Abstract
NLRP1 inflammasome is one of the best-characterized inflammasomes in humans and other mammals. However, the existence of this inflammasome in nonmammalian species remains poorly understood. In this study, we report the molecular and functional identification of an NLRP1 homolog, Danio rerio NLRP1 (DrNLRP1) from a zebrafish (D. rerio) model. This DrNLRP1 possesses similar structural architecture to mammalian NLRP1s. It can trigger the formation of a classical inflammasome for the activation of zebrafish inflammatory caspases (D. rerio Caspase [DrCaspase]–A and DrCaspase-B) and maturation of D. rerio IL-1β in a D. rerio ASC (DrASC)–dependent manner. In this process, DrNLRP1 promotes the aggregation of DrASC into a filament with DrASCCARD core and DrASCPYD cluster. The assembly of DrNLRP1 inflammasome depends on the CARD–CARD homotypic interaction between DrNLRP1 and DrASCCARD core, and PYD–PYD interaction between DrCaspase-A/B and DrASCPYD cluster. The FIIND domain in DrNLRP1 is necessary for inflammasome assembly. To understand the mechanism of how the two DrCaspases are coordinated in DrNLRP1 inflammasome, we propose a two-step sequential activation model. In this model, the recruitment and activation of DrCaspase-A/B in the inflammasome is shown in an alternate manner, with a preference for DrCaspase-A followed by a subsequent selection for DrCaspase-B. By using morpholino oligonucleotide–based knockdown assays, the DrNLRP1 inflammasome was verified to play important functional roles in antibacterial innate immunity in vivo. These observations demonstrate that the NLRP1 inflammasome originated as early as in teleost fish. This finding not only gives insights into the evolutionary history of inflammasomes but also provides a favorable animal model for the study of NLRP1 inflammasome-mediated immunology and diseases.
Introduction
Inflammasomes are cytoplasmic molecular platforms that trigger the activation of inflammatory caspases (such as caspase-1/4/5/11) and processing of proinflammatory cytokines (such as IL-1β and IL-18) (1, 2). Members of the NOD-like receptor (NLR) family, including NLRP1, NLRP3, NLRP6, NLRP7, and NLRC4, and the apoptosis-associated speck-like protein containing caspase activation and recruitment domain (ASC) adaptor protein are core components of the inflammasomes that link microbial and endogenous signals to effector caspase-1 (3, 4). After being activated in mammalian cells, ASC is used to form a single, compact speck structure in the cytoplasm, which is important for the oligomerization and autohydrolyzation of caspase-1 in certain inflammasomes. According to the requirement of ASC, inflammasomes are classified into ASC-dependent, such as mammalian NLRP3 and AIM2, and ASC-independent types, such as mammalian NLRP1b and NLRC4 (5–8).
Among the numerous inflammasomes that have been identified in humans and other mammalian species, the NLRP1 inflammasome is the first and most extensively studied because of its critical roles in innate immunity and pathogenesis of various diseases (9, 10). NLRP1, also known as NALP1, DEFCAP, and CARD7, was initially identified from human (Homo sapiens) macrophages (11). The human NLRP1 (HsNLRP1) is structurally characterized by the presence of an N-terminal pyrin domain (PYD), a NAIP, CIIA, HET-E, and TP1 (NACHT) nucleotide-binding domain, a leucine-rich repeat (LRR) domain, a function to find domain (FIIND), and a C-terminal caspase recruitment domain (CARD) (12–14). Differing slightly from HsNLRP1, the murine NLRP1s lack the N-terminal PYD but retain the other domains, including the C-terminal CARD (15). The N-terminal PYD of HsNLRP1 was reported to be an autoinhibitory domain for the regulation of HsNLRP1 inflammasome activation (16, 17). In general, the human and murine NLRP1s directly recruit caspase-1 by homotypic CARD–CARD interaction, allowing mammalian NLRP1s to activate the inflammatory caspases and the downstream IL-1β maturation in an ASC-independent manner (18, 19).
Although their occurrence and existence in humans and other mammalian species are numerously investigated, those of inflammasomes in ancient vertebrates, such as teleost fish, remain poorly understood. Several previous studies have shown that there are two homologs of proinflammatory caspase, namely Danio rerio Caspase (DrCaspase)–A (Caspy) and DrCaspase-B (Caspy2) in zebrafish (20, 21). The pro–IL-1β proteins in fish (as well as in amphibian and bird) lack a conserved caspase-1 recognition site, implying the existence of a unique mechanism for IL-1β maturation in lower vertebrates (22, 23). In fact, both DrCaspase-A and DrCaspase-B participate in the cleavage of zebrafish pro–IL-1β (pro–DrIL-1β) but with different specificities. The DrCaspase-A first cleaves the pro–DrIL-1β at the D104 residue; then, the pro–DrIL-1β was converted into a partially processed form of ∼20 kDa. Thereafter, the 20 kDa mediate-formed zebrafish IL-1β (DrIL-1β) was further cleaved by DrCaspase-B at the D122 and transformed into a fully processed mature form with a molecular mass of ∼18 kDa (24). However, whether the activation of DrCaspase-A and DrCaspase-B depends on an upstream inflammasome is still unclear. According to several studies on zebrafish, the effector caspases, such as DrCaspase-A, are recruited by D. rerio ASC (DrASC) through the PYD domain. This finding suggests that DrASC can only be associated with NLRs by the CARD domain left over from the two domains (25). Among the NLR family members, NLRP1 is the only member that contains a C-terminal CARD domain (26). Thus, whether an NLRP1 homolog and an NLRP1 inflammasome exist in zebrafish is an interesting topic of research. Moreover, the maturation of DrIL-1β relies on the close cooperation between DrCaspase-A and DrCaspase-B, thus posting the question of how these two caspases were coordinated in such an NLRP1 inflammasome. This coordination is a unique feature distinct from that of mammalian inflammasomes, in which usually only one caspase (caspase-1) is included.
In the current study, we identified an NLRP1 homolog, D. rerio NLRP1 (DrNLRP1) and a DrNLRP1-associated inflammasome from zebrafish (27, 28). DrNLRP1 was characterized by a number of conserved structural architectures and functional roles in DrASC nucleation, inflammasome formation, proinflammatory caspases activation, and DrIL-1β maturation, which were typically seen in mammalian NLRP1 counterparts. We also found that the DrNLRP1 inflammasome plays an important role in the innate defense against bacterial infection. Importantly, we proposed a sequential activation model to explain how a pair of proinflammatory caspases (DrCaspase-A and DrCaspase-B) is organized and activated in DrNLRP1 inflammasome instead of one (caspase-1) in mammals. In this model, the recruitment and activation of the two DrCaspases in DrNLRP1 inflammasome may be in an alternate manner, with a preference for DrCaspase-A, followed by a subsequent selection of DrCaspase-B. To our knowledge, this study is the first to identify an inflammasome from teleost fish, providing a cross-species understanding of the evolutionary history of inflammasome immunology.
Materials and Methods
Experimental fish and embryo
Wild-type AB zebrafish (D. rerio) were bred and maintained in circulating water at 28°C under standard conditions as previously described (29). The fish were held in the laboratory for at least 2 wk before the experiments. Male and female zebrafish with body lengths of 3–4 cm and weights of 0.5–1.0 g that exhibited healthy appearance and activity were used for the study. Zebrafish embryos were collected at different stages of embryonic development according to previously established protocols (30). The experiments were conducted in accordance with legal regulations and ethical approval.
Bacterial strain
An Edwardsiella tarda (TL5m) strain isolated from Trionyx sinensis was kindly provided by Prof. J. Shen (Zhejiang Institute of Freshwater Fisheries, Huzhou, China). The E. tarda strain was inoculated from 15% (v/v) glycerol stock cultures stored at −80°C, and it was grown in tryptic soy broth (0.85% tryptone, 0.15% Soytone, 0.25% NaCl, 0.125% K2HPO4, and 0.125% glucose, pH 7.5). The liquid cultures were cultured for 8 h in a 30°C shaker to reach the exponential phase (OD600 = 0.6) (31). CFU assay of the bacteria was performed by plating dilutions of the culture on tryptic soy agar.
Molecular cloning
The Genome and Expressed Sequence Tags databases maintained by the National Center for Biotechnology Information (NCBI), the University of California Santa Cruz, and Ensembl were used to predict NLRP1 homolog in zebrafish as previously described (32). Total RNA was isolated from zebrafish embryos and tissues by using an RNAiso Plus kit (Takara Bio). The cDNAs of DrNLRP1 were amplified by RT-PCR according to the homologous sequences predicted before. Full-length encoding sequence of DrCaspase-A, DrCaspase-B, DrASC and D. rerio IL-1β (DrIL-1β) were generated on the basis of the published sequences in the NCBI genome database (accession no. NM_131505.2, NM_152884.2, NM_131495.2, AY340959.1). The primers used in cloning are listed in Supplemental Table I. PCR products were purified and inserted into the pGEM-T EASY vector (Promega) as previously described (33).
Bioinformatics analysis
The Map Viewer in the NCBI was used for retrieving the genome assemblies and locations (34). By comparing DrNLRP1 cDNAs with genome sequences, gene organizations (intron/exon boundaries) were elucidated and drawn by using GeneMapper 2.5. Multiple alignment of DrNLRP1 was analyzed by using the Clustal X program (version 2.0). Phylogenetic trees were constructed by using MEGA 5.0 with the neighbor-joining or maximum likelihood (ML) method (35). The ML tree was also constructed using IQ-TREE web server (http://iqtree.cibiv.univie.ac.at) with default settings (36). The potential functional motifs were predicted using the Pfam 31.0 and Conserved Domains Database of NCBI. The domain structures of DrNLRP1 were analyzed using SWISS-MODEL, and the potential DrNLRP1 protein structure was predicted using I-TASSER (37). The tertiary structural figures were reviewed and colored in PyMOL software. The evolutionary correlation was assessed by coevolutional coefficient, which was performed by Pearson correlation analysis in SPSS (version 22) software (38).
