Stress-induced nuclear depletion of Entamoeba histolytica 3′-5′ exoribonuclease EhRrp6 and its role in growth and erythrophagocytosis

The 3′-5′ exoribonuclease Rrp6 is a key enzyme in RNA homeostasis involved in processing and degradation of many stable RNA precursors, aberrant transcripts, and noncoding RNAs. We previously have shown that in the protozoan parasite Entamoeba histolytica, the 5′-external transcribed spacer fragment of pre-rRNA accumulates under serum starvation–induced growth stress. This fragment is a known target of degradation by Rrp6. Here, we computationally and biochemically characterized EhRrp6 and found that it contains the catalytically important EXO and HRDC domains and exhibits exoribonuclease activity with both unstructured and structured RNA substrates, which required the conserved DEDD-Y catalytic-site residues. It lacked the N-terminal PMC2NT domain for binding of the cofactor Rrp47, but could functionally complement the growth defect of a yeast rrp6 mutant. Of note, no Rrp47 homologue was detected in E. histolytica. Immunolocalization studies revealed that EhRrp6 is present both in the nucleus and cytosol of normal E. histolytica cells. However, growth stress induced its complete loss from the nuclei, reversed by proteasome inhibitors. EhRrp6-depleted E. histolytica cells were severely growth restricted, and EhRrp6 overexpression protected the cells against stress, suggesting that EhRrp6 functions as a stress sensor. Importantly EhRrp6 depletion reduced erythrophagocytosis, an important virulence determinant of E. histolytica. This reduction was due to a specific decrease in transcript levels of some phagocytosis-related genes (Ehcabp3 and Ehrho1), whereas expression of other genes (Ehcabp1, Ehcabp6, Ehc2pk, and Eharp2/3) was unaffected. This is the first report of the role of Rrp6 in cell growth and stress responses in a protozoan parasite.

whereas in human, it is concentrated in the nucleoli and also found in nucleoplasmic and cytoplasmic exosome (24). Although RRP6 is not essential for viability, its deletion in S. cerevisiae leads to temperature sensitivity, slow growth, and accumulation of 5Ј-ETS sequences (25).
The Rrp6 domain structure has been extensively studied in yeast and human by crystal structure analysis. The exonuclease (EXO) domain of yeast and human RRP6, and that of bacterial RNase D belongs to the DEDD superfamily (DEDD-Y subfamily) of exonucleases that act by a hydrolytic mechanism involving two divalent metal ions (26 -28). The EXO domain is flanked by a single C-terminal helicase and RNase D C-terminal (HRDC) domain (29). These two domains are sufficient for catalytic activity in yeast (30). However, both yeast and human RRP6 contain additional domains. These include an N-terminal PMC2NT domain that is needed for Rrp6 to bind to its cofactor Rrp47 (a dsRNA-and DNA-binding protein) (31)(32)(33)(34); a region C-terminal to HRDC required for interaction with the core exosome and with RNA (35); and a putative NLS domain at the C terminus (28).
We have been studying the regulation of ribosomal biogenesis in the primitive parasitic protist, Entamoeba histolytica, which causes amoebiasis in humans. We have earlier shown that transcription continued in E. histolytica cells subjected to growth stress by serum starvation, but pre-rRNA processing was inhibited, leading to accumulation of unprocessed pre-rRNA and partially processed fragments of the 5Ј-ETS (36). The removal of 5Ј-ETS subfragments in model organisms is done by the 3Ј-5Ј exonuclease activity of Rrp6 (3,9,12). To investigate whether Rrp6 might be performing a similar function in a primitive eukaryote like E. histolytica we biochemically characterized EhRrp6. Here we show that although EhRrp6 sequence differs from the S. cerevisiae and human homologs as it has large deletions at both the N and C termini, the enzymatic properties of EhRrp6 are conserved, and Ehrrp6 could complement the growth defect of a Scrrp6⌬ deletion mutant. Ehrrp6 down-regulation led to increase in levels of 5Ј-ETS subfragments. Furthermore, we show that EhRrp6 is essential for E. histolytica growth and acts as a stress sensor. It is lost from the nuclei during growth stress and is required to maintain the transcript levels of key genes involved in phagocytosis, a process important for E. histolytica pathogenesis.

Identification of exosome core subunits of E. histolytica
The focus of this study is the characterization of EhRrp6, which is implicated in 5Ј-ETS processing, and is functionally associated with the core exosome. We undertook a preliminary analysis to computationally identify the exosome subunits of E. histolytica. By performing a sequence homology search we identified 8 of the 9 proteins corresponding to the eukaryotic Exo9-core ring and cap subunits in E. histolytica. These have also been reported in an earlier study (37). For further categorization, we performed a phylogeny construction using the Exo9 protein sequences from Homo sapiens, S. cerevisiae, and Trypanosoma brucei. As expected, the ring and cap proteins clustered into two different groups (Fig. S1), with the latter containing three members; Rrp4, Rrp40, and Csl4. Of these we could identify the E. histolytica homologs of Rrp4 and Rrp40 (EHI_163510 and EHI_004770, respectively), but the Csl4 homolog could not be identified. This corroborates with the earlier study (37). The remaining six E. histolytica proteins grouped with the six eukaryotic ring subunits (Rrp41, Rrp42, Rrp45, Rrp46, Rrp43, and Mtr3). However, it was not possible to identify the individual E. histolytica homologs for each of these six subunits. Rather the sequences grouped into two categories: Rrp41-like (EHI_040320 and EHI_086520) and Rrp42-like (EHI_000580 and EHI_188080). The remaining two sequences (EHI_126330 and EHI_166910) also grouped in the Rrp42-like category but with low confidence. This was unlike the previous study where all six had been classified as Rrp45-like (37). Our analysis shows overall conservation of the core exosome structure in E. histolytica; the major difference being absence of Csl4 subunit.

