Tachyzoites of T. gondii strains ME49 and TgHB2 were propagated in human foreskin fibroblasts (HFF), which were cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 μg/ml streptomycin and 10 mM L-glutamine. The genetically modified parasites were propagated and cultured under the same conditions.

Gene expression datasets available from ToxoDB were used for this purpose. First, the “Oocyst, tachyzoite, and bradyzoite developmental expression profiles (M4)” dataset was searched for protein coding genes whose expression had a fold change greater than 10 between the tachyzoite (2 days in vitro) and bradyzoite (8 days in vitro) stages. Second, the “Transcriptome during acute or chronic infection in mouse brain” dataset was searched using the same parameters to find genes that had 10-fold difference between “acute infection 10 days p.i.” and “chronic infection 28 days p.i.”. Third, the “Bradyzoite Differentiation Expression Profiles (ME49, GT1, CTGara)” dataset was used to find genes that had 10-fold expression difference between “ME49 tachyzoite” and “ME49 pH 8.2.” Subsequently, genes that were identified in all three searches were collected and analyzed manually. There were 18 genes identified in total, which included lactate dehydrogenase LDH2, bradyzoite antigen BAG1, bradyzoite rhoptry protein BRP1 and other well described bradyzoite specific genes. Judging from the “Product Production” provided by ToxoDB, two (ANK1 with the gene ID TgME49_216140 and DnaK-TPR with the gene ID TgME49_202020) out of the 18 proteins contained tetratricopeptide repeats (TPR), which then became the focus of this study.

To further characterize the TPR containing proteins, their primary sequences were used in protein Blast searches¹, as well as in the Motif Scan² server to search the Prosite and Pfam databases, to identify potential function domains.

To predict the number and positions of the TPR motifs in TPR containing proteins, their sequences were analyzed by the profile based method TPRpred³ (Karpenahalli et al., 2007), using the P-value threshold of 10^-4.

To construct homologous templates for ANK1 or DnaK-TPR knockouts, the 5′- and 3′- homology arms of the corresponding genes were cloned into the pUC19 vector along with the selection marker DHFR, using multi-fragment cloning. The primers used to amplify the homology arms from genomic DNA of ME49 were listed in Table 1. After obtaining each fragment, 5′ homology arm, DHFR and 3′ homology arm were ligated into the linearized pUC19 (achieved through SacI and HindIII digestion) vector by the One Step Cloning Kit (ClonExpress II One Step Cloning Kit, Vazyme, United States) according to manufacturer’s instructions. The gene specific CRISPR plasmids were constructed by site directed mutagenesis of the UPRT targeting CRISPR plasmid, as previously described (Shen et al., 2014, 2017). Primers used for mutagenesis were also listed in Table 1. All constructs were confirmed by Sanger sequencing before use.

To generate gene knockouts in desired strains, corresponding homology construct and gene specific CRISPR plasmid were co-transfected into freshly egressed tachyzoites, as described previously (Shen et al., 2017). Subsequently transfectants were selected with 1 μM pyrimethamine for 3 – 4 passages until the drug resistant pools became stable. Single clones were isolated through limiting dilution in 96-well plates and examined by diagnostic PCRs (PCR1, PCR2 and PCR3; primers are listed in Table 1) to check the disruption of corresponding genes.

Approximately 1 × 10⁷ freshly egressed parasites (tachyzoites or bradyzoites) of the ME49 strain were collected and purified through polycarbonate membranes with the pore size of 3 μm. Parasites were washed with PBS and total RNA was extracted from them using the trizol methods, according to manufacturer’s instructions (TransGen Biotech, China). Bradyzoites were obtained by growing parasites in differentiation medium (RPMI 1640 medium supplemented with 1% FBS and 50 mM HEPES, pH 8.2) along with ambient CO₂ for 4 days (Walker et al., 2013). After RNA isolation, 2 μg total RNA from each sample was reverse transcribed into cDNA using the PrimeScript RT reagent Kit (Takara Bio, Japan).

