Phosphorylation of Toxoplasma gondii Secreted Proteins during Acute and Chronic Stages of Infection

Toxoplasma gondii is a common parasite that infects up to one-third of the human population. Initially, the parasite grows rapidly, infecting and destroying cells of the host, but subsequently switches to a slow-growing form and establishes chronic infection. In both stages, the parasite lives within a membrane-bound vacuole within the host cell, but in the chronic stage, a durable cyst wall is synthesized, which provides protection to the parasite during transmission to a new host. Toxoplasma secretes proteins into the vacuole to build its replicative niche, and previous studies identified many of these proteins as phosphorylated. We investigate two secreted proteins and show that a phosphorylated region plays an important role in their regulation in acute stages. We also observed widespread phosphorylation of secreted proteins when parasites convert from acute to chronic stages, providing new insight into how the cyst wall may be dynamically regulated.

date. It is substantially phosphorylated after secretion (20) and contains a localized cluster of phosphorylation sites in its C terminus, allowing for targeted genetic mutagenesis. TGGT1_289540 contains a signal peptide and is predicted to localize to the dense granules (8). It has four predicted coiled-coil domains and an unstructured C terminus that contains 8 phosphorylation sites (Fig. 1A). Three further phosphorylation sites are located within the rest of the protein. To first verify that TGGT1_289540 is a secreted protein, we expressed a myc-tagged version of the protein, which localized to the PV. Interestingly, in contrast to most GRAs that either are found to fill the space between parasites or are found associated with the PV membrane, TGGT1_289540 appeared in strand-or filament-like structures (Fig. 1B). Western blotting confirmed the presence of a single isoform of myc-tagged protein at the predicted size of ϳ100 kDa in transgenic parasites but not in a wild-type (WT) control (Fig. 1C).
To verify this unusual distribution of TGGT1_289540 in the PV, we generated rabbit polyclonal antibodies against the recombinant protein. As was the case with the myc-tagged isoform, analysis of the endogenous protein showed a filamentous distribution in the PV (Fig. 1D). Both the signal detected by immunofluorescence analysis (IFA) and a band recognized by the antibody in Western blots were absent upon the disruption of the TGGT1_289540 locus ( Fig. 1D and E).
As the Toxoplasma PV contains a distinct IVN of membranous tubules, we tested whether the strands were related to these structures. The formation of the IVN is dependent on GRA2, so we assessed TGGT1_289540 localization in RHΔgra2 parasites that lack the IVN (15). No obvious difference in filamentous TGGT1_289540 localization was observed (Fig. 1F), indicating that the filaments are distinct from, and their formation does not depend on, the IVN. We therefore named the protein strandforming protein 1 (SFP1 here).
We hypothesized that SFP1 strands could function in transport between the parasites and the host cell. Toxoplasma endocytoses proteins from the host cell through an unknown mechanism (17), so we assessed whether SFP1 plays a role in the uptake of proteins from the host cell into the parasite at 3 h postinfection. As endocytosed material is rapidly digested, this can be observed only upon inhibition of cathepsin protease L function with the inhibitor morpholinurea-leucine-homophenylalanine-vinyl phenyl sulfone (LHVS) (17). Under these conditions, there was no difference in the proportions of parasites that had taken up Venus fluorescent protein from the host cell, indicating that SFP1 does not contribute to protein uptake (see Fig. S1A in the supplemental material). Toxoplasma upregulates host c-Myc in infected cells, a phenotype which is dependent on protein translocation from the PV into the host cell (11,24). To test whether SFP1 plays a role in protein export into the host cell, we compared the abilities of wild-type and Δsfp1 parasites to upregulate host c-Myc. No differences were observed in host cell c-Myc upregulation, indicating that SFP1 does not play a role in protein export from the PV (Fig. S1B).
It was recently shown that long filaments of actin extend into the PV from the parasites during Toxoplasma infection (25). We therefore tested whether treatment with cytochalasin D, to depolymerize actin, would disrupt SFP1 strand formation. We still saw substantial filamentation of SFP1 in the presence of cytochalasin D, indicating that actin is not important in this process (Fig. 1G).
While the presence of the IVN is not important for SFP1 filament formation, the opposite could be true, that is, that SFP1 may be important for the formation of the IVN. However, IVN formation was not affected in the absence of SFP1, as shown by transmission electron microscopy (EM) (Fig. 1H). No obvious abnormalities in the overall PV organization or PVM structure were observed in RHΔsfp1 parasites.
Collectively, these data indicate that SFP1, a protein that was reported to be phosphorylated after secretion, localizes to the PV and forms novel, IVN-and actinindependent filamentous structures. These structures, as suggested by our data, do not contribute to protein uptake from or protein export into the host cell.
A second SFP1-related protein, GRA29, forms related structures. As secreted protein families can form by gene duplication and diversification, we looked for potential paralogs of SFP1 in the Toxoplasma genome using BLASTP. We identified a single protein (TGGT1_269690 [GRA29]) with 24% identity and a similar architecture ( Fig. 2A and Fig. S2) of a signal peptide, coiled-coil domains, and, like SFP1, C-terminal phosphorylation sites that appear phosphorylated after secretion (20). GRA29 is predicted to localize to the dense granules (8) and has been shown to be secreted into the PV (26). To compare the localization of this protein with that of SFP1, we generated a hemagglutinin (HA)-tagged version, which by IFA localized to the PV, and Western blotting showed the expected size of ϳ94 kDa ( Fig. 2B and C). In contrast to SFP1, however, we saw smaller filamentous structures and also large puncta associated with the parasites (Fig. 2B), consistent with the localization reported by Nadipuram et al. (26). To investigate whether SFP1 and GRA29 were within the same structures, we performed superresolution microscopy using anti-SFP1 antibodies and anti-HA antibodies to localize GRA29. At this improved resolution, GRA29 puncta appeared as round spheres or doughnuts, often with SFP1 filaments associated with or branching out from them (Fig. 2D). We also observed filamentous SFP1 with smaller dots and strands of GRA29. To further investigate these unusual sphere-like structures, we performed correlative light electron microscopy (CLEM) to identify the subcellular structures. This revealed that GRA29::HA accumulates in electron-dense particles in the PV that appear to contain no membrane and resemble proteinaceous aggregates (Fig. S3).