Comparative sequence and structure analysis of EhRrp6
We looked for the E. histolytica homologue of Rrp6 byj NCBI-BLAST search performed against nonredundant protein database of E. histolytica. Only one protein (XP_650756) with a Rrp6-like EXO domain was found and it was referred to as EhRrp6. EhRrp6 has a deduced size of 517 amino acids and is encoded by a single copy gene (EHI_021400) lacking any introns. Multiple sequence alignment of the EhRrp6 sequence with its homologues from other organisms showed well-conserved EXO and HRDC domains (Fig. 1A). Across its entire sequence EhRrp6 shared 30.5 and 32.9% sequence identity with ScRrp6 and HsRRP6, respectively. The EhRrp6 EXO domain (183-373 amino acids) was of comparable length to the EXO domains of ScRrp6 and HsRRP6, and shared 50 and 51% sequence identity, respectively. The EhRrp6 HRDC domain (398 -463 amino acids) was of comparable length to the HRDC domain of ScRrp6, whereas it was 15 amino acids shorter than HRDC of HsRRP6 and the sequence identity was 45 and 38%, respectively. Unlike Rrp6 homologues in other organisms we were unable to find a conserved PMC2NT domain in EhRrp6 (Fig. 1A). We also checked for the presence of the PMC2NT domain in E. histolytica as a separate protein, or as part of some other protein by taking the PMC2NT domain of HsRrp6, ScRrp6, and TbRrp6 as query and searching the database as described for EhRrp6 identification. No significant hit was obtained. Phylogenetic analysis using the amino acid sequence of conserved domains (EXO and HRDC) of EhRrp6 showed that it was closer to the protein from lower eukaryotes (Fig. 1B).
Amino acid sequence analysis showed that the active site residues in the EXO domain, known to play a critical role in RNA degradation, were well conserved in EhRrp6 (Asp-212, Glu-214, Asp-270, Tyr-335, and Asp-339) ( Fig. 2A). The structure of EhRrp6 (residues 183 to 463) containing the EXO and HRDC domains was modeled through comparative homology modeling (Fig. 2B). The DOPE score (Ϫ33247.70), ProSA web Z score (Ϫ7.49), and PROCHECK results (Ramachandran plot; 93.6 and 6.4% residues in most favorable and additional allowed Exonuclease EhRrp6 in growth stress region, respectively) suggested that the EhRrp6 modeled structure was of good quality. Its overall structure was similar to its homologs and had root mean square deviation values of 0.531, 0.763, and 0.714 Å over the aligned residues with structures of RRP6 from H. sapiens, S. cerevisiae, and T. brucei, respectively (Fig. 2C). Like Rrp6 of H. sapiens and T. brucei, the linker of EhRrp6 was shorter (9 amino acids) than the linker of S. cerevisiae (26 amino acids) (Fig. 2D). The shorter linker has been suggested to make the active site more accessible for both structured and nonstructured RNA substrates in H. sapiens and T. brucei Rrp6 (38) suggesting a similar behavior for EhRrp6.
Furthermore, we performed MD simulation of 12 in silico systems to determine the effect of the absence of Mg 2ϩ ions or introduction of DEDD-Y mutations on the catalytic activity of EhRrp6 (Fig. 2E). In the absence of Mg 2ϩ ions, the structure of EhRrp6 was deviated significantly from its native (modeled) structure, suggesting that Mg 2ϩ ions stabi-

Exonuclease EhRrp6 in growth stress
lize the structure and are crucial for catalytic activity (data not shown).

Enzymatic activity of EhRrp6
To assay for the exoribonuclease activity of EhRrp6 we expressed the WT protein in Escherichia coli (SHuffle), as a His 6 fusion protein and purified it by Ni-NTA affinity chromatography and gel filtration. Mutants in the highly conserved DEDD-Y active site residues were obtained to determine their effect on enzyme activity. Two double mutants (D212A,E214A; Y335A,D339A) and a single mutant (D270A) were generated and the recombinant proteins purified. The purity was checked by SDS-PAGE electrophoresis (Fig. S2). To assay for exoribonuclease activity a 50-nt AU-rich RNA substrate was used, which lacked secondary structure, because double-stranded regions could potentially restrict enzyme activity (28). The purified protein (0.1 M) was incubated with 5Ј-radiolabeled AU-rich RNA (1 M). A time course assay was performed for the indicated time periods (Fig. 3A). There was loss of fulllength RNA substrate with concomitant appearance of progressively shorter molecules and final accumulation of an end product Ͻ10 nt. This pattern is very similar to that reported for the distributive 3Ј-5Ј exoribonuclease of yeast and human Rrp6 (28,39). The EhRrp6 mutants in the conserved active site residues (mentioned above) were assayed with the same substrate under the same reaction conditions. None of the mutant proteins showed any activity (Fig. 3B), confirming the conserved role of these residues in the active site of EhRrp6. Activity of the WT enzyme was also checked with another RNA substrate expected to have secondary structure. This was a 60-nt RNA modified from an E. histolytica sequence (3Ј-end of EhSINE1) (40). The sequence was predicted by RNA-fold to have a 30-nt 3Ј-overhang followed by a stem-loop (Fig. 3C). This RNA was treated with EhRrp6 under the conditions described for the AU-rich RNA substrate. Unlike the latter substrate, which was progressively degraded to shorter fragments, the 60-nt structured substrate showed accumulation of a 30-nt intermediate, in addition to the Ͻ10 nt end product (Fig. 3C). This could be due to stalling of EhRrp6 at the base of the stem-loop, as shown for Rrp6 from yeast and human (28). To further characterize the properties of EhRrp6 we determined the conditions of temperature and pH with the 50-nt AU-rich RNA substrate and found maximal activity at 37°C, pH 8.0 (Fig. 3, D and E). The requirement for Mg 2ϩ was absolute as no activity was found in

Exonuclease EhRrp6 in growth stress
the absence of Mg 2ϩ or presence of EDTA (Fig. 3F). The nuclease activity of EhRrp6 was specific to RNA as no activity was observed with either ss-or dsDNA substrates (Fig. 3G).

Ehrrp6 complements the growth defect of rrp6⌬ yeast strain
To determine whether Ehrrp6 could functionally complement the S. cerevisiae rrp6 mutant, we expressed Ehrrp6 in a temperature-sensitive rrp6⌬ yeast strain (3). This strain grows normally at 30°C but is severely growth restricted at 37°C. Successful transformation of yeast cells with the WT and mutant Ehrrp6 was confirmed by PCR with Ehrrp6-specific primers and expression of EhRrp6 was confirmed by Western blot analysis (Fig. S3). Cells transformed with Ehrrp6 grew normally at both 30 and 37°C, whereas those transformed with the Ehrrp6 catalytic domain double mutant (D212A,E214A), or with vector alone were unable to grow at the nonpermissive temperature (Fig. 4). These data demonstrate that Ehrrp6 indeed encodes a functional protein that can complement the growth defect in the yeast rrp6 mutant strain. Although Ehrrp6 lacks the N-terminal PMC2NT domain and has a shorter C-terminal region (by 164 amino acids) compared with yeast sequence, its ability to complement the yeast mutant strain indicates that Ehrrp6 contains the essential sequences required for Rrp6 activity and the missing sequences could be dispensable or redundant under the experimental conditions.

