(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Toxoplasma gondii exploits the host ESCRT machinery for parasite uptake of host cytosolic proteins

['Yolanda Rivera-Cuevas', 'Department Of Microbiology', 'Immunology', 'University Of Michigan Medical School', 'Ann Arbor', 'Michigan', 'United States Of America', 'Joshua Mayoral', 'Department Of Pathology', 'Albert Einstein College Of Medicine']

Date: 2022-01 

Toxoplasma gondii is a master manipulator capable of effectively siphoning the resources from the host cell for its intracellular subsistence. However, the molecular underpinnings of how the parasite gains resources from its host remain largely unknown. Residing within a non-fusogenic parasitophorous vacuole (PV), the parasite must acquire resources across the limiting membrane of its replicative niche, which is decorated with parasite proteins including those secreted from dense granules. We discovered a role for the host Endosomal Sorting Complex Required for Transport (ESCRT) machinery in host cytosolic protein uptake by T. gondii by disrupting host ESCRT function. We identified the transmembrane dense granule protein TgGRA14, which contains motifs homologous to the late domain motifs of HIV-1 Gag, as a candidate for the recruitment of the host ESCRT machinery to the PV membrane. Using an HIV-1 virus-like particle (VLP) release assay, we found that the motif-containing portion of TgGRA14 is sufficient to substitute for HIV-1 Gag late domain to mediate ESCRT-dependent VLP budding. We also show that TgGRA14 is proximal to and interacts with host ESCRT components and other dense granule proteins during infection. Furthermore, analysis of TgGRA14-deficient parasites revealed a marked reduction in ingestion of a host cytosolic protein compared to WT parasites. Thus, we propose a model in which T. gondii recruits the host ESCRT machinery to the PV where it can interact with TgGRA14 for the internalization of host cytosolic proteins across the PV membrane (PVM). These findings provide new insight into how T. gondii accesses contents of the host cytosol by exploiting a key pathway for vesicular budding and membrane scission.

Intracellular pathogens exploit their host to gain the resources necessary to sustain infection; however, precisely how the intracellular parasite Toxoplasma gondii acquires essential nutrients from its host remains poorly understood. Previous work showed that T. gondii is capable of internalizing host derived cytosolic proteins and delivering them to its lysosome-like compartment for degradation. However, the mechanism by which the material is trafficked across the membrane delimiting the replicative vacuole in which the parasite resides remained unclear. Here, we report a role for the parasite effector protein TgGRA14 in the recruitment of the host ESCRT machinery and in the uptake of host cytosolic proteins. Important human pathogens have developed strategies for exploiting the host ESCRT machinery for intracellular subsistence. Our study sheds lights on the strategy used by a eukaryotic pathogen in to exploit the host ESCRT machinery for the internalization of resources from its host cell.

Funding: We gratefully acknowledge the funding support from U.S. National Institutes of Health grants T32AI007528 (Y.R.C.), F31AI152297 (Y.R.C), F31AI36401 (J.M.), T32GM007288 (J.M.), the University of Perugia Fondo Ricerca Di Base 2019 program of the Department of Chemistry, Biology, and Biotechnology (M.D.C.), the U.S. National Institutes of Health grant T32AI007414 (E.B.O.), University of Michigan Life Sciences Fellowship Program (E.B.O.), U.S. National Institutes of Health grants R37AI071727 (A.O.), R01AI158501 (S.L.), and R01DK120623 (J.Z.S.), the University of Michigan Institute for Clinical and Health Research grant UL1TR002240 (J.Z.S.), U.S. National Institutes of Health grant P30DK034933 (J.Z.S.), the Edward Mallinckrodt, Jr. Foundation (S.L), and U.S. National Institutes of Health grants R01AI34753 (L.M.W.) and R01AI120607 (V.B.C). The funders played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Herein we show that disruption of the host ESCRT machinery impairs host cytosolic protein uptake in T. gondii. Furthermore, we identified TgGRA14, a PVM resident dense granule protein [ 27 ] that possesses ESCRT recruitment motifs, as a host ESCRT-interacting protein. The ESCRT recruitment motifs encoded in TgGRA14 are exposed to the host cytosol and are capable of recruiting ESCRT in the context of HIV-1 virus-like particle release. Additionally, replicating TgGRA14-deficient parasites have a markedly lower percentage of internalized host cytosolic protein compared to wildtype, demonstrating that TgGRA14 contributes to ingestion. Nonetheless, recruitment of host ESCRT components to the PV is not completely dependent on TgGRA14 given that the recruitment of the accessory ESCRT protein ALIX is not affected by disruption of TgGRA14. Importantly, recruitment of TSG101 relies on TgGRA14, and specifically involved the PTAP motif encoded in TgGRA14. Collectively, our data supports a model in which T. gondii uses TgGRA14 and likely other GRA proteins to usurp the host ESCRT machinery for the vesicular transport of host cytosolic proteins across the PVM.

A variety of important human pathogens have developed mechanisms for recruiting host cell ESCRT machinery to ensure an efficient infection and transmission. Intracellular pathogens that differ in pathogenesis, such as the human immunodeficiency virus-1 (HIV-1) [ 15 – 19 ], dengue virus [ 20 ], hepatitis C virus (HCV) [ 21 , 22 ], Ebola virus [ 23 , 24 ], and Mycobacterium tuberculosis [ 25 , 26 ] have all evolved mechanisms to interact with the host ESCRT machinery. However, the contribution of ESCRT machinery in the pathogenesis of an intracellular eukaryotic microorganism remains limited to the aforementioned role in T. gondii cell invasion.

