A proteomic and cellular analysis of uropods in the pathogen Entamoeba histolytica

Abstract

Exposure of Entamoeba histolytica to specific ligands induces cell polarization via the activation of signalling pathways and cytoskeletal elements. The process leads to formation of a protruding pseudopod at the front of the cell and a retracting uropod at the rear. In the present study, we show that the uropod forms during the exposure of trophozoites to serum isolated from humans suffering of amoebiasis. To investigate uropod assembly, we used LC-MS/MS technology to identify protein components in isolated uropod fractions. The galactose/N-acetylgalactosamine lectin, the immunodominant antigen M17 (which is specifically recognized by serum from amoeba-infected persons) and a few other cells adhesion-related molecules were primarily involved. Actin-rich cytoskeleton components, GTPases from the Rac and Rab families, filamin, α-actinin and a newly identified ezrin-moesin-radixin protein were the main factors found to potentially interact with capped receptors. A set of specific cysteine proteases and a serine protease were enriched in isolated uropod fractions. However, biological assays indicated that cysteine proteases are not involved in uropod formation in E. histolytica, a fact in contrast to the situation in human motile immune cells. The surface proteins identified here are testable biomarkers which may be either recognized by the immune system and/or released into the circulation during amoebiasis.

Figure 1. Spatiotemporal analysis of the redistribution of Con A associated with the surface of E. histolytica.

Trophozoites (of 20–30 µm size) were incubated under the microscope at 37°C and fluorescent Con A was added at the starting time point. In vivo imaging was performed and the uropod formation process was detected by frames (indicated by numbers) recorded in a confocal microscope with a Nipkow disk device. A: the micrograph represents 100 images recorded as 10 images every second. Note the polarisation of fluorescent Con A over time and the increase in the extrusion of particles into the medium (white arrow). B: Enlarged frames from a chosen cell are presented, with the white star marking the end of the trophozoite at which the uropod is formed. The entire sequence lasted 11 seconds. See Video S1 for visualisation of details in real time.

Figure 2. Serum from amoebiasis patients promotes uropod formation in E. histolytica.

The micrographs represent confocal microsopy sections of fixed parasites following incubation with serum from patients with amoebic liver abscesses (A) or from healthy donors (B). The left panels show immunofluorescence images obtained after incubation with anti-human antibodies. The right panels show the overlaid phase contrast/fluorescence images of entire parasites. The detector gain for fluorescence was increased in images shown in panel B. Scale bars: 10 µm.

Figure 3. Distribution in functional categories of the proteins present in the E. histolytica uropod extruded fraction.

A. Electrophoretic analysis of proteins from the ConA-uropod complex and from crude extract. A sample of UEF or amoebic extracts (10 µg, U = uropod; A = amoebae) were resolved by SDS-PAGE. The Gal/GalNAc lectin heavy chain (170 kDa) and the Con A (24 kDa) were revealed by western blot. B. Protein identification with LC-MS/MS was followed by proteome comparisons using the BLAST computer program, GO annotations and manual annotations. Two LC-MS/MS experiments were performed. Only proteins identified by at least two peptides in each experiment were taken into account. In all, 104 proteins were present in both experiments and could be analyzed. The entire data set was submitted to Tranche (https://proteomecommons.org/tranche/) database.

Figure 4. Cysteine proteinases are present as pro-enzymes and active enzymes in the uropod extruded fractions.

A. Substrate gel electrophoresis of uropod extruded fractions (1 µg of proteins), which were separated by electrophoresis in SDS-PAGE co-polymerized with gelatine. To visualize the cysteine proteinase activity, gels were stained with Coomassie blue. The figure shows the inverted image. B. Cellular localisation of CP-A5, -A1 and -A2 in E. histolytica. Trophozoites were incubated with Con A (green). Upon incubation, the cells were fixed and stained for CP-A5 (up panel) or CP-A1 and -A2 (low panel) with specific antibodies (red). Scale bar: 10 µm.

Figure 5. Conservation of carbohydrate-binding residues in the galactose-binding-like domain of M17 homologues of Entamoeba.

A. The domain architecture of M17 (EHI_015380). The transmembrane domain and the galactose binding-like domain (IPR008979) were identified using Philius and InterProScan software packages, respectively. B. The amino acid sequence alignment of the carbohydrate-binding domain of M17 homologues in Entamoeba and GH84C of C. perfringens (PBD ID: 2V5D_A). Residues with >75% identity are highlighted. M17 homologues of E. dispar (prefix EDI) and E. invadens (prefix EIN) were identified using BLASTP analysis of their proteomes, with M17 as the query. The arrows indicate the carbohydrate binding residues in GH84C . C. Predicted structural model of the galactose-binding like domain of M17. The model was predicted from the 3D jury meta-server , with C. perfringens GH84C as the best-hit template (i.e. the template with the highest 3D-jury score = 81.56; score of 50 is the default cut-off, which results in a prediction accuracy of above 90%). The side-chains of the three conserved carbohydrate binding residues are coloured and labelled as in panel B.

Figure 6. The cellular localization of the immunodominant antigen M17 in E. histolytica.

Trophozoites were incubated with serum from healthy patients (left panels), with serum from patients with amoebic liver abscesses (middle panes) or with green fluorescent Con A (right panels). Upon incubation, the cells were fixed and stained for M17 with a specific antibody (red) and co-stained with a serum recognizing anti-human IgG (left and middle panels). Scale bar: 10 µm.