Plasmid constructions
The open reading frame of DrNLRP1 was inserted into pCMV (Stratagene) to construct eukaryotic expression vectors with Flag- and HA-tags. The encoding sequences of DrNLRP1 with deletion of NACHT, FIIND, and CARD domains were constructed into pCMV-Tag2B and were named as pCMV-DrNLRP1-ΔNACHT, pCMV-DrNLRP1-ΔFIIND, and pCMV-DrNLRP1-ΔCARD. In addition, the DrASC, DrCaspase-A, DrCaspase-B, and DrIL-1β precursor (pro–DrSupplemental Table I. The quick mutation site-directed mutagenesis kit (Beyotime) was used to construct a quinary mutant of DrCaspase-A (named as DrCaspaseA-5DA), in which five asparagic acids (D106, D282, D293, D294, and D295) were substituted by alanines. Plasmids for transfection and microinjection were prepared free of endotoxin by using the Endo-Free Plasmid Mini Kit II (Omega Bio-Tek).
Quantitative real-time PCR for expression analysis
The transcripts of DrNLRP1 and its coordinator (DrASC, DrCaspase-A, DrCaspase-B) genes in zebrafish tissues and embryos were analyzed via quantitative real-time PCR (qRT-PCR) on a Mastercycler ep realplex instrument (Eppendorf). In brief, all PCR experiments were performed in a total volume of 10 μl by using a SYBR Premix Ex Taq kit (Takara Bio). The reaction mixtures were incubated for 2 min at 95°C, followed by 40 cycles of 15 s at 95°C, 15 s at 60°C, and 20 s at 72°C. The relative expression levels were calculated using the 2−∆Ct and 2−∆∆Ct method with β-actin for normalization. Each PCR trial was run in triplicate parallel reactions and repeated three times. The related primers are listed in Supplemental Table I. The primer efficiency was checked.
Preparation of polyclonal Ab
Ab against DrIL-1β was produced by an antigenic epitope-based protocol as previously described (39). Briefly, the epitope sequences of DrIL-1β were predicted through ABCpred, BepiPred, and IEDB online software. The hydrophilic and Ag indices were evaluated by using DNAStar. The predicted epitope peptide of DrIL-1β (DRKDTERIINFELC) was chemically synthesized and coupled to keyhole limpet hemocyanin (KLH) at a ratio of 10:10 mg (carrier/peptide) by InvivoGen. Subsequently, 6-wk-old New Zealand White rabbits with a weight of 1.5–2.0 kg were immunized with the synthetic peptides (0.5 mg/kg) in CFA or IFA four times (40). Antiserum was collected after the last immunization when Ab titer reached above 1:10,000 as determined by microplate-based ELISA. The specificity of the Ab was further validated through Western blot analysis.
Constitution of DrNLRP1 inflammasome in HEK293T cells
HEK293T cells were seeded into six-well plates at 5 × 105 per well in DMEM culture medium (HyClone) with 10% FBS (Bovogen Biologicals) at 37°C in 5% CO2. After 24 h, cells were transfected with plasmids expressing pro–DrIL-1β (800 ng), DrCaspase-A (200 ng), DrCaspase-B (200 ng), DrASC (200 ng), and Dr41). Twenty-four to forty-eight hours later, cell lysates were used for Western blot or caspase fluorogenic assay.
Caspase assay with fluorogenic substrates
HEK293T cells (one well in a six-well plate) or zebrafish embryos (∼20 embryos) were harvested and lysed with 100 μl of Caspase Cell Lysis Buffer (Enzo Life Sciences). A total of 100 μg of protein lysate was added to the Caspase Assay Buffer (Enzo Life Sciences) containing 100 μM of acetyl-Tyr-Val-Ala-Asp-amido-4-trifluoromethylcoumarin (Ac-YVAD-AFC) (specific to DrCaspase-A in zebrafish) or acetyl-Trp-Glu-His-Asp-AFC (Ac-WEHD-AFC) (specific to DrCaspase-B in zebrafish) (Alexis, San Diego, CA) as described (20, 24, 25, 42, 43). After incubation at 37°C for 2 h, the cleavage of caspase-type-specific substrate emitted a fluorescent signal that was measured with excitation at 400 nm and emission at 505 nm on a Synergy H1 Hybrid Reader (BioTek Instruments). The activation proportion of caspases was calculated by [(experimental group − control group)/control group] × 100%. The control group included lysates from cells that received no transfection.
Western blot analysis
HEK293T cells or zebrafish embryos were treated with cell lysis buffer for Western blot and IP (Beyotime) containing protease inhibitor mixture (Roche) (44). The proteins were separated by 12% SDS-PAGE and then transferred onto PVDF transfer membranes (Millipore Sigma). The blots were blocked with 5% nonfat dry milk (BioBasic), following by incubating with mouse anti-Flag/Myc/HA mAbs (Abcam). After adding HRP-conjugated goat anti-rabbit/mouse IgG Ab (Abcam), the objective proteins were visualized with ECL reagents (GE Healthcare) by a digital gel image analysis system (Tanon). The grayscale quantization of the protein stripe was performed by ImageJ software.
Coimmunoprecipitation assay
Coimmunoprecipitation (Co-IP) was performed to detect the interaction among DrNLRP1, DrASC, DrCasapse-A, and DrCaspase-B. For this, HEK293T cells were plated in 10-cm dishes (Corning) and were cotransfected of 6 μg recombinant eukaryotic plasmids or empty vector as negative control. At 48 h posttransfection, cells were lysed with precooling cell lysis buffer as described above. Lysates were incubated with mouse or rabbit Abs (1:200 dilution) at 4°C overnight. The next day, the mixture of cell lysates and Ab was incubated with 50 μl protein A–agarose beads (Roche) for 4 h. The beads were washed three times with lysis buffer, and the obtained samples were analyzed by Western blot assay (45). Expression of the transfected plasmids was also analyzed in the whole cell lysates as an input control.
Immunofluorescence analysis of DrNLRP1-dependent DrASC nucleation
A density of 1 × 105 HEK293T cells per well was seeded on coverslips in a 24-well plate. After 24 h, cells were transfected with plasmids expressing DrNLRP1 (400 ng/ml), DrASC (100 ng/ml), and DrCaspases (100 ng/ml) (4647). DrNLRP1-dependent DrASC nucleation was quantified by calculating the DrASC speck-forming rate [(number of cells with specks/number of all the cells) × 100%]. More than 100 cells with DrASC specks from immunofluorescence images were counted in each experimental group.
Electroporation for ZF4 cells transfection
6 cells were suspended in 200 μl of DMEM-F12 in a 0.4-cm electroporation cuvette, and 20 μg of plasmid DNA was added (48). After electroporation (square wave pulses, 270 V, 25 ms, MicroPulser electroporator; Bio-Rad), the cells were suspended into 1 ml of DMEM-F12 medium and transferred onto two wells of a 24-well plate containing coverslips. At 30 h posttransfection, ZF4 cells were fixed with 2% paraformaldehyde and analyzed by immunofluorescence staining as described above.
Intracellular bacterial infection/stimulation model
For examination of DrCaspases activation in vivo, a bacterial infection/stimulation model was developed by using zebrafish embryos. For this, E. tarda, an intracellular virulent pathogen for various aquatic animals, was cultured and reached the exponential phase (OD600 = 0.6) with a concentration of 109 CFU/ml. The bacteria solution was centrifuged at 1700 × g for 3 min, and E. tarda cells were harvested in the precipitate and resuspend by PBS (pH 7.4). Six and forty-eight hours postfertilization (hpf), zebrafish embryos were exposed to 1 × 108 CFU/ml E. tarda in a 10-cm dish (49). After immersion in the bacterial suspension for 40 min to 4 h, the embryos were collected and detected the activation of DrCaspases by Ac-YVAD/WEHD-AFC and the maturation of DrIL-1β by the rabbit anti–DrIL-1β polyclonal Ab. DNA from E. tarda was extracted with a bacterial DNA preparation kit (Omega Bio-Tek). DNA (200 pg/embryo), LPS (Escherichia coli O55:B5, 2 ng/embryo; Sigma-Aldrich), and muramyl dipeptide (MDP) (2 ng/embryo; InvivoGen) were microinjected into 6-hpf embryos (50). After 40 min, embryos were collected, and the activation of DrCaspases and the maturation of DrIL-1β were detected.
Visualization of DrNLRP1 inflammasome in vivo
In vivo examination of DrNLRP1 inflammasome in response to bacterial infection was conducted in zebrafish embryos. For this procedure, pCMV-Tag2B-DrNLRP1 and pCMV-Myc-DrASC were comicroinjected into the one-cell stage embryos at a concentration of 100 pg/embryo. Then, the 48-hpf zebrafish larvae were infected by E. tarda (1 × 108 CFU/ml) for 4 h, collected, and sliced into 6-μm thick frozen sections by a freezing microtome (CM1950; Leica) (44). For visualization of DrNLRP1 inflammasome, immunofluorescent staining was performed on the sections as described above.
