Evolution-inspired redesign of the LPS receptor caspase-4 into an interleukin-1β–converting enzyme

Despite bioinformatic predictions, canine inflammatory caspases are functional homologs of Casp-1

The gene encoding the inflammasome stimulatory protein NLRP3 is conserved in carnivorans (15). To determine whether NLRP3 is operational in carnivoran cells, we primed canine monocyte-derived macrophages (MDMs) (fig. S1A) with LPS and stimulated them with the K⁺ ionophore nigericin, an inducer of NLRP3 activation in murine and human cells. Stimulated cells were then stained with the membrane-impermeable dye propidium iodide (PI), which binds to intracellular nucleic acids upon plasma membrane disruption, and determined cellular adenosine 5′-triphosphate (ATP) levels as a proxy for viability. Treatment of canine MDMs with nigericin stimulated an increase in PI fluorescence and a decrease in cellular ATP, both of which are indicators of pyroptosis (Fig. 1C and fig. S1B). Nigericin induced these responses in the presence or absence of LPS, as is observed in human monocytes (16). LPS priming correlated with the release of IL-1β (Fig. 1D), which was processed into the bioactive p17 fragment (Fig. 1E; fragment at ~17 kDa). The NLRP3 inhibitor MCC950 (17) prevented nigericin-induced cell death, as cells treated with this inhibitor exhibited lower PI staining and higher ATP levels, as compared with noninhibitor-treated cells (Fig. 1C and fig. S1B). In addition, the pan-caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (zVAD-FMK) and disulfiram, an inhibitor of GSDMD pore formation (18), reduced nigericin-induced death (Fig. 1C and fig. S1B). All three inhibitors diminished IL-1β release (Fig. 1D). Overall, these data indicate that the NLRP3 inflammasome pathway is intact in canine cells.

Our finding that IL-1β can be cleaved and released from canine MDMs was notable, as Casp-1 is missing in dogs. To explain these findings, we determined whether canine Casp-1/4 (cCasp-1/4a and cCasp-1/4b) proteins can operate as Casp-1. We designed an experimental system based on stable, heterologous expression of a caspase in immortalized bone marrow–derived macrophages (iBMDMs) from mice deficient in mCasp-1 and mCasp-11 (hereafter referred to as Casp-1/11^−/− iBMDMs). We reconstituted Casp-1/11^−/− iBMDMs with cCasp-1/4 isoforms (cCasp-1/4a or cCasp-1/4b), mCasp-1, mCasp-11, or hCasp-4. cCasp-1/4a and cCasp-1/4b could be detected by immunoblot using a Casp-4-specific, but not a Casp-1-specific, antibody, underscoring their identity as structural Casp-4 homologs (Fig. 1F).

We primed cells with LPS and subsequently stimulated them with nigericin. Within wild-type (WT) cells, these treatments induced pyroptosis, as assessed by the release of the cytosolic enzyme lactate dehydrogenase (LDH) and IL-1β (Fig. 1G). Both processes were abrogated in Casp-1/11^−/− iBMDMs expressing green fluorescent protein (GFP) or mCasp-11 as a transgene but were restored by expression of mCasp-1 (Fig. 1G). Similar to mCasp-1, cells expressing cCasp-1/4a or cCasp1/4b released LDH and IL-1β upon LPS priming and nigericin treatment. These phenotypes coincided with the cleavage of GSDMD into its pore-forming N-terminal domain (Fig. 1H). Similar results were obtained when cells were transfected with poly(dA:dT) (fig. S1, C and D), which stimulates the absent in melanoma 2 (AIM2) inflammasome (19, 20). Poly(dA:dT)-induced LDH release was independent of LPS priming (fig. S1C), as expected (19).

In addition to stimulating pyroptosis, select inflammasome activators can induce IL-1β release from living (hyperactive) cells (21, 22). Infection of macrophages with Staphylococcus aureus lacking O-acetyltransferase A (SA113 ΔoatA) causes hyperactivation in a Casp1-dependent manner (22). We infected transgene-expressing Casp-1/11^−/− iBMDMs with SA113 ΔoatA. WT iBMDMs, as well as cells expressing mCasp-1, cCasp-1/4a, or cCasp-1/4b, responded to infection with the release of IL-1β, but not LDH, and the amount of IL-1β could be boosted if cells were primed with LPS (Fig. 1I). No IL-1β was released from Casp-1/11^−/− iBMDMs expressing GFP or mCasp-11 upon infection (Fig. 1I).

Within mCasp-1, the N-terminal CARD is critical for its recruitment into inflammasomes (23, 24). To determine the role of the CARD within cCasp-1/4a and cCasp-1/4b, we reconstituted Casp-1/11^−/− iBMDMs with a canine caspase construct lacking its Casp-1–like CARD. Cells expressing this mutant caspase (cCasp-1/4bΔCARD) failed to release LDH or IL-1β after LPS priming followed by nigericin treatment (fig. S1, E to H). These collective data demonstrate that both cCasp-1/4 isoforms can operate similar to Casp-1 in the context of multiple inflammasome stimuli.

Canine caspases are not LPS sensors and canine cells cannot respond to intracellular LPS

While the first CARD of cCasp-1/4b is similar to that of Casp-1, the second CARD shares homology with the LPS-sensing CARDs of hCasp-4 and mCasp-11. We therefore investigated whether cCasp-1/4b might respond to exposure to LPS. Transgenic cells were primed with LPS before delivery of LPS into the cytosol via electroporation. LPS electroporation stimulated LDH release from WT iBMDMs and Casp-1/11^−/− iBMDMs expressing mCasp-11 or hCasp-4, indicating the induction of pyroptosis (Fig. 2A). Cells expressing mCasp-1 or GFP did not lyse upon LPS electroporation. None of our reconstituted Casp-1/11^−/− iBMDMs released significant amounts of IL-1β after LPS electroporation (Fig. 2B). This finding validates current dogma, which predicts that mCasp-11 and mCasp-1 are both needed for LPS to induce cleavage and release IL-1β (11). Because our iBMDMs produce either mCasp-1 or mCasp-11 (not both), IL-1β release is diminished.

