(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Role of Rabenosyn-5 and Rab5b in host cell cytosol uptake reveals conservation of endosomal transport in malaria parasites [1]

['Ricarda Sabitzki', 'Molecular Biology', 'Immunology Section', 'Bernhard Nocht Institute For Tropical Medicine', 'Hamburg', 'Anna-Lena Roßmann', 'Marius Schmitt', 'Sven Flemming', 'Andrés Guillén-Samander', 'Hannah Michaela Behrens']

Date: 2024-06 

Vesicular trafficking, including secretion and endocytosis, plays fundamental roles in the unique biology of Plasmodium falciparum blood-stage parasites. Endocytosis of host cell cytosol (HCC) provides nutrients and room for parasite growth and is critical for the action of antimalarial drugs and parasite drug resistance. Previous work showed that PfVPS45 functions in endosomal transport of HCC to the parasite’s food vacuole, raising the possibility that malaria parasites possess a canonical endolysosomal system. However, the seeming absence of VPS45-typical functional interactors such as rabenosyn 5 (Rbsn5) and the repurposing of Rab5 isoforms and other endolysosomal proteins for secretion in apicomplexans question this idea. Here, we identified a parasite Rbsn5-like protein and show that it functions with VPS45 in the endosomal transport of HCC. We also show that PfRab5b but not PfRab5a is involved in the same process. Inactivation of PfRbsn5L resulted in PI3P and PfRab5b decorated HCC-filled vesicles, typical for endosomal compartments. Overall, this indicates that despite the low sequence conservation of PfRbsn5L and the unusual N-terminal modification of PfRab5b, principles of endosomal transport in malaria parasite are similar to that of model organisms. Using a conditional double protein inactivation system, we further provide evidence that the PfKelch13 compartment, an unusual apicomplexa-specific endocytosis structure at the parasite plasma membrane, is connected upstream of the Rbsn5L/VPS45/Rab5b-dependent endosomal route. Altogether, this work indicates that HCC uptake consists of a highly parasite-specific part that feeds endocytosed material into an endosomal system containing more canonical elements, leading to the delivery of HCC to the food vacuole.

Funding: This work was funded by the European Research Council (ERC, grant 101021493 to TS) and the DFG funded research training group GRK2771 (to TS). MS thanks for funding from the Jürgen Manchot Stiftung and AGS thanks for funding from EMBO (Postdoctoral Fellowship EMBO ALTF Number 166-2022). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we identified a P. falciparum Rbsn5-like (PfRbsn5L) protein and showed that it interacts and functions with PfVPS45 and PfRab5b in the transport of HCC to the food vacuole. Our data provide evidence that the P. falciparum Rab5-Rbsn5-VPS45 fusion complex—and thus elements of this part of the endosomal pathway—is evolutionarily conserved although the binding specificity of the PfRbsn5L FYVE domain remains unknown. Additionally, double conditional inactivation of PfRbsn5L together with a cytostomal collar protein involved in the early phase of endocytosis at the PPM suggests that the HCC-containing vesicles originated from the cytostome. Overall, our data suggest that HCCU consists of a parasite-specific initial part at the PPM that delivers endocytosed material into an endosomal system that contains more canonical aspects.

The presence and function of VPS45 in the parasite may indicate that endosomal transport in the parasite follows a more canonical pathway than endocytosis initiation at the PPM. PfVPS45, PI3P kinase, the phosphoinositide-binding protein PfPX1, host Peroxiredoxin-6, and actin [ 13 , 14 , 21 – 23 ] have so far been implicated in endosomal transport of HCC. However, the identification of proteins involved in this process is difficult because in apicomplexan parasites, many homologs of endolysosomal proteins appear to have been repurposed for functions associated with the specialized secretory organelles needed for host cell invasion [ 24 – 29 ]. Hence, similarity to endolysosomal proteins has limited predictability to identify such proteins in malaria parasites. It is for instance still unclear whether PfRab5 isoforms are involved in HCCU. P. falciparum Rab5a, initially thought to function in HCCU [ 10 ], is only essential in schizonts, not trophozoites, indicating no role in HCCU [ 30 ]. PfRab5b has been shown to localize to the parasite food vacuole and the PPM [ 31 , 32 ]. However, direct functional data is lacking for both PfRab5b and PfRab5c and it is at present unknown if they are involved in HCCU or not. In model organisms, VPS45 typically functions together with Rab5 and the Rab5-effector rabenosyn5 (Rbsn5) (in mammals) or Vac1/PEP7 (in yeast) [ 33 – 35 ] in a fusion complex important for endosome maturation [ 33 ]. However, if and which Rab5 is involved in HCCU in malaria parasites is unknown and an Rbsn5 has not been detected in the parasite’s genome.

