Genetic targeting of Card19 is linked to disrupted NINJ1 expression, impaired cell lysis, and increased susceptibility to Yersinia infection

Cell death plays a critical role in inflammatory responses. During pyroptosis, inflammatory caspases cleave Gasdermin D (GSDMD) to release an N-terminal fragment that generates plasma membrane pores that mediate cell lysis and IL-1 cytokine release. Terminal cell lysis and IL-1β release following caspase activation can be uncoupled in certain cell types or in response to particular stimuli, a state termed hyperactivation. However, the factors and mechanisms that regulate terminal cell lysis downstream of GSDMD cleavage remain poorly understood. In the course of studies to define regulation of pyroptosis during Yersinia infection, we identified a line of Card19-deficient mice (Card19lxcn) whose macrophages were protected from cell lysis and showed reduced apoptosis and pyroptosis, yet had wild-type levels of caspase activation, IL-1 secretion, and GSDMD cleavage. Unexpectedly, CARD19, a mitochondrial CARD-containing protein, was not directly responsible for this, as an independently-generated CRISPR/Cas9 Card19 knockout mouse line (Card19Null) showed no defect in macrophage cell lysis. Notably, Card19 is located on chromosome 13, immediately adjacent to Ninj1, which was recently found to regulate cell lysis downstream of GSDMD activation. RNA-seq and western blotting revealed that Card19lxcn BMDMs have significantly reduced NINJ1 expression, and reconstitution of Ninj1 in Card19lxcn immortalized BMDMs restored their ability to undergo cell lysis in response to caspase-dependent cell death stimuli. Card19lxcn mice exhibited increased susceptibility to Yersinia infection, whereas independently-generated Card19Null mice did not, demonstrating that cell lysis itself plays a key role in protection against bacterial infection, and that the increased infection susceptibility of Card19lxcn mice is attributable to loss of NINJ1. Our findings identify genetic targeting of Card19 being responsible for off-target effects on the adjacent gene Ninj1, disrupting the ability of macrophages to undergo plasma membrane rupture downstream of gasdermin cleavage and impacting host survival and bacterial control during Yersinia infection.

infection, we identified a line of Card19-deficient mice (Card19 lxcn ) whose macrophages were protected from cell lysis and showed reduced apoptosis and pyroptosis, yet had wildtype levels of caspase activation, IL-1 secretion, and GSDMD cleavage. Unexpectedly, CARD19, a mitochondrial CARD-containing protein, was not directly responsible for this, as an independently-generated CRISPR/Cas9 Card19 knockout mouse line (Card19 Null ) showed no defect in macrophage cell lysis. Notably, Card19 is located on chromosome 13, immediately adjacent to Ninj1, which was recently found to regulate cell lysis downstream of GSDMD activation. RNA-seq and western blotting revealed that Card19 lxcn BMDMs have significantly reduced NINJ1 expression, and reconstitution of Ninj1 in Card19 lxcn immortalized BMDMs restored their ability to undergo cell lysis in response to caspase-dependent cell death stimuli. Card19 lxcn mice exhibited increased susceptibility to Yersinia infection, whereas independently-generated Card19 Null mice did not, demonstrating that cell lysis itself plays a key role in protection against bacterial infection, and that the increased infection susceptibility of Card19 lxcn mice is attributable to loss of NINJ1. Our findings identify genetic targeting of Card19 being responsible for off-target effects on the adjacent gene Ninj1, disrupting the ability of macrophages to undergo plasma membrane rupture downstream of gasdermin cleavage and impacting host survival and bacterial control during Yersinia infection.

Introduction
Regulated cell death is an evolutionarily conserved mechanism by which multicellular organisms control fundamental biological processes ranging from tissue development to microbial infection. Apoptosis and pyroptosis represent two distinct forms of regulated cell death that are activated in response to diverse stimuli, including developmental signals [1,2], tissue stress or injury [3], and microbial infection [4][5][6][7]. While apoptosis is an immunologically quiescent or suppressive form of cell death during tissue homeostasis and development, certain stimuli, including infections and chemotherapeutic agents, induce apoptosis that is accompanied by inflammatory signals that contribute to anti-pathogen and anti-tumor immunity [8][9][10]. Apoptosis is activated by specific apoptotic initiator caspases, following certain intrinsic or extrinsic signals, and mediates organized disassembly of the cell via a process that limits cellular permeability and enables phagocytosis of the dying cell [11,12]. Alternatively, pyroptosis is a lytic form of regulated cell death characterized by caspase-1-or -11-dependent plasma membrane disruption mediated by cleavage and activation of the pore forming protein Gasdermin D (GSDMD), and release of interleukin-1 (IL-1) family cytokines and other intracellular components [13][14][15][16]. However, inhibition of the receptor-proximal signaling kinases TAK1 or IKK by chemotherapeutic drugs or during infection by bacterial pathogens such as Yersinia, can trigger a caspase-8-dependent apoptosis pathway that is also associated with GSDMD cleavage [17][18][19]. Interestingly, a recent study demonstrated that only caspase-8 activated in the RIPK1-containing complex IIb can efficiently cleave GSDMD in macrophages [20]. These findings suggest a potential point of intersection of cell death pathways previously viewed as distinct [21].
During pyroptosis, secretion of IL-1 family cytokines and release of intracellular contents are temporally and genetically linked, and it has been suggested that IL-1 cytokines are released from cells when they undergo caspase-1 or -11-dependent lysis [22,23]. However, IL-1 secretion and cell lysis can be uncoupled in certain cell types or in response to specific stimuli. For example, the osmoprotectant glycine prevents release of lactate dehydrogenase (LDH), a key indicator of cell lysis, but does not affect IL-1β cytokine release [24,25]. Conversely, genetic ablation of IL-1 does not prevent cell lysis following pyroptotic stimuli. Moreover, IL-1 secretion can occur by living cells in the absence of apparent cytotoxicity [26,27]. Notably, neutrophils and monocytes display evidence of caspase-1 processing and IL-1 secretion in the absence of cell lysis [28,29]. Furthermore, dendritic cells treated with the oxidized lipid oxPAPC, or macrophages treated with the N-acetyl glucosamine fragment of bacterial peptidoglycan, release IL-1 without undergoing cell death [30,31]. Interestingly, GSDMD is genetically required for the release of IL-1 cytokines in response to non-pyroptotic stimuli, a state termed hyperactivation [26,31,32]. Collectively, these data imply that while GSDMD processing and membrane insertion are critical for ultimate cell lysis, a cell fate decision checkpoint exists that distinguishes GSDMD-dependent IL-1 secretion from terminal cell death.
In our efforts to identify regulators of cell death during bacterial infection, we investigated the caspase activation and recruitment domain (CARD)-containing protein CARD19 [33]. CARD19 is a mitochondrial membrane protein that contains an N-terminal CARD and C-terminal transmembrane domain, suggesting that it could be involved in the regulation of cell death or inflammatory responses [33,34]. Multiple mitochondria-associated proteins regulate inflammasome activation and inflammatory signaling [35][36][37]. Notably, the mitochondrial CARD-containing protein, MAVS, plays a critical role in anti-viral immune signaling and IL-1 cytokine release [35,38], and the mitochondrial outer membrane lipid cardiolipin plays an important role in regulating NLRP3 inflammasome assembly [39].
Intriguingly, primary macrophages from Card19-deficient mice (Card19 lxcn ) were resistant to cell lysis and release of lytic cell markers in response to apoptotic and pyroptotic stimuli, but exhibited wild-type levels of caspase activation, IL-1 secretion, and gasdermin processing. Unexpectedly however, two independent CRISPR/Cas9 knockout mouse lines that we generated did not show the same phenotype; furthermore, expression of CARD19 in immortalized BMDMs from the original Card19 lxcn line did not restore their ability to undergo lysis, indicating that CARD19 itself is unlikely to be responsible for regulating terminal cell lysis in response apoptotic and pyroptotic stimuli. The cell lysis defect in Card19 lxcn mice was unrelated to Sterile alpha and heat armadillo motif-containing protein (SARM1), which was reported to regulate NLRP3 inflammasome-dependent cell death independently of IL-1β cytokine release [40], as BMDMs from four independent Sarm1 -/lines showed no defects in cell death, IL-1β release, and GSDMD cleavage. RNA-seq identified a handful of differentially regulated genes in Card19 lxcn BMDMs, including a recently reported regulator of plasma membrane lysis, Ninj1 [41], which is located adjacent to Card19 on chromosome 13. Notably, Card19 lxcn BMDMs showed significantly reduced Ninj1 mRNA expression at baseline and diminished NINJ1 protein levels both at baseline and upon LPS stimulation. Critically, reconstitution of immortalized Card19 lxcn BMDMs rescued the defect in cell lysis, similar to Ninj1 -/-BMDMs, indicating that the phenotype of Card19 lxcn mice results from a defect in NINJ1 protein expression. Card19 lxcn mice were more susceptible to infection with Yersinia, exhibiting enhanced mortality and systemic bacterial burdens relative to wildtype mice, implicating NINJ1-mediated lysis in host-pathogen interactions. Our data extend recent findings describing NINJ1 as a regulator of cell lysis downstream of caspase activation and gasdermin cleavage, and highlight the role of lytic cell death in host defense against bacterial infection.

