several inflammatory diseases. Activation of the NLRP3 inflammasome induces the secretion of proinflammatory cytokines IL-1β and IL-18 and pyroptotic cell death. NLRP3 contains a leucine-rich repeat (LRR) domain at its C terminus. Although posttranslational modifications in this LRR domain have been shown to regulate NLRP3 inflammasome activation, the role of the entire LRR domain in NLRP3 inflammasome activation remains controversial. Here, we generated mouse macrophages that express an endogenous NLRP3 mutant lacking the LRR domain. Deletion of the LRR domain diminished NLRP3 inflammasome activation in macrophages. Furthermore, using NLRP3-deficient macrophages that are reconstituted with NLRP3 mutants lacking the LRR domain, we found that deletion of the LRR domain inhibited NLRP3 inflammasome activation. Mechanistically, deletion of the LRR domain inhibited NLRP3 self-association, oligomerization, and interaction with the essential regulator NEK7. Our results demonstrate a critical role for the LRR domain in NLRP3 inflammasome activation.
Inflammasomes are cytoplasmic multiprotein complexes that form in response to microbial infection or cell damage (1). In most cases, an inflammasome contains a pattern recognition receptor as the sensor, the adaptor protein apoptosisassociated speck-like protein containing a caspase-activation and recruitment domain (ASC), and the inflammatory caspase-1 (2,3). Inflammasome assembly results in caspase-1 autocleavage and activation. Activated caspase-1 processes the proinflammatory cytokines pro-interleukin (IL)-1β and pro-IL-18 into their mature forms and cleaves gasdermin D (GSDMD) to release its active N terminus, which forms pores in the plasma membrane to induce cytokine release and pyroptotic cell death (4)(5)(6). To date, six inflammasomes have been well documented, including the nucleotide-binding oligomerization domain-like receptor (NLR) family members NLRP1, NLRP3, NLRP6, NLRC4, pyrin, and absent in melanoma 2 (AIM2) inflammasomes (2,7). Among these inflammasomes, the NLRP3 inflammasome has been intensively investigated due to its protective role in host defense against various infections as well as its pathogenic role in several inflammatory disorders (8,9).
NLRP3 consists of three domains: an N-terminal pyrin domain (PYD), a central NAIP, CIITA, HET-E, and TP1 (NACHT) domain, and a C-terminal leucine-rich repeat (LRR) domain (21). NLRP3 interacts with the adaptor protein ASC through PYD-PYD interactions to induce the formation of filamentous aggregates known as ASC specks (22,23). The NACHT domain has an adenosine triphosphatase activity required for NLRP3 conformational change and oligomerization (24,25). In contrast, the role of the LRR domain in NLRP3 inflammasome activation is more complex. Earlier studies have suggested that the LRR domain has an autoinhibitory function through intramolecular interactions (26). Recently, several posttranslational modifications, including phosphorylation and ubiquitination, have been found in the LRR domain to regulate NLRP3 inflammasome activation (27)(28)(29)(30). However, there are conflicting reports on whether the entire LRR domain plays a role in NLRP3 inflammasome activation. Some studies have shown that the replacement of the LRR domain by the LacZ protein or its deletion abolished NLRP3 inflammasome activation (31,32). Furthermore, structural studies have shown that the LRR domain mediated the NLRP3-NEK7 interaction and the formation of the NLRP3 cage structure, which disperses the trans-Golgi network at the early stage of the inflammasome pathway (33,34). However, others have shown that the LRR domain is dispensable for NLRP3 inflammasome activation (35,36).
In this study, we have used mouse knock-in (KI) macrophages that express an endogenous NLRP3 mutant lacking the LRR domain and NLRP3-deficient macrophages that are reconstituted with NLRP3 mutants lacking the LRR domain. Our results demonstrate that the LRR domain is required for NLRP3 inflammasome activation. Furthermore, our results show that the LRR domain mediates NLRP3 self-association, oligomerization, and interaction with its essential regulator NEK7.

