Toxoplasma DJ-1 Regulates Organelle Secretion by a Direct Interaction with Calcium-Dependent Protein Kinase 1

We previously discovered that TgDJ-1 functions within the pathway underlying the exocytosis of specialized virulence-associated organelles called micronemes (18), but the molecular mechanism by which it asserted control over this event remained unclear. To get a better sense of structural features important for TgDJ-1 function, we solved the crystal structure of TgDJ-1 to 2.1 Å (Fig. 1A and Table S4 in the supplemental material). As with other orthologues, TgDJ-1 is a constitutive homodimer. The target of WRR-086, Cys127, sits next to the nucleophilic pocket that contains Cys104 (Fig. 1A). As reported for most DJ-1 structures, the reactive Cys104 was found exclusively in the oxidized sulfinic acid state. The function of DJ-1 as a redox sensor has been linked to this highly conserved and reactive cysteine at position 106 in humans (12). To independently assess the reactivity of Cys104 in TgDJ-1, we applied a recently described method that uses cysteine sensitivity to iodoacetamide alkylation (at two concentrations, high and low) to profile the reactivity of cysteines (19). Using this approach, the more reactive a cysteine residue is, the closer its ratio score (high concentration/low concentration) is to 1. We found that TgDJ-1 Cys104 had a cysteine reactivity ratio score of 2.03 (standard deviation [SD] = 0.03), compared with the nearby reactive cysteine, cysteine 127, that is the target of WRR-086, which had a score of 3.44 (SD = 0.21). These data indicate that TgDJ-1 Cys104 is highly reactive, potentially explaining its hyperoxidized state in the crystal structure. To precisely define the reactivity of Cys104, we measured its pK_a. We monitored the absorption of the thiolate anion at 240 nm as a function of pH (20, 21) and determined that it has a pK_a of 5.0 ± 0.1 (see Fig. S1 in the supplemental material). This result further confirms the high reactivity of this residue and suggests that at physiological pH, this cysteine would be highly sensitive to oxidation and could function as a ROS sensor as suggested in other systems (8).

Comparing our structure to previously deposited structures of other DJ-1 family members, we noted that the two structures of human DJ-1 in which Cys106 is reduced contain a tetragonal anion (sulfate [PDB accession no. 2OR3 ] [22] or phosphate [PDB accession no. 3BWE ] [23]). This anion is most prominent in structures of reduced DJ-1, as it would clash with the oxidized cysteine. We noted that the ion is coordinated by a highly conserved arginine dyad present in the DJ-1 superfamily of proteins (12) (Fig. 1B and C). An interesting structural feature is that each arginine of the dyad is contributed by a different individual monomer unit of the DJ-1 dimer. Despite this high level of conservation, no function for these residues has been reported for any of the DJ-1 orthologues, including human.

DJ-1 has been suggested to interact with a variety of proteins (8), leading to speculation that it functions as a scaffold or chaperone. One of the first proteins shown to directly interact with human DJ-1 was named DJ-1 binding protein (DJBP) (24), later renamed EF-hand calcium-binding domain-containing protein 6. That HsDJ-1 can bind to a calcium-responsive protein is particularly interesting as TgDJ-1 plays a role in the calcium-triggered process of microneme secretion. CDPK1 is essential for microneme secretion (15) and has an autoinhibitory C-terminal calmodulin-like domain containing four tandem EF-hand domains (25) that undergo a dramatic conformational rearrangement upon calcium binding to activate the kinase. This led us to hypothesize that TgDJ-1 could regulate microneme secretion directly through an interaction with CDPK1. To test this, we performed in vitro association studies using recombinant glutathione S-transferase (GST)-tagged CDPK1 and recombinant TgDJ-1 (rTgDJ-1). These studies confirmed that GST-CDPK1, and not GST alone, was able to precipitate rTgDJ-1, demonstrating that these two proteins can interact directly and that the complex requires no additional factors for formation (Fig. 1D).

Many kinases are regulated by autophosphorylation (26). CDPK1 contains at least three major phosphosites (serine 25, 61, and 349), identified in a recent phosphoproteomic study (27). In light of this and the insights gleaned from the crystal structure, we investigated whether phosphorylation of CDPK1 was critical for its interaction with TgDJ-1. Coexpression of CDPK1 with lambda phosphatase in Escherichia coli produces homogenously dephosphorylated recombinant kinase for crystallographic studies (25). The previous successful crystallization of CDPK1 coexpressed with lambda phosphatase reassured us that the protein would likely be correctly folded but lack phosphorylation modifications. Using this protocol, we expressed and purified dephosphorylated GST-tagged CDPK1. Conversion to the dephosphorylated state was confirmed by a modest SDS-PAGE shift of the GST-CDPK1 expression product (Fig. 1E). We then tested the ability of either phosphorylated or dephosphorylated CDPK1 to precipitate rTgDJ-1. Remarkably, the interaction displayed absolute dependence upon the phosphostate of the kinase, with TgDJ-1 associating only with a phosphorylated form of CDPK1 (Fig. 1E).