Down-regulation of Ehrrp6 leads to accumulation of 5-ETS
The role of RRP6 in the removal of 5Ј-ETS subfragments has been well documented both in yeast and human (6,21,26). To demonstrate a direct correlation of EhRrp6 levels with 5Ј-ETS processing in E. histolytica we down-regulated the expression of Ehrrp6 by using the antisense expression approach that has been successful in expression knockdown of a variety of E. histolytica genes (41,42). Cell lines were constructed for ectopically overexpressing Ehrrp6 in the sense or antisense orientation using a vector with tetracycline (tet)-inducible promoter (43,44) (Fig. 5A). The levels of EhRrp6 were determined by Western blotting. In antisense Ehrrp6-overexpressing cells the levels of EhRrp6 came down by ϳ1.8-fold of cells transfected with vector alone, after tet-induction; whereas conversely the levels went up ϳ1.5-fold in sense-overexpressing cells (Fig. 5B). The level of 5Ј-ETS subfragments was determined in these cell lines by Northern hybridization. We have earlier shown accumulation of 5Ј-ETS subfragments (0.7-0.9 kb), which migrate as a broad band in serum-starved E. histolytica cells (36). Very strong accumulation of these fragments was seen in the antisense cell line grown for 48 h with tet ( Fig. 5C). There was no accumulation of 5Ј-ETS subfragments in the sense cell line after 48 h of growth with tet. Rrp6 is known to be involved in generating the mature 3Ј-end of 5.8S rRNA. We checked for the accumulation of the 5.8S 3Ј-extended precursor (equivalent to 5.8Sϩ30 in (25, 45)) by quantitative RT-PCR using a primer 30 nt downstream of 5.8S (Fig. 5D). There was ϳ5-fold accumulation of 5.8S extended precursor in EhRrp6 down-regulated cells. Down-regulation of Ehrrp6 resulted in a severe growth defect in E. histolytica cells (Fig. 5E). Antisense expressing cells showed slow growth phenotype even in the absence of tet, presumably due to some leaky expression. These cells attained stationary phase at lower cell density compared with the sense cell line. Growth of antisense cell line was very slow in the presence

Exonuclease EhRrp6 in growth stress
of tet, although cell lysis was not observed. Overexpression of Ehrrp6 did not seem to elicit a growth phenotype. These data demonstrate a direct role of Ehrrp6 in degradation of processed 5Ј-ETS subfragments and show that EhRrp6 performs an essential function in E. histolytica.

Subcellular localization of EhRrp6 in normal and serum-starved E. histolytica cells
To determine the subcellular distribution of the EhRrp6 protein in E. histolytica we used polyclonal antibody raised against the recombinant protein. The specificity of the antibody was

Exonuclease EhRrp6 in growth stress
checked by Western blot analysis of E. histolytica total cell lysate (Fig. S4). The antibody cross-reacted with a single band of 60 kDa, corresponding to the expected size of EhRrp6. Trophozoites were stained with this antibody and co-localization was determined in nuclei stained with Hoechst. In normal E. histolytica cells EhRrp6 was concentrated in the nuclei and there was substantial staining of the cytosol as well (Fig. 6A). The relative intensity of nuclear versus cytoplasmic staining was determined by selecting five random regions of the cytosol and nucleus per cell and averaging out the data for 10 cells (Fig. 6C). The nuclear staining intensity was ϳ3 times higher than cytosol in normal cells. Thus our data show dual localization of Rrp6 in nuclei and cytosol of normal E. histolytica trophozoites. Such dual localization was also seen in human cell lines and in T. brucei (12,24,46,47), whereas in yeast cells the protein was detected only in the nucleus (26,48). Interestingly, when E. histolytica cells were subjected to serum starvation for 24 h there was ϳ3-fold reduction in the total levels of EhRrp6 in cell lysates, as measured by Western blotting (Fig. 6D). Transcript levels, determined by RNA-seq (two biological replicates), were reduced by ϳ1.7fold, indicating a greater reduction of protein levels (Fig. 6D). In both cases normalization was done with EhCabp1, which did not change in starved cells. Immunofluorescence analysis of these cells showed almost complete loss of EhRrp6 from nuclei of serum-starved cells, whereas the cytosolic staining was maintained ( Fig. 6A). Upon replenishment of serum to starved cells there was a gradual increase in EhRrp6 levels in nuclei, and by 12 h the nuclear:cytoplasmic ratio was restored to that in normal cells (Fig. 6B). The disappearance of the accumulated 5Ј-ETS subfragments in 24-h serum-starved cells after serum replenishment paralleled the nuclear restoration of EhRrp6 (Fig. 6E). The preferential loss of EhRrp6 from nuclei of serumstarved cells was also demonstrated by subcellular fractionation to obtain lysates from cytoplasmic and nuclear fractions. The protein was detected by Western blotting (Fig. 6F). The E. histolytica calcium-binding proteins, EhCabp1 and EhCabp6, were used as cytoplasmic and nuclear markers, respectively, as they are known to be exclusively localized in these compartments (49,50). The data showed a ϳ3.5-fold drop in total EhRrp6 levels in starved cells compared with control ( Fig. 6F). The protein concentration in the cytosolic fraction was not significantly altered, whereas there was ϳ30-fold reduction of protein in the nuclear fraction. Our data suggest that the loss of EhRrp6 from nuclear fraction was not due to increased localization to the cytosol, as the cytosolic fraction showed no increase. Interestingly, the mechanism is totally reversible, with the nuclear:cytoplasmic ratio being restored within 12 h of serum replenishment.
Furthermore, we checked whether depletion of EhRrp6 from the nucleus was specific to serum starvation or was a more general stress response. Immunofluorescence analysis was done with cells subjected to heat stress or oxidative stress, using the methods described (51). In both cases the nuclear staining of EhRrp6 dropped significantly within 60 min of stress induction and by 90 min the protein was almost completely lost from the nucleus (Fig. 6, G-I). There was no significant change in cytoplasmic staining intensity. Hoechst staining showed that nuclei retained their integrity in stressed cells and the loss of staining was not due to nuclear breakdown. Thus the nuclear loss of EhRrp6 appears to be a general response to growth stress. The depletion of EhRrp6 from the nucleus in serumstarved cells provides an explanation for our earlier observation that the 5Ј-ETS subfragments, which are known substrates of Rrp6, accumulate to high levels in serum-starved E. histolytica cells (36).

The proteasome system is involved in nuclear loss of EhRrp6 in serum-starved cells
The presence of proteasome has earlier been demonstrated in E. histolytica (52). Treatment of cells with specific proteasome inhibitors like lactacystin and MG-132 has been shown to result in growth defect and also inhibition of encystation in Entamoeba invadens (53). To check whether proteasome inhibition could stall the loss of EhRrp6 during serum starvation we treated serum-starved cells with lactacystin or MG-132 and checked EhRrp6 levels by Western blotting. In untreated cells EhRrp6 levels began to decline within 8 h of serum starvation and were Ͻ28% of control by 12 h. This decline was not observed in serum-starved cells treated with lactacystin or MG-132 (Fig. 7A) at the concentrations of the inhibitors shown to block Entamoeba growth (53). EhRrp6 levels were Ͼ79% of control in cells treated with the inhibitors. This suggests the role of proteasome in targeting EhRrp6 for degradation during growth stress.

