Entamoeba histolytica: Quantitative Proteomics Analysis Reveals Putative Virulence-Associated Differentially Abundant Membrane Proteins

Abstract.

Entamoeba histolytica is a protozoan parasite that causes amebiasis and poses a significant health risk for populations in endemic areas. The molecular mechanisms involved in the pathogenesis and regulation of the parasite are not well characterized. We aimed to identify and quantify the differentially abundant membrane proteins by comparing the membrane proteins of virulent and avirulent variants of E. histolytica HM-1:IMSS, and to investigate the potential associations among the differentially abundant membrane proteins. We performed quantitative proteomics analysis using isobaric tags for relative and absolute quantitation labeling, in combination with two mass spectrometry instruments, that is, nano-liquid chromatography (nanoLC)-matrix-assisted laser desorption/ionization-mass spectrometry/mass spectrometry and nanoLC-electrospray ionization tandem mass spectrometry. Overall, 37 membrane proteins were found to be differentially abundant, whereby 19 and 18 membrane proteins of the virulent variant of E. histolytica increased and decreased in abundance, respectively. Proteins that were differentially abundant include Rho family GTPase, calreticulin, a 70-kDa heat shock protein, and hypothetical proteins. Analysis by Protein ANalysis THrough Evolutionary Relationships database revealed that the differentially abundant membrane proteins were mainly involved in catalytic activities (29.7%) and metabolic processes (32.4%). Differentially abundant membrane proteins that were found to be involved mainly in the catalytic activities and the metabolic processes were highlighted together with their putative roles in relation to the virulence. Further investigations should be performed to elucidate the roles of these proteins in E. histolytica pathogenesis.

Introduction

Entamoeba histolytica is an enteric protozoan parasite that causes amebiasis in human and nonhuman primates. The invasive trophozoites of E. histolytica are capable to inhabit host tissues and provoke diseases such as intestinal colitis and liver abscess. However, the molecular mechanism and pathogenesis of amebiasis need further investigations. Entamoeba histolytica strains have been reported to exhibit different degrees of virulence, resulting in various clinical symptoms in patients.¹ Several amebic virulence determinants have been characterized, such as galactose/N-acetyl galactosamine-inhibitable lectin (Gal/GalNAc lectin), cysteine proteases (CPs), amebapore, and phagosome-associated proteins. Galactose/N-acetyl galactosamine-inhibitable lectin was reported to be responsible for the adherence of trophozoites to host cells, cytolysis, parasite invasion, and phagocytosis.^1–3 Cysteine proteases have been implicated in cytopathic activities such as damage of host cells and host tissue, induction of intestinal inflammation, and digestion of some components of the extracellular matrix.^4,5 Phagosome-associated proteins such as EhRac A, EhPAK (p21-activated serine/threonine protein kinase), and actin have been reported to be involved in endocytosis and disease pathogenesis.⁶

Membrane surface is the first outer layer where the trophozoites and host cell come into contact. The adherence of trophozoites to the host cells, which is mediated by Gal/GalNAc lectin, is an important step for the disease development. Furthermore, cell–cell adhesion is initiated by cell adhesion molecules, which are also members of integral membrane proteins.⁷ Membrane proteins have transmembrane (TM) domains and/or signal peptides to facilitate the biological functions of the organism. For instance, TM domain-containing proteins are responsible for cellular recognition and adhesion, molecular receptors, substrate transportation through membranes, signal transduction, protein secretion, and enzymatic activity.⁸ Meanwhile, signal peptides are responsible for leading the arrangement of proteins into cellular organelles or cell membrane.⁹ The important functions exhibited by the TM domains and signal peptides suggested that membrane proteins play essential roles in membrane trafficking and vesicular trafficking machinery in relation to the expression of virulence factors of E. histolytica. This can be related to a study by Nozaki and Nakada-Tsukui in 2006¹⁰ in which membrane protein candidates including CPs and amebapores were found to be responsible for cytopathic activities during phagocytosis. Other membrane-associated proteins such as small GTPase and Rab proteins also play important roles in membrane trafficking because they serve as molecular switches that coordinate the sequential vesicle fusion with target membranes during phagosome maturation.¹¹

The single-cell E. histolytica lacks rough endoplasmic reticulum (ER) and Golgi apparatus. However, similar to other eukaryotic cells, it contains the essential components that encode the basic vesicular transport system.^12,13 This system, which comprises membrane-associated proteins, plays a vital role in the pathogenesis of amebiasis, although the transportation of pathogenic factors or virulence determinants into the host remains unknown. According to the findings by Biller et al.,¹⁴ the plasma membrane of E. histolytica is greatly interconnected and persistently exchanges with the intracellular membrane systems; thus, proteins may be present in the extracellular system without any hint of their plasma membrane association. To date, the numbers of in-depth studies of membrane proteomes of E. histolytica are limited, and there are still many E. histolytica hypothetical proteins that remain unknown. In this study, we performed differential membrane protein abundance profiling analysis between virulent (vEh) and avirulent (avEh) variants of E. histolytica (HM-1:IMSS) to further understand their functions and interactions. Because virulence varies among strains, we restricted this comparison to E. histolytica strain HM-1:IMSS.

Material and Methods

Ethics statement.

All experiments involving animals were strictly performed according to the Mexican Law for the Production, Care, and Use of Laboratory animals (NOM-062-ZOO-1999). All animal procedures were performed under the protocol number 091-2016, approved by the Institutional Animal Care and Use Committees of the Facultad de Medicina, Universidad Nacional Autónoma de México. All efforts were made to minimize animal suffering.

Entamoeba histolytica (HM-1:IMSS) amebic culture.

