mRNA Decay Proteins Are Targeted to poly(A)+ RNA and dsRNA-Containing Cytoplasmic Foci That Resemble P-Bodies in Entamoeba histolytica

In higher eukaryotes, mRNA degradation and RNA-based gene silencing occur in cytoplasmic foci referred to as processing bodies (P-bodies). In protozoan parasites, the presence of P-bodies and their putative role in mRNA decay have yet to be comprehensively addressed. Identification of P-bodies might provide information on how mRNA degradation machineries evolved in lower eukaryotes. Here, we used immunofluorescence and confocal microscopy assays to investigate the cellular localization of mRNA degradation proteins in the human intestinal parasite Entamoeba histolytica and found evidence of the existence of P-bodies. Two mRNA decay factors, namely the EhXRN2 exoribonuclease and the EhDCP2 decapping enzyme, were localized in cytoplasmic foci in a pattern resembling P-body organization. Given that amoebic foci appear to be smaller and less rounded than those described in higher eukaryotes, we have named them “P-body-like structures”. These foci contain additional mRNA degradation factors, including the EhCAF1 deadenylase and the EhAGO2-2 protein involved in RNA interference. Biochemical analysis revealed that EhCAF1 co-immunoprecipitated with EhXRN2 but not with EhDCP2 or EhAGO2-2, thus linking deadenylation to 5′-to-3′ mRNA decay. The number of EhCAF1-containing foci significantly decreased after inhibition of transcription and translation with actinomycin D and cycloheximide, respectively. Furthermore, results of RNA-FISH assays showed that (i) EhCAF1 colocalized with poly(A)+ RNA and (ii) during silencing of the Ehpc4 gene by RNA interference, EhAGO2-2 colocalized with small interfering RNAs in cytoplasmic foci. Our observation of decapping, deadenylation and RNA interference proteins within P-body-like foci suggests that these structures have been conserved after originating in the early evolution of eukaryotic lineages. To the best of our knowledge, this is the first study to report on the localization of mRNA decay proteins within P-body-like structures in E. histolytica. Our findings should open up opportunities for deciphering the mechanisms of mRNA degradation and RNA-based gene silencing in this deep-branching eukaryote.


Introduction
The control of messenger RNA (mRNA) degradation plays a key role in the posttranscriptional regulation of gene expression [1].In eukaryotic cells, mRNA degradation is initiated by shortening of the 39 end-poly(A) + tail by the CCR4/CAF1/ NOT1-5 complex and the PAN2, PAN3 and PARN deadenylases [2].Next, the DCP1 and DCP2 decapping enzymes remove the cap from the 59 end of the deadenylated mRNA in a process stimulated by Dhh1, Edc3, Pat1/Mrt1, Sm-like (Lsm1p-7p) and Hedls proteins [3].Lastly, the uncapped mRNA body is degraded by XRN1 and XRN2 exoribonucleases in the 59-to-39 direction.
Alternatively, mRNA is degraded in the 39-to-59 direction by a complex of PH-like ribonucleases known as the exosome [4].
There is growing body of evidence to show that mRNA degradation occurs in specialized cytoplasmic foci variously referred to as mRNP granules, mRNA-decay foci, GW182-bodies and mRNA processing bodies (P-bodies).These structures are enriched in RNA substrates and mRNA turnover proteins.Pbodies are mRNA processing centers within which non-translating transcripts are sorted and either silenced or degraded [5].Although the protein inventory of P-bodies has not been defined in detail, around 25 different factors have been detected within these small cytoplasmic domains [6].The P-bodies observed in yeast, insect, nematode and mammalian cells have critical roles in mRNA degradation, mRNA storage, mRNA surveillance and RNA-based gene silencing mechanisms [7].In fact, it has been demonstrated that P-bodies have high concentration of target transcripts and the AGO2, DICER, and GW182 proteins involved in RNA interference by microRNAs (miRNAs) and small interfering RNAs (siRNAs) [8,9,10].In addition, P-bodies also contain the UPF and SMG families of proteins involved in the nonsense mediated decay (NMD) mechanism that degrades aberrant mRNAs containing premature translation termination codons.[11].In mammals, most cells contain from three to nine Pbodies, although this number appears to depend on the cell's proliferation ratio and stress conditions, such as nutrient deprivation, oxidative stress and heat shock.Cell treatments with inhibitors of transcription or translation (such as actinomycin D and cycloheximide, respectively) lead to the disassembly of Pbodies [12].In contrast, treatment with puromycin increases the number and size of P-bodies; this observation emphasizes the relevance of active mRNA and protein synthesis in the assembly of these structures [10].Even though P-body components clearly play crucial roles in mRNA degradation, there is also evidence to show that NMD and silencing mediated by miRNAs and siRNAs occur in cells lacking detectable P-bodies.This suggests that Pbody formation is a consequence, rather than a cause, of mRNA degradation activity [10,13].Furthermore, translational repression by miRNAs does not require P-body structures; the localization of miRNAs and RNA-induced silencing complex (RISC) proteins within P-bodies appears to be a consequence of translational repression [14].
Human amebiasis is caused by infection with the intestinal protozoan parasite Entamoeba histolytica.The disease is endemic in many developing countries and affects around 50 million people worldwide [15].The symptoms of amebiasis include colitis, dysentery and extra-intestinal abscesses.Despite the importance of post-transcriptional events in the regulation of gene expression, the parasite's mRNA degradation and RNA-based gene silencing mechanisms are poorly understood.Some proteins involved in the RNA interference pathway, including EhAGO2-2, EhRNAseIII and EhRdRP, have been identified and characterized in E. histolytica [16,17].However, proteins involved in mRNA decay pathways have not previously been studied and the potential role of P-bodies in mRNA decay and RNA interference has not been clearly established in protozoan parasites.Here, we sought to localize mRNA degradation factors within E. histolytica and found evidence of the existence of P-body-like structures.Our findings open up opportunities for deciphering the mechanisms of mRNA decay and RNA gene silencing in this parasitic, deep-branching eukaryote.

