To determine host and parasite requirements for eliciting antigen-specific CD8 T cell responses to T. gondii, we took advantage of ‘T57’ transnuclear mice which were cloned from the nucleus of a single tetramer-positive T. gondii-specific CD8 T cell. T57 T cells from these mice have a single T cell receptor (TCR) specificity for the TGD057_96-103 epitope presented by H-2K^b MHC 1 [49,57], and when adoptively transferred, confer resistance to infection with a type II strain [57]. TGD057 is a protein with unknown function but is predicted to be in the parasite’s secretory pathway [58], and when deleted does not negatively impact parasite fitness, as inferred from a genome-wide CRISPR-CAS9 loss-of-function screen (phenotype score 2.1) [59], and from similar growth rates in tissue culture and plaque sizes observed between Δtgd057 and parental strains. Importantly, TGD057 (TG_215980) is highly expressed (ToxoDB.org) and the peptide epitope is conserved between strains (S1 Fig), facilitating comparative analyses of naïve CD8 T cell responses to parasite strains which may differ in immune modulation. In our experimental setup (i.e. the T cell activation ‘T57 assay’) (Fig 1A), bone marrow-derived macrophages (BMDMs) are infected with T. gondii, co-cultured with splenocytes and lymph node cells from naïve transnuclear T57 mice, and CD8 T cell activation markers or effector cytokines in the supernatant are measured. Reflecting early T cell activation events culminating in calcium-dependent NFAT activation of the IL-2 gene [60], this cytokine is produced as early as 24 hours post addition of T cells to parasite-infected BMDMs (Fig 1B). In contrast and consistent with their naïve state, the T57 IFNγ response develops with time, and is maximally detected at 48 hours (Fig 1C). Its magnitude is influenced by the parasite genotype, as evidenced by consistently high cytokine responses to the ME49 and MAS strains, and low responses to the RH strain (Fig 1B and 1C). T. gondii infection elicits strong CD8 T cell IFNγ responses to TGD057 [49] and other antigens [61], as such IL-17 is only marginally detected in this system (S2A Fig). The measured phenotypes are antigen-specific, because the cytokine response is abolished in response to Δtgd057 strains which do not express the antigen (Fig 1D and 1E). Additionally, we noted that T57 CD8 T cells fail to upregulate the early activation marker CD69 in response to Δtgd057. Previous observations from sub-cellular fractionation and immunofluorescence studies have identified TGD057 to be both within the PV of infected cells [41] and to the cytoskeleton region of the parasite [48]. Three-dimensional mass spectrometry LOPIT analysis (Location of Organelle Proteins by Isotype Tagging) posits TGD057 to dense granules but lacks a strict assignment to any one organelle, the latter observation being consistent with most cytoskeleton network associated proteins [62]. An endotagged RH_tgd057-HA strain was generated and TGD057 is always found within PVM defined by GRA7 staining vacuoles (Fig 1F), demonstrating TGD057, in its natural state, stays inside the PV. In summary, the T57 system allows analysis of naïve CD8 T cell responses to an endogenous vacuolar antigen of T. gondii.

To understand how T. gondii vacuolar antigens might escape from the vacuole and enter the host’s MHC 1 antigen presentation pathway, the parasite’s export machinery was explored. One way for vacuolar proteins to enter the host cytosol is through parasite-mediated export across the PV membrane (PVM). Dense granule proteins that reside within the PV can leave the vacuole through T. gondii’s export machinery, which includes the Golgi-resident protein aspartyl protease, ASP5 [63,64], and the PVM-integrated translocon protein MYR1 [65]. T. gondii ASP5 is an orthologue of Plasmodium protease Plasmepsin V which recognizes a Plasmodium export element (PEXEL) motif (RxLxE/Q/D) [66] and cleaves after the leucine (RxL↓xE/Q/D) preparing PEXEL-bearing proteins for export across the PVM into the host erythrocyte [67]. Like Plasmepsin V, T. gondii ASP5 recognizes and cleaves a PEXEL-like motif (RRL↓XX) (Toxoplasma export element or ‘TEXEL’ motif) [64], and its protease function is necessary for the export of all known exported PV proteins [68]. For example, GRA16 contains an RRL↓XX sequence, is cleaved by ASP5, and utilizes the MYR1 translocon complex for protein export [64,69]. Other GRAs including GRA24 [63,65] and TEEGR/HCE1 [70,71], while being fully dependent on ASP5 and MYR1 for their export, lack a functional TEXEL sequence. As TGD057 contains an RRL sequence, we hypothesized ASP5 and/or MYR1 may be involved in the export of TGD057 from the PV, leading to MHC 1 antigen presentation and T57 antigen-specific CD8 T cell responses. To test this, the TGD057-specific CD8 T cell response to Δasp5 and Δmyr1 strains was measured as previously described in Fig 1A. In contrast to the hypothesis, ME49 Δasp5 induced a higher CD8 T cell response compared to that of wildtype (WT) ME49 (Fig 2). Moreover, T57 responses to ME49, ME49 Δmyr1 and MYR1 complementation strains (ME49 Δmyr1::MYR1) were comparable (Fig 2). Since the T57 cytokine response to the type I RH strain was uniformly low (Fig 1B and 1C), inferring requirements for export machinery using these parasite strains was uninformative. Nonetheless, CD8 T cell responses to the type II strain do not require ASP5 and MYR1, suggesting protein export from the PV is not necessary for MHC 1 antigen presentation of the vacuolar TGD057 antigen.

