A forward genetic screen identifies a negative regulator of rapid Ca2+-dependent cell egress (MS1) in the intracellular parasite Toxoplasma gondii

Toxoplasma gondii, like all apicomplexan parasites, uses Ca2+ signaling pathways to activate gliding motility to power tissue dissemination and host cell invasion and egress. A group of “plant-like” Ca2+-dependent protein kinases (CDPKs) transduces cytosolic Ca2+ flux into enzymatic activity, but how they function is poorly understood. To investigate how Ca2+ signaling activates egress through CDPKs, we performed a forward genetic screen to isolate gain-of-function mutants from an egress-deficient cdpk3 knockout strain. We recovered mutants that regained the ability to egress from host cells that harbored mutations in the gene Suppressor of Ca2+-dependent Egress 1 (SCE1). Global phosphoproteomic analysis showed that SCE1 deletion restored many Δcdpk3-dependent phosphorylation events to near wild-type levels. We also show that CDPK3-dependent SCE1 phosphorylation is required to relieve its suppressive activity to potentiate egress. In summary, our work has uncovered a novel component and suppressor of Ca2+-dependent cell egress during Toxoplasma lytic growth.

The phylum Apicomplexa comprises a large group of obligate intracellular parasites, many of which have medical and agricultural significance. Plasmodium spp. are the causative agents of malaria, whereas Cryptosporidium spp. cause severe diarrhea. Toxoplasma gondii is the most ubiquitous of all apicomplexan parasites, chronically infecting Ͼ30% of the human population. Acute Toxoplasma infection can cause congenital birth defects, epilepsy, hydrocephalus, and stillbirth. Chronic infection is a leading cause of progressive blindness in some countries and can lead to neurological disease in immunocompromised patients (1)(2)(3).
To survive in the host, Toxoplasma, like all apicomplexan parasites, must actively invade host cells to initiate replication. After several rounds of division, unknown signals trigger tachyzoites to actively exit (egress from) host cells, allowing for subsequent rounds of invasion and intracellular growth. Thus, invasion and egress are critical for the establishment and persistence of infection and therefore represent potential therapeutic targets to treat infection.
Like all apicomplexan species, Toxoplasma initiates motility, invasion, and egress via an intracellular signaling cascade that activates exocytosis of adhesins from the microneme organelles and stimulates the glideosome, an actomyosin-based motor, required to drive motility. Ca 2ϩ , cGMP, and phospholipid signaling pathways all play important and interconnected roles in activating microneme release and parasite motility (4 -6). Critical to transducing intracellular Ca 2ϩ flux is a group of "plantlike" Ca 2ϩ -dependent protein kinases (CDPKs) 2 (7). CDPKs are direct fusions between a kinase and calmodulin-like domains, an architecture that is not found in metazoans. CDPKs are localized to several intracellular cellular compartments where they presumably act on a defined set of substrates, altering their activity upon phosphorylation. CDPK1, a cytosolically localized ortholog, activates exocytosis of micronemes, releasing adhesins onto the tachyzoite surface, thus initiating gliding motility (8). CDPK3, on the other hand, localizes to the inner side of the parasite plasma membrane and potentiates rapid, Ca 2ϩ ionophore-induced egress (9 -11). In related Plasmodium species, Plasmodium falciparum CDPK1 has been implicated in parasite development and host cell invasion in blood stages (12)(13)(14), CDPK5 is required for parasite egress from erythrocytes (15), and Plasmodium berghei CDPK3 is necessary for motility during sexual stage development (16,17). Interestingly, CDPK orthologs also play important roles in other cellular processes, such as parasite division, carbon metabolism, and stage differentiation (18 -21).
Little is understood about how CDPKs activate motility. Recent efforts have utilized techniques such as substrate capture to identify the direct targets of CDPKs (22), and proximity ligation to identify putative interacting partners, the latter indi-cating that CDPK3 may directly phosphorylate components of the glideosome to regulate motility (23). Quantitative global phosphoproteomics have identified phosphoproteins that may be regulated by CDPK3 during egress, revealing a diversity of signaling pathways (24). In this study, we utilize a forward genetic screen to reveal novel components of Ca 2ϩ -dependent egress in Toxoplasma. By screening for gain-of-function mutants in a cdpk3-deficient background, we identify suppressor of Ca 2ϩ -dependent egress 1 (SCE1), which appears to act as a suppressor during egress. Further, we show that the functions of CDPK3 and SCE1 are interconnected and that phosphorylation of SCE1 by CDPK3 controls its suppressive activity during host cell egress.

Identification of a suppressor of egress
To identify new factors involved in Toxoplasma Ca 2ϩ -dependent egress, we took advantage of previously described CDPK3 knockout (⌬cdpk3) parasites (supplemental Table S1) (11), which show negligible egress following stimulation with the Ca 2ϩ ionophore A23187. We devised a strategy to enrich and isolate parasites from the egress-deficient ⌬cdpk3 background that reverted to an egress-competent phenotype (hereafter referred to as ⌬cdpk3 rev ) through the acquisition of random mutations (Fig. 1A). Egress-deficient ⌬cdpk3 parasites were split into 10 independent populations and mutagenized using the chemical mutagen ethyl methane sulfonate (EMS) (hereafter referred to as ⌬cdpk3 EMS.1 -⌬cdpk3 EMS.10 ). Mutagenized populations were then treated with the Ca 2ϩ ionophore A23187, which stimulates Ca 2ϩ -dependent egress in wild-type parasites, in medium supplemented with dextran sulfate to block reinvasion. The host cell monolayer was then washed to remove any parasites that had regained the capacity to egress, which were then transferred into a fresh culture flask for subsequent expansion.

