Cytosolic FeS cluster assembly occurs at the mitochondrion of Toxoplasma gondii

We set out to determine the subcellular localization of the previously uncharacterized candidate CIA pathway proteins in T. gondii. Using CRISPR/Cas9 genome editing, we introduced a 3× hemagglutinin (HA) epitope tag into the 3′ end of the coding sequences of the proposed electron transfer chain proteins TgTah18 (www.ToxoDB.org gene identifier TGGT1_249320; [17]) and TgDre2 (TGGT1_216900), and all three of the proposed CTC proteins, TgCIA1 (TGGT1_313280), TgCIA2 (TGGT1_306590) and TgMMS19 (TGGT1_222230), in an anhydrotetracycline (ATc)-regulatable TgNBP35-cMyc strain that we had generated previously [15]. We validated correct integration of the tag in each parasite line by PCR analysis (S1A–S1E Fig). We were unsuccessful in integrating a HA tag into the 3′ end of the coding sequence for the proposed FeS transfer protein TgNar1 (TGGT1_242580). Instead, we introduced a HA-mini-Auxin Inducible Degron (HA-mAID) tag at the 5′ end of the coding region of TgNar1 in a tdTomato and Tir1-expressing parasite strain (RH∆ku80/Tir1-FLAG/tdTomato; [18]; S1F Fig). To test whether each protein was expressed in the disease-causing tachyzoite stage of the parasite life cycle, we extracted proteins from the TgTah18-HA, TgDre2-HA, HA-mAID-TgNar1, TgCIA1-HA, TgCIA2-HA, and TgMMS19-HA lines, separated them by Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE), and performed western blotting with anti-HA antibodies. All six proteins were expressed and were of approximately the expected masses for the epitope-tagged proteins (Fig 1A–1F).

Download:

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

Fig 1. The CIA pathway localizes to the mitochondrion of Toxoplasma gondii.

(A–F) Western blots of proteins extracted from (A) TgTah18-HA, (B) TgDre2-HA, (C) HA-mAID-TgNar1, (D) TgCIA1-HA, (E) TgCIA2-HA, or (F) TgMMS19-HA expressing parasites, separated by SDS-PAGE and probed with anti-HA antibodies. TgMMS19-HA appears as two separate bands, the lower of which may represent a degradation product. (G–L) Immunofluorescence assays of (G) TgTah18-HA, (H) TgDre2-HA, (I) HA-mAID-TgNar1, (J) TgCIA1-HA, (K) TgCIA2-HA, or (L) TgMMS19-HA expressing parasites, probed with anti-HA antibodies (green) to label the proteins-of-interest and anti-TgTom40 antibodies (magenta) to label the mitochondrion. Scale bars are 2 µm. DIC, differential interference contrast transmission image. Right, corresponding fluorescence profiles depicting the intensity of anti-HA (green) and anti-TgTom40 (magenta) labeling along the yellow line on merged images. The numerical data underlying this Figure can be found in S1 Data. (M) Model for the CIA pathway in T. gondii parasites, with Tah18, NBP35 and Nar1 positioned on the outer face of the mitochondrion, Dre2 in the cytosol, and the CIA Targeting Complex (CTC; consisting of CIA1, CIA2 and MMS19), exhibiting dual cytosolic and mitochondrial localization. Electrons (e⁻) are sourced from NADPH via an electron transfer chain consisting of Tah18 and Dre2. Sulfur (yellow) and possibly iron (red) are sourced from the iron–sulfur cluster (ISC) synthesis pathway in the mitochondrion, assembled as [4Fe-4S] clusters on the NBP35 scaffold, then trafficked via Nar1 to the CTC. The CTC donates FeS clusters to client FeS proteins.

https://doi.org/10.1371/journal.pbio.3003520.g001

To determine the subcellular localization of each protein, we performed immunofluorescence assays. We found that TgTah18-HA and HA-mAID-TgNar1 both co-localized with the mitochondrial marker TgTom40, whereas TgDre2-HA localized throughout the parasite cytosol (Fig 1G–1I). Curiously, all three proteins of the putative CTC exhibited a dual localization, overlapping with TgTom40 in the mitochondrion and also exhibiting diffuse labeling through the cytosol (Fig 1J–1L). Taken together, these data indicate that all the putative CIA pathway proteins in T. gondii, with the exception of TgDre2 but including the previously characterized scaffold protein TgNBP35 [15,16], localize to the mitochondrion. We showed previously that TgNBP35 localizes to the cytosolic face of the outer mitochondrial membrane [15], and our data are therefore consistent with the CIA pathway occurring on the cytosolic face of the outer mitochondrial membrane in T. gondii (Fig 1M).

The CIA pathway is critical for parasite proliferation and protein synthesis in Toxoplasma gondii

We set out to test the importance of each of the candidate CIA pathway proteins for parasite proliferation. Using CRISPR/Cas9 genome editing, we introduced mAID-HA epitope tags at the 3′ ends of the coding regions for TgTah18, TgDre2, TgCIA1, TgCIA2 and TgMMS19 of tdTomato/Tir1 parasites (S2 Fig). The mAID tagging system facilitates proteasomal degradation of mAID-fused proteins upon the addition of the small molecule 3-indoleacetic acid (IAA; [19]). Since the abundance of the resulting proteins can be conditionally regulated, we termed the resulting lines expressing mAID-tagged CIA pathway proteins regulatable (r)-TgTah18-mAID-HA, rTgDre2-mAID-HA, rHA-mAID-TgNar1, rTgCIA1-mAID-HA, rTgCIA2-mAID-HA, and rTgMMS19-mAID-HA. To determine whether the mAID-tagged proteins could be regulated upon the addition of IAA, we cultured each parasite line in the absence or presence of IAA for 24 h. We extracted the resulting proteins, separated them by SDS-PAGE and performed western blotting. We found that the abundances of all six candidate CIA pathway proteins were substantially depleted upon IAA addition (Fig 2A–2F).

Download:

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

Fig 2. Candidate CIA pathway proteins are important for parasite proliferation.

(A–G) Fluorescence proliferation assays and western blotting of (A) rTgTah18-mAID-HA, (B) rTgDre2-mAID-HA, (C) rHA-mAID-TgNar1, (D) rTgCIA1-mAID-HA, (E) rTgCIA2-mAID-HA, (F) rTgMMS19-mAID-HA, and (G) WT (RH∆ku80/Tir1-FLAG/tdTomato) parasites cultured in the absence (black) or presence (red) of IAA. Parasite proliferation is expressed as a percentage of the final fluorescence measurement in the -IAA condition for each line. Individual data points and error bars represent the mean ± SD of three technical replicates. Data are representative of three independent experiments. Error bars not visible are smaller than the symbol. The numerical data underlying this Figure can be found in S1 Data. For the western blotting, parasites were cultured in the absence (−) or presence (+) of IAA for 24 h, with protein samples separated by SDS-PAGE then subjected to western blotting with anti-HA antibodies to detect the mAID-HA-tagged protein of interest and anti-TgTom40 antibodies as a loading control (A–F, left).

https://doi.org/10.1371/journal.pbio.3003520.g002

To explore the contribution of each CIA pathway protein to parasite proliferation, we performed fluorescence-based proliferation assays, measuring tdTomato fluorescence in parasites across time as a function of parasite proliferation. We found that in the absence of IAA, all parasite lines proliferated normally, exhibiting a typical sigmoidal growth curve (Fig 2). The parental RH∆ku80/Tir1-FLAG/tdTomato (WT) line proliferated normally in the presence of IAA, indicating that IAA is not toxic to parasites (Fig 2G). By contrast, the rTgTah18-mAID-HA, rTgDre2-mAID-HA, rHA-mAID-TgNar1, rTgCIA1-mAID-HA, rTgCIA2-mAID-HA, and rTgMMS19-mAID-HA lines all exhibited a severe proliferation defect when cultured in the presence of IAA (Fig 2A–2F), with the phenotype upon the depletion of TgNar1 consistent with that reported in another recent study [20]. These data indicate that TgTah18, TgDre2, TgNar1, TgCIA1, TgCIA2 and TgMMS19 are all critical for parasite proliferation.

Next, we wanted to explore whether the candidate proteins function in the CIA pathway. Given its curious dual localization to the mitochondrion and cytosol, we initially focused on the CTC protein TgCIA1. Previous studies have found that impairment of the cytosolic FeS assembly pathway in T. gondii leads to a depletion in the abundances of some cytosolic and nuclear FeS cluster-containing proteins, including the ribosome recycling protein TgABCE1 [15,20–22]. We integrated a 3× cMyc tag at the 3′ end of the TgABCE1 open reading frame in the rTgCIA1-mAID-HA line using CRISPR/Cas9 genome editing (S3A and S3B Fig). We cultured these parasites for 0, 24, or 48 hours on IAA, separated proteins by SDS-PAGE, and performed western blotting to detect the TgABCE1-cMyc protein. We found a small but significant decrease in TgABCE1-cMyc abundance after 24 h of IAA incubation and a strong ~75% decrease after 48 h (Fig 3A and 3B).

Download:

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

Fig 3. Characterizing the effects of depleting candidate CIA pathway proteins in Toxoplasma gondii parasites.

(A) Western blots of proteins extracted from rTgCIA1-mAID-HA/TgABCE1-cMyc parasites cultured for 0, 24 or 48 h in IAA and separated by SDS-PAGE. Samples were probed with anti-HA, anti-cMyc, and anti-TgTom40 antibodies. (B) Relative abundance of the TgABCE1-cMyc protein in the western blot was determined as a percentage of the 0 h control, with abundances normalized using the TgTom40 loading control. Data points represent the mean ± SD of five independent experiments for the 0 and 24 h conditions and three independent experiments for the 48 h condition. Data were analyzed using a one-way ANOVA followed by Tukey’s multiple comparisons test, with p values shown. (C) Western blots measuring the incorporation of puromycin into proteins from WT (RH∆ku80-Tir1-FLAG/tdTomato) or rTgCIA1-mAID-HA parasites cultured in the absence (−) or presence (+) of IAA for 24 h, separated by SDS-PAGE and probed with anti-puromycin, anti-HA, and anti-TgTom40 antibodies. (D) Relative abundance of puromycin incorporation was determined as a percentage of the -IAA control for each parasite line, with abundances normalized using the TgTom40 loading control. Data points represent the mean ± SD of three independent experiments. Data were analyzed using a one-way ANOVA followed by Tukey’s multiple comparisons test, with relevant p values shown. (E) Multiple sequence alignment of a region of the T. gondii CIA2 protein (TgCIA2) with CIA2 homologs in Saccharomyces cerevisiae (ScCIA2), Homo sapiens (HsCIA2), and Drosophila melanogaster (DmCIA2). The cysteine residue proposed to function in FeS cluster binding is highlighted by a red box. (F) Western blot of proteins extracted from rTgCIA2-mAID-HA/cTgCIA2_WT-Ty1 or rTgCIA2-mAID-HA/cTgCIA2_C524A-Ty1 parasites, separated by SDS-PAGE and probed with anti-Ty1 antibodies to detect the cTgCIA2_WT-Ty1 and cTgCIA2_C524A-Ty1 proteins, and anti-TgTom40 antibodies as a loading control. (G) Fluorescence proliferation assays of rTgCIA2-mAID-HA, rTgCIA2-mAID-HA/cTgCIA2_WT-Ty1, or rTgCIA2-mAID-HA/cTgCIA2_C524A-Ty1 parasites cultured in the absence (black) or presence (red) of IAA. Parasite proliferation is expressed as a percentage of the -IAA condition on the final day of the experiment for each parasite line, with values depicting the mean ± SD of three technical replicates. Error bars not visible are smaller than the symbol. Data are representative of three independent experiments. The numerical data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003520.g003