mRNA degradation machineries in Entamoeba histolytica
In order to identify homologous genes potentially involved in mRNA decay in E. histolytica, we screened the parasite genome at AmoebaDB (http://amoebadb.org/amoeba/)by using the amino acid sequences of yeast and human proteins involved in mRNA turnover as probes.Heuristic searches identified several putative protein-encoding genes for mRNA degradation machineries, including deadenylation, decapping, NMD, and RNA interference activities (Figure 1, Table S1).Our data indicated that although several important genes are lacking, mRNA decay machineries are generally well conserved in E. histolytica.We detected putative Ehdcp2, Ehlsm1-6, Ehedc3 and Ehdhh1 genes (coding for mRNA decapping factors) and an Ehxrn2 gene (coding for a putative 59-to-39 exoribonuclease).In contrast, we did not find any dcp1 or xrn1 homologs.We also identified genes for NOT1-5 subunits and for EhCAF1 and EhCAF1-like deadenylases.In contrast, the lack of a ccr4 deadenylase homolog suggested the existence of a CAF1/ NOT complex lacking one of the major deadenylases.Intriguingly, we did not detect pan2, pan3 and parn genes; this indicates that E. histolytica's deadenylation machinery is smaller than those described in yeast and in the human.Amoeba genome also contains genes for the NMD machinery, as represented by three Ehupf genes (Table S1).RNA interference pathway members, including EhRdRP, EhRNAseIII and EhAGO2-2 proteins, have already been reported in this parasite [16,17].We (like others) did not detect DICER and GW182 homologs, suggesting that RNA interference may occur by small interfering RNAs generated by DICER-independent mechanisms in E. histolytica [18].

