EhPgp5 mRNA Stability Is a Regulatory Event in theEntamoeba histolytica Multidrug Resistance Phenotype*

The multidrug resistance (MDR) phenotype inEntamoeba histolytica is characterized by the overexpression of the EhPgp5 gene in trophozoites grown in high drug concentrations. Here we evaluated the role ofEhPgp5 mRNA stability on MDR using actinomycin D.EhPgp5 mRNA from trophozoites growing without emetine had a half-life of 2.1 h, which augmented to 3.1 h in cells cultured with 90 μm and to 7.8 h with 225 μm emetine. Polyadenylation sites were detected at 118-, 156-, and 189-nucleotide (nt) positions of the EhPgp5mRNA 3′-untranslated region. Interestingly, trophozoites grown with 225 μm emetine exhibited an extra polyadenylation site at 19 nt. The 3′-untranslated region sequence is AU-rich and has putative consensus sequences for RNA-binding proteins. We detected a RNA-protein complex in a region that contains a polypyrimidine tract (142–159 nt) and a cytoplasmic polyadenylation element (146–154 nt). A longer poly(A) tail in the EhPgp5 mRNA was seen in trophozoites grown with 225 μm emetine. Emetine stress may affect factors involved in mRNA turnover, including polyadenylation/deadenylation proteins, which could induce changes in the EhPgp5 mRNA half-life and poly(A) tail length. Novel evidence on mechanisms participating in E. histolytica MDR phenotype is provided.

Entamoeba histolytica, the protozoan parasite responsible for human amoebiasis, presents the multidrug resistance (MDR) 1 phenotype (1) described first in mammalian cells (2) and then in several protozoan parasites (3,4). MDR is associated with the overexpression of a 170-kDa membrane molecule known as P-glycoprotein (PGP), an energy-dependent pump that extrudes drugs from the cells (5,6). In E. histolytica, MDR phenotype is given mainly by overexpression of the EhPgp1 and EhPgp5 genes, which are finely regulated by transcriptional factors (7)(8)(9). Although EhPgp1 is constitutively expressed in drug-resistant trophozoites of clone C2, EhPgp5 gene is overexpressed only when C2 cells are grown in a high emetine concentration (10,11). Both genes are also amplified in the presence of a high drug concentration (12).
Transcriptional regulation of eukaryotic mdr genes has been considered as the major control point for PGP synthesis, although gene amplification mechanisms also participate in this event (12,13). Moreover, there is growing evidence of pivotal post-transcriptional (14 -17) and post-translational (18 -20) regulation of the PGP expression. On the other hand, mRNA stability has recently emerged as a critical control step in determining cellular stationary mRNA levels. The abundance of a particular mRNA can fluctuate many folds due to alterations in mRNA stability without any change in the transcription rate (21). The mRNA half-life is determined by a complex set of protein interactions at the 3Ј-untranslated region (3Ј-UTR) depending on conserved cis-element sequences and secondary structures (for review, see Ref. 22). The 3Ј-UTR also contains consensus sequence elements that mediate mRNA nuclear export, cytoplasmic localization, translation efficacy, and polyadenylation control (23,24). The pre-mRNAs are polyadenylated in a reaction involving 3Ј endonucleolytic cleavage followed by poly(A) tail synthesis (25). Poly(A) tail is also a modulator of mRNA stability and translation (26,27). Strict control of poly(A) tail length is achieved by the concerted interplay of key factors, including poly(A) polymerase, deadenylases, and poly(A)-binding protein activities (25).
Several reports have addressed the importance of mRNA stability on the mdr genes expression regulation. Pgp1, Pgp2, and Pgp3 mRNAs have a higher half-life in rat tumor cells than in normal cells (15), whereas rat MDR hepatocytes in culture present a higher amount of PGP2 protein due to a post-transcriptional mechanism controlling mRNA stability (14). Human MDR1 mRNA has a half-life of 30 min, which is prolonged to more than 20 h upon treatment with cycloheximide, suggesting that protein synthesis inhibition may influence the stability of certain mRNAs (16,17). However, molecular mechanisms controlling mdr mRNA stability remains to be elucidated.