Card19 lxcn BMDMs are deficient for caspase-dependent cell death
Our efforts to understand how cell death is regulated during bacterial infection led us to investigate the mitochondrial CARD-containing protein, CARD19, formerly known as 34]. To test the possible contribution of CARD19 to cell death during pyroptosis, we compared the kinetics of cell death in primary bone marrow derived macrophages (BMDMs) from B6 mice and mice with a targeted deletion in Card19 generated by Lexicon Genetics (designated Card19 lxcn ), in response to infection by Salmonella enterica serovar Typhimurium (S. Tm) or LPS +ATP treatment. S. Tm infection of BMDMs activates the well-characterized NAIP/NLRC4/caspase-1 inflammasome pathway, leading to pyroptosis and IL-1 cytokine release [4,5,42,43]. In contrast, LPS+ATP treatment activates the NLRP3/ASC/caspase-1 inflammasome, which induces pyroptosis and IL-1 cytokine release via activation of the P2X7 receptor and potassium efflux [44][45][46][47][48]. Card19 lxcn BMDMs displayed a striking and significant defect in cell death following S. Tm infection or LPS+ATP treatment, as measured by release of LDH (Figs 1A, 1B and S1A and S1B). Consistently, Card19 lxcn BMDMs exhibited significantly reduced levels of membrane permeability as determined by PI uptake, both in kinetics and amplitude (Fig 1C and 1D).
Peritoneal macrophages from Card19 lxcn mice have reduced levels of cell death while 2 days post-injection of LPS or other inflammatory triggers, such as thioglycolate, the primary population in the peritoneal cavity shifts to the SPM [56,57]. Interestingly, thioglycolate-elicited peritoneal macrophages exhibited significantly higher levels of LDH than peritoneal macrophages isolated from PBS-treated mice following infection with either S.Tm or Yptb (Fig 3A-3C). These data indicate that SPM undergo elevated levels of cell death relative to LPMs. Moreover, thioglycolate-elicited PMs from Card19 lxcn mice had significantly lower levels of LDH release that C57BL/6J PMs. It is possible that the results could be explained by a defect in SPMs in Card19 lxcn mice following thioglycolate treatment; however, these results are consistent with our findings in primary BMDMs that ex vivo isolated macrophages from Card19 lxcn animals also have a defect in cell lysis (Fig 3D and 3E). Card19 lxcn BMDMs are deficient for caspase-dependent cell death. Primary C57BL/6J (B6) or Card19 lxcn BMDMs were infected with Salmonella enterica serovar Typhimurium (S. Tm) or Y. pseudotuberculosis (Yptb), or treated with LPS+ATP or staurosporine (Sts), as indicated, and the kinetics of cell death was assayed by release of lactate dehydrogenase (LDH) (A, B, E, F) or propidium iodide (PI) uptake (C, D, G, H). Each figure is representative of three or more independent experiments. LDH release was assayed at specific times post-infection. PI uptake was measured over a two or ten-hour timecourse with fluorometric measurements taken at 2.5 minute (C, D) or 10 minute (G, H) intervals. Mean ± SEM of triplicate wells is displayed. Each panel is representative of three or more independent experiments. ��� p < 0.001, �� p < 0.01, � p < 0.05. n.s. not significant, analyzed by 2-way ANOVA with Bonferroni multiple comparisons post-test. https://doi.org/10.1371/journal.ppat.1009967.g001

PLOS PATHOGENS
NINJ1 deficiency in Card19 line abrogates cell lysis in response to cell death stimulii Card19 lxcn BMDMs retain intracellular alarmin HMGB1 following activation of cell death HMGB1 is a nuclear chromatin-associated protein that is released from cells upon loss of plasma membrane integrity [58]. As expected in B6 cells, S. Tm or LPS+ATP treatment led to significant HMGB1 release into the supernatant by one-hour post-infection, and by 4 hours post-exposure to Yptb or staurosporine (Sts) (Fig 4A). In contrast, Card19 lxcn cells exhibited significantly delayed release of HMGB1 for each respective timepoint, consistent with our findings that they exhibit increased resistance to membrane rupture (Fig 4A). HMGB1 was