We found that neither cCasp-1/4a– nor cCasp-1/4b–expressing cells died upon cytosolic LPS delivery (Fig. 2A). These results imply that cCasp-1/4b cannot be activated by LPS. To further investigate this lack in LPS responsiveness, we determined whether the Casp-4–like CARD from cCasp-1/4b can functionally replace the LPS-sensing CARD from either mCasp-11 or hCasp-4. We designed chimeric caspases where the CARDs from cCasp-1/4b were attached to the enzymatic domain from mCasp-11 (mEnz11) or hCasp-4 (hEnz4) (cCARD1 + 4/mEnz11 and cCARD1 + 4/hEnz4). These chimeric proteins were stably produced in Casp-1/11^−/− iBMDMs (Fig. 2, C and D). Cells expressing these chimeras failed to promote pyroptosis when electroporated with LPS (Fig. 2E). We further considered the possibility that the Casp-1–like CARD prevents LPS detection by cCasp-1/4b. However, deletion of this additional CARD did not render cells responsive to LPS electroporation (fig. S2A). Even within primary canine MDMs, we observed no evidence of cell death or IL-1β release upon LPS electroporation (Fig. 2, F to H). Altering electroporation conditions did not reveal any LPS-specific changes in canine MDMs (fig. S2, B to D). In human and pig monocytes (but not MDMs), the TLR4 pathway activates the NLRP3 inflammasome to promote IL-1β release (25). We observed similar responses when we stimulated canine monocytes with LPS, as these treatments promoted IL-1β release (fig. S2, E and F). Canine cells are therefore not generally unresponsive to LPS but are specifically unresponsive to cytosolic LPS.

Canine caspase activities reveal distinct mechanisms of IL-1β and GSDMD substrate selection by murine and human inflammatory caspases

Our findings suggest that cCasp-1/4 proteins operate most similarly to Casp-1 homologs in humans and mice. We therefore investigated mechanisms that underlie this symmetry of activities. Cysteine 285 (C285) is the main catalytic residue within human and murine Casp-1 (26), which is conserved in the canine caspases. We introduced Cys-to-Ala mutations at the equivalent sites in cCasp-1/4a and cCasp-1/4b (C285A and C370A, respectively) and expressed these mutants in Casp-1/11^−/− iBMDMs (Fig. 3A). We did not observe any LDH release, secretion of IL-1β, or processing of GSDMD after LPS + nigericin treatment of cells expressing these cCasp-1/4 mutants (Fig. 3, B to F). The catalytic activity of cCasp-1/4a and cCasp-1/4b is therefore required to induce pyroptosis and IL-1β release.

The ability of cCasp-1/4 proteins to induce IL-1β processing was unexpected, as their catalytic domain is most similar to that found in hCasp-4, a caspase that has minimal ICE activity (27, 28). The presence of cleaved IL-1β in the supernatants of the reconstituted macrophages could be explained if Casp-4 homologs actually do have ICE activity, but this activity is dormant under natural conditions and can only be stimulated upon recruitment into an inflammasome.

To address this possibility, we created a scenario whereby the enzymatic domains of hCasp-4 or mCasp-11 can be recruited into inflammasomes via a mechanism similar to Casp-1. This was accomplished by generating fusions between the CARD of hCasp-1 or mCasp-1 to full-length hCasp-4 or mCasp-11 or the isolated enzymatic domains of these caspases (Fig. 3G). When expressed in Casp-1/11^−/− iBMDMs, these chimeric enzymes caused a similar degree of pyroptosis after LPS + nigericin treatment (Fig. 3, H and I). However, this pyroptotic cell death was not accompanied by the release of mature IL-1β (Fig. 3, J and K). These findings eliminate the possibility that recruitment of hCasp-4 or mCasp-11 into an inflammasome stimulates a latent ICE activity.

On the basis of the above-described phenotypes, we considered the possibility that cCasp-1/4 display intrinsic ICE properties. We therefore purified recombinant catalytic domains from mCasp-1, hCasp-4, and cCasp-1/4. Analysis by SDS–polyacrylamide gel electrophoresis (PAGE) confirmed that the recombinant enzymes are autocatalytically processed into the large (p20) and small (p10) catalytic subunits (Fig. 4A). We then characterized their ability to cleave peptide-based and full-length protein substrates. Enzyme kinetic analyses using N-a-acetyl-Tyr-Val-Ala-Asp-p-nitroanilide ( YVAD-pNA), a chromogenic tetrapeptide substrate optimized for cleavage by Casp-1, revealed similar Michaelis-Menten constants (K_m) for mCasp-1 and cCasp-1/4, whereas the K_m of hCasp-4 for this substrate was so high that it was not possible to calculate (Fig. 4, B and C). The turnover number (k_cat) and catalytic efficiency of mCasp-1 were only two- to threefold higher than those of cCasp-1/4 (Fig. 4C). Similar results were observed when we examined whether the caspases can process pro–IL-1β. As expected, pro–IL-1β was cleaved far more efficiently by mCasp-1 than by hCasp-4 (Fig. 4, D and E). Our results are consistent with earlier reports that hCasp-4 can process pro–IL-1β in vitro, albeit at very slow rates, which imply that this reaction is unlikely to be of physiological relevance (29). cCasp-1/4 processed murine pro–IL-1β at a near-comparable rate to mCasp-1 (Fig. 4, F and G). Equivalent results were obtained when the canine variant of this cytokine was used as a substrate (fig. S3, A to D). We also investigated the ability of our recombinant caspases to process another IL-1 family member, pro–IL-18 (30). As expected (28, 31), mCasp-1 and hCasp-4 efficiently cleaved pro–IL-18 and produced the mature 18-kDa fragment (p18) and cCasp-1/4 cleaved pro–IL-18 with a similar efficiency (fig. S3, E to H). These findings demonstrate that cCasp-1/4 enzymes can process IL-1 family cytokines comparably with mCasp-1.

