AOX and mtG3PDH are localized inside T. hominis mitosomes

To determine the precise subcellular localization of AOX and mtG3PDH in T. hominis, we generated specific rabbit polyclonal antisera against these enzymes as well as rat antisera against the known mitosomal marker ThmtHSP70 [10]. To enrich for the specific antibodies, ThAOX- and ThmtG3PDH-sera were affinity purified against the respective recombinant proteins. In immunofluorescence microscopy experiments, the affinity purified ThAOX and ThmtG3PDH antibodies detected punctate signals that co-localized with those detected using ThmtHSP70 polyclonal sera (Figs 1A and S1), consistent with their localization to the mitosome. To further investigate their mitosomal localization, the ThAOX- and ThmtG3PDH-antibodies were used to probe western blots of the protein fractions obtained after separation by differential centrifugation (Figs 1B and S2). Similar to signals detected using antibodies against the known mitosomal marker proteins ThTOM70 [13] and ThmtHSP70 [10], ThAOX and ThmtG3PDH signals were detected mainly in the low-speed (1000 g) membrane fraction and in the mitosome-enriched [13] (25,000 g) fraction. In the mitochondria of model organisms, AOX localizes to the matrix face of the inner mitochondrial membrane [29], while mtG3PDH localizes on the intermembrane face of the inner membrane [30]. In order to investigate the location of ThAOX and ThmtG3PDH, the mitosome-enriched fraction was subjected to a proteinase-K protection assay (Figs 1B and S2). In agreement with previous reports [13], the mitosomal protein import receptor ThTOM70, which is located on the mitosome surface, was digested by proteinase (Fig 1B). By contrast, ThmtHSP70, ThAOX and ThmtG3PDH signals were protected from proteinase activity consistent with a protected location inside the mitosome.

Download:

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

Fig 1. The components of the microsporidian alternative respiratory pathway localise inside T. hominis mitosome and are enriched in T. hominis spores rather than intracellular stages from infected host rabbit kidney cells.

(A) Confocal immunfluorescence images of single T. hominis cells inside the infected host RK13 cells. Monolayer of T. hominis infected host cells was fixed with methanol-acetone and double labelled with the affinity purified polyclonal rabbit antibodies against T. hominis AOX (red, top panel) or T. hominis mtG3PDH (red, bottom panel) and polyclonal rat antisera raised against T. hominis mtHSP70 (green). Nuclei of the parasite and the host (asterisk) were labelled with DAPI (blue), scale bars correspond to 1 μm. (B) Fractionation by differential ultracentrifugation was performed on lysates of T. hominis infected RK13 cells lysed with a glass homogeniser. Western blots of protein fractions were probed with the affinity purified antibodies against T. hominis AOX and mtG3PDH, as well as polyclonal sera against characterized T. hominis proteins that localize to: outer mitosomal membrane (ThTOM70) [13], mitosomal matrix (mtHSP70) [10], cell membrane (ThNTT3) [5], and cytosol (ThGRX) [13]. Mitosome-enriched 25,000g (25k) fraction was treated with proteinase K (+PK–TX), or proteinase K and Triton X-100 (+PK +TX). Supernatant fraction (super.) is a supernatant after the final 100,000g (100k) centrifugation. (C) Analyses of the intensities of the specific bands detected in protein extracts from purified T. hominis spores (Th spores), T. hominis infected RK13 cells (RKTh) and non-infected RK13 (RK) cells control. The western blots were probed with affinity purified antibodies against ThAOX or ThmtG3PDH, and polyclonal sera against ThPGK, ThmtHSP70, ThNFS, and ThTOM70. The intensities of the detected bands were measured with FIJI [31]. Significantly different values were indicated (T-test; ***, P ≤ 0.001). Error bars represent standard deviation of intensities values measured in three replicate western blotting experiments. 10 μg (BCA assay) of each protein extract were loaded onto gels (S3 Fig). (D) Hypothetical model of the ATP producing pathway consisting of glycolysis and mitosomal alternative respiratory pathway based on bioinformatics analyses of T. hominis genome [8,15]. The mitosomal alternative respiratory pathway (light orange) provides the mechanism for oxidation of cytosolic NADH to NAD⁺; and the NAD⁺ serves as an electron acceptor required to maintain generation of ATP in glycolysis. (E) Model of the main mitosomal function at different stages of the T. hominis life cycle. The intracellular stage of the life cycle is initiated by the injection of the spore’s contents (sporoplasm) into the host cell, followed by the growth of the parasite cell into a large multinucleate proliferative cell (meront). The inset figures show the magnified section of the parasite cells containing a mitosome (orange). Mitosomes of the spores and the sporoplasms contain the machinery for oxidation of NADH to NAD⁺ via the alternative respiration (AOX and mtG3PDH), required for the generation of the ATP in the glycolytic pathway (glycolysis) [15]. Later stages of the T. hominis life cycle can utilize the membrane nucleotide transporters (NTTs) to steal ATP from their host [3–5], while using their mitosomes mainly for the generation of the Fe-S clusters [11,13] required by essential proteins such as DNA polymerases and helicases [13]. The Fe-S clusters are inserted into the apoproteins (APO) via the cytosolic Fe-S cluster assembly system (CIA) forming a functional holoenzyme (HOLO) [13]. Mitosomal proteins enriched in the spores are coloured in pink (AOX, mtG3PDH) whereas the protein detected across the parasite life cycle are coloured in brown (Fe-S biosynthesis including NFS), green (chaperone proteins e.g. mtHSP70), or blue (mitosomal protein import including TOM70). Fe-S containing nuclear proteins involved in DNA replication (DNA polymerases and helicases) were indicated in violet.