The TgDJ-1 crystal structure led us to consider a possible function for the highly conserved arginine dyad as a phosphate-coordinating motif located on the edge of the nucleophilic trough. As both arginine residues are expected to be required for correct coordination of the tetragonal anion, we predicted that mutation of a single arginine of the dyad would prevent correct coordination of the charged phosphate group and abrogate binding. To directly test this, we mutated the two arginine residues at positions 27 and 47 to alanine. We expressed and purified a single mutant, TgDJ-1^R47A, and the double mutant, TgDJ-1^R27A/R47A, and then tested whether these residues were involved in the interaction. Our attempts to generate the single TgDJ-1^R27A mutant were unsuccessful. As we had previously observed, the TgDJ-1 wild-type strain (TgDJ-1^WT) interacted only with phosphorylated CDPK1 (Fig. 1F). As expected, when a single arginine of the dyad was mutated to an alanine in TgDJ-1^R47A, CDPK1 was unable to interact regardless of its phosphorylation state (Fig. 1F). In contrast, the double mutant (TgDJ-1^R27A/R47A) was able to bind to CDPK1 regardless of its phosphorylation state (Fig. 1F). These data are consistent with the TgDJ-1 arginine dyad functioning as a phosphorylation recognition motif or gatekeeper, facilitating recognition of phosphorylated CDPK1 by TgDJ-1.

In light of the published role for DJ-1 in sensing intracellular reactive oxygen species (28) and for CDPK1 as a calcium sensor (25), we sought to test if these naturally occurring signaling molecules could influence the interaction of the two proteins. For this purpose, we used a more authentic “cell extract” system in which we manipulated the calcium and redox conditions. We prepared soluble cell extracts from naturally egressed CDPK1-hemagglutinin (CDPK1-HA) tachyzoites and adjusted the extracts to make them either calcium rich (+calcium) or calcium depleted (+EGTA), as well as reducing (+dithiothreitol [+DTT]) or oxidizing (+H₂O₂). We used H₂O₂ as the oxidizing agent as it is broadly regarded as the most relevant candidate for the transduction of ROS signals (29). After addition of recombinant TgDJ-1 (rTgDJ-1), we immunoprecipitated (IP) CDPK1-HA and assessed the ability of the kinase to coprecipitate rTgDJ-1. Interestingly, the interaction between rTgDJ-1 and CDPK1-HA was both calcium and H₂O₂ dependent (Fig. 2A), with rTgDJ-1 predominantly interacting with CDPK1 in reducing conditions and in the absence of calcium.

We next sought to determine if specific signals could trigger dissociation of the TgDJ-1/CDPK1 complex. We isolated the complex from soluble cell extracts using conditions that favored its assembly (reducing and calcium depleted) and then incubated with calcium or H₂O₂. Interestingly, the addition of either calcium or H₂O₂ alone was not sufficient to trigger dissociation of the complex (Fig. 2B). However, in the presence of both calcium and ROS, the complex dissociated (Fig. 2B). These data suggest that the TgDJ-1/CDPK1 complex could function as a sensor for these two secondary messengers.

To determine if DJ-1 binding had any direct effect on CDPK1 function, we performed an in vitro kinase assay, measuring residual ATP levels present in a given reaction following phosphorylation of a peptide substrate as an endpoint read (low signal = high kinase activity). We optimized the concentration of CDPK1 based on its 50% effective concentration (EC₅₀) of ~15 nM, hypothesizing that this concentration of kinase would provide sufficient dynamic range to distinguish between inhibition and activation in subsequent experiments (Fig. 3A). CDPK1 has been structurally confirmed to exist in two distinct conformers (25): apo (calcium-free, inactive) and calcium-bound (active). Intracellular concentrations of calcium in tachyzoites have been estimated to be between 50 and 100 nM (30); however, we observed kinase activity at relatively high concentrations of between 5 and 50 µM calcium (Fig. 3B). On the basis of these data, we induced binding of TgDJ-1 under low (50 nM)-calcium conditions, followed by the addition of ATP and peptide substrate and a high calcium concentration (500 µM) to stimulate full activation of the kinase. To test if TgDJ-1 affected the activity of the kinase, we titrated wild-type TgDJ-1 against CDPK1 and discovered that TgDJ-1 increased the activity of CDPK1 in a dose-dependent manner (Fig. 3C). We did not observe this effect in the absence of the peptide substrate, indicating that the effect was most likely due to increased phosphorylation of the peptide rather than to increased CDPK1 autophosphorylation or phosphorylation of TgDJ-1.