Deletion of the LRR domain from endogenous NLRP3 protein impairs nigericin-induced NLRP3 inflammasome activation in macrophages
Previous studies showed conflicting results on the role of the LRR domain in NLRP3 inflammasome activation. Notably, most of these studies overexpressed NLRP3 mutants lacking the LRR domain by transfection or viral infection or replaced the LRR domain of the endogenous NLRP3 protein with a large LacZ protein, which may interfere with NLRP3 function (31,32,35,36). To determine the exact role of this LRR domain in NLRP3 inflammasome activation, we used the CRISPR-Cas9 gene-editing system to generate immortalized mouse bone marrow-derived macrophages (iBMDMs) that endogenously express an NLRP3 mutant lacking the LRR domain. An NLRP3 mutant [NLRP3(1-720)] was chosen as it lacks the entire LRR domain and has been reported to fully restore NLRP3 inflammasome activation in NLRP3-deficient mouse macrophages (Fig. 1A) (35). We selected three KI macrophage clones (Nlrp3 C721Stop/C721Stop ) in which a stop codon (TGA) has successfully replaced the cysteine (C721) codon (TGC) at the coding region of mouse Nlrp3 gene (Fig. 1B). These KI macrophages are expected to express an NLRP3 mutant that consists of the first 720 amino acid residues and lacks the entire LRR domain.
We assessed the protein expression of endogenous NLRP3 protein from both wildtype (parent cells) and KI macrophages by Western blot. An antibody that recognized an epitope in the Actin was used as a control. Representative blots (n = 2). The band density of NLRP3 was measured by ImageJ and normalized to the full-length NLRP3. D, macrophages were stimulated with LPS (4 h) alone or plus 5 μM nigericin (1 h). Cell lysates and supernatants were immunoblotted with indicated antibodies. Representative blots (n = 3). Measurement of IL-1β (E) and TNF-α (F) in the supernatants from stimulated macrophages by ELISA. Representative data(n = 3). *p < 0.05 (one-way ANOVA). Data are the mean ± SD of triplicate wells. CL, cleaved; FL, full-length; LPS, lipopolysaccharide; LRR, leucine-rich repeat.
N-terminal pyrin domain of NLRP3 protein (Cryo-2) was used to detect both full-length and truncation mutant NLRP3 from these macrophages. In addition, macrophages derived from Nlrp3 −/− mice were included as controls in our experiments. The full-length NLRP3 (118 kDa) was detected in wildtype macrophages at the resting state (Fig. 1C). As expected, all three KI macrophage clones expressed the short version of NLRP3 protein (79 kDa), further confirming our sequencing results (Fig. 1C). Of note, no protein band appeared in the lane for Nlrp3 −/− macrophages, indicating the specificity of our antibody (Fig. 1C). Importantly, the detected protein level of this NLRP3 mutant from each clone was markedly lower (less than 50%) when compared with that of full-length NLRP3 expressed in wildtype macrophages (Fig. 1C). Additionally, the mRNA levels of this NLRP3 mutant in KI macrophages were also lower than those of the full-length NLRP3 in wildtype macrophages (Fig. S1A). However, the expression levels of other NLRP3 inflammasome components, including caspase-1, NEK7, and ASC, were comparable between wildtype and KI macrophages (Fig. 1C). Next, we assessed NLRP3 inflammasome activation in these macrophages under the conditions of lipopolysaccharide (LPS) priming or LPS priming plus nigericin treatment. Previous studies have suggested that the LRR domain keeps an inactive conformation of NLRP3 through intramolecular interaction (26). Similar to wild-type macrophages, KI macrophages did not exhibit caspase-1 activation, GSDMD cleavage, or IL-1β secretion after LPS priming (Fig. 1, D and E). Nigericin, a classic NLRP3 stimulus, induced NLRP3 inflammasome activation in wildtype macrophages, as revealed by caspase-1 activation, GSDMD cleavage, and IL-1β secretion ( Fig. 1, D and E). In contrast, KI macrophages, similar to NLRP3-deficient macrophages, were defective in NLRP3 inflammasome activation after LPS priming plus nigericin treatment ( Fig. 1, D and E). The secretion of TNF-α, an inflammasome-independent cytokine, was comparable among these macrophages (Fig. 1F). Collectively, these results indicate that the LRR domain is required for nigericin-induced NLRP3 inflammasome activation in macrophages.