https://doi.org/10.1371/journal.ppat.1011024.g001

Mitosomal AOX and mtG3PDH are most abundant in T. homnis spores

To investigate the relative abundance of ThAOX and ThmtG3PDH in the dispersive extracellular T. hominis spores and in the intracellular stages of the parasite’s life cycle, we performed semi-quantitative western blotting analyses of protein extracts from purified spores and from a population of T. hominis-infected RK13 cells enriched in the intracellular stages (Figs 1C and S3), and immunofluorescence localization experiments in cryosection samples containing T. hominis spores (S15 Fig). The specific bands detected using antibodies against ThPGK (phosphoglycerate kinase), ThAOX, and ThmtG3PDH had significantly higher intensity in the extracts from spores than in the extracts from samples enriched in the intracellular stages of the parasite. This result supports a role for the mitosome in sustaining energy metabolism inside T. hominis spores. By contrast, signals for antibodies to components of the mitosomal ISC-biosynthesis pathway (ThNFS, ThmtHSP70) and the mitosomal protein import machinery (ThTOM70, ThmtHSP70) (Figs 1C and S3), were strongest in extracts from the intracellular stages of the parasite. These data are consistent with an increased requirement for iron-sulphur proteins in DNA replication and protein synthesis during rapid growth of the parasite, and with an increase in mitosome number prior to cell division into daughter cells, each requiring their own complement of ISC-producing mitosomes (Fig 1E).

To test expression levels of the mitosomal proteins across different stages of the parasite’s life cycle inside the host, we used RNAseq analyses of RNA samples from a synchronised time course of T. hominis infection (Fig 2C). The transcriptomics analysis demonstrated that ThNFS and ThmtHSP70 were most highly transcribed during the proliferative stages of the parasite’s life cycle, consistent with the antibody data. ThPGK, ThAOX and ThmtG3PDH were most highly transcribed during formation of dispersive spores, further corroborating a role for these proteins in sustaining energy metabolism inside the spore (Fig 1D and 1E).

Mitosomal respiratory enzymes are present inside mitosomes in early meront stages of T. hominis infection

In the RNAseq analyses, higher levels of AOX transcription were observed in the first sampled time point (Fig 2C), and AOX and mtG3PDH fluorescence was only detected in the smallest T. hominis early meront stage in the non-synchronised infection samples (S4E Fig). This suggested a possible role for the AR pathway in the initial stages of infection. To test this hypothesis, we investigated the mitosomal localization of AOX and mtG3PDH during the parasite life cycle in a synchronised time course of infection using immunofluorescence microscopy (Figs 2A–2B, S4 and S5). Antibodies against both proteins detected punctate signals co-localizing with mitosomal mtHSP70 in tiny (cell diameter of 2.25 μm ± 0.29 μm, S4B Fig) intracellular vegetative parasites shortly (0.5 hours post infection, hpi; Figs 2A–2B and S4A) after the addition of T. hominis spores to the uninfected host cell monolayer. These data suggest that AOX and mtG3PDH are already present in cells at the very early stages of infection, including both sporoplasm and early meronts through de novo expression and/or the prior presence of both proteins inside the spores. In previously published work, ThAOX and ThmtG3PDH were detected by proteomics of T. hominis spores, alongside glycolytic proteins including ThPGK, the enzyme catalysing the first ATP-yielding step of the pathway [15]. In the same study, ThPGK was also shown to be enriched in the spores of T. hominis using quantitative immunoelectron microscopy [15]. AOX and mtG3PDH were previously detected in the mitosomes of spores of P. locustae, using immunoelectron microscopy, but not in the intracellular stages [16]. Taken together the available data suggest that mitosomal ThAOX and ThmtG3PDH are present inside the infective spores and in the mitosomes of meronts, during the early hours of the infection process.

Download:

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

Fig 2. Detection of the AOX and mtG3PDH inside the T. hominis mitosomes at the early meront stages but not the late proliferative stages of the parasite’s infection cycle inside the RK13 host cells.