To verify the GRA29::HA structures, we raised polyclonal antibodies against recombinant GRA29. With the antibodies against endogenous protein, we observed only short strands and small puncta in the PV that did not resemble the HA-tagged parasite line ( Fig. 2E and F). The specificity of the antibodies was verified by Western blotting and IFA, which showed the expected molecular weight and positive IFA signal in WT but not in GRA29 knockout (KO) parasites ( Fig. 2F and G). This indicates that the addition of the C-terminal HA tag disrupts GRA29 localization, inducing protein aggregates, suggesting a potentially important role of the C-terminal region. In order to colocalize endogenous nontagged SFP1 and GRA29, we generated antibodies against SFP1 peptides in mice, which showed the expected reactivity in IFAs and Western blots ( Fig. 2H and I). The colocalization of SFP1 and GRA29 revealed that both proteins appear in similar structures, with GRA29 often being present at either end of SFP1 filaments (Fig. 2I).
Collectively, these data show that SFP1 and GRA29 form novel filamentous structures in the PV. C-terminal tagging of GRA29 alters its localization, either causing or stabilizing unusual structures that SFP1 can still associate with.
GRA29 regulates SFP1 strand formation even in the absence of other Toxoplasma proteins. As SFP1 and GRA29 appeared to localize within the same structures, we hypothesized that they may show an interdependence. To address this, we localized SFP1 in the RHΔgra29 parasite line and GRA29 in the RHΔsfp1 parasite line using the specific antibodies raised. Whereas no alteration in GRA29 was observed in RHΔsfp1, SFP1 formed fewer and longer filaments in the absence of GRA29 (Fig. 3A). To identify whether any other Toxoplasma proteins are required for (i) SFP1 filament formation and (ii) SFP1 filament initiation, we expressed both SFP1 and GRA29, lacking their respective signal peptides, in human foreskin fibroblasts (HFFs). Both SFP1 and GRA29 formed long filamentous structures in HFFs (Fig. 3B), often filling the complete cytoplasm and, in the case of GRA29, sometimes associated with the nucleus. When coexpressed in HFFs, GRA29 frequently formed smaller puncta and short strands, likely due to the reduced expression levels of each protein. Strikingly, SFP1 filaments radiated from GRA29 structures, suggesting that GRA29 can initiate SFP1 strand formation (Fig. 3C).
These data indicate that both SFP1 and GRA29 can each form filament-like structures independently in the absence of other Toxoplasma proteins. Although we were unable to confirm the interaction by coimmunoprecipitation from infected cells or with recombinant proteins (data not shown), the colocalization of SFP1 and GRA29 even in this nonphysiological context suggests that they can directly interact. Furthermore, the localization pattern observed in HFFs and the extended SFP1 filaments in parasites lacking GRA29 suggest that GRA29 regulates SFP1 strands, perhaps initiating strand formation or controlling their length.
The SFP1/GRA29 C-terminal tail is required for strand formation. The change of the localization of GRA29 upon C-terminal HA tagging and the cluster of phosphorylation sites in the C termini of SFP1 and GRA29 suggested that the C-terminal tail may play an important role in the regulation of strand formation. To address this, we expressed truncated SFP1 and GRA29 lacking the unstructured C-terminal region in HFFs. In contrast to forming filaments under these conditions, both proteins appeared distributed throughout the HFF cytoplasm, with sparse foci of aggregation observed (Fig. 4A). This strongly suggested that the C terminus of the protein plays an important role in filament formation. Consistently, during Toxoplasma infection with RHΔsfp1ϩSFP1ΔCt, there was a lack of typical SFP1 strand formation in the PV (Fig. 4B). However, SFP1ΔCt appeared aggregated within the PV rather than dispersed, suggesting that other PV-resident proteins impact SFP1 distribution.
Having established that the C terminus of SFP1 is important for its subcellular organization, we asked whether the phosphorylation of SFP1 is important for filament formation. To investigate this, we first turned to the expression of phosphomutants in HFFs, in which all phosphorylation sites in the SFP1 C terminus were mutated to alanine (SFP1_ALA), and phosphomimetics, in which all phosphorylation sites were mutated to glutamic acid (SFP1_GLU). While SFP_ALA mutants displayed normal filaments in HFFs, mimicking phosphorylation led to an inability of SFP1 to form filaments. This indicates that phosphorylation may be a negative regulator of filament formation (Fig. 4C). This was replicated when SFP1_ALA and SFP1_GLU were expressed in RHΔsfp1 parasites (Fig. 4D), where extended strands were observed with phosphomutant but not phosphomimetic mutations. If phosphorylation was preventing filament formation, we hypothesized that SFP1 expressed in HFFs should not be phosphorylated. To verify that assumption, we immunoprecipitated SFP1 from HFFs and analyzed the proteins by mass spectrometry (MS). Despite the presence of many nonmodified peptides from SFP1, we observed no phosphopeptides (data not shown), indicating that, indeed, SFP1 is not phosphorylated by human kinases when expressed in HFFs.