We then tested whether the presence of TgDJ-1 was sufficient to activate CDPK1, whether activation still required calcium, and if any effect observed was additive or synergistic. We performed an assay where different concentrations of TgDJ-1 (low, 0.26 µM; high, 2.6 µM) were added to CDPK1, followed by the addition of the peptide substrate, ATP, and 500 µM EDTA or a low (50 nM) or high (500 µM) concentration of calcium. Neither the EDTA-treated samples nor the low-calcium-treated samples exhibited any appreciable activity, and this was unaffected by the presence of TgDJ-1 (Fig. 3D, EDTA and 50 nM sections of scatter plots). We once again observed that high concentrations of calcium activated the kinase and that TgDJ-1 stimulated this activity in a dose-dependent manner (Fig. 3D; 500 µM section of scatter plot). Thus, the calcium and TgDJ-1 effects on CDPK1 activity appear to be synergistic in nature. Assays performed with TgDJ-1 in the absence of CDPK1 produced no signal, confirming that the observed signals for the TgDJ-1-plus-CDPK1 (TgDJ-1+CDPK1) samples were not due to contaminants in the TgDJ-1 recombinant protein preparations (Fig. 3D).

We hypothesized that the activation observed with different concentrations of calcium was a reflection of CDPK1 transitioning from the apo state to the fully active conformer (Fig. 3B) (25). To investigate how these different calcium-bound CDPK1 conformers responded to TgDJ-1 potentiation, TgDJ-1 was allowed to associate with CDPK1 in the presence of a titration of calcium. We observed that in high concentrations of calcium, the potentiation effect of TgDJ-1 was blocked, with kinase activity returning to the level seen with kinase alone (Fig. 3E).

Combining our insights from the crystal structure and the precipitation data, we hypothesized that if the association of TgDJ-1 with CDPK1 were necessary to potentiate kinase activity, TgDJ-1 mutations predicted or directly shown to affect the interaction could alter its ability to influence kinase activity. We assayed a panel of TgDJ-1 mutants for their ability to potentiate CDPK1 activity. These included the arginine dyad mutants (R47A, R27A/R47A) and an additional mutant where the reactive cysteine C104 was mutated to an alanine (C104A) that we predicted could participate in the interaction on the basis of the crystal structure. Consistent with our precipitation data, the R47A mutant that failed to associate with CDPK1 (Fig. 1F) was defective in its ability to potentiate CDPK1 activity (Fig. 3F). The double-arginine mutant (R27A/R47A) that constitutively associated with CDPK1 irrespective of its phosphorylation state (Fig. 1F) potentiated kinase activity to a degree similar to that seen with wild-type TgDJ-1 (Fig. 3F). Interestingly, the C104A mutant displayed a similar loss in its ability to potentiate kinase activity, indicating that this critical and highly conserved residue may also contribute to the interaction.

Given the structural and functional data suggesting a potential biological role for the DJ-1/CDPK1 complex, we moved to an intact cell system to validate the relevance of the complex. We began by deleting the native dj-1 (TGGT1_214290) locus using double homologous recombination in the RHΔku80 parasite line to replace the gene with the HXGPRT drug selection cassette (Fig. 4A). We confirmed successful deletion of the dj-1 locus by diagnostic PCR analysis (Fig. 4B). To better assess the knockout parasites (the ΔTgdj-1 mutants), we raised a murine polyclonal antibody against E. coli-expressed recombinant TgDJ-1 (18). When used to probe Western blots of whole tachyzoite lysates, the antibody recognized a single species of ~22 kDa (Fig. S2A), consistent with its predicted size of 19.5 kDa, and biochemical fractionation showed that TgDJ-1 partitioned exclusively to the cytosolic compartment (Fig. S2B). This result was supported by its localization in immunofluorescence assays (Fig. S2C). We isolated clonal populations of knockout parasites by limiting dilution, and Western blot analysis confirmed the absence of the TgDJ-1 protein (Fig. 4C). Plaque assays indicated that ΔTgdj-1 parasites grew normally in terms of both plaque number and size (Fig. 4D), suggesting that, in this genetic background and under the conditions used for cell culture, TgDJ-1 is nonessential or can be readily compensated for. Likewise, the motility of ΔTgdj-1 parasites in two dimensions (2D) was indistinguishable from the motility of the wild type (Fig. 4E).

