Proximity biotinylation reveals novel secreted dense granule proteins of Toxoplasma gondii bradyzoites

Toxoplasma gondii is an obligate intracellular parasite which is capable of establishing life-long chronic infection in any mammalian host. During the intracellular life cycle, the parasite secretes an array of proteins into the parasitophorous vacuole (PV) where it resides. Specialized organelles called the dense granules secrete GRA proteins that are known to participate in nutrient acquisition, immune evasion, and host cell-cycle manipulation. Although many GRAs have been discovered which are expressed during the acute infection mediated by tachyzoites, little is known about those that participate in the chronic infection mediated by the bradyzoite form of the parasite. In this study, we sought to uncover novel bradyzoite-upregulated GRA proteins using proximity biotinylation, which we previously used to examine the secreted proteome of the tachyzoites. Using a fusion of the bradyzoite upregulated protein MAG1 to BirA* as bait and a strain with improved switch efficiency, we identified a number of novel GRA proteins which are expressed in bradyzoites. After using the CRISPR/Cas9 system to characterize these proteins by gene knockout, we focused on one of these GRAs (GRA55) and found it was important for the establishment or maintenance of cysts in the mouse brain. These findings highlight new components of the GRA proteome of the tissue-cyst life stage of T. gondii and identify potential targets that are important for maintenance of parasite persistence in vivo.


Introduction
Toxoplasma gondii is an apicomplexan parasite that chronically infects nearly every animal and approximately one-third of the world's human population [1][2][3]. While the infection is typically asymptomatic in healthy persons, infection in immunocompromised patients (such as those with AIDS or patients taking immunosuppressive drugs) can result in life-threatening PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0232552 May 6, 2020 1 / 20 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 central nervous system disease [3,4]. While therapies exist that can combat the acute infection consisting of rapidly growing tachyzoites, there are no effective treatments that can clear the chronic infection which is mediated by slow-growing bradyzoite cysts. Thus, patients who are chronically infected with bradyzoites live under a life-long threat of reactivation of the parasite if a lapse in immune surveillance occurs [4]. A mechanistic understanding of how T. gondii bradyzoite cysts are formed and able to maintain lifelong persistence in the host is critical for the development of novel therapies that target this important intracellular pathogen.
T. gondii actively invades its host cells and replicates inside of a membrane-bound parasitophorous vacuole (PV) within the host cell cytoplasm [5]. Host cell invasion, PV formation and maintenance are mediated by a set of specialized secretory organelles known as micronemes, rhoptries, and dense granules [6][7][8][9]. While micronemes and rhoptries play roles in the initial stages of attachment and invasion, the dense granules secrete proteins called GRAs into the vacuolar space that participate in the remodeling and maintenance of the PV during intracellular replication [10][11][12][13][14][15][16]. While many GRAs function within the vacuole after secretion, some GRAs are able to cross the vacuolar membrane into the host cell and hijack cellular immune functions [17][18][19][20][21][22]. Most of the currently known GRAs have been characterized in the context of the acutely infectious tachyzoite life-cycle stage of the parasite. As expected, some of these have been found to have important roles during both stages of infection. GRA17, which operates as a small-molecule transporter at the PVM in tachyzoites, has also been shown to be important for growth and maintenance of bradyzoites [23,24]. Several other well-studied GRAs such as GRA1, GRA2, GRA4, GRA5, GRA6, GRA9 and GRA12 have been shown to express during the bradyzoite phase and localize to the cyst wall and matrix at different points of maturation [16,[25][26][27][28]. Deletion of GRA2 results in defective formation of the cyst matrix and the failure of GRA4 and GRA6 to accumulate at the cyst periphery; this finding was in line with previous data which revealed the importance of GRA2, GRA4 and GRA6 in the formation of the intravacuolar network in tachyzoites [12,16,29].
While cyst formation and maintenance are likely to involve an array of bradyzoite-specific GRA proteins, only a few such proteins have been identified and little is known how they might enable the parasite to establish and maintain the chronic infection. These include CST1, bradyzoite pseudokinase 1 (BPK1) and the newly discovered CST2/GRA50, CST3/GRA51, CST4, CST5/GRA52, and CST6/GRA53 [30][31][32][33][34]. The deletion of the glycoprotein CST1 results in fragile cysts with thin cyst walls and lower cyst burden in a chronic-infection model [31]. BPK1 was found to be important in oral infectivity, and the disruption of this protein led to small in vivo cysts whose walls were more easily broken down by pepsin-acid treatment [34]. The purification of the cyst wall using Percoll gradient followed by immunoprecipitation with an anti-CST1 monoclonal antibody was used to identify cyst components. This study revealed 5 new bradyzoite-upregulated GRAs which localized to the tissue cyst wall and cyst matrix (named CST2 -CST6). Of these, CST2/GRA50 was important for parasite virulence and establishment of cyst burden [33]. Although this study revealed new proteins which are bradyzoite-specific cyst-wall components, it is likely that there are many more GRAs which participate in the formation and maintenance of bradyzoite tissue cysts during the chronic T. gondii infection.
To further bridge this gap in our knowledge, we have developed new approaches to rapidly identify and determine the function of novel GRAs. One important advance is our adaptation of the BioID technique to T. gondii for identifying proteins in specific subcellular compartments [35,36]. This approach utilizes a bait protein fused to a promiscuous biotin ligase (BirA � ) that biotinylates interacting and proximal proteins in subcellular compartments in vivo. The resulting biotinylated proteins are isolated by streptavidin chromatography and are then identified by mass spectrometry. We have recently utilized this approach to identify a robust dataset of previously known and candidate vacuolar proteins in the tachyzoite stage of the parasite [36]. Our follow-up endogenous tagging of a subset of these candidate GRAs resulted in the verification of many new GRAs, demonstrating the effectiveness of this approach. This dataset thus provided an initial tachyzoite GRA proteome of T. gondii, many of which are also expressed in bradyzoites.
In this manuscript, we build on these results by extending the GRA BioID experiments to bradyzoites, utilizing a bradyzoite-upregulated protein as a bait and a background strain which enables improved switching to bradyzoites in vitro. Using this approach, we have identified five novel GRA proteins expressed in bradyzoite cysts. We subsequently used CRISPR/ Cas9 to disrupt all five of the genes encoded these GRAs, demonstrating that they are not essential for growth in vitro. We then focused on one of these knockouts (GRA55) which is dispensable for growth in vitro and virulence of T. gondii during the acute phase of infection but is important for establishing or maintaining chronic infection in a mouse model. Together, this research identifies a group of novel bradyzoite-upregulated GRA proteins that are novel potential targets for the chronic T. gondii infection.

