TgCRT-deficient parasites lose endolysosomal system integrity due to altered VAC morphology

Previous studies have localized TgCRT in the VAC by gene epitope-tagging and immunofluorescence microscopy, and utilized an anhydrotetracycline-regulated system to reduce levels of expression of TgCRT in a Type I Toxoplasma RH strain. These results revealed that TgCRT is involved in volume control of the VAC [15]. However, incomplete depletion of TgCRT limits further characterization of its function. Additionally, lack of detailed phenotypic characterization restricts our understanding of how the swollen VAC affects parasite fitness and virulence. Here, we adopted a genetically tractable RH-derived strain, termed RHΔku80 (hereafter referred to as WT), to produce a complete TgCRT knockout (refer to S1 Text, Table 1, and S1A and S1B Fig for more details). The RHΔku80 strain lacks non-homologous end-joining DNA repair, thus enhancing homology-dependent DNA recombination [24]. Due to the increased homologous recombination efficiency, this strain has been widely used as a wild type Toxoplasma strain.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Parasite strains used in this study.

https://doi.org/10.1371/journal.ppat.1007775.t001

Upon generating RHΔku80Δcrt, we observed that purified extracellular Δcrt parasites exhibited large “concave” subcellular structures under differential interference contrast (DIC) microscopy, whereas WT and ΔcrtCRT strains did not display this phenotype (Fig 1A). This subcellular structure was also observed in pulse-invaded Δcrt parasites (Fig 1B). To identify the swollen structures, we stained the WT, Δcrt, and ΔcrtCRT parasites with anti-TgCPL antibodies. TgCPL is a major luminal endoprotease in the VAC of Toxoplasma [4,25]. Immunofluorescence microscopy showed that TgCPL staining co-localized with concave subcellular structures in Δcrt (Fig 1B). The TgCPL staining in Δcrt was larger than that in WT and ΔcrtCRT parasites, indicating that the VAC becomes swollen when TgCRT is absent. We quantified the VAC sizes based on TgCPL staining as described previously [13,15]. VAC diameter for the Δcrt parasites (1.12 ± 0.07 μm) is approximately 2.6-fold larger than for WT parasites (0.43 ± 0.03 μm), while the ΔcrtCRT (0.46 ± 0.02 μm) VAC was similar to that measured for WT parasites (Fig 1B). If we assume the VAC is approximately spherical, then the Δcrt parasite VAC is approximately 15-fold larger than the WT VAC. In contrast to pulse-invaded parasites, the swollen concave structure was not observed in replicated Δcrt parasites (Fig 1C). However, TgCPL staining showed differences between WT and Δcrt parasites (Fig 1C). For WT, the VAC displayed dynamic structures during replication that appeared as small fragmented puncta upon TgCPL staining [4]. However, TgCPL staining revealed fewer small fragmented punctate structures in replicating Δcrt parasites and instead featured one or more larger and intensely stained puncta (Fig 1C). Overall, we found that loss of TgCRT severely alters the morphology of the VAC in pulse-invaded as well as replicating Toxoplasma.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The TgCRT-deficient parasites displayed a swollen vacuole and a disrupted endolysosomal system.

(A) Extracellular Δcrt parasites showed an enlarged concave subcellular structure, indicated by the arrow, under the differential interference contrast (DIC) microscopy. Scale bar = 5 μm. (B) The swollen subcellular structure indicated by the arrow was also observed in pulse-invaded Δcrt parasites and colocalized with a major luminal peptidase of the VAC, cathepsin L-like protease (TgCPL). The TgCPL staining was used to assess the morphology of the VAC. The mean VAC size ± standard error of mean (SEM) was calculated for three independent measurements and is listed in the figure. Statistical significance was determined using unpaired two-tailed Student’s t-test. Scale bar = 2 μm. (C) In pulse-invaded parasites, the TgCPL (VAC marker) and proTgM2AP or TgVP1 (ELC markers) staining were juxtaposed in the WT and ΔcrtCRT strains, but co-localized in the Δcrt strain. During replication, the VAC in WT and ΔcrtCRT parasites became fragmented. However, the aberrant overlap between the VAC and ELC significantly decreased the extent of VAC fragmentation in the Δcrt mutant. TgCPL staining presented as one or more larger and intensely stained puncta, which co-localized with proTgM2AP and TgVP1 (indicated by arrows in the insets). The scale bars in the images of pulse invaded and replicated parasites are 2 μm and 5 μm, respectively. The colocalization of the TgCPL with proTgM2AP or TgVP1 was analyzed by a Coloc2 colocalization analysis software. The Pearson’s correlation coefficient (PCC) of both VAC and ELC staining was measured within 10 individual parasites for each replicate in a total of three replicates and presented as mean ± standard deviation (SD). Statistical significance was determined using unpaired two-tailed Student’s t-test. ****, p<0.0001; n.s., not significant.

https://doi.org/10.1371/journal.ppat.1007775.g001

The VAC is a lysosome-like organelle, participating in the parasite’s endolysosomal system. It provides an environment for maturation of TgCPL and delivers activated TgCPL to its adjacent endosome-like compartment (ELC) to assist in processing microneme proteins required for parasite invasion [4]. It also serves as a digestive organelle to degrade endocytosed proteins [4,11,26]. We hypothesized that the dramatic swelling of the VAC might affect the integrity of the parasite’s endolysosomal system. We stained WT, Δcrt, and ΔcrtCRT parasites with antibodies recognizing markers of the VAC (anti-TgCPL) and of the ELC (anti-proTgM2AP or TgVP1) [4,10], and captured a series of deconvolved Z-stack images for individual co-staining. In pulse-invaded parasites, the VAC and ELC displayed distinct subcellular staining in WT and ΔcrtCRT strains, whereas in Δcrt parasites both markers partially co-localized (Fig 1C). Similarly, the non-fragmented TgCPL puncta in replicating Δcrt parasites also showed partial colocalization with both proTgM2AP and TgVP1 (Fig 1C). We quantified the colocalization of TgCPL with proTgM2AP and TgCPL with TgVP1 by measuring their Pearson’s correlation coefficient (PCC). Our analysis revealed that TgCPL showed significantly higher colocalization with proTgM2AP or TgVP1 in Δcrt parasites than WT and ΔcrtCRT strains in both stages of pulse invasion and replication (Fig 1C). A sodium/proton exchanger, named TgNHE3, was previously identified in the ELC [27]. We also co-stained WT, Δcrt, and ΔcrtCRT strains with anti-TgCPL and anti-TgNHE3 antibodies, and did not observe partial colocalization in Δcrt mutant (S2 Fig). These results mirror a previous observation that the ingested host proteins in the TgCPL-deficient parasites overlapped with proTgM2AP to a greater extent than TgNHE3 during their trafficking through the parasite’s endolysosomal system [26], suggesting that the TgNHE3 occupies a distinct subdomain of the ELC. Overall, our findings suggest that altered VAC morphology due to the absence of TgCRT affects the integrity of the parasite’s endolysosomal system.

