(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

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



Anaerobic peroxisomes in Entamoeba histolytica metabolize myo-inositol

['Zdeněk Verner', 'Department Of Parasitology', 'Faculty Of Science', 'Charles University', 'Biocev', 'Vestec', 'Czech Republic', 'Vojtěch Žárský', 'Tien Le', 'Ravi Kumar Narayanasamy']

Date: 2021-11

Entamoeba histolytica is believed to be devoid of peroxisomes, like most anaerobic protists. In this work, we provided the first evidence that peroxisomes are present in E. histolytica, although only seven proteins responsible for peroxisome biogenesis (peroxins) were identified (Pex1, Pex6, Pex5, Pex11, Pex14, Pex16, and Pex19). Targeting matrix proteins to peroxisomes is reduced to the PTS1-dependent pathway mediated via the soluble Pex5 receptor, while the PTS2 receptor Pex7 is absent. Immunofluorescence microscopy showed that peroxisomal markers (Pex5, Pex14, Pex16, Pex19) are present in vesicles distinct from mitosomes, the endoplasmic reticulum, and the endosome/phagosome system, except Pex11, which has dual localization in peroxisomes and mitosomes. Immunoelectron microscopy revealed that Pex14 localized to vesicles of approximately 90–100 nm in diameter. Proteomic analyses of affinity-purified peroxisomes and in silico PTS1 predictions provided datasets of 655 and 56 peroxisomal candidates, respectively; however, only six proteins were shared by both datasets, including myo-inositol dehydrogenase (myo-IDH). Peroxisomal NAD-dependent myo-IDH appeared to be a dimeric enzyme with high affinity to myo-inositol (Km 0.044 mM) and can utilize also scyllo-inositol, D-glucose and D-xylose as substrates. Phylogenetic analyses revealed that orthologs of myo-IDH with PTS1 are present in E. dispar, E. nutalli and E. moshkovskii but not in E. invadens, and form a monophyletic clade of mostly peroxisomal orthologs with free-living Mastigamoeba balamuthi and Pelomyxa schiedti. The presence of peroxisomes in E. histolytica and other archamoebae breaks the paradigm of peroxisome absence in anaerobes and provides a new potential target for the development of antiparasitic drugs.

E. histolytica colonizes the human large intestine upon ingestion of amebic cysts. Trophozoites dwell in the hypoxic milieu of the intestinal lumen and can invade the epithelium to cause potentially lethal amebiasis. As an anaerobe, E. histolytica is dependent on glycolysis to generate ATP, whereas mitochondria are present in a greatly reduced form called mitosomes. Peroxisomes are believed to be absent in E. histolytica, as in other anaerobes, supported by the lack of typical peroxisomal pathways, including β-oxidation and catalase. However, we identified seven proteins involved in peroxisome biogenesis, the peroxins, and our investigations of their cellular localization indicated that E. histolytica possesses the anaerobic form of peroxisomes that we recently discovered in free-living anaerobic relatives. Proteomic analysis of anaerobic peroxisomes revealed the presence of highly active homodimeric myo-inositol dehydrogenase (myo-IDH), which catalyzes NAD-dependent oxidation of myo-inositol and other cyclitols. Since upstream and downstream pathways are absent in anaerobic peroxisomes, the role of myo-IDH is most likely to maintain intraperoxisomal redox potential. Discovery of anaerobic peroxisomes in E. histolytica and identification of myo-IDH that is absent in host cells provide a new avenue for the development of strategies against the parasite.

Peroxisomes are generally thought to be absent in anaerobic organisms, including E. histolytica [ 26 ]. Recently, however, anaerobic peroxisomes were described in the free-living anaerobic amoeba Mastigamoeba balamuthi, which bear an anaerobic form of mitochondria named hydrogenosomes [ 27 ]. M. balamuthi peroxisomes contain a major part of the pyrimidine biosynthesis pathway, enzymes required for coenzyme A synthesis, and carbohydrate metabolism, while β-oxidation and catalase are absent [ 27 ]. Moreover, the presence of peroxisomes was noted in another anaerobic member of the Archamoebae, Pelomyxa schiedti [ 28 ]. The discovery of anaerobic peroxisomes in free-living archamoebae prompted us to search for peroxisomes in evolutionarily related parasitic Entamoeba species. Indeed, we identified a minimal set of pex genes that provide E. histolytica with the capacity to form peroxisomes, and we partially characterized the peroxisomal proteome. The common feature of anaerobic peroxisomes in Archamoebae appears to be the metabolism of myo-inositol.