Deletion of the LRR domain inhibits NLRP3 inflammasome activation by diverse NLRP3 stimuli
NLRP3 inflammasome activation is induced by potassium efflux-dependent (e.g., ATP, nigericin, Nano-SiO 2 ) or independent stimuli (e.g., imiquimod) (11,12). Given our data showing that KI macrophages were defective in nigericininduced NLRP3 inflammasome activation, we next examined whether this defect was common to other NLRP3 stimuli. As they secreted a similar level of TNF-α after LPS priming, the KI macrophage clone KI-C2 (KI henceforth) and wildtype macrophages were used for the following experiments (Fig. 1F). Macrophages were stimulated with ATP, nigericin, Nano-SiO 2 , or imiquimod after LPS priming. KI macrophages showed markedly reduced caspase-1 activation, GSDMD cleavage, and IL-1β secretion when compared to wildtype macrophages (Fig. 2, A and B). In contrast, KI macrophages had comparable AIM2 inflammasome activation induced by poly(dA:dT) or NLRC4 inflammasome activation by Salmonella when compared to wildtype macrophages (Fig. 2, A and B). Inflammasome activation induces the formation of ASC aggregates called ASC specks, which are composed of ASC oligomers in the cytosol (22). Both ATP and nigericin induced ASC speck formation in wildtype macrophages, whereas almost no ASC speck was detected in KI macrophages for both stimuli (Fig. 2, C and D). However, KI macrophages had comparable ASC speck formation in response to Salmonella as compared to wildtype macrophages (Fig. 2, C and D). Biochemical analysis revealed that KI macrophages had much less ASC oligomerization as compared to wildtype cells after treatment of ATP or nigericin, although they showed a comparable level of ASC oligomerization in response to Salmonella (Fig. 2E). To exclude whether other nonspecific factors besides this LRR domain deletion in endogenous NLRP3 attributed to the defect of NLRP3 inflammasome activation observed in KI macrophages, we expressed C-terminal HAtagged NLRP3 full-length [NLRP3(FL)], LRR deletion mutant [NLRP3(1-720)], or the LRR domain [NLRP3(730-1033)] in KI macrophages by a lentiviral vector. Macrophage clones that expressed higher levels of the LRR deletion mutant protein than those of full-length NLRP3, indicated by anti-HA immunoblots, were selected for assessing nigericin-induced NLRP3 inflammasome activation (Fig. 2F). Expression of full-length NLRP3 restored NLRP3 inflammasome activation, indicated by caspase-1 activation and IL-1β secretion, in KI macrophages, while both the LRR deletion mutant and LRR domain alone failed to rescue NLRP3 inflammasome activation (Fig. 2, F and G). Taken together, these results indicated that the LRR domain is required for NLRP3 inflammasome activation in macrophages.
NLRP3 mutants lacking the LRR domain fail to restore NLRP3 inflammasome activation in NLRP3-deficient macrophages Several LRR domain-lacking mutants of NLRP3, including NLRP3 1-686 and 1-720, have been reported to fully restored inflammasome activation in mouse NLRP3-deficient macrophages (35). As shown in our data, the protein level of endogenously expressed protein NLRP3(1-720) in KI macrophages was lower than that of the full-length NLRP3 in macrophages. As a complementary approach, we constitutively expressed SFP (S-tag, Flag, and streptavidin-binding tag)-tagged full-length NLRP3, NLRP3 mutant 1-686, or 1-720 in mouse NLRP3-deficient macrophages and selected macrophage clones that expressed comparable protein levels of NLRP3 (Fig. 3A). In contrast to the previous study in which macrophages were primed with LPS for 11 h in the presence of doxycycline, we primed reconstituted macrophages with LPS for 4 h before stimulation with ATP, nigericin, nano-SiO 2 , or imiquimod (35). Consistent with our results from KI macrophages, macrophages reconstituted with NLRP3 mutant 1-686 or 1-720 showed a significant reduction in caspase-1 activation, GSDMD cleavage, and IL-1β secretion when compared to macrophages with full-length NLRP3 (Fig. 3, A and B). The levels of TNF-α secretion were comparable between Critical role for the LRR domain in NLRP3 activation macrophages reconstituted with wildtype or mutant NLRP3 under the same conditions (Fig. 3C). Accordingly, macrophages reconstituted with NLRP3 mutant 1-686 or 1-720 had less ASC oligomerization in response to ATP or nigericin when compared to macrophages with full-length NLRP3 (Fig. 3D). Furthermore, NLRP3-deficient macrophages that expressed an untagged NLRP3 mutant 1-720 through a constitutive or doxycycline-inducible system also failed to restore caspase-1 activation after nigericin stimulation (Fig. S1, B and C). Taken together, these results further indicated that the LRR domain is also essential for NLRP3 inflammasome activation in reconstituted NLRP3-deficient macrophages.

