A Toxoplasma gondii patatin-like phospholipase contributes to host cell invasion

Loading metrics

Open Access
Peer-reviewed

Research Article

Sarah K. Wilson, 

Justine Heckendorn, 

Bruno Martorelli Di Genova, 

Lindsey L. Koch, 

Peggy J. Rooney, 

Naomi Morrissette, 

Maryse Lebrun, 

Laura J. Knoll

x

Published: July 6, 2020

https://doi.org/10.1371/journal.ppat.1008650

?This is an uncorrected proof.

AbstractToxoplasma gondii is an obligate intracellular parasite that can invade any nucleated cell of any warm-blooded animal. In a previous screen to identify virulence determinants, disruption of gene TgME49_305140 generated a T. gondii mutant that could not establish a chronic infection in mice. The protein product of TgME49_305140, here named TgPL3, is a 277 kDa protein with a patatin-like phospholipase (PLP) domain and a microtubule binding domain. Antibodies generated against TgPL3 show that it is localized to the apical cap. Using a rapid selection FACS-based CRISPR/Cas-9 method, a TgPL3 deletion strain (ΔTgPL3) was generated. ΔTgPL3 parasites have defects in host cell invasion, which may be caused by reduced rhoptry secretion. We generated complementation clones with either wild type TgPL3 or an active site mutation in the PLP domain by converting the catalytic serine to an alanine, ΔTgPL3::TgPL3S1409A (S1409A). Complementation of ΔTgPL3 with wild type TgPL3 restored all phenotypes, while S1409A did not, suggesting that phospholipase activity is necessary for these phenotypes. ΔTgPL3 and S1409A parasites are also virtually avirulent in vivo but induce a robust antibody response. Vaccination with ΔTgPL3 and S1409A parasites protected mice against subsequent challenge with a lethal dose of Type I T. gondii parasites, making ΔTgPL3 a compelling vaccine candidate. These results demonstrate that TgPL3 has a role in rhoptry secretion, host cell invasion and survival of T. gondii during acute mouse infection.
Author summary
Toxoplasma gondii is a eukaryotic parasite commonly found in warm-blooded animals worldwide. Successful replication of T. gondii within the host relies on its sophisticated cell invasion mechanism. Previous studies found that disruption of the T. gondii gene TgME49_305140 generated a mutant that could not establish a chronic infection in mice. The protein product of this gene, named TgPL3, is large with both patatin-like phospholipase and microtubule binding domains. Here we show that TgPL3 is localized to the apical end of T. gondii, which is used for invasion. A mutant T. gondii strain with a deletion of TgPL3 (ΔTgPL3) was generated. ΔTgPL3 parasites are defective for host cell invasion and do not cause disease in mice, even at high doses. Complementation of the TgPL3 gene back into ΔTgPL3 parasites repairs the invasion and mice infection defects. Moreover, complementation of ΔTgPL3 with a point mutation in the active site of the phospholipase domain (S1409A) did not rescue the invasion and mouse infection defects, suggesting the PLP domain is responsible for these phenotypes. Vaccination of mice with ΔTgPL3 and S1409A parasites protected them against lethal challenge with T. gondii, highlighting the potential of ΔTgPL3 as a vaccine strain.

Citation: Wilson SK, Heckendorn J, Martorelli Di Genova B, Koch LL, Rooney PJ, Morrissette N, et al. (2020) A Toxoplasma gondii patatin-like phospholipase contributes to host cell invasion. PLoS Pathog 16(7):
e1008650.

https://doi.org/10.1371/journal.ppat.1008650Editor: Michael J. Blackman, Francis Crick Institute, UNITED KINGDOMReceived: August 20, 2019; Accepted: May 22, 2020; Published: July 6, 2020Copyright: © 2020 Wilson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.Data Availability: All relevant data are within the manuscript and its Supporting Information files.Funding: This research was supported by the National Institutes of Health (NIH) National Research Service Award T32 AI007414 (SKW), 1F32AI065023-01A2 (PJR), and R01AI144016-01 (LJK), Department of Defense, Air Force office of Scientific Research, National Defense Science and Engineering Graduate Fellowship (32 CFR 168a, SKW), LABoratoires d’EXcellence ARCANE, ParaFrap ANR-11-LABX-0024 (ML), and the Morgridge Metabolism Interdisciplinary Fellowship from the Morgridge Institute for Research (BMDG). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.Competing interests: PJR is currently employed by the Stratatech Corporation. Her affiliation with this company did not affect this study in any way. This affiliation does not alter our adherence to all PLOS Pathogens policies on sharing data and materials. The other authors have declared that no competing interests exist.