The original identification of TgDJ-1 as the target of the small molecule WRR-086 suggested that it was involved in the regulation of microneme secretion (18). With this in mind, we investigated if microneme secretion was altered in ΔTgdj-1 parasites. Microneme secretion can be monitored over time by detection of microneme protein 2 (MIC2) in culture supernatant following its release by organelle exocytosis. Basal microneme secretion in two knockout clones selected from two independent transfections (clones C10 and F6) was increased compared to that seen with the RHΔku80 parent (WT) line (Fig. 4F and G and 5A and B). As both of the ΔTgdj-1 clones exhibited the same phenotype, we continued our studies focusing on clone F6. Consistent with atypical microneme secretion, ΔTgdj-1 parasites were moderately less proficient at invading host cells during a time-restricted attachment/invasion assay (Fig. 4H). As microneme secretion provided the clearest readout for the phenotype resulting from the loss of TgDJ-1, we focused on this assay for the remainder of the study. To determine if the absence of TgDJ-1 was the only factor responsible for the microneme secretion defect, we performed a genetic rescue through complementation performed with an N-terminally FLAG-tagged TgDJ-1 protein expressed under the control of the GRA1 promoter and targeted to the UPRT locus of ΔTgdj-1 parasites (ΔTgdj-1^WTcomp) (Fig. S3A). Expression of FLAG-tagged DJ-1 was confirmed by Western blotting (Fig. S3B), and a single clone (D3) selected for further characterization. Surprisingly, ΔTgdj-1^WTcomp parasites had reduced levels of basal microneme secretion compared to both wild-type and knockout parasites (Fig. 5A and B). As there are two distinct modes of microneme secretion, we next tested whether induced secretion was similarly affected by complementing the transgene. Ethanol-stimulated microneme secretion in wild-type parasites was induced to a higher level in ΔTgdj-1 parasites than in the parental strain (Fig. 5C and D). Furthermore, the rescue of TgDJ-1 expression in the complemented ΔTgdj-1^WTcomp cell line decreased the response of secretion to ethanol stimulation, reducing it to levels below those seen with WT parasites (Fig. 5C and D). In an attempt to better understand these results, we examined the expression levels of TgDJ-1 and found that it was dramatically overexpressed in the ΔTgdj-1^WTcomp parasites compared to native TgDJ-1 expression levels (Fig. S3C). This was likely due to the use of the GRA1 promoter to drive its expression. Despite the observed differences in microneme secretion, ΔTgdj-1^WTcomp parasites grew normally in terms of plaque number and size (Fig. S3D). Combined, these data confirm that TgDJ-1 plays a regulatory role in both basal and induced microneme secretion.

To confirm that the TgDJ-1/CDPK1 interaction occurred in cells, we performed coimmunoprecipitations (co-IPs) using parasite strains expressing epitope-tagged versions of TgDJ-1 (FLAG-TgDJ-1 in ΔTgdj-1^WTcomp parasites) to pull down endogenous CDPK1 or parasite strains expressing hemagglutinin-tagged CDPK1 (CDPK1-HA) (31) to pull down endogenous TgDJ-1. Though suggestive of an interaction, the results from these coimmunoprecipitations were inconsistent, leading us to speculate that the interaction may be of low affinity or that the context dependence of the interaction was too sensitive or transient to capture in these coimmunoprecipitations.

On the basis of the known role for CDPK1 as an essential regulator of microneme secretion (15), our data showing that TgDJ-1 interacts with CDPK1 (Fig. 1D and E and 2), and the observed microneme secretion phenotype in the ΔTgdj-1 and ΔTgdj-1^WTcomp parasites (Fig. 4F and G and 5A and B and C and D), we examined the levels of CDPK1 expression in these different parasite lines. Intriguingly, CDPK1 protein levels showed a striking inverse relationship to the amount of TgDJ-1 present in the cell; in the absence of TgDJ-1, CDPK1 was more abundant at the protein level, with the amount of CDPK1 being ~3 times greater in ΔTgdj-1 parasites than in the wild type (Fig. 5E and F). To determine if this was a general mechanism by which parasites coped with the loss of TgDJ-1, we assessed the level of CDPK1 expression in a second ΔTgdj-1 clone isolated from an independent transfection and found that it was similarly increased (Fig. S4). These data suggested that compensating for the loss of TgDJ-1 by increasing the amount of CDPK1 present in the cell is a potential mechanism of adjustment within the pathway. In agreement with this, CDPK1 levels in ΔTgdj-1^WTcomp parasites were lower than in the knockout strain (Fig. 5E and F).

