An Extracellular Redox Signal Triggers Calcium Release and Impacts the Asexual Development of Toxoplasma gondii

The ability of an organism to sense and respond to environmental redox fluctuations relies on a signaling network that is incompletely understood in apicomplexan parasites such as Toxoplasma gondii. The impact of changes in redox upon the development of this intracellular parasite is not known. Here, we provide a revised collection of 58 genes containing domains related to canonical antioxidant function, with their encoded proteins widely dispersed throughout different cellular compartments. We demonstrate that addition of exogenous H2O2 to human fibroblasts infected with T. gondii triggers a Ca2+ flux in the cytosol of intracellular parasites that can induce egress. In line with existing models, egress triggered by exogenous H2O2 is reliant upon both Calcium-Dependent Protein Kinase 3 and diacylglycerol kinases. Finally, we show that the overexpression a glutaredoxin-roGFP2 redox sensor fusion protein in the parasitophorous vacuole severely impacts parasite replication. These data highlight the rich redox network that exists in T. gondii, evidencing a link between extracellular redox and intracellular Ca2+ signaling that can culminate in parasite egress. Our findings also indicate that the redox potential of the intracellular environment contributes to normal parasite growth. Combined, our findings highlight the important role of redox as an unexplored regulator of parasite biology.


INTRODUCTION
Toxoplasma gondii is a single-cell obligate intracellular parasite from the Apicomplexa phylum that can infect any warm-blooded animal. Its seroprevalence is estimated at more than one-third of the human population (Halonen and Weiss, 2013;Flegr et al., 2014). Within the host, the asexual lifecycle of T. gondii exists as two distinct stages: rapidly proliferating tachyzoites that characterize the acute infection, and slower replicating encysted bradyzoites that are associated with chronic infection (Sheffield and Melton, 1968;Dubey, 1996). While infections are usually benign, in immunocompromised patients and foetuses (Luft and Remington, 1992;Weiss and Dubey, 2009) the lytic tachyzoite lifecycle is responsible for severe clinical pathology. During a lytic cycle, tachyzoites attach onto and actively penetrate host cells, forming a permissive replication niche called the parasitophorous vacuole (PV). Parasites then replicate by endodyogeny (Goldman et al., 1958) until they eventually egress from the host cell, leading to its lytic destruction. The interconversion of virulent tachyzoites with persistent encysted bradyzoites is influenced by host immune pressure and is key to understanding disease recrudescence (Montoya and Liesenfeld, 2004). The ability of T. gondii to infect diverse host species relates to its remarkable capacity to resist host defences, efficiently recognize and quickly respond to myriad environmental clues. Among biological signal cascades activated by external clues, Ca 2+ signaling is the most notable and best studied in T. gondii [reviewed in (Lourido and Moreno, 2015)]. Ca 2+ is a ubiquitous signaling molecule with a vital role in tachyzoite host-cell invasion (Vieira and Moreno, 2000;Pace et al., 2014), motility (Wetzel et al., 2004) and egress (Endo et al., 1982) through activation of effector proteins such as the plant like Ca 2+ -dependent Protein Kinase 1 (CDPK1) (Murphy et al., 2010;Ojo et al., 2010) and CDPK3 (Garrison et al., 2012). Alongside Ca 2+ , other intracellular signaling molecules are known to play important roles in the tachyzoite lytic cycle. These include phosphatidic acid (PA) (Bullen et al., 2016), and the activation of protein kinase G by cyclic guanidine monophosphate (cGMP) (Brown et al., 2016).
Contrasting with Ca 2+ , the study of reactive oxygen species (ROS) as signaling molecules is incipient in the Apicomplexa. Hydrogen peroxide (H 2 O 2 ) is a neutral ROS molecule with an ability to cross membranes (Guzik et al., 2000), and its roles in signaling are diverse. These include its interaction with glutathione (GSH), and the oxidation of cysteine residues leading to allosteric changes in a variety of proteins such as phosphatases, transcription factors and ion channels (Winterbourn and Hampton, 2008). Typically, the oxidation of redox-sensitive cysteines is a reversible process catalysed by enzymes that use GSH or nicotinamide adenine dinucleotide phosphatase (NADPH) (Xiong et al., 2011;Pastore and Piemonte, 2012;Sabens Liedhegner et al., 2012;Popov, 2014) as redox cofactors. Most studies of H 2 O 2 and T. gondii tachyzoites have focused on the ability of host cells to use the damaging oxidative properties of ROS as a component of innate defence, and strategies employed by tachyzoites to overcome this defence (Murray and Cohn, 1979;Ding et al., 2004;Akerman and Muller, 2005;Luder and Gross, 2005;Luder et al., 2009).
A T. gondii orthologue of the H 2 O 2 -detoxifying enzyme catalase is expressed in the parasite cytosol, and confers protection against host oxidative stress (Kaasch and Joiner, 2000;Ding et al., 2004). In Ding et al. (2004, compiled a group of 14 genes related to the T. gondii antioxidant system. These typically localized to the parasite cytosol and mitochondria, and were assigned to one of five major redox system classifications: (1) metabolic genes (e.g. superoxidedismutase and catalase); (2) thioredoxins (Trxs, proteins that promote cysteine thiol-disulfide exchange); (3) Protein Disulfide Isomerases (PDIs, proteins that disrupt or form cysteine disulfide bonds to assist protein folding); (4) glutaredoxin-glutathione (Grx-GSH, small proteins that use GSH as a cofactor for thiol-disulfide exchange) and (5) peroxiredoxins (Prxs, enzymes that detoxify hydroperoxides like H 2 O 2 and organic hydroperoxides).
Expanding this broad description of the parasite's antioxidant system, other studies have tested the association of redox with T. gondii signaling and cell cycle regulation. Exogenous treatment of intracellular parasites with the reducing agent dithiothreitol (DTT) triggers Ca 2+ mobilization and parasite egress (Stommel et al., 1997). This egress response was triggered by the depletion of host-cell ATP resulting from the activity of an exported parasite ATPase (Silverman et al., 1998). In a separate study, the oxidation-sensitive protein TgDJ-1 was found to associate with CDPK1 and promote microneme secretion in T. gondii (Child et al., 2017). More recently, oxidative stress (generated by sodium arsenite) has been shown to trigger tachyzoite differentiation into bradyzoites following phosphorylation of T. gondii eIF2a (TgIF2a) by the translation initiation factor kinase TgIF2K-B (Augusto et al., 2020). Together, these data suggest that T. gondii modulates biological processes in response to changes in redox homeostasis.
Here, we use an in silico approach to establish a compendium of redox-associated genes and provide an updated view of these genes in T. gondii. We then investigate the impact of H 2 O 2 upon parasite biology, demonstrating that a H 2 O 2 signal outside the boundaries of the infected host cell can be received and interpreted deep within the cytosol of intracellular parasites. We find that exogenous treatment of H 2 O 2 triggers mobilization of Ca 2+ culminating in CDPK3-dependent egress. Finally, we use a genetically encoded redox reporter to dissect redox oscillations prior to egress. Unexpectedly, we discover that overexpression of the active catalytic domain of a human Grx in the parasite's cytosol or PV delays parasite asexual replication. Our results corroborate the existence of a rich variety of antioxidant proteins located in multiples cellular compartments and highlight importance of redox in the basic biology of T. gondii.