Several parasite-host interactions involve the recruitment and exploitation of host factors [ 6 ] and organelles to the PVM [ 7 – 9 ]. For example, T. gondii is known to exploit a subset of the host endosomal sorting complex required for transport (ESCRT) machinery during its invasion of host cells [ 10 ]. More recent work has shown that host ESCRT is also recruited to the PVM during parasite replication [ 11 ]. This latter recruitment likely occurs via a distinct mechanism since it involves many more components of the ESCRT machinery than that for cell invasion. ESCRT is involved in various cellular processes including formation of the multivesicular body (MVB), plasma membrane repair, cytokinesis, exosome release, and autophagy [ 12 , 13 ]. The machinery is composed of five complexes ESCRT-0, ESCRT-I, ESCRT-II, ESCRT-III, and the Vps4 complex, which sequentially interact with each other to bud vesicles away from the cytosol [ 14 ]. Although the localization of ESCRT components at the PVM is consistent with a possible role in facilitating the budding of vesicles into the PV lumen, to date there are no reports of parasite proteins that interact with host ESCRT during parasite replication.

T. gondii is capable of internalizing host-derived cytosolic proteins [ 3 ]. Once endocytosed, the material is trafficked throughout the endolysosomal system of the parasite and delivered to its lysosomal vacuolar-like compartment (VAC) for degradation by the cysteine protease cathepsin L (TgCPL) [ 4 , 5 ]. However, the molecular mechanisms by which this material is trafficked across the PV membrane (PVM) and is further endocytosed by the parasite remained poorly understood.

Intracellular pathogens have evolved diverse strategies that rely on the exploitation of host factors to ensure intracellular subsistence, replication, and immune evasion. The protozoan parasite Toxoplasma gondii is a successful intracellular pathogen that can infect a wide range of nucleated cells while residing in a non-fusogenic replicative vacuole that helps protect it from host intrinsic defenses. This replicative vacuole, known as the parasitophorous vacuole (PV), is remodeled by a subset of specialized secretory proteins called dense granules proteins (GRAs) [ 1 ]. Although this compartment also segregates the parasite from the nutrient-rich environment of the host cytosol, T. gondii has developed multiple mechanisms for acquiring the resources it needs to sustain infection [ 2 ].

A. Representative images from 3 biological replicates for the GFP-TSG101 recruitment in RΔgra14GRA14 complemented strains. PVM was labeled using an antibody against TgGRA8. Images were analyzed by confocal microscopy. Scale bar is 5 μm. B. Quantification of TSG101 recruitment to the PV between WT, RΔgra14 and RΔgra14 complementation mutants. Data represent the mean TSG101 intensity in the PV relative to the cytosol (PV/cytosol). Each point represents the well average. (~18 wells per biological replicates (n = 3)). Statistical analysis was done using Kruskal-Wallis test. Statistical analysis between complement strains, RH with Δgra14 and Δgra14 with Δgra14:GRA14 WT are shown. ****p<0.0001. C. Quantification of host cytosolic mCherry uptake by RΔgra14 and RΔgra14GRA14 mutants treated with LHVS for 20 h. At least 200 parasites were analyzed per blinded sample. Data represents the mean from ≥ 3 biological replicates. Statistical analysis was by Student’s t-test.

Complementation of RΔgra14 with wildtype TgGRA14 (RΔgra14GRA14 WT ) or TgGRA14 encoding mutations in the YPX (n) L (RΔgra14GRA14 ALIX- ) restored recruitment of GFP-TSG101 (Figs 8A and 8B and S7 ) . However, complementation of RΔgra14 with TgGRA14 constructs encoding PTAP mutations (RΔgra14GRA14 TSG101- and RΔgra14GRA14 TSG101-ALIX- ) were impaired in the recruitment of GFP-TSG101 to the PV (Figs 8A and 8B and S7A ) . These findings are consistent with TgGRA14 interacting with TSG101 through the PTAP motif. Although TgGRA14 deficient parasites do not appear to have a virulence defect based on in vivo [ 45 – 47 ] or in vitro [ 45 , 46 ] data, we decided to measure replication of TgGRA14 late domain mutant strains by quantifying the number of parasites per PV. We also measured the PV size as another approach to assess replication, a strategy previously used [ 48 ]. Consistent with the literature, there was no difference in growth between RH,RΔgra14, and the TgGRA14 complementation strains ( S6B and S6C Fig ) .

Next, to determine if TgGRA14 late domain motifs contribute to the recruitment of ESCRT and uptake of host cytosolic proteins, we complemented the RΔgra14 strain with TgGRA14-HA encoding late domain motif mutations ( S5 Fig ) . To verify that the mutations in TgGRA14 did not affect its topology at the PV, we stained semi-permeabilized infected cells with antibodies against HA to label the TgGRA14 C-terminus and an antibody that recognizes the TgGRA14 N-terminus [ 44 ]. Whereas signal for both the TgGRA14 N-terminus and C-terminus (HA) was detected in samples that were fully permeabilized with 0.1% saponin ( S6A Fig ), only signal for the C-terminus was detected following semi-permeabilization with 0.00001% saponin, a concentration of saponin that will permeabilize the host plasma membrane but not the PVM ( S6B Fig ). These results, which are consistent with previous findings for wildtype TgGRA14 [ 27 ], confirm that the topology of TgGRA14 is not affected by late domain motif mutations at the C-terminus.