Morpholino oligonucleotide and capped mRNA
The morpholino oligonucleotides (MOs) against DrNLRP1 mRNA (NLRP1-MO: 5′-TGAGGTCAGTGGGTTTGATTGGACA-3′) and DrCaspase-A mRNA (CaspaseA-MO: 5′-CCATGTTTAGCTCAGGGCGCTG-3′) and the standard control MO (5′-CTCTTACCTCAGTTACAATTTATA-3′) were designed and synthesized by Gene Tools (Philomath, OR). The MOs were dissolved with nuclease-free H2O to 1 mM as stock solutions (51). For MO-resistant mRNA synthesis, DrNLRP1, DrCaspase-A, or DrCaspaseA-5DA full-length cDNA sequence was constructed into the pcDNA3.1 vector by using the primers as shown in Supplemental Table I. The capped mRNA was synthesized using the mMESSAGE kit (Ambion), purified with Mini Quick Spin RNA columns (Roche), and then solubilized in DEPC water. The one-cell stage embryo was microinjected with NLRP1-MO (1.5–4.5 ng) or CaspaseA-MO (3.0 ng) for gene knockdown and MO-resistant mRNA (200 pg) for the sake of rescue. The phenotype of zebrafish embryos were visualized via an inverted microscopy (Zeiss Axiovert 40 CFL; Carl Zeiss).
Functional evaluation of DrNLRP1 inflammasome in innate antibacterial immunity
The functional role of DrNLRP1 inflammasome in innate immunity was evaluated through its antibacterial activity in zebrafish embryo model. For this evaluation, the DrNLRP1 or DrCaspase-A knockdown, rescue, and control embryos were challenged with E. tarda (1 × 108 CFU/ml) at 6 or 48 hpf. Mortality in each group was monitored during the 12 h period at one interval. The relative survival rate (RSR) was calculated using the following formula: RSR (%) = (survival rate of the infected group/survival rate of the mock PBS-administered control group) × 100%. Infection experiments were performed in triplicate with 100 embryos per group.
Statistical analysis
The data in the study were presented as the mean ± SD of each group. Statistical significance between experimental and control groups was assessed by two-tailed Student t test and was considered at *p < 0.05, **p < 0.01, and ***p < 0.001. The sample number for each group of zebrafish exceeded 20–100 embryos. More than 100 cells with immunofluorescence speck were counted for quantifying DrASC nucleation. All experiments were replicated at least three times.
Results
Identification of DrNLRP1 gene
With HsNLRP1 NACHT domain or Mus musculus NLRP1b (MmNLRP1) full-length sequences as queries, a homologous NLRP1 gene (DrNLRP1) was retrieved from zebrafish genome and Expressed Sequence Tags databases by using Genscan and BLAST programs. Moreover, by searching DrASC CARD sequence in zebrafish database, DrNLRP1 was also the first one in the result list (Fig. 1A). DrNLRP1 gene comprised 13 exons and 12 introns and was located on chromosome 2 within a 19.2 kb genomic fragment (Fig. 1B). Genes adjacent to HsNLRP1 locus (i.e., SDHC, PFN1, HES6, UBA52, and PER2) at chromosome 17 were found to be clustered around DrNLRP1 gene, although several gene loci, such as UBA52 and PER2, were in a reverse order (Fig. 1C). These observations indicated an overall coincident chromosomal synteny of NLRP1 gene on human and zebrafish genomes. In order to further analyze the phylogenic relationship of DrNLRP1, the ML tree was constructed with bootstrap analysis (1000 replicates) and SH-aLRT branch test. The DrNLRP1 was found to be grouped with mammalian NLRP1 counterparts with high bootstrap support (Fig. 2), verifying that DrNLRP1 is homology with regard to vertebrate NLRP1. The cloned DrNLRP1 cDNA consists of 5278 bp comprising a 183 bp 5′-untranslated region, a 1027 bp 3′-untranslated region, and a 4068 bp open reading frame that encodes 1355 aa (Supplemental Fig. 1, GenBank accession no. MH118554). According to the above results, the identification of DrNLRP1 gene is reliable, and this DrNLRP1 was subjected to further characterization.
Molecular characterization of DrNLRP1 gene. (A) Schematic of the search process in the molecular cloning of DrNLRP1 gene. Dashed boxes represent the domains whose sequences were used as queries for the zebrafish database search. (B) Exon/intron organization of DrNLRP1 gene. Exons and introns are indicated by black boxes and lines, respectively. (C) Gene synteny and chromosomal location analysis of HsNLRP1 and DrNLRP1 and genes adjacent to NLRP1 loci on human chromosome 17 (top) and zebrafish chromosome 2 (bottom). Arrows indicate gene orientation.
Phylogenetic analysis of the relationship of NLRP1 between fish and other species. The ML tree was constructed by IQ-TREE web server with default settings and the autoselected substitution model. The credibility of each branch was estimated by bootstrap analysis (1000 replicates) and SH-aLRT branch test. The accession numbers for the sequences included in the phylogenetic analysis are listed as follows: H. sapiens NLRP1, AAH51787.1; Pan troglodytes NLRP1, JAA27823.1; Canis lupus familiaris NLRP1, XP 005619937.1; Ailuropoda melanoleuca NLRP1, XP 011215410.2; Hipposideros armiger NLRP1, XP 019520352.1; Physeter catodon NLRP1, XP 007102896.1; Rattus norvegicus NLRP1, NP 001139227.2; M. musculus NLRP1, NP 001155886.1; Phascolarctos cinereus NLRP1, XP 020831609.1; Charadrius vociferus NLRP1, KGL99904.1; Leptosomus discolor NLRP1, KFQ03745.1; Amazona aestiva NLRP1, KQL45381.1; Chelonia mydas NLRP1, EMP24960.1; Xenopus tropicalis NLRP1, XP 004911151.1; Xenopus laevis NLRP1, XP 018095534.1; DrNLRP1, XP 009297081.1; Fundulus heteroclitus NLRP1, JAR78043.1; Maylandia zebra NLRP1, XP 023011105.1; Branchiostoma belcheri NLRP1, XP 019646667.1; Strongylocentrotus purpuratus NLRP1, XP 011666919.1; Hyalella Azteca NLRP1, XP 018018216.1; Orbicella faveolata NLRP1, XP 020600786.1.
Structural characterization of DrNLRP1 protein
DrNLRP1 protein was predicted to have 1355 aa with a putative molecular mass of 154.1 kDa. It possesses an NACHT (246–463 aa), an LRR (752–913 aa), a FIIND (1003–1233 aa), and a CARD (1269–1350 aa) domain, which is similar to the domain architecture of mammalian NLRP1s (Fig. 1A). By SWISS-MODEL program, the individual domain (NACHT, LRR, and CARD) and the combined domain (NACHT plus LRR domains) of DrNLRP1 apparently exhibit overall conserved tertiary structures along with those of HsNLRP1 and MmNLRP1 (Fig. 3A, 3B). The DrNLRP1 NACHT domain contains hallmark motifs of mammalian NLRP1s, such as the ATP/GTPase-specific P-loop (Walker A motif; GxxGxGKS/T, where x is any residue) and the Mg2+-binding site (Walker B motif; hhhhD/E, where h is a hydrophobic residue). Moreover, the His552 residue of DrNLRP1 and the adjacent motif (FxHxxxQEF), which contribute to ATP or ADP binding and protein oligomerization, were also conserved with mammals (Supplemental Fig. 1). The LRR domain in DrNLRP1 contains a cap helix followed by six LRR, with each repeat bearing a structural unit containing a β-strand and an α-helix with a conserved amino acid sequence motif (LxxLxLxxN/CxL). The LRR form a well-defined “horseshoe”-shaped structural scaffold (Fig. 3C), which is believed to be for ligand recognition. In DrNLRP1 FIIND domain, there are Phe1109 and Ser1110 corresponding to the autolytic cleavage sites (Phe1212 and Ser1213) in HsNLRP1 for its activation. However, the residue His1186 in HsNLRP1 that plays an important role during the posttranslational modification event cannot be found in DrNLRP1 (Supplemental Fig. 1). The CARD domain of DrNLRP1 comprises a helical bundle with six antiparallel α-helices (H1 to H6) (Fig. 3D). Same as the mammalian counterparts, there are two conservative and charged residues in DrNLRP1 CARD domain (Supplemental Fig. 1). The negatively charged Asp1290 is on the slightly concave surface formed by the H2 and H3 helices, whereas the positively charged Arg1318 is on the convex surface formed by the H1 and H4 helices (Fig. 3D), which may contribute to the highly specific homotypic interaction of CARD domains through mutual recognition between concave and convex surfaces.
Structural characterization of DrNLRP1 protein. (A) Tertiary structures of the individual domains (NACHT, LRR, and CARD) of NLRP1s among human, mouse, and zebrafish predicted by SWISS-MODEL with crystal structures of NACHT (PDB ID: 5irl.1.A), LRR (PDB ID: 4im6.1.A), and CARD (PDB ID: 3kat.1.A) as models. (B) Tertiary structures of the combined domain (NACHT plus LRR domains) of NLRP1s among human, mouse, and zebrafish modeled by SWISS-MODEL with 5irm.1.B as template. (C) Domain architecture and tertiary structure of DrNLRP1 modeled by I-TASSER. The top five threading templates are 5irlA, 5irmA, 5irl, 4kxfB, and 4kxfK. (D) Analysis of the structure of the CARD domain in DrNLRP1. The top five threading templates used by I-TASSER are 2l9 mA, 3katA, 1dgnA, 2l9m, and 1z6tA. The multiple alignments of DrNLRP1, HsNLRP1, MmNLRP1, and DrASC are presented beside the charged residues (Asp1290 and Arg1318) of CARD domain.