PLOS PATHOGENS
NINJ1 deficiency in Card19 line abrogates cell lysis in response to cell death stimulii virtually undetectable in Card19 lxcn macrophage supernatants one-hour post-treatment with S. Tm or LPS+ATP, and minimally detectable at 4 hours post-treatment in response to Yptb or Sts (Fig 4A). To obtain insight into the dynamics of HMGB1 release from cells, we analyzed HMGB1 release by confocal microscopy following exposure of Wt and Card19 lxcn cells to Sts. We observed three morphologies of HMGB1 staining-nuclear, in which the majority of HMGB1 was located in the nucleus; cellular, in which HMGB1 was absent from the nucleus but still contained within the confines of a cell; and extracellular, in which a cloud of HMGB1 was detectable surrounding what was likely the remnants of apoptotic cells and cellular debris. Notably, and consistent with intracellular retention of HMGB1 in Card19 lxcn cells, a higher frequency of Card19 lxcn cells relative to B6 BMDMs retained nuclear HMGB1 (Fig 4B and 4C), while B6 BMDMs exhibited extracellular HMGB1 cloud formation much more frequently than Card19 lxcn BMDMs (Fig 4B white arrows and quantified in Fig 4D). Altogether, these findings indicate that Card19 lxcn BMDMs resist cell lysis and release of intracellular contents during induction of cell death triggered by caspase-1 or caspase-8.

Card19 lxcn macrophages are not deficient in activation of caspase-1 or -8
Most currently known regulators of cell death act at the level of caspase activation or assembly of caspase-activating inflammasome complexes [59]. Pyroptosis is mediated by activation of caspase-1 or -11 [15,59,60], which cleave Gasdermin D (GSDMD), thereby triggering cell lysis by enabling formation of oligomeric N-terminal GSDMD pores in the plasma membrane [14,16,61,62]. Surprisingly, despite the significant reduction in cytotoxicity in Card19 lxcn cells, processing of caspase-1 was equivalent in B6 and Card19 lxcn cells 15 minutes post-infection, at which time cell death is virtually undetectable in either genotype (Fig 5A). Moreover, B6 cells released significant amounts of cleaved and pro-caspase-1 into cell supernatants at later timepoints, in contrast to Card19 lxcn BMDMs (Fig 5A), consistent with the reduced levels of pyroptosis in Card19 lxcn cells in response to S. Tm infection. Reduced levels of cleaved caspase-1 in Card19 lxcn BMDM supernatants was matched by increased amounts of cleaved caspase-1 in Card19 lxcn whole cell lysates (Fig 5A). Similarly, initial processing of GSDMD to the cleaved p30 form was equivalent in B6 and Card19 lxcn cell lysates, while the release of processed GSDMD into the supernatant was significantly reduced in Card19 lxcn cells (Fig 5A). Thus, Card19 lxcn cells have no defect in either caspase-1 activation, or in the cleavage of caspase-1 targets.
Cell-extrinsic apoptosis is mediated by oligomerization and auto-processing of caspase-8 [63,64]. Active caspase-8 directly cleaves its pro-apoptotic downstream targets, including the Bcl-2 family member Bid [65,66], and the executioner caspases-3 and -7. Similarly to caspase-1, cleavage of caspase-8 as well as the processing of caspase-8 substrates, including Bid, caspase-3, the caspase-3 target poly-ADP ribose polymerase (PARP), and the pore-forming protein GSDME also known as DFNA5 [18,67], were equivalent in B6 and Card19 lxcn cells in response to Yptb infection or staurosporine treatment (Fig 5B and 5C). Notably, we observed GSDMD processing in response to Yptb infection as well as Sts treatment, in both B6 and Card19 lxcn BMDMs (Fig 5C). Importantly, direct comparison of caspase-1 and GSDMD processing in response to multiple stimuli revealed that Card19 lxcn cells had reduced release of cleaved caspase-1 and N-terminal GSDMD into the supernatant, and correspondingly increased retention in cell lysates, consistent with reduced terminal lysis of the cells in the absence of CARD19 (Fig 5D). Card19 lxcn BMDMs had wild-type levels of caspase-8 activity in response to either Yptb infection or Sts treatment, confirming that Card19 lxcn BMDM do not have a defect in caspase activity (Fig 5E). Altogether, these findings demonstrate that Card19 lxcn BMDMs do not have a defect in caspase activation or downstream target cleavage, and therefore likely have a defect in cell lysis downstream of caspase activation. How Card19 lxcn cells might retain membrane integrity downstream of both inflammatory and apoptotic caspase activation is not clear. However, caspase-8 and caspase-1 can both cleave IL-1β [51,68,69], as well as GSDMD [17,18,20] raising the question of whether Card19 lxcn cells and Gsdmd -/cells have similar defects in cell lysis in response to Yptb infection. Surprisingly, although Gsdmd -/-BMDMs exhibited a significant defect in cell death in response to Yptb infection (S2A Fig), this defect was not as pronounced as in Card19 lxcn BMDMs, suggesting that Card19 lxcn cells likely have a defect in additional GSDMD-independent mechanisms of cell lysis, potentially involving GSDME/DFNA5, which was also processed in response to Yptb infection (S2B and S2C Fig). Consistent with recent findings [17][18][19], we observed cleaved GSDMD in Casp1/11 -/lysates following either Yptb infection or staurosporine treatment. Although levels of cleaved GSDMD were significantly lower than in Wt cells, this finding is consistent with findings that caspase-8 cleaves GSDMD less efficiently than caspase-1 (S2B and S2C Fig) [19]. The caspase-8-selective inhibitor, IETD, abrogated GSDMD and GSDME cleavage in response to Yptb or staurosporine (S2B and S2C Fig), suggesting that Yersinia and staurosporine induce GSDMD and GSDME cleavage in Casp1/11 -/cells in a caspase-8-dependent manner. Consistently, Ripk3 -/-BMDMs showed robust cleavage of GSDMD and GSMDE while Ripk3 -/-Casp8 -/-BMDMs showed a near-complete absence of N-terminal GSDMD and GSDME p30 following Yptb infection and significant reduction in GSDMD and GSDME pro-