(A) and (B) Representative immunofluorescence images of the T. hominis cells observed at the early time points of the time course of a synchronised T. hominis infection inside the RK13 cells. Methanol-acetone fixed samples were double labelled with the affinity purified antibodies against ThAOX (red, A top panel), or polyclonal sera against ThNFS (red, B top panel); and polyclonal rat sera against ThmtHSP70 (green; A, and B top panel). Presented time points correspond to: the first detection of the intracellular stages of the parasite 30 minutes after addition of the spores (0.5 hpi), increase of the parasite’s cell diameter (12 hpi, 16 hpi), and observation of double nucleate T. hominis cells following the first nuclear division (22 hpi). The observed stages of the parasite life cycle: sporoplasms, and meronts (early, intermediate, and late) were annotated on the images and described in detail in S4 and S5 Figs. The fluorescent images are maximum intensity projections of Z-stack across the parasite cells. Nuclei of the parasite and the host (asterisk) were labelled with DAPI, scale bars correspond to 1 μm. Analyses of the numbers of the fluorescent point signals detected with the specific antibodies against ThAOX (red, A bottom panel), ThNFS (red, B bottom panel); and ThmtHSP70 (green) inside the T. hominis cells at different stages of the parasite life cycle (A and B top panels, and S4 and S5 Figs). Extended version of the analyses including phase contrast images of the parasite cells, statistical analyses, and results obtained using antibodies against ThG3PDH are presented in S5 Fig. (C) Transcription levels of ThAOX, ThmtG3PDH, ThPGK, ThmtHSP70, ThNFS, and ThTOM70 from RNAseq analyses of samples from a synchronized time course of T. hominis infection sampled at the key intracellular stages of the parasite’s life cycle (D). The transcription levels were presented as TPM (transcripts per million) values (yellow-red scale), or as Z–score values (blue-red scale) to normalize for the differences between relatively highly transcribed PGK and the less abundant transcripts. (D) The model is based on previously published data [5] and presents the key stages of the life cycle of T. hominis: 0 hpi—the extracellular environmental spore (Mature spore) injects its contents (Sporoplasm) into the host cell using a specialised infection apparatus called a polar tube, 3 hpi–observation of small ovoid early meront cells, 14 hpi–increase in the parasite’s cell diameter and observed decreased number of the mitosomes labelled with anti-AOX/mtG3PDH antibodies, 22 h–first division of the parasite’s nuclei, 40 h–first division of the parasite’s cells, 70 h–initiation of the spore formation, 96 h–mature spores infect new uninfected host cells. During the Proliferation stage, the parasite is capable of stealing ATP from its host via nucleotide NTT transporters (inset figure). Dashed line indicates the membrane vesicle (parasitophorous vesicle) that surrounds parasite cells residing inside the host cytosol.

https://doi.org/10.1371/journal.ppat.1011024.g002

Late-stage infection T. hominis mitosomes are enriched in ISC-biosynthesis proteins rather than respiratory enzymes

We next investigated if AOX and mtG3PDH could be detected inside mitosomes during the later stages of T. hominis infection. The number of fluorescent point signals of ThAOX and ThmtG3PDH decreased significantly in the larger parasite cells (Fig 2A bottom panel, and S5D–S5E Fig) after 12 hours post-infection (S4A Fig). By contrast, the number of mitosomal mtHSP70 signals observed in these larger cells was significantly higher than in cells from the earlier time points (Fig 2A bottom panel, and S5D–S5E Fig). The essential ISC biosynthesis protein Nfs was also localized to mitosomes across the life cycle of the parasite (Figs 2B and S5F), providing further support for the importance of mitosomes in this essential cellular process [11,13]. The absence of signals for AOX and mtG3PDH inside the later intracellular stages (12, 16, and 22 hpi; Figs 2A and S4A) could be caused by decreased protein expression levels combined with a dilution effect from division of mitosomes. This hypothesis is consistent with the decrease in the number of AOX transcripts (Fig 2C) after 14h post infection. The presence of AOX and mtG3PDH in the initial, but not the late stages of infection, may also explain the very low intensities of AOX and mtG3PDH, relative to the strong signals for ThmtHSP70 and ThTOM70, in western blots of the mitosome-enriched fraction (Figs 1B and S2). The samples used for fractionation were enriched in the intracellular stages of the parasite (based on microscopy observations), and the method used to lyse the cells (glass homogeniser) does not efficiently lyse spores. This suggests that the weak ThAOX and ThmtG3PDH bands in the 25k pellet may correspond to proteins present in the mitosomes of sporoplasm and early meront intracellular stages of T. hominis. These observations are consistent with a reduced role for the alternative respiratory pathway during the later stages of T. hominis cell proliferation.

In summary, the RNAseq and IFA evidence for ThAOX expression at early time points suggests that the enzyme can play a role in maintaining an optimal NAD⁺/NADH ratio to support glycolysis during the very early stages of infection when the parasite is establishing itself inside the host cell (Fig 1E). This may reflect a requirement of T. hominis to supplement ATP imported from the host by plasma membrane located nucleotide transport proteins [3–6] during host colonization. In addition to ATP generation, it is possible that early meronts of T. hominis may also use glycolysis and AOX to metabolise glyceraldehyde-3-phosphate from the pentose phosphate pathway, in order to generate NADPH required for thioredoxin, folate metabolism, and/or sphingolipid biosynthesis [8]. In contrast, late-stage infection T. hominis intermediate and late meronts seem to use their mitosomes mainly in the biosynthesis of essential iron-sulphur clusters rather than for alternative respiration (Fig 1E).

Anti-microbial agents perturb the life cycle of T. hominis infecting mammalian cells

To investigate the possible role of mitosomal AR in T. hominis spores and early stages of infection we developed a methodology for testing the effects of specific inhibitors on parasites living inside the host. Functional studies of obligate intracellular parasites such as Microsporidia pose exceptional challenges because of the intimate association between the host and the parasite cells [32]. The use of specific inhibitors could potentially help in the functional characterization of microsporidian proteins that are important for infection, including enzymes mediating central roles of specific metabolic pathways. We used a recently developed methodology for establishing a synchronised time course of infection by T. hominis [5], to develop a protocol for testing the effects of specific drugs on parasites in conjunction with immunofluorescence microscopy and specific antibodies. We first tested the effects of two clinically used anti-microsporidian drugs [33]: namely albendazole [34,35] and fumagillin [21,36]. The concentrations of the drugs were guided by the published effects of fumagilin and albendazole on the microsporidian Encephalitozoon intestinalis in an in vitro RK13 host cell infection system [37]. In the previous study [37] the reported MIC50 values were: albendazole 8–55 ng/ml (30–207 nM), and fumagillin 0.5–0.8 ng/ml (1.1–1.9 nM), and the maximum inhibition of the infection was observed at albendazole 100 ng/ml (377 nM), and fumagillin 10 ng/ml (21 nM). In our experiments fumagillin was used at 9.2-higher concentration (200 nM, 92 ng/ml) to ensure observation of the maximum inhibition effects. The albendazole concentration used in our study was 92 ng/ml (350 nM). This was slightly below the previously reported concentration for full inhibition (100 ng/ml) of E. intestinalis as in the previous study higher concentrations of albendazole had deleterious effects on the host RK13 cells [37,38]. Albendazole was previously used against T. hominis in another study [38] where the concentration of 100 ng/ml (377 nM) appeared to be the minimum dose for control of T. hominis in mouse myoblast cell cultures. Similarly, we observed no secondary infections using 350 nM albendazole in our experimental system of T. hominis cultured in RK13 host cells. The concentration of the CCB (50 μM) was selected as the lowest concentration we tested where no progression of the T. hominis life cycle (parasite nuclear division, cellular division, or spore formation) was observed (S6C Fig).

