Murine Guanylate Binding Protein 2 (mGBP2) controls Toxoplasma ...

Murine Guanylate Binding Protein 2 (mGBP2) controls Toxoplasma gondii replication Daniel Degrandia,1, Elisabeth Kravetsa,1, Carolin Konermanna,1, Cornelia Beuter-Guniaa, Verena Klümpersa, Sarah Lahmea, Eva Wischmanna, Anne K. Mausbergb, Sandra Beer-Hammera,c,d, and Klaus Pfeffera,2 a Institute of Medical Microbiology and Hospital Hygiene, Heinrich-Heine-University Duesseldorf, Duesseldorf 40225, Germany; bDepartment of Neurology, Heinrich-Heine-University Duesseldorf, Duesseldorf 40225, Germany; cDepartment of Pharmacology and Experimental Therapy, Institute of Experimental and Clinical Pharmacology and Toxicology, Eberhard Karls University, Tübingen 72074, Germany; and dInterfaculty Center of Pharmacogenomic and Pharmaceutical Research, Tübingen 72074, Germany

Edited by Tak W. Mak, The Campbell Family Institute for Breast Cancer Research, Ontario Cancer Institute at Princess Margaret Hospital, University Health Network, Toronto, ON, Canada, and approved November 27, 2012 (received for review April 12, 2012)

IFN-γ orchestrates the host response against intracellular pathogens. Members of the guanylate binding proteins (GBP) comprise the most abundant IFN-γ–induced transcriptional response. mGBPs are GTPases that are specifically up-regulated by IFN-γ, other proinflammatory cytokines, toll-like receptor agonists, as well as in response to Listeria monocytogenes and Toxoplasma gondii infection. mGBP2 localizes at the parasitophorous vacuole (PV) of T. gondii; however, the molecular function of mGBP2 and its domains in T. gondii infection is not known. Here, we show that mGBP2 is highly expressed in several cell types, including T and B cells after stimulation. We provide evidence that the Cterminal domain is sufficient and essential for recruitment to the T. gondii PV. Functionally, mGBP2 reduces T. gondii proliferation because mGBP2-deficient cells display defects in the replication control of T. gondii. Ultimately, mGBP2-deficient mice reveal a marked immune susceptibility to T. gondii. Taken together, mGBP2 is an essential immune effector molecule mediating antiparasitic resistance. pathogen defense

| cell-autonomous immunity | host-pathogen interaction

I

nterferons (IFNs) and tumor necrosis factor provide fundamental cellular defense mechanisms against intracellular pathogens (1–3). IFNs regulate the expression of several hundred genes, including four families of GTPases: Mx proteins, immunity related GTPases (IRGs; p47 GTPases), VLIGs, and p65 GBPs (4–7). Mx proteins have been shown to possess a potent antiviral activity against a wide range of paramyxoviruses, e.g., influenza and vesicular stomatitis virus (VSV) (8). Recent studies have attributed this effect to direct interaction of the protein with viral particles, thereby inhibiting viral replication (9, 10). The IFN-γ–induced IRGs comprise now a gene family with 23 functional members in mice (11). The investigation of gene-deficient mouse strains, such as Irgm1 (LRG-47), Irgd (IRG-47), and Irgm3 (IGTP), showed that these IRG proteins are essential during infection with intracellular pathogens, such as Listeria monocytogenes, Mycobacterium tuberculosis, or Toxoplasma gondii in mice (5, 12). Although IRGs exist in several mammalian species, the gene family appears to be degenerated in humans (11). In contrast, the 65-kDa GBPs are highly conserved throughout the vertebrate lineage (13, 14). In humans, seven orthologs and at least one pseudogene have been identified (13, 15). In mice, five GBPs had been initially described (16–18). Using Affymetrix and in silico analyses, we recently identified the complete mGBP gene family with the additional members mGbp6, mGbp7, mGbp8, mGbp9, mGbp10, and mGbp11 (14, 19). The genes of the mGBP family were located on clusters on chromosomes 3 (mGbp1, mGbp2, mGbp3, mGbp5, and mGbp7) and 5 (mGbp4, mGbp6, mGbp8, mGbp9, mGbp10, and mGbp11) (14). Additionally, we found that mGbp4 is a nonfunctional allele, at least in the C57BL/6 background (20). Furthermore, we described that mGBP1, mGBP2, mGBP3, mGBP6, mGBP7, and mGBP9 localize around the parasitophorous vacuole (PV) of T. gondii (ME49) (19). Even though 65-kDa GBPs have been known for more than 20 y, most