RHΔku80Δcrt shows reduced invasion and acute virulence

Toxoplasma utilizes exocytosis and endocytosis via endolysosomes to release invasion effectors, and to ingest host proteins during intracellular growth, respectively [11,28–30]. We therefore characterized invasion, replication, and egress for the Δcrt strain. First, we measured the invasion efficiency of parasites at 30–120 min post-infection. At 30 min post-infection, the Δcrt mutant showed ~50% reduction in invasion compared to WT and ΔcrtCRT (Fig 2A). Differences in invasion efficiency between WT and Δcrt were reduced by ~20% at 60 min post-infection, and were not seen at 120 min post-infection (S3A Fig), suggesting that Δcrt parasites have slower invasion kinetics relative to the WT strain. To further understand the basis of this invasion deficiency, we compared parasite attachment to host cells using previously published methods [3,31]. We found that the Δcrt mutant showed ~50% reduction in host cell attachment compared to WT and ΔcrtCRT strains (Fig 2B). In contrast, we observed no significant differences in both gliding distance and types (S3B Fig). Second, we used immunofluorescence microscopy to quantify parasite replication. Infected cells were stained with DAPI and anti-TgGRA7 antibodies to define individual parasite nuclei and parasitophorous vacuolar (PV) membranes, respectively. The average number of parasites per PV was calculated for each strain to compare replication rates. There were no statistical differences in parasite replication between WT and Δcrt parasites at 28 and 40 h post-infection (Fig 2C). We also introduced NanoLuc luciferase into WT, Δcrt, and ΔcrtCRT parasites, and measured the fold-change of luciferase activity for 72 h post-infection to calculate relative growth rates. Similarly, we did not observe growth differences between WT and Δcrt at 24, 48, and 72 h post-infection (Fig 2C). Third, the egress efficiency of each strain was determined by a lactate dehydrogenase release-based assay. The parasites were incubated with 0.5 mM Zaprinast for 5 min to induce egress. The egressed parasites disrupt host cell membranes to release lactate dehydrogenase, which is subsequently quantified to extrapolate to the number of egressed PVs. We did not observe egress defects in the TgCRT-deficient parasites (Fig 2D). Last, we determined the acute virulence of Δcrt parasites in a murine model. Outbred CD-1 mice were infected with a subcutaneous or intravenous inoculum of 100 WT, Δcrt, or ΔcrtCRT parasites. Thirty percent of mice infected with the Δcrt mutant survived when mice were infected subcutaneously, whereas WT and ΔcrtCRT infections led to quantitative mortality at 10–12 days post-infection (Fig 2E). Mice receiving WT parasites by intravenous inoculation showed mortality starting at 13 days post-infection and all expired at 20 days. The Δcrt and ΔcrtCRT parasites caused death in 40% and 80% of infected mice, respectively (Fig 2E). Statistical analysis showed that mice infected with WT and Δcrt parasites have significant difference in their survival time. Seroconversion of the surviving mice was confirmed by ELISA. We also challenged surviving Δcrt mice with 1000 WT parasites by subcutaneous injection and did not observe lethality after 30 days post-challenge. These findings indicate that the pre-inoculation of Δcrt parasites conferred immunological protection against subsequent acute toxoplasmosis. Given the hyper-virulent nature of Type I Toxoplasma strain in a murine model, the Δcrt mutant dramatically lost its acute virulence compared to WT parasites. Collectively, our findings revealed that Toxoplasma parasites require the TgCRT protein for optimal invasion and acute virulence but not for replication and egress.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Parasite invasion and acute virulence were reduced in the Δcrt parasites.

(A) At 30 min post-infection, the Δcrt parasites showed 0.90 ± 0.31 (mean ± SD) parasites per host cell, which is approximately a ~50% reduction in invasion compared to the WT (1.88 ± 0.10) and ΔcrtCRT (2.10 ± 0.40) strains. Six fields of view were counted per strain per replicate. The assay was performed in triplicate. (B) The Δcrt mutant exhibited ~50% reduction in attachment compared to WT and ΔcrtCRT strains. Eight view fields were captured and quantified per strain per replicate. The assay was performed in triplicate. (C) The Δcrt displayed comparable replication rates as WT and ΔcrtCRT parasites. Both fluorescence- and luminescence-based replication assays showed similar results. The average number of parasites per parasitophorous vacuole derived from the fluorescence-based replication assay was also calculated and listed above the plots as mean ± SD. (D) A lactate dehydrogenase release assay was used to measure the egress efficiency of parasites. No differences were observed in parasite egress between WT and Δcrt parasites. Statistical significance in the assays listed in panel A through D was determined by using unpaired two-tailed Student’s t-test. (E) The acute virulence of TgCRT-deficient parasites was evaluated in a murine model via subcutaneous and intravenous infections. One hundred parasites from each strain were used to infect outbred CD-1 mice (n = 10 mice for each strain). The mortality of the mice was monitored for 30 days. Seroconversion of the surviving mice was evaluated by ELISA to confirm successful infection. Additionally, the surviving mice were allowed to rest for 10 days before subsequent challenge with 1,000 WT parasites by subcutaneous inoculation. The Δcrt mutant exhibited reduced acute virulence compared to the WT and ΔcrtCRT strains and conferred immunological protection in the surviving mice. Data were recorded and are presented using the Kaplan-Meier plot. Statistical analysis was performed using the Log-rank (Mantel-Cox) test. For all statistical significance calculation, **, p<0.01; ****, p<0.0001; n.s, not significant.

https://doi.org/10.1371/journal.ppat.1007775.g002

Δcrt parasites show impaired microneme secretion

During infection, Toxoplasma parasites sequentially secrete proteins to facilitate host invasion. Microneme proteins are the first to be secreted. These proteins traffic through the parasite’s endolysosomal system and undergo intracellular maturation before being trimmed and released from the parasite surface by intramembrane cleavage [3–8,32]. To test which step(s) is affected in the Δcrt parasites, we probed cell lysates and excretory secretory antigen fractions (ESAs) of each strain with anti-TgMIC2, anti-TgM2AP, and anti-TgMIC5 by immunoblotting to measure abundances and secretion patterns. The migration patterns of these three microneme proteins in cell lysates were similar among the strains (Fig 3A). The abundances of the individual microneme proteins were normalized against the protein level of the Toxoplasma actin protein by densitometry and plotted for quantification. All three strains showed comparable steady-state abundances of these proteins (Fig 3A). To further evaluate abundances of secreted microneme proteins, we probed constitutive and induced ESAs with the same antibodies. The constitutive and induced ESAs were generated by incubating purified parasites in D10 medium (DMEM medium supplemented with 10% (v/v) cosmic calf serum) for 30 min at 37°C or D10 medium supplemented with 1% (v/v) ethanol for 2 min at 37°C, respectively. In the ESAs secreted by WT parasites, TgMIC2 exhibited two bands migrating at 100 kDa and 95 kDa, while TgM2AP showed 4 proteolytically processed polypeptides along with pro- and mature forms. However, TgMIC2 only existed as a 100 kDa band in Δcrt parasites. Furthermore, mature TgM2AP was not processed in the constitutive ESAs of the Δcrt strain and showed significantly reduced processing in the induced ESAs of Δcrt parasites (Fig 3B). The secreted TgMIC5 protein displayed similar migration patterns among these strains (Fig 3B). Secretion of these microneme proteins was also quantified by normalizing the relative abundances of the proteins against the protein level of secreted TgGRA7, a dense granule protein. The secretion of TgMIC2, TgM2AP, and TgMIC5 were reduced by ~80%, 50%, and 40%, respectively, in the induced ESAs of Δcrt parasites, compared to the WT strain. The differences in the amount of microneme secretion were less significant in the constitutive ESAs. Secretion of TgMIC2 and TgM2AP was decreased by ~40% and 25%, respectively, in Δcrt parasites compared to the WT strain, whereas TgMIC5 did not show a difference (Fig 3B). Lower ESA secretion induced by ethanol in the Δcrt mutant suggests that Δcrt parasites are deficient in rapid release of certain proteins from the micronemes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. The deletion of TgCRT altered microneme secretion, without affecting the microneme steady abundance, intracellular trafficking, or intramembrane cleavage on the parasite surface.

(A) The steady-state level of microneme proteins was not altered in the Δcrt parasites. Freshly lysed parasites were filter-purified, lysed, and subjected to SDS-PAGE electrophoresis and immunoblotting. The blots were probed with anti-TgMIC2, TgM2AP, and TgMIC5 antibodies, along with anti-TgActin as a loading control. Individual microneme proteins were normalized against the corresponding TgActin to quantify their steady-state expression. Four independent preparations and probing of parasite lysates were conducted in this assay. (B) Δcrt parasites secreted fewer microneme proteins than WT and ΔcrtCRT parasites, and altered the microneme secretion patterns. Freshly filter-purified parasites were incubated in medium at 37°C for 30 min to generate the constitutive ESAs, or were treated with 1% (v/v) ethanol in medium to produce induced ESAs. The ESA fractions were separated and probed with anti-TgMIC2, TgM2AP, and TgMIC5 antibodies for quantification of the secreted forms of these microneme proteins. The ESA fractions were also probed with anti-TgGRA7 antibody, a dense granule protein, as a loading control. Five independent preparations and probing of both constitutive and induced ESAs were performed in this assay. (C) Pulse-invaded and replicated parasites were stained with anti-TgMIC2 and anti-TgM2AP antibodies to examine their intracellular trafficking. No defects were detected in their intracellular trafficking. (D) Pulse-invaded and replicated parasites were also stained with anti-TgROP7 to examine the morphology of the rhoptry and intracellular trafficking of TgROP7. The rhoptries maintained similar morphology and trafficking patterns among WT, Δcrt, and ΔcrtCRT strains. The scale bars in the images of pulse invaded and extracellular parasites are 2 μm, and the scale bars in the images of replicated parasites are 5 μm. The statistical significance among the steady expression and secretion of microneme proteins in WT, Δcrt, and ΔcrtCRT strains was determined using unpaired two-tailed Student’s t-test. **, p<0.01; ****, p<0.0001; n.s., not significant.