Deletion of the LRR domain inhibits NLRP3 self-association and oligomerization
Recent structural studies with purified NLRP3 oligomers have revealed that NLRP3 self-associates through the LRR-LRR interaction (34,37). To investigate how the LRR domain controls NLRP3 inflammasome activation, we assessed the effects Figure 2. The LRR domain is required for NLRP3 inflammasome activation induced by diverse stimuli. A, LPS-primed WT and NLRP3 1-720 knock-in (KI) macrophages were primed with LPS (4 h) and then stimulated with PBS (mock), 5 mM ATP (1 h), 5 μM nigericin (1 h), 200 μg/ml Nano-SiO 2 (4 h), 15 μg/ml Imiquimod (4 h), 4 μg/ml poly(dA:dT) (4 h) or Salmonella (m.o.i= 10, 2 h). Cell lysates and supernatants were immunoblotted with indicated antibodies. Representative blots (n = 3). B, measurement of IL-β in the supernatants from (A) by ELISA. Representative data (n = 3). *p < 0.05 (unpaired two-sided t test). Data are the mean ± SD of triplicate wells. C, ASC immunostaining in wildtype or KI macrophages primed with LPS and stimulated with PBS, ATP, nigericin, or Salmonella. Scale bars, 10 μm. ASC specks are indicated by white arrows. Representative images (n = 3). D, quantification of ASC specks from (C). The percentage of ASC speck-containing cells was calculated from three different fields with at least 100 cells each. *p < 0.05 (unpaired two-sided t test). of the LRR domain deletion on NLRP3 self-association and oligomerization in cells. We co-expressed SFP or HA-tagged full-length or NLRP3 mutants in HEK 293T cells. The fulllength NLRP3 showed self-association as SFP-tagged NLRP3 pulled down HA-tagged NLRP3. In contrast, self-association for both NLRP3 mutants 1-686 and 1-720 was markedly reduced (Fig. 4A). In macrophages, after stimulation of NLRP3 stimuli, NLRP3 forms high-molecular-mass complexes, as indicated by protein native gels (17). We compared these highmolecular-mass forms of NLRP3 in cells reconstituted with fulllength or the LRR deletion mutant. Macrophages reconstituted with the full-length NLRP3 displayed the high-molecular-mass NLRP3 complexes (>1200 kDa) after stimulation of ATP and nigericin. In contrast, these high-molecular-mass forms of NLRP3 were markedly reduced in macrophages reconstituted with NLRP3 truncation mutant 1-686 or 1-720 (Fig. 4B). Accordingly, in 2-D native gels, these high-molecular-mass forms of NLRP3 proteins were markedly reduced in macrophages reconstituted with NLRP3 mutant 1-686 or 1-720 as compared with the full-length NLRP3 (Fig. 4C). Taken together, our results indicate that the LRR domain mediates NLRP3 selfassociation and oligomerization.

