Zebrafish (Danio rerio H.) were obtained from the Zebrafish International Resource Center and mated, staged, raised, and processed as described (24). The lines roy^a9/a9; nacre^w2/w2 (casper) (25), Tg(mpx:eGFP)ⁱ¹¹⁴ (26), Tg(mpeg1:eGFP)^gl22 and Tg(mpeg1:GAL4)^gl25 (27) have been previously described.

Zebrafish Caiap was identified by searching the CARD protein family (PF00619) in the PFAM database.¹ Zebrafish full-length Caiap sequence was then compared with other known Caiap sequences, obtained from The Universal Protein Resource (UniProt) database,² and with the newly identified variants by multiple sequence alignment carried out with the ClustalX version 2.1 program (28). The molecular weights were estimated using the Protein Molecular Weight tool from The Sequence Manipulation Suite.³ The domains of the proteins deduced from the nucleotide sequences were determined using the Simple Modular Architecture Research Tool (SMART), from the European Molecular Biology Laboratory (EMBL) website⁴ (29, 30). Finally, three-dimensional structure predictions were performed using The IntFOLD Integrated Protein Structure and Function Prediction Server⁵ (31) and visualized with The PyMOL Molecular Graphics System, Version 1.8 Schrödinger, LLC.⁶ The ModFOLD Quality Assessment Server (Version 4.0) was used to check the accuracy of the models (32).

The genes encoding zebrafish Caiap (NM_001025492), CaiapΔCARD (deletion from I12 to Y97), wild-type (WT) Flag-Caspa (NM_131505), catalytic inactive Flag-Caspa mutant (C230A) and WT Asc-eGFP (NM_131495) were synthesized by GenScript Corporation. The cmv/sp6:caiap-mCherry, cmv/sp6:caiapΔCARD-mCherry, and uas:caiap-mCherry; cmlc2:eGFP constructs were generated by MultiSite Gateway assemblies using LR Clonase II Plus (Life Technologies) according to standard protocols and using Tol2kit vectors described previously (33). The zebrafish Caspa and Asc-Myc expression constructs were previously described (18).

Specific morpholinos (Gene Tools) were resuspended in nuclease-free water at 1 mM (Table S1 in Supplementary Material). In vitro-transcribed RNA was obtained following the manufacturer’s instructions (mMESSAGE mMACHINE kit, Ambion). Morpholinos and RNA were mixed in microinjection buffer and microinjected into the yolk sac of one-cell-stage embryos using a microinjector (Narishige) (0.5–1 nl per embryo). The same amount of MOs and/or RNA was used in all experimental groups. The efficiency of the MOs was checked by assessing caspase-1 activity.

In some experiments, Tg(mpeg1:GAL4)^gl25 one-cell stage embryos were injected with a solution containing 100 pg uas:caiap-mCherry; cmlc2:eGFP construct and 50 pg Tol2 RNA in microinjection buffer (0.5× Tango buffer and 0.05% phenol red solution). Embryos were sorted at 2 dpf according to the presence or absence of green fluorescence in their heart before being infected (see below).

For crispant experiments, sgRNAs obtained by in vitro transcription using the MAXIscript T7 Kit (Ambion) were first checked in vitro using 100 ng of an amplicon containing the target sequence, 30 nM sgRNA and 30 nM EnGen^® Cas9 NLS from Streptococcus pyogenes (New England Biolabs). Injection mixes were then prepared with 500 ng/µl Cas9 and 100 ng/µl control (5′-CGTTAATCGCGTATAATACG-3′) or caiap (5′-GGGCCACACCGCTGTTGCTG-3′) sgRNA in 300 mM KCl buffer, incubated for 5 min at 37°C and used directly without further storage (34).