Peroxisomes have been widely reported from aerobic organisms, and their evolution was suggested to be tightly connected with oxygen-dependent metabolic pathways, particularly β-oxidation of fatty acids, redox homeostasis, and bioenergetics [ 23 , 24 ]. Moreover, it has been shown in human mutant fibroblasts that peroxisomes could be formed by a fusion of vesicles derived from both the ER and mitochondria [ 25 ]. In this case, Pex3 and Pex14 were first inserted into the mitochondrial outer membrane, which later buds in the form of a vesicle; moreover, Pex16-containing vesicles were shown to originate from the ER. Mature peroxisomes were determined to be a union of the two distinct classes of pre-peroxisomal vesicles [ 25 ].

Despite the wide range of metabolic diversity, all peroxisomes share a common de novo biogenesis pathway mediated by a set of specific proteins named peroxins (PEXs). In the model organism Saccharomyces cerevisiae, peroxisomal matrix proteins are synthesized in the cytosol and recognized by the receptor proteins Pex5 and Pex7 with C-terminal peroxisomal targeting signal (PTS)1 and N-terminal PTS2, respectively [ 21 ]. The receptor-cargo is then recruited by docking proteins (Pex13, Pex14 and Pex17) and a transient pore composed of Pex5-Pex14 is formed. The cargo is released into the lumen of peroxisomes while RING finger proteins (ubiquitin ligases Pex2, Pex10 and Pex12) are assembled onto the transition pore. The receptor is recycled from the membrane upon monoubiquinylation of a conserved cysteine or degraded upon polyubiquitinylation of a conserved lysine. The process of ubiquitinylation involves RING proteins, and ubiquitin-conjugated Pex4 is anchored to the membrane via Pex22. Release of the receptor from the membrane is facilitated by the Pex1/6 heterohexamer anchored to the membrane via Pex15, which is also responsible for deubiquitinylation [ 21 ]. Membrane proteins are incorporated into the peroxisomal membrane by the cycling chaperone Pex19, which recognizes the membrane PTS and recruits the protein to a membrane-anchored Pex3 [ 18 ]. The trafficking of glycosylated proteins from the ER is mediated by Pex16, which is absent in Saccharomyces species [ 18 , 22 ]. Similar to mitochondria, peroxisomes can undergo fission, exploiting dynamin-like proteins that are recruited by Pex11 [ 18 ].

Entamoeba histolytica is a causative agent of amoebiasis, one of the most prevalent parasitic diseases of humans. Over 65,000 lethal cases of amoebiasis per year have been reported worldwide [ 1 , 2 ], and E. histolytica prevalence has been estimated to reach 3.55% globally [ 3 ]. E. histolytica colonizes the human large intestine, and upon a not-well-understood trigger, the trophozoites invade the mucous barrier and cause dysentery and eventually extraintestinal amoebiasis [ 4 ]. Due to its adaptation to oxygen-poor environments and its parasitic lifestyle, the cellular organelles and metabolism of E. histolytica are highly modified [ 5 ]. Entamoeba does not possess classical mitochondria; instead, the cells harbor minimal versions, called mitosomes. The organelles do not contain DNA or exhibit classical energy metabolism, yet they are surrounded by a double membrane [ 6 – 8 ]. The only known mitosomal function is sulfate activation [ 9 , 10 ]. Energy metabolism is based on glycolysis in the cytosol, where glucose is converted to pyruvate, and extended glycolysis, that includes conversion of pyruvate to acetyl-CoA by pyruvate:ferredoxin oxidoreductase; acetyl-CoA is then utilized for ATP production by acetyl-CoA synthase [ 11 ]. The E. histolytica cytosol contains numerous vesicles, lysosomes, endosomes, and multivesicular bodies, whereas the endoplasmic reticulum (ER) and Golgi apparatus (GA) are difficult to recognize and, in the case of the Golgi, were once thought to be absent [ 12 ]. However, later studies showed that the ER forms a tubular structure similar to the ER of other eukaryotic cells [ 13 , 14 ], and a GA was visualized in the form of separated vesicles bearing some GA markers, such as Golgi-associated coatomer protein ɛ-COP [ 15 , 16 ]. Peroxisomes are single membrane-bound multifunctional organelles. A defining feature of classical peroxisome is the presence of peroxide-generating and detoxifying pathways with the marker enzyme catalase [ 17 , 18 ]. However, some specialized forms of the organelle do not contain these key pathways and play various different roles: glycosomes of trypanosomatids are known to contain the first six or seven glycolytic enzymes [ 19 ], and plant glyoxysomes engage in the glyoxalate cycle, whereas Woronin bodies of filamentous fungi serve as a physical barrier between two cells upon hyphal wounding [ 20 ].

