Host PI(3,5)P₂ Activity Is Required for Plasmodium berghei Growth During Liver Stage Infection

Abstract

Malaria parasites go through an obligatory liver stage before they infect erythrocytes and cause disease symptoms. In the host hepatocytes, the parasite is enclosed by a parasitophorous vacuole membrane (PVM). Here, we dissected the interaction between the Plasmodium parasite and the host cell late endocytic pathway and show that parasite growth is dependent on the phosphoinositide 5‐kinase (PIKfyve) that converts phosphatidylinositol 3‐phosphate [PI(3)P] into phosphatidylinositol 3,5‐bisphosphate [PI(3,5)P₂] in the endosomal system. We found that inhibition of PIKfyve by either pharmacological or non‐pharmacological means causes a delay in parasite growth. Moreover, we show that the PI(3,5)P₂ effector protein TRPML1 that is involved in late endocytic membrane fusion, is present in vesicles closely contacting the PVM and is necessary for parasite growth. Thus, our studies suggest that the parasite PVM is able to fuse with host late endocytic vesicles in a PI(3,5)P₂‐dependent manner, allowing the exchange of material between the host and the parasite, which is essential for successful infection.

Malaria is one of the most deadly diseases worldwide, killing nearly 700 000 people per year, mainly children under the age of 5 1. The infection cycle commences with the bite of a female Anopheles spp. mosquito, which, while feeding, inoculates Plasmodium spp. parasites into the skin of a vertebrate host. The parasite, in the form of a sporozoite, then travels from the skin dermis to the blood, where it is carried with the blood flow until it reaches the liver 2-4. In the liver stage of the life cycle, the sporozoites glide through several hepatocytes before actively infecting a single one and forming a parasitophorous vacuole (PV) 5, 6. It is inside this compartment that each sporozoite lives and replicates through schizogony into thousands of exo‐erythrocytic forms (EEFs), the merozoites, which are capable of infecting red blood cells (RBCs) 7. Strikingly, the parasite is able to switch from a form that interacts with one of the most metabolically active cells in the body, to a form that is able to develop inside a RBC. However, contrary to the interaction with RBCs, the development of Plasmodium parasites within hepatocytes remains poorly characterized. Importantly, since the liver stage is clinically silent, it is considered an ideal point for prophylactic intervention, including vaccination 8. Indeed, it has been shown that it is possible to induce protective immune responses against the liver stage of the life cycle, especially during late liver stage development 9. Thus, efforts to better understand this stage are essential.

Recently, we and others have shown that in the liver stage, EEFs are able to interact with late endosomes/lysosomes 10 and to scavenge mitochondrial lipoic acid from host mitochondria 11. Moreover, an interaction with the endoplasmic reticulum has been described, although this finding remains controversial 12, 13.

Phosphoinositides (PIs) are a family of interconvertible phospholipids that play pivotal roles in membrane trafficking and cell signaling. They are key components of all eukaryotic cell membranes and, although present in minor concentrations, PIs are specifically localized to different organelles, contributing to the identity of intracellular membranes and establishing the so‐called ‘PI code’ 14, 15. For this reason, they serve as versatile regulators in organelle‐specific interactions. The seven known PIs stem from the same phospholipid backbone, phosphatidylinositol that can be reversibly phosphorylated at the 3, 4 and/or 5 positions of the inositol ring through the action of different kinases and phosphatases 16. Each one of these PIs exerts its regulatory role by specifically recruiting different effector proteins that mediate different functions. One of the best studied PI species is phosphatidylinositol 3‐phosphate [PI(3)P]. PI(3)P was shown to be enriched in the membranes of early endosomes where it regulates endosome homotypic fusion by recruiting early endosome‐antigen 1 (EEA1) and other Rab5 effector proteins. Furthermore, through the action of phosphoinositide 5‐kinase (PIKfyve), a type III PI kinase that is also recruited to PI(3)P‐enriched endosomal membranes in a Rab5‐dependent manner 17, PI(3)P can be phosphorylated at position 5, resulting in phosphatidylinositol 3,5‐bisphosphate [PI(3,5)P₂] that localizes to late endosomes/lysosomes 17-19.

Interestingly, it has been shown that several bacteria are able to modulate host PI(3)P metabolism either by avoiding its recruitment to the microbial vacuole, as in the case of Mycobacterium tuberculosis, or in contrast by actively recruiting it, as in the case of Salmonella 20. Moreover, parasites like Toxoplasma gondii recruit PI(3)P to the neighborhood of the vacuole, as part of a fundamental role for autophagy in parasite replication 21, while in Trypanosoma cruzi infection, an essential role for PI(3)P was described during host cell invasion 22. Finally, PIKfyve activity was shown to have an essential role in Salmonella replication 23.

In order to elucidate the mechanism by which Plasmodium subverts the host membrane trafficking machinery, we analyzed the dynamics of PIs, in particular PI(3)P and its conversion to PI(3,5)P₂, by PIKfyve, in infected hepatocytes. We show that host PIKfyve activity is crucial for Plasmodium replication inside host hepatocytes, since parasite growth is impaired when host kinase activity is disrupted during late liver stage. Interestingly, we also found that the PI(3,5)P₂ effector protein TRPML1 partially colocalizes with the parasitophorous vacuole membrane (PVM) and is required for efficient parasite growth. Thus, we propose that PI(3)P is converted to PI(3,5)P₂ by PIKfyve in the neighborhood of the PV, and that this conversion is necessary for the fusion of late endocytic vesicles with the PV.