Parasite and Host Cell Culture
T. gondii tachyzoites from strain RH (Type I) lacking hypoxanthineguanine phosphoribosyl transferase gene (HXGPRT) were cultivated in vitro on monolayers of primary human foreskin fibroblast (HFF, ATTC ® ) in a humidified incubator at 37°C, 5% CO 2, 3% O 2 atmosphere and maintained in Dulbecco's Modified Eagle medium (DMEM) supplemented with 10% of foetal bovine serum (FBS) and 2mM L-glutamine, without antibiotics. All culture were tested against Mycoplasma infection on a monthly basis.

Identification of T. gondii Genes Related to Redox
Within the Toxoplasma database (ToxoDB (Gajria et al., 2008), release 50 beta 17 Dec 2020), a list of genes related to redox signaling in T. gondii were obtained using the keywords "thioredoxin", "glutathione", "glutaredoxin", "peroxiredoxin" and "protein disulfide" on gene text search option. The function domains of the canonical antioxidant groups Trxs, Grx-GSH, Prxs and PDIs were confirmed for each gene using Basic Local Alignment Search Tool (BLAST) at National Center for Biotechnology information (NCBI) database (www.ncbi.nlm. nih.gov/). The spatial protein localization related to each redox gene were extracted from the localisation of organelle proteins by isotope tagging (hyperLOPIT) (Crook et al., 2018;Barylyuk et al., 2020) dataset available through ToxoDB.

Plaque and Replication Assays
For plaque assays, tachyzoites were harvested from infected HFF by syringe passage followed by filtration (5mm).~100 tachyzoites were added per well of a 6-well plate prepared with confluent HFF monolayers, and allowed to grow undisturbed for 6 days. Plates were washed with phosphate-buffer saline (PBS) and fixed with cold methanol, stained with crystal violet and scanned. Plaque counts and area measurements were performed using FIJI software by drawing region of interesting (ROI). For replication assays, the freshly lysed tachyzoites were added to HFF monolayers grown on a m-slide 8 well glass bottom chamber (Ibidi ® ) with multiplicity of infection (MOI) 1. To synchronize the infection, parasites were allowed to settle onto chilled host cells for 20 minutes, and then allowed to invade for 2 hours at 37°C. Cells were washed with PBS to remove extracellular tachyzoites. Cells were incubated for a further 18 hours, and subsequently fixed with 3% paraformaldehyde at room temperature (RT) for 20 min, and blocked with PBS supplemented with 2% FBS. Parasites per vacuole were counted by excitation with 470nm laser (4% intensity) widefield Nikon Eclipse Ti-E inverted microscope equipped with an ORCA-Flash4.0 camera (Hamamatsu, Japan) and NIS-Elements Viewer software (Nikon), 60x-oil objective. All parasite strains were genetically encoding for GFP or roGFP2. All strains were tested tree independent times, each with three technical replicates.

