The parasite Toxoplasma gondii is one of the most common parasites worldwide and is known for its unusual life cycle. It reproduces sexually inside its primary host—the cat—and produces eggs that are released in faeces. Other animals, most often rodents, can then become infected when they unknowingly eat the eggs while foraging. Once inside its new host, the parasite reproduces asexually until the rodent’s immune system begins to fight back. It then becomes semi-dormant and forms cysts within the brain and muscle cells of its host. In an added twist, the parasite also causes rodents to lose their fear of cats. This increases their chances of being caught and eaten, thereby helping the parasite to return to its primary host and complete its life cycle.

Previous work has shown that virulent strains of T. gondii can evade the host immune system in mice by secreting enzymes that inactivate immune-related proteins called IRG proteins. This prevents the infection being cleared and leads to death of the host within a few days. The existence of these virulent strains is intriguing because parasites that kill their host, and thus prevent their own reproduction, should be eliminated from the population. The fact that they are fairly common suggests that there must be a hitherto unknown mechanism that allows rodents to survive these virulent strains.

Lilue et al. now report the existence of such a mechanism in strains of mice found in the wild. In contrast to laboratory mice, wild mice produce IRG proteins that inhibit the enzymes secreted by the virulent strains of T. gondii. Moreover, the IRG genes in wild mice are highly variable, whereas laboratory mice all have virtually identical IRG genes.

By uncovering the complexity and variability of IRG genes in wild mice—complexity that has been lost from laboratory strains—Lilue et al. solve the conundrum of how highly virulent T. gondii strains can persist in the mouse population, and offer an explanation for the evolution of parasitic strains with differing levels of virulence.

A virulent parasite that overcomes the immune system and kills its host may seem to have won the confrontation, but it is a Pyrrhic victory when the early death of the host reduces the probability of parasite transmission. Indeed it is in the interests of all hosts and most parasites to prolong the encounter. In this sense, virulence is a failure of co-adaptation. Haldane’s conjecture (Haldane, 1949) that intense and fluctuating selection imposed by parasites will generate host protein polymorphism is widely accepted (Woolhouse et al., 2002; Clark et al., 2007; Kosiol et al., 2008; Fumagalli et al., 2011). The presence in a population of multiple host resistance alleles confronting multiple parasite virulence alleles may reflect a dynamic equilibrium permissive for the persistence of both parties. However this equilibrium is achieved only at the expense of individual interactions fatal for either the parasite or the host, as a consequence of confrontations of inappropriate alleles. In mammals, however, the ability of the adaptive immune system to respond within the time scale of an individual infection and to remember for a lifetime, buffers individuals against dangerous genetic novelty arising from parasites. As a result, life or death outcomes for common infectious diseases in mammals are not generally determined by single, highly penetrant, polymorphic genes. We here report such a case, involving infection of the house mouse, Mus musculus, with the ubiquitous intracellular protozoan parasite, Toxoplasma gondii. T. gondii has a complex life cycle (Dubey, 1998) (Figure 1). The sexual process occurs in true cats (Felidae) and intermediate hosts become infected by ingesting oocysts spread in cat faeces. A phase of fast intracellular replication and spread (tachyzoite phase) stimulates immunity in the intermediate host, and this in turn induces parasite encystment in brain and muscle cells and lifelong persistence. Predation of the infected host by a cat completes the life cycle. If immunity fails, tachyzoite replication continues uninterrupted, killing the infected host within a few days (Deckert-Schlüter et al., 1996). Thus the probability that T. gondii completes its life cycle, which is roughly linear with duration of infection of the intermediate host, depends on early immune control.

Figure 1

Download asset Open asset

Mus musculus is probably the evolutionarily most important intermediate host for T. gondii, because it is very abundant worldwide and sympatric with a uniquely abundant felid, the domestic cat. Early immune control of T. gondii in mice depends on a family of IFNγ-inducible cytoplasmic effector proteins, the 47 kDa immunity-related GTPases (IRG proteins; for nomenclature of IRG genes and proteins see Bekpen et al. (2005); Martens and Howard (2006) and ‘Materials and methods’) (Taylor et al., 2000; Collazo et al., 2001; Liesenfeld et al., 2011). These assemble on the cytosolic face of the parasitophorous vacuole membrane (PVM), causing its rupture and killing the included parasite (Martens et al., 2005; Zhao et al., 2009b). In the C57BL/6 (BL/6) laboratory mouse strain about 20 IRG genes occur in two adjacent clusters on chromosome 11 and one cluster on chromosome 18 (Bekpen et al., 2005) (Figure 2A). The whole 47 kDa sequences of IRG proteins are typically translated from a single exon. Exceptional are certain ‘tandem’ IRGB proteins with a molecular weight of about 94 kDa. These genes are transcribed across two chromosomally adjacent IRG coding units and the intergenic spacer is spliced out as an intron resulting in a single open reading frame (Bekpen et al., 2005. See also ‘Materials and methods’ for a note on nomenclature of the tandem genes and proteins). IRG proteins fall into two major functional and sequence sub-families, the GKS group (IRGA, IRGB and IRGD proteins) that are effector proteins at the PVM, and the GMS group (Irgm1, Irgm2 and Irgm3) that are negative regulators of the GKS proteins (Hunn et al., 2008). Many strains of T. gondii (e.g., the abundant Eurasian strains designated types II and III) are well controlled by the IRG system in laboratory mice, encyst, and are considered avirulent. But others (e.g., type I) are highly virulent (Sibley and Boothroyd, 1992), killing the mouse host during the tachyzoite phase of infection. It has very recently been shown that virulence differences between T. gondii strains are largely due to polymorphic variation in ROP18 and ROP5 (Saeij et al., 2006; Taylor et al., 2006; Khan et al., 2009; Behnke et al., 2011), members of a family of kinases and pseudokinases (El Hajj et al., 2006). These proteins are secreted during parasite entry and accumulate on the cytosolic face of the PVM (Boothroyd and Dubremetz, 2008). Virulent ROP allotypes inactivate IRG effector proteins by phosphorylating essential threonines in the nucleotide-binding domain (Fentress et al., 2010; Steinfeldt et al., 2010). In view of the unique importance of Mus musculus for the transmission of T. gondii, parasite strains acutely lethal for mice should be counterselected. Their presence at a significant frequency therefore demands an explanation. In this paper, we reveal that the IRG resistance system in mice has a scale of polymorphic complexity that rivals the MHC, and that a resistant IRG haplotype in the mouse can counter polymorphic virulence factors in the parasite, generating a host phenotype that is permissive for encystment, and thus for the propagation, of virulent strains.

Figure 2 with 2 supplements see all

Download asset Open asset

We compared the IRG genes from a number of mouse strains, mostly those re-sequenced in the Mouse Genomes Project (Keane et al., 2011) with the canonical BL/6 sequences (Figure 2B). The re-sequenced mice are largely from established laboratory strains. Others are recently derived from the wild and not admixed with laboratory mice. Mus spretus is a wild species distinct from Mus musculus with a divergent history of 1–3 million years (Suzuki et al., 2004). Within laboratory strains we found relatively little IRG protein polymorphism on Chr 11. In contrast, the wild-derived strains showed variation in IRG gene number and remarkable protein polymorphism. For example Irgb6, a protein of 406 residues, had an allele in the CAST/EiJ strain with 47 amino acid substitutions relative to BL/6. On Chr 18, IRG protein sequence variation between laboratory strains was more apparent, and again, wild-derived strains showed extensive sequence polymorphism and copy number variation. The IRG proteins of the outgroup, M. spretus, showed considerable protein divergence from M. musculus sequences. Pseudogenes occurred in every haplotype and some (e.g. Irga5 and Irgb7) were preserved between M. musculus and M. spretus. Other IRG sequences were pseudogenes in some haplotypes and apparently intact in others, for example Irga3 and Irga8. Even within this limited group of strains we can distinguish eleven distinct IRG gene haplotypes on Chr 11 and thirteen on Chr 18, already yielding a theoretical population diversity of several hundred IRG genotypes. Diagonal dot-plot comparisons between IRG gene clusters of wild-derived CAST/Ei and MSM/Ms with BL/6 showed IRG genes within tracts of duplication and deletion, associated with numerous repeats in both orientations, a genomic configuration that would be expected to be dynamic even on short time-scales, and doubtless responsible for the two very different dot-plots (Figure 2C and Figure 2—figure supplement 2).