Results

PI(3)P transiently associates with Plasmodium berghei vacuoles

We have previously shown that Plasmodium berghei sporozoites interact with the host late endocytic pathway 10. Since PIs are major regulators of this pathway, we hypothesized that sporozoites are also able to interact with and/or subvert host PIs. To investigate this, we determined the localization of PI(3)P, a known early endosome marker 24 in P. berghei‐infected Hepa 1‐6 mouse hepatoma cell line. In order to visualize the localization of host PI(3)P, a well‐described PI(3)P‐binding probe, consisting of two FYVE domains from the hepatic growth factor‐regulated tyrosine kinase substrate (Hrs) fused with GFP (GFP‐2xfyve^Hrs) 25 was used in live cell imaging studies. To increase the likelihood of observing a parasite within a GFP‐2xfyve^Hrs‐expressing cell, we used an adenovirus‐based expression system that has been shown to be an efficient method for gene delivery into hepatocytes 26. Interestingly, a transient accumulation of PI(3)P around the P. berghei vacuole could be observed, at 16 h post‐infection (hpi), the onset of parasite replication (Figure 1A and Video S1, Supporting Information). To compare the kinetics of association of PI(3)P with the recruitment of known endosomal markers, we also analyzed by live cell imaging the distribution of Rab5, an early endosomal marker and Rab7, a late endosomal marker. The results show that Rab5 does not constitutively associate with the parasite (Figure 1B), whereas Rab7 is clearly seen persistently surrounding the parasite (Figure 1C), at 16 hpi. Nevertheless, we cannot exclude that a transient interaction with Rab5‐positive endosomes also occurs. Importantly, these results are in accordance with our previous studies showing that host early endosomes do not accumulate around the parasites, whereas late endosomes/lysosomes associate with the PV throughout liver stage infection 10. Since PIs can be rapidly phosphorylated to a different PI species, by the action of different kinases, these results suggest that PI(3)P could be converted to PI(3,5)P₂ in the late endocytic vesicles that are present in the vicinity of the PV.

Figure 1

PI(3)P vesicles transiently associate with Plasmodium berghei vacuoles during liver stage infection. Time‐lapse images of live cell confocal microscopy of Hepa 1‐6 cells transduced with (A) the PI(3)P probe, GFP‐2xfyve^Hrs (B) GFP‐Rab5 or (C) GFP‐Rab7 (green) and infected with RFP‐P. berghei parasites for 16 hpi (red). Scale bars, 10 µm.

PIKfyve is present in the vicinity of P. berghei vacuoles

In mammalian cells, the conversion of PI(3)P to PI(3,5)P₂ is mediated by PIKfyve 17-19, 27. To investigate if PI(3)P is converted to PI(3,5)P₂ in the late endocytic vesicles that surround the parasite 10, we assessed the localization of PIKfyve in Hepa 1‐6 cells transduced with adenoviruses expressing GFP‐PIKfyve and infected with GFP‐P. berghei. Notably, the latter exhibited only a faint GFP signal that became undetectable under the conditions and settings used. The infection was terminated at different time‐points and following staining with the PVM marker UIS4, a confocal microscopy analysis was performed (Figure 2). Interestingly, we observed several GFP‐PIKfyve‐positive vesicles in the vicinity of the parasite. Nevertheless, no specific accumulation of the kinase around the parasites was detected, in contrast with other markers such as GFP‐Rab7 (Figure 1B), CD63 and LAMP1 10. This difference can be explained by the fact that PIKfyve binds to PI(3)P on early endosomes and is highly mobile and dynamically associated with membranes 27, 28. Thus, although the distribution of PIKfyve was not restricted to the vicinity of the parasite, the presence of the kinase in vesicles in the neighborhood of the PV can explain why PI(3)P is only transiently, rather than permanently, detected around the PV, and further suggests a mechanism by which PI(3)P is converted to PI(3,5)P₂ in the late endocytic vesicles surrounding the parasite.

Figure 2

PIKfyve‐positive vesicles are present in the vicinity of Plasmodium berghei vacuoles throughout liver stage infection. Hepa 1‐6 cells were transduced with GFP‐PIKfyve (green), infected with GFP‐P. berghei parasites and stained with anti‐UIS4 antibody (red). Cells were fixed at different times post‐infection (6 hpi, 16 hpi, 24 hpi, 30 hpi and 40 hpi) and prepared for confocal microscopy analysis. Nuclei were labeled with DAPI (blue). Scale bars, 10 µm.