To identify the molecular determinants for the ICE activity of cCasp-1/4, we first generated chimeric enzymes consisting of the p20 domain of cCasp-1/4 and the p10 domain of hCasp-4, and vice versa. Both chimeras exhibited intermediate catalytic efficiencies of pro–IL-1β cleavage (fig. S4, A to C). This finding suggests that sites distributed across the large and small subunit contribute to pro–IL-1β cleavage, with sites in the small subunit being dominant.

We reasoned that critical residues should be conserved among carnivoran Casp-1/4 homologs, but not in hCasp-4. By applying this rationale, we identified several conserved residues (E288/P290, H342, K345, and K379) in cCasp-1/4, which we individually replaced with the respective amino acids from hCasp-4. Each of the single point mutations (E288A/P290R, H342D, K345M, and K379T) led to a slight decrease in cleavage efficiency compared with WT cCasp-1/4 (Fig. 4H). When combining three (triple mutant: H342D, K345M, and K379T) or all four of the mutations (quadruple mutant), we observed additive effects causing pro–IL-1β cleavage efficiency to drop to a level similar to hCasp-4 (Fig. 4H and fig. S4, E to J). The ability of the quadruple mutant to cleave GSDMD remained unaffected, indicating a specific impact of the mutations on cleavage of pro–IL-1β, without affecting the overall activity of the caspase (Fig. 4H and fig. S4K). Introduction of reciprocal canine-specific mutations into hCasp-4 had the opposite effect. While individual point mutations led to modest increases (~4- to 6-fold) in the ability to cleave pro–IL-1β, by combining mutations at several sites, we generated an hCasp-4 quadruple mutant, which is as efficient as cCasp-1/4 at processing pro–IL-1β (Fig. 4I and fig. S5, A to H). We observed only minor effects of these mutations on GSDMD cleavage (Fig. 4I). Consistent with these findings, the cCasp-1/4 triple mutant maintained the ability to support LPS + nigericin–induced pyroptosis when expressed in Casp-1/11^−/− iBMDMs (Fig. 4, J and K). In this situation, however, cell lysis was not accompanied by the release of mature IL-1β (Fig. 4, L and M). These findings therefore identify amino acids that determine ICE activity in human and canine caspases.

When mapping the amino acids of interest onto a homology model of cCasp-1/4 in complex with the peptide inhibitor zVAD-FMK, we found that the conserved positively charged residues H342, K345, and K379 are located in the same region in the p10 subunit: a cleft formed by two loops proximal to the tetrapeptide binding site (fig. S5I). This charged site is spatially distinct from the hydrophobic recognition site for GSDMD (highlighted in yellow in fig. S5I) (32, 33). This observation suggests that positive charges in this cavity promote recognition of pro–IL-1β as a cleavage substrate. To test this model, we created additional caspases carrying charge reversal mutations within this region. A D381R mutation in cCasp-1/4 increased its ability to cleave pro–IL-1β by 4-fold (fig. S6, A and B), whereas a R354D mutation in hCasp-4 further decreased its cleavage efficiency by 10-fold (fig. S6, A and C). Similar to its canine counterpart, mCasp-1 E379R was ~3-fold better at cleaving pro–IL-1β than WT mCasp-1 (fig. S6, A and D).

H342 drew our attention, as it is positioned in a manner that suggests a direct involvement in coordinating the tetrapeptide in the active site (fig. S5I). H342 is present in most carnivoran Casp-1/4 proteins and is conserved in Casp-1. Casp-4 homologs that contain no ICE activity, such as hCasp-4, display a negative charge at this site (fig. S6E). Introducing a charge-swap H340D mutation into mCasp-1 greatly decreased its ability to process pro–IL-1β while only slightly affecting GSDMD cleavage (~31-fold and ~2-fold reduction compared with WT mCasp-1, respectively) (Fig. 5A and fig. S6, F to H). To determine whether this cleavage deficiency extended to activities within cells, we reconstituted Casp-1/11^−/− iBMDMs with an mCasp-1 H340D mutant (Fig. 5B). After LPS + nigericin treatment, we detected less processed IL-1β released from cells expressing mCasp-1 H340D, as compared with WT mCasp-1–expressing cells (Fig. 5, C and D). LDH release, GSDMD processing, and autocatalytic cleavage of mCasp-1 H340D were unimpaired, indicating an IL-1β–specific effect of the H340D mutation (Fig. 5, E to G). We confirmed these results in cells expressing hCasp-1 (or the respective H340D mutant), thereby establishing the importance of H340 across species (Fig. 5, H to K). These collective findings establish that the charge of amino acids within the catalytic site determines ICE activity but has minimal impact on the GSDMD-cleaving activity of inflammatory caspases.

Recognition and cleavage of GSDMD are mediated by hydrophobic interactions between a conserved protease exosite present in the L2 loop region of all inflammatory caspases and the C-terminal domain of GSDMD (fig. S7A) (32, 33). We investigated whether mutations in this site affect the ability of a caspase to process pro–IL-1β. We generated cCasp-1/4 and mCasp-1 mutants in which the exosite has been disrupted by introducing a hydrophilic side chain (cCasp-1/4 W294N and mCasp-1 L294N). cCasp-1/4 W294N and mCasp-1 L294N exhibited a marked decrease in the ability to process GSDMD compared with their WT counterparts (fig. S7, B and C). In contrast, we observed only weak effects on pro–IL-1β cleavage (~40- versus 4-fold reduction; fig. S7, B and D). Overall, these data suggest that recognition of GSDMD and pro–IL-1β is mediated by spatially and functionally distinct regions of the caspase.