mRNA degradation genes are differentially expressed under diverse stress conditions
We next sought to establish whether mRNA degradation genes are transcribed in E. histolytica trophozoites grown in basal culture conditions and how these genes are modulated in response to the stress induced by heat shock, UV-C-induced DNA damage and nitric oxide treatments, as described elsewhere [19,20,21].We designed specific primers (Table S2) and performed quantitative real time-PCR assays for representative genes from the various mRNA degradation machineries (Figure 2).After heat shock, Ehxrn2, Ehlsm1, Ehupf1, and Ehrnase III gene expression was found to be upregulated 1.5-to 12-fold relative to non-treated cells, whereas the Ehcaf1, Ehnot1, Ehdcp2 and Ehago2-2 genes were repressed 2-fold (Figure 2).After UV-C-induced DNA damage, Ehupf1 was upregulated by a factor of 6, whereas the expression of Ehcaf1, Ehnot1, Ehdcp2, Ehlsm1, Ehedc3 and Ehrnase III genes was significantly repressed.Ehxrn2 and Ehago2-2 mRNA levels were similar in treated and non-treated cells.Ehnot1 and Ehdcp2 genes were significantly repressed after exposure to nitrosative stress in comparison with untreated trophozoites, whereas Ehcaf1, Ehlsm1, Ehupf1, Ehago2-2 and Ehrnase III were upregulated 2-to 16-fold.We quantified the mRNA expression of the L31 ribosomal protein-encoding gene as an endogenous control; its expression did not change significantly under any of the tested stress conditions.
EhXRN2 and EhDCP2 have the typical architecture of a 59-39 exoribonuclease and a decapping protein, respectively To initiate the study of mRNA degradation proteins in E. histolytica, we focused on the putative EhXRN2 exoribonuclease and the EhDCP2 decapping enzyme.We selected these proteins because their homologs have been described as P-body markers in the human and in yeast [22].Homologous XRN2 and DCP2 proteins from a variety organisms were assigned by BLAST analysis and ClustalW multi-alignments of predicted amino acid sequences (Figure S1A, S2A, and Table S3, S4).Ehxrn2 is an intronless gene (2813 nt) that predicts a 938 amino acids polypeptide.Sequence similarity searches revealed that EhXRN2 exhibits 46% identity (e = 4.9e-139) and 45% identity (e = 1.2e-108) with Homo sapiens XRN2 and Schizosaccharomyces pombe RAT1, respectively.Phylogenetic inference of full-length XRN2 sequences revealed that EhXRN2 is closely related to homologous proteins from other protozoa (Figure S1B).Comparison of the amino acids sequence of EhXRN2 with those of HsXRN2 and SpRAT1 homologs indicate the presence of conserved functional and structural motifs as represented by the XRN_N nuclease domain and the internal tower domain (Figure 3A), which is responsible for catalytic activity in SpRAT1 [23].Importantly, the active site motif KX 2 QQX 2 RR, which is critical for SpRAT1 ribonuclease function, is well conserved in EhXRN2.The high degree of homology between EhXRN2 and SpRAT1 allowed us to construct a structural model of EhXRN2 by using SpRAT1 crystal data as template.The predicted tertiary structure suggested that the EhXRN2 N-terminal domain is constituted by a six a-helix bundle and a two-stranded b-sheet, the orientation and length of which is similar to that found in the SpRAT1 structural model (Figure 3B-C).
The Ehdcp2 gene (699 nt) contains two introns and predicts an open reading frame of 233 amino acids.EhDCP2 exhibits 28% (e = 8.4e-25) and 29% (e = 1.6e-19) identity with human DCP2 and yeast DCP2, respectively.Phylogenetic inference revealed that EhDCP2 is closely related to homologous proteins from other protozoa (Figure S2B).EhDCP2 contains the conserved DCP2 box A domain (Figure 3D) and the consensus sequence for the nudix motif GX 5 EX 7 REUXEEXGU (including the three catalytic glutamate (E) residues).In mammalian cells, both these domains are responsible for cap structure removal [24].Modeling of EhDCP2's tertiary structure using the ScDCP2 structure as a template indicated that the two proteins are similar in terms of conformation and length (Figure 3E-F).