Phylogenetic analysis of the components of DrNLRP1 inflammasome complex
Other potential components of DrNLRP1 inflammasome complex, including DrASC, DrCaspase-A, DrCaspase-B, and DrIL-1β, were cloned from zebrafish cDNA libraries in this study. The structure of DrASC, which is composed of conservative PYD and CARD domains, is similar to those in mammals. However, DrCaspase-A and DrCaspase-B contained PYD domains in their N-terminal regions, which was distinct from the mammalian counterpart (caspase-1) that contains the conventional CARD domain. The structure of DrIL-1β is overall similar to mammals, but two inflammatory cleavage sites (104D-X and 122D-X) in DrIL-1β were observed compared with only one site in mammalian IL-1β (Fig. 4A). As a result, shown in Fig. 4B, both DrCaspase-A and DrCaspase-B belong to the branch of inflammatory caspases in the phylogenetic tree, whereas these two caspases cannot be completely clustered together with mammalian caspase-1 or caspase-4/5/11. Thus, DrCaspase-A and DrCaspase-B belong to the inflammatory caspase family and are the paralogs of caspase-1/4/5/11 instead of the orthologs. Furthermore, the phylogenetic trees of caspase-1 and IL-1β possessed the same topological structure (Supplemental Fig. 2A, 2B). By counting the evolutionary distance and performing the Pearson correlation analysis, the coevolutional coefficient between caspase-1 and IL-1β was calculated as 0.925 (p < 0.01; a value exceeding 0.8 denotes a distinct correlation relationship), meaning that the two genes are relatively correlated in the evolutionary process (Supplemental Fig. 2C). Moreover, the coevolutional coefficient of ASC and NLRP1 in vertebrates is 0.901 (p < 0.01) (Supplemental Fig. 2D), and the two conservative and charged residues in DrNLRP1 CARD domain are also found in DrASC CARD domain (Fig. 3C), indicating that those two genes are correlated through CARD domains in evolutionary history. All results indicate that the components of DrNLRP1 inflammasome are conserved and coevolutional during vertebrate evolution.
Characterization of the components of DrNLRP1 inflammasome complex. (A) Schematic of the components of NLRP1 inflammasome complex in mammals (Mal.) and zebrafish. (B) Phylogenetic analysis of the relationship of inflammatory caspases from Arthropoda to Vertebrata. The phylogenetic tree was constructed by MEGA (version 5.0) using the ML method. The apoptotic caspases of vertebrates served as outgroups. The reliability of each node was estimated by bootstrapping with 2000 replications. The number before each species name indicates the order of caspase. For instance, 1-Homo_sapiens represents caspase-1 in human.
Involvement of DrNLRP1 in DrCaspase-A and DrCaspase-B activation
To evaluate whether a functional DrNLRP1 inflammasome can exist in cells, the involvement of DrNLRP1 in DrCaspase-A and DrCaspase-B activation was initially detected in HEK293T cells that naturally show minimal expression of NLRP1 and other inflammasome component molecules. For this procedure, the expression constructs for DrNLRP1, DrASC, DrCaspase-A, and DrCaspase-B were transfected into the cells in different combinations. The activation of DrCaspase-A and DrCaspase-B was first examined by using DrCaspase-A/B–specific fluorescent substrate (Ac-YVAD/WEHD-AFC). Slight activation of DrCaspase-A/B was observed in the cells transfected with pCMV–DrCaspase-A/B alone or cotransfected only with pCMV–DrASC. However, when DrCaspase-A/B and DrASC were coexpressed with DrNLRP1, the protease activity of DrCaspase-A/B was significantly upregulated by ∼75% for DrCaspase-A and 50% for DrCaspase-B in a DrNLRP1 dose-dependent manner (Fig. 5A, 5B). Notably, when DrNLRP1 was coexpressed with DrCaspase-A/B without DrASC, DrCaspase-A/B was still not activated. As shown in Fig. 5C, the DrNLRP1- and DrASC-mediated activation of DrCaspase-A was higher by 40% than that of DrCaspase-B. This increment indicates that DrNLRP1 and DrASC preferentially activate DrCaspase-A when DrCaspase-A and DrCaspase-B were expressed simultaneously. Next, Western blot analysis was performed to further characterize the involvement of DrNLRP1 in DrCaspase-A/B proteolytic activation. Results showed that the 45/47 kDa pro–DrCaspase-A/B was activated and self-cleaved into a 35 kDa protein (p35, the cleavage product of active DrCaspase-A/B) if cells were coexpressed with DrNLRP1 and DrASC; the p35 product of DrCaspase-A was relatively stronger than that derived from DrCaspase-B (Fig. 5D, 5E). By contrast, no cleavage of the precursor of DrCaspase-A/B was detected in cells without the expression of either DrASC or DrNLRP1. These results verify the functional role of DrNLRP1, which activates both DrCaspase-A and DrCaspase-B in an ASC-dependent manner, with a preference for activating DrCaspase-A to some extent.
DrNLRP1 activates DrCaspase-A and DrCaspase-B in a DrASC-dependent manner in HEK293T cells. (A) DrNLRP1 and DrASC activate DrCaspase-A detected by specific Ac-YVAD-AFC fluorescent substrate. (B) DrNLRP1 and DrASC activate DrCaspase-B detected by specific Ac-WEHD-AFC fluorescent substrate. (C) DrNLRP1 activates DrCaspase-A and DrCaspase-B simultaneously, which was detected by Ac-YVAD-AFC or Ac-WEHD-AFC fluorescent substrates. (D and E) Western blot assay of the autohydrolyzation of DrCaspase-A (D) and DrCaspase-B (E) when coexpressed with DrNLRP1 and DrASC. Caspase activity was detected and expressed as the fold induction over the control as described in Materials and Methods. Each data point shows the mean ± SD with three replicates. *p < 0.05, **p < 0.01.
DrNLRP1 inflammasome contributes to pro–DrIL-1β maturation
Based on the above results, DrNLRP1 is an initiator for organizing a DrNLRP1–DrASC–DrCaspase-A/B–mediated inflammasome (referred to as DrNLRP1 inflammasome), which may further contribute to the maturation of DrIL-1β. For clarification, the eukaryotic expression plasmids of DrNLRP1, DrASC, DrCaspase-A, DrCaspase-B, and pro–DrIL-1β were cotransfected into HEK293T cells in different combinations. As shown in Fig. 6, the pro–DrIL-1β protein (31 kDa) expressed in cells coexisted with several midforms (∼25–30 kDa), which were the cleavage products of several other proteases existing in cells, including neutrophil elastase, proteinase 3, cathepsins G and D, granzyme A, and matrix metalloproteinases. The pro–DrIL-1β can be partially processed into a 20 kDa product (the first cleavage product at D104 residue of pro–DrIL-1β) by the activated DrCaspase-A alone (Fig. 6A). However, the activated DrCaspase-B alone was ineffective for the pro–DrIL-1β cleavage (Fig. 6B). When pro–DrIL-1β was coexpressed with DrCaspase-A, DrCaspase-B, DrASC, and DrNLRP1, in which all the DrNLRP1 inflammasome components were orchestrated in the complex, the pro–DrIL-1β can be cleaved into an 18 kDa mature form (Fig. 6C). This outcome indicates that a functional DrNLRP1 inflammasome occurs in cells, finally contributing to pro–DrIL-1β maturation.
DrNLRP1 inflammasome contributes to partial maturation of pro–DrIL-1β. HEK293T cells were transfected with a pcDNA3.1–DrIL-1β construct alone or with pCMV–DrNLRP1, pCMV–DrASC, pCMV–DrCaspase-A, and pCMV–DrCaspase-B. At 24 h posttransfection, immunoblot analysis was performed on the cell lysates with mouse anti-Flag or anti-Myc mAb. (A) Cleavage of pro–DrIL-1β triggered by DrNLRP1–DrASC activated DrCaspase-A alone. (B) Cleavage of pro–DrIL-1β triggered by DrNLRP1–DrASC activated DrCaspase-B alone. (C) Cleavage of pro–DrIL-1β triggered by activated DrCaspase-A and DrCaspase-B. (D) Cleavage of pro–DrIL-1β triggered by activated aPYDCasB and bPYDCasA, which had the exchanged PYD domains. Blots were reprobed for GAPDH as a loading control. Bar charts in (C) and (D) showed the relative density of the cleavage product of DrCaspases and DrIL-1β in the blots. The results are representative of three independent experiments, as described in Materials and Methods.
DrNLRP1 triggers the formation of a classical inflammasome complex in vitro
Given that ASC nucleation (also called ASC speck) is a convenient upstream readout for inflammasome activation and can be easily observed under light microscopy, the DrNLRP1-triggered assembly of inflammasome complex was detected by the formation of DrASC speck through immunofluorescence examination. When DrASC or DrNLRP1 was expressed alone, the fluorescence was a weak signal that diffused throughout the cell (Fig. 7A). With the coexpression of DrASC and DrNLRP1, the fluorescent signal of DrASC became one bright spot in a paranuclear area (Fig. 7B). In the amplified images, the DrASC speck reached a size of ∼1–2 μm in diameter, and DrNLRP1 formed an outer ring around DrASC. These results indicate that DrNLRP1 promotes DrASC-aggregate speck formation. As further support, Co-IP assay was performed to examine the protein–protein interaction involved in DrASC nucleation. Interaction clearly existed between DrNLRP1 and DrASC but not among the CARD-lacking mutants (DrNLRP1–ΔCARD and DrASC–ΔCARD) (Fig. 8A). This observation suggests that DrNLRP1 interacts with DrASC through CARD domains, triggering DrASC nucleation to form a classical ASC-dependent inflammasome. Furthermore, the ΔFIIND and ΔCARD mutants of DrNLRP1 blocked the DrNLRP1-dependent DrASC nucleation, whereas the ΔNACHT mutant did not influence the formation of DrASC speck (Fig. 8B, Supplemental Fig. 3). Similarly, the ΔNACHT mutant can activate DrCaspase-A/B, whereas the ΔFIIND and ΔCARD mutants cannot (Fig. 8C, 8D). These results suggest that instead of the NACHT domain, the FIIND and CARD domains of DrNLRP1 contribute to the nucleation of DrASC and the activation of DrCaspase-A/B. Finally, zebrafish ZF4 cell line was used to verify the DrNLRP1-dependent DrASC nucleation. Typical DrNLRP1–DrASC specks clearly existed in ZF4 cells, which were coexpressed with DrNLRP1 and DrASC (Supplemental Fig. 3C). The ZF4 cells that owned the DrNLRP1–DrASC speck presented pyroptotic morphology with crushed nuclei and irregular membranes.