Adaptation of BioID for Toxoplasma bradyzoite GRAs
Our previous success with using BioID to identify secreted GRAs in tachyzoites suggested that this approach would also work for bradyzoites. To improve our changes of obtaining bradyzoite GRAs, we made several modifications to our approach. First, we used as our background the PrugniaudΔku80 strain of T. gondii that contains GFP driven by the bradyzoite specific promoter LDH2, which has previously been shown to be effective at monitoring switching to bradyzoites [37]. Second, we sought to improve the efficiency of in vitro switching to bradyzoites, thus we deleted AP2IV-4 (S1 Fig), which encodes one of the nuclear Apetala-2 (AP2) proteins which has previously been shown to be important for the maintenance of the tachyzoite phenotype in vitro [38]. Lastly, we used the bradyzoite-upregulated GRA protein MAG1 [39,40] as our bait which we fused to the biotin ligase BirA � (Fig 1A). Deletion of AP2IV-4 and obtaining the PruΔku80::Δap2iv-4 and MAG1-BirA � ::Δap2iv-4 parasites enables increased bradyzoite switching efficiency without disrupting the lytic cycle of the tachyzoites as previously described [38].
When we expressed the MAG1-BirA � fusion driven from its endogenous promoter, the fusion was readily detectable in tachyzoites and trafficked appropriately to the PV when examined by immunofluorescence assay (IFA) (Fig 1B). While MAG1 is considered a bradyzoite GRA, this detection was expected as it is clearly expressed in tachyzoites and is one of the stronger hits in our tachyzoite GRA BioID experiments [36]. Upon in vitro switching to bradyzoites, the fusion expressed in bradyzoites and trafficked appropriately to cyst vacuole where it co-localized with the known GRA marker GRA14 (Fig 1B) [41]. The fusion protein migrated at a molecular weight of~105 kDa as predicted [42] In vivo biotinylation of the PV using MAG1-BirA � parasites After producing switch-efficient MAG1-BirA � parasites, we assessed whether the fusion protein was active in the PV. To do this, we switched MAG1-BirA � parasites in vitro using high pH media for 5 days and supplemented biotin to the media on the final day [43]. Transition from the tachyzoite to bradyzoite stage was confirmed by the presence of GFP in the parasite cytosol ( Fig 1C). As a control, MAG1-BirA � parasites were switched on a separate monolayer, without biotin supplementation. While the control parasites just showed the expected apicoplast background staining, we observed streptavidin staining overlapping with the fusion protein in the parasite vacuole when we supplemented biotin (Fig 1C). This demonstrated that the biotin was able to penetrate the PV compartment and the MAG1-BirA � was active during the induced bradyzoite stage. We additionally demonstrated by western blot that there was an increase in the number of biotinylated proteins in MAG1-BirA � lysates when excess biotin was supplemented to the media ( Fig 1D). We then proceeded with a large-scale MAG1-BioID experiment by infecting human foreskin fibroblasts with a high multiplicity of infection, switching them in vitro to bradyzoites over the course of 5 days, supplementing biotin on day 4. We then harvested the cells, lysed the samples using stringent conditions, and purified the biotinylated protein fraction using streptavidin-affinity chromatography.
As with previous experiments, mass spectrometry analysis revealed a large number of tachyzoite and bradyzoite GRAs highly ranked by peptide spectrum matches when we examined the MAG1-BirA � lysate in comparison to the parental lysate (Table 1 and S1 Table) [36]. We found the bait protein (MAG1) as well as almost all numbered GRAs, MYR proteins, TgIST1, WNG1, WNG2, bradyzoite GRAs (such as CST proteins, MCP3, MCP4, and BPK1), cyclophilin, NTPaseI and NTPaseII. BPK1 did appear, although slightly lower ranking in abundance. There were several expected proteins that were missing from our dataset including the exported proteins GRA19, GRA20 and GRA21 [44]. In total, the MAG1-BioID experiment revealed 537 proteins that were not present in the control lysate, 675 proteins in total including some GRAs that were present in comparatively low levels in the control lysate (S1 Table). GRAs constituted many of the highest-ranking proteins in this experiment, however, as  expected many other secreted proteins including rhoptry proteins (ROPs/RONs), and microneme proteins (MICs) were also found concentrated in the top hits of the dataset.