EhXRN2, EhDCP2 and EhCAF1 are cytoplasmic proteins
We next investigated the cellular localization of EhXRN2 and EhDCP2 factors in E. histolytica.An Ehxrn2 gene fragment (351 nt) and the full-length Ehdcp2 gene were PCR-amplified from genomic DNA and cloned in-frame into the pRSET-A expression vector.Recombinant 6xHis-tagged EhXRN2 and EhDCP2 polypeptides were expressed in bacteria (Figure 4A and B; lane 2), purified by affinity chromatography and used as immunogens in the mouse to obtain specific antibodies.In Western blot assays based on anti-EhXRN2 and EhDCP2 antibodies, we detected the corresponding purified recombinant proteins (Figure 4A and B, lane 5).The subcellular localization of EhXRN2 and EhDCP2 were investigated by Western blotting cytoplasmic and nuclear protein extracts from E. histolytica trophozoites (Figure 4C).In the cytoplasmic fraction, specific antibodies recognized bands at 103 kDa and 27.5 kDa, which correspond to the predicted molecular weights for endogenous EhXRN2 and EhDCP2 proteins, respectively (Figure 4C, lanes 2 and 4).In contrast, no proteins in the nuclear fraction (lanes 1 and 3) were recognized.The results of control assays with used pre-immune serum (which did not give any signal; data not shown) and antibodies that recognized the EhPC4 transcription factor (found mainly in the nuclear fraction (lane 5), as reported elsewhere [25]) confirmed the absence of crosscontamination between cytoplasmic and nuclear fractions.
We then used immunofluorescence and confocal microscopy assays to study the distribution of mRNA degradation proteins within trophozoites.As shown in Figure 4, EhXRN2 and EhDCP2 proteins were diffusely immunodetected throughout the cell (D-H and I-M) but were mainly concentrated in bright cytoplasmic foci with a diameter ranging from 50 to 300 nm; this resembled the organization into P-bodies seen in higher eukaryotes.To further investigate the localization of additional mRNA degradation proteins within trophozoites, we studied the cellular distribution of EhCAF1 deadenylase using specific rabbit antibodies generated in-house (manuscript in preparation).In particular, EhCAF1 had a similar pattern to EhXRN2 and EhDCP2 within cytoplasmic foci and a more diffuse general staining (Figure 4N-R).EhXRN2, EhDCP2 and EhAGO2-2 colocalize within EhCAF1-containing cytoplasmic foci The observed colocalization of EhXRN2, EhDCP2 and EhCAF1 prompted us to look for common focal, cytoplasmic structures.We performed co-immunolocalization and confocal microscopy assays in trophozoites by using EhCAF1 as a marker for the P-body-like structures (Figure 5).Our data showed that both EhXRN2 and EhDCP2 co-immunolocalized within the majority of EhCAF1-containing-cytoplasmic foci (Figure 5, A-D  and E-H).We next wondered whether additional factors related to mRNA degradation might also accumulate with EhXRN2 and EhDCP2 in a unique set of EhCAF1-containing foci.Hence, we performed immunofluorescence experiments using antibodies raised against EhAGO2-2 an Argonaut protein involved in RNA interference [18].Indeed, the EhAGO2-2 signal also colocalized with EhCAF1 in trophozoites (Figure 5, I-L).None of these proteins gave a significant nuclear signal.Taken as a whole, our data indicate that these mRNA degradation factors are concentrated within cytoplasmic P-body-like structures.
We next sought to establish whether the colocalized EhXRN2, EhDCP2, EhCAF1, and EhAGO2-2 proteins could interact physically by performing co-immunoprecipitation assays with anti-EhCAF1 or -EhXRN2 antibodies as bait and trophozoite cytoplasmic proteins as prey (Figure 5M).Our results showed that EhCAF1 co-immunoprecipitated with EhXRN2 and vice versa; these data fit with the co-immunolocalization studies and suggest a novel functional link between deadenylation and 59-to-39 mRNA degradation (Figure 5M, upper panel, lanes 2 and 4).In contrast, we did not detect co-immunoprecipitation of EhCAF1 with either EhDCP2 or EhAGO2-2 -suggesting that despite their colocalization in cytoplasmic foci in trophozoites, these proteins do not interact.Surprisingly, we also found co-immunoprecipitation of actin protein with both baits (Figure 5M, bottom panel, lanes 2 and 4).In agreement with these results, we also identified actin as an EhCAF1-interacting protein in pull-down assays using GST-EhCAF1 recombinant fusion protein as bait and cytoplasmic proteins as prey (manuscript in preparation).It has been reported that mammalian P-body proteins are anchored to the cytoskeleton and associate with tubulin and actin-bundles in human cells, suggesting that P-bodies require a dynamic interaction with the cytoskeleton [26].

The formation of cytoplasmic foci depends on the presence of active transcription
To evaluate the impact of active ongoing transcription on Pbody size and number, we measured the change over time in the number of EhCAF1-containing foci after transcription inhibition with actinomycin D, as described elsewhere [27].Our results showed that abrogation of transcription considerably altered the number of cytoplasmic foci (Figure 6A).We observed that number of cytoplasmic foci per cell diminished significantly after 4 h of incubation with actinomycin D. After 12 h treatment, the number of foci per cell had fallen by 50%.This effect was not observed in cells treated with the vehicle DMSO alone.The results of Western blot assays showed that the expression level of EhCAF1 was not significantly affected by actinomycin D treatment (data not shown).Hence, the fall in the number of foci was not due to the inhibition of EhCAF1 expression.Our data indicate that cytoplasmic foci formation is linked to active transcription -a feature that is conserved in P-bodies in higher eukaryotes.