The cell death defect in Card19 lxcn BMDMs is independent from cytokine release
Consistent with our findings that Card19 lxcn BMDMs are not defective in their ability to regulate proteolytic activity of caspase-1, CARD19-deficient cells released wild-type levels of IL-1 cytokines in response to S. Tm, LPS+ATP, or infection with the ΔyopEJK Yptb strain, which triggers a caspase-11/NLRP3-dependent pathway of IL-1 cytokine release [70,71] (Fig 6A and  6B). These data indicate that Card19 lxcn cells most likely do not have a defect at the level of GSDMD-dependent pore formation and cytokine release, and that the Card19 lxcn cells decouple secretion of IL-1β from pyroptosis, which normally occurs in response to these stimuli [26,32]. Consistent with a recent finding that Card19 lxcn cells do not have altered NF-κB activation [33], Card19 lxcn BMDMs did not have a defect in secretion of IL-6, TNF, or IL-12p40, (Fig 6C-6E). Altogether, these findings suggest that Card19 lxcn cells have decoupled cell lysis as determined by LDH and HMGB1 release and the release of caspase-1-dependent cytokines.

PLOS PATHOGENS
NINJ1 deficiency in Card19 line abrogates cell lysis in response to cell death stimulii indicate that lack of death in Card19 lxcn cells is not evidence of a hyperactivation phenotype, but suggests Card19 lxcn BMDMs are likely resistant to cell lysis due to a defect at a terminal stage of pyroptosis. Curiously, ΔoatA-infected cells exhibited significantly lower levels of GSDMD processing, indicating that threshold levels of GSDMD processing or membrane insertion may underlie the distinction between hyperactivation and cell lysis ( Fig 6F).

Card19 lxcn BMDMs are hypomorphic for NINJ1 gene and protein expression
Numerous mitochondrial proteins have been implicated in recent years as essential regulators of caspase activity, GSDMD and terminal events in during lysis [35,37,73]. While these studies were in progress, Sterile alpha and armadillo motif-containing protein (SARM1) was reported to regulate terminal cell lysis following stimulation of the NLRP3 inflammasome [40], and like CARD19, also localizes to the mitochondrial outer membrane [33]. We therefore considered the possibility that SARM1 might contribute to cell lysis via interaction with CARD19. Unexpectedly however, BMDMs isolated from multiple independent lines of Sarm1 -/mice (described further in Materials and Methods), including Sarm1 -/mice reported by Carty et al. Moreover, TNF release was unaffected in cells lacking SARM1 compared to WT controls, indicating that the NLRP3 inflammasome priming step (NF-κB pathway) is not dysregulated by the absence of SARM1 (S3E Fig). Furthermore, in contrast to Card19 lxcn BMDMs, the Sarm1 -/macrophage lines had wild-type levels of cell death and cytokine release in response to non-canonical inflammasome activation (S3F and S3G Fig). Notably, some Sarm1 -/lines were reported to contain a passenger mutation in the closely-linked gene Xaf1, indicating that some previous cell death-related phenotypes attributed to SARM1 may be due to alteration of Xaf1 [75]. Altogether, these findings indicate that the phenotype of Card19 lxcn BMDMs is not related to potential interactions with SARM1.
Next, to test the sufficiency of CARD19 to induce cell death, we generated immortalized Card19 +/+ and Card19 lxcn bone marrow hematopoietic progenitors using the ER-HOXB8 system [76], and transduced the progenitors with CARD19-expressing lentiviral constructs. Unexpectedly, Card19 lxcn iBMDMs transduced with CARD19 displayed comparable death to vector control-transduced iBMDMs, raising questions as to whether CARD19-deficiency was directly responsible for the observed defect in caspase-dependent death in Card19 lxcn BMDMs (Fig 7A). , Card19 lxcn , Casp1/11 -/or Gsdmd -/-BMDMs were left untreated or infected with S. Tm, and cell lysates (Lys) and TCA-precipitated supernatants (Sups) were run on SDS-PAGE and analyzed by western blotting for cleaved Casp1, GSDMD, or Actin (cell lysate loading control). Cell death from these cells was assayed in parallel by LDH release (indicated below blots). Mean ± SEM is displayed. 2-way ANOVA with Bonferroni multiple comparisons post-test. ��� p < 0.001, �� p < 0.01, � p < 0.05. n.d. not detectable. (B) B6 (+) or Card19 lxcn (-) BMDMs were infected with Yptb or treated with Sts or 2, 3, or 4 hours, or left untreated (Un) as indicated. Cell lysates and TCA-supernatants (HMGB1 sup) were prepared at indicated times and analyzed for presence of CARD19, cleaved caspase-8, tBid, cleaved caspase-3, cleaved PARP, HMGB1 and Actin (loading control). (C) B6 and Card19 lxcn BMDMs were left untreated, infected with Yptb or treated with staurosporine. Lysates were harvested at the indicated time points, run on SDS-PAGE and analyzed by western blotting for CARD19, GSDMD, GSDME/DFNA5, and actin (loading control).  Passenger mutations in genes closely linked with the gene of interest in backcrossed C57BL/6 mice can confound interpretation of immune responses and cell death pathways in these lines [23,77,78]. Notably, Card19 lxcn mice were originally generated using 129SvEvBrd ESCs, and backcrossed for 10-12 generations [33]. Card19 +/lxcn and 129/SvIm/J BMDMs phenocopy wildtype BMDMs with respect to their ability to undergo terminal cell lysis in response to S. Tm., Yptb, and Sts, indicating that the phenotype is linked to the Card19 locus (S4A- S4C  Fig). To rule in or rule out the possibility that a passenger mutation linked to the Card19 locus might be responsible, we next generated two independent Card19-deficient murine lines directly on the C57BL/6J background using CRIPSR/Cas9. The first of these removed exons 2-3, removing the entire CARD domain, but leaving a truncated mRNA that could produce a peptide fragment containing the transmembrane domain, which we termed Card19 ΔCARD (S4D Fig and S1 Table). The second was a complete deletion of the Card19 locus, which we termed Card19 Null (see Materials and Methods for further details of construction of these lines). Critically, Card19 ΔCARD and Card19 Null BMDMs exhibited no observable defect in their ability to undergo cell lysis following infection with either Yptb or treatment with LPS+ATP, in contrast to BMDMs derived from the original Card19 lxcn line (Fig 7B and 7C). Altogether, these data indicate that loss of CARD19 is not the mechanistic basis of the defect in cell death seen in the original Card19 lxcn cells.
To identify the regions of the Card19 lxcn chromosome that may contain the original 129SvEvBrd lineage, we performed SNP genotyping analysis of the Card19 lxcn mice which indicated that approximately six megabases adjacent to the Card19 locus were derived from the 129SvEvBrd lineage (S5 Fig and S2 Table). We additionally performed whole exome sequencing on B6 and Card19 lxcn mice to address the possibility that Card19 lxcn mice contain a functionally important polymorphism or have an expression defect in a gene closely linked to Card19. Although whole exome sequencing identified potential mutations in eight genes on chromosome 13, they were not high probability candidates (S3 Table).
We further performed RNA-seq on B6 and Card19 lxcn BMDMs, with the hypothesis that transcriptional alteration of genes closely linked to Card19 lxcn might result in the observed loss of cell lysis (Fig 7D and S4 Table). Intriguingly, Ninj1, recently identified as a key regulator of plasma membrane rupture during lytic cell death in response to apoptotic, pyroptotic and necrotic stimuli [41] is directly upstream of Card19, making it a likely candidate for passenger mutations or off-target effects (S5 Fig, blue arrow). Card19 lxcn and Ninj1 -/-BMDMs have a strikingly similar phenotype, as they both have a defect in cell lysis that is independent of IL-1 cytokine release and processing of caspases and gasdermins ([41] , Figs 1-6). SNP mapping with a high-density SNP array identified a 129SvEvBrd-derived region in Card19 lxcn cells downstream of both Ninj1 and Card19 that also contained several B6 SNPs (S2 Table and S5  Fig). It is therefore likely that complex recombination occurred at the Ninj1/Card19 locus. To test whether any mutations in Ninj1 were present in the Card19 lxcn mice, we directly sequenced all 5 exons of Ninj1, including the 3' splice acceptor site for exon 2, reported to be mutated in the ENU-mutagenesis screen that first identified Ninj1, but were not able to detect any deviations from the reference genome in the Ninj1 sequence in Card19 lxcn mice. However, Ninj1 mRNA levels were reduced at baseline in Card19 lxcn BMDMs relative to wild-type BMDMs, and a knockout-validated antibody [41] demonstrated that NINJ1 protein levels were significantly reduced but not entirely absent in Card19 lxcn BMDMs (Fig 7E). Importantly, NINJ1 protein levels were not reduced in Card19 Null BMDMs (Fig 7F). To directly test whether loss of NINJ1 in the Card19 Lxcn BMDMs might be responsible for the defect in cell lysis, we reconstituted Card19 lxcn iBMDMs with Ninj1 and electroporated with LPS. Critically, Ninj1 reconstitution restored cell lysis to wildtype levels, similarly to the observations for Ninj1 -/cells (Figs 7G and S4B). Altogether, our findings indicate that in addition to loss of CARD19, Card19 lxcn mice also have substantially reduced NINJ1 protein levels, most likely as a result of Card19 targeting, leading to a defect in cell lysis consistent with the loss of NINJ1 function.

