Not a Simple Tether: Binding of Toxoplasma gondii AMA1 to RON2 during Invasion Protects AMA1 from Rhomboid-Mediated Cleavage and Leads to Dephosphorylation of Its Cytosolic Tail

ABSTRACT Apical membrane antigen 1 (AMA1) is a receptor protein on the surface of Toxoplasma gondii that plays a critical role in host cell invasion. The ligand to which T. gondii AMA1 (TgAMA1) binds, TgRON2, is secreted into the host cell membrane by the parasite during the early stages of invasion. The TgAMA1-TgRON2 complex forms the core of the “moving junction,” a ring-shaped zone of tight contact between the parasite and host cell membranes, through which the parasite pushes itself during invasion. Paradoxically, the parasite also expresses rhomboid proteases that constitutively cleave the TgAMA1 transmembrane domain. How can TgAMA1 function effectively in host cell binding if its extracellular domain is constantly shed from the parasite surface? We show here that when TgAMA1 binds the domain 3 (D3) peptide of TgRON2, its susceptibility to cleavage by rhomboid protease(s) is greatly reduced. This likely serves to maintain parasite-host cell binding at the moving junction, a hypothesis supported by data showing that parasites expressing a hypercleavable version of TgAMA1 invade less efficiently than wild-type parasites do. Treatment of parasites with the D3 peptide was also found to reduce phosphorylation of S527 on the cytoplasmic tail of TgAMA1, and parasites expressing a phosphomimetic S527D allele of TgAMA1 showed an invasion defect. Taken together, these data suggest that TgAMA1-TgRON2 interaction at the moving junction protects TgAMA1 molecules that are actively engaged in host cell penetration from rhomboid-mediated cleavage and generates an outside-in signal that leads to dephosphorylation of the TgAMA1 cytosolic tail. Both of these effects are required for maximally efficient host cell invasion.

between AMA1 and invasion was first suggested when antibodies against AMA1 were found to block invasion (10,13,14). Subsequently, parasites either conditionally depleted of AMA1 (12) or completely lacking AMA1 (8) showed reduced host cell attachment and a severe invasion defect. AMA1 localizes to the "moving junction," a ring-like zone of tight apposition between the parasite and host cell plasma membranes, through which an invading parasite squeezes as it pushes its way into the host cell (15)(16)(17)(18). At the moving junction, the extracellular domain of AMA1 physically connects the host cell and parasite membranes by binding to RON2, a protein secreted from the rhoptries into the host cell membrane during invasion (15,(19)(20)(21)(22). Several independent studies identified domain 3 (D3) of RON2 as the region that binds to AMA1 with high (6 nM) affinity (16,(23)(24)(25)(26). RON2, in turn, binds to other rhoptry-derived proteins (RON4, -5, and -8) secreted into the host cell (15,21,22,(27)(28)(29)(30). Binding of this heterooligomeric RON protein complex to the host cell cytoskeleton would create the "anchor" against which the parasite exerts force during internalization (30)(31)(32)(33). This widely accepted model for the role of AMA1 at the moving junction was questioned recently by the observation that the small number of ama1 knockout parasites that are able to invade do so with normal kinetics, suggesting that AMA1 functions early in invasion, i.e., in host cell attachment, but is not required at the moving junction for parasite internalization (8). However, it was subsequently shown that the parasite can make use of functional homologs of AMA1 to partially compensate for the loss of AMA1 (17). Despite this functional redundancy, parasites lacking AMA1 are completely avirulent in immunocompetent mice (34), reinforcing the importance of AMA1 in the parasite's lytic cycle.
Like other microneme proteins, after T. gondii AMA1 (TgAMA1) has been trafficked onto the parasite surface, its transmembrane domain is cleaved, and its ectodomain is shed from the parasite (9,10,(35)(36)(37). The intramembrane proteolysis of microneme proteins is mediated by rhomboid proteases, primarily T. gondii ROM4 (TgROM4), which is distributed over the entire parasite surface, and to a lesser extent TgROM5, which is also localized on the parasite surface but is concentrated at the posterior end (35)(36)(37)(38)(39)(40). The function of microneme protein shedding is not entirely clear. It was recently proposed that shedding facilitates the establishment of an anterior-to-posterior concentration gradient of intact micronemal adhesins on the parasite surface (since new adhesins are continually secreted from the micronemes at the anterior end of the parasite), helping the parasite to orient properly during invasion (37). Whatever its function, constitutive intramembrane cleavage presents a problem for the model described above, since AMA1 cannot serve as an effective tether to the host cell if it is constantly being cleaved and shed from the parasite surface. The data presented here resolve this conundrum. We show that binding of TgAMA1 to TgRON2 significantly reduces its cleavage; thus, the subset of TgAMA1 molecules on the parasite surface that is actively engaged in host cell penetration is protected from rhomboid-mediated cleavage, stabilizing the interaction between the two cells. Furthermore, we report that binding of TgAMA1 to TgRON2 leads to the dephosphorylation of the TgAMA1 cytosolic tail (C-tail), suggesting that TgAMA1 plays a direct role in outside-in signaling during host cell invasion.