We added sequences amplified from genomic DNA of wild mice from several Eurasian sites (Figure 3—figure supplement 1) to the data from inbred strains to generate nearest neighbour phylograms (Figure 3A). We analysed five IRG genes, namely Irgm1, a regulatory IRG protein of the GMS class (Hunn et al., 2008), and Irga6, Irgb2, Irgb6, and Irgb10, all effector GKS proteins localizing to the T. gondii PVM during infection (Khaminets et al., 2010). Irgb2 forms the N-terminal half of the tandem protein, Irgb2-b1. The nearest-neighbour phylograms of Irgm1, Irgb10 and Irga6 are shallow and the M. spretus sequences fall into outgroups. Thus most of the sparse polymorphic variation in these sequences has been acquired since the divergence of M. musculus from M. spretus. This conclusion is supported by a tendency for individual minor variants to be concentrated in one or other of the three recent, geographically separated and partially isolated subspecies, Mus musculus musculus, Mus musculus domesticus and Mus musculus castaneus (Figure 3A, Figure 3—figure supplements 2–4). In contrast, the phylograms for Irgb2 and Irgb6 have a depth similar to the mouse-rat divergence (about 20 million years) (Gibbs et al., 2004), the M. spretus sequences are embedded in the M. musculus trees, and any tendency to correlation with subspecies is seen only in the outermost branches. Thus the polymorphism in these two genes is ancient and has persisted through a number of speciation events. The scale of polymorphism in the IRG system estimated by Tajima’s π from 7 mouse strains resembles that of classical MHC genes (Figure 3B). That there is also considerable local polymorphism within geographic ranges was confirmed by the identification of numerous heterozygotes at all loci examined by PCR among the wild mouse captures.

Figure 3 with 4 supplements see all

Download asset Open asset

Figure 3—source data 1

Nucleotide diversities of 50 random genes in seven mouse strains.

Seven laboratory and wild-derived inbred mouse strains were analysed. The ORFs of 50 random genes were selected based on their position in the C57BL/6 genome (NCBI reference assembly build 37). In addition, selected IRG and MHC members were assembled, and Tajima's π values were calculated. Klra4 is closest to 130M in Chr 6, but lost in many mouse strains (Cutler and Kassner, 2008). The adjacent gene Klra5 was used instead.

https://doi.org/10.7554/eLife.01298.008

Download elife-01298-fig3-data1-v1.xls

Figure 3—source data 2

Alignment of Irgm1 alleles, in FASTA format.

https://doi.org/10.7554/eLife.01298.009

Download elife-01298-fig3-data2-v1.fas

Figure 3—source data 3

Alignment of Irga6 alleles, in FASTA format.

https://doi.org/10.7554/eLife.01298.010

Download elife-01298-fig3-data3-v1.fas

Figure 3—source data 4

Alignment of Irgb2 alleles, in FASTA format.

https://doi.org/10.7554/eLife.01298.011

Download elife-01298-fig3-data4-v1.fas

Figure 3—source data 5

Alignment of Irgb6 alleles, in FASTA format.

https://doi.org/10.7554/eLife.01298.012

Download elife-01298-fig3-data5-v1.fas

Figure 3—source data 6

Alignment of Irgb10 alleles, in FASTA format.

https://doi.org/10.7554/eLife.01298.013

Download elife-01298-fig3-data6-v1.fas

With such striking polymorphic diversity in the sequences of proteins known to be involved in resistance against T. gondii, it was appropriate to assay the ability of different IRG genotypes to resist virulent T. gondii. The wild-derived Indian strain CIM, which has a number of divergent IRG alleles (Figure 2B), proved to be remarkably resistant to the type I virulent T. gondii strain, GT-1 (Figure 4A). All CIM mice survived intraperitoneal infection with GT-1 tachyzoites while all laboratory mice (NMRI strain) died within 15 days. In vitro, tachyzoite proliferation of avirulent strains in mouse cells is inhibited by induction of IRG proteins with IFNγ (Könen-Waisman and Howard, 2007). In contrast, proliferation of type I virulent strains is not inhibited (Zhao et al., 2009c). By this assay, IFNγ-induced CIM diaphragm-derived cells (DDC see ‘Materials and methods’) inhibited proliferation of type I virulent RH-YFP strain tachyzoites as efficiently as they inhibited the proliferation of avirulent strains, while BL/6 DDC inhibited only the avirulent strains (Figure 4B). Cells from two other M. m. castaneus strains, CAST/EiJ and CTP were almost as resistant as CIM. Additionally, IFNγ-induced CIM cells died by reactive cell death after infection with both virulent RH-YFP and avirulent ME49, while BL/6 cells died only after infection with avirulent strains (Figure 4C). Reactive death of T. gondii-infected mouse cells after induction with IFNγ is associated with IRG protein-mediated host resistance, as previously reported (Zhao et al., 2009b).

Figure 4

Download asset Open asset

Resistance of CIM mice to the type I virulent RH-YFP strain in vivo was exploited to test linkage of resistance to the IRG system in a (CIM×BL/6)F₁×BL/6 backcross. For typing purposes the IRG gene clusters from CIM and BL/6 were differentiated by RFLPs in Irga1 (Chr 18) and Irgb6 (Chr 11). Fluorescent-tagged tachyzoites (RH-YFP) were injected intraperitoneally into mice and the frequency of infected peritoneal cells measured by FACS after 5 days. In five independent experiments involving a total of 65 typed animals (Figure 5A), resistance by this assay was almost completely dominant and was tightly linked to the IRG_CIM gene cluster on Chr 11 (p<<10⁻⁶). There was no detectable association of virulence with the IRG_CIM cluster on Chr 18. In a similar analysis of 53 F₂ progeny, mice typed as homozygous IRG_CIM at the Chr 11 gene cluster were as resistant as wild-type CIM mice. Thus the reduced resistance of some mice heterozygous for IRG_CIM on Chr 11 may be due to a gene dosage effect at the IRG locus. Assayed directly by survival, 51 backcross and F₂ mice infected with RH-YFP followed the same pattern as in the peritoneal cell assay (Figure 5B), with the Chr 11 IRG_CIM homozygotes (cc in Figure 5B) showing complete resistance, the heterozygotes (bc) substantial but incomplete resistance and the IRG_BL/6 homozygotes (bb) complete and acute susceptibility.

Figure 5

Download asset Open asset

Recent results have implicated the inflammasome core component Nlrp1 (NLR family, pyrin domain containing 1) in resistance to T. gondii in human (Witola et al., 2011) and rat (Sergent et al., 2005; Cavailles et al., 2006). Since the Nlrp1 complex locus is about 20 Mb telomeric to the IRG system on Chr 11 it was possible that polymorphism at this locus (Boyden and Dietrich, 2006) was responsible for the differential resistance apparently linked to the IRG complex. We therefore typed all the backcross progeny shown in Figure 5A by PCR designed to distinguish the BL/6 and CIM allotypes of the Nlrp1b locus (see ‘Materials and methods’). We obtained 17% recombinants between Irgb6 and Nlrp1. There was no correlation between Nlrp1 genotype and the ability to clear virulent T. gondii (Figure 5A). It is anyway unlikely that Nlrp1 plays a role in the resistance polymorphism we describe since it is not detectably expressed in the IFNγ-induced DDC transcriptomes (data not shown).

The existence of mouse genotypes resistant to virulent T. gondii strains is consistent with a co-evolutionary explanation for the evolution of virulence. To sustain the argument, however, it would be necessary to show that T. gondii strains that are lethal in laboratory mice, and thereby suffer a major cost, can form functional cysts in resistant CIM mice, permitting their propagation. We therefore searched for brain cysts in CIM mice infected 6–8 weeks earlier with virulent type I or avirulent type III strains of T. gondii. Two type I virulent strains, GT-1 and BK, formed cysts in CIM mice (Figure 5C, Table 1). Thus the resistance of the CIM mouse provides an adaptive niche for highly virulent T. gondii strains. The avirulent type III strain, NED, encysted in the laboratory mice >20× more efficiently than in CIM mice (Table 1).

The IRG resistance mechanism operates cell-autonomously, and in BL/6 cells the loading of IRG proteins onto the parasitophorous vacuole occurs only with avirulent strains even in cells infected simultaneously with virulent and avirulent strains (Zhao et al., 2009a; Khaminets et al., 2010). Consistently, the resistance of the CIM strain against virulent T. gondii was reflected in the behaviour of IRG proteins in IFNγ-induced CIM-derived cells infected with virulent RH-YFP strain T. gondii. In BL/6 cells Irgb6_BL/6 loaded onto only about 8% of vacuoles, while in CIM cells the highly divergent Irgb6_CIM protein loaded onto more than 50% of vacuoles (Figure 6A). Even more extreme, the tandem protein, Irgb2-b1_BL/6, which is poorly expressed in BL/6 cells (Figure 6B), was not detectable on the RH-YFP PVM, while the highly divergent Irgb2-b1_CIM was well-expressed in CIM cells and loaded onto over 95% of vacuoles (Figure 6A).

