Trophozoites of clones A and C2 (strain HM1: IMSS) (Orozco et al., 1985) were axenically cultured in TY1-S-33 medium (Diamond et al., 1978). Emetine stress was exerted by the incubation of trophozoites from clones A or C2 in 8 or 20 μM emetine, respectively, for 24 h.

The TRANSFACT database (MatInspector software) was used to identify potential transcription factor binding sites in the EhPgp5 promoter (Matys et al., 2006).

To perform transfection experiments, several plasmids were constructed by inserting PCR-amplified EhPgp5 gene core promoter fragments into multiple cloning sites of the pBSCAT-ACT plasmid (Gómez et al., 1998). The constructs performed are shown in Figures 1A,B. Constructs with specific mutations in the CdxA, YY1, and HSE regions were performed by PCR amplification using the primers shown in Figure 1C. Promoter-less (pBSCAT-ACT) and minimal p259Pgp5 promoter constructs (Pérez et al., 1998) were used as negative and positive controls, respectively. The pA5′A3′CAT plasmid (Nickel and Tannich, 1994), which contains a 480 bp fragment from the actin gene promoter, was used as an internal control of transfection efficiency. Correct sequence orientation of the cloned fragments was determined by sequencing analysis.

Transfection assays were performed by electroporation (Nickel and Tannich, 1994). Briefly, 10⁶ trophozoites of clones A and C2 were transfected with 100 μg of the different plasmid constructs (ΔCdxA-Pgp5, ΔCdxA-YY1-Pgp5, Δ5HSE-Pgp5, pmuCdxAPgp5, pmuYY1cPgp5, pmuYY1gPgp5, pmuHSEPgp5, pmHSE-TTCPgp5, p-sinHSEPgp5, p259Pgp5, pBSCAT-ACT, and pA5′A3′CAT). Electroporated trophozoites were transferred into plastic flasks (Nalgene, NY, USA) containing 50 ml of TYI-S-33 medium and were incubated for 48 h at 37°C. Two hours after electroporation, we added 8 μM emetine to clone A or 20 μM emetine to clone C2. CAT activities were measured by two-phase diffusion assay using 5 μg of trophozoite extracts and 200 μl of chloramphenicol (1.25 mM), which were incubated with [¹⁴C]-butyril-CoA (NEN Life Science Products, MA, USA) for 2 h. Protein concentrations were determined using Bradford's methods (Bradford, 1976). CAT activities were expressed as the cpm of the butyrylated derivatives. The background obtained from the trophozoites transfected with the pBSCAT-ACT plasmid was subtracted from the measurements given by the plasmids containing the different promoter constructions. CAT activities were determined in the linear range of assays, representing three independent experiments performed in triplicate. The efficiency of the transfection experiments was monitored by the activity of pA5′A3′CAT plasmid (Nickel and Tannich, 1994).

NEs were prepared from trophozoites of clones A and C2 grown in the presence [clones A₍₈₎ and C2₍₂₀₎] or absence (clones A and C2) of emetine, following Schreiber et al.'s protocol (Schreiber et al., 1989), modified by Gómez et al. (1998). Protein concentrations were determined by Bradford's methods (Bradford, 1976).

Double-stranded oligonucleotides of the EhPgp5 HSE sequence from −150 to −137 bp (HSE-S-16, 5′-ATAGAAATTTTTCATA-3′) were synthesized and labeled using T4 polynucleotide kinase (Invitrogen, CA, USA) in the presence of [γ-³² P]-ATP. Specific activity was determined by scintillation counting. Gel-shift assays were performed in triplicate as described previously (Gómez et al., 1998). Briefly, we used 0.5–1 ng of labeled probes (10,000 cpm), 1 μg of poly [d (I-C)] (Amersham Pharmacia Biotech, NJ, USA), 20 μg of NE from trophozoites and 10% glycerol in DNA–protein binding buffer. For competition assays, 150-fold molar excess of unlabeled oligonucleotides was incubated with NE at 4°C for 10 min prior to adding the radiolabeled probes. As competitors, we used poly [d (I-C)] and the double-stranded oligonucleotides S-34 (5′-TATAATAACCATTTTGGGTATAGAAATTTTTCAT-3′) and HSEm (5′-ATATCGATTTTTCATA-3′).

UV-crosslinking assays were performed according to Ausubel et al. (1994) with some modifications. Standard gel-shift mixtures using NE from trophozoites of both clones grown in the absence or presence of emetine and the radiolabeled HSE probe were scaled up five-fold and were exposed to irradiation in a 312 nm UV-transilluminator (Bio-Rad, CA, USA) for 10 min at 4°C; then, the proteins were separated by 15 % SDS-PAGE (Laemmli, 1970). The gels were dried and analyzed by autoradiography. DNA cross-linked proteins were analyzed by western blot using the standard procedure (Ausubel et al., 1994). The membranes were incubated with 2 μg of human anti-HSF1 (H-311, Santa Cruz Biotechnology, TX, USA) in 5% non-fat dry milk and 0.05% Tween-20 in phosphate-buffered saline (PBS), with a pH of 7.4, overnight at 4°C. The proteins were revealed by peroxidase-conjugated anti-rabbit secondary antibodies (ZYMED Laboratories, CA, USA) (1:3000) and were immunodetected by a chemiluminescence system (ECL Plus ™, Amersham Pharmacia Biotech, NJ, USA). The assays were repeated at least three times for each studied condition.

