IMC1 is recruited onto developing daughter buds concurrently with IMC4 and later than IMC3, 6, and 10

Among the fourteen alveolins that comprise the SPN, IMC1, IMC3, IMC4, IMC6, and IMC10 have been shown to be present in the IMC body of both maternal and budding parasites [9,23,24]. While IMC1 and IMC4 are expressed equally in both the maternal and daughter cells, IMC3, IMC6, and IMC10 are enriched in the daughter cells during bud formation. Despite their localizations being previously identified, the timing of the recruitment of these proteins to the daughter bud cytoskeleton has yet to be characterized in T. gondii. To explore the precise timing of the expression and loading onto daughter buds of these key alveolins, we co-stained wild-type parasites with IMC1 and IMC3, IMC4, IMC6, and IMC10. We determined that IMC1 and IMC4 are recruited onto the growing daughter buds at equivalent times and always have overlapping localization patterns (Fig 1A). In contrast, we found that IMC3 (Fig 1B), IMC6 (Fig 1C), and IMC10 (Fig 1D) are all detected in daughter buds prior to the appearance of IMC1, suggesting that IMC1 and IMC4 are recruited to the growing IMC network later in replication than IMC3, IMC6, and IMC10. We also observed that IMC1 and IMC4 extend into the apical cap region of the IMC, whereas IMC3, IMC6 and IMC10 are restricted to the body portion of the IMC and are not present in the apical cap (Fig 1B–1D, yellow arrowheads).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. IMC1 is expressed later than IMC3, IMC6, and IMC10 and concurrently with IMC4.

A) IFA showing IMC1 is expressed similar to IMC4 in early, middle, and late stages of daughter bud formation. Magenta, anti-IMC1; Green, anti-IMC4. B) IFA showing IMC3 is expressed earlier than IMC1 during daughter bud formation. Magenta, anti-IMC1; Green, anti-IMC3. C) IFA showing IMC6 is expressed earlier than IMC1 during daughter bud formation. Magenta, anti-IMC1; Green, anti-IMC6. D) IFA showing IMC1 is expressed earlier than IMC10 during daughter bud formation. Magenta, anti-IMC1; Green, anti-IMC10. White arrows highlight the earlier expression in the early daughter bud panels. Yellow arrowheads indicate the apical end of parasite where IMC1 is present but IMC3, IMC6, and IMC10 are not. All scale bars are 2 µm.

https://doi.org/10.1371/journal.ppat.1014080.g001

Disruption of IMC1 results in severe morphological defects in vitro

In the original T. gondii genome-wide CRISPR screen, IMC1 was assigned a phenotype score of -4.0 [26], placing it among the lowest of the 14 alveolins. This highly negative score suggests essentiality; however, each gene must be individually assessed to determine whether it is truly essential. Thus, we attempted to disrupt IMC1 in RH∆ku80∆hxgprt parasites and were surprisingly able to generate a line that lacked IMC1 staining by immunofluorescence assays, which also showed that the intracellular parasites appeared rounded and swollen (∆imc1, Fig 2A). The IMC1 knockout was confirmed by PCR which showed an absence of the coding region and the presence of the selectable marker in the IMC1 locus in the knockout (Fig 2B). We then rescued the loss of expression by inserting a full-length IMC1 complementation construct driven by its endogenous promoter into the UPRT locus of the parasite (Fig 2C) [29]. We determined that the complemented protein localized similar to wild-type parasites by IFA (strain denoted IMC1^c, Fig 2D) and was also expressed at similar levels compared to the parental strain by western blot analysis (Fig 2E).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. IMC1 knockout and complementation, loss of IMC1 affects parasite morphology.

A) IFA of intracellular WT parasites showing proper localization of IMC1 (top). IFA of intracellular Δimc1 parasites showing the absence of IMC1 and swollen morphology (bottom, arrows). B) PCR verification of the genomic loci from WT (Rh∆hxgprt∆ku80) and ∆imc1 parasites. Diagram demonstrates primers used to amplify regions of the IMC1 coding sequence (blue arrows, coding region check) and the regions containing the 5’ and 3’ sites of recombination for the knockout locus (recombination check, red and green arrows). C) Schematic of the full-length IMC1 complementation construct driven by the endogenous IMC1 promoter. The full-length protein is 609 amino acids, with residues 98-422 (yellow) encompassing the alveolin domain. The five predicted palmitoylation sites with a CSS-Palm 4.0 cutoff score > 5 are shown (maroon) [36]. D) IFA of intracellular Δimc1 parasites expressing the IMC1 complementation construct, demonstrating restoration of IMC1 localization (IMC1^c strain). E) Western blots of whole cell lysates confirming similar protein levels expressed by the IMC1^c strain, as well as an absence of signal in the Δimc1 strain. Stained with anti-IMC1 with anti-catalase serving as a load control. All scale bars are 2 µm. WT, wild-type; CRC, Coding Region Check; 5’RC, 5’ Recombination Check; 3’RC, 3’ Recombination Check.

https://doi.org/10.1371/journal.ppat.1014080.g002

Disruption of IMC1 results in defects in the lytic cycle

To assess the effects of disrupting IMC1 on parasite growth, we first performed standard plaque assays (Fig 3A). We found that the ∆imc1 strain was dramatically hindered in the lytic cycle, displaying a severe ~85.7% reduction in plaque area, which was fully rescued by complementation (Fig 3B). We also observed a ~ 71.8% reduction in plaque efficiency, indicating that many of the knockout parasites suffered defects in the lytic cycle that resulted in lethality (Fig 3C). The plaque efficiency defect was again fully rescued in the IMC1^c parasites. This extreme decrease in plaque area and efficiency demonstrates that IMC1 is important for one or more stages of the lytic cycle.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. ∆imc1 parasites have severe defects in the lytic cycle.