Figure 6

Download asset Open asset

The loading of the effector IRG protein Irga6 onto the parasitophorous vacuole of virulent T. gondii strains is prevented by a parasite-derived kinase complex (ROP5/ROP18) that phosphorylates two threonines that are essential for IRG protein function (Steinfeldt et al., 2010). Remarkably, in CIM cells infected with RH-YFP, phosphorylated Irga6_CIM was barely detectable with an antiserum specific for Irga6_BL/6 phosphorylated at T108 (Figure 6C). Irga6_CIM differs from Irga6_BL/6 at only two residues, both distant from the phosphorylation sites and was phosphorylated normally when transfected into IFNγ-induced L929 cells infected with virulent RH-YFP (Figure 6D). Thus Irga6_CIM apparently remains unphosphorylated in CIM cells as a result of active inhibition of the parasite kinase complex. The following experiment showed that the highly polymorphic tandem protein, Irgb2-b1_CIM, is largely responsible. Irgb2-b1_CIM was transfected into IFNγ-induced BL/6 mouse embryonic fibroblasts (MEFs) infected with RH-YFP virulent strain T. gondii. Phosphorylated Irga6 was measured at the PVs in transfected cells expressing Irgb2-b1_CIM and in untransfected cells. Figure 6E,F show that the amount of phosphorylated Irga6 was strikingly reduced on vacuoles loaded with Irgb2-b1_CIM while the amount of total Irga6 loaded onto Irgb2-b1_CIM positive vacuoles was increased (Figure 6G). Thus the decreased signal of phosphorylated Irga6 on the PVM was not caused by competition for loading between Irga6 and Irgb2-b1, but rather by inhibition of phosphorylation. Transfected Irgb2-b1_CIM also stimulated the loading of endogenous Irgb6_BL/6 onto RH-YFP vacuoles (Figure 6H). Transfected Irgb2-b1_BL/6, which is well expressed unlike the endogenous protein, had little or no effect on the loading of Irgb6_BL/6 (Figure 6H). Thus the essential difference between Irgb2-b1_BL/6 and Irgb2-b1_CIM in determining resistance lies in the amino acid sequence polymorphism rather than in the protein expression level.

Preliminary results suggest that Irgb2-b1_CIM may bind directly to the protein product of the virulent allele of ROP5. Thus loading of Irgb2-b1_CIM onto the PVM of ROP5-deficient RH strain parasites in CIM cells was greatly reduced, and consistently, the amount of loading on to different RH-related strains correlated with strain-specific variation in the amount of ROP5 (Figure 7A). Irgb2-b1_CIM itself becomes phosphorylated during infection with virulent T. gondii (Figure 7B), thus is also a target for the active ROP5-ROP18 kinase complex, suggesting that Irgb2-b1_CIM may block phosphorylation of Irga6 by binding ROP5 pseudokinase at the vacuole, thereby distracting rather than inhibiting ROP18 kinase. ROP5 binds to Irga6 via helix 4 (H4) of the Irga6 nucleotide binding domain (Fleckenstein et al., 2012), so a homologous structure on Irgb2-b1 may be involved in ROP5 interaction. Indeed the putative H4 and αD structural domains of the Irgb2 subunit are highly polymorphic and show recent divergent selection (Figure 7C), indicating possible co-evolution with ROP kinases and pseudokinases. The high polymorphism of H4 in Irgb2-b1 is also consistent with a direct interaction with a polymorphic component of the pathogen. If Irgb2-b1 interacts with a host protein to bridge to ROP5 the interaction surface between the two host proteins would not be expected to evolve rapidly under divergent selection.

Figure 7

Download asset Open asset

We have shown that the IRG protein system essential for resistance against T. gondii infection in the mouse has a complex polymorphism on the scale of the MHC, and that at least one IRG haplotype, found in the wild-derived CIM strain mouse, is strikingly resistant to T. gondii strains that are highly virulent for laboratory mice. We also provide a mechanistic explanation for the resistance of the CIM mouse against type I virulent T. gondii strains. Resistance is determined by the presence of the polymorphic tandem IRG protein, Irgb2-b1_CIM encoded on Chr 11, which blocks the ROP5/ROP18 kinase complex of the virulent parasite, preventing phosphorylation and consequent inactivation of IRG effector proteins. Taken together, our results suggest a selective explanation for the evolution of T. gondii strains that are highly virulent for certain mice. If the mouse is an evolutionarily significant host for T. gondii the parasite must balance its virulence against mouse resistance, in order to allow encystment. To minimise the cost of infection, selection on the mouse favours the evolution of strong resistance alleles. This in turn leads to the selection of parasite strains able to counteract heightened resistance sufficiently to allow encystment in these mouse genotypes, as type I T. gondii strains in CIM mice. However such virulent T. gondii strains are counterselected by acute lethality in less resistant mice, while less resistant mice are better hosts for less virulent T. gondii strains.

Why, however, are not all mice highly resistant? Loss of the entire IRG system in several vertebrate groups, for example higher primates and birds (Bekpen et al., 2005), suggests that possession of the IRG system may be costly. Highly resistant genotypes may be more costly than less resistant ones. Alternatively, molecular specificity in interactions between polymorphic parasite virulence factors and mouse IRG proteins may favour IRG genotypes that are highly resistant to some T. gondii genotypes but more susceptible to others. Such an allele-specific system is familiar from polymorphic plant disease resistance (R) genes and strain-specific pathogen virulence and avirulence (Flor 1971; Dodds et al., 2006), where exact molecular matches or mismatches determine the outcome of a given infection. These alternative models can now be analysed formally and tested experimentally.

Much of the argument in this paper has focused on the importance of the house mouse as an intermediate host and vector for T. gondii, resting on the global abundance of the species and its sympatry with the cat, and obviously supported by the intimate antagonistic relationship between T. gondii virulence factors and mouse resistance factors. However, although many bacteria and protozoal parasites are not resisted by the IRG system, T. gondii is not the only organism that is. At least two members of the Chlamydia species complex are resisted by the IRG system in mice, and polymorphic variation on Chr 11, possibly associated with Irgb10, affects the level of resistance (Bernstein-Hanley et al., 2006; Miyairi et al., 2007). There will surely be other organisms that may contribute to the genomic complexity and polymorphism of the IRG system. For example, the massive and ancient polymorphism of Irgb6 is not directly accounted for by our experiments. In in vitro experiments in BL/6 cells transfected Irgb2-b1_CIM protects Irga6 from phosphorylation in the absence of the CIM allotypes of the other IRG proteins including Irgb6 (Figure 6) and in the CAST/Ei strain, which is almost as resistant as CIM against T. gondii, Irgb6 is barely expressed (unpublished results). It may be that Irgb6 polymorphism reflects a pattern of resistance against another abundant mouse pathogen. However it is also not excluded that Irgb6 polymorphisms relate to further virulence polymorphism in T. gondii not expressed in the limited range of strains used in the present study. Likewise, mice are not the only intermediate hosts for T. gondii. The parasite infects all known mammals and birds, and many other species than M. musculus are prey for domestic cats. Furthermore, individuals of other species, for example rats (Jacobs and Jones, 1950) and American deer mice (Peromyscus) (Frenkel, 1953) have been shown to resist infection with strains virulent for laboratory mice. Thus some of the genomic complexity of the ROP system and other virulence-associated T. gondii secretory proteins may be relevant to immunity modification in other host species. Nevertheless, despite these distractions, mice and T. gondii have clearly had a large selective impact on each other.

The close association of cat and mouse with humans has led to large-scale transformations in the ecology of the parasite over the last 10,000 years. The impact of the dramatic narrowing of the parasite’s effective host range on the genetics of the virulence-resistance relationship may help to understand better the co-evolutionary forces at work in this system. The patterns of genetic resistance against T. gondii in its most important intermediate hosts will determine the relative abundance of strains contaminating the environment, and thereby the strains available for human infection.

1. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus clone MSMg01-329H21

   ID KF705682. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705682
2. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus clone MSMg01-362F04

   ID KF705684. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705684
3. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus clone MSMg01-544P17

   ID KF705686. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705686
4. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus clone MSMg01-494M12

   ID KF705680. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705680
5. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus clone MSMg01-419B05

   ID KF705685. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705685
6. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus strain MSM/Ms clone MSMg01-148E20

   ID KF705681. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705681
7. 1. Lilue J
   2. Müller UB
   3. Steinfeldt T
   4. Howard JC

   (2013) Mus musculus molossinus clone MSMg01-355D01

   ID KF705683. Publicly available at GenBank (http://www.ncbi.nlm.nih.gov/genbank/).

   http://www.ncbi.nlm.nih.gov/nuccore/?term=KF705683

1. 1. Yalcin B
   2. Adams DJ
   3. Flint J
   4. Keane TM

   (2012) Mouse Genomes Project

   These data are released in accordance with the Fort Lauderdale agreement and Toronto agreements.

   http://www.sanger.ac.uk/resources/mouse/genomes/
2. 1. Abe K
   2. Noguchi H
   3. Tagawa K
   4. Yuzuriha M
   5. Toyoda A
   6. Kojima T
   7. Ezawa K
   8. Saitou N
   9. Hattori M
   10. Sakaki Y
   11. Moriwaki K
   12. Shiroishi T

   (2004) NIG Mouse Genome Database

   Available at NIG Mammalian Genetics Laboratory, Japan.

   http://molossinus.lab.nig.ac.jp/msmdb/
3. 1. Holt K

   (2008) Mus musculus castaneus strain CAST/Ei clone CH26-332A4

   ID CU695228. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695228
4. 1. Holt K