To investigate the evolution of E. histolytica myo-IDH, we performed phylogenetic analysis using eukaryotic (33) and prokaryotic (70) protein sequences, including those with known structures and functions (Figs 12 and S8 ). The eukaryotic sequences formed three distinct branches nested within bacterial orthologs. The IDH sequences identified in four Entamoeba species, M. balamuthi and P. schiedti, formed a monophyletic cluster of peroxisomal IDHs with high statistical support. All these sequences possessed PTS1 signals (-PKL, -SKL, -AKL), except for one of two P. schiedti IDH paralogs. The peroxisomal IDHs were placed within broader mostly eukaryotic branches (Group I). The closest orthologs appeared to be putative IDHs in ochrophytes and oomycotes, and we also identified IDH orthologs in the amoebozoan Planoprotostelium fungivorum and members of Hemochordata, Tunicata, Echinodermata, and Arthropoda. None of these orthologs possessed a PTS1 signal, although a rarePTS1 signal was noted for the IDH of P. fungivorum(-SKF). The cellular localization of most eukaryotic IDHs of Group I is likely cytosolic; however, IDHs of Stramenopiles that formed a sister group to peroxisomal IDHs were predicted to localize to mitochondria. In addition, IDHs of red algae with primary plastids and the haptophyte Emiliania huxleyi and two members of Ochrophyta with secondary plastids of red algal origin were predicted to target chloroplasts [ 49 ] (Group III) ( S8 Table ).

Identification of putative IDH in peroxisomes prompted us to investigate the substrate specificity of this enzyme. An interproscan search revealed the presence of the N-terminal NAD-binding Rossmann fold domain of the glucose–fructose oxidoreductase (GFO)_IDH_MocA protein family (PFAM: PF01408). These enzymes can utilize a broad spectrum of substrates, such as myo-inositol, scyllo-inositol, D-glucose, D-xylose, D-fructose, and 1,5-anhydro-D-fructose [ 43 – 47 ]. Protein sequence alignment of E. histolytica IDH to IDHs of other archamoebae and structurally characterized bacterial paralogs revealed the presence of conserved residues required for the binding of NAD + (motifs I and II) and key residues of motifs III-VI that define the substrate-binding pocket ( S3 Fig ) [ 48 ]. However, diversity within motif IV did not allow reliable estimation of the enzyme substrate specificity ( S3 Fig ). Thus, recombinant E. histolytica IDH was expressed in E. coli, affinity purified ( S4 Fig ), and used to determine its kinetic parameters for various substrates ( Table 1 and S5 Fig ). The enzyme revealed dehydrogenase activity preferentially toward myo-inositol (hereafter myo-IDH), and this activity was dependent on NAD + , while negligible activity was detected with NADP + . The Michaelis constant K m for scyllo-inositol was approximately 10-fold higher than that for myo-inositol and those for D-glucose and D-xylose two orders of magnitude higher. The molecular mass of recombinant myo-IDH under reducing conditions using SDS-PAGE was 37.4 kDa, which corresponded well with the theoretical weight of 36.4 kDa, including the 6x-His tag. Molecular mass of the native enzyme was initially determined by size-exclusion chromatography. The peak of enzymatic activity was recovered at the elution volume corresponding to 41 kDa suggesting a momomeric structure ( S6 Fig ). However, because molecular mass determined by this approach might be affected by the protein shape, we also used multiangle light scattering (MALS). This analysis revealed that the molecular mass of the native recombinant enzyme was 77.4 (±4) kDa, suggesting a homodimeric structure with a minor contribution from the trimer (118.1± 22 kDa) ( S7 Fig ).

Predicted peroxisomal localization was evaluated for seventeen selected proteins ( S7 Table ). The proteins were expressed with an N-terminal mCherry tag in the yeast strain BY4742:POX1-EGFP expressing the integrated GFP-tagged peroxisomal marker protein acyl-CoA oxidase (Pox1). Three proteins, putative IDH ( Fig 9 ), and hypothetical proteins EHI_051440 and EHI_045060 were observed in round vesicles in which they colocalized with Pox1, whereas other localizations were observed for the rest of the tested proteins ( Fig 9 and S7 Table ). The localization of IDH in E. histolytica was studied using a specific polyclonal antibody raised against the corresponding recombinant protein ( S4 Fig ). Confocal immunofluorescence microscopy revealed IDH-labeled punctual structures distinct from mitosomes visualized by an antibody against APSK (PCC r = -0.159)( Fig 10A ). Furthermore, we investigated the colocalization of IDH in cells expressing His-tagged PEXs (Figs 10B and S2 and S5 Tables). IDH was detected in similar structures as above and partially colocalized with Pex14 (PCC r = 0.512, Fig 10B ) and Pex16 (PCC r = 0.260). The strongest correlation (PCC r = 0.547) was observed between IDH and Pex11 ( Fig 10B ). The vesicular localization of IDH was confirmed by immunoelectron microscopy ( Fig 10C ). The double-labeling experiments revealed IDH signals with those of Pex16 within the vesicular structures( Fig 10C ).