In E. histolytica, mRNA stability mechanisms have not been studied yet. The presence of higher levels of EhPGP5 protein in the multidrug-resistant trophozoites of clone C2 could be influenced by both transcriptional activation and increased mRNA stability. In this paper, we measured the EhPgp5 mRNA halflife in trophozoites of clone C2 grown at different emetine concentrations. Our data showed that EhPgp5 mRNA stability is increased at high emetine concentrations, indicating that mRNA half-life is also regulating the MDR phenotype. In addition, here we initiated the study of the mechanisms involved in mRNA turnover in this parasite.
Transcriptional Inhibition by Actinomycin D-Actinomycin D (Roche Molecular Biochemicals) dissolved in dimethyl sulfoxide (Me 2 SO) (0.5 mg/ml) was added to the trophozoites cultures to a final concentration of 10 g/ml of medium, and cells were incubated at 37°C for different times. Fresh medium supplemented with [ 3 H]UTP (10 Ci/ml) was added to the actinomycin D-treated trophozoites for 2 more hours in the absence of actinomycin D. Immediately, total RNA was isolated by TRIzol (Invitrogen). Incorporation of [ 3 H]UTP in 20 g of total RNA was measured by liquid scintillation counting system (Beckman) in duplicate samples, and data obtained were plotted. Cytotoxicity of actinomycin D and Me 2 SO was checked by cell viability using trypan blue and measuring the growth rate of the treated cultures.
Reverse Transcriptase (RT)-PCR Experiments-100 ng of total RNA from trophozoites of clones A, C2, C2(90), and C2(225) were preincubated at 37°C for 15 min with 10 units of RNase-free DNase I (Stratagene). Single-stranded cDNAs were synthesized using 10 mM each dNTP and 100 ng of oligo(dT 18 ) in diethyl pyrocarbonate-treated water. The mixture was heat-denatured at 65°C for 5 min and quick-chilled on ice. Then we added buffer used to synthesize the first-strand (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl 2 ), 100 mM dithiothreitol, 200 units of Superscript TM II RNase H Ϫ reverse transcriptase (Invitrogen), and 40 units of SUPERase-in ribonuclease inhibitor (Ambion). This mixture was incubated at 42°C for 1 h. To remove the excess RNA template, 2 units of RNase H (Amersham Biosciences) were added, and the mixture was incubated at 37°C for 15 min. Quantitative multiplex PCR for EhPgp5 and actin cDNAs were performed with 1 ⁄5 volume of the reverse transcription mixture, 10 mM each dNTPs, 5 mM MgCl 2 , 2.5 units of Taq DNA polymerase (Invitrogen), and the EhPgp5 (5Ј-GTAG-GAGGTGCAGTATTTCC-3Ј) sense and (5Ј-CCATCCTATTTCTTGTTT-GAC-3Ј) antisense internal primers (30). The actin (5Ј-AGCTGTTCTT-TCATTATATGC-3Ј) sense and (5Ј-TTCTCTTTCAGCAGTAGTGGT-3Ј) antisense internal primers (31) were used in the same sample as an internal control. PCR was done in 22 cycles at 95°C for 30 s, 52°C for 35 s, 72°C for 90 s, and a final extension step at 72°C for 7 min. Amplified products were separated by 6% PAGE.