Colocalization of DrNLRP1 and DrASC. (A) Transient transfection of pCMV-Myc-DrASC or pCMV-Tag2B-DrNLRP1 in HEK293T cells and diffuse florescent signals were detected in the cells. (B) Transient transfection of pCMV-Myc-DrASC and pCMV-Tag2B-DrNLRP1 in HEK293T cells. Flag-DrNLRP1 signal (red) and Myc-DrASC signal (green) accumulate in the same speck (white arrowheads). Scale bar, 10 μm. Images were captured under a laser-scanning confocal microscopy (LSM-710; original magnification ×630; Carl Zeiss). The results are representative of three independent experiments.
Functional evaluation of the domains in DrNLRP1. (A) Co-IP assay shows that DrNLRP1 interacts with DrASC through CARD domain. The black lines indicate where parts of the images were joined from a single experiment. (B) The statistics of DrASC speck-forming rates induced by DrNLRP1 and its mutants. More than 100 cells with DrASC speck were counted in each experimental group to quantify DrNLRP1-dependent DrASC nucleation. The original immunofluorescence images are shown in Supplemental Fig. 3B. (C) Function of DrNLRP1 mutants in the activation of DrCaspase-A detected by the specific Ac-YVAD-AFC fluorescent substrate. (D) Function of DrNLRP1 mutants in the activation of DrCaspase-B detected by the specific Ac-WEHD-AFC fluorescent substrate. Values represent the mean ± SD calculated in triplicate cultures. *p < 0.05, **p < 0.01.
DrASCCARD-mediated DrASC nucleation
The ASC nucleation in mammals is started by an ASC filament with ASCPYD core and ASCCARD cluster, the ASCPYD core combines with NLRPs, and the ASCCARD cluster recruits caspase-1. For understanding the mechanism of DrASC nucleation, two mutant DrASC proteins that contained only the PYD domain (named as DrASC–PYD) and the CARD domain (named as DrASC–CARD) were constructed. Interestingly, overexpression of DrASC–PYD or DrASC–CARD in cells induces the filament (ASCPYD or ASCCARD) formation instead of the speck construction (Fig. 9A, 9B). The results represented the assembly of incomplete DrASC that can be induced by the interaction between PYD–PYD or CARD–CARD domains. The ASCCARD filament, not the ASCPYD filament, was colocalized with DrNLRP1 upon the introduction of DrNLRP1 (Fig. 9C, 9D), suggesting that DrNLRP1 is associated with DrASC by their CARD–CARD interaction. However, the formation of DrASCCARD filament (with or without introducing DrNLRP1) cannot activate DrCaspase-A or DrCaspase-B (Fig. 9E). Thus, the typical DrASC nucleation in DrNLRP1 inflammasome comprised the DrASCCARD core and DrASCPYD cluster, the latter of which contributes to the recruitment and activation of DrCaspases by their PYD–PYD interaction. Next, the tertiary structures of ASCPYD core (Protein Data Bank [PDB] identifier [ID]: 2N1F) and ASCCARD core (PDB ID: 5FNA) were analyzed between zebrafish and mammalian counterparts (Fig. 9F). The Asn14 and Arg38 residues in ASCPYD and the Asp26, Arg54, and Val33 residues in ASCCARD are predicted to be highly conserved from fish to mammals, suggesting that these amino acids may perform important functional roles in the ASC fibrillation. However, the mammalian ASC’s Tyr59 residue, which plays an essential role in the PYD-homotypic interaction, was replaced with Phe59 in DrASC. This replacement implies an evolutionary issue, which results in a slight difference of ASC filaments between fish and mammals.
DrNLRP1 triggers the DrASCCARD-mediated DrASC nucleation. (A) Immunofluorescence examination of DrASCPYD filament when transiently transfected by Myc-tagged DrASC–PYD (DrASC-ΔCARD) in HEK293T cells. (B) Immunofluorescence examination of DrASCCARD filament when transiently transfected by Myc-tagged DrASC–CARD (DrASC-ΔPYD) in HEK293T cells. (C and D) Flag-DrNLRP1 was colocalized with the DrASCCARD filament (D) but not with DrASCPYD filament (C). Images were captured under a laser scanning confocal microscope (LSM-710; original magnification ×630; Carl Zeiss). (E) DrNLRP1 coexpressed with DrASCCARD or DrASCPYD cannot activate DrCaspase-A/B when being detected by the specific Ac-YVAD/WEHD-AFC. (F) Tertiary structure of ASCPYD (PDB ID: 2N1F) and ASCCARD (PDB ID: 5FNA) filaments. The conserved and important residues in the fibrillation of ASCPYD and ASCCARD are colored and listed by the side. The data are the average caspase activity ± SD of three independent experiments (*p < 0.05, **p < 0.01).
DrNLRP1–DrASC complex recruits DrCaspase-A and DrCaspase-B
Given that ASC nucleation provides a platform for the recruitment of caspase-1, the correlation between DrASC and DrCaspase-A/B was further examined. The Co-IP assay showed that the interaction clearly existed between DrASC and DrCaspase-A (Fig. 10A). However, the interaction existing between DrASC and DrCaspase-B was slightly detectable (Fig. 10B) and occurred only with the coexpression of DrNLRP1 (Fig. 10C). These results suggest that DrASC exhibits preference for DrCaspase-A association. According to immunofluorescence analysis, both DrCaspase-A and DrCaspase-B were perfectly colocalized with DrASC depending on the coexpression of DrNLRP1 (Fig. 11). Nevertheless, when DrCaspase-A and DrCaspase-B were coexpressed with DrNLRP1 and DrASC in cells, these two DrCaspases were almost independently localized in the inflammasome, and only a few DrCaspase-A– and DrCaspase-B–colocalized inflammasomes can be detected (Fig. 12A). Moreover, the percentage of the DrCaspase-A–associating cells (∼23%) was higher than those of the DrCapsase-B–associating cells (∼9%). These observations suggest that DrCaspase-A and DrCaspase-B may be sequentially recruited into the DrNLRP1–DrASC complex, with a preferential recruitment of DrCaspase-A first, followed by the replacement of DrCaspase-B after DrCaspase-A was self-cleaved and released from the inflammasome. As a support for this concept, the detection of DrCaspase-A– and DrCaspase-B–colocalized inflammasomes may represent an intermediate form during the alternate transition of these two caspases in the inflammasome (Fig. 12A, 12B). However, the percentage of this DrCaspase-A/B–colocalized inflammasome was extremely low, suggesting rapid sequential replacement, which was only detectable in a narrow window period.
Protein interactions between DrASC and DrCaspases in HEK293T cells. (A and B) DrASC interacts with DrCaspase-A (A) but not DrCaspase-B (B). (C) Protein–protein interactions among DrCaspase-B, DrASC, and DrNLRP1. HEK293T cells were transfected with pCMV-Myc-DrASC, pCMV-Tag2B–DrCaspase-B, and pCMV-HA-NLRP1 for 48 h. Cell lysates were immunoprecipitated with rabbit anti-Flag Ab and analyzed by Western blot using mouse anti-Flag, anti-Myc, or anti-HA Ab against DrCaspase-B, DrASC, or DrNLRP1, respectively (top panel). Expression of the transfected plasmids was analyzed with anti-Flag, anti-Myc, or anti-HA Ab in the whole cell lysates (bottom panels).
Visualization of DrNLRP1–DrASC–DrCaspases inflammasome complex. (A) Diffuse florescent signal (top panel) emerged with the transient transfection of pCMV-Myc-DrASC and pCMV-Tag2B–DrCaspase-A in HEK293T cells. With the coexpression of DrNLRP1, DrASC recruited DrCaspase-A and formed specks in the cells (bottom panels with white arrowheads). (B) Diffuse florescent signal (top panel) emerged with the transient transfection of pCMV-Myc-DrASC and pCMV-Tag2B–DrCaspase-B in HEK293T cells. With the coexpression of DrNLRP1, DrASC also recruited DrCaspase-B and formed specks in the cells (bottom panels with white arrows). Scale bar, 10 μm (LSM-710; original magnification ×630; Carl Zeiss). The results are representative of three independent experiments.
Sequential activation of DrCaspase-A and DrCaspase-B in DrNLRP1 inflammasome. (A) Coexpression of DrNLRP1, DrASC, DrCaspase-A, and DrCaspase-B in HEK293T cells elicited the formation of DrCaspase-A (white arrowheads) or DrCaspase-B (white arrows) specks in the cells. (B) Coexpression of DrNLRP1, DrASC, bPYDCasA, and aPYDCasB in HEK293T cells elicited the formation of bPYDCasA (white arrowheads) or aPYDCasB (white arrows) specks in the cells. (C) Amino acid alignments of the PYD domains in DrASC, DrCaspase-A, and DrCaspase-B. (D) Tertiary structures of the PYD domains predicted by SWISS-MODEL. The same residues among DrASC, DrCaspase-A, and DrCaspase-B were red in color. The same residues between DrASC and DrCaspase-A or DrASC and DrCaspase-B were colored by orange or yellow, respectively. (E) Schematic of the five potential cleavage sites (colored by red) of DrCaspase-A. (F) Coexpression of DrNLRP1, DrASC, CaspaseA-5DA, and DrCaspase-B in HEK293T cells triggered the formation of CaspaseA-5DA (white arrowheads) specks but blocked the nucleation of DrCaspase-B in the cells. The rates of cells developing the ASC speck are marked on the upper-left corner of each panel, as conducted by calculating more than 100 cells from immunofluorescence images in each experimental group. Data are represented as the mean ± SD for three coverslips. Scale bar, 10 μm (LSM-710; original magnification ×630; Carl Zeiss).