Instead, we hypothesized the T57 CD8 T cell response requires PV disruption and is IRG-mediated, as implicated from studies of OT1 CD8 T cell responses, or hybridoma derivatives, to parasite strains that express the model OVA antigen in the PV lumen [45,72,73]. To this end, T57 assays were performed with various BMDMs that are defective in IRG function. In the experimental setup, IFNγ is derived from activated T57 cells and is predicted to induce IRG expression in WT but not Stat1-/- or Ifngr-/- macrophages. Consistent with this supposition, the TGD057-specific CD8 T cell response to the ME49 strain was nearly abolished in the absence of IFNγ-STAT1 signaling (Fig 3A). In mice, there are 23 IRGs that can be separated into two subfamilies: 1) the effector IRGs (or ‘GKS class’ based on an amino acid motif in their GTP binding P-loop) and 2) the regulatory IRGs (‘GMS class’) [74]. Whereas effector IRGs bind the vacuolar membrane of the PV [75,76] and mediate membrane destruction via GTP hydrolysis [77], regulatory IRGs (or ‘IRGMs’) localize to host cellular organelles preventing effector IRGs from destroying host membranes [78,79]. Regulatory IRGMs bind to effector IRGs keeping them in their GDP bound inactive state [80], in a manner similar to T. gondii ROP5 [3,4,81]. In the absence of IRGMs, effector IRGs fail to localize to pathogen PVs [80,82] and pathogen restriction is lost [83]. In mice there are three regulatory IRGs (IRGM-1, -2 and -3) and Irgm1-/- or Irgm1/3-/- double knockout macrophages were analyzed. Similar to the OVA system, the TGD057-specific CD8 T cell response to stimulatory parasite strains, such as type II ME49 and the atypical strain MAS, require the activity of regulatory IRGMs (Fig 3A and 3B). In addition, IFNγ-inducible Guanylate-Binding Proteins (GBPs) localize to the PV in an IRGM-dependent manner [78] and mediate T. gondii resistance [15]. GBPs encoded on murine chromosome 3 (GBP^chr3) are involved in PV disruption, promote effector IRG recruitment to the PV of T. gondii [73,84], and once compromised will attack the parasite’s plasma membrane, decreasing its fitness [14]. To test whether GBPs are required for TGD057-specific CD8 T cell responses, GBP^chr3-deficient BMDMs lacking Gbp1, Gbp2, Gbp3, Gbp5, and Gbp7 were screened [73]. GBP^chr3 were not significantly involved in promoting TGD057-specific CD8 T cell IFNγ responses to the ME49 strain (Fig 3C), but were partially required for the response to the atypical strain MAS (Fig 3D). Altogether, these observations are consistent with a model in which the PVM is compromised by host machinery downstream of regulatory IRGs, including GBPs [78] and likely other immunity genes, that mediate vacuolar antigen escape from the PV and entry into the host’s MHC 1 antigen-presentation pathway.

Previous studies have shown that T. gondii virulence in mice is determined by parasite effectors that protect the PV from host immune attack. Specific alleles and isoforms of the rhoptry pseudokinase ROP5 [3,5,9,12,13,85], ROP18 [4,7,10,81,86] and ROP17 kinases [4,81] encoded by virulent parasite strains, are noted for their ability to inhibit the destructive functions of IRGs at the PVM, including Irgb6 and Irga6 [2–4,80,81]. These rhoptry proteins also impact the association of GBPs with the PV [84,87]. Thus, we reasoned the host CD8 T cell response to TGD057 would be antagonized by some or all of these secreted effectors. Indeed, when ROP5 is not expressed in the type I RH strain (clade A, RH Δrop5), the CD8 T cell IFNγ and IL-2 response is robust (Fig 4, S3A Fig), and on average, the IFNγ response is half of that elicited by the stimulatory type II strains (Fig 4). The ROP5 locus consists of ROP5A, ROP5B and ROP5C genes, which differ in copy number between strains, and is under diversifying selection [3,9,12,13,81,85]. ROP5B and ROP5C isoforms of virulent strains bind to and prevent accumulation of effector IRGs on the parasite’s PVM, including Irga6 and Irgb6 [3,4,81], which are required for host resistance to primary infection [88,89]. In contrast, ROP5A lacks a defined host binding partner and function, but does promote virulence by an unknown mechanism that is sensitive to copy number. For example, increasing copy number of ROP5A from one to two copies promotes virulence of RH Δrop5 complementation strains, whereas RH Δrop5+ROP5B+ROP5A phenocopies the virulence of the parental RH strain [9]. Therefore, a series of RH Δrop5 strains complemented with one or two copies of ROP5B or ROP5A isoforms from clade A, or a single ROP5A isoform from the type II genetic background (clade D) was analyzed. Importantly, all RH Δrop5+ROP5 complementation strains phenocopied the T57 response to the parental RH strain (Fig 4). Since ROP5A inhibits the T57 IFNγ response (Fig 4), we infer Irgb6 and Irga6 are likely not responsible for this phenotype because the ROP5A isoform does not inhibit Irgb6 or Irga6 coating of the PVM [4], nor Irga6 oligomerization [3]. Consistent with this supposition, all ROP5A complementation strains failed to inhibit Irgb6-PV association (Fig 4B and 4C). Of note, the sequestosome-1 (p62) associates with the PV in an IFNγ-induced IRGM-dependent manner to promote OT1 T cell responses to OVA-expressing T. gondii strains [44]. However, p62-PV association was not inhibited by ROP5A but instead phenocopied the PV-association patterns of Irgb6 (Fig 4D and 4E). Irgb6 has recently been shown to be required for p62-PV association in IFNγ-stimulated cells, which is consistent with these observations [89]. The ROP18 [4] and ROP17 kinases [6] phosphorylate and inactivate Irga6 and Irgb6. Although there were slight increases in the T57 IFNγ response to the RH Δrop17 and RH Δrop18 strains, these responses were not significantly different compared to that of the parental RH strain (Fig 4). Altogether, the data show that multiple ROP5 alleles and isoforms can suppress the T57 response to the TGD057 antigen, and this most likely by a mechanism independent of host Irgb6, Irga6 and p62 localization to the PV.