IntroductionToxoplasma gondii is an obligate intracellular parasite that can infect all warm-blooded animals and any nucleated cell within the host. T. gondii is a member of the Apicomplexa phylum, which includes Plasmodium spp., the causative agent of malaria, and Cryptosporidium, one of the leading causes of waterborne disease outbreaks. Worldwide, T. gondii infects an estimated 30–50% of humans [1], which are dead-end intermediate hosts. The sexual cycle of T. gondii is restricted to the feline intestinal epithelium, from which T. gondii is excreted as oocysts in feces. The asexual cycle of T. gondii can occur in any warm-blooded animal and has two developmental stages: a rapidly replicating form called the tachyzoite and a slow growing stage called the bradyzoite. T. gondii is acquired orally either by ingestion of oocyst-contaminated vegetables or other foods, or by eating bradyzoite cyst-harboring meat products. Upon infection of an intermediate host, cyst and oocyst stages differentiate into the tachyzoite form and disseminate throughout the host before transitioning to the bradyzoite form. Bradyzoite-containing cysts reside in the brain and muscle tissue, and represent the long-term chronic phase of infection. Most infections are asymptomatic, but immunocompromised persons are at risk for severe primary infections or reactivation of a chronic infection that can lead to encephalitis or death. To date, no medications are available that can clear the chronic stage of T. gondii.
The T. gondii lytic cycle begins with a complex invasion mechanism. Gliding motility propels the parasite to find permissive host membranes [2]. Upon initial attachment, a tighter interaction is enabled by the formation of the moving junction (MJ). The MJ is a complex of the microneme protein apical membrane antigen 1 (AMA1) and rhoptry neck (RON) proteins, which are secreted from apical parasite organelles called micronemes and rhoptries, respectively [3]. While the RON proteins are typically part of the invasion machinery, the rhoptry bulb (ROP) proteins secreted during this process can modulate the host immune and metabolic pathways to be more favorable for parasite survival [4]. The parasite moves through the MJ, invaginating the host cell membrane to create a parasitophorous vacuole in which the parasite will replicate.
Previously, a signature tagged mutagenesis (STM) screen identified insertional T. gondii mutants that showed a decrease in chronic cyst burden compared to wild type parasites in vivo [5]. One of the identified mutants corresponded to TgME49_305140, annotated as a patatin-like phospholipase (PLP) domain containing protein, and reported here as TgPL3. PLP enzymes were first discovered in potato tubers and played a role in storage [6, 7], but these enzymes have since been implicated in both lipid metabolism and inflammation in plants, animals, and more recently microorganisms. Bacterial PLPs have phospholipase activity and are virulence factors [8, 9]. There are many PLP domain containing proteins across the Apicomplexa phylum and six putative PLP enzymes in T. gondii, two of which have been investigated [8]. The first described T. gondii PLP, named TgPL1, is critical for survival of parasites in activated macrophages [10]. During macrophage activation and bradyzoite cyst development, TgPL1 changes location from within punctate vesicles in the cytoplasm to the parasitophorous vacuole membrane [11, 12]. In late chronic infection, mice infected with ΔTgPL1 parasites had increased expression of inflammatory cytokines and fewer brain lesions from reactivated cysts compared to wildtype-infected mice [13]. Another T. gondii PLP, TgPL2, was shown to localize to the apicoplast and play a critical role in lipid homeostasis within that organelle [14]. Knockdown of TgPL2 led to excessive accumulation of membranes around the apicoplast, altered the lipid profile of the parasite, and resulted in parasite death over time.
Here we present the functional characterization of TgPL3. Along with the PLP domain, TgPL3 contains a microtubule binding domain and is localized in the apical end of the parasite. ΔTgPL3 parasites have reduced host cell invasion, which may be caused by a defect in rhoptry secretion. ΔTgPL3 parasites are also virtually avirulent in vivo, with reduced cytokine stimulation and parasitemia. Mutation of the active site serine recapitulated the in vivo and in vitro defects, suggesting that phospholipase activity is responsible for rhoptry secretion and invasion. These results show that TgPL3 plays a role in invasion of host cells, unique from the first two characterized PLPs.

