Identification of Four Entamoeba histolytica Organellar DNA Polymerases of the Family B and Cellular Localization of the Ehodp1 Gene and EhODP1 Protein

We report the identification of a family of four active genes (Ehodp1, Ehodp2, Ehodp3, and Ehodp4) encoding putative DNA polymerases in Entamoeba histolytica, the protozoan parasite responsible of human amoebiasis. The four Ehodp genes show similarity to DNA polymerases encoded in fungi and plant mitochondrial plasmids. EhODP polypeptides conserve the 3′-5′ exonuclease II and 5′-3′ polymerization domains, and they have the I, II, and III conserved boxes that characterize them as DNA polymerases of family B. Furthermore, we found in EhODP polymerases two novel A and B boxes, present also in DNA polymerases encoded in fungi mitochondrial plasmids. By in situ PCR, Ehodp1 gene was located in nuclei and in DNA-containing cytoplasmic structures. Additionally, using polyclonal antibodies against a recombinant rEhODP1-168 polypeptide, and confocal microscopy, EhODP1 was located in cytoplasmic DNA-containing structures.


Introduction
Entamoeba histolytica is the protozoan parasite causative of human amoebiasis [1]. Replication in E. histolytica is inhibited by aphidicolin [2,3], a specific inhibitor of mammalian α, δ and ε DNA polymerases. Additionally, EhMCM2, EhMCM3 and EhMCM5 genes, whose products are part of the helicase complex, have been cloned and characterized [4,5]. Although nuclear α and δ DNA polymerase sequences are present in E. histolytica genome [6], DNA polymerase encoding genes have not been isolated or characterized and DNA replication processes are poorly understood in this parasite.
In eukaryotes, replicative DNA polymerases are grouped in two families: (1) family A, which includes γ DNA polymerases of animals and fungi, and Pol I-like DNA polymerases responsible for mitochondrial DNA replication in plants and slime mold. (2) family B comprises the α, δ and ε DNA polymerases involved in nuclear DNA replication [7][8][9], archaebacterial, viral, bacteriophage DNA polymerases such as those present in phages T4 and RB69 and DNA polymerases encoded in fungi and plant mitochondrial plasmids [10]. Commonly, the fungal plasmids are linear and they have been frequently found in filamentous fungi [11]. Transcription and replication of linear plasmids are initiated in terminal inverted repeats by a plasmid encoded phage-like single subunit RNA polymerase and by a DNA polymerase of the family B, respectively [12]. Replication in these plasmids is thought to occur by a protein-primed mechanism, similar to that described for Bacillus subtilis phage phi29 [13]. The B DNA polymerases are distinguished by the presence of up to six common regions in their amino acid sequences (boxes I to VI). The most conserved regions (I and II) include aspartic acid residues essential to catalytic polymerase activity [10]. Although it has been reported that E. histolytica had a secondary mitochondrial lost [14], no genes encoding γ DNA polymerase responsible for mitochondrial DNA replication have been detected [6]. However, trophozoites carry mitosomes, a mitochondrial cytoplasmic remnant organelle lacking DNA [15,16], and crypton and EhkO [17][18][19][20], two DNA-containing cytoplasmic organelles, with a double membrane. Crypton is a 0.5 to 1 μm organelle that carries the mitochondrial chaperonin Hsp60 [17,18], whereas EhkO varies from 0.5 to 5 μm and it has the EhPFO enzyme [21]. Some authors have suggested that crypton and EhkO could be the same structure [18]; however, their morphological and biochemical characteristics need to be better studied to define this.
The mechanism of DNA replication and the proteins and genes involved in this process in E. histolytica are unknown. To better understand the DNA replication process in this parasite we have initiated the search and study of its DNA polymerase genes. Here, we report the identification of a gene family (Ehodp1, Ehodp2, Ehodp3 and Ehodp4) encoding putative E. histolytica DNA polymerases. All of them correspond to the family B, with a high similarity to fungi and plant DNA polymerases encoded in mitochondrial plasmids. RT-PCR experiments indicated that the four genes are expressed in trophozoites. Additionally, in situ PCR assays demonstrated that Ehodp1 gene is located in nuclei and in cytoplasmic DNA-containing structures. By confocal microscopy using polyclonal antibodies against a recombinant EhODP1 fragment, the EhODP1 polypeptide was only detected in cytoplasmic structures, but not in nuclei.