Next, we performed in silico prediction of PTS1 signals in proteins predicted in all available Entamoeba species using a machine learning algorithm (PTS1 ML) optimized according to the M. balamuthi peroxisomal proteome [ 27 ]. We rationalized that a high score for PTS1 ML prediction in E. invadens that seems to lack peroxisomes likely represents a false positive signal and that the absence of proteins in E. invadens in comparison to other species might be related to the absence of peroxisomes. Thus, the PTS1 ML score for E. invadens proteins was subtracted from the corresponding PTS1 ML score for E. histolytica ( S7 Table ). In addition, we used three other available tools for PTS1 prediction [ 41 , 42 ]. These analyses provided a set of 56 proteins with putative PTS1 predicted by at least two tools ( S7 Table ). The comparison of this dataset and proteomic data ( S6 Table ) revealed only six proteins present in the proteome of the putative peroxisome-enriched fraction with the predicted PTS1 signal using our strict criteria. These proteins include five hypothetical proteins of unknown function (EHI_045060, EHI_050510, EHI_183900, EHI_185440, and EHI_161040) and putative IDH (EHI_125740). The absence of genes for IDH and the hypothetical protein EHI_045060 in the E. invadens genome further support its possible associations with putative peroxisomes of E. histolytica ( S7 Table ).

Comparison with the proteome of M. balamuthi peroxisomes [ 27 ] revealed eight proteins in common, such as putative inositol dehydrogenase (IDH), long-chain fatty acid-CoA ligase, and malate dehydrogenase ( S6 Table ). We were also interested in the overlap between our dataset and the previously reported mitosomal proteome [ 9 ]. This comparison revealed 35 common proteins, of which four proteins detected in the mitosome-enriched fraction were between our top 100 peroxisomal candidates with the highest cumulative score ( S6 Table ). These proteins included putative IDH, which was previously localized in the cytosol [ 9 ], a hypothetical protein EHI_103470, which we recognized as Pex11, a hypothetical protein (EHI_170120) that was reported as the mitosomal membrane protein [ 40 ], and a microtubule-binding protein ( S6 Table ).

Initial experiments to isolate putative peroxisomes based on differential and gradient centrifugation did not allow effective separation of Pex14-localized organelles and mitosomes. Therefore, we decided to use affinity purification of the organelles. Pex14 was expressed with a C-terminal poly-His or V5 tag facing the cytosol [ 39 ], and putative peroxisomes were isolated from cell homogenates using magnetic beads conjugated with the corresponding antibody. In a pilot experiment, all steps were monitored by western blotting to assess enrichment of putative peroxisomes using an anti-His antibody and mitosome contamination using an anti-Cpn60 antibody (Figs 8 and S1B ). Western blot analysis showed approximately 40-fold enrichment of the Pex14-His signal after differential centrifugation (150,000 x g pellet, Fig 8B ), and the Pex14-His signal was separated from Cpn60 using anti-His antibody-conjugated beads ( Fig 8A ). A summary of LFQ mass spectrometry analyses of putative peroxisome-enriched fractions with His-tagged Pex14 (seven experiments) and V5-tagged Pex14 (six experiments) is given in S6 Table . In two experiments with V5-tagged Pex14, we omitted Tween 20 to estimate, which proteins are loosely associated with the organelle surface. Altogether, 655 proteins were identified in the peroxisome-enriched fraction, of which 24 were observed only in experiments without Tween 20 ( S6 Table ). In addition to tagged Pex14, the dataset contained two other PEXs, Pex5 and Pex11. Interestingly, although orthologs of RING complex E3 ubiquitin ligases Pex2, Pex10 and Pex12 and E2 ubiquitin-conjugating protein Pex4 are absent in E. histolytica, a single putative E3 ubiquitin ligase EHI_030770 and two ubiquitin-conjugating enzyme family proteins, EHI_083560 and EHI_048700, were identified in the proteome.