SCE1 during Toxoplasma egress
likely arisen from an initial single ⌬cdpk3 rev mutant and, thus, that the etiological mutation selected for by our strategy was rare. This is congruent with our finding that only 2 of the 10 mutagenized populations gave rise to ⌬cdpk3 rev parasites.
To identify the etiological mutation for the rescue-of-egress phenotype in ⌬cdpk3 rev parasites, we applied a stepwise ranking process with the following assumptions: the mutation must be non-synonymous; given the low rate of reversion across 10 populations and the identical egress levels in rescued populations, the same gene was mutated between populations; the mutated gene has different mutations in the different populations; the mutation is an SNP resulting in loss of protein function; and the mutated gene encodes a protein that localizes to the plasma membrane, like CDPK3. The full list of SNPs and the results from each of the ranking steps are presented in supplemental Table S2. Following this, independent mutations were identified in the coding sequence of the hypothetical protein TGME49_269260 in all EMS-mutagenized populations. TGME49_269260 is conserved in the closely related apicomplexan parasites Neospora caninum and Hammondia hammondi but contains no computationally predicted domains other than five C-terminal transmembrane domains (Fig. 1C). In ⌬cdpk3 EMS.7 , ⌬cdpk3 EMS.7.1 , and ⌬cdpk3 EMS.7.2 , the mutation in the TGME49_269260 gene resulted in the missense mutation H565R, whereas, in ⌬cdpk3 EMS.10 , ⌬cdpk3 EMS.10.1 and ⌬cdpk3 EMS.10,2 , the mutation introduced a stop codon near the N terminus (Gln106STOP). We therefore hypothesized that mutation of the coding sequence of TGME49_269260 in both ⌬cdpk3 EMS.7 and ⌬cdpk3 EMS.10 parental populations resulted in an inactive gene product and, therefore, that this gene likely encoded a factor that inhibited Ca 2ϩ -dependent egress in ⌬cdpk3 parasites. The protein product of TGME49_269260 was therefore named suppressor of Ca 2ϩ egress 1 (SCE1).

SCE1 is a suppressor of CDPK3-dependent egress
To begin investigation of the role of SCE1 in Toxoplasma, the endogenous locus was tagged with a 5ϫTy epitope tag in RH:⌬ku80 parasites. By IFA, SCE1-Ty localized to the parasite periphery, including the residual body (a pocket of plasma membrane left behind during tachyzoite cell division), suggesting that SCE1 is associated with the plasma membrane (Fig.  2Ai). This localization is similar to that seen previously for CDPK3 (11). SCE1 is predicted to be ϳ75 kDa, but, by reducing Western blot, this protein runs at approximately double the size (Fig. 2Aii).
To investigate the role of SCE1 in parasite egress, we genetically deleted sce1 in both RH:⌬ku80 and ⌬cdpk3 lines (giving rise to the lines ⌬sce1 and ⌬cdpk3⌬sce1, respectively) (supplemental Fig. S1 and Table S1). We then tested whether the loss of SCE1 in ⌬cdpk3 parasites recapitulated the ⌬cdpk3 rev phenotype. Following stimulation with A23187 for 3 min, parasites egressed from ϳ65% of ⌬cdpk3⌬sce1 parasitophorous vacuoles, similar to what was seen in EMS-mutagenized strains (Fig.  2B). ⌬sce1 parasites showed no significant difference in egress compared with wild-type parasites under these conditions.
We then tested the effect of complementation on egress capacity. ⌬sce1:SCE1 wt and ⌬cdpk3:SCE1 wt showed no significant difference in A23187 sensitivity relative to their parental strains, demonstrating that overexpression of SCE1 wt did not modify egress competence (Fig. 2Cii). However, complementation with SCE1 wt -HA restored the egress-deficient phenotype in ⌬cdpk3⌬sce1, ⌬cdpk3 EMS.7.1 , and ⌬cdpk3 EMS.10.1 . Together, these data strongly suggest that loss of SCE1 activity in ⌬cdpk3 rev confers the rescue-of-egress phenotype.

