Unraveling the relevance of the polyadenylation factor EhCFIm25 in Entamoeba histolytica through proteomic analysis

We recently reported that silencing of the polyadenylation factor EhCFIm25 in Entamoeba histolytica, the protozoan which causes human amoebiasis, affects trophozoite proliferation, death, and virulence, suggesting that EhCFIm25 may have potential as a new biochemical target. Here, we performed a shotgun proteomic analysis to identify modulated proteins that could explain this phenotype. Data are available via ProteomeXchange with identifier PXD027784. Our results revealed changes in the abundance of 75 proteins. Interestingly, STRING analysis, functional GO‐term annotations, KEGG analyses, and literature review showed that modulated proteins are mainly related to glycolysis and carbon metabolism, cytoskeleton dynamics, and parasite virulence, as well as gene expression and protein modifications. Further studies are needed to confirm the hypotheses emerging from this proteomic analysis, to thereby acquire a comprehensive view of the molecular mechanisms involved.

represents one of the major leading causes of death from parasitic diseases worldwide, mainly in developing countries [1]. Relevant DNA motifs in promoters and several transcription factors have been reported; capping has not been described yet, but most of the main components of splicing and polyadenylation machineries found in higher eukaryotic cells, have been identified [2]. Our group reported that E. histolytica presents the six subunits of the cleavage and polyadenylation specificity factor (CPSF160, CPSF100, CPSF73, CPSF30, Fip1, and WDR33), but only two of the three subunits of the cleavage stimulation factor (CstF77 and CstF64); it has both CLP1 and PCF11 subunits of the cleavage factor IIm (CFIIm), but only the small 25 kDa subunit of the CFIm; the poly(A) polymerase (PAP), the poly(A) binding protein (PABP), RBPP6, and PP1a were also found [3].
In humans, CFImI is a heterotetramer complex with two 25 kDa subunits bound to a dimer of larger subunits (59 or 68 kDa). Notably, CFIm25 is essential for polyadenylation factor recruitment, poly(A) site selection, pre-mRNA cleavage, and poly(A) tail synthesis [4][5][6][7]. The fact that the 25 kDa subunit is the only CFIm component identified in E. histolytica suggests that mRNA polyadenylation could involve a different mechanism in this parasite, and EhCFIm25 could have a central role in this event. Molecular characterization of EhCFIm25 confirmed that it is a Nudix protein although three of the four glutamate residues (E154, E157, and E158) in the conserved Nudix box are replaced by lysine, and the last glycine residue (G160) is substituted by the hydrophilic residue serine. As the human protein, EhCFIm25 interacts with EhPAP [8], but also with the transcriptional coactivator EhPC4 (J. D. Ospina-Villa, our unpublished data) related to virulence, DNA replication, and multinucleation in E. histolytica [9,10]. Although three-dimensional structure prediction indicates the absence of a classical RNAbinding domain, EhCFIm25 can bind the 3 0 untranslated region (3 0 UTR) of E. histolytica transcripts through the participation of the conserved Leu135 and Tyr236 residues [11,12]. By using the SELEX strategy, we isolated two RNA aptamers that contain the GUUG motif recognized by EhCFIm25. Importantly, EhCFIm25 sequester by aptamers, rapidly induced parasite death, which confirms that targeting the polyadenylation process, namely EhCFIm25, represents an effective strategy for controlling E. histolytica [13]. On the other hand, EhCFIm25 silencing produced a significant acceleration in parasite proliferation and cell death; moreover, cells were larger and multinucleated, and their ability to move and phagocyte erythrocytes was significantly reduced, indicating loss of virulence. Additionally, functional experiments showed that EhCFIm25 controls the selection of the distal poly(A) site in mRNA 3 0 UTR [14]. However, the relationships between EhCFIm25 inhibition and these phenotypical characteristics remain unknown. Here, we performed a shotgun proteomic analysis to evidence changes in protein expression that can explain the observed phenotype.