Inhibition of translation affects cytoplasmic foci formation
A common feature of mammalian P-bodies is that they are disrupted by drugs such as cycloheximide that stabilize mRNAs in polysomes, whereas their number and size are increased by drugs such as puromycin that release ribosomes from mRNA [10,12].To evaluate the effect of inhibiting protein synthesis on the number and size of EhCAF1-containing cytoplasmic foci, we treated E. histolytica trophozoites with cycloheximide or puromycin.The results showed that the number of foci per cell rapidly fell after 15 min of incubation with cycloheximide and almost disappear after 120 min of treatment (Figure 6B).In contrast, puromycin treatment led to a slight increase in the number of cytoplasmic foci after 120 min (Figure 6C).Control assays showed that DMSO alone did not have a significant effect on the number of EhCAF1-containing foci.
EhCAF1 and EhAGO2-2 colocalize with poly(A) + RNA and dsRNA substrates We next looked at whether polyadenylated RNA substrates accumulate in P-body-like foci under basal culture conditions.Poly(A) + RNAs were detected in trophozoites by fluorescence in situ hybridization (FISH) with a FITC-labeled oligo(dT) 30 probe.EhCAF1 foci were evidenced by immmunofluorescence with specific antibodies.Interestingly, our results showed that although the FITC-oligo(dT) 30 signal was dispersed throughout the cytoplasm, a small proportion of probe colocalized with immunostained EhCAF1 in cytoplasmic foci (Figure 7E).Control experiments showed that the FITC-oligo(dT) 30 signal was specific for RNA, since it disappeared after RNAse A treatment (Figure 7F-J).
To gain insights into the potential role of P-body-like structures in RNA interference, we sought to determine whether a specific double-stranded RNA (dsRNA) accumulated in cytoplasmic foci and colocalized with EhAGO2-2 during silencing of an endogenous gene.The long dsRNA was designed to specifically knock down the Ehpc4 gene expression and its efficacy was validated by using the soaking approach described for amoeba [28].Briefly, dsRNA was generated in bacteria, purified, Cy3-labeled and transfected into trophozoites to produce fluorescent siRNA.Our data showed that the dsRNA-Cy3 signal was found in discrete, cytoplasmic foci seven days after transfection of the trophozoites (Figure 7M).Moreover, coupled FISH and immunofluorescence assays using the dsRNA-Cy3 probe and EhAGO2-2 antibodies showed that the two signals colocalized in large, cytoplasmic foci (Figure 7K-O).Remarkably, the focus size and signal intensity for EhAGO2-2 were greater in cells than in non-transfected, control cells -suggesting that EhAGO2-2 protein expression was induced by siRNA transfection.Western blot assays confirmed that EhPC4 protein levels were reduced by 50% and 80% three and seven days after dsRNA-Cy3 transfection, respectively.This indicated that RNA interference had been successful (Figure 7P-Q).We can rule out the possibility that the cytoplasmic foci containing siRNA against the Ehpc4 gene were non-transfected vesicles or endosomes, since we used the soaking method to deliver the dsRNA-Cy3 probe to trophozoites.However, we cannot be sure that only the siRNA guide strand was located in the cytoplasmic foci, since it has been reported that siRNAs can localize in P-bodies as dsRNA [29].