EhPgp5 mRNA Stability Assays-Total RNA from trophozoites of clones C2, C2(90), and C2(225) was obtained at 0, 2, 4, 8, and 12 h after actinomycin D-induced transcriptional blockage. EhPgp5 and actin mRNAs were measured by multiplex RT-PCR as described above, and intensity of the bands in ethidium bromide-stained gels was quantified by densitometric analysis in a PhosphorImager apparatus (Personal Molecular Imager FX, Bio-Rad). The pixels given by the actin transcript in trophozoites of clones C2, C2(90), and C2(225) without treatment (t 0 ) were taken as 100% in each clone. EhPgp5 mRNA levels were normalized with respect to the actin amount in each lane. Experimental EhPgp5 mRNA half-life (the time at which 50% of mRNA molecules remained intact) was determined by plotting the EhPgp5 mRNA amount at different times on a semilogarithmic scale. In these estimations the EhPgp5 mRNA amount at t 0 was taken as 100% in each clone. Theoretical half-life of the EhPgp5 mRNA was obtained from the logarithmically transformed best-fit line by linear regression analysis using the decay equation t1 ⁄2 ϭ ln 2/K, where K corresponds to the decay constant (32). S1 Nuclease Mapping Experiments-20 g of total RNA from trophozoites of clones C2, C2(90), and C2(225) were hybridized with a 697 bp of [␣-32 P]dATP (PerkinElmer Life Sciences)-labeled probe corresponding to the last 100 bp of the EhPgp5 coding region and 597 bp of the EhPgp5 3Ј-UTR genomic sequence. The probe was PCR-amplified from the P4 plasmid, which contains the last 1466 bp of the EhPgp5-coding region and ϳ1500 bp of its 3Ј-UTR genomic sequence (7), using the EhPgp5-3Ј-UTR-S (5Ј-AAAATAGTAGAACAAGGA-3Ј) sense and Eh-Pgp5-3Ј-UTR-AS (5Ј-CGAACAAAGGCTTAAA-3Ј) antisense primers. Amplified EhPgp5-3Ј-UTR fragment was sequenced in a ABI PRISM automatic sequencer. Purified DNA probe was heat-denatured at 95°C for 15 min and cooled at 4°C for 3 min. Then, an RNA aliquot was added, and immediately, the DNA/RNA hybrid mixture was co-precipitated with ethanol at Ϫ20°C overnight. A DNA/RNA pellet was collected by centrifugation, air-dried, and resuspended in 20 l of S1 nuclease hybridization buffer (80% deionized formamide, 40 mM MOPS, pH 7, 400 mM NaCl, and 1 mM EDTA). The DNA/RNA hybrids were fully denatured at 85°C for 10 min and subsequently incubated at 42°C overnight. Then, samples were 10-fold diluted and treated with 1250 units of S1 nuclease (Invitrogen) at 37°C for 1 h in the reaction buffer (300 mM sodium acetate, pH 4.6, 1 mM NaCl, and 10 mM zinc acetate). At the same time, we performed a sequencing reaction of the EhPgp5 3Ј-UTR using the EhPgp5-3Ј-UTR-S primer. The sequencing products and RNA fragments protected of the S1 nuclease digestion were resolved through denaturing 8% PAGE at room temperature, vacuum-dried, and visualized in a PhosphorImager apparatus.
Analysis of Poly(A) Tail Length-Total RNA was used for ligasemediated poly(A) test (LM-PAT) according to the method described (35).

RESULTS
Actinomycin D Inhibits E. histolytica Transcription-To study the mechanisms controlling mRNA decay in E. histolytica we first investigated the effect of the transcription inhibitor actinomycin D on viability, growth, and mRNA synthesis in trophozoites of the drug-sensitive clone A and drug-resistant clone C2. Cell viability of trophozoites of clone C2 incubated 12 h with or without actinomycin D was 98%. At this time cell growth was slightly delayed in the actinomycin D-treated trophozoites in comparison with untreated cells (Fig. 1A). The effect of actinomycin D on cell growth and viability was similar in trophozoites of all clones tested (data not shown). Then we performed experiments to determine whether mRNA synthesis was affected by actinomycin D and the time required to inhibit at least 90% of mRNA synthesis. Results showed that actinomycin D affects the [ 3 H]UTP incorporation into new synthesized RNA in a time-dependent manner (Fig. 1B). Untreated trophozoites of clones A and C2 incubated for 2 h (t 0 ) with [ 3 H]UTP incorporated 16,450 and 14,660 cpm, respectively. Each value was taken as 100% incorporation for the corresponding clone (Fig. 1B). Trophozoites of clones A and C2 preincubated for 30 min with the drug and then incubated with [ 3 H]UTP for 2 h incorporated 50.4 and 59.9% of radioactivity, respectively. One hour later, both clones presented only 12% incorporation of [ 3 H]UTP (Fig. 1B). These low levels of RNA synthesis were maintained in trophozoites of both clones incubated with actinomycin D for up to 8 h (Fig. 1B). As a control for mRNA integrity, equivalent amounts of total RNA were isolated at each time, and rRNA were visualized in ethidium bromide-stained agarose gels (Fig. 1C). No appreciable changes in the rRNA amount were observed, indicating no significant RNA degradation. These data provided us a temporal window in which the mRNA decay could be analyzed.