Fumagillin is currently the only available broad spectrum anti-microsporidian drug, although it can also have toxic side effects [1]. The target of fumagillin inhibition, methionine aminopeptidase 2 (Met-AP2), catalyses the cleavage of N-terminal methionine [39], one of the most common post-translational protein modifications. Two classes of Met-AP enzymes (Met-AP1 and Met-AP2) are present in many eukaryotes, including humans and Saccharomyces cerevisiae [40]. In yeast, the deletion of both Met-AP homologues is lethal [41]. Fumagillin binds covalently to a histidine residue within the active site of human Met-AP2 [42] and inhibits growth of a Met-AP1 yeast deletion strain [36]. Although both Met-AP1 (Map1) and Met-AP2 (Map2) are present in the genome of Rozella allomycis, the cryptomycota outgroup to Microsporidia [9], only a homologue of the fumagillin-sensitive Map2 is present in microsporidian genomes. A crystal structure of the fumagillin-bound E. cuniculi Map2 is available [39] and 8 out of 10 residues contacting with fumagillin in E. cuniculi Map2, are also conserved in the T. hominis homologue (S7 Fig).

To test the effects of fumagillin on T. hominis, we investigated the progression of the parasite life cycle inside RK13 host cells in fumagillin-treated and non-treated control cultures. In the non-treated control, T. hominis progressed through its life cycle, leading to the formation of mature spores 120 hpi (Fig 3A, non-treated). The spores that formed in the non-treated samples were able to infect new host cells, as evidenced by the observation of early meronts inside the newly infected host cells, and a significant increase in the numbers of infected host cells at 120 hpi. By contrast, only the very early stages of the T. hominis life cycle (early meronts) were observed (Fig 3A) in all time points from the samples treated with fumagillin. No parasites were observed undergoing nuclear division, cell division, or spore formation, suggesting arrest of the T. hominis life cycle at a very early stage of the infection. Consistent with the observed absence of newly formed spores, no new infections were observed at 120 hpi in the fumagillin-treated samples (Fig 3A). Despite the disruption of the parasite’s life cycle, drug treated T. hominis cells persisted inside the host cytosol and significantly increased their cell diameter, with the highest measured average change of 1 μm between the time points 14 hpi and 70 hpi (Fig 3B). In the corresponding time points of the non-treated control, the parasite’s cell diameter increased by on average 2.5 μm (the highest recorded change of 2.8 μm was measured between 14 hpi and 40 hpi) (Fig 3B). The measured increase in the parasite cell diameter in the fumagillin-treated samples indicated that T. hominis can still grow at a limited rate at this drug concentration. These data suggest that the inhibition was not complete—potentially due to the complex pharmacokinetics of the drug in the T. hominis/RK13 co-culture system. The observed inhibition of the parasite life cycle is consistent with an important role of T. hominis Map2 in post-translational protein modifications during the very early stages of infection, including early meronts.

Download:

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

Fig 3. AOX inhibitor supresses growth of T. hominis cells inside the RK13 host cells at early meront stages of the infection, and affects the ability to infect by the spores formed in the presence of the drugs.

Effects of the two clinically used anti-microsporidian drugs (albendazole, 350 nM; and fumagillin, 200 nM), and an inhibitor of the AOX Colletochlorin B (CCB, 50 μM) on progression of the parasite life cycle were tested in a synchronised time course of infection in T. hominis-RK13 infection system. The drugs were added to infective spores suspended in culture medium 24 hours prior to the initiation of infection. (A) Confocal immunofluorescence images of representative T. hominis cells labelled with rabbit antibodies against ThNTT4 (red) and rat antibodies against mtHSP70 (green); observed in samples collected at different time points of incubation with the drugs. The time points corresponded to: observed decreased number of the mitosomes labelled with anti-AOX/mtG3PDH antibodies (14h), first nuclear division (24h), first cellular division (40h), initiation of the spore formation (70h), infection of new host cells by newly formed mature spores (120h). New infections were first observed at 96h (Fig 2C); but dominated cultures and were quantified at 120 hours post infection. Nuclei of the parasite and the host were labelled with DAPI (blue), scale bars correspond to 2 μm (non-treated, albendazole, fumagillin) or 1 μm (CCB). (B) Violin plots of the numbers of the T. hominis infected RK cells (top row), or the diameter of T. hominis cells (bottom row) observed across a time course of infection in presence of the drugs (albendazole, fumagillin, and CCB), and in the non-treated control (No drug). Datapoints were displayed as black circles. ANOVA analysis was used to test significance of the differences between the measured values (****, P ≤ 0.0001; ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05). The same no drug control was used for comparisons with the different treatment groups (S6 Fig). All experiments were performed in three biological replicates. (C) To test the effects of the CCB on the parasite development after the observed loss of the AOX and mtG3PDH mitosomal fluorescent signals (Figs 1D and 1F and S5) the drug was added 22 hours after the initiation of the infection (S8 Fig). Violin plots represent the numbers of the: T. hominis-infected RK13 cells (infected RK13), the RK13 cells containing bags of T. hominis spores (RK13 with spores), early T. hominis meronts (early meronts), and intermediate/late T. homnis meronts (inter./late meronts); observed across the time course of infection in the CCB-treated samples (+), or in the non-treated control (-).