Axenic cultures of the vEh and avEh E. histolytica HM-1:IMSS trophozoites were maintained in TYI-S-33 medium according to standard protocols by Diamond et al.¹⁵ Trophozoites were harvested after 72 hours and washed with phosphate-buffered saline, pH 7.2. A total of 1 × 10⁷ parasites were aliquoted into each tube and immediately lyophilized. The virulence of the vEh trophozoites was maintained by periodically recovering them from amebic liver abscess of infected hamsters according to the protocol from Olivos et al.¹⁶ Meanwhile, the avEh trophozoites were vEh trophozoites that lost the ability to cause amebic liver abscess because of the long-term in vitro maintenance of the culture (> 18 years).^16,17

Membrane protein extraction.

The membrane proteins of vEh and avEh trophozoites were extracted using the ProteoExtract^™ Native Membrane Protein Extraction kit (Calbiochem, San Diego, CA) according to the manufacturer’s protocol. The separated cytosolic fraction from the membrane fraction was kept for further analysis. The extracted vEh and avEh membrane proteins were each subjected to acetone precipitation by adding six times volume of 100% ice-cold acetone to the protein sample, then vortexed, and incubated at −20°C for 4 hours. The precipitated protein was redissolved in 100 µL of dissolution buffer (pH 8.5, 0.5 M triethylammonium bicarbonate, 0.1% sodium dodecyl sulfate [SDS]) before isobaric tags for relative and absolute quantitation (iTRAQ^®) labeling. The amount of extracted membrane protein was determined using RCDC assay (Bio-Rad Laboratories, Hercules, CA). The absorbance of the protein sample was read at 750 nm, using a NanoDrop^™ 2000 spectrophotometer (Thermo Scientific, Wilmington, DE).

To assess the extracted membrane protein profile, 40 µg each of vEh and avEh membrane proteins were separated by 10% SDS-PAGE using the Mini-PROTEAN^® electrophoresis system (Bio-Rad Laboratories) for 90 minutes at 100 V. The gel was washed using water with gentle agitation for 5 minutes, for three times. Subsequently, overnight staining was performed using RAMA stain (containing 0.05% Coomassie brilliant blue R250, 10% acetic acid, 15% methanol, and 3% ammonium sulfate). The gel was destained with water until the protein bands were clearly visible.

Protein digestion and iTRAQ labeling.

Duplex iTRAQ analysis was performed using two isobaric tags (114 and 117). Hundred microgram each of vEh and avEh membrane proteins was treated with 1 µL denaturant (2% SDS), followed by adding 2 µL reducing reagent (50mM tris-[2-carboxyethyl] phosphine), and then incubated at 60°C for 1 hour. The reduced cysteine residues were alkylated with cysteine blocking reagent (200 mM methyl methanethiosulfonate [MMTS] in isopropanol) at room temperature for 10 minutes. The protein samples were then digested by adding 10 µL trypsin solution and incubated at 37°C for 12–16 hours. The digested peptides were labeled with iTRAQ reagent (AB Sciex, Foster City, CA) at room temperature for 1 hour, whereby the avEh and vEh peptides were labeled with 114 and 117 isobaric tags, respectively. Finally, the labeled peptides of avEh and vEh were pooled into a tube. Cation exchange was then performed using ICAT Cation Exchange hardware and ICAT Cation Exchange Buffer Pack (Applied Biosystems, Foster City, CA) by following the manufacturer’s protocol.

Nano-liquid chromatography coupling MALDI-MS/MS analysis.

Peptides separation was performed using a NanoLC-2D Ultra system (Eksigent, Sprockhövel, Germany) connected to an automated Eksigent EKSpot MALDI spotter (Eksigent). In the tray of the autosampler, 5 µL of peptide sample was first auto-loaded and packed into a ProteCol C18 trap column (internal diameter: 300 µm; length: 50 mm; pore size: 300 Å; and particle size: 3 µm). Mobile phase buffer A (2% acetonitrile [ACN], 0.1% trifluoroacetic acid [TFA]) and buffer B (98% ACN, 0.1% TFA) were used during reversed phase gradient elution. The flow profile for gradient elution of peptides was set with 20–80% of acetonitrile for a duration of 165 minutes and a flow rate of 300 nL/minutes. The eluted peptide fractions were spotted directly on the 384-well MALDI plate, with a concentration of 5 mg/mL α-cyano-4-hydroxycinnamic acid matrix (Sigma Aldrich, St. Louis, MO) continuously being added to the column effluent via a mixing tee piece at a flow rate of 1.8 µL/minute, from the 55th to the 135th minute of the gradient phase. The chromatogram was recorded at a wavelength of 214 nm by using a SONNTEK UV detector 2500 (SONNTEK, Upper Saddle River, NJ).

Tandem mass spectrometry analysis was performed using a MALDI-TOF/TOF^™ 5800 system (AB Sciex, Framingham, MA). MS spectra were acquired between m/z range 800 and 4,000, and the precursor ion selections were set by accumulating up to 500 laser shots per spectrum. Signal-to-noise (S/N) ratio was set at a minimum of 10, and spots with most intense precursor ions were subjected to tandem mass spectrometry (MS/MS) analysis. Mass spectrometry spectra were acquired with a maximum of 10 precursors, by accumulating up to 2000 laser shots per spectrum, and set to a minimum ratio of 15 S/N. External calibration of mass spectra was performed using calibration mixture 1 from AB Sciex. Five technical runs were performed on every biological replicate to ensure the result consistency of the protein identification.

Nano-liquid chromatography coupling electrospray ionization tandem mass spectrometry (ESI-MS/MS) analysis.

The analysis was performed at the Protein and Proteomics Centre, National University of Singapore. Briefly, peptide separation was performed using nanoLC Ultra and ChiPLC-nanoflex (Eksigent, Dublin, CA) in trap-elute configuration. The samples were first loaded on a 200-μm × 0.5-mm column (ChromXP C18-CL, 3 μm) and eluted on an analytical 75-μm × 15-cm column (ChromXP C18-CL, 3 μm). Mobile phase A (2% ACN and 0.1% formic acid) and mobile phase B (98% ACN and 0.1% formic acid) were used for peptide separation at a flow rate of 0.3 μL/minutes. The gradient elution was set as follows: 5–5% of mobile phase B in 1 minute, 5–12% of mobile phase B in 19 minutes, 12–30% of mobile phase B in 120 minutes, 30–70% of mobile phase B in 10 minutes, 70–90% of mobile phase B in 2 minutes, 90–90% in 7 minutes, 90–5% in 6 minutes, and held at 5% of mobile phase B for 10 minutes.