ABCE1 and other cytosolic FeS proteins are important for protein translation from cytosolic ribosomes. Previous studies have found that depletion of proteins that contribute to the CIA pathway result in a global impairment of protein synthesis in T. gondii [20,21]. We cultured WT and rTgCIA1-mAID-HA parasites in the absence or presence of IAA for 24 h, then treated parasites with the protein synthesis inhibitor puromycin for 15 min. Puromycin becomes incorporated into nascent polypeptides during translation and can be detected by anti-puromycin antibodies, thus providing a measure for newly synthesized proteins. We separated proteins from puromycin-treated parasites using SDS-PAGE and performed western blotting. We observed a significant ~80% depletion in puromycin-labeled proteins in IAA-treated rTgCIA1-mAID-HA parasites, whereas puromycin labeling was unchanged in IAA-treated WT parasites (Fig 3C and 3D). We next performed puromycin incorporation assays in rTgTah18-mAID-HA, rTgDre2-mAID-HA, rHA-mAID-TgNar1, rTgCIA2-mAID-HA and rTgMMS19-mAID-HA parasites cultured for 24 h in the absence or presence of IAA. We observed significantly less puromycin incorporation upon the depletion of TgTah18-mAID-HA, HA-mAID-TgNar1, TgCIA2-mAID-HA and TgMMS19-mAID-HA but not upon the depletion of TgDre2-mAID-HA (S4A and S4B Fig).

To test whether TgDre2 knockdown leads to the depletion of TgABCE1, we integrated a cMyc epitope tag into the TgABCE1 locus of rTgDre2-mAID-HA parasites (S3C Fig). We cultured these parasites for 0, 24, or 48 hours on IAA, separated proteins by SDS-PAGE, and performed western blotting to detect the TgABCE1-cMyc protein. We found a ~40% decrease in TgABCE1-cMyc abundance after 24 h of IAA incubation and a ~60% decrease after 48 h (S4C Fig). Loss of TgDre2, therefore, does impact the abundance of the cytosolic Fe–S protein TgABCE1, leaving open the possibility that TgDre2 has a role in cytosolic Fe–S synthesis. Taken together, our data indicate that the putative CIA pathway proteins of T. gondii are essential for parasite proliferation, and that most of the candidate CIA pathway proteins in T. gondii (with the possible exception of TgDre2) are critical for protein translation, a process that relies on cytosolic FeS proteins like ABCE1.

The CIA2 protein in animals and fungi contains a reactive cysteine residue that is thought to function in FeS cluster binding and facilitating the transfer of FeS clusters to client proteins [10,23,24]. This cysteine residue has been shown to be critical for CIA pathway function and cell survival in these eukaryotes [23,24]. Alignments of yeast and animal CIA2 proteins with TgCIA2 revealed that the reactive cysteine residue is conserved as residue 524 in the TgCIA2 protein (Figs 3E and S5). We hypothesized that if TgCIA2 functions like its counterpart in other eukaryotes, mutating the TgCIA2_C524 residue would ablate CIA pathway function and, subsequently, lead to impaired parasite proliferation. We constitutively expressed either Ty1 epitope-tagged WT TgCIA2 (constitutive (c)TgCIA2_WT-Ty1) or a Ty1-tagged TgCIA2 mutant wherein the cysteine at residue 524 was mutated to Ala (cTgCIA2_C524A-Ty1) in rTgCIA2-mAID-HA parasites. Western blotting revealed that both Ty1-tagged TgCIA2 proteins were expressed at similar levels (Fig 3F) and fluorescence proliferation assays revealed that expression of cTgCIA2_WT-Ty1 fully rescued parasite proliferation upon knockdown of the TgCIA2-mAID-HA protein (Fig 3G). By contrast, we observed minimal proliferation of rTgCIA2-mAID-HA/cTgCIA2_C524A-Ty1 parasites cultured in the presence of IAA (Fig 3G). These data indicate that the proposed FeS cluster binding Cys-524 residue of TgCIA2 is critical for its function, consistent with the hypothesis that the TgCIA2 protein of T. gondii functions in a similar manner to the CIA2 protein of other eukaryotes in coordinating FeS cluster binding on the CTC [10,23,24].

TgCIA1, TgCIA2, and TgMMS19 are components of the same protein complex

We set out to explore the candidate CTC proteins of T. gondii in more detail. First, we asked whether the TgCIA1, TgCIA2 and TgMMS19 proteins exist in a protein complex. We extracted proteins from TgCIA1-HA, TgCIA2-HA and TgMMS19-HA expressing parasites, separated them by blue native (BN)-PAGE, and performed western blotting with anti-HA antibodies. All three proteins were present in a complex of slightly larger than 720 kDa (Fig 4A, black arrowhead). The TgCIA1-HA protein, but not the others, was also observed in smaller protein complexes of ~400 and ~240 kDa (Fig 4A, red and blue arrowheads).

Download:

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

Fig 4. Characterizing the protein composition of the CIA Targeting Complex (CTC) in Toxoplasma gondii parasites.

(A) Western blot of proteins extracted from TgCIA1-HA, TgCIA2-HA and TgMMS19-HA expressing parasites, separated by BN-PAGE and probed with anti-HA antibodies or anti-TgTom40 as a loading control. (B, D) Western blots of proteins extracted from rTgCIA1-mAID-cMyc/TgCIA2-HA and rTgCIA1-mAID-cMyc/TgMMS19-HA parasites cultured in the absence (-) or presence (+) of IAA for 24 h, separated by either SDS-PAGE (B) or BN-PAGE (D) and probed with anti-HA, anti-cMyc or anti-TgTom40 antibodies. Data are representative of three independent experiments. (C) Quantification of the western blots depicted in (B), depicting relative abundance of the TgCIA2-HA or TgMMS19-HA proteins determined as a percentage of the -IAA control, with abundances normalized using the TgTom40 loading control. Data points represent the mean ± SD of three independent experiments. Data were analyzed using a paired t test with p values shown. The numerical data underlying this Figure can be found in S1 Data. (E, F) Western blot of proteins extracted from rTgCIA2-mAID-HA/TgCIA1-smFP-cMyc and rTgMMS19-mAID-HA/TgCIA1-smFP-cMyc parasites, cultured for 24 h in the absence (−) or presence (+) of IAA, separated by either (E) SDS-PAGE or (F) BN-PAGE and probed with anti-cMyc, anti-HA, or anti-TgTom40 antibodies. Data are representative of three independent experiments. In the BN-PAGE western blots (A, D, and F), the black arrowheads indicate the >720 kDa candidate CIA Targeting Complex; red and blue arrowheads indicate the lower mass complexes containing the TgCIA1 protein.

https://doi.org/10.1371/journal.pbio.3003520.g004

To test whether the other candidate CIA pathway proteins were present in protein complexes, we also undertook BN-PAGE western blotting on proteins extracted from the rTgTah18-mAID-HA, rTgDre2-mAID-HA and rHA-mAID-TgNar1 lines. We observed TgTah18-mAID-HA in a protein complex of ~330 kDa, but were unable to clearly detect the TgDre2-mAID-HA or HA-mAID-TgNar1 proteins in complexes (S6A Fig). These data suggest that TgTah18, TgDre2, and TgNar1 are not components of the >720 kDa CTC.

If the TgCIA1, TgCIA2, and TgMMS19 proteins exist in the same complex, we reasoned that depleting one would result in alterations to the protein complex in which the other two proteins reside. We initially introduced Ty1-epitope tags into the native TgCIA2 and TgMMS19 loci of the rTgCIA1-mAID-HA line, but were subsequently unable to detect robust TgCIA2-Ty1 and TgMMS19-Ty1 expression, possibly because the antibody we have to the Ty1 epitope is not of sufficiently high affinity. Since we knew we could detect the HA-tagged versions of these proteins, we generated a new IAA-regulatable TgCIA1-expressing line by fusing a mAID-cMyc tag into the 3′ region of the TgCIA1 open reading frame (S7A Fig). We then introduced a HA tag into the 3′ region of the TgCIA2 or TgMMS19 open reading frames of this parasite line (S7B and S7C Fig). We cultured the resulting rTgCIA1-mAID-cMyc/TgCIA2-HA and rTgCIA1-mAID-cMyc/TgMMS19-HA parasites in the absence or presence of IAA for 24 h, separated proteins by SDS-PAGE, and performed western blotting. We observed robust depletion of the TgCIA1-mAID-cMyc protein upon the addition of IAA (Fig 4B). We also a observed an ~80% depletion of the TgCIA2-HA protein, and a small but significant ~20% depletion of the TgMMS19-HA protein (Fig 4B and 4C). To determine the timing of the observed TgCIA2-HA depletion upon TgCIA1 knockdown, we cultured rTgCIA1-mAID-cMyc/TgCIA2-HA parasites in IAA for a range of times between 0 and 12 h. We observed a depletion of the TgCIA1-mAID-cMyc protein as soon as 3 h after IAA addition, and a concomitant and significant decrease in the abundance of the TgCIA2-HA protein (S6B Fig). These data indicate that TgCIA1 knockdown impacts the stability of the TgCIA2 and, to a lesser extent, the TgMMS19 proteins.

We next extracted proteins from rTgCIA1-mAID-cMyc/TgCIA2-HA and rTgCIA1-mAID-cMyc/TgMMS19-HA parasites cultured in the absence or presence of IAA for 24 h, separated proteins by BN-PAGE, and performed western blotting. TgCIA1-mAID-cMyc knockdown resulted in a depletion and reduced molecular mass in both the TgCIA2-HA- and TgMMS19-HA-containing protein complexes (Fig 4D), with the depletion of the TgCIA2-HA complex occurring concomitantly with TgCIA1-mAID-cMyc knockdown (S6C Fig).

We next undertook the reverse experiment, asking whether depletion of TgCIA2 or TgMMS19 resulted in changes in the abundance of TgCIA1 and impairment of the TgCIA1-containing protein complexes. We integrated a spaghetti monster-fluorescent protein-cMyc (smFP-cMyc) tag into the TgCIA1 locus of rTgCIA2-mAID-HA and rTgMMS19-mAID-HA parasites (S7D Fig). We cultured the rTgCIA2-mAID-HA/TgCIA1-smFP-cMyc and rTgMMS19-mAID-HA/TgCIA1-smFP-cMyc lines in the absence or presence of IAA for 24 h and performed both SDS-PAGE and BN-PAGE western blotting. We observed no changes to TgCIA1-smFP-cMyc protein abundance upon the knockdown of either TgCIA2-mAID-HA or TgMMS19-mAID-HA (Fig 4E). However, we observed a complete loss of the >720 kDa TgCIA1-smFP-cMyc-containing complex upon the knockdown of both TgCIA2-mAID-HA and TgMMS19-mAID-HA (Fig 4F, black arrowhead). The two smaller TgCIA1-containing complexes were unaffected by TgCIA2 or TgMMS19 depletion (Fig 4F, red and blue arrowheads).