For most infection experiments, ST 12023 (wild type) and the isogenic derivative mutants for SPI-1/SPI-2 (prgH020:Tn5lacZY ssaV:aphT) (kindly provided by Prof. D. Holden) were used. For some experiments, the ST strains used were: 14028s (wild type) and its isogenic derivatives fliC/fljB mutant and FliCON, which persistently expresses the flagellin protein FliC (35, 36) (kindly provided by Dr. E.A. Miao). Overnight cultures in Luria-Bertani medium (LB) were diluted 1/5 in LB with 0.3 M NaCl, incubated at 37°C until 1.5 optical density at 600 nm was reached, and finally diluted in sterile PBS. Larvae of 2 dpf were anesthetized in embryo medium with 0.16 mg/ml tricaine and 10 or 50 bacteria were injected into the yolk sac (survival curves and caspase-1 activity, respectively) or 100 into the otic vesicle (WISH). Larvae were allowed to recover in egg water at 28–29°C, and monitored for clinical signs of disease or mortality over 5 days. At least three independent experiments were performed with a total number of 300 larvae.

Tail fin amputation was performed at 3 dpf as previously described (37) in casper larvae.

The caspase-1 activity was determined with the fluorometric substrate Z-YVAD-AFC (caspase-1 substrate VI, Calbiochem) as described previously (15, 16, 19). In brief, 25–35 larvae were lysed in hypotonic cell lysis buffer [25 mM 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), 5 mM ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), 5 mM dithiothreitol (DTT), 1:20 protease inhibitor cocktail (Sigma-Aldrich), pH 7.5] on ice for 10 min. For each reaction, 80 µg protein were incubated for 90 min at 23°C with 50 µM Z-YVAD-AFC and 50 µl of reaction buffer [0.2% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 0.2 M HEPES, 20% sucrose, 29 mM DTT, pH 7.5]. After the incubation, the fluorescence of the AFC released from the Z-YVAD-AFC substrate was measured with a FLUOstart spectrofluorometer (BGM, LabTechnologies) at an excitation wavelength of 405 nm and an emission wavelength of 492 nm. One representative caspase-1 activity assay out of the three carried out is shown accompanying each survival assay.

Approximately 300–500 non-infected and infected larvae from the lines Tg(mpx:eGFP)ⁱ¹¹⁴ and Tg(mpeg1:eGFP)^gl22 were anesthetized in tricaine at 24 hpi, minced with a razor blade, incubated at 28°C for 30 min with 0.077 mg/ml Liberase (Roche). The resulting cell suspension was passed through a 40-µm cell strainer. Cell sorting was performed on a FACSCalibur (BD Biosciences) and a SH800Z (Sony).

Total RNA was extracted from whole embryos/larvae, larval heads, or sorted cells with TRIzol reagent (Invitrogen) following the manufacturer’s instructions and treated with DNase I, amplification grade (1 U/μg RNA; Invitrogen). SuperScript III RNase H⁻ Reverse Transcriptase (Invitrogen) was used to synthesize first-strand cDNA with oligo(dT)18 primer from 1 µg of total RNA at 50°C for 50 min. Real-time PCR was performed with an ABI PRISM 7500 instrument (Applied Biosystems) using SYBR Green PCR Core Reagents (Applied Biosystems). Reaction mixtures were incubated for 10 min at 95°C, followed by 40 cycles of 15 s at 95°C, 1 min at 60°C, and finally 15 s at 95°C, 1 min 60°C, and 15 s at 95°C. For each mRNA, gene expression was normalized to the ribosomal protein S11 (rps11) content in each sample using the Pfaffl method (38). Non-infected samples were used as calibrator. The primers used are shown in Table S2 in Supplementary Material. In all cases, each PCR was performed with triplicate samples and repeated with at least two independent samples.

Transparent Casper embryos were used for WISH (39). caiap sense and antisense RNA probes were generated using the DIG RNA Labeling Kit (Roche Applied Science) from linearized plasmids. Embryos were imaged using a Scope.A1 stereomicroscope equipped with a digital camera (AxioCam ICc 3, Zeiss).