Subsequently the cDNA was used as template for real-time PCR (RT-PCR) to estimate the expression levels of target genes. Three pairs of specific primers were designed for this purpose: ANK1-Fw and ANK1-Rv; DnaK-TPR-Fw and DnaK-TPR-Rv; GAPDH-Fw and GAPDH-Rv (Sequences are listed in Table 1). RT-PCR was performed on the ViiA-7 Real-Time system (Life Technologies, Camarillo, CA, United States) using the SYBR Green method, according to the manufacturer’s instructions. Cycling conditions for RT-PCR were set as: initial denaturation at 95°C for 30 s, followed by 40 cycles of 95°C for 15 s, 56°C for 30 s and 72°C for 20 s. Each sample was prepared and examined three times independently. Gene expression levels was estimated by the Δ^C_T method using the GAPDH gene as internal reference, as described previously (Selseleh et al., 2012).

Freshly egressed tachyzoites (2 × 10⁷) were purified through membrane filtration and washed with PBS. Subsequently genomic DNA was extracted from the purified parasites using the EasyPure Genomic DNA Kit (Transgen Biotech, China). Purified genomic DNA was sheared into short fragments of about 350 bps by sonication and then used for library construction using the TruSeq Library Construction Kit (Illumina, San Diego, CA, United States) according to the manufacturer’s instructions. The libraries were then sequenced by the Illumina HiSeq platform (Illumina, San Diego, CA, United States). The average sequencing depth was around 110X. To detect genetic alterations, clean reads obtained from genome sequencing were mapped to the reference genome of GT1, and SAMTOOLs, BreakDancer and CNVnator were used to detect short indels, structural variations (insertion, deletion, and inversion) and copy number variations, respectively (Chen et al., 2009; Li et al., 2009; Abyzov et al., 2011). The mapping results were visualized by the Integrative Genomics Viewer (Thorvaldsdottir et al., 2013).

Human foreskin fibroblasts cells were seeded into six-well plates and grown at 37°C until confluent. Subsequently 100 purified tachyzoites were added to each well and cultured for 10–12 days without disturbance under standard growth conditions. Then the samples were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet to develop the plaques. Finally the plates were scanned by a scanner (FileScan 380, Microtek, China) and the number and sizes of plaques in each well were determined as described previously (Shen and Sibley, 2014). Each strain was tested three times independently, each with triplicates.

Coverslips in 24-well plates seeded with confluent HFF monolayers were infected with freshly harvested tachyzoites (10⁵ parasites per well) for 1 h at 37°C. Extracellular parasites that failed to invade HFF cells were washed away and invaded ones were grown for another 24 h under standard tachyzoite growth conditions. Subsequently the samples were fixed with 4% paraformaldehyde and stained with swine anti-Toxoplasma (gifts from the Zhao lab at Huazhong Agricultural University, to stain extracellular parasites). After extensive washing, samples were permeabilized with 0.1% Triton X-100 for 10 min and stained with rabbit anti-TgALD (gifts from the Sibley lab at Washington University in St. Louis, to stain all parasites). Primary antibodies were detected by Alexa Fluor 594-conjugated goat anti-rabbit IgG and FITC conjugated goat anti-swine IgG secondary antibodies, respectively (Life Technologies, Camarillo, CA, United States). Parasites that were stained red but not green (these were intracellular parasites) were analyzed to determine the number of parasites in each parasitophorous vacuole (PV). Each sample was tested three times independently, each with triplicates.

Alkaline medium was used to induce bradyzoite differentiation in vitro, as described previously (Walker et al., 2013). Briefly, Purified parasites were used to invade HFF monolayer (10⁴ parasites/well) seeded on coverslips in 24-well plates for 1 h under standard growth conditions. Subsequently non-invaded parasites were washed away with PBS and the rest of the cultures were grown in differentiation medium with ambient CO₂ for 4 days. After treatment, samples were fixed with 4% paraformaldehyde, permeablized with 0.1% Triton X-100, and stained with rabbit anti-TgALD and mouse anti-TgBAG1 (gifts from the Zhao lab at Huazhong Agricultural University) antibodies. Alexa Fluor 594-conjugated goat anti-rabbit IgG and Alexa Fluor 488-conjugated goat anti-mouse IgG secondary antibodies (Life Technologies, Camarillo, CA, United States) were used to detect primary antibodies. Hoechst 33342 was used to stain cell nuclei. Parasites that were stained both green and red were counted as bradyzoites, whereas parasites that were stained only red were treated as tachyzoites. Bradyzoite differentiation efficiency was determined by dividing the number of green vacuoles by the number of red vacuoles. At least 100 total vacuoles were counted for each sample in each experiment and each sample was repeated three times independently.