Egress Assay
Freshly lysed tachyzoites from RH-GFP-t2a-jRCaMP1b or RH-GFP-T2A-jRCaMP1bDCDPK3 parasites were harvested and inoculated (MOI:1) onto confluent HFF cells grown on 24 well plates and grown for 18 hours. Cells were washed with PBS and the growth medium replaced with phenol red free DMEM without FBS and incubate for a further three hours. Medium was removed and cells were incubated for 30 minutes with PBS plus drugs (A23187, R59022, H 2 O 2 ) or vehicle controls (water, DMSO). At the end of incubation, the supernatant of the wells was carefully aspirated and cells were detached by 10 min incubation with 200 mL of RT Accutase ™ . Accutase-release cell suspensions were fixed with 200 mL of 8% paraformaldehyde (20 minutes, RT). Fixed cells were transferred into falcon 5 mL tube with 35 mm nylon mesh cap. Free fluorescent parasite and infected HFF population were analysed and quantified in a BD LRSFortessa ™ . A total of 5000 events were collected for each tube. Blue laser (488nm) with 530/30 nm filter was used to detect the GFP signal. An uninfected HFF control was used to assess the effect of each drug treatment on host cell gating. Each drug condition was tested three independent times, each with six technical replicates.

Fluorescence Microscopy
For cytosolic Ca 2+ fluorometric measurements, RH-GFP-T2A-jRCaMP1b or RH-GFP-T2A-jRCaMP1bDCDPK3 parasites were added (MOI:1) to confluent HFF grown on 8 well glass bottom chamber and allowed to grow for 18 hours. Wells were washed with PBS and media replaced by phenol red-free DMEM medium without FBS, and incubate for a further two hours. Images of live infected HFFs were captured at 37°C, using a 60xoil objective in the same widefield microscope previously describe on replication assay. Only vacuoles containing 4 to 8 parasites were considered for data analyses. GFP (470 nm excitation/520 nm emission) and jRCaMP1b (555 nm excitation/605 nm emission) with 100 milliseconds acquisition rate signal were collected every second for up to a maximum 10 minutes. Drugs were applied on the wells after one minute of acquisition by pipetting. Image analyses were performed using FIJI software. Raw fluorescence readout (F) for each vacuole on each channel was normalised against the average of the baseline signal before adding the drug (F 0 ) using the ratio F/F 0 bringing the resting baseline value to one. To distinguish intracellular Ca 2+ oscillation signal from vacuole movement, the F (jRCaMP1b) /F 0 (jRCaMP1b) values were normalised against the GFP signal F (GFP) / F 0(GFP) . Ca 2+ response during a specific time is given by the formula: (F (jRCaMP1b) /F 0(jRCaMP1b) )/(F (GFP) /F 0(GFP) ).

Chemical and Reagents
Hydrogen peroxide 30% (w/w) solution with stabilizer; diacylglycerol kinase inhibitor (R59022); A23187, Ionomycin, anhydrous methanol, and crystal violet solution were obtained from Sigma-Aldrich Company Ltd. a-tocopherol phosphate disodium salt from Merk. Dimethylsulfoxide (DMSO) anhydrous and paraformaldehyde 16% solution from Life Technologies. Accutase cell detachment solution from Fisher Scientific Ltd.

Statistical Analysis
The data are represented as the mean ± SEM and analysed using two-tailed paired Student t test between two groups and one-way or 2way analyses of variance (ANOVA) with Bonferroni's multiple comparisons test for comparing means between ≥3. All data were analysed using GraphPad Prism 9 software (California, USA). The data were considered statistically significant when P values <0.05.

Redox-Associated Genes
We initially sought to update our understanding of the antioxidant response and redox-signaling network in T. gondii. We mined gene sequence and annotation information, as well as proteomic datasets present on ToxoDB (Gajria et al., 2008) to update the list of 14 redox associated genes previously summarised by Ding et al. (2004). Our bioinformatic approach screened for genes containing at least one functional domain of the major redox signaling groups (Trxs, PDIs, Grx-GSHs and Prxs). We identified a total of 58 redox-associated genes (Supplementary Table 1), including 26 Trxs, 16 GRX-GSHs, six PDIs, six Prxs, and four metabolic genes (including three distinct superoxide dismutases and one catalase). With the exception of the metabolic genes (Sibley et al., 1986;Kaasch and Joiner, 2000;Odberg-Ferragut et al., 2000;Ding et al., 2004), the majority of genes representing the other redox signaling groups remain uncharacterized. For an improved view of the subcellular distribution of these gene products, we extracted their primary localisation from the recently published hyperLOPIT dataset (Barylyuk et al., 2020) (Figure 1 and Supplementary Table  1). This indicated a broad distribution of these proteins throughout the cell and suggested the existence of spatially distinct mechanisms of redox regulation for different subcellular compartments.