Collectively, these data show that the C termini of SFP1 and GRA29 are important for their regulation and that the phosphorylation of SFP1 appears to be a negative regulator, with the nonphosphorylated protein forming strands. SFP1 and GRA29 are dispersed and do not form strands in chronic-stage cysts. To determine whether SFP1 and GRA29 are required in vivo, we generated gene KOs in the type II strain Pru and verified gene disruption by Western blotting (Fig. S4A). Mice infected with PruΔsfp1 and PruΔgra29 parasites succumbed to infection similarly to WT PruΔku80 parasites, indicating that neither protein is essential for parasite survival in vivo ( Fig. S4B and C). During our experiments with the Pru strains, we observed that in some vacuoles, SFP1 and GRA29 appeared distributed throughout the PV space instead of forming characteristic strands (Fig. 5A). Type II strains differ from the type I RH strain in that they more frequently convert to slow-growing bradyzoites in tissue culture. Bradyzoite cysts can be distinguished from tachyzoite vacuoles using fluorescently labeled lectin (Dolichos biflorus agglutinin [DBA]) that binds to the cyst wall (27). Parasites with a disperse SFP1/GRA29 pattern also stained positive for the cyst wall marker (Fig. 5A). Spontaneous conversion to bradyzoites occurs at low levels in standard tissue culture, but high levels of conversion can be induced by pH stress (28). Under bradyzoite-inducing conditions, parasite vacuoles were positive for the cyst wall marker CST1 and showed disperse SFP1 and GRA29, indicating that both proteins undergo a major change in localization between the tachyzoite and bradyzoite stages (Fig. 5B).
Secreted proteins are differentially phosphorylated in the chronic stage. We hypothesized that increased phosphorylation of SFP1 and GRA29 under bradyzoite conditions could lead to their dispersion. This phenomenon of redistribution in the bradyzoite cyst has been described previously for other GRAs (22,29), and so we wanted to determine if these were also differentially phosphorylated. We therefore used comparative phosphoproteomics to determine differences in the phosphorylation of SFP1 and GRA29 and, more broadly, between acute (tachyzoite) and chronic (bradyzoite) conditions. Triplicate samples of Toxoplasma parasites were grown under tachyzoite conditions for 27 h or bradyzoite conditions for 3 days (to give comparable numbers of parasites per vacuole) (Fig. 6A). We performed quantitative mass spectrometry to compare protein and phosphosite abundances under the two conditions using tandem mass tags (TMTs) as described previously (30,31). First, we analyzed differential Peptides from triplicate tachyzoite and bradyzoite samples were labeled with tandem mass tags (TMTs), enriched for phosphopeptides, and analyzed by mass spectrometry. (B) Volcano plot showing differential phosphorylation between tachyzoite and bradyzoite conditions after normalization to protein abundance. Each circle represents a phosphosite, and those that change significantly (FDR, 1%; Ϫ0.9 Ͼ log 2 -fold change Ͼ 0.9) are shown in green. (C) Dense granule proteins are enriched in the subset of proteins that show significantly changing phosphorylation between tachyzoites and bradyzoites. The chart shows the predicted localization of phosphoproteins using LOPIT (8). PM, plasma membrane; IMC, inner membrane complex; ER, endoplasmic reticulum. protein abundances and identified 171 differentially regulated proteins, with 103 proteins with higher abundances in bradyzoites and 68 with higher abundances in tachyzoites. As expected, known bradyzoite markers such as BAG1, ENO1, LDH2, and MAG1 were more abundant in the bradyzoite samples than in the tachyzoite samples (Data Set S1 and Fig. S5).
In the phosphopeptide-enriched samples, we quantified 7,650 phosphorylation sites on 2,235 Toxoplasma proteins. For 3,730 of the phosphorylation sites, we also obtained quantitative proteome data to which phosphorylation site changes were normalized. This allowed us to assess true differential phosphorylation rather than changes originating from differential protein abundances between conditions. After this normalization step, we identified 337 phosphorylation sites that were significantly different between tachyzoite (144 sites more phosphorylated) and bradyzoite (193 sites more phosphorylated) stages ( Fig. 6B and Data Set S1). The 337 phosphorylation sites are found on 170 proteins, 51 of which are predicted to be secreted (ROPs or GRAs in the LOPIT data set [8]). While GRAs made up 11% of proteins with nonchanging phosphosites, they represented 26% of differentially phosphorylated proteins (Fig. 6C). This enrichment of GRAs in the subset of proteins that are differentially phosphorylated between tachyzoite and bradyzoite conditions indicates that these are indeed subject to extensive differential phosphorylation during stage conversion.
In this experiment, we also identified phosphosites corresponding to the C termini of SFP1 and GRA29, but they did not significantly change between conditions (although a previously undetected N-terminal site on GRA29 was detected as more phosphorylated in bradyzoites). This indicates that while mimicking phosphorylation disrupts SFP1 strands in the PV in tachyzoites, this modification is not sufficient for the dispersion observed in the forming cyst. However, the GRAs that were detected as differentially phosphorylated included IVN-localized GRA2, GRA4, GRA6, and GRA12 and PV membrane GRA1 and GRA5 (Table 1). We also detected differential phosphorylation of the cyst wall proteins CST1, CST3, CST4, and CST6. Furthermore, of the 38 recently identified putative cyst wall components (32), 24 were differentially phosphorylated, suggesting that phosphorylation could play an important role in the formation of this structure.
It is worth noting that we also observed differential phosphorylation on proteins important for DNA regulation. This includes the AP2 domain transcription factors AP2VIII-2 and AP2XII-4 and 32 predicted nuclear proteins. These may be important for transcriptional regulation during stage conversion, but their role is not further discussed here.