A) Plaque assays of wild-type, ∆imc1, and IMC1^c parasites. Representative individual plaques are also shown. B) Quantification of plaque area, showing a severe defect in growth in the ∆imc1 strain. C) Quantification of plaque efficiency. D) Quantification of invasion assay, showing % invaded. ** P = 0.0011. E) Quantification of the egress assay, showing percent total vacuoles egressed. F) Quantification of the motility assay, showing % motile parasites. G) Quantification of the motility assay, showing average trail lengths. H) Quantification of the replication assay, showing the breakdown of parasites per vacuole for the three strains. B-G) Statistical analysis was performed using two-tailed t tests. H) Statistical analysis was performed using the Kolmogorov-Smirnov test. ** P = 0.0011. *** P = 0.0007. **** P < 0.0001.

https://doi.org/10.1371/journal.ppat.1014080.g003

To determine which stages of the lytic cycle were disrupted by the loss of IMC1, we individually examined invasion, replication and egress. We first carried out invasion assays in which we allowed wild-type, ∆imc1, and IMC1^c parasites to settle on a monolayer in invasion-restrictive media, then altered the conditions to invasion permissive media [30]. We then quantified intracellular and extracellular parasites and found that ∆imc1 parasites showed a severe failure to invade, with only 12 ± 3% being found intracellular, compared to the wild-type or IMC1^c strains, with 64 ± 8% and 65 ± 5% intracellular, respectively (Fig 3D). We also examined the efficiency of parasite egress in the IMC1 knockout using ionophore-induced egress assays [31]. We allowed the three strains to form large vacuoles and then induced egress using the calcium ionophore A23187 and observed that 56 ± 15% of the wild-type and 53 ± 13% of the IMC1^c vacuoles were induced to egress, while only 12 ± 7% of the ∆imc1 were (Fig 3E). To test whether parasite motility was impacted by absence of IMC1, trail deposition assays were performed [32]. We observed a decrease in motile parasites between the wild-type and ∆imc1 strains, with an average of 35 ± 12% and 32 ± 13% of wild-type and IMC1^c parasites, respectively, whereas only 12 ± 7% of ∆imc1 parasites formed trails (Fig 3F). We additionally saw a decrease in trail length between strains, with average measurements of 24 ± 13 and 23 ± 14 µm for the wild-type and IMC1^c parasites and 11 ± 7 µm for ∆imc1 (Fig 3G). These results indicate that motility likely plays a key role in the defects seen in the ∆imc1’s ability to undergo invasion and egress.

We then performed replication assays where we allowed the strains to grow for 30 hours [25]. We found that the majority of wild-type and complemented parasites had replicated intracellularly to 16 parasites per vacuole, with an average of 44 ± 1% and 44 ± 2% respectively. The ∆imc1 strain, on the other hand, had only replicated mostly to four-parasite vacuoles each, with an average of 35 ± 3% of total vacuoles being at this stage (Fig 3H). While this difference was not statistically significant, a trend towards a decrease in the knockout was apparent. The deficiencies in invasion, egress, and replication fidelity indicate that the growth defects caused by the absence of IMC1 are due to a failure to efficiently undergo the lytic cycle at all stages.

Disruption of IMC1 results in defects in endodyogeny and parasite morphology

To determine exactly how the loss of IMC1 impacts replication, we examined intracellular parasites for defects in morphology and daughter cell formation. We observed multiple defects in endogeny, including asynchronous replication, maternal parasites with more than two daughter buds (multidaughters), disorganized vacuoles, breaks in the IMC network, and incomplete septation (Fig 4A). We quantified the prevalence of these defects in intracellular parasites which strikingly resulted in an average of 99.4 ± 0.3% of the ∆imc1 vacuoles having one or more of these defects while only 2.2 ± 0.3% and 2.4 ± 0.3% of wild-type and IMC1^c vacuoles had any defects, respectively (Fig 4B). We also noticed that, in extracellular parasites, many individual knockout parasites retained the “incomplete septation” defect, with 30 ± 1% of all ∆imc1 parasites having more than one apical cap, as seen by ISP1 staining, compared to 0.0 ± 0.0% and 0.3 ± 0.4% of wild-type and IMC1^c respectively (Fig 4C and 4D).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. ∆imc1 displays severe defects in endodyogeny and parasite morphology.

A) IFAs of representative images of five different abnormalities in endodyogeny seen in ∆imc1 parasites. Magenta, anti-ISP1; Green, anti-IMC6. B) Quantification of vacuoles in WT, ∆imc1, and IMC1^c strains that display at least one or more of the endodyogeny abnormalities shown in Fig 4A. C) IFA of extracellular ∆imc1 parasites showing multiheaded parasites (arrowheads). Magenta, anti-ISP1; Green, anti-IMC6. D) Quantification of extracellular parasites from WT, ∆imc1, and IMC1^c strains that are multiheaded. E) Phase contrast image of extracellular wild-type and ∆imc1 parasites demonstrating extreme rounding in the knockout. F) Representative images captured by ImageStream flow cytometry of WT, ∆imc1 and IMC1^c parasites. The parasites were stained with anti-IMC6 in green. G) ImageStream analysis of population aspect ratios. Higher aspect ratio values indicate parasite roundness. H) ImageStream analysis of population circularity. Higher circularity values indicate more circular cells. I) ImageStream analysis of population elongatedness. Higher elongatedness values indicate more elongated cells. **** P < 0.0001. Statistical significance was determined using two-tailed t tests. All scale bars are 2 µm.