https://doi.org/10.1371/journal.ppat.1011024.g003

The anti-microsporidian activity of albendazole is most likely due to its ability to bind to microsporidian beta-tubulin and inhibit the formation of microtubules [43]. In the albendazole treated samples, T. hominis cells increased their cell diameter and divided at 40 hpi, as in the non-treated control (Fig 3A). However, unlike the control, in drug-treated samples the cell division was asymmetric, leading to the formation of a single large ‘mother’ cell and single small ‘daughter’ cell (Fig 3A). DAPI-stained nuclei were observed in the ‘mother’ cells but not in the ‘daughter’ cells, indicating inhibition of the nuclear division process (Fig 3A). This is consistent with the proposed mechanism of albendazole activity, whereby binding of the drug to beta-tubulin inhibits formation of nuclear microtubules required for the division of the nucleus in the process of closed mitosis (where the nuclear envelope stays intact) [44] observed in Microsporidia [43]. The process of asymmetric division with the formation of anucleate ‘daughter’ cells continued throughout the time course of the infection (Fig 3A). We observed no evidence for newly formed spores inside the infected host cells, or any new infections, in the albendazole-treated 120 hpi samples. Although the albendazole treatment did not inhibit the ability of T. hominis to divide or increase in biomass, it inhibited nuclear division and the formation of new spores (Fig 3A and 3B) required for the propagation of the infection.

AOX antagonists inhibit progress of the life cycle of T. hominis but do not inhibit AOX-negative E. cuniculi

We investigated if AOX antagaonist, colletochlorin B (CCB) [45] affected the life cycle of T. hominis infecting RK13 cells (Fig 3). CCB inhibited growth of the parasite in the early meront stages of the infection, reminiscent of the effects observed in the fumagillin-treated samples (Figs 3 and S9). In the treated samples, the parasite cell diameter increased significantly over the time course of infection; however, no nuclear division, cell division, spore formation, or new infections were observed; suggesting a microstatic rather than a microcidal effect of the drug.

To investigate if the observed inhibitory effect of CCB was reversible, the drug was removed by washing with culture medium without the drug 22h after the start of the infection (S10 Fig). After drug removal, the parasite continued to progress through its life cycle, suggesting that the microstatic effect is reversible, which was unexpected given the reported tight-binding of CCB to AOX in vitro [45]. This might be consistent with a continuous low-level expression of AOX in the early stages of infection as suggested by the RNAseq analyses (Fig 1D). The observed inhibition of the parasite’s life cycle in the presence of a specific AOX inhibitor suggests that mitosomal AOX plays an important role during the early meront stages of the infection.

In order to test if CCB has an effect on the progression of the parasite life cycle after the observed loss of mitosomal AOX fluorescent signals, CCB was added to the culture 22 hours after the addition of T. hominis spores to the host cells (Figs 3C and S8). Parasite cells in the CCB-treated samples grew, divided and made spores similar to the non-treated control. However, no new infections were observed in the presence of CCB, despite observing spore bags inside the host cells (S8 Fig). This suggests a functionally important role for AOX during spore formation and/or during the infection process. In the non-treated samples, the parasite life cycle was completed within 120 hpi and new infections were observed 120 and 240 hpi (Figs 3C and S9).

The negative effects on the parasite cells were observed at relatively high concentrations of CCB (50 μM) which is much higher than the IC50 value of 14.8 nM for the AOX enzyme in vitro [45]. The in vitro activity of T. hominis ThAOX was also previously shown to be sensitive to 10 nM of the structurally related AOX inhibitor ascofuranone [18]. Under the in cellulo conditions of our assay it was not possible to measure the concentration of CCB inside the mitosome. To reach its target the drug has to cross at least four membranes (host cell membrane, parasite cell membrane, and two mitosomal membranes). The diffusion properties of CCB through these membranes and through the additional parasitophorous vesicle membrane, which surrounds T. hominis are unknown. That the cell cycle arrest we observe for T. hominis is due to the specific inhibition of ThAOX by CCB is supported by the absence of inhibition for the AOX-deficient microsporidian Encephalitozoon cuniculi (Figs 4A and S12) grown under the same conditions as T. hominis (Fig 4B and 4C).

Download:

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

Fig 4. The AOX inhibitor CCB has no apparent effect on AOX-deficient microsporidian E. cuniculi infecting RK13 cells.