Results
TgPL3 has predicted microtubule binding and patatin-like phospholipase domains
Sequence and structural alignment of the 277kDa TgPL3 protein revealed a PLP domain predicted to have phospholipase A2 (PLA2) activity and a MIP-T3 microtubule-binding domain (MtBD) surrounded by polyserine stretches (Fig 1A), which may act as flexible linkers allowing the binding domain to be accessible to its target [15]. PLP domains include a glycine-rich oxyanion hole, a catalytic serine-aspartate dyad where the serine falls within the lipase motif G-X-S-X-G and a conserved proline between the serine and aspartate (Fig 1B). Humans have nine predicted PLPs and all of these PLPs contain the catalytic serine-aspartate dyad [16]. To further analyze the PLP potential of TgPL3, we threaded the protein sequence of its PLP domain onto the crystal structure for the plant PLP [17]. The TgPL3 PLP domain aligns with the plant model to form an active site (Fig 1C).
Fig 1. Predicted domains of TgPL3 based on sequence and structural alignments.(A) In orange, two serine-rich domains (S) flank the microtubule binding domain (MtBD) in green. The first serine-rich domain contains 34 consecutive serines and the second contains 37 serines interrupted by 2 prolines and 2 alanines. The putative PLP domain is shown in black. (B) The PLP domain contains all the necessary conserved amino acids for activity based on sequence alignment. The known conserved patatin motifs are underlined and catalytic S/A dyad are colored red. The serine to alanine mutation (S1409A) is indicated by an arrow. (C) Structural alignment of the TgPL3 PLP domain (orange) to the patatin PLA2 crystal structure (blue) predicts correct folding to form the active site. The secondary and tertiary structures of the TgPL3 PLP domain were predicted using I-TASSER and the resulting model was aligned to the PDB model of patatin (1OWX) in PyMOL. The catalytic serine and aspartate dyad are shown in green in each model.

https://doi.org/10.1371/journal.ppat.1008650.g001

TgPL3 localizes to the apical end of T. gondii
The previous two PLP genes characterized in T. gondii had unique localizations and functions. To localize TgPL3, we generated antibodies against TgPL3 by immunizing mice with either a wheatgerm-purified PLP domain (Fig 2) or three different synthesized peptides (S1A Fig). All antibodies generated showed TgPL3 localizing to the apical end of intracellular RHΔKu80ΔHPT parasites. Counter staining with MIC2 antibodies revealed TgPL3 localizes above the micronemes (Fig 2A, S1B). We next examined the apical ring and conoid regions further using RNG1-YFP and CAP1-YFP fusion RH T. gondii strains [18,19]. In intracellular parasites, TgPL3 localizes below both TgRNG1 and TgCAP1 (Fig 2B & 2C). However, in extracellular parasites where conoid extrusion is induced, TgPL3 partially extends out past the apical ring along with the conoid (Fig 2D).
Fig 2. TgPL3 localizes to the apical end of tachyzoites.Top three panels show immunofluorescence of TgPL3 (red) in intracellular tachyzoites co labeled with (A) MIC2, (B) RNG1 or (C) CAP1 (green). Bottom two panels show TgPL3 localization in extracellular tachyzoites after (D) induced conoid extrusion or (E) deoxycholate extraction (DOC). (F) Schematic showing costain localization in intracellular parasites (created with BioRender.com). Images for panels A, B, C and D were taken under the same magnification and the scale bars in the lower right corners equal 5 μm.

https://doi.org/10.1371/journal.ppat.1008650.g002As described above, TgPL3 has a predicted microtubule-binding domain (Fig 1A). To examine cytoskeleton association, we extracted the cytoskeleton of extracellular parasites in deoxycholate (DOC) detergent. When T. gondii are lysed with DOC, only the cytoskeleton and its tightly bound proteins remain [18]. After DOC extraction, TgPL3 colocalized to the apical end, identified using counterstaining for ß-tubulin (Fig 2E and S2 Fig). Thus, TgPL3 is localized to the apical cap region of the parasite (schematized in Fig 2F).