Collectively, these data show that SFP1 and GRA29 undergo a dramatic change in localization between tachyzoite and bradyzoite conversion but that their phosphorylation is transient or not required for this transition, at least at the experimental time point selected here. Furthermore, we show widespread changes in phosphorylation states in a substantial number of secreted proteins, indicating the activity of stagespecific kinases and phosphatases in regulating PV-resident proteins.

DISCUSSION
Here, we have analyzed the localization of two secreted phosphoproteins, SFP1 and GRA29, that relocalize during stage conversion to bradyzoites. Both proteins form stranded structures in the tachyzoite PV that are distinct from the IVN and actin filaments previously observed. We did not identify a function of SFP1 and GRA29 during the lytic cycle or in vivo infections, preventing us from assessing the impact of their phosphorylation on Toxoplasma biology. It could be that the proteins have a subtle effect on growth that we have not been able to measure or that they are important in another host species, parasite stage, or genetic background that we have not tested here. However, neither SFP1 nor GRA29 showed a reduction in fitness in vivo (33) in our recent more sensitive in vivo CRISPR screen, supporting a nonessential role under these conditions.
We show the importance of the C termini of both proteins for strand formation.  Using mutational analysis, we show that SFP1 phosphorylation is not required for filament formation and that mimicking phosphorylation disrupts filament formation. As such, mimicking phosphorylation mirrored the phenotype observed with the removal of the C-terminal tail. Phosphorylation negatively regulates the intrinsic behavior of SFP1 and abrogates its ability to form strands. However, as both SFP1 and GRA29 were identified to be phosphorylated after secretion into the PV (20), where, as we show here, they form fibril-like structures, we assume that their phosphorylation is a dynamic process, potentially leading to the regulation of their length and location. It could also be that phosphorylated and nonphosphorylated SFP1 and GRA29 localize to different structures in the PV. How both proteins relocalize during stage conversion is unclear.
While mimicking phosphorylation appears sufficient to promote the dispersal of SFP1 HFFs, this is not observed when expressed in tachyzoites, where aggregation is observed. This indicates that other proteins or posttranslational modifications are involved in SFP1 regulation in parasites upon stage conversion.
Recently, the PV-localized kinase WNG1/ROP35 was shown to phosphorylate several GRAs, including GRA6, and to regulate their membrane association (21). As many GRAs form complexes during trafficking (34), WNG1 phosphorylation is hypothesized to release them from association with a chaperone, freeing them to associate with membranes. As such, WNG1-dependent phosphorylation of GRAs was shown to promote their normal behavior and localization. We now propose that in addition to WNG1 phosphorylation promoting the release and membrane association of GRAs, the phosphorylation of secreted proteins can negatively regulate their behavior. In this way, the parasite could use phosphorylation by kinases within the PV lumen to regulate PV-resident proteins. This would allow the parasite to tailor its niche and respond to differing conditions or requirements. Several secreted kinases localize within the PV lumen and are upregulated during the chronic stages of infection (for example, ROP21, ROP27, ROP28, and, indeed, WNG1/ROP35). Importantly, secreted kinases have been shown to contribute to cyst formation in mice, indicating that phosphorylation within the PV/forming cyst is required for cyst development. ROP21/27/28 contribute to the parasite establishing chronic infection, but their function and targets remain unknown (35). Additionally, a triple knockout of ROP38/29/19 showed a severe defect in cyst formation (36). Here, we expand our knowledge of cyst formation by reporting the differential phosphorylation of proteins within the forming cyst. Our data also identify phosphorylation that is reduced upon stage conversion. The expression of many ROP family kinases is reduced in the chronic stage of infection, which could account for some changes, but it is likely that secreted phosphatases are contributing to the differential phosphorylation. The necessity of deleting several ROPs to observe strong phenotypic effects on bradyzoites may reflect that the system is flexible with a high degree of redundancy. Alternatively, the combined effect of multiple kinases could allow the parasite to integrate several signals leading to stage differentiation.
PV-resident GRAs relocalize as the parasite converts to the chronic stage and remodels the PV to form a cyst. Guevara et al. recently described the dynamic nature of this relocalization over the course of cyst formation (for GRA1, -2, -4, -6, -9, and -12) (22). Furthermore, they showed that IVN-localized GRAs were required for normal formation of the cyst wall. Here, we show that these GRAs, with the exception of GRA9 (which is upregulated in bradyzoites), are differentially phosphorylated in the acute and chronic stages, suggesting that phosphorylation could play a role in their regulation. Interestingly, GRA2 showed phosphorylation sites that were enriched under each condition; two sites within a predicted amphipathic helix were more phosphorylated under tachyzoite conditions, while a third site was more phosphorylated in the bradyzoite samples. It is tempting to speculate that dual phosphorylation within the polar side of the amphipathic helix (a1) of GRA2 could alter its membrane binding. Further work assessing the phosphorylation of secreted proteins over time is required to decipher the impact of this modification and the role of both kinases and phosphatases in remodeling the PV. However, the regulation of proteins after parasite secretion adds an intriguing extra layer of complexity to pathogen effector biology and the establishment of Toxoplasma chronic infection.
Sequence analysis. Gene sequences and predicted protein localizations (LOPIT data set) were obtained from ToxoDB. Alignments were generated using ClustalW, and schematics were generated using DOG2.0 (41). Protein sequences were analyzed by Multicoil2 (42) and Globplot (43) to identify coiled-coil and unstructured regions, respectively.