More detailed localization of PEXs was investigated using confocal microscopy. First, we attempted to distinguish putative peroxisomes from mitosomes ( Fig 5A ). We observed that both PEXs and APSK, a marker protein of mitosomes, labeled small round vesicles of similar sizes, but none of the PEXs colocalized with APSK. The Pearson’s correlation coefficient in colocalized volume (PCC)ranged between -0.072 to 0.023, which indicates no or negligible correlation between PEXs and APSK signals ( S2 and S5 Figs and S5 Table ). The only exception was Pex11, which partially localized to mitosomes with weak PCC r = 0.241. This observation was further supported by structured illumination microscopy (SIM)( Fig 5B ). While Pex14 did not colocalize with the APSK signal, Pex11 was partially associated with mitosomes ( Fig 5B ). Pex14 also did not colocalize with other cellular vesicles and vacuoles labeled with Atg8(PCC r = 0.033) [ 36 ] and the ER marker BiP1(PCC r = 0.021) [ 37 ] (Figs 6 and S2 and S5 Tables). The number of putative peroxisomes counted for Pex14-labeled organelles was approximately 267±60 per 100 μm 2 (mean ± S.D., n = 25). The specific vesicular localization of Pex14 and Pex16 was further supported by immunoelectron microscopy. Weak but highly specific signals for His-tagged Pex14 were associated with the membranes of round vesicles of approximately 90–100 nm in diameter ( Fig 7A and 7B ). Similarly, His-tagged Pex16 was observed at the vesicular membrane. We also observed Pex16-labeled vesicles that were partially surrounded by two more membranes, with Pex16 signals on the proximal and distal membranes. The character of these structures is unclear though they remind lamellar derivatives of the endoplasmic reticulum that deliver Pex16 and other membrane proteins for de novo formation of putative peroxisomes ( Fig 7C and 7D ) [ 38 ].

The subcellular localization of PEXs was studied in transfectants expressing Pex5, 11, 14, 16, and 19. First, we analyzed recombinant proteins using immunoblot analysis of these dimentable (organelle) and soluble (cytosolic) fractions of transfectants (Figs 4 and S1 ). This analysis confirmed that all PEXs are expressed on the protein level, and more importantly, Pex5, 11, 14, and 16 were predominantly present in the sedimentable fraction, whereas the signal for Pex19 was comparably strong in the organellar fraction and the cytosol. Iron-containing superoxide dismutase (Fe-SOD) was used as a cytosolic marker ( Fig 4A ). For the soluble receptor Pex5, we expected to observe higher amounts of this protein in the cytosol. Thus, we performed a more detailed cell fractionation using differential centrifugation in five consecutive steps at 380–150,000 x g, and we used 0.05% Tween-20 in the buffer to limit protein aggregation. Under these conditions, a majority of Pex5 was in the soluble fraction, while the majority of the peroxisomal membrane Pex11 was present in sedimentable fractions ( Fig 4B ). This result indicates that Pex5 is transiently associated with organelles.

Identification of the putative PEX proteins prompted us to test whether the proteins are expressed and to what extent. Total RNA was isolated using cells upon reaching a monolayer in the culture (approximately 48 hours). RT-qPCR showed that under these conditions, all PEXs were expressed, but the relative expression level was lower than that of the selected housekeeping enzymes pyruvate kinase, phosphofructokinase, and alcohol dehydrogenase ( S3 Table ). The low expression level of putative PEXs corresponded to previously published transcriptomic data [ 34 ]. The highest expression was observed for Pex11, while Pex19 was the least expressed. Because the regulation of genes involved in the same function might be linked, we decided to clone Pex5, 11, 14, 16, and 19 into a vector that allowed their overexpression with a polyHis-tag under the control of a strong lectin promotor [ 34 , 35 ]. The transfectants revealed a relative PEX overexpression from 31 (Pex11) to 580 (Pex5)-fold ( S4 Table ). However, overexpression of a particular PEX gene had a negligible effect on native expression of other PEXs, with the exception of Pex14, overexpression of which was accompanied by slightly increased expression of Pex16 (3-fold).

