Toxoplasma gondii microneme secretion involves intracellular Ca(2+) release from inositol 1,4,5-triphosphate (IP(3))/ryanodine-sensitive stores.

Calcium-mediated microneme secretion in Toxoplasma gondii is stimulated by contact with host cells, resulting in the discharge of adhesins that mediate attachment. The intracellular source of calcium and the signaling pathway(s) triggering release have not been characterized, prompting our search for mediators of calcium signaling and microneme secretion in T. gondii. We identified two stimuli of microneme secretion, ryanodine and caffeine, which enhanced release of calcium from parasite intracellular stores. Ethanol, a previously characterized trigger of microneme secretion, stimulated an increase in parasite inositol 1,4,5-triphosphate, implying that this second messenger may mediate intracellular calcium release. Consistent with this observation, xestospongin C, an inositol 1,4,5-triphosphate receptor antagonist, inhibited microneme secretion and blocked parasite attachment and invasion of host cells. Collectively, these results suggest that T. gondii possess an intracellular calcium release channel with properties of the inositol 1,4,5-triphosphate/ryanodine receptor superfamily. Intracellular calcium channels, previously studied almost exclusively in multicellular animals, appear to also be critical to the control of parasite calcium during the initial steps of host cell entry.

charged during Toxoplasma gondii invasion. Microneme proteins possess a variety of adhesive domains and include soluble and transmembrane forms (2,3). Their release appears highly regulated in that host cell contact triggers a burst of microneme release (1). Binding studies suggest that microneme proteins participate in attachment to host cell surfaces (4 -7) and that the transmembrane form of the adhesin MIC2 may link parasite and host cell membranes during invasion.
Previous studies demonstrated that increases in intracellular calcium ([Ca 2ϩ ] i ) mediate microneme secretion in T. gondii. For example, association of T. gondii tachyzoites with host cells results in increases in parasite [Ca 2ϩ ] i (8) and chelation of T. gondii [Ca 2ϩ ] i blocks microneme secretion and invasion (9). Moreover, reagents that raise T. gondii [Ca 2ϩ ] i levels stimulate microneme discharge in the absence of host cells (10). These include Ca 2ϩ ionophores ionomycin and A23187, as well as the sarco-endoplasmic reticulum ATPase inhibitor thapsigargin (10). Identification of alcohols as potent artificial triggers of microneme secretion led to studies demonstrating that ethanol also raises parasite [Ca 2ϩ ] i levels (11). T. gondii possess several intracellular stores of Ca 2ϩ , including the acidocalcisomes (12), mitochondria, and the endoplasmic reticulum (13). Whether all or some of these stores participate in microneme secretion during host cell invasion is unknown. The mechanism of Ca 2ϩmediated microneme release in both artificially triggered and natural microneme secretion during invasion remains largely undefined.
Regulated exocytosis is associated with fluxes in [Ca 2ϩ ] i , although the exact step requiring Ca 2ϩ is unclear and may depend on the system being studied (14). Ca 2ϩ signaling involves the mobilization of Ca 2ϩ from two sources: intracellular stores and the extracellular medium. In mammalian cells, the sarco-endoplasmic reticulum is a commonly utilized source of readily accessible [Ca 2ϩ ] i , which can be rapidly mobilized under a variety of stimuli. Two types of sarco-endoplasmic reticulum Ca 2ϩ release channels have been identified in multicellular animals: the IP 3 1 receptor and the ryanodine receptor, which are structurally similar and are thought to be evolutionarily related (15). Binding of the second messenger IP 3 to its receptor releases Ca 2ϩ into the cytosol, whereas ryanodine receptors are modulated in some cells by cyclic ADP-ribose (16). This release of intracellular Ca 2ϩ can in some manner signal the activation of Ca 2ϩ entry, a process known as capacitative Ca 2ϩ entry (17).