Plasmid construction. All primers are shown in Table S1 in the supplemental material. Vectors were generated by inverse PCR using KOD hot-start polymerase (Novagen) or by Gibson  ). Gibson reaction mixtures were incubated for 60 min at 50°C. All vectors were verified by sequencing. sfp1 and gra29 sequences were amplified from Toxoplasma genomic DNA (gDNA) prepared from 1E7 RHΔku80::diCre parasites using a Qiagen blood and tissue DNA extraction kit, unless otherwise stated.
To generate pSFP1, the TGGT1_289540 promoter and coding sequence were amplified using primers 1 and 2 and inserted into HindIII/PacI-digested pGRA (13). A C-terminal myc tag was inserted by inverse PCR using primers 3 and 4. pSFP1ΔCterm, which lacks residues 863 to 937, was generated by inverse PCR on pSFP1 with primers 4 and 5. To generate phosphomutant versions of pSFP1, initially, a silent ApaI site was inserted into pSFP1 (at nucleotide [nt] 2529) by inverse PCR with primers 6 and 7. Four-hundredbase-pair synthetic DNA fragments (IDT) (Text S1) encoding the SFP1-myc C terminus or versions with Ser/Thr-to-Ala or Ser/Thr-to-Glu mutations were mixed with ApaI/PacI-digested pSFP1 ApaI in Gibson reaction mix. The sites mutated to Ala or Glu were S862, S868, S871, S887, S891, S897, S906, S913, S924, and T933. pGRA29-HA was generated by amplifying the GRA29 promoter and coding sequence using primers 8 and 9 and inserting the amplicon into HindIII/NcoI-digested pGRA.
For pcDNA-NTAP-Flag SFP1 constructs, the SFP1 coding sequence excluding the signal peptide (amino acids 1 to 53) was amplified with primers 10 and 11 for full-length or primers 10 and 12 for C-terminally truncated (lacking amino acids 863 to 937) versions. For phosphosite mutant versions, primers 10 and 11 were used with pSFP1_ALA or pSFP1_GLU as the template. All amplicons were inserted into NotI/EcoRI-digested pcDNA-NTAP, a pcDNA3.1 derivative for expression with a 3ϫ Flag tag, a tobacco etch virus (TEV) cleavage site, and a calmodulin binding peptide. For the pcDNA-HA-GRA29 construct, GRA29 lacking its predicted signal peptide (residues 1 to 25) was initially amplified with primers 13 and 14 and inserted into NotI/EcoRI-digested pcDNA-NTAP. The Flag tag was removed, and an N-terminal HA tag was inserted by inverse PCR with primers 15 and 16. pcDNA-GRA29ΔCterm, lacking residues 771 to 865, was generated by inverse PCR with primers 17 and 18.
Parasite transfection. Parasites were released from HFFs by syringe lysis, pelleted at 650 ϫ g, and resuspended in cytomix (10 mM KPO 4 , 120 mM KCl, 0.15 mM CaCl 2 , 5 mM MgCl 2 , 25 mM HEPES, 2 mM EDTA). Approximately 5E6 parasites were combined with 40 g DNA in 400 l cytomix in a 2-mm cuvette and electroporated using a Bio-Rad Gene Pulser Xcell instrument at 1,500 V and 25 F. Repair templates were linearized by overnight digestion, purified by phenol-chloroform precipitation, and resuspended in H 2 O. At 24 h posttransfection, 50 g/ml mycophenolic acid (Merck) and xanthine (Sigma) (M/X) was added to select for transfectants.
Generation of parasite lines. RHΔku80 lacking SFP1 was generated by transfecting parasites with pSAG1-CAS9gSFP1 targeting the 5= end of the gene (starting 84 nt after ATG) with a 183-nt repair template to remove the protospacer-adjacent motif (PAM) site and introduce stop codons (Text S1). Parasites expressing the transfected Cas9-green fluorescent protein (GFP) were sorted by flow cytometry into 96-well plates containing HFF monolayers using a BD Influx cell sorter (BD Biosciences). This line was used for complementing pSFP1myc constructs. Alternatively, gra29 and sfp1 were disrupted in the RHΔku80 and PruΔku80 backgrounds by cotransfecting a double gRNA pCAS9 vector targeting the 5= and 3= ends of the gene and a synthetic repair construct with 500-nt homology arms and a hypoxanthine-xanthine-guanine phosphoribosyl transferase (HXGPRT)-thosea asigna virus 2A (T2A)-mCherry cassette (GeneArt; Life Technologies). Transfectants were selected by the addition of M/X, and clones were screened by immunofluorescence and verified by Western blotting.
Antibody generation. N-terminally 6ϫHis-tagged SFP1 recombinant protein was expressed from pET-28a-SFP1 in Escherichia coli BL21 cells. His-SFP1 was purified using Ni-nitrilotriacetic acid (NTA) affinity purification under native conditions according to the standard manufacturer's protocol (Qiagen). Recombinant His-SFP1 was used to immunize female New Zealand White rabbits (Covalab) for the generation of polyclonal antibodies. To generate the mouse anti-SFP1 antibody, phosphopeptides corresponding to the C terminus of SFP1 were conjugated to keyhole limpet hemocyanin and used to immunize 4 mice (Covalab). Day 53 sera showed reactivity against SFP1. While mice were immunized with phosphorylated isoforms, the antibodies recognize nonphosphorylated peptides.
Parasite infections. For wide-field, confocal, or EM imaging, HFFs were seeded on 13-mm no. 1.5 coverslips in 24-well plates. Confluent monolayers were infected at a multiplicity of infection (MOI) of 1 for 16 to 24 h. For cytochalasin D treatment, infected cells were incubated with 1 M cytochalasin D or dimethyl sulfoxide (DMSO) for 1 h before fixation. Samples were washed with PBS before fixation in 3% formaldehyde (FA) for 15 min. For EM analysis, infected cells were fixed in 4% FA for 15 min followed by 2.5% glutaraldehyde-4% FA for 30 min and storage in 1% FA.