Initially, we searched for PEXs in the genomes of E. histolytica and four other Entamoeba species, the human parasites E. dispar and E. moshkovskii, a parasite of non-human primates E. nuttalli, and a reptile pathogen E. invadens. The searches, which were performed using PEXs of a related amoebozoan M. balamuthi and HMMs of other eukaryotic PEX proteins as queries, revealed the presence of a reduced set of seven PEXs in all primate species ( Fig 1A and S1 Table ). The identified PEXs include members of each functional category ( Fig 1B ). Components of matrix protein imports were represented by Pex5 and Pex14. Pex5 is a soluble receptor that recognizes the PTS1 of proteins to be imported by binding to a domain formed by tetratricopeptide repeats within its C-terminal half ( Fig 2 ). A transient pore constituting of Pex14 and Pex5 is formed after the interaction of Pex5 with the docking complex by binding to characteristic motifs in the N-terminal half of Pex14 ( Fig 2 ) [ 29 ]. Members of membrane protein import include Pex16, a protein involved in embedding of peroxisomal membrane proteins, and Pex19, a membrane protein receptor. Although E. histolytica Pex19 is considerably shorter than the human ortholog, its classification to the Pex19 family (PF04614) is strongly supported, although it lacks the N-terminal domain that is required for interaction with Pex3 [ 30 , 31 ] ( Fig 2 ). We did not identify any member of the RING finger proteins of the receptor recycling machinery (Pex2, Pex10, Pex12); however, we detected putative Pex1 and Pex6, which are responsible for receptor deubiquitinylation ( Fig 1B ). Because Pex1 and Pex6 possess AAA domains that are present in other proteins and may lead to false-positive identification, we also performed a phylogenetic analysis of Pex1/Pex6 orthologs and related AAA proteins with different functions ( Fig 3 ). This analysis supported the correct identification of putative Entamoeba Pex1 and Pex6. Finally, we identified Pex11, a multi-purpose protein, which is involved in peroxisomal proliferation by elongation and fission and metabolism via the formation of a membrane channel [ 32 , 33 ]. Surprisingly, we did not identify any PEXs in E. invadens except for Pex19 ( Fig 1A ). All identified proteins showed apparent divergence from canonical sequences of yeast and human orthologues, with protein sequence identity ranging from 8–23% ( S2 Table ). Taken together, our searches suggested that E. histolytica and three other Entamoeba species have a minimal set of seven PEXs that might be able to facilitate peroxisomal biogenesis, while E. invadens most likely lacks these organelles, with only Pex19 being present.

Discussion

E. histolytica is believed to be devoid of peroxisomes, like most anaerobic protists [26,50]. In this work, we provided the first evidence that peroxisomes are present in this parasite, although E. histolytica peroxisomes seem to be remarkably reduced in comparison to their known peroxisome counterparts. E. histolytica contains only seven homologs of known PEXs (Pex1, Pex6, Pex5, Pex11, Pex14, Pex16, and Pex19) that may participate in organelle biogenesis. Targeting of matrix proteins to peroxisomes is reduced to the PTS1-dependent pathway mediated via the soluble Pex5 receptor, while the PTS2 receptor Pex7 is absent. Proteomic analyses of affinity-purified peroxisomes and in silico PTS1 predictions led to the identification of peroxisomal myo-IDH. Importantly, E. histolytica myo-IDH with PTS1 was shared with E. dispar, E. nutalli and E. moshkovskii and with evolutionarily related free-living M. balamuthi and P. schiedti, which all belong to the Archamoebae group of anaerobic protists, whereas E. invadens most likely lost peroxisomal myo-IDH together with peroxisomes.

The presence of peroxisomes in E. histolytica is supported by the localization of Pex5, Pex11, Pex14, Pex16, and Pex19 to numerous vesicles that are distinct from the structures of the ER, endosomal/lysosomal vesicles and mitosomes. Pex14 and Pex16 were detected in single membrane-bound vesicles of approximately 90–100 nm in diameter, which is within the range of anaerobic peroxisomes (80–440 nm) observed in M. balamuthi [27]. The matrix protein myo-IDH colocalized with Pox1 in yeast peroxisomes and within organelles labeled with Pex14, Pex16, and Pex11 in E. histolytica. Deletion of PTS1 tripeptide caused partial mistargeting of myo-IDH to the cytosol.