As a direct test for whether TgCIA1 and TgCIA2 interact, we performed co-immunoprecipitation experiments on the rTgCIA2-mAID-HA/TgCIA1-smFP-cMyc parasite line. Immunoprecipitation of TgCIA1-smFP-cMyc with anti-cMyc beads co-purified TgCIA2-mAID-HA but not the mitochondrial outer membrane protein TgTom40, and immunoprecipitation of TgCIA2-mAID-HA with anti-HA beads co-purified TgCIA1-smFP-cMyc but not TgTom40 (S6D Fig), providing further evidence that TgCIA1 and TgCIA2 are components of the same protein complex.

Taken together, our data are consistent with TgCIA1, TgCIA2, and TgMMS19 existing in the same complex of >720 kDa. The CTC of animals exists as a heteromeric complex consisting of two copies of each of the three proteins [10]. The combined mass of a similarly arranged complex in T. gondii would be ~850 kDa, which is conceivably the mass we observe for the TgCIA1, TgCIA2, and TgMMS19 complexes in BN-PAGE. Our data also indicate that TgCIA1 exists in smaller protein complexes that do not include TgCIA2 or TgMMS19.

A novel loop in the TgCIA1 protein facilitates the dual cytosolic and mitochondrial localization of the CIA targeting complex in T. gondii

The candidate CTC proteins of T. gondii exhibit an interesting dual localization to the mitochondrion and cytosol that is suggestive of a dynamic role for this complex in parasite biology. We set out to characterize the mitochondrial targeting of the CTC in more detail. We first asked whether TgCIA1 was required for the mitochondrial targeting of TgCIA2 and TgMMS19. We performed immunofluorescence assays on rTgCIA1-mAID-cMyc/TgCIA2-HA or rTgCIA1-mAID-cMyc/TgMMS19-HA parasites cultured for 3 h in the absence or presence of IAA, probing for the HA-tagged TgCIA2 and TgMMS19 proteins. As expected, both TgCIA2-HA and TgMMS19-HA exhibited dual mitochondrial and cytosolic localization in parasites cultured in the absence of IAA (Fig 5A and 5B). Notably, however, the mitochondrial localization of both TgCIA2-HA and TgMMS19-HA was significantly reduced upon TgCIA1-mAID-cMyc depletion (Fig 5A and 5B). This indicates that both TgCIA2 and TgMMS19 depend on TgCIA1 for their mitochondrial targeting.

Download:

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

Fig 5. Localization of the CTC to the mitochondrion of Toxoplasma gondii is mediated by TgCIA1.

(A–D) Immunofluorescence assays of (A) rTgCIA1-mAID-cMyc/TgCIA2-HA, (B) rTgCIA1-mAID-cMyc/TgMMS19-HA, (C) rTgCIA2-mAID-HA/TgCIA1-smFP-cMyc, and (D) rTgMMS19-mAID-HA/TgCIA1-smFP-cMyc parasites, cultured in the absence (top) or presence (bottom) of IAA for either 3 h (A, B) or 24 h (C, D). Samples were probed with anti-HA (A, B) or anti-cMyc (C, D) antibodies to detect the protein-of-interest (green) and anti-TgTom40 antibodies to detect the mitochondrion (magenta). Scale bars are 2 µm. DIC, differential interference contrast. Middle, corresponding fluorescence profiles depicting the intensity of anti-HA or anti-cMyc labeling (green) and the anti-TgTom40 labeling (magenta) along the yellow line on merged images. Right, the correlation between the protein of interest and TgTom40 was quantified using the Pearson correlation coefficient (r) and analyzed using an unpaired t test, with the p values shown. The numerical data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003520.g005

We next asked whether the mitochondrial localization of TgCIA1 is dependent on either TgCIA2 or TgMMS19. We cultured rTgCIA2-mAID-HA/TgCIA1-smFP-cMyc or rTgMMS19-mAID-HA/TgCIA1-smFP-cMyc parasites in the absence or presence of IAA for 24 h and performed immunofluorescence assays. Neither the depletion of TgCIA2-mAID-HA nor TgMMS19-mAID-HA significantly altered the localization of TgCIA1-smFP-cMyc (Fig 5C and 5D), indicating that the mitochondrial localization of TgCIA1 is not dependent on either TgCIA2 or TgMMS19. Taken together, these data indicate that targeting of the CTC to the mitochondrion of T. gondii is mediated by TgCIA1.

Next, we asked whether TgCIA1 targeting to the mitochondrion is the result of interactions with other mitochondrially-localized components of the CIA pathway. We showed previously that TgNBP35, the proposed scaffold of the CIA pathway, is anchored to the mitochondrial outer membrane courtesy of an N-terminal TMD [15]. We cultured rTgNBP35-cMyc/TgCIA1-HA parasites in the absence or presence of ATc for 2 days to deplete TgNBP35-cMyc abundance, then performed immunofluorescence assays with anti-HA antibodies to determine TgCIA1-HA localization. We found that TgCIA1-HA continued to localize to the mitochondrion following TgNBP35-cMyc depletion (S8A Fig). We next tested whether TgTah18, another mitochondrially-localized component of the CIA pathway in T. gondii (Fig 1G), is important for the mitochondrial localization of TgCIA1. We integrated a smFP-cMyc tag into the TgCIA1 locus of rTgTah18-mAID-HA parasites (S8B Fig). We cultured rTgTah18-mAID-HA/TgCIA1-smFP-cMyc parasites in the absence or presence of IAA for 24 h and performed immunofluorescence assays. We found that the localization of TgCIA1-smFP-cMyc was unchanged upon TgTah18-mAID-HA depletion (S8C Fig). Taken together, these data indicate that the mitochondrial targeting of TgCIA1 is independent of TgNBP35 and TgTah18.

A recent study found that Nar1 homologs from other eukaryotes harbor a C-terminal amino acid motif that facilitates Nar1 interaction with the CTC [25]. This motif, which includes a C-terminal tryptophan residue, appears to be conserved in TgNar1 (S8D Fig), which could explain why we were unable to C-terminally tag this protein. To test whether mitochondrially-localized TgNar1 facilitates the mitochondrial localization of TgCIA1, we introduced a smFP-cMyc tag into the TgCIA1 locus of rHA-mAID-TgNar1 parasites (S8B Fig). We cultured the resulting rHA-mAID-TgNar1/TgCIA1-smFP-cMyc parasites in the absence or presence of IAA for 24 h and performed both SDS-PAGE and BN-PAGE western blotting. Depleting HA-mAID-TgNar1 did not affect the abundance of the TgCIA1-smFP-cMyc protein or the formation of the TgCIA1-smFP-cMyc-containing protein complexes (S8E and S8F Fig). We were also unable to detect TgCIA1-smFP-cMyc protein upon immunoprecipitation of HA-mAID-TgNar1 protein (S8G Fig), suggesting that these proteins do not stably interact. This is consistent with our BN-PAGE data that TgNar1 is not part of the CTC (S6A Fig), although our data do not rule out that transient interactions occur between these proteins. Finally, we cultured rHA-mAID-TgNar1/TgCIA1-smFP-cMyc parasites in the absence or presence of IAA for 24 h and performed immunofluorescence assays to detect the TgCIA1-smFP-cMyc protein. We found that the dual mitochondrial/cytosolic localization of TgCIA1-smFP-cMyc was unchanged upon HA-mAID-TgNar1 depletion (S8H Fig). Taken together, these data indicate that the mitochondrial localization of TgCIA1 is not due to interactions with mitochondrially localized TgNar1.

We looked to the AlphaFold2 predicted structure of TgCIA1 for clues into what could be mediating the targeting of TgCIA1 (and by extension, the CTC) to the mitochondrion. CIA1 belongs to the WD40 protein family, which commonly mediate protein-protein interactions [26,27]. The CIA1 proteins from yeast and animals contain characteristic WD40 repeat domains folded into a 7-bladed β-propeller arranged around a central axis (Fig 6A; [10,28]). Each WD40 domain (or “blade”) is comprised of four antiparallel β-strands (denoted A–D in Fig 6A, pink) with small loops of 5–12 amino acids connecting the C and D strands (CD loops; Fig 6A; [10,28]). In the AlphaFold2 predicted structure of TgCIA1, the 7-bladed β-propeller is structurally conserved. However, the TgCIA1 protein contains three large insertions compared to the yeast and animal proteins (Figs 6A and S9; [10,28]). These insertions are located in the CD loops of the first, third, and fifth β-propeller blades, and we termed these the CD1, CD3, and CD5 loops, respectively (Fig 6A). The CD1 loop is 122 amino acids long in TgCIA1, the CD3 loop is 82 amino acids long, and the CD5 loop is 289 amino acids (S9 Fig). Extended CD loops are also present in the CIA1 homologs of other myzozoans (S9 Fig), a eukaryotic taxon that includes apicomplexans and their closest free-living relatives such as chrompodellids and dinozoans [13,14]. This is most apparent in the CD5 loop of myzozoan CIA1 proteins, which ranges in length from 45 amino acids in the oyster parasite P. marinus to 313 amino acids in the malaria-causing parasite P. falciparum (S9 Fig). This is in contrast to the much shorter, six-amino acid-long CD5 loop of CIA1 in the ciliate Paramecium tetraurelia, a sister taxon to the myzozoans. The extended CD loops in the TgCIA1 structure are predicted with poor confidence by AlphaFold2, and for the most part lack clear structural elements (Fig 6A).

Download:

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

Fig 6. An elongated CD5 loop is necessary for targeting TgCIA1 to the mitochondrion of Toxoplasma gondii.

(A) Structure of the Saccharomyces cerevisiae protein ScCIA1 (PDB: 2HES) and the predicted AlphaFold2 structure of TgCIA1(UniProt ID: A0A125YRQ0), with the A, B, C, and D strands of the third propellor blade in the ScCIA1 protein labeled in pink. The loops connecting the C and D strands of blade 1 (CD1 loop; red), blade 3 (CD3 loop, yellow), and blade 5 (CD5 loop; green) in each structure were colored using EzMol software [63]. (B, E) Western blots of proteins extracted from rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1 (WT-Ty1), rTgCIA1-mAID-HA/cTgCIA1_ScCD1-Ty1 (ScCD1-Ty1), rTgCIA1-mAID-HA/cTgCIA1_ScCD3-Ty1 (ScCD3-Ty1), and rTgCIA1-mAID-HA/cTgCIA1_ScCD5-Ty1 (ScCD5-Ty1) parasites cultured in the absence or presence of IAA for 24 h, separated by (B) SDS-PAGE or (E) BN-PAGE, and probed with anti-Ty1, anti-HA, or anti-TgTom40 antibodies. The masses of the predicted CTC (black arrowhead) and smaller TgCIA1-containing complexes (red and blue arrowheads) are depicted in the BN-PAGE data. (C) Immunofluorescence assays of rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1 (WT-Ty1), rTgCIA1-mAID-HA/cTgCIA1_ScCD1-Ty1 (ScCD1-Ty1), rTgCIA1-mAID-HA/cTgCIA1_ScCD3-Ty1 (ScCD3-Ty1), and rTgCIA1-mAID-HA/cTgCIA1_ScCD5-Ty1 (ScCD5-Ty1) parasites, probed with anti-Ty1 (green) and anti-TgTom40 (magenta) antibodies. Scale bars are 2 µm. DIC, differential inference contrast. Right, corresponding fluorescence profiles depicting the intensity of anti-Ty1 (green) and anti-TgTom40 (magenta) labeling along the yellow line on merged images. (D) The correlation between Ty1-tagged proteins and TgTom40 was quantified using the Pearson correlation coefficient (r). The data were analyzed using a one-way ANOVA followed by Tukey’s multiple comparisons test with p values shown. (F) Fluorescence proliferation assays of rTgCIA1-mAID-HA, rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1, rTgCIA1-mAID-HA/cTgCIA1_ScCD1-Ty1, rTgCIA1-mAID-HA/cTgCIA1_ScCD3-Ty1, and rTgCIA1-mAID-HA/cTgCIA1_ScCD5-Ty1 parasites cultured in the absence (black) or presence (red) of IAA. Parasite proliferation is expressed as a percentage of the final fluorescence measurement in the -IAA condition for each line. Individual data points and error bars represent the mean ± SD of three technical replicates. Error bars not visible are smaller than the symbol. Data are representative of three independent experiments. The numerical data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003520.g006