(A) Heatmap representing distribution of enzymes involved in energy metabolism including glycolysis, alternative respiration (AR), and nucleotide transporters used to steal ATP from the host (NTT, and MFS) was mapped onto a species tree of Microsporidia, Rozellomycota, and Fungi. AOX-negative microsporidia including E. cuniculi were annotated in blue and grey. Presence or absence of protein homologues was verified in genomes using protein homology network analysis. The homologues of the two studied mitosomal AR proteins (AOX, G3PDH) are coloured in orange. The cytosolic glycerol-3-phosphate dehydrogenase (GUT) was indicated in black to distinguish it from the mitosomal AR enzymes. (B) Representative immunofluorescence images of the E. cuniculi cells, observed inside RK13 cells at different time points across the time course of synchronised infection in the presence of the Colletochlorin B derivative (CCB, 50 μM), or in the non-treated control (Non-treated). Sampled time point corresponded to: nuclear division (24h), cell division (48h), spore formation (72h), infection of new host cells by newly formed mature spores (96h). The methanol-acetone fixed samples were labelled with rat antibodies against E. cuniculi NTT2 [3] (green), and rabbit antibodies against E. cuniculi mtHSP70 [11] (red). Nuclei of the parasite and the host (asterisk) were labelled with DAPI (blue), scale bars correspond to 2 μm. (C) Box and whiskers plots of the numbers of the E. cuniculi-infected RK cells observed across a time course of infection in presence of the drug (CCB) or in the non-treated control (ND). T-test was used to test significance of the differences between the measured values (non-significant, ns).

https://doi.org/10.1371/journal.ppat.1011024.g004

Protein expression and generation of antibodies

Full length T. hominis mtHSP70, and T. homnis mtG3PDH (orf_1639, GenBank: ELQ75445.1) gene excluding the first 393 nucleotides on the 5’ terminus and 21 nucleotides on the 3’ terminus, were cloned into a pET100 expression plasmid and expressed in the E. coli BL21 (DE3) as His-tagged recombinant proteins. The first 108 nucleotides excluded from the ThmtG3PDH construct, corresponded to the predicted N-terminal targeting signal which may be cleaved off during the import to the mitosomes [18], and thus may not be present in the mature mitosomal mtG3PDH. The remainder of the ThmtG3PDH sequence excluded from the construct, corresponded to the region of the ThmtG3PDH highly identical to the mammalian mtG3PDH, that could potentially cause cross-reactivity of the produced antibodies with the mammalian homologue in the rabbit kidney (RK13) host cells. A synthetic gene encoding T. hominis AOX sequence (orf_2158, GenBank: ELQ74912.1), excluding two regions (positions 421–510, and 601–690 of the nucleotide sequence; corresponding to the regions Tyr141-Leu170, and the Leu201-Phe230 in the amino acid sequence) homologous to the membrane interaction domains in the solved structure of T. brucei AOX [19] that may be poorly accessible to antibodies, was synthesized (GeneScript, USA), cloned into pQE40 expression plasmid, and expressed in the E. coli M15 as a DHFR fusion protein. Expressed proteins were purified with the BugBuster protein extraction reagent (Novagen), following the manufacturers protocol with the addition of Benzonase (Novagen) and lysozyme. 1 mg of each purified protein was separated by SDS-PAGE, and the Instant Blue stained protein bands were cut out from the polyacrylamide gels and used for the commercial (Agrisera, Sweden) generation of rabbit (ThAOX, and ThmtG3PDH), and rat (ThmtHSP70) polyclonal antisera. Each of the animals was immunized three times (week 1, week 5 and week 9) followed by the ELISA test and test bleeding (20 ml rabbit; 0.1 ml rat; week 11). The final immunization (week 13) was followed by the final bleed (50–70 ml rabbit; 3–4 ml rat; week 15).

To affinity-purify the antibodies, approximately 500 μg of protein used for immunisation of animals was separated on SDS polyacrylamide gels (100 μg/gel), transferred onto a nitrocellulose membrane, and blocked using mTBST (Tris-buffered saline; 0.1% Tween 20; 5% milk). 5 ml of the ThmtG3PDH, or ThAOX rabbit antisera were diluted 1:10 in mTBST and incubated with the membrane overnight at 4°C with rocking. After washing the membrane three times with TBST and once with TBS, the antibodies were eluted from the membrane with 0.2 M glycine-HCl, pH 2.5 for 10 minutes, followed by neutralisation to pH 7.0 using unbuffered 1M TRIS, and the addition of NaCl to final concentration of 50 mM. The purified antibodies were concentrated to the concentrations of 0.1–0.5 mg/ml (measured with NanoDrop, ThermoFisher) and the buffer was changed for PBS, using protein concentrator ultrafiltration centrifugal tubes (100,000 MWCO, ThermoFisher). The antibodies were aliquoted, stored frozen at -20°C, or in 50% glycerol at -20°C; and used at 5 μg/ml for IFA, or 1 μg/ml for Western blotting.

Western blotting analyses of protein extracts from T. hominis spores and T. hominis infected RK cells