Tandem mass spectrometry analysis was performed using 5600 TripleTOF systems (AB SCIEX, Concord, ON). Mass spectrometry (MS) spectra were acquired between m/z range 400 and 1,800, and the precursor ion selections were set at accumulation times of 250 ms per spectrum. Mass spectrometry analysis was performed on the 20 most intense precursor ions, with an accumulation time of 100 ms per cycle, with 15 seconds dynamic exclusion. The MS/MS was acquired in the high-sensitivity mode using rolling collision energy.

Database search, criteria, and parameters.

Protein identification analysis for the iTRAQ experiment was executed using ProteinPilot^™ software version 4.5 (AB SCIEX). Paragon algorithm was used as the default program to search against AmoebaDB version 4.1 database (accessed October 15, 2017) and common Repository of Adventitious Proteins (cRAP) for peptide identification and isoform-specific quantification. The cRAP includes the possible contaminant proteins in this study such as bovine serum albumin and keratin (http://www.thegpm.org/crap/). For protein identification, a strict cutoff was applied to minimize false-positive results. Proteins with unused ProtScore ≥ 1.3 (confidence interval of 95%) and at least two peptides with 95% confidence were used as the identification criterion. Bias correction was also applied to the resulting data set to minimize any discrepancies conveyed because of the unequal mixing when the differently labeled samples were pooled. For iTRAQ quantification analysis, Pro Group algorithm automatically selects the peptide for quantification with the criterion of at least two peptides with 99% confidence, for the reporter peak area, error factor, and P-value calculations. The false discovery rate was set at < 1%, while a detected protein threshold of > 0.47 (confidence interval of 66%) and a competitor error margin of 2.00 were applied.

In Paragon algorithm configuration, parameters were set as follows: digestion: trypsin; cysteine alkylation: MMTS; instrument: TOF/TOF 5800 (for MALDI) and TripleTOF 5600 (for ESI); search effort: thorough; ID focus: biological modifications; false detection rate (FDR) analysis: yes; background correction: yes; bias correction: yes. The measurement used in ProteinPilot was derived from the 95% confidence interval (error factor) and calculated from the standard deviation in logspace.

For the selection of differentially abundant proteins, we considered the following criteria: 1) the proteins must contain at least two unique peptides (at least 95% confidence) and 2) the proteins must have P < 0.05 and the proteins identified with mass tag changes ratio must be > 2. A protein that was detected in two replicates by either LC-MALDI-MS/MS or LC-ESI-MS/MS was regarded as significant.

Cytosolic fraction.

In-solution digestion of the proteins in the cytosolic fraction was performed by following the method published by Ujang et al. (2016).¹⁸ The digested samples were analyzed via LC-ESI-MS/MS at Proteomics Core Facility, Malaysia Genome Institute, National Institutes of Biotechnology Malaysia. The peptides were separated using the Dionex UltiMate^™ 3000 RSLCnano (Thermo Fisher Scientific, San Jose, CA) nano-liquid chromatography system, and the MS analysis was performed using an Orbitrap Fusion^™ (Thermo Fisher Scientific, Bremen, Germany) mass spectrometer. To achieve spatial discrimination of the peptides, they were packed into and eluted with EASY-Spray Column PepMap^® RSLC, C18, 2 µm particle size, 50 µm id × 150 mm coupled with pre-column (µ-Precolumn PepMap 100, C18, 3 µm particle size, 300 µm id × 5 mm) at a flow rate of 0.3 µL/minutes with mobile phases A (deionized distilled water with 0.1% formic acid) and B (ACN with 0.1% formic acid). The elution gradient of the sample was set for 101 minutes with a gradient of mobile phase B, 5–95% for 93 minutes, 95% for 2 minutes, and then to 5% in 2 minutes. For MS analysis, the instrument was set in the data-dependent mode. The parameters for the full scan spectra were as follows: a scan range of 310–1,800 m/z, a resolving power of 120,000, an AGC target of 4.0e5 (400,000), and a maximum injection time of 50 ms. The method consisted of 3 seconds at the top speed mode where precursors were selected for a maximum of a 3-second cycle. Only precursors with an assigned monoisotopic m/z and a charge state of 2–7 were further analyzed for MS/MS. All precursors were filtered using a 20-second dynamic exclusion window and an intensity threshold of 5,000. The MS/MS spectra were analyzed using the following parameters: a rapid scan rate with a resolving power of 60,000, an AGC target of 1.0e2 (100), 1.6 m/z isolation window, and a maximum injection time of 250 ms. Precursors were fragmented by collision-induced dissociation and higher-energy collisional dissociation at a normalized collision energy of 30% and 28%, respectively. Data analysis and database matching against AmoebaDB version 4.1 and cRAP were performed using Proteome Discoverer^™ software version 2.1 (Thermo Scientific, Waltham, MA). The parameters used in the analysis were fixed modifications: carbamidomethylation (C), variable modification: oxidation (M), deamidation of asparagine (N) and glutamine (Q), maximum missed cleavage set at two, FDR < 0.1%, and parent mass and precursor mass tolerance at 10 ppm and 0.6 Da, respectively. The significant score (−10 lgP) for protein acceptance was set at > 20, whereas the minimum unique peptide was set at one.

Gene ontology, membrane protein prediction, and protein–protein interaction prediction.

The differentially abundant proteins were subjected to bioinformatics analysis using the Protein ANalysis THrough Evolutionary Relationships (PANTHER) DB classification system (http://www.pantherdb.org/) (accessed April 6, 2017), to identify their subcellular localization, molecular function, or biological process.