https://doi.org/10.1371/journal.ppat.1014080.g004

Examination of extracellular ∆imc1 parasites by phase contrast microscopy showed severe morphological changes, which was observed as a loss of their normal crescent shape and the formation of strikingly rounded parasites (Fig 4E). To quantify these shape defects, we used ImageStream flow cytometry to image ~10,000 extracellular parasites of wild-type, ∆imc1, and IMC1^c that were stained with antibodies to IMC6 to label the periphery of the parasites (Fig 4F) [33]. We specifically examined changes in aspect ratio, circularity, and elongatedness and found that ∆imc1 parasites had an average aspect ratio value of 0.93, compared to 0.63 for both the wild-type and IMC1^c strains (Fig 4G). A value closer to 1 suggests a rounder cell, indicating that ∆imc1 displays a significantly rounder morphology. Similarly, ∆imc1 parasites had an average circularity score of 9.60 while the wild-type and IMC1^c strains parasites had scores of 5.24 and 5.26, respectively (Fig 4H). Finally, we observed ∆imc1 parasites were less elongated than the wild-type and IMC1^c strains (Fig 4I). These results show that the loss of IMC1 results in rounder and shorter parasites that have substantial defects in endodyogeny.

Disruption of IMC1 in type II parasites disrupts the acute and chronic infection in vivo

To determine how the loss of IMC impacts the acute and chronic infection in vivo, we disrupted IMC1 in the Prugnaiud (Pru∆hxgprt∆ku80) strain of T. gondii, which also contains a GFP reporter that is driven by the bradyzoite-specific LDH2 promoter for assessing switching (S2A Fig) [34,35]. The knockout was verified by PCR (S2B Fig) and the resulting ∆imc1_II strain showed similar morphological defects as the type I knockout (S2A Fig). To verify that ∆imc1_II parasites retained their ability to form cysts, we performed an in vitro bradyzoite switching assay, in which parasites are induced to switch via culturing in a high pH, low CO₂ environment. The proportion of bradyzoite cyst formation in wild-type and ∆imc1_II parasites was quantified, which confirmed the knockout strain’s cyst forming ability (Fig 5A). In fact, we found that the ∆imc1_II strain was more able to switch in vitro with an average of 90 ± 19% of ∆imc1 vacuoles becoming GFP⁺ while only 48 ± 30% of the wild-type strain were GFP⁺ (Fig 5B). This increase in pH driven switching efficiency may be the result of increased stress already occurring in the knockout strain.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. ∆imc1_II fails to develop a chronic infection in vivo.

A) IFA of in vitro switched GFP⁺ parasite tissue cysts in both WT (top) and ∆imc1_II (bottom) parasites. Magenta, anti-IMC6. B) Quantification of the percentage of GFP⁺ vacuoles after being induced to switch at 93 hpi. C) Time course of average weights of mice infected with WT (Pru∆hxgprt∆ku80) parasites (blue circles) and ∆imc1_II parasites (red squares) through the duration of the infection. E) Quantification of brain cyst burden of mice infected with WT and ∆imc1 parasites at 30 days post infection. Cyst burden value was normalized for brain weights. Significance was determined using two-tailed t tests. *** P = 0.0003. ** P = 0.0086. **** P < 0.0001. All scale bars are 2 µm. hpi, hours post infection.

https://doi.org/10.1371/journal.ppat.1014080.g005

To assess the ∆imc1_II strain’s ability to infect and form cysts in vivo, we infected 5 mice each with an average of 114 and 1235 plaque-forming units of wild-type and ∆imc1_II parasites, respectively. The mice infected with the wild-type parasites showed a typical weight loss pattern of infection while those infected with the knockout continually gained weight (Fig 5C). After 30 days, the mice were euthanized and serum and brains were collected for serology and cyst enumeration. We found that the mice infected with wild-type parasites resulted in an average of 520 cysts per brain, while no cysts could be found in those infected with ∆imc1_II parasites (Fig 5D). In addition, we attempted to infect fibroblasts with 50% of the brain homogenate and parasites could not be recovered. We also probed western blots of wild-type parasite lysates with serum from each mouse and found no reactivity from the knockout while sera from the wild-type infection showed a typical signal of infection (S2C and S2D Fig). To determine if higher doses were better able to establish an infection, we injected three mice with ~200,000 pfu of the ∆imc1_II strain and found that the mice were able to seroconvert but again no cysts could be observed or parasites recovered from the brains at 30 days post infection (S2E Fig). Taken together, this indicates that ∆imc1_II parasites are capable of forming tissue cysts in vitro, but are cleared early in infection and fail to establish a chronic infection in vivo.

Deletion and mutation analyses elucidates functional domains of IMC1

IMC1 is composed of a conserved alveolin domain in the center of the protein that is flanked by N and C terminal extensions that contain predicted palmitoylation sites [36]. Previous work from our lab and others have shown that the alveolin domain is important for proper protein trafficking and function in the alveolins that have been studied [23,25]. To identify regions of IMC1 that are necessary for localization and function, we first assessed the predicted palmitoylation sites at residues C9, C12, and C13 by mutating them from cysteines to alanines in our full-length IMC1 complementation construct (Fig 6A). We then generated an N-terminal deletion series of IMC1, guided by secondary structure and homology to Plasmodium spp. (S3A and S3B Fig) [37]. Each of the mutants were targeted to the UPRT locus, as was done for the IMC1^c strain. IFA analysis showed that these mutants all localize to the IMC, although each of the mutants showed a slight enrichment in the apical cap. These constructs also generally appear to rescue the swollen morphology of the knockout (Fig 6B). Plaque assays showed that the growth defects caused by the knockout are largely rescued by each of the mutants, demonstrating that the N-terminal half of the protein is mostly dispensable (Fig 6C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. N-terminal mutation and deletion series of IMC1.