294–299 | PNAS | January 2, 2013 | vol. 110 | no. 1

aspects of their functions remain enigmatic and only scarce data exist on their function in cell lines. With respect to immune function, hGBP1 and mGBP2 have been shown to mediate antiviral activities in HeLa cells and NIH 3T3 fibroblasts, respectively, after infection with VSV or encephalomyocarditis virus (EMCV) (21, 22). The viral growth inhibition by mGBP2 was dependent on GTP binding for EMCV but not for VSV, suggesting different effector mechanisms of the protein depending on the type of virus. Recent findings suggest an important role of distinct GBP family members in the resistance against bacterial pathogens like L. monocytogenes and Mycobacteria by transporting antimicrobial cargo to bacteriacontaining phagosomes (23). Up to now, little is known about the role of mGBPs in parasitic infections. Previously, we could show rapid recruitment of mGBP2 to the PV of intracellular T. gondii (19). Now, we have characterized the function of mGBP2 in the immune defense against the parasite T. gondii in detail. mGBP2 protein expression was extensively analyzed in various cell types. Additionally, we further analyzed the subcellular distribution of mGBP2 and the role of the C-terminal, middle, and N-terminal domain with regard to the correct localization and redistribution in T. gondii-infected cells. Our results clearly show that the interaction between mGBP2 and the PV prevents the parasite from replication, thus inhibiting parasite spread. By generation of mGBP2−/− mice, we could define an increased susceptibility of mGBP2−/− mice in T. gondii infection, which is characterized by an increased parasite burden in the brain. Results mGBP2 Protein Expression. Previous studies have shown that mGBPs, in particular mGBP2, are highly up-regulated upon IFN-γ and TLR stimulation on transcriptional and protein level (17–19). To confirm whether this is reflected on the protein level in primary cells, macrophages were matured from bone marrow cells of C57BL/6 mice (BMDM) (Fig. 1A). After 16 h, the mGBP2 expression was substantially up-regulated by IFN-γ, LPS, and polyinosinic-polycytidylic acid (pI:C). Also, strong up-regulation of mGBP2 in BMDM was observed after stimulation with type I IFN and TNF. To elucidate whether spleen lymphocytes are able to express mGBP2, B (B220+) and T (CD90+) cells were purified out of spleens of C57BL/6 mice via cell sorting (Fig. 1B). Clearly, in both cell types mGBP2 was up-regulated after IFN-γ stimulation,

Author contributions: D.D., E.K., C.K., S.B.-H., and K.P. designed research; D.D., E.K., C.K., C.B.-G., V.K., S.L., E.W., and A.K.M. performed research; D.D., E.K., C.K., S.B.-H., and K.P. analyzed data; and D.D. and K.P. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1

D.D., E.K., and C.K. contributed equally to this work.

2

To whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1205635110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1205635110

mGBP genes. The expression of mGBP1, mGBP3, and mGBP5 in mGBP2−/− mice was assayed by infecting mGBP2+/− and mGBP2−/− mice i.p. with 20 cysts T. gondii (ME49). Five days after infection, the protein levels were measured in spleen, liver and lung (Fig. S1D). Clearly, mGBP1, -3, and -5 were expressed at normal levels in mGBP2−/− mice, ensuring that the targeting of the mgbp2 locus did not affect the expression of the neighboring genes. Also, the correct localization of other mGBPs was confirmed by infecting IFN-γ–treated mGBP2−/− MEFs with T. gondii. As shown in infected WT cells (19), the mGBP1 protein localizes correctly to the PV of the parasite (Fig. S1E). Thus, we conclude that the deletion of mGBP2 does not affect the expression and localization of other mGBPs.

Fig. 1. Expression of mGBP2. (A) mGBP2 protein expression in BMDM after 16-h stimulation with the indicated stimuli. (B) Protein expression of mGBP2 in various cell subtypes. Splenocytes of C57BL/6 mice were stained using B220+- or CD90+-specific antibodies (BD) and purified by cell sorting (BD FACS Aria), then stimulated for 16 h with the indicated stimuli. (C) Induction of mGBP2 is IRF1 dependent. Embryonic fibroblasts of IRF1-deficient mice were stimulated with IFN-γ for 16 h. (D) Protein expression of mGBP2 in the brain of T. gondii infected mice. C57BL/6 mice were infected i.p. with 20 cysts of freshly prepared T. gondii (ME49). After 5, 7, or 12 d, mice were killed and the brains were prepared. As positive control for Western blotting, cell lysate of IFN-γ–stimulated WT MEFs was used (+). All experiments were done in triplicate; one representative experiment is shown.