EhPgp5 mRNA Stability Is Higher in Trophozoites of the Clone C2 Cultured with Emetine-To investigate the EhPgp5 mRNA stability in trophozoites growing with different emetine concentrations, we determined the EhPgp5 mRNA half-life in actinomycin D-treated trophozoites of clones C2, C2(90), and C2(225). EhPgp5 mRNA was measured from 0 to 12 h by semiquantitative RT-PCR assays in total RNA. We included actin primers in all reactions as internal control. Results showed that EhPgp5 mRNA was present in untreated trophozoites of clone C2, and the signal diminished progressively at 2 and 4 h after the transcriptional blockage ( Fig. 2A). In clone C2(90), the EhPgp5 transcript was detected up to 8 h after actinomycin D treatment (Fig. 2B). Interestingly, in clone C2(225) the EhPgp5 mRNA was detected even 12 h after the transcriptional blockage (Fig. 2C). These data indicated that EhPgp5 mRNA amounts are reduced in a time-dependent manner in actinomycin D-treated trophozoites, but they are maintained for a longer time in the emetine-cultured cells. In contrast, we did not detect the EhPgp5 mRNA in untreated trophozoites of the wild type drug-sensitive clone A (Fig. 2G), confirming that EhPgp5 gene is not transcribed in the drugsensitive trophozoites (36).
The EhPgp5 and actin mRNAs were quantified by densitometry and arbitrary expressed in pixels (Fig. 2, D-F). Pixels given by actin at t 0 were taken as 100% in each clone, and the EhPgp5 percentage was expressed with respect to the actin mRNA levels. At t 0 , the actin amount appeared almost unaltered, whereas EhPgp5 varied in the three clones. In trophozoites of clone C2, the EhPgp5 mRNA was 63% of the actin mRNA (Fig. 2D), and in C2(90) it was 81% (Fig. 2E), whereas clone C2(225) exhibited similar levels of both transcripts (Fig.  2F). The other bands seen in the gels were not related to the EhPgp5 transcript, as probed by Southern blot hybridization using the EhPgp5 probe (data not shown).
To determine the experimental mRNA half-life, the results of normalized EhPgp5 mRNA levels were plotted in a semilogarithmic scale against the exposure time to actinomycin D (Fig.  2H). In these calculations, the amount of EhPgp5 and actin mRNA at t 0 was taken as 100% in each clone. In trophozoites of clone C2, the experimental EhPgp5 mRNA half-life was estimated in 2.1 h, whereas in C2(90) it was 3.1 h and 7.8 h in C2(225), confirming significant variations in the decay rates of the three clones. In addition, these experiments showed that actin mRNA decay remained with minimal changes during the 12-h course of the transcription inhibition in all clones. There are reports proposing that actin mRNA has a half-life between 24 and 33 h in mammalian cells (37). According to our experiments, E. histolytica actin mRNA has a half-life longer than 12 h.
Experimental values were close to the theoretical EhPgp5 mRNA half-life predicted from the decay equation described under "Experimental Procedures." The theoretical EhPgp5 mRNA half-life was 1.2 h in the trophozoites of clone C2, 2.7 h in C2(90), and increased to 5.6 h in C2(225) ( Table I).
A Specific RNA-Protein Complex Was Detected at the EhPg5 3Ј-UTR mRNA-Stability of mRNA is also regulated through site-specific binding of cytoplasmic proteins to consensus sequences at the 3Ј-UTR mRNA. The presence of putative Eh-Pgp5 mRNA binding regulatory proteins was investigated by RNA electrophoretic mobility shift assays using the PSI 19 , PSII 118 PSIII 156 and PSIV 189 fragments as RNA probes (Fig.  5A). The results showed that PSIII 156 and PSIV 189 transcripts were able to form a RNA-protein complex with CE from C2, C2(90), and C2(225) clones (Fig. 5, B-D, lanes 2 and 3). RNAprotein complex was specifically competed by a 350-fold molar excess of the same unlabeled transcript, but they were maintained in the presence of tRNA, used as nonspecific competitor (Fig. 5, E and F, lanes 3 and 4). The intensity and migration of the RNA-protein complex was similar in all clones (Fig. 5, B-F). In contrast, we did not find any RNA-protein complex when we used PSI 19 and PSII 118 fragments (Fig. 5, B-D, lanes 4 and 5).