Due to the conserved nature of the TGD057 peptide epitope (S1 Fig), a unique opportunity arose to explore the development of CD8 T cell IFNγ responses to multiple parasite strains spanning the genetic diversity of T. gondii [90]. Any observed trend between T. gondii virulence, genetic background, and the T cell response may offer clues to possible parasite immune modulation and adaptation to immune pressure incurred by CD8 T cells. Among the Eurasian clonal strains (types I, II, III) and North American isolates from haplogroups (HG) XI and XII, the virulent type I strains (GT1, RH) induced the lowest CD8 T cell response while intermediate virulent types II (Pru, ME49), XI (COUGAR) and XII (B73) and low virulent type III (CEP) strains, induced relatively high CD8 T cell IFNγ (Figs 1C and 5A) and IL-2 responses (Fig 1B and S3B Fig). ‘Atypical’ strains, many of which are endemic to South America and highly virulent in laboratory mice (FOU, CAST, MAS, TgCatBr5, P89, GUY-MAT, GUY-DOS, GUY-KOE, RUB, and VAND), differed dramatically in eliciting T57 cytokine responses (Fig 5A and S3B Fig), signifying that parasite virulence is not a sole predictor of CD8 T cell activation. Instead a unique phenotypic pattern emerged, in which clade A strains (type I, HG VI and VII) [90] conferred low T57 cytokine responses while most other strains from clades B, C, D and F had potential to induce high cytokine responses (Fig 5A and S3B Fig). Consistent with previous results, T57 IFNγ responses did not correlate with known Irgb6- and/or Irga6-PV associations of these strains. For example, a low percentage (~10%) of Irgb6 recruitment to the PV is observed for the MAS strain, yet a high CD8 T cell IFNγ response is induced (Fig 5A), similar in magnitude to that of type II clade D strains (Figs 1C and 5A) whose PVs are highly decorated with Irgb6 (~45%) [81]. Even among highly virulent type I GT1 and Guyanan strains (i.e. GUY-KOE, GUY-MAT, GUY-DOS, VAND), where approximately 25% or less Irgb6-PV coating is observed [81], the CD8 T cell responses to these strains differ dramatically (Fig 5A and S3B Fig). Furthermore, one notable outlier among clade A strains is the relatively high T57 response to BOF (Fig 5B and S3B Fig). BOF encodes a single copy of ROP5B that is marginally expressed [81]. When BOF is complemented with the LC37 cosmid, which encodes the entire ROP5 locus from the clade A type I genetic background, the CD8 T cell IFNγ response is largely reduced (Fig 5B). We infer from these assays the clade A genetic background inhibits T57 IFNγ responses, but the identity of the ROP5-host interacting partner and why this genetic background leads to repressed responses is currently unknown.

Following antigen-driven TCR stimulation (or ‘signal 1’), early activated T cells receive secondary cues from the environment including co-stimulation (‘signal 2’) and cytokines (‘signal 3’) to commit to the production of cytokines such as IFNγ. Whether clade A strains, through ROP5 or other effectors, intersect one or several of these activation steps to lower T57 IFNγ responses is unclear. To explore this issue further, we generated a ‘T-GREAT’ IFNγ reporter mouse line by crossing T57 with GREAT mice [91]. GREAT mice report IFNγ transcription with an internal ribosomal entry site (IRES)-eYFP reporter cassette inserted between the stop codon and endogenous 3’UTR with the poly-A tail of the Ifng gene. Without abrogating translation of IFNγ, GREAT mice allow faithful detection of Ifng transcription via YFP fluorescence and flow cytometry [91]. In addition, surface expression of the activation marker CD69 is a proxy for early TCR signaling events, and is one of the first markers expressed by naïve T lymphocytes after activation [92]. In this way, the relative amount of TGD057 that has escaped the PV and ultimately presented by MHC 1 molecules can be inferred by T cell upregulation of CD69, and this can be measured independently of IFNγ transcription in T-GREAT cells. Naïve T-GREAT cells were co-cultured with BMDMs infected with RH (clade A), RH Δrop5, and ME49 (clade D) strains (Fig 6A), and the frequency of activated CD8 T cells (CD62L-CD69+ CD8+ T cells) (Fig 6B) as well as YFP levels (Ifng:YFP+ of CD62L-CD69+ CD8+ T cells) (Fig 6C) were measured by flow cytometry at 18 hours. Without parasite, few CD69+ or Ifng:YFP+ T-GREAT CD8 T cells were detected in the co-culture, consistent with their naïve beginnings (Fig 6B and 6C). In contrast and over a range of multiplicity of infections (MOIs), there was approximately three-fold more CD69+ CD8 T cells elicited by the ME49 compared to the RH strains (Fig 6B and 6D). Among T cells that have been activated (CD69+), a comparison of the Ifng transcript level revealed a six-fold increase in response to the ME49 compared to the RH strain (Fig 6C and 6E). Although removal of ROP5 enhances the activation and differentiation of T-GREAT cells to the RH Δrop5 strain, Ifng transcript levels never equaled that elicited by ME49, especially at lower MOIs (Fig 6E). Therefore, the data show T57 cells do in-fact recognize the RH strain, but once activated this genetic background fails to elicit other signals necessary for full induction of the T57 IFNγ response. Moreover, the data implicate T. gondii genetic determinants other than ROP5 intersect CD8 T cell IFNγ responses at the activation and likely differentiation steps.