showing that mGBP2 is broadly expressed in innate and adaptive immune cells. Also activation of the T-cell receptor signaling in CD90+ cells by activation via CD3/CD28 resulted in increased expression of mGBP2, showing that the activation of mGBP2 is regulated by multiple pathways. Furthermore, a significant reduction of mGBP2 protein expression in IRF1−/− BMDM was observed, indicating that mGBP2 is under the transcriptional control of IRF1 (Fig. 1C). In previous studies, we examined the expression of mGBP2 in liver und lungs of T. gondii infected mice (19). mGBP2 expression could also be measured in the brain of infected mice. Expression of mGBP2 was first detectable after 12 d of infection (Fig. 1D), which correlates with the time window where T. gondii disseminates into the brain. Taken together, our data show that mGBP2 is expressed in several primary cells after appropriate activation and is subject to temporal and spatial regulation during the course of infection with T. gondii. Generation of mGBP2-Deficient Mice. To analyze the in vivo role of mGBP2 in infection we generated mGBP2−/− mice by homologous gene targeting in embryonic stem cells (Fig. S1A). The correct recombination was verified by Southern blot. Loss of mGBP2 expression in mGBP2−/− MEFs was confirmed by Western blotting (Fig. S1B). The mgbp2 locus is found on chromosome 3, adjacent to other mGBPs (Fig. S1C). Hence, we tested whether the inactivation of mGBP2 would affect the regulation of proximal Degrandi et al.

ization of mGBP2 with the parasitophorous vacuole of intracellular T. gondii ME49 but not of the virulent T. gondii strain BK in IFN-γ–activated MEFs (19). In addition, mGBP2 expression is highly induced upon infection with T. gondii and L. monocytogenes in several organs of C57BL/6 mice. These observations suggest an important role of mGBP2 during infection, together with other mGBPs, which may involve a direct function of mGBP2 at the PV of intracellular T. gondii. To test the in vivo role of mGBP2 in T. gondii infection, we infected WT, IRF1−/− and mGBP2−/− mice with high doses (40 cysts i.p.) of freshly prepared T. gondii ME49 and monitored the survival of the mice over 60 d (Fig. 2A). WT and mGBP2−/− mice showed an initially comparable mortality up to 12 d after infection, whereas all IRF1−/− died within 12 d. We measured the parasite burden in the acute phase of infection using a T. gondii-specific quantitative PCR (qPCR; Fig. S2). No significant differences in the parasite load during the acute phase of infection were measured in lung, liver, spleen, and brain of mGBP2−/− mice compared with WT mice. In the chronic phase of infection, WT mice did not show any increased mortality until day 45 after infection, whereas mGBP2−/− mice showed a steadily increased mortality over the whole observation period. To test the parasite load during the chronic phase, we counted cysts out of the brain of WT and mGBP2−/− mice at day 30 after infection. Strikingly, in mGBP2−/− mice, more than double the amount of cysts was enumerated compared with WT mice (Fig. 2B). Because we observed increased mortality of mGBP2−/− mice in T. gondii infections, we raised the question whether mGBP2 confers a cell autonomous inhibition of T. gondii replication. To address this possibility we infected WT and mGBP2−/− BMDMs, primary astrocytes, and MEFs and determined the ratio between PVs containing rosettes and single parasites 32 h after infection (see Figs. 2 C and D and 4A and Fig. S3A). In mGBP2−/− cells, we found significantly more rosettes than in WT cells, indicating a substantially increased parasite replication in the absence of mGBP2. This finding is further substantiated by the normal expression of other mGBPs, such as mGBP3 and mGBP5, in the absence of mGBP2 (Fig. S3 B and C). Interestingly, a decreased amount of vesiculated T. gondii PVs was observed in mGBP2−/− MEFs 5 h and 8 h after infection (Fig. 2 E and F). In summary, we observed a higher parasitic burden in the brain, which proves an increased susceptibility toward T. gondii in mGBP2−/− mice. Previously, we showed that mGBP2 is also highly expressed in several tissues of L. monocytogenes infected mice (19). To assay the in vivo significance of mGBP2 in L. monocytogenes infection, we infected WT, mGBP2−/−, and IRF1−/− mice with 1× LD50 L. monocytogenes i.p. and monitored the survival of the mice for 15 d. No significant differences in morbidity and mortality could be observed between WT and mGBP2−/− mice, whereas IRF1−/− mice succumbed early after infection (Fig. S4A). Also, mGBP2 failed to localize to intracellular L. monocytogenes (Fig. S4B). These data imply that mGBP2 is highly important for immunity against T. gondii but not for L. monocytogenes. PNAS | January 2, 2013 | vol. 110 | no. 1 | 295

IMMUNOLOGY

mGBP2-Deficiency Results in Susceptibility Toward T. gondii Infection but Not L. monocytogenes Infection. We observed a rapid colocal-