   (2008) Mus musculus castaneus strain CAST/Ei clone CH26-226N16

   ID CU695224. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695224
5. 1. Holt K

   (2008) Mouse DNA sequence from clone CH26-243M20 on chromosome 11, complete sequence

   ID CU695226. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695226
6. 1. Holt K

   (2008) Mouse DNA sequence from clone CH26-455J9 on chromosome 11, complete sequence

   ID CU695230. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695230
7. 1. Holt K

   (2008) Mouse DNA sequence from clone CH26-333C17 on chromosome 11, complete sequence

   ID CU695229. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695229
8. 1. Holt K

   (2008) Mus musculus castaneus strain CAST/Ei clone CH26-240F21

   ID CU695225. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695225
9. 1. Holt K

   (2008) Mouse DNA sequence from clone CH26-316B17 on chromosome 18, complete sequence

   ID CU695227. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

   http://www.ncbi.nlm.nih.gov/nuccore/CU695227
10. 1. Holt K

    (2008) Mouse DNA sequence from clone CH26-76B7 on chromosome 18, complete sequence

    ID CU695231. Publicly available at the NCBI Nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore).

    http://www.ncbi.nlm.nih.gov/nuccore/CU695231

1. 1. Abe K
   2. Noguchi H
   3. Tagawa K
   4. Yuzuriha M
   5. Toyoda A
   6. Kojima T

   et al. (2004) Contribution of Asian mouse subspecies Mus musculus molossinus to genomic constitution of strain C57BL/6J, as defined by BAC-end sequence-SKIP analysis

   Genome Res 14:2439–2447.

   https://doi.org/10.1101/Gr.2899304

   • Google Scholar
2. 1. Albert B
   2. Sabin MD

   (1941) Toxoplasmic encephalitis in children

   JAMA 116:.

   https://doi.org/10.1001/jama.1941.02820090001001

   • Google Scholar
3. 1. Antony VB
   2. Owen CL
   3. Hadley KJ

   (1989) Pleural mesothelial cells stimulated by asbestos release chemotactic activity for neutrophils in vitro

   Am Rev Respir Dis 139:199–206.

   https://doi.org/10.1164/ajrccm/139.1.199

   • Google Scholar
4. 1. Behnke MS
   2. Khan A
   3. Wootton JC
   4. Dubey JP
   5. Tang KL
   6. Sibley LD

   (2011) Virulence differences in Toxoplasma mediated by amplification of a family of polymorphic pseudokinases

   Proc Natl Acad Sci USA 108:9631–9636.

   https://doi.org/10.1073/pnas.1015338108

   • Google Scholar
5. 1. Bekpen C
   2. Hunn JP
   3. Rohde C
   4. Parvanova I
   5. Guethlein L
   6. Dunn DM

   et al. (2005) The interferon-inducible p47 (IRG) GTPases in vertebrates: loss of the cell autonomous resistance mechanism in the human lineage

   Genome Biol 6:R92.

   https://doi.org/10.1186/Gb-2005-6-11-R92

   • Google Scholar
6. 1. Bernstein-Hanley I
   2. Coers J
   3. Balsara ZR
   4. Taylor GA
   5. Starnbach MN
   6. Dietrich WF

   (2006) The p47 GTPases Igtp and Irgb10 map to the Chlamydia trachomatis susceptibility locus Ctrq-3 and mediate cellular resistance in mice

   Proc Natl Acad Sci USA 103:14092–14097.

   https://doi.org/10.1073/pnas.0603338103

   • Google Scholar
7. 1. Boothroyd JC
   2. Dubremetz JF

   (2008) Kiss and spit: the dual roles of Toxoplasma rhoptries

   Nat Rev Microbiol 6:79–88.

   https://doi.org/10.1038/Nrmicro1800

   • Google Scholar
8. 1. Boyden ED
   2. Dietrich WF

   (2006) Nalp1b controls mouse macrophage susceptibility to anthrax lethal toxin

   Nat Genet 38:240–244.

   https://doi.org/10.1038/ng1724

   • Google Scholar
9. 1. Cavailles P
   2. Sergent V
   3. Bisanz C
   4. Papapietro O
   5. Colacios C
   6. Mas M

   et al. (2006) The rat Toxo1 locus directs toxoplasmosis outcome and controls parasite proliferation and spreading by macrophage-dependent mechanisms

   Proc Natl Acad Sci USA 103:744–749.

   https://doi.org/10.1073/pnas.0506643103

   • Google Scholar
10. 1. Clark RM
    2. Schweikert G
    3. Toomajian C
    4. Ossowski S
    5. Zeller G
    6. Shinn P

    et al. (2007) Common sequence polymorphisms shaping genetic diversity in Arabidopsis thaliana

    Science 317:338–342.

    https://doi.org/10.1126/science.1138632

    • Google Scholar
11. 1. Collazo CM
    2. Yap GS
    3. Sempowski GD
    4. Lusby KC
    5. Tessarollo L
    6. GFV Woude

    et al. (2001) Inactivation of LRG-47 and IRG-47 reveals a family of interferon gamma-inducible genes with essential, pathogen-specific roles in resistance to infection

    J Exp Med 194:181–187.

    https://doi.org/10.1084/jem.194.2.181

    • Google Scholar
12. 1. Cutler G
    2. Kassner PD

    (2008) Copy number variation in the mouse genome: implications for the mouse as a model organism for human disease

    Cytogenet Genome Res 123:297–306.

    https://doi.org/10.1159/000184721

    • Google Scholar
13. 1. Darde ML
    2. Bouteille B
    3. Pestre-Alexandre M

    (1992) Isoenzyme analysis of 35 Toxoplasma gondii isolates and the biological and epidemiological implications

    J Parasitol 78:786–794.

    https://doi.org/10.2307/3283305

    • Google Scholar
14. 1. Deckert-Schlüter M
    2. Rang A
    3. Weiner D
    4. Huang S
    5. Wiestler OD
    6. Hof H

    et al. (1996)

    Interferon-gamma receptor-deficiency renders mice highly susceptible to toxoplasmosis by decreased macrophage activation

    Lab Invest 75:827–841.

    • Google Scholar
15. 1. Din W
    2. Anand R
    3. Boursot P
    4. Darviche D
    5. Dod B
    6. Jouvin-Marche E

    et al. (1996) Origin and radiation of the house mouse: clues from nuclear genes

    J Evol Biol 9:519–539.

    https://doi.org/10.1046/j.1420-9101.1996.9050519.x

    • Google Scholar
16. 1. Dodds PN
    2. Lawrence GJ
    3. Catanzariti AM
    4. Teh T
    5. Wang CI
    6. Ayliffe MA

    et al. (2006) Direct protein interaction underlies gene-for-gene specificity and coevolution of the flax resistance genes and flax rust avirulence genes

    Proc Natl Acad Sci USA 103:8888–8893.

    https://doi.org/10.1073/pnas.0602577103

    • Google Scholar
17. 1. Dubey JP

    (1980)

    Mouse pathogenicity of Toxoplasma gondii isolated from a goat

    Am J Vet Res 41:427–429.

    • Google Scholar
18. 1. Dubey JP

    (1998) Advances in the life cycle of Toxoplasma gondii

    Int J Parasitol 28:1019–1024.

    https://doi.org/10.1016/S0020-7519(98)00023-X

    • Google Scholar
19. 1. El Hajj H
    2. Demey E
    3. Poncet J
    4. Lebrun M
    5. Wu B
    6. Galeotti N

    et al. (2006) The ROP2 family of Toxoplasma gondii rhoptry proteins: proteomic and genomic characterization and molecular modeling

    Proteomics 6:5773–5784.

    https://doi.org/10.1002/pmic.200600187

    • Google Scholar
20. 1. Fentress SJ
    2. Behnke MS
    3. Dunay IR
    4. Mashayekhi M
    5. Rommereim LM
    6. Fox BA

    et al. (2010) Phosphorylation of immunity-related GTPases by a Toxoplasma gondii-secreted kinase promotes macrophage survival and virulence

    Cell Host Microbe 8:484–495.

    https://doi.org/10.1016/j.chom.2010.11.005

    • Google Scholar
21. 1. Fleckenstein MC
    2. Reese ML
    3. Konen-Waisman S
    4. Boothroyd JC
    5. Howard JC
    6. Steinfeldt T

    (2012) A Toxoplasma gondii pseudokinase inhibits host IRG resistance proteins

    PLOS Biol 10:e1001358.

    https://doi.org/10.1371/journal.pbio.1001358

    • Google Scholar
22. 1. Flor HH

    (1971) Current status of the gene-for-gene concept

    Annu Rev Phytopathol 9:275–296.

    https://doi.org/10.1146/annurev.py.09.090171.001423

    • Google Scholar
23. 1. Frenkel JK

    (1953)

    Host, strain and treatment variation as factors in the pathogenesis of toxoplasmosis

    Am J Trop Med Hyg 2:390–414.