Conversion to bradyzoites. To induce parasite switching to bradyzoites, HFF monolayers were infected at an MOI of 0.5 for 3.5 h and then washed four times with PBS before the addition of switch medium (RPMI 1640 [catalog no. R4130; Sigma] with 1% FBS [pH 8.1]) and incubation at 37°C with ambient CO 2 for 3 days. Cells were fixed in 3% FA, quenched in 50 mM ammonium chloride, and permeabilized in a solution containing 0.2% Triton X-100, 0.1% glycine, and 0.2% bovine serum albumin (BSA) in PBS for 20 min on ice.
Intracellular fluorescent-protein ingestion assay. Intracellular fluorescent-protein ingestion was assessed as described previously by McGovern et al. (45). Briefly, 70 to 80% confluent CHO᎑K1 cells were transfected in 35-mm dishes using X-TremeGENE 9 transfection reagent (catalog no. 6365787001; Roche). Each dish was transfected with 2 g of the pcDNA 3.1 Venus plasmid using a 3:1 ratio of plasmid to transfection reagent in Opti᎑MEM (catalog no. 31985062; Gibco) and a total final volume of 200 l. The CHO K1 cells were then incubated overnight at 37°C and synchronously infected with T. gondii parasites using the Endo buffer method at 18 to 24 h posttransfection (46). T. gondii parasites were treated with 1 M LHVS or an equal volume of DMSO for 48 h prior to synchronous invasion, replacing the LHVS-and DMSO-treated media every 6 to 18 h, and also during the 3-h incubation period after synchronous invasion. T. gondii parasites were purified at 3 h postinvasion as previously described (17), pelleted by centrifugation at 1,000 ϫ g for 10 min at 4°C, and subjected to a protease protection assay by resuspension in 250 l of freshly prepared 1 mg/ml pronase (catalog no. 10165921001; Roche)-0.01% saponin-PBS at 12°C for 1 h. The parasites were then pelleted by centrifugation at 1,000 ϫ g for 10 min and washed 3 times in ice-cold PBS before depositing them on Cell᎑Tak (catalog no. 354240; Corning)coated chamber slides. Parasites were fixed with 4% paraformaldehyde and imaged with an AxioCam MRm camera᎑equipped Zeiss Axiovert Observer Z1 inverted fluorescence microscope. Ingestion of host cytosolic Venus was scored manually as Venus positive or Venus negative without sample blinding. Superresolution microscopy. For superresolution microscopy, 24-by 24-mm no. 1.5H coverslips were washed in 1 M HCl overnight and rinsed in water before coating with 1% (wt/vol) porcine gelatin (catalog no. G1890; Sigma) in 6-well plates. Confluent HFF monolayers were infected at an MOI of 0.5 and centrifuged at 180 ϫ g for 3 min to synchronize infection. Samples were fixed and processed as described above except that antibody dilutions were passed through 0.45-m filters and an additional final fixation step (10 min, 2% FA) was included after secondary antibody incubation. Superresolution microscopy imaging was performed using an OMX v3 structured illumination microscope (Applied Precision/GE Healthcare) equipped with 405-, 488-, and 592.5-nm diode lasers; electron-multiplying charge-coupleddevice (CCD) cameras (Cascade II 512 by 512; Photometrics); and a 100ϫ/1.40-numerical-aperture (NA) PlanApochromat oil immersion lens objective (Olympus). Three-dimensional structured illumination microscopy (3D-SIM) image stacks were reconstructed and aligned using softWoRx (Applied Precision/GE Healthcare).
Correlative light and electron microscopy. HFFs were seeded on a 3.5-cm dish with a gridded coverslip (MatTek) and infected, fixed, and processed as described above. Regions of interest with GRA29-HA cage structures were selected and imaged by SIM. Tiled images of the surrounding area were generated by confocal imaging to aid in identifying the region by EM. The cells were then fixed and prepared for EM as described below. The light and electron microscopy overlay was generated by using the TurboReg plug-in of Fiji software, with an affine transformation using the parasite nuclei and apicoplasts as landmarks, and by selecting high-intensity pixels in Photoshop using the magic wand tool with tolerance set to 40.
EM sample preparation and imaging. Cells were fixed in a solution containing 2.5% glutaraldehyde and 4% formaldehyde in 0.1 M phosphate buffer (PB) at room temperature (RT) for 30 min, postfixed in 1% reduced osmium tetroxide for 1 h, stained with 1% tannic acid in 0.05 M PB for 45 min, and quenched with 1% sodium sulfate in 0.05 M PB for 5 min. For CLEM experiments, the gridded coverslip was then removed from the MatTek dish. After washing in PB and water, the cells were dehydrated in 70, 90, and 100% ethanol (twice for 5 min each) before infiltration with a 1:1 mix of propylene oxide and Epon resin (Taab Embed 812) for 1 h and then twice for 90 min with Epon resin. Finally, coverslips were embedded in resin for 24 h at 60°C, sectioned (70 to 75 nm), and stained with lead citrate. Images were acquired using a Tecnai Spirit Biotwin transmission electron microscope (Thermo Fisher Scientific).
In vivo infections. C57BL/6 (wild-type) mice were bred and housed under pathogen-free conditions in the biological research facility at The Francis Crick Institute in accordance with the Home Office UK Animals (Scientific Procedures) Act 1986. All work was approved by the UK Home Office (project license PDE274B7D) and The Francis Crick Institute Ethical Review Panel and conformed to European Union directive 2010/63/EU. All mice used in this study were male and between 7 and 9 weeks old.