EhRrp6 is required for E. histolytica growth and phagocytosis
As already shown, EhRrp6 down-regulation led to a severe growth defect (Fig. 5E). To further understand the involvement of EhRrp6 in cell growth we determined the time taken for EhRrp6 levels to decrease following tet addition in antisense cell lines. Tet was added to the culture 6 h after inoculating cells into fresh growth medium. Cells were removed at different time points and the levels of EhRrp6 were measured by Western blotting. Tet addition to antisense cells resulted in a gradual decline of EhRrp6 levels, and by 36 h the levels in antisense (ϩtet) cells were ϳ25% of cells grown without tet (Fig. 7B). Conversely, the EhRrp6 levels were elevated ϳ1.3-fold in sense cells (ϩtet). Subsequent experiments were carried out after 36 h of tet addition. We checked whether EhRrp6-depleted cells showed any hallmarks of apoptosis. From Hoechst staining we found nuclear integrity to be maintained in these cells (Fig. 7C). We determined the extent of blebbing, which generally increases in apoptotic cells. We found ϳ2-fold reduction in the average number of blebs per cell in antisense (ϩtet) cells (Fig.  7D), which may be due to reduced cell motility. Indeed, we found ϳ30% reduction in cell motility in the antisense (ϩtet) cells (Fig. 7E), as determined by transwell cell migration assay (51). No appreciable change in cell size was observed. Thus it is unlikely that EhRrp6-depleted cells underwent apoptosis. Rather, the reduction in cell growth may be due to block in cell proliferation.
Because EhRrp6 levels were severely reduced during growth stress and conversely, reduction of EhRrp6 led to growth arrest (Figs. 5 and 6), we checked whether overexpression of EhRrp6 could have a protective effect during growth stress. The EhRrp6 sense cell line was grown with tet, and at 48 h the cells were Exonuclease EhRrp6 in growth stress subjected to serum starvation. There was significant increase in cell number in the EhRrp6 overexpressing cells even after serum starvation, compared with control cell lines in which the cell number increased very slightly after starvation (Fig. 7F).
This suggests that the presence of EhRrp6 could delay the onset of stress signaling.
Phagocytosis is an essential process for E. histolytica growth and is required for pathogenesis. We checked whether EhRrp6

Exonuclease EhRrp6 in growth stress
down-regulation had any effect on erythrophagocytosis. Cells were incubated with RBCs, 36 h after tet addition, by which time the EhRrp6 levels had declined in the antisense (ϩtet) cells. RBC uptake, determined by measuring the absorbance at different time points, was severely reduced in the EhRrp6 antisense cells (Fig. 8A). After 10 min, uptake was only ϳ50% compared with cells transfected with vector (ϩtet). The average number of phagocytic cups per cell was ϳ40% in the antisense cells compared with vector control (Fig. 8B). The sense (ϩtet) cells behaved the same as control cells. This showed that EhRrp6 down-regulation had a marked inhibitory effect on erythrophagocytosis.
To further understand the involvement of EhRrp6 in erythrophagocytosis we checked localization of EhRrp6 in normal cells during RBC uptake. EhRrp6 did not localize to the phagocytic cups, which were strongly stained with TRITC-phalloidin due to enrichment of F-actin (Fig. 8C), nor was it associated with the phagosome. This indicates that physical involvement of EhRrp6 with the phagocytic machinery was unlikely.

A subset of phagocytosis-related genes are down-regulated in EhRrp6-depleted cells
A large number of E. histolytica genes have been demonstrated to have direct roles in phagocytosis (41, 49, 54 -57). We checked the expression levels of some of these genes in EhRrp6 down-regulated cells by Western blotting of total cell lysates with gene-specific antibodies (Fig. 9A). Expression of two of the genes (Ehcabp3 and Ehrho1) was significantly reduced in the antisense (ϩtet) cells compared with control cells (TOC vector ϩtet). Expression was reduced by 1.8-and 1.6-fold of control, respectively. Three of the tested genes (Ehcabp1, Ehcabp6, and Ehc2pk) showed no change in expression, whereas Eharp2/3 expression was slightly reduced (85% of vector control). However, this apparent decrease in Eharp2/3 is probably not significant as the AS (Ϫtet) cells also showed the same expression (Fig. 9A). Immunolocalization studies were done with EhCabp3 and EhRho1 (which were down-regulated in antisense cells) and EhCabp1, which remained unchanged, was used as control. In the vector (ϩtet) and sense (ϩtet) cell lines EhCabp3, EhRho1, and EhCabp1 all co-localized with Ehactin at the phagocytic cups, as expected for normal E. histolytica cells. All of these proteins were significantly enriched in phagocytic cups compared with cytosol as determined quantitatively (Fig. 9B). However, in the antisense (ϩtet) cell line the pattern was different. As expected, staining for EhRrp6 was very low in these cells and it was low both in the nucleus and cytosol. Staining intensity of EhRho1 and EhCabp3 was also extremely low in these cells, whereas the level of EhCabp1 was comparable with control. No enrichment of EhRho1 and EhCabp3 could be seen at the phagocytic cups, whereas Ehactin and EhCaBP1 were enriched at phagocytic cups in these cells and were colocalized (Fig. 9B). We conclude that EhRrp6 down-regulation specifically affected the levels of selected proteins in the phagocytic pathway rather than a generalized inhibitory effect on all genes.
To determine whether the drop in protein levels of EhRho1 and EhCabp3 could be due to a translational defect we checked the transcript levels of these genes by quantitative RT-PCR and found ϳ6.3and ϳ3.2-fold reduction of EhCabp3 and EhRho1 transcripts, respectively, and no change in EhC2pk. EhCabp1 was used as internal control (Fig. 9C). Because the fold-reduction in protein levels did not exceed the drop in transcript levels, it appears unlikely that EhRrp6 down-regulation affects the translation of EhCabp3 and EhRho1. We conclude that EhRrp6 is required for maintaining the transcript levels of these key phagocytosis-related genes.

Discussion
The 3Ј-5Ј exoribonuclease encoded by RRP6 is involved in 3Ј-end processing of a variety of stable RNA precursors including pre-rRNAs, snRNAs, and snoRNAs and in transcription termination and regulation of poly(A) tail length. It is also required for removal of aberrant transcripts and cryptic unstable transcripts and is thus a key enzyme in RNA homeostasis (58,59). Here we have undertaken a detailed study of this exonuclease in the protozoan parasite E. histolytica, in which the 5Ј-ETS subfragments of pre-rRNA are found to accumulate under growth stress. The excised 5Ј-ETS is a known target of degradation by Rrp6 in other organisms and accumulates in rrp6 mutant cells (2,12,60).
Rrp6 works in conjunction with the exosome. We found overall conservation of the core exosome in E. histolytica; the major difference being absence of Csl4 subunit. Although CSL4 is essential in yeast, mutant strains with truncated versions that lack NTD or S1 domain are viable, and the zinc-ribbon domain is not essential in vivo. In addition it has been shown that Csl4 is not stably associated with exosomes in vitro (61). Absence of csl4 has also been reported in another protozoan parasite, Giardia lamblia (37), and may be a more common feature in early eukaryotic evolution.
Sequence analysis and comparative modeling of EhRrp6 showed high conservation of structure in the catalytic EXO and HRDC domains. EhRrp6 lacked the N-terminal PMC2NT domain found in both yeast and human Rrp6. This domain is also present in the T. brucei Rrp6 (38) and in one of the three