    • Google Scholar
24. 1. Fumagalli M
    2. Sironi M
    3. Pozzoli U
    4. Ferrer-Admettla A
    5. Pattini L
    6. Nielsen R

    (2011) Signatures of environmental genetic adaptation pinpoint pathogens as the main selective pressure through human evolution

    PLOS Genet 7:.

    https://doi.org/10.1371/journal.pgen.1002355

    • Google Scholar
25. 1. Gibbs RA
    2. Weinstock GM
    3. Metzker ML
    4. Muzny DM
    5. Sodergren EJ
    6. Scherer S

    et al. (2004) Genome sequence of the Brown Norway rat yields insights into mammalian evolution

    Nature 428:493–521.

    https://doi.org/10.1038/Nature02426

    • Google Scholar
26. 1. Gubbels MJ
    2. Li C
    3. Striepen B

    (2003) High-throughput growth assay for Toxoplasma gondii using yellow fluorescent protein

    Antimicrob Agents Chemother 47:309–316.

    https://doi.org/10.1128/AAC.47.1.309-316.2003

    • Google Scholar
27. 1. Guenet JL
    2. Bonhomme F

    (2003) Wild mice: an ever-increasing contribution to a popular mammalian model

    Trends Genet 19:24–31.

    https://doi.org/10.1016/S0168-9525(02)00007-0

    • Google Scholar
28. 1. Haldane JBS

    (1949)

    Disease and evolution

    Ric Sci Suppl A 19:68–76.

    • Google Scholar
29. 1. Hunn JP
    2. Koenen-Waisman S
    3. Papic N
    4. Schroeder N
    5. Pawlowski N
    6. Lange R

    et al. (2008) Regulatory interactions between IRG resistance GTPases in the cellular response to Toxoplasma gondii

    EMBO J 27:2495–2509.

    https://doi.org/10.1038/emboj.2008.176

    • Google Scholar
30. 1. Jacobs L
    2. Jones FE

    (1950) The parasitemia in experimental toxoplasmosis

    J Infect Dis 87:78–89.

    https://doi.org/10.1093/infdis/87.1.78

    • Google Scholar
31. 1. Keane TM
    2. Goodstadt L
    3. Danecek P
    4. White MA
    5. Wong K
    6. Yalcin B

    et al. (2011) Mouse genomic variation and its effect on phenotypes and gene regulation

    Nature 477:289–294.

    https://doi.org/10.1038/Nature10413

    • Google Scholar
32. 1. Khaminets A
    2. Hunn JP
    3. Konen-Waisman S
    4. Zhao YO
    5. Preukschat D
    6. Coers J

    et al. (2010) Coordinated loading of IRG resistance GTPases on to the Toxoplasma gondii parasitophorous vacuole

    Cell Microbiol 12:939–961.

    https://doi.org/10.1111/j.1462-5822.2010.01443.x

    • Google Scholar
33. 1. Khan A
    2. Taylor S
    3. Ajioka JW
    4. Rosenthal BM
    5. Sibley LD

    (2009) Selection at a single locus leads to widespread expansion of Toxoplasma gondii lineages that are virulent in mice

    PLOS Genet 5:e1000404.

    https://doi.org/10.1371/journal.pgen.1000404

    • Google Scholar
34. 1. Könen-Waisman S
    2. Howard JC

    (2007) Cell-autonomous immunity to Toxoplasma gondii in mouse and man

    Microbes Infect 9:1652–1661.

    https://doi.org/10.1016/j.micinf.2007.09.005

    • Google Scholar
35. 1. Kosiol C
    2. Vinar T
    3. da Fonseca RR
    4. Hubisz MJ
    5. Bustamante CD
    6. Nielsen R

    et al. (2008) Patterns of positive selection in six mammalian genomes

    PLOS Genet 4:e1000144.

    https://doi.org/10.1371/journal.pgen.1000144

    • Google Scholar
36. 1. Librado P
    2. Rozas J

    (2009) DnaSP v5: a software for comprehensive analysis of DNA polymorphism data

    Bioinformatics 25:1451–1452.

    https://doi.org/10.1093/bioinformatics/btp187

    • Google Scholar
37. 1. Liesenfeld O
    2. Parvanova I
    3. Zerrahn J
    4. Han SJ
    5. Heinrich F
    6. Munoz M

    et al. (2011) The IFN-gamma-inducible GTPase, Irga6, protects mice against Toxoplasma gondii but not against Plasmodium berghei and some other intracellular pathogens

    PLOS ONE 6:e20568.

    https://doi.org/10.1371/journal.pone.0020568

    • Google Scholar
38. 1. Lunde MN
    2. Jacobs L

    (1983) Antigenic differences between endozoites and cystozoites of Toxoplasma gondii

    J Parasitol 69:806–808.

    https://doi.org/10.2307/3281034

    • Google Scholar
39. 1. Martens S
    2. Howard JC

    (2006) The interferon-inducible GTPases

    Annu Rev Cell Dev Biol 22:559–589.

    https://doi.org/10.1146/annurev.cellbio.22.010305.104619

    • Google Scholar
40. 1. Martens S
    2. Parvanova I
    3. Zerrahn J
    4. Griffiths G
    5. Schell G
    6. Reichmann G

    et al. (2005) Disruption of Toxoplasma gondii parasitophorous vacuoles by the mouse p47-resistance GTPases

    PLOS Pathog 1:e24.

    https://doi.org/10.1371/journal.ppat.0010024

    • Google Scholar
41. 1. Martens S
    2. Sabel K
    3. Lange R
    4. Uthaiah R
    5. Wolf E
    6. Howard JC

    (2004)

    Mechanisms regulating the positioning of mouse p47 resistance GTPases LRG-47 and IIGP1 on cellular membranes: retargeting to plasma membrane induced by phagocytosis

    J Immunol 173:2594–2606.

    • Google Scholar
42. 1. Miyairi I
    2. Tatireddigari VRRA
    3. Mahdi OS
    4. Rose LA
    5. Belland RJ
    6. Lu L

    et al. (2007)

    The p47 GTPases Iigp2 and Irgb10 regulate innate immunity and inflammation to murine Chlamydia psittaci infection

    J Immunol 179:1814–1824.

    • Google Scholar
43. 1. Pfefferkorn ER
    2. Guyre PM

    (1984)

    Inhibition of growth of Toxoplasma gondii in cultured fibroblasts by human recombinant gamma interferon

    Infect Immun 44:211–216.

    • Google Scholar
44. 1. Reese ML
    2. Zeiner GM
    3. Saeij JP
    4. Boothroyd JC
    5. Boyle JP

    (2011) Polymorphic family of injected pseudokinases is paramount in Toxoplasma virulence

    Proc Natl Acad Sci USA 108:9625–9630.

    https://doi.org/10.1073/pnas.1015980108

    • Google Scholar
45. 1. Reischl U
    2. Bretagne S
    3. Kruger D
    4. Ernault P
    5. Costa JM

    (2003) Comparison of two DNA targets for the diagnosis of Toxoplasmosis by real-time PCR using fluorescence resonance energy transfer hybridization probes

    BMC Infect Dis 3:7.

    https://doi.org/10.1186/1471-2334-3-7

    • Google Scholar
46. 1. Saeij JP
    2. Boyle JP
    3. Coller S
    4. Taylor S
    5. Sibley LD
    6. Brooke-Powell ET

    et al. (2006) Polymorphic secreted kinases are key virulence factors in toxoplasmosis

    Science 314:1780–1783.

    https://doi.org/10.1126/science.1133690

    • Google Scholar
47. 1. Sergent V
    2. Cautain B
    3. Khalife J
    4. Deslee D
    5. Bastien P
    6. Dao A

    et al. (2005) Innate refractoriness of the Lewis rat to toxoplasmosis is a dominant trait that is intrinsic to bone marrow-derived cells

    Infect Immun 73:6990–6997.

    https://doi.org/10.1128/IAI.73.10.6990-6997.2005

    • Google Scholar
48. 1. Sibley LD
    2. Boothroyd JC

    (1992) Virulent-strains of Toxoplasma gondii comprise a single clonal lineage

    Nature 359:82–85.

    https://doi.org/10.1038/359082a0

    • Google Scholar
49. 1. Southern PJ
    2. Berg P

    (1982)

    Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter

    J Mol Appl Genet 1:327–341.