To assess the quality of the experiment, the predictions of membrane protein topology on the proteins were analyzed using a bioinformatics tool, TOPCONS (http://topcons.cbr.su.se/) (accessed April 6, 2017). A protein was considered as a membrane protein if it contained TM helix and/or signal peptide.

The protein–protein interaction prediction was performed using STRING software (version 10.5) (https://string-db.org/) (accessed March 13, 2018) to summarize the network of predicted protein interactions. The association in the protein–protein interaction network predicted by STRING covered either direct physical binding or indirect interaction, for example, involvement in the same metabolic pathway or cellular process. The network analysis of differentially abundant membrane proteins of the vEh variant of E. histolytica was performed separately, presented by the confidence view. The protein network under the confidence view which showed stronger associations was represented by thicker lines.

Radar plot was created with the differentially abundant proteins using Microsoft Excel to allow a clear depiction of the differentially abundant proteins with various fold changes.

Results

Protein identification.

In total, 1,307 proteins were identified from both biological replicates, where 194 and 1,113 proteins were identified by LC-MALDI-TOF/TOF and LC-ESI-MS/MS, respectively (Figure 1A). A total of 41.2% and 41.5% membrane proteins were predicted from the total proteins identified from LC-MALDI-TOF/TOF and LC-ESI-MS/MS, respectively (Figure 1B).

Figure 1.

(A) Venn diagram indicating the two sets of proteins that were analyzed independently by LC-MALDI-TOF/TOF and LC-ESI-MS/MS. (B) Membrane protein prediction using TOPCONS on the total identified proteins analyzed by LC-MALDI-TOF/TOF and LC-ESI-MS/MS. LC-ESI-MS/MS = liquid chromatography coupling electrospray ionization tandem mass spectrometry.

Cytosolic fraction.

To assess the specificity of the membrane extraction method in this experiment, the membrane protein prediction was performed by TOPCONS on the proteins identified from the cytosolic fraction (Figure 2A). A total of 556 proteins were identified from the cytosolic fraction by LC-ESI-MS/MS. TOPCONS predicted 54 of 556 (9.7%) of identified proteins were regarded as membrane proteins (Supplemental Information). The Venn diagram revealed that there were 43 proteins (4.4%) commonly detected in both membrane and cytosolic fractions (Figure 2B).

Figure 2.

(A) Membrane protein prediction using TOPCONS on total protein identified from cytosolic fraction. (B) Venn diagram indicating the comparison of total identified proteins from membrane fraction and cytosolic fraction.

Differentially abundant proteins.

In total, by combining the results from both mass spectrometer analyses, 75 differentially abundant proteins (36 increased and 39 decreased abundance) fulfilled our filtering criterion in the vEh versus avEh variants. When membrane protein prediction analysis was applied to all the identified differentially abundant proteins, there were 19 increased and 18 decreased proteins that were predicted as membrane proteins (Figure 3A and B). We also compared the identified differentially abundant proteins across both the biological replicates. There were three (30.0%) and 12 (16.9%) proteins commonly identified by LC-MALDI-TOF/TOF (Figure 3C) and LC-ESI-MS/MS, respectively (Figure 3D).

Figure 3.

(A) Membrane protein prediction using TOPCONS on the differentially abundant proteins that exhibit ≥ 2-fold changes in virulent versus avirulent variants. (B) Differentially abundant membrane proteins identified by LC-MALDI-TOF/TOF and LC-ESI-MS/MS. M1 and M2 represent two replicates detected from LC-MALDI-TOF/TOF and E1 and E2 represent two replicates detected from LC-ESI-MS/MS. (C) Differentially abundant membrane proteins identified by LC-MALDI-TOF/TOF in two technical replicates represented by M1 and M2. (D) Differentially abundant membrane proteins identified by LC-ESI-MS/MS in two technical replicates represented by E1 and E2.

The complete details of the analyses of differentially abundant proteins including membrane protein prediction, gene ontology functional classification, and protein–protein interaction prediction analysis are provided in the Supplemental Information.

Virulent versus avEh membrane proteins with increased abundance.

The combination of differentially abundant membrane proteins identified by both mass spectrometers revealed that 19 membrane proteins were identified with increased abundance in the vEh versus avEh variants (Table 1). The top five increased membrane proteins that exhibited the highest fold change were the hypothetical protein (4.7-fold); NAD(P) transhydrogenase subunit alpha, putative (4.4-fold); calreticulin (4.3-fold); beta-N-acetylhexosaminidase, beta subunit (4.2-fold); and dipeptidyl peptidase, putative (4.0-fold). Radar plot was used to represent the massive number of the increased membrane proteins with different fold changes in a clear visualization (Figure 4A).

Table 1.

List of identified increased abundance membrane proteins with fold change ≥ 2 in virulent versus avirulent variants