Exonuclease EhRrp6 in growth stress
Rrp6 isoforms of A. thaliana (12). It interacts with Rrp47, which promotes the catalytic activity of Rrp6 and may also have a role in maintaining appropriate expression levels of Rrp6 (33,62). Rrp47, both from yeast and human (called C1D), binds structured nucleic acids and is thought to promote Rrp6 activity by facilitating its binding to structural elements within RNA, for example, helices at the 3Ј termini (32,33). We could not find any sequence homologous to rrp47 in the E. histolytica data-

Exonuclease EhRrp6 in growth stress
base. It appears that this gene may be missing in E. histolytica and consequently its interacting domain in Ehrrp6 is also absent. In this context it was interesting that Ehrrp6 could complement the ts growth defect of the Scrrp6⌬ mutant. The same observation was also made in A. thaliana in which the isoform that lacked the PMC2NT domain (Atrrp6l1) could complement the ts growth defect of Scrrp6⌬, whereas Atrrp6l2 in which this domain is present could not (11,12). It is possible that Rrp6 enzymes that have evolved to function in the absence of the PMC2NT domain could use alternative mechanisms for structured RNA recognition. In yeast the N-terminal domains of Rrp6 and Rrp47 interact to provide a surface for the binding of MTR4 helicase (63), which is involved in unwinding the 3Ј-tail of RNA substrate so that it can be threaded into the Figure 8. EhRrp6 is required for E. histolytica phagocytosis. A, spectrophotometric assay for erythrophagocytosis. The RBC uptake assay was performed in the indicated cell lines. Cells were incubated with RBCs for different time points and the amount of RBC uptake was determined spectrophotometrically using RBC solubilization assay as described under "Experimental procedures." The experiments were carried out three times independently. B, quantitative analysis of phagocytic cups. Thirty cells were randomly selected (in three independent experiments) and the number of phagocytic cups were counted for each cell line grown with tet. C, to look at EhRrp6 localization with respect to phagocytic cups, cells grown for 48 h were harvested and incubated with RBCs for indicated time periods at 37°C and subsequently fixed for further processing. Cells were immunostained with EhRrp6-specific antibody, followed by Alexa 488conjugated secondary antibody (green). F-actin was stained with TRITC-conjugated phalloidin (red). Slides were viewed using Nikon confocal microscope. Arrowhead indicates phagocytic cups, the cross marks the closure of cups before scission, and the star marks the phagosome. Figure 9. Effect of EhRrp6 depletion on expression of phagocytosis related genes. A, selected genes known to be involved in phagocytosis (Ehcabp3, Ehrho1, Ehcabp1, Ehcabp6, Ehc2pk, Eharp2/3) and EhCoactosin were studied. Protein levels were determined by Western blotting in the indicated cell lines. Tet addition (30 g/ml) was for 36 h. Ehcoactosin was used as loading control. Band intensity in each lane was determined by densitometry using ImageJ software. Normalized values were obtained from an average of three experiments and values relative to TOC (ϩtet) cells are indicated at the bottom of each lane. B, immunolocalization of phagocytosis-related proteins in EhRrp6-depleted cells and control cells, grown for 36 h in the presence of tet and incubated with RBCs for 10 min. Phagocytic cups were visualized by Ehactin staining with TRITC-phalloidin (red). Indicated proteins were immunostained with specific antibodies, followed by secondary antibodies conjugated with Alexa 488 (green) (for EhRrp6) and Alexa 405 (blue) (for EhCabp1, EhRho1, and EhCabp3). Arrow shows the phagocytic cups. Scale bar indicates 10 m. For quantitative analysis (shown in panels on the right), relative pixel intensity of fluorescent signals from cells stained with the indicated proteins were calculated from actively phagocytosing amoebic cells. Relative intensities were calculated for each marker by NIS-Elements AR 3.0. This analysis was carried out with 30 randomly selected cells. (n ϭ 30, bar represents standard error.) C, quantitative real-time PCR analysis of relative transcript levels of Ehrrp6, Ehcabp3, Ehrho1, and Ehc2pk genes in the indicated cell lines. Ehcabp1 was used as internal control. (Group data represent mean Ϯ S.D.)