Cloning of a 504 bp DNA Fragment from the Ehodp1 Gene.
A 504 bp DNA fragment from contig1 sequence (from 232 to 735 bp) was PCR amplified from E. histolytica total DNA using 200 nM of each primer (odp1-f and odp1-r), 400 μM dNTPs, 2 mM MgCl 2 and 2 U of Taq DNA polymerase. PCR was performed using the conditions mentioned above. Amplified DNA was purified and cloned into pRSET A vector (Invitrogen) to generate the recombinant pRSET A-Ehodp1-504 plasmid (prEhodp1). Sequencing of cloned DNA was carried out using the Big Dye Terminator kit version 3.1 (Applied Biosystems) in an Automated DNA Sequencer (310 Genetic Analyzer, Applied Biosystems). Nucleotide sequence data of the Ehodp1 gene is available in the GenBank database under the accession number EU423197.

Two Dimensional-Gel Electrophoresis and MALDI-TOF
Analysis. Purified fractions, obtained from the Ni 2+ -NTAagarose affinity chromatography, were analyzed in 2D-gels. Isoelectric focusing was performed with ZOOM strips (linear pH 3-10 gradient) in an XCell SureLock Mini-Cell system (Invitrogen) at 200 V (20 minutes), 450 V (15 minutes), and 2,000 V (30 minutes). Second dimension was done through a 15% SDS-PAGE with a Mini-Protean II system (Bio-Rad). Then, gels were Coomassie Brilliant Blue stained and selected spots were cut and sent to the Protein Chemistry Core Facility at Columbia University for analysis by mass spectrometry in a MALDI-TOF system.

Generation of Rat Polyclonal Antibodies Against rEhODP1-168 Polypeptide and Western Blot Assays.
Wistar rats were three times intramuscularly immunized with 15 μg of purified rEhODP1-168 polypeptide mixed with diluted 1 : 10 (v/v) Titer Max Gold (CytRx Corporation) in PBS at 15 days intervals. Then, immune serum was collected 14 days after last immunization. Rats were bled before the first immunization to obtain preimmune serum. For Western blot assays, purified rEhODP1-168 polypeptide was separated by 15% SDS-PAGE and transferred to a nitrocellulose membrane. Membranes were blocked with 3% (w/v) nonfat milk in PBS for one hour [24]. Immunodetection of His-tagged polypeptide was done by incubation with 0.3 μg/mL mouse anti-6His monoclonal antibodies (Roche) for 1 hour at 37 • C. Membranes were washed 3 times with PBS, followed by incubation with horseradish peroxidaseconjugated goat anti-mouse IgG secondary polyclonal antibodies (Zymed) (1 : 2,000) at room temperature (RT) for 1 hour. Immunoreactive bands were visualized using 3, 3 -diaminobenzidine and 0.025% (v/v) H 2 O 2 [25]. For Western blot of trophozoite total extracts supplemented with 1 × complete protease inhibitors (Roche) and EhkOenriched fraction, proteins were separated through 10% SDS-PAGE and transferred to nitrocellulose membranes that were treated as mentioned above. Then, membranes were incubated with either rat anti-rEhODP1-168 polyclonal antibodies or preimmune serum (1 : 1,000) overnight at 4 • C, and revealed as described [25].

