Mitosomes in Trophozoites and Cysts of the Reptilian Parasite Entamoeba invadens

Abstract

Heat shock protein genes led to the discovery of mitosomes in Entamoeba histolytica, but mitosomes have not been described for any other Entamoeba species, nor have they been identified in the cyst stage. Here, we show that the distantly related reptilian pathogen Entamoeba invadens contains mitosomes, in both trophozoites and cysts, suggesting all Entamoeba species contain these organelles.

TEXT

Mitochondria have played a crucial role during eukaryotic evolution. They enabled and facilitated the development of multicellular life by relaxing the energetic constraints facing prokaryotes (11). The discovery of genes encoding proteins that are normally targeted to mitochondria in other eukaryotes (3) in the genome of Entamoeba histolytica, an organism previously thought to represent an earlier phase of eukaryotic life, raised doubts about the presumed premitochondrial status of this intestinal parasite. Antibodies raised against one of these, the mitochondrial chaperonin Hsp60, clearly showed the presence of an organelle, which was called a mitosome (21) or crypton (13). When the mitochondrion-like leader sequence of Hsp60 was removed, the protein accumulated in the cytosol, a phenotype that could be reversed by replacing the presequence with a genuine mitochondrial targeting sequence from another species, suggesting that the discovered organelle was indeed mitochondrial in nature (21).

Although mitosomes have been discovered in other former Archezoa (22, 25), the function of these organelles is not ob-vious. Analyses of the genomes of these human pathogens only suggested a handful of genes whose products are targeted to mitosomes. Most of these are “structural” in nature and encode heat shock proteins and metabolite or protein importers. Only a few “functional” enzymes have been discovered, and those involved in iron-sulfur cluster assembly seem to be a common denominator for all mitosomes. Unexpectedly, these proteins of clear mitochondrial ancestry have been replaced by lateral gene transfer with a much simpler system in E. histolytica (1, 24). Whether these proteins are genuinely mitosomal in this organism remains a matter of dispute (14, 16). The most thorough attempt to understand mitosomal function in E. histolytica employed mass spectroscopy on Percoll-purified mitosomes (16). Frustratingly, two-thirds of the 95 identified proteins were hypothetical proteins. However, this study indicated that E. histolytica mitosomes are involved in sulfate activation (16), a thus far unique mitosomal trait of the Entamoeba organelles.

There are two stages in the E. histolytica life cycle: a motile trophozoite stage found inside the human host and a resistant infectious cyst which is excreted by infected individuals. The factors controlling encystation and excystation in E. histolytica are poorly understood, and our lack of knowledge is further hampered by our inability to induce cyst formation in vitro. As a consequence, nothing is known about what happens to mitosomes in cysts. The reptilian pathogen Entamoeba invadens acts as a proxy for the study of cyst formation as it is relatively straightforward to induce encystation in this Entamoeba species (18). As mitosomes have never been identified in any other member of the genus Entamoeba and there is no knowledge regarding the fate of mitosomes in cysts, we have studied the distribution of mitosomes in the reptilian pathogen Entamoeba invadens.