Notably, the CD loops are located on the opposite side of the CIA1 protein from that which interacts with CIA2 [10]. This places the CD loops in position to be involved in interactions external to the CTC. We therefore tested whether the extended CD1, CD3, or CD5 loops of TgCIA1 could have a role in mitochondrial targeting. We generated TgCIA1 transgenes in which the endogenous CD1, CD3, or CD5 loops of TgCIA1 were substituted for the equivalent, but substantially shorter, loops from the cytosolic CIA1 protein of yeast. We fused the resulting transgenes to a Ty1 epitope tag to enable their detection, and overexpressed them from the constitutive α-tubulin promoter in the rTgCIA1-mAID-HA line. We termed the resulting proteins cTgCIA1_ScCD1-Ty1, cTgCIA1_ScCD3-Ty1, and cTgCIA1_ScCD5-Ty1, respectively, and also included a cTgCIA1_WT-Ty1 control in our analyses. SDS-PAGE western blot analysis confirmed that each protein was expressed in T. gondii and were of the expected masses (Fig 6B). We observed some differences in expression levels between the constitutively-expressed proteins, with the cTgCIA1_WT-Ty1 and cTgCIA1_ScCD1-Ty1 proteins more abundant than the cTgCIA1_ScCD3-Ty1 and cTgCIA1_ScCD5-Ty1 proteins (S10A Fig). To test the localization of the modified proteins, we performed immunofluorescence assays. As expected, the cTgCIA1_WT-Ty1 protein exhibited dual localization to the cytosol and mitochondrion (Fig 6C). Both cTgCIA1_ScCD1-Ty1 and cTgCIA1_ScCD3-Ty1 proteins also localized dually to the mitochondrion and cytosol (Fig 6C). This infers that the native CD1 and CD3 loops of TgCIA1 are not important for mitochondrial targeting, although we observed that rTgCIA1-mAID-HA/cTgCIA1_ScCD1-Ty1 parasites exhibited aberrant mitochondrial morphology, with ~75% of parasites containing “tadpole”-like mitochondria instead of the typical “lasso” and ‘branched’ structures that mitochondria adopt in intracellular parasites (S10B Fig; [29]). In contrast to the other CD loop mutants, the mitochondrial localization of the cTgCIA1_ScCD5-Ty1 protein was significantly reduced, with the protein localized predominantly to the cytosol (Fig 6C and 6D). This indicates that the extended CD5 loop of TgCIA1 is necessary for mitochondrial targeting.

We next explored the importance of the extended CD loops for TgCIA1 function. We first performed BN-PAGE western blotting. Each of the CD loop mutants associated with a complex of ~720 kDa, which likely represents the CTC (Fig 6E; black arrowhead). This implies that the CD1, CD3, and CD5 loops are not required for TgCIA1 to assemble into the CTC. The cTgCIA1_WT-Ty1, cTgCIA1_ScCD1-Ty1, and cTgCIA1_ScCD3-Ty1 proteins were all found in the smaller TgCIA1 complexes (Fig 6E; red and blue arrowheads), similar in mass to those we had observed previously with the natively tagged TgCIA1-HA protein (Fig 4A). The smaller TgCIA1 complexes in the cTgCIA1_WT-Ty1 and cTgCIA1_ScCD1-Ty1 lines were, in proportion to the ~720 kDa complex, more abundant than for the natively tagged TgCIA1-HA protein (compare Figs 4A to 6E), possibly an artifact of protein overexpression from the non-native α-tubulin promoter. Curiously, the cTgCIA1_ScCD5-Ty1 protein did not appear to be present in the smaller mass complexes (Fig 6E), suggesting a role for the CD5 loop in assembly of these complexes.

Next, we tested whether the CD1, CD3, or CD5 loops of TgCIA1 are important for parasite proliferation. We conducted fluorescence proliferation and plaque assays on rTgCIA1-mAID-HA parasites constitutively expressing WT TgCIA1 or the CD loop mutants in the absence or presence of IAA. As expected, the severe proliferation defect observed upon TgCIA1-mAID-HA depletion was rescued by constitutive expression of cTgCIA1_WT-Ty1 (Figs 6F and S10C). cTgCIA1_ScCD1-Ty1-expressing parasites exhibited a moderate proliferation defect upon TgCIA1-mAID-HA depletion, resulting in fewer (rather than smaller) plaques (Figs 6F and S10C). This suggests a possible role for the CD1 loop in processes that affect the viability of extracellular parasites or the ability of parasites to invade host cells. Parasites expressing the cTgCIA1_ScCD3-Ty1 protein exhibited only minor defects in proliferation when the TgCIA1-mAID-HA protein was depleted (Figs 6F and S10C). Notably, cTgCIA1_ScCD5-Ty1-expressing parasites exhibited a severe impairment of parasite proliferation when TgCIA1-mAID-HA was depleted, indistinguishable from the proliferation defect observed in non-complemented rTgCIA1-mAID-HA parasites (Figs 6F and S10C). This indicates that the CD5 loop is critical for TgCIA1 protein function.

Having demonstrated that the CD5 loop of TgCIA1 is necessary for mitochondrial targeting, we wondered whether the CD5 loop alone could mediate mitochondrial protein targeting. We inserted the CD5 loop of TgCIA1 into a green fluorescent protein (GFP)-Ty1 reporter at an internal site in GFP shown previously to tolerate insertions (between the eighth and ninth β-strands of GFP; [30]). We expressed GFP_TgCD5-Ty1 in T. gondii parasites and attempted to select parasites stably expressing the transgene. We found that, following selection, very few parasites expressed the GFP_TgCD5-Ty1 protein. In those that did, the GFP_TgCD5-Ty1 protein exhibited a dual localization to both the mitochondrion and cytosol (S11 Fig). However, we noticed that the mitochondrial morphology in these parasites appeared aberrant, suggesting a potential toxic effect of GFP_TgCD5-Ty1 overexpression that complicates our interpretation of the data.

We next generated a transgene encoding the structurally characterized Drosophila melanogaster CIA1 protein (DmCIA1; [10]) in which we replaced the native CD5 loop of DmCIA1 with the TgCIA1 CD5 loop. We expressed cDmCIA1_TgCD5-Ty1 or a corresponding cDmCIA1_WT-Ty1 protein in rTgCIA1-mAID-HA parasites and performed immunofluorescence assays to determine protein localization. As expected, cDmCIA1_WT-Ty1 localized in the cytosol and did not overlap with the mitochondrion (Fig 7A). By contrast, the cDmCIA1_TgCD5-Ty1 protein co-localized with the mitochondrial marker (Fig 7A). Taken together, these data indicate that the CD5 loop of TgCIA1 is sufficient to mediate mitochondrial localization.

Download:

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

Fig 7. The CD5 loop of TgCIA1 is sufficient for mitochondrial targeting, with the positioning of the loop important for TgCIA1 function.

(A) Immunofluorescence assays of parasites constitutively expressing cDmCIA1_WT-Ty1 (top) or cDmCIA1_TgCD5-Ty1, (bottom) probed with anti-Ty1 (green) and anti-TgTom40 (magenta; mitochondrion) antibodies. Scale bars are 2 µm. DIC, differential interference contrast. Right, corresponding fluorescence plots depicting the intensity of anti-Ty1 (green) and anti-TgTom40 (magenta) labeling along the yellow line in merged images. (B) Constitutive expression of a cTgCIA1_CD5-to-CD1-Ty1 protein variant in rTgCIA1-mAID-HA parasites, shown in schematic (top), and probed in an immunofluorescence assay with anti-Ty1 (green) and anti-TgTom40 (magenta; mitochondrion) antibodies (bottom). Scale bars is 2 µm. DIC, differential interference contrast. Right, corresponding fluorescence plot depicting the intensity of anti-Ty1 (green) and anti-TgTom40 (magenta) labeling along the yellow line in merged image. Left, western blot of proteins extracted from rTgCIA1-mAID-HA/cTgCIA1_CD5-to-CD1-Ty1 parasites, separated by SDS-PAGE and probed with anti-Ty1 antibodies. (C) Fluorescence proliferation assays of rTgCIA1-mAID-HA, rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1, and rTgCIA1-mAID-HA/cTgCIA1_CD5-to-CD1-Ty1 parasites, cultured in the absence (black) or presence (red) of IAA. Parasite proliferation is expressed as a percentage of the fluorescence measurement in the -IAA condition on the final day of the assay for each line. Individual data points and error bars represent the mean ± SD of three technical replicates. Error bars not visible are smaller than the symbol. Data are representative of three independent experiments. The numerical data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003520.g007

We next asked whether the position of the CD5 loop in the TgCIA1 protein was important for mitochondrial localization or protein function. We engineered a TgCIA1 construct in which the amino acid sequence encoding the CD5 loop of TgCIA1 was located in the CD1 loop of the protein, with the native CD5 loop replaced by the equivalent CD5 loop of yeast (Fig 7B). We constitutively expressed the resulting protein, which we termed cTgCIA1_CD5-to-CD1-Ty1, in rTgCIA1-mAID-HA parasites, validated expression by western blotting (Fig 7B), and performed an immunofluorescence assay to determine localization of the protein. Interestingly, we observed that the cTgCIA1_CD5-to-CD1-Ty1 protein localized exclusively to the mitochondrion, no longer exhibiting the dual localization we observed in the wild type TgCIA1 protein (Fig 7B). We also found that constitutive expression of the cTgCIA1_CD5-to-CD1-Ty1 protein was unable to rescue the proliferation defect observed upon rTgCIA1-mAID-HA knockdown (Figs 7C and S10D).

Taken together, our data indicate that the mitochondrial targeting of the CTC is mediated by TgCIA1. Specifically, the CD5 loop of TgCIA1 is both necessary and sufficient for mitochondrial targeting, and this targeting is independent of the other mitochondrially-localized CIA pathway proteins TgNBP35, TgTah18, and TgNar1. Our data also indicate that, while the position of the CD5 loop in the protein is not critical for mitochondrial targeting, it is critical for facilitating the dual localization of TgCIA1 to the cytosol and mitochondrion. Finally, we have shown that the CD5 loop of TgCIA1, and its positioning within the protein, is critical for TgCIA1 to carry out its functions in parasites.