Neither IP 3 , ryanodine receptors, or capacitative Ca 2ϩ entry have been identified in unicellular organisms, including yeast and protozoa. However, evidence linking IP 3 to [Ca 2ϩ ] i flux has been described in several unicellular systems. Treatment of membrane vesicles from Candida albicans (18) or Plasmodium chabaudi (19) with IP 3 results in Ca 2ϩ release and is blocked by the IP 3 receptor antagonist heparin. IP 3 induces Ca 2ϩ release from Euglena gracilis microsomes in a dose-dependent manner (20). Carbachol stimulation of Trypanosoma cruzi results in increased [Ca 2ϩ ] i and IP 3 (21). In characean algae, introduction of IP 3 produces action potentials, an event known to involve increases in cytoplasmic Ca 2ϩ (22). Collectively, these studies indicate a potential role for IP 3 in Ca 2ϩ -mediated signaling events among early branching unicellular eukaryotes. We were interested in determining whether IP 3 functions as a mediator of microneme secretion by increasing parasite [Ca 2ϩ ] i . In this work, we present evidence that T. gondii may possess intracellular release channels of the IP 3 /ryanodine receptor superfamily.

EXPERIMENTAL PROCEDURES
Chemicals-Caffeine, EGTA, and ryanodine were purchased from Sigma. BAPTA-AM and xestospongin C were purchased from Calbiochem. The Biotrak TM [ 3 H]IP 3 assay system was purchased from Amersham Biosciences. All other reagents were analytical grade. Drugs were resuspended as 100ϫ stock solutions in either distilled water or Me 2 SO.
Parasite Culture-T. gondii strain RH and the lacZ-expressing clone 2F were maintained as tachyzoites in human foreskin fibroblast cells as described (23). For Ca 2ϩ measurement experiments, RH tachyzoites were maintained in bovine turbinate cells (ATCC CRL 1390) (12). All experiments used freshly lysed out parasites purified by passage through a 23-gauge needle and filtration through a 3 micron Nucleopore membrane. All cultures tested free of Mycoplasma using the Gen-Probe TM mycoplasma detection system.

SDS-Polyacrylamide Gel Electrophoresis and Western
Blotting-SDS-PAGE was performed in 7% mini-gels under reducing conditions and transferred to nitrocellulose as previously described (10). Western blotting was performed with mouse anti-TgMIC2 monoclonal antibody 6D10 (ascites, 1:10,000) (10), mouse anti-TgMIC4 monoclonal antibody 5B1 (supernatant, 1:10) (24), and rabbit polyclonal anti-TgACT1 actin antibody (1:10,000) (25). Mouse anti ␤-galactosidase monoclonal antibody 40 -1a (supernatant, 1:50) was a gift from Dr. Joshua Sanes (University of Washington, Seattle, WA). IP 3 Quantification-Tachyzoites were resuspended at 3 ϫ 10 9 /ml in 10 mM HEPES, 1 mM EGTA in Hanks' balanced salt with no calcium chloride, magnesium chloride, magnesium sulfate, or sodium bicarbonate and split into two tubes. After equilibration at 37°C for 5 min, secretagogues or a solvent control were added (time 0). At 30 -60 s intervals 0.1 ml of parasites were removed and added to an equal volume of ice-cold 15% trichloroacetic acid. Tubes were vortexed and spun at 4°C for 15 min at 2000 ϫ g. Supernatants were extracted three times with 10 volumes of water-saturated ether and adjusted to pH 7.5 with 1 M sodium bicarbonate. Pellets were resuspended in SDS-PAGE sample buffer for confirmation of secretion stimulation by Western blotting of TgMIC2. IP 3 in ether-extracted supernatants was measured in duplicate based on the potential of the parasite samples containing IP 3 to displace [ 3 H]IP 3 from a bovine IP 3 receptor as described in the Amersham Biotrak TM protocol. Basal levels of IP 3 ranged from 0.4 to 0.6 pmol/10 8 parasites.
Microneme Secretion Assay-Tachyzoites were resuspended at 10 8 /ml in Dulbecco's modified Eagle's medium, 1% fetal bovine serum, and 20 mM HEPES, pH 7.5 at 18°C in 1.5 ml microcentrifuge tubes. Pretreatment with stock solutions of drugs proceeded for 15 min before transferring parasites to 37°C for 2 min to stimulate secretion followed by transfer to a wet ice bath. In parallel, 100 M BAPTA-AM was loaded for 10 min at 18°C, while 10 mM EGTA was added immediately prior to secretagogue stimulation to minimize leaching of intracellular Ca 2ϩ stores. Parasite supernatants were separated from pellets by centrifugation at 4°C and run on 7% SDS-PAGE gels next to pellet dilutions to indicate the secreted and cellular forms of TgMIC2. Inadvertent lysis of parasites was monitored by the release of constitutively expressed ␤-galactosidase in clone 2F or actin in strain RH.