https://doi.org/10.1371/journal.ppat.1007775.g003

To examine whether the abnormal secretion of microneme proteins alters their intracellular trafficking patterns, we stained pulse-invaded and replicated parasites with TgMIC2 and TgM2AP antibodies. Both microneme proteins trafficked to the apical end of the parasites and showed normal staining patterns (Fig 3C). Prior to secretion, some transmembrane microneme proteins are released via proteolytic cleavage by the intramembrane rhomboid protease TgROM4. Deletion of TgROM4 leads to retention of some microneme proteins on the parasite’s plasma membrane, such as TgMIC2 and TgAMA1 (Toxoplasma apical membrane antigen 1) [6–8,32]. To test whether the aberrant endolysosomal system alters the retention of microneme proteins on the surface of parasites, we stained purified, non-permeabilized extracellular parasites with anti-TgMIC2 antibodies. Immunofluorescence microscopy did not reveal excess TgMIC2 on the plasma membrane of Δcrt parasites (S4 Fig), suggesting that the reduced secretion of microneme proteins is not due to their inefficient intramembrane cleavage on the parasite’s plasma membrane.

The endosome-like compartment is involved not only in the trafficking of microneme proteins, but also rhoptry contents [33]. We stained pulse-invaded and replicated parasites with anti-TgROP7 antibodies to examine the trafficking of rhoptry proteins and the morphology of rhoptries. TgROP7 staining revealed typical rhoptry patterns located at the apical end of the parasites (Fig 3D), excluding the possibility of aberrant trafficking of rhoptry contents and possible defects in rhoptry biogenesis. Taken together, our data suggest that the invasion defects for Δcrt parasites are caused by incomplete trimming and consequent inefficient secretion of microneme proteins at the parasite’s plasma membrane, but not by altered intracellular maturation, trafficking, or intramembrane cleavage of microneme proteins, nor by altered rhoptry morphology.

TgSUB1 transcript and protein levels are decreased for Δcrt parasites

The inefficient proteolytic processing of TgMIC2 and TgM2AP in Δcrt ESAs led us to investigate whether these phenotypic observations were caused by the abnormal expression patterns or subcellular trafficking of Toxoplasma subtilisin 1 (TgSUB1) in Δcrt parasites. A previous publication reported that parasites lacking TgSUB1 showed defective trimming of secreted microneme proteins, such as TgMIC2 and TgM2AP [3], which we noted resemble the altered patterns of TgMIC2 and TgM2AP products observed for the Δcrt mutant. Therefore, we quantified secreted TgSUB1 in both constitutive and induced ESAs by probing them with an anti-SUB1 antibody, previously found to specifically react against TgSUB1 and PfSUB1 [34]. Immunoblotting analysis revealed that there was no detectable TgSUB1 in the ESAs of Δcrt parasites (Fig 4A). TgSUB1 stays on the plasma membrane via its GPI-anchor prior to its release by self-shedding [12,35]. Given that there was no detectable TgSUB1 in the ESAs of the Δcrt strain, we measured the abundance of TgSUB1 on the parasite surface by immunofluorescence microscopy to assess if the Δcrt retained TgSUB1 on the plasma membrane. We stained non-permeabilized extracellular parasites with anti-SUB1 to evaluate the amount of surface-anchored TgSUB1. Similarly, no detectable TgSUB1 was observed on the plasma membrane of Δcrt parasites (Fig 4B). These data suggest that there is lower expression of TgSUB1 on the surface of Δcrt parasites.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Transcript and protein levels of a subtilisin-related protease, TgSUB1, were reduced in Δcrt parasites.

(A) The abundance of secreted TgSUB1 in the constitutive and induced ESAs was measured by immunoblotting. There was no detectable TgSUB1 in the constitutive and induced ESAs secreted by Δcrt parasites via immunoblotting. (B) The abundance of plasma membrane-anchored TgSUB1 was measured by probing non-permeabilized parasites with antibodies recognizing TgSUB1. The Δcrt mutant significantly reduced TgSUB1 on its plasma membrane. (C) The intracellular TgSUB1 expression level was significantly reduced in Δcrt as shown by immunofluorescence microscopy. TgSUB1 was identified as a microneme protein. TgSUB1 and TgMIC5 were both found to localize in the micronemes of pulse-invaded and replicated WT and ΔcrtCRT parasites. However, TgSUB1 signal was not detected in the Δcrt parasites. (D) The cell lysates of the Δcrt parasites showed that the steady level of TgSUB1 was reduced by ~85% in the Δcrt strain. (E) The transcript level of TgSUB1 was quantified by quantitative PCR. It was also decreased by approximately 90% in the Δcrt mutant. All assays listed in this figure were repeated in triplicate. Statistical significance was calculated using unpaired two-tailed Student’s t-test: *, p<0.05; ***, p<0.001; n.s., not significant.

https://doi.org/10.1371/journal.ppat.1007775.g004

To further dissect the basis for reduced TgSUB1 on the surface, we tested two possibilities: 1) TgSUB1 traffics aberrantly within the parasite to prevent its delivery to parasite surface, and 2) the expression level of TgSUB1 is reduced. TgSUB1 is a microneme protein that also traffics through the parasite’s endolysosomal system [3,12]. The aberrant endolysosomal system in Δcrt parasites potentially alters intracellular trafficking and/or maturation of TgSUB1 that then reduces its expression. To test these two possibilities, first, we stained pulse-invaded and replicated parasites with anti-SUB1 to examine TgSUB1 intracellular trafficking patterns. TgMIC5 localization was used as a reference for typical expected microneme staining. Surprisingly, we observed much less TgSUB1 staining in Δcrt parasites compared to the WT strain (Fig 4C). Next, we quantified abundance of TgSUB1 in parasite cell lysates and found that TgSUB1 was decreased by approximately 85% in Δcrt parasites compared to WT parasites (Fig 4D). To further understand how TgSUB1 expression is suppressed in the Δcrt mutant, we performed qPCR to measure TgSUB1 mRNA for WT, Δcrt, and ΔcrtCRT parasites. TgSUB1 transcript was reduced ~10-fold upon deletion of TgCRT (Fig 4E). Collectively, our findings suggest that the arrested overlap of the VAC and ELC dramatically decreases the abundance of TgSUB1 protein, which then alters the proteolytic processing of normally secreted micronemal protein invasion effectors, thereby reducing invasion efficiency.

VAC alterations reduce endolysosomal protease proteins and transcripts

The swollen VAC and its aberrant overlap with the ELC in the Δcrt parasites could conceivably lead to altered gene transcription to assist in the adaptation of these parasites. We conducted transcriptome sequencing to detect global alterations in gene transcription for Δcrt parasites relative to WT. Differential gene expression analysis identified 102 genes whose transcript levels changed greater than 1.5-fold in the Δcrt strain. Forty-six and fifty-six genes had increased and reduced transcripts, respectively (Fig 5A and S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. The transcript and protein abundances of several VAC-residing proteases were decreased in Δcrt.