A myzozoan-specific amino acid motif in the CD5 loop mediates the mitochondrial localization of TgCIA1

We next set out to uncover the features of the CD5 loop of TgCIA1 that facilitate mitochondrial targeting. Alignments of the CIA1 protein from a range of eukaryotes identified the presence of a short, conserved motif consisting of three aromatic and one positively charged amino acid in the CD5 loop of the CIA1 protein in T. gondii and other myzozoans (Figs 8A and S9). This motif (and the extended CD5 loop generally) was not present in other eukaryotic clades, including ciliates such as P. tetraurelia, which are the nearest relatives of the myzozoans [14].

Download:

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

Fig 8. An aromatic amino acid motif in the CD5 loop of TgCIA1 facilitates mitochondrial targeting.

(A) Left, a region of a multiple sequence alignment of CIA1 homologs highlighting a motif of amino acid residues within the CD5 loop of the protein that is conserved in myzozoans. Asterisks denote key residues of the motif, with green denoting aromatic residues and orange denoting a positively charged residue. Right, a schematic of the motif observed in the dinozoans Perkinsus marinus and Symbiodinium microadriaticum, the apicomplexans Plasmodium falciparum, Cryptosporidium parvum, and Toxoplasma gondii, and the chrompodellids Vitrella brassicaformis and Chromera velia. Green = aromatic, orange = positively charged, gray = not conserved. Amino acid length of the CD5 loop in each species is listed on the right. (B, C) Western blots of proteins extracted from TgCIA1_WT-HA (WT-HA), TgCIA1_W526A-HA (W526A-HA), TgCIA1_Y527A-HA (Y527A-HA), and TgCIA1_R533A-HA (R533A-HA) expressing parasites, separated by (B) SDS-PAGE or (C) BN-PAGE, and probed with anti-HA or anti-TgTom40 antibodies. The black arrowhead indicates the >720 kDa CIA Targeting Complex; red and blue arrowheads indicate the lower mass complexes containing the TgCIA1 protein. (D) Immunofluorescence assays of TgCIA1_WT-HA (WT-HA), TgCIA1_W526A-HA (W526A-HA), TgCIA1_Y527A-HA (Y527A-HA), and TgCIA1_R533A-HA (R533A-HA) parasites. The proteins of interest (green) and the mitochondrion (magenta) were labeled with anti-HA and anti-TgTom40 antibodies, respectively. Schematics depicting the modified amino acid sequence in the CD5 motif of the proteins from each panel are included next to images (left). Scale bars are 2 µm. DIC, differential interference contrast. Right, corresponding fluorescence profile depicting intensity of anti-HA (green) and anti-TgTom40 (magenta) labeling along the yellow lines of the merged images. (E) The correlation between HA-tagged proteins of interest and TgTom40 was quantified using the Pearson correlation coefficient (r) and the data were analyzed using a one-way ANOVA followed by Tukey’s multiple comparisons test. (F) Plaque assays of TgCIA1_WT-HA, TgCIA1_W526A-HA, TgCIA1_Y527A-HA, and TgCIA1_R533A-HA parasites. Parasites were cultured in the absence (top) or presence (bottom) of IAA for 8 days and are representative of three independent experiments. (G) Immunofluorescence assays of rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1 and rTgCIA1-mAID-HA/cTgCIA1_F532A-Ty1 parasites. The cTgCIA1_WT-Ty1 protein (green) and the mitochondrion (magenta) were labeled with anti-Ty1 and anti-TgTom40 antibodies, respectively. Schematics depicting the amino acid sequence in the CD5 motif of the proteins from each panel are included next to images (left). Scale bars are 2 µm. DIC, differential interference contrast. Right, corresponding fluorescence profile depicting intensity of anti-Ty1 (green) and anti-TgTom40 (magenta) labeling along the yellow lines of the merged images. (H) The correlation between Ty1-tagged proteins and TgTom40 was quantified using the Pearson correlation coefficient (r). The data were analyzed using a one-way ANOVA (alongside the r values depicted in S12F Fig) followed by Tukey’s multiple comparisons test, with the p value shown. (I) Fluorescence proliferation assays of rTgCIA1-mAID-HA, rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1, and rTgCIA1-mAID-HA/cTgCIA1_F532A-Ty1 parasites, grown in the absence (black) or presence (red) of IAA. Parasite proliferation is expressed as a percentage of the fluorescence measurement in the -IAA condition on the final day of the assay for each line. Individual data points and error bars represent the mean ± SD of three technical replicates. Error bars not visible are smaller than the symbol. Data are representative of three independent experiments. The numerical data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003520.g008

We hypothesized that the conserved motif of the CD5 loop could facilitate the mitochondrial targeting of TgCIA1. To test this, we used CRISPR/Cas9-based genome editing to individually substitute each residue in the motif for alanine in the native TgCIA1 locus of the TgCIA1-HA/rTgNBP35-cMyc parasite line. We were successful in generating mutants in the W526, Y527, and R533 residues (but not the F532 residue), terming the resulting proteins TgCIA1_W526A-HA, TgCIA1_Y527A-HA, and TgCIA1_R533A-HA (S12A Fig). SDS-PAGE western blot analyses confirmed that the mutated proteins were expressed at similar abundances to a TgCIA1_WT-HA control (Figs 8B and S12B). We next performed BN-PAGE western blotting and found that all mutated proteins were present in the >720 kDa CTC as well as in the smaller TgCIA1-containing complexes we had observed previously (Fig 8C).

To test whether the motif is important for mitochondrial targeting, we performed immunofluorescence assays. Notably, all mutations in the CD5 loop motif resulted in the TgCIA1 protein localizing predominantly to the cytosol (Fig 8D), although quantifications revealed that the TgCIA1_R533A-HA protein exhibited significantly greater mitochondrial co-localization than the TgCIA1_W526A-HA and TgCIA1_Y527A-HA proteins (Fig 8E).

Next, we investigated whether the W526, Y527, or R533 residues of the conserved CD5 loop motif contribute to the role of TgCIA1 in parasite proliferation. We performed plaque assays comparing the proliferation of parasites expressing TgCIA1_W526A-HA, TgCIA1_Y527A-HA, and TgCIA1_R533A-HA to parasites expressing TgCIA1_WT-HA. We found that TgCIA1_W526A-HA and TgCIA1_Y527A-HA expressing parasites exhibited severe proliferation defects, indicating that these residues are critical for TgCIA1 function (Fig 8F). By contrast, the proliferation of parasites expressing TgCIA1_R533A-HA was indistinguishable from those of parasites expressing TgCIA1_WT-HA, suggesting the R533 residue is largely dispensable for TgCIA1 function (Fig 8F).

We were unable to modify the TgCIA1 locus to express a F532A mutation using the genome editing approach. As an alternative, we constitutively expressed TgCIA1_F532A-Ty1 from the α-tubulin promoter in rTgCIA1-mAID-HA parasites, generating a line we termed rTgCIA1-mAID-HA/cTgCIA1_F532A-Ty1. We found that the cTgCIA1_F532A-Ty1 protein was expressed at a similar abundance, and in protein complexes of similar masses, to the cTgCIA1_WT-Ty1 protein (S12C and S12D Fig). Notably, we found that the cTgCIA1_F532A-Ty1 protein localized predominantly to the cytosol, exhibiting significantly less mitochondrial co-localization than the cTgCIA1_WT-Ty1 protein (Fig 8G and 8H). We also constitutively-expressed TgCIA1_W526A-Ty1, TgCIA1_Y527A-Ty1 and cTgCIA1_R533A-Ty1 isoforms in rTgCIA1-mAID-HA parasites, and found that the localization of these matched what we observed in the genome-edited point mutants, with all TgCIA1 variants localizing predominantly to the cytosol, although the cTgCIA1_R533A-Ty1 again exhibited greater mitochondrial co-localization than the other variants (S12E and S12F Fig).

Finally, we undertook fluorescence proliferation assays and plaque assays to test whether the F532 residue of the CD5 loop is important for TgCIA1 function. We compared proliferation of rTgCIA1-mAID-HA/cTgCIA1_F532A-Ty1 parasites to rTgCIA1-mAID-HA and rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1 parasite lines cultured in the absence or presence of IAA. We observed that proliferation of rTgCIA1-mAID-HA/cTgCIA1_F532A parasites was severely impaired when TgCIA1-mAID-HA was knocked down (Figs 8I and S10E), indicating that the F532 residue is essential for TgCIA1 protein function. We also tested the proliferation of rTgCIA1-mAID-HA/TgCIA1_W526A-Ty1, rTgCIA1-mAID-HA/TgCIA1_Y527A-Ty1, and rTgCIA1-mAID-HA/cTgCIA1_R533A-Ty1 parasites in the absence or presence of IAA. This revealed that rTgCIA1-mAID-HA/cTgCIA1_Y527A-Ty1 and rTgCIA1-mAID-HA/cTgCIA1_R533A-Ty1 parasites proliferated normally upon knockdown of the TgCIA1-mAID-HA protein, whereas proliferation of rTgCIA1-mAID-HA/cTgCIA1_W526A-Ty1 parasites was substantially reduced upon TgCIA1-mAID-HA depletion, although not to the same extent as observed in rTgCIA1-mAID-HA/cTgCIA1_F532A-Ty1 parasites (S12G and S10E Figs). These findings suggest that constitutive overexpression of the TgCIA1_W526A and TgCIA1_Y527A mutant isoforms can partially or fully rescue the severe proliferation defects observed in the genome-edited point mutants.

Taken together, these data indicate that the W526, Y527, and F532 residues of the aromatic amino acid motif in the CD5 loop of TgCIA1 are critical for both mitochondrial localization of the TgCIA1 protein and for the functional role of TgCIA1 in parasite proliferation. The R533 residue contributes to the mitochondrial localization of TgCIA1, although not to the same extent as the other residues of this motif that we tested. Surprisingly, despite its role in mitochondrial targeting, the R533 residue of TgCIA1 appears to be dispensable for TgCIA1 function.

Given the conservation of the CD5 loop motif in myzozoans, we asked whether CD5 loops from other myzozoans could complement the function of T. gondii CD5 loop. We replaced the CD5 loop of TgCIA1 with the equivalent CD5 loop from the chrompodellid V. brassicaformis or the dinozoan S. microadriaticum, generating proteins we called cTgCIA1_VbCD5-Ty1 or cTgCIA1_SmCD5-Ty1. The V. brassicaformis CD5 loop contains the same residues in the conserved aromatic motif as the TgCIA1 protein (Fig 8A), but is considerably shorter than the T. gondii CD5 loop (87 amino acids versus 249 amino acids). The S. microadriaticum CD5 loop only encodes two amino acids of the motif (the phenylalanine and arginine residues; Fig 8A), and is also considerably shorter than the T. gondii CD5 loop (97 amino acids). We expressed the cTgCIA1_VbCD5-Ty1 and cTgCIA1_SmCD5-Ty1 proteins in rTgCIA1-mAID-HA parasites, validated expression by SDS-PAGE western blotting (Fig 9A and 9B), and performed immunofluorescence assays to determine protein localization. Both the cTgCIA1_VbCD5-Ty1 and cTgCIA1_SmCD5-Ty1 proteins localized predominantly to the cytosol (Fig 9A–9C), although the cTgCIA1_SmCD5-Ty1 protein also exhibited some observable co-localization with the mitochondrion (Fig 9B, arrowheads).

Download:

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

Fig 9. The CD5 loop of the CIA1 protein is functionally conserved in myzozoans.