There are multiple metabolic interplays between mitochondria and peroxisomes, and in some cases, mitochondria may participate in peroxisome biogenesis [25]. However, the metabolism of E. histolytica mitochondria (mitosome) is reduced to a single pathway, and their protein import machinery contains only 2 out of 30 components known in model yeast mitochondria [9,51,52]. In this view, the observed reduction in peroxisomes is consistent with overall changes during the course of reductive evolution that formed Entamoeba species [5]. Most eukaryotes possess 13–17 PEXs, including humans and M. balamuthi, and four PEXs, Pex3, Pex10, Pex12, and Pex19, are considered peroxisomal markers absent only in organisms devoid of peroxisomes [53]. Three of these markers (Pex3, Pex10, and Pex12) were not identified in E. histolytica (Fig 1A). Pex3, together with Pex16 and Pex19, are involved in peroxisomal membrane protein (PMP) import.Pex3 facilitated the membrane docking of newly synthesized PMPs delivered via the shuttling receptor Pex19 directly from the cytosol (class I pathway for PMP import). Pex3 itself and other PMPs are recruited to the ER membrane by Pex16, which is cotranslationally inserted into the ER and subsequently delivered to peroxisomes through vesicular transport (class II pathway) [25,38]. We speculate that Pex16 might be sufficient for PMP import without a Pex3 contribution in E. histolytica. In support of this possibility, the formation of membrane structures with PMPs was observed in cells in which pex3was deleted [54], and the budding of pre-peroxisomal vesicles was dependent on Pex19, a homolog of which we found in Entamoeba [55]. Moreover, the presence of Pex16 and Pex19 without Pex3 was noticed in Tetrahymena thermophila [50]. Nevertheless, we cannot rule out the possibility that E. histolytica possesses a highly divergent Pex3 that was not recognized by current bioinformatic tools. Indeed, the highly divergent trypanosomal Pex3 remained elusive for a long time, and it has been identified only recently [56,57]. However, the absence of the N-terminal domain of E. histolytica Pex19, which is essential for its interaction with Pex3 [30,31] is more consistent with the lack of Pex3 in this organism. Interestingly, Pex19 is the only peroxin, homolog of which we identified in E. invadens that likely lack peroxisomes. Thus, Pex19 may have a more general function in protein sorting, which is unrelated to peroxisomes. Pex10 and Pex12 (RING complex) belong to the Zn-RING finger E3 ubiquitin-protein ligase family that, together with the E2 conjugating protein Pex4, participate in Pex5 recycling.Pex5 with peroxisomal cargo associates with membrane-docking components to form a transient pore with Pex14. Upon the release of the cargo, Pex5 is dislocated from the membrane in an ATP-dependent manner by Pex1 and Pex6 (dislocase complex) [58], which we both identified in E. histolytica. Recognition of Pex5 by the dislocase complex is dependent on Pex5 ubiquitination via thiol ester bond at the conserved cysteine at the N-terminus by the RING complex [59]. The absence of Pex10, Pex12, and Pex4 in E. histolytica suggests that Pex5 might be recycled without a ubiquitination step. In support of this view, E. histolytica Pex5 lacks an N-terminal cysteine residue. Moreover, in Pichia pastoris, PTS receptors and Pex14 have been shown to form the minimal translocation machinery that can facilitate peroxisomal import independent of the RING complex [60]. Alternatively, other E3 and E2 family proteins may compensate for the absence of the RING complex and Pex4. E. histolytica possesses genes for at least six E3 ligases, of which the E3 ligase EHI_030770 was found in the proteome of the peroxisome-enriched fraction, although, unlike Pex10 and Pex12, EHI_030770 does not possess any predicted transmembrane domain (S7 Table). There are also two E2 ubiquitin conjugating (Ubc) proteins (EHI_083560, EHI_048700) in the proteome, of which EHI_083560 (named EhUbc5) was partially characterized and crystalized, however, its target proteins are not known [61]. In metazoans and fungi, Pex4 is absent and its function is compensated by Ubc-conjugating proteins [62,63]. However, replacement of Pex10 and Pex12 has not been observed thus far. In contrast, the absence of the PTS2 receptor Pex7 observed in E. histolytica is not unprecedented, and when Pex7 is present, PTS2 targets peroxisomes much less frequently than PTS1s [64]. For example, Pex7 was found in related M. balamuthi in which PTS2 was predicted in 15% of putative peroxisomal proteins [27]. The absence of Pex7 was noticed in nematodes, several arthropods, the rhodophyte Galdieria sulpharia, and the stramenopile Thalassiosira pseudonana [50,65].