Extracellular Hydrogen Peroxide Induces Cytosolic Ca 2+ Flux in Intracellular Tachyzoites
In other eukaryotic systems there are clear associations between calcium and redox signaling. For example, oxidation of cysteines on the ryanodine receptor stimulates calcium release from intracellular compartments (Meissner, 2002;Oda et al., 2015). To investigate if a similar overlap exists between these two signaling networks in T. gondii, we tested the ability of H 2 O 2 to trigger a Ca 2+ response in tachyzoites within a host cell. We generated a Type 1 parasite (RH strain) expressing the Ca 2+ sensor protein jRCamP1b (Dana et al., 2016) fused to GFP via the T2A peptide for bicistronic protein expression (referred to as RH-GFP-T2A-jRCaMP1b). The jRCaMP1b is a bright red fluorescent sensor for Ca 2+ with an improved dynamic range and kinetic compared to jRCaMP1a (Dana et al., 2016). The ratio of these two distinct fluorescent proteins provided an elegant tool to distinguish changes in fluorescence due to cell movement from the Ca 2+ signal (Figure 2 and Supplementary Figures 1A, B). Upon exogenous addition of 100 mM H 2 O 2 we observed a distinct increase in the intracellular Ca 2+ reporter signal (Figure 2A and Supplementary Videos 1-3). To determine whether the oxidative properties of H 2 O 2 are responsible for triggering parasite Ca 2+ release, we repeated the experiment in the presence of the antioxidant a-tocopherol ( Figure 2B).
Pre-treatment of infected host cell monolayers with 50 mM atocopherol abolished H 2 O 2 -stimulated Ca 2+ release in intracellular parasites. The addition of the Ca 2+ ionophore A23187 caused an intense Ca 2+ signal spike in the parasite cytosol in both the presence and absence a-tocopherol (Figures 2A, B, respectively), suggesting that the presence of the antioxidant does not interfere with Ca 2+ stores in intracellular parasites. Vehicle solvent for H 2 O 2 (water) and for A231877 (DMSO) did not trigger Ca 2+ mobilisation (Figure 2 and Supplementary Figures 1C, D, respectively).
We compared the magnitude of the Ca 2+ signal increase following H 2 O 2 treatment with the increase triggered by A23187, a small molecule known to induce parasite egress by Ca 2+ -dependent mechanisms  ( Figure 2C). H 2 O 2 induced an overall Ca 2+ signal increase of 35% ± 1.7. This was lower than the Ca 2+ signal increase observed when parasites were treated with A23187 alone (140 ± 4.3), or A23187 on cells pre-treated with 100 mM H 2 O 2 (61% ± 0.9). Within restricted timeframe of this experiment (10 minutes), a single egress event was observed for cells treated with 100 mM H 2 O 2 ( Figure 2C and Supplementary Video 2). Observation of egress induced by H 2 O 2 was more frequent when host cells were infected with a high multiplicity of infection (MOI = 5) (Figure 2 and Supplementary Video 3). We also measured the magnitude of Ca 2+ signal increase induced by H 2 O 2 in DCDPK3 parasites ( Figure 2D). The magnitude of the Ca 2+ signal induced by H 2 O 2 in RH-GFP-T2A-jRCaMP1bDCDPK3 was lower (18.1% ± 4.3) compared to RH-GFP-T2A-jRCaMP1b (32.5% ± 1.1). This could reflect previously observed differences in the resting Ca 2+ levels of these lines (Treeck et al., 2014). Together, these data suggest that tachyzoites within the PV can perceive and respond to oxidation events initiated outside the host cell.