RESULTS
Treatment of parasites with the D3 peptide of TgRON2 reduces the amount of TgAMA1 shed from the parasite surface. Parasites expressing FLAG-tagged TgAMA1 were generated by allelic replacement at the endogenous TgAMA1 locus (see Fig. S1 in the supplemental material). The tag was placed within the extracellular domain (ectodomain) of TgAMA1 at a position that does not disrupt its function (35). These ARAMA1 WT parasites were pretreated either with glutathione S-transferase (GST) alone or with the D3 domain of RON2 fused to glutathione S-transferase (GST-D3) and used in constitutive microneme secretion assays. Parasites treated with 5 M GST-D3 showed, on average, a 5.7-fold decrease in the amount of FLAG-tagged TgAMA1 ectodomain shed into the assay supernatant compared to GST-treated controls ( Fig. 1A and B). The effect was dose dependent, increasing steadily from 0.01 to 1 M GST-D3 ( Fig. 1C and D). In contrast, GST-D3 treatment caused no significant decrease in the shedding of another microneme protein, TgMIC2, at concentrations as high as 5 M (Fig. 1A to D). Similar results were observed when microneme secretion was induced with calcium ionophore: treatment of parasites with GST-D3, but not GST, resulted in a decrease in TgAMA1 ectodomain shedding with little or no effect on the shedding of TgMIC2 (see Fig. S2 in the supplemental material).
To control for any artifactual effects of the GST fusion, we also tested the effect of TgRON2-2, a cysteine-dicyclized synthetic peptide that encompasses the TgAMA1-binding residues within the TgRON2 D3 domain (23). Like GST-D3, treatment with TgRON2-2 resulted in a dose-dependent decrease in the amount of TgAMA1 ectodomain shed into the secretion assay supernatant (see Fig. S3A to S3D in the supplemental material), and TgRON2-2 concentrations as high as 5 M had no effect on the shedding of either TgMIC2 (Fig. S3A, S3C, and S3D) or another microneme protein, TgMIC8 (Fig. S4).
Treatment of parasites with GST-D3 inhibits the cleavage of TgAMA1, not its trafficking to the parasite surface. The reduced amount of TgAMA1 ectodomain recovered in the assay supernatant after GST-D3 treatment (shown again in Fig. 2A, third and fourth lanes) could be due to either reduced trafficking of fulllength TgAMA1 from the micronemes onto the parasite surface or reduced intramembrane cleavage of TgAMA1 once it reaches the surface. If GST-D3 inhibits trafficking, we would expect to find less TgAMA1 on the parasite surface, but if it inhibits cleavage, we would expect to find more (Fig. 2B, left panel). To discriminate between these possibilities, we made use of parasites that contain an anhydrotetracycline (ATc)-repressible copy of wild-type TgAMA1 and a second, FLAG-tagged copy of TgAMA1 that is either wild type (AMA1 WT ) or contains mutations within its transmembrane domain (AMA1 AG/FFϩGG/FF ) that render it resistant to cleavage by rhomboid proteases (35). Following ATc treatment, AMA1 WT parasites (which express the FLAG-tagged wildtype allele of TgAMA1) were treated with either GST or GST-D3, and the amount of TgAMA1 on the surface of the parasites was quantified by flow cytometry (37,39). There was a significant increase in the amount of TgAMA1 on the surface of GST-D3treated parasites compared to parasites treated with GST alone (Fig. 2B, middle and right panels). Similar to the Western blotbased measurements of microneme secretion, the effect of GST-D3 on TgAMA1 surface abundance showed a clear dose dependence by flow cytometry (see Fig. S5 in the supplemental material). For a control, we measured the amount of glycosylphosphatidylinositol (GPI)-anchored TgSAG1 on the parasite surface (39) and found it to be unaffected by GST-D3 treatment (see Fig. S6 in the supplemental material).
These data are consistent with GST-D3 inhibiting surface cleavage of TgAMA1, rather than its trafficking from the micronemes. To independently confirm this, we did a similar experiment using the AMA1 AG/FFϩGG/FF parasites, which, after ATc treatment, express the cleavage-resistant TgAMA1 allele ( Fig. 2A, AMA1 AG/FFϩGG/FF ). In these mutant parasites, we would expect less TgAMA1 on the surface if GST-D3 reduces its trafficking, but little or no difference in the amount of TgAMA1 on the surface if GST-D3 treatment inhibits cleavage, since cleavage is already low in the mutant (Fig. 2C, left panel). We observed the latter result: no difference in the amount of AMA1 AG/FFϩGG/FF was detected on the parasite surface after treatment with GST-D3 (Fig. 2C, middle and right panels). Taken together, these data argue that the binding of TgRON2 to TgAMA1 reduces TgAMA1 intramembrane cleavage.