Measurement of Parasite [Ca 2ϩ ] i -After being harvested, tachyzoites were washed twice at 500 ϫ g for 10 min at room temperature in buffer A (116 mM NaCl/5.4 mM KCl/0.8 mM MgSO 4 /5.5 mM D-glucose/50 mM HEPES, pH 7.4). Parasites were resuspended to a final density of 10 9 cells/ml in loading buffer, which consisted of buffer A plus 1.5% (w/v) sucrose and 6 M fura-2/AM. The suspensions were incubated for 30 min at 25°C. Subsequently, the cells were washed twice with buffer A to remove extracellular dye. Parasites were resuspended to a final density of 10 9 cells/ml in buffer A and kept in ice. Parasites were viable for several hours under these conditions. For fluorescence measurements, a 50-l aliquot of the cell suspension was diluted into 2.5 ml of buffer A (2 ϫ 10 7 cells/ml final density) in a cuvette placed into a Hitachi F-2000 spectrofluorometer at room temperature. For measurements with fura-2, excitation was at 340 and 380 nm and emission was at 510 nm. The fura-2 fluorescence response to [Ca 2ϩ ] i was calibrated from the ratio of fluorescence values at 340 and 380 nm after subtraction of the background fluorescence of the cells at 340 and 380 nm. Attachment and Invasion Assays-Tachyzoite invasion of human foreskin fibroblasts grown to confluence in 24-well plates was quantified by colorimetric detection of ␤-galactosidase expressed by strain 2F as previously described (25). To assess the effects of drugs on attachment, plates were fixed with 2.5% (w/v) formaldehyde for 20 min at 4°C, washed three times with complete media (Dulbecco's modified Eagle's medium, 10% FBS, 10 mM HEPES, 2 mM glutamine, 20 g/ml gentamicin), and then equilibrated at 37°C with complete media. This fixation protocol prevents the majority of parasites from invading while still allowing adhesion to occur. Parasites were resuspended at 10 7 /ml in invasion medium and treated for 10 min at room temperature with stock solutions of drugs. Treated parasites were added in 0.2 ml volume to warm 24-well plates of human foreskin fibroblasts in quadruplicate and incubated for 20 min at 37°C for both assays.

Ryanodine Stimulates Ca 2ϩ -mediated Microneme Secretion-Ryanodine has been shown to mediate increases in [Ca 2ϩ
] i in a variety of multicellular organisms (26). To determine whether ryanodine stimulated Toxoplasma microneme secretion, we treated tachyzoites with ryanodine at concentrations ranging from 1 nM to 100 M, removed the cells by centrifugation, and examined secreted proteins MIC2 and MIC4 present in supernatants by Western blotting. MIC2 is a convenient marker for microneme secretion because the secreted 95-100-kDa form (sMIC2) is released into supernatants, while the cell-associated 115-kDa form (cMIC2) remains in the parasite pellet (27). Similarly, the secreted form of MIC4 is 70-kDa, while cell-associated MIC4 is 72 kDa. Constitutive expression of ␤-galactosidase in the 2F strain was used to control for inadvertent lysis of the parasites during the experiment and was typically between 1-5%, as indicated by comparison with dilutions of a parasite cell standard (% cell stds).