Fig. 2. Infection of mGBP2-deficient mice and cells with T. gondii. (A) C57BL/6, IRF1−/−, and mGBP2−/− mice were infected i.p. with T. gondii (40 cysts of strain ME49) and monitored for 60 d. The survival rate of WT (n = 14), IRF1−/− (n = 5), and mGBP2−/− (n = 19) mice is shown. A combination of two independent experiments is shown. (B) Cysts per brain of WT and mGBP2−/− mice at day 30 after i.p. infection with 40 cysts T. gondii (ME49) 30 d after infection (n = 3). ***P < 0.0005. (C) WT and mGBP2−/− BMDMs pretreated overnight with IFN-γ and infected with T. gondii ME49 for 32 h. T. gondii were stained with antiSAG1 mAb and anti–mouse-Cy3 (red). The amount of intracellular T. gondii rosettes and the amount of PVs containing a single parasite were enumerated microscopically, and the ratio was built (n = 3). ***P < 0.0005. (D) Astrocytes isolated from the brain of newborn WT and mGBP2−/− mice were treated as described in C (n = 3). ***P < 0.0005. (E) C57BL/6 and mGBP2−/− MEFs transduced with GFP were infected with T. gondii ME49 for 5 h and stained with anti-SAG1 mAb and anti–mouse-Cy3 (red). One representative image is shown. Arrows indicate PV disruption. (Scale bars: 5 μM.) (F) WT and mGBP2−/− MEFs transduced with GFP were infected with T. gondii ME49 for the indicated time periods and stained with anti-SAG1 mAb and anti–mouse-Cy3. The percentage of disrupted intracellular PVs out of total intracellular PVs from approximately 100 infected cells was enumerated microscopically (n = 6). **P < 0.005; ***P < 0.0005.

Kinetics of mGBP2 Acquisition to the T. gondii PV. The kinetics of mGBP2 acquisition at the PV of intracellular T. gondii was studied by live-cell confocal microscopy using GFP–mGBP2 transduced mGBP2−/− MEFs preincubated with IFN-γ before infection with T. gondii. mGBP2 was already found at the PV approximately 10 min after infection with a maximum observed after approximately 20 min (Fig. S5 and Movies S1 and S2). Functional Analysis of mGBP2 Protein Domains. The human GBP1 protein has been described to be composed of three distinct domains: the globular N-terminal domain, which includes all motifs responsible for nucleotide binding and hydrolysis (G domain), and two helical domains: the middle domain (M domain) and the C-terminal effector domain (E domain) (24). Based on sequence similarity of mGBP2 and hGBP1 and computer based protein structure prediction we identified the corresponding domains of the mGBP2 protein. To better understand the functional role of the aforementioned domains on the subcellular localization and function we stably transduced mGBP2−/− MEFs with N-terminally GFP-tagged deletion constructs of mGBP2 (Fig. S6A). The expression of the corresponding deletion constructs was verified by Western blot (Fig. S6B). The GFP-tagged mGBP2 full-length protein showed a vesicular-like localization comparable to the WT protein and readily relocalized to the T. gondii PV after infection in IFN-γ–stimulated cells (Fig. 3A). However, the mGBP2 deletion mutants showed no vesicularlike distribution. Mutants lacking the E domain (G, GM) were 296 | www.pnas.org/cgi/doi/10.1073/pnas.1205635110

ubiquitously distributed all over the cells. Mutants containing the E domain (E, ME, GE) were found to be homogenously localized in the cytosol. Thus, all three domains are required for correct localization of the mGBP2 protein in uninfected cells. Surprisingly, the deletion mutants bearing the C-terminal E domain (E, ME, GE) were able to accumulate at the T. gondii PV, whereas mutants lacking the E domain did not show this capacity (Fig. 3A). To determine whether the N-terminal domains (G, M) play a role in the effector function of mGBP2 once the protein reaches the T. gondii PV, we quantified the percentage of intracellular T. gondii, which showed colocalization with the GFP-tagged mGBP2 constructs over 48 h after infection (Fig. 3B). After 2.5 h, we observed only a slight reduction in colocalization of E domain containing mutants compared with endogenous or full-length mGBP2 protein. However, although full-length mGBP2 association with the PV peaked at 5 h after infection, the colocalization rates of deletion mutants at the PV quickly dropped. After 24 h, the amount of mGBP2-positive PVs containing single parasites dropped quickly with all mutants. We conclude that the E domain of mGBP2 is required and sufficient for association of the protein with the T. gondii PV. However, the G and M domains bear important functions for a sustained presence of mGBP2 at the PV. Isoprenylation of the C Terminus of mGBP2 Is Required for Accumulation at the PV. To verify whether the C-terminal region

of mGBP2 is required and sufficient for accumulation to the Degrandi et al.