To delimitate the region in which the RNA-protein complex was formed, we carried out cross-competition experiments using the PSIV 189 transcript as labeled probe and the PSI 19 , PSII 118 , and PSIII 156 RNAs as competitors. In the three clones the complex formed in the PSIV 189 region was specifically competed by the same probe and by PSIII 156 (Fig. 6, A-C, lanes 5 and 6) but not by PSI 19 and PSII 118 RNA fragments, as expected (Fig. 6, A-C, lanes 3 and 4). These results suggest that RNA-protein interaction takes place in a region of 38 nt (nt 119 -156), which is shared by PSIII 156 and PSIV 189 fragments. This region contains a PS (nt 127-131) and a 15-nt Py (nt 142-156) sequences, including a CPE (nt 146 -154) motif, which could be targets for regulatory RNA-binding proteins (43)(44)(45)(46).
The EhPgp5 mRNA Poly(A) Tail Is Longer in C2(225) Trophozoites-Poly(A) tail is an important modulator of mRNA turnover, and its length is subjected to cellular control throughout the life span of the mRNA (25). We investigated the poly(A) tail length of EhPgp5 mRNA from trophozoites of clones C2, C2(90), and C2(225) by LM-PAT (35), as described under "Experimental Procedures." We observed three well defined bands corresponding to the PSII 118 , PSIII 156 , and PSIV 189 predicted transcripts plus 100 bp of the EhPgp5 open reading frame, respectively (Fig. 7A). In these assays, we could not amplify the PSI 19 variant, probably because EhPgp5 PSI 19 transcript has a very short poly(A) tail. The identity of the amplified products was confirmed by Southern blot hybridization with a DNA probe containing the last 100 bp of the open reading frame and the first 19 nt of the EhPgp5 mRNA 3Ј-UTR (Fig. 7B). The signal appeared as a smear ranging from 218 to 300 nt in mRNA from trophozoites of clones C2 and C2(90) (Fig. 7B). Interestingly, in clone C2(225) the hybridization signal showed a longer smear spanning from 218 to 500 nt. The same membrane hybridized with the actin 3Ј-UTR probe gave no signal (Fig. 7C). In contrast, in the actin control assays, we detected a defined 130-bp band corresponding to actin 3Ј-UTR (ϳ30 nt) plus 99 bases of the 3Ј-actin open reading frame and a short smear spanning 130 -150 nt in all clones (Fig. 7, D and E). The actin control membrane gave no signal with the EhPgp5 3Ј-UTR probe (Fig. 7F).
These LM-PAT patterns represent an enlargement of the poly(A) tail length of the EhPgp5 mRNA in the trophozoites of clone C2(225) or, alternatively, a shortening in C2 and C2(90) cells, suggesting that changes in poly(A) tail length are involved in EhPgp5 mRNA half-life. DISCUSSION Previously, we demonstrated that the EhPgp5 gene is overexpressed in E. histolytica trophozoites grown in the presence of high drug concentration (30). Transcriptional factors partic- ipate in the EhPgp5 gene promoter activation (11). Results presented here show novel evidence that post-transcriptional EhPgp5 gene regulation occurs in the drug-resistant trophozoites. EhPgp5 mRNA is more stable in trophozoites grown in 225 M emetine than in those grown in 90 M or without drug (Fig.  2). Additionally, the EhPgp5 mRNA 3Ј-UTR length is heterogeneous (Fig. 3A), which may influence the mRNA half-life. The PSIII 156 and PSIV 189 mRNA variants augmented when the emetine dose was increased (Fig. 3, A-E). Their predicted secondary structure suggests that they have exposed a Py tract and a CPE motif (data not shown). Furthermore, a RNA-binding protein complex was detected in their 39-nt shared region (Figs. 5 and 6). In other organisms, polypyrimidine tract-binding proteins have been involved in splicing and stability control of the mRNA, whereas CPE-motif interacting proteins target specific mRNAs to cytoplasmic polyadenylation, producing the translational activation of the transcripts (43)(44)(45)(46). Interestingly, the EhPgp5 mRNA presents a longer poly(A) tail in clone C2(225) (Fig. 7), and it is well known that large poly(A) tails give higher stability to mRNA and promote a more efficient translation (39). Emetine stress could affect the expression of many factors, including the polyadenylation/deadenylation proteins involved in the poly(A) tail length control.