Card19 lxcn mice exhibit increased susceptibility to Yersinia
To assess whether the impact on cell death was biologically relevant during an in vivo infection in which cell death is important for systemic bacterial control, we infected Card19 lxcn mice with the gram-negative Yersinia pseudotuberculosis (Yptb). Consistent with findings that mice lacking essential mediators of cell death are more susceptible to bacterial infection [49,[79][80][81], Card19 lxcn mice were significantly more susceptible to oral Yptb infection and had a defect in controlling systemic Yptb tissue burdens (Fig 8A and 8B). Interestingly, Card19 lxcn mice were equally susceptible to oral infection with S. Typhimurium (S. Tm), suggesting that cytokines released during GSDMD pore formation independent of cell lysis are sufficient to mediate control of S. Tm in vivo (S6A Fig). Importantly, Card19 Null mice did not have increased susceptibility to Yptb, indicating that loss of NINJ1 and not CARD19 is responsible for the phenotype of Card19 lxcn mice (S6B Fig). Yptb-infected Card19 lxcn mice exhibited increased splenomegaly (Fig 8C and 8D), suggesting either failure to control infection and elevated recruitment of inflammatory cells to the spleen, or a failure of innate cells to undergo cell death during bacterial infection in vivo. Card19 lxcn mice had similar levels of serum IL-6 and IL-12, although they had increased levels of TNF, consistent with their increased bacterial burden (Fig 8E). Yptb-induced cell death is dependent on caspase-8 [49,80]. We therefore also infected Ripk3 -/-Casp8 -/mice with Yptb to determine whether the phenotype of mice deficient in cell lysis due to hypomorphic Ninj1 expression is similar to that of mice lacking caspase-8. Notably, and consistent with our prior studies on caspase-8, Yptb oral infection led to rapid acute lethality in Ripk3 -/-Casp8 -/mice (S6C Fig), suggesting that while NINJ1-dependent cell lysis contributes to host defense against Yersinia, additional caspase-8-dependent factors likely play a role as well. Altogether these data indicated that NINJ1-mediated cell lysis is not important for control of cell-intrinsic cytokine production but contributes to host defense against bacterial infection.

Discussion
Our study has revealed that Card19 lxcn BMDMs are resistant to cell lysis during both pyroptosis and caspase-8-mediated apoptosis while maintaining caspase activation and release of IL-1 family cytokines. This data highlights the existence of a regulatory checkpoint downstream of gasdermin processing that controls the cell fate choice to undergo terminal cell lysis. Until recently, caspase-dependent cleavage of gasdermin D, and the subsequent release of the