We sought to test if mutations that affected the DJ-1/CDPK1 complex and the ability of TgDJ-1 to potentiate CDPK1 activity also influenced its ability to regulate microneme exocytosis. Initial attempts to introduce mutant dj-1 alleles into parasites proved unsuccessful, suggesting that the arginine dyad mutations were not tolerated when expressed in ΔTgdj-1 parasites, eliminating the possibility of examining the function of these mutations by direct gene replacement strategies. To overcome this technical problem, we attempted to generate dj-1 alleles N-terminally tagged with an FKBP-derived destabilization domain (32). However, the incorporation of this domain resulted in a fusion protein that was not stably expressed. In our previous work, we demonstrated that the small molecule WRR-086 targets a nonconserved cysteine at position 127 (Cys127) (18). Mutation of this cysteine in the native dj-1 locus protected parasites from the effect of the drug. This led us to consider an alternative strategy using the C127A mutation to engineer drug-insensitive dj-1 alleles that could be expressed in the background of wild-type TgDJ-1, making the parasites merodiploid for the gene. The wild-type allele could then be selectively targeted with WRR-086, allowing any phenotype associated with other mutations carried on the C127A background to be revealed through complementation (Fig. 6A). We anticipated that these mutant proteins would be phenotypically dominant in the presence of drug, allowing us to test the functional contribution of the mutations found to affect the ability of TgDJ-1 to potentiate CDPK1 activity in vitro with kinase function in cells. We expected that any mutation that affected the nature of the interaction would be unable to functionally complement the loss of the wild-type protein.

We transfected tachyzoites with constructs to overexpress a FLAG-tagged TgDJ-1^C127A mutant allele in the background of the endogenous TgDJ-1 through use of the GRA1 promoter to drive the transgene. Using this approach, we generated the TgDJ-1/TgDJ-1^C127A merodiploid line (Fig. 6A), as well as lines that expressed the C104A, R47A, and R27A/R47A mutations maintained in the background of the drug-insensitive C127A allele, referred to as TgDJ-1/TgDJ-1^C127A/C104A, TgDJ-1/TgDJ-1^C127A/R47A, and TgDJ-1/TgDJ-1^{C127A/R27A/R47A}, respectively. We isolated isogenic populations of parasites expressing FLAG-tagged TgDJ-1 mutant alleles and confirmed equivalent relative expression levels of the mutant proteins by Western blotting (Fig. 6B). Unlike the results seen with the ΔTgDJ-1 parasites, we found that the levels of CDPK1 were unaffected by the introduction of these mutant proteins (Fig. 6B). Indirect immunofluorescence assay (IFA) results confirmed a diffuse cytosolic distribution for the overexpressed mutant proteins that matched that observed for native TgDJ-1 (Fig. 6C and S2C). All merodiploid TgDJ-1 parasite lines grew normally, generating plaques that were equal in number and size to wild-type plaques (Fig. 6D). Together, these data confirmed that the mutant proteins were expressed and localized appropriately to potentially complement the native gene product. We next tested the ability of the mutant alleles to rescue inhibition of microneme secretion by WRR-086, and, importantly, microneme secretion was similar to wild-type secretion levels in the absence of drug for all merodiploid lines (Fig. S5). As anticipated, TgDJ-1/TgDJ-1^C127A parasites were resistant to the effect of WRR-086, exhibiting relatively normal levels of microneme secretion in the presence of the drug (Fig. 6E). It should be noted that, although TgDJ-1/TgDJ-1^C127A parasites were resistant to WRR-086, the level of rescue achieved was not complete, suggesting that the overexpression of the C127A allele could not fully complement the loss of function of the native protein. Parasites expressing the TgDJ-1^C127A/C104A allele were partially resistant to the effects of WRR-086, but the level of phenotypic rescue achieved was lower than that seen with the TgDJ-1^C127A allele (Fig. 6E), supporting the kinase assay data indicating a likely critical role for this residue in the potentiation of CDPK1 activity and TgDJ-1 function in microneme secretion. As expected, and in accordance with our in vitro biochemical data showing that the mutation of one arginine of the dyad prevented association of TgDJ-1 with the kinase, the TgDJ-1^C127A/R47A allele was incapable of rescuing microneme secretion in the presence of drug (Fig. 6E). Intriguingly, TgDJ-1/TgDJ-1^{C127A/R27A/R47A} parasites displayed drug sensitivity that was similar to that seen with parasites expressing the single arginine mutant (TgDJ-1/TgDJ-1^C127A/R47A; Fig. 6E). These data confirm that mutations that disrupt or change the features of the TgDJ-1/CDPK1 complex also influence the ability of the protein to regulate microneme exocytosis in vivo.