Interestingly, E. histolytica Pex11 displayed dual localization in peroxisomes and mitosomes. Previously, mitochondrial localization of Pex11 was observed only in cells such as yeast and human fibroblasts that lack peroxisomes upon deletion of Pex3 and Pex19 [66,67]. It seems that Pex11 has some affinity to mitochondria, possibly via mitochondrial proteins, with which Pex11 interacts. Pex11 is involved in elongation and proliferation of peroxisomes, which share with mitochondria dynamin-related proteins (DRPs) for the peroxisomal division;Pex11 can interact with mitochondria via the ERMES complex [66] and the translocase of the outer mitochondrial membrane (TOM) complex receptor Tom22 [68]. Neither ERMES nor Tom22 are present in mitosomes of E. histolytica [69,70]; however, E. histolytica possesses two DRPs that are involved in the fission of mitosomes and may potentially interact with Pex11 [71]. Interestingly, Pex11 was previously identified in the mitosomal proteome, supporting its dual localization [9]. However, these proteomic data need to be considered with caution, as in our experiments, peroxisomes and mitosomes comigrated using standard density gradient separation and it is likely that peroxisomes contaminated the mitosomal proteome. We cannot rule out the possibility that the partial mitosomal localization of Pex11 was a result of its mislocalization due to protein overexpression. However, overexpression of Pex11 in E. histolytica estimated by RT-qPCR was the lowest compared to four other recombinant PEXs for which no association with mitosomes was observed.

The matrix of E. histolytica peroxisomes contains myo-IDH, which catalyzes reversible NAD+-dependent conversion of myo-inositol to keto-2-inositol.Myo-inositol is synthesized from glucose-6-phosphate by the activity of inositol 3-phosphate synthase and inositol-3-phosphatase, which are both present in E. histolytica. Myo-inositol is utilized for the synthesis of phosphatidylinositol and a spectrum of phosphoinositide derivatives that are produced by a set of phosphatidylinositol kinases and phosphatases [72,73]. These compounds have multiple roles as components of membrane lipids, they are used in cell signaling pathways, energy homeostasis, and as cytoprotective solutes [74,75]. In mammals, insects, yeast and Chlorophyta, the catabolism of myo-inositol is initiated by oxygen-dependent myo-inositol oxygenase that converts myo-inositol to glucuronic acid [76]. However, none of these eukaryotic biosynthetic or catabolic pathways required the activity of myo-IDH. Bacteria such as Klebsiella aerogenes can grow on myo-inositol and other cyclitols as sole carbon sources. Here, myo-IDH catalyzes the first step of the catabolic pathway, leading to the production of acetyl-CoA [77]. However, our searches did not identify any enzyme downstream from myo-IDH in this pathway in E. histolytica. The distribution of myo-IDH in eukaryotes is limited to a few lineages of algae [78] and invertebrates. The function of myo-IDH remains enigmatic; however, algae contain two isoenzymes that were predicted to localize in the cytosol and mitochondria. It could be hypothesized that predicted dual localization may allow shuttling of a reducing power between mitochondria and the cytosol via myo-inositol/keto-2-inositol transport across the mitochondrial membrane [78]. Interestingly, dual localization of myo-IDH in peroxisomes and the cytosol was predicted in M. balamuthi and P. schiedti, suggesting the possibility of NAD-linked redox shuttling between the cytosol and peroxisomes to maintain intraperoxisomal redox as in the case of NAD+-linked malate dehydrogenase and glycerol-3-phosphate dehydrogenase-dependent shuttle systems that maintain the redox balance in yeast peroxisomes [79]. In Entamoeba species, we found only peroxisomal myo-IDH, which makes this explanation less plausible, although dual localization of proteins with peroxisomal targeting under specific physiological conditions has been observed [80]. Although the biochemical context of E. histolytica myo-IDH needs further investigation, the enzyme displayed two unusual features. First, E. histolytica myo-IDH has a notably higher affinity for myo-inositol characterized by a Km value that is three and four orders of magnitude lower (Km 0.044 mM) than that of the characterized orthologs in Bacillus subtilis (Km 18 mM) and the red alga Galdieria sulphuraria (Km 430 mM). Similar to the bacterial enzyme, D-glucose and D-xylose can serve as substrates for E. histolytica myo-IDH with Km values increased to 8.5 mM and 13.2 mM, respectively. These values are still two orders of magnitude lower than those in the bacterial enzyme (167 mM and 190 mM) [43,81]. Second, E. histolytica myo-IDH appeared to be highly active as a homodimer. In B. subtilis and G. sulphuraria, the molecular weight of native myo-IDH was approximately 160 kDa, which indicated a homotetrameric structure, which has been confirmed by structural studies in B. subtilis [48] and Lactobacillus casei [82].

In conclusion, findings of peroxisomes in E. histolytica and previously in M. balamuthi erode the paradigm of peroxisome absence in anaerobes. It also suggested that a minimal set of only seven peroxins might be sufficient to build these organelles. Moreover, the specific parasite peroxisomes with a functional myo-IDH that is absent in host cells might be an interesting target for the development of antiparasitic drugs. However, this report represents only a starting point for further functional investigations of anaerobic peroxisomes in this important human parasite.

[END]

[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010041

(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/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.