EhCFIm25 gene silencing
Double-stranded RNA (dsRNA)-based EhCFIm25 gene silencing was performed as described [14]. Briefly, E. histolytica trophozoites (5.0 9 10 4 ) were grown for 4 days in TYI-S-33 complete medium in the presence of EhCFIm25-dsRNA molecules (100 lgÁmL À1 ) obtained from E. coli HT115 (rnc14:DTn10) cells transformed with the pL4440-EhCFIm25 plasmid. Every day, a 10 lL aliquot of culture was taken to determine parasite number in a Neubauer chamber and living cells stained by Trypan blue. Experiments were performed twice in triplicate, and results were expressed as mean AE SD. Trophozoites growing in standard conditions (without dsRNA) and treated with gfp-dsRNA were used as controls. Data corresponding to the EhCFIm25-dsRNA condition were compared with both control conditions using the two-way ANOVA test. P < 0.05 was considered as statistically significant.

RNA and protein isolation
Total proteins and RNA of E. histolytica trophozoites treated with EhCFIm25-dsRNA and control cells (day 4) were obtained by the TRIzol reagent (Invitrogen) according to manufacturer's instructions. Protein quality and amount were verified by 10% SDS/PAGE and the Bradford method before being used in Mass spectrometry assays. RNA integrity and concentration were assessed by agarose gel electrophoresis and spectrophotometric analysis in a spectrophotometer (NanoReady Touch, Hangzhou Lifereal Biotechnology). RNA was used in Real-Time qRT-PCR to assess inhibition of EhCFIm25 mRNA expression and validate proteomic data as described below.

Mass spectrometry analysis LC-ESI-HDMSE
The volume corresponding to 150 lg total proteins obtained from the three culture conditions was delivered to the Laboratorio Nacional de Servicios Experimentales (LaNSE), CINVESTAV (Mexico), for protein identification and absolute quantitation by mass spectrometry analysis LC-ESI-HDMSE as described [16,17]. Raw files containing MS and MS/MS spectra were deconvoluted, compared, and quantified by PROTEINLYNX GLOBAL SERVER (PLGS) v3.0.3 software against E. histolytica (Strain: ATCC 30459/ HM-1:IMSS, downloaded from UniProt, 7959 protein sequences, 3 February 2021) concatenated with reverse database. Workflow parameters were included as described [17]. [All protein identifications reported had a percentage of ≥ 95% reliability (protein autocurate green)]. Full data are included in Table S1.

Data analysis and bioinformatics
Protein amounts (fmol) in trophozoites treated with EhCFIm25-dsRNA were compared with trophozoites treated with gfp-dsRNA and nontreated control cells; all data were expressed as a base 2 logarithm [18]. Protein abundance changes were used to construct a hierarchical clustering (Euclidean distance) by the heatmapper program (http:// www.heatmapper.ca). Proteins displaying at least AE 1.0 absolute fold change in both comparisons, EhCFIm25-dsRNA vs. gfp-dsRNA and EhCFIm25-dsRNA vs. nontreated cells, were considered as differentially expressed and selected for further analyses; proteins that were detected in control cells but not in the EhCFIm25-dsRNA condition were also considered in the study. Selected proteins were used Gene Ontology (GO) enrichment analysis. The GO terms and biochemical pathways of proteins were obtained using DAVID v6.8 (https://david.ncifcrf.gov/) and corroborated in AmoebaDB database (https://amoebadb.org/ amoeba/app/), UniProt database (https://www.uniprot.org/), and KEGG PATHWAY database (https://www.genome.jp/ kegg/pathway.html). All data were visualized using the CytoScape tools software (https://cytoscape.org/). Additionally, modulated proteins were submitted to STRING analysis using the MCL clustering tool (https://string-db.org/).