Sequential activation of DrCaspase-A and DrCaspase-B by PYD–PYD interaction
For further support, the possible mechanism underlying the sequential activation of DrCaspase-A and DrCaspase-B in DrNLRP1 inflammasome was explored. In multiple alignment analysis, the sequence identity of the PYD domains between DrASC and DrCaspase-A was 82.92%, and the identity of the domains between DrASC and DrCaspase-B was 55.46%. Thus, we supposed that the preferential association of DrCaspase-A with the DrNLRP1 inflammasome may be determined by the increased sequence homology between the PYD of DrASC and DrCaspase-A. To clarify this hypothesis, two chimera molecules (bPYD-CasA and aPYD-CasB) were constructed, in which the PYD domains of DrCaspase-A and DrCaspase-B were replaced by each other. After bPYD-CasA, aPYD-CasB, DrASC, and DrNLRP1 were coexpressed in HEK293T cells, both bPYD-CasA and aPYD-CasB were colocalized with the DrASC speck individually (Fig. 12B, Supplemental Fig. 3D); these two caspases were capable of self-cleaving into a 35 kDa cleavage product (Fig. 6D). However, the percentage of the aPYD-CasB-DrASC speck-containing cells (∼18%) exceeded that of the bPYD-CasA-DrASC speck-forming cells (∼6%), and the self-cleavage activity of aPYD-CaspB was more apparent than that of bPYD-CaspA. These observations indicated that the priority of DrCaspase-A in DrNLRP1 inflammasome was substituted by DrCaspase-B when their PYD domains were exchanged. As a result, the pro–DrIL-1β in bPYD-CasA, aPYD-CasB, DrASC, and DrNLRP1 coexpressed cells can only be processed into a faint 18 kDa mature form (Fig. 6D). It indicated that the processing of pro–DrIL-1β was significantly impaired with the disorder of the sequential activation of DrCaspase-A and DrCaspase-B in DrNLPR1 inflammasome. This outcome suggests that the sequence identity of PYD domains among DrASC and DrCaspase-A/B contributes to the preferential association and activation of the two caspases.
Structurally, a total of 65 and 38 aa in the PYD domains of DrCaspase-A and DrCaspase-B, respectively, were consistent with DrASC (Fig. 12C). These conserved amino acids formed potential interactive interfaces of PYD domains in DrASC, DrCaspase-A, and DrCaspase-B (Fig. 12D). Among these amino acids, a considerable amount of the hydrophobic residues, such as Leu16, Leu21, Ile49, Val57, and Ile75, and surface-charged residues, such as Arg22, Lys23, Glu43, and Asp50, were observed in DrASC and DrCaspase-A. It suggests that the hydrophobic and charge effects provided by these residues may play important roles in homotypic PYD interactions. Notably, the positively charged Lys23 and hydrophobic Ile49 residues, which were conserved between the DrASC and DrCaspase-A PYD domains, were absent in the DrCaspase-B PYD domain, implying that the two residues may largely contribute to the differential association of the two DrCaspases with DrNLRP1 inflammasome. Moreover, five potential autocleavage sites (D106, D282, D293, D294, and D295) of DrCaspase-A were predicted, and the asparagic acids in these sites were substituted by alanine to construct a quinary mutant (named as DrCaspaseA-5DA) for functional evaluation (Fig. 12E). Although being coexpressed with DrCaspase-B, DrASC and DrNLRP1, the DrCaspaseA-5DA majorly predominated the DrASC nucleation, and DrCaspase-B was mostly diffused in the cells (Fig. 12F). Thus, when DrCaspase-A cannot be autocleaved and released from DrNLRP1 inflammasome, DrCaspase-B lost the chance to combine with the molecular platform. This result supports for the sequential activation model of DrCaspase-A and DrCaspase-B in DrNLRP1 inflammasome. The schematic of the model is presented in Fig. 13. Evidently, DrNLRP1 proteins combine with DrASC through CARD–CARD interaction and form a DrASC fiber with a DrASCCARD core and a DrASCPYD cluster (Fig. 13A). The DrNLRP1–DrASC substructures interact with each other to assemble a speck structure in the cytoplasm with a diameter of 1–2 μm (Figs. 7B, 13B). The DrNLRP1–DrASC multiprotein complex recruits DrCaspase-A and DrCaspase-B by PYD–PYD interaction in a sequential manner. DrCaspase-A is the first to be activated in the complex, which cleaves pro–DrIL-1β at D104 into the midformed DrIL-1β. The second activated DrCaspase-B cleaves midformed DrIL-1β at D122 and finally results in DrIL-1β maturation (Fig. 13C).
Schematic diagram of DrNLRP1 inflammasome and the proposed sequential activation model for DrCaspase-A/B in DrNLRP1 inflammasome. (A) DrASC-dependent aggregation of DrNLRP1 inflammasome from side view. (B) DrNLRP1–DrASC–DrCaspase-A/B multiprotein complex and formation of the cytoplasmic speck from top view. (C) Sequential activation of DrCaspase-A and DrCaspase-B in DrNLRP1 inflammasome for the maturation of DrIL-1β in a DrCaspase-A– and DrCaspase-B–directed cleavage order.
Expression analysis of the components of DrNLRP1 inflammasome complex
To evaluate the existence and functional roles of DrNLRP1 inflammasome in vivo, the expression of DrNLRP1, DrASC, DrCaspase-A, and DrCaspase-B was initially analyzed in zebrafish tissues and embryos by qRT-PCR. All of the genes were ubiquitously expressed and distributed similarly in the examined immune-relevant tissues, such as spleen, head kidney, gill, and skin (Fig. 14A). These genes were also constitutively expressed in the early developmental embryos (2–96 hpf) (Fig. 14B–E). Infection of E. tarda (108 CFU/ml) for 40 min promoted the expression of these genes in zebrafish embryos as determined at 6 hpf (Fig. 14F). This event was followed by the activation of DrCaspase-A/B and the maturation of DrIL-1β when detected by caspase-specific fluorescent substrates and anti–DrIL-1β polyclonal Ab (Fig. 14G, 14H). This observation indicates that zebrafish embryo with intracellular bacterial infection is a favorable in vivo model system for the functional study of DrNLRP1 inflammasome.
Analyses for the expression of DrNLRP1 inflammasome–related molecules and the activation of DrNLRP1 inflammasome in zebrafish in response to bacterial infection. (A) qRT-PCR analysis of the expression patterns of DrNLRP1 inflammasome-related genes in adult zebrafish tissues. (B–E) Expression patterns of DrNLRP1 (B), DrASC (C), DrCaspase-A (D), and DrCaspase-B (E) in zebrafish embryos at different developmental stages. (F and H) E. tarda infection induced DrNLRP1 inflammasome activation in 6-hpf zebrafish embryo, as examined by the fold change of the mRNA levels (F), the activation of DrCaspase-A/B (G), and the maturation of DrIL-1β (H). Immunoblot analysis was performed on the lysate of embryos for DrIL-1β by using a rabbit polyclonal Ab and was reprobed for GAPDH as a loading control. The relative expression levels (A–E) were calculated using the 2−ΔCt method with β-actin for normalization,16 and the fold change (F) was calculated by the method of 2−∆∆Ct. Data points represent the mean ± SD from three independent experiments (**p < 0.01, ***p < 0.001).
In vivo determination of DrNLRP1 inflammasome
Next, the structural and functional characterizations of DrNLRP1 inflammasome were further examined in zebrafish embryo model to provide in vivo evidence. For this purpose, direct visualization of the inflammasome formation and functional evaluation of DrNLRP1 inflammasome-mediated DrCaspase-A/B activation and DrIL-1β maturation were initially conducted. The eukaryotic expression vectors of DrNLRP1 and DrASC were comicroinjected into the one-cell stage embryos at a concentration of 100 pg/embryo. The DrNLRP1 inflammasome was clearly observed after 48 hpf when stimulated by E. tarda for 4 h, as determined by the typical DrASC speck structures colocalized with DrNLRP1 (Fig. 15A). Moreover, DrCaspase-A and DrCaspase-B activities were significantly upregulated with the overexpression of DrNLRP1 and DrASC, accompanied by the markedly increased maturation of DrIL-1β (Fig. 15B). These findings suggest that DrNLRP1 triggers the formation of classical inflammasome complexes in vivo, inducing the activation of endogenous DrCaspase-A and DrCaspase-B and leading to the maturation of DrIL-1β.
In vivo structural and functional determinations of DrNLRP1 inflammasome by overexpression or gene knockdown assays. (A) Coexpression of Flag-tagged DrNLRP1 and Myc-tagged DrASC in vivo triggers the nucleation of DrNLRP1–DrASC speck in zebrafish embryos. (B) Coexpression of DrNLRP1 and DrASC in vivo increased the activation level of DrCaspase-A/B and promoted the maturation of DrIL-1β under E. tarda infection. (C and D) Knockdown of DrNLRP1 by NLRP1-MO decreased the activation of DrCaspase-A (C) and DrCaspase-B (D) in 6-hpf embryos after E. tarda (E. tard) infection at 106 CFU/ml for 40 min. (E) MO-resistant DrNLRP1 mRNA rescued the activation of DrCaspase-A/B in 6-hpf embryos after E. tarda infection. (F) The phenotype of 48-hpf zebrafish larvae after gene knockdown and rescue. (G and H) The RSRs of 6-hpf embryos (G) and 48-hpf larvae (H) after E. tarda infection at 106 CFU/ml for 12 h. Zebrafish embryos were microinjected with standard-MO (Ctrl-MO), NLRP1-MO, or both with the corresponding mRNA [NLRP1 − (MO + mRNA)]. Data are representative of three independent experiments as mean ± SD (*p < 0.05, **p < 0.01).