One of us previously demonstrated that the mitochondrion-type Hsp70 (mHsp70) chaperone is enriched in mitosomal fractions (20). In order to validate that work, which was performed using a heterologous mHsp70 antibody (23), we obtained a homologous antibody raised against recombinant E. histolytica mitosomal Hsp70. To maximize recombinant protein production, codon usage was converted from E. histolytica to Escherichia coli using JCat (7). Subsequently, a synthetic gene was constructed, which included a C-terminal histidine tag for protein purification and the restriction site BamHI at both termini to enable cloning into the BamHI site of the pET-3c expression vector. The poly-His-tagged recombinant protein was produced in Escherichia coli BL21(DE3)(pLysY) cells and purified under nondenaturing conditions by immobilized-metal ion-affinity chromatography using Ni-nitrilotriacetic acid (NTA). Correct protein identity was verified by mass spectroscopy, and this protein was subsequently used for immunization. E. histolytica HM-1:IMSS and E. invadens IP-1 were grown using standard conditions. Cyst formation was induced according to established protocols (18). For E. invadens trophozoite localization experiments, cells were washed in phosphate-buffered saline (PBS) and fixed with 3% formaldehyde in PBS for 45 min while mature cysts were fixed using 3% formaldehyde in PBS overnight at 4°C. The dehydrated specimens were rehydrated with PBS for 30 min, permeabilized with 0.2% Triton X-100 in PBS for 20 min at room temperature, and blocked for 2 h with 2% bovine serum albumin (BSA) in PBS. The cell preparations were incubated with titrated E. histolytica Hsp60 (1:300) (a kind gift of C. Graham Clark) and E. histolytica mHsp70 (1:100) (this study) antibodies in PBS with 2% BSA and 0.2% Triton X-100 for 1 h at room temperature in a humid chamber. Secondary antibodies coupled to Alexa Fluor 594 and 488 (Invitrogen), respectively, were used to detect bound antibodies. Specimens were thoroughly washed in PBS with 0.5% BSA and 0.05% Triton X-100 between incubations and finally embedded with Vectashield (Vector Labs) or Glycergel (Dako) mounting medium. Nuclear DNA was detected with the intercalating agent 4′,6-diamidino-2-phenylindole (DAPI). Immunofluorescence image data collection was performed on a Leica SP2 AOBS confocal laser-scanning microscope (Leica Microsystems, Wetzlar, Germany) with an oil immersion objective (Leica, HCX PL APO 63 × 1.4) and a pinhole setting of Airy 1 with 2-fold oversampling. Image stacks were further processed using the Huygens deconvolution software package, version 2.7 (Scientific Volume Imaging, Hilversum, Netherlands). Three-dimensional reconstruction, volume rendering, and colocalization analysis were done with the Imaris software suite, version 7.2 (Bitplane, Zurich, Switzerland).

In order to understand the phylogenetic relationship of amoebozoan mHsp70, a data set of 39 protein sequences with 25 eukaryotic and 14 prokaryotic taxa was assembled. Protein sequences were aligned using ClustalW in SeaView version 4.2.12. The data set contained 620 informative patterns from a total of 704 sites. Phylogenies were calculated using the model-based maximum likelihood approach (ML) using PhyML (8) and the Bayesian inference approach using MrBayes (17). For ML analyses, modelgenerator v.0.85 (10) suggested the model LG+G+F, with 8 rate categories and an alpha shape parameter of 0.48 to fit the observed data best. Four Bayesian analyses were run using a mixed-amino-acid model accommodating 4 rate+inv categories containing 4 chains each. One million generations were calculated, and trees were sampled every 1,000 generations. The model stabilized rapidly, and 250 trees were discarded as burn-in. Mitochondrial targeting signals were analyzed using the localization prediction tools WoLF PSORT (9) and Mitoprot (4).

The presence of mitosomes in E. histolytica is well documented, but there is no information regarding the presence of these organelles for any other Entamoeba species. As there is (partial) genome information available for several other Entamoeba species, we decided to screen these genomes for the presence of Hsp60 and mHsp70. Putative Hsp60 and mHsp70 sequences were identified in Entamoeba dispar and E. invadens, but incomplete Hsp60 sequences lacking their N termini could only be identified for Entamoeba terrapinae and Entamoeba moshkovskii, while no mHsp70 sequences could be identified with reasonable certainty in these two species. As shown before (6), targeting signal prediction programs have difficulties identifying Entamoeba mitosomal presequences, but alignment of the N termini of Hsp60 and mHsp70 clearly indicates the presence of presequences that are upstream of the analogous prokaryotic N terminus and which have been shown to be genuine targeting signals in E. histolytica (Fig. 1).

Fig. 1.

Analyses of the amino-terminal regions of the mitochondrial chaperones Hsp60 (A) and mHsp70 (B) from the Amoebozoa. The bacterial homologue from Rickettsia prowazekii is shown for comparison. Acanthamoeba castellanii Hsp60 is not shown due to being incomplete at the N terminus. Predicted or confirmed cleavage sites are indicated by a dash, while the asterisk denotes an incomplete N terminus. The underlined residues for E. histolytica Hsp60 have been shown to be required for mitosomal targeting (21).