A) Schematic for mutants created from the full length N-terminal 3xV5 IMC1 complementation construct (IMC1^c). IMC1^C9,12,13A, IMC1^∆2-84, IMC1^∆2-217, and IMC1^∆2-312. B) IFA of the IMC1 mutants expressed in the ∆imc1 strain. Each of the constructs targets similar wild-type to the IMC. Magenta, anti-V5; Green, anti-IMC6. C) Quantification of plaque size for each ∆imc1 line containing one of the mutant copies of IMC1. IMC1 mutants largely rescue growth defects seen in ∆imc1 parasites. Statistical significance was determined using two-tailed t tests. **** P < 0.0001. ** P = 0.0063. * P = 0.0131. All scale bars are 2 µm.

https://doi.org/10.1371/journal.ppat.1014080.g006

Since these mutants removed approximately two-thirds of the alveolin domain, we wanted to determine if we could remove the entire alveolin domain before losing proper IMC1 function (IMC1^∆2-423, Fig 7A). We found that deletion of the N-terminal region including the alveolin domain fails to target to the IMC and mislocalizes to the cytoplasm (Fig 7B). We also assessed the growth of these parasites and found that this construct was unable to rescue the ∆imc1 strain growth defects (Fig 7C). To determine if residues 313–423 are sufficient for targeting and function, we expressed just this region of the protein in the ∆imc1 strain and found that this portion localizes properly but the parasites retain the swollen morphology shared by the knockout (IMC1^313-423, Fig 7D and 7E). We then generated a mutant that contains all of IMC1 except this region (IMC1^∆313-423, Fig 7F) and found that this deletion partially targets to the IMC, but also localizes to the cytoplasm (Fig 7G). When assessing the growth of ∆imc1 parasites with these mutants, we found that the IMC1^313-423 strain formed slightly larger plaques compared to the ∆imc1 strain but still had a sharp defect compared to wild-type parasites (Fig 7H). In contrast, IMC1^∆313-423 was unable to rescue the defects of the ∆imc1 strain. Taken together, these results demonstrate that IMC1^313-423 is sufficient for IMC targeting and is required for function of the full-length protein.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Deletion analyses identify regions of IMC1 important for targeting and function.

A) Schematic for the N-terminal deletion removing the entire alveolin domain (IMC1^∆2-423). B) IFA showing the IMC1^∆2-423 protein is mislocalized to the cytoplasm of the parasite and ∆imc1 parasite rounding is retained. C) Quantification of plaque area showing the IMC1^∆2-423 strain does not rescue growth. D) Schematic for the IMC1^313-423 mutant which removes all but the C-terminal portion of the alveolin domain. E) IFA of the IMC1^313-423 mutant within the ∆imc1 background showing correct localization but parasite rounding retained. F) Schematic of the IMC1^∆313-423 mutant that removes the C-terminal portion of the alveolin domain. G) IFA of the IMC1^∆313-423 mutant within the ∆imc1 background showing partial IMC localization and parasite rounding. H) Quantification of the plaque area for the IMC1^∆313-423 and IMC1^313-423 strains. I) Schematic of the IMC1^∆574-609 construct. J) IFA of the IMC1^∆574-609 strain showing proper localization of IMC1 lacking its C-terminal region and that parasite morphology appears restored. K) Quantification of plaque area showing the IMC1^∆574-609 mutant restores the growth defects of ∆imc1 parasites. Significances were determined using two-tailed t tests. **** P < 0.0001. ** P = 0.0010. Magenta, anti-V5; Green, anti-IMC6. All scale bars are 2 µm.

https://doi.org/10.1371/journal.ppat.1014080.g007

Previous work has shown that the C-terminal region of IMC1 contains predicted palmitoylation sites and that this region is proteolytically processed, an event which is important for incorporation into the cytoskeletal network of the IMC [28]. To determine if this region is important for trafficking and function in ∆imc1 parasites, we deleted the C-terminal 35 amino acids that encompass the palmitoylation sites from our full-length complementation construct and expressed this deletion mutant in the knockout (Fig 7I). We found that the protein targets to the IMC and mostly rescues the knockout (Fig 7J and 7K). This data shows that this region is largely dispensable for targeting and function of IMC1.

Previous work on IMC1 demonstrated that mutagenesis of a cysteine residue in the C-terminal portion of the protein (C593) was sufficient to disrupt processing and incorporation into the network [28]. Because our C-terminal deletion might be mimicking this cleavage event, we recreated this mutant in our full-length complementation construct (IMC1^C593T) and expressed the protein in ∆imc1 parasites. We found that the IMC1^C593T mutant trafficked properly to the IMC, but only partially rescued the growth phenotype (S5A and S5B Fig). However, we were unable to visualize any migration difference in IMC1^C593T parasites by western blot and the protein was fully incorporated into the cytoskeleton as shown by detergent fractionation (S5C and S5D Fig). This data indicates that the IMC1^C593T mutation does not affect IMC1 trafficking or cytoskeletal association, but it does diminish IMC1 function.

Disruption of IMC1 weakens the stability of the parasite cytoskeleton

To determine whether the absence of IMC1 impacts the stability of other alveolins in the cytoskeleton, we performed detergent fractionation experiments to separate the membrane and cytoskeletal elements of the parasite [38,39]. Wild-type and ∆imc1 strains were individually extracted with 1% Triton X-100, and the solubilized membrane and insoluble cytoskeletal fractions were run on western blots probing for ISP3, a marker of the membrane fraction, and the alveolins IMC3, IMC4, IMC6, and IMC10 for the cytoskeleton. We found that in wild-type parasites, IMC3, IMC4, IMC6, and IMC10 remain in the insoluble pellet while ISP3 is released into the supernatant, as expected (Fig 8A) [39]. However, in the ∆imc1 strain, IMC3, IMC4, IMC6, and IMC10 can be seen in both the supernatant and pellet fractions, suggesting that the lack of IMC1 results in the partial disassociation of these proteins off the cytoskeleton when exposed to detergent (Fig 8B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Loss of IMC1 results in heightened cytoskeletal sensitivity to detergent.