IL-12 signaling is essential for IFNγ-mediated control of T. gondii [93–95], and is required for full induction of IFNγ-producing KLRG1+ effector CD8 T cells following T. gondii type II infections in vivo [48,49]. Moreover, the RH strain fails to induce robust IL-12 secretion in infected macrophages [96,97], perhaps underpinning the low T57 IFNγ responses to clade A strains observed in this system. To understand what extent IL-12 influences IFNγ-production, the T57 assay was performed with IL-12p40 deficient Il12b-/- BMDMs. The T57 IFNγ response to ME49- (Fig 7A), or MAS-infected Il12b-/- BMDMs (Fig 7B) was reduced but not entirely abrogated compared to that of infected WT BMDMs. A partial reduction was also reported for adoptively transferred Il-12rβ2-/- CD8 T cells that specifically lack IL-12 signaling during primary infection [48]. Next, the T57 assay was performed with parasite strains known to regulate host IL-12 production. Three T. gondii effector proteins—GRA15, GRA24 and ROP16—modulate IL-12 production in infected BMDMs [98–101]. Polymorphisms in GRA15 and ROP16 largely account for parasite strain differences in alternative (M2) and classical activation (M1) of macrophages [100]. Specifically, polymorphisms in GRA15 render type II strains able to activate the NF-ĸB pathway through direct association with TRAF2 and TRAF6 [102], and its expression is an absolute requirement for IL-12p70 [100] and largely responsible for IL-12p40 production by type II-infected BMDMs [98]. Through activation of host p38 MAPK, GRA24 promotes IL-12p40 and chemokine secretion by T. gondii-infected BMDMs [101]. Although no consistent difference between parental type II (Pru) and GRA15-deficient or GRA24-deficient strains was observed, T57 IFNγ production was decreased but not abolished in response to a double deletion Pru Δgra15 Δgra24 strain (Fig 7C). With respect to ROP16, in all T. gondii strains except those of clade D [103], the ROP16 kinase activates host STAT3, STAT5, STAT6 transcription factors [104–107], leading to the suppression of NF-ĸB signaling by an unknown mechanism [104,107]. When activating alleles of ROP16 are expressed as a transgene within the type II strain (Pru +ROP16_A), it reduces IL-12 production in T. gondii-infected BMDMs and induces the expression of many M2 associated genes [100]. The T57 IFNγ response to the Pru +ROP16_A was reduced to half that of the parental strain (Fig 7D). Thus, although IL-12 and T. gondii GRA15, GRA24, and ROP16 have some impact in regulating TGD057-specific CD8 T cell IFNγ-production, the IL-12 axis does not fully account for IFNγ commitment in this system.

Both NLRP3 and NLRP1 inflammasome activation occur following T. gondii infection [52,53,108], and IL-1 and IL-18 are known regulators of IFNγ production in a variety of cell types [109], including cells of the adaptive immune system [110]. Therefore, BMDMs deficient at various steps in the inflammasome activation cascade were analyzed. In brief, most NLRP proteins undergo ASC-driven oligomerization, causing auto-activation of caspase-1, that in turn lead to the cleavage and maturation of IL-1β and IL-18 [111]. Inflammasome activated caspase-1 and -11 also activate Gasdermin D, a key pore forming protein responsible for pyroptosis and extracellular release of IL-1/18 in several biological contexts [112–117]. The T57 IFNγ response was largely reduced to parasite-infected Nlrp1-/- BMDMs, and completely absent to infected Nlrp3-/- BMDMs (Fig 8A and 8B). In contrast, the IFNγ response was only partially decreased to Asc-/-, Casp1/11-/-, and Gsdmd-/- BMDMs infected with ME49 (Fig 8A), and no consistent difference was observed between knockout and WT BMDMs infected with MAS (Fig 8B). These results indicate CD8 T cell IFNγ differentiation, though entirely dependent on NLRs, only partially involves inflammasome matured IL-1 and/or IL-18. Consistent with this supposition, the CD8 T cell IFNγ response to parasite-infected macrophages is reduced but not abrogated in the absence of IL-18 and IL-1R-signaling, as assessed with neutralizing antibodies that block IL-1β and IL-1α engagement with its receptor, IL-1R, and Il18-/- BMDMs (Fig 8C). Low levels of IL-1β and IL-18 detected in the co-culture may underscore the limited role these cytokines play in promoting IFNγ responses in this system (S2B Fig).