reciprocal mGBP2-5 chimeric protein showed no vesicular-like distribution and no accumulation at the PV (Fig. S7A). Additionally, mGBP2 proteins containing the C-terminal CaaX-motif of mGBP5, which is responsible for posttranslational modification, still were able to relocalize to the T. gondii PV (Fig. S7B), although at reduced rates. Interestingly, mutating the CaaX motif of mGBP5 to the one of mGBP2 leads to association of mGBP5 to the PV. To clearly determine the role of the isoprenylation motif of mGBP2 (CTIL) we mutated the motif to STIL, disrupting the recognition sequence of the Geranygeranyltransferase I (25, 26). Interestingly, the subcellular localization as well as the potential to recruit to the T. gondii PV was lost (Fig. S7C). Thus, we conclude that mGBP2 specific isoprenylation is required for recruitment, but other motifs within the C-terminal helices of the proteins affect the efficiency of recruitment. C-Terminal E Domain of mGBP2 Is Required for the Control of T. gondii Proliferation. To characterize the functional role of the G,

Fig. 3. Subcellular localization of mGBP2 truncation mutants. (A) mGBP2−/− MEFs were transduced with the indicated GFP–mGBP2 constructs. Transduced cells were pretreated o/n with IFN-γ and infected with T. gondii ME49 for 2 h. T. gondii were stained with anti-SAG1 and anti–mouse-Cy3 (red). One representative experiment out of two is shown. (B) mGBP2−/− MEFs were transduced with the indicated mGBP2 constructs. Cells were pretreated overnight with IFN-γ and infected with T. gondii ME49 for the indicated time points. T. gondii were stained with anti-SAG1 and anti–mouse-Cy3. The amount of T. gondii PVs colocalizing with GFP was enumerated (n = 2).

T. gondii PV we cloned a chimeric protein composed of the G domain of mGBP5 and the C-terminal region of mGBP2 (mGBP52) (Fig. S7D). WT mGBP5 protein is not able to recruit to the PV and shows no vesicular-like distribution (19). However, the mGBP5-2 chimeric protein did not localize in vesicle-like structures, but accumulated at the T. gondii PV of infected cells. The Degrandi et al.

Discussion T. gondii is an obligate intracellularly replicating parasite that infects a large range of cells of warm-blooded animals. The replication of T. gondii within cells is effectively confined by the antimicrobial effects of IFN-γ, which triggers the expression of approximately 2,000 genes, many of which are poorly functionally characterized up to now (5, 27). In mice, a few effector mechanisms have been shown to control stasis and killing of intracellular parasites (28). In recent years, several members of the IRGs have been shown to play a major role in controlling T. gondii replication (29). The mechanisms mediated by IRGs are not well understood. The current hypothesis implies a direct stripping of the parasitophorous vacuole membrane after assembly and oligomerization of IRGs at the PV, releasing the parasite into the cytosol making it accessible to further digesting mechanisms and immune responses (5, 30). We have previously shown that the GBP proteins are able to accumulate at the T. gondii PV very similarly to the IRG proteins (19), suggesting that GBP proteins may act in an analogous way. Recently, an important role of GBPs in T. gondii defense was shown by deletion of the complete genomic GBP cluster on PNAS | January 2, 2013 | vol. 110 | no. 1 | 297

IMMUNOLOGY

M, and E domains of mGBP2 on T. gondii control we examined the replication of parasites in mGBP2−/− MEFs stably transduced with N-terminally GFP-tagged deletion constructs of mGBP2 (Fig. 3 and Fig. S6) and determined the ratio between PVs containing rosettes and single parasites 32 h after infection (Fig. 4A). In GFP expressing mGBP2−/− MEFs, we found significantly more rosettes than in WT cells or mGBP2−/− cells reconstituted with GFP–mGBP2, confirming our results obtained from BMDMs and astrocytes and proving full functionality of the GFP-tagged mGBP2 protein. Reconstitution of mGBP2−/− cells with deletion constructs containing the E domain led to an intermediate level of replication control. Loss of mGBP2 expression or reconstitution with deletion mutants had no effect on the amount of PVs per cell or the amount of infected cells (Fig. S8 A and B, respectively). Interestingly, only PVs containing a single parasite showed colocalization with mGBP2 or the E domain containing deletion mutants (Fig. 4B). A significant amount of PVs with single parasites was still found in mGBP2−/− cells 32 h after infection, although considerably lower than in WT cells. This might be due to the targeting of other mGBP family members to the PV. We always observed colocalization of mGBP1 and mGBP2 at the same T. gondii PV, whereas mGBP9 partly targeted other intracellular parasites (Fig. S9). Thus, some mGBPs appear to act together at specific PVs, other are accumulated to different intracellular PVs. Taken together, these findings indicate that recruitment of the mGBP2 protein at the T. gondii PV inhibits parasite replication induced by a block of rosette formation. This process is mediated by the C-terminal E domain of the protein.