SCE1 does not significantly affect sensitivity to egress stimuli
We then wanted to determine what role SCE1 plays during egress. We previously hypothesized that CDPK3 acts to amplify egress capacity upon A23187 stimulation (11). We therefore reasoned that SCE1 may act in the opposite manner to dampen Ca 2ϩ or cGMP signaling and the sensitivity of parasites to agonists of these pathways.
To determine whether SCE1 controls the rate of response to egress stimulation, we measured host cell exit at 2.5, 5, and 15 min after A23187 stimulation (Fig. 3A). At 2.5 min, egress capacity mimicked the results above (Fig. 2); however, by 5 min, ⌬cdpk3⌬sce1 showed egress equivalent to wild-type parasites ( Fig. 3A). At 15 min, ⌬cdpk3 showed some egress capacity, and ⌬cdpk3⌬sce1 and wild-type egress remained equivalent. We observed no significant difference in egress capacity at any time point between ⌬sce1 and the wild type (Fig. 3A). We also tested whether loss of SCE1 altered the sensitivity of parasites to a range of concentrations of A23187. Here, although SCE1 knockout parasites (⌬sce1 and ⌬cdpk3⌬sce1) showed a trend toward increased sensitivity to low concentrations of A23187, this did not reach statistical significance, as analyzed by twoway ANOVA (Fig. 3B).
We then investigated whether SCE1 plays a role in egress regulated by cGMP signaling. We previously described a potent Apicomplexa-specific PDE inhibitor, BIPPO, which stimulates egress through PKG (42). Using this inhibitor, we and others have also demonstrated a link between cGMP and Ca 2ϩ signaling (21,25,26). To explore the rate of parasite egress following cGMP signaling, we monitored the egress of ⌬cdpk3, ⌬sce1, and ⌬cdpk3⌬sce1 parasites at 2.5, 5, and 15 min after BIPPO treatment (Fig. 3C). ⌬cdpk3⌬sce1 tachyzoites again showed significantly more egress than parental ⌬cdpk3 parasites, demonstrating that SCE1 also acts as a suppressor in cGMP-stimulated egress through a CDPK3-dependent mechanism (Fig. 3C). Although ⌬sce1 tachyzoites appeared to egress slightly higher than wild-type parasites, this was not statistically significant, as assessed by two-way ANOVA. Interestingly, using BIPPO, we found that, unlike A23187 treatment, there was no difference between wild-type and ⌬cdpk3⌬sce1 at 2.5 min. We also examined the sensitivity of parasites to a range of BIPPO concentrations. We again observed a non-statistically significant trend (as analyzed by two-way ANOVA) toward increased egress of ⌬sce1 and ⌬cdpk3⌬sce1 parasites . Sensitivity of SCE1 mutants to egress agonists. Shown is egress of wild-type, ⌬sce1, ⌬cdpk3, and ⌬cdpk3⌬sce1 parasites stimulated with either A23187 or BIPPO. A, egress assay measured at 2.5, 5, and 15 min. There was no difference in egress capacity between the wild type and ⌬sce1 at any time point. B, egress at 5 min over a concentration gradient of A23187. The only statistically significant differences observed are by comparison between all strains and with ⌬cdpk3 for the two highest concentrations of A23187. C, egress in response to BIPPO, measured at 2.5, 5, and 15 min. There was no statistically significant difference between the wild type and ⌬sce1 at any time point. D, egress in response to BIPPO at 5 min over a concentration gradient. The only statistically significant differences observed are by comparison between all strains and ⌬cdpk3 for the two highest concentrations of A23187. Error bars show mean Ϯ S.D. All statistical tests were performed using two-way ANOVA and Tukey correction for multiple comparisons. The data were derived from three independent biological replicates.

SCE1 and CDPK3 regulate phosphorylation signaling during egress, as demonstrated by global quantitative proteomics
CDPK3 influences the phosphorylation of hundreds of proteins, either as direct substrates or downstream of CDPK3 activity (24). We hypothesized that loss of SCE1 allows for CDPK3-independent egress in ⌬cdpk3⌬sce1 by either up-regulating an alternative egress signaling pathway or by alleviating a cellular event that occurs downstream of CDPK3 that is required for subsequent signaling steps to occur. Although SCE1 is not predicted to be a kinase, we hypothesized that it may indirectly control egress by regulating the phosphorylation state of important Ca 2ϩ signaling proteins that act downstream of CDPK3. Investigating this may therefore also identify the relationships between CDPK3 and other Ca 2ϩ -signaling proteins.
lating host cell egress occur within Toxoplasma, so HFF cell debris was removed by differential centrifugation, and human peptides were excluded from analysis (Fig. 4B). A total of 7483 unique peptides were identified, of which 5775 (77.2%) were phosphorylated. 5235 (70.0%) of the total peptides mapped to Toxoplasma proteins, of which 4423 (84.5%) were phosphorylated. This suggests that contaminating human peptides were largely dephosphorylated. Mediancentered SILAC ratios of all quantified Toxoplasma sites showed a normal Gaussian distribution, indicating that the presence of host cell proteins, which were mostly unlabeled, did not significantly skew the expected distributions of SILAC ratios (Fig. 4C).
We then filtered SILAC data for phosphopeptides showing evidence of regulation by CDPK3 or SCE1, according to the log 2 -converted SILAC ratios M/L (⌬cdpk3⌬sce1/wild type), H/L (⌬cdpk3/wild type), and H/M (⌬cdpk3/⌬cdpk3⌬sce1) (Fig.  4D). Only phosphopeptides with a minimum phosphosite localization score of 0.75 were considered (supplemental Table  S3, tab 1). These were first determined to be misregulated (significantly up-or down-regulated relative to the wild type) in ⌬cdpk3 (supplemental Table S3, tab 2). Phosphopeptides were then filtered for those whose phosphorylation state was restored in ⌬cdpk3⌬sce1 to a wild type-like phosphorylation state (e.g. down-regulated in H/L, up-regulated in H/M) (supplemental Table S3, tab 3). Following these criteria, a total of 308 phosphopeptides were misregulated in ⌬cdpk3 parasites relative to the wild type. 182 phosphopeptides that were misregulated in ⌬cdpk3 showed rescue in ⌬cdpk3⌬sce1, representing 59% of all ⌬cdpk3 misregulated phosphopeptides. Compellingly, this represented 84% of all ⌬cdpk3⌬sce1 misregulated phosphopeptides (214 phosphopeptides). Therefore, rescue of egress following knockout of SCE1 in ⌬cdpk3 is associated with the restoration of phosphorylation to a significant proportion of phosphorylation sites to near wild-type levels. This suggests that SCE1 does indeed indirectly regulate the phosphorylation of a large proportion of proteins downstream of CDPK3 activity during egress or that loss of this protein activates an alternative pathway that converges at common substrates for activation of egress.