Finally, to test whether NLRP3 impacts T cell activation or differentiation, T-GREAT CD8 T cell activation profiles to parasite-infected Nlrp3-/- BMDMs were explored. Whereas the percentage of early activated CD69+ T-GREAT cells in response to ME49 and MAS infections was slightly decreased to Nlrp3-/- compared to WT BMDMs (Fig 9A), IFNγ transcript levels in the activated CD69+ population dropped by 50–70% (Fig 9B), suggesting NLPR3 induces a macrophage-derived signal required for IFNγ transcription in activated CD8 T cells. Co-stimulation determines transcriptional regulation of IFNγ in tumor-infiltrating T cells [118], and in its absence, enforces post-transcriptional silencing of IFNγ in anergic self-reactive T cells [119], perhaps underscoring the blunted transcriptional (Fig 9B) and translational CD8 T cell IFNγ response to parasite-infected Nlrp3-/- BMDMs (Fig 8). To determine whether NLRP3 regulates co-stimulatory pathways, a variety of co-stimulatory ligands and the PD-L1 inhibitory receptor were measured on infected BMDMs. As previously demonstrated for B7 family members CD80, CD86 [120], PD-L1 [100], and as demonstrated here, ICOSL, are readily induced on the surface of infected BMDMs (Fig 9C). Receptors and ligands of the TNF superfamily are also expressed transcriptionally following infection in BMDMs including CD70, CD40, OX40L and 41BBL [100], of which surface expression of OX40L appears most sensitive to induction by T. gondii infection (Fig 9C and S5B Fig). However, none of the measured receptors or ligands bore evidence for NLRP3-dependent regulation (Fig 9D and S5C Fig), nor does loss of NLRP3 influence MHC 1 K^b expression (S5C Fig), as has been described for NLRC5-dependent induction of MHC class 1 genes [121]. To summarize, although inflammasome matured cytokines may play some role in promoting T57 IFNγ production, there is no substitute for NLRP3, and to a similar extent NLRP1 in this system. Importantly, our data identify a novel NLR-dependent but NLR-ASC inflammasome complex-independent pathway that regulates CD8 T cell IFNγ responses to an intracellular pathogen.

All animal protocols were approved by UC Merced’s Committee on Institutional Animal Care and Use Committee (IACUC) (AUP17-0013). All mouse work was performed in accordance with the recommendations in the Guide to the Care and Use of Laboratory Animals of the National Institutes of Health and the Animal Welfare Act (assurance number A4561-1). Inhalation of CO₂ to effect of 1.8 liters per minute was used for euthanasia of mice.

Tachyzoites of Toxoplasma gondii strains were passaged in ‘Toxo medium’ [4.5 g/liter D-glucose in DMEM with GlutaMAX (Gibco, cat#10566024), 1% heat-inactivated fetal bovine serum (FBS) (Omega Scientific, cat#FB-11, lot#441164), 1% penicillin-streptomycin (Gibco, cat#15140122)], in confluent monolayers of human foreskin fibroblasts (HFFs). HFFs were cultured in ‘HFF medium’ [4.5 g/liter D-glucose in DMEM with GlutaMAX (Gibco), 20% heat-inactivated FBS (Omega Scientific), 1% penicillin-streptomycin (Gibco), 0.2% Gentamicin (Gibco, cat#15710072), 1X L-Glutamine (Gibco, cat#21051024)]. Strains assayed include GT1 (type I, clade A), BOF (Haplogroup “HG” VI, clade A), FOU (HG VI, clade A), CAST (HG VII, clade A), MAS (HG IV, clade B), TgCatBr5 (HG VIII, clade B), CEP hxgprt- (type III, clade C), P89 (HG IX, clade C), ME49 Δhxgprt::Luc [139] (type II, clade D), Pru Δhxgprt (type II, clade D), COUGAR (HG XI, clade D), B73 (HG XII, clade D), GUY-KOE (HG V, clade F), GUY-MAT (HG V, clade F), RUB (HG V, clade F), GUY-DOS (HG X, clade F), and VAND (HG X, clade F). Other strains used include BOF+LC37 [81], RH Δhxgprt [140], RH Δhxgprt Δku80 [141], RH Δhxgprt Δku80 Δrop5::HXGPRT (RH Δrop5) [9], RH Δhxgprt Δku80 Δtgd057::HXGPRT (RH Δtgd057) [26], RH Δhxgprt Δku80 TGD057-HA::HXGPRT (RH_tgd057-HA) (generated here), Pru Δhxgprt Δku80, Pru Δhxgprt Δku80 Δtgd057::HXGPRT (Pru Δtgd057) (generated here), Pru A7 Δhxgprt::gra2-GFP::tub1-FLUC (Pru A7) [142], Pru A7 Δhxgprt Δgra15::HXGPRT (Pru Δgra15) [98], Pru A7 Δhxgprt Δgra24 (1E9) (Pru Δgra24) (generated here), Pru Δhxgprt Δgra15 Δgra24 (1F8) (Pru Δgra15 Δgra24) (generated here), Pru A7 Δhxgprt +ROP16_A::HXGPRT (Pru +ROP16_A) [100], RH Δhxgprt Δmyr1::HXGPRT (RH Δmyr1) [125], and RH Δku80 Δasp5-ty::DHFR (RH Δasp5) [69]. ME49 Δasp5::DHFR (ME49 Δasp5) was a generous gift from Dominique Soldati-Favre (University of Geneva) [63]. ME49 Δhxgprt::Luc Δmyr1::HXGPRT (ME49 Δmyr1), ME49 Δhxgprt::Luc Δmyr1 +MYR1-HA::HXGPRT (ME49 Δmyr1::MYR1) [65], and RH Δhxgprt Δrop17::HXGPRT (PCRE) (RH Δrop17) [143] were generous gifts from John Boothroyd (Stanford University). The ROP5 complementation strains used in this study are as followed: RH Δhxgprt Δku80 Δrop5 +ROP5A_II-ME49His6-3xFlag (C1A6) (RH Δrop5 +ROP5A_D) (generated here), RH Δhxgprt Δku80 Δrop5 +ROP5B_III-CTG-His6-3xFlag (H2) (RH Δrop5 +ROP5B) (generated here), RH Δhxgprt Δku80 Δrop5 +ROP5A_III-CTGHis6-3xFlag (C1B1) (RH Δrop5 + ROP5A), RH Δhxgprt Δku80 Δrop5 Δuprt::ROP5A_III-HA +ROP5B_III-CTG-His6-3xFlag (AC11) (RH Δrop5 +ROP5A+ROP5B), RH Δhxgprt Δku80 Δrop5 Δuprt::ROP5A_III-HA +ROP5A_III-CTG-His6-3xFlag (AA12) (RH Δrop5 +ROP5A+ROP5A) [9].