Secretion of MIC2 and MIC4 was stimulated by ryanodine at doses from 1 nM to 10 M but inhibited at 100 M when compared with parasites treated with Me 2 SO (Fig. 1A). The inhibition of microneme secretion at 100 M is consistent with previous studies indicating that high doses keep the ryanodine receptor in a closed conformation (26). To determine whether ryanodine stimulation of microneme secretion involved an increase of [Ca 2ϩ ] i from either intracellular stores or the extracellular medium, we pretreated parasites with the membranepermeable calcium chelator BAPTA-AM to prevent [Ca 2ϩ ] i increases (8) or added EGTA to the parasites immediately prior to addition of ryanodine to chelate extracellular Ca 2ϩ and prevent Ca 2ϩ entry (12). BAPTA-AM pretreatment inhibited microneme secretion upon stimulation with 100 nM ryanodine (Fig. 1A), indicating that ryanodine-mediated secretion required a rise in parasite [Ca 2ϩ ] i . Additionally, EGTA also blocked ryanodine stimulation of microneme secretion, showing that extracellular Ca 2ϩ was required by the parasite. Monitoring of [Ca 2ϩ ] i using fura-2-loaded parasites confirmed these results. The increase in [Ca 2ϩ ] i produced by ryanodine in the absence of extracellular Ca 2ϩ (with 1 mM EGTA added to the incubation medium; Fig. 1B, ϩEGTA) confirmed the hypothesis that ryanodine is able to release Ca 2ϩ from intracellular stores of T. gondii. Under these conditions [Ca 2ϩ ] i rapidly decreased, probably through Ca 2ϩ efflux to the external medium or Ca 2ϩ uptake into other intracellular stores. In the presence of extracellular Ca 2ϩ a greater and sustained response was observed upon addition of ryanodine (Fig. 1B). Both the magnitude and duration of this elevated [Ca 2ϩ ] i response were greater in the presence of extracellular Ca 2ϩ (Fig. 1B). This sustained [Ca 2ϩ ] i elevation in the presence of extracellu-lar Ca 2ϩ could be due either to capacitative Ca 2ϩ entry (17) or to inhibition of Ca 2ϩ efflux by ryanodine as has been postulated to occur in smooth muscle cells (28). Because treatment of tachyzoites with EGTA prevented microneme secretion ( Fig.  1A) a sustained [Ca 2ϩ ] i increase is apparently responsible for this process.
Caffeine Stimulates Ca 2ϩ -mediated Microneme Secretion-Caffeine is a pharmacological agonist of ryanodine receptor Ca 2ϩ release (26). Based on the above evidence that T. gondii has a ryanodine-sensitive Ca 2ϩ response, we tested caffeine for its ability to stimulate microneme secretion. Tachyzoites were incubated with caffeine at concentrations ranging from 1 M to 1 mM before collection of supernatants for analysis by Western blot. Caffeine stimulated microneme secretion at all doses, while only causing minimal parasite lysis ( Fig. 2A). Additionally, BAPTA-AM pretreatment to prevent parasite [Ca 2ϩ ] i increase (8) abrogated caffeine-induced microneme secretion, indicating that a rise in [Ca 2ϩ ] i was required for caffeine's effect. Unlike ryanodine, caffeine's ability to stimulate microneme secretion did not require extracellular Ca 2ϩ , as chelation with EGTA prior to addition of caffeine did not affect secretion of TgMIC2 and TgMIC4 ( Fig. 2A).
Caffeine treatment also increased parasite [Ca 2ϩ ] i , as shown by direct monitoring of Ca 2ϩ using fura-2-loaded tachyzoites (Fig. 2B). Unlike ryanodine, caffeine resulted in a more sus- The lack of inhibition of microneme secretion by addition of EGTA could be related to either this more sustained response or to a closer spatial association between the caffeine-sensitive channels and the micronemes. Collectively, these results indicate that caffeine increased [Ca 2ϩ ] i , which in turn stimulated microneme secretion in T. gondii.
Toxoplasma appears very sensitive to the effects of caffeine, as stimulation of secretion was observed down to micromolar levels while this compound is typically used at millimolar doses in other systems (29). Caffeine has at least two major effects on mammalian cells: stimulation of ryanodine receptor-mediated Ca 2ϩ release and inhibition of cyclic nucleotide phosphodiesterases (reviewed in Ref. 29). To address the possibility that caffeine stimulated microneme secretion through inhibition of cyclic nucleotide phosphodiesterases, we tested several additional compounds known to act on this pathway. Treatment with isobutylmethylxanthine, a nonspecific phosphodiesterase inhibitor, stimulated microneme secretion at 100 M. 2 However, treatment with forskolin, which elevates cyclic AMP in T. gondii (30), 8-bromo-cyclic-AMP (also expected to increase cyclic AMP), and 8-bromo-cyclic GMP (expected to increase cyclic GMP) had no effect on microneme secretion. 2 Ethanol Increases Parasite IP 3 -Ethanol is a potent trigger of microneme secretion in the absence of host cells (11). Because the effects of ethanol correlate with increases in parasite [Ca 2ϩ ] i , we were interested in determining whether the second messenger IP 3 was generated. Parasite IP 3 was measured using a commercially available radioreceptor assay. A rapid increase in IP 3 was observed after addition of 1% (171 mM) ethanol to extracellular tachyzoites at 37°C (Fig. 3A). Peak IP 3 levels were reached within the first minute after secretagogue addition in three independent experiments, with a mean maximal fold increase of 2.27 Ϯ 1.31 compared with 0.84 Ϯ 0.48 for the buffer-treated control. In parallel, secretion of microneme proteins was assayed by Western blot (Fig. 3B). IP 3 release often preceded secretion of micronemal proteins, which is consistent with its signaling activity mediating stimulation of Ca 2ϩ release from intracellular stores.