Less is known about proteins in later phases of HCCU, the transport of internalized HCC to the parasite food vacuole. Inactivation of the parasite’s orthologue of VPS45 leads to an accumulation of HCC-filled vesicles in the parasite [ 13 ]. This indicated that PfVPS45 is involved in HCC transport, resembling the function of its orthologues in model organisms that are needed for the transport of endosomal cargo to the lysosome [ 19 , 20 ]. The HCC-filled vesicles induced after PfVPS45 inactivation are enclosed by 1 or 2 membranes, can contain smaller internal vesicles similar to endosomes in model organisms, and often harbor phosphatidylinositol 3-phosphate (PI3P) in their membrane facing the cytosol, overall suggesting endosomal characteristics [ 13 ].

Compared to eukaryotic model organisms, endocytosis in malaria blood stages faces particular challenges arising from the unique environment in which the intracellular parasite resides [ 8 ]. The parasites develop surrounded by a milieu of high protein density (mainly hemoglobin), from which they are separated not only by their plasma membrane (PPM) but also by an additional membrane, the parasitophorous vacuolar membrane (PVM). Morphological studies implicated the cytostome, an invagination of the PPM and the surrounding PVM, as the site where HCCU is initiated at the PPM [ 9 – 12 ]. However, the mechanism of how endocytic structures are formed remains unclear [ 8 ], and only recently functional data directly implicated specific parasite proteins in this process [ 3 , 13 – 15 ]. Most of these proteins are located at an electron dense collar surrounding the cytostomal neck and are involved in the early phase of HCCU (i.e., for the presumed initiation and formation of endocytic vesicular structures at the PPM) [ 3 , 4 , 15 , 16 ]. The majority of the proteins at the cytostomal collar do not resemble typical endocytosis proteins [ 3 , 15 ], indicating that the initiation of endocytosis for HCCU displays strong parasite-specific adaptations that—based on recent work—are conserved in apicomplexans [ 17 , 18 ].

Malaria caused by Plasmodium falciparum parasites remains an important cause of infectious disease-related death [ 1 ]. The pathology of malaria is caused by the asexual development of the parasite within red blood cells (RBCs) of the host. Progress in reducing the global impact of malaria has slowed recently and drug resistance was identified as one factor jeopardizing malaria control [ 1 ]. The action of several antimalarial drugs critically depends on the degradation of host cell hemoglobin in the parasite’s lysosome-like compartment termed the food vacuole [ 2 ]. The hemoglobin derives from cytosol the parasite takes up from the host cell in an endocytic process. Hemoglobin degradation products activate the current first-line drug artemisinin (and derivatives), and decreased susceptibility to these drugs is associated with a reduced host cell cytosol uptake (HCCU) [ 3 , 4 ]. The endocytosed hemoglobin is a source of amino acids for the parasite [ 5 ]. Consequently, amino acid availability is a growth restricting factor in parasites with a reduced susceptibility to artemisinin, indicating a trade-off between artemisinin susceptibility and HCCU levels [ 6 ]. HCCU is also critical for providing space for parasite growth and for the osmotic stability of the infected host cell [ 7 ]. The uptake and digestion of hemoglobin hence constitute a vulnerability for the parasite. However, despite its importance, the molecular basis for HCCU is not well understood [ 8 ].