    • Google Scholar
50. 1. Steinfeldt T
    2. Konen-Waisman S
    3. Tong L
    4. Pawlowski N
    5. Lamkemeyer T
    6. Sibley LD

    et al. (2010) Phosphorylation of mouse immunity-related GTPase (IRG) resistance proteins is an evasion strategy for virulent Toxoplasma gondii

    PLOS Biol 8:e1000576.

    https://doi.org/10.1371/journal.pbio.1000576

    • Google Scholar
51. 1. Suzuki H
    2. Shimada T
    3. Terashima M
    4. Tsuchiya K
    5. Aplin K

    (2004) Temporal, spatial, and ecological modes of evolution of Eurasian Mus based on mitochondrial and nuclear gene sequences

    Mol Phylogenet Evol 33:626–646.

    https://doi.org/10.1016/j.ympev.2004.08.003

    • Google Scholar
52. 1. Tamura K
    2. Peterson D
    3. Peterson N
    4. Stecher G
    5. Nei M
    6. Kumar S

    (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods

    Mol Biol Evol 28:2731–2739.

    https://doi.org/10.1093/molbev/msr121

    • Google Scholar
53. 1. Taylor S
    2. Barragan A
    3. Su C
    4. Fux B
    5. Fentress SJ
    6. Tang K

    et al. (2006) A secreted serine-threonine kinase determines virulence in the eukaryotic pathogen Toxoplasma gondii

    Science 314:1776–1780.

    https://doi.org/10.1126/science.1133643

    • Google Scholar
54. 1. Taylor GA
    2. Collazo CM
    3. Yap GS
    4. Nguyen K
    5. Gregorio TA
    6. Taylor LS

    et al. (2000) Pathogen-specific loss of host resistance in mice lacking the IFN-gamma-inducible gene IGTP

    Proc Natl Acad Sci USA 97:751–755.

    https://doi.org/10.1073/pnas.97.2.751

    • Google Scholar
55. 1. Winsser J
    2. Verlinde JD
    3. Vanthiel PH
    4. Davel J
    5. Vanderelst P

    (1948) Isolation of Toxoplasma from cerebrospinal fluid of a living infant in Holland

    Proc Soc Exp Biol Med 67:292–294.

    https://doi.org/10.3181/00379727-67-16279

    • Google Scholar
56. 1. Witola WH
    2. Mui E
    3. Hargrave A
    4. Liu S
    5. Hypolite M
    6. Montpetit A

    et al. (2011) NALP1 influences susceptibility to human congenital toxoplasmosis, proinflammatory cytokine response, and fate of Toxoplasma gondii-infected monocytic cells

    Infect Immun 79:756–766.

    https://doi.org/10.1128/IAI.00898-10

    • Google Scholar
57. 1. Woolhouse MEJ
    2. Webster JP
    3. Domingo E
    4. Charlesworth B
    5. Levin BR

    (2002) Biological and biomedical implications of the co-evolution of pathogens and their hosts

    Nat Genet 32:569–577.

    https://doi.org/10.1038/Ng1202-569

    • Google Scholar
58. 1. Zhao Y
    2. Ferguson DJ
    3. Wilson DC
    4. Howard JC
    5. Sibley LD
    6. Yap GS

    (2009a) Virulent Toxoplasma gondii evade immunity-related GTPase-mediated parasite vacuole disruption within primed macrophages

    J Immunol 182:3775–3781.

    https://doi.org/10.4049/jimmunol.0804190

    • Google Scholar
59. 1. Zhao YO
    2. Khaminets A
    3. Hunn JP
    4. Howard JC

    (2009b) Disruption of the Toxoplasma gondii parasitophorous vacuole by IFN gamma-inducible immunity-related GTPases (IRG proteins) triggers necrotic cell death

    PLOS Pathog 5:e1000288.

    https://doi.org/10.1371/journal.ppat.1000288

    • Google Scholar
60. 1. Zhao YO
    2. Rohde C
    3. Lilue JT
    4. Konen-Waisman S
    5. Khaminets A
    6. Hunn JP

    et al. (2009c) Toxoplasma gondii and the immunity-related GTPase (IRG) resistance system in mice: a review

    Mem Inst Oswaldo Cruz 104:234–240.

    https://doi.org/10.1590/S0074-02762009000200016

    • Google Scholar

Thank you for sending your work entitled “Toxoplasma and the mouse: reciprocal virulence and resistance polymorphism” for consideration at eLife. Your article has been favorably peer reviewed by Detlef Weigel, eLife Deputy editor, and two other reviewers.

The Deputy editor and the other reviewers discussed their comments before we reached this decision, and the Deputy editor has assembled the following comments to help you prepare a revised submission.

This manuscript reports on the role of different mouse IRG genotypes in host defense against Toxoplasma gondii. The complexity and level of polymorphism in the IRG system are comparable to that at the MHC, and in a wider collection of mouse strains IRG genotypes can be found that confer resistance to strains of T. gondii that are highly virulent in laboratory mouse strains. In addition, the authors provide evidence that specific IRGs from resistant mouse strains function by dysregulating the activity of parasite virulence factors. The experiments are well designed and conducted, and the results are properly interpreted. The authors go on to speculate that the IRG polymorphism in mice (and loss of the entire IRG system in several vertebrate groups) reflects a fitness trade-off. In addition, the authors acknowledge that the mouse–T. gondii interaction is probably only a part of the evolutionary picture, and that IRG evolution is likely to be driven by other pathogens as well, such as Chlamydia.

The reviewers, however, felt that some of the evolutionary inferences are overstated. For example, the abstract begins with “Virulence in T. gondii for its natural intermediate hosts, the mouse, is an evolutionary paradox…” This would only be the case if one takes laboratory mice, known to be highly derived, as representative of wild mice. In fact, prior studies have shown that deer mice (Peromyscus) are resistant to virulent strains of T. gondii (Am. J. Trop. Med. Hyg. 2, 390-415 (1953), as is the case for other common hosts such as chickens and rats. It is therefore reasonable to assume that wild house mice might express resistance mechanisms not found in laboratory mice, given the extremely narrow genetic diversity of the latter. The authors find such a mechanism and this is certainly interesting, but perhaps not as surprising as claimed. Hence, one could conclude that the virulence mechanisms of the parasite have evolved for “resistant hosts” in general and not wild mice per se. Although the system seems to resemble the R gene-avirulence gene systems for plants and their pathogens, the co-evolution arguments should be toned down considerably, as the work stands on its own.

Major comments:

1) There is a concern about the mechanism proposed for Irgb2-b1 function. In cells from CIM mice infected with RH parasites, the authors demonstrate a substantial increase in Irgb6 and Irgb2-b1 loading onto the parasite vacuole, compared to cells from B6 mice. In addition, although there is undetectable p-Irga6_CIM in CIM cells infected with RH parasites, Irga6_CIM can be phosphorylated if transfected into other cells (L929 cells), suggesting that Irga6 phosphorylation is actively inhibited in CIM cells. The authors state that Irgb2-b1_CIM inhibits the phosphorylation of endogenous Irga6 in CIM mice. This statement is based on experiments in which they transfect Irgb2-b1_CIM into B6 cells stimulated with IFN-γ and observe a reduction in p-Irga6 localization to the parasite vacuole membrane, which they quantify. However, they do not directly show that Irgb2-b1_CIM inhibits Irga6 phosphorylation. It seems that an alternate interpretation of these data (from Figure 6E,F) is that Irgb2-b1_CIM may inhibit the localization of Irga6 to the vacuole, where it is then phosphorylated. This point should be addressed/clarified.

Overall, the data implicating ROP5 in the loading of Irgb2-b1_CIM onto the vacuole surrounding virulent parasites is very indirect. Granted it is decreased in Δrop5 parasites, but this does not equate to a model where Irgb2-b1 binds to ROP5 or inhibits its activity. Other indirect mechanisms are possible. This portion of the manuscript is too speculative. Either concrete data need to be provided or this should be extensively modified to indicate it might result from direct or indirect binding. Does Irgb2-b1 have a direct effect on ROP18? It would look the same phenotypically (Δrop5 parasites would phenocopy this). The model promoted here suggests that ROP5 binds to IRGs. If this interaction is robust, it should be straightforward to demonstrate this for Irgb2-b1. If such data are elusive, the authors should probably rethink their model. Figure 7 is somewhat speculative, and is also not well explained. For example, what is the significance of the IIa/IIs ratios in the graphic?

In the last paragraph of the Results when the authors describe Figure 7, they state, “Preliminary results suggest that Irgb2-b1_CIM may bind directly to the protein product of the virulent allele of ROP5.” What are these preliminary results? Including this interaction data would significantly strengthen the authors’ model that Irgb2-b1_CIM binding to ROP5 may contribute to blocking Irga6 phosphorylation in CIM mice.

2) The authors rely at least in part on Illumina resequencing data for their phylogenetic studies; such data are notoriously problematic for highly polymorphic and complex, duplicated regions of the genome. It would be appropriate for the authors to discuss how the BAC sequences they generated (which are only mentioned in the Materials and methods) as well as their extensive PCR data compared with the Illumina resequencing data.

3) It might be worth pointing out that type I strains are relatively rare, yet they share the “acute lab mouse” phenotype with many South American strains that not coincidentally share a ROP18 allele (and also ROP5 although this is less well characterized). If it were only clonal type I strains at play, the evolutionary argument that natural hosts must somehow be resistant to such strains would be much less meaningful (type 1 strains are less than 5% of all strains encountered in the wild).