HEK293T cells (CRL-11268; American Type Culture Collection) were maintained in DMEM:F12 (1:1) supplemented with 10% FCS, 2 mM Glutamax, and 1% penicillin-streptomycin (Life Technologies). Plasmid DNA was prepared using the Mini-Prep procedure (Qiagen). DNA pellets were resuspended in water and further diluted, when required, in PBS. Cells grown on coverslips were transfected with Lipofectamine (Thermofisher), fixed with 4% paraformaldehyde in PBS, incubated 20 min at room temperature with 20 mM glycin, permeabilized with 0.5% NP40 and blocked for 1 h with 2% BSA. Cells were then labeled with anti-FLAG monoclonal (1:7,000) or anti-Myc polyclonal (1:2,000) both from (Sigma-Aldrich), followed by Alexa 488-conjugated secondary antibody (Thermofisher). Samples were mounted using a mounting medium from Dako and examined with a Leica laser scanning confocal microscope AOBS and software (Leica Microsystems). The images were acquired in a 1,024 × 1,024 pixel format in sequential scan mode between frames to avoid cross-talk. The objective used was HCX PL APO CS × 63 and the pinhole value was 1, corresponding to 114.73 µm.

Pull down assays were also performed as described previously (40) with small modifications. HEK293T cells were transfected with the indicated plasmids in each figure, washed twice with PBS, solubilized in lysis buffer (50 mM Tris–HCl, 150 mM NaCl, 1% NP40 and protease inhibitors) during 30 min in agitation and centrifuged (13,000 × g, 10 min). Cell lysate (1 mg) was incubated for 2 h at 4°C under gentle agitation with 40 µl of slurry of ANTI-FLAG^® M2 or Myc Affinity Gels (Sigma-Aldrich). The immunoprecipitates were washed four times with lysis buffer containing 0.15 M NaCl and then twice with PBS. Finally, the resin was boiled in SDS sample buffer and the bound proteins were resolved on 10 or 15% SDS-PAGE and transferred to nitrocellulose membranes (BioRad) for 50 min at 200 mA. Blots were probed with specific antibodies to FLAG (Sigma-Aldrich), Myc (ThermoFisher), and mCherry (TermoFisher), and then developed with enhanced chemiluminescence reagents (GE Healthcare) according to the manufacturer’s protocol. In some experiments, protein were resolved under non-reducing conditions by omitting SDS and β-mercaptoethanol for the loading buffer.

Data are shown as mean ± SEM and were analyzed by analysis of variance and a Tukey multiple range test to determine differences among groups. The differences between two samples were analyzed by the Student’s t-test. A log rank test with the Bonferroni correction for multiple comparisons was used to calculate the statistical differences in the survival of the different experimental groups. A chi-square test was used to determine differences in the number of specks formed by Caiap in HEK293T cells.

A PFAM search to identify proteins harboring CARD domains revealed the presence of Caiap (CARD-ANK Inflammasome Adaptor Protein) in the zebrafish. The caiap gene contains two exons and a single open reading frame encoding a putative polypeptide of 744 amino acids (predicted molecular mass of 80.9 kDa) with an N-terminal CARD domain and 16 C-terminal ANK repeats (Figure S1 in Supplementary Material; Figure 1A). Strikingly, uncharacterized orthologs of zebrafish Caiap were found in many organisms, including phylogenetically distant ray-finned fish species, cartilaginous fish (elephant shark), lobe-finned fish (coelacanth), amphibian, reptiles, birds, and marsupials. However, we failed to find a Caiap ortholog in placental mammals, invertebrates, protochordates (amphioxus and sea squirts), jawless fish (lamprey), and lung fish by using homology and synteny searches. Phylogenetic analysis confirmed that the origin of Caiap predated the split of fish and tetrapods more than 450 million years ago, suggesting, therefore, that Caiap was lost in lung fish and placental mammals during evolution (Figure 1B).