Redesign of hCasp-4 to intrinsically link LPS detection to IL-1β cleavage and release without the need for inflammasomes

In mice, mCasp-11 and mCasp-1 link LPS detection to the release of IL-1β, with inflammasomes serving as an intermediate between these enzymes. On the basis of the finding that we can redesign hCasp-4 into an enzyme with the ability to process IL-1β in vitro, it should be possible to condense the natural multistep LPS-mediated pathway to IL-1β release into one single protein. To fulfill this task, hCasp-4 would need to exhibit LPS sensing, GSDMD, and IL-1β cleavage functionalities. As a proof of concept of this idea, we first designed chimeric caspases consisting of the LPS-binding CARD from hCasp-4 or mCasp-11 and the enzymatic domain of cCasp-1/4 (named hCARD4/cEnz and mCARD11/cEnz, respectively). These chimeric enzymes were expressed in Casp-1/11^−/− iBMDMs (Fig. 6A and fig. S8A). Cells expressing either caspase underwent pyroptosis, as assessed by LDH release after LPS electroporation (Fig. 6, B and C). Cells expressing either chimeric caspase also released higher amounts of IL-1β, as compared with cells expressing WT caspases (Fig. 6, D and E). The cleaved p17 fragment of IL-1β was exclusively detected in the supernatants of cells expressing hCARD4/cEnz or mCARD11/cEnz but not of cells expressing hCasp-4 or mCasp-11 (Fig. 6, F and G). The ability of these chimeric caspases to promote IL-1β cleavage and release was independent of NLRP3, as production of this cytokine was insensitive to NLRP3 inhibition by MCC950 (Fig. 6, H and I, and fig. S8, B and C). In contrast, MCC950 prevented IL-1β release (but not LDH release) from WT iBMDMs after LPS electroporation (fig. S8, D and E), as expected (10). These findings demonstrate that caspases can be engineered to bypass the need for inflammasomes to promote IL-1β cleavage and release.

To determine whether the chimera-based strategy of inflammasome bypass can be extended to a subtler engineering approach, we examined the redesigned hCasp-4 quadruple mutant that displays ICE activity in vitro. Within Casp-1/11^−/− cells expressing this LPS receptor, which contains only four amino acid substitutions, LPS electroporation stimulated IL-1β and LDH release. In contrast, cells expressing WT hCasp-4 released LDH, but not IL-1β (Fig. 6, J to L). Consistent with IL-1β release by redesigned hCasp-4 being an inflammasome-independent process, IL-1β release was insensitive to MCC950 (Fig. 6, K and L). To further verify that redesigned hCasp-4 could bypass inflammasomes and cleave and release IL-1β, we determined whether these cells could respond to LPS + nigericin treatment. Consistent with the fact that these cells lack endogenous mCasp-1 and mCasp-11, LPS + nigericin treatment was unable to promote IL-1β release (fig. S8, F and G). Thus, IL-1β cleavage after LPS electroporation is driven by the intrinsic activities of redesigned hCasp-4, independent of inflammasomes.

We investigated whether this inflammasome-independent pathway can be engaged by Gram-negative bacteria. We focused on Salmonella enterica serovar Typhimurium, which activates inflammasomes by mCasp-11, NLRP3, and NAIP-NLRC4 (34). We eliminated NLRP3 activities from consideration in these studies, as the iBMDMs used are unresponsive to NLRP3 agonists. In addition, we used a flagellin-deficient strain of Salmonella (SL1344 fliC/fljB) (35) to diminish NAIP-NLRC4 inflammasome activities (36). Consistent with this experimental setup focusing attention on LPS-induced activities, infection with SL1344 fliC/fljB induced LDH release in an hCasp-4 or mCasp-11 transgene-dependent manner (Fig. 6, N and O). Consistent with our findings with electroporated LPS, infection-induced IL-1β release was strongly enhanced within cells expressing redesigned caspases that contain LPS-binding and ICE activity (hCasp-4 quadruple mutant, hCARD4/cEnz, or mCARD11/cEnz) (Fig. 6, P to S). In contrast, cells expressing WT hCasp-4 or mCasp-11 were unable to release cleaved IL-1β after infection (Fig. 6, R and S). These collective data demonstrate the redesign of a normally inflammasome-dependent process into a one-protein signaling pathway that intrinsically links LPS detection to IL-1β cleavage and release.

Multiple animal species encode a one-protein signaling pathway that bypasses inflammasomes to link LPS detection with IL-1 cleavage and release

Our finding that the inflammasome pathway can be bypassed by the simple introduction of four mutations into hCasp-4 raises the question of whether similar caspases and pathways exist in nature. To address this possibility, we performed a cell-based mini-screen, where we introduced Casp-4 homologs from species spanning a range of mammalian orders (namely, mouse, cat, lemur, rabbit, sheep, horse, bat, manatee, and wombat) into Casp-1/11^−/− iBMDMs (Fig. 7A). We electroporated cells with LPS and assessed IL-1β release and pyroptosis. With the exception of the manatee and wombat genes, all Casp-4 homologs responded to LPS electroporation with the induction of pyroptosis (Fig. 7B). The observation that manatee and wombat Casp-4 do not support LPS-induced pyroptosis is interesting, as these species are evolutionarily most distant from the rest of the animals in our panel. This finding hints that recognition of LPS might not be a primordial function of Casp-4 but was acquired after these species diverged. Another caspase that stands out is the Casp-4 homolog from the cat (feline Casp-1/4b). Unlike other caspases examined, Casp-1/11^−/− iBMDMs expressing feline Casp-1/4b released large amounts of mature IL-1β (Fig. 7, C and D). As observed with our synthetically redesigned pathways, IL-1β release induced by feline Casp-1/4b was insensitive to MCC950 (Fig. 7E). Feline primary MDMs behaved in a similar manner: When LPS was delivered to the cytosol, pyroptosis and the release of mature IL-1β were evident. Unlike in murine iBMDMs, MCC950 cotreatment did not inhibit IL-1β processing and secretion (Fig. 7F). We eliminated the possibility that MCC950 is ineffective in feline cells, as it inhibited cell death and IL-1β release when NLRP3 was activated by nigericin (Fig. 7G). These findings suggest that IL-1β is processed and secreted upon activation of feline Casp-1/4b by LPS without the need for inflammasomes. Intrigued by the apparent dichotomy in the behavior of canine and feline Casp-1/4b, we characterized Casp-4 homologs from additional carnivoran species. When introduced into Casp-1/11^−/− iBMDMs, most caspases examined, including those found in carnivorous and herbivorous bears (polar bear and panda), marine pinnipeds (harbor seal), and diverse felines (tiger and cheetah), conferred the ability to cause LDH and IL-1β release in response to LPS electroporation (Fig. 7H). The only exceptions in this panel were made by Casp-4 homologs from dingo and fox, which, like the common dog, are members of the Canidae family. These proteins failed to respond to intracellular LPS (Fig. 7H). When we correlated our functional analyses with their evolutionary relationship, we could divide the examined Casp-4 homologs into three distinct classes: Class 1, represented by mCasp-11 and hCasp-4, are Casp-4 homologs, which sense LPS, but do not exhibit ICE activity; class 2, which includes most carnivoran Casp-1/4 proteins, both sense LPS and have ICE activity; and class 3 is represented by Casp-1/4 proteins from canines, which have ICE activity, but are unresponsive to LPS. Most carnivorans, but not dogs and closely related species, therefore have the ability to mount an inflammatory response to cytosolic LPS, and they do so by using a natural one-protein signaling pathway (fig. S9).