Exonuclease EhRrp6 in growth stress
exosome channel (64,65). The homologue of mtr4 is present in E. histolytica and it may be recruited to the exosome by a different mechanism independent of Rrp47.
In yeast the Rrp6 C-terminal domain (CTD) is required for binding to the core exosome and for degrading poly(A) ϩ rRNA processing products (35). The CTD encompasses amino acid residues 518 -733, which contain two distinct elements. Residues 518 -616 constitute the exosome-associating region, and adopt the structure when associated with Exo9 (23). Residues 634 -733 are disordered and are rich in lysine and arginine, with a calculated pI of 10.3. This C-terminal tail is called lasso as it binds RNA and stimulates RNase activities of Rrp44 and Rrp6 within the exosome. The CTD of EhRrp6 is very short and encompasses only 54 amino acids. Although its calculated pI is basic (8.32), it is not as much enriched in lysine and arginine residues as the yeast and human Rrp6 lasso. The prokaryotic RNase D family also lacks a highly basic lasso (27). Further analysis of Rrp6 from a variety of eukaryotes is required to establish the extent of conservation of highly basic lasso in eukaryotic Rrp6 (23).
Human Rrp6 has been shown to more efficiently degrade generic RNA substrate, with secondary structure, beyond the stem-loop compared with yeast Rrp6. This is attributed to the structure of the yeast catalytic domain in which the active site is located in a deep cleft. In comparison, the structure of the human catalytic domain shows the active site to be more solvent exposed, which could allow access to the 3Ј-end of structured RNA (28). This structural difference in Rrp6 is due to differences in the length of linker that connects the EXO and HRDC domains (28,29). In yeast the linker is 26 residues, but in humans, the linker is only 10 residues, and in T. brucei it is 12 residues (38). The shorter linker results in a more solventexposed Rrp6 active site in human and T. brucei Rrp6. Our comparative modeling analysis showed the linker length in EhRrp6 to be 9 residues and the enzyme is predicted to have a more accessible active site. Accordingly, EhRrp6 could efficiently degrade a generic RNA substrate with secondary structure.
Under normal growth conditions EhRrp6 was located both in the nucleus and cytosol. Interestingly, we show that EhRrp6 is almost completely lost from the nucleus when cells are subjected to growth stress by serum starvation. Conversely, downregulation of this protein resulted in severe growth stress (Fig.  5). It is possible that this protein could be a stress sensor in E. histolytica and its nuclear loss may trigger cellular reprogramming in response to stress. A similar observation has been reported in budding yeast cells in response to changes in nutritional status, like nitrogen starvation in the presence of a nonfermentable carbon source, which induces the cells to undergo meiosis and sporulation. The Rrp6 protein, which is stable during mitotic growth, declines progressively as the developmental program shifts from mitotic growth (fermentation) to respiration and sporulation (66). Although serum starvation is not known to induce E. histolytica cells to differentiate in culture, this nutritional limitation stops cell division and possibly triggers pathways that prepare the cell to survive in a maintenance mode. The loss of Rrp6 in sporulating yeast cells leads to elevated levels of ncRNAs like MUTs, CUTs, and rsSUTs which are direct targets of Rrp6 during vegetative growth. It is thought that Rrp6 negatively regulates meiotic development by maintaining these ncRNAs at low levels. It is possible that EhRrp6 is also involved in degradation of specific ncRNAs, which needs to be explored. The decline in EhRrp6 protein levels in stressed cells could be reversed with proteasomal inhibitors, suggesting that the loss of EhRrp6 may be through proteasomal degradation. It is not known whether the same happens in sporulating yeast as well.
The down-regulation of EhRrp6 resulted in a severe growth phenotype, showing that this protein is essential for E. histolytica cell proliferation. The requirement of Rrp6 in proper mitotic cell division has been demonstrated in Drosophila as well. It was shown that dRrp6 is needed for mitosis (in exosome-independent manner), and that down-regulation of dRrp6 led to chromosome segregation defects and reduction in the number of dividing cells (67). dRrp6 depletion did not induce apoptosis but caused a loss of cell proliferation. In our study also we did not find any evidence of apoptosis in cells depleted of EhRrp6. The mechanism by which Rrp6 depletion leads to mitotic defect is not yet understood. Its depletion could stabilize specific mRNAs or small regulatory RNAs with roles in mitosis (67). Interestingly, Rrp6 could also be involved in physical stabilization of mitotic structures (68,69).
One of the determinants of E. histolytica pathogenesis is its ability to phagocytose target cells. The presence of ingested erythrocytes is a hallmark of virulent strains (70). Depletion of EhRrp6 resulted in marked reduction in erythrophagocytosis, which was not due to general down-regulation of genes known to be essential for E. histolytica phagocytosis (41, 49, 54 -57). Of the six genes tested, the expression of two of them (Ehcabp3 and EhRho1) came down (Fig. 9). Both the mRNA and protein levels were reduced. Because both proteins have been shown to modulate actin dynamics it is not surprising that reduction in their levels led to reduced phagocytosis (54,55). How EhRrp6 specifically targets selected genes is intriguing and needs to be further investigated.
In conclusion, we have biochemically characterized EhRrp6, a 3Ј-5Ј exoribonuclease that is lost from the nucleus during growth stress in E. histolytica. Its down-regulation adversely affected amoebic growth and phagocytosis, a process required for E. histolytica pathogenesis. This is the first report of subcellular changes in Rrp6 levels in a parasite system responding to growth stress. This important regulatory system may well mediate parasite response to the host environment and in pathogenesis.

Ethics statement
Mice used for generation of antibodies were approved by the Institutional Animal Ethics Committee (IAEC), Jawaharlal Nehru University (IAEC code number 19/2013), New Delhi, India. All animal experimentations were performed according to the National Regulatory Guidelines issued by CPSEA (Committee for the Purpose of Supervision of Experiments on Animals), Ministry of Environment and Forests, Government of India.

Cell culture, maintenance, and stable transfection of E. histolytica trophozoites
Trophozoites of E. histolytica strain HM-1:IMSS and all transformed strains were maintained and grown in TY1-S-33 medium supplemented with 125 l of 250 units ml Ϫ1 of penicillin G (potassium salt from Sigma) and 0.25 mg ml Ϫ1 of streptomycin per 100 ml of medium as described before (71). For serum starvation E. histolytica trophozoites growing for 48 h in medium containing 15% adult bovine serum were transferred to low-serum (0.5%) medium. To make sense and antisense EhRrp6 expressing cell lines, E. histolytica was transfected by electroporation (42). Drug selection was initiated after 2 days of transfection in the presence of 10 g ml Ϫ1 of hygromycin B (for tetracycline inducible vector) (44). For induction, tetracycline (30 g ml Ϫ1 ) was added 6 to 12 h after subculture to the medium for the indicated time periods. Cells were harvested after 48 h for Western blotting.

Comparative sequence analysis, comparative modeling, and molecular dynamics simulation of EhRrp6
To identify the Rrp6 protein in E. histolytica, we took the Rrp6 protein sequences from H. sapiens (NP_001001998), S. cerevisiae (NP_014643), and T. brucei (XP_844313) and performed searches using NCBI protein BLAST (two methods: blastp and PSI-BLAST with default parameters) against the nonredundant (nr) protein database of E. histolytica HM-1: IMSS-A (taxid:885318) (72). Two matrices, BLOSSUM62 and PAM250, and word size of 2, 3, and 6 were used to search for suitable proteins. The threshold value for cutoff was set to 0.005 and all other parameters were kept as default. Multiple sequence alignment and phylogenetic tree construction methods are explained under supporting Methods SM1.1. EhRrp6 structure (EXO and HRDC domain, amino acid sequence range 183 to 463) was modeled using two templates in Modeler (version 9.14) (73). Two Rrp6 proteins, one from H. sapiens (PDB ID 3SAF) and the other from S. cerevisiae (PDB ID 2HBK) having sequence identity 44 (query coverage 99%) and 45% (query coverage 90%), respectively, were selected as templates for modeling after performing protein BLAST against Protein Data Bank PDB through NCBI BLAST web interface with default settings (72,74). EhRrp6 best modeled structure was selected on the basis of DOPE score and qualitatively evaluated using the ProSA web server (https://prosa.services.came.sbg.ac.at/prosa. php) 4 (75) and PROCHECK (76). The two magnesium ions (Mg 2ϩ ) in the catalytic pocket of EhRrp6 were placed at the same position where manganese (Mn 2ϩ ) and magnesium (Mg 2ϩ ) are located in PDB codes 2HBK and 3SAF, respectively, through structural alignment. Molecular dynamics (MD) simulation of native and mutant EhRrp6 forms is described under supporting Methods SM1.2. Graphs were plotted using Python Matplotlib library (77). UCSF chimera (1.10.2) was used for structural comparison and visualization (78).