Caiap was seen to be well conserved across vertebrate species. Thus, zebrafish Caiap showed from 47 to 66% amino acid identity and from 64 to 69% amino acid similarity with Caiap from other ray-finned fish species (Table 1). In addition, a similar degree of conservation was found between zebrafish Caiap and those from other vertebrate groups, with the exception of marsupials (35% identity and 55% amino acid similarity) (Table 1). Multiple alignment of all Caiap identified revealed that the CARD domain and ANK repeats were the best conserved regions of the protein (Figure S1 in Supplementary Material). More interestingly, three-dimensional structural prediction revealed an identical tertiary structure of Caiap in all vertebrate groups (Figures 2A,B). In fact, all the structures perfectly fitted when superimposed (Figure 2C).

The expression profile of zebrafish Caiap was examined using RT-qPCR and WISH. It was found that the mRNA levels of caiap transcripts were maternally transferred, since they peaked at fertilization time and then rapidly declined (Figure 3A). In adult fish, caiap transcripts were detected in kidney, the hematopoietic organ of adult fish, heart, and skin, but not in muscle, ovary, gills, eye, or brain (Figure 3B). In addition, the mRNA levels of caiap were seen to be weakly higher 24 h post-infection (hpi) in the infection site of zebrafish larvae infected with ST (Figure 3C). As expected, the gene encoding pro-inflammatory IL-1β robustly increased in infected larvae (Figure 3C).

To further confirm the RT-qPCR results, WISH was performed in infected and wounded larvae. Although no positive cells were observed in non-infected larvae, a few small round caiap⁺ positive cells were found at the infection site (the otic vesicle), the number of positive caiap⁺ cells increasing from 4 to 24 hpi (Figure 4A). Similarly, a few caiap⁺ cells were also observed in the wound 24 h after transection of the tail fin tip (Figure 4B). As expected, no positive cells were observed when using the caiap sense probe (Figure S2 in Supplementary Material). This result suggested that both infection and wounding are able to induce the expression of caiap in immune cells recruited to the infection and wounding sites, namely, macrophages (41, 42) and neutrophils (19, 37, 43). Therefore, macrophages and neutrophils were sorted from mpeg1:eGFP (27) and mpx:eGFP (26) transgenic larvae, respectively, upon infection with ST and the transcript levels of caiap were analyzed by RT-qPCR in both cells types. The results showed that while caiap transcripts drastically increased in macrophages upon ST infection (Figure 3D), they remain unaltered in neutrophils upon infection (Figure 3E).

To further characterize zebrafish Caiap we used a morpholino (MO)-mediated gene knockdown strategy whereby the MO was able to bind the translation start of the caiap mRNA and, therefore, to inhibit its translation (Figure S1A in Supplementary Material). The efficiency of the MO was validated by using a caspase-1 activity assay with a fluorogenic substrate, Z-YVAD-AFC (Figures 5A,C), which has previously been shown to be processed by gilthead seabream caspase-1 (15, 19, 44) and zebrafish Caspa (18). Caiap morphant animals injected with 1 pg/egg MOs showed reduced caspase-1 activity but they developed normally and were viable (Figures 5C,D). However, Caiap-deficient larvae showed increased susceptibility to WT ST compared with their control siblings (Figure 5D) and impaired caspase-1 activity in response to the infection (Figure 5C). Additionally, infection susceptibility and decreased caspase-1 activity were both fully reversed by injection of non-targetable caiap mRNA, confirming the specificity of the MO (Figures 5C,D). In contrast, forced expression of CaiapΔCARD failed to rescue the high susceptibility and impaired caspase-1 activation of Caiap-deficient larvae (Figures 5E,F), suggesting that the CARD domain is indispensable for Caiap function. Furthermore, Caiap crispant larvae also showed higher susceptibility to ST infection and reduced caspase-1 activity upon infection than their WT siblings (Figure S3 in Supplementary Material).