To test whether the binding of TgRON2 inhibits rhomboidmediated cleavage of TgAMA1 directly, we co-expressed TgAMA1 and the rhomboid protease TgROM5 in mammalian HEK293 cells and measured the amount of TgAMA1 released into the culture supernatant in the presence or absence of GST-D3. We used brefeldin A in these experiments to block secretion of any potential intracellularly processed TgAMA1 (Fig. 3A, left blot). GST-D3 bound to the surface of the TgAMA1-expressing cells (data not shown) but had no detected effect on the amount of TgAMA1 shed (Fig. 3A, right blot). Although TgROM4 is the protease responsible for the majority of TgAMA1 cleavage in the parasite (36,37), no one has yet succeeded in expressing functional TgROM4 in mammalian cells (38,40). However, TgROM5 is responsible for most of the residual TgAMA1 cleavage in parasites lacking TgROM4 (36,37), and we found that GST-D3 treatment inhibits this TgROM5-mediated cleavage of TgAMA1 (Fig. 3B). Thus, GST-D3 inhibits rhomboid-mediated TgAMA1 cleavage in parasites but not mammalian cells, suggesting the involvement of additional parasite-specific factors in the underlying mechanism (see Discussion).
Functional consequences of altered TgAMA1 cleavage. If interaction of TgAMA1 with TgRON2 at the moving junction inhibits cleavage of TgAMA1, this could stabilize the junction and enable efficient parasite penetration into the host cell. To test this hypothesis, we used a parasite line expressing a hypercleavable mutant form of TgAMA1 (AMA1 L/G ), which shows a Ͼ20-fold increase (35) in constitutive cleavage compared to AMA1 WT (Fig. 4A, compare third and seventh lanes). Treatment of the AMA1 L/G parasites with saturating amounts of GST-D3 inhibits this cleavage by~50% ( Fig. 4A seventh and eighth lanes, and Fig. 4B). However, because the baseline cleavage of the mutant protein is so high, significantly more shedding of mutant TgAMA1 is observed in the presence of GST-D3 than is seen FIG 2 GST-D3 causes reduced cleavage of TgAMA1 but not reduced trafficking to the parasite surface. (A) A microneme secretion assay was performed using conditional AMA1 WT and AMA1 AG/FFϩGG/FF parasites treated with either 1 M GST-D3 (ϩ) or GST (Ϫ). Assay pellet (P) and supernatant (S) fractions are indicated. The decrease in shedding of TgAMA1 was evident in AMA1 WT parasites treated with GST-D3 (compare AMA1 WT supernatants with and without GST-D3). No shed ectodomain was detected using either GST-or GST-D3-treated TgAMA1 AG/FFϩGG/FF parasites (TgAMA1 AG/FFϩGG/FF supernatants with and without GST-D3). Full-length protein in the supernatant is likely due to parasite lysis. TgMLC1 was used as a loading control. (B, left) Schematic showing the two possible outcomes from treatment of AMA1 WT parasites with GST-D3. If GST-D3 reduces cleavage of TgAMA1, we would expect more surface TgAMA1 (red spikes) after GST-D3 treatment than after GST treatment. If GST-D3 affects trafficking of TgAMA1 to the parasite surface, we would expect less TgAMA1 on the surface after GST-D3 treatment. (Middle) By flow cytometry, GST-D3 treatment resulted in an increase in surface TgAMA1 in AMA1 WT parasites, consistent with an effect on cleavage. Representative histograms are shown. The dashed-line histogram represents unstained parasites, the gray histogram represents parasites stained with primary antibody only, the purple histogram represents parasites treated with GST, and the orange histogram represents parasites treated with GST-D3. Despite this increase in TgAMA1 shedding, the AMA1 L/G parasites are capable of normal levels of invasion in a 60-min endpoint assay (35) and show no obvious difference in the moving junction compared to that seen in invading AMA1 WT parasites (data not shown). However, when live invasion assays were performed to compare the kinetics of invasion, AMA1 WT parasites took on average 16.1 s to internalize (range, 10.33 to 24.34 s), whereas the AMA1 L/G parasites took longer, on average 20.7 s (range, 13.04 to 39.36 s; Fig. 4C).
Taken together, these data suggest a model in which the binding of TgRON2 to TgAMA1 at the moving junction protects the TgAMA1 molecules that are actively engaged in host cell penetration from rhomboid-mediated cleavage, enabling maximally efficient host cell invasion.