Cloning, protein expression, and purification
The E. histolytica rrp6 gene (EHI_021400) was PCR-amplified from genomic DNA using Phusion DNA polymerase. Digested PCR products were cloned in pET-30(b) vector. Point mutations of Ehrrp6 were introduced by site-directed mutagenesis. The presence of mutations was confirmed by nucleotide sequencing. Table S1 shows the list of primers used to generate the WT and mutant constructs. E. coli Shuffle (DE3) (Novagen) was transformed with the pET30b-Ehrrp6 constructs. For protein expression, 1% of overnight grown cultures of a single colony was added to 2 liters of Luria-Bertani (LB) broth (Sigma) containing kanamycin (50 g/ml). Cells were grown to A 600 of 0.6 at 37°C. The cell culture was induced with 0.25 mM isopropyl ␤-D-thiogalactopyranoside and shaken overnight at 18°C. The cells were harvested at 5,000 rpm for 20 min and stored at Ϫ80°C for later use. The first step of purification was carried out using Ni-NTA (Qiagen, Germany) affinity chromatography. The cell pellets stored at Ϫ80°C were thawed and mixed with lysis buffer (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10 mM imidazole, 5% glycerol, 1 mM ␤-mercaptoethanol (␤ME), 1 mM phenylmethylsulfonyl fluoride). After the addition of lysozyme (0.15 mg/ml), the cell suspension was incubated at 4°C for 30 min. The cell suspension was sonicated (QSonica Ultrasonic Systems) in an ice-water mixture at 25% of the amplitude with a pulse (three to four cycles) of 30 s each interspersed with a 1-min interval. Cell lysate was treated with 0.1% Triton X-100, followed by incubation for 30 min on a rotating rocker (Nulife) at 4°C. The lysate was then centrifuged at 15,000 rpm for 30 min at 4°C. The supernatant was passed through a Ni-NTA column (Qiagen) equilibrated with buffer (20 mM Tris-HCl, pH 8.0, 300 mM NaCl, 1 mM ␤ME, and 10 mM imidazole). After washing (washing buffer: 20 mM Tris-HCl, pH 8.0, 300 mM NaCl, 50 mM KCl, 2 mM ATP, 10 mM MgSO 4 , 1 mM ␤ME) with a gradient of imidazole from 25 to 150 mM, protein was eluted with buffer (20 mM Tris-HCl, 150 mM NaCl, 1 mM ␤ME, 5% glycerol, and 250 mM imidazole). The concentrated protein was loaded on a gel permeation chromatography Superdex 200 10/300 GL column (GE Healthcare), which was previously equilibrated with buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5% glycerol, and 1 mM ␤ME). The protein fractions were pooled and dialyzed against (20 mM Tris-HCl, pH 8.0, 50 mM NaCl, 5% glycerol, and 1 mM ␤ME) then concentrated using an Amicon Ultra-15 Centrifugal Filter Unit with Ultracel-30 membrane (Millipore) to ϳ2.5 mg/ml. The protein concentration was determined by Bradford method and stored at Ϫ80°C. For tetracycline-inducible expression of Ehrrp6 in E. histolytica, CAT gene of the shuttle vector pEhHYG-tet-O-CAT (44) was excised using KpnI and BamHI and the Ehrrp6 gene was inserted in its place in either the sense or the antisense orientation.

EhRrp6 exonuclease assays
To check the 3Ј-5Ј exoribonuclease activity of purified EhRrp6, 5Ј 32 P-labeled RNA was taken as substrate. The 50-nt AU-rich RNA and 60-nt generic RNA had the sequences: 5Ј-GAAUUAUUUAUUAUUUAUUUAUUAUUUAUUUAUUU-AUUAUUUAUUUAUUA-3Ј and 5Ј-GAGCUAGGAAGAAU-AGAUGAAAAAUCUAUUAAUAUAUAAUUAAUUACUU-UUUUUUUUUUU-3Ј. RNAs were heated at 95°C for 2 min and cooled to room temperature prior to determining activities. To label the in vitro-transcribed RNA at its 5Ј end, it was dephosphorylated using antarctic phosphatase, followed by rephosphorylation catalyzed by T4 polynucleotide kinase in the presence of [␥-32 P]ATP. Exoribonuclease activities were performed in a 10-l reaction mixture of 10 mM Tris-HCl (pH 8.0), 10 mM DTT, 50 mM KCl, 5 mM MgCl 2 , 1 unit/l of RNase inhibitor (New England Biolabs), 10 M RNA, and 1 M protein at 37°C. Single-point assays of mutant EhRrp6 were conducted for 120 min. Reactions were quenched by the addition of 10 l of loading buffer (95% formamide, 20 mM EDTA, 1% DNA loading dye) and snap chilled in ice-water. Samples were loaded onto a 15% polyacrylamide, 8 M urea gel for electrophoresis. Gels were imaged using a Fuji FLA-5000 scanner. For DNA substrates the following sequence (40) was used as ssDNA and was annealed with its homologue to obtain ds DNA, which was checked on native gel: SINE-1ssDNA, 5Ј-CCCCTGAGCTAGGAAG-AATAGATGAAAAATCTATTAATACTTAATTAATTACT-TTTTTCTTTTTA-3Ј.

Yeast complementation
Ehrrp6 cDNAs were cloned into pRS 426-GPD and transformed into WT (BY4742) and rrp6⌬ (Dharmacon 11777) yeast strains using the lithium acetate method (79). Following selection at 30°C, the cells were tested for growth at 30 and 37°C.

Immunofluorescence staining
Immunofluorescence staining of E. histolytica cells was performed as described before (41). In brief, E. histolytica cells were collected by centrifugation and washed before re-suspending in TYI-S-33 medium. The cells were transferred onto acetone-cleaned coverslips placed in a Petri dish and allowed to adhere for 10 min at 35.5°C. The culture medium was removed and cells were fixed with 3.7% pre-warmed paraformaldehyde for 30 min. Cells were permeabilized with 0.1% Triton X-100/ PBS for 3 min. The fixed cells were then washed with PBS and quenched for 30 min in PBS containing 50 mM NH 4 Cl. The coverslips were blocked with 1% BSA/PBS for 30 min, followed by incubation with primary antibody at 37°C for 1 h. The coverslips were washed with PBS followed by 1% BSA/PBS before incubation with secondary antibody for 30 min at 37°C. Antibody dilutions used were: anti-EhRrp6/anti-EharpC2/anti-Eh-Rho1 at 1:100, anti-EhCabp1/anti-EhCabp3/anti-EhC2pk/anti-EhCabp6 at 1:200, and anti-r-EhCabp6 at 1:300, anti-rabbit/ mice Alexa 488, and Pacific blue-410 (Molecular Probes) at 1:250, TRITC-Phalloidin at 1:250. The preparations were further washed with 1ϫ PBS and stained with Hoechst (20 g/ml) for 10 min at 37°C. The cells were washed thoroughly with 1ϫ PBS and mounted on a glass slide using DABCO (1,4diazbicyclo(2,2,2)octane (Sigma), 10 mg/ml in 80% glycerol). The edges of the coverslip were sealed with nail polish to avoid drying. Confocal images were visualized using a Nikon Real Time Laser Scanning Confocal Microscope (Model A1R). The raw images were processed using NIS element 3.20 (Nikon) that included merging of Alexa Fluor 488, Hoechst, and differential interference contrast (DIC) channels, acquisition of intensity profile, determination of intensity at the region of interest.