(A, B) Immunofluorescence assay of (A) rTgCIA1-mAID-HA/cTgCIA1_VbCD5-Ty1 and (B) rTgCIA1-mAID-HA/cTgCIA1_SmCD5-Ty1 parasites, probed with anti-Ty1 (green) and anti-TgTom40 (magenta; mitochondrion) antibodies. Schematics depicting the amino acid sequence in the CD5 motif of the proteins from each panel are included below the Ty1 labeled images. Scale bars are 2 µm. DIC, differential interference contrast. Arrowheads highlight regions where the cTgCIA1_SmCD5-Ty1 protein exhibits visible mitochondrial localization. Right, corresponding fluorescence profile depicting intensity of anti-Ty1 (green) and anti-TgTom40 (magenta) labeling along the yellow line on merged image. Left, western blots of the proteins of interest, separated by SDS-PAGE and probed with anti-Ty1 antibodies. (C) The correlation between the Ty1-tagged proteins of interest and TgTom40 was quantified using the Pearson correlation coefficient (r). The data were analyzed using a one-way ANOVA followed by Tukey’s multiple comparisons test with p values shown. (D) Fluorescence proliferation assays of rTgCIA1-mAID-HA, rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1, rTgCIA1-mAID-HA/cTgCIA1_VbCD5-Ty1, and rTgCIA1-mAID-HA/cTgCIA1_SmCD5-Ty1 parasites, grown in the absence (black) or presence (red) of IAA. Parasite proliferation is expressed as a percentage of the fluorescence measurement in the -IAA condition on the final day of the assay for each line. Individual data points and error bars represent the mean ± SD of three technical replicates. Error bars not visible are smaller than the symbol. Results are representative of three independent experiments. The data for the rTgCIA1mAID-HA and rTgCIA1-mAID-HA/cTgCIA1_WT-Ty1 lines are identical to those depicted in Fig 7C, the experiments for which were performed simultaneously. The numerical data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003520.g009

Next, we investigated whether the CD5 loops of the SmCIA1 and VbCIA1 homologs were functionally equivalent to CD5 loop of TgCIA1. We cultured rTgCIA1-mAID-HA/cTgCIA1_VbCD5-Ty1 and rTgCIA1-mAID-HA/cTgCIA1_SmCD5-Ty1 parasites in the absence or presence of IAA and performed fluorescence proliferation and plaque assays. These revealed that both the cTgCIA1_SmCD5-Ty1- and cTgCIA1_VbCD5-Ty1-complemented lines proliferated normally when cultured in the presence of IAA (Figs 9D and S10D), indicating that the CD5 loops from the S. microadriaticum and V. brassicaformis CIA1 proteins can functionally replace the equivalent loop in the TgCIA1 protein of T. gondii.

Taken together, our data demonstrate that the CD5 loop of TgCIA1 contains a motif that is conserved throughout the myzozoans (Fig 8A), and which contributes to targeting the CIA1 protein to the mitochondrion. Numerous residues of this motif, including a phenylalanine residue from this motif that is found in all the analyzed myzozoan CIA1 sequences except Cryptosporidium parvum (Fig 8A), are critical for TgCIA1 to carry out its biological role.

The CD5 loop of TgCIA1 is not required for cytosolic protein synthesis

We have shown that the CTC of T. gondii exhibits dual localization to the mitochondrion and cytosol, courtesy of a myzozoan-specific CD5 loop in the CIA1 protein of the complex (Fig 10A). It is conceivable that the mitochondrial localization of TgCIA1 is important for the transfer of [4Fe-4S] clusters from mitochondrially-localized Nar1 to the CTC. TgCIA1 localizes to the mitochondrion independently of its association with TgNar1 (S8H Fig), but it is possible that the CD5 loop of TgCIA1 interacts with mitochondrial outer membrane lipids or an accessory protein of the mitochondrial outer membrane (Fig 10Ai–10Aii). This could place the CTC in position on the outer membrane to interact with TgNar1 and enable [4Fe-4S] cluster transfer to occur. In these scenarios, the CD5 loop of TgCIA1 functions as a mitochondrial targeting signal to facilitate FeS cluster transfer from TgNar1. Notably, T. gondii expresses the FeS proteins TgELP3 and TgRlmN on the outer face of outer mitochondrial membrane [31]. An alternative possibility is, therefore, that the CD5 loop instead functions in enabling TgCIA1 and the CTC complex to interact with these client FeS proteins on the outer membrane (Fig 10Aiii). In this scenario, the CD5 loop functions not in enabling [4Fe-4S] cluster transfer from TgNar1, but instead to enable [4Fe-4S] cluster transfer from the CTC to client mitochondrial outer membrane proteins.

Download:

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

Fig 10. A model for the spatial organization and evolution of the CIA pathway in Toxoplasma gondii and related eukaryotes.

(A) Models for the spatial organization of the CIA pathway in T. gondii and related organisms. A sulfur-containing product of the mitochondrial ISC pathway (possibly a [2Fe-2S] cluster) is exported from the mitochondrion. [4Fe-4S] clusters assemble on the NBP35 scaffold (N35) at the mitochondrial outer membrane, with electrons for this process donated from NADPH via mitochondrial Tah18 (T18) and possibly Dre2 (D2). [4Fe-4S] clusters are transferred from the NBP35 scaffold to Nar1 (N1), and further to the CIA targeting complex (CTC) comprised of CIA1 (C1), CIA2 (C2) and MMS19 (M19). CIA1 contains an extended CD5 loop (green) that mediates localization of the entire CTC to the mitochondrion. The CD5 loop may bind to (i) outer mitochondrial membrane lipids or (ii) outer mitochondrial membrane accessory proteins, and place the CTC in position to receive FeS clusters from Nar1. Alternatively, (iii) the CD5 loop may mediate interactions with FeS client proteins that are anchored on the mitochondrial outer membrane, thus facilitating FeS transfer from the CTC to these proteins. Elements of the diagram were created in BioRender. Hodgson, E. (2025) https://BioRender.com/ecxz792. (B) Western blots measuring the incorporation of puromycin into proteins from rTgCIA1-mAID-HA/cTgCIA1_ScCD5-Ty1 (ScCD5) or rTgCIA1-mAID-HA/cTgCIA1_CD5-to-CD1-Ty1 (CD5-to-CD1) parasites cultured in the absence (−) or presence (+) of IAA for 24 h, separated by SDS-PAGE and probed with anti-puromycin and anti-TgTom40 antibodies. (C) Relative abundance of puromycin incorporation from (B) was determined as a percentage of the -IAA control for each parasite line, with abundances normalized using the TgTom40 loading control. Data points represent the mean ± SD of three independent experiments. Data were analyzed using a one-way ANOVA followed by Tukey’s multiple comparisons test, with relevant p values shown. The numerical data underlying this Figure can be found in S1 Data. (D) An illustrative phylogenetic tree adapted from currently accepted models of eukaryotic evolution [63–65]. Mitochondrial targeting domains for key components of the CIA machinery, including the extended CD5 loop of CIA1, the N-terminal (N-term) transmembrane domain (TMD) of NBP35 and the C-terminal (C-term) TMDs of the FeS proteins Elp3 and RlmN, all evolved early in myzozoan evolution, subsequent to their divergence from the ciliate lineage. The endosymbiotic acquisition of a chloroplast also occurred early in myzozoan evolution.

https://doi.org/10.1371/journal.pbio.3003520.g010

If the mitochondrial localization of TgCIA1 is critical for FeS cluster transfer from TgNar1, we predicted that impairing the mitochondrial localization of TgCIA1 will impair all downstream processes that require the CTC, such as cytosolic protein translation. We therefore measured protein translation in a parasite strain expressing a TgCIA1 variant that is unable to target to the mitochondrion (cTgCIA1_ScCD5-Ty1, which lacks the mitochondrial targeting CD5 loop and localizes exclusively to the cytosol; Fig 6C). We compared this to protein translation in a parasite strain expressing a TgCIA1 variant that is targeted exclusively to the mitochondrion (cTgCIA1_CD5-to-CD1-Ty1; Fig 7B). Remarkably, we still observed robust cytosolic protein translation in parasites expressing only the cytosolically localized cTgCIA1_ScCD5-Ty1 protein (Fig 10B and 10C). By contrast, we observed a significant depletion of cytosolic protein translation when parasites expressed only the mitochondrially-localized cTgCIA1_CD5-to-CD1-Ty1 protein (Fig 10B and 10C), similar to the defect we observed when TgCIA1 is depleted (Fig 3C and 3D). These data indicate that CD5 loop-dependent mitochondrial localization of TgCIA1 is not required for the FeS cluster-dependent process of protein translation in the cytosol. These data are, therefore, inconsistent with the hypothesis that the CD5 loop of TgCIA1 is important for FeS cluster transfer from Nar1 to the CTC at the mitochondrial outer membrane (Fig 10Ai–10Aii). Instead, our data indicate that some functions facilitated by TgCIA1 (such as cytosolic protein translation) are independent of the CD5 loop. Our data also suggest that the cytosolic localization of TgCIA1 is required to enable cytosolic protein translation, since we observed a strong translation defect in parasites expressing only the mitochondrially-localized cTgCIA1_CD5-to-CD1-Ty1 protein, although we cannot rule out this defect is due to the aberrant positioning of the CD5 loop in this protein interfering with CTC function.

Parasite culturing

Tachyzoite-stage T. gondii parasites were cultured in human foreskin fibroblasts (HFFs) using Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 1% (v/v) fetal bovine serum, 2 g/L NaHCO₃, 0.2 mM L-glutamine, 50 units/mL penicillin, 50 µg/mL streptomycin, 10 µg/mL gentamicin, and 0.25 µg/mL amphotericin b. Cultures were kept in humidified incubators at 37°C and 5% CO₂. Where appropriate, 0.1 mM IAA (Merck catalog number I2886), 0.5 µg/mL ATc (Merck catalog number 37919), or an equivalent volume of 100% ethanol (vehicle control) was added.

Generation of genetically modified parasites

We used CRISPR/Cas9 genome editing to introduce 3× HA, 3′ mini-auxin-inducible degron-3× HA (mAID-HA), 3′ mAID-3× cMyc (mAID-cMyc), 5′ HA-mAID, 3′ Ty1, or 3′ spaghetti monster fluorescent protein (smFP)-cMyc epitope tags into the 3′ or 5′ regions of the open reading frames of TgTah18 (www.ToxoDB.org identifier TGGT1_249320; [17]), TgDre2 (TGGT1_216900), TgNar1 (TGGT1_242580) TgCIA1 (TGGT1_313280), TgCIA2 (TGGT1_306590), TgMMS19 (TGGT1_222230), or TgABCE1 (TGGT1_216790). To enable, these modifications, we introduced DNA encoding single guide (sg)RNAs targeting the 3′ or 5′ regions of these genes into the vector pSAG1::Cas9-U6-UPRT (Addgene plasmid 54467, [48]) using Q5-site directed mutagenesis (New England Biolabs). We performed the Q5 reaction using the gene-specific “3′ or 5′ CRISPR fwd” primer and a “universal CRISPR rvs” primer (S1 Table). We amplified a donor DNA sequence encoding the epitope tag plus 50 bp flanks of the target gene using gene specific “tag fwd” and “tag rvs” primers and HA, C-terminal mAID-HA, N-terminal HA-mAID, mAID-cMyc, cMyc, Ty1, or smFP-cMyc gBlocks (Integrated DNA Technologies, IDT) as templates (S1 Table).