Interaction of TgAMA1 with TgRON2 reduces phosphorylation on the C-tail of TgAMA1.
Given that binding of GST-D3 to the extracellular domain (23,26) of TgAMA1 affects cleavage of residues within the TgAMA1 transmembrane domain, we asked whether the binding might have other distal effects, altering some aspect of TgAMA1 C-tail function. Phosphorylation of S610 on the C-tail of Plasmodium falciparum AMA1 (PfAMA1) is known to be important for invasion (41,42). The C-tail of TgAMA1 was also recently shown to be phosphorylated, but on a different residue, S527 (43). To test whether TgRON2 binding to TgAMA1 affects the phosphorylation state of S527, we metabolically labeled parasites by stable isotope labeling with amino acids in cell culture (SILAC) (44,45) with either "heavy" or "light" isotopic versions of L-arginine and L-lysine. The heavy-labeled parasites were then treated with GST-D3, and the light-labeled parasites were treated with GST. TgAMA1 was immunoprecipitated from each sample, and mass spectrometry was used to quantify the heavy and light tryptic peptides recovered. The relative heavy/light (H/L) ratio of the phosphorylated S527-containing peptide in the two samples was compared to the H/L ratio of all TgAMA1 tryptic peptides recovered. The analysis revealed a consistent 30 to 40% reduction in S527 phosphorylation following treatment with GST-D3 ( To test whether the phosphorylation state of S527 is functionally important, parasites expressing nonphosphorylatable (S527A) or phosphomimetic (S527D) forms of TgAMA1 were generated by allelic replacement. Diagnostic PCR confirmed integration of the mutant alleles at the endogenous TgAMA1 locus, and the mutant proteins localized properly and were expressed at levels similar to those of ARAMA1 WT (see Fig. S9A to S9C in the supplemental material). The ARAMA1 S527A and ARAMA1 S527D parasite lines shed similar amounts of TgAMA1, and shedding was reduced in both lines following treatment with GST-D3 (Fig. 6A), demonstrating that GST-D3-induced dephosphorylation of S527 does not play a direct role in the GST-D3-induced reduction in TgAMA1 shedding. In invasion assays, parasites expressing the nonphosphorylatable S527A allele invaded host cells to levels indistinguishable from those of the wild-type parasites. In contrast, parasites expressing the phosphomimetic S527D allele showed a consistent 25% decrease in host cell invasion (Fig. 6B), suggesting that the TgRON2-induced dephosphorylation of TgAMA1 is necessary for optimal host cell invasion.
The site on PfAMA1 that needs to be phosphorylated for invasion to occur, Ser610, does not align with S527 in TgAMA1, but instead corresponds to a potentially phosphomimetic TgAMA1 residue, D558. To determine whether the negative charge on D558 is important for C-tail function and invasion in T. gondii, parasites expressing a D558A mutant allele of TgAMA1 were generated by allelic replacement (see Fig. S9B and S9C in the supplemental material). These parasites showed no defect in invasion (Fig. 6B). This result, together with the observation that C-tail dephosphorylation rather than phosphorylation is required for optimal invasion in T. gondii, provides compelling evidence that AMA1 C-tail function is regulated differently in T. gondii and P. falciparum.

DISCUSSION
During invasion, AMA1 (embedded in the parasite plasma membrane) binds with high affinity to RON2 (embedded in the host cell plasma membrane) to form the core of the moving junction. We show here that when TgAMA1 engages with TgRON2, it be- comes less susceptible to rhomboid-mediated intramembrane cleavage. We therefore propose that TgAMA1 cleavage is regulated spatially on the parasite surface. Most of the surface TgAMA1 is constitutively cleaved and shed, establishing an anterior to posterior gradient of TgAMA1 that may be important for attachment or junction formation (37), but the TgAMA1 molecules that are actively engaged with TgRON2 at the moving junction are resistant to cleavage, enabling them to serve as effective tethers between the two cells. Consistent with this model, parasites expressing a mutant form of TgAMA1 in which the transmembrane domain is more readily cleaved by surface rhomboids showed a modest but significant delay in their invasion kinetics (Fig. 4C). Note that while TgAMA1 cleavage is greatly enhanced in these mutant parasites, it is still partially inhibited by GST-D3 (Fig. 4B), which may explain why the effect on invasion is not more pronounced, i.e., even though the mutant TgAMA1 is more susceptible to cleavage than the wild type, binding to TgRON2 still provides some degree of protection against cleavage and a proportion of the parasite-host cell tethers therefore likely remain intact.