Real-Time qRT-PCR (Real-Time Quantitative Reverse Transcription-PCR)
Total RNA was used to synthesize cDNA using the Super-Script III Reverse Transcriptase (Invitrogen) according to manufacturer's instructions in a GeneQ Thermal Cycler (BIOER, Hangzhou, China). The qPCR assay was completed using SensiFAST TM SYBR Hi-ROX (Bioline) with specific primers for selected genes as follows: EhCFIm25: sense 5 0 -TGGAGAAGATGATCCTGTTGAAG-3 0 and antisense 5 0 -TCTTTGACTTGACTTACATGAACTG-3 0 primers [14]; pyruvate phosphate dikinase ( [19]. All reactions were performed in a StepOne real-time PCR system (Applied Biosystems) with the following cycling conditions: enzyme activation at 95°C for 2 min, followed by 40 cycles of denaturation at 95°C for 5 s and annealing/extension at 62°C for 30 s. The experiment was carried out by triplicated, and the relative expression of mRNA was determined by the 2 ÀDDCt method. Data corresponding to the EhCFIm25-dsRNA condition were analyzed by Prism GraphPad software using paired Student's t-test to compare the EhCFIm25-dsRNA condition with both control conditions. P < 0.05 was considered as statistically significant.

Silencing of EhCFIm25 by dsRNA
To evidence how the absence of EhCFIm25 affects protein expression, we first induced downregulation of EhCFIm25 gene expression by soaking trophozoites with purified EhCFIm25-dsRNA as described [14]. As expected, real-time qRT-PCR assays confirmed that on day 4, EhCFIm25 gene transcription was reduced by about 75% in parasite cultures containing EhCFIm25-dsRNA, when compared with gfp-dsRNA-treated cells and nontreated cells. Additionally, trophozoites exhibited an accelerated proliferation from day 2 when compared with both control cells, as well as a higher number of dead cells (Fig. S1). All these observations agreed with our previous report [14] and strongly supported the downregulation of EhCFIm25 expression following EhCFIm25-dsRNA soaking.
Identification of modulated proteins following EhCFIm25 silencing Protein extracts were obtained from EhCFIm25-dsRNA-treated trophozoites and both control cultures (gfp-dsRNA condition and nontreated cells) and analyzed by mass spectrometry analysis LC-ESI-HDMSE as described [16][17][18]. A total of 139 proteins (including 11 uncharacterized proteins) was identified and quantified, considering 99.9% accuracy and detection of at least two peptides for protein identification (Table S1). Hierarchical clustering evidenced that many proteins showed similar changes in abundance when comparing EhCFIm25-dsRNA-treated cells vs. gfp-dsRNA condition and nontreated cells, indicating that EhCFIm25 silencing had an effect on the gene expression ( Fig. 1). Because of the relevance of EhCFIm25 in polyadenylation and the role of poly(A) tail in translation and mRNA stability, the absence of EhCFIm25 probably affected mRNA turnover and translation; however, we cannot discard that the turnover of some proteins could also be changed.
Of these 139 proteins, 75 showed a significant log2 fold change value when comparing EhCFIm25-dsRNA-treated cells vs. both control conditions. These included 48 less abundant proteins (comprising four uncharacterized proteins) and 12 more abundant proteins (with two uncharacterized proteins); notably, 15 proteins (including four uncharacterized proteins) were not detected in the absence of EhCFIm25 ( Fig. 2A). The four most reduced proteins were the PPi-type phosphoenolpyruvate carboxykinase 3 (C4LWQ8) (PPi-PEPCK3/PEPCK3), the pyrophosphatefructose 6-phosphate 1-phosphotransferase (C4LZC2) (PPi-PFK/PFK), and the 3-ketoacyl-CoA synthase 4 (B1N502) that are related to metabolic pathways, such as glycolysis, and lipid metabolism, respectively, as well as a leucine-rich (LRR) repeat containing protein (C4M553). On the other hand, three of the highly enriched proteins corresponded to the 60S ribosomal protein L14 (C4LTQ0), 60S ribosomal protein L27 (C4M727), and ubiquitin (C4LY85), indicating that EhCFIm25 silencing had an impact on protein synthesis. The fourth one is an uncharacterized protein (C4M429). To validate changes in protein abundance, we selected genes corresponding to three proteins with a reduced abundance, PFOR, PPDK, and PRX, and four proteins with an increased amount, ACD, ARP2/ 3, myosin heavy chain ubiquitin, and evaluated mRNA expression by real-time RT-qPCR in the three groups of trophozoites. In agreement with proteomics data, PFOR, PPDK, and PRX genes were downregulated in response to EhCFIm25 silencing in comparison with the untreated cells and the gfp-dsRNA condition, while ACD, ARP2/3, and myosin heavy chain ubiquitin genes were upregulated (Fig. 2B,C). We previously showed that EhCFIm25 interacts with EhPAP and EhPC4, two proteins involved in polyadenylation and transcription, respectively. Although these nuclear proteins did not appear in the proteomes, we investigate whether there were any changes in the mRNA expression of EhPAP and EhPC4 genes upon silencing the EhCFIm25. Results showed mRNA expression of both genes was increased in EhCFIm25-silenced trophozoites (Fig. S2).