Impaired host PIKfyve kinase activity delays P. berghei growth

In order to address the functional relevance of the transient interaction between PI(3)P and the PV as well as the role of PIKfyve during parasite liver stage growth, we used a specific PIKfyve inhibitor, namely YM201636 29, 30. Since the Plasmodium genome does not contain any PIKfyve ortholog 31, we reasoned that this inhibitor does not directly interfere with the parasite PI metabolism. To confirm this, we tested the YM201636 compound in infected red blood cells (iRBCs) and observed that parasite maturation in vitro and survival were not affected (Figure S1). To address the effect of the inhibitor in parasite replication in Hepa 1‐6 cells, two different concentrations were used: 1 µm and 0.2 µm. At 1 µm, Hepa 1‐6 cells showed vacuolation after 1 h of adding the inhibitor, indicating that the formation of PI(3,5)P₂ was being impaired 29, 30, 32. Moreover, at lower concentrations of YM201636, where PIKfyve‐catalyzed synthesis of PI(5)P is preferentially inhibited over that of PI(3,5)P₂, the vacuolation phenotype was not observed 32, even after 24 h of incubation (Figure S2A,B). Importantly, cell viability at both concentrations used was not compromised (Figure S2C). Next, P. berghei‐infected Hepa 1‐6 cells were treated with either 0.2 µm or 1 µm of YM201636 at 16 hpi, a stage when parasite replication commences 10. The infection was stopped at 48 hpi, by fixation and samples were processed for immunofluorescence analysis. Parasites were labeled using anti‐P. berghei Hsp70 antibody, which stains the parasite cytoplasm, and the size of parasites was measured. Strikingly, parasites in cells treated with 1 µm of YM201636, but not with 0.2 µm, showed a significant decrease in size, when compared to control samples (Figure 3A). Notably, while the size of the parasites decreased, the number of parasites per microscope field was not affected in cells treated with 1 µm of YM201636, indicating that at this concentration the inhibitor does not interfere with parasite survival (Figure S3).

Figure 3

Host PIKfyve inhibition by YM201636 delays parasite growth. A) Hepa 1‐6 cells, infected with GFP‐P. berghei parasites, were treated with 1 µm or 0.2 µm of YM201636 (YM) from 16 hpi to 48 hpi. Where indicated, the inhibitor was washed out at 48 hpi and infection was allowed to proceed until 60 hpi. Widefield microscopy images of parasites, stained with anti‐GFP and anti‐UIS4 antibodies (green), were acquired and sizes were measured using ImageJ software (***p < 0.0001; n.s., non‐significant). Representative confocal microscopy images are shown for each condition. Nuclei were labeled with DAPI (blue). B) Mouse primary hepatocytes, infected with GFP‐P. berghei parasites, were treated with 1 µm of YM201636 (YM) from 16 hpi to 44 hpi and parasite sizes measured as in (A). Representative confocal microscopy images are shown for each condition (***p < 0.0001). C) Hepa 1‐6 cells, infected with Plasmodium yoelii parasites, were treated with 1 µm of YM201636 from 16 hpi to 44 hpi and parasite sizes measured as in (A). Representative confocal microscopy images are shown for each condition. Parasites were stained at 44 hpi with anti‐Hsp70 antibody (green) and nuclei with DAPI (blue) (***p < 0.0001). Scale bars, 10 µm.

To better understand the effect of PIKfyve inhibition in parasite growth, we allowed the parasites to grow from 16 hpi to 60 hpi in the presence of the inhibitor, or the inhibitor was washed out at 48 hpi (Figure 3A). In both cases, parasites were able to continue growing, showing that parasite growth was delayed but not blocked, even in the continuous presence of the inhibitor. Importantly, in the case where the inhibitor was not washed out, enlarged vacuoles were detected by LAMP1 staining, indicating that the inhibitor was exerting its effects (Figure S4A). Furthermore, we observed that there was no significant size difference between control, 0.2 µm and 1 µm of YM201636 at 60 hpi. At this time point, the parasites reach their maximum growth limit and their size starts to diminish, in an asynchronous way, as some parasites start to release merosomes filled with blood stage infective forms, the merozoites (Figure S4B,C) 41, 7. However, although not significantly, at 60 hpi and 1 µm of YM201636, the size of the parasites tends to be smaller compared to control or cells treated with 0.2 µm of YM201636, suggesting that the recovery is delayed. Notably, at lower concentrations of YM201636, parasite size is not affected even when cells are treated for longer periods (16–60 hpi), suggesting a predominant role of PI(3,5)P₂ over PI(5)P (Figure 3A). Together, these results indicate that the inhibition of host PIKfyve function by a specific chemical inhibitor delays parasite growth but parasites remain viable and able to continue growing.

Importantly, a similar decrease in parasite growth was observed when isolated mouse primary hepatocytes were infected with GFP‐P. berghei and treated with 1 µm of YM201636 (Figure 3B) or when another rodent malaria parasite species, P. yoelli, was used (Figure 3C), suggesting that our findings are not species‐specific and are physiologically relevant.

Non‐pharmacological inhibition of host PIKfyve decreases parasite growth

In order to confirm our observations with the PIKfyve inhibitor, we first attempted to use either siRNA or shRNA against host PIKfyve. However, only 20% silencing was achieved as assessed by RT‐qPCR. In order to circumvent this, three additional approaches were used. Firstly, we performed a prolonged overexpression (48 h) of GFP‐2xfyve^Hrs, the PI(3)P‐binding domain, which blocks PIKfyve activity by sequestering PI(3)P. Using this method, the FYVE domain‐directed binding of PIKfyve to PI(3)P and consequent conversion to PI(3,5)P₂ become impaired 33, 34. Secondly, we overexpressed the catalytically inactive mutant GFP‐PIKfyve^K1831E that can bind PI(3)P but is unable to catalyze the shift to PI(3,5)P₂, which induces a dominant‐negative effect 17, 34, 35. We confirmed that the blockade of PIKfyve by both approaches induces vacuolation, similar to that seen with the PIKfyve inhibitor, without significantly affecting cell viability or replication (Figure S5A,B,D). Next, Hepa 1‐6 cells were transduced with adenovirus vectors encoding GFP‐2xfyve^Hrs, GFP‐PIKfyve^wt or GFP‐PIKfyve^K1831E, at 1 hpi. The infection was allowed to proceed until 48 hpi, and the cells were processed for immunofluorescence microscopy to analyze parasite size. Parasites were stained using anti‐UIS4 or anti‐Hsp70 antibodies. Strikingly, using these methods to inhibit PIKfyve, we observed a significant decrease in parasite size when compared to non‐transduced cells or cells transduced with adenoviruses encoding GFP alone or GFP‐PIKfyve^wt (Figure 4A–D).