The mechanism by which TgRON2 binding inhibits TgAMA1 cleavage is not known. GST-D3 clearly inhibits cleavage by TgROM4, since TgROM4 is responsible for the great majority of TgAMA1 cleavage in parasites (36,37), and treatment with the peptide inhibits 80 to 90% of total TgAMA1 cleavage ( Fig. 1; see Fig. S3 in the supplemental material). GST-D3 treatment also inhibits TgAMA1 cleavage by TgROM5 (Fig. 3B). However, the effect of the peptide on TgAMA1 cleavage is unlikely to be mediated through a general inhibition of rhomboid protease activity, since treatment with GST-D3 did not alter the processing of TgMIC2 and TgMIC8 ( Fig. 1; see Fig. S4 in the supplemental material), which are cleaved by the same proteases (36,37). Binding of TgRON2 to TgAMA1 is known to induce a conformational change in the TgAMA1 ectodomain (23) that could somehow be propagated to the transmembrane domain, making it a less suitable substrate for rhomboid protease cleavage. This also seems unlikely, since we expressed TgAMA1 and TgROM5 in mammalian cells and saw no difference in the amount of TgAMA1 shed in the presence or absence of GST-D3 (Fig. 3A), whereas GST-D3 treatment did inhibit TgROM5-mediated shedding in parasites (Fig. 3B). Inhibition of rhomboid-mediated TgAMA1 cleavage by GST-D3 in parasites but not mammalian cells may reflect the involvement of an additional protein(s) that is present in the parasite but not the mammalian cell. Alternatively, the physicochem- ical environment within the parasite and mammalian cell plasma membranes may be different in some way that alters the dynamics of the TgAMA1 transmembrane helix and/or its propensity to enter into the rhomboid active site, affecting cleavage (46).
The phosphorylation of S610 on the C-tail of PfAMA1 is important for invasion of erythrocytes by P. falciparum merozoites (41,42). Intriguingly, alignment of the orthologous AMA1 C-tails from 13 different apicomplexan parasites reveals that S610 is highly conserved across the phylum except in T. gondii, where it is replaced by potentially phosphomimetic D558 (19). T. gondii tachyzoites expressing TgAMA1 with a D558A mutation invaded cells normally (Fig. 6B), indicating that C-tail function is regulated via different mechanisms in T. gondii and P. falciparum. Indeed, the TgAMA1 C-tail is phosphorylated on a different residue, S527, and we show here that GST-D3 treatment of parasites results in a 30 to 40% decrease in S527 phosphorylation ( Fig. 5; see Fig. S8 in the supplemental material). When wild-type TgAMA1 was replaced with a TgAMA1 S527D phosphomimetic allele, we found that invasion was partially inhibited (Fig. 6B). The effects of GST-D3 are very likely induced through direct binding to TgAMA1, since in the absence of TgAMA1, no detected peptide binds to the parasite surface (24). Taken together, these data suggest that when TgAMA1 binds to TgRON2, an outside-in signal is generated that leads to dephosphorylation of the AMA1 C-tail, and this dephosphorylation is required for maximally efficient invasion. The stoichiometry of phosphorylation on S527 appears to be low, based on the relative abundance of the phospho and dephospho forms of the S527-containing peptide (e.g., compare the areas of the extracted ion chromatograms in Fig. 5) and the observation that a decrease in the abundance of the phospho form following GST-D3 treatment was not associated with a corresponding increase in the abundance of the dephospho form ( Fig. 5; see Fig. S8 in the supplemental material). Further studies will be required to determine whether the low stoichiometry reflects localized phosphorylation of a specific subset of TgAMA1 molecules within the parasite, e.g., those that have been trafficked from the micronemes onto the parasite surface. If this were the case, dephosphorylation of the TgAMA1 C-tail at the moving junction could serve to distinguish TgAMA1 molecules that are actively involved in invasion from the rest of the surface TgAMA1. It is worth noting that at least two protein phosphatases have been previously implicated in attachment and invasion, one of which localizes to the apical end of the parasite prior to invasion (47,48).
In summary, our data show that the binding of the ectodomain of TgAMA1 to TgRON2 does more than just tether the parasite to the host cell. This interaction has two additional effects on TgAMA1, distal to the site of binding: it protects the transmembrane domain of TgAMA1 from cleavage by cell surface rhomboid protease(s), and it generates an outside-in signal that leads to dephosphorylation of the TgAMA1 C-tail, enhancing invasion efficiency. These effects may contribute to the ability of exogenously added GST-D3 to reduce the efficiency of host cell invasion by T. gondii (19). Inhibition was originally presumed to be due to simple competition between the added GST-D3 and native TgRON2 in the host cell plasma membrane for binding of TgAMA1, but the results presented here indicate that GST-D3 binding has additional important effects on TgAMA1 function. Future studies will focus on the mechanism of TgRON2-induced cleavage resistance and the role of TgAMA1 dephosphorylation in invasion and subsequent events of the parasite's lytic cycle.