Phosphopeptides regulated by CDPK3 and SCE1 are predicted to be involved in cell signaling
It has been shown previously that loss of CDPK3 influences the phosphorylation of diverse sets of proteins (24). We therefore aimed to determine whether particular sets of proteins or particular subcellular compartments were disproportionately influenced by CDPK3 and SCE1 activity. Filtered phosphopeptides described above were further grouped as follows: "⌬cdpk3-misregulated," all that fit the minimum criteria of being up-or down-regulated in ⌬cdpk3 relative to the wild type; "⌬sce1-restored," all that showed rescue to near wild-type levels in ⌬cdpk3⌬sce1 relative to ⌬cdpk3; "⌬sce1-unrestored," all that showed no rescue in ⌬cdpk3⌬sce1; and "⌬cdpk3-independent," all that showed no significant change in ⌬cdpk3 (supplemental Table S3, tabs 2-5, respectively). GO terms were annotated for all proteins where phosphopeptides were detected from within these groups. GO term enrichment analysis was performed, comparing representation of GO terms within each phospho-protein group against a reference set of the whole Toxoplasma annotated proteome.
Several GO process terms were significantly enriched among proteins containing ⌬cdpk3-misregulated phosphosites compared with a reference set of the whole Toxoplasma annotated proteome (Fig. 5A). These included terms that are consistent with involvement in regulation of signaling processes, such as cell signaling processes ("signal transduction," "cell communication," and "intracellular signal transduction"), turnover of cyclic nucleotides ("phosphoric diester hydrolase activity"), and signaling through phospholipases ("phospholipase activity"). Interestingly, a number of GO process terms associated with nuclear transport were enriched in this dataset ("nuclear transport" and "nucleocytoplasmic transport"). Nuclear processes such as transcription are unlikely to play a direct role in the up-regulation of egress, which normally acts very rapidly (1-2 min) but may be altered as a downstream consequence of changes to egress efficiency or Ca 2ϩ signaling following constitutive knockout of CDPK3.
Many of these same GO terms were enriched for proteins containing ⌬sce1-restored phosphosites (Fig. 5B). However, certain terms were lacking, most noticeably those having to do with sphingolipid metabolism ("glycosphingolipid metabolic process" and "sphingolipid metabolic process"), organic substance transport, phosphoric diester hydrolase activity, sphingomyelin phosphodiesterase activity, and phospholipase activity. Instead, several of these terms are enriched in the ⌬sce1-unrestored group (Fig. 5C). This suggests that, although phosphorylation of proteins contributing to these processes may change in ⌬cdpk3 parasites, this is unlikely to directly relate to CDPK3-or SCE1-dependent egress signaling. This group of proteins was therefore not investigated further.
By contrast, proteins in the ⌬cdpk3-independent group showed no significant enrichment in GO terms relating to cell signaling (Fig. 5D). A slight enrichment in terms relating to protein and metabolite binding ("binding," "nucleic acid binding," "nucleotide binding," "heterocyclic compound binding," and "organic cyclic compound binding") was seen, however. We also note an underrepresentation of membrane-associated proteins ("membrane," "intrinsic component of membrane," "integral component of membrane," and "membrane part"); however, this may be a consequence of the general difficulty of solubilizing membrane proteins in proteomics.
We next interrogated our hits for phosphopeptides belonging to proteins that are known or predicted to be involved in apicomplexan motility and invasion signaling pathways. One phosphosite on each of the PDEs TGME49_202540 and TGME49_280410 appeared to be misregulated in ⌬cdpk3 and restored in ⌬cdpk3⌬sce1, indicating that these may be crucial CDPK3-and SCE1-regulated sites in egress signaling (supplemental Table S4). In addition, the abundance of two phosphosites on the cAMP-dependent protein kinase (PKA) regulatory subunit shows regulation by CDPK3 and SCE1, indicating that such cyclic nucleotide signaling pathways may play a role in egress. Furthermore, the abundance of a single phosphopeptide of phospholipase C (PLC) was dependent on CDPK3 and SCE1. PLC has been implicated in regulating Ca 2ϩ flux in Toxoplasma (6,27). One phosphosite of the signaling kinase casein kinase 1

SCE1 during Toxoplasma egress
showed regulation by CDPK3 and SCE1. Casein kinase has been implicated in the phosphorylation of adhesin tails in P. falciparum (28). Interestingly, the uncharacterized Ca 2ϩ -dependent kinase CDPK2A showed several phosphosites regulated by CDPK3 that were independent of SCE1. CDPK2A has been implicated previously in egress, making it a very strong candidate going forward (24). Several hypothetical proteins also showed a high abundance of phosphosites regulated by CDPK3 and SCE1, such as TGME49_279100. For further discussion of identified CDPK3-and SCE1-dependent phosphopeptides see supplemental information. Together, these data highlight several cellular processes wherein phosphorylation is regulated by CDPK3 and indicate that SCE1 plays an indirect antagonistic role in regulating several of these, particularly those pertaining to cell signaling. This supports the hypothesis that CDPK3 and SCE1 regulate a common signaling pathway during Ca 2ϩ -stimulated egress.