Study design

The aim of this study was to investigate Casp-1/4 fusion proteins found in carnivoran mammals and use them as a tool to obtain general insights into inflammatory caspase substrate selection. We investigated the function of these enzymes in vitro using recombinant proteins, in reconstituted murine macrophages and feline and canine primary cells. Sample sizes for each experiment are indicated in the figure legends.

Cell lines

All cells were cultured in humidified incubators at 37°C and 5% CO₂. Immortalized iBMDMs from WT or Casp-1/11^−/− mice were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin (Pen-Strep), l-glutamine, and sodium pyruvate, hereafter referred to as complete DMEM (cDMEM). cDNA sequences for WT or chimeric caspase constructs of interest carrying an N-terminal Myc-tag were ordered as gBlocks (Integrated DNA Technologies) and cloned into pMSCV IRES EGFP using Not I and Sal I restriction sites. Point mutations were introduced using the QuikChange mutagenesis kit (Agilent) or the Q5 mutagenesis kit (NEB) according to the manufacturer’s protocols. All constructs generated here were sequence-confirmed by Sanger sequencing. Human embryonic kidney (HEK) 293T cells were cultured in cDMEM and used as packaging cells for retroviral vectors. For the production of retroviral particles, 2.5 × 10⁶ HEK293T cells were seeded in a 10-cm cell culture dish. After overnight incubation at 37°C, cells were transfected with 10 μg of pMSCV IRES EGFP encoding the protein of interest, 6 μg of pCL-ECO, and 3 μg of pCMV-VSVG using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. After 18 to 24 hours at 37°C, media were changed to 6 ml of fresh cDMEM, and virus-containing supernatant was collected 24 hours after media change. Supernatants were clarified from cellular debris by centrifugation (400g, 5 min) and filtered through a 0.45-μm polyvinylidene difluoride syringe filter. Caspase-1/11^−/− iBMDMs (~2 × 10⁶) were transduced twice on two consecutive days in a six-well plate by adding 4.5 ml of viral supernatant supplemented with polybrene (1:2000; EMD Millipore) per well, followed by centrifugation for 1 hour at 1250g and 30°C. GFP⁺ cells were sorted twice on a FACSAria or FACSMelody cell sorter (BD Biosciences) to obtain cell lines with stable and homogeneous expression of the target protein. Transgene expression was confirmed by immunobloting using a rabbit anti–Myc-tag or mouse anti–Myc-tag primary antibody (both from Cell Signaling Technologies) at a 1:1000 dilution.

Primary cell culture

Canine whole blood from healthy beagle dogs or feline whole blood from healthy cats of unspecified breed was purchased from BioIVT and processed within 24 hours after blood draw. Primary cells were cultured in RPMI supplemented with 10% FBS, l-glutamine, sodium pyruvate, and Pen-Strep (referred to as complete RPMI). Whole blood was diluted at a 1:1 ratio with sterile phosphate-buffered saline (PBS) pH 7.4 + 2.5 mM EDTA before layering 30 ml of diluted blood over 15 ml of FiColl Paque PLUS density gradient media (GE Healthcare). Density gradient centrifugation was performed at 800g for 35 min at 20°C. Total peripheral blood mononuclear cells (PBMCs) were harvested from the interphase and washed twice with wash buffer [PBS (pH 7.4), 2.5 mM EDTA, and 1% FBS]. Red blood cells were lysed by resuspending the pellet in ammonium-chloride-potassium (ACK) lysis buffer and incubating for 5 to 10 min at room temperature (RT). After a final washing step in wash buffer, total PBMCs were seeded in T75 cell culture flasks in 15 ml of complete RPMI (40 × 10⁶ PBMCs per flask). After incubation for 2 hours at 37°C for canine PBMCs or overnight for feline PBMCs, nonadherent PBMCs were removed by three washes with prewarmed, sterile PBS (pH 7.4). Adherent cells (monocytes) were either used for experiments right away or cultured in complete RPMI supplemented with recombinant human macrophage colony-stimulating factor (M-CSF) (50 ng/ml) (R&D Systems) for 5 to 7 days. Media were replenished with fresh complete RPMI containing M-CSF every 2 to 3 days.