Deletion of the LRR domain abolishes the NEK7-NLRP3 interaction
Previous studies have shown that the LRR domain in NLRP3 is required for the NEK7-NLRP3 interaction in HEK 293T cells (17,18). Furthermore, a structural study of the NEK7-NLRP3 complex has shown that NEK7 interacts with the LRR domain of NLRP3 (33). Surprisingly, NLRP3 mutants lacking the LRR domain including 1-686 were reported to pull down NEK7 in macrophages (35). Therefore, we examined whether the LRR domain is required for NLRP3-NEK7 interaction in our reconstituted macrophages. Both NLRP3 mutants 1-686 and 1-720 failed to pull down NEK7 from macrophages stimulated with LPS or LPS plus ATP (Fig. 5A). The NLRP3 mutant 1-720 or the LRR domain alone also failed to pull down NEK7 from HEK 293T cells (Fig. 5B). According Critical role for the LRR domain in NLRP3 activation to the Uniprot/Swiss-Prot, there are nine LRRs in the LRR domain of NLRP3. To further define the exact LRR(s) of the LRR domain that is essential for NLRP3-NEK7 interaction, we generated a series of NLRP3 mutants which progressively lack one LRR starting from the C terminus. Co-expression of the Flag-tagged NLRP3 mutant and HA-tagged NEK7 in HEK 293T cells was used to assess the interaction between NEK7 and NLRP3 mutants. All of these NLRP3 mutants that lack one or several LRRs in the LRR domain lost the ability to bind NEK7 (Fig. 5C). Furthermore, the NLRP3 mutant lacking one LRR (Δ LRR 9) failed to restore nigericin-induced caspase-1 activation in reconstituted NLRP3-deficient macrophages (Fig. 5D). Collectively, these results indicate that the intact LRR domain is essential for NEK7-NLRP3 interaction and NLRP3 inflammasome activation.

Discussion
NLRP3 activation triggers the assembly of an inflammasome that plays a critical role in innate immunity. In addition to its N-terminal pyrin domain and central NACHT domain, NLRP3 contains an LRR domain at its C terminus. Several posttranslational modifications have been found on the LRR domain to regulate NLRP3 inflammasome activation (11,12). In this study, we investigated the role of this LRR domain in NLRP3 inflammasome activation in mouse macrophages. Our results show that the LRR domain is required for NLRP3 inflammasome activation in macrophages. Furthermore, we show that the LRR domain mediates NLRP3 self-association, oligomerization, and interaction with NEK7, which are critical molecular events for NLRP3 inflammasome activation.
Our results showed that the protein level of NLRP3 1-720 in KI macrophages was much less than that of full-length NLRP3 in wildtype cells. This reduction might result from the nonsense-mediated decay of NLRP3 1-720 mRNA after the introduction of the stop codon in the Nlrp3 gene (Fig. S1A) (38). However, our reconstitution experiments showed that the protein level of this NLRP3 LRR deletion mutant (without a tag) was consistently lower than that of the full-length NLRP3, suggesting that the LRR domain promotes NLRP3 stability in macrophages (Fig. S1, B and C). This role of the LRR domain in NLRP3 protein stability, although it has not been explored further, has been suggested in previous studies (31). In addition, previous studies report that the chaperone HSP90 interacts with NLRP3 via the LRR domain and stabilizes the NLRP3 protein (15). Therefore, we hypothesize that the deletion of the LRR domain in NLRP3 might cause the loss of this HSP90-mediated protection and thus render its degradation by the proteasome. In support of this hypothesis, the E3 ubiquitin ligase Cbl-b has been reported to ubiquitinate the NACHT domain at K496 for proteasome-mediated degradation of NLRP3 (28). Interestingly, short NLRP3 transcripts that encode NLRP3 LRR domain truncation forms have been reported in mouse and human cells (39,40). It remains unknown whether those truncation proteins of NLRP3 undergo degradation in cells.
NLRP3 inflammasome activation in NLRP3-deficient human monocytic U937 cells. In addition, BMDMs expressing an NLRP3 mutant, in which the LRR domain was replaced by a LacZ protein, were defective in NLRP3 inflammasome activation induced by monosodium urate crystals (31). Consistent with the latter two studies, our results indicate that the LRR domain is required for NLRP3 inflammasome activation in mouse macrophages. In our KI macrophages that express an endogenous NLRP3 mutant lacking the LRR domain, NLRP3 inflammasome activation was completely abrogated in response to all tested NLRP3 stimuli. This defect may also partially be attributed to the low protein level of this endogenous NLRP3 mutant in KI macrophages. However, overexpression of this NLRP3 mutant failed to restore NLRP3 inflammasome activation in these KI macrophages, suggesting an intrinsic role for the LRR domain in NLRP3 inflammasome activation. In addition, macrophages expressing NLRP3 truncation mutant 1-686 or 1-720, to a similar level as the full-length protein, were defective in NLRP3 inflammasome activation in our reconstitution experiments. We do not know the exact reasons for the above disparities. One possibility is that Hafner-Bratkovi c et al. (35) induced a high level of NLRP3 proteins by using a longer time (11 h) of LPS priming in a doxycycline-inducible system. This experimental setting might induce a high level of protein expression and thus bypass the requirement of the LRR domain in NLRP3 inflammasome activation in macrophages. This is likely also applied to the reconstituted NLRP3 inflammasome in HEK 293 T cells (36). However, we failed to reproduce these results in macrophages using a similar doxycycline-inducible system (Fig. S1C). Importantly, we found that LRR deletion reduced NLRP3 selfassociation and oligomerization. Consistent with this observation, NLRP3 mutants with mutations in the LRR domain have been reported to be defective in the formation of a double-ring cage structure of mouse NLRP3 (34). Therefore, it is conceivable that deletion of the LRR will abolish the NLRP3 cage formation and subsequent inflammasome activation. Additionally, we further confirm and extend previous findings that an intact LRR domain in NLRP3 is required for its interaction with NEK7 (17,18). Since the NLRP3 cage structure is not compatible with NEK7 binding, it remains to be determined when and where the LRR domain is involved in the interaction with NEK7 in this inflammasome activation pathway.