Although HIV-1 Gag encodes several late domain motifs that can interact with multiple ESCRT components to initiate assembly of the machinery, recruitment of the ESCRT machinery at the site of viral budding through its interaction with TSG101 appears to be the most critical [ 43 ]. Given that TgGRA14 similarly encodes several late domain motifs, we wanted to identify the main late domain motif facilitating the recruitment of the host ESCRT machinery. To that end, we made alanine substitutions in the TgGRA14 PTAP and YPX (n) L motifs ( Fig 7A ) . Late domain motif mutations in the hybrid GagGRA14 protein disrupted the release of HIV-1 VLPs, suggesting that these motifs encoded in TgGRA14 are important in the context of ESCRT-dependent HIV-1 budding process ( Fig 7B ) .

A. Experimental design for the analysis of the internalization of host cytosolic proteins of TgGRA14-deficient parasites. (1) Inducible mCherry HeLa cells were infected with parasites for 4 hours, (2) at 4 hpi, extracellular parasites were removed, and the infected monolayer was treated with LHVS for 20 h, (3) parasites were harvested at 24 hpi and analyzed by microscopy. B. Quantification of host cytosolic mCherry uptake at 24 hpi by RH or RΔgra14 type I strains treated with DMSO or LHVS for 20 h. C. Quantification of host cytosolic mCherry uptake at 24 hpi by ME49 or MΔgra14 type II strains treated with DMSO or LHVS for 20 h. At least 200 parasites were analyzed per blinded sample. Data represents the mean from ≥ 3 biological replicates. Statistical analysis was by Student’s t-test. *p<0.05, **p<0.01, ***p<0.001.

To further test the model, we next assessed the role of TgGRA14 in the uptake of host cytosolic proteins using type I RH parasites lacking TgGRA14 (RΔgra14). Parasites were treated with the CPL inhibitor LHVS prior to infection of mCherry inducible HeLa cells and harvested at 4 hours post-infection (hpi) ( S3A Fig ) . mCherry inducible HeLa cells have a high percentage of positive cells (>80%) but have lower expression than transiently transfected cells. Since TgGRA14-deficient parasites showed normal ingestion at this time point ( S3B Fig ) , we decided to begin treating the parasites with LHVS at 4 hpi to determine if TgGRA14 is necessary for the uptake of host cytosolic proteins later in infection ( Fig 6A ) . Quantification of ingestion at 24 hpi in this scheme revealed a significant reduction in uptake of host cytosolic proteins in TgGRA14-deficient parasites ( Fig 6B ) . To assess whether this was conserved between strains, we generated TgGRA14-deficient parasites in the type II ME49Δku80 background (MΔgra14) ( S4 Fig ) . Like RΔgra14, we observed a significant reduction in the percentage of mCherry-containing parasites in the MΔgra14 at 24 hpi ( Fig 6C ) . From these data we conclude that TgGRA14 is important for the uptake of host cytosolic proteins late during the later stages of intracellular infection.

A. Volcano plot of the host proteins immunoprecipitated with TgGRA14 under tachyzoite infection conditions showing the ESCRT-accessory proteins (orange), ESCRT-I components (blue), ESCRT-III components (green) and other non-ESCRT associated proteins (purple). Colored dots represent proteins with >1 log2 fold change in tagged vs control lysates and a negative log2 p-value >3.32 (equivalent to p<0.1). B. Immunoblots for the analysis of immunoprecipitation samples confirming the TgGRA14 interaction with host ESCRT components. C. Volcano plot representative of the T. gondii proteins immunoprecipitated with TgGRA14. Some of the enriched dense granule proteins are highlighted in purple. Colored dots represent proteins with >1 log2 fold change in tagged vs control lysates and a negative log2 p-value >3.32 (equivalent to p<0.1). D. Immunoblot analysis of immunoprecipitation samples confirming the TgGRA14 interaction with some of the enriched dense granule proteins.