Purification of the TgPL3 patatin-like domain and testing for PLA2 activity
To test if TgPL3 had PLA2 activity we first purified the 67 kDa PLP domain. Attempts to express the PLP domain in E. coli were unsuccessful, likely due to the toxicity of the protein in live cells. We then switched to an in vitro transcription and translation system to express the TgPL3 PLP domain. The domain was successfully expressed in wheat germ extract; however, multiple attempts to show PLA2 activity using a thin layer chromatography assay were unsuccessful. It is likely that expressing the 67 kDa PLP domain apart from the rest of the 277kDa TgPL3 protein creates a misfolded protein or otherwise inactive protein. For example, the Pseudomonas aeruginosa PLP requires ubiquitin or ubiquitinated proteins as cofactor for in vitro activity [20]. To address potential PLA2 activity for this manuscript, we engineered a TgPL3 complementation construct with TgPL3 containing a single point mutation that changes the predicted catalytic serine in the lipase domain to an alanine (Fig 1B).

Rapid fluorescence activated cells sorting (FACS) needed to obtain TgPL3 deletion
The TgPL3 mutant discovered in the STM screen [5] contained an insertion in the TgME49_305140 promoter that created a fusion transcript of the chloramphenicol acetyl transferase (CAT) and TgPL3 coding regions controlled by the constitutively active α-tubulin promoter (S3 Fig). To assess the role of TgPL3 in pathogenesis, we deleted the entire open reading frame (ORF) in RHΔKu80ΔHPT parasites by CRISPR. Initial attempts to delete the TgPL3 ORF from T. gondii strain ME49ΔHPT using hypoxanthine-xanthine-guanine phosphoribosyl transferase (HPT) as the positive selectable marker and uracil phosphoribosyl transferase (UPT) as the negative selectable marker were unsuccessful. While PCR indicated a population of successful TgPL3 deletion mutants in newly transfected parasites, these deletion mutants were rapidly outcompeted by drug-resistant random insertion mutants during drug selection. The recent genome-wide CRISPR screening database revealed this gene has a phenotype score of -3.48, which predicts that this gene is fitness conferring during infection of human fibroblasts [21].
Therefore, we created a rapid screening process using fluorescent markers in place of drug-selectable markers. This method uses mCherry as a positive marker and GFP as a negative marker (Fig 3A). After transfection, the parasites were allowed to invade and replicate for 2–3 days without serial passage, whereupon they were sorted by flow cytometry. Parasites that were mCherry positive and GFP negative were sorted directly into 96-well plates for clone isolation. In addition, the RHΔKu80ΔHPT T. gondii strain, which lack the non-homologous end-joining protein Ku80, was used for knockout generation to virtually eliminate random insertion of the DNA construct [22]. After transfection into the RHΔKu80ΔHPT strain, we saw mCherry positive and GFP negative populations during FACS (S4 Fig), whereas transfection into to ME49ΔHPT parasites yielded primarily random insertion mutants rather than knockout clones. We isolated several TgPL3 deletion (ΔTgPL3) clones from the RHΔKu80ΔHPT transfection. These clones were confirmed to be TgPL3 deletions by Southern blot using a probe specific to the 5’ flanking region of TgME49_305140, which was present in both the endogenous and knockout loci (Fig 3B). We also confirmed the deletion by immunofluorescence assay using the anti-TgPL3 antibody and found that TgPL3 was present in the apical end in parental parasites but not the ΔTgPL3 mutants (Fig 3C).
Fig 3. Deletion of TgPL3 from RHΔKu80 parasites using FACS sorting.(A) Schematic for the targeted knockout of the TgPL3 gene in T. gondii by double homologous crossover. Parental strain RHΔku80ΔHXGPRT parasites were electroporated with the CRISPR/Cas9 plasmid containing gRNA targeting the 5’ end of TgPL3 and the knockout plasmid containing TgPL3 5’ and 3’ flanking regions surrounding mCherry which allows for sorting mCherry positive parasites using FACS. Sorting GFP negative clones in tandem ensures a double crossover event. (B) Genomic DNA of clones were screened by southern blot using a probe targeting the 5’ upstream (US) flanking region of TgPL3. Parental RHΔku80ΔHXGPRT (WT), complement (C1) and S1409A (SA) parasites show a 12kb band, ΔTgPL3 (KO) strains show a 3kb band. (C) Parasite strains were further analyzed by immunofluorescence of TgPL3 (red) co-labeled with MIC2 (green). Images were taken under the same magnification and the scale bar in the lower right corner equals 5 μm.