In summary, our findings demonstrated a critical role for the LRR domain in NLRP3 inflammasome activation in macrophages. Our results reveal that the LRR domain controls inflammasome activation by mediating NLRP3 selfassociation, oligomerization, and interaction with NEK7.
cDNA synthesis and real-time RT-PCR iBMDMs were stimulated with LPS (200 ng ml −1 ) for 4 h or left unstimulated. Total RNA was extracted from iBMDMs using E.Z.N.A. Total RNA kit I (Omega Bio-Tek, R6834-01) according to the manufacturer's instructions. cDNA was synthesized using a High-Capacity-RNA-to-cDNA kit (Fisher Scientific, 43-874-06). Real-time PCR was performed using SYBR Green Master Mix (Fisher Scientific, 43-444-63) at a 7500 Real-Time PCR system. The primers sequence was: Nlrp3 forward, CCCTTGGAGACACAGGACTC; Nlrp3 reverse, GAGGCTGCAGTTGTCTAATTCC; GAPDH forward, AGCTTGTCATCAACGGGAAG; GAPDH reverse, TTTGATGTTAGTGGGGTCTCG; The relative expression level of Nlrp3 to GAPDH was calculated using the 2 (−ct) method and normalized to the level of unstimulated wildtype iBMDMs.
Reconstitution of the full-length NLRP3 and its truncation mutants in iBMDMs SFP-tagged, HA-tagged, or untagged full-length mouse NLRP3 or its truncation mutant [NLRP3(1-686), NLRP3(1-720), or NLRP3(730-1033)] were cloned into lentiviral plasmid pHIV-EGFP (Addgene, 21373) or pIN-DUCER21 (Addgene, 46948). Each lentiviral construct was co-transfected with package plasmids pCMV-VSV-G (Addgene, 8454) and pCMV-dR8.2 dvpr (Addgene, 8455) into HEK 293T cells for 96 h. Lentiviruses in the culture supernatants were concentrated by using lenti-X concentrator (Takara Bio, 631231). iBMDMs were seeded into 6-well plates overnight at a density of 1 × 10 5 per well. Cells were then transduced with lentiviruses in the presence of 8 μg ml −1 polybrene (Sigma, H9268) for 24 h before replacement with the fresh culture medium. After 3 to 4 days, transduced cells were sorted by flow cytometry using GFP as a marker. The expression of reconstituted proteins was determined by immunoblotting with an anti-HA, anti-NLRP3, or anti-Flag antibody. Macrophages expressing comparable levels of GFP were used in bulk for experiments with untagged NLRP3 proteins. For experiments with tagged NLRP3 proteins, cell clones of transduced macrophages that expressed different levels of NLRP3 proteins were selected by serial dilution and western blots.