A, B) Detergent fractionation of A) wild-type and B) ∆imc1 parasites, demonstrating that IMC3, IMC4, IMC6, and IMC10 are present in both soluble and insoluble fractions of ∆imc1 parasites treated with 1% Triton X-100. C) IFA of detergent extraction of wild-type (left), ∆imc6 (middle), ∆imc1 (right) parasites demonstrating the change in detergent sensitivity between ∆imc1 and ∆imc6 strains after being treated with 1% Triton X-100 + 1% DOC solution. DOC, deoxycholate. D) TEM images of negative-stained, detergent-extracted wild-type (left), ∆imc6 (middle), ∆imc1 (right) parasites showing the ∆imc1 strain SPN is more disrupted by detergent extraction. Magenta, anti-SAG1; Green, anti-PolyE. T, total cell lysate. S, detergent soluble fraction. P, detergent insoluble fraction. Scale bars for IFAs are 2 µm. Scale bars for TEM are 0.5 µm.

https://doi.org/10.1371/journal.ppat.1014080.g008

We next wanted to determine how sensitive the ∆imc1 parasites were to detergents in general and also compare this to our previous ∆imc6 strain [25]. We thus treated extracellular wild-type, ∆imc1, and ∆imc6 parasites with a solution of 1% Triton X-100 and 1% deoxycholate and stained for tubulin for the cytoskeleton and SAG1 for the plasma membrane. We observed that this treatment resulted in the expected ghosts in both wild-type and ∆imc6 parasites, in which the cytoskeleton remained relatively intact and SAG1 was released, although the ∆imc6 strain was misshapen as we previously observed [25]. However, using these conditions, ∆imc1 parasites were clearly more disrupted, often forming umbrella-like structures with the microtubules splayed outwards, suggesting that the cytoskeleton of ∆imc1 is more sensitive to detergent extraction (Fig 8C). We also examined these strains by transmission electron microscopy and saw similar results with the ∆imc1 strain being more readily disrupted upon detergent extraction (Fig 8D).

IMC4 is mislocalized in ∆imc1 parasites and IMC1 and IMC4 directly interact

Given the morphological changes and increased sensitivity to detergent extraction in ∆imc1 parasites, we next examined whether a series of other alveolins (IMC3, IMC4, IMC6, IMC7, IMC10, and IMC12) were potentially impacted upon loss of IMC1. While IMC3, IMC6, IMC7, IMC10 and IMC12 localize normally to the IMC (S4B–S4D Fig), IMC4 – which normally overlaps perfectly with IMC1 - is mislocalized exclusively in the forming daughter buds of replicating ∆imc1 parasites (Fig 9A and 9B). We then tested this with the IMC1^∆313-423 and IMC1^313-423 complemented strains as well and found that, while IMC4 is still mislocalized in the IMC1^∆313-423 strain (Fig 9C), its localization is restored to wildtype in the IMC1^313-423 parasites (Fig 9D). This indicates that residues 313–423 of IMC1 play an important role in IMC4 recruitment during daughter bud formation. Since IMC1 and IMC4 co-localize with each other in wild-type parasites, this defect in the absence of IMC1 demonstrates a coordination of IMC1 and IMC4 within the parasite cytoskeleton.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. IMC4 is mislocalized in ∆imc1 daughter buds and directly binds IMC1.

A) IFA of non-budding (top) and budding (bottom) wild-type parasites showing proper localization of IMC4. B) IFA of ∆imc1 parasites showing mislocalization of IMC4 in daughter buds. C) IFA of IMC1^∆313-423 showing IMC4 remains mislocalized in daughter buds. D) IFA of IMC1^313-423 demonstrating that residues 313-423 are sufficient for the association of IMC4 to the growing daughter bud IMC. E-I) Images of yeast strain growth on restrictive (left) and permissive media (right) demonstrating IMC1 and IMC4 interact. J) AlphaFold prediction of binding interaction between IMC1 and IMC4 (top) and confidence prediction (bottom). Red indicates higher confidence, blue indicates lower confidence, pLDDT score ranges from 20.8 to 92.3. Magenta, anti-IMC4; Green, anti-IMC6.

https://doi.org/10.1371/journal.ppat.1014080.g009

Based on these results, we wanted to determine if IMC1 binds directly to IMC4. To address this question, we performed pairwise yeast two-hybrid assays. To do this, the IMC1 coding region lacking the regions containing its predicted palmitoylation sites (IMC1^14-591) was cloned into the pP6 (GAL4 activation domain) prey vector and residues 13–314 of IMC4 were cloned into the pB27 (LexA) bait vector [16]. The constructs were co-transformed into yeast and growth was assessed on either permissive or restrictive media. We found that the IMC1/IMC4 transformed yeast grew on restrictive media, demonstrating a direct interaction (Fig 9E). To determine whether the smallest functional N-terminal deletion can bind IMC4 (see Fig 6A, IMC1^∆2-312), we cloned this portion into the bait vector and found that it could also interact with IMC4 (Fig 9F). We then generated an IMC1 construct that lacks the essential residues (aa313–423) and found that this also interacted with IMC4 (IMC1^∆313-423, Fig 9G). We also wanted to see if the fully mislocalized fragment of IMC1 could still bind IMC4, so we tested the binding of this deletion and found that this did not allow direct binding between IMC1 and IMC4 (IMC1^∆2-423, Figs 9I and S4E). Finally, to determine if the IMC1 essential domain is sufficient for this interaction, we expressed just this region (IMC1^313-423) and found that that it is indeed able to interact with IMC4 (Fig 9H). Autoactivation was tested and confirmed that none of these constructs autoactivated when paired against the corresponding empty vector (S4F Fig). Together, this data demonstrates that the IMC1^313-423 is sufficient for IMC4 interaction, but other regions of the protein are also able to interact with IMC4. To better understand how IMC1 and IMC4 may interact to form filament-like structures, we used AlphaFold to predict their potential organization [40]. This suggested that the beta sheets of the alveolin domain interact with one another, and the proteins are organized in a parallel fashion when forming the IMC network filaments (Fig 9J). In addition, the interwoven alveolin domains of IMC1 and IMC4 appear to fold in upon each other and form electrostatic interactions between beta sheets of the alveolin domains. While there are likely other binding partners that may change the overall structure of the proteins in the IMC network, the IMC4 localization in ∆imc1 parasites, yeast two-hybrid, and AlphaFold predictions all indicate a direct binding relationship between IMC1 and IMC4.