Identification of novel bradyzoite dense granule proteins from BioID datasets
Our dataset also contained a large number of hypothetical proteins. We narrowed this list of proteins as likely GRAs by selecting those proteins that were highly ranked in the dataset and lacked a C-terminal endoplasmic reticulum retention signal (K/HDEL) [11,45]. We prioritized those proteins with predicted signal peptides, however as many secreted proteins do not contain these sequences, we considered GRA candidates from the hypothetical proteins that also lacked this feature. Because we wished to focus on those proteins that were upregulated in bradyzoites, we also narrowed our candidate pool by focusing on proteins that were shown to be highly transcribed in bradyzoites. However, we also chose some candidates whose expression was similar between tachyzoites and bradyzoites [46]. We chose 9 candidates and used endogenous gene tagging to recombine a C-terminal 3xHA epitope tag for each gene product and examined their localization by IFA. We found that 5 of these trafficked to the PV in both tachyzoites and bradyzoites (Fig 2). We named these GRA55 (TgME49_309760), GRA56 (TgME49_309930), GRA57 (TgME49_217680), and GRA58 (TgME49_268790). In addition, while reviewing earlier datasets from a prior GRA-BioID experiments [36], we found another GRA-GRA59 (TgME49_313440)-which was also upregulated in bradyzoites. Of these confirmed GRA proteins, only GRA56 had an identifiable domain which has homology to melibiase, an alpha-galactosidase that catabolizes the disaccharide melibiose.

Targeted deletion of novel GRA proteins and functional characterization
Once we identified these novel GRAs, we assessed their function by gene deletion using CRISPR/Cas9 and homologous recombination [47][48][49]. Gene knockouts were assessed by loss of the epitope tag from the parental line and replacement of the DHFR selectable marker used in tagging for HXGPRT used in the knockout (e.g. by growth in mycophenolic acid and xanthine but failure to grow in pyrimethamine) ( Fig 3A). Each of the knockouts was then confirmed by PCR ( Fig 3B). We examined the deletion mutants by plaque assay and discovered none of the lines had impaired in vitro growth (Fig 4, S2 Fig). We then proceeded with a pilot study of each of the knockouts to identify severe changes in cyst burden. This pilot study suggested that Δgra55 parasites produced substantially lower cyst numbers per mouse brain, thus we focused on this strain for complementation and more detailed in vivo analyses.