IP 3 Receptor Antagonist Inhibits Attachment, Invasion, and Microneme Secretion-The possibility that Toxoplasma contains intracellular Ca 2ϩ release channels of the IP 3 /ryanodine receptor family prompted us to examine additional drugs that could block invasion and microneme secretion through modulation of Ca 2ϩ channels. Xestospongin C is a membrane-permeable IP 3 receptor antagonist that does not affect IP 3 binding in rabbit cerebellar microsomes (31). Tachyzoites were pretreated with xestospongin C at doses from 10 to 0.1 M before stimulation with ethanol or caffeine. Secretion induced by ethanol and caffeine was reduced upon treatment with Ͼ1 M xestospongin C, as were the basal levels of secretion (Fig. 4B). In agreement with these results, Ca 2ϩ release by addition of either caffeine (Fig. 4C) or ethanol (Fig. 4D) was reduced upon treatment with xestospongin C (Fig. 4, C and D). We also tested xestospongin C in attachment and invasion assays on host cells. In parallel, tachyzoites were pretreated with different drug concentrations before addition to host cells for 20 min. Attachment assays were performed on formaldehyde-fixed host cells for the same length of time. Xestospongin C inhibited both attachment and invasion in a dose-dependent manner with an approximate IC 50 of 2 M for attachment and 8 M for invasion ( Fig 4A). Treatment of fura-2/AM-loaded parasites with 1 M xestospongin C reduced the increase in [Ca 2ϩ ] i upon stimulation with caffeine ( Fig 4C) and ethanol (Fig 4D), paralleling the inhibition of microneme secretion observed with this dose in Fig. 4B. Collectively, these studies show that xestospongin C is an antagonist of Ca 2ϩ -mediated microneme secretion, attachment, and invasion in T. gondii.

DISCUSSION
Calcium-mediated exocytosis of T. gondii microneme proteins is a key step in the invasion of host cells. The control of intracellular Ca 2ϩ levels in T. gondii is not well characterized, prompting our search for modulators of secretion whose target was well studied in other Ca 2ϩ signaling systems. We identified ryanodine and caffeine, two ryanodine receptor agonists, as stimuli of Ca 2ϩ -mediated microneme secretion. We also found that the second messenger IP 3 was induced upon ethanol stimulation of parasites. Finally, the IP 3 receptor antagonist xestospongin C inhibited the caffeine-and ethanol-induced increases in [Ca 2ϩ ] i , as well as preventing microneme secretion, attachment, and invasion. Together, these results strongly suggest that channels of the IP 3 /ryanodine receptor superfamily in T. gondii mediate increases in [Ca 2ϩ ] i , which in turn promote microneme secretion.
Ryanodine is an active component of the botanical insecticide ryania, an alkaloid found in the shrub Ryania speciosa. Ryanodine has potent effects on a variety of multicellular organisms (25). Receptors that release calcium from intracellular stores in response to ryanodine treatment have been described in vertebrates, where they usually are expressed in specialized neurosecretory or muscle cells (15). Ryanodine receptors also exist in a variety of invertebrates including crustaceans, insects, and worms; however, they have not been described in unicellular organisms (32). At micromolar levels ryanodine stabilizes ryanodine receptors in the open form, allowing Ca 2ϩ liberation from the sarco-endoplasmic lumen (26). At higher concentra- tions, ryanodine receptor channels in rabbit skeletal sarcoplasmic reticulum vesicles are persistently blocked (33). We observed a similar biphasic dose response in T. gondii, and ryanodine stimulation of microneme secretion was dependent on both intracellular and extracellular calcium. Ryanodine stimulation of [Ca 2ϩ ] i release has been shown to require extracellular calcium as part of its Ca 2ϩ -induced Ca 2ϩ release mechanism, suggesting it may activate flux through plasma membrane channels and capacitative entry (17,26). The induction of calcium release in T. gondii treated with ryanodine parallels that in well studied systems, suggesting the existence of ryanodine-responsive calcium channels in this early branching eukaryote.