Transfection, immunoprecipitation, and pull-down assay HEK 293T cells were seeded into 6-well plates overnight at a density of 6.25 × 10 5 cells per well. Plasmids expressing HAtagged NEK7, Flag, or SFP-tagged wild-type NLRP3, or its truncation mutant were single-or co-transfected into HEK 293T cells by Lipofectamine LTX (ThermoFisher, 15338100) for 16 h. After transfection, cells were washed with ice-cold PBS twice and lysed in ice-cold lysis buffer [50 mM Tris-HCl (pH 7.4), 2 mM EDTA, 150 mM NaCl, 0.5% (vol/vol) Nonidet P-40, 1 × EDTA-free protease inhibitor cocktail) at 4 C for 10 min. Cell debris was removed by centrifugation at 12,000g for 10 min at 4 C. Cell lysates were precleared by incubation with protein G agarose beads for 1 h at 4 C. Precleared lysates were then incubated with an anti-Flag (1:200) antibody at 4 C overnight. The proteins bound by the antibody were pulled down with protein G beads and subjected to immunoblotting analysis.
ASC speck staining and imaging iBMDMs were seeded overnight into an 8-well permanox chamber slide at a density of 1 × 10 5 cells per well. Cells were primed with 200 ng ml −1 LPS for 4 h in serum-free IMDM and then stimulated with 5 mM ATP (30 min), 5 μM nigericin (1 h), or Salmonella (m.o.i = 10, 4 h). After stimulation, cells were washed once with PBS and then fixed in 4% paraformaldehyde for 20 min at room temperature. Fixed cells were permeabilized with 0.2% Triton X-100 for 5 min and blocked with PBS buffer containing 5% BSA for 1 h. Cells were incubated overnight with an anti-ASC antibody (1:200) in the blocking buffer. After washing three times with PBS, cells were incubated for 1 h with Alexa-Fluor-488 conjugated anti-rabbit secondary antibody (1:5000) in the blocking buffer. Cells were washed with PBS three times and mounted with Prolong Diamond Antifade Mountant with DAPI (Thermo Fisher, P36965). Cell images were taken using a Nikon E-800 microscope system and processed with ImageJ.

ASC oligomerization assay
iBMDMs were seeded overnight in 6-well plates at a density of 1 × 10 6 cells per well. Cells were primed with 200 ng ml −1 LPS for 4 h in serum-free IMDM and then stimulated with 5 mM ATP (30 min), 5 μM nigericin (1 h), or Salmonella (m.o.i = 10, 4 h). After removing the medium, cells were directly lysed in wells for 15 min at 4 C with 300 μl of ice-cold PBS buffer supplemented with 0.5% Triton X-100, 0.5 mM PMSF, and 1 × EDTA-free protease inhibitor cocktail. Cell lysates were collected by scrapping from each well and transferred to Eppendorf tubes. Lysates were separated into supernatants (TritonX-100-soluble fraction) and pellet (Triton X-100-insoluble fraction) by centrifugation at 6000g for 15 min at 4 C. The Triton X-100-insoluble fractions were washed with PBS twice and cross-linked for 30 min at room temperature with 2 mM bis[sulfosuccinimidy] suberate (BS 3 ) (Thermo Fisher, 21580). The cross-linked pellets were spun down at 6000g for 15 min and dissolved in an SDS sample buffer for subsequent immunoblotting analysis.

Blue native PAGE and 2D PAGE
Blue native gel electrophoresis was performed as previously described (42). Briefly, iBMDMs were seeded overnight into 6-well plates at a density of 1 × 10 6 cells per well and stimulated as indicated. After stimulation, cells were washed once with cold PBS and then lysed in ice-cold native lysis buffer (20 mM Bis-tris, 500 mM ε-aminocaproic acid, 20 mM NaCl, 10% (w/v) glycerol, 0.5% digitonin, 0.5 mM Na 3 VO 4 , 1 mM PMSF, 0.5 mM NaF, 1 × EDTA-free Roche protease inhibitor cocktail, pH 7.0) for 15 min on ice. Cell lysates were clarified by centrifugation at 20,000g for 30 min at 4 C; proteins were separated in 4 to 12% blue native PAGE and then analyzed by Western blot. For the two-dimensional (2D) PAGE, the natively resolved gel slice was loaded into the well of 4 to 12% SDS-PAGE gel after being soaked in 1 × Laemmli buffer as previously described (42).

Statistical analysis
Data are represented as mean ± SD. Statistical analysis was performed using unpaired two-tailed Student's t test or oneway ANOVA with GraphPad 9.0. A p value less than 0.05 was considered statistically significant.

Data availability
All data generated for this study are included in this article.
Supporting information-This article contains Supporting information.