With the aim of having a general view of the impact of change in protein abundance in EhCFIm25-silenced trophozoites, the identified proteins were categorized using functional GO-term annotations. The 10 uncharacterized proteins were not included since BLAST analyses failed to assign them a possible identity and function (data not shown). Many of the less abundant proteins corresponded to cellular and intracellular compartments; concerning biological process, they were mainly associated with metabolism, including metabolic process, organic substance metabolic process, and cellular metabolic process; finally, the main function was binding. Regarding proteins that were more abundant in the absence of EhCFIm25, they were mainly found as part of the ribosome and cytoplasm, and mostly related to Arp2/3 complex-mediated actin nucleation and translation; interestingly, the main molecular functions were related to binding, including RNA, ATP, and nucleotide binding, as well as ATPase activity. Lastly, proteins that were not detected in EhCFIm25-silenced trophozoites corresponded to cytosol and ribosome; they were mainly associated with translation and related to nucleotide and RNA binding, as well as part of the structural constituent of ribosome (Fig. 3).
KEGG analyses evidenced that four of the proteins with a reduced abundance, C4LXE8, C4M192, C4M230, and C4LVD3, have an impact on at least five biochemical pathways. They also confirmed that reduced proteins are mainly related to metabolism, namely carbon metabolism, metabolic pathways, glycolysis/gluconeogenesis, and pyruvate metabolism; they were also related to gene expression regulation by acting in spliceosome, protein processing and endoplasmic reticulum, ribosome, RNA transport, aminoacyl-t RNA biosynthesis, and biosynthesis of amino acids. Regarding proteins with an increased abundance, three proteins (C4M8B0, C4M2Y7, and C4LY85) were involved in two biochemical pathways. Regulation of actin cytoskeleton was the most affected pathway. On the other hand, three of the proteins that were not detected in EhCFIm25-silenced trophozoites (C4M4T9, B1N3F1, and C4M6Y2) participate in three biochemical pathways. Gene expression regulation, including RNA transport and mRNA surveillance, spliceosome, and ribosome, was the most affected process (Fig. 4).