To determine if TgGRA14 physically associates with other ESCRT components, we analyzed interacting proteins by immunoprecipitation (IP) and liquid chromatography tandem mass spectrometry (LC-MS/MS). To avoid potential artifacts due to overexpression, we generated a strain with TgGRA14 endogenously epitope-tagged at the C-terminus with HA in the genetically tractable ME49Δku80 background using CRISPR/Cas9 gene editing ( S2 Fig ) . Protein lysates were harvested either from infected human fibroblast cultures (either under tachyzoite (Tz-HFF) or bradyzoite growth conditions (Bz-HFF)) or infected mouse primary cortical neuron cultures (Tz-Neuron). Anti-HA coated magnetic beads were used to enrich for GRA14 protein from lysates, and eluates were harvested for LC-MS/MS. As a control for IP experiments, non-tagged ME49Δku80 parasites were used and processed in parallel to cell cultures infected with TgGRA14-HA tagged parasites. Notably, results from the IP-MS/MS analysis of tachyzoite infected human fibroblast (Tz-HFF) cultures revealed that most of the significantly enriched host proteins were related to the ESCRT machinery ( Fig 5A and S2 Table ) . Several of these interactions were verified by IP/immunoblot ( Fig 5B ) . As predicted, both TSG101 and ALIX immunoprecipitated with TgGRA14, along with most components of the ESCRT-I complex, several components of the ESCRT-III complex and the ESCRT-adaptor protein Programed cell death 6 (PDCD6/ALG-2, ALG-2 hereafter) [ 40 , 41 ]. Interestingly, the ALG-2 interacting protein Peflin (PEF1) [ 42 ] also immunoprecipitated with TgGRA14; however, a role for PEF1 in ESCRT mediated vesicle formation has not been reported. The IP analysis also identified TgGRA14 interactions with several other parasite proteins including soluble (TgGRA1, TgMAG1) and transmembrane (TgGRA8, TgGRA15 and TgGRA4) dense granule proteins ( Fig 5C and S2 Table ) , which were confirmed by IP/immunoblot ( Fig 5D ) . Altogether, this data supports a working model in which TgGRA14 interacts with host ESCRT machinery at the PVM where it can potentially trigger vesicle formation to enclose host cytosolic proteins.

A. Representative images from three biological replicates comparing the recruitment of ALIX and GFP-TSG101 between WT and RΔgra14. Images were analyzed by confocal microscopy. Scale bar is 5 μm. B. Quantification of ALIX recruitment to the PV between RH, RΔgra14, R:GRA14 OE . Data represent the mean ALIX intensity at the PV relative to the cytosol (PV/cytosol). Each point represents the well average. (~18 wells per biological replicates (n = 3)). C. Quantification of TSG101 recruitment to the PV between RH, RΔgra14, R:GRA14 OE parasites. Data represent the mean TSG101 intensity in the PV relative to the cytosol (PV/cytosol). Each point represents the well average. (~18 wells per biological replicates (n = 3)). Statistical analysis was done using Kruskal-Wallis test. Only statistical differences are shown. ****p<0.0001.

To determine if recruitment of host ESCRT components to the PVM is dependent on TgGRA14 we infected GFP-TSG101 HeLa cells with WT or RΔgra14 parasites and probed for host ALIX. After comparing the recruitment of both ESCRT components, we observed defects in the recruitment of GFP-TSG101 but not ALIX in TgGRA14-deficient parasites ( Fig 4A ) . Using a spinning disk confocal high content imaging instrument, we quantified the abundance of GFP-TSG101 and ALIX in the PV between WT, RΔgra14 and R:GRA14 OE to determine if overexpression of TgGRA14 could also affect their recruitment. More specifically, we quantified the ratio of GFP-TSG101 or ALIX associated with the PV (demarcated with anti-GRA1) to that of the cytoplasm (area of the cell occluding the PV and nucleus). This analysis revealed that recruitment of ALIX to the PVM is not affected by overexpression or loss of TgGRA14 ( S1 and 4B Figs) . In contrast, overexpression of TgGRA14 increased recruitment of GFP-TSG101 and ablation of TgGRA14 significantly reduced GFP-TSG101 recruitment to the PVM ( S1 and 4D Figs) . Whether ALIX recruitment is entirely independent of TgGRA14 or through a redundant mechanism in the absence of TgGRA14 remains to be elucidated. Regardless, these findings suggest that TgGRA14 contributes to the PVM recruitment of TSG101.

A. Recruitment of host ALIX and TSG101 with TgGRA14. GFP-TSG101 HeLa cells infected with TgGRA14-HA over expressing parasites (R:GRA14 OE ) were stained for the anti-ALIX and anti-HA. Representative images analyzed by structured illumination microscopy. 3D Projections. Scale bar is 5 μm. B. Proximity Ligation assay (PLA) reaction from samples infected with either WT or R:GRA14 OE strains. PLA reactions were performed using antibodies for the host ALIX and the HA-tag to analyze the interaction of TgGRA14 with ALIX. PLA reaction using antibodies against host MAPK7 and HA-tag, was used a negative control. Representative images from at least 3 biological replicates. Scale bar is 5 μm. C. PLA reaction from GFP-TSG101 HeLa infected with either WT or TgGRA14-HA strains. PLA reactions were performed using antibodies for GFP and the HA-tag to analyze the interaction of TgGRA14 with TSG101. PLA reaction using antibodies against GFP and the HA-tag in WT HeLa cells was used a negative control. Representative images from at least 3 biological replicates. Scale bar is 5 μm.