4) The extreme polymorphism of the IRG family of proteins suggests that other pathogens might also impinge on this pathway. This is only briefly alluded to in the Discussion, but in fact it seems relevant from the opening and should be mentioned in the Introduction already. The present work reveals the role of one component of the IRG family in combating Tg. And yet, the role of many other components remains undefined. Again, how important co-evolution with T. gondii is for IRG diversification is still unknown. Along these lines, the observation that the virulent strain GT1 can form cysts in resistant CIM mice is interesting, but not a very strong argument for co-evolution. Virulent strains of T. gondii can form cysts in any resistant species. In this regard, there does not seem anything particularly different about the wild mouse. It is the lab mouse that is the exception. It remains an interesting historical artefact that the extreme susceptibility of the lab mouse led to ROP kinase discovery. Along these lines, little is said about the coincidence between different IRG genotypes and T. gondii strains in the wild, which is obviously important to determine the significance of the laboratory observations.

5) The life cycle cartoon contains several errors. Oocysts contain two sporozoites each with four sporozoites, not two as shown. Tachyzoites are shown outside cells but should be replicating within them. Asexual stages also occur in the cat.

The reviewers […] felt that some of the evolutionary inferences are overstated. For example, the abstract begins with “Virulence in T. gondii for its natural intermediate hosts, the mouse, is an evolutionary paradox…”

Well, paradoxes are there to be resolved. While there is a literature with which we are familiar, and now cited, showing that individuals of other species (including man, of course) are resistant to many virulent T. gondii strains, the belief is general and has been reiterated for years, that mice are peculiarly vulnerable to the parasite. This generalisation has however been based on a limited database: until our own study, which is now pretty well known in the community, everybody has used classical laboratory mouse strains to examine this question. Yes, it's pretty obvious, but as said, (A) it can't be found in the published literature and (B) paradoxes are there to be resolved: this paper presents experimental data that resolve the paradox. Incidentally it is far from obvious that the laboratory strains of mouse should be so homogeneous in the IRG gene clusters. The MHCs of the same strains are strikingly polymorphic. Their shared vulnerability to type I Toxoplasma strains is the case simply because they have homogenised specific haplotypes of the IRG system. One may wonder whether this is due to an undiscovered fitness cost of certain IRG haplotypes rather than to a limited genetic basis among the lab strains.

This would only be the case if one takes laboratory mice, known to be highly derived, as representative of wild mice. In fact, prior studies have shown that deer mice (Peromyscus) are resistant to virulent strains of T. gondii (Am. J. Trop. Med. Hyg. 2, 390-415 (1953), as is the case for other common hosts such as chickens and rats. It is therefore reasonable to assume that wild house mice might express resistance mechanisms not found in laboratory mice, given the extremely narrow genetic diversity of the latter. The authors find such a mechanism and this is certainly interesting, but perhaps not as surprising as claimed.

We do not say anywhere that it is surprising: indeed it was the original apparently general “susceptibility” that was surprising, that was the “paradox”.

Hence, one could conclude that the virulence mechanisms of the parasite have evolved for “resistant hosts” in general and not wild mice per se.

Not sure we understand this statement. Considering what virulence and resistance consist of, i.e. specific interactions between polymorphic molecules of host and pathogen, we do not really grasp what is meant by “‘resistant hosts’ in general”. As we show in Figure 2, there are multiple alleles for relevant resistance molecules: we have explored the resistance properties of a minute sample of the available allelic variation against a minute sample of virulence alleles. It is plausible that further resistance variants may confer protection against further virulence variants.

Although the system seems to resemble the R gene-avirulence gene systems for plants and their pathogens, the co-evolution arguments should be toned down considerably, as the work stands on its own.

We have lowered the prominence of our remarks on plant R-genes. Still, it is a conjecture based on properties the two systems have in common, but not implausible and good to think about, we find. To instance some of the similarities: both are encoded in short, uninterrupted clusters; both are deeply polymorphic; both show evidence of dynamic genomic behaviour within clusters; both show simple resistance/susceptibility behaviour confronted with strain-specific pathogen variants. The MHC shows these properties too, but MHC-related associations with infectious disease resistance and susceptibility are usually much less clear-cut, for a number of reasons.

In summary, we appreciate the various reservations raised above and have made some modifications to the text. Nevertheless the views are ours and, unlike the experimental results themselves (hopefully), can be strengthened or weakened by future experiments. We feel that we should be entitled to express our views clearly, so long as they are not evidently wrong, even if a referee is not of the same opinion.

1) There is a concern about the mechanism proposed for Irgb2-b1 function. In cells from CIM mice infected with RH parasites, the authors demonstrate a substantial increase in Irgb6 and Irgb2-b1 loading onto the parasite vacuole, compared to cells from B6 mice. In addition, although there is undetectable p-Irga6_CIM in CIM cells infected with RH parasites, Irga6_CIM can be phosphorylated if transfected into other cells (L929 cells), suggesting that Irga6 phosphorylation is actively inhibited in CIM cells. The authors state that Irgb2-b1_CIM inhibits the phosphorylation of endogenous Irga6 in CIM mice. This statement is based on experiments in which they transfect Irgb2-b1_CIM into B6 cells stimulated with IFN-γ and observe a reduction in p-Irga6 localization to the parasite vacuole membrane, which they quantify. However, they do not directly show that Irgb2-b1_CIM inhibits Irga6 phosphorylation. It seems that an alternate interpretation of these data (from Figure 6E,F) is that Irgb2-b1_CIM may inhibit the localization of Irga6 to the vacuole, where it is then phosphorylated. This point should be addressed/clarified.

The referee is correct, the data presented is open to another interpretation; thanks for pointing it out. Indeed we already have the data, which we now present as an additional image in Figure 6G, that tests the alternative hypothesis suggested by the referee. In the images originally presented we showed only the deficit of p-Irga6 at vacuoles loaded with Irgb2-b1_CIM. However using an antibody against total Irga6 we also measured the loading of total Irga6 at vacuoles loaded with Irgb2-b1_CIM. New Figure 6G shows that, so far from being reduced as predicted by the referee's hypothesis, the loading of un-phosphorylated, and thus active, Irga6 is significantly increased by the presence of Irgb2-b1_CIM. This result is expected if the absence of p-Irga6 shown in Figure 6F is because the Irgb2-b1_CIM acts, as proposed, as a decoy for the parasite kinase complex, drawing it away from its intended target, rather than as an inhibitor of Irga6 binding. Note that binding of Irgb6 at the vacuole is also increased by the presence of Irgb2-b1_CIM (Figure 6H).

Overall, the data implicating ROP5 in the loading of Irgb2-b1_CIM onto the vacuole surrounding virulent parasites is very indirect. Granted it is decreased in Δrop5 parasites, but this does not equate to a model where Irgb2-b1 binds to ROP5 or inhibits its activity. Other indirect mechanisms are possible. This portion of the manuscript is too speculative. Either concrete data need to be provided or this should be extensively modified to indicate it might result from direct or indirect binding. Does Irgb2-b1 have a direct effect on ROP18? It would look the same phenotypically (Δrop5 parasites would phenocopy this). The model promoted here suggests that ROP5 binds to IRGs. If this interaction is robust, it should be straightforward to demonstrate this for Irgb2-b1.

The referee is correct that the evidence we present is indirect, though it is internally consistent, and externally consistent with the already published evidence for the direct interaction between virulent ROP5 and Irga6, with a precise structural basis, and shortly to be published as a crystal structure of the binary complex (from the Boothroyd lab and shortly to be published as a crystal structure of the binary complex [from Michael Reese and John Boothroyd's lab, PDB codes 4LV5 (ROP5Bi:Irga6) and 4LV8 (ROP5Ci:Irga6) to be released 25/09/2013]). It is important to stress that the IRG proteins are a homologous family and there is every reason to expect that they have homologous behaviour at a structural level. The participation of helix 4 of Irga6 in this interaction is thus also consistent with the evidence we present that helix 4 of Irgb2-b1 is under divergent selection (Figure 7C), if Irgb2-b1 acts as a competitor of Irga6 for ROP5 and interacts directly with the virulence factor. The analysis of the mechanism by which Irgb-2b1 inhibits phosphorylation of Irga6 is not the main theme of this paper. On the basis of our results we consider that a direct interaction is the likeliest hypothesis, but a fuller analysis is beyond the scope of this paper. While these results are suggestive, further experimental analysis is needed to establish whether Irgb2-b1 indeed interacts directly with ROP5.

If such data are elusive, the authors should probably rethink their model.

To be clear, this is not yet a model; it is our present interpretation of the data and we consider it to be the most parsimonious. Certainly other interpretations are possible. The detail of the inhibitory mechanism is not central to the main argument of our paper, though of course is of great interest in itself.

Figure 7 is somewhat speculative, and is also not well explained. For example, what is the significance of the IIa/IIs ratios in the graphic?