E. histolytica proteins that bind to the HSE sites of the EhPgp5 gene core promoter were partially purified three separate times under non-denaturing conditions using a DNA-binding protein purification kit (Roche, CA, USA) and NE from trophozoites of both clones A_[8] and C2_[20]. Concatemeric polynucleotides were prepared by PCR amplification of the HSE probe. The oligomer was coupled to magnetic particles coated with streptavidin, as described by the manufacturer. Seventy-five micrograms of NE from trophozoites were mixed with the magnetic particles in protein buffer, poly [d (I-C)] and poly-L-lysine. Then, NE was incubated at 4°C for 60 min. After extensive washing, proteins bound to the particles were eluted with elution buffer containing 2 M KCl. All of the fractions were analyzed by 10% SDS-PAGE (Laemmli, 1970). The gels were silver stained or transferred to nitrocellulose membranes for western blotting assays. Purified fractions were added to DNA–protein binding reactions to perform EMSA, as described above.

All of the data are expressed as the means ± SDs. IBM SPSS software, version 23, was used to compare all of the groups to each other. For all tests, p < 0.05 was considered significant.

Previously, a functional region of 59 bp (−170 to −111 bp) of the EhPgp5 gene promoter was involved in the induction of EhPgp5 gene expression (Nieto et al., 2005). To detect specific transcription elements within this region, we performed a bioinformatic search for the presence of consensus sequences of different transcription factors within the 59-bp EhPgp5 gene promoter region, using the TransFact database. The structural analysis of this sequence revealed the presence of putative consensus sequences for CdxA (−170 to −164 bp) and YY1 (−163 to −153 bp) transcription factors and, interestingly, a HSE at −151 to −136 bp (Figure 1A).

To identify whether these cis-regulatory elements direct specific EhPgp5 gene expression in E. histolytica drug-sensitive [A₍₈₎] and drug-resistant [C2₍₂₀₎] clones grown in the presence of 8 and 20 μM emetine, respectively, we performed a series of mutations and deletions within the 59 bp region. Different constructs were generated, as described in the Materials and Methods section (Figures 1B,C). These plasmids contain internal deletions of CdxA, YY1, or HSE sequences.

Transfection assays of all constructs were performed in trophozoites from clones A and C2. In transfected trophozoites of both clones grown without emetine, no CAT activities were detected with any constructs, compared to the CAT activity developed by pA5′A3′CAT plasmid (Figure 2). However, when emetine was added to the culture medium after electroporation, we observed significantly higher CAT activities with all the constructs, similar to the activity shown by the EhPgp5 gene core promoter (p259Pgp5 plasmid) (Figure 2), except with the pΔ5-HSEPgp5 plasmid. No significant differences were observed in the CAT activities shown by ΔCdxAPgp5, ΔCdxA-YY1Pgp5, pmuCdxAPgp5, pmuYY1cPgp5, and pmuYY1gPgp5, suggesting that CdxA and YY1 putative consensus sequences were not involved in EhPgp5 gene transcriptional activation. In contrast, the construct pΔ5-HSEPgp5 that does not contain six of 16 bases that form the putative HSE element (ataGAAatttTTCata) displayed significant CAT activity reductions of 80 and 78% in trophozoites from clones A_[8] and C2_[20], respectively.

To demonstrate that the putative HSE (−151 to −136 bp) is responsible for activating EhPgp5 gene expression by emetine, we generated three new constructs introducing point mutations into the putative HSE. pmuHSEPgp5 contains three nucleotide replacements in the central inverted repeated GAA of HSE (ATAGAAATTTTTCATA → ATATCGATTTTTCATA); pmHSE-TTCPgp5 had three base changes in the inverted repeated TTC and its flanking bases (ATAGAAATTTTTCATA → ATAGAAATTAGATGAA), and in p-sinHSEPgp5, the HSE sequence was completely replaced (GACTGCTCGACGTGAC). We transfected these constructs into trophozoites from clones A and C2 and then cultured them in the absence or presence of 8 or 20 μM emetine, respectively. The results showed that all of the constructs presented CAT activities in trophozoites from both clone cultures without emetine (Figure 3). However, a dramatic difference in expression was observed when emetine was added to the cultures, and the plasmid pmuHSEPgp5 presented higher CAT activity, similar to that shown by the p259Pgp5 minimal promoter (Figure 3). The activities observed were similar in both clones [A₍₈₎ and C2₍₂₀₎], suggesting that the changes introduced into the GAA sequence from the HSE did not affect the transcriptional activation. Interestingly, the plasmid pmHSE-TTCPgp5 presented a strong CAT activity reduction in trophozoites from both clones, suggesting that the second repeat of the HSE was indispensable to EhPgp5 gene expression. Moreover, when the HSE was eliminated, CAT reporter gene activity was not detected, either in trophozoites of clone A_[8] or in trophozoites of clone C2_[20].