Phosphorylation of SCE1 regulates Toxoplasma egress
During our interrogation of SILAC data, we noted that two phosphosites of SCE1 (Ser-641 and Ser-645) were consistently less abundant in ⌬cdpk3 (supplemental Table S4). Data from previous phosphoproteomic investigations of Toxoplasma egress also show lower levels of SCE1 phosphorylation in CDPK3-deficient parasites (24). Furthermore, recent experiments using the substrate capture technique sug-gested that SCE1 is likely to be a direct substrate of CDPK3 3 (supplemental Table S5 and Fig. 6A).
To test the importance of phosphorylation in regulating SCE1 activity, we designed a set of complementation constructs expressing cDNA copies of SCE1 containing substitution mutations within detected phosphosites. Mutations were either phosphomimetic (Ser 3 Asp, SCE1 M ) or phospho-null (Ser 3 Ala, SCE1 N constructs) (Fig. 6A). Constructs included one of two sets of mutations. Set 1 (i.e. M 1 and N 1 ) contained all mutations at phosphosites that were detected in a least one substrate capture experiment (supplemental Table S5). Set 2 (i.e. M 2 and N 2 ) contained only mutations at sites detected in more than one substrate capture experiment. Both sets 1 and 2 included mutations in Ser-641 and Ser-645 that were less abundant in global analyses (24) (supplemental Table S4).
To investigate the role of phosphorylation of SCE1 during egress, phosphosite mutants (and SCE1 wt ) were introduced ectopically into ⌬sce1, ⌬cdpk3, and ⌬cdpk3⌬sce1 parasites (supplemental Table S1). Expression was confirmed by IFA, showing identical localization to SCE1 wt (described above) (Fig.  6B). Therefore, phosphomutants did not affect the membrane localization of SCE1. Western blotting showed that the expression levels of the ectopic protein were comparable across all lines (Fig. 6C).   Error bars show mean Ϯ S.D. *, p Ͻ 0.05; **, p Ͻ 0.001. All statistical tests were performed using two-way ANOVA and Tukey correction for multiple comparisons. The data were derived from three independent biological replicates.

SCE1 during Toxoplasma egress
To assess the role of SCE1 phosphorylation during CDPK3dependent egress, we assayed the response of phosphosite mutants to stimulation with 2 M A23187 for 3 min. Expression of the phosphomimetic mutants in ⌬sce1 (parasite strains ⌬sce1:SCE1 M1 and ⌬sce1:SCE1 M2 ) did not affect A23187-stimulated egress relative to ⌬sce1:SCE1 wt (Fig. 6Di). However, expression of phospho-null mutants (parasite strains ⌬sce1: SCE1 N1 and ⌬sce1:SCE1 N2 ) significantly inhibited ⌬sce1 parasite egress efficiency, with the percentage of parasitophorous vacuoles from which parasites egressed decreasing to ϳ35% and 25% for ⌬sce1:SCE1 N1 and ⌬sce1:SCE1 N2 , respectively. This suggests that phosphorylation of SCE1 is required for efficient egress even in the presence of functional CDPK3. Expression of wild-type or phosphomutant SCE1 in the ⌬cdpk3 background did not result in any changes in egress capacity (Fig.  6Dii). This suggests that mimicking phosphorylation of SCE1 alone cannot overcome loss of CDPK3 during egress.
Overall, our data strongly suggests that phosphorylation of SCE1 is very important in relieving its suppressive activity during Ca 2ϩ -stimulated egress. It is interesting to note that, within all genetic backgrounds (⌬sce1, ⌬cdpk3, and ⌬cdpk3⌬sce1), expression of either phosphomimetic SCE1 (SCE1 M1 and SCE1 M2 ) or either phospho-null mutant (SCE1 N1 and SCE1 N2 ) yielded essentially the same egress phenotype. This demonstrates that the key phosphorylation sites that regulate SCE1 activity are captured in both groups of mutants (Ser-8, Ser-73, Ser-225, Ser-641, and Ser-645).