Having established that TgGRA14 can exploit the ESCRT machinery in the context of viral budding we next turned our attention to determining if TgGRA14 is linked to host ESCRT during T. gondii infection of host cells. First, we stained for ALIX in HeLa cells expressing endogenously tagged GFP-TSG101[ 37 ] to determine if these ESCRT proteins are recruited to the PVM in parasites overexpressing GRA14 C-terminally tagged with HA [ 27 ] (R:GRA14 OE ). Analysis by structured illumination microscopy (SIM) showed that ALIX is recruited to the PV ( Fig 3A ) , consistent with what has been observed by Cygan, A. M., et al [ 11 ]. This analysis further revealed that GFP-TSG101 is also abundantly recruited to the PVM where it and ALIX both colocalize with TgGRA14-HA ( Fig 3A ) . To further analyze the juxtaposition of ALIX and GFP-TSG101 with TgGRA14, we performed a proximity ligation assay (PLA) for which a positive reaction indicates that two proteins are within 40 nm of one another. This approach has been used previously to study the interactions of ESCRT components in Kaposi’s Sarcoma-Associated Herpesvirus (KSHV) infected cells [ 38 , 39 ]. We used the R:GRA14 OE parasite strain and probed with anti-HA to detect TgGRA14. Whereas no signal was observed in untagged parasites (WT, RH parasites), we consistently detected a reaction between ALIX and TgGRA14-HA in tagged parasites ( Fig 3B ) . As an additional test of specificity, we did not see a reaction between TgGRA14-HA and MAPK7, a host kinase localized throughout the cytosol. We also observed positive reactions between TgGRA14-HA and GFP-TSG101 in infected endogenously tagged GFP-TSG101 HeLa cells that were absent in infected untagged HeLa cells ( Fig 3C ) . Together these findings suggest that ALIX and TSG101 are proximal to TgGRA14 in infected cells.

A. Schematic of TgGRA14 topology at the PVM. The TgGRA14 C-terminus encoding the late domain motifs PTAP and YPX (n) L is exposed to the host cytosol whereas the N-terminus is exposed to the PV lumen. B. Schematic representation of the predicted recruitment of the ESCRT recruitment by TgGRA14 through the late domain motifs in comparison with their known function in HIV-1 budding. The PTAP and YPX (n) L motifs can mediate interactions with host ESCRT-I components TSG101 and the ESCRT accessory protein ALIX, respectively. C. Experimental design for the substitution of the HIV-1 Gag p6 domain for the TgGRA14 C-terminus portion encoding late domain motifs to generate the GagGRA14. aa, amino acid. Note that the aa numbering in the hybrid Gag/GRA14 is for GRA14. D. Analysis of VLP release by HIV-1 Gag, GagΔp6 and GagGRA14. Representative immunoblots of VLP release include top: VLP release sample (100%) and bottom: cellular lysate sample (4%). The role for the ESCRT machinery in GagGRA14 release was assessed by disruption of the ESCRT machinery using the VPS4A dominant negative form (VPS4A EQ ). Data represents the mean from ≥ 3 biological replicates. Statistical analysis was by Student’s t-test. Only statistical differences are shown. *p<0.05, **p<0.01, ***p<0.001.

Among the 33 identified candidates, we prioritized the PV transmembrane protein TgGRA14 because previous work showed that its C-terminus, which encodes a PTAP motif for putative binding to TSG101, is exposed to the host cytosol [ 27 ]. Just downstream of the PTAP motif we also identified a YPX (n) L motif for the potential recruitment of ALIX ( Fig 2A ) . Since HIV-1 Gag uses both a PTAP and a YPX (n) L motif to direct virus budding [ 18 ], we predicted that, like HIV-1 Gag, TgGRA14 would associate with ESCRT components at the PVM to promote vesicle formation ( Fig 2B ) . If so, we further reasoned that the C-terminal domain of TgGRA14 would interact with ESCRT and could be functionally equivalent to the late-domain containing p6 domain of HIV Gag. To test this, we transfected cells with constructs expressing HIV-1 Gag, HIV-1 Gag lacking the p6 late domain (GagΔp6) and HIV-1 Gag with the p6 domain substituted with a portion of the TgGRA14 C-terminal domain ( Fig 2C ) . As expected, expression of HIV-1 Gag sustained the release of virus-like particles (VLPs) in an ESCRT dependent manner, and deletion of the p6 domain diminished VLP production, with residual release being independent of ESCRT ( Fig 2D ) . Interestingly, substitution of the late-domain motif portion of TgGRA14 for the HIV-1 p6 domain allowed VLP production that equaled that of native HIV-1 Gag, with release being similarly dependent on ESCRT. Thus, when coupled with HIV-1 Gag, the C-terminal region of TgGRA14 encoding late domain motifs can function for exploitation of host ESCRT for the release of HIV-1 VLPs.

Enveloped viruses including HIV-1 can exploit the host ESCRT machinery for viral budding [ 32 ]. In the case of HIV-1, the host ESCRT machinery is recruited to the site of viral budding by the Gag protein through the late-domain motifs, such as PTAP and YPX (n) L, which are capable of interacting with the ESCRT-I component Tumor suppressor gene 101 (TSG101) and the ESCRT accessory Programmed cell death 6-interacting protein (PDCD6IP/ALIX, ALIX hereafter), respectively [ 16 , 33 ]. To identify T. gondii effector proteins as candidates for the recruitment of the host ESCRT machinery, we performed a bioinformatic search using the vEuPathDB Toxoplasma Informatics Resources Database (ToxoDB). The first step in our research strategy consisted of a motif search for proteins within the T. gondii genome encoding the late domain motif P(S/T)AP. As a second step we looked for proteins that also encode a signal peptide to center our search on secreted proteins. Through this search strategy we identified 33 secretory proteins with a putative P(S/T)AP late domain motif ( S1 Table ) . The identified candidates included the rhoptry neck protein RON5, which has been shown to interact with the host TSG101 ESCRT-I protein during cell invasion [ 34 ]. However, RON5 is only secreted during invasion and is not present at the host-parasite interface during parasite replication. We considered GRA proteins as better candidates since they are secreted continuously during parasite replication [ 35 ]. Three of the 33 genes identified were predicted to be GRA proteins based on data from a recent extensive subcellular proteomics study [ 36 ] ( S1 Table ) .