We transfected the sgRNA-expressing plasmids and HA-containing donor DNAs targeting the 3′ region of the TgTah18, TgDre2, TgCIA1, TgCIA2, or TgMMS19 into an ATc regulatable TgNBP35-cMyc strain described previously [15]. This generated TgTah18-HA, TgDre2-HA, TgCIA1-HA, TgCIA2-HA, and TgMMS19-HA expressing parasites. We transfected the sgRNA-expressing plasmids targeting TgTah18, TgDre2, TgNar1, TgCIA1, TgCIA2, or TgMMS19 and C-terminal mAID-HA, N-terminal HA-mAID, or mAID-cMyc-containing donor DNAs into RH∆ku80-Tir1-FLAG parasites expressing tdTomato red fluorescent protein (described previously, [18]). This generated the IAA-regulatable rTgTah18-mAID-HA, rTgDre2-mAID-HA, rHA-mAID-TgNar1, rTgCIA1-mAID-HA, rTgCIA2-mAID-HA, rTgMMS19-mAID-HA, and rTgCIA1-mAID-cMyc parasite lines. We transfected the sgRNA-expressing plasmid targeting TgABCE1 and the corresponding cMyc donor DNA into the rTgCIA1-mAID-HA and rTgDre2-mAID-HA parasite lines to generated the rTgCIA1-mAID-HA/TgABCE1-cMyc and rTgDre2-mAID-HA/TgABCE1-cMyc lines, respectively. We transfected the sgRNA-expressing plasmids targeting TgCIA2 and TgMMS19 and the corresponding HA donor DNA into the rTgCIA1-mAID-cMyc parasite line to generate the rTgCIA1-mAID-cMyc/TgCIA1-HA and rTgCIA1-mAID-cMyc/TgMMS19-HA lines. We transfected the sgRNA-expressing plasmid targeting TgCIA1 and the corresponding smFP-cMyc donor DNA into the rTgCIA2-mAID-HA, rTgMMS19-mAID-HA, rTgTah18-mAID-HA, and rHA-mAID-TgNar1 parasite lines to generate the rTgCIA2-mAID-HA/TgCIA1-smFP-cMyc, rTgMMS19-mAID-HA/TgCIA1-smFP-cMyc, rTgTah18-mAID-HA/TgCIA1-smFP-cMyc, and rHA-mAID-TgNar1/TgCIA1-smFP-cMyc lines, respectively.

Two days after each of the transfections described above, GFP-positive parasites (which express Cas9-GFP encoded on the modified pSAG1::Cas9-U6 vector) were sorted into wells of a 96-well plate using a FACSMelody cell sorter (BD Biosciences). After approximately one week, we identified wells containing clonal parasites (i.e., wells with single plaques), extracted genomic DNA from parasites in these wells, and performed PCR screening analyses to identify genetically modified parasites using gene-specific “3′ or 5′ scrn fwd” and “3′ or 5′ scrn rvs” primers (S1 Table).

To complement the rTgCIA2-mAID-HA line with a constitutively-expressed copy of wild type TgCIA2, we synthesized the TgCIA2_WT open reading frame and an in-frame 3′ Ty1 epitope tag as a gBlock (S1 Table). We PCR amplified the TgCIA2_WT-Ty1 open reading frame with the primers CIA2 comp fwd and Ty1 rvs, digested the resulting PCR product with BglII and XmaI and ligated this into the equivalent restriction enzyme cut sites of pUDTG [49]. The resulting vector, which we termed pUDTTy(TgCIA2_WT), expresses TgCIA2_WT-Ty1 from the constitutive α-tubulin promoter, in a vector that contains a pyrimethamine-resistant TgDHFR selectable marker and a flanking sequence of the non-essential TgUPRT gene for genomic integration. To complement the rTgCIA2-mAID-HA line with a constitutively-expressed copy of TgCIA2 in which the putative FeS cluster-binding residue Cys-524 was mutated to Ala, we synthesized the TgCIA2_C524A open reading frame and an in-frame 3′ Ty1 tag as a gBlock (S1 Table). We PCR amplified the TgCIA2_C524A-Ty1 open reading frame with the primers CIA2 comp fwd and Ty1 rvs, digested the resulting PCR product with BglII and AvrII (excising the TgCIA2_C524A open reading frame) and ligated this into the equivalent sites of pUDTTy(TgCIA2_WT) generating a plasmid we termed pUDTTy(TgCIA2_C524A). We linearized the pUDTTy(TgCIA2_WT) and pUDTTy(TgCIA2_C524A) vectors with MfeI, transfected into rTgCIA2-mAID-HA parasites, selected on 1 µM pyrimethamine and cloned parasites by limiting dilution. This generated the rTgCIA2-mAID-HA/cTgCIA2_WT-Ty1 and rTgCIA2-mAID-HA/cTgCIA2_C524A-Ty1 parasite lines.

To complement the rTgCIA1-mAID-HA line with constitutively-expressed copies of wild type TgCIA1 (TgCIA1_WT), TgCIA1 with the CD1 loop replaced with the CD1 loop from yeast (TgCIA1_ScCD1), TgCIA1 with the CD5 loop replaced with the CD5 loop from yeast (TgCIA1_ScCD5), TgCIA1 with the CD1 loop replaced with the T. gondii CD5 loop and the native T. gondii CD5 loop replaced with the CD5 loop from yeast (TgCIA1_CD5-to-CD1), TgCIA1 with the CD5 loop replaced with the CD5 loop from V. brassicaformis (TgCIA1_VbCD5), or TgCIA1 with the CD5 loop replaced with the CD5 loop from S. minutum (TgCIA1_SmCD5), we synthesized the open reading frames of the desired proteins as gBlocks (S1 Table). We PCR amplified the TgCIA1 variants with the primers CIA1 comp fwd and Ty1 rvs, digested the resulting PCR product with BglII and AvrII, and ligated this into the equivalent restriction enzyme sites of the pUDTTy(TgCIA2_WT) vector. This generated vectors we termed pUDTTy(TgCIA1_variant). We linearized the resulting vectors with MfeI, transfected into rTgCIA1-mAID-HA parasites, selected on 1 µM pyrimethamine, and cloned parasites by limiting dilution. This generated the rTgCIA1-mAID-HA/cTgCIA1_{CD variant}-Ty1 cell lines.

To complement the rTgCIA1-mAID-HA line with constitutively-expressed copies of TgCIA1 with the CD3 loop replaced with the CD3 loop from yeast (TgCIA1_ScCD3), or where residues in the CD5 aromatic amino acid motif were modified to alanines (TgCIA1_W526A, TgCIA1_Y527A, TgCIA1_F532A, or TgCIA1_R533A), we undertook a site-directed mutagenesis approach. We initially digested the pUDTTy(TgCIA1_WT) vector with SpeI and NsiI, and religated the vector using the compatible sticky ends generated in the digest. This removed a large portion of the vector (the DHFR selectable marker and some of the UPRT flank), and made the resulting TgCIA1-encoding vector smaller and, consequently, easier to use for site-directed mutagenesis. We performed Q5 site-directed mutagenesis with the primers CIA1_ScCD3, CIA1_W526A, CIA1_Y527A, CIA1_F532A, or CIA1_R533A fwd and rvs as per the manufacturer’s instructions (New England Biolabs). After verifying successful mutagenesis by Sanger sequencing, we excised the regions of the resulting vectors encoding the TgCIA1 variant with BglII and AvrII, and ligated these into the equivalent sites of the pUDTTy(TgCIA2_WT) vector. This generated vectors we termed pUDTTy(TgCIA1_variant). We linearized the resulting vectors with MfeI, transfected into rTgCIA1-mAID-HA parasites, selected on 1 µM pyrimethamine, and cloned parasites by limiting dilution. This generated the rTgCIA1-mAID-HA/cTgCIA1_ScCD3-Ty1 and rTgCIA1-mAID-HA/cTgCIA1_{aromatic motif mutant}-Ty1 cell lines.

To constitutively express GFP containing the CD5 loop of TgCIA1 (GFP_TgCD5), we synthesized a GFP-TgCD5 gBlock and PCR amplified the GFP-TgCD5 open reading frame with the primers GFP fwd and rvs (S1 Table). We digested the resulting PCR product with BglII and AvrII, and ligated this into the equivalent restriction enzyme sites of the pUDTTy(TgCIA2_WT) vector. This generated vectors we termed pUDTTy(GFP_TgCD5). We linearized the resulting vector with MfeI, transfected into RH∆ku80-Tir1-FLAG/tdTomato parasites, and selected on 1 µM pyrimethamine.

To constitutively-express wild type D. melanogaster CIA1 (DmCIA1_WT) or DmCIA1 with the CD5 loop replaced with the CD5 loop from T. gondii (DmCIA1_TgCD5), we synthesized both variants as gBlocks, then PCR amplified them using the primers DmCIA1 fwd and Ty1 rvs. We digested the resulting PCR products with BglII and AvrII, and ligated them into the equivalent restriction enzyme sites of the pUDTTy(TgCIA2_WT) vector. This generated vectors we termed pUDTTy(DmCIA1_WT/TgCD5). We linearized the resulting vectors with MfeI, transfected into rTgCIA1-mAID-HA parasites, selected on 1 µM pyrimethamine, and cloned parasites by limiting dilution. This generated the rTgCIA1-mAID-HA/cDmCIA1_WT/TgCD5-Ty1 cell lines.

To engineer point mutations in the aromatic amino acid motif of the CD5 loop of the native TgCIA1 locus of T. gondii, we adopted a CRISPR/Cas9-based genome editing approach. We introduced DNA encoding sgRNAs targeting the genome near the region encoding the TgCIA1_W526 and TgCIA1_Y527 residues or near the region encoding the TgCIA1_F532 and TgCIA1_R533 residues into the pSAG1::Cas9-U6-UPRT vector using Q5-site directed mutagenesis (New England Biolabs). We performed the Q5 reaction using the gene specific “CIA1-W526/Y527 CRISPR fwd” or “CIA1-F532/R533 CRISPR fwd” primers and a “universal CRISPR rvs” primer (S1 Table). To generate repair templates, we annealed complementary oligonucleotides termed W526A fwd and rvs, Y527A fwd and rvs or R533A fwd and rvs (S1 Table) by mixing 2 nmol of the fwd and rvs oligos then heating to 98°C for 3 min, before allowing the samples to cool to room temperature. We co-transfected the annealed W526A or Y527A oligos with the plasmid encoding the CIA1-W526/Y527-targeting sgRNA, and the annealed R533A oligos with the plasmid encoding the CIA1-F532/R533 sgRNA, into rTgNBP35-cMyc/TgCIA1-HA parasites, then sorted and cloned GFP-positive parasites 2 days post-transfection. We extracted genomic DNA from resulting clones and amplified the region encoding the TgCIA1 aromatic amino acid motif with the primers CIA1 motif seq fwd and rvs (S1 Table), and identified clones harboring successful genetic modifications by Sanger sequencing of the resulting PCR products.

Transfections were performed as described previously [50], using 2 mm gap cuvettes and a single 1.5 kV pulse at 25 µF capacitance and 50 Ω resistance using a Bio-Rad Gene Pulser II electroporator.