SCE1 does not affect cytosolic Ca 2؉ dynamics during egress
Our results suggest that SCE1 may regulate generic egress signaling events. We therefore wondered whether SCE1 may act as a Ca 2ϩ transporter. Indeed, previous reports suggest that CDPK3 influences the resting cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] cyt ) (24) and, thus, that loss of SCE1 may restore this imbalance to promote egress. Although SCE1 appears to have no primary amino acid homology to Ca 2ϩ transporters, the online structural homology prediction software iTasser suggests that SCE1 has greatest structural homology to transient receptor potential cation subfamily v (TRPV)-like channels (29). TRPV channels are typically found on the plasma membrane that typically transport cations such as Ca 2ϩ , Mg 2ϩ , and Na ϩ in response to a range of extracellular stimuli (30). To investigate the possibility that SCE1 could be a Ca 2ϩ transporter or channel, we first measured [Ca 2ϩ ] cyt in extracellular parasites suspended either in a Ca 2ϩ -free buffer or in buffers containing a range of free Ca 2ϩ concentrations (Fig. 7A)  It should be noted that a previous study reported that extracellular ⌬cdpk3 knockout parasites suspended in a Ca 2ϩ -free buffer had an elevated basal Ca 2ϩ level compared with wildtype parasites, although these measurements were not calibrated (24). In the presence of [Ca 2ϩ ] o , the values we obtained for [Ca 2ϩ ] cyt were slightly higher in ⌬cdpk3 compared with the wild type. It is not clear why these differences are seen between our studies.
So far, we have only detected a phenotypic difference upon deletion of SE1 during egress, when tachyzoites are still intracellular. It therefore might be the case that SCE1 only functions when tachyzoites are intracellular. We therefore tested for differences in the levels of [Ca 2ϩ ] cyt at this stage, when tachyzoites are egress-competent. To do this, we introduced a plasmid expressing the genetically encoded Ca 2ϩ biosensor GCaMP6 into the uprt locus of wild-type, ⌬cdpk3, ⌬sce1, and ⌬cdpk3⌬sce1 parasites (25). We observed negligible differences in resting [Ca 2ϩ ] cyt between strains (Fig. 7B). Egress was then stimulated using A23187 or BIPPO or by semipermeabilization of host cells with saponin, and the subsequent rate of Ca 2ϩ increase (Fig. 7C) and peak Ca 2ϩ concentration (Fig. 7D) were monitored. We saw no significant difference in these parameters between any of the lines for any of these conditions. Overall, these data suggest that SCE1 does not have a role in maintaining [Ca 2ϩ ] cyt while extracellular or intracellular and plays no role in Ca 2ϩ levels during stimulated egress.

Discussion
Plant-like Ca 2ϩ -dependent protein kinases are important signal transducers that regulate motility, invasion, and host cell egress in apicomplexan parasites. We have shown previously that CDPK3 in Toxoplasma is critical for host cell egress (11), but how this kinase (or indeed any other CDPK) regulates egress remains unexplained. To uncover novel signaling proteins that participate in CDPK3-regulated egress, we designed a forward genetic screen to identify mutants that circumvent their dependence on CDPK3 during A23187-stimulated egress. Here we identified a novel regulator of Toxoplasma egress, the transmembrane protein SCE1. We demonstrated that SCE1 is a suppressor of Ca 2ϩ -dependent Toxoplasma host cell egress whose suppressive activity is relieved by phosphorylation. Furthermore, we demonstrated that SCE1 acts, through unknown mechanisms, to suppress the phosphorylation of an array of proteins predicted to be involved in CDPK3-dependent signal transduction pathways during egress.
Combining these findings with functional analysis of phosphorylation sites has allowed us to put together a model of CDPK3-dependent egress (Fig. 8). Previous work suggested that CDPK3 acts in an amplification loop to promote egress (30) (Fig. 8A, thicker arrows), and we propose here that CDPK3 also directly phosphorylates SCE1, relieving the suppressive activity of this protein, allowing for maximal stimulated host cell egress (Fig. 8A). The loss of CDPK3 results in the loss of phosphorylation of SCE1 and inhibition of egress. Under this condition, CDPK3/SCE1-independent egress still takes place, which is offset by the suppressive activity of SCE1 (Fig. 8B).
Upon additional deletion of SCE1, the alternative egress pathway becomes more prominent, promoting greater but not full egress capacity (Fig. 8C). Prevention of SCE1 phosphorylation (by making phospho-null mutations) prevents relief of egress suppression even in the presence of CDPK3, thus dampening egress compared with the wild type (Fig. 8D). The suppressive activity of SCE1 can be somewhat overcome when phosphorylation events are mimicked with aspartic acids, resulting in increased egress compared with wild-type SCE1 (Fig. 8E).
Interestingly, loss of SCE1 alone did not appear to result in premature egress of parasites, demonstrating that exogenous stimuli are still required to trigger egress in its absence. Our data demonstrate that egress signaling in Toxoplasma requires both activation of positive regulators and relief of suppression by negative regulators. Therefore, SCE1 knockout may, in fact, increase the speed and sensitivity of wild-type egress to exogenous stimuli. However, in our experiments, wild-type and ⌬sce1 egress are indistinguishable because wild-type egress is already so efficient in vitro, occurring just 1-2 min following A23187 or BIPPO treatment. In the future, it would be interesting to study the importance of SCE1 for the ability of Toxoplasma to infect cell types that are typically infected in vivo, such as macrophages. Such cells are more prone to immune pressure; thus, parasites within them may have to rapidly initiate egress earlier to escape targeting by the immune system. SCE1 appears to localize to the tachyzoite plasma membrane, thus potentially co-localizing with CDPK3 and cGMP signaling proteins like PKG (5,11). Our data presented here support a