In order to confirm the mitochondrial nature of the identified putative mHsp70, we conducted phylogenetic analyses, including extended amoebozoan sampling (Fig. 2). Our analyses clearly confirm the mitochondrial ancestry and monophyly of amoebozoan mHsp70s. Although we correctly recovered the sister group relationship with the opisthokonts, this node was only weakly supported. As shown previously (2), alphaproteobacteria are basal to all eukaryotes in accordance with their supposed role as donor of the mitochondrial endosymbiont.

Fig. 2.

Phylogenetic relationships of eukaryotic mHsp70 and prokaryotic DnaK homologues. An unrooted maximum likelihood tree produced by PhyML is shown. Circles at the nodes represent bootstrap values as determined using PhyML and posterior probabilities (pp) as determined by MrBayes (black circle, bootstrap of >60% and pp of >0.9; half-filled circle, black on the left, bootstrap of >60% and pp of <0.9; half-filled circle, black on the right, bootstrap of <60% and pp of >0.9). The long branch leading to the Entamoeba spp. has been shortened by 50%. Acanthamoeba castellanii mHsp70 was removed from the analyses due to its erratic placement in the analyses, in addition to its failing the compositional χ² test in PUZZLE. Accession numbers of the sequences used are available on request.

To verify our in silico analyses, which clearly predict the Entamoeba mHsp70 to be mitochondrial, we conducted laser-scanning confocal microscopy and three-dimensional image rendering using the homologous Hsp60 and mHsp70 antisera on E. histolytica whole-cell preparations. The Hsp60 antibody localizes in a discrete punctate pattern and abundance, similar to previous reports (12, 16), and the homologous mHsp70 antiserum colocalizes to the same areas (Fig. 3 A to D), confirming the earlier fractionation data (20). When these antisera were used on the distantly related E. invadens, a similar localization pattern was observed. However, as the representative images in Fig. 3 and 4 show, we regularly detected at least 10-fold fewer organelles in E. invadens. This clearly suggests that this distant Entamoeba species (19) contains mitosomes, as well allowing us to suggest all species in the genus do contain this organelle. Although perhaps discounted by most, some still entertain the possibility of genuine extant Archezoa without mitochondria (5, 15). Systematic demonstration of mitochondria (or mitosomes/hydrogenosomes) in all branches of eukaryotes therefore has its merits.

Fig. 3.

Localization of Entamoeba histolytica and E. invadens mitosomes using Hsp60 and mHsp70 antibodies. Presented are orthogonal projection views of the volume images reconstructed from image stacks showing E. histolytica (A to D) and E. invadens (E to H) trophozoites with DNA labeling using DAPI. (A and E) Hsp60 labeling. (Inset) Corresponding bright-field image. (B and F) mHsp70 labeling. (C and G) Merged image. (D and H) Projection images of the colocalization channel generated with the Imaris colocalization module (gray) depicting voxels in which significant overlap of both labels was detected. (Inset) Scatterplots indicating the extent of colocalization within the volume image. Bars = 3 μm.

Fig. 4.

Orthogonal projection views of the volume images reconstructed from image stacks showing the localization of Entamoeba invadens mitosomes in mature cysts using Hsp60 and mHsp70 antibodies. DNA was labeled using DAPI. (A) Hsp60 labeling. (Inset) Corresponding bright-field image. (B) mHsp70 labeling. (C) Merged image. (D) Depiction of voxels with signal overlap as in Fig. 3D and H. Bars = 2 μm.

Finally, as no information about mitosomes in cysts exists for any species, we decided to address this issue using E. invadens, where encystation can easily be induced in vitro (18). The distribution and abundance of mitosomes resemble those of E. invadens trophozoites (Fig. 4). No mitosome-organizing center can be identified in cysts, suggesting that the inheritance of these organelles in Entamoeba is stochastic.

In conclusion, our work clearly demonstrates the presence of mitosomes in the reptilian parasite E. invadens, which is distantly related to E. histolytica. The presence of these organelles in both E. histolytica and E. invadens suggests that all Entamoeba spp. contain this organelle. We also show that mitosomes are abundant in the infectious cysts, suggesting that these enigmatic organelles may play a role in this important life cycle stage as well.