GRA55 is not important for in vitro growth or virulence but affects chronic infection of Type II T. gondii
We engineered GRA55-complemented parasites by re-introducing the wild-type gene driven from its endogenous promoter into the uracil phosphoribosyltransferase (UPRT) locus of the Δgra55 parasite genome [51] (Fig 4A). We verified that GRA55 targeted properly to the PV by IFA and showed similar expression levels to wild type parasites by western blot and thus we named this strain GRA55c (Fig 4B and 4C). We verified that that Δgra55 parasites did not have any decreased growth by plaque assays (Fig 4D) compared to wild-type and complemented strains, confirming that this protein does not meaningfully participate in parasite growth in vitro. Upon examination of the role of GRA55 in acute infection in mice, we found that Δgra55 parasites killed mice in a dose-dependent manner, identical to that of mice infected with wild-type and GRA55c (Fig 4E, 4F and 4G). We then conducted an investigation into  whether GRA55 was important for chronic toxoplasmosis in mice. We injected 3 separate groups of mice intraperitoneally with 250 parasites per mouse, allowed the infection to proceed to 30 days, and then euthanized the mice and examined tissue cyst burden in the brains. We found that mice infected with Δgra55 parasites suffered a consistent 10-fold decrease in cyst burden compared to the mice infected with wild-type and GRA55c parasites (Fig 4H and  4I). In summary, this data demonstrated that although GRA55 is not important for the acute infection and mortality, it is an important secreted protein for the establishment or maintenance of the chronic infection in mice.