The Pru Δhxgprt Δku80 Δtgd057::HXGPRT strain was generated using the same primers and strategy as previously described [26]. The RH Δhxgprt Δku80 TGD057-HA::HXGPRT endotagged strain was generated as previously described [144]. In brief, for endogenous tagging [141] of TGD057 with an HA tag, the gene (TGGT1_215980) was amplified with a forward primer internal to the ATG start site of TGD057 [“AC_tgd057endoF2” 5’-CACCAACTACGTCGGAGCGCCTGTACG-3’], containing a 5′-CACC-3′ sequence required for directional TOPO cloning in pENTR/D-TOPO (Invitrogen), and a reverse primer [“Tgd057 R HA stop” 5’-TTACGCGTAGTCCGGGACGTCGTACGGGTACTCGACCTCAATGTTGTATTC-3’], containing the hemagglutinin (HA) tag sequence (underlined) followed by a stop codon. The resulting TGD057 HA-tagged DNA fragment was then cloned into the pTKO-att parasite expression vector [98] by Gateway Recombination Cloning Technology (Invitrogen). The resulting vector was linearized and transfected into RH Δhxgprt Δku80 parasites by electroporation in a 2 mm cuvette (Bio-Rad Laboratories) with 2 mM ATP (MP Biomedicals) and 5 mM glutathione (EMD) in a Gene Pulser Xcell (Bio-Rad Laboratories), with the following settings: 25 μFD, 1.25 kV, ∞ Ω. Stable integrants were selected in media with 50 μg/ml of mycophenolic acid (Axxora) and 50 μg/ml of xanthine (Alfa Aesar) and cloned by limiting dilution. The correct tagging was confirmed by PCR, using a primer upstream of the plasmid integration site and a primer specific for the HA tag (5’-CGCGTAGTCCGGGACGTCGTACGGGTA-3’), and by an immunofluorescence assay (IFA) using an HA-specific antibody (Sigma, clone 3F10).

For the generation of RH Δrop5 +ROP5A_D and RH Δrop5 +ROP5B complementation strains, RH Δhxgprt Δku80 Δrop5::HXGPRT tachyzoites were transfected by electroporation with linearized pTKO-“ROP5KO” -ROP5A_II-ME49-His6-3xFlag plasmid [9], or a similarly generated pTKO -ROP5B_III-CTG-His6-3xFlag plasmid as described in [9]. In brief, for both strains the complementation allele, ROP5-His6-3xFlag, is flanked by homology arms to the Δrop5::HXGPRT locus, whereby the transfected population is selected for removal of HXGPRT and replacement with the complementation allele in 6-thioxanthine selection medium [177 μg/mL of 6-thioxanthine (TRC, cat# T385800) in 4.5 g/liter D-glucose in GlutaMAX DMEM (Gibco), with 1% dialyzed FBS (Omega Scientific, cat#FB-03, lot#463304)]. Post-selection and limiting dilution cloning, ROP5-His6-3xFlag complementation strains were assessed by IFA with a mouse anti-Flag primary antibody (Sigma, clone M2) at 1:500 dilution and Alexa Flour 594 goat anti-mouse IgG secondary antibody (Life Technologies) at 1:3000 dilution. Both clones expressed the ROP5-HF in the expected rhoptry organelles by IFA.

For generating the Pru A7 Δhxgprt Δgra24 strain, Pru A7 Δhxgprt parasites were transfected with a NotI-linearized plasmid expressing a loxP-flanked pyrimethamine selectable cassette (loxP-DHFR-mCherry-loxP) (Addgene plasmid #70147, was a gift from David Sibley, Washington University in St. Louis) and a CRISPR-CAS9 construct targeting GRA24 (TGME49_230180). Transfectants were selected and cloned in medium containing pyrimethamine, and screened for the disruption of GRA24. For generating the Pru A7 Δhxgprt Δgra15 Δgra24 strain, first the Pru A7 Δhxgprt Δgra15::HXGPRT strain [98] was gene-edited with CRISPR-CAS9 targeting HXGPRT and selected against its expression with 6-thioxanthine. Then, a Pru A7 Δhxgprt Δgra15::hxgprt- clone was used to make a Pru A7 Δhxgprt Δgra15 Δgra24 double knockout strain using the same method as described above. Finally, the loxP-DHFR-mCherry-loxP cassette was removed from both Δgra24 and Δgra15/Δgra24 strains by transfection with a Cre recombinase parasite-expression plasmid [145], and immediately cloned by limiting dilution. The details of which can be found in a later manuscript by Mukhopadhyay et al.