Discussion
mRNA decay pathways and machineries have been extensively studied in S. cerevisiae and H. sapiens.In contrast, little is known about mRNA degradation factors in protozoan parasites.Here, we identified protein-encoding genes for mRNA degradation machineries in E. histolytica.Our identification of decapping, deadenylation, NMD and RNA interference activities suggests that mRNA degradation components and mechanisms are conserved in this parasite.However, the composition of these machineries is simpler than that seen in higher eukaryotic cells -a frequent feature in the small genomes of divergent parasites.Importantly, we found that the deadenylation machinery is incomplete in E. histolytica because we did not find ccr4, pan2, pan3, or parn deadenylases; this result suggests that EhCAF1 is the parasite's main poly(A) ribonuclease.Furthermore, the decapping machinery lacks dcp1 and xrn1 genes but contains DCP2 and XRN2 homologs, which are typical P-body markers.Interestingly, the EhXRN2 and EhDCP2 proteins involved in 59-to-39 mRNA decay accumulated in discrete cytoplasmic foci in a manner similar to that seen in P-bodies in higher eukaryotes.This localization pattern was also observed for EhCAF1 deadenylase and EhAGO2-2.Given that DCP2, XRN2, CAF1 and AGO2 are resident Pbody proteins in human cells, our data suggest that the foci observed in E. histolytica correspond to P-bodies.However, the amebic foci are smaller and less rounded than P-bodies and so we prefer to refer to them as P-body-like structures.Given the similarity between proteins involved in mRNA decay in the human, in yeast and in E. histolytica and the colocalization of EhCAF1 with EhXRN2 and EhAGO2-2 in cytoplasmic structures, our results also suggest that mRNA degradation occurs at these foci.However, this hypothesis requires formal confirmation through additional experiments designed to evidence the degradation of RNA substrates and the localization of other mRNA decay proteins in these P-body-like structures.
Cytoplasmic foci resembling P-bodies have previously been described in the protozoan parasite Trypanosoma.Cassola et al. [30] reported that mRNA and proteins related to mRNA metabolism were concentrated in cytoplasmic mRNA granules in starved T. cruzi and T. brucei parasites and suggested that the said granules provided transient protection of transcription during periods of stress.However, these structures may not represent typical Pbodies.A complex containing CAF1, NOT1, NOT2, NOT5, DHH1 and a possible homolog of Caf130 was found in T. brucei, in which CAF1 deadenylation activity was shown to be essential for cell survival [31].In particular, CAF1 protein appeared to be predominantly distributed in the cytoplasm and the characteristic, punctate staining indicative of association with RNA storage or degradation granules was not found.Holetz et al. [32] showed that the TcDhh1 helicase protein was present in polysome-independent complexes and in discrete cytoplasmic foci resembling P-bodies.The latter varied in number as a function of the cell's nutritional status and following cycloheximide and puromycin treatments.Despite these initial efforts, the existence and potential roles of Pbodies in mRNA degradation events in protozoan parasites have not been comprehensively addressed.
We cannot rule out the existence of other RNA-containing foci in E. histolytica, such as the stress granules described in eukaryotic cells.Stress granules are cytoplasmic aggregates of stalled 48S preinitiation complexes.They include small ribosomal subunits, translation initiation factors (eIF2, eIF3, eIF4, eIF4G), PABP, p54/Rck and the TTP and CPEB RNA binding proteins that regulate mRNA translation and decay [33,34].Although P-bodies and stress granules share some proteins (e.g.TTP, p54/Rck, Lsm1, and Xrn1), there are some notable differences; P-bodies contain components of the mRNA decay pathway, whereas stress granules contain components of the translation initiation machinery and other specific markers (including TIA1) [35,36].The presence of stress granules in E. histolytica remains to be established; studies with specific markers would be required to gain a comprehensive view of cytoplasmic structures containing RNA binding proteins in this parasite.
In mammalian cells, treatment with actinomycin D induces the loss of P bodies [12], suggesting that the latter disaggregate once mRNA transcripts are no longer available for degradation.In contrast, the fact that trypanosome mRNA granules persist after actinomycin D treatment suggests that transcript degradation does not take place in these structures [30].P-body formation in human and yeast also depends on the presence of elevated levels of nontranslating mRNAs, which indicates that mRNAs exit the translation cycle before they enter P-bodies [37,38].The observed effects of transcriptional and translational inhibitors that are known to inhibit mRNA decay suggest that cytoplasmic foci represent sites in which mRNA decay occur.Our data revealed a similar situation in E. histolytica trophozoites, since the number of P-body-like structures number fell after treatment with actinomycin D and cycloheximide.It has been reported that the translation inhibitor puromycin induces an increase in the number and size of P-bodies by releasing ribosomes from mRNA.In E. histolytica trophozoites, the number of P-body-like structures increase slightly after puromycin treatment.Hence, our data indicate that P-bodylike structures are dynamic and are associated with ongoing mRNA synthesis and translation in E. histolytica.
The co-immunolocalization of the deadenylase EhCAF1 and the exoribonuclease EhXRN2 in cytoplasmic foci in E. histolytica supports the hypothesis whereby P-body-like structures have an important role in mRNA decay.Furthermore, these results constitute the first evidence of a functional interaction between deadenylation and 59-to-39 mRNA degradation.The interaction between actin and both EhCAF1 and EhXRN2 further suggests that the proteins in amoebic P-body-like structures are associated with the cytoskeleton, as is the case in the human [26].Additional, ongoing experiments should help us to understand the implications of these interactions within mRNA degradation machineries.
A small proportion of polyadenylated RNAs was concentrated in cytoplasmic foci, along with EhCAF1.In addition, fluorescent siRNAs were rapidly localized with EhAGO2-2 protein in cytoplasmic foci during RNA interference of Ehpc4 gene; this observation implies that P-body-like structures may have an important role in siRNA-mediated gene silencing by facilitating the interaction between siRNAs and Argonaut proteins.It was recently reported that EhAGO2-2 protein was localized in nuclei with siRNAs containing 59-polyphosphate termini during transcriptional gene silencing in E. histolytica strain G3 [39].Surprisingly, we did not detect a nuclear signal for EhAGO2-2.This may be due to the fact that depending on the siRNAs generated during RNA interference, Argonaut proteins can downregulate gene expression by transporting specific classes of small regulatory RNAs to distinct cellular compartments.During transcriptional silencing, EhAGO2-2 specifically transports small regulatory RNAs to nuclei in which gene silencing is mediated by histone modifications [39].Our results indicate that both EhAGO2-2 and specific siRNAs are preferentially targeted to the cytoplasm during RNA interference.
Taken as a whole, our present results provide evidence to suggest that mRNA degradation proteins involved in deadenylation and decapping pathways are organized within cytoplasmic, P-body-like structures that can be observed under the microscope.Future work is needed to understand the composition of these structures and define the localization of other mRNA decay proteins from NMD and RNA interference machineries and thus provide a better understanding of mRNA decay in this protist.This research should also provide us with an opportunity to answer intriguing, fundamental questions about the posttranscriptional regulation of gene expression in this deep-branching eukaryotic parasite.