Discussion
We previously reported that EhCFIm25 is essential for E. histolytica survival and virulence properties, since its silencing affected trophozoite proliferation, produced parasite death, altered cell size and nucleus number, and reduced mobility and erythrophagocytosis capacity of parasites, which prompted us to propose this polyadenylation factor as a new biochemical target in this human pathogen [14]. Our proteomic analysis revealed that the absence of EhCFIm25 produced changes in the abundance of 75 proteins. Among these, we focused on proteins related to glycolysis and carbon metabolism, cytoskeleton dynamics and parasite virulence, and gene expression and protein modifications.   enzymes were also affected, such as PPi-type phosphoenolpyruvate carboxykinase 3 (C4LWQ8) (PPi-PEPCK3), malic enzyme (Q9NH04) (ME), and malate dehydrogenase_ putative (B1N2Y0) (MDH) that control the PEP-oxaloacetate-malate-pyruvate cycle, the L-myo-inositol-1-phosphate synthase (C4M5S9) that catalyzes the conversion of glucose 6-phosphate to 1-l-myo-inositol-1-phosphate, and the anaerobic glycerol-3-phosphate dehydrogenase subunit A_ putative (B1N3H7) (G3PDH) that converts glycerol-3phosphate to dihydroxyacetone phosphate [21]. Interestingly, the increased abundance of the acetyl-CoA synthetase_ putative (C4LUV9) (ACD) and the reduced amount of bifunctional ADHE (C4M230), and C4LVY0 and C4M7F5, two putative alcohol dehydrogenases of 42 kDa and 46 kDa, respectively, that did not correspond to the reported ADH1 (39 kDa) and ADH3 (43 kDa), suggested that parasites direct the last steps of glycolysis to the conversion of acetyl-CoA into acetate with the production of ATP in an attempt to restore, at least partially, the lack of energy that results from downregulation of upstream glycolytic enzymes (Fig. 6). However, energy deficiency is likely the main cause of parasite death following EhCFIm25 silencing.
In agreement with this assumption, it has been reported that ADHE activity is an important control point for the glycolytic flux and its inhibition can decrease the parasite energy load and survival [22].  The cytoskeleton dynamics is essential for various cellular functions in E. histolytica, from cytokinesis to morphogenesis, including adhesion, endocytosis, phagocytosis, motility, and migration, that are pivotal activities in amoebic pathogenesis [23]. Although it is the key cytoskeletal protein, actin dynamics requires a coordinated cooperation with a large number of actinbinding proteins (ABPs) (including myosin II).
Recently, a bioinformatic approach identified a total of 390 ABPs, most of them corresponding to uncharacterized proteins [24]. Here, we found that several proteins linked to the actin-rich cytoskeleton were modulated in EhCFIm25-silenced trophozoites. Notably, eight ABPs (C4M126, C4M9I2, B1N322, C4M1S5, C4LW6, C4M295, C4M553, and C4LWF5) were downregulated in EhCFIm25-silenced trophozoites. C4M126 (F-BAR domain-containing protein) and C4M9I2 (SH3 domain protein) contain a BAR (bin-amphiphysin-Rvs) domain that has been involved in membrane curvature and actin binding. In E. histolytica, phagocytosis was inhibited by EhBAR (C4M128) silencing, indicating its participation in actin dynamics [25]. In addition, the SH3 domain is enriched in proteins that are upregulated in phagocytic trophozoites [26]. The atypical filamin (B1N322) [27], a-actinin 1 (C4M1S5) [28], and aactinin 2 (C4LWU6) [29] are members of the calponin family that are able to cross-link actin to stimulate an active net; filamins, particularly filamin A also known as ABP120, have been localized in pseudopods and uropods [30], while a-actinin 2 (EhActn2) was involved in phagocytic cup formation [31]. The adenylyl cyclaseassociated protein (C4M295) of the cyclase-associated protein (CAP) family is involved in actin remodeling [32] and nutritional response signaling [33]. The leucine-rich repeat containing protein (C4M553), an I/LWEQ domain-containing protein, has two leucinerich repeats that are thought to promote protein interactions with surface proteins [34] and a FERM domain that links cytoskeleton proteins to the plasma membrane [35]. EhLimA (C4LWF5) is a LIM domaincontaining protein that colocalized to the plasma membrane with actin filaments in E. histolytica [36]. These alterations in ABP amount can be related to the reduced abundance of a small Ras family GTPase (C4M8B0) and a Rab GDP dissociation inhibitor (B1N309) known to regulate the GDP-GTP exchange reaction of members of the Rab family, leading to a reduction in vesicular trafficking.