https://doi.org/10.1371/journal.ppat.1008650.g003

Complementation of ΔTgPL3 parasites
To ensure the in vitro phenotypes associated with loss of TgPL3 were not due to CRISPR off-target effects, we complemented the ΔTgPL3 parasites. We generated ΔTgPL3::TgPL3 clones (complement) by targeting TgPL3 cDNA back into the endogenous locus using a CRISPR/Cas9 plasmid with a mCherry specific guide RNA and screened for loss of mCherry expression. We also created a ΔTgPL3::TgPL3S1409A complement strain (S1409A) with an active site mutation in the PLP domain by converting the catalytic serine to an alanine (Fig 1C). Complement clones were sorted by flow cytometry and screened by southern blot (Fig 3B). Restoration of TgPL3 protein was confirmed by immunofluorescence assay (Fig 3C).

ΔTgPL3 parasites have an invasion but not an attachment defect
The difficulty obtaining ΔTgPL3 parasites and a -3.48 CRISPR score [21] suggested there would be a severe growth defect with the loss of TgPL3. Indeed, we regularly passaged more ΔTgPL3 parasites than parental strain to maintain synchronized cultures. We first analyzed the internal replication rate of each strain by counting the number of parasites per vacuole at 18 and 26 hours post-invasion. We did not observe significant differences in the replication rates of invaded parasites between the parental, knockout and complement strains (Fig 4A and S5 Fig). However, plaque assay, which measures parasite growth over multiple lytic cycles, showed fewer and smaller plaques in ΔTgPL3 and S1409A compared to WT and complement strains (Fig 4B and S6A Fig). The localization of TgPL3 to the apical end of the parasite suggested involvement in invasion, so we then evaluated the ability of the mutants to invade using the red/green invasion assay [23]. The parasites were allowed to infect fibroblasts for 5 minutes prior to fixation and extracellular and intracellular parasite were quantified by differential staining. A significantly lower percent of ΔTgPL3 parasites invaded the host cell compared to the parental strain (Fig 4C and S7 Fig). This invasion defect was restored by complementation with TgPL3 but not S1409A, highlighting that phospholipase activity is likely involved in the invasion phenotype. Because the invasion assay is biased based on the ability of the parasite to successfully attach prior to invasion, we analyzed the ability of the parasites to attach using the glutaraldehyde fixed host cell assay [24]. We saw no difference in the attachment between the parental, knockout and complement strains (Fig 4D and S8 Fig). As a negative control we incubated all parasites strains with BAPTA-AM and saw no parasite attachment (S8 Fig). Finally, when we induced the egress of intracellular parasites, there was no difference in the ability of each strain to successfully egress from the host cell (S6B Fig).
Fig 4. TgPL3 is critical for invasion but not attachment.(A) Triplicate monolayers of HFFs were infected with 3.4 x 104T. gondii parasites. At 18 or 26 hours post infection, tachyzoites per vacuole were scored in at least 100 randomly encountered vacuoles per replicate. Average tachyzoites per vacuole is shown at each time point. P value was nonsignificant in a two tailed t-test compared to WT. (B) 9 days post infection, HFF monolayers were stained with crystal violet to reveal T. gondii plaque formation. White arrows indicate the small plaques created by the KO and SA parasites. (C) The percentage of parasites that were successfully able to invade the host cell was determined using the red/green invasion assay [21]. Four independent experiments were performed in triplet, and the combined results are shown here. *** p
Read More