Immunofluorescence confocal laser assays
Trophozoites were cultured on cover slides, fixed with 4% paraformaldehyde at 37uC for 40 min and permeabilized with 0.5% Triton-X100 in PBS 16at 37uC for 40 min.Next, cells were blocked in 2% BSA for 60 min at 37uC, washed three times with PBS 16 and incubated with either anti-EhCAF1 and anti-EhAGO2-2 antibodies generated in the rabbit or anti-EhXRN2, and anti-EhDCP2 antibodies generated in the mouse (1:200) at 37uC for 1 h.Cells were washed three times and incubated with fluorescein-labeled secondary antibodies (1:200) at 37uC for 1 h.Lastly, DNA was counterstained with 49-6-diamidino-2-phenylindole (DAPI, 5 mg/ml) for 1 min.Samples were analyzed under an inverted microscope connected to a laser confocal scanning system (TCS SP2, Leica Microsystems).

Immunoprecipitation assays
The immunoprecipitation protocol was essentially that described in [44] with minimal modifications.Briefly, cytoplasmic proteins (50 mg) were diluted in 300 ml HNTG buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 0.1% Triton, 10% glycerol) and incubated with 10 ml of protein G beads (Sigma) for 30 min at 4uC.Non-specifically interacting proteins were excluded by centrifugation at 3000 rpm at 4uC for 5 min.The supernatant was incubated with either anti-EhCAF1 antibodies (1:100) or anti-EhXRN2 antibodies (1:100) for 2 h at 4uC.Next, protein G-Sepharose beads were added to the samples and incubated for 16 h at 4uC.Immunoprecipitated proteins were collected by centrifugation, washed three times with HNTG buffer and then separated using 12% SDS-PAGE.Proteins were transferred to nitrocellulose membranes and detected by Western blot assays with rabbit anti-EhCAF1 or mouse anti-EhXRN2 antibodies.

Bacterial expression of double-stranded RNA and trophozoites transfection
Double-stranded RNA was obtained as described in [28].Briefly, the Ehpc4 full coding sequence (EHI_192520) was cloned into the SmaI and XhoI restriction sites of the pL4440 plasmid vector, which is flanked by T7 promoters [45].Competent E. coli HT115 cells were transformed with pL4440-Ehpc4 and spread onto LB-agar plates containing ampicillin and tetracycline.Isolated bacterial colonies were obtained after overnight culture at 37uC.Double-stranded RNA synthesis was induced by incubation with 2 mM IPTG for 4 h at 37uC.Cy3-labeling of purified dsRNA was performed with the Silencer siRNA labeling kit (Ambion), according to the manufacturer's instructions.The reaction mixture (containing 5 mg dsRNA, nuclease free-water, labeling buffer and Cy3 labeling reagent) was incubated for 1 h at 37uC in the dark.Labeled dsRNA was precipitated with ethanol, resuspended in DEPC-treated water and immediately used in colocalization assays.

RNA fluorescence in situ hybridization
To detect poly(A) + RNA in fluorescence in situ hybridization (FISH) assays, parasites were harvested and allowed to adhere to microscope slides.Cells were fixed with 4% paraformaldehyde in PBS, denatured in 70% formamide at 70uC and incubated for 10 min with 25 mM NH 4 Cl.The cells were permeabilized and blocked for 30 min in 0.5% Triton X-100/2% BSA in PBS, followed by 2 h of prehybridization at room temperature in hybridization buffer (2% BSA, 46 SSC, 5% dextran sulfate, 35% deionized formamide and 10 U RNAse inhibitor).Hybridization was performed overnight at room temperature in a humidified culture chamber in the presence of 300 ng/ml FITC-conjugated oligo-(dT) 30 in hybridization buffer.Slides were washed once in 46 SSC plus 35% deionized formamide and once in 46 SSC and were then counterstained with DAPI (1 mg/ml).Slides were mounted with Vectashield antifade (Vector).The subcellular localization was analyzed with a Leica TCS SP2 microscope.Merged images were obtained by superimposing image files using LCS Lite software (Leica Microsystems).Table S1 Predicted mRNA degradation proteins in Entamoeba histolytica.

(PDF)
Table S2 Primers used in quantitative real-time PCR assays.