Isolation of an EhkO-Enriched Fraction.
EhkOs were purified from [ 3 H]-Thymidine labeled trophozoites as described [23]. Cells were washed with PBS and resuspended in 8 volumes of buffer A (10 mM EDTA, 10 mM DTT, 10 mM HEPES, pH 7.9) containing 1× complete protease inhibitors (Roche) and 250 mM sucrose. Then, cells were gently disrupted on ice using a Potter homogenizer and centrifuged at 160 ×g for 10 minutes. The supernatant was centrifuged at 10,000 ×g for 10 minutes at 4 • C, and the pellet was resuspended in 15% (v/v) Nycodenz (Axis-Shield) in buffer A and top loaded on a Nycodenz discontinuous gradient (30%, 40% and 50%, all v/v). Then, pellet was centrifuged at 13,000 ×g for 60 minutes at 4 • C. Fractions of 0.5 mL were collected with a DensiFlow II C system (Buchler Instruments) and a RediFrac 1,000 fraction collector (Bio-Rad). 100 μL aliquots of each fraction were 10% (w/v) TCA precipitated and their radioactivity content was determined in a LS6500 liquid scintillation counter (Beckman). EhkO-containing fractions were identified by [ 3 H]-Thymidine incorporation.

In Situ PCR.
Assays were performed according to protocols included in the In Situ-PCR (IS-PCR) manual (Perkin-Elmer). Exponentially growing trophozoites were attached to glass slides, washed with PBS at 37 • C and paraformaldehyde fixed (4% w/v) at RT for 60 minutes. Fixed trophozoites were incubated in 20 mM HCl (15 minutes), washed twice in PBS (5 minutes), incubated in 0.01% (v/v) Triton X-100 (90 seconds), washed in PBS and treated with 1 μg/mL proteinase K at 40 • C (25 minutes). Next, cells were dehydrated in ethanol solutions (30%, 50%, 70% and 100%, v/v) at 4 • C (5 minutes). For IS-PCR, samples were covered with 50 μL of reaction mixture containing 200 μM dNTPs, 400 μM of each odp1-f and odp1-r primers (described in Materials and Methods, Section 2.3), 4.5 mM MgCl 2 , 10 U of AmpliTaq DNA Polymerase IS (Perkin-Elmer) and 0.1 μL of Cy5-dCTP (Amersham). PCR conditions were as follows: one cycle at 94 • C (3 minutes), followed by 30 cycles of annealing at 59 • C (60 seconds) and extension at 72 • C (60 seconds). A final extension step was carried out at 72 • C (7 minutes). Then, samples were fixed with 2% (w/v) paraformaldehyde at RT (5 minutes), washed with PBS, incubated with 25 mg/mL RNase A (Roche) (20 minutes), washed with PBS, counterstained with 20 μg/mL PI (5 minutes), and observed through a Leika DM-IRE2 confocal microscope. IS-PCR negative controls were carried out, one without Taq DNA polymerase and the second containing all the reaction components except for oligonucleotides.   [12,27]. Boxes I and II contain part of the catalytic domain, which includes two aspartic acid residues that interact with Mg 2+ ions [28]. In EhODP1, EhODP2, EhODP3 and EhODP4, the corresponding aspartic acid residues are D1040 and D795, D1093 and D848, D952 and D707, and D1140 and D895, respectively (Figures 1 and 3(a)). Interestingly, we also identified here two other novel boxes (A and B) in EhODP sequences that we also located in fungi mitochondrial plasmids and in the putative T. vaginalis DNA polymerase (Figures 1, 2 and 3(b)). In EhODP1, Box A (IIFKDTLALIPTSISNFKTFFKLDGKYEKEIFPY) spans from I616 to Y649, in EhODP2 from I669 to Y702, in EhODP3 from I528 to Y561 and in EhODP4 from I716 to Y749. Box B in EhODP1 (FNTEKYCEYYCLRDVLVL-REGFLKYK) spans from F697 to K722, while in EhODP2, EhODP3 and EhODP4, box B corresponds to the regions located at F750 to K775, F609 to K634 and F797 to K822, respectively.