Fig. 4. The C-terminal domain of mGBP2 is required to inhibit T. gondii proliferation. (A) mGBP2−/− MEFs transduced with the indicated constructs were pretreated o/n with IFN-γ and infected with T. gondii ME49 for 32 h. T. gondii were stained with anti-SAG1 mAb and anti–mouse-Cy3 (red). The amount of intracellular T. gondii rosettes and the amount of PVs containing a single parasite were enumerated microscopically, and the ratio was built (n = 3). *P < 0.05. (B) Cells described in A were analyzed microscopically. Rosettes containing PVs show no colocalization with mGBP2 constructs. One representative experiment out of three is shown.

chromosome 3 in mice, inactivating mgbp1, mgbp2, mgbp3, mgbp5, and mgbp7 (31). Here, we show in vivo that genetic loss of one single GBP protein in mice, mGBP2, is sufficient to induce susceptibility to T. gondii, demonstrating that mGBP proteins have specific nonredundant functions in T. gondii infection. This increased morbidity and mortality was particularly pronounced during the chronic phase of infection. Although we could not observe increased parasite load in the first days of infection in mGBP2−/− mice, formation of cysts in the brain during the chronic phase was markedly heightened, implying that control of T. gondii replication in the brain is severely affected in the absence of mGBP2. Increased replication of T. gondii could be observed in in vitro infections of macrophages, astrocytes, and MEFs from mGBP2−/− mice, accompanied by a decreased level of PV vesiculation. In contrast, we could not determine any 298 | www.pnas.org/cgi/doi/10.1073/pnas.1205635110

significance for the mGBP2 protein in controlling L. monocytogenes infection. mGBP2 mice survived the infection similarly to WT. Additionally, we could not detect any colocalization of mGBP2 with intracellular L. monocytogenes organisms. A recent publication, however, has shown increased Listeria and Mycobacteria replication in mGBP1−/− mice (23). It is likely that the support of host resistance to a specific pathogen spectrum is different for each GBP family member. A similar observation was made for members of the IRG family, where, e.g., genetic loss of LRG-47 (Irgm3) results in acute susceptibility to T. gondii and L. monocytogenes infection, whereas IRG-47−/− (Irgd) mice showed susceptibility to T. gondii in the chronic phase of infection but not to L. monocytogenes (32). Nonetheless, partial functional redundancy or synergism between mGBP family members cannot be excluded, because we observed concomitant recruitment of mGBP1 and mGBP2 to the same T. gondii PV. Yamamoto et al. (31) could not find any rescue of T. gondii control in cells with the deleted GBP cluster on chromosome 3 reconstituted with mGBP2. This observation, together with our data, suggests that the antiparasitic effect of mGBP2 is essential but not sufficient to effectively control T. gondii infection and that synergism with other mGBPs is required. mGBP2 expression can be induced in hematopoietic and nonhematopoietic cells (ref. 19 and this study). It is reasonable to assume that mGBP2 mediates its antimicrobial role in a cell autonomous way, similarly to IRG proteins. This is mirrored by our observation that mGBP2−/− MEFs are impaired in controlling T. gondii replication within the PV, resulting in increased rosette formation. The mode of action of mGBP2 that leads to replication control is not clear up to now. Kim et al. (23) have suggested that GBP proteins, such as mGBP1, mGBP7 and mGBP10, promote killing of intracellular pathogens by delivering subunits of the phagocyte oxidase complex (Phox) or antimicrobial peptides to the pathogen containing vacuole. Production of reactive oxygen species by Phox has not been shown to be essential in T. gondii control (33). However, targeting of the autolysosomal machinery to the parasite has been implicated by several studies to play a major role in the control of infection with contribution of IFN-inducible IRGs (34, 35). An involvement of mGBP2 in targeting of membraneous structures to the T. gondii PV, possibly of autolysosomal nature, is conceivable particularly considering the requirement for the C-terminal isoprenylation motif and thus association with membraneous compartments for correct localization and recruitment of mGBP2 to the PV. These concepts are currently under investigation. mGBP2 belongs to the superfamily of Dynamin-related proteins (36). Although the crystal structure of mGBP2 is not available yet, the sequence similarity to human GBP1 allows the distinction of the three typical GBP domains: the globular G domain, bearing the motifs responsible for GTP-binding and hydrolysis, and the two helical middle (M) and effector (E) domains. Our data clearly reveal that the E domain of mGBP2 is essential for the recruitment of the protein to the T. gondii PV and for the control of parasite replication. This reflects the necessity of membrane integration which is probably mediated via the C-terminal isoprenylation motif. Surprisingly, the E domain alone was sufficient to partially restore parasite replication control in mGBP2−/− MEFs. Interestingly, another study showed that the C-terminal helices of mGBP2 alone are able to mediate effector functions by influencing cell spreading (37). These findings raise the question for the relevance of the G domain and GTPase activity in the antimicrobial effects of mGBP2. The data obtained in this study suggest that the G domain is on the one hand responsible for the localization of mGBP2 within vesicle-like structures in uninfected cells and on the other hand for stable association of the protein with the T. gondii PV. The significance of the vesicle-like distribution of mGBP proteins has not been understood until now, complicated by the fact that beyond containing GBPs, the nature of this structures could not be characterized by us or other groups, Degrandi et al.