Although forced ubiquitous expression of caiap mRNA alone barely increased the resistance to ST infection (Figures 5D, 6A and 7A) and caspase-1 activity (Figures 5C, 6B and 7B), forced expression of caiap in macrophages, using the macrophage-specific promoter mpeg1 (27), resulted in increased resistance to ST infection (Figure 5G). Notably, Caiap levels did not affect fish susceptibility (Figures 6A,C) or caspase-1 activity (Figures 6B,D) in response to a syngenic double ST mutant for SPI-1 and SPI-2, which contains a large number of genes encoding a type 3 secretion system that is required for virulence in mouse (45) and zebrafish (19), comparing with the controls. Moreover, the impact of Caiap on ST infection seemed to depend on flagellin, since the virulence of the flagellin mutant strain (FliC/FljB) of ST (35) was not affected by Caiap (Figures 6E,F), while the increased larval resistance and caspase-1 activity to ST which overexpresses flagellin (FliCON) (36) were found to be Caiap-dependent (Figures 6E,F). These results suggest that Caiap functions downstream of flagellin, although other PAMPs may also regulate Caiap activation.

Most inflammasomes require the adaptor protein ASC for their assembly, activation, and function. However, some NLRs that contain a CARD domain can bind directly to caspase-1. To this group belongs NLRC4, which has been shown to take part in host defenses against bacterial infection (46, 47). However, it has recently been found that ASC can strengthen the signal of the NLRC4 inflammasome (48) or being required for some specific responses (49). The presence of a CARD domain in Caiap suggests that it may directly recruit and activate Caspase-1 without the need for ASC, at least in most evolutionarily advanced species where true orthologs of mammalian caspase-1 with CARD domain exist. To check this, we ablated Asc using a previously validated translation-blocking MO (19). Genetic depletion of Asc in zebrafish drastically increased the susceptibility to ST of both WT and Caiap-deficient larvae (Figure 7A). In addition, Asc deficiency inhibited basal and Caiap-induced caspase-1 activity in non-infected and infected larvae (Figure 7B). Conversely, forced expression of Asc strongly increased larval resistance to the infection and caspase-1 activity, both effects being largely independent of Caiap (Figures 7C,D). Unexpectedly, however, reconstitution of Caiap and Asc complexes in HEK293T cells, which lack each of these components, showed that Caiap and CaiapΔCARD fused to mCherry were both diffusely distributed in the cytosol in the absence of Asc but formed a ring around the speck in the presence of Asc (Figures 7E,F). In addition, co-immunopreciptation assays confirmed that Caiap was unable to physically interact with Asc (Figure 7G).

Caspase a (Caspa, also known as Caspy) is considered the functional homolog of mammalian caspase-1, since it preferentially cleaves AcYVAD-AMC, a caspase-1 substrate, and interacts and co-localizes at the speck with zebrafish Asc in HEK293T cells (18). In addition, recent studies have demonstrated in vivo that Caspa is the effector enzyme of the zebrafish inflammasome in both macrophages (20) and neutrophils (19). Therefore, we next analyzed whether the forced expression of Caspa was able to rescue the high ST susceptibility of Caiap-deficient larvae. The results showed that Caspa was indeed able to rescue the infection susceptibility (Figure 8A) but only partially caspase-1 activity (Figure 8B) of Caiap-deficient larvae. In addition, forced expression of Caspa alone increased the resistance of larvae to ST infection (Figure 8A), confirming previous results (19).

The above results show that Caiap acts through Asc and Caspa to promote ST resistance, although this does not clarify the mechanism by which Caiap plays this function. We therefore reconstituted Caiap and Caspa complexes in HEK293T cells and, surprisingly, WT Caiap and CaiapΔCARD were both able to form specks in the presence of wild type but not of catalytic inactive mutant Caspa (Figures 8C,D). These results were further confirmed by pull down assays, in which WT Caiap and CaiapΔCARD were able to physically interact with WT Caspa but not with catalytic inactive Caspa (Figure 8E). In addition, while WT Caiap overexpressed in HEK293T cells self-oligomerized, assayed under non reducing conditions, CaiapΔCARD required active Caspa to fully oligomerize (Figure 8F). Taken together, these results show that Caiap interacts with catalytic active Caspa (P20/P10) through its ANK domain and self-oligomerizes via its CARD domain.