To determine whether T. gondii relies on the host ESCRT machinery during infection we analyzed replication following disruption of the host ESCRT machinery through the expression of VPS4A EQ . Compared to cells expressing exogenous VPS4A WT , there was a moderate replication defect in parasites growing in cells expressing VPS4A EQ ( Fig 1D and 1E ) . To test if the host ESCRT machinery plays a role in the uptake of host cytosolic proteins by T. gondii, host cells were transiently transfected with a cytosolic fluorescent reporter protein (Venus) in addition to either VPS4A WT or VPS4A EQ . The transfected cells were infected with parasites lacking the major lysosome-like VAC protease TgCPL (RΔcpl) to reduce the degradation of the internalized Venus and allow visualizing its signal within the parasite. As expected, based on previous studies [ 3 ], RΔcpl parasites are capable of internalizing host-derived Venus. Because only 30–40% of the host cells express Venus, the percentage of parasites with ingested Venus is likely an underestimate of the ingestion efficiency. Interestingly, we found that exogenous expression of VPS4A EQ , but not VPS4A WT , impaired RΔcpl accumulation of Venus in newly invaded ( Fig 1F ) and replicating ( Fig 1G ) parasites. These findings suggest that disruption of host ESCRT limits parasite replication and uptake of host cytosolic proteins.

A. Schematic of the ESCRT machinery function in MVB formation. Ubiquitinated cargo targeted for degradation is recognized by ESCRT-0 and ECSRT-I, which further recruits ESCRT-II and the ESCRT accessory protein ALIX. At the last steps, ESCRT-III forms spirals at the membrane to mediate membrane constriction and the VPS4 complex facilitates scission and disassembly of the machinery by hydrolyzing ATP. B. Distribution of the ESCRT-III component CHMP4B at the host cytosol (i) and the PV (ii) in HeLa cells transiently transfected with VPS4A WT -mCherry and infected with RH parasites for 24 h. C. Distribution of the ESCRT-III component CHMP4B at the host cytosol (i) and the PV (ii) in HeLa cells transiently transfected with VPS4A EQ -mCherry and infected with RH parasites for 24 h. The PV was labeled with anti-TgGRA7. Images were analyzed by structural illuminated microscopy (SIM). Scale bar is 5 μm. D. Growth assay of RH parasites in HeLa cells transiently transfected with exogenous VPS4 WT or VPS4A EQ compared to untransfected control. The percentage of PV with 1, 2, 4 or 8+ parasites was calculated for each cell subset. At least 20 PV were counted per blinded sample. Data represents means from 7 biological replicates. E. Mean number of parasites per PV from growth assay. Statistical analysis was by Student’s t-test. F. Quantification of ingestion of host cytosolic Venus at 30 mpi in RH or RΔcpl parasites harvest from cells were transiently co-transfected with a plasmid encoding cytosolic Venus fluorescent protein expression and exogenous expression of either VPS4A WT or VPS4A DN . Representative images for parasites with ingested host-derived cytosolic Venus (shown in magenta) at 30 mpi included at the top. Scale bar is 5 μm. G. Quantification of host cytosolic Venus ingestion at 24 hpi in RH or RΔcpl parasites. At least 200 parasites were analyzed per blinded sample. Representative images for parasites with ingested host-derived cytosolic Venus (magenta) at 24 hpi are included at the top. Scale bar is 5 μm. Data represents the mean from ≥ 3 biological replicates. Statistical analysis was by Student’s t-test. Only statistical differences are shown. *p<0.05, **p<0.01.

The ESCRT machinery is composed of five complexes (ESCRT-0 –ESCRT-III and VPS4) that subsequently interact with one another to form vesicles that bud away from the cytosol ( Fig 1A ) . At the last step, the AAA ATPases, VPS4A and VPS4B, hydrolyze ATP to disassemble the machinery at the membrane and facilitate scission [ 28 , 29 ]. Expression of a dominant negative form of VPS4A (VPS4A EQ ), which is unable to hydrolyze ATP [ 30 ], results in the accumulation of the ESCRT-III component CHMP4B at endosomal structures due to failed scission events. We found that whereas overexpression of VPS4A wildtype (VPS4A WT ) did not cause overt accumulation of VPS4A WT or CHMP4B at the PV ( Fig 1B ), overexpression of VPS4A EQ resulted in marked association of VPS4A and CHMP4B with the PV ( Fig 1C ) . That late acting ESCRT components CHMP4B and VPS4A are not observed to the same extent in VPS4A WT is expected since the dynamics of VPS4 appear to be transient, with puncta relating to VPS4B recruitment in MVB lasting for <4 s [ 31 ].