No	Accession number	Protein name	Peptides (95%)	% Cov	iTRAQ ratio (117:114)	Detected by*
1	EHI_007330	Beta-N-acetylhexosaminidase, beta subunit	23	36.6	2.109	LC-ESI-MS/MS-2
1	EHI_007330	Beta-N-acetylhexosaminidase, beta subunit	25	35.6	4.169	LC-ESI-MS/MS-1
2	EHI_007930	Hypothetical protein	14	38.3	2.070	LC-ESI-MS/MS-1
3	EHI_010710	Hypothetical protein	13	30.3	4.699	LC-ESI-MS/MS-1
4	EHI_014030	NAD(P) transhydrogenase subunit alpha, putative	43	25.0	4.406	LC-ESI-MS/MS-2
5	EHI_015380	Immuno-dominant variable surface antigen	40	39.7	2.606	LC-ESI-MS/MS-1
6	EHI_059830	Hypothetical protein	31	31.9	2.399	LC-ESI-MS/MS-2
7	EHI_065330	Gal/GalNAc lectin subunit Igl2	95	58.0	2.089	LC-ESI-MS/MS-1
8	EHI_068090	Serine carboxypeptidase (S28) family protein	15	28.7	2.377	LC-ESI-MS/MS-1
9	EHI_069560	Hypothetical protein	25	55.1	2.512	LC-ESI-MS/MS-1
10	EHI_097630	Phospholipase B, putative	19	39.1	2.109	LC-ESI-MS/MS-1
11	EHI_100320	Multidrug resistance protein, putative	34	18.4	2.489	LC-ESI-MS/MS-2
12	EHI_120590	Hypothetical protein	24	31.5	2.780	LC-ESI-MS/MS-1
12	EHI_120590	Hypothetical protein	19	23.0	3.467	LC-ESI-MS/MS-2
13	EHI_126310	Rho family GTPase	5	29.7	3.467	LC-ESI-MS/MS-1
14	EHI_133900	Galactose-inhibitable lectin 170-kDa subunit, putative	143	58.4	2.399	LC-ESI-MS/MS-1
15	EHI_136160	Calreticulin, putative	5	20.6	2.655	LC-MALDI-TOF/TOF-1
15	EHI_136160	Calreticulin, putative	7	38.1	3.221	LC-MALDI-TOF/TOF-1
15	EHI_136160	Calreticulin, putative	6	24.9	3.767	LC-MALDI-TOF/TOF-2
15	EHI_136160	Calreticulin, putative	12	32.4	4.325	LC-MALDI-TOF/TOF-2
16	EHI_136440	Dipeptidyl peptidase, putative	14	26.5	4.018	LC-ESI-MS/MS-1
17	EHI_149860	Hypothetical protein T24C12.3, putative	11	21.6	2.089	LC-ESI-MS/MS-1
18	EHI_161040	Hypothetical protein	16	60.6	2.355	LC-ESI-MS/MS-1
19	EHI_182720	Dipeptidyl peptidase, putative	25	40.4	3.192	LC-ESI-MS/MS-1

Gal/GalNAc lectin = galactose/N-acetyl galactosamine inhibitable lectin; iTRAQ = Isobaric tags for relative and absolute quantitation; LC-ESI-MS/MS = liquid chromatography electrospray ionization tandem mass spectrometry.

^*Note: Detection in replicate 1 or 2 is represented by “1” or “2” after the name of MS; Peptides (95%) indicates the number of distinct peptides having at least 95% confidence; % Cov (coverage) indicates the percentage of matching amino acids from identified peptides having confidence greater than 0 divided by the total number of amino acids in the sequence.

Figure 4.

Radar plot of increased abundance (A) and decreased abundance (B) membrane proteins in virulent versus avirulent variants.

Functional analysis of the membrane proteins with increased abundance.

The proteins with increased abundance in vEh variants were involved in molecular functions such as catalytic activity (eight proteins), binding (three proteins), signal transducer activity (one protein), and transporter activity (one protein) (Figure 5A). These proteins were involved in biological processes such as metabolic process (nine proteins), cellular process (three proteins), cellular component organization or biogenesis (one protein), biological regulation (one protein), and response to stimulus (one protein) (Figure 5B). These proteins were classified into six protein classes, which included hydrolase (six proteins), enzyme modulator (one protein), transporter (one protein), calcium-binding protein (one protein), transferase (one protein), and oxidoreductase (one protein), as shown in Figure 5C.

Figure 5.

Pie chart representation of the Protein ANalysis THrough Evolutionary Relationships functional classification of increased abundance membrane proteins in virulent versus avirulent variants according to their molecular function (A), biological process (B), and protein class (C) involvement.

Virulent versus avEh membrane proteins with decreased abundance.

A total of 18 proteins with decreased abundance were identified in vEh variants when compared with the avEh variant by using both the mass spectrometers (Table 2). The top five decreased membrane proteins that exhibited the highest fold change were thioredoxin, putative (4.8-fold); hypothetical protein, conserved (4.8-fold); hypothetical protein (3.9-fold); long-chain fatty-acid CoA ligase, putative (3.6-fold); and hypothetical protein (3.4-fold). These proteins were represented in the radar plot as shown in Figure 4B.

Table 2.

List of identified decreased abundance membrane proteins with fold change ≥ 2 in virulent versus avirulent variants

No	Accession number	Protein name	Peptides (95%)	% Cov	iTRAQ ratio (117:114)	Detected by*
1	EHI_001100	Hypothetical protein	11	29.4	0.488	LC-ESI-MS/MS-2
2	EHI_017000	Hypothetical protein	26	36.8	0.483	LC-ESI-MS/MS-2
3	EHI_021560	Thioredoxin, putative	37	47.8	0.207	LC-ESI-MS/MS-2
3	EHI_021560	Thioredoxin, putative	31	59.6	0.483	LC-ESI-MS/MS-1
4	EHI_022130	Hypothetical protein, conserved	12	17.0	0.387	LC-ESI-MS/MS-1
5	EHI_025090	Long-chain fatty-acid CoA ligase, putative	22	28.5	0.275	LC-ESI-MS/MS-1
5	EHI_025090	Long-chain fatty-acid CoA ligase, putative	26	28.8	0.308	LC-ESI-MS/MS-2
6	EHI_057700	Hypothetical protein, conserved	8	10.0	0.209	LC-ESI-MS/MS-2
7	EHI_077500	Galactose-specific adhesin 170-kD subunit	9	12.8	0.406	LC-MALDI-TOF/TOF-1
8	EHI_093970	Hypothetical protein	42	53.5	0.483	LC-ESI-MS/MS-1
9	EHI_111990	CXXC-rich protein	19	17.2	0.457	LC-ESI-MS/MS-1
10	EHI_114140	Hypothetical protein	4	39.0	0.319	LC-ESI-MS/MS-1
11	EHI_117690	Hypothetical protein	14	65.9	0.413	LC-ESI-MS/MS-2
12	EHI_128100	Peptidyl-prolyl cis-trans isomerase, putative	12	55.2	0.461	LC-ESI-MS/MS-2
13	EHI_138080	Hypothetical protein	12	52.2	0.294	LC-ESI-MS/MS-2
14	EHI_148990	Heat shock protein 70, putative	15	23.3	0.409	LC-ESI-MS/MS-2
15	EHI_160980	Hypothetical protein	15	33.9	0.258	LC-ESI-MS/MS-1
16	EHI_163540	Estradiol 17-beta-dehydrogenase, putative	11	21.9	0.413	LC-ESI-MS/MS-2
17	EHI_182530	Hypothetical protein	5	46.1	0.353	LC-ESI-MS/MS-2
18	EHI_199590	70-kDa heat shock protein, putative	10	26.9	0.366	LC-MALDI-TOF/TOF-2
18	EHI_199590	70-kDa heat shock protein, putative	130	61.7	0.466	LC-ESI-MS/MS-2