Infected cells containing large intracellular vacuoles were detached from the flask using a cell scraper, and parasites were released by passage through a blunt 26-gauge needle. Host cell debris was removed using a sterile 3-m Nuclepore (Whatman) filter. Unless otherwise indicated, parasites were centrifuged at 2,500 rpm at 25°C for 4 min and resuspended in medium prior to use.
Generation of FLAG-tagged allelic replacement parasites. Phleomycin-resistant parasites containing either wild-type or mutant FLAG-tagged TgAMA1 at the TgAMA1 endogenous locus were generated using the vector pA/TgAMA1 WT Flag.BLE as described in reference 35 and outlined in Fig. S1 in the supplemental material. QuikChange mutagenesis was used to introduce the desired point mutation(s) into pA/ TgAMA1 WT Flag.BLE using the primers listed in Table S1. Following parasite transfection and selection, single clones were screened for integration of Flag-TgAMA1 at the correct locus using the primers listed in Table S1 (p1 to p4). The four allelic replacement parasite lines used in this study are designated ARAMA1 WT , ARAMA1 527A , ARAMA1 527D , and ARAMA1 D558A .
Immunofluorescence analysis. Confluent monolayers of HFFs on 25-mm circular glass coverslips were infected with parasites for 12 h, fixed for 15 min in phosphate-buffered saline (PBS) containing 2.5% (vol/vol) paraformaldehyde and permeabilized for 15 min in PBS containing 0.25% (vol/vol) Triton X-100 (TX-100). After the coverslips were blocked for 30 min in blocking buffer (PBS containing 1% [wt/vol] bovine serum albumin [BSA]), they were incubated for 15 min in blocking buffer containing mouse anti-FLAG (Sigma-Aldrich) or rabbit anti-TgIMC1 (50), each at a dilution of 1:1,000, followed by incubation for 15 min with Alexa Fluor 488-conjugated goat anti-mouse IgG (Invitrogen) or Alexa Fluor 546-conjugated goat anti-rabbit IgG (Invitrogen), each at a 1:1,000 dilution in blocking buffer. The coverslips were mounted on glass slides and imaged using the 100ϫ objective of a Nikon Eclipse TE 300 epifluorescence microscope.
Microneme secretion assays. GST and GST-D3 were purified from Escherichia coli (Rosetta strain; Novagen/EMD Millipore) as previously described (24,51) except that no detergent was used in the lysis or wash buffers. The purified proteins were dialyzed against 10 mM Tris-HCl (pH 8.0)Ϫ150 mM NaCl and concentrated in Amicon Ultra 0.5-ml 10,000 (10K) membrane centrifugal filters (Millipore). Glycerol was added to 20% (vol/vol), and the stocks were stored at Ϫ80°C.
Microneme secretion assays were performed as described previously Parasites were incubated at 37°C with CO 2 for 10 and 30 min for induced and constitutive secretion, respectively, placed on ice for 5 min, and centrifuged at 1,200 ϫ g for 5 min. The pellet and supernatant fractions were analyzed by SDS-PAGE/Western blotting. Secretion assays with TgRON2-2 (generously provided by Marty Boulanger) were done in DMEM with 1% dialyzed FBS (Invitrogen) and 20 mM HEPES (pH 7.2), and the total volume of parasites and peptide was 120 l (52). For secretion assays with AMA1 KD i parasites (AMA1 WT , AMA1 L/G , and AMA1 AG/FFϩGG/FF) , the assay medium was HBSS with 100 mM HEPES (pH 7.2), and the parasites were incubated for 15 min at 37°C (35). Immunoblots were probed with mouse anti-TgMIC2 (generous gift from Vern Carruthers; 1:10,000), mouse anti-FLAG (1:10,000), rabbit anti-TgMIC8 (generous gift from Markus Meissner; 1:500), and rabbit anti-TgMLC1 (1:1,000) in Odyssey LI-COR block buffer (LI-COR Biosciences), followed by incubation with IRDye 680-conjugated anti-rabbit IgG and IRDye 800-conjugated anti-mouse IgG (LI-COR), each at 1:20,000 in PBS containing 0.5% BSA. The blots were washed in PBS and scanned using an Odyssey CLx infrared imager (LI-COR). Images were processed using Image Studio software (LI-COR). After adjusting for parasite equivalents loaded (in all cases the percentage of the pellet fraction loaded was half that of the supernatant fraction), the signal intensity ratio of the band in each pair of pellet and supernatant fractions was calculated and plotted using GraphPad Prism 6.