Hydrogen Peroxide Induces Parasite Egress in a Mechanism Dependent on CDPK3
Calcium flux accompanies natural egress (Stewart et al., 2017), and calcium ionophores are well-characterized inducers of egress (Endo et al., 1982). To better investigate the potential of H 2 O 2 to FIGURE 1 | Schematic representation of a T. gondii tachyzoite displaying 58 redox-associated genes, and their primary protein location as determined by HyperLOPIT (Barylyuk et al., 2020). Gene ID accession numbers are provided, and genes categorized into five groups: black (Trxs); red (PDIs); green (Prxs); blue (Grx-GSHs) and orange (metabolic genes). Further details are provided in Supplementary Table 1. induce parasite egress, we analysed populations of infected host cells by flow cytometry. The GFP signal from parasites expressing the Ca 2+ sensor and particle size were used to distinguish infected host cells from free tachyzoites (Figure 2 and Supplementary Figure 2A). We tested whether treatment of infected host cells with different concentrations of H 2 O 2 could increase the proportion of free tachyzoites. Incubation of infected host cells with H 2 O 2 resulted in a dose-dependent increase in the number of free tachyzoites ( Figure 2E). We sought to understand how the H 2 O 2 stimulated egress integrates into our current molecular understanding of this process. The small molecule R59022 is an inhibitor of diacylglycerol kinase (Cooke et al., 1987) that inhibits egress by disrupting the formation of phosphatidic acid (Bullen et al., 2016). For all concentration tested, treatment of cells with H 2 O 2 resulted in a pharmacological rescue of R59022 egress inhibition. However in the presence of R59022, H 2 O 2 did not stimulate egress above the baseline value. This could relate to the discrete points of activity for the two targets of R59022: diacylglycerol kinase 1 and 2 (Bullen et al., 2019)  After confirming that exogenous addition of H 2 O 2 stimulated Ca 2+ flux in intracellular T. gondii parasites, we investigated the redox state within the parasite cytosol and PV. We reasoned that if calcium release was the direct result of an oxidative signal, the redox status of cellular compartments separating intracellular parasites from the extracellular environment should also be affected. To test this, we generated transgenic parasite strains constitutively expressing the redox sensor protein GRX1-roGFP2, in two different cellular compartments: the parasite cytosol, or targeted to the PV as a consequence of an N-terminal fusion with the GRA8 signal (C) Intensity of parasite Ca 2+ signal increase in RH-GFP-T2A-jRCaMP1b parasites (expressed as a percentage over baseline) following addition of: 1 mM A23187 alone, 1 mM of A23187 following 100 mM H 2 O 2 pre-treatment, and 100 mM H 2 O 2 alone. Gray dots indicate vacuole data points where parasite egress was observed during the measurement period, blue dots represent vacuoles where egress was not observed. (D) Intensity of parasite Ca 2+ signal increase (%) over the baseline upon addition of 100 mM H 2 O 2 in RH-GFP-T2A-jRCaMP1b versus RH-GFP-T2A-jRCaMP1bDCDPK3 parasites. For (C, D) histograms present data mean ± SEM of three independent experiments (five vacuoles measured in each experiment), with individual vacuole data points also shown. (E, F) Egress assay measuring tachyzoite release after compound treatment in RH-GFP-T2A-jRCaMP1b and RH-GFP-T2A-jRCaMP1bDCDPK3, respectively. Data represent the mean ± SEM of three independent experiments (except for DMSO that has two independent experiments), with six technical replicates for each. All data were normalised to the water control. Significance was calculated using one-way Anova, Bonferroni's multiple comparisons test. P values: *< 0.05; **< 0.01, ***< 0.001 and ****< 0.0001. sequence (Carey et al., 2000). GRX1-roGFP2 is a ratiometric redox reporter that detects changes in both reduced (GSH) and oxidized (GSSG) glutathione (Aller et al., 2013) (Figures 3A, B). Importantly, the redox relay system underpinning GRX1-roGFP2 is not affected by pH, which is known to confound data interpretation with other redox sensor proteins (Gutscher et al., 2008). We used GRX1-roGFP2 parasites to track dynamic changes in GSH/GSSG by fluorescence microscopy, and used the normalized GSH/GSSG signal ratio to measure the intensity of oxidation events ( Figure 3C and Supplementary Figure 3A). Intracellular parasites expressing the redox sensor targeted to the PV (RH-GRA8-GRX1-roGFP2) or cytosol (RH-GRX1-roGFP2) detected an oxidation event upon exogenous addition of 100 mM H 2 O 2 ( Figures 3D, E, respectively). The oxidation event within the PV was of greater magnitude compared to that detected within the parasite cytosol, indicating that the strength of the oxidative signal was diminished as it crossed the biological membrane separating these compartments. The water control did not affect the redox signal from GRX1-roGFP2 sensor (Figure 3 and Supplementary Figure 3B).
Having observed that H 2 O 2 could stimulate calcium flux, we investigated the reciprocal nature of this relationship by testing the ability of calcium ionophores to trigger an oxidative event. Using GRX1-roGFP2 parasites, we measured the resting redox state prior to ionomycin-induced egress. Ionomycin-induced egress was not accompanied by detectable changes in redox within either the PV or parasite cytosol ( Figures 3F, G, respectively). The ionophore A23187 could not be used because of saturating autofluorescence associated with this small molecule in the fluorescence channel used to measure oxidation (Figure 3 and Supplementary Figure 3C). These data suggested that no significant change in the GSH/GSSG ratio occurs before parasite egress induced by ionomycin.