Figure 4

Non‐pharmacological inhibition of host PIKfyve activity impairs parasite growth. A) Hepa 1‐6 cells were infected with GFP‐P. berghei parasites and transduced with adenoviruses encoding either GFP as a control or GFP‐2xfyve^Hrs, at 1 hpi. The infection was stopped at 48 hpi. Widefield microscopy images of parasites were acquired and sizes were measured using ImageJ software (***p < 0.0001; n.s., non‐significant). B) Representative confocal microscopy images of parasites for each condition are shown. PVMs were stained with anti‐UIS4 antibody (red) and nuclei were labeled with DAPI (blue). GFP and GFP‐2xfyve^Hrs are shown in green. C) Hepa 1‐6 cells were infected with GFP‐P. berghei parasites and either non‐transduced or transduced with adenoviruses encoding control GFP‐PIKfyve^wt or the catalytically inactive dominant‐negative mutant GFP‐PIKfyve^K1831E, at 1 hpi. The infection was stopped at 48 hpi and parasite sizes measured as in (A) (***p < 0.0001; n.s., non‐significant). D) Representative confocal microscopy images of parasites for each condition are shown. The parasites were stained with anti‐Hsp70 antibody (red) and nuclei were labeled with DAPI (blue). GFP‐PIKfyve^wt and GFP‐PIKfyve^K1831E are shown in green. E) PIKfyve^flox/flox MEFs transduced with adenoviruses encoding either control GFP or GFP‐Cre to knockdown (KD) PIKfyve were infected with GFP‐P. berghei parasites. Infection was stopped at 50 hpi and parasite sizes measured as in (A) (***p < 0.0001; n.s., non‐significant). F) Representative confocal microscopy images of parasites for each condition are shown. The parasites were stained with anti‐Hsp70 (red) and nuclei were labeled with DAPI (blue). GFP and GFP‐Cre are shown in green. Scale bars, 10 µm.

Finally, we induced the knockdown (KD) of PIKfyve in mouse embryonic fibroblasts (MEFs), upon Cre‐recombinase overexpression 36. For this, PIKfyve^flox/flox MEFs were transduced with GFP‐Cre or GFP control adenoviruses and transduction was allowed to proceed for 72 h before infection with GFP‐P. berghei. At 50 hpi, parasite size was quantified by microscopy by staining UIS4. As expected, enlarged vacuoles accumulated in MEFs where the PIKfyve gene expression was disrupted, although the degree of vacuolation varied, since the expression levels of the Cre‐recombinase differ from cell to cell (Figure S5C). Strikingly, parasite size was significantly decreased in PIKfyve KD cells when compared to control cells (Figure 4E,F).

Thus, these results corroborate the observations derived from the experiments with the PIKfyve inhibitor and highlight the importance of host PIKfyve for Plasmodium liver growth.

The interaction of late endocytic vesicles with P. berghei parasites is abrogated when PIKfyve is inhibited

Previous results from our group have shown that late endosomes/lysosomes surround P. berghei parasites throughout liver stage infection and that interaction with late endocytic vesicles is required for parasite growth 10. Importantly, others have shown that there is an impairment of the late stages of maturation of late endocytic organelles when the function of PIKfyve and production of PI(3,5)P₂ are perturbed, with consequent accumulation of PI(3)P in the membrane of early endosomes 17, 34. Therefore, we characterized in detail the interaction of PI(3)P and late endocytic organelles with the PV, upon disruption of the host PIKfyve function. First, we analyzed the interaction of PI(3)P with the PV when PIKfyve was inhibited. For that, Hepa 1‐6 cells were transduced with adenoviruses encoding GFP‐2xfyve^Hrs, 24 h before infection with GFP‐P. berghei for 16 h. This approach mimics the prolonged overexpression of GFP‐2xfyve^Hrs, described above, which blocks the PIKfyve kinase activity and PI(3)P conversion. As expected, the results show that the PI(3)P‐positive enlarged vacuoles start to accumulate around the PV upon PIKfyve inhibition, further suggesting that under normal conditions PI(3)P is converted to PI(3,5)P₂ in the neighborhood of the parasite (Figure 5A). Similar results were obtained when P. berghei‐infected Hepa 1‐6 transduced at 16 hpi with adenovirus vectors encoding GFP‐2xfyve^Hrs (low expression levels to avoid PIKfyve displacement from PI(3)P‐enriched membranes, Ref. 34) were treated with 1 µm of YM201636, and infection was stopped at 30 hpi, in the middle of the replication phase. The results showed again a progressive enlargement and a slight increase in the number of PI(3)P‐positive vacuoles throughout the cell, a consequence of impaired conversion of PI(3)P to PI(3,5)P₂, but also in the close vicinity of the PV (Figure 5B–D).