Flow cytometry. Parasites were harvested and resuspended in motility buffer (1ϫ minimal essential medium [MEM], 1% FBS, 10 mM Glu-taMAX [Thermo Fisher], 10 mM HEPES [pH 7.2]) or microneme secretion medium with 1% ovalbumin and 0.5% FetalPlex. A total of 3 ϫ 10 7 parasites were incubated with either GST or GST-D3 at 37°C on a nutator for 30 min and prepared for flow cytometry (39). Briefly, parasites were fixed using 4% (vol/vol) paraformaldehyde in PBS for 20 min on ice, washed three times in blocking buffer (PBS containing 1% [wt/vol] BSA and 1% [vol/vol] goat serum) and incubated for 20 min in this buffer. Parasites were then incubated for 15 min with 5 g/ml mouse anti-FLAG, 20 g/ml fluorescein isothiocyanate (FITC)-conjugated rabbit anti-T. gondii (ab20907 [Abcam] [Western blotting revealed that Ͼ95% of the immunoreactivity of this antibody is directed against TgSAG1 {data not shown}]) or 2 g/ml rabbit anti-GST (Immunology Consultants Laboratory [ICL] antibodies). After four washes at 1,000 ϫ g for 2 min each, parasites were incubated with Alexa Fluor 546-conjugated goat antimouse IgG or Alexa Fluor 488-conjugated goat anti-rabbit IgG at a 1:500 dilution for 15 min. Parasites were washed four times and resuspended in 200 l of blocking buffer. Flow cytometric analysis was performed on a MACSQuant VYB instrument (Miltenyi Biotech) equipped with 561-nm and 488-nm lasers. The Alexa Fluor 546 signal (by 561-nm laser) was detected using a 586/15 bandpass filter and Alexa Fluor 488/FITC signals (by 488-nm laser) using a 525/50 bandpass filter. Data were acquired using MACSQuantify v2.5 software and analyzed with FlowJo software (TreeStar v10).
Live imaging of parasite invasion. Parasite invasion kinetics were measured as described previously (35), except that the invasion medium was 1ϫ MEM containing 1% FBS, 10 mM GlutaMAX, and 10 mM HEPES (pH 7.2).
Mass spectrometry. (i) SILAC and immunoprecipitation. SILAC DMEM medium (Thermo Fisher) was supplemented with 10% dialyzed FBS, 10 mM HEPES (pH 7.2), and 10 units/ml each of penicillin and streptomycin sulfate. Heavy SILAC medium also contained stable isotopic forms of "heavy" L-arginine-HCl ( 13 C 6 , 15 N 4 ) at 0.398 mM and L-lysine-2HCl ( 13 C 6 , 15 N 2 ) at 0.798 mM (Cambridge Isotope Laboratories). The light media contained naturally occurring "light" isotopic forms of L-arginine and L-lysine at 0.398 mM and 0.798 mM, respectively (53). L-Proline was added to the media to 40 mg/liter in order to prevent arginine-to-proline conversion (54). HFFs were grown in heavy or light SILAC medium for five or six passages. Two days prior to the experiment, 12 T75 flasks of heavy-and light-labeled host cells were infected with parasites that had been cultured in heavy-or light-labeled host cells for one complete lytic cycle.
(ii) Trypsin digestion. The band containing 75-kDa TgAMA1 was excised and subjected to reduction, alkylation, and in-gel trypsin digestion, as previously described (55).
(iii) Liquid chromatography-mass spectrometry. The tryptic digestion products were dissolved in 20 l of 0.1% formic acid and 2.5% acetonitrile, and 6 l was loaded onto a fused silica microcapillary liquid chromatography (LC) column (12 cm by 100 m [inner diameter]) packed with C 18 reversed-phase resin (5-m particle size; 20-nm pore size; Magic C 18 AQ, Michrom Bioresources). Peptides were separated by applying a gradient of 3 to 45% acetonitrile in 0.1% formic acid at a flow rate of 500 nl/min for 75 min. Nanospray was used to introduce peptides into a linear ion trap LTQ Orbitrap mass spectrometer (Thermo Fisher) via a nanospray ionization source. Mass spectrometry (MS) data were acquired in a data-dependent acquisition mode, in which an Orbitrap survey scan from m/z 360 to 1,700 (resolution, 30,000 full width at half maximum [FWHM] at m/z 400) was paralleled by 10 LTQ tandem mass spectrometry (MS/MS) scans of the most abundant ions. The "lock mass" option was utilized in all full scans. After a liquid chromatography-mass spectrometry (LC-MS) run was completed and spectra were obtained, the spectra were searched against the T. gondii proteome database, v8 (http:// www.toxodb.org/toxo/) using SEQUEST in Proteome Discoverer 1.4 (Thermo Electron). The search parameters permitted a 20-ppm precursor MS tolerance and a 1.0-Da MS/MS tolerance. Carbamidomethylation of cysteines was set as fixed modifications, and oxidation of methionine (M), phosphorylation on serine (S), threonine (T) and tyrosine (Y), SILAC labels of 13 C 6 15 N 2 at lysine and of 13 C 6 15 N 4 at arginine were allowed as variable modifications. Up to three missed tryptic cleavages of peptides were considered, and the false-discovery rate was set at 1% at the peptide level. The relative abundances of heavy-labeled (GST-D3-treated) and light-labeled (GST-treated) peptides (expressed as H/L ratios) were quantified by integrating the intensities of peptide ion elution profiles of the isotopologues with the Precursor Ions Quantifier node, and the probability of phosphorylation for each S/T/Y site on each peptide was calculated by the phosphoRS 3.0 node in Proteome Discoverer 1.4.