GRX1-roGFP2 Redox Sensor Affects Parasite Fitness During Asexual Replication
We noted that one unavoidable result of using the GRX1-roGFP2 redox relay sensors would be the associated overexpression of the catalytic domain of glutaredoxin (as a consequence of it being fused to roGFP2). GRX1 is a small redox enzyme that confers protection against oxidative stress (Gutscher et al., 2008;Laporte et al., 2012), and the overexpression of this enzyme would be expected to shift GSH/GSSG ratios in favour of the reduced form (GSH). Correspondingly, we hypothesized that T. gondii strains overexpressing GRX1 would have their normal redox status shifted to a more reduced potential. To test whether the overexpression of this redox protein affected parasite asexual growth, we generated a transgenic parasite strain expressing a version of the redox sensor where we had mutated the key catalytic cysteine residue of GRX to render it enzymatically inactive (GRX1 ser -roGFP2). As a result, GRX1ser-roGFP2 sensor did not respond changes in GSH/GSSG following H 2 O 2 treatment (Figure 4 and Supplementary Figure 4). We compared the growth of this strain with parasites expressing the catalytically active version of the sensor. As before, both active and inactive versions of the GRX-fusion sensor were targeted to either the parasite cytosol or PV. As a control group, we used a transgenic parasite strain expressing redox-insensitive GFP (RH-GFP-Luc) (Ploemen et al., 2009). This strain was used as a control in preference to RHDku80DHXGPRT parasites. We reasoned that it would better control for any metabolic cost and fitness impact resulting from the expression of a large non-native sensor. It would also allow for live imaging of both control and redox sensor strains in parallel. Plaque assays confirmed that RH-GFP-Luc parasites were able to form plaques on confluent monolayers of HFFs with equal efficiency to the RHDku80DHXGPRT ( Figure 4A and Supplementary Figure 5). Parasites expressing catalytically active GRX1 in the cytosol presented fewer plaques compared to RH-GFP-Luc ( Figure 4A). Despite being able to successfully maintain PV-targeted sensor strains in culture, all parasites lines where the sensor was targeted to the PV did not form clear measurable plaques after six days of growth. Intensely stained plaque-shaped boundaries visible on the HFF monolayer suggested these parasites had successfully grown, but that parasite lytic growth had not outcompeted host cell monolayer recovery sufficiently to produce a clear zone of lysis ( Figure 4B). When either the catalytically active or inactive sensors were targeted to the parasite cytosol, parasites formed smaller plaques compared to the RH-GFP-Luc control ( Figures 4B, C, respectively). These data suggested that the presence roGFP2 alone might be sufficient to affect parasite growth. To directly test the influence of redox environment on the lytic cycle, we grew RH-GFP-Luc parasites with 10 mM N-acetyl cysteine (NAC), a small antioxidant molecule that functions by donating cysteine to increase GSH biosynthesis (Ezerina et al., 2018). Addition of NAC decreased the size of plaques generated by RH-GFP-Luc parasites ( Figure 4C), suggesting that GSH/GSSG imbalance compromised parasite growth. To better understand the impact of overexpressing the sensor on parasite growth, we counted the number of parasites per vacuole after 20 hours of asexual parasite replication ( Figure 4D). All strains expressing the redox sensor had reduced replication compared to RH-GFP-Luc ( Figure 4D and Supplementary Figure 6), with catalytic inactivation of the GRX domain providing a partial rescue of the replication defect. Strains exhibiting the slowest replication were those where the sensor was targeted to the PV, supporting the plaque growth data in Figure 4B. Together, these data suggest parasite growth is sensitive to changes in GSH/ GSSG ratios.