(A) RNA-Seq was performed in WT and Δcrt parasites. Each sample was sequenced in duplicate for statistical comparison. A volcano plot was used to summarize the genes with altered transcription greater than 1.5-fold and with statistical significance less than 0.05 in the Δcrt mutant relative to the WT strain. Forty-six and fifty-six genes labeled in the blue and red dots became up- and down-regulated in the Δcrt mutant, respectively. The blue and red dash lines represent the borderline of 1.5-fold change in gene transcripts, and the genes above the black dash line had p values of statistical significance below 0.05. (B) qPCR was used to validate 4 down-regulated proteases that were identified by RNA-Seq analysis, along with two known VAC proteases, TgCPL and TgCPB. TgAMN, TgSCP, TgASP1, and TgCPB displayed down-regulated transcription in Δcrt parasites compared to WT. The qPCR assay was repeated in three technical replicates per biological replicate in a total of three biological replicates. Data shown in the figure are represented as mean ± SEM. The fold change of individual genes was calculated using a ΔΔCT (cycle threshold) method for relative quantification by normalization against WT values using TgActin as a normalization control. The variation shown in WT parasites was determined by normalization of individual CT values per replicate against the input of total RNA. (C) The steady protein abundances of TgCPL and TgCPB were quantified in the lysates of parasites by immunoblotting. The protein levels of TgCPL and TgCPB in the Δcrt mutant were reduced by ~40% and 60%, respectively, compared to WT parasites. Immunoblotting assays were repeated in 3–4 replicates for statistical comparison. (D) TgSCP and TgAMN were endogenously tagged with 3xmyc and 3xHA, respectively, at their C-termini. The expression of the epitope-tagged proteins was confirmed by immunoblotting. The parasites were co-stained with antibodies recognizing their respective epitope tags as well as the VAC and ELC markers. Both TgSCP and TgAMN were localized in the VAC and ELC by immunofluorescence. Statistical significance was performed using unpaired two-tailed Student’s t-test. *, p<0.05; ***, p<0.001; n.s., not significant.

https://doi.org/10.1371/journal.ppat.1007775.g005

Four proteases were among the list of genes showing reduced transcripts in the Δcrt mutant, including a putative aminopeptidase N protein (TgAMN, TGGT1_221310), a putative Pro-Xaa serine carboxypeptidase (TgSCP, TGGT1_254010), aspartyl protease 1 (TgASP1, TGGT1_201840), and an ICE family protease-like protein (TgICEL, TGGT1_243298). We validated transcript levels for these proteases, as well as two known VAC luminal proteases (TgCPL and TgCPB), in WT, Δcrt, and ΔcrtCRT strains by qPCR. The qPCR analysis showed that the transcript levels of TgAMN, TgSCP, TgASP1, and TgCPB were decreased by 50%, 20%, 47%, and 14%, respectively, in Δcrt parasites (Fig 5B). Protein levels of TgCPL and TgCPB were quantified by immunoblotting and compared for WT, Δcrt, and ΔcrtCRT parasites. Although TgCPL transcript levels did not differ, the abundance of TgCPL protein was decreased ~40% in the Δcrt mutant (Fig 5C). TgCPB transcript levels were reduced in Δcrt, and both the pro- and mature forms of TgCPB protein were decreased relative to WT parasites (Fig 5C). Densitometry analysis showed that the expression level of mature TgCPB was reduced by ~60% in the Δcrt parasites (Fig 5C).

To determine the subcellular locations of the down-regulated proteases, we tagged endogenous TgAMN and TgSCP with 3xHA and 3xmyc epitope tags at their C-termini in WT parasites, respectively (S5 Fig and Table 1). After drug-selection, we probed cell lysates with anti-HA and anti-myc antibodies, respectively, to test expression. Immunoblotting revealed that the observed molecular masses of both proteins were similar to the predicted sizes based on the primary sequences (Fig 5D). Next, the tagged strains were co-stained with antibodies recognizing the epitope tags along with anti-TgCPL or anti-VP1 antibodies to determine their subcellular locations. Immunofluorescence microscopy revealed that both TgSCP and TgAMN were localized to the VAC/ELC (Fig 5D). TgASP1 subcellular location was also determined to be within the VAC (data were deposited in a public repository, www.toxodb.org; Dou, Z. et al., in preparation). Collectively, these data suggest that the swollen VAC in Δcrt parasites causes reduced transcription and translation of several endolysosomal proteases.

Suppression of proteolysis within the swollen Δcrt VAC partially restores VAC size, organellar separation, and invasion

Given that CRT is a putative small solute transporter, the deletion of TgCRT potentially results in the accumulation of small nutrient molecules generated by proteolysis within the VAC, further swelling its size. Therefore, we speculated that inhibition of proteolytic activity within the swollen VAC might reduce its size. We tested this hypothesis by chemically or genetically suppressing VAC proteolysis. First, we treated WT, Δcrt, and ΔcrtCRT parasites with 1 μM LHVS, an irreversible inhibitor of TgCPL protease [25]. TgCPL is a major endopeptidase involved in the maturation of microneme proteins and digestion of host proteins [4,11]. Infected host cells were incubated with LHVS for 48 h to allow full inhibition of TgCPL. Treated parasites were liberated from host cells and used to infect new host cells for 30 min, followed by TgCPL staining to quantify the size of the VAC. As expected, LHVS-treated Δcrt parasites displayed smaller VACs than DMSO-treated Δcrt parasites (Fig 6A). To validate these findings, we genetically ablated TgCPL in Δcrt parasites to create the ΔcrtΔcpl double knockout by CRISPR-Cas9-based genome editing. The replacement of TgCPL with a pyrimethamine resistance cassette was confirmed by PCR and immunoblotting (S6A Fig and Table 1). The ΔcrtΔcpl mutant showed a significantly smaller concave structure than Δcrt parasites under DIC microscopy (S7 Fig). We also compared the size of the VAC in WT, Δcrt, and ΔcrtΔcpl based on TgCPB staining by immunofluorescence microscopy using similar methods previously described, and found that similarly, the ΔcrtΔcpl partially reversed its VAC size compared to Δcrt parasites, but still showed a larger VAC than the WT strain (Fig 6B). Moreover, we deleted TgCPB in Δcrt to test whether such phenotype of partial VAC size restoration was independent of the deletion of specific VAC/ELC-localizing proteases. TgCPB was previously identified as another known VAC-localizing protease, displaying both endo- and exo-peptidase activities [14,25]. Due to its carboxypeptidase activity, it is expected that TgCPB generates more small solutes relative to TgCPL. We used CRISPR-Cas9 genome editing to generate a ΔcrtΔcpb double knockout (S6B Fig and Table 1). The successful gene ablation was confirmed by PCR and immunoblotting (S6B Fig). The resulting ΔcrtΔcpb mutant also showed a smaller concave subcellular structure compared to the Δcrt mutant (S7 Fig). The size of the VAC in WT, Δcrt, and ΔcrtΔcpb was quantified based on the TgCPL staining as described above and the ΔcrtΔcpb parasite VAC was reduced by ~35% compared to Δcrt parasites (Fig 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Reduction of proteolytic activity within the swollen Δcrt VAC reduced its size and partially restored integrity of the parasite’s endolysosomal system and parasite invasion.

(A) The Δcrt parasites were incubated with 1 μM LHVS, an irreversible inhibitor of TgCPL, for one lytic cycle, followed by a pulse invasion. Parasites were stained with anti-TgCPL antibodies to determine the size of the VAC. The swollen VAC phenotype was significantly reduced in the LHVS-treated parasites. Scale bar = 2μm. (B) TgCPL and TgCPB, two VAC-residing peptidases, were genetically deleted in the Δcrt strain individually to create ΔcrtΔcpl and ΔcrtΔcpb double knockouts, respectively. Both ΔcrtΔcpl and ΔcrtΔcpb parasites displayed a partial reduction in the size of the VAC compared to WT parasites. Additionally, we genetically deleted TgSUB1 within the Δcrt mutant to create the ΔcrtΔsub1 double knockout. Similarly, the VAC size was partially recovered in the ΔcrtΔsub1 mutant. The VAC sizes are listed in the figure as mean ± SEM. The VAC size measurement was repeated in triplicate. Scale bar = 2 μm. (C) The pulse-invaded parasites showing colocalization between TgCPL (VAC marker) and proTgM2AP (ELC marker) in WT, Δcrt, ΔcrtCRT, and ΔcrtΔcpb strains were quantified. At least 100 parasites were quantified for each replicate in a total of three replicates. The ΔcrtΔcpb parasites had a significantly lower percentage of parasites having arrested colocalization between the TgCPL and proTgM2AP compared to the Δcrt mutant. (D) The lysates of WT, Δcrt, and ΔcrtΔcpl or ΔcrtΔcpb parasites were probed with antibodies recognizing TgSUB1. The steady expression of TgSUB1 was restored in both ΔcrtΔcpl and ΔcrtΔcpb parasites. (E) The constitutive and induced ESAs of WT, Δcrt, and ΔcrtΔcpb strains were collected and probed with the antibodies indicated in the figure. The secretion and processing of microneme proteins were also recovered in the ΔcrtΔcpb mutant. (F) The invasion efficiency of WT, Δcrt, and ΔcrtΔcpb strains was determined using the procedures mentioned above. The ΔcrtΔcpb parasites showed increased invasion efficiency (1.63 ± 0.10 parasites per host cell) compared to the Δcrt strain (1.05 ± 0.12 parasites per host cell), albeit still a lower efficiency than the WT parasites (2.00 ± 0.12 parasites per host cell). Statistical significance in all assays listed in this figure was calculated using unpaired two-tailed Student’s t-test. *, p<0.05; **, p<0.01; ***, p<0.001; n.s., not significant.