SCE1 during Toxoplasma egress
role for the plasma membrane as a key signaling hub for regulation of Ca 2ϩ signaling and egress in Toxoplasma, with CDPK3 and SCE1 as key participants. We hypothesized that SCE1 may be a Ca 2ϩ transporter or channel not easily identifiable by primary sequence homology; however, our data suggest that SCE1 does not significantly affect cytosolic Ca 2ϩ levels before or during egress. Alternatively, SCE1 could suppress the activity of other signaling proteins or ion channels by blocking or sequestering them at the plasma membrane. In this model, phosphorylation of SCE1 could relieve its suppressive function, releasing sequestered proteins and thus promoting downstream signaling events. We attempted to identify interaction partners of SCE1 by using quantitative proteomics to identify proteins that differentially associate with SCE1 phospho-null versus phosphomimetic mutants, but these studies did not yield conclusive results (data not shown).
SCE1 is conserved in the closely related parasites N. caninum and H. hammondi but has no obvious homologs in other apicomplexan species. It is interesting to note that this group of closely related parasites are the only members of Apicomplexa to have the capacity to egress the host cell at any stage during intracellular growth. A checkpoint on egress, such as SCE1, may thus be essential to inhibit premature activation in this clade of apicomplexan parasites. By contrast, other apicomplexans utilize a range of asexual replication strategies (e.g. schizogony, endopolyogeny, and merogeny) wherein daughter cells are only formed after multiple rounds of nuclear division are completed. This places a developmental constraint on egress, meaning that signaling checkpoints may be largely unrequired.
Quantitative phosphoproteomics showed that a large portion of phosphosites misregulated in ⌬cdpk3 parasites were rescued to a wild type-like state in ⌬cdpk3⌬sce1, supporting the notion that CDPK3 and SCE1 antagonistically regulate egress signaling events. These data also identified potentially important regulatory phosphosites on several proteins implicated in Ca 2ϩ -dependent signaling and egress. One phosphosite of the important Ca 2ϩ flux regulator PLC (Ser-88) was observed to be down-regulated in ⌬cdpk3 parasites and rescued in ⌬cdpk3⌬sce1 parasites, and it will be interesting to test the role of this phosphorylation site in Ca 2ϩ signaling now that PLC mutants are available (6). Phosphosites on two PDEs and the regulatory subunit of PKA were observed to be down-regulated in ⌬cdpk3 parasites and rescued in ⌬cdpk3⌬sce1 parasites, supporting a larger role for cyclic nucleotide kinases in egress. It is unknown whether these PDEs have cGMP hydrolysis activity or whether they are targeted by the known cGMP PDE inhibitors zaprinast and BIPPO. These data suggest that knocking out SCE1 may enhance egress by altering the levels of cyclic nucleotides in Toxoplasma, thereby increasing the activity of cyclic nucleotide-dependent kinases involved in egress, like PKG.
This study has shown that forward genetic screens using non-lethal knockout mutants with tightly defined phenotypes (such as ⌬cdpk3) can be a powerful means of unraveling signaling pathways in Toxoplasma and potentially other apicomplexan parasites. Moving forward, techniques such as this will be particularly useful to identify proteins such as SCE1 that display no phenotype when knocked out individually and are thus invisible by standard reverse genetics. As our appreciation for the complexity of pathogen cell signaling systems grows, these tools will become ever more useful, particularly when facilitated by whole-cell techniques such as quantitative phosphoproteomics and high-throughput genetic manipulation systems such as CRISPR/Cas9 (33,34).

Parasite culture and transfection
Toxoplasma parasites were grown in human foreskin fibroblasts (HFFs) in D1 medium (DMEM supplemented with 1% FCS (Invitrogen) and 1% v/v Glutamax (Invitrogen)) at 37°C with 10% CO 2 . Before parasite infection, HFFs were grown to confluency and maintained in DMEM supplemented with 10% Cosmic calf serum (Thermo Scientific). Transfection of parasites with DNA was performed by electroporation using a Bio-Rad Gene Pulser II module with conditions as described previously (35). Positive selection of recombinant parasites was performed using mycophenolic and xanthine, chloramphenicol, or phleomycin as described previously (35)(36)(37) or with 5Ј-fluo-2Ј-deoxyuridine-negative selection for integration at the uracil phosphoribosyl transferase locus (38).

DNA cloning
All sequence models were based on the Toxoplasma ME49 genome from ToxoDB.org (39). Oligonucleotide primers and synthesized gene fragments (gBlocks) were ordered from Integrated DNA Technologies. Assembly of DNA constructs was performed using the Gibson assembly cloning kit (New England Biolabs), according to the instructions of the manufacturer. Details regarding the construction of specific plasmids are presented in the supplemental information.

Forward genetic screen
Ten populations of ⌬cdpk3 parasites were mutagenized with EMS as described previously (40). After HFF lysis, ϳ80,000 parasites from each population were added to fresh HFFs and allowed to invade for 2 h. Excess parasites were washed away three times with PBS. The medium was replaced with D1 supplemented with 25 mg/ml dextran sulfate (Sigma-Aldrich), and cultures were grown for 28 h. Flasks were washed twice with PBS, egress was stimulated for 2 min (4 M A23187, 25 mg/ml dextran sulfate in DMEM-HEPES, 37°C), and then cells were quenched with ice-cold PBS. Egressed parasites were washed off twice with ice-cold PBS; supernatant were removed and centrifuged to pellet parasites. Pellets were resuspended and added to fresh HFFs. These steps were repeated until an egress assay showed rescue of the ⌬cdpk3 egress-deficient phenotype. Clonal parasite lines were derived from rescue-of-egress populations using limited dilution. Details regarding sequencing of ⌬cdpk3 rev mutants and egress assays are included in the supplemental information.