mRNA half-life and translation are linked in ways that are not completely understood. In cells exposed to translation inhibitors some mRNAs are stabilized in several ways, including alterations in polyadenylation rates (22). For example, in mammalian cells, cycloheximide prolongs c-myc mRNA half-life by slowing the deadenylation process but does not promote degradation of the mRNA body once deadenylation is being completed (47). The heterogeneity of the EhPgp5 transcripts is explained by the alternative usage of several polyadenylation signals detected in the 3Ј-UTR (Fig. 4), as has been well documented for other systems, including other mdr genes (40,48). Mouse mdr1a mRNA shows length variations at both 5Ј and 3Ј ends, and mRNA variants have very large 3Ј-UTRs, which are differentially overexpressed in multidrug-resistant cell lines (48). Interestingly, the EhPgp5 mRNA also presents heterogeneity in the 5Ј end of trophozoites of clones C2 and C2(225) (11). All these data indicate that EhPgp5 gene is a complex transcriptional unit whose regulation produces multiple transcript sizes at the 3Ј and 5Ј ends.
Emetine partially inhibits protein synthesis in trophozoites of clone C2(225) (data not shown). This could induce a stabilizing mRNA effect (22,49) affecting certain EhPgp5 transcript variants. The PSIV 189 , PSIII 156 , and PSII 118 mRNA variants were detected in all clones, whereas PSI 19 was observed only in clone C2(225). PSI 19 , PSIII 156 , and PSIV 189 transcripts were more abundant in the trophozoites of clone C2(225). PSI 19 transcript, which has a very short poly(A) tail length or is not polyadenylated, was almost 2-fold the amount of PSIII 156 and PSIV 189 in clone C2(225) (Fig. 3). It is possible that the expression of undetermined factors linked to the emetine effect and whose function could be independent of the poly(A) tail contribute to the PSI 19 transcript stability. However, additional experiments are required to confirm this hypothesis.
AURE motifs have been involved in destabilization of mRNAs with a short half-life (42). However, Prokipcak et al. (16) find that AUREs at the 3Ј-UTR of human MDR1 mRNA is an inefficient promoter of mRNA decay, which suggests that AURE-dependent mRNA stability regulation may not operate in certain cases, such as the MDR1 and EhPgp5 mRNAs. This assumption is supported because under the experimental conditions reported here, we did not detect any RNA-protein complex in the AURE motifs present in the EhPgp5 3Ј-UTR mRNA, suggesting that they do not act as cis-regulatory elements in the EhPgp5 mRNA half-life control.
We found a RNA-protein complex in the proximity of the Py tract in the PSIII 156 and PSIV 189 transcripts (Figs. 5 and 6) that may contribute to their stability in C2(225) cells. However, the same complex was also detected in clones C2 and C2(90), suggesting that other factors present only in clone C2(225) are required to stabilize certain EhPgp5 mRNA variants. The identity of the 3Ј-UTR EhPgp5 mRNA-interacting protein(s) detected here remains to be elucidated. Interestingly, the EhPgp5 mRNAs from trophozoites of clone C2(225) present longer poly(A) tails than those from C2 and C2(90) cells, suggesting that polyadenylation and deadenylation events, occurring at different rates, could be affecting the EhPgp5 mRNA half-life.
Our working hypothesis assumes that trophozoites of clone C2 grown without emetine have some factors that maintain short poly(A) tails, which may contribute to a shorter EhPgp5 mRNA half-life (Fig. 8). Some of these factors could be affected by emetine in C2(225) cells, and emetine-responsive factors could both induce an enhanced polyadenylation of EhPgp5 mRNAs. The expression and activity of some proteins involved in 3Ј to 5Ј exonucleolytic mRNA degradation and polyadenylation may also be participating in the longer EhPgp5 mRNA half-life in C2(225) cells. 2 Hence, the putative role of other 3Ј end processing and polyadenylation/deadenylation factors cannot be discarded.