https://doi.org/10.1371/journal.ppat.1007775.g006

To test if, like TgCPL and TgCPB, residual expression of TgSUB1 in Δcrt parasites contributed to VAC swelling, we deleted TgSUB1 in Δcrt to create and validate a ΔcrtΔsub1 double knockout mutant (S6C Fig and Table 1). We found that targeted deletion of TgSUB1 in Δcrt partially reversed the swollen VAC phenotype (Fig 6B and S7 Fig). Altogether, our findings from the chemical or genetic inhibition of different proteases strongly suggest that the proteolysis within the VAC plays a key role in its swelling in parasites lacking TgCRT.

Next, we tested whether the partial restoration of VAC size was also associated with the reversal of other phenotypes observed in the Δcrt mutant. Here, we chose the ΔcrtΔcpb strain as a representative strain for further testing. ΔcrtΔcpb showed fewer parasites with partial overlap between the markers for the VAC (TgCPL) and ELC (proTgM2AP) compared to Δcrt. Approximately 44% of ΔcrtΔcpb parasites showed partial overlap between TgCPL and proTgM2AP staining compared to 62% in the Δcrt strain (Fig 6C), with both significantly higher than the 19% and 25% seen for WT and ΔcrtCRT strains, respectively. We also tested whether the partial resegregation of the VAC and ELC in both ΔcrtΔcpb and ΔcrtΔcpl mutants was associated with the recovery of TgSUB1 expression. By immunoblotting, TgSUB1 showed comparable expression in both the WT and ΔcrtΔcpb strains (Fig 6D). Similarly, such recovery of TgSUB1 expression was also seen in the ΔcrtΔcpl strain (Fig 6D). Moreover, TgSUB1 secretion was observed in both constitutive and induced ESA fractions in the ΔcrtΔcpb parasites (Fig 6E). TgM2AP and TgMIC2 were correctly cleaved by TgSUB1 in ΔcrtΔcpb, and their secretion patterns were similar to those seen in the WT strain (Fig 6E). This partial restoration of phenotypes in the ΔcrtΔcpb mutant resulted in an increase of invasion efficiency up to ~60% compared to the Δcrt strain, although invasion was still significantly lower than that of WT parasites (Fig 6F). Collectively, these data show a close association between the size of the VAC, altered morphology of the parasite’s endolysosomal system, protein abundance of TgSUB1, microneme protein secretion and processing, and parasite invasion.

TgCRT is a functional transporter

Finally, we attempted to express TgCRT in S. cerevisiae yeast following previously described strategies for PfCRT [36,37]. Native TgCRT cDNA was not expressed well in S. cerevisiae yeast (S8 Fig). Via alignment with PfCRT (S9A Fig), removing the 300 N-terminal residues of the TgCRT gene that are non-homologous to PfCRT preserves all putative transmembraneous domains and inter-helical loop regions. Thus, following a previously published strategy for difficult to express PfCRT mutants [22] we created a fusion gene that replaced the 300 most N-terminal residues of the TgCRT sequence with the 111 most N-terminal residues from the S. cerevisiae plasma membrane ATPase (PMA), which harbors a yeast plasma membrane localization sequence (S9B and S9C Fig). The fusion protein was well expressed in the plasma membrane of S. cerevisiae (S8 Fig) with vacuolar-to-cytosol topology preserved, similar to the expression of PMA-PfCRT in yeast as described in a previous publication [38]. Following an approach previously described for PfCRT and PfCRT mutants [22,39], we assayed PMA-TgCRT expressing yeast for chloroquine (CQ) transport (Fig 7A). Via alignment with PfCRT (S9A Fig), TgCRT T369 corresponds to the well-studied K76 residue within PfCRT; previously, mutation of PfCRT K76 to T has been shown to increase the efficiency of CQ transport by PfCRT, although it is not the sole determining factor [22,40,41]. We expressed both WT TgCRT and T369K TgCRT in yeast to measure their transport efficiencies. Both the wild type TgCRT protein and a TgCRT T369K mutant were found to transport CQ, albeit slower than PfCRT^HB3 and PfCRT^Dd2 derived from CQ-sensitive and resistant malaria strains, respectively, under similar conditions (Fig 7A). Also, we noted that TgCRT required higher external [CQ] (80 mM versus 16 mM for PfCRT) to achieve similar levels of transport, suggesting increased CQ K_m for this transporter (Fig 7A). Analysis of PfCRT isoforms and the role that individual amino acid substitutions play in modifying transport activity has been ongoing for well over a decade (e.g. Callaghan et al. 2015 [22]). Since PfCRT—TgCRT identity is modest (S9A Fig) many more TgCRT mutagenesis studies will be required to fully define catalytically critical sites in TgCRT. Regardless, these initial data show that mutation of a TgCRT threonine, analogous to the well-known position 76 T for PfCRT [22,40,41], also affects CQ transport as is the case for PfCRT.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. TgCRT is a functional transporter and its transport efficiency is correlated with VAC size in the parasites.

(A) CQ transport by PfCRTs and TgCRT (PMA-TgCRT) expressed in S. cerevisiae. Transport was extrapolated from CQ-induced growth delays as described in [37,39]. Results are the average of at least four independent experiments ± SEM. EV, empty vector; PfCRT^Dd2, a PfCRT expression construct derived from CQ-resistant malaria strain; PfCRT^HB3, a PfCRT expression construct derived from CQ-sensitive malaria strain; PMA-TgCRT, plasma membrane ATPase-TgCRT fusion (see S9 Fig); PMA-TgCRT^T369K, TgCRT fusion protein harboring T to K substitution at the position analogous to residue 76 in PfCRT (see text). (B) A mutation of T369K was introduced into WT TgCRT complementation construct by site-directed mutagenesis before it was electroporated into the Δcrt mutant. The VAC sizes were determined based on TgCPL staining using the methods mentioned above. The Δcrt mutant complemented with TgCRT^T369K partially restored its VAC size, but it was still significantly bigger than that transfected with WT TgCRT. Statistical significance was calculated using unpaired two-tailed Student’s t-test. *, p<0.05; **, p<0.01.

https://doi.org/10.1371/journal.ppat.1007775.g007

We also exchanged threonine for lysine in the WT TgCRT complementation construct and transfected Δcrt to examine the extent to which TgCRT^T369K affects VAC size in Toxoplasma parasites. Interestingly, in contrast to full recovery of VAC size in the WT TgCRT complementation strain, TgCRT^T369K only partially restored the swollen VAC (Fig 7B). These findings, along with the TgCRT transport data, strongly suggest that the swollen VAC is caused by luminal osmolyte excess, similar to findings for PfCRT as described in "Discussion".

In summary, our findings strongly suggest a role for TgCRT in small solute transport that impacts VAC volume in the absence of drug transport, similar to the role proposed for PfCRT [18,19]. The identity of the relevant osmoregulatory transport substrate(s) remains to be determined. However, at least for T. gondii, TgCRT expression is also required for proper segregation of other organelles within the endolysosomal system that, in turn, indirectly facilitates microneme secretion and parasite invasion. The data also indicate that the invasion deficiency exhibited by the Δcrt mutant is likely due to multiple factors, since the recovery of TgSUB1 expression and trimming of microneme proteins in ΔcrtΔcpb did not completely reverse the invasion defects. To the best of our knowledge, this is the first observation of the regulation of the endolysosomal protease expression in apicomplexan parasite by a CRT protein.