Results

Identification of a putative P. falciparum Rbsn5-like protein If PfVPS45-dependent endosomal transport is evolutionary conserved, it is expected to also depend on an equivalent of Rbsn5 or Vac1/PEP7 which up to now had been elusive in malaria parasites. In order to identify possible Rbsn5 candidates, we conducted in silico searches. Rbsn5 from other organisms contain an FYVE-type zinc finger but only a single FYVE-domain containing protein was previously detected in the P. falciparum genome and named FYVE-containing protein (FCP) [36]. BLAST searches using human Rbsn5 (Q9H1K0) identified FCP as the top hit. However, the BLAST-detected similarity (46% identity) was restricted to 37 amino acids of the FYVE domain (score 51.2; E value 4e-07). It was therefore unclear whether FCP corresponds to the P. falciparum Rbsn5 or to a different FYVE-domain protein such as, e.g., the early endosomal antigen 1 (EEA1). To clarify this, we used HHPred [37,38] to query the P. falciparum proteome with human Rbsn5 which identified a different protein, PF3D7_1310300, as the top hit. This protein displayed similarity to HsRbsn5 over 212 (E value 5e-12) of its 247 amino acids and contains an FYVE/PHD zinc finger in its N-terminal half (Figs 1A and S1A). As the HHPred detected similarity went beyond the FYVE domain, we reasoned that this was the most likely candidate for PfRbsn5. Alignment of PF3D7_1310300 with HsRbsn5 showed that the P. falciparum protein missed the region containing the N-terminal C2H2 zinc finger and the NPF repeat region in the C-terminal half of human Rbsn5 (Fig 1A) but displayed 52.2% similarity over its entire sequence with the corresponding region of HsRbsn5, whereas FCP showed 43.4% similarity over its entire sequence with the corresponding region of HsRbsn5. PF3D7_1310300 was also the top hit if the HHPred search was repeated using yeast VAC1/PEP7 (E-value 8.2e-12) which is considered the likely single equivalents of Rbsn5 and EEA1 in yeast [33]. FCP was the second-best hit of this HHPred search but covered less sequence (E value 3.3e-10). Nevertheless, FCP and PF3D7_1310300 show only a low level of conservation between themselves. As malaria parasites contain 2 FYVE proteins with similarities to VAC1/PEP7 (FCP and PF3D7_1310300), we followed the nomenclature for mammalian cells and tentatively named the one with the higher similarity and longer region of similarity to human Rbsn5 (PF3D7_1310300) P. falciparum Rabenosyn-5-like protein (PfRbsn5L). Besides the FYVE domain, PfRbsn5L also contained a region with some similarity to the Rab5-binding domain present in other Rabenosyn5s (S1B Fig). Interestingly, a closer inspection of the amino acid sequence of the PfRbsn5L FYVE domain revealed differences to conserved PI3P binding residues although the general fold of its AlphaFold2 predicted structure [39,40] closely matched experimental FYVE domain structures (S1A and S1C Fig). The amino acid positions deviating from the consensus in the PfRbsn5L FYVE were the same that also deviated in the FYVE domain of human protrudin which—in contrast to other FYVE domains—does not bind PI3P but other phosphoinositides [41]. In contrast, the FYVE domain of FCP matched the consensus (S1A Fig). We expressed a tandem of the PfRbsn5L FYVE domain fused to mCherry in parasite also expressing the PI3P reporter P40X fused to GFP [42]. The 2xFYVE domain construct appeared uniformly distributed in the parasite cytoplasm and we did not find any accumulation at PI3P containing structures (S1D Fig). We conclude that the PfRbsn5L FYVE domain does not bind PI3P or other membrane lipids. Alternatively, it might bind rare phosphoinositide species that are not sufficiently abundant to lead to a recruitment detectable above background or if requires cooperativity with other regions or interactors of PfRbsn5L to mediate binding. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Conditional Inactivation of PfRbsn5L leads to vesicles in the parasite and parasite death. (A) Comparison of domain architecture of the putative PfRbsn5L, H. sapiens HsRbsn5, and S. cerevisiae VAC1/PEP7. BD, binding domain C2H2. (B) Live-cell microscopy images of the indicated stages of PfRbsn5L-2xFKBP-GFP-2xFKBPendo parasites (white arrow, nucleus-proximal PfRbsn5Lendo foci; light blue arrow, faint dispersed signal in the nucleus; dark blue arrow, signal at the food vacuole; purple arrow, intense focus at the food vacuole in schizont stage parasites). (C) Live-cell microscopy images of PfRbsn5L-2xFKBP-GFP-2xFKBPendo parasites, co-expressing Graspepi. Arrows show PfRbsn5Lendo foci close to nuclei and are color coded based on overlap with the closest GRASP focus. Overlap: full overlap; near: less than GRASP focus diameter apart, far: more than one focus diameter apart. Pie shows proportion of these foci (n = 45 with cells from 3 independent imaging sessions). (D) Live-cell microscopy images of PfRbsn5L (PfRbsn5L-2xFKBP-GFP-2xFKBPendo+ nmd3’NLS-FRB-mChepi parasites) using a nuclear mislocalizer (mislocalizer) [30] 1 h after induction of knock-sideway (1 h rapalog) compared to the control (control). Knock-sideway was classified (pie chart) as complete (no signal detected outside the nucleus), partial (signal in the nucleus but also at original site), or absent (no) mislocalization in n = 30 parasites from 2 independent experiments. (E) Flow cytometry-based growth curve over 2.5 growth cycles of PfRbsn5L knock sideways (rapalog) compared to the control parasites. One representative of n = 3 independent experiments, all replicas shown in (S1F Fig). (F) Representative DIC live-cell images of parasites 0 h, 2 h, 6 h, and 8 h after induction of knock-sideway of PfRbsn5L (+ rapalog) compared to control. Blue arrows, vesicular structures. (G) Quantification of number of vesicles in synchronous trophozoites 0 h, 2 h, 6 h, and 8 h after induction of PfRbsn5L knock-sideway. Data shown as superplot [77] from n = 3 independent experiments (individual experiments: blue, yellow, and black with 147, 176, and 144 parasites (small dots), respectively; average of each experiment as large dot); two-tailed unpaired t test; red lines, mean; black lines, error bar (SD); p-values indicated. Scale bars, 5 μm and 1 μm in the magnifications. Nuclei were stained with DAPI. DIC, differential interference contrast; endo, endogenous; epi, episomal; Rbsn5Lendo, 2xFKBP-GFP-2xFKBP-tagged Rbsn5L expressed from endogenous locus. The data underlying this figure can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3002639.g001