PLOS PATHOGENS
NINJ1 deficiency in Card19 line abrogates cell lysis in response to cell death stimulii gasdermin D N-terminal pore-forming domain was viewed as the terminal step in lytic cell death. However, some cell types or stimuli trigger IL-1β release without cell lysis, indicating that these are distinct cellular responses that can be decoupled [28,29,31,82]. Moreover, the recent observation that GSDMD-dependent release of IL-1β is regulated by charge-charge interactions highlights the selective nature of the GSDMD pore. That gasdermin-dependent secretion of IL-1 cytokines can be uncoupled from cell lysis further points to regulatory steps subsequent to cleavage and activation of gasdermin D that can be engaged to promote or limit cell lysis.
While our initial observations that Card19 lxcn macrophages were resistant to cell lysis despite having wild-type levels of GSDMD processing suggested that CARD19 might regulate cell lysis, two CRISPR/Cas9-generated lines of mice that we generated directly on a B6 background (Card19 ΔCARD and Card19 Null ) did not recapitulate this phenotype. Moreover, expressing CARD19 in Card19 lxcn cells did not complement the lysis phenotype. These findings indicated that a CARD19-independent factor that was presumably affected in Card19 lxcn mice was most likely responsible for the resistance of Card19 lxcn cells to cell lysis. Intriguingly, Ninj1 which is immediately upstream of Card19 on chromosome 13, regulates cell lysis in response to pyroptotic and apoptotic stimuli [41]. Ninj1 encodes a protein whose function was previously linked to axonal guidance, and was identified in a forward genetic screen for regulators of cell death [41]. The phenotype of Ninj1 -/cells is strikingly similar to that of Card19 lxcn BMDMs [41]. Although we were unable to identify any mutations in the NINJ1 coding sequence or intronic splice acceptor that would indicate a potential defect in NINJ1 in Card19 lxcn mice, Card19 lxcn BMDMs have substantially diminished levels of NINJ1 gene and protein expression, implicating NINJ1 as the likely driver for the phenotype in the Card19 lxcn line. Furthermore, reconstitution of Ninj1 in Card19 lxcn iBMDMs restored wildtype levels of cell lysis following LPS electroporation. Together, these results indicate that genetic targeting of Card19 resulted in an off-target effect on NINJ1 expression, leading to impaired terminal cell lysis and anti-Yersinia defense.
Unintended off-target effects on neighboring genes following genetic targeting have previously been reported. Notably, insertion of a PGK-Neo cassette into the granzyme B locus and β-globin locus control regions was found to disrupt multiple genes in the surrounding area [83]. Targeting of exons 13-16 of SH3 and multiple ankyrin repeat domains 3 (Shank3B) with a neomycin resistance cassette resulted in altered expression of neighboring genes as well as increased expression of exons 1-12 and the formation of an unusual Shank3 isoform not expressed in wildtype mice [84]. Furthermore, RNA-seq analysis of target and neighboring gene expression in a number of homozygous mutants revealed increased frequency of downregulated neighboring genes that would be expected due to local transcription dysregulation [85]. Our finding that genetic targeting of Card19 resulted in decreased expression of Ninj1 are consisted with these reports. Our findings underscore the complexity of genetic manipulation in mice and the importance of complementation analyses in validating knockout phenotypes. In addition to the finding that NINJ1 regulates cell lysis, two separate CRISPR-based screens revealed the Ragulator/Rag/mTORC pathway as being important for GSDMD-dependent pore formation as well as Yersinia-induced activation of caspase-8, although they propose distinct roles for the function of the Ragulator complex in cell death [86,87]. Intriguingly, Evavold et al. propose that the Ragulator complex acts downstream of GSDMD cleavage to mediate pore formation, but upstream of NINJ1-induced cell lysis, whereas NINJ1 acts subsequent to pore formation to induce membrane rupture and ultimate cytolysis. Thus, in contrast to initial models proposing that release of the N-terminal portion of GSDMD was necessary and sufficient to trigger cell lysis, these findings collectively support a multi-step regulated process, downstream of GSDMD cleavage, that ultimately triggers lytic rupture of the cell.
How terminal cell lysis occurs downstream of GSDMD cleavage or plasma membrane insertion remains enigmatic. The mitochondrial protein SARM1 was previously reported to regulate cell lysis in response to NLRP3 inflammasome activation independently of IL-1β cytokine release. Mitochondrial disruption is a feature of both lytic and apoptotic forms of cell death, and the N-terminal GSDMD fragment is recruited to the mitochondrial membrane as well [73,88]. Unexpectedly however, primary macrophages generated from four independent lines of Sarm1 -/-BMDMs had wild-type levels of cell lysis and IL-1β cytokine release. Whether passenger mutations, which have been reported in Sarm1 -/murine lines [75], may account for the reported phenotype of Sarm1 -/-BMDMs, remains to be determined. Precisely how cell lysis can be uncoupled from IL-1 cytokine release remains a key unanswered question in this field. BMDMs from Card19 lxcn , Ninj1 -/-, and Rag -/mice all display reduced cell lysis despite substantial levels of IL-1β cytokine release, similar to the hyperactivation phenotype previously described by Kagan and colleagues [30,31]. However, we observed that BMDMs induced to undergo hyperactivation following infection with ΔoatA mutant S. aureus exhibit reduced levels of total caspase-1 and GSDMD cleavage, suggesting that the regulated step in hyperactivation occurs at the level of caspase processing or complex assembly, rather than at the level of GSDMD pore formation.
In addition to demonstrating that terminal cell lysis is a regulated step downstream of GSDMD activation, this study extends our understanding of the role of lytic cell death in antibacterial host defense. Card19 lxcn mice showed significantly increased susceptibility to Yersinia infection, with higher systemic burdens and reduced ability to survive infection, consistent with our previous studies demonstrating that the RIPK1-Casp8-dependent cell death pathway enables activation of cytokine production from uninfected bystander cells. Interestingly, Ripk3 -/-Casp8 -/mice were even more susceptible to Yptb infection, which could be due to the fact that caspase-8-deficient cells are completely resistant to Yptb-induced cell death, or that caspase-8 also has cell-death-independent functions in regulating inflammatory cytokine production [89][90][91]. Altogether, our work highlights cell lysis and the release of inflammatory mediators from ruptured cells as a component of inflammatory responses that play an important role in mediating anti-microbial immune defense.

Ethics statement
All animal studies were performed in compliance with the federal regulations set forth in the Animal Welfare Act (AWA), the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health, and the guidelines of the University of Pennsylvania Institutional Animal Use and Care Committee. All protocols used in this study were approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania, (Animal Welfare Assurance Number D16-00045/A3079-01, Protocol #804523).
All reagents and resources are listed in S5 Table. Generation of Card19 ΔCARD mice Card19 ΔCARD mice were produced by the CRISPR/Cas9 Mouse Targeting Core (PSOM-U-Penn) using CRISPR/Cas9 technology (flanking sgRNA and Cas9mRA microinjected in 1-cell mouse embryos), and a targeting strategy resulting in removal of exons 2-3. The resulting CARD19 deletion removes amino acids 3-99 including the entire CARD and giving a predicted protein product of 87 AA in length.

Generation of Card19 Null mice
Card19 Null mice were produced by the CRISPR/Cas9 Mouse Targeting Core (PSOM-UPenn) using CRISPR/Cas9 technology (flanking sgRNA and Cas9mRA microinjected in 1-cell mouse embryos), and a targeting strategy resulting in deletion of exon 1, resulting in the premature appearance of a stop codon and degradation by nonsense-mediated decay. CARD19 5' Target Sequence: TAGGCGAAGGGACGCCGACCCGG. CARD19 3' Target Sequence: GTTTCTGGTACGAGCTGGCAGGG. Mice were backcrossed onto C57BL/6J for at least 2 consecutive generations.