Ethical statement

This study was performed in compliance with the Public Health Service Policy on Humane Care and Use of Laboratory Animals and Association for the Assessment and Accreditation of Laboratory Animal Care guidelines. The animal protocol was approved by Clemson University’s Institutional Animal Care and Use Committee (Animal Welfare Assurance A3737-01, protocol number AUP2016-012). All efforts were made to minimize discomfort. CO₂ overdose was used to euthanize mouse. This method of euthanasia is consistent with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association.

Chemicals and reagents

Morpholine urea-leucyl-homophenyl-vinyl sulfone phenyl (LHVS) was kindly provided by the Bogyo lab at Stanford University. Fluo-4-AM was purchased from Invitrogen (Life Technologies). Ionomycin was purchased from Sigma Aldrich. All oligonucleotide primers used in this study were listed in S2 Table. Other chemicals used in this work were analytical grade and were purchased from VWR unless otherwise indicated.

Parasite culture

Toxoplasma gondii parasites were cultured in human foreskin fibroblast (HFF) cells (ATCC, SCRC-1041) or hTERT cells in Dulbecco’s Modified Eagle Medium (DMEM) media supplemented with 10% cosmic calf serum at 37°C with 5% CO₂. The parasites were harvested by membrane filtration as described previously [14].

Generation of transgenic parasites

To generate the TgCRT-deficient strain, ~1.5 kilobases (kb) of the 5’- and 3’-UTR of the TgCRT gene, respectively, were PCR-amplified and flanked at both ends of the bleomycin resistance cassette (BLE) to assemble a TgCRT deletion construct. The resulting plasmids were introduced into WT parasites by electroporation. The transfected parasites were selected with 50 μg/ml bleomycin twice, while in their extracellular stage as described previously [14]. Clones of the TgCRT-deficient parasites were isolated by limiting dilution. The correct replacement of TgCRT with the BLE cassette was confirmed by PCR (see S1 Text).

To complement Δcrt parasites, we modified the plasmid pTub-TgCRT-mCherry-3xmyc (a gift from the van Dooren lab), which expresses a C-terminally mCherry-3xmyc epitope-tagged TgCRT under the Toxoplasma tubulin promoter. The plasmid was digested with HpaI and MfeI restriction endonucleases to remove the tubulin promoter and a segment of TgCRT. The remaining DNA fragment served as the backbone for subsequent Gibson assembly to incorporate a PCR amplified ~1 kb region upstream of the Tgku80 gene, the ~1 kb fragment of the TgCRT 5’-UTR region, and the removed partial TgCRT coding sequence to produce the TgCRT complementation plasmid, pCRT-TgCRT-mCherry-3xmyc. The complemented TgCRT is driven by its cognate promoter to maintain physiologic similarity to native TgCRT expression in WT parasites. The 1 kb region located ~6 kb upstream of the TgKu80 gene was used to facilitate a single integration of the TgCRT complementation plasmid into this specific locus by single crossover homologous recombination. The TgCRT complementation construct was digested with SwaI restriction endonuclease enzyme, gel-extracted, purified, and transfected into Δcrt parasites by electroporation.

To introduce NanoLuc luciferase (nLuc) into parasites, we PCR-amplified and assembled the TgTubulin promoter, the coding sequence of the nLuc luciferase, and an HXG selection marker into a nLuc expression construct. The resulting plasmid was transfected into WT, Δcrt, and ΔcrtCRT strains. The transfectants were selected with 25 μg/ml mycophenolic acid and 50 μg/ml xanthine. Stable populations were subjected to limiting dilution to generate individual clones of WT::nLuc, Δcrt::nLuc, and ΔcrtCRT::nLuc and clones were confirmed via luciferase activity.

To generate the ΔcrtΔcpb mutant, the TgCPB gene was replaced with a pyrimethamine resistance cassette using the CRISPR-Cas9 genome editing system [50,51]. The pyrimethamine resistance cassette was PCR-amplified and flanked by 50 bp regions upstream and downstream of the start and stop codons of the TgCPB gene for homologous recombination. A 20 bp region located at the beginning of the coding region of the TgCPB gene was used to design guide RNA and replace the guide RNA targeting TgUPRT gene in the plasmid pSAG1-Cas9::UPRTsgRNA using Q5 site-directed mutagenesis (NEB). The Cas9-GFP and guide RNA constructs were co-transfected into Δcrt parasites with the corresponding repair PCR product. The guide RNA and Cas9 generated a gap within the TgCPB gene to facilitate double crossover homologous recombination. Correct gene replacement was confirmed by PCR. Similar strategies were used to create the ΔcrtΔcpl and ΔcrtΔsub1 mutants. Please refer to the figure legend of S6 Fig for more details.

To epitope-tag TgAMN, we again used CRISPR-Cas9 editing tools to modify the corresponding gene. Guide RNA recognizing the 20 bp region near the TgAMN stop codon was generated using the methods above. The 50-bp homologous regions upstream and downstream of the stop codon of the TgAMN gene were cloned at the 5’- and 3’-ends of the DNA sequence containing the 3xHA epitope tag and the pyrimethamine resistance cassette, respectively, by PCR. The plasmid encoding the guide RNA targeting TgAMN and Cas9-GFP and the PCR product were co-transfected into WT parasites. The stop codon of TgAMN was replaced by the 3xHA epitope tag and pyrimethamine resistance cassette. Stable populations were generated after multiple rounds of pyrimethamine selection and the TgAMN-3xHA fusion protein was confirmed by immunoblotting analysis.

TgSCP was endogenously tagged with a 3xmyc epitope tag via single crossover. An approximately 1 kb region upstream of the TgSCP stop codon was PCR amplified and fused in frame with a 3xmyc epitope to assemble TgSCP-3xmyc. A pyrimethamine resistance cassette was also included, the resulting plasmid was linearized and transfected into WT parasites. The correct tagging was confirmed by immunoblotting.

Site-directed mutagenesis

Threonine 369 was mutated to lysine in the WT TgCRT complementation construct, via site-directed mutagenesis according to the Q5 site-directed mutagenesis procedure (NEB). Linear PCR product was phosphorylated, circularized, and transformed into E. coli. Correct clones were identified by direct DNA sequencing.

Transfection of Toxoplasma parasites

T. gondii parasites were allowed to grow in HFF cells for 48 h at 37°C with 5% CO₂. Freshly egressed parasites were syringed, filter purified, and harvested in Cytomix buffer (25 mM HEPES, pH 7.6, 120 mM KCl, 10 mM K₂HPO4/ KH₂PO4, 5 mM MgCl₂. 0.15 mM CaCl₂, and 2 mM EGTA). Parasites were pelleted at 1,000x g for 10 min, washed once in Cytomix buffer, and resuspended in Cytomix buffer at 2.5 x 10⁷ parasites per ml. 400 μl of parasite suspension was mixed with 20 μg DNA and 2 mM ATP/5 mM reduced glutathione to a final volume of 500 μl. The mixture was electroporated at 2 kV and 50 ohm resistance using the BTX Gemini X2 (Harvard Apparatus). Transfectants were inoculated into a T25 flask pre-seeded with a confluent monolayer of HFF cells. The transfected parasites were allowed to recover prior to drug selection applied 24 h post transfection.

Immunofluorescence and colocalization analysis

Freshly lysed parasites were used to infect confluent HFF cells pre-seeded in an 8-well chamber slide for 1 hr (pulse-invaded parasites) or 18–24 h (replicated parasites). The extracellular parasites were attached to chamber slides using 0.1% (w/v) poly-L-lysine. Immunofluorescence was performed as described previously [11,14]. Images were viewed and digitally captured using a Leica CCD camera equipped with a DMi8 inverted epifluorescence microscope and processed with Leica LAS X software. For deconvolution microscopy, a series of Z-stack images were captured and processed by the 3-D deconvolution operation embedded in the Leica LAS X software using the following parameters: Total iterations: 10; Refractive index: 1.52; Method: Blind; Remove background: Yes; Rescale intensity: Yes; AutoQuant Deconvolution algorithms licensed from Media Cybernetics Inc. The fluorescence in final images was presented as maximum projection. The colocalization analysis of two proteins of interests was performed by the Coloc2 plugin embedded in the Fuji image processing suite using the following parameters [52]: Threshold regression, Bisection; Algorithms, Li for both channels. Pearson’s correlation coefficient above threshold was recorded and plotted for comparison. The colocalization analysis was derived from 10 parasites per replicate for a total of three replicates. The statistical analysis was calculated by unpaired two-tailed Student’s t-test.