All 4–6 weeks old female ICR mice and 10–12 weeks old Kunming mice were purchased from the Hubei Provincial Centers for Disease Control. To test the virulence of parasites in mice, freshly harvested tachyzoites were purified by filtration through 3 μm polycarbonate membranes and then used to infect mice by intraperitoneal injection. Each parasite strain was tested by infection of 10 mice at the dose of 100 tachyzoites per mouse. After infection, mice were monitored daily for 30 days. The cumulative mortality was calculated by dividing the number of mice that died by the number of mice infected. At the conclusion of the virulence tests, mice that survived the infection at day 30 were sacrificed and brain tissues were homogenized to determine the number of Toxoplasma cysts by DBA-FITC staining, as described previously (Buchholz et al., 2013). All animal experiments were approved by The Scientific Ethic Committee of Huazhong Agricultural University (permit #: HZAUMO-2016-041).

Both ANK1 and DnaK-TPR have been reported previously as proteins whose expression was drastically increased during tachyzoite to bradyzoite conversion. ANK1 was initially reported as 35.m00032 and DnaK-TPR was reported as 20.m00351 (Friesen et al., 2008). According to the current gene model on ToxoDB, ANK1 is a multi-domain protein with 563 amino acids. Sequence analysis using BLAST, Pfam and PROSITE database searching indicated that it has an ankyrin repeat domain in the middle with 5 – 7 ankyrin repeats next to each other, depending on the program used (Figure 1). The TPR prediction program TPRpred (Karpenahalli et al., 2007) predicted that it likely has 3 TPR motifs near the C-terminus. The positions and sequences of these TPR motifs were shown in Figure 1. DnaK-TPR is also a multi-domain protein with 763 amino acids. Sequence analysis suggested that it has a DnaK/HSP70 family domain with molecular chaperone function in the central part (roughly between residues 248 and 396). TPRpred predicted that it may have two TPR motifs near the end of the protein, the one between A711 and S744 is of high probability (Figure 1).

Previous studies (Friesen et al., 2008; Ueno et al., 2011), as well as gene expressing data (Micro-array and RNA-Seq) from ToxoDB, all suggested that the expression of ANK1 and DnaK-TPR was significantly different between the tachyzoite and bradyzoite stages. To quantitatively analyze the expression differences of these two genes between the tachyzoite and bradyzoite stages, real-time PCR (RT-PCR) was performed, using cDNA reverse transcribed from total RNA of tachyzoites or in vitro induced bradyzoites of the ME49 strain. Bradyzoites were obtained by growing parasites in differentiation medium for 4 days with ambient CO₂. Using GAPDH expression levels as reference, RT-PCR results showed that expression of ANK1 was barely detectable at the tachyzoite stage, but increased almost 500-fold when cultured under bradyzoite inducing conditions (Figure 2). Similarly, there was a 16-fold increase in the transcript level of DnaK-TPR upon bradyzoite induction (Figure 2). Therefore, expressions of both ANK1 and DnaK-TPR were indeed significantly up-regulated upon bradyzoite induction, consistent with the data from ToxoDB and previous studies (Friesen et al., 2008; Ueno et al., 2011). In addition, using alkaline media or CO₂ starvation to treat the parasites, the Roos lab at University of Pennsylvania looked at gene expression changes every 6 h upon bradyzoite induction. Their data deposited in ToxoDB suggest that obvious increase in DnaK-TPR expression was observed 36 h post either treatment. Whereas for ANK1, rise in expression was seen as early as 12 h upon alkaline stress, but did not become apparent until 36 h after CO₂ starvation. These results imply that both ANK1 and DnaK-TPR are early induced genes during bradyzoite conversion.

To study the functions of ANK1 and DnaK-TPR during T. gondii growth and development, the CRISPR/CAS9 genome editing technique was used to delete these genes in type 2 strain ME49 and type 1 strain TgHB2. The strategy for deleting target genes by CRISPR/CAS9 mediated homologous recombination was illustrated in Figure 3A. After transfecting the gene specific CRISPR plasmid and homology template into corresponding parental strain (ME49 or TgHB2), transfectants were selected with pyrimethamine and single cloned by limiting dilution. Single clones were then screened by diagnostic PCR to check the correct replacement of target genes by the selection marker DHFR (PCR1/2/3 in Figure 3A). Using this strategy, ANK1 and DnaK-TPR single deletion mutants were generated in both ME49 and TgHB2. Diagnostic PCR results for one clone of ME49 Δank1 and one clone of ME49 ΔdnaK-tpr were shown in Figures 3B,C, respectively. Results for the same mutants produced in TgHB2 were identical to that made in ME49, therefore were not shown here. To further confirm the successful knockout of these genes, we performed whole genome sequencing on the two mutants made in the TgHB2 background. For both TgHB2 Δank1 and TgHB2 ΔdnaK-tpr, complete absence of target gene was observed (Figure 4). Meanwhile, we detected increased sequencing coverage for the exons of the DHFR-TS gene, which corresponds to the increased copy number of DHFR as a consequence of replacing target genes with DHFR (Figure 4). Together, these results confirmed the complete deletion of the ANK1 and DnaK-TPR genes, successful disruption of these genes indicated that neither gene is required for tachyzoite survival, which is consistent with their low expression at the tachyzoite stage.