The parasite strains used in this study are listed in Table 1.

WRR-086 was synthesized as previously described (18).

Type I T. gondii strain RH was maintained by passage through confluent monolayers of human foreskin fibroblasts (HFFs). Host cells were cultured as previously described (35). Parasites were harvested for use in assays by either syringe lysis of infected HFF monolayers or collection of parasites from culture supernatant after spontaneous lysis of the monolayer.

Transgenic parasite strains were made by electroporating the T. gondii strain RHΔku80 (36) with 15 μg of linearized plasmid encoding the construct of interest and selecting for mycophenolic acid (MPA)-xanthine- or 5-fluorodeoxyuridine (FUdR)-resistant parasites, as described previously (37, 38). Clonal parasites were selected by limiting dilution. Integration was verified using a locus-specific PCR. Details of all strains used can be found in Table 1.

Plasmids for the construction of vectors described were obtained from the Boothroyd laboratory. These include pTKO2 (used to generate the ΔTgDJ-1 parasite strain), pUPRT (used to complement ΔTgDJ-1 parasites), and pGRA (used to generate the merodiploid strains). For the generation of the ΔTgDJ-1 parasite strain, the T1 sequence was amplified using primers 5′-TGTTTCATGACGTACCCGACACAGCAC-3′ and 5′-GCGACGGGCTGCGGTAGG-3′ and the T2 sequence was amplified using primers 5′-AAGCTTAAGCTTGAGTCGTTACAGCCTAATGAAGCG-3′ and 5′-GGGCCCCAAGTCCGAATCTGCCTAACTCC-3′.

In all cases, mutagenesis was performed by site-directed mutagenesis (Stratagene) or using the Phusion site-directed mutagenesis technique (Thermo).

For all of the cell assays, compound treatments were performed as follows: intracellular parasites were released from heavily infected host cell monolayers by syringe lysis. Hanks balanced salt solution (HBSS)-washed parasites were incubated with hydrogen peroxide or compound for 15 min at 37°C at the concentration described in the text (unless otherwise stated) before being used for the appropriate cell assay.

Toxoplasma plaque formation was assayed as described previously (39).

Microneme secretion assays quantifying the release of MIC2 into the extracellular medium were performed as described previously (40), with modifications described previously (35). For quantification, the following approach was used: mean pixel intensities for GRA7 loading control bands were measured in ImageJ. Band intensities were initially normalized to the background intensity measurements of a sample region from the same lane on the blot. GRA7 background subtracted intensity readings were then compared to the readings from the control lane (e.g., zero time point, dimethyl sulfoxide [DMSO] treatment). This generated the loading adjustment factor to normalize for loading against constitutively secreted GRA7. Mean pixel intensities for MIC2 bands were then measured in ImageJ. Band intensities were initially normalized to the background of a sample region from the same lane. Background subtracted intensity readings were then multiplied by the adjustment factor determined above to adjust for loading. Mean values are presented for 3 replicates.

2D motility and the attachment/invasion of host cells were assayed as described previously (35).

RHΔku80 tachyzoites (50 × 10⁸) were washed with HBSS, pelleted, and snap-frozen. The cell pellet was thawed and resuspended in 200 µl of hypotonic lysis buffer (50 mM HEPES, pH 7) and incubated on ice for 30 min. Samples were then spun at 100 × g for 1 h. The supernatant was separated as the “soluble” fraction, and the pellet was resuspended in 200 µl of phosphate-buffered saline (PBS) as the “insoluble” fraction. Fractions were solubilized with equivalent amounts of reducing Laemmli SDS-PAGE sample buffer, and equivalent volumes were analyzed by SDS-PAGE and Western blotting.

Immunofluorescence assays were performed as described previously (35) with the antibodies indicated in relevant figure legends.

Cell pellets of HBSS-washed extracellular tachyzoites were obtained for the relevant parasite strains. Pellets were thawed in lysis buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 0.5% NP-40) and incubated on ice for 30 min. Extracts were clarified by centrifugation at 14,000 rpm and 4°C. Supernatants were separated and retained as the cell extracts. The protein concentration of extracts was quantified by bicinchoninic acid (BCA) assay to ensure that equivalent amounts of protein were available for use in the respective comparative assays.