Identification of Four Genes
Then, using the Simple Modular Architecture Research Tool (SMART), we found in EhODPs the characteristic DNA pol B 2 (PF03175) domain of organellar and viral DNA polymerases of family B (Figure 1). In EhODP1 this domain is located at amino acids 570 to 1042 (e value 1.1 × 10 −11 ), in EhODP2 at amino acids 621 to 1095 (e value 1.9 × 10 −15 ), in EhODP3, at amino acids 480 to 954 (e value 1 × 10 −10 ) and in EhODP4 at amino acids 670    to 1142 (e value 4.3 × 10 −10 ). In addition, we found in them the 3 -5 exonuclease II domain [29] (Figure 3). The presence of boxes I, II, III, A and B in the four EhODP polypeptides, conserved in DNA polymerases encoded in mitochondrial plasmids, as well as the detection of the 3 -5 exonuclease II domain, strongly suggests that these E. histolytica polymerases constitute a family of putative organellar DNA polymerases belonging to family B.

Transcription of the Ehodp1, Ehodp2, Ehodp3 and Ehodp4
Genes in E. histolytica Trophozoites. To determine if Ehodp genes were expressed in trophozoites, we performed RT-PCR assays with specific oligonucleotide pairs for each gene. Results showed that the amplified products presented the expected 524, 317, 149 and 376 bp sizes for Ehodp1, Ehodp2, Ehodp3 and Ehodp4 genes, respectively (Figure 4(a)), suggesting that the four Ehodp genes are transcriptionally active. The expression level of each gene was measured by semiquantitative RT-PCR (Figures 4(b)-4(d)) and data were normalized against the amount of actin transcript simultaneously obtained. We calculated the relative abundance of the Ehodp genes considering Ehodp1 gene expression level as 100%. Thus, Ehodp2 exhibited 96.7%, Ehodp4 showed 35%, while Ehodp3 was expressed only at 3.7%. Negative RT-PCR controls with the same RNA were performed for each gene in the same conditions, but without reverse transcriptase (Figure 4(a)). As negative controls, we alternatively performed reactions without Taq DNA polymerase and without Ehodp1 specific oligonucleotides. No fluorescent signals were obtained in these cases (Figures 5(c), 5(d), Cy5). Integrity of trophozoites was confirmed by Nomarsky microscopy (Figure 5, MN). These results demonstrate that Ehodp1 gene is located in nuclei and in cytoplasmic DNA-containing structures that may correspond to EhkOs [19][20][21] or cryptons [17,18], or both.

Expression of a Recombinant EhODP1 Polypeptide.
To initiate the characterization of the EhODP gene family, we cloned an Ehodp1 504 bp DNA fragment to obtain the recombinant plasmid prEhodp1 that was used to transform E. coli BL21(DE3)pLysS cells to produce a histidine-tagged   (Figure 6(a)). Both polypeptides copurified when extracts from induced bacteria were passed through a Ni 2+ -NTA-agarose column under denaturing conditions (Figure 6(b)). Then, the recombinant protein was detected by mouse monoclonal antibodies against the 6His tag. Antibodies only recognized the 33 kDa band (Figures 6(c), 6(d)), and they did not react with the 23 kDa band. To obtain further data to identify the rEhODP1-168 polypeptide, we analyzed the purified fraction by 2D gels. Both proteins in the fraction presented a closely related isoelectric point to the expected value of 4.5 (Figure 6(e)). Therefore, we excised the spots from the 2D gel and performed a MALDI-TOF mass analysis. Amino acid sequences indicated that the 33 kDa polypeptide corresponded to rEhODP1-168 (Figures 6(e), 6(f)), whereas the 23 kDa polypeptide (Figure 6(e)) was a histidine-rich bacterial protein, with similarity to a bacterial FKBP-type peptidyl-prolyl cis-trans isomerase. The molecular weight showed in SDS-PAGE by rEhODP1-168 could be explained by the presence of acidic residues in its sequence, which affects electrophoretic migration, as it has been described for caldesmon, tropomyosin and calsequestrin proteins [30].