Caffeine also induced microneme secretion by T. gondii, and was effective at doses that are 100-fold lower than its activity on mammalian cells. Unlike ryanodine, caffeine did not require extracellular Ca 2ϩ , a result similar to that reported for rat PC12 cells (35). In addition to acting on ryanodine-like calcium channels, caffeine and related alkylxanthines have been shown to inhibit cyclic nucleotide phosphodiesterases (25,26,33). The resulting increases in cyclic AMP and activation of protein kinase A can give rise to an increase in [Ca ϩ2 ] i in mammalian cells (36). Although isobutylmethylxanthine, a nonselective phosphodiesterase inhibitor, also stimulated microneme secretion by T. gondii, two drugs expected to directly increase cyclic AMP, forskolin and 8-bromo-cyclic AMP, had no affect on microneme secretion. Similarly, treatment of parasites with 8-bromo-cyclic GMP did not affect microneme secretion. Thus, while we cannot rule out possible inhibition of phosphodiesterases, the effect of caffeine in elevating [Ca 2ϩ ] i in T. gondii does not seem to be due to this mechanism. A third possible mode of action is that caffeine may inhibit the sarco-endoplasmic reticulum type ATPase involved in refilling calcium stores. This would be consistent with the results showing a more sustained [Ca 2ϩ ] i elevation with caffeine than with ryanodine in the absence of extracellular Ca 2ϩ (Fig. 2B). While this pump is not yet characterized in T. gondii, studies in the ciliate Paramecium tetraurelia have shown that the sarco-endoplasmic reticulum ATPase is exquisitely sensitive to caffeine inhibition (37). Ciliates are phylogenetically close relatives of the Apicomplexa. Determination of whether this mechanism for caffeine also operates in T. gondii and related parasites will require further study. Unlike paramecium, the [Ca 2ϩ ] i pool in T. gondii is sensitive to thapsigargin (12); however, ethanol stimulation of thapsigargin-pretreated parasites generates an additional increase in [Ca 2ϩ ] i (11), implying that they also possess a thapsigargininsensitive [Ca 2ϩ ] i pool. Similarly, pretreatment with thapsigargin prior to stimulation with caffeine or ryanodine did not inhibit microneme secretion (data not shown), suggesting that this same thapsigargin-insensitive pool may be stimulated by caffeine or ryanodine.
Our studies demonstrate that IP 3 acts as a second messenger during the stimulation of tachyzoites with ethanol, a known Ca 2ϩ -mediated trigger of microneme secretion (11). In higher eukaryotes, IP 3 induces calcium release from channels located in the sarco-endoplasmic reticulum in a variety of cell types (38). Typically, IP 3 is produced along with diacylglycerol after hydrolysis of phosphatidylinositol 4,5-bisphosphate by inositol phospholipid-specific phospholipase C (PLC). In higher eukaryotes, this process is mediated by PLC-␥ signaling through tyrosine kinase receptors and PLC-␤ activated via heterotrimeric G-proteins (39). Comparisons of PLC isoforms ␦, ␥, and ␤ suggest that PLC-␦ was present in single-celled eukaryotes, while the ␤ and ␥ forms arose after the development of multicellular organisms ϳ940 million years ago (40). Consequently, parasites likely contain only the ancestral PLC-␦ isoform. Consistent with this idea, T. cruzi contains a PLC-␦ isoform (41) and PLC-␦ candidate sequences are present in the Plasmodium falciparum and Plasmodium vivax genomes (42). While not conventionally associated with Ca 2ϩ signaling, PLC-␦ has recently been linked to Ca 2ϩ transients associated with fertilization in mice (43). Thus, it is possible that PLC-␦ plays an important role in protozoan calcium regulation.