Differentiation and immortalization of iBMDMs

BMDMs were immortalized as described before with slight modifications (40). L929 fibroblasts secreting M-CSF were cultured in cDMEM and M-CSF–containing supernatants were harvested, cleared by centrifugation (400g, 5 min), and passed through a 0.22-μm filter. CreJ2 cells were cultured in cDMEM and J2 retrovirus–containing supernatants were harvested, cleared by centrifugation (400g, 5 min), and passed through a 0.45-μm filter. Femurs from freshly euthanized Casp-1/11^−/− mice (B6N.129S2-Casp1tm1Flv/J from the Jackson Laboratory) were dissected and flushed with sterile PBS (pH 7.4). The obtained cell suspension was passed through a 70-μm cell strainer before plating cells in non–tissue culture–treated 10-cm dishes (1 × 10⁷ cells per dish). Cells were cultured in cDMEM supplemented with 30% M-CSF–containing supernatants from L929 cells for 4 days. On days 4 and 6 after isolation, bone marrow cells were transduced with J2 virus by swapping media for immortalization media (30% L929 supernatant + 70% CreJ2 supernatant) for 24 hours. After day 7, cells were passaged in cDMEM supplemented with gradually decreasing doses of L929 conditioned media (starting at 30%) until they divided normally in unsupplemented cDMEM. When the concentration of supplemented L929 supernatant was below 5%, cells were cultured in tissue culture–treated plates for better adherence.

Ligand and chemical reconstitution

Escherichia coli LPS (serotype O:111 B4) was purchased from Enzo Life Biosciences as a ready-to-use stock solution of 1 mg/ml and used at a working concentration of 1 μg/ml. Nigericin was bought from Invivogen, dissolved in 100% ethanol to 6.7 mM, and used for NLRP3 stimulation at a concentration of 10 μM. Poly(dA:dT) from Invivogen was dissolved to a stock concentration of 1 mg/ml. Poly(dA:dT) was transfected at a final concentration of 5 μg/ml. MCC950 and zVAD-FMK (both from Invivogen) were resuspended in sterile dimethyl sulfoxide (DMSO) to a concentration of 20 mM and used at a final concentration of 10 and 20 μM, respectively. Disulfiram was purchased from Tocris Biosciences, dissolved in DMSO to generate a stock solution at 20 mM. During inhibition experiments, disulfiram was present at a concentration of 50 μM. To inhibit pyroptosis, cells were pretreated with inhibitors for 30 min and the respective inhibitor was present during signal 2 of inflammasome stimulation. PI solution (1 mg/ml) was purchased from Millipore Sigma. Recombinant human M-CSF (Chinese hamster ovary–expressed, carrier-free) was bought from R&D Systems and dissolved in sterile PBS (pH 7.4) (stock concentration, 50 μg/ml). YVAD-pNA was obtained from Enzo Biosciences and was prepared as a 20 mM stock solution in DMSO.

Inflammasome activation, cell death assays, and ELISAs

Inflammasome activation assays were performed in a 96-well format. To activate the NLRP3 inflammasome, 1 × 10⁵ cells were seeded in duplicate wells in 100 μl of cDMEM and incubated for 1 hour at 37°C and 5% CO₂ for cells to adhere to the plate before adding 100 μl of cDMEM with or without 2× LPS (final concentration, 1 μg/ml) for priming. After 4 hours at 37°C and 5% CO₂, cells were stimulated with 10 μM nigericin in 200 μl of cDMEM for 3 hours. Cell lysis was quantified using the CyQuant LDH cytotoxicity assay kit (Thermo Fisher Scientific). Fifty microliters of supernatants was mixed with 50 μl of LDH assay buffer and incubated for 15 to 30 min at RT. Absorbance at 490 and 680 nm was measured on a Tecan Spark plate reader, and signal was normalized to lysis controls. Because primary cells yielded low signal intensities in LDH assays, PI staining was used as a readout to quantify membrane perforation in experiments involving these cells. A total of 1 × 10⁵ canine cells or 0.5 × 10⁵ feline cells were treated as described in a black 96-well plate with clear bottom and PI was added to the media at a 1:300 dilution 30 min before the end of the stimulation. Plates were spun for 5 min at 400g and fluorescence was measured on a Tecan Spark device at an excitation wavelength of 530 nm and an emission wavelength of 617 nm. Cellular ATP levels as a measure of viability were determined using the Celltiter-Glo Luminescent Cell Viability kit (Promega). Supernatant was completely removed from the cells stimulated in a 96-well format before adding 30 μl of Opti-MEM and 30 μl of Celltiter-Glo solution per well. After incubation for 5 min at RT, the mixture was transferred into a white 96-well plate, and luminescence signal was quantified using a Tecan Spark device. Release of IL-1β was assessed by enzyme-linked immunosorbent assay (ELISA) using the IL-1 beta Mouse Uncoated ELISA Kit (Thermo Fisher Scientific), or the feline or canine IL-1 beta/IL1F2 DuoSet kit (both from R&D Systems), respectively, according to the manufacturer’s protocols. To activate the AIM2 inflammasome, 5 × 10⁴ cells were seeded in duplicate wells in 100 μl of cDMEM and incubated overnight at 37°C and 5% CO₂. After priming with LPS (1 μg/ml) for 4 hours, cells were transfected with poly(dA:dT) (5 μg/ml) (Inivivogen) using Lipofectamine 2000 in a total volume of 200 μl of cDMEM per well. Supernatants for LDH assay and ELISA analysis were collected 6 hours after transfection.

Processing of GSDMD and caspase-1 after inflammasome activation was analyzed by immunoblotting using a rabbit anti-GSDMD (Abcam) or mouse anti-murine caspase-1 p20 (Adipogen) primary antibodies, diluted at 1:1000. iBMDMs (1 × 10⁶ to 2 × 10⁶) of the cell line of interest were seeded in a six-well plate and primed with LPS (1 μg/ml) for 3 to 4 hours. In order to capture proteins present in the lysate and the supernatant, nigericin was administered in 1.5 ml of Opti-MEM media for 2 hours. Samples for immunoblotting were prepared by adding 375 μl of 5× SDS loading buffer directly to the well and heated to 65°C for 10 min.