Figure 5

PI(3)P‐positive enlarged vesicles surround parasitophorous vacuoles when PIKfyve is inhibited. A) Hepa 1‐6 cells were transduced with adenoviruses encoding GFP‐2xfyve^Hrs, 24 h prior to infection with GFP‐P. berghei parasites for 16 hpi. Parasites were stained with anti‐UIS4 antibody (red) and nuclei were labeled with DAPI (blue). Four representative confocal microscopy images are shown. B) Hepa 1‐6 cells were infected with P. berghei parasites, transduced with adenoviruses encoding GFP‐2xfyve^Hrs (green) together or not with 1 µm of YM201636 (YM). Infection was stopped at 30 hpi and samples were prepared for confocal microscopy. The PVM was stained with anti‐UIS4 antibody (red) and nuclei with DAPI (blue). C) The area occupied by PI(3)P‐positive vesicles within 2 µm from the PVM was quantified using ImageJ software (***p < 0.0001). D) The number of PI(3)P‐positive vesicles within 2 µm from the PVM was quantified using ImageJ software (n.s., non‐significant). Scale bars, 10 µm.

Next, we addressed the interaction of the PV with late endocytic vesicles, upon blocking PIKfyve kinase activity with the YM201636 compound. In control infected cells, LAMP1 and CD63 were distributed in a rim around the PV, which colocalized with the PVM, labeled by UIS4 (Figure 6A,B). Thus, in order to evaluate if PIKfyve inhibition, and consequently the blockade of PI(3)P to PI(3,5)P₂ conversion, could perturb the interaction of late endocytic vesicles with the PV, we scored for the presence or absence of a LAMP1/CD63‐positive rim under conditions of PIKfyve inhibition. We decided to look at 16 hpi, the onset of parasite replication, since at late time points, such as 48 hpi, the parasite occupies the majority of the cytoplasm and the host cell organelles become concentrated in a small area of the cell. Moreover, at 16 hpi there are no differences in parasite size between control and YM201636‐treated cells that could account for differences in aggregation of host late endosomes/lysosomes around the PV (Figure S6). Thus, Hepa 1‐6 cells were treated with either 1 µm or 0.2 µm of YM201636 at 1 hpi, and the infection was allowed to proceed until 16 hpi. Host cell late endosomes/lysosomes were stained using anti‐LAMP1 or anti‐CD63 antibodies and PVMs with anti‐UIS4 antibody. We observed that the colocalization of late endosome/lysosome markers with the PVM marker in a rim around parasites was impaired when the function of PIKfyve and production of PI(3,5)P₂ were disrupted pharmacologically by the PIKfyve inhibitor YM201636 (Figure 6A,B).

Figure 6

PIKfyve inhibition interferes with the interaction between Plasmodium berghei parasites and late endosomes/lysosomes. A and B) Hepa 1‐6 cells were infected with GFP‐P. berghei parasites for 16 h and treated with 1 µm of YM201636 (YM), at 1 hpi. The PVMs were stained with anti‐UIS4 antibody (red), late endosomes/lysosomes with anti‐LAMP1 (A) or anti‐CD63 (B) antibodies (green) and nuclei with DAPI (blue). Parasites were scored for the presence or absence of a LAMP1‐ or CD63‐positive rim around the PV. Representative confocal microscopy images are shown below each condition (***p < 0.0001; n.s., non‐significant). Scale bars, 10 µm. C and D) Hepa 1‐6 cells were transduced with adenoviruses encoding GFP or GFP‐2xfyve^Hrs, 24 h prior to infection with GFP‐P. berghei parasites for 16 h. The PVMs were stained with anti‐UIS4 antibody and late endosomes/lysosomes with anti‐LAMP1 (C) or anti‐CD63 antibodies (D). Images were acquired in a confocal microscope. Parasites were scored for the presence or absence of a LAMP1‐ or CD63‐positive rim around the PV (*p < 0.05; n.s., non‐significant). E and F) Hepa 1‐6 cells were transduced with adenoviruses encoding GFP‐PIKfyve^wt or GFP‐PIKfyve^K1831E, 24 h prior to infection with GFP‐P. berghei parasites for 16 h. The PVMs were stained with anti‐UIS4 and late endosomes/lysosomes with anti‐LAMP1 (E) or anti‐CD63 antibodies (F). Images were acquired in a confocal microscope. Parasites were scored for the presence or absence of a LAMP1‐ or CD63‐positive rim around the PV (*p < 0.05; n.s., non‐significant).

Moreover, we scored for the presence of a LAMP1/CD63‐positive rim in infected cells expressing GFP‐2xfyve^Hrs for 40 h and compared to infected cells expressing GFP, or in GFP‐PIKfyve^K1831E‐expressing infected cells compared to infected cells expressing GFP‐PIKfyve^wt (Figures 6C–F and S7). We were able to detect a significant decrease in the formation of a rim of LAMP1 around the PV, in the case of the former and CD63 in the case of the latter. Nevertheless, even in the cases where the differences are not significant, the formation of the rim was still noticeably decreased, when PIKfyve function was impaired.