Previous evidence for IP 3 -responsive Ca 2ϩ release channels in unicellular organisms has been indirect. Entamoeba histolytica crude membranes bind radiolabeled IP 3 (44), and a crossreacting IP 3 receptor antibody detects an appropriately sized protein in the ciliate Blepharisma japonicum (45,46). Combined with the present findings, these results indicate that single-celled eukaryotes also likely have intracellular calcium release channels that respond to IP 3 . However, our attempts to identify IP 3 or ryanodine receptor channels in T. gondii using a cross-reactive antibody (mouse monoclonal antibody 34C antichicken ryanodine receptor) (47) have not been successful. IP 3 and ryanodine receptors are ϳ40% homologous between Caenorhabditis elegans and vertebrates, and they are likely to be significantly more divergent in protozoa due to their early branching phylogenetic position (48). We have analyzed the nearly complete genomes of P. falciparum (42) and Cryptosporidium parvum (www.parvum.mic.vcu.edu/) and the Apicomplexan expressed sequence tag databases (paradb.cis.upenn. edu/) and found that these organisms do not contain well recognized motifs for IP 3 or ryanodine receptors such as SPRY and MIR domains. Consequently, it is likely that protozoa such as T. gondii possess intracellular calcium release channels that are unconventional in sequence and/or function.
Our observations suggest that IP 3 catalyzes [Ca 2ϩ ] i release in the parasite through an intracellular channel related to the IP 3 ryanodine receptor superfamily during T. gondii invasion of host cells. To test whether production of IP 3 was necessary for microneme secretion and host cell attachment, we utilized xestospongin C, a compound that inhibits IP 3 -dependent Ca 2ϩ release from rat cerebellar microsomes (31) and Dictyostelium vesicles (34). Xestospongin C acts primarily as a competitive inhibitor of IP 3 to block calcium release from IP 3 receptors; however, it can also inhibit ryanodine receptors with somewhat lower potency (31). Xestospongin C treatment of T. gondii inhibited ethanol-and caffeine-mediated secretion of micronemes and prevented attachment and invasion of host cells. Together, these results imply that xestospongin C inhibits microneme secretion by modulating a T. gondii [Ca 2ϩ ] i release channel that responds to IP 3 . These findings suggest that the engagement of natural ligands during host cell invasion may also give rise to IP 3 to promote the observed increases in [Ca 2ϩ ] i in the parasite reported previously (10).
The results of our studies using Ca 2ϩ release channel agonists and antagonists strongly suggest that the protozoan parasite T. gondii possess intracellular calcium release channels of the IP 3 /ryanodine superfamily. Relatively low concentrations of caffeine, ryanodine, and xestospongin C affected microneme release, indicating that protozoa may be unusually sensitive to these agents that affect calcium channels. While we cannot rule out secondary effects on other pathways, the primary response observed to all these compounds is consistent with their acting on a calcium release channel. A model integrating our findings with a hypothetical signaling cascade is presented in Fig. 5. We show that ethanol increases IP 3 production, a second messenger that could act to liberate the [Ca 2ϩ ] i pool. We also show that ryanodine increases [Ca 2ϩ ] i , suggesting it interacts directly with a putative [Ca 2ϩ ] i release channel via a Ca 2ϩ -induced Ca 2ϩ release mechanism. The calcium-dependent stimulation of microneme secretion by ryanodine, caffeine, or ethanol was thapsigargin-insensitive, indicating that this unique pool may be responsible for calcium signaling in parasites. During contact with host cells, it is likely that a similar cascade is triggered to release Ca 2ϩ from an intracellular store and stimulate microneme secretion. A role for a putative parasite PLC in the conversion of phosphatidylinositol bisphosphate to IP 3 and diacylglycerol is suggested by the available data but has not yet been directly demonstrated. The responsiveness to both IP 3 and ryanodine suggests that calcium release channels may exist as two separate entities. However, we have not excluded the intriguing possibility that protozoa possess a single ancient channel displaying hybrid properties of both. Analysis of calcium channels in protozoa will be illuminating to studies of the function and evolution of eukaryotic calcium signaling proteins.