Ethics statement

Our protocol was approved by the University of California, Los Angeles Institutional Animal Care and Use Committee (Chancellor’s Animal Research Committee protocol: #2004–005). Mice were euthanized when the animals reached a moribund state and euthanasia was performed following AVMA guidelines.

T. gondii culture

Parental T. gondii RH∆hxgprt∆ku80 [50], PruΔku80Δhxgprt (ldh2GFP) [35], and subsequent strains were cultured on confluent monolayers of human foreskin fibroblasts (HFF) host cells at 37°C and 5% CO₂ in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 5% fetal bovine serum (Gibco), 5% Cosmic Calf serum (HyClone), and 1x penicillin-streptomycin-l-glutamine (Gibco). Strains were selected using media supplemented with either 50 µg/mL of mycophenolic acid/xanthine (MX) [51] or 5µg/mL of 5-Fluoro-5’-deoxyuridine (FUDR) [29].

Gene knockout

To generate a knockout of IMC1, a protospacer was designed to target an exon within IMC1 (TgGT1_231640) and ligated it into the pU6-Universal plasmid utilizing primers P1-2 (Table I) [52]. The homology-directed repair (HDR) template included 40 bp of homology immediately upstream of the start codon and 40 bp of homology approximately 200 bp downstream of the stop codon (P3-4). Along with an HXGPRT selectable marker driven by a NcGRA7 promoter, the HDR template was amplified from a pJET vector using P3-4. For transfection, the PCR-amplified HDR template was purified by phenol-chloroform extraction and this plus ~50 µg of the gRNA-carrying pU6-Universal plasmid were ethanol precipitated. Both constructs were electroporated into the RHΔhxgprtΔku80 or PruΔku80Δhxgprt (ldh2GFP) parental parasite strains. Transfected parasites were allowed to invade a confluent monolayer of HFFs overnight, and appropriate selection was subsequently applied. Successful knockout was confirmed by IFA, and clonal lines were obtained through limiting dilution.

IMC1 complementation

All IMC1 complementation constructs were modified from the previously generated IMC29 complementation construct [53]. Each construct contains either the full-length or truncated version of the IMC1 coding sequence and UPRT homology regions to drive this cassette to the UPRT locus. For this study, the IMC29 promoter was replaced with the IMC1 promoter using Gibson assembly. The IMC1 promoter was amplified from genomic DNA (P13-14). The full length IMC1 cDNA and the vector plasmid were amplified using P9-12. The product was purified and ligated using the NEBuilder HiFi DNA Assembly kit, resulting in the final plasmid. The full-length construct was also remade with an additional N-terminal 3xV5 epitope tag. To make truncation constructs, the same process was followed for each truncation using the tagged version: IMC1^∆2-84 was amplified using P24/P23, IMC1^∆2-217 was amplified using P25/P23, IMC1^∆2-312 was amplified using P26/P23, IMC1^∆2-423 was amplified using P27/P23, and IMC1^∆574-609 was amplified using P28/P29 off of the full-length constructs. Additionally, IMC1^313-423 was amplified using P28/P30 from IMC1^∆2-312 and IMC1^∆313-423 was amplified using P31/P32 from the full-length plasmid. To make the IMC1^C593T mutant, a similar process was followed using the full length untagged version and primers P33/P34. After plasmid construction, 100 µg of each complementation plasmid was linearized with PsiI-HF and transfected into Δimc1 parasites with a pU6-gRNA that targets the UPRT coding region. Selection was performed with 5ug/ml FUDR for replacement of UPRT and clones were screened by IFA [29].

Antibodies

The V5 epitope tag was detected using mouse mAb anti-V5 (catalog no. R96025; Invitrogen). T. gondii-specific antibodies used include mouse mAb 45.36 anti-IMC1 [54], pAb rabbit anti-IMC3 [55], rabbit pAb anti-IMC6 [22], rat pAb anti-IMC6 [22], rat pAb anti-IMC7 [24], rat pAb anti-IMC10 [27], rabbit pAb anti-IMC12 [56], mAb mouse anti-SAG1 DG52 [57], mouse mAb anti-MIC2 [58], and rabbit pAb anti-SAG2 [59], mouse mAb anti-ROP7 [60], mouse mAb anti-ISP1 [50], mouse mAb anti-F₁β subunit [61], rabbit pAb anti-CENTRIN1 (Kerafast; EBC004), rat pAb DrpB [62], mouse mAb anti-Atrx1 [63], mouse pAb anti-ISP3 [50], and rabbit pAb anti-catalase [64].