For TGD057 visualization, HFFs were seeded on coverslips with HFF medium in 24-well tissue culture-treated plates. The confluent monolayer HFFs were infected with T. gondii and incubated at 37°C, 5% CO₂ overnight. For Irgb6- and p62-PV localization, BMDMs were plated on coverslips with ‘BMDM medium’ [4.5 g/liter D-glucose in DMEM with GlutaMAX (Gibco), 20% heat-inactivated FBS (Omega Scientific), 1% penicillin-streptomycin (Gibco), 1X non-essential amino acids (Gibco, cat#11140076), 1mM sodium pyruvate (Gibco, cat#11360070)] supplemented with 20% L929 conditioned medium in 24-well tissue culture-treated plates. The BMDMs were treated with 20 ng/ml of IFNγ overnight. The cells were then infected with T. gondii and incubated at 37°C, 5% CO₂ for 3–4 hours. The samples were fixed with 3% formaldehyde in phosphate buffered saline (PBS) for 20 minutes and blocked with blocking buffer (3% BSA, 5% normal goat serum or fetal bovine serum depending the species of antibody used, 0.2% Triton X-100, 0.1% sodium azide in PBS). To visualize TGD057-HA, RH or RH_tgd057-HA were stained with rat anti-HA primary antibody (Sigma, clone 3F10) at 1:500 dilution, followed by Alexa Fluor 594 goat anti-rat IgG (Life Technologies) secondary antibody (1:3000 dilution). To visualize p62, infected cells were stained with mouse anti-p62 (anti-SQSTM1) primary monoclonal antibody (Abnova, clone 2C11) at 1:50 or 1:100 dilutions, followed by Alexa Fluor 594 goat anti-mouse IgG (Life Technologies) secondary antibody at 1:3000 dilution. To visualize Irgb6, infected BMDMs were stained with TGTP goat polyclonal primary antibody (Santa Cruz Biotechnology, sc-11079) at 1:100 dilution, followed by Alexa Fluor 594 donkey anti-goat IgG (Life Technologies) at 1:3000 dilution. T. gondii PVM was stained with polyclonal rabbit anti-GRA7 primary antibody (gift from John Boothroyd, Stanford University) and Alexa Fluor 488 anti-rabbit IgG (Life Technologies) at 1:3000 dilution. Host nuclei were visualized with DAPI (Thermo Fisher, cat#62248) at 1:10,000 dilution or Hoechst (Life Technologies, cat# H3075) at 1:3000 dilution.

Six-week-old female Stat1-/- (colony 012606), Ifngr-/- (colony 003288), Nlrp1-/- (colony 021301), Nlrp3-/- (colony 021302), Casp1/11-/- (colony 016621), Il18-/- (colony 004130), and Il12b-/- (colony 002693) and wildtype C57BL/6J (B6) (colony 000664) mice were purchased from Jackson Laboratories, and all of the C57BL/6 background. C57BL/6 Asc-/- mice were generous gifts from Vishva Dixit (Genentech). Hind bones from C57BL/6 Gsdmd-/- mice [146] were generous gifts from Igor Brodsky (University of Pennsylvania). Irgm1-/- and Irgm1/m3-/- hind bones were provided from Gregory Taylor (Duke University). GBP^chr3-/- hind bone marrow cells were provided by Masahiro Yamamoto (Osaka University). Bone marrow cells were obtained and cultured in BMDM medium supplemented with 20% L929 conditioned medium. After 6–7 days of differentiation, BMDMs were harvested and were 98% pure CD11b+ CD11c- macrophages by FACS. Asc-/-, Gsdmd-/-, and Nlrp3-/- BMDMs were stained with anti-mouse MHC 1 K^b-PE labeled antibodies (BioLegend, clone AF6-88.5) and they were positive by FACS analysis.

Transnuclear T57 mice [57] were bred in-house under specific pathogen free (SPF) conditions. ‘T-GREAT’ mice were generated by back- and inter-crossing between T57 and IFNγ-stop-IRES:eYFP- endogenous poly-A tail reporter mice (GREAT mice) [91], such that breeders obtained from F3 intercrossed mice were homozygous at three alleles: T57 TCRα (TRAV6-4 TRAJ12 rearrangement), T57 TCRβ (TRBV13-1 TRBJ2-7 rearrangement), and the GREAT reporter. T-GREAT mice were then maintained in our SPF facility with no overt fitness defects observed. Genotyping primers were as followed: GREAT allele (FW 5’-CCATGGTGAGCAAGGGCGAGG-3’; RV 5’-TTACTTGTACAGCTCGTCCAT-3’); wildtype Ifng allele (FW 5’-CAGGAAGCGGAAAAGGAGTCG-3’; RV 5’-GTCACTGCAGCTCTGAATGTT-3’); T57 TCRα (TRAV6-4 TRAJ12 rearrangement: “146-alpha” FW 5’-GATAAGGGATGCTTCAATCTGATGG-3’; “108-alpha” RV 5’-CTTCCTTAGCTCACTTACCAGGGCTTAC-3’); endogenous non-rearranged TRAV6-4 and TRAJ12 loci (“191-alpha” FW 5’-GAGGCTTTACGTTAGTGATCTAAAC-3’; “108-alpha” RV); T57 TCRβ (TRBV13-1 TRBJ2-7 rearrangement: “91-beta” FW 5’- CTTGGTCGCGAGATGGGCTCCAG-3’; “103-beta” RV 5’- GTGGAAGCGAGAGATGTGAATCTTAC-3’); endogenous non-rearranged TCRBV13-1 and TRBJ2-7 loci (“142-beta” FW 5’-GCACTCGGCTCCTCGTGTTAGGTG-3’; “103-beta” RV).