On the other hand, only two cytoskeletal-related proteins were found with an increased abundance: the heavy chain of myosin II (C4LU72) which binds Factin and is responsible for cell contraction and motility [37] and ARP2 (C4M5I6), one of the seven subunits of the ARP2/3 complex that mediates actin nucleation and has been localized in phagosomes [38]. Globally, changes observed in the abundance of these cytoskeletal proteins likely affect cytoskeleton organization and dynamics, which can be related to the augmented size of trophozoites and the increased number  of nuclei in EhCFIm25-silenced trophozoites, probably due to failure in cytokinesis; cytoskeleton alterations could also be responsible for the reduced mobility and erythrophagocytosis capacity in trophozoites, affecting their virulence capacity [14]. Congruently with these observations, our results evidenced the downregulation of other proteins previously related to phagocytosis, an essential process for parasite nutrition and pathogenesis. Members of the cysteine protease-binding protein family (CPBF), a single-transmembrane carrier/receptor family, interact with lysosomal hydrolytic enzymes and regulate their trafficking; notably, CPBF6 (C4LUR3) was identified in phagosomes, participating in the transport of cargo proteins, namely a-amylase and c-amylase [39]. EhP3 (C4M0F4), a homolog of 14-3-3 family of protein, participates in initiation/formation of phagocytic cups and formation of phagosome, by acting as an adaptor molecule to recruit proteins that allow dynamics of Factin rearrangement during phagocytosis [40]. Several heat-shock proteins (HSP) have also been involved in parasite virulence; notably, the activity of HSP70 is necessary for amoebic liver abscess formation in hamsters [41], while HSP90 controls the process of phagocytosis [42]. Then, the modulation of HSP90 (CELVB2) and HSP70 (C4M770, C4M3S5) could contribute to the reduced virulence capacity of EhCFIm25-silenced parasites. In addition to phagocytosis, the virulence process of E. histolytica includes adherence to target cells, cytolysis, and evasion of host immune response with the participation of various key proteins [43]. Therefore, the reduced amount of the galactose-inhibitable lectin 170 kDa subunit (B1N3C7) (Gal/GalNAc lectin) located on the surface of trophozoite contributes to reduce adherence to galactose and N-acetylgalactosamine of host cells. Moreover, we also found lower amount of peroxiredoxin (B1N5Y9) (PRX) and superoxide dismutase (SOD) (C4LX20), two enzymes involved in resistance of host oxidative defenses, that are important components of amoebic virulence [44]. The reduced amount of 3-ketoacyl-CoA synthase 4 or fatty acid elongase (B1N502) could also have an impact on parasite virulence properties that require an active lipid membrane dynamic [45].

EhCFIm25 silencing has an impact on gene expression
The last group of proteins with an altered amount in EhCFIm25-silenced trophozoites can be generally clustered around gene expression and protein modifications, which may explain the proteomic changes described above that promotes parasite death and virulence diminution. Most of these proteins showed a reduced abundance or were not detected, in agreement with the high number of downregulated proteins. They include several ribosomal proteins (C4LU56, B1N384, B1N3R9, B1N2Z3, C4M4I3, C4LTQ5, B1N442), aminoacyl-tRNA synthetases (C4LWW8, C4LXL0, C4LZN0), elongation factors (C4M7D4, B1N306), and an aminopeptidase (C4M0L4), a protein kinase domain-containing protein (B1N2E5) and a protein disulfide isomerase (C4MB38). Additionally, two proteins that are important for RNA metabolism were not detected: the ATP-dependent RNA helicase DDX39 (B1N3F1) whose homologue in humans participates in the regulation of transcription, splicing, and RNA export [46] and the polyadenylate-binding protein_ putative (C4M6Y2) that interacts with the PAP to enhance polyadenylation at the 3 0 -end of mRNA and bind the poly(A) tail to avoid transcript degradation and promotes translation through mRNA circularization and interaction with eIF4G [47]. Altogether, our results demonstrate that the absence of EhCFIm25 had an impact on gene expression and proteome, and as a result, on E. histolytica survival and virulence properties, however, the molecular mechanisms involved remain unclear. In our previous work, we showed that EhCFIm25 controls