T. hominis infected- (ThRK13) and non-infected-Rabbit kidney cells (RK13) were routinely cultured in 175 cm² flasks at 37°C, and 5% CO₂ in Dulbecco’s Modified Eagle Medium (DMEM), containing 10% fetal bovine serum (Sigma), kanamycin 100 μg/ml, penicillin 100 μg/ml, streptomycin 100 μg/ml and fungizone 1 μg/ml. For total protein extraction, a confluent monolayer of ThRK13 or RK13 was washed three times with PBS and lysed with a 2% SDS in PBS lysis buffer containing protease inhibitor cocktail (Sigma P8340), PMSF (Sigma), MgCl₂ (0.6 mM final concentration), and Benzonase (25U, Novagen). T. hominis spores were purified from the culture medium collected from ThRK13 culture, with density gradient centrifugation in Percoll as described previously [5]. The Percoll purified spores were lysed by boiling for 10 minutes in the 2% SDS in PBS lysis buffer with protease inhibitors. The concentration of the samples for western blotting was determined with Pierce BCA assay (ThermoFisher Scientific), and 10 μg of protein extracts in 1 x Laemmli sample buffer (1xSB) was loaded per lane for SDS–PAGE and transferred onto a nitrocellulose membrane for Western blotting analysis. All blocking and antibody incubations (primary overnight at 4°C, secondary HRP-conjugated 1h at RT) steps were performed in mTBST, and washes (3 x 10 minutes) were performed using TBST with the final wash in TBS. Proteins were visualized by chemiluminescence using a ChemiDoc XRS imager system (Bio-Rad). Protein transfer efficiency was verified by Coomassie Blue staining of the gels and Red Ponceau S staining of the membranes after blotting. Band intensities were analysed using Image Lab (Bio-Rad) and in FIJI [31].

Immunofluorescence assay localization of microsporidian proteins and image analyses

Immuno-fluorescence assay (IFA) was done as described previously with minor modifications [5]. Shortly, RK13 cells infected with T. hominis or E. cuniculi grown on coverslips were fixed in methanol/acetone 50:50 v/v at -20°C for at least 10 minutes. After blocking with 5% (w/v) milk in PBS (mPBS), slides were incubated overnight with a mPBS solution containing the relevant affinity purified antibodies or antisera at 4°C, washed three times in PBS, and then incubated for 1 h with the secondary goat anti- rat or anti-rabbit antibodies (ThermoFisher) conjugated to Alexa Fluor 594 (red) or 488 (green). For the double labelling experiments, the incubation with the rat primary and the incubation with anti-rat secondary antibodies, were followed by the incubation with the rabbit primary and the incubation with the anti-rabbit secondary antibodies. Cells were incubated with DAPI in PBS for 5 min to visualize the host and the parasite nuclei. Slides were mounted using ProLong Gold (ThermoFisher). All cells were visualized using either a Nikon A1R confocal microscope with X63 or a Zeiss Axioimager II microscope with a X100 phase contrast objective lens. The images were processed in FIJI [31]. Fluorescent point signals were detected and quantified using “Find Spots” function implemented in Volocity (PerkinElmer). Colocalization of fluorescent point signals was tested using Coloc2 plugin in the FIJI [31] as described in S1 Fig.

Testing effects of specific drugs across the time course of microsporidian infection in the RK13 cells

The fresh spores were isolated on the day of the experiments by lysing the 175 cm² cultures of the ThRK13 or EcRK13 cells in 0.2% triton X-100 in PBS, followed by a sonication (3 x 45 seconds at low power intensity of 40 W, and 0.5 second cycling, on ice), and Percol purification of the highly resistant microsporidian spores. Gradient separation and removal of different stages of the spore formation, present in the lysates in addition to the infective mature spores, was avoided in order to minimise the loss of the infective spores. For T. hominis infection usually at least 1 flask of heavily infected RK culture was required for the infection of a single 24-well plate. In order to achieve the high levels of infection required for the isolation of high numbers of microsporidian spores, the RK13 cells already infected with the parasite had to be cultured for at least one month with the regular culture splitting (1:2–1:3 dilution at least once a week) and without the addition of the non-infected host cells, as the trypsinization of the ThRK13 or EcRK13 had a positive effect on the infection levels. To initiate the synchronised time course of the microsporidian infection, the Percol purified spores of T. hominis (5–20 x 10⁶ /cm²) or E. cuniculi (1–5 x 10⁶ /cm²) were added to the 80% confluent RK13 monolayer, and incubated for 2 hours at 37°C. The incubation of the spores was followed by a thorough washing (at least 3 washes) with the DMEM to remove the excess spores. Washing the cultures with the DMEM culture media instead of PBS was crucial to maximise the removal of the excess spores and minimise the stress to the host cells. For the immunofluorescence microscopy, theThRK13 or EcRK13 cells were cultured in 24-well plates on round cover slips, and time-points were collected by fixing the cells in methanol/acetone followed by the IFA. To test the effects of the drugs, the purified spores were added to a DMEM containing a drug (DMEMd) and incubated overnight at 4°C. After the overnight incubation, the spores in DMEMd were incubated for 1 hours at 37°C and added to a monolayer of the RK13 cells to initiate the infection (as described above). The DMEMd and DMEM media were changed every 48 hours. The drugs were dissolved in the DMSO (Sigma-Aldrich) and used at the following concentrations: albendazole (Alfa Aesar, 98% purity), 350 nM; fumagillin (Cayman Chemical; 95% purity), 200 nM; Colletochlorin B (CCB), 50μM. Non-treated controls were always grown in the presence of the DMSO concentration equivalent to that used in the treated samples. Every experiment was performed in triplicates and was independently replicated three times.

For the quantitative analyses of the infection, five, 4x4 mosaic images were taken using a Zeiss Axioimager II microscope with a X100 phase contrast objective lens at the top, the bottom, right side, left side, and in the centre of each cover slip. The parasite meronts were identified based on the labelling with rat anti-mtHSP70, rabbit anti-NTT4 antibodies, and DAPI. Due to the impermeability of the microsporidian spore to the antibodies, the intracellular bags of spores were identified based on the phase contras images. A single focus of the parasite infection (a single parasite cell, in early time points; or a group of parasite cells within a single parasitophorous vesicle, in later time points) was counted as a single infected host cell. The diameter of the T. hominis cells was measured at the widest area of the parasite cell labelled with the cell membrane marker (NTT4). Significance of the observed effects was tested using T-test; and ordinary two-way ANOVA (Tukey) in GraphPad Prism 9 (GraphStats), where tested drugs were the first factor and the time points of the infection were the second factor.