Immunofluorescence assay and western blot

For immunofluorescence assays, HFF cells were seeded on glass coverslips and allowed to grow until confluency. Parasites were then allowed to infect these monolayers and 18–36 hours post infection, the coverslips were fixed with 3.7% formaldehyde in PBS and processed for immunofluorescence as described [60]. Primary antibodies are added to blocking buffer (1x PBS, 5% BSA, 0.2% TX-100) and incubated on the coverslips for one hour, washed with PBS five times for two minutes each, species-specific secondary antibodies (Alexa Fluor 488/594) were incubated for one hour, and washed with PBS 5 times. Coverslips were then mounted on a microscope slide in VectaShield (Vector Labs, Burlingame, CA), viewed with an Axio Imager.M2 fluorescent microscope (Zeiss), and processed with Zeiss ZEN 3.7 software (Zeiss), which included deconvolution. Images with the 594-fluorophore channel were pseudocolored to magenta.

For western blotting, parasites were lysed using 1x Laemmli Sample Buffer (50 mM Tris-HCL [pH 6.8], 10% glycerol, 2% SDS, 100 mM DTT, and 0.1% bromophenol blue) and boiled for 10 minutes. Lysates were run on SDS-PAGE gels and transferred to nitrocellulose membranes. The membranes were blocked using 5% milk and 1xPBS supplemented with 0.1% Tween-20 mixture for half an hour before probing with the appropriate primary antibody and the corresponding secondary antibody conjugated to horse radish peroxidase (HRP). Probing was visualized on ChemiDoc XRS+ (Bio-Rad) using SuperSignal chemiluminescence substrate (Thermo Scientific). Seroconversion was determined using infection sera collected at 30 days post infection (diluted at 1:200).

Detergent fractionation

To separate parasites into detergent soluble and insoluble fractions, parasites were collected, washed with 1xPBS, and resuspended in 1% Triton X-100 supplemented with Complete Protease Inhibitor Cocktail (Roche) [38,39]. This was incubated on ice for 20 minutes before being centrifuged at 14,000g for 15 minutes to separate the sample. The supernatant was mixed with 2x Laemmli Sample Buffer and the pellet was resuspended in an equivalent volume of 1x Sample Buffer. The fractions were then visualized using SDS-PAGE and Western Blotting. Antibodies probing for IMC6 were used to detect the insoluble cytoskeletal fraction and ISP3 was used to indicate the solubilized membrane fraction.

Detergent extraction of extracellular parasites

For extracellular detergent extraction the parasites were washed in 1xPBS were first allowed to settle on coverslips coated in poly-L-lysine for 30 minutes. The remaining liquid was then aspirated and replaced with 1% Triton X-100 and 1% DOC and allowed to incubate for 30 minutes. The coverslips were then washed and fixed with 3.7% formaldehyde in 1xPBS for 15 minutes. The coverslips were then washed again, stained with primary and secondary antibodies, and visualized by fluorescence microscopy. For transmission electron microscopy, parasite ghosts were prepared using an on-grid extraction technique for wild-type, Δimc1 and Δimc6 parasites [25]. Briefly, 4 μl of parasite resuspension was applied onto a glow-discharged continuous carbon-coated EM grid. After allowing the sample to settle for 2 minutes, the excess buffer was blotted from the edge of the grid. The grid was then placed face-down onto a 20 μl drop of detergent extraction buffer (1% Triton X-100, 0.5% DOC, 50 mM Tris, 150 mM NaCl) and left to float for 5 minutes to lyse the parasites. After this initial extraction, the grid was blotted again and transferred to a fresh drop of the same buffer for another 5 minutes. This extraction step was repeated six times in total. Subsequently, the grid was washed four times using 20 μl drops of PBS, following the same procedure. Finally, the grid was stained with 2% uranyl acetate for negative staining and imaged using a FEI Tecnai T12 transmission electron microscope equipped with a CCD camera and operated at 120 kV.

Plaque assay

Six-well plates hosting HFF monolayers were allowed to become confluent. Then, equivalent number of parasites were allowed to infect the monolayer for seven days, forming plaques. The samples were fixed with ice-cold methanol and stained with crystal violet [24]. For each condition, the area of around 50 plaques were measured using ZEN blue software. Additionally, the number of plaques formed by the parasites added was quantified. All plaque assays were performed with three biological replicates. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

In vitro bradyzoite switching assay

Parasites were allowed to invade 24-well plates with confluent HFFs for three hours before changing media into a high pH media designed to induce in vitro switching [65]. The coverslip plate was then allowed to sit at 5% CO₂ and 37°C for three hours. The plate was then wrapped in parafilm to seal any remaining CO₂ inside of the wells and moved to an incubator at 0% CO₂ and 37°C for 12 hours. After 12 hours, the plate was returned to an incubator at 5% CO₂ and 37°C for one hour before repeating the process until 93 hours had elapsed. The coverslips were then probed with anti-IMC6 antibodies and the experiment was quantified by determining the percentage of vacuoles that had turned into cysts (GFP⁺) out of the total vacuoles in each field. Triplicate experiments were performed, with each replicate spanning at least 15 fields and 50 vacuoles. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

Mouse infections and brain cyst quantification

Wild-type Pru∆ku80∆hxgprt and ∆imc1 parasites were counted and resuspended in Opti-MEM prior to intraperitoneal injection into groups of five female CBA/J mice each. A total of 114 wild-type and 1235 ∆imc1 plaque forming units/mouse were used for the innoculum. For high dose infections, ~ 200,000 pfu were used of the ∆imc1 strain for infection. The mice were observed for 30 days and then sacrificed. Mouse brains were collected, homogenized, and examined for the presence of LDH2-GFP+ tissue cysts [34]. Quantification was performed by examining 25-µl aliquots of each brain homogenate using fluorescence microscopy until approximately 25% of the brain by volume was counted. Total cyst burden was then extrapolated. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