Mice were infected with PruΔku80, PruΔsfp1, or PruΔgra29 tachyzoites by intraperitoneal (i.p.) injection with either 3E5 or 5E4 parasites in 200 l PBS on day 0 to assess mouse survival. Mice were monitored and weighed regularly for the duration of the experiments.
Mass spectrometry. (i) SFP1 transfection and immunoprecipitation. A total of 2.8 ϫ 10 6 HFFs were seeded into a 15-cm dish 24 h before transfection with 44 g pcDNA-SFP1 using Lipofectamine 3000 (Thermo Fisher) according to the manufacturer's instructions. At 18 h posttransfection, cells were lysed on ice with 2 ml IP buffer (Thermo Scientific) for 30 min, passed through a 23-gauge needle, and clarified by centrifugation. The supernatant was incubated with 60 l of polyclonal SFP1 antibody-coated Fastflow protein G-Sepharose (GE Healthcare). Sepharose bead complexes were washed in IP buffer followed by Tris-buffered saline, and proteins were eluted by boiling in Laemmli buffer.
(ii) In-gel digestion and LC-MS/MS. Eluted proteins were separated by SDS-PAGE until the running front had migrated approximately 2 cm into the gel (12% NuPAGE; Invitrogen) and stained with colloidal Coomassie (InstantBlue; Expedeon). After the excision of 8 horizontal gel slices per lane, proteins were in-gel digested with trypsin (Promega/Pierce) using a Janus liquid-handling system (PerkinElmer). Tryptic peptides were analyzed by liquid chromatography-mass spectrometry (LC-MS) using an Orbitrap Velos mass spectrometer coupled to an Ultimate 3000 ultrahigh-performance liquid chromatography (uHPLC) instrument equipped with an Easy-Spray nanosource (Thermo Fisher Scientific) and acquired in datadependent mode.
(iii) MS data processing and analysis. The data were searched against Toxoplasma gondii and Homo sapiens (both UniProt) databases using the Andromeda search engine. Raw data were processed with MaxQuant (version 1.5.2.8). Cysteine carbamidomethylation was selected as a fixed modification. Methionine oxidation, acetylation of the protein N terminus, and phosphorylation (S, T, and Y) were selected as variable modifications. The enzyme specificity was set to trypsin with a maximum of 2 missed cleavages. The data sets were filtered on posterior error probability to achieve a 1% false-discovery rate (FDR) on the protein, peptide, and site levels. Data were further analyzed with Perseus (version 1.5.0.9). The data were filtered to remove common contaminants and identifications (IDs) originating from reverse decoy sequences and identified only by site. The log 10 values for peptide intensities were then determined, and all SFP1 peptides (58% sequence coverage) were analyzed based on their modification state. No phosphorylated peptides were found.
Tachyzoite versus bradyzoite phosphoproteomes. (i) Infection conditions. Confluent HFF monolayers were grown in 15-cm dishes and infected with PruΔhxgprt. For tachyzoite samples, three dishes were infected with 1.3 ϫ 10 7 parasites and incubated for 27 h. For bradyzoite samples, three plates were infected with 1.1 ϫ 10 7 parasites for 3.5 h before extensive washing and the addition of switch medium (described above). Bradyzoite infection samples were incubated as described above for 3 days, with fresh switch medium added daily.
(ii) Cell lysis and protein digestion. Lysis was performed in a solution containing ice-cold 8 M urea, 75 mM NaCl, and 50 mM Tris (pH 8.2) supplemented with protease (complete mini; Roche) and phosphatase (Phos Stop; Roche) inhibitors. Lysis was followed by sonication to reduce sample viscosity (30% duty cycle, with three 30-s bursts, on ice). The protein concentration was measured using a bicinchoninic acid (BCA) protein assay kit (Pierce). Lysates (1 mg each) were subsequently reduced with 5 mM DTT for 30 min at 56°C and alkylated in the dark with 14 mM iodoacetamide for 30 min at RT. Following iodoacetamide quenching with 5 mM DTT for 15 min in the dark, the lysates were diluted with 50 mM ammonium bicarbonate to Ͻ4 M urea and digested with LysC (Promega) for 2 to 3 h at 37°C. The lysates were further diluted with 50 mM ammonium bicarbonate to Ͻ2 M urea and digested with trypsin (Promega) overnight at 37°C. After digestion, samples were acidified with trifluoroacetic acid (TFA) (Thermo Fisher Scientific) to a final concentration of 1% (vol/vol). All insoluble material was removed by centrifugation, and the supernatant was desalted on Sep-Pak C 18 cartridges (Waters).
(iii) TMT labeling. Samples were dissolved at 1 mg/ml in a solution containing 50 mM Na-HEPES (pH 8.5) and 30% (vol/vol) acetonitrile and labeled with the respective TMT reagents (Thermo Fisher Scientific) (2.4 mg reagent/1 mg sample) for 1 h at RT. Labeling was then quenched with 0.3% hydroxylamine for 15 min at RT, and samples were acidified (pH ϳ2) with formic acid. After verification of labeling efficiency via mass spectrometry, the lysates were mixed in a 1:1 ratio, vacuum dried, and desalted on Sep-Pak C 18 cartridges.