iTRAQ = isobaric tags for relative and absolute quantitation; LC-ESI-MS/MS = liquid chromatography electrospray ionization tandem mass spectrometry.

^*Note: Detection in replicate 1 or 2 is represented by “1” or “2” after the name of MS; Peptides (95%) indicates the number of distinct peptides having at least 95% confidence; % Cov (Coverage) indicates the percentage of matching amino acids from identified peptides having confidence greater than 0 divided by the total number of amino acids in the sequence.

Functional analysis of the membrane proteins with decreased abundance.

The membrane proteins with decreased abundance in vEh variants are involved in substantial molecular functions such as catalytic activity (three proteins), binding (one protein), signal transducer activity (one protein), and transporter activity (one protein), as shown in Figure 6A. These proteins are also involved in biological processes including metabolic process (three proteins), cellular process (two proteins), localization (two proteins), response to stimulus (two proteins), biological regulation (two proteins), and cellular component organization or biogenesis (one protein) (Figure 6B). These proteins are assigned into five protein classes, which comprise the transfer/carrier protein (one protein), oxidoreductase (one proteins), hydrolase (one protein), ligase (one protein), and membrane traffic protein (one protein), as shown in Figure 6C.

Figure 6.

Pie chart representation of the Protein ANalysis THrough Evolutionary Relationships functional classification of decreased abundance membrane proteins in virulent versus avirulent variants according to their molecular function (A), biological process (B), and protein class (C) involvement.

Protein–protein interaction prediction among differentially abundant membrane proteins.

Membrane proteins with increased abundance that are involved in the catalytic activity and metabolic process are shown in Figure 7A and B, respectively, whereas membrane proteins with decreased abundance that are involved in the catalytic activity and metabolic process are shown in Figure 7C and D, respectively. The details of the STRING analyses are tabulated in Supplemental Information.

Figure 7.

Interaction network of differentially abundant membrane proteins generated with STRING. Nodes highlighted with red rings represent increased abundance membrane proteins involved in (A) molecular function and (B) biological process and decreased abundance membrane proteins involved in (C) molecular function and (D) biological process according to gene ontology analysis via Protein ANalysis THrough Evolutionary Relationships. This figure appears in color at www.ajtmh.org.

Discussion

In this study, the applications of two kinds of mass spectrometers enhanced the proteome coverage and generated complementary results. The analyzer systems in both mass spectrometers were configured to function with their own acquisition modes and, hence, maximize the number of precursors interrogated. Therefore, unique proteins can be obtained when different mass spectrometer systems were combined. This was proven when proteins were uniquely detected by either LC-MALDI-TOF/TOF or LC-ESI-MS/MS (Figure 2). For example, calreticulin (EHI_136160) and Ras family GTPase (EHI_058090) were unique proteins detected solely by LC-MALDI-TOF/TOF and LC-ESI-MS/MS, respectively. A previous study by Bodnar et al. using shotgun analysis also showed that the combination of different ionization techniques improved the proteome coverage and maximized the number of peptide fragmentation spectra.¹⁹

The pathogenic strain of E. histolytica HM1:IMSS is known as a causative agent of amebiasis, where they express their pathogenicity by invasion and ability to lyse the epithelial tissues of the host. But the reasons or conditions responsible for human pathogenicity have remained enigmatic. Several proteomic comparison studies have been performed to compare the proteomes among E. histolytica strains and Entamoeba species.^20,21 This is because the genetic variations exhibited by different species, strains, and variants could influence the gene and protein expression.²² To narrow down the minute variation between variants in the pathogenic strains, this study provided an indepth understanding on the comparison of proteome profiles between vEh and avEh variants of a single strain.

The putative roles in virulence of the membrane proteins with increased abundance in the vEh variant.

Protein ANalysis THrough Evolutionary Relationships DB analysis demonstrated that most of the membrane proteins with increased abundance were involved in the catalytic activity (molecular functions) and metabolic process (biological process), as highlighted in Figure 7. The proteins include Rho family GTPase (EHI_126310), NAD(P) transhydrogenase subunit alpha (EHI_014030), multidrug resistance (MDR) protein (EHI_100320), calreticulin (EHI_136160), beta-N-acetylhexosaminidase, beta subunit (EHI_007330), serine carboxypeptidase (EHI_068090), dipeptidyl peptidase (EHI_136440, EHI_182720), and uncharacterized protein (EHI_149860). These membrane proteins with increased abundance may play specific roles in contributing to E. histolytica virulence.