(iv) Parallel reaction monitoring. Parallel reaction monitoring (PRM) was carried out on the Q-Exactive mass spectrometer coupled to an EASY-nLC system (Thermo Fisher). Peptides were separated on a fused silica capillary (100 m by 120 mm) packed with Halo C 18 (2.7-m particle size, 90-nm pore size, Michrom Bioresources) at a flow rate of 300 nl/min. Peptides were introduced into the mass spectrometer via a nanospray ionization source at a spray voltage of 2.0 kV. Mass spectrometry data were acquired with alternating MS-selected ion monitoring (SIM) scans and PRM (two scan groups), and lock mass function was activated (m/z, 371.1012; use lock masses, best; lock mass injection, full MS). Full scans were acquired from m/z 300 to 2,000 at 70,000 resolution (automatic gain control [AGC] target, 1e 6 ; maximum ion time [max IT], 100 ms; profile mode). PRM was carried out with higher-energy collisional dissociation (HCD) MS/MS scans at 17,500 resolution on the precursors of interest (their m/z values were preimported into the inclusion list), with the following settings: AGC target, 5e 4 ; max IT, 100 ms; isolation width of 1.6 m/z; and a normalized collisional energy of 35%. The light-and heavy-labeled TgAMA1 peptides were monitored with various charge states. The raw data were searched against the T. gondii proteome database, and the search files (.msf) imported into Skyline for selecting the precursor or transitions for quantitation. The raw data were exported from XCalibur to GraphPad Prism 6 for chromatogram plotting.
Laser scanning cytometer-based invasion assay. The two-color invasion assay was performed as previously described (56) with the following modifications. First, 3 ϫ 10 6 parasites were used to infect confluent HFF cells on 25-mm circular coverslips. The parasites were allowed to settle for 20 min at 21°C and then incubated for 1 h at 37°C. After fixation, samples were blocked for 1 h in PBS containing 2% BSA. The assays were performed 9 or 10 times, and each biological replicate was performed in duplicate.
Heterologous TgAMA1 cleavage assay. Human HEK293T cells were grown at 37°C and 5% CO 2 in DMEM supplemented with 10 mM HEPES (pH 7.2), 2 mM L-glutamine, and 10% FBS and transfected with XtremeGENE HP DNA transfection reagent (Roche) as described previously (35). Transfected cells were pretreated for 5 min with DMEM containing 10 M brefeldin A, followed by conditioning in DMEM in the presence of 10 M brefeldin A and 5 M GST or GST-D3 for 2.5 h. Cleaved GFP-TgAMA1 released into the culture supernatant was then detected by quantitative anti-green fluorescent protein (anti-GFP) Western blot analysis using the Odyssey imager.

ACKNOWLEDGMENTS
We thank members of the Ward lab for helpful comments on the manuscript, Marty Boulanger for helpful discussions and for generously providing the RON2-2 peptide, L. David Sibley for the TgROM4 knockout parasites, Julia Ganister Fields for proteomic support, Jan Schwarz for assistance with laser scanning cytometry, Lauren Hinkel for carrying out preliminary experiments, and Anne Kelsen for excellent technical assistance. Automated DNA sequencing was performed in the VT Cancer Center DNA Analysis Facility, laser scanning cytometry in the UVM Microscopy Imaging Center, and flow cytometry in the Harry Hood Bassett Flow Cytometry and Cell Sorting Facility, all at the University of Vermont. Genomic-scale data sets and ancillary information were obtained from the Toxoplasma Genome Database (ToxoDB.org). ToxoDB is a component of the Eukaryotic Pathogen Genomics Resource (EuPathD-B.org), a Bioinformatics Resource Center (BRC) supported by the Na-tional Institutes of Allergy and Infectious Diseases; we gratefully acknowledge the staff responsible for developing and maintaining this resource. This work was supported by Public Health Service grant AI105191 to G.E.W., grants AI21423 and AI97633 to J.B., grant AI066025 to S.U., and grant GM103496, which supports core facilities at the University of Vermont. M.T. was supported by an American Heart Association Postdoctoral Fellowship. The Vermont Genetics Network Proteomics Facility is supported through NIH grant P20GM103449 from the INBRE Program of the National Institute of General Medical Sciences.
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.