Measurement of cytosolic pH in Toxoplasma via pHluorin 2 (PHL2)

To express PHL2 in the cytoplasm of the parasites, the gene encoding PHL2 was cloned under the Toxoplasma tubulin promoter in a plasmid carrying a chloramphenicol resistance cassette for selection. The resulting plasmids were introduced into WT, Δcrt, and ΔcrtCRT strains by electroporation. After drug selection, the parasite strains expressing PHL2 were cloned out by limiting dilution. Individual clones were confirmed by fluorescence observation via immunofluorescence microscopy. Prior to pH measurement, a calibration curve was generated by measuring the ratio of pHL2 fluorescence excited at 405 nm and 485 nm in the WT strain expressing cytosolic pHL2 according to previously published methods [53]. Briefly, the PHL2-expressing WT parasites were filter-purified, resuspend in PBS at 1x10⁷ parasites per ml. One hundred microliters of parasites were pelleted and resuspended in buffer with different pH values ranging from 6.2 to 7.8. The detergent, Triton X-100, was added into the parasite resuspensions to 0.1% and incubated at room temperature for 10 min to permeabilize parasite cell membranes for PHL2 release. Released PHL2 was excited at 405 and 485 nm and the emitted fluorescence at 528 nm was recorded via a BioTek Synergy H1 fluorescence plate reader. The ratio of the emission intensities at 528 nm (I405/I485) for both excitation wavelengths were plotted, which yields a calibration equation by linear regression analysis. Cytosolic PHL2 expressing WT, Δcrt, and ΔcrtCRT parasites were harvested, resuspended in PBS, and their I405/I485 ratios were measured. The pH of their cytosol was calculated by applying their I405/I485 ratios to the calibration equation. All assays were repeated in three technical replicates per biological replicate in a total of three biological replicates. Data were presented as mean ± SEM.

Excretory secretory antigens (ESAs) preparation

Freshly egressed parasites were syringed, filter purified, and resuspended at 5 x 10⁸ parasites/ml in D1 medium (DMEM medium supplemented with 1% FBS). One hundred microliters of parasite suspension were transferred to a microfuge tube and incubated at 37°C for 30 min to prepare constitutive ESAs. To isolate induced ESAs, the parasite suspension was incubated in D1 medium supplemented with 1% ethanol for 2 min at 37°C. ESAs were separated from intact parasites by centrifugation at 1,000 x g for 10 min. ESA fractions were transferred to a new microfuge tube, mixed with SDS-PAGE sample loading buffer, and boiled for 5 min for immunoblotting analysis.

SDS-PAGE and immunoblotting

Parasite lysates and ESA fractions were prepared in 1x SDS-PAGE sample buffer and boiled for 5 min before resolving on standard SDS-PAGE gels. For immunoblotting, gels were transferred to PVDF membranes by semi-dry protein transfer methods. Blots were blocked with 5% non-fat milk and incubated with primary antibody diluted in 1% non-fat milk. Goat anti-mouse or anti-rabbit IgG antibodies conjugated with horseradish peroxidase were used as the secondary antibody. Immunoblots were developed with SuperSignal WestPico chemiluminescent substrate (Thermo). The chemiluminescence signals were captured using the Azure Imaging System. Bands were quantified by densitometry using LI-COR Image Studio software.

Parasite invasion and attachment assay

The red-green invasion assay was used to measure the efficiency of parasite invasion. Freshly purified parasites were syringed, filter purified, and resuspended at 5 x 10⁷ parasites/ml in invasion medium (DMEM supplemented with 3% FBS). Two hundred microliters of parasite resuspension were inoculated into each well of an 8-well chamber slide pre-seeded with HFF cells, and parasites were allowed to invade host cells for 30, 60, and 120 min before fixation with 4% formaldehyde for 20 min. Before membrane permeabilization, slides were stained with mouse anti-TgSAG1 monoclonal antibody (1:1,000) for 1 h to label attached parasites. After treatment with 0.1% Triton X-100 for 10 min, the parasites were stained with rabbit polyclonal anti-TgMIC5 antibody (1:1,000) for 1 h to stain both invaded and attached parasites. Subsequently, slides were stained with goat anti-mouse IgG conjugated with Alexa 594 (red) (Invitrogen, 1:1,000) and goat anti-rabbit IgG conjugated with Alexa 488 (green) (Invitrogen, 1:1,000) along with DAPI for nuclear staining. After staining, slides were mounted with anti-fade Mowiol solution and observed by immunofluorescence. Intracellular parasites only showed green fluorescence, whereas extracellular parasites exhibited both red and green fluorescence. Six fields of view from individual invasion experiments were captured by a Leica DMi8 inverted epifluorescence microscope and processed with ImageJ software. The invasion efficiency of each strain was quantified using the following equation ([sum of green parasites] − [sum of red parasites]) ⁄ total host nuclei.

For attachment assay, the HFF cells pre-seeded in 8-well chamber slides were fixed with 2% glutaraldehyde for 10 min at 4°C followed by PBS wash and quench with 1M glycine overnight. The fixed HFF cells were incubated with D3 medium (DMEM supplemented with 3% FBS) for 1 h at 37°C before parasite inoculation. Freshly filter-purified parasites were resuspended at 5x10⁷ parasites per ml in D3 medium (DMEM supplemented with 3% FBS). Two hundred microliters of parasite resuspension (1x10⁷) were added into each well and allowed to attach onto fixed HFF cells for 15 min at 37°C. Unattached parasites were washed off by PBS rinse 10 times. Parasites attached to HFF cells were fixed with 4% paraformaldehyde for 20 min at room temperature and stained with mouse anti-TgSAG1 antibodies for 1 h at room temperature. Goat anti-mouse IgG conjugated with Alexa 594 (red) (Invitrogen, 1:1,000) was used as secondary antibody. Eight view fields in each well for individual strains were captured at 200x magnification and quantified. The assay was repeated in three biological replicates. Final data were combined and plotted using Prism software. The unpaired two-tailed Student’s t-test was used to calculate statistical significance.

Immunofluorescence-based replication assay

Freshly egressed parasites were filter-purified and inoculated into individual wells of an 8-well chamber slide pre-seeded with HFF cells at approximately 1 x 10⁵ cells per well. Non-invaded parasites were washed off at 4 h post-infection. Invaded parasites were allowed to infect host cells for an additional 24 and 36 h before fixation. The infected host cells were stained with monoclonal anti-TgGRA7 (1:1,000) antibody and DAPI to help distinguish individual parasitophorous vacuoles (PVs) and the nuclei of parasites, respectively. Slides were subjected to standard immunofluorescence microscopy for imaging. One hundred parasitophorous vacuoles were enumerated for each strain and plotted as the distribution of different sized PVs. In addition, replication was also expressed as the average number of parasites per PV.

Luminescence-based growth assay

Parasites expressing NanoLuc luciferase were inoculated into a white 96-well tissue culture plate with a flat, solid bottom (Greiner Bio-One) pre-seeded with confluent HFF cells at 1.5 x 10³ cells/well. Each strain was inoculated into 4 individual wells to monitor the fold-change of luciferase activity versus time, which is proportional to intracellular growth. At 4 h post-infection, the individual wells were aspirated to remove non-invaded parasites. The first well was treated with 100 μl of lysis buffer containing NanoLuc luciferase substrate and incubated for 10 min, and a luminescence reading was taken by using the BioTek multimode H1 hybrid plate reader. The remainder of the 3 wells were replenished with fresh D10 medium without phenol red for an additional 24, 48, and 72 h. Subsequent luminescence readings were all performed via the methods above. Luminescence readings versus time were normalized against the reading at 4 h post-infection to calculate the fold-change of parasite growth.