Immunofluorescence assays

Confluent HFFs growing on glass coverslips were inoculated with freshly egressed parasites and cultured for ~24 hours at 5% CO₂ and 37°C. Infected host cells were then fixed in 3% (v/v) paraformaldehyde in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH 7.4) for 15–20 min and permeabilized in 0.25% (v/v) Triton X-100 (TX-100) in PBS for 10 min. Samples were blocked in 2% or 4% (w/v) bovine serum albumin (BSA) in PBS overnight at 4°C. Samples were probed with either rat anti-HA (1:200 dilution; clone 3F10 Sigma-Aldrich, catalog number 11867423001), mouse anti-cMyc (1:100 or 1:200 dilution; clone 9E10 Santa Cruz Biotechnology, catalog number SC-40) or mouse anti-Ty1 (1:200 dilution; clone BB2, [51]) primary antibodies, together with rabbit anti-TgTom40 (1: 2,000 dilution; [52]) primary antibody. Samples were subsequently probed with donkey anti-rat AlexaFluor 488 (1:500 dilution; Thermo Fisher Scientific, catalog number A21208), goat anti-mouse AlexaFluor 488 (1:250 or 1:500 dilution; Thermo Fisher Scientific, catalog number A11029), or goat anti-mouse AlexaFluor 546 (1:250 dilution; Thermo Fisher Scientific, catalog number A11030) secondary antibodies together with goat anti-rabbit AlexaFluor 546 (1:500 dilution; Thermo Fisher Scientific, catalog number A11035) or goat anti-rabbit AlexaFluor 647 (1:500 dilution; Thermo Fisher Scientific, catalog number A21245) secondary antibodies. Images were acquired on a DeltaVision Elite deconvolution microscope (GE Healthcare) using a Photometrics CoolSNAP HQ² camera. Brightness and contrast were adjusted linearly using SoftWoRx suit 2.0 software. Images were artificially colored and the fluorescence intensity profiles, measuring pixel intensity along a line of interest, were generated using FIJI software [53]. The fluorescence intensity values were normalized to the maximum value of either channel using the equation x normalized = (x − x minimum)/(x maximum − x minimum) and plotted using GraphPad Prism 9 software. To analyze the degree of signal overlap between two channels within the boundary of a parasite vacuole Pearson’s correlation coefficients were determined. Pearson’s r values were generated using the plugin Coloc2 with Costes regression using FIJI software [53] and plotted using GraphPad Prism 10 software. To test the extent of non-specific antibody labeling, we performed immunofluorescence assays on TATi∆ku80 parasites lacking epitope tags. These data revealed minimal levels of non-specific labeling with the epitope tag antibodies at the antibody concentrations and imaging conditions that we used in the study (S13 Fig).

SDS-PAGE, BN-PAGE, and western blotting

For SDS-PAGE, parasite proteins were solubilized in LDS sample buffer (Thermo Fisher Scientific) containing 2.5% v/v β-mercaptoethanol to a concentration of 2.5 × 10⁵ parasites/µL. Proteins from 2.5 × 10⁶ parasite equivalents were separated on a pre-cast NuPage 12% Bis-Tris polyacrylamide gel (Thermo Fisher Scientific) and transferred onto a 0.45 µm pore-sized nitrocellulose membrane.

For BN-PAGE, parasite proteins were solubilized in BN-PAGE lysis buffer (Thermo Fisher Scientific NativePAGE sample buffer containing 1% v/v TX-100, 2 mM EDTA, and 1× cOmplete protease inhibitor cocktail, Merck, catalog number 11873580001) to a concentration of 2.5 × 10⁵ parasites/µl. Proteins from 2.5 × 10⁶ parasite equivalents were separated on a pre-cast NativePAGE 4%–16% Bis-Tris polyacrylamide gel (Thermo Fisher Scientific), transferred onto a polyvinylidene difluoride (PVDF) membrane, fixed in 10% (v/v) acetic acid, and de-stained in methanol.

Nitrocellulose and PVDF membranes were blocked overnight in “Blotto,” a Tris-buffered saline (TBS; 137 mM NaCl, 2.7 mM KCl, 25 mM Tris-HCl, pH 7.4) solution containing 4% (w/v) skim milk powder. Membranes were subsequently probed with rat anti-HA (1:200–1:400 dilution; clone 3F10 Sigma-Aldrich, catalog number 11867423001), mouse anti-cMyc (1:100 or 1:200 dilution; clone 9E10 Santa Cruz Biotechnology, catalog number SC-40), mouse anti-Ty1 (1:200 or 1:400 dilution; clone BB2 [51]), mouse-anti-puromycin (1:3,000 dilution; clone 12D10, Merck catalog number MABE343), or rabbit anti-TgTom40 (1:2,000 dilution [52]) primary antibodies diluted in Blotto. Secondary antibodies used were horseradish peroxidase-conjugated goat anti-rat IgG (1:5,000 or 1:10,000 dilution; Abcam, catalog number ab97057), goat anti-mouse IgG (1:2,500, 1:5,000, or 1:10,000 dilution; Abcam, catalog number ab6789), or goat anti-rabbit IgG (1:5,000 or 1:10,000 dilution; Abcam, catalog number ab97051).

Antibody-labeled membranes were incubated in enhanced chemiluminescence solution (0.04% w/v luminol, 0.007% w/v coumaric acid, 0.01% H₂O₂, 100 mM Tris pH 9.35) and imaged using a ChemiDoc MP imaging system (Bio-Rad).

In western blotting experiments comparing the expression of different proteins in a particular parasite line (e.g., where band intensities were quantified), the same membrane was probed multiple times with different antibodies, with bound antibodies removing by stripping between probings. For stripping, membranes were treated twice in stripping buffer (200 mM glycine, 3.5 mM SDS, 1% v/v Tween-20, pH 2) for 15 min, twice in PBS for 10 min, twice in TBS containing 0.05% (v/v) Tween-20 for 5 min, and then blocked in Blotto for at least 1 h. Where relevant, band intensities were quantified using ImageJ (1.53k; [54]).

Puromycin incorporation assays

Puromycin incorporation assays were adapted from a previously described methodology [21]. Parasites were cultured in the absence or presence of IAA for 24 h. Intracellular parasites were mechanically egressed from host cells by passage through a 26-gauge needle, and host cell debris removed by filtering samples through a 5 µm PVDF filter. Parasites were pelleted by centrifugation at 1,500g and resuspended in pre-warmed complete culture medium (±IAA) to a final concentration of approximately 1.5 × 10⁷ parasites/mL. Puromycin (Merck, catalog number P8833) was added to the parasite samples to a final concentration of 10 µg/mL, and parasites were incubated in a humidified 37°C incubator at 5% CO₂ for 15 min with loosened lids to allow gas exchange. Puromycin-labeled parasites were pelleted by centrifugation at 12,000g for 1 min, washed in room temperature PBS, re-pelleted by centrifugation at 12,000g for 1 min, and resuspended in reducing LDS sample buffer (Thermo Fisher Scientific) to a final concentration of 2.5 × 10⁵ parasites/µl. Proteins were separated by SDS-PAGE and detected by western blotting as described above.

Co-immunoprecipitations

Immunoprecipitations were performed as described previously [49,52]. Briefly, parasites were solubilized on ice for at least 30 min in a lysis buffer containing 1% (v/v) TX-100, 150 mM NaCl, 2 mM EDTA, 1× cOmplete protease inhibitor cocktail, and 50 mM Tris-HCl (pH 7.4). Insoluble proteins were removed by centrifugation at 21,000g and the clarified supernatant was incubated overnight with anti-HA affinity matrix (Sigma-Aldrich, catalog number 11815016001) or Myc-Trap anti-cMyc agarose beads (Chromotek, catalog reference yta) at 4°C on a spinning wheel. Unbound proteins were precipitated with trichloroacetic acid. Beads containing bound proteins were washed in a wash buffer containing 0.01% or 1% v/v TX-100, 150 mM NaCl, 2 mM EDTA, 50 mM Tris-HCl pH 7.4, and proteins were eluted from the beads using LDS sample buffer (Thermo Fisher Scientific) containing 2.5% v/v β-mercaptoethanol. The unbound and bound fractions were resuspended in reducing LDS sample buffer, with protein samples subsequently separated by SDS-PAGE and detected by western blotting as described above.

Fluorescence proliferation assays

Fluorescence proliferation assays were performed as described previously [55]. Optical bottom 96-well plates (Corning or Greiner) containing confluent HFFs were inoculated with 2,000 tdTomato-expressing parasites, and cultured in the absence or presence of IAA in phenol-red-free DMEM. Each condition was performed in triplicate. Well fluorescence was analyzed using a FLUOstar Optima plate reader (BMG Labtech) once daily, or twice daily during the expected exponential growth phase, over a 5–6 day period. Fluorescence values (with average background fluorescence subtracted) were expressed as a percentage of the average final fluorescence measurement in the -IAA condition for each cell line and were plotted using GraphPad Prism 9 or 10, with a sigmoidal curve fitted to the data using the equation:

.

Plaque assays

Confluent HFFs in 25 cm² tissue culture flasks were inoculated with 500 parasites and cultured in the absence or presence of IAA for 6 to 8 days at 37°C and 5% CO₂. Infected HFFs were stained with Gram’s crystal violet solution (Merck, catalog number 94448) for 1–3 hours and rinsed with PBS. Flasks were air-dried before imaging with a CanoScan 9000F scanner (Canon).

Multiple sequence alignment and structural analyses

CIA1 homologs were identified by screening the CIA1 ortholog group (OG6_102166) on OrthoMCL (https://orthomcl.org/orthomcl; [56]). If a CIA homolog sequence was not found in an organism-of-interest on OrthoMCL, additional BLAST searches were conducted on NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi; [57]) or UniProt (https://www.uniprot.org; [58]). We performed reciprocal BLAST searches with each candidate hit on ToxoDB (www.toxodb.org; [17]) and excluded sequences where TgCIA1 was not recognized as the top hit. Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/; [59]) was used to generate a multiple sequence alignment. Sequences that exhibited significant misalignment were excluded. The alignments were plotted using pyBoxshade (https://github.com/mdbaron42/pyBoxshade). Multiple sequence alignments of TgCIA2 and TgNar1 with homologs from other organisms were also performed using Clustal Omega, and plotted using pyBoxshade, with protein sequences acquired from VEuPathDB (www.veupathdb.org; [60]), ToxoDB and UniProt. The CIA2 alignment was manually adjusted using Jalview (version 2.11.4.1; [61]).

The structure of S. cerevisiae CIA1 was generated previously [28] and was obtained from the Protein Data Bank (PDB: 2HES). The predicted structure of TgCIA1 was obtained from AlphaFold2 (UniProt ID A0A125YRQ0; [62]). The structural analyses of CIA1 homologs were viewed and colored using EzMol software (http://www.sbg.bio.ic.ac.uk/ezmol/; [63]).

Data analysis and availability

All statistical tests were performed in GraphPad Prism 9 or 10 software and are described in the figure legends. Most fluorescence proliferation assays, plaque assays and western blots depicted in the figures are representative of at least three independent experiments (as indicated in the figure legends). These replicate proliferation, plaque assay and western blotting data, as well as additional representative immunofluorescence images from the microscopy experiments, are included in S2 Data. Source images for all western blots, PCRs, and plaque assays are included in a S1 Raw Images, and the numerical values (both raw and normalized) for the graphed data are included in the S1 Data.