Discussion
Toxoplasma bradyzoites are characterized by slow growing parasites encased in a cyst wall that enables transmission of the parasite to new hosts upon predation [25, 52,53]. While the cysts are formed from secreted components from the parasite, the protein composition of the cyst has been challenging to determine as it is difficult to purify robust numbers of cysts and then separate the secreted components from the parasites. In addition, while switching from tachyzoites to bradyzoites can be accomplished in vitro, rapidly growing tachyzoites in the culture routinely overwhelm the slow growing bradyzoites and further complicate purifications. The use of the CST1 monoclonal antibody to purify fragments of the cyst wall and identify the purified components has been successful at identifying many cyst constituents [33]. However, purifying wall fragments requires relatively low stringency lysis conditions that may result in some contamination in these purifications. The adaptation of proximity biotinylation in T. gondii provides a new approach to label proteins prior to isolation and is ideal for compartments hard to access like the secreted GRAs of the cyst. We were additionally able to improve the in vitro switch efficiency using a strain which disrupted for the bradyzoite repressor AP2IV-4 [38]. While this approach helped us identify new GRAs, we found that challenges still remain with significant tachyzoite contamination, particularly at the high multiplicity of infection which we used in these experiments to minimize host biotinylated protein contamination in our streptavidin pulldowns. The recent identification of a master regulator of bradyzoite differentiation (BFD1) is likely to enable more robust cyst production without the stress of in vitro stage switching culture conditions and may yield a deeper insight into the cyst composition using proximity labeling [54]. As expected, our proximity labeling experiments with MAG1 as a bait protein identified many of the same proteins we found in the tachyzoite proteome. While MAG1 is one of the earliest described bradyzoite-enriched GRAs that is abundant in cysts, similar to CST1 there is also substantial MAG1 expression in tachyzoites (both proteins ranked high in our GRA--BioID of tachyzoites) [36]. Regardless, we were able to find many of the known bradyzoite upregulated proteins including CST1, BPK1, and the newly discovered CST proteins CST2, CST3, CST4, and CST5 [30,33,34]. We also found a group of candidate GRAs, five of which we verified by endogenous gene tagging. Interestingly, each of these new GRA proteins could be clearly detected in tachyzoites. This agrees with CST2, CST3, CST 4, and CST6 recently identified in the CST1 pulldown experiments which can also be observed in tachyzoites [33]. This may be due to the in vitro bradyzoite switching conditions used in these experiments and could indicate that these proteins are more important for transitioning to the bradyzoite stage rather than maintenance of mature cysts. In agreement with in vitro cysts not containing the full complement of proteins present in in vivo cysts, neither this study nor the CST1 pulldown identified BCP1, which has only been identified in the cyst wall of in vivo derived cysts [55]. It will be interesting to determine if cysts induced via BFD1 more faithfully replicate in vivo cysts, and thereby enable new categories of bradyzoite GRAs to be obtained, or if these are still lacking important characteristics of cysts isolated from mice.
In agreement with the T. gondii genome-wide CRISPR screen, we were able to disrupt all five of the new GRAs that we identified here [56]. While gene knockouts using CRISPR/Cas9 are generally efficient in Δku80 strains of Toxoplasma, we have found that replacing the Toxoplasma promoter driving HXGPRT with the Neospora GRA7 promoter improves the frequency of knockouts, frequently reaching 100% of the transfected population.
Most GRA proteins lack any identifiable homology to known eukaryotic proteins. GRA56, which we identified in this study, is a large protein with a domain that has homology to the melibiase family proteins. Melibiase proteins are alpha-galactosidases first discovered in plants, then found in several bacteria, animal tissues and fungi [57][58][59][60][61][62][63][64][65]. The enzyme acts to break apart melibiose, a disaccharide composed of galactose and glucose by a 1,6-alpha bond. In many fungi including Saccharomyces species, this enzyme is a large glycoprotein which contains an N-terminal signal sequence and is secreted outside of the cell, similar to GRA56 identified here [66][67][68]. We hypothesize that T. gondii may use this secreted enzyme to catabolize melibiose for energy utilization after uptake into the PV. While GRA56 does have a negative (-1.13) fitness score in the genome-wide CRISPR screen [56], it is possible that this knockout has compensated by upregulating other pathways as has previously been observed in Toxoplasma [69].
Our pilot in vivo assays suggested that disruption of GRA55 resulted in a defect in cyst counts, thus we complemented the strain with the wild-type gene. We found that the knockout showed no defect in plaque size or in acute virulence in vivo, but the chronic infection resulted in a~10 fold decrease in cyst number in vivo, which was rescued by complementation. We did not observe any gross change is cyst size or morphology, and did not observe fragility of the cysts while manipulating the samples for counting as has been seen in the CST1 deletion mutants [31]. We performed an in vitro examination of switching efficiency between GRA55-HA, Δgra55, and GRA55c parasites. Although there was a general trend of increased efficiency from GRA55-HA to Δgra55 parasites, we found that there was no significantly increased switch efficiency between the three strains. Further, complementation of GRA55 did not reverse the trend (S3 Fig). This suggests that the defect may be in tissue tropism to the brain or in the formation or maintenance of cysts in vivo, although we cannot exclude the possibility that subtle changes in growth in vivo over the 30 day time period used for cyst development play a role in the decreased cyst count in GRA55 knockouts. Since GRA55 lacks homology to known proteins, the mechanism of action will be best be determined by mutagenesis experiments to identify functional domains and additional proximity biotinylation experiments to identify binding partners and regulators. Together, this research improves our understanding of the composition and function of GRA proteins in bradyzoites which will ultimately lead to new therapies to treat this poorly understood yet critical stage of infection.

Host cell and parasite cultures
PruΔku80Δhxgprt (Type II, strain Prugniaud) and modified strains of T. gondii were used to infect human foreskin fibroblast (HFF) cells. HFFs were grown in Dulbecco's Modified Eagle's Medium supplemented with 10% fetal bovine serum and 2 mM glutamine, and maintained as previously described [70].

IFA and Western blot
For IFA, HFFs were grown to confluence on coverslips and infected with T. gondii parasites. After 18 to 36 h, the coverslips were fixed and processed for indirect immunofluorescence as previously described [71]. Primary antibodies were detected by species-specific secondary antibodies conjugated to Alexa 594/488. The coverslips were mounted in ProLong Gold antifade reagent (ThermoFisher) and viewed with an Axio Imager Z1 fluorescence microscope (Zeiss) as previously described [41]. For Western blotting, parasites were lysed in Laemmli sample buffer (50 mM Tris-HCl [pH 6.8], 10% glycerol, 2% SDS, 1% 2-mercaptoethanol, 0.1% bromophenol blue), and lysates were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. Blots were probed with the indicated primary antibodies, followed by secondary antibodies conjugated to horseradish peroxidase (HRP). Target proteins were visualized by chemiluminescence.