RNA isolation and Northern hybridization
Cells were removed at different time points. Total RNA from ϳ5 ϫ 10 6 cells was purified using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. For Northern blot analysis 10 g of total RNA was resolved on a 1.2% formaldehyde-agarose gel in gel running buffer (0.1 M MOPS (pH 7.0), 40 mM sodium acetate, 5 mM EDTA (pH 8.0)) and 37% formaldehyde at 4 V/cm. The RNA was transferred on to Gene-Screen plus R membrane (PerkinElmer). [␣-32 P]dATP-labeled probe was prepared by random priming method using the DecaLabel DNA labeling kit (Thermo Scientific). Hybridization and washing conditions for RNA blots were as per the manufacturers instructions.

Total cell lysate preparation
E. histolytica trophozoites were harvested at 280 ϫ g for 7 min/4°C. The pellet was washed with cold PBS (pH 7.2), resuspended in 50 mM Tris-Cl (pH 7.0), 5% glycerol, 2% SDS and 1ϫ protease inhibitor mixture (Sigma) and kept at 95°C for 5 min. The sample was centrifuged at 13,000 ϫ g for 5 min. The supernatant was collected and quantification was done by bicinchoninic acid (BCA). Yeast cell lysate was prepared according to Shirai et al. (80). Cells were harvested, 0.7 N NaOH solution was added, and sample was incubated for 3 min at room temperature. The cells were centrifuged and SDS-PAGE sample buffer was added. The sample was heated at 95°C for 10 min and the supernatant obtained after centrifugation was loaded.

Subcellular fractionation
Separation of nuclear and cytosolic fractions was essentially done as described (81). Briefly, ϳ10 7 cells growing in log phase were harvested at 280 ϫ g for 7 min at 4°C and the cell pellet was washed with PBS number 8. The washed pellet was resuspended in 2 ml of lysis buffer (10 mM HEPES, pH 7.5, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM DTT, 0.2% Nonidet P-40 detergent and protease inhibitor mixture) and incubated on ice, 15 min followed by centrifugation at 3000 ϫ g for 10 min at 4°C. The pellet contained the nuclear fraction and the supernatant contained the cytoplasmic fraction. Nuclear integrity was checked by microscopy. Nuclear pellet was resuspended in 50 l of lysis buffer added with 0.5% Triton X-100 and protein content of each fraction was estimated by BCA assay.

Western blotting
Samples were separated on 15% SDS-PAGE and the gel was transferred on to a polyvinylidine fluoride membrane by wet transfer method and processed using standard protocols. The antigens were detected with polyclonal anti-EhRrp6 (1Ϻ2,000), EhCoactosin (1:5,000) (raised in mice) or anti-EhCabp1, EhCabp6 raised in rabbits (1Ϻ5,000, raised in our laboratory), and anti-␣-tubulin antibody (Sigma number T6199) followed by secondary anti-rabbit or anti-mouse Igs conjugated to HRPO (horseradish peroxidase) at 1Ϻ10,000 dilution (Sigma, A6667 or A2554). ECL reagents were used for visualization (Millipore). Protein concentration was estimated by BCA assay Exonuclease EhRrp6 in growth stress using BSA as a standard. Quantification of band intensity in each lane was done by densitometry using ImageJ software and values were normalized using respective internal control.

RNA-Seq transcriptome sequencing
Total RNA was purified from exponentially growing E. histolytica cells (and 24-h serum-starved cells) using TRIzol reagent (Invitrogen) and used for selection of poly(A) plus RNA and library preparation was done after oligo(dT) selection. RNA-Seq libraries were generated by performing RNA fragmentation, random hexamer primed cDNA synthesis, linker ligation, and PCR enrichment. These libraries were then sequenced on the Illumina HiSeq 2500 (version 3 Chemistry) platform. From paired-end reads low quality sequences were removed including non-poly(A)-tailed RNAs using bowtie2 (version 2.2.2) and in-house Perl scripts. About 35 million reads were obtained and on an average, ϳ90.13% of total reads passed Ͼ ϭ 30 Phred score. The reads were aligned to the E. histolytica (HM1:IMSS) total genome, downloaded from AmoebaDB, using RSEM version 1.2.31 with default parameters and commands: "rsem-prepare-reference" and "rsem-calculate-expression." The data were obtained for two biological replicates.

Erythrophagocytosis assay
The assay was performed essentially as described (41). E. histolytica trophozoites were harvested in 1ϫ PBS (pH 7.2) and equal numbers (10 5 cells) were incubated with 10 7 RBCs for the indicated times, at 37°C in 0.5 ml of culture medium. The amoebae and erythrocytes were centrifuged and cold distilled water was added to lyse the nonengulfed RBCs and re-centrifuged at 1,000 ϫ g for 2 min. This step was repeated twice, followed by re-suspension in 1 ml of formic acid to lyse Entamoeba cells containing engulfed RBCs. The absorbance was measured at 400 nm with formic acid as blank.

Cell proliferation and cell migration assay
Growth kinetics was studied by inoculating an equal number of cells (3 ϫ 10 4 cells/ml). The cells were allowed to grow at 35.5°C in 5-or 7-ml tubes and harvested at 12, 24, 36, 48, and 72 h post-inoculation. The cells were harvested in 1ϫ PBS (pH 7.2) by chilling on ice followed by centrifugation at 1,500 ϫ g for 5 min. The cell pellet was resuspended in 1 ml of 1 ϫ PBS (pH 7.2). The cells were counted with hemocytometer by mixing cells with 0.4% trypan blue in a ratio of 1:1. At every time point, the cells were harvested and counted as mentioned above. For motility assay, experiments were carried out as described before (51). 1.5 ϫ 10 5 harvested amoebic cells were added to top chamber of a transwell containing 8 M pore size (Costar, USA) in incomplete TYI medium. TYI medium with 15% serum was placed in the lower chamber. A 24-well plate was sealed with parafilm and placed in an anaerobic bag for 2 h at 35.5°C. Parasites migrated in the bottom chamber were chilled, transferred to a microcentrifuge tube, centrifuged at 1,000 ϫ g for 5 min, resuspended in TYI medium, and quantified with a hemocytometer. Each experiment was performed in triplicate. The standard error bars were calculated and represented.

Statistical analysis
Statistical comparisons were made using analysis of variance test. Experimental values were reported as the mean Ϯ S.D. Differences in mean values were considered significant at "one black star" p value Յ0.05, "two black star" p value Յ0.01, and "three black star" p value Յ0.001. All calculations of statistical significance were made using the GraphPad InStat software package (GraphPad Prism 7).