Subcellular localization and conditional inactivation of PfRbsn5L To investigate the localization and function of PfRbsn5L, we tagged the rbsn5l gene with the sequence encoding 2xFKBP-GFP-2xFKBP using the selection linked integration (SLI) system to modify the endogenous locus [30]. The resulting PfRbsn5L-2xFKBP-GFP-2xFKBPendo cell line (short PfRbsn5Lendo) (Figs 1B and S2A) showed PfRbsn5L in foci and accumulations in addition to a general cytosolic distribution. The most prominent accumulations were foci in proximity to the nucleus (Fig 1B, white arrow) that increased in number with the increasing number of nuclei during progression of the parasite blood cycle. In addition, a signal was present at the food vacuole in trophozoite stages (Fig 1B, dark blue arrow) with a more intense accumulation in proximity of the food vacuole in schizont stage parasites (Fig 1B, purple arrows). In some cells, a faint dispersed signal overlapping with the DAPI-stained nuclei was observed (Fig 1B, light blue arrow). Co-expression of a fluorescently tagged Graspepi showed that only about half of the PfRbsn5L foci at the nuclei were in close proximity or overlapped with the Golgi-apparatus (Figs 1C, white arrows and S2B) and apart from the nucleus proximal foci, PfRbsn5L foci did not regularly overlap with the ER (S2C Fig). Overall, the localization of PfRbsn5L was similar to the one we previously observed for PfVPS45 in the PfVPS45-2xFKBP-GFPendo cell line [13]. To investigate its function, we conditionally inactivated the parasite’s Rbsn5L using knock-sideways [30,43,44], an approach particularly suited for rapid inactivation of target proteins [45]. This method is based on the FRB-FKBP dimerization system. One domain of this system is fused to the protein of interest (POI) and the other domain is fused to a trafficking signal, e.g., a nuclear localization signal (NLS) (the so-called mislocalizer). Upon addition of a small ligand (rapalog), the POI and the mislocalizer dimerize and by virtue of the trafficking signal on the mislocalizer, the POI is removed from its site of action, e.g., to the nucleus if an NLS is used. We episomally expressed the mislocalizer (nmd3’1xNLS-FRB-mChepi) in the PfRbsn5L-2xFKBP-GFP-2xFKBPendo cell line. Upon addition of rapalog, PfRbsn5L was efficiently mislocalized to the nucleus within 1 h (Fig 1D). To determine the relevance of PfRbsn5L for parasite blood stage development, we monitored the parasitemia in cultures grown in presence or absence (control) of rapalog over 5 days using flow cytometry. PfRbsn5L inactivation led to a substantial growth defect in comparison to the control parasites (Figs 1E and S2D), indicating an important function of PfRbsn5L for the asexual blood stages of P. falciparum parasites. Assessment of growth in synchronous parasites showed that inactivation of PfRbsn5L in ring stages prevented development into trophozoites in the majority of parasites, whereas inactivation at the transition to the trophozoite stage resulted in a marked accumulation of aberrant late stage parasites of which most failed to give rise to new rings (S3 Fig).

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002639

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/