Immunoprecipitations (IPs) were performed as follows (unless otherwise specified in the text). Typically, 200 µg cell extract protein was added to 200 µl of IP buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 0.5% NP-40). Anti-HA beads (Pierce) (40 μl) were then added to the sample, and the IPs were performed according to the instructions of the manufacturer. For IPs using conditioned extracts and recombinant FLAG-tagged TgDJ-1, the samples were adjusted to be reducing and calcium depleted (2 mM DTT, 5 mM EGTA), oxidizing and calcium depleted (H₂O₂ [at concentrations specified in the figure legend or text], 5 mM EGTA), reducing and calcium rich (2 mM DTT, 5 mM CaCl₂), or oxidizing and calcium rich (H₂O₂ [at concentrations specified in the figure legend or text], 5 mM CaCl₂). Once adjusted, IP samples containing the extracts were incubated for 15 min on ice. FLAG-TgDJ-1 (5 μg) was then added to each, and the samples were incubated for a further 5 min on ice before the addition of 40 µl of anti-HA agarose beads (Pierce), with the IP performed according to the instructions of the manufacturer. Specific extract conditions were maintained for all wash steps unless otherwise stated.

The Toxoplasma gondii DJ-1 gene (TgDJ-1) (18) was PCR amplified using primers that introduced a 5′ NdeI restriction site and a 3′ BamHI site. The amplified insertion was digested with NdeI and BamHI and cloned between these restriction sites in the bacterial expression vector pET15b (Novagen). The C104S and C127S point mutations were made using standard site-directed mutagenesis with mutagenic primers. The constructs were transformed into chemically competent E. coli strain BL21(DE3) (Novagen Merck, Darmstadt, Germany,) for protein expression. The recombinant protein has a thrombin-cleavable N-terminal hexahistidine tag to facilitate purification using Ni²⁺ metal affinity chromatography. Protein expression and purification were performed as described previously (41), except that all buffers contained 2 mM DTT. The absorbance of the purified sample was measured at 280 nm in DTT-free buffer and converted to a protein concentration using a calculated extinction coefficient for TgDJ-1 at 280 nm (ε₂₈₀) of 7,825 M⁻¹ cm⁻¹ (Expasy) and a molecular mass of 29,296 Da. The purified protein was concentrated to 20 mg/ml in storage buffer (25 mM HEPES [pH 7.5], 100 mM KCl, 2 mM DTT) using a stirred pressure cell concentrator with a 10-kDa molecular mass cutoff. The purified protein ran as a single band on an overloaded SDS-PAGE gel stained with Biosafe Coomassie blue (Bio-Rad, Hercules, CA). The concentrated protein was divided into aliquots, rapidly frozen in liquid nitrogen, and stored at −80°C.

For GST-CDPK1 constructs, CDPK1 was amplified from cDNA and cloned into pGEX-6P1 for expression with an N-terminal GST tag. Protein was expressed as described above. For preparations of dephosphorylated GST-CDPK1, the vector was coexpressed with a construct for the expression of lambda phosphatase. Following expression, GST-tagged recombinant proteins were purified using glutathione-Sepharose beads. Following binding, beads were washed extensively with lysis buffer with 5 mM EGTA and 2 mM DTT and then stored at −20°C as a 50% slurry in lysis buffer with 5 mM EGTA and 2 mM DTT and 50% glycerol.

Glutathione beads (10 μl) coated with GST, GST-CDPK1, or GST-CDPK1 coexpressed with lambda phosphatase were added to 200 µl of pulldown buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 5 mM EGTA, 2 mM DTT) and 1 µg of recombinant TgDJ-1 (wild-type or mutants as specified in the text or the figure legends). Samples were incubated with agitation performed at 4°C for 1 h. Beads were then washed three times with 500 µl of pulldown buffer, and associated proteins were solubilized with equal volumes of reducing Laemli sample buffer and submitted for analysis by SDS-PAGE, Coomassie staining, and Western blotting.

Polyclonal murine antibodies were raised against TgDJ-1 as previously described (31). Polyclonal murine antibodies against CDPK1 were generated in a previous study (31).