Cellular assays
Immunofluorescence assays (IFAs) were performed on 4% of formaldehyde-fixed samples using standard procedures. Images were taken with a DeltaVision Elite microscope (GE Healthcare) using a CoolSnap2 charged coupled device (CCD) SCE1 during Toxoplasma egress detector and SoftWoRx software (GE Healthcare). Images were then processed and assembled using ImageJ, Adobe Photoshop, and Adobe Illustrator software. Extended methods are included in the supplemental information.

Phosphoproteomics of egress sample preparation
RH:⌬ku80, ⌬cdpk3, and ⌬sce1:⌬cdpk3 parasites were grown on confluent HFFs in SILAC labeling medium for 1 week (DMEM without Arg or Lys (Gibco) with 1% dialyzed FCS, 1% Glutamax, 1% penicillin/streptomycin (Gibco) and supplemented with either (light) L-[U- 12 Labs)). After complete HFF lysis, 7.5 ϫ 10 7 labeled parasites of each strain were added to HFFs in 2 ϫ 175 cm 2 culture flasks each and allowed to invade in SILAC labeling medium for 2 h. Excess parasites were washed away with 3ϫ PBS, the SILAC medium was replaced, and parasites were grown for a further 26 h. Egress stimulation medium was then added (2 M A23187 in 5 ml of DMEM without Arg and Lys). After 1.5 min, egress was quenched by addition of 20 ml of ice-cold PBS, and the flasks were placed in an ice water bath for 1 min. Host cells and parasites were released from the surface of the culture flasks by cell scraping. Excess parasites and host cells were washed from the flasks using 5 ml of ice-cold PBS, and scraped cells and PBS wash were added to a 30-ml syringe and passed through a 27-gauge needle. Host cell debris was pelleted by centrifugation (600 ϫ g, 2 min, 0°C), the supernatant was transferred to a fresh centrifuge tube, and parasites were pelleted (1200 ϫ g, 10 min, 0°C). Parasite pellets were lysed immediately. More detailed methods for phosphoproteomic analysis are included in the supplemental information.

Ca 2؉ measurements with extracellular tachyzoites using Fura-2
Measurements of [Ca 2ϩ ] cyt were performed with extracellular tachyzoites that had recently emerged from their host cells or that were harvested from their host cells by passage through a 26-gauge needle. The parasites were passed through a 3-m filter to remove host cell material and then centrifuged (1500 ϫ g for 10 -20 min at room temperature). The supernatant medium was discarded, and the parasites were resuspended and then washed once (12,000 ϫ g for 1 min) in a saline solution consisting of 125 mM NaCl, 5 mM KCl, 25 mM HEPES, 20 mM glucose, and 1 mM MgCl 2 (pH 7.10). The parasites were resuspended in 2 ml of the saline solution and loaded with Fura-2 dye by incubation for 30 min at 37°C in the presence of 6 M Fura-2/AM and 0.04% pluronic acid F-127. The parasites were then washed twice (12,000 ϫ g for 1 min) in saline solution to remove extracellular dye.
Fluorescence measurements were performed on Fura-2loaded parasites (5-10 ϫ 10 7 ) suspended either in Ca 2ϩ -free saline solution (identical to the saline solution described above but containing 1 mM EGTA) or in Ca 2ϩ -free saline solution to which sufficient CaCl 2 was added to achieve a final free Ca 2ϩ concentration of 1 M, 100 M, or 1 mM. The measurements were performed at 37°C using a PerkinElmer Life Sciences LS 50B fluorescence spectrometer with a dual excitation Fast Filter accessory using excitation wavelengths of 340 nm and 380 nm and an emission wavelength of 515 nm. The measurements of [Ca 2ϩ ] cyt were calibrated in situ (41) using a K d of 120 nM for the Fura-2⅐Ca 2ϩ complex.
For each external Ca 2ϩ concentration, the [Ca 2ϩ ] cyt value was determined from fluorescence measurements made within 40 min of the start of fluorescence recording (and within ϳ60 min of having loaded the dye into the parasites). The order in which different [Ca 2ϩ ] o values were tested varied between experiments.

Ca 2؉ measurements in intracellular tachyzoites and during egress
Intracellular Ca 2ϩ measurements were undertaken by introducing GCaMP6 into the different parasite lines as described previously (25). This plasmid co-expresses mCherry to allow for internal normalization of the signal. Briefly, plasmids were integrated into the uprt locus by selecting parasites on fluorodeoxyribose. Relative Ca 2ϩ levels were then measured on resting intracellular tachyzoites as well as on those stimulated to egress using either A23187, BIPPO, or saponin. Fluorescence intensities were then calculated using ImageJ, and ratios were calculated in an Excel spreadsheets and plotted using Prism software.