Egress assay

A lactate dehydrogenase release assay was used to measure the egress efficiency of parasites. Freshly lysed parasites were filter-purified and resuspended in D10 medium at 5 x 10⁵ parasites/ml. One hundred microliters of parasite suspension were inoculated into each well of a 96-well plate pre-seeded with HFF cells. The parasites were allowed to replicate for 18–24 h, washed, and incubated with 50 μl of Ringer’s buffer (10 mM HEPES, pH 7.2, 3 mM HaH₂PO₄, 1 mM MgCl₂, 2 mM CaCl₂, 3 mM KCl, 115 mM NaCl, 10 mM glucose, and 1% FBS) for 20 min. Subsequently, an equal volume of 1 mM Zaprinast dissolved in Ringer’s buffer was added to the wells and incubated for 5 min at 37°C and 5% CO₂. Uninfected wells were treated with 50 μl of Ringer’s buffer containing 1% Triton X-100 or normal Ringer’s buffer, serving as positive and negative controls, respectively. The released lactate dehydrogenase was centrifuged at 1,000 x g for 5 min twice to pellet insoluble cell debris. Fifty microliters of supernatant were subjected to the standard lactate dehydrogenase release assay as described previously [54]. The egress efficiency of each strain was calculated using the following equation, ([LDH activity derived from individual parasites]− [LDH activity of negative control]) ⁄ ([LDH activity of positive control]−[LDH activity of negative control]).

Chemically induced motility assay

Twenty-four hours prior to microscopy, 35 mm MatTek dishes (MatTek Corporation) were incubated with 10% fetal bovine serum (FBS) to provide sufficient protein to form a surface conducive to motility. MatTek dishes were washed once with PBS and loaded with 2 ml of Ringer buffer without Ca²⁺. MatTek dishes were chilled on ice and parasites were added and allowed to equilibrate for 15 min. To remove parasites that have not attached, dishes were washed twice with 2 mL of Ringer buffer without Ca²⁺. Dishes were then placed in the General Electric Delta Vision environmental chamber set to 37°C and allowed to equilibrate for 5 min. Time-lapse videos were taken and photographed using an Olympus IX-71 inverted fluorescence microscope with a Photometrix CoolSnapHQ CCD camera driven by Delta Vision software. The exposure duration, gain, laser intensity, and filter settings were identical for all videos taken for quantification. After 30 seconds, 1 μM Ionomycin was added to stimulate motility. Tracings were quantified for two different measurements: A) for circular motility, the total numbers of parasites in the field of view were divided by the total number of parasites completing at least one full circle. Data presented were from 6 independent trials. B) for calculating total distance traveled, ImageJ software with the MTrackJ plugin was used to track and calculate distance. We report the average distance of a minimum of three parasites (out of a maximum of 5) from 5–6 independent biological trials.

Measurement of cytoplasmic calcium with Fura2

Intracellular calcium was determined fluorometrically by loading parasites with Fura 2-AM as described previously [55]. Briefly, parasites were harvested, washed with buffer A (116 mM NaCl, 5.4 mM KCl, 0.8 mM MgSO₄, 5.5 mM D-glucose, and 50 mM HEPES, pH 7.4) with glucose (BAG), and filtered through a 5 μm filter. Parasites were resuspended to 1 x 10⁹ parasites/ml in BAG supplemented with 1.5% sucrose and 5 μM Fura 2-AM. Parasite suspensions were incubated for 26 min in a 26°C water bath with mild agitation. Cells were washed twice with BAG to remove extracellular dye and were resuspended to a final density of 1 x 10⁹ cells/ml. A 50 μl aliquot of the cell suspension was added to 2.45 ml of Ringer buffer without calcium in a cuvette placed in an Hitachi F-7000 fluorescence spectrometer (Hitachi). The excitations were set at 340 and 380 nm, and emission at 510 nm. The Fura2 response was calibrated from the ratio of 340/380 nm fluorescence values [55]. To determine the initial calcium levels via the chemical indicator, initial calcium readings from parasites were determined from 4 independent trials by averaging the first 50 seconds of each tracing.

Size measurement of the VAC

The size of the VAC was quantified based on TgCPL or TgCPB staining (both TgCPL and TgCPB are VAC luminal proteases). Freshly purified parasites were inoculated into pre-seeded HFF chamber slides, allowed to invade host cells for 30 min prior to fixation, stained with polyclonal rabbit anti-TgCPL antibody (1:100) or mouse anti-TgCPB antibody (1:500), and VAC diameter measured by immunofluorescence microscopy. The distance of the widest diagonal of TgCPL or TgCPB staining was used as the diameter of the VAC and was quantified using Leica LAS X software. Measurements for at least 50 individual parasites were performed for each replicate in a total of three replicates. The measurements are presented as mean ± SEM (standard error of mean).

Transcriptome sequencing and quantitative PCR (qPCR) assay

WT and TgCRT-deficient parasites were grown in HFF cells for 2 days. The parasites were syringed, filter-purified, and resuspended in ice-cold PBS buffer. A trizol-based method was used to extracted total RNA by using the Direct-zol RNA MiniPrep Plus kit (Zymo). Purified total RNA was converted to sequencing read libraries using the TruSeq Stranded mRNA sequencing kit (Illumina). The prepared libraries were subjected to 2 x 125 bp paired-end Illumina HiSeq2500 sequencing. Each sample was sequenced to a depth of at least 20 million reads. The sequencing reads per sample were trimmed and mapped to the genome of Toxoplasma GT1 strain (release 34) for gene differential expression profiling by the Clemson University Genomics Computational Lab.

Approximately 500 ng of total RNA was used to measure the steady levels of transcripts for individual genes by using the Luna Universal One-Step RT-PCR kit (NEB). The qPCR assay was performed using the BioRad CFX96 Touch Real-Time PCR detection system. The cycle threshold (CT) values for individual genes were used for double delta CT (ΔΔCT) analysis to calculate their relative abundances to that of WT parasites using the Bio-Rad CFX Maestro software. TgActin was used as the housekeeping gene for normalization.

Mouse studies

Six- to eight-week-old, outbred CD-1 mice were infected by subcutaneous or intravenous injection with 100 WT or mutant parasites diluted in PBS. The infected mice were monitored for symptoms daily for a total of 30 days. Mice that appeared moribund were humanely euthanized via CO₂ overdose, in compliance with Clemson University IACUC’s approved protocol. The seroconversion of the surviving mice was tested by enzyme-linked immunosorbent assay (ELISA). The surviving mice were allowed to rest for 10 days, prior to subcutaneous injection with a challenge dose of 1000 WT parasites, and were monitored daily for survival for additional 30 days.

Generation of TgCRT expression construct in yeast

TgCRT cDNA was PCR-amplified from pTub-TgCRT-mCherry-3xmyc plasmid using a forward primer that introduced a 5’ KpnI site and S. cerevisiae Kozak sequence, and a reverse primer that omitted the mCherry-3xmyc tag and introduced a 3’ XmaJI site. The PCR amplified DNA was digested with KpnI and XmaJI and subcloned into pYES2-6xHis-BAD-V5 (hexa His, biotin acceptor domain, V5 tags) plasmid behind the GAL1 promoter and in front of the His-BAD-V5 epitope tags to generate the plasmid pYES/TgCRT-hbv. To generate the plasmid pYES/PMA-TgCRT-hbv, DNA encoding TgCRT-hbv was PCR amplified using a forward primer that omitted the first 900 bases of TgCRT and introduced a 5’ SacI site, and a reverse primer that included a 3’ NotI site and His-BAD-V5 tags. The amplified DNA was digested with SacI and NotI and subcloned into a SacI/NotI-digested pYES/PfHB3_PMA (from [37]; modified via site-directed mutagenesis to introduce a SacI site at the PMA-PfCRT interface). Mutagenesis reactions were performed using reagents obtained from Agilent (Santa Clara, CA).

Preparation of yeast membrane and immunoblotting

Isolation of yeast membranes and detection of proteins by Western blot were completed using previously published methods [22].

Measurement of CQ incorporation in yeast expressing CRT

Quantitative growth rate analysis was used to calculate CQ transport as previously described in detail elsewhere [22,37,39]. Briefly, growth under each condition was measured in duplicate at an initial cell density of OD₆₀₀ = 0.1 in 96-well plates placed in a Tecan (Durham, NC) M200Pro or BioTek (Winooski, VT) Epoch2 plate reader. CQ-induced growth delays at 80 mM CQ, pH 6.75 were calculated as the difference in time taken to reach maximal growth rate in PMA-TgCRT non-inducing versus inducing media (see [39]).

Statistics

Statistical analysis was performed using Prism software (GraphPad version 8). The methods used in different assays were indicated in the figure legends. All statistical significance analysis in this paper used the raw data sets for calculation (refer to S2 Text for raw data sets).