E. coli lysates overexpressing TgDJ-1 were diluted to a 2 mg protein/ml solution in PBS. Each sample (2× 0.5-ml aliquots) was treated with an iodoacetamide probe at a concentration of either 10 or 100 µM using 5 µl of a 1 mM or 10 mM stock in DMSO. The labeling reaction mixtures were incubated at room temperature for 1 h. Click chemistry was performed by the addition of a 150 µM concentration of either a light-tobacco etch virus (TEV) tag (for 10 µM iodoacetamide [IA] samples) or a heavy TEV tag (for 100 µM IA samples) (15 µl of a 5 mM stock) as described in reference 19. Samples were allowed to react at room temperature for 1 h. After the click chemistry step, the light-tag- and heavy-tag-labeled samples were mixed together and centrifuged (5,900 × g, 4 min, 4°C) to pellet the precipitated proteins. The pellets were washed twice in cold MeOH, after which the pellet was solubilized in PBS containing 1.2% SDS via sonication and heating (5 min and 80°C). Upon resolubilization, the proteome sample was subjected to streptavidin enrichment and on-bead trypsin and TEV digestion as previously described (19). The resulting probe-labeled peptides obtained after TEV digestion were analyzed on a linear trap quadrupole (LTQ)-Orbitrap Discovery mass spectrometer (Thermo, Fisher) using the Mudpit protocol as previously reported (19). Peptide identification was achieved using the SEQUEST algorithm and quantified using CIMAGE as reported previously (19).

The pK_a value of the reactive cysteine residue (Cys104) in TgDJ-1 was measured by monitoring the absorption of the thiolate anion at 240 nm as a function of pH (20, 21). A triple buffer (10 mM boric acid, 10 mM sodium citrate, 10 mM sodium phosphate) was prepared, and TgDJ-1 protein was added to reach a final concentration of 10 μM. The pH of each sample was changed by the addition of 0.5- to 1.0-μl aliquots of 0.5 N or 5.0 N NaOH. After each addition of NaOH, the absorbance at 240 and 280 nm was measured using a Cary 50 spectrophotometer (Varian, Inc., Palo Alto, CA). The A₂₈₀ value was used to normalize the A₂₄₀ measurement to the protein concentration in the cuvette. After spectrophotometric measurement, the pH of each sample was measured using an Orion micro-pH electrode (Thermo Fisher, Waltham, MA, USA) that had been calibrated at the beginning of the experiment. Measured absorbance values were converted to an extinction coefficient at 240 nm (ε₂₄₀) using a calculated ε₂₈₀ value for TgDJ-1 of 7,825 M⁻¹ cm⁻¹ (Expasy). The experiment was performed in triplicate for wild-type protein and once for the C127S and C104S mutants. The data were fitted to a modified version of the Henderson-Hasselbalch equation (22) in Prism (GraphPad Software, Inc.). The reported pK_a values and associated errors are derived from the fit procedure.

All recombinant protein was expressed in E. coli Rosetta2(DE3) (EMD Biosciences). N-terminally His₆-tagged TgDJ-1 was purified on nickel-nitrilotriacetic acid (Ni-NTA) resin (Qiagen), and the His₆ tag was removed by overnight cleavage using thrombin (Hematologic Technologies) at 4°C. All proteins were further purified by anion exchange and gel filtration chromatography. After elution from Ni-NTA, TgDJ-1 was kept in fresh 10 mM DTT for all subsequent steps.

High-quality crystals of TgDJ-1 grew from a 1:1 mixture of protein (5 mg/ml)–10 mM HEPES (pH 7.0)–100 mM NaCl–10 mM DTT and 1.5 M ammonium citrate tribasic (Hampton Research). Crystals were flash frozen in mother liquor for data collection.

The diffraction data were collected at beamline 11.1 of SSRL (the Stanford Synchrotron Radiation Laboratory) at a wavelength of 0.979 Å and a temperature of 100 K. Indexing, integration, and scaling of the diffraction data were performed using the XDS suite (42). Initial phases were obtained by molecular replacement using Phaser (43) and searching with a homology model of TgDJ-1 created from HsDJ-1 with Modeller (44). Manual rebuilding in Coot (45) and refinement in refmac (46, 47) led to a final 2.08-Å structure which was deposited in the Protein Data Bank. The structure showed good stereochemistry (97.5% favored) from Ramachandran plots as validated by the program MOLPROBITY (48).

All structural images were created using PyMOL 1.7.0 (Schrödinger, LLC).

Kinase assays were performed using a kinase Glo kit (Promega) and were optimized according to the instructions of the manufacturer with the following modifications. A standard assay would include two incubation steps: an initial 30-min incubation at 37°C of a reaction mixture typically containing recombinant CDPK1 kinase, 50 nM CaCl₂, kinase buffer (50 mM Tris [pH 7.4], 10 mM MgCl₂, 2 mM DTT), and TgDJ-1 (unless otherwise stated). Following this initial incubation, 10 µM ATP, 10 µM syntide, and 500 µM CaCl₂ were added (unless otherwise stated) and the reaction mixture was incubated for a further 30 min at 37°C. Assays were then processed according to the instructions of the manufacturer.

The 2.08-Å structure determined in this work was deposited in the Protein Data Bank (PDB accession no. 4XLL ).