Discussion

Herein we show that host ESCRT and TgGRA14 are necessary for parasite ingestion of host cytosolic protein, that TgGRA14 has motifs that support ESCRT-dependent budding of HIV-1 VLPs, and that TgGRA14 co-immunoprecipitates with components of host ESCRT and other parasite secretory proteins in the setting of different parasite growth conditions (e.g., tachyzoite vs. bradyzoite) and host cell environments (e.g., human fibroblasts vs. mouse neurons). We also establish a contribution of TgGRA14 to the PVM recruitment of the ESCRT-I protein TSG101 in a manner dependent upon a PTAP motif in the cytosolically exposed C-terminal domain of TgGRA14. Further, our findings suggest that expression of TgGRA14 is not necessary for PVM recruitment of ALIX despite TgGRA14 having a putative ALIX interaction motif. Finally, we found that although GRA14 is necessary for parasite uptake of host protein, its TSG101 and ALIX motifs are dispensable for this process.

Whereas ESCRT and TgGRA14 are both required for the ingestion pathway late during infection, parasite ingestion of host cytosolic protein soon after invasion depends on ESCRT, but not TgGRA14. This suggests that early in infection T. gondii interacts with the host ESCRT machinery through a different mechanism. Although it has been shown that GRA proteins are secreted into the PV within 30 min post-invasion [35], it is unclear how long it takes them to reach their correct destinations and become functional thereafter. By contrast at least some proteins discharged from the parasite rhoptries during invasion are immediately functional since several are required for entry including those released from the rhoptry neck (RON proteins). Indeed, prior work showed that RON4 and RON5 interact with host ALIX and TSG101, respectively [10]. The function of TSG101 and ALIX during T. gondii invasion appears to be mainly as adaptor proteins for interaction with the host cytoskeleton and is likely independent of their roles in vesicle budding [10]. However, RON proteins remain associated with a discrete region of the nascent PVM for some time after invasion [49], raising the possibility that they could function in ESCRT-dependent parasite uptake of host cytosolic protein. Additional studies are necessary to test this possibility.

The abundant strategies used by pathogens to exploit the host ESCRT machinery highlight the importance of this host pathway in microbial pathogenesis. The multiple functions of the machinery in the cells provide numerous possibilities for an intracellular pathogen to benefit from it. This is not limited to mammalian host cells and their pathogens, since ESCRT-III and VPS4A homologs can be found in archaea Sulfolobus spp.[50,51] and they are exploited by the archaeal Sulfolobus turreted icosahedral virus (STIV) for replication [52]. The best described examples of a pathogen usurping the ESCRT machinery are seen in viral egress from the host cell, this is the case for multiple enveloped viruses such as HIV-1[18], the equine infectious anemia virus (EIAV) [43], and the Ebola virus [23,24]. Although a role for the ESCRT-III homologs has not been defined, a similar observation of ring-like structures at budding sites has been described for the archaeal Sulfolobus spindle-shaped virus-1 (SSV1) [53]. The known mechanisms for enveloped virus recruitment of eukaryotic host cell ESCRT machinery in viral budding is through pathogen proteins encoding late domain motifs, PTAP, YPX (n) L and PPXY, which mimic motifs present in ESCRT components and are necessary for the assembly of the machinery [16,54]. These motifs are present in the C-terminus of our candidate T. gondii ESCRT-interacting protein TgGRA14. Because the most predominant late domain motifs in viral budding are PTAP and YPX (n) L, we decided to elucidate their function in TgGRA14.

Our data demonstrates that the C-terminus portion of TgGRA14 is sufficient to substitute for the HIV-1 p6 domain, providing evidence that these late domain motifs encoded in TgGRA14 can recruit the host ESCRT machinery for HIV-1 VLP budding. Mutations in TgGRA14 late domain motifs impaired HIV-1 VLP release and mutations in the PTAP motif impaired GFP-TSG101 recruitment. Nonetheless, mutations in the late domain motifs had no apparent effect in the uptake of host cytosolic proteins suggesting alternative mechanisms for the ESCRT recruitment by TgGRA14 or additional effector proteins that function in concert with or in parallel to TgGRA14 for protein ingestion. It is, however, also possible that our current assay for ingestion is not sensitive enough to detect changes in the dynamics of this process, like slower rates of uptake.