The legend of Figure 7C has been modified to explain this data more clearly. Here is a more detailed account:

The πa/πs ratio gives an indication of the frequency with which coding variants occur in a sequence compared with the frequency of non-coding variants. It is related to the dN/dS parameter often used. The analysis is based on a moving window along the sequence and the colour coding records the regions with different frequencies of coding substitutions. The extended “red” domain on and adjacent to helix 4 identifies a region that appears to be under recent positive selection for a divergent sequence. This would occur, e.g., if this region of the molecule was under pressure to interact with new variants of a pathogen protein. Normally one would not expect to see this effect if Irgb2-b1 interacts with a host protein to “bridge” to ROP5 because the interaction surface between two host proteins would not be expected to evolve rapidly under divergent selection. We have added an additional comment on this to the text.

In the last paragraph of the Results when the authors describe Figure 7, they state, “Preliminary results suggest that Irgb2-b1_CIM may bind directly to the protein product of the virulent allele of ROP5.” What are these preliminary results? Including this interaction data would significantly strengthen the authors’ model that Irgb2-b1_CIM binding to ROP5 may contribute to blocking Irga6 phosphorylation in CIM mice.

The “preliminary results” are those specified in the paper and shown in Figure 7. We are working on this issue at the biochemical level as part of a further study on the details of the mechanism by which Irgb2-b1 interferes with Irga6 phosphorylation. As stated above, we feel we have made a good indirect case for a direct interaction with the ROP5 & ROP18 kinase complex, but a conclusive test is not straightforward. We consider it outside the scope of our paper and hope that the editors are prepared to leave this as it is.

2) The authors rely at least in part on Illumina resequencing data for their phylogenetic studies; such data are notoriously problematic for highly polymorphic and complex, duplicated regions of the genome. It would be appropriate for the authors to discuss how the BAC sequences they generated (which are only mentioned in the Materials and methods) as well as their extensive PCR data compared with the Illumina resequencing data.

To answer the referees in short, the Illumina resequencing is exceptionally reliable. To be clear, however, the only Illumina-derived transcriptome data of our own (at CCG Cologne) that we present here is the CIM transcriptome from IFNγ-induced DDC. The MSM BAC was sequenced by Illumina at Beijing Genomics Institute, de novo assembled then aligned to the C57BL/6 canonical genome. To test the validity of using Illumina resequencing of complete transcriptomes as a general approach for the rapid determination of unknown IRG haplotypes, we made an Illumina run on RNA of IFNγ-induced C57BL/6 fibroblasts to compare with the canonical genomic sequence of this strain that was established by extended Sanger sequencing on shotgun genomic fragments and from BACs. The C57BL/6 haplotype on Chr 11 is peculiarly difficult because it contains 4 very recently divergent tandem Irgb genes and additionally a duplicated Irgb6 gene, the two isoforms differing by only 3 non-coding nucleotides in the ORF. The resequencing results were not only easy to align against the canonical sequence, but also acted as a welcome confirmation of the sequences of the C57BL/6 genome: it was unclear for earlier releases of the C57BL/6 genome whether indeed there was an additional duplication in the IRGB region of Chr 11. The Illumina results show beyond doubt that both the extra copy of Irgb6 and the two extra tandem Irgb genes are real. All the intact IRG genes were recovered in up to 12,000 fold coverage; only degraded pseudogenes were not found. A second comparison was made with the CAST/Ei sequence. In that case, the data from the Sanger mouse genome project was based on Illumina resequencing while our data was based on Sanger sequencing of shotgun genomic fragments from BACs. Again there is perfect conformity between the two sequence sets.

Inevitably, Illumina resequencing of a transcriptome is unable to establish haplotypic linkage structure in heterozygotes carrying two IRG haplotypes, as is common in wild mice. The Sanger genomic Illumina data presented in this paper, however, is all obtained from inbred mice, whether of wild origin or laboratory strains.

Our resequencing reads and alignments are available if the referees would like to confirm their quality for themselves. The strength of the data is not only its consistency, but also the extreme depth of coverage, enabling us to identify and fully sequence IRG genes with expression levels over a 50x range or more. This favourable sensitivity is presumably achieved by the massive induction of IRG genes by IFNγ: we collect RNA from cells stimulated for 24 hr with high concentrations of IFNγ. At this time, IRG gene sequences may reach 1.0% of the total transcriptome.

3) It might be worth pointing out that type I strains are relatively rare, yet they share the “acute lab mouse” phenotype with many South American strains that not coincidentally share a ROP18 allele (and also ROP5 although this is less well characterized). If it were only clonal type I strains at play, the evolutionary argument that natural hosts must somehow be resistant to such strains would be much less meaningful (type 1 strains are less than 5% of all strains encountered in the wild).

The referees are correct that Type I strains are rather uncommon in much of the Eurasian domain; however, they appear to be relatively abundant in some parts of the Far East, e.g., (Puvanesuaran, et al. 2013) Malaysia, 100% type I; (Kyan, et al. 2012) Japan, 6/14 type I; (Zakimi, et al. 2006) Japan 22/49 type I. It is perhaps relevant that the highly resistant and very similar IRG haplotypes of the CIM and CAST/Ei strains are both from the Mm castaneus subspecies local to Southeast Asia. We feel, however, that to stress such correlations on such limited data may encourage various speculations when what is needed is a detailed ecological analysis from the wild environment that we are presently undertaking, coupled with direct experimental support.

4) The extreme polymorphism of the IRG family of proteins suggests that other pathogens might also impinge on this pathway. This is only briefly alluded to in the Discussion, but in fact it seems relevant from the opening and should be mentioned in the Introduction already.

With all respect for the referees’ view, we feel we deal with this important point quite comprehensively, devoting more than a third of the entire Discussion to it. It is a very interesting issue, although it does not bear directly on the results that we present in this paper. It is clear from the data that polymorphism in at least one of the IRG proteins on chromosome 11 is key to control of Toxoplasma virulence. In our view it is premature to speculate intensively on what might be the adaptive meaning of the rest of the polymorphism. Maybe it has to do with other Toxoplasma virulence genotypes; maybe with resistance properties of other pathogen species. Certainly the complete set of mouse pathogens under IRG control is not yet known, but equally certainly it is a surprisingly small subset of the entire range.

Unless the editors insist, we would rather not raise this very interesting but in our view secondary point in the Introduction when we devote a considerable amount of space to it in the Discussion.

The present work reveals the role of one component of the IRG family in combating Tg. And yet, the role of many other components remains undefined. Again, how important co-evolution with T. gondii is for IRG diversification is still unknown. Along these lines, the observation that the virulent strain GT1 can form cysts in resistant CIM mice is interesting, but not a very strong argument for co-evolution.

Agreed. Still, if only certain wild haplotypes are permissive for encystment by “virulent” strains, this does have, let us say, co-evolutionary potential.

Virulent strains of T. gondii can form cysts in any resistant species. In this regard, there does not seem anything particularly different about the wild mouse. It is the lab mouse that is the exception. It remains an interesting historical artefact that the extreme susceptibility of the lab mouse led to ROP kinase discovery.

A wild mouse homozygous for lab mouse chromosome 11 haplotypes would almost certainly be susceptible to virulent strains of T. gondii (though perhaps there are some resistant haplotypes on chromosome 18, even if not in CIM). Note that no other locus in the (CIM x BL/6) x BL/6 cross was able to compensate for the homozygous susceptible Chr 11 locus. Yet these mice are mongrels for the rest of the genome. We haven't found a perfect lab haplotype in a wild mouse yet although some are pretty close, but we haven't looked at very many yet either. Are the referees proposing that the haplotype itself is a laboratory artefact?

Presumably mice with these haplotypes are out there somewhere. We would rather say that the lab mouse a haplotypes, which are highly susceptible to virulent strains, constitute an accidental outcome of the breeding history of the mouse rather than an artefact (unless, as mentioned above, other haplotypes carry a cost under lab conditions).

Along these lines, little is said about the coincidence between different IRG genotypes and T. gondii strains in the wild, which is obviously important to determine the significance of the laboratory observations.

To our knowledge, nothing is known about this. The relevance and importance of the IRG system in controlling T. gondii is a recent discovery; its genetics is the content of this paper and has not been reported before. The referees raise what is obviously an important development from the present study, a development in which the present authors are presently fully engaged.

Author details

1. Jingtao Lilue

   Institute for Genetics, University of Cologne, Cologne, Germany

   Contribution

   JL, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Urs Benedikt Müller

   Competing interests

   The authors declare that no competing interests exist.

2. Urs Benedikt Müller

   Institute for Genetics, University of Cologne, Cologne, Germany

   Contribution

   UBM, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Jingtao Lilue

   Competing interests

   The authors declare that no competing interests exist.

3. Tobias Steinfeldt

   Institute for Genetics, University of Cologne, Cologne, Germany

   Contribution

   TS, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Jonathan C Howard

   1. Institute for Genetics, University of Cologne, Cologne, Germany
   2. Fundação Calouste Gulbenkian, Instituto Gulbenkian de Ciência, Oeiras, Portugal

   Contribution

   JCH, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   jhoward@igc.gulbenkian.pt

   Competing interests

   The authors declare that no competing interests exist.

• 3,422

  Page views

• 347

  Downloads

• 75

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.