DISCUSSION
Investigations of ROS and T. gondii biology initially focused on the host innate immune response. Host macrophages can generate oxidative bursts, creating a toxic microenvironment to fight microbial infection (McLeod et al., 1983;Sibley et al., 1985;Dupre-Crochet et al., 2013), with early studies interested in mechanisms used by T. gondii to evade oxidative stress (Wilson et al., 1980;Shrestha et al., 2006;Lima and Lodoen, 2019). However, increasing evidence from multiple organisms has demonstrated that H 2 O 2 has functions as a signaling molecule [reviewed in (Veal and Day, 2011)], suggesting the likelihood of a more complex role for ROS in the pathophysiology of T. gondii. To sense, respond, and protect against potential oxidative damage from ROS stimuli, cells employ a network of redox-associated proteins such as Trxs, PDIs, Prxs, Grxs, superoxide dismutase and catalase. It was previously shown that T. gondii possesses all these elements (Ding et al., 2004), and we provide an update on the number, diversity and cellular distribution of these redox-associated proteins (Figure 1). The function of some of these redox-associated gene products in T. gondii parasites has been examined: disruption of catalase decreases parasite virulence in mice and in vitro tolerance to H 2 O 2 (Ding et al., 2004); peroxiredoxin-1 interacts with histone lysine methyltransferase to likely regulate gene expression by chromatin rearrangement (Sautel et al., 2009); two thioredoxins have essential roles in apicoplast biogenesis (Biddau et al., 2018), and another thioredoxin with cytosolic localization has been associated with parasite virulence (Xue et al., 2017). Nevertheless, functional information for most of the redox-associated proteins identified in this work remains elusive. The ability of low concentrations of H 2 O 2 to initiate an intracellular Ca 2+ response in mammalian systems is well documented (Roveri et al., 1992;Volk et al., 1997;Avdonin et al., 2017) and has helped to solidify H 2 O 2 as a signaling molecule. Addition of antioxidants blocks intracellular Ca 2+ release induced by H 2 O 2 in smooth muscle cells (Roveri et al., 1992) suggesting that the oxidative properties of H 2 O 2 are required for Ca 2+ signaling. In human endothelial cells, oxidation induced by non-toxic concentrations of H 2 O 2 target a Ca 2+ channel located in acid compartments (Avdonin et al., 2017). Distinct from these models, this work is the first to report that an oxidation event induced by H 2 O 2 mobilizes intracellular Ca 2+ in T. gondii tachyzoites within the infected host cell (Figure 2A). This is the first time this has been shown for any apicomplexan parasite. Studying the parasite within the host cell provides the best approximation of physiological conditions to observe both Ca 2+ and redox signaling. Moreover, using parasites expressing a genetically encoded fluorescent Ca 2+ sensor abrogated the need to use fluorescent Ca 2+ indicator dyes. Use of these dyes can be damaging to the cells, and typically require the use of other small molecules to avoid dye loss and compartmentalization (Di Virgilio et al., 1990). GFP coexpressed with Ca 2+ sensor allows Ca 2+ responses to be distinguished from parasite movement which avoids the need to use inhibitors of parasite motility such as cytochalasin D (Carter, 1967). Finally, the concentration of 100 mM H 2 O 2 has been showed to be non-toxic for human fibroblast (Hyslop et al., 1995). Altogether, our protocol to investigate intracellular Ca 2+ release following H 2 O 2 treatment sought to avoid cellular stress that could compromise the true redox/Ca 2+ dynamic within the parasite. As previously report in muscle cells (Roveri et al., 1992), the presence of antioxidant inhibited the parasite Ca 2+ response triggered by H 2 O 2 . Using the GRX1-roGFP2 redox sensor, we confirmed that intracellular parasites directly sense an oxidation event following the exogenous addition of 100 mM H 2 O 2 . It is not clear whether the exogenous addition of 100 mM of H 2 O 2 results in H 2 O 2 reaching the intracellular parasites as the host cell contains an extensive network of antioxidant proteins that would be expected to scavenge and neutralise H 2 O 2 . This could be directly addressed in future experiments using T. gondii strains expressing a sensor to specifically detect H 2 O 2 (Gutscher et al., 2009).
Regardless of whether Ca 2+ is mobilized within the parasite as a direct result of an interaction with H 2 O 2 , or via secondary oxidation signal from the host, Ca 2+ regulates all aspects of parasite host cell invasion (Lourido and Moreno, 2015) included egress (Garrison et al., 2012). For the short time period used to track parasites by microscopy, the intensity of the Ca 2+ signal response induced by H 2 O 2 was quite modest compared to A23187, and that likely explains why egress events were rare. Longer incubations with H 2 O 2 induced parasite egress through a mechanism that requires CDPK3 and likely phosphatidic acid.
Introduction of the redox sensors described in this work into the DCDPK3 parasite background could provide direct evidence its role in the crosstalk between calcium and redox. Our work provides the first evidence that suggests that oxidation can trigger T. gondii egress. Interestingly, at the other end of the redox spectrum, the reductive molecule DDT can also mobilize Ca 2+ and induce parasite egress (Stommel et al., 1997). Although the mechanism by which H 2 O 2 and DDT lead to parasite egress appears to be distinct, these data indicate that redox can influence the parasite's lytic cycle. We anticipate that parasite biology is tuned to a specific environmental redox potential, and that the perturbation of redox homeostasis with either oxidative or reductive stress elicits a phenotypic response.
The mechanism for how H 2 O 2 mobilizes Ca 2+ in animals is better understood but the evolutionary distance between vertebrates and T. gondii makes direct comparisons more challenging. T. gondii is more closely related to plants, and shares a more similar signaling toolkit (Nagamune and Sibley, 2006). It has been recently reported that plants possess a cell surface H 2 O 2 receptor that once activated, triggers a Ca 2+ influx into the cells through a Ca 2+ ion channel (Wu et al., 2020). T. gondii also has a H 2 O 2 -sensitive protein that associates with the CDPK1 to promote microneme secretion (Child et al., 2017), an event that requires Ca 2+ mobilization to allow parasite invasion (Lovett et al., 2002).
Our work suggests a connection between oxidation and parasite Ca 2+ release but our data did not find evidence that ionophore-induced Ca 2+ release changes the redox state of either the parasite's cytosol or PV. The use of the GRX1-roGFP2 sensor to track redox changes based on GSH/GSSG allows a fast and selective assessment of redox fluctuations in real time (Aller et al., 2013), with improved dynamics compared to other redox sensors like roGFP1 or roGFP2 (Hanson et al., 2004). Considering that Ca 2+ mobilization is also a fast event, GRX1-roGFP2 is a suitable tool to investigate the relationship between ROS and Ca 2+ . Moreover, the fact that GRX1 does not directly interact with H 2 O 2 (Gutscher et al., 2008) implies this sensor would not interfere with any eventual interaction between H 2 O 2 and its potential targets within the parasite or PV. Unexpectedly, expression of GRX1-roGFP2 within T. gondii parasites was detrimental to asexual replication. This is surprising as the GRX1-based sensor is generally well tolerated, and presents negligible toxicity in neurons (Gutscher et al., 2008;Hasel et al., 2015). Inactivation of the catalytic domain of GRX1 only partially recovered parasite growth, suggesting that roGFP2 alone is sufficient to influence parasite replication. Cells contain millimolar concentrations of GSH within their cytoplasm and organelles, and under physiological conditions the cell maintains the majority of this redox buffer molecule in a reduced form ([GSH]>[GSSG]) (Brieger et al., 2012). Parasites overexpressing GRX1 would be expected to affect the normal GSH/GSSG balance. The addition of NAC, a small molecule that can be used to generate GSH, also slowed parasite growth. This supports the hypothesis that the GSH/GSSG ratio can influence parasite replication. Parasites expressing GRX1ser-roGFP2 are likely have altered redox potential due ability of roGFP2 not fused with an active GRX1 to interact directly with oxidizing molecules like H 2 O 2 .
Targeting of the redox sensor to the PV had the greatest effect upon parasite asexual growth. During asexual replication within host cells, T. gondii resides within the PV. This compartment separates the parasite from the host cytoplasm, providing a niche for parasite survival and replication (Clough and Frickel, 2017). Signaling molecules from the host must cross the PV in order to reach the parasite. Should an oxidative signal from the host encounter an unusually reductive environment within the PV due the buffering effect of the redox sensor, the signal could be lost before reaching the parasite. Our work increases our understanding of the complex redox system in T. gondii, and provides the first evidence that parasite replication is sensitive to redox imbalance. Redox clearly has an important role in T. gondii pathophysiology and we anticipate that the unveiling of this network will guide future covalent drug discovery targeting redox-sensitive chemically reactive cysteines.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.    The graph presents the independent GFP trace (green) and Ca 2+ signal from jRCaMP1b sensor (red). Note the parasite movement at 420s. (B) Graph of the Ca 2+ signal is normalized to GFP to minimize artefact on the Ca 2+ measurements due to parasite movement. Red arrow indicates the moment of parasite egress. (C) Water, the vehicle solvent for H 2 O 2 , does not mobilize Ca 2+ . The graph displays the trace of three independent vacuoles from the same field of view. Data are representative of 15 infected vacuoles from four independent experiments. (D) DMSO, the vehicle solvent for ionophore A23187, does not mobilize Ca 2+ . The graph displays the trace of four independent infected vacuoles from the same field of view. Data are representative of 12 rosettes from three independent experiments. (A-D) black arrows indicate the time of drug/solvent addition.