Generation of MAG1-BirA � fusion
To generate the MAG1-BirA � fusion, the MAG1 gene and promoter was PCR amplified from genomic DNA (primers: p5-p6 -S2 Table). This was cloned into the p.LIC.BirA � .3xHA.DHFR vector by ligation independent cloning as previously described [35,72]. 30 μg of the construct was linearized with HpaI and transfected into PruΔku80Δhxgprt::Δap2IV-4.hxgrprt parasites. The parasites were selected with medium containing 1 μM pyrimethamine, cloned by limiting dilution, and screened by IFA and Western blot against the HA tag. A clone expressing the fusion protein was selected and designated MAG1-BirA � :Δap2IV-4.hxgrprt.

Affinity capture of biotinylated proteins
HFF monolayers were infected with parasites expressing MAG1-BirA � fusions or the parental line (PruΔku80Δhxgprt:::Δap2IV-4.hxgrprt) for 3h. Media was changed to bradyzoite "switch" media (DMEM with 1%FBS, 50mM HEPES, 1%PSG, pH 8.1). [35,72]. Parasites were incubated at 37˚C in ambient air (0.05% CO2) for 5 days to induce switch to tissue cysts, exchanging media every 2 days to maintain alkaline pH [33,43]. 150 μM biotin was supplemented in the media on the final 24 hrs. Cyst formation was verified by expression of LDH2-GFP. Intracellular parasites were collected, washed in PBS, and lysed in radioimmunoprecipitation assay (RIPA) buffer (50mM Tris pH 7.5, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-40) supplemented with Complete Protease Inhibitor Cocktail (Roche) for 30 min on ice. Lysates were centrifuged for 15 min at 14,000 × g to pellet insoluble material and the supernatant was incubated with Streptavidin Plus UltraLink Resin (Pierce) at room temperature for 4 h under gentle agitation. Beads were collected and washed five times in RIPA buffer, followed by three washes in 8M urea buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl). Ten percent of each sample was boiled in Laemmli sample buffer, and eluted proteins were analyzed by Western blotting by streptavidin-HRP while the remainder was used for mass spectrometry.

Mass spectrometry
The proteins bound to streptavidin beads were reduced and alkylated via sequential 20 minute incubations of 5mM TCEP 10mM iodoacetamide at room temperature in the dark while being mixed at 1200 rpm an Eppendorf thermomixer. Proteins were then digested by the addition of 0.1μg Lys-C (FUJIFILM Wako Pure Chemical Corporation, 125-05061) and 0.8μg Trypsin (Thermo Scientific, 90057) while shaking 37˚C overnight. The digestions were quenched via addition of formic acid to a final concentration of 5% by volume. Each sample was desalted via C18 tips (Thermo Scientific, 87784) and then resuspended in 15μL of 5% formic acid before analysis by LC-MS/MS. Peptide samples were then separated on a 75μM ID, 25cm C18 column packed with 1.9μM C18 particles (Dr. Maisch GmbH) using on a 140-minute gradient of increasing acetonitrile and eluted directly into a Thermo Orbitrap Fusion Lumos instrument where MS/MS spectra were acquired by Data Dependent Acquisition (DDA). Data analysis was performed using the ProLuCID and DTASelect2 algorithms as implemented in the Integrated Proteomics Pipeline-IP2 (Integrated Proteomics Applications, Inc., San Diego, CA). Protein and peptide identifications were filtered using DTASelect and required a minimum of two unique peptides per protein and a peptide-level false positive rate of less than 1% as estimated by a decoy database strategy.
The transfected parasites were grown in media containing 1 μM pyrimethamine and selected parasites were cloned by limiting dilution. Clones were designated as GRA(X)-HA (e.g. GRA55-HA). All confirmed GRAs were examined in silico by BLAST to search for other T. gondii proteins with sequence similarity.