1. Huang S, et al. (1993) Immune response in mice that lack the interferon-gamma receptor. Science 259(5102):1742–1745. 2. Pfeffer K, et al. (1993) Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell 73(3): 457–467. 3. Endres R, et al. (1997) Listeriosis in p47(phox-/-) and TRp55-/- mice: protection despite absence of ROI and susceptibility despite presence of RNI. Immunity 7(3):419–432. 4. Boehm U, Klamp T, Groot M, Howard JC (1997) Cellular responses to interferongamma. Annu Rev Immunol 15:749–795. 5. Martens S, Howard J (2006) The interferon-inducible GTPases. Annu Rev Cell Dev Biol 22:559–589. 6. Shenoy AR, et al. (2007) Emerging themes in IFN-gamma-induced macrophage immunity by the p47 and p65 GTPase families. Immunobiology 212(9-10):771–784. 7. Taylor GA, Feng CG, Sher A (2007) Control of IFN-gamma-mediated host resistance to intracellular pathogens by immunity-related GTPases (p47 GTPases). Microbes Infect 9 (14-15):1644–1651. 8. Haller O, Stertz S, Kochs G (2007) The Mx GTPase family of interferon-induced antiviral proteins. Microbes Infect 9(14-15):1636–1643. 9. Haller O, Kochs G (2002) Interferon-induced mx proteins: Dynamin-like GTPases with antiviral activity. Traffic 3(10):710–717. 10. Gao S, et al. (2011) Structure of myxovirus resistance protein a reveals intra- and intermolecular domain interactions required for the antiviral function. Immunity 35(4): 514–525. 11. Bekpen C, 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 (11):R92. 12. Taylor GA, Feng CG, Sher A (2004) p47 GTPases: regulators of immunity to intracellular pathogens. Nat Rev Immunol 4(2):100–109. 13. Olszewski MA, Gray J, Vestal DJ (2006) In silico genomic analysis of the human and murine guanylate-binding protein (GBP) gene clusters. J Interferon Cytokine Res 26 (5):328–352. 14. Kresse A, et al. (2008) Analyses of murine GBP homology clusters based on in silico, in vitro and in vivo studies. BMC Genomics 9:158. 15. Tripal P, et al. (2007) Unique features of different members of the human guanylatebinding protein family. J Interferon Cytokine Res 27(1):44–52. 16. Cheng YS, Colonno RJ, Yin FH (1983) Interferon induction of fibroblast proteins with guanylate binding activity. J Biol Chem 258(12):7746–7750. 17. Boehm U, et al. (1998) Two families of GTPases dominate the complex cellular response to IFN-gamma. J Immunol 161(12):6715–6723. 18. Nguyen TT, Hu Y, Widney DP, Mar RA, Smith JB (2002) Murine GBP-5, a new member of the murine guanylate-binding protein family, is coordinately regulated with other GBPs in vivo and in vitro. J Interferon Cytokine Res 22(8):899–909. 19. Degrandi D, et al. (2007) Extensive characterization of IFN-induced GTPases mGBP1 to mGBP10 involved in host defense. J Immunol 179(11):7729–7740. 20. Konermann C, et al. (2007) In silico and in vitro characterization of mGBP4 splice variants. DNA Cell Biol 26(12):847–851.

Degrandi et al.

host resistance effector system providing protection against intracellular pathogens. Materials and Methods C57BL/6 mice were purchased from Charles River and maintained in the animal facility of the Heinrich-Heine-University under SPF conditions. All procedures performed on animals in this study have been approved by the Animal Care and Use Facility of the Heinrich-Heine University of Duesseldorf and the government of Nordrhein-Westfalen and were in accordance with the German animal laws. L. monocytogenes infection in C57BL/6 and mutant mice was performed by intraperitoneally injection of approximately 0.1 × LD50 or 1 × LD50, and survival was monitored over 15 d. For T. gondii infection mice were injected intraperitoneally with 0.2 mL PBS containing 40 cysts of the T. gondii strain ME49. The cysts were prepared out of brains of infected CD1 mice as described (19). The organs were removed at the indicated time points post infection. Additional materials and methods are described in SI Materials and Methods. ACKNOWLEDGMENTS. We thank Julia Hartmann and Nicole Küpper for excellent experimental assistance and Thomas Hartung for providing Listeria LTA. This study was generously supported by the Deutsche Forschungsgemeinschaft (Grants FOR729, SFB590, and GRK1045) and the Jürgen Manchot Foundation.