The P(S/T)AP late domain motif described in retroviral exploitation of the host ESCRT machinery and encoded by ESCRT components such as HRS (ESCRT-0) and ALIX is critical for interaction with TSG101[54]. This highlights the role of proline-rich sequences in host ESCRT assembly. Furthermore, ALG-2 interacts with ALIX and PEF1, another protein enriched in our immunoprecipitation data, through proline-rich regions including a PPYPXnYP motif, where X represents any amino acid and n can vary, but is most often 4[55]. Interestingly, TgGRA14 has a putative PPYPXnYP pseudomotif (PPYVPPMYP) in its C-terminus, which could mediate an interaction with the host ALG-2 as an alternative mechanism for TgGRA14 engagement of the host ESCRT machinery. The ESCRT adaptor protein ALG-2 is capable of interacting with both TSG101[56] and ALIX [57,58], and it can also stabilize an interaction between these two proteins in a Ca2+-dependent manner [59]. ALG-2 functions as an adaptor protein for the assembly of the ESCRT machinery at sites of membrane injury [41] and it regulates ALIX function in MVB cargo sorting [40]. This host protein is highly enriched in the PV based on our immunoprecipitation data [11] and although its role at this site remains to be elucidated, it could be acting as a bridge to promote the interaction between ESCRT components. The involvement of ALG-2 could suggest a role for Ca2+ in the function of the host ESCRT at the PVM. Further analysis of our TgGRA14 immunoprecipitation data looking for interacting parasite effector proteins encoding putative late domain motifs identified TgGRA8 and TgGRA4 as alternative candidates for the recruitment of ALG-2. However, the topology of these dense granule proteins at the PVM, their ability to interact with ESCRT components and function in ingestion remain to be elucidated. Also, additional dense granule proteins that do not encode putative late domain motifs could still be playing a role in ESCRT function at the PVM. This is the case for several intracellular pathogens which manipulate host ESCRT machinery via mechanisms that appear to be independent of late domain motif mediated interactions. Thus, it is conceivable that other T. gondii effector proteins could be interacting with the host ESCRT machinery in conjunction or independently of TgGRA14. This may explain the ALIX recruitment and residual GFP-TSG101 recruitment to the PV in TgGRA14-deficient parasites. Although our findings suggest an interaction between TgGRA14 and ESCRT, evidence of direct interaction is lacking. Thus, we cannot rule out that disruption of TgGRA14 indirectly affects ESCRT-dependent ingestion of host cytosolic protein by affecting other parasite proteins. Future studies should focus on measuring direct interactions and defining if TgGRA14 functions in concert with other dense granule proteins for the recruitment of the host ESCRT machinery. This would help us understand why there is no virulence defect in TgGRA14-deficient parasites [27] and the importance of the uptake of host cytosolic proteins in T. gondii pathogenesis.

Since ALIX is still recruited to the PVM of TgGRA14-deficient parasites, it is conceivable that other transmembrane dense granule proteins are contributing to the interaction with the host ESCRT components at the PVM and compensating for the loss of TgGRA14. TgGRA15 is an interesting TgGRA14-interacting protein since it has also been demonstrated by immunoprecipitation to interact with several ESCRT components [60]. Although its role in the recruitment of the ESCRT machinery remains to be further elucidated, the known function of TgGRA15 in type II strains is to activate NF-κB signaling [61], a host immune response that is also moderately stimulated by TgGRA14[47]. The extent to which TgGRA15 or other TgGRA14-interacting dense granule proteins take part in the uptake of host cytosolic proteins remains unknown. The role of TgGRA7 in the context of host cytosolic protein uptake by T. gondii has been studied, but it does not seem to be contributing to this pathway. In contrast, TgGRA2, which is critical for the integrity of the intravacuolar network (IVN), is necessary for efficient ingestion [3]. TgGRA14 is present in the IVN but is not necessary for IVN structure [27]. It is possible that IVN disruption in TgGRA2 impairs TgGRA14 distribution at the PVM and IVN, thus resulting in less ingestion. Alternatively, these proteins could be involved in the uptake of host cytosolic proteins through different mechanisms. The emerging picture is that T. gondii has the potential to exploit host ESCRT through a variety of means, which could indicate that it has evolved multiple mechanisms because of a critical reliance on host ESCRT.

Blood stage malaria parasites internalize and digest in their food vacuole large amounts of hemoglobin from the cytosol of infected erythrocytes as a critical source of amino acids for their replication. Although hemoglobin laden vesicles derived from the PVM have been seen directly entering malaria parasites through the mouth-like cytostome and other structures [62–64], little is known about how such vesicles are generated. Components of the ESCRT machinery have been identified in the proteome of mature erythrocytes [65] and in exosomes derived from immature erythrocytes (reticulocytes) [66] Although this raises the possibility that intraerythrocytic malaria parasites similarly exploit host ESCRT for protein ingestion, there is currently no evidence that ESCRT is functional for vesicle biogenesis in erythrocytes. Hepatocytes, in which liver stage malaria parasites replicate, have functional ESCRT [67–69]. However, the extent to which liver stage parasites internalize host protein remains to be determined. Recent work has identified a role for malaria homologs of ESCRT in the generation of extracellular vesicles derived from infected erythrocytes [70]. This study, however, did not distinguish whether such parasite ESCRT homologs function within the parasite, at the PVM, in cytoplasmic membranous structures, or at the erythrocyte plasma membrane. Homology searches fail to detect TgGRA14 homologs in other apicomplexan parasites except for some other tissue cyst-forming coccidian parasites (Hammondia hammondi, Neospora caninum, and Besnoitia besnoiti), but this could be due to the rapid evolution of proteins that reside at the parasite-host interface [71]. Attention should be given to future studies dissecting the commonalities and distinctions of host protein uptake by apicomplexan parasites.

Based on our findings, we propose a working model in which T. gondii acquires host cytosolic proteins by vesicular trafficking across the PVM happening through the interaction of the host ESCRT machinery with TgGRA14 (Fig 9). This model posits that TgGRA14 functions as an ESCRT-adaptor for the recruitment of ESCRT-I components that subsequently interact with the accessory ESCRT protein ALIX, involved in bridging the ESCRT-I and ESCRT-III complexes. Although our work exposes the first mechanistic insight into how T. gondii recruits host ESCRT to the PVM during intracellular replication, much remains to be learned about the role of TgGRA14 and other parasite secretory proteins in this process.

[END]

[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010138

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


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