Mouse strains
Card19 lxcn mice were previously described [33] and maintained as a breeding line in-house. This strain has been deposited at the Jackson Laboratory and is available as B6.129-Card19tm1Lex/SchaJ, Stock No. 036463. B6 and Casp1/Casp11 -/mice obtained from Jackson Laboratories and subsequently maintained as a breeding line in-house. Previously published knockout mouse lines that were used to generate BMDMs are indicated in S1 Table. Gsdmd -/-BMDMs were previously described [94], provided by Russell Vance, and maintained as a breeding line in-house. Ripk3 -/mice [95] were a gift of Kim Newton and Vishva Dixit (Genentech) and Ripk3/Casp8 -/mice [96] were a gift of Doug Green (St. Jude Children's Hospital). Casp11 -/mice were originally generated by Junying Yuan [97] (Harvard University) and kindly provided by Tiffany Horng (Harvard University). Sarm1(MSD) -/mice were a gift from Adriano Aguzzi (University Hospital Zurich). Breeders were routinely genotyped. Mice were maintained in a specific pathogen-free facility by University Laboratory Animal Resources (ULAR) staff in compliance with IACUC approved protocols.

Generation of immortalized Card19 lxcn myeloid progenitors using ER-HOXB8
Card19 lxcn murine myeloid progenitors were immortalized using the ER-HoxB8 system [76]. Bone marrow was harvested from female mice and myeloid progenitors were isolated using a Percoll density gradient. Progenitors were plated in 5% SCF-conditioned media supplemented with 10 ng/mL IL-3 and IL-6 for three days. Progenitors were spinfected on fibronectin coated plates with ER-HoxB8 retrovirus in 25 ug/mL polybrene at 1000g for 90 minutes. Progenitors were supplemented with 0.5 uM estrogen and 10 ng/mL GM-CSF to induce macrophage progenitor differentiation. Progenitors were replated with fresh estrogen and GM-CSF for two days. Retrovirally infected cells were selected for with the addition of 1 mg/mL geneticin for 2-3 days. Progenitors were harvested and passaged until immortalization was confirmed by the death of all uninfected control cells. Mature macrophages were generated from immortalized progenitors by plating cells in 30% macrophage media and growing for 5-6 days. Cells were fed with additional macrophage media on day 3.

Transduction of Card19 lxcn immortalized progenitors
HEK293T cells were transfected using Lipofectamine 2000 with pCL-Eco retroviral packaging vector and MSCV2.2-CARD19 or empty vector (MSCV2.2). Successful transfection was confirmed by imaging under a widefield as MSCV contains IRES-GFP. Card19 lxcn ER-HoxB8 immortalized progenitors were spinfected with filtered viral supernatants from transfected HEK293T cells at 2500 RPM for 90 minutes. Transduced progenitors were transduced a second time the following day as before to increase infection efficiency. Progenitors were allowed to recover and proliferate prior to 2.5 ug/mL puromycin selection. A 95%+ pure population was isolated by flow cytometry sorting on GFP+. Mature macrophages were generated from immortalized, transduced progenitors by plating cells in 30% macrophage media and growing for 5-6 days, with additional media supplementation on day 3.

Plasmids and constructs
All constructs are listed in S5 Table. pcDNA3.1/CARD19-FLAG was obtained from GenScript (Refseq NM_026738.2). CARD19 was amplified and inserted into MSCV2.2 using Gibson cloning. All constructs were confirmed by sequencing prior to experimentation.
Cell death assays-LDH, PI uptake, & Dextran Dye-150 release LDH. Triplicate wells of BMDMs were seeded in TC-treated 96 well plates. BMDMs were infected with indicated bacterial strains as indicated above. BMDMs were primed with 100 ng/ mL LPS for 3 hours followed by 2.5 mM ATP treatment. BMDMs were treated with 10 uM staurosporine or 100 uM etoposide. BMDMs were pretreated with 100 uM zVAD(OMe)-FMK or 100 ug/mL cycloheximide for 1 hour before treatment with 100 ng/mL LPS. For non-canonical inflammasome cell death, BMDMs were primed with 400 ng/mL Pam3CSK4 for 3 hours and then infected with DH5α E. coli, MOI 30 for 16-20 hours. 100 ug/mL gentamycin was added one hour post treatment to all infectious experimental conditions. BMDMs were primed with LPS (100 ng/ml) for 4 hours and stimulated with nigericin (5 μM). BMDMs were primed with Pam3CSK4 (1 μg/ml) for 4 h and were transfected with 2 μg/ml E. coli O111:B4 LPS with Fugene HD. At indicated time points, plates were spun down at 250g and supernatants were harvested. Sups were combined with LDH substrate and buffer according to the manufacturer's instructions and incubated in the dark for 35 min. Plates were read on a spectrophotometer at 490 nm. Percent cytotoxicity was calculated by normalizing to maximal cell death (1% triton) and no cell death (untreated cells).
PI uptake. Propidium Iodide uptake was performed as previously described [26]. Briefly, triplicate wells of BMDMs were seeded in TC-treated black-walled 96 well plates. BMDMs were infected or treated as described above in 50 uL HBSS plus 10% FBS. Propidium Iodide (2x, 10 uM) was added in 50 uL HBSS to each well and incubated for 5 minutes in the dark to allow for stabilization of the signal in the maximal cell death wells (1% triton). PI uptake was detected by fluorescence on a BioTek Synergy HT Multi-Detection Microplate Reader (540/25 excitation, 590/35 emission, sensitivity 35, integration time 1 sec) every 2.5 minutes (S. Tm., ATP) or every 10 minutes (Yptb, Sts) for the indicated time points. Gentamycin (100 ug/mL) was added one hour post infection (Yptb, Sts). Percent cytotoxicity was calculated as described above.

Confocal microscopy
BMDMs were seeded as described above, treated with staurosporine for indicated time points, fixed, permeabilized, and blocked. BMDMs were stained for HMGB1 (1:200) at 37˚C for 1 hr, Alexa Fluor rabbit 488 (1:4000) at RT for 1 hr, and Hoechst and Phalloidin at RT for 30 min. Slides were imaged as described above, with a single z-plane taken per field. Percent nuclear HMGB1 was quantified by identifying total HMGB1 staining per cell and comparing the overlap with nuclear staining (Hoechst). 65-200 cells were analyzed per genotype and condition. Cloud analysis was completed by counting the number of HMGB1 clouds per field of view with 5-8 fields per condition.