21. Anderson SL, Carton JM, Lou J, Xing L, Rubin BY (1999) Interferon-induced guanylate binding protein-1 (GBP-1) mediates an antiviral effect against vesicular stomatitis virus and encephalomyocarditis virus. Virology 256(1):8–14. 22. Carter CC, Gorbacheva VY, Vestal DJ (2005) Inhibition of VSV and EMCV replication by the interferon-induced GTPase, mGBP-2: Differential requirement for wild-type GTP binding domain. Arch Virol 150(6):1213–1220. 23. Kim BH, et al. (2011) A family of IFN-γ-inducible 65-kD GTPases protects against bacterial infection. Science 332(6030):717–721. 24. Prakash B, Praefcke GJ, Renault L, Wittinghofer A, Herrmann C (2000) Structure of human guanylate-binding protein 1 representing a unique class of GTP-binding proteins. Nature 403(6769):567–571. 25. Sinensky M, Lutz RJ (1992) The prenylation of proteins. Bioessays 14(1):25–31. 26. Stickney JT, Buss JE (2000) Murine guanylate-binding protein: Incomplete geranylgeranyl isoprenoid modification of an interferon-gamma-inducible guanosine triphosphate-binding protein. Mol Biol Cell 11(7):2191–2200. 27. Zhang SY, et al. (2008) Inborn errors of interferon (IFN)-mediated immunity in humans: insights into the respective roles of IFN-alpha/beta, IFN-gamma, and IFNlambda in host defense. Immunol Rev 226:29–40. 28. Scharton-Kersten TM, Yap G, Magram J, Sher A (1997) Inducible nitric oxide is essential for host control of persistent but not acute infection with the intracellular pathogen Toxoplasma gondii. J Exp Med 185(7):1261–1273. 29. Howard JC, Hunn JP, Steinfeldt T (2011) The IRG protein-based resistance mechanism in mice and its relation to virulence in Toxoplasma gondii. Curr Opin Microbiol 14(4): 414–421. 30. Martens S, et al. (2005) Disruption of Toxoplasma gondii parasitophorous vacuoles by the mouse p47-resistance GTPases. PLoS Pathog 1(3):e24. 31. Yamamoto M, et al. (2012) A cluster of interferon-γ-inducible p65 GTPases plays a critical role in host defense against Toxoplasma gondii. Immunity 37(2):302–313. 32. Collazo CM, 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(2):181–188. 33. Halonen SK, Weiss LM (2000) Investigation into the mechanism of gamma interferonmediated inhibition of Toxoplasma gondii in murine astrocytes. Infect Immun 68(6): 3426–3430. 34. Subauste CS (2009) Autophagy in immunity against Toxoplasma gondii. Curr Top Microbiol Immunol 335:251–265. 35. Hunn JP, Feng CG, Sher A, Howard JC (2011) The immunity-related GTPases in mammals: A fast-evolving cell-autonomous resistance system against intracellular pathogens. Mamm Genome 22(1-2):43–54. 36. Praefcke GJ, McMahon HT (2004) The dynamin superfamily: Universal membrane tubulation and fission molecules? Nat Rev Mol Cell Biol 5(2):133–147. 37. Balasubramanian S, Messmer-Blust AF, Jeyaratnam JA, Vestal DJ (2011) Role of GTP binding, isoprenylation, and the C-terminal alpha-helices in the inhibition of cell spreading by the interferon-induced GTPase, Mouse Guanylate-Binding Protein-2. J Interferon Cytokine Res 31(3):291–298. 38. Kravets E, et al. (2012) The GTPase activity of murine guanylate-binding protein 2 (mGBP2) controls the intracellular localization and recruitment to the parasitophorous vacuole of Toxoplasma gondii. J Biol Chem 287(33):27452–27466.

PNAS | January 2, 2013 | vol. 110 | no. 1 | 299

IMMUNOLOGY

yet. However, binding and hydrolysis of GTP to GDP and GMP appears to enforce and prolong localization of mGBP2 at the T. gondii PV; thus, we suggest a regulatory function of the G domain on the antimicrobial effector function of the protein. This hypothesis is also supported by the fact that the E domain alone spontaneously accumulates at the T. gondii PV even in cells not stimulated with IFN-γ, whereas full-length mGBP2 required additional stimulation of the cells with IFN-γ for efficient recruitment to the parasite, probably by interaction with other IFN-γ–induced proteins. We recently could show that GTP binding and hydrolysis is pivotal for correct subcellular localization of full-length mGBP2 in uninfected and T. gondii infected cells (38). These observations suggest that after GTP hydrolysis a recruitment motif within the E domain of mGBP2 becomes unmasked, which enables the association with the T. gondii PV. It will be interesting to study in detail the regulatory role of the G domain of mGBP2 in regard to its antiparasitic function. In summary, we have defined mGBP2 as an important effector molecule in the IFN-γ–triggered cell autonomous resistance to T. gondii in mice. Thus, the GBP family emerges as a very important