After obtaining the knockout mutants, we sought to determine the contribution of these genes to the growth of T. gondii tachyzoites in detail. First, plaque assay was used to assess the overall growth rates of these mutants. As shown in Figure 5, after 10 or 12 days’ growth, both ANK1 and DnaK-TPR deletion mutants formed similar number and sizes of plaques on HFF host cell monolayer as the parental strain did. Mutants made in both ME49 and TgHB2 produced plaques with similar efficiencies as their corresponding parental strain. These results suggested that neither ANK1 nor DnaK-TPR had a significant impact on tachyzoite growth in vitro.

To further check the parasite growth in vitro, intracellular replication assay was performed to check the proliferation efficiency. Tachyzoites of each strain were used to infect HFF cells for 1 h and successfully invaded parasites were allowed to replicate for 24 h under standard growth conditions. Then the number of parasites in each parasitophorous vacuole (PV) was determined by immunofluorescent staining of Toxoplasma aldolase (Figure 6A). The results showed that both ANK1 or DnaK-TPR deletion mutants displayed very similar replication dynamics as the parental strains (Figures 6B,C), suggesting that ANK1 and DnaK-TPR do not play critical roles for tachyzoite replication in vitro.

As genes whose expressions were significantly up-regulated at the bradyzoite stage, they were hypothesized to play roles in bradyzoite differentiation or development. To check these possibilities, mutant strains were induced with alkaline conditions to form bradyzoites in vitro, which were detected by positive staining of a bradyzoite marker BAG1 (Figure 7A). After growing parasites under the differentiation condition for 4 days, about 80% of ME49 PV turned BAG1 positive, indicating a high efficiency of bradyzoite transition (Figure 7B). ME49 Δank1 and ME49 ΔdnaK-tpr displayed very similar transition rates as the parental strain ME49. Under this condition, about 70% of the TgHB2 PV became BAG1 positive (Figure 7B), which is slightly lower than that of ME49, consistent with the type 1 nature of this strain. Nonetheless, ANK1 or DnaK-TPR deletion mutants made in this strain also had similar bradyzoite formation efficiencies as the parental strain, suggesting that indeed these two genes did not have significant roles in bradyzoite formation in vitro.

To check the effect of ANK1 or DnaK-TPR disruption on parasite virulence in vivo, tachyzoites of mutant strains were used to infect ICR mice and the survival of mice were monitored. The results for the ME49 serials of strains (ME49, ME49 Δank1, Me49 ΔdnaK-tpr) were shown in Figure 8. At the infection dose of 100 tachyzoites per mouse, the parental strain ME49 killed all the mice, whereas both ANK1 and DnaK-TPR single deletion mutants killed about 70% of infected mice, suggesting that the virulence of both ANK1 and DnaK-TPR mutants was slightly attenuated in laboratory mice.

Since the expression of both ANK1 and DnaK-TPR are dramatically up-regulated in bradyzoites, we sought to determine whether they have any role in chronic infection establishment. To do that, 10–12 weeks old Kunming mice (genetically similar to ICR mice) were infected with 100 tachyzoites of each strain. Survival of mice was monitored daily and the number of cysts in the brain of mice that survived 30 days post infection was estimated. The use of 10–12 weeks old Kunming mice is because typical 4–6 weeks old ICR mice were susceptible to ME49 infection and no survivals could be obtained 30 days post infection. DBA-FITC staining and cyst counting suggested that there was no significant difference in brain cyst formation between the ANK1 or DnaK-TPR knockout mutants and parental strain ME49 (Figure 9), indicating that neither gene directly contributes to chronic infection establishment.