Since in the absence of functional PIKfyve kinase activity the maturation of late endocytic organelles is arrested, these results suggest that late endosomes/lysosomes are unable to interact with the PV in the absence of PI(3,5)P₂ formation. Thus, our observations point toward a fundamental role for host PIKfyve and consequently, the presence of PI(3,5)P_2, in the interaction of the PV with late endocytic organelles.

TRPML1 partially colocalizes with the parasitophorous vacuole membrane and is required for parasite growth

Mucolipin transient receptor potential 1 (Mucolipin1 or TRPML1) is an endolysosomal protein and a PI(3,5)P₂ effector 37, 38. When TRPML1 binds to PI(3,5)P₂, it activates and promotes membrane fusion by stimulating intraluminal Ca²⁺ release. To investigate if TRPML1 is present in the vesicles that surround the PV, we transfected Hepa 1‐6 cells with GFP‐TRPML1 and infected these cells with P. berghei parasites. Infected cells were fixed at 16 hpi, 24 hpi, 44 hpi and 48 hpi and samples were processed for immunofluorescence analysis. PVMs were labeled using anti‐UIS4 antibody. Strikingly, GFP‐TRPML1 was found in vesicles surrounding the PV, with partial colocalization being more apparent at 16 hpi and 24 hpi (Figure 7A). To address the functional relevance of this association, Hepa 1‐6 cells were stably transduced with lentiviruses encoding shRNAs targeting TRPML1 and further infected with P. berghei parasites. Two stable cell lines with different silencing efficiencies, called shTRPML1‐1 and shTRPML1‐2 (Figure 7C) were used along with a non‐targeting shRNA control cell line. The infection was stopped at 48 hpi by fixation and samples were processed for immunofluorescence analysis. PVMs were stained using anti‐UIS4 antibody and the size of parasites was quantified. Strikingly, we observed that parasite growth was significantly impaired when the expression level of TRPML1 was reduced by 80% (Figure 7B). Therefore, these results suggest that host late endocytic vesicles are able to fuse with P. berghei PVM in a PI(3,5)P₂‐ and TRPML1‐dependent manner.

Figure 7

Host TRPML1 partially colocalizes with the parasitophorous vacuole membrane and is required for normal parasite growth. A) Hepa 1‐6 cells were infected with GFP‐P. berghei parasites and transfected with GFP‐TRPML1. Infection was stopped at 16 hpi, 24 hpi, 44 hpi or 48 hpi and samples prepared for confocal microscopy analysis. The PVM was stained with anti‐UIS4 antibody (red). GFP‐TRPML1 is shown in green and nuclei were labeled with DAPI (blue) (confocal Z‐projections). Scale bars, 10 µm and 2 µm (insets for 16 hpi and 24 hpi). B) TRPML1‐deficient Hepa 1‐6 cells (TRPML1‐1 or TRPML1‐2) were infected with GFP‐P. berghei parasites and the infection was stopped at 48 hpi. Widefield microscopy images of parasites were acquired and sizes measured using ImageJ software (**p < 0.001; n.s., non‐significant). Representative images of parasites are shown below each condition. C) Hepa 1‐6 cells were stably transduced with shRNA‐encoding lentiviruses targeting mouse TRPML1 (TRPML1‐1 or TRPML1‐2). The silencing efficiency was measured by RT‐qPCR using specific primers.

Discussion

The findings described here extend our previous studies showing that P. berghei interacts with the late endocytic pathway of host hepatocytes and that this interaction is an essential step for the fast growth of parasites during the late liver stage of infection 10. While this manuscript was under review, another interesting study using mostly live cell imaging of transgenic parasites expressing the PVM markers UIS4 and IBIS1, showed that the PVM extends throughout the host cell cytoplasm and colocalizes with late endosomal markers such as CD63 and LAMP1, further supporting our model 13. In particular, the work presented here suggests a model where the parasite is able to specifically recognize and fuse with host late endocytic organelles in a PI(3,5)P₂‐ and TRPML1‐dependent manner. First, albeit rarely, an association of PI(3)P with the PV was seen at 16 hpi, a time‐point when the parasite starts to replicate 10. Moreover, PIKfyve‐positive vesicles were detected in the vicinity of the PV, suggesting that PI(3)P is converted to PI(3,5)P₂, which could explain why PI(3)P is only transiently detected there. Strikingly, when host PIKfyve function was impaired either pharmacologically, by dominant‐negative interference, by PI(3)P substrate sequestration or genetically, we observed a delay in parasite growth. These results are in agreement with our previous observations showing that when acidification and consequently fusion of host late endocytic compartments was inhibited using ammonium chloride or concanamycin A, parasite growth was significantly impaired 10. Taken together, these data reinforce the notion that an interaction between growing parasites and the host late endocytic pathway is essential for parasite growth.