Fractionation of infected rabbit kidney cells, and proteinase K treatment of the mitosome-enriched fraction

Fractionation and proteinase K protection assay were performed as described previously [5] with modifications. Shortly, T. hominis infected RK13 cells were washed in PBS, and scraped off the surface of the culture flask and suspended in an ice cold HSDP lysis buffer (0.25 M sucrose, 10 mM HEPES-KOH pH 7.2) containing a protease inhibitor cocktail (Sigma), PMSF (Sigma), DNase (1 U/ml), and RNase (10 mg/ml). All subsequent manipulations were performed on ice. Cells were lysed using a Dounce homogenizer, the lysate was incubated until not viscous, and subsequently subjected to ultracentrifugation at 1,000g for 10 min, 10,000g for 30 min, 25,000g and 100,000g for 1 h. The 100,000g supernatant represented the cytosolic fraction. The pellets were resuspended in PBS containing proteinase inhibitors and PMSF, with an exception of the 25,000g samples for the proteinase K protection assay, which were resuspended in PBS. The concentration of the samples for western blotting was determined with BCA assay, the samples were boiled in the 1xSB, and 10 μg of protein from each fraction was loaded per lane for SDS–PAGE and Western blot analysis using HRP-conjugated secondary antibodies (Jackson Laboratories). Proteins were visualized by chemiluminescence using a ChemiDoc XRS imager system (Biorad).

Proteinase K protection assay was performed on the 25,000g (mitosome-enriched) pellets. The 25,000g pellet was split into three aliquots which were: either incubated with 50 μg/ml of proteinase K (Roche) for 20 minutes, or incubated 0.2% (v/v) Triton X-100 for 10 min at room temperature followed by the incubation with the proteinase for 20 min; or incubated without the addition of the proteinase and the detergent (non-treated control). The proteinase treated samples as well as the non-treated control were then incubated with trichloroacetic acid at 20% (v/v) final concentration for 30 min on ice. Precipitated protein pellets were washed in ice cold acetone and solubilized in 1xSB containing proteinase inhibitors and PMSF. The protein fractions were subsequently analysed by western blotting.

RNA sequencing of the samples collected across the synchronised time course of T. hominis infection

Fresh T. hominis spores were isolated from heavily infected RK cells on the day of the experiment, used to inoculate RK cells grown to 80% confluence in 150 cm² round tissue culture dishes, washed at 2h post inoculation, and grown as previously described for testing the effects of drugs across the time-course. At 3, 14, 22, 40, 70 and 96 h post-inoculation, cells from two culture dishes were separately harvested in RNAprotect cell reagent and immediately frozen at -80°C. A third culture dish at each time point contained round cover slips, which were used for IFA to monitor the progression of infection. Total RNA was extracted from the cell pellet using a combination of TRIzol reagent and bead beating, as previously described [46]. Paired-end libraries were prepared using the Illumina mRNA sample preparation kit and sequenced on two lanes of an Illumina HISeq 2500. Transcript abundances were calculated with kallisto [47] and analysed using sleuth [48]. The sequence data were submitted to GenBank SRA database: RNA-Seq data: BioProject PRJNA278775 with the BioSample accession numbers SAMN11265032-SAMN11265043 (one accession for each of the two samples per time point post infection) [5,6].

Bioinformatic identification of microsporidian protein homologues

Protein similarity networks were generated out of 184,600 protein sequences encoded in 9 fungi and 21 microsporidia/cryptomycota genomes (downloaded from the NCBI server); with BLAST [49], and quick edge file optimization implemented in EGN [50] using the following thresholds: BLAST (a = 12, e = 1e-05), simple link parameter (E-value threshold = 1e-05, hit identity threshold = 20%, minimum hit coverage = 20%, without the best reciprocal hit condition enforcement), and with the quick edge file creation (E-value threshold = 1e-05, hit identity threshold = 20%, minimum hit coverage = 20%, minimal hit length = 30 amino acids). The 12,022 identified networks were analysed using igraph [51] and dnet [52]⁠ packages in R [53], and visualized in Cytoscape [54]. Communities of the densely connected subgraphs within the largest protein network that contained 61,903 protein sequences (the second largest network contained 734 protein sequences), were isolated using the walktrap community detection algorithm implemented in the igraph package [51]. Sequences of the proteins involved in the metabolic pathways of interest in the analysed fungi were downloaded from KEGG. Networks containing proteins of interest were extracted and annotated using eggnog-Mapper v1 [55]. Heatmaps of the protein networks were generated using gplots package (https://CRAN.R-project.org/package=gplots) ⁠within R.

Microsporidian species tree

To generate a species tree of the analysed Microsporidia and Fungi, single orthologues from 51 networks that contained only a single protein sequence from each species, were aligned using MAFFT [56], trimmed with trimAl [57], and concatenated with the Nexus module from the Biopython package [58] to a final alignment of 8076 amino-acid sites. Best fitting phylogenetic model was selected using ProtTest3 [59], and phylogenetic tree was generated under PROTGAMMALG model with 100 bootstrap replicates in RaxML [60]. Phylogenetic trees were visualised and annotated using FigTree [61].