AUTHOR CONTRIBUTIONS
Supplementary Figure 2 | Quantification of RH-GFP-T2A-jCamP1b parasite egress by flow cytometry. (A) Representative gating using GFP (Blue laser, 488nm, filter 530/30) against Forward scatter (FSC-A). HFF water control: uninfected human foreskin fibroblast (HFF) are mainly localised in quadrant c. Lysed tachyzoites: free RH-GFP-T2A-jCamP1b parasites are mainly localised in quadrant a. iHFF water control: infected host cells are detected in quadrant b. By using the GFP it is possible to distinguish free fluorescent parasite from iHFF in a population and assess the egress rates throughout different treatments. (B) Effect of drug incubation on non-infected HFF. The graph presents a change on event number within the non-fluorescent HFF gate. Data represent the mean ± SEM of three independent experiments (except for DMSO treatment that has two independent experiment), six technical replicates on each. Significance was calculated using one-way Anova, Bonferroni's multiple comparisons P value ***< 0.001. RHDku80DHXGPRT plaque with similar efficiency. Bar graphs presenting plaque count data from a six-day plaque assay. The data presented is from three independent biological experiments, each with three technical replicates. No statistically significant difference was found between the two strains using two-tailed paired Student t test. (F) % of vacuoles with more than 32 parasites. Significance was calculated using two-way Anova, Bonferroni's multiple comparisons test. P values: *< 0.05; **< 0.01, ***< 0.001 and ****< 0.0001.