For further clarity, antisense MO-based knockdown assays were performed in the 6-hpf embryos, in which DrCaspase-A and DrCaspase-B were significantly induced (p < 0.01) by E. tarda immersion infection (108 CFU/ml) for 40 min (Fig. 14G). The MO against DrNLRP1 was microinjected into the one-cell stage embryos with various concentrations (1.5–4.5 ng). The activation of DrCaspase-A and DrCaspase-B significantly decreased (from 55 to 15% and from 50 to 25%) in a MO dose-dependent manner (Fig. 15C, 15D). The impaired DrCaspase-A and DrCaspase-B activation was notably rescued (p < 0.01) by the MO-resistant DrNLRP1 mRNAs (Fig. 15E). In addition, the deficient phenotype of 48-hpf zebrafish larvae after knockdown was rescued by the DrNLRP1 mRNAs (Fig. 15F). As a result, DrNLRP1 knockdown disturbed the antibacterial immunity, leading to a significant decrease in survival compared with the control group. However, the high mortality of 6- and 48-hpf zebrafish embryos in DrNLRP1 morphants can be restored by the administration of MO-resistant mRNA (Fig. 15G, 15H). These results indicate the important function of DrNLRP1 inflammasome in the innate antibacterial immunity of zebrafish.
Similarly, the sequential activation model of DrCaspase-A and DrCaspase-B in DrNLRP1 inflammasome was evaluated by MO-based knockdown and MO-resistant mRNA rescue assays in the 6-hpf embryos under E. tarda infection as described above. The MO against DrCaspase-A (3.0 ng) and MO-resistant mRNAs (200 pg) was microinjected into the one-cell stage embryos in different combinations. As expected, the activation of DrCaspase-A significantly declined (from 60 to 10%; p < 0.01) in DrCaspase-A morphants, and the impaired DrCaspase-A activation was rescued by MO-resistant DrCaspase-A mRNA but not by MO-resistant DrCaspase-5DA mRNA (Fig. 16A). Meanwhile, the activation of DrCaspase-B was not decreased in DrCaspase-A morphants or in morphants rescued by MO-resistant DrCaspase-A mRNA but was significantly reduced (from 50 to 30%; p < 0.05) in the morphants rescued by DrCaspase-5DA mRNA, encoding a mutant DrCaspase-A that lost the autocleavage sites and occupied the DrNLRP1 inflammasome (Fig. 16A). This observation is consistent with that from HEK293T cells, again demonstrating that the recruitment of DrCaspase-B into DrNLRP1 inflammasome does not require the help of DrCaspase-A but is dependent on the cleavage and release of DrCaspase-A from the inflammasome. Correspondingly, the percentage of survival of zebrafish embryos with E. tarda infection was dramatically decreased in DrCaspase-A morphants, which was restored by the administration of MO-resistant DrCaspase-A mRNA but not by CaspaseA-5DA mRNA (Fig. 16B).
Sequential activation model of DrCaspase-A/B in DrNLRP1 inflammasome and the mechanism underlying DrNLRP1 inflammasome activation in response to bacterial infection in zebrafish embryos and HEK293T cells. (A) Sequential activation model examined by in vivo knockdown and rescue assays. Zebrafish embryos were microinjected with standard-MO (Ctr-MO), CaspaseA-MO (CasA-MO), CaspaseA-MO with the corresponding mRNA (CasA-mRNA), or CaspaseA-MO with CaspaseA-5DA mRNA (CasA5DA-mRNA). The activation of DrCaspase-A/B was determined in 6-hpf embryos after E. tarda infection at 108 CFU/ml for 40 min. (B) Sequential activation of DrCaspase-A/B evaluated by in vivo antibacterial activity against E. tarda infection through knockdown and rescue assays. Zebrafish embryos received various MOs and MO-resistant mRNAs. RSRs were determined by using 6-hpf embryos after E. tarda infection at 108 CFU/ml for 12 h. (C) Evaluation of bacterial LPS, MDP, and DNA for the activation of DrNLRP1 inflammasome by in vivo knockdown assay. Zebrafish embryos were microinjected with Ctr-MO or NLRP1-MO (P1-MO) and LPS, MDP, or bacterial DNA in different combinations. The activation of DrNLRP1 inflammasome was determined by the increased activity of DrCaspase-A/B using specific Ac-YVAD/WEHD-AFC fluorescent substrate (top) and IL-1β maturation using Western blot (bottom). (D) Evaluation of ROS for the activation of DrNLRP1 inflammasome in HEK293T cells coexpressed with DrASC and DrNLRP1 under treatment of ROS or N-acetyl-l-cysteine. DrNLRP1 inflammasome activation was determined through increased activity of DrCaspase-A by using specific Ac-YVAD-AFC fluorescent substrate. All data are representative of three independent experiments as mean ± SD (*p < 0.05, **p < 0.01).
Finally, the embryo model was used to explore how DrNLRP1 is activated by bacterial infection. We chose three typical pathogen-associated molecular patterns (PAMPs) from Gram-negative bacteria (LPS, MDP, and genomic DNA) that function as activators for certain mammalian inflammasomes to test whether they can activate DrNLRP1 inflammasome. In vivo stimulation and MO-based knockdown assays were used. Results showed that the activation of DrNLRP1 inflammasome was overall induced by bacterial LPS, MDP and DNA, as determined by the increased activities of DrCaspase-A/B and maturation of DrIL-1β in embryos administered with PAMPs compared with that of control embryos that did not receive any stimulation (Fig. 16C). By knockdown assay, the activation of DrCaspase-A/B and maturation of DrIL-1β were significantly abrogated (from 50 to 10% for DrCaspase-A, and from 40 to 10% for DrCaspase-B; p < 0.05) in DrNLRP1 morphants stimulated by MDP but not by LPS and bacterial DNA (Fig. 16C). These results suggest that MDP largely contributes to the activation of DrNLRP1 inflammasome in zebrafish. In addition, other inflammasomes may exist in zebrafish, which can be activated in response to bacterial LPS and DNA stimulation.
Discussion
NLRP1 inflammasome is one of the best-characterized inflammatory mediators in innate immunity and pathogenesis of various diseases in mammals (52–54). However, the existence of NLRP1 inflammasome in ancient vertebrates is poorly understood. In this study, an NLRP1 homolog (DrNLRP1) was characterized from zebrafish. Structurally, this DrNLRP1 shares an overall coincident gene organization and chromosomal synteny, as well as protein domain architecture and tertiary structure, with human and mouse NLRP1s. With the coexpression of DrNLRP1 or its mutants and DrASC in cells, DrNLRP1 is able to trigger the formation of a classical inflammasome complex, as determined by the assembly of the DrASC speck structure under fluorescence microscopy. The Co-IP assay reveals that the assembly of DrNLRP1 inflammasome depends on the CARD–CARD homotypic interaction between DrNLRP1 and DrASC proteins. In addition, both the FIIND and CARD domains of DrNLRP1 are necessary for DrASC nucleation, which is similar to that of mammalian NLRP1 inflammasome (16, 55). Functionally, DrNLRP1 can trigger the activation of both DrCaspase-A and DrCaspase-B while being detected by specific fluorescent substrates and Western blot assay. The activated DrCaspase-A/B is autohydrolyzed into a p35 product by the DrNLRP1 inflammasome. This finding is supported by a recent study, proving that the active caspase-1 is a p33/p10 dimer (56). Subsequently, the activated DrCaspase-A and DrCaspase-B work together in the maturation of DrIL-1β. The above results indicate that DrNLRP1 inflammasome acts as a platform for the recruitment and activation of proinflammatory caspases and pro–IL-1β, which is a hallmark functional performance typically seen in mammalian inflammasome.
Despite these similarities, DrNLRP1 inflammasome acquires distinct features from its mammalian counterparts. For example, the inflammatory DrCaspase-A and DrCaspase-B are both PYD-containing caspases instead of CARD-containing caspases in mammals (57), whereas DrNLRP1 owns a CARD that cannot directly bind with such caspases. Thus, the recruitment of proinflammatory caspases into DrNLRP1 inflammasome needs the help of the adaptor DrASC. In other words, DrNLRP1 inflammasome triggers the activation of DrCaspase-A and DrCaspase-B in an ASC-dependent way, whereas mammalian NLRP1 inflammasomes can activate caspase-1 in an ASC-independent way (58). Accordingly, DrNLRP1 triggers the aggregation of DrASC into a filament with DrASCCARD core and DrASCPYD cluster, which is opposite to that of mammalian ASC filament (5, 59). To explain how DrCaspase-A and DrCaspase-B are coordinated in DrNLRP1 inflammasome, we assumed that the two DrCaspases are recruited and activated in the inflammasome in a sequential manner, with a preference for DrCaspase-A and a subsequent choice for DrCaspase-B. This sequential activation is determined by the PYD–PYD homotypic interaction between the DrASC and DrCaspases. Several experimental lines apparently support this hypothesis. For instance, DrCaspase-A and DrCaspase-B always independently exist in DrNLRP1 inflammasome. However, when the autocleavage site of DrCaspase-A is depleted, the DrCaspase-B associated inflammasome is rarely visible. The activation of DrCaspase-B is extremely impaired in the HEK293T cells and E. tarda–infected embryos, in which DrCaspase-A is largely replaced by a mutant DrCaspase-A (DrCaspaseA-5DA) without any potential autocleavage sites. These findings indicate that the retention of DrCaspase-A in DrNLRP1 inflammasome inhibits the recruitment of DrCaspase-B into the inflammasome. In other words, DrCaspase-B is recruited into the inflammasome, followed by the cleavage and release of DrCaspase-A from the inflammasome, suggesting that these two DrCaspases are sequentially activated. In addition, the occurrence of the DrCaspase-A–associated inflammasome in cells is more apparent than the DrCaspase-B–associated one, suggesting that DrCaspase-A preferentially associates with the inflammasome. However, this preferential association of DrCaspase-A can be replaced by DrCaspase-B when the PYD domains of these two caspases are exchanged with each other. Given the higher similarity of the PYD domain of DrASC with DrCaspase-A compared with DrCaspase-B, the homotypic degree between the PYD domains governs the preference. By multiple alignment and homology modeling analyses, a number of hydrophobic and charged amino acid residues are conspicuously conserved on the extensive surfaces of PYD domains among DrASC, DrCaspase-A, and DrCaspase-B. These results suggest the involvement of hydrophobic and charge effects in homotypic PYD–PYD interactions (60, 61). Further study (such as site-directed mutagenesis) is needed to clarify this notion. To date, the mechanisms underlying homotypic interactions, such as PYD–PYD and CARD–CARD interactions and even others, are still poorly understood; thus, the DrNLRP1 inflammasome is anticipated to be a favorable model system for such kind of study.