(iv) Phosphopeptide enrichment. Desalted and vacuum-dried samples were solubilized in 1 ml of loading buffer (80% acetonitrile, 5% TFA, 1 M glycolic acid) and mixed with 5 mg of TiO 2 beads (Titansphere, 5 m; GL Sciences Japan). Samples were incubated for 10 min with agitation, followed by a 1-min spin at 2,000 ϫ g to pellet the beads. The supernatant was removed and used for a second round of enrichment as explained below. Beads were washed with 150 l loading buffer, followed by two additional washes, the first with 150 l 80% acetonitrile-1% TFA and the second with 150 l 10% acetonitrile-0.2% TFA. After each wash, beads were pelleted by centrifugation (1 min at 2,000 ϫ g), and the supernatant was discarded. Beads were dried in a vacuum centrifuge for 30 min, followed by two elution steps at high pH. For the first elution step, beads were mixed with 100 l of 1% (vol/vol) ammonium hydroxide, and for the second elution step, beads were mixed with 100 l of 5% (vol/vol) ammonium hydroxide. Each time, the beads were incubated for 10 min with agitation and pelleted at 2,000 ϫ g for 1 min. The two eluates were removed following each spin and subsequently pooled before undergoing vacuum drying. The supernatant from the TiO 2 enrichment was desalted on Sep-Pak C 18 cartridges, and the High Select Fe-NTA phosphopeptide enrichment kit (Thermo Fisher Scientific) was used for a second round of enrichment according to the manufacturer's instructions.
(v) Sample fractionation and desalting. Combined TiO 2 and Fe-NTA phosphopeptide eluates (phosphoproteome) as well as 100 g of the postenrichment supernatant (total proteome) were fractionated using the Pierce high-pH reversed-phase kit (Thermo Fisher Scientific) according to the manufacturer's instructions. The resulting fractions, 8 for each phospho-and total proteome, were taken to dryness by vacuum centrifugation and further desalted on a stage tip using Empore C 18 discs (3M). Briefly, each stage tip was packed with one C 18 disc and conditioned with 100 l of 100% methanol, followed by 200 l of 1% TFA. The sample was loaded in 100 l of 1% TFA, washed 3 times with 200 l of 1% TFA, and eluted with 50 l of 50% acetonitrile-5% TFA. The desalted phosphoproteome and total proteome fractions were vacuum dried in preparation for LC-tandem MS (MS/MS) analysis.
(vi) LC-MS/MS and data processing. Samples were resuspended in 0.1% TFA and loaded onto a 50-cm Easy-Spray PepMap column (75-m inner diameter, 2-m particle size; Thermo Fisher Scientific) equipped with an integrated electrospray emitter. Reverse-phase chromatography was performed using the RSLC nano U3000 instrument (Thermo Fisher Scientific) with a binary buffer system (solvent A, 0.1% formic acid and 5% DMSO; solvent B, 80% acetonitrile, 0.1% formic acid, and 5% DMSO) at a flow rate of 250 nl/min. The samples were run on a linear gradient of 5 to 60% solvent B in 150 min, with a total run time of 180 min, including column conditioning. The nanoscale LC (nanoLC) instrument was coupled to an Orbitrap Fusion Lumos mass spectrometer using an Easy-Spray nanosource (Thermo Fisher Scientific). The Orbitrap Fusion Lumos instrument was operated in data-dependent mode using 3 acquisition methods. For the phospho MS2 method, high-energy collisional dissociation (HCD) MS/MS scans (R ϭ 50,000) were acquired after an MS1 survey scan (R ϭ 120,000) using an MS1 target of 4E5 ions and an MS2 target of 2E5 ions. The number of precursor ions selected for fragmentation was determined by the "Top Speed" acquisition algorithm with a cycle time of 3 s and dynamic exclusion of 60 s. The maximum ion injection time utilized for MS2 scans was 86 ms, and the HCD collision energy was set at 38. For the phospho MS3 method, collision-induced dissociation (CID) MS/MS scans (R ϭ 30,000) were acquired after an MS1 survey scan with the same parameters as the ones described above. The MS2 ion target was set at 5E4, with multistage activation of neutral loss (H 3 PO 4 ) enabled. The maximum ion injection time utilized for MS2 scans was 80 ms, and the CID collision energy was set at 35. The HCD MS3 scan (R ϭ 60,000) was performed with synchronous precursor selection enabled to include up to 5 MS2 fragment ions. The ion target was 1E5, the maximum ion injection time was 105 ms, and the HCD collision energy was set at 65. The proteome MS2 method was the same as the one for phospho MS2 with the following modification to MS2 target 5E4 ions. Acquired raw data files were processed with MaxQuant (47) (version 1.5.2.8), and peptides were identified from the MS/MS spectra searched against Toxoplasma gondii (ToxoDB) and Homo sapiens (UniProt) proteomes using the Andromeda (48) search engine. TMT-based experiments in MaxQuant were performed using the "reporter ion MS2 or MS3" built-in quantification algorithm, with the reporter mass tolerance set to 0.003 Da. Cysteine carbamidomethylation was selected as a fixed modification. Methionine oxidation, acetylation of the protein N terminus, deamidation (N, Q), and phosphorylation (S, T, and Y) were selected as variable modifications. The enzyme specificity was set to trypsin with a maximum of 2 missed cleavages. The precursor mass tolerance was set to 20 ppm for the first search (used for mass recalibration) and to 4.5 ppm for the main search. The "Match between runs" option was enabled for fractionated samples (time window of 0.7 min). The data sets were filtered on posterior error probability to achieve a 1% false-discovery rate on the protein, peptide, and site levels. Data were further analyzed as described in Results and in Data Set S1 using Microsoft Office Excel 2016 and Perseus (49) (version 1.5.0.9).
Data availability. The data behind the manuscript are freely available. The processed proteome and phosphoproteome data are included in Data Set S1, and the raw mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE (50) partner repository with the data set identifier PXD019729.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, DOCX file, 0.02 MB.