LPS electroporation

The Neon Transfection System (Thermo Fisher Scientific) was used to deliver bacterial LPS into the cytoplasm of cells. Canine MDMs or murine iBMDMs cells were primed with LPS for 3 to 4 hours in six-well plates, lifted with PBS + EDTA, and resuspended in R buffer at a density of 10 × 10⁶ cells/ml. Feline MDMs were primed equivalently in non–tissue culture (TC)–treated six-well plates and resuspended in R buffer at a density between 3 × 10⁶ and 5 × 10⁶ cells/ml. LPS or sterile PBS (negative control) was mixed with the cell suspension (1 μg of LPS or 1 μl of PBS per 1 × 10⁶ cells, respectively) before aspirating the cell suspension into the electroporation pipette equipped with a 100-μl tip and performing electroporation with two pulses with a width of 10 ms and a voltage of 1400 V, unless stated otherwise. Cells were then dispensed into media at a density of 5 × 10⁵ cells/ml (murine and canine cells) or 2.5 × 10⁵ cells/ml (feline cells) and seeded in 96-well (200 μl per well) or 6-well tissue culture plates. Cell death and IL-1β release were quantified 3 hours after electroporation by LDH or PI and CelltiterGlo assay and ELISA.

Bacterial infections

S. aureus strain lacking the oatA gene (SA113 ΔoatA) was a gift from D. Underhill (Cedars Sinai). Bacteria were streaked on tryptic soy agar (TSA) agar plates containing sheep blood (Thermo Fisher Scientific) and incubated overnight at 37°C. Single colonies were picked to inoculate 5 ml of Todd-Hewitt broth (Millipore Sigma) containing kanamycin (25 μg/ml) and grown at 37°C and 250 rpm for 24 hours. Cultures were washed three times in PBS (pH 7.4), and OD₆₀₀ (optical density at 600 nm) was determined before appropriately diluting bacteria in cDMEM without antibiotics. A total of 1 × 10⁵ iBMDMs per well in a 96-well plate were primed with LPS (1 μg/ml) or left untreated followed by infection with SA113 ΔoatA at a multiplicity of infection (MOI) of 30 in a total volume of 200 μl per well. Synchronized infection was facilitated by spinning the plates at 500g for 5 min right after addition of bacteria-containing media. After 1 hour of infection, extracellular bacteria were killed by changing the media to cDMEM (with Pen-Strep) and supplemented with gentamicin (50 μg/ml). Cell culture supernatants for LDH release assay and ELISA to quantify levels of IL-1β were collected 12 hours after initial infection.

Salmonella strain deficient for flagellin (SL1344 fliC/fljB) was a gift from I. Brodsky (University of Pennsylvania), and infections were performed as described before with minor modifications (35). Bacteria were streaked on LB agar plates containing kanamycin (25 μg/ml), grown overnight at 37°C, and plates were stored at 4°C for later use. Overnight cultures [3 ml of LB + kanamycin (25 μg/ml) and chloramphenicol (25 μg/ml)] were inoculated with a single bacterial colony and grown at 37°C while shaking at 250 rpm. On the next morning, bacterial culture was diluted into high salt LB (3 ml of LB + 100 μl of overnight culture + 78 μl of sterile 5 M NaCl) and incubated for another 3 hours at 37°C without shaking. Infections were performed in a 96-well (for LDH and ELISA analyses) or 12-well format (for IL-1β immunoprecipitations) using 1 × 10⁵ and 1 × 10⁶ cells per well, respectively. Cells were seeded in cDMEM without Pen-Strep and primed with LPS (1 μg/ml) for 3 hours. Bacteria were washed three times with prewarmed cDMEM without Pen-Strep and added to cells at an MOI of 100 in 200 μl (96-well) or 2 ml (12-well) of cDMEM without Pen-Strep. Synchronized infection was facilitated by spinning the plates at 500g for 5 min right after addition of bacteria-containing media, and cells were incubated at 37°C and 5% CO₂. After 1 hour, gentamicin was added to a final concentration of 100 μg/ml to kill extracellular bacteria. Cell culture supernatants for downstream analyses were harvested at 4 hours after infection.

Immunoprecipitation of IL-1β from cell culture supernatants

Immunoprecipitation (IP) of murine IL-1β from cell culture supernatants was performed as described before (7). Supernatants from 0.5 × 10⁶ to 1.0 × 10⁶ cells (cell number consistent within each individual experiment) stimulated as indicated were transferred into 5-ml flow cytometry tubes with snap cap and depleted of cells and debris by spinning at 400g for 5 min. Cell-free supernatants were transferred into new tubes and rotated overnight at 4°C in the presence of 0.5 μg of biotinylated goat anti-murine IL-1β antibody (R&D Systems) and 20 μl of neutravidin agarose beads (Thermo Fisher Scientific). Remaining cells in the well were lysed in 1× SDS loading dye and served as input control. Beads were subsequently washed three times with PBS (pH 7.4) before eluting bound proteins in 50 μl of 1× SDS loading dye. Immunoprecipitated and cell-associated IL-1β was detected by immunoblotting using a rabbit anti-murine IL-1β antibody (GeneTex). Cell-associated actin or tubulin was detected as a loading control using a mouse anti-tubulin antibody (1:100 dilution; DSHB hybridoma bank) or a mouse anti-actin antibody from Sigma-Aldrich at a dilution of 1:5000. IP of canine or feline IL-1β was performed equivalently from pooled supernatants from 2 wells of a 96-well plate containing 0.5 × 10⁵ to 1.0 × 10⁵ cells per well using 0.18 μg of biotinylated canine/feline IL-1β ELISA detection antibody (R&D Systems) as bait and rabbit anti-canine IL-1β antibody from Bio-Rad or goat anti-feline IL-1β antibody from R&D Systems for immunoblot detection. Unless stated otherwise, all primary antibodies were used at a concentration of 1:1000.