the efficient selection of distal (or downstream) poly(A) sites in E. histolytica transcripts and hypothesized that its silencing may have an impact on the 3 0 -end mRNA processing and therefore gene expression of the 20 identified genes with two polyadenylation sites [14,48]. Surprisingly, proteins that are modulated in our proteomics analysis did not correspond to any of these genes, which agrees with the limited impact of alternative polyadenylation on gene expression regulation in this parasite. In humans, CFIm25 is also essential for polyadenylation factors recruitment, pre-mRNA cleavage, and poly(A) tail synthesis [4][5][6][7]. Therefore, it is possible that changes in protein abundance following EhCFIm25 silencing may result more from alterations in these events that from an upstream shift in poly(A) site selection; future experiments required to be performed to confirm this assumption. On the other hand, it is important to point out that a proteome does not represent the full set of proteins expressed in a cell, and it is an approximation that depends on the identification of peptides and their amount; moreover, the shotgun analysis presented here has been performed in total protein extracts, which may limit the detection of less abundant proteins. Notably, several genes with two poly(A) sites correspond to nuclear proteins that participate in DNA condensation, DNA binding, translation, splicing, and mRNA binding; similarly, the nuclear EhCFIm25 protein was not identified in any of the three proteomes. The absence of nuclear proteins among the identified proteins is likely related to the lower amount of nuclear proteins in comparison with cytoplasmic proteins. It is possible that a proteomic analysis of nuclear proteins would give some insights about the abundance of these specific proteins. Additionally, the increased expression of EhPAP and EhPC4 that interact with EhCFIm25 suggests the existence of compensatory mechanisms in gene expression regulation that may help to overcome the absence of EhCFIm25 and reduce, at least partially, alterations in transcription and RNA processing, for parasite survival during the first days.
In conclusion, the present proteomic approach confirmed the significance of EhCFIm25 as a biochemical target in E. histolytica by providing interesting insights about the proteins and biological pathways that are altered following the absence of this polyadenylation factor. Notably, our results suggested that EhCFIm25 silencing affected energy metabolism and cytoskeleton dynamics causing parasite death and reducing parasite virulence, respectively, by altering gene expression, which explains parasite death, and alteration in virulence properties that were observed in the absence of EhCFIm25. The next step will involve additional studies to confirm the hypotheses emerging from our proteomic analysis and get a comprehensive view of the molecular mechanisms involved.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig S1. Effect of EhCFIm25 silencing on E. histolytica trophozoites proliferation and viability. Trophozoites (5.0 9 10 4 ) were soaked with EhCFIm25-dsRNA (100 lLÁmL À1 ) at 37°C. At day 4, total RNA was obtained to evaluate EhCFIm25 mRNA expression by Real-time quantitative reverse transcription polymerase chain reaction (Real-time qPCR). The EhRNAPII gene mRNA expression was determined and used as normalization control. Data corresponding to the EhC-FIm25-dsRNA condition were compared to both control conditions using the paired Student's t-test. **P < 0.01; ***P < 0.001 (A). Each day, cells were counted (B) and parasite viability was assessed by the Trypan blue assay (C). Data corresponding to the EhCFIm25-dsRNA condition were compared to both control conditions using the two-way ANOVA test. Federation of European Biochemical Societies *P < 0.05 and ***P < 0.001. n = 3. Error bars represent SD. Fig S2. Real-time quantitative reverse transcription polymerase chain reaction (Real-time qRT-PCR) for EhPAP and EhPC4 genes in EhCFIm25-dsRNA trophozoites compared with (A) non-treated cells and (B) gfp-dsRNA treated parasites. The EhRNAPII gene mRNA expression was determined and used as normalization control for all qRT-PCR assays. Data corresponding to the EhCFIm25-dsRNA condition were compared to both control conditions using the paired Student's t-test. *P < 0.05; ***P < 0.001. n = 3. Error bars represent SD.