Rho family GTPase is a member of G protein and is responsible for the regulation of actin organization, cytoskeleton dynamics, cell cycle, and gene expression of E. histolytica.²³ The overexpressed Rho family GTPase genes have been reported to play an important role in the adaptation of trophozoites to the host’s intestinal environment by altering their cytoskeleton-mediated processes.²⁴ Therefore, we postulate that the increased abundance of this protein in the vEh variant may play a role in sustaining the adaptation proficiency of trophozoites in the host, hence increasing the chance of successful colonization and invasion. This scenario also concurred with a previous study which indicated that the adaptation of E. histolytica to the host intestinal environment was attended by the increased subset of the cell signaling genes including the TM kinases, Ras and Rho family GTPases, and calcium-binding proteins.²⁵

NAD(P) transhydrogenase subunit alpha, putative is responsible for molecular functions including binding and catalytic activity, and biological processes including cellular process, localization, and metabolic process. This integral membrane protein is involved in the proton transport of hydride ions reversibly between nicotinamide adenine dinucleotide and nicotinamide adenine dinucleotide phosphate across the membrane, for the synthesis of the nicotinamide adenine dinucleotide phosphate hydrogen, glutathione reduction, and regulation of flux through the tricarboxylic acid cycle.²⁶ This protein (alternately known as pyridine nucleotide transhydrogenase) was previously detected in the phagosomes of E. histolytica during phagocytosis.²⁷ Apart from that, an increased expression of the gene-encoding protein NAD(P) transhydrogenase subunit alpha, putative was also observed in the E. histolytica Rahman strain after contact with human mucus compared with Rahman in the in vitro culture.²⁵ Hence, we hypothesize that the increased abundance of NAD(P) transhydrogenase in the vEh variant may indicate a higher potential of phagocytosis than that by the avEh variant.

Multidrug resistance protein is a stress response protein that is involved in TM transport and responsible for the ATP binding and ATPase activity. The transcriptomic regulation of the phenotype of MDR in E. histolytica has been proven to be regulated by the P-glycoprotein (Pgp) (EhPgp1 and EhPgp5) genes and these genes have been shown to participate in the MDR in E. histolytica, leading to an ATP-dependent efflux.^28–31 This is because the MDR is mediated by the overexpression of genes encoding Pgps, whereby Pgp functions as a multidrug efflux pump which is capable to discharge drugs out of the cell.^32–34 In this case, we hypothesize that the increased abundance of MDR protein in the vEh variant of E. histolytica may indicate that it could induce an efflux response even without the presence of drugs.

Calreticulin is a calcium-binding protein that is located mostly in the ER,³⁵ which has a role in binding and metabolic processes. The increased abundance of calreticulin observed in this study was also seen at the gene expression level, whereby the gene-encoding E. histolytica calreticulin demonstrated an overexpression in the vEh trophozoites during the development of experimentally induced amebic liver abscess in hamsters.³⁶ It was reported to engage in the initial stage of the host–parasite relationship as this protein triggered an immunogenic response in the human host, mainly in amebic liver abscess patients.³⁷ In addition, ameba developed an evasion mechanism against the host immune responses, by expressing the ER-associated calreticulin, which inhibited the classical serum complement pathway.³⁸ This showed that calreticulin may play a significant role in the host–parasite relationship during the early infection phases.³⁸ Therefore, we postulate that the increased abundance of calreticulin in the vEh variant of E. histolytica may indicate its involvement in the regulatory mechanism in relation to pathogenesis between the host and the parasite to ensure its successful survival in the host.

Beta-N-acetylhexosaminidase, beta subunit is a glycosidase which is involved in hydrolysis and mainly takes part in glycolysis. It has been postulated by researchers that this protein might participate with supplementary components and induce the carbohydrate meshwork damaging of intestinal epithelial lining, facilitating the parasite invasion.³⁹ The association of glycosidase and mucus depletion that caused intestinal disruption was also reported, where gene-encoded beta-N-acetylhexosaminidase and beta-galactosidase in the vEh strain were found to be upregulated during the colon invasion.²⁵ This consequence is also supported by previous findings where glycosidase activity diminished the intestinal mucin in E. histolytica.⁴⁰ Therefore, the increased abundance of beta-N-acetylhexosaminidase protein, shown in the present study, may support the invasive behavior of the vEh variant during the intestinal invasion.

Serine carboxypeptidase is classified under the S28 serine peptidase family. This protein was also found in the uropods proteome of E. histolytica.¹⁴ The potential biological function and mode of action of this particular protein in relation to virulence have not been reported and specifically characterized. However, the serine peptidase from Entamoeba invadens has been postulated to play a significant role in the encystation process.⁴¹ Fortunately, Makioka et al. in 2009 demonstrated the involvement of serine protease in the E. invadens excystation and metacystic development using inhibitors specific for serine proteases.⁴² The protease activity of serine proteases from the S28, S9, and S26 families was inhibited in the cyst lysates by phenylmethanesulfonyl fluoride. Their findings indicated that the participation of Entamoeba serine proteases in the excystation and metacystic development was related to pathogenicity.

Dipeptidyl peptidase is from the lipase family which is categorized under the clans of serine peptidase. Dipeptidyl peptidase was previously detected in the isolated uropod extruded fractions when E. histolytica trophozoites were exposed to the serum of amebiasis patients.⁴³ This enzyme is involved in lipid metabolism such as hydrolysis of triglycerides, phospholipids, and cholesterol esters.⁴⁴ At the gene-expression level, the gene encoding E. histolytica dipeptidyl peptidase demonstrated the highest level of expression among all serine peptidases,⁴⁵ although the actual role of this enzyme in contributing to the E. histolytica pathogenicity still remains unclear. On the other hand, the involvement of dipeptidyl peptidase IV of Giardia in encystation during Giardia differentiation in an in vitro model was confirmed.⁴⁶ The increased abundance of dipeptidyl peptidase in the vEh variant of E. histolytica may not directly contribute to the virulence factor, but the discovery of this protein in the uropod of E. histolytica provides an enlightening information as several components of the uropod such as Gal/GalNAc lectin may trigger immune responses against E. histolytica from the hosts.⁴³ Therefore, further investigation of the function of this enzyme will be needed.