We also found that the PI(3,5)P₂ effector TRPML1 38, 39 is present in the vesicles surrounding the PVM and also partially colocalizes with it. TRPML1 is a Ca²⁺ channel that has known functions in late endocytic membrane fusion 37-39. Moreover, it mediates intraluminal Ca²⁺ release, which is necessary to trigger heterotypic and homotypic fusion of late endocytic vesicles. Furthermore, it has been shown to be necessary for the fusion of amphisomes, which are organelles that are formed from the convergence between the late endocytic and autophagic pathways 40. Thus, we speculate that, by binding to PI(3,5)P₂ in the host late endocytic vesicles, TRPML1 is necessary for their fusion with the PVM. This would also explain why the distribution of host LAMP1/CD63 on a rim that colocalizes with the PVM is lost when the conversion of PI(3)P to PI(3,5)P₂ is blocked. Importantly, under these conditions we also observed an accumulation of late endocytic vesicles around the PV, further supporting the existence of a blockade in organelle fusion. The TRPML1 channel belongs to a family of proteins that, in mammals, is composed of two other members, TRPML2 and TRPML3. Interestingly, TRPML1 and TRPML2 are highly expressed in the liver, which may reflect their importance in this organ 41. Nevertheless, functional redundancy between TRPML1 and TRPML2 might also occur, besides the unique functions they fulfill, which may explain why there was not a complete blockade of parasite growth when TRPML1 was silenced by 80%. Alternatively, the remaining 20% expression level could explain the incomplete effect.

A role for TRPML1 in a host‐pathogen interaction is a novel finding, but the fact that functional PIKfyve and PI(3,5)P₂ are necessary for the fusion of the Salmonella vacuole with macropinosomes and for intracellular replication of the bacteria 23, suggests a possible common mechanism of interaction between intracellular pathogens and their host cells. However, it should be emphasized that the approaches that affect PIKfyve protein levels or enzymatic activity also impair synthesis of PI(5)P along with that of PI(3,5)P₂ 18, 32, 46. Therefore, whereas our data are consistent with a role for PI(3,5)P₂ in the reduction of parasite growth, additional inputs by PI(5)P cannot be excluded.

Hence, we hypothesize that the parasite vacuole mimics the membrane composition of a late endosome/lysosome and uses this mechanism to fuse with host late endocytic vesicles. Importantly, direct fusion with host cell organelles would provide the necessary membrane components for the fast expansion of the parasite vacuole. However, the parasite proteins required for these specific membrane interactions remain obscure, as is the mechanism by which TRPML1 is activated to promote the fusion of two membranes 38. Hence, further elucidation of the molecular players involved in these interactions will be crucial to help understand how Plasmodium parasites are able to develop so successfully within host cells.

Materials and Methods

Tissue culture and inhibitors

Mouse hepatoma cell line Hepa 1‐6 (ATTC) was generously provided by M. M. Mota (Instituto de Medicina Molecular, Lisbon). Cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Gibco/Invitrogen) supplemented with 10% fetal calf serum (FCS) (Gibco/Invitrogen), 100 U/mL penicillin and 100 µg/mL streptomycin (Gibco/Invitrogen). Cells were maintained in a humidified incubator at 37°C and 10% CO₂. Primary mouse hepatocytes were isolated from female C57BL/6 mice (8–10 weeks old) according to the protocol described 42. PIKfyve^flox/flox MEFs were isolated as described elsewhere 36. PIKfyve inhibitor YM201636 was purchased from Symansis (Cell Signaling Science).

Parasite strains and culture

Plasmodium yoelli 17X, green fluorescent protein (GFP) (parasite line 259cl2) or red fluorescent protein (RFP) (parasite line 733 cl1) expressing P. berghei ANKA salivary gland sporozoites were collected from infected female Anopheles stephensi mosquitoes. Mosquitoes were bred in the insectarium of the Instituto de Medicina Molecular, Lisbon, and were kindly provided by M. M. Mota. Infected mosquitoes were dissected between days 20 and 24 after the infectious blood meal. Mean number of sporozoites was determined using a hemocytometer.

Culture of parasite schizonts

Infected mice, at day 6 after infection, were bled and the blood used for in vitro culture for 18–20 h so that parasites could develop into schizonts. This was achieved after overnight culture at 37°C in RPMI medium containing FCS and gassed with a mixture of 10% CO₂, 5% O₂ and 85% N₂. Parasite cultures were performed in medium with 1 µm of YM201636, or with DMSO or only medium as controls. Schizonts obtained from the different culture conditions were injected into mice and parasitemia monitored by flow cytometry.

Constructs, cell transfection, virus recombination and cell transduction

For production of pENTR‐GFP‐2xfyve^Hrs construct, the 2xfyve^Hrs binding domain was excised from pEGFP‐2xfyve^Hrs using EcoRI/SalI and inserted into the modified mammalian expression vector pENTR‐GFP‐C2, previously described 43. For production of pENTR‐GFP‐PIKfyve, mouse PIKfyve was excised from pCMV5‐HA‐mPIKfyve 44 using EagI/SalI and inserted into the modified mammalian expression vector pENTR‐GFP‐C2, previously described 43. The pENTR‐GFP‐PIKfyve^K1831E was produced by site‐directed mutagenesis using the specific primers: PR1 5′‐GAGCAAATGCCTCGTTTGGAAGTC‐3′ and PR2 5′‐CAGAATGAATCTATCATCTTCGGTGGCATA‐3′. All constructs were confirmed by sequencing. The resulting plasmids were used to generate pAd adenoviral vectors and viral particles by LR recombination, according to the manufacturer's instructions (Invitrogen). Adequate virus titers to result in expression in 80–90% of cells were used. The GFP‐TRPML1 was a kind gift from Dr. Haoxing Xu (University of Michigan) and was transfected into Hepa 1‐6 cells using Lipofectamine™2000 (Invitrogen) according to the manufacturer's instructions. The shRNA‐encoding lentiviral particles to silence mouse TRPML1 were purchased from Sigma (#SHCLNV‐NM_053177). The viral particles were added to cells at a MOI of 5 and the transduced cells were selected using puromycin. Silencing efficiency was measured by real‐time quantitative PCR (RT‐qPCR) using the specific primers: forward, 5′‐AACACCCCAGTGTCTCCAG‐3′ and reverse, 5′‐GAATGACACCGACCCAGACT‐3′. Total RNA was extracted using the RNeasy Mini Kit (Qiagen) and 1 µg of total RNA was reversed‐transcribed using SuperScriptII RNase H‐reverse transcriptase (Invitrogen) and random hexamer primers (Invitrogen). Reactions were incubated at 65°C for 5 min, then at 25°C for 10 min, followed by 42°C for 50 min and finally at 70°C for 15 min. RT‐qPCR was performed in an ABI Prism 7900HT using ABI Power SYBR Green PCR Master Mix. RT‐qPCR reactions were performed in triplicates. The gene expression was calculated relative to control wells and normalized for GAPDH expression.