2x10⁵ BMDMs cells were plated per well in a 96-well tissue culture-treated plate, in BMDM medium supplemented with 10% L929 conditioned medium. The following day, these BMDMs were infected with T. gondii tachyzoites in ‘T cell medium’ [(RPMI 1640 with GlutaMAX (Gibco, cat#61870127), 20% heat-inactivated FBS (Omega Scientific, cat#FB-11, lot#441164), 1% penicillin-streptomycin (Gibco, cat#15140122), 1 mM sodium pyruvate (Gibco, cat#11360070), 10 mM HEPES (Gibco), 1.75 μl of β-mercaptoethanol (Gibco, cat#21985023) per 500 mL RPMI 1640 with GlutaMAX]. The infections were performed in triplicates, at MOI 0.6, 0.2, and 0.07. Then, lymph nodes and spleens were obtained from either T57 or T-GREAT transnuclear mice. The lymph node cells and splenocytes were combined and red blood cells were lysed with ammonium chloride-potassium (ACK) lysis buffer. 5x10⁵ cells were added into each well of the infected BMDMs (approximately 2 hours post-infection). For IL-1R neutralization, 50 μg/mL of anti-mouse IL-1R antibody (BioXCell, clone JAMA-147) or 50 μg/mL of isotype control (BioXCell, cat#BE0091) were added when BMDMs were infected.

Confluent monolayer HFFs, seeded in 24-well plates, were infected with 100 and 300 parasites. Plaques were counted 4–6 days after infection. Displayed results are from MOIs with similar viability, the equivalent of ~MOI 0.2 was chosen for most assays.

The concentration of cytokines in the 24h and 48h supernatants from the T57 T cell activation assay was measured by ELISA according to the manufacturer’s instructions (IFNγ: Invitrogen eBioscience, cat#88731477, IL-2: Invitrogen eBioscience, cat#88702477, IL-17A: Invitrogen eBioscience, cat# 88737188, IL-1β: R&D Systems, cat#DY401-05, IL-18: Invitrogen, cat#BMS618-3). The supernatants were analyzed at various dilutions (1:2, 1:20, and 1:200) to obtain values within the linear range of the manufacture’s ELISA standards.

At 18h after T-GREAT T cell activation, samples were harvested for FACS analysis. With preparations all done on ice, cells were washed with ‘FACS buffer’ [PBS pH 7.4 (Gibco, cat#10010049), 2% heat-inactivated FBS (Omega Scientific)] and blocked with ‘blocking buffer’ [FACS buffer with 5% normal Syrian hamster serum (Jackson Immunoresearch, cat#007-000-120), 5% normal mouse serum (Jackson Immunoresearch, cat#015-000-120), and anti-mouse CD16/CD32 FcBlock (BD Biosciences, clone 2.4G2) at 1:100 dilution)]. Then, the samples were stained at 1:120 dilution with fluorophore-conjugated anti-mouse monoclonal antibodies against CD8α PE (eBioscience, clone 53–6.7), CD3ε APC-eFlour780 (eBioscience, clone 17A2), CD62L eFlour450 (eBioscience, clone MEL-14), and CD69 APC (BioLegend, clone H1.2F3). For analysis of GFP+ Pru A7-infected BMDMs, cells were harvested at 18h, washed and blocked as previously described, and stained at 1:100 dilution with PE-labeled anti-mouse antibodies against CD40 (BioLegend, clone 3/23), CD70 (eBioscience, clone FR70), CD80 (eBioscience, clone 16-10A1), CD86 (eBioscience, clone GL1), CD252 (eBioscience, OX40L clone RM134L), CD274 (eBioscience, PD-L1 clone MIH5), CD275 (eBioscience, ICOSL clone HK5.3), or rat IgG2a kappa isotype control antibodies (eBioscience, clone eBR2a). Tissue culture plates containing the infected BMDMs were placed on ice prior to harvesting and washing as described above. All samples were then stained with propidium iodide (PI) at 1:1000 dilution (Sigma, cat#P4170). Flow cytometry was performed on an LSRII (Becton Dickinson) and analyzed with FlowJo software; PI+ cells were excluded from analysis.

For all bar graphs, dots represent values obtained from an individual experiment. Results between parasite strains were often expressed relative to the response elicited by the type II strain (equal 1), or the response to infected knockout macrophages normalized to infected wildtype macrophages (equal 1). All statistical analyses (one-way or two-way ANOVA with Bonferroni’s correction, Kruskal-Wallis test with Dunn’s correction for non-parametric data in Fig 5, and unpaired two-tailed t-test) were performed with GraphPad Prism version 8.3.0.