Complementation of GRA55
GRA55-HA was re-introduced into the genome of the Δgra55 line in the uracil phosphoribosyltransferase (UPRT) locus [51]. This was done by amplifying the GRA55-3xHA gene as well as its native promoter from genomic DNA of the GRA55-HA parasite line (p13-p14) (S2 Table). The PCR product was cloned into the UPRT knockout vector with a tubulin promoter added using PacI and SpeI, forming a construct with the GRA55-3xHA cassette inserted between the UPRT 5' and 3' flanking regions. This plasmid was linearized using SacI-HF and transfected into Δgra55 parasites and selected with 5-Fluoro-5'-deoxyuridine (FUDR) to isolate parasites with the construct integrated at the UPRT locus [76]. Clones were selected by limiting dilution, examined by IFA (anti-HA antibody), and a positive clone was designated GRA55c.

Plaque assay
Serial dilutions of parasites were infected into separate wells of a 6-well plate with an HFF monolayer and allowed to form plaques. The dilutions were calibrated to infect 100 parasites/ well, 200 parasites/well and 300 parasites/well. 6 days after infection, HFF monolayers were fixed in 3.7% formaldehyde for 15 minutes, washed with PBS, dried and stained with safranin. Plaques were then visualized with a Zeiss upright light microscope (Zeiss Axio Imager Z1). Plaque areas were measured using ZEN software (Zeiss) and plaque sizes were compared. 2 separate experiments of 30 plaques for each parasite line were measured to generate a mean plaque size along with interquartile ranges. Statistical significance comparing lines was calculated using two-sample two-tailed t-test.

Quantitation of in vitro bradyzoite differentiation efficiency in GRA55-HA, Δgra55, and GRA55c parasites
HFFs were grown on coverslips for seven days and infected with GRA55-HA, Δgra55 or GRA55c parasites for 3h. Media was changed to bradyzoite "switch" media (DMEM with 1% FBS, 50mM HEPES, 1%PSG, pH 8.1). [35,72] and parasites were incubated at 37˚C in ambient air (0.05% CO 2 ) for 3 days to induce switch to tissue cysts [33,43]. The Coverslips were then fixed for IFA analysis. To quantify bradyzoite switch efficiency, GFP positive and negative vacuoles were counted and percentages of GFP positive vacuoles were calculated. Two biological replicates were reported for each strain. Statistical significance was calculated comparing the means of GRA55-HA with Δgra55 or GRA55c with Δgra55 using a one-way ANOVA test (using p < 0.05 for significance).

Mouse virulence assays
Intracellular GRA55-HA, Δgra55 and GRA55c parasites mechanically liberated from infected HFF monolayers and resuspended in Opti-MEM Reduced Serum Medium (Ther-moFisher) prior to intraperitoneal injection into groups of 4 female C57BL/6 mice each. Injected parasites were confirmed to be live and viable parasites by plaque assays with HFF monolayers. Mice were monitored for symptoms of infection, weight loss, and mortality for 21 days. Survival was plotted on a Kaplan-Meier curve. To confirm infection, surviving mice were sacrificed at 30 days after infection, and mouse brains were collected in aseptic fashion, homogenized, and examined by phase microscopy and fluorescence microscopy (for detection of GFP under LDH2 promoter for parasite encystation). In parallel, 1/3 of each brain was homogenized and incubated with HFF monolayers to qualitatively assess viability of parasites.

Mouse brain cyst quantitation
Intracellular GRA55-HA, Δgra55, and GRA55c parasites were mechanically liberated from infected HFF monolayers and resuspended in Opti-MEM prior to intraperitoneal injection into groups of 4 female CBA/J mice each. 250 parasites/mouse of GRA55-HA, Δgra55 and GRA55c were injected. The mice were monitored for 30 days after infection, and then sacrificed. Mouse brains were collected, homogenized and examined for T. gondii cysts (as above). Quantitation of cysts was performed by examining 25 μL aliquots of homogenate by fluorescence microscopy until approximately 25% (by volume) of each brain was examined. Total cyst burden was then extrapolated. Statistical significance of cyst burden between GRA55-HA, Δgra55, and GRA55c experiments was calculated by two-sample 2-tailed t-test [77].

Animal experimentation ethics statement
Specific details of our protocol were approved by the UCLA Institutional Animal Care and Use Committee, known as the Chancellor's Animal Research Committee (protocol: 2004-005). Mice were euthanized for two purposes. In the acute infection model, they were euthanized when the animals reached a moribund state. For the chronic infection model, these mice were sacrificed at the conclusion of the experiment at 30 days postinfection. Euthanasia per AVMA guidelines-slow (20-30% per minute) displacement of chamber air with compressed CO2. This is followed by confirmatory method of cervical dislocation'.