Hypothetical protein (EHI_149860) was predicted as a glycosylase that is involved in the hydrolase activity which acts on the glycosyl bonds. This protein was also involved in several metabolic pathways including the pyrimidine metabolism and glycosaminoglycan degradation based on information available in the AmoebaDB database. Based on Simple Modular Architecture Research Tool (SMART) sequence analysis, the intrinsic feature of this protein consists of a TM region between 7 and 29 amino acid residues, two low complexity regions between 2 and 27 amino acid residues, and 93–207 amino acid residues. This protein was also found in the purified uropod fractions from trophozoites.⁴³ However, we are unable to postulate the potential role of this protein in relation to virulence and pathogenicity because of the limited information. Hence, further investigations are needed to identify its potential role as E. histolytica virulence factor.

The putative roles in virulence of the membrane proteins with decreased abundance.

On the other hand, the membrane proteins with decreased abundance were also mainly involved in the catalytic activity (molecular functions) and metabolic process (biological process) including the 70-kDa heat shock protein (HSP) (EHI_199590), long-chain fatty-acid CoA ligase (EHI_025090), B-keto acyl reductase (EHI_163540), CXXC-rich protein (EHI_111990), and uncharacterized protein (EHI_138080).

The decreased abundance of 70-kDa heat shock protein in the membrane of the vEh variant was interesting as this may provide an insight into the nature of this protein in the vEh strain (HM-1:IMSS). Heat shock protein is a stress response protein toward heat; when there is a sudden increase in temperature, HSP is synthesized and activated to protect the ameba.⁴⁷ In this study, the decreased abundance of 70-kDa HSP in the membrane of the vEh variant probably showed that the trophozoites were under a stress-free condition in the culture. The decreased abundance of 70-kDa HSP was also observed in glucose-starved trophozoites,⁴⁸ which may indicate that only certain stresses activate the HSP to induce a protective response. Under hyperoxia conditions, high transcriptional level of HSP genes in the vEh strain (HM-1:IMSS) was observed when compared with the non-vEh strain, and this was correlated with a high protein level of HSP70.¹⁷ The discrepancy in our results compared with the previous study by Santos et al. could also be because of the fact that they analyzed the total lysates of vEh and non-vEh E. histolytica HM-1:IMSS strains under hyperoxia conditions, whereas we only analyzed the membrane fractions of cultured amebae.¹⁷ The HSP70 is necessary for amebic virulence such as in the formation of amebic liver abscess in hamsters.¹⁷ Furthermore, the presence of these HSPs in the membrane fraction can be linked with the involvement of HSP70 in facilitating chaperones, intracellular protein transportation, and secretion functions.⁴⁹

Long-chain fatty-acid CoA ligase is an enzyme that is responsible mainly for the ligase activity and is involved in the lipid metabolic processes. Under a glucose-starved condition, the trophozoites of E. histolytica showed a remarkably reduced abundance in the proteins associated with metabolism, particularly in the fatty acid pathway.⁴⁸ In a normal E. histolytica pyruvate-to-ethanol pathway, acetyl-CoA is converted to acetaldehyde and then reduced to ethanol.⁵⁰ But the decreased level of the long-chain fatty-acid CoA ligase under the glucose-starved condition may be significant, whereby acetyl-CoA was shifted from the fatty acid pathway into the pyruvate-to-ethanol pathway,⁵¹ to deliver an alternative energy source for sustaining the survival of trophozoites. By referring to the reported statements,^50,51 we speculate that the decreased abundance of this protein in the vEh variant may have been associated with the E. histolytica virulence by altering the fatty acid pathway to increase their susceptibility toward stresses during glucose starvation.

Beta-ketoacyl reductase, also alternately known as estradiol 17-beta-dehydrogenase, is an oxidoreductase that is responsible for the oxidoreductase activity and is involved in the lipid metabolic process. This protein was previously found in the phagosome proteome⁵² and uropod fractions⁴³ of E. histolytica. Beta-ketoacyl reductase is one of the proteins involved in the elongation of long-chain fatty acids, and catalyzes the reduction of β-ketoacyl-CoA to L-β-hydroxyacyl CoA.⁵³ In this present study, the decreased abundance of this protein in the vEh variant of E. histolytica may be linked to the downregulation of long-chain fatty-acid CoA ligase as both are involved in the lipid metabolism. The correlation of the function of this protein with virulence is unclear. However, we postulate the decreased abundance of the beta-ketoacyl reductase could attempt to reduce the lipid metabolic processes to minimize energy usage.

CXXC-rich protein has a signal peptide and seven furin-like cysteine-rich regions that are found in a wide range of proteins and is involved in signal transduction via receptors tyrosine kinase.⁵⁴ The association between Gal-GalNAc lectin complex and Glycosylphosphatidylinositol-surface protein, which belongs to the CXXC family of surface proteins, was reported by Cheng et al.⁵⁵ Some CXXC motifs of the variant surface glycoproteins might implicate them in protein–protein interactions as they have signal peptide for cellular signaling. The decreased abundance of CXXC-rich protein in the vEh variant could be due to the low signal transduction among trophozoites and host receptor, particularly during the host–parasite interaction.

Hypothetical protein (EHI_138080) is one of the integral components of the membrane that is responsible for the cell transport. Based on the SMART sequence analysis, this protein contains short signal peptides between 1 and 22 amino acid residues and a TM region between 170 and 192 amino acid residues. This protein was also found in the uropod fraction of E. histolytica.³⁹ According to the limited information, we postulate that the potential role of this protein in TM transport and signaling pathway is because of the presence of TM and signal peptide. However, further investigation is required to identify its actual role in E. histolytica.

In conclusion, this study demonstrated that the application of iTRAQ-based quantitative approach could efficiently identify the differentially abundant membrane proteins between the vEh and avEh variants of E. histolytica HM-1:IMSS. Furthermore, this study has provided new knowledge on the putative roles of differentially abundant membrane proteins of the pathogenic E. histolytica HM-1:IMSS. Further investigations on the differentially abundant membrane proteins will be needed to elucidate the specific functional roles including those of hypothetical proteins to understand their roles in disease pathogenesis.

Note: Supplemental information and figure appear at www.ajtmh.org.