(PDF)
Table S3 Comparison of EhXRN2 with related proteins from several organisms.(PDF) Table S4 Comparison of EhDCP2 with related proteins from several organisms.(PDF)

Figure 1 .Figure 2 .
Figure 1.A comparative representation of mRNA decay machineries in E. histolytica (black shapes) and H. sapiens (white and black shapes).Homologous proteins were assigned via BLAST and ClustalW multiple alignments.The undulated line denotes mRNA.doi:10.1371/journal.pone.0045966.g001

Figure 3 .
Figure 3.Comparison of XRN2 and DCP2 from eukaryotic organisms and E. histolytica.(A) A schematic representation of XRN2 protein from E. histolytica, H. sapiens and S. pombe.Red box: the XRN domain; cyan box: the tower domain (TD).The numbers are relative to the initial methionine in each protein.Lower panel: multiple alignments of amino acid residues from the TD of homologous proteins.Black box: a conserved residue; grey box: a conserved substitution.Asterisks mark the conserved residues in the active site.(B, C) Ribbon diagrams of ScRAT1's tertiary structure (B) and the predicted model of EhXRN2 (C).Red: the N-terminal XRN_N domain; cyan: the TD.(D) A schematic representation of the DCP2 protein from E. histolytica, H. sapiens, and S. cerevisiae.Red box: the DCP2 domain; blue box: the nudix domain.The numbers are relative to the initial methionine in each protein.Lower panel: multiple alignments of amino acid residues from the nudix motif (PF00293) of homologous proteins.Asterisks indicate the conserved residues involved in catalytic activity.Black box: a conserved residue; grey box: a conserved substitution.(E, F) Ribbon diagrams of ScDCP2's tertiary structure (E) and the predicted model of EhDCP2 (F).Red: the DCP2 domain; blue: the nudix domain; yellow: the nudix motif.Three-dimensional models were displayed and refined using the RCSB PBD Protein Workshop 3.9 viewer.doi:10.1371/journal.pone.0045966.g003

Figure 6 .
Figure 6.Determination of the number of EhCAF1-containing cytoplasmic foci after treatment with actinomycin D, cycloheximide or puromycin.After treatment with actinomycin D (5 mg/ml, panel A), cycloheximide (10 mg/ml, panel B) or puromycin (200 mg/ml, panel C), trophozoites were fixed, incubated with anti-EhCAF1 and FITC-labeled secondary antibodies, counterstained with DAPI and analyzed under an immunofluorescence confocal microscope.Left panel: change over time in the mean number of foci per trophozoite, in response to inhibition of transcription (A) or translation (B-C).The data correspond to six randomly chosen fields.Right panel: representative images of cells treated with DMSO vehicle (control) and actinomycin D, cycloheximide or puromycin (observed in the green channel).doi:10.1371/journal.pone.0045966.g006

Figure 7 .
Figure 7. Colocalization of poly(A) + RNA and dsRNA substrates with EhCAF1 and EhAGO2-2 in cytoplasmic foci.(A-E) Poly(A) + RNA and EhCAF1 colocalization assays.Trophozoites were immunostained with EhCAF1 (B) antibodies.Poly(A) + RNAs were detected by hybridization with FITC-conjugated oligo-(dT) 30 (C).The cells were counterstained with DAPI (D) and analyzed with immunofluorescence confocal microscopy.The merged image (E) shows the overlapping signals.White arrowheads indicate the accumulation of poly(A) + RNA and EhCAF1 signals.Yellow arrowheads indicate FITC-poly(A) + signal that does not overlap with EhCAF1-containing foci.(F-J) RNAse A treatment, performed as a control experiment.Immunodetection of EhCAF1 (G), detection of poly(A) + (H) and DNA counterstaining with DAPI (I).Merged image (J).(K-O) EhAGO2-2 and Cy3-dsRNA colocalization assays.Trophozoites were transfected with Cy3-dsRNA targeting Ehpc4 (K-O) and then immunostained with EhAGO2-2 antibodies on day seven after transfection (L).Cells were counterstained with DAPI (N) and analyzed with immunofluorescence confocal microscopy.The merged image (O) shows both signals.Arrowheads mark colocalized signals in cytoplasmic foci.(P) Western blot analysis for EhPC4 and EhActin from proteins extracts obtained on day seven after Cy3-dsRNA transfection.(Q) A densitometric analysis of the bands in P. doi:10.1371/journal.pone.0045966.g007

Figure
Figure S1 XRN2 related proteins.(A) Multiple alignments of related XRN2 proteins.(B) Phylogenetic relationships between EhXRN2 and XRN2 family members.The rooted tree was created with a neighbor-joining algorithm in the Mega 5.1 program, on the basis of ClustalW alignments of full amino acid sequences.Each protein's accession number is given in brackets.(PDF) Figure S2 DCP2 related proteins.(A) Multiple alignments of related DCP2 proteins.(B) Phylogenetic relationships between EhDCP2 and DCP2 family members.The rooted tree was created with a neighbor-joining algorithm in the Mega 5.1 program, on the basis of ClustalW alignments of full amino acid sequences.Each protein's accession number is given in brackets.(PDF)