Caspase 8 activity
BMDMs were seeded as described above in a white-walled 96 well plate and treated with Yptb or Sts for the indicated times. Z-IETD-fmk (500 uM, SM Biochemicals) was added 1 hour prior to infection/treatment as a control to block caspase-8 activity. Caspase-8 activity was detected using Caspase-8 Glo (Promega) according to the manufacturer's instructions. Luminescence was read on a Gen5 plate reader.

Cytokine release
Triplicate wells of BMDMs were seeded in TC-treated 48 well plates. All conditions except untreated were primed with 100 ng/mL LPS for 3 hrs or unless otherwise indicated. BMDMs were infected with bacterial strains or treated with 2.5 mM ATP. BMDMs were primed with LPS (100 ng/ml) for 4 hours and stimulated with nigericin (5 μM). IL-1β release were measured at the indicated time points. BMDMs were primed with LPS (100 ng/ml) and TNF release was measured after 4 hours. BMDMs were primed with Pam3CSK4 (1 μg/ml) for 4 h and were transfected with 2 μg/ml E. coli O111:B4 LPS with Fugene HD. IL-1β release were measured after 16 h. ELISA supernatants were added to IL-1α, IL-1β, IL-6, IL-12, and TNF-α capture antibody-coated 96-well plates and incubated at 4˚C overnight. Plates were incubated with the appropriate biotinylated antibodies in 1% BSA, followed by streptavidin. ELISA was developed with o-phenylenediamine dihydrochloride in citric acid buffer and stopped with 3M sulfuric acid. Plates were read on a spectrophotometer at 490 nm.

SNP sequencing
Tails from four Card19 lxcn mice were sent to Dartmouse for a genetic background check. Each sample was interrogated over 5300 SNPs space along the mouse genome with an average density of 0.5 Mbp using an Illumina Infinium Genotyping Assay.

Whole exome sequencing
Genomic DNA was extracted from 1 B6 and two Card19 lxcn BMDMs derived from three independent mice using a DNeasy blood and tissue kit (Qiagen) according to the manufacturer's instructions. gDNA was shipped to Genewiz for Illumina HiSeq 2x150 bp sequencing. C57BL/ 6J was used as a reference genome.

RNA-seq
B6 and Card19 lxcn BMDMs were prepared as above from 3 independent mice per genotype. 1×10 6 BMDMs were treated with 100 ng/mL LPS for 3 hours or left untreated. RNA was extracted using the RNeasy Plus Mini Kit (Qiagen) according to the manufacturer's instructions. Sequence-ready libraries were prepared using the Illumina TruSeq Total Transcriptome kit with Ribo-Zero Gold rRNA depletion (Illumina). Quality assessment and quantification of RNA preparations and libraries were carried out using an Agilent 4200 TapeStation and Qubit 3, respectively. Samples were sequenced on an Illumina NextSeq 500 to produce 75-base pair single end reads with a mean sequencing depth of 35 million reads per sample. Raw reads from this study were mapped to the mouse reference transcriptome (Ensembl; Mus musculus GRCm38) using Kallisto version 0.46.2 [98]. Raw sequence data are available on the Gene Expression Omnibus (GEO; accession no. GSE168489) All subsequent analyses were carried out using the statistical computing environment R version 4.0.3 in RStudio version 1.2.5042 and Bioconductor [99]. Briefly, transcript quantification data were summarized to genes using the tximport package [100] and normalized using the trimmed mean of M values (TMM) method in edgeR [101]. Genes with <1 CPM in 2 samples were filtered out. Normalized filtered data were variance-stabilized using the voom function in limma [102], and differentially expressed genes were identified with linear modeling using limma (FDR � 0.10; absolute log 2 FC � 1) after correcting for multiple testing using Benjamini-Hochberg. Volcano plots were made using the Enhanced Volcano package [103].

Mouse infections
Age and sex-matched B6, Card19 lxcn , Card19 Null , Ripk3 -/-, and Ripk3 -/-Casp8 -/mice between 8-12 weeks were orally infected with 10 8 CFU of Yersinia pseudotuberculosis strain IP2777 after fasting overnight. Mice were monitored for survival for 30 days or were euthanized using CO 2 followed by cervical dislocation, in compliance with IACUC approved euthanasia protocols. Blood was harvested by cardiac puncture following death. Serum cytokine levels (IL-6, IL-12, TNF-α) were measured by ELISAs as described above. Spleen, liver, mesenteric lymph nodes, and Peyer's patches were removed, weighed, homogenized, diluted in PBS, and plated on LB agar plates with irgasan for CFUs. B6 and Card19 lxcn were orally infected with 10 7 CFUs of Salmonella enterica serovar Typhimurium strain SL1344 after fasting overnight. Mice were monitored for survival for 30 days.

Thioglycolate injections
Age and sex-matched B6 and Card19 lxcn mice between 9-11 weeks were injected intraperitoneally with 4% aged thioglycolate or PBS. 48 hours after injection, mice were euthanized using CO 2 . The peritoneum was lavaged with 10 mL cold PBS and the peritoneal exudate cells were spun down. RBCs were lysed using RBC lysis buffer. PECS were counted using trypan blue exclusion on a hemocytometer and plated in complete DMEM without antibiotics in 96 well plates (1×10 5 cells for PBS PECS, 0.7×10 5 cells for TG PECS). Cells were washed with warm PBS to remove non-adherent cells and infected with Yptb, S. Tm. or treated with staurosporine as described above.

Statistical analysis
Statistical analysis was completed using GraphPad Prism. Two-tailed Student's t test or paired Student's t test were used for comparisons of two groups. One-way analysis of variance (ANOVA) with pairwise comparisons and Bonferroni post-test correction was used for multiple group comparisons. Repeated-measures ANOVA or paired t tests were used for matched samples. Studies were conducted without blinding or randomization. Values of p < 0.05 were considered statistically significant.
Supporting information S1 Fig. CARD19 is required for optimal caspase-dependent cell death. Primary C57BL/6J (B6), Card19 -/-, Card19 +/+ , and Casp11 -/-BMDMs were treated with (A) S. Tm (B) Yptb, and cell death was assayed by LDH release. BMDMs were treated, supernatants were harvested from triplicate wells at indicated time points and measured for cytotoxicity. The mean ± SEM of means from triplicate wells from 3-19 independent experiments as indicated in parenthesis.  Table. BMDM and Murine Sources. Sources of BMDMs and their derivation are listed in S1 Table. (DOCX) S2