Recombinant protein expression and purification

DNA sequences encoding residues 131 to 404 of cCasp-1/4a, residues 94 to 377 of hCasp-4, and residues 130 to 404 of mCasp-1 were amplified by polymerase chain reaction using gBlocks of the full-length proteins as a template and cloned into a pET28A(+) vector with an N-terminal His₆-tag using Bam HI and Eco RI restriction sites. cDNA encoding full-length murine pro–IL-1β, canine pro–IL-1β, and human pro–IL-18 was synthesized by Integrated DNA Technologies and cloned equivalently. Chemically competent Rosetta (DE3) pLysS cells (EMD Millipore) were transformed with the plasmid of interest and plated on LB agar plates with kanamycin (25 μg/ml). Overnight precultures inoculated with a single colony were grown at 30°C and 250 rpm in 2× YT media containing kanamycin (25 μg/ml) and chloramphenicol (50 μg/ml). Expression cultures of 500 to 1500 ml were inoculated with overnight cultures at a ratio of 1:100 and incubated at 37°C and 250 rpm until the OD₆₀₀ reached a value between 0.7 and 0.8. After a cooling step on ice, protein expression was induced by adding isopropyl-β-d-thiogalactopyranoside to a final concentration of 0.25 mM, and expression was allowed to proceed overnight at 18°C. Bacterial pellets were harvested by centrifugation (5000g for 20 to 30 min at 4°C) and stored at −20°C if not immediately used for protein purification.

To purify recombinant proteins, bacterial pellets were resuspended in resuspension buffer [25 mM Hepes-NaOH (pH 7.4), 150 mM NaCl, and 10 mM imidazole] and lysed by ultrasonication. Cell lysates were clarified by centrifugation (30 min, 20,000g, 4°C) and passed through a 0.22-μm syringe filter before pouring them into a gravity flow column containing a bed of nickel–nitrilotriacetic acid agarose beads (Qiagen). Beads were washed with at least 10 bed volumes of wash buffer [25 mM Hepes (pH 7.4), 400 mM NaCl, and 25 mM imidazole] and bound protein was eluted stepwise in resuspension buffer supplemented with 40 to 250 mM imidazole. Fractions containing the protein of interest were pooled and buffer-exchanged into storage buffer [25 mM Hepes-NaOH (pH 7.4), 150 mM NaCl, and 10% glycerol] using a PD-10 desalting column (GE Healthcare). Last, protein was concentrated by centrifugal ultrafiltration. Aliquots were snap-frozen in liquid nitrogen and stored at −80°C for later use. Purity and integrity of the purified proteins were analyzed by SDS-PAGE followed by InstantBlue staining (Expedeon). Recombinant human GSDMD was provided by Hao Wu (Harvard Medical School) and purified as previously described (18).

In vitro protein substrate cleavage assays

Twofold dilution series of the indicated recombinant caspase were incubated with purified murine/canine pro–IL-1β, human pro–IL-18, or human GSDMD at a final concentration of 50 nM in 40 μl of caspase assay buffer [10 mM Pipes (pH 7.2), 10% sucrose, 10 mM dithiothreitol, 100 mM NaCl, 1 mM EDTA, and 0.1% CHAPS] for 30 min at 37°C. Reactions were stopped by adding 15 μl of 5× SDS loading dye and boiling at 65°C for 10 min. Cleavage products were analyzed via immunoblot using rabbit anti-murine IL-1β (Genetex), rabbit anti-human IL-18 (MBL), or rabbit anti-GSDMD (Cell Signaling Technologies) primary antibody at a 1:1000 dilution. Band intensities were quantified using ImageJ to determine EC₅₀ (median effective concentration) values and catalytic efficiencies were calculated as described before (8) using the following equationkcatKm=ln(2)(EC50 × t)

In vitro peptide cleavage assays

For peptide cleavage assays, recombinant mCasp-1, hCasp-4, or cCasp-1/4 were first diluted to a concentration of 100 nM in caspase assay buffer. To start the reaction, 20 μl of the diluted caspase was then mixed with 80 μl of a serial dilution of the chromogenic tetrapeptide substrate YVAD-pNA in the same buffer (final concentration of caspase was 20 nM in a total volume of 100 μl) in a clear 96-well plate. Absorbance at a wavelength of 405 nm was measured every 20 s for 20 min using a Tecan Spark plate reader with temperature control set to 37°C. Substrate solution was prewarmed to 37°C before adding to the caspase to ensure homogeneous assay conditions. Absorbance values were plotted in dependence of time, and initial velocities were determined by performing linear fits of the resulting curves. A pNA standard curve was generated to transform absorbance values into molar concentrations. Initial velocities were plotted as a function of the substrate concentration, and kinetic parameters (K_m, V_max, and k_cat) were determined by performing a fit according to the Michaelis-Menten equation in GraphPad Prism.

Sequence alignments and homology modeling

Mammalian caspase homologs were identified by NCBI BLASTp search using amino acid sequences of cCasp-1/4a or hCasp-1 as search queries. Sequences of interest were aligned using the Clustal Omega Multiple Sequences Alignment tool and visualized in ESPript 3.0 (41, 42).

To generate a homology model of the catalytic domain of cCasp-1/4, we used tools available on the SWISS-MODEL online server (43). A crystal structure of the catalytic domain of hCasp-1 in complex with the inhibitor zVAD-FMK (sequence identity, 57 and 63% for the p20 and p10 fragment, respectively) was used as an input template (Protein Data Bank ID: 2H51) (44). The obtained homology model was visualized in PyMOL (Schroedinger Inc.).

Quantification and statistical analysis

Statistical significance was determined by two-way analysis of variance (ANOVA) or one-way ANOVA with Tukey’s multiple comparison test, or unpaired, two-tailed Student’s t test as appropriate for each experimental dataset. The statistical method used is indicated in the respective figure legends. P < 0.05 was seen as statistically significant. All statistical analyses were performed using GraphPad Prism data analysis software. Data are representative of at least three independent repeats. Data with error bars are presented as means ± SEM.