Replication defect quantification

HFFs were seeded on glass coverslips until fully confluent and were infected at low MOI, incubated for 3 hours at 37˚C, and extracellular parasites were washed away by media changes. Then, 30 hours post-infection, coverslips were fixed with 3.7% formaldehyde, processed for immunofluorescence, and labeled with anti-ISP1 and anti-IMC6. To score normal vs. abnormal vacuoles, > 300 vacuoles across at least 15 fields were counted. Vacuoles were categorized as abnormal if any one of the replication defects shown in Fig 4 were present in a vacuole. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

Invasion assay

Parasites in large vacuoles were mechanically lysed using a syringe and resuspended in invasion restrictive media and settled onto coverslips with a HFF monolayer for 20 minutes [30]. The invasion restrictive media was then replaced with warm growth media supplemented with 20 mM HEPES and incubated in 37°C and 5% CO₂ for 30 minutes. Coverslips were then fixed, blocked with non-permeabilizing 3% BSA in 1x PBS for 30 minutes. The extracellular parasites were stained with anti-SAG1 antibodies for an hour. The coverslips were washed with 1xPBS and permeabilized with 3% BSA, 0.2% Triton X-100, 1x PBS for 30 minutes, and all parasites were stained with anti-SAG2. Coverslips were washed with 1xPBS and mounted on microscope slides using the mounting medium VectaShield (Vector Labs) and viewed with an Axio Imager.M2 fluorescence microscope (Zeiss). Parasites were assigned with SAG1-,SAG2+ (Invaded) or SAG1 + ,SAG2+ (Uninvaded). Three biological replicates were performed with at least 300 parasites across 15 fields per triplicate. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

Egress assay

HFFs were infected at a low MOI and grown until vacuoles reached 16–32 parasites per vacuole. The media was then changed to warm HBSS -/ + 1uM A23187 ionophore [31]. Fixative was added to the coverslips after 4 minutes and the coverslips were stained with the rabbit anti-IMC12. Percentage of egressed vacuoles per total vacuoles was enumerated in triplicates with at least 100 vacuoles and 15 fields per replicate. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

Motility assay

Trail deposition assays were performed as previously described [32]. HFFs were infected at medium MOI and grown until large vacuoles. Parasites were then mechanically lysed and resuspended in warm HBSS with Calcium and Magnesium supplemented with 0.01M HEPES buffer. Parasites were allowed to settle on the surface of Poly-L-Lysine coated coverslips for 1 hour at 37°C. The coverslips were then fixed, blocked, and probed for SAG1 positivity using standard IFA protocol. Trails were imaged with the Axio Imager.M2 fluorescent microscope (Zeiss) and measured using Zeiss ZEN 3.7 software (Zeiss). Percentage of motile parasites was enumerated in triplicates with at least 250 parasites across 15 fields per replicate. Trail lengths were measured in triplicates with at least 25 trails per replicate. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

Pairwise yeast 2-hybrid assay

Using Gibson Assembly, genes of interest were cloned into pB27 (N-LexA-bait-C fusion) and pP6 (N-GAL4^AD-prey-C fusion) vectors containing either a N-terminal GAL4 Activating Domain or a LexA DNA Binding Domain. Smaller fragments were built using Q5 site-directed mutagenesis using primers (P35-47). Pairs of constructs containing both a pB27 and a pP6 were co-transformed into the L40 strain of S. cerevisiae [MATa his3D200trp1–901 leu2–3112 ade2 LYS2::(4lexAop-HIS3) URA3::(8lexAop-lacZ) GAL4]. Stains were allowed to grow overnight in liquid permissive media (-Leu/-Trp), re-diluted to OD₆₀₀ = 2 and then spotted in serial dilutions on permissive (-Leu/-Trp) and restrictive (-Leu/-Trp/-His) plates. After growing for 3–5 days, the plates were imaged. Auto-activation was tested for by co-transforming individual constructs with their corresponding empty vector into yeast and performing spot assays as described previously [16].

Replication defects by IFA

Parasites were allowed to invade HFFs seeded on a coverslip at a low MOI for 1 hour at 37°C, then remaining extracellular parasites were washed away. Coverslips were then fixed using 3.7% formaldehyde and stained with mouse anti-ISP1 and rabbit anti-IMC6. To score normal vs. abnormal vacuoles, > 300 vacuoles across at least 15 fields were counted. Vacuoles were categorized as abnormal if one or more of the replication defects shown in Fig 4A were present in a vacuole. All IFAs were performed in triplicate. For extracellular parasites, fully lysed wild-type, Δimc1, and IMC1^c strains were collected and washed in 1xPBS before being settled onto coverslips coated with poly-L-lysine (Sigma Aldrich). These coverslips were then stained using IFA with mouse anti-ISP1 and rabbit anti-IMC6, taking caution not to disturb the parasites. Quantifications were performed by counting three replicates of >100 parasites per strain. Statistical analysis, graphs, and figures were generated using Prism GraphPad 10.

ImageStream flow cytometry

Mechanically lysed extracellular parasites were stained using indirect immunofluorescence in solution. The parasites were fixed using 3.7% formaldehyde, permeabilized and blocked in solution. These parasites were then stained with rabbit anti-IMC6 and anti-rabbit Alexa Fluor 588. For each strain, around 10,000 individual images were taken by the INSPIRE application on the ImageStream MKII (Cytek Biosciences, Seattle, WA) flow cytometer. Images were acquired using a 6 μm width core and 60x magnification. Analysis of the samples were performed using the IDEAs v6.4 software (Cytek Biosciences, California). The aspect ratio of each cell was calculated using the ratio of the minor axis divided by the major axis. The circularity of each cell was measured by a cell’s degree of deviation from a circle, with higher values indicating higher circularity and lower values indicating a higher degree of variation from a circle. Elongatedness was calculated using the ratio of height over width of a cell, with lower values indicating less elongatedness.