As a second approach to gaining insight into the role of HSEs in the promoter activity of the EhPgp5 gene, we analyzed DNA–protein interactions using gel shift and competition assays. Gel electrophoretic mobility shift assays were performed to obtain evidence of nuclear protein binding to the putative HSE. Double-stranded synthetic oligonucleotides containing the EhPgp5 HSE were incubated with NE from clones A, A_[8], C2, and C2_[20]. Additionally, oligonucleotides containing the wild-type consensus HSE and the HSEm mutated at the repeated GAA and the region of 34 bp of the EhPgp5 promoter from −170 to −136 bp, which contains the CdxA, YY1, and HSE sequences, were used as specific competitors. As demonstrated in Figures 4A,B, the HSE was specifically recognized by nuclear proteins from E. histolytica clones grown in the presence or absence of emetine, forming one DNA–protein complex that had identical electrophoretic mobility using NE from both clones. Interestingly, the intensity of the complex was enhanced when the DNA probe was incubated with NE from the trophozoites from clones A_[8] and C2_[20] (Figures 4A–D, lanes 2 and 7). Competition assays showed that formation of the complex with the HSE probe was efficiently blocked by the wild-type HSE and the S34 oligonucleotides when a 150-fold molar excess was used (Figures 4A–E, lanes 3, 5, 8, and 10) but not with the HSEm, for which the DNA–protein complexes only diminished (Figures 4A–E, lanes 4 and 9), suggesting that the DNA–protein interaction was partially affected by changing the GAA repeated sequence. However, this change did not interfere with the EhPgp5 promoter activity (Figures 4A,B). In contrast, when poly [d(I-C)] was used as a non-specific competitor (350-fold molar excess), the DNA–protein complex formation was not affected (Figures 4A–D, lanes 6 and 11).

To identify the nuclear proteins binding to the EhPgp5 HSE, we performed UV-crosslinking assays after gel shift. Two main proteins of 51 and 34 kDa after correction for the bound probe (61 and 44 kDa) were found in both clones incubated with or without emetine (Figure 5). The HSE-protein interactions were specific due to complex formation being prevented by a 150-fold molar excess of cold oligonucleotide (lane 5), while binding proteins were not inhibited using the unspecific competitor poly [d (I-C)] (lane 6), indicating the specificity of the complexes. The bands obtained using the NE trophozoites from both clones grown in the presence of emetine (Figures 5A,C,E,F), were more intense than the bands obtained using the NE trophozoites from both clones grown in the absence of emetine (Figures 5B,D–F).

To identify the proteins interacting with the EhPgp5 HSE, we searched for putative heat shock transcription factors (HSTF) by western blot assays of the UV-crosslinked complexes from trophozoites of clones A, A_[8], C2, and C2_[20] using the human heterologous antibody anti-HSF1. As shown in Figure 5G, a band of 51 kDa was detected after correction for the bound probe from the DNA–protein complex formed with the HSE element. The band was stronger (20%) in C2_[20], trophozoites (Figure 5H). In trophozoites from both clones grown without emetine, a very slight signal was detected (data not show).

We performed partial purification of proteins that bind to the HSE using the DNA-binding protein purification kit from Roche Molecular Biochemical (Roche, CA, USA), NE from trophozoites of clones A_[8] and C2_[20] and a concatamerized oligonucleotide containing the HSE. Two main and defined enrichment purified proteins of ~94 and 66 kDa, respectively, were detected on a silver-stained SDS-polyacrylamide gel in NE from trophozoites of clone A_[8], (Figure 6A, lane 3), while two proteins of 62 and 51 kDa were purified from the NE from the trophozoites of clone C2_[20] (Figure 6D, lane 3). To determine whether purified proteins were also detected by the HSTF1 heterologous antibody, we performed western blot assays. The results showed that both pairs of proteins, 94 and 66 kDa and 62 and 51 kDa from A_[8] and C2_[20] trophozoites, respectively, were recognized by the human anti-HSTF1 (Figures 6B,E). These results supported the idea that the proteins that binding to the EhPgp5 HSE could be putative heat shock like transcription factors.

Furthermore, the semipurified proteins from both clones were used to determine whether these proteins could bind to the HSE sequence. As shown in Figures 6C,F,G,H (lane 2), two specific DNA–protein complexes were observed with the semipurified proteins of clones A_[8] and C2_[20]. The complexes were highly specific due to being completely inhibited by a 150-fold molar excess of the same unlabeled wild-type HSE and S34 oligonucleotides (lanes 3 and 5), and 350-fold molar excess of poly [d(I-C)] did not affect complex formation (lane 6). Interestingly, the HSEm did not affect complex formation (Figures 6C,F–H lane 4).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling Editor declared a shared affiliation, though no other collaboration, with all the authors.