Antibodies and immunostaining

Isolated primary hepatocytes and Hepa 1‐6 cells (9 × 10⁴) were seeded on coverslips on 24‐well plates one day before the addition of ∼4 × 10⁴ freshly dissected sporozoites. Cells were centrifuged at 1600 × g for 5 min at 4°C to ensure the contact of the parasites with the cells. At the appropriate times, cells were washed 2× with PBS, and fixed with 4% paraformaldehyde (PFA; Electron Microscopy Sciences) in PBS for 15 min at room temperature. Cells were blocked and permeabilized with 1% BSA (Sigma), 0.05% saponin (Sigma) and 1% FCS (Gibco/Invitrogen) in PBS for 30 min. Cells were then incubated with primary antibodies in the same buffer for 1 h. After washing in PBS, cells were incubated with secondary antibodies in the same buffer for 30 min. To visualize the nuclei, cells were incubated with DAPI (Invitrogen) for 1 min. Samples were mounted using MOWIOL mounting medium (Calbiochem). Antibodies and dilutions used: anti‐P. berghei HSP70 (2E6, kindly provided by M. M. Mota, 1:2500), anti‐UIS4 (SicGen, 20 µg/mL), anti‐GFP (Invitrogen, 20 µg/mL), anti‐LAMP1 (obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA), anti‐CD63 (MBL, 2 µg/mL). Specific Alexa fluor^® labeled secondary antibodies (Invitrogen) were all used at 1:400.

Confocal microscopy

Confocal microscopy images were acquired in a Leica TCS SP5 microscope (^©Leica Microsystems), using a 63× immersion oil lens (objective type: HCX PL APO CS, NA: 1.40–0.60). For live cell confocal imaging, Hepa 1‐6 cells (4 × 10⁴) were seeded on glass bottom culture dishes (MatTek Corporation) and infected with 5 × 10⁴ RFP‐P. berghei sporozoites. Samples were visualized using a 40× immersion oil lens (objective type: PLAN APO, NA: 1.30) of an inverted spinning disk microscope system (Andor technology™) fitted with a temperature and CO₂ control chamber. Time‐lapse images were acquired at 3 seconds per frame, during 10 min. When necessary, images were post‐processed for overall brightness and contrast and figures were assembled using ImageJ (NIH) and Photoshop (CS5, Adobe) software. The inverted spinning disk microscope was controlled by μManager software 45.

Parasite number and size quantification

To measure parasite size and number, images for each condition were acquired in a Nikon Eclipse TE2000‐S automated widefield screening microscope using a 40× air lens (CFI PLAN APO, NA: 0.95) or in a Leica TCS SP5 microscope (^©Leica Microsystems), using a 63× immersion oil lens (HCX PL APO CS, NA: 1.40–0.60). Parasite area (in µm²) was measured by drawing manually the circumference around the parasite and automatically calculated using ImageJ software (NIH). The number of parasites was counted and corrected for the number of images acquired for each condition (200–400 images) and is presented as total number of parasites per microscope field. The Nikon Eclipse TE2000‐S automated widefield screening microscope was controlled by μManager software 45.

Statistical analysis

Statistical analysis was performed using Prism software (GraphPad Software Inc.) and an unpaired Student t‐test. P values < 0.05 were considered statistically significant.

Acknowledgments

We would like to thank Dr. Maria M. Mota (Instituto de Medicina Molecular, Portugal) for advice and critical reading of the manuscript and all the members of her laboratory for mosquito rearing, antibodies and very helpful discussions during the course of this work. We also wish to thank Dr. Haoxing Xu (University of Michigan) for the GFP‐TRPML1 construct. Finally, we would like to thank Dr. Ligia A. Gonçalves (Instituto Gulbenkian de Ciência, Portugal) for valuable help with primary hepatocyte isolation. This work was supported by Fundação para a Ciência e Tecnologia (FCT, PTDC/SAU‐MII/108206/2008), National Institutes of Health DK58058 (to A. S.) and American Diabetes Association #7‐13‐BS‐161 (to A. S.). C. T. M. was funded by FCT fellowship SFRH/BD/45458/2008, M. L. S by FCT fellowship SFRH/BD/27705/2006, L. C. S. by FCT fellowship SFRH/BPD/65764/2009, and C. F. P by FCT fellowship SFRH/BD/33541/2008. The authors declare that there are no conflicts of interest.