The maturation of DrIL-1β depends on the cleavage of pro–DrIL-1β at D104 and D122 by activated DrCaspase-A and DrCaspase-B in a strict order (24, 62). Thus, we assume that the sequential activation of DrCaspase-A and DrCaspase-B in DrNLPR1 inflammasome enables the orderly and efficiently processing of pro–DrIL-1β. Actually, when the sequential activation of DrCaspase-A and DrCaspase-B was reversed by coexpressing DrNLRP1 and DrASC with bPYD-CasA and aPYD-CasB instead of the wild-type DrCaspase-A/B, only extremely faint 18 kDa mature DrIL-1β can be detected. This observation means that the processing of pro–DrIL-1β in this experiment is disorganized and significantly impaired. The prioritized activation of DrCaspase-A provides a guarantee for the full maturation of DrIL-1β. Given that the DrCaspase-A– and DrCaspase-B–like caspases (also referred to caspase-1A and caspase-1B) can be widely predicted in numerous other teleost fish, the cooperation of the two inflammatory caspases may be universal in fish species. Thus, our work may make a favorable guide for such a study in teleost species. Notably, two classes of caspase were known in zebrafish and other Cyprinidae fish. Inflammatory caspases, such as DrCaspase-A and DrCaspase-B, own a PYD domain in their N terminus, whereas the other apoptotic caspases, such as caspase-3/8/9, own a CARD domain in their N-terminal region (63, 64). This observation indicates that intracellular inflammatory and apoptotic signaling pathways may be occurred more independently in ancient vertebrates (at least in cyprinid fish), whose biological significance is worth further study.
Finally, the structural and functional characterizations of DrNLRP1 inflammasome were determined in zebrafish both in vitro and in vivo with the help of zebrafish ZF4 cells and embryos. As expected, typical DrNLRP1–DrASC inflammasomes were clearly detected when cells and embryos were overexpressed with DrNLRP1 and DrASC. Compared with HEK293T cells containing DrNLRP1 inflammasome exhibiting normal morphology, ZF4 cells containing DrNLRP1 inflammasome usually present pyroptotic morphology with nuclear condensation and plasma membrane rupture (65). This event may depend on the expression of endogenous pyroptosis-related molecules in ZF4 cells that are activated by DrNLRP1 inflammasome. In the embryo model, the DrNLRP1–DrASC nucleation triggers the endogenous DrCaspase-A and DrCaspase-B activation and DrIL-1β maturation under stimulation of E. tarda infection. These activities can be diminished by antisense MO-based knockdown of DrNLRP1 and rescued by MO-resistant mRNAs. In accordance with these observations, the high mortality of DrNLRP1 morphants in E. tarda immersion-infection model is restored by the administration of the MO-resistant mRNAs. These results verify the functional roles of DrNLRP1 inflammasome in antibacterial innate immunity in vivo. To investigate the mechanism of DrNLRP1 inflammasome activation in response to bacterial infection in vivo, we administer three typical PAMPs (LPS, MDP, and bacterial DNA) into zebrafish embryos. The MDP largely contributes to the activation of DrNLRP1 inflammasome. This result is consistent with the observation that MDP is a potent stimulator for mammalian NLRP1 inflammasome activation (66). Mammalian inflammasomes (such as NLRP1 and NLRP3) can be activated by various cellular metabolites, such as ATP, short-chain fatty acids, uric acid, cholesterol crystals, and reactive oxygen species (ROS) (67, 68), the latter of which frequently increasing during viral and intracellular bacterial invasions. Therefore, we test the potential stimulatory role of ROS in the activation of DrNLRP1 inflammasome. As expected, the activation of DrNLRP1-dependent DrCaspase-A in HEK293T cells is substantially stimulated when the cells are treated with H2O2 (0.1 mM) but is inhibited when the cells are treated with N-acetyl-l-cysteine (5 mM), a widely used ROS scavenger (Fig. 16D). These observations suggest that the alteration of cellular metabolic homeostasis (such as redox state) during bacterial infection may stimulate the activation of DrNLRP1 inflammasome. However, further study is needed to clarify this notion.
In addition to sensing foreign infectious or endogenous danger-related triggers, NLRP1 inflammasome must avoid spontaneous and aberrant activation to ensure the stabilization of cellular homeostasis. Therefore, the negative feedback regulatory mechanism underlying NLRP1 inflammasome activation is a particular research focus. A novel endogenous inhibitor of NLRP1 inflammasome, named as dipeptidyl dipeptidase 9 (DPP9), is recently identified from diverse primary cell types of humans and mice (Ref. 69 and F.L. Zhong, K. Robinson, C. Lim, C. R. Harapas, C. Yu, W. Xie, R. M. Sobota, V. B. Au, R. Hopkins, J. E. Connolly, S. Masters, and B. Reversade, manuscript posted on bioRxiv). This DPP9 shares a similar domain structure with other DPP-IV family members consisting of an N-terminal β-barrel (DPP-IV N) and a C-terminal S9 hydrolase domain, the latter of which binds to the autocleaving FIIND module and maintains NLRP1 in its monomeric and inactive state. Interestingly, a DPP9 homolog (D. rerio DPP9) is clearly predicted from zebrafish. This D. rerio DPP9 shares high sequence identity with human (79.79%) and mouse (78.49%) counterparts and possesses converted N-terminal β-barrel and C-terminal S9 hydrolase domains (data not shown), suggesting its potential functional involvement in the negative regulation of DrNLRP1 inflammasome activation. As a support, the FIIND domain in DrNLRP1 is critical for the assembly and activation of DrNLRP1 inflammasome. In summary, the DPP9-mediated NLRP1 regulatory pathway is likely to exist in zebrafish and is highly conserved from fish to mammals throughout the vertebrate evolution. Further clarification on this issue would aid in understanding the molecular and functional evolutionary history of NLRP1 inflammasome.
Collectively, our present study demonstrates that the NLRP1 inflammasome originates as early as in teleost fish. As an ancient protein apparatus, NLRP1 inflammasome in zebrafish possesses conserved structural hallmarks and functional roles in inflammatory responses, as well as some distinct features from its mammalian counterparts. Given the crucial roles of NLRP1 inflammasome in innate immunity and disease pathogenesis (70), our results would not only provide insight into the evolutionary history of inflammasomes but also present a new animal model for the investigation of NLRP1 inflammasome-mediated immunology and disorders.
Disclosures
The authors have no financial conflicts of interest.
Footnotes
This work was supported by grants from the National Natural Science Foundation of China (31630083, 31372554, 31472298, 31572641, 31272691), Stem Cell and Translational Research, the National Key Research and Development Program of China (2016YFA0101001), the Open Fund of the Laboratory for Marine Biology and Biotechnology, Qingdao National Laboratory for Marine Science and Technology, Qingdao, China (OF2017NO02), and the Zhejiang Major Special Program of Breeding (2016C02055-4).
The sequences presented in this article have been submitted to GenBank (http://www.ncbi.nlm.nih.gov/genbank/) under accession number MH118554.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- Ac-WEHD-AFC
- acetyl-Trp-Glu-His-Asp-AFC
- Ac-YVAD-AFC
- acetyl-Tyr-Val-Ala-Asp-amido-4-trifluoromethylcoumarin
- ASC
- apoptosis-associated speck-like protein containing caspase activation and recruitment domain
- CARD
- caspase recruitment domain
- Co-IP
- coimmunoprecipitation
- DPP9
- dipeptidyl dipeptidase 9
- DrASC
- D. rerio ASC
- DrCaspase
- Danio rerio caspase
- DrIL-1β
- zebrafish IL-1β
- DrNLRP1
- D. rerio NLRP1
- FIIND
- function to find domain
- hpf
- hour postfertilization
- HsNLRP1
- human NLRP1
- ID
- identifier
- LRR
- leucine-rich repeat
- MDP
- muramyl dipeptide
- ML
- maximum likelihood
- MmNLRP1
- Mus musculus NLRP1b
- MO
- morpholino oligonucleotide
- NACHT
- NAIP, CIIA, HET-E, and TP1
- NCBI
- National Center for Biotechnology Information
- NLR
- NOD-like receptor
- PAMP
- pathogen-associated molecular pattern
- PDB
- Protein Data Bank
- PYD
- pyrin domain
- qRT-PCR
- quantitative real-time PCR
- ROS
- reactive oxygen species
- RSR
- relative survival rate.
- Received April 4, 2018.
- Accepted July 26, 2018.
- Copyright © 2018 by The American Association of Immunologists, Inc.
References
In this issue