Immunodetection of EhODP1 in Trophozoite Total Extracts and in EhkO-Enriched Fractions.
To search for the presence of EhODPs in EhkOs, we purified these organelles (Figure 7(a)) as described in Materials and Methods [23] and carried out Western blots with rat anti-EhODP1 polyclonal antibodies. As a control, we used trophozoite total extracts obtained at the same time and similar conditions than EhkOs and kept at −20 • C during the EhkOs purification process [23] and fresh prepared trophozoite total extracts. In all samples we used the proteinase inhibitor cocktail [23]. The antibodies recognized a 150 kDa band surrounded by a fuzzy region (Figure 7, line 2) only in fresh trophozoite extracts. In frozen trophozoites extracts they immunodetected two bands of 105 and 70 kDa, and in the EhkO-enriched fraction antibodies only detected the 70 kDa band. No signal was obtained in total protein extracts when preimmune serum was used as a negative control (Figure 7(b), lane 5). Theoretical molecular weight of EhODP1 is 135.5 kDa. However, we observed a 150 kDa band. This difference in the molecular weight may be due to the presence of acidic residues in the protein [30], although posttranslational modifications cannot be disregarded. On the other hand, the 105 and 70 kDa bands that appeared in gels could be degradation products of the EhODP1 polypeptide. Hübscher et al. [31] reported that prokaryotic and eukaryotic replicative DNA polymerases are extremely sensitive to proteolytic cleavage, even in the presence of protease inhibitors. They detected major polymerase activity in the 110, 74 and 35 kDa polypeptides in extracts of calf thymus, human fibroblasts and HeLa cells, while in Ustilago maydis, Drosophila melanogaster and E. coli extracts they detected major activity in 110 and 74 kDa bands. These authors suggested that the remarkable similarity in sizes and numbers of polypeptides generated by storage of extracts may result from the conservation of localized amino acid sequences or polypeptide conformations that are particularly susceptible to proteolytic cleavage, generating active fragments of defined sizes. Spanos et al. [32] found a 109 kDa band corresponding to active DNA polymerase in freshly prepared homogeneous E. coli DNA polymerase I, but when they kept this preparation at −20 •

Cellular Location of EhODP1 in Cytoplasmic DNA-Containing Structures of Fixed Trophozoites.
We carried out the immunodetection of EhODP1 in fixed and permeabilized trophozoites, using the anti-rEhODP1-168 antibodies. Fixed trophozoites were contrasted with PI to stain DNA-containing structures. Through confocal microscope, nuclei and cytoplasmic DNA-containing structures appeared stained by PI (Figure 8). Interestingly, the anti-rEhODP1-168 antibodies reacted with structures of 4 μm, but they did not stain nuclei (Figure 8(a)), giving support to the assumption that EhODP1 protein is located in DNAcontaining structures that probably correspond to EhkOs. Merging images confirmed the colocalization of both red and green fluorescent signals in these structures but not in nuclei. As negative controls, we used preimmune serum (Figure 8(b)) or we omitted the first antibody (data not shown). In both cases, red fluorescence was evident but no green fluorescent signals were obtained. Cellular integrity was verified through Nomarsky microscopy.

Discussion
In this work we report the existence of four genes (Ehodp1, Ehodp2, Ehodp3, and Ehodp4) encoding DNA polymerases in E. histolytica. Proteins encoded by these genes have the pol B 2 domain characteristic of organellar and viral DNA polymerases, which includes the 3 -5 exonuclease II domain and the conserved boxes I, II and III, that characterize them as members of family B (Figures 1-3). In addition, these polymerases have two non previously described novel boxes, named here A and B that are shared by DNA polymerases of T. vaginalis and by those encoded by fungi mitochondrial plasmids (Figures 1-3). It is known that some DNA polymerase encoding plasmids can integrate into mitochondrial DNA as a consequence of DNA rearrangements produced in the mitochondrial genome. Interestingly, different phenotypes have been observed in some fungi such as senescence in Neurospora [33,34] or an increase in longevity in P. anserina [35,36] that contain this type of plasmids. Additionally, it is known that the DNA polymerase encoded in K. lactis pGKL-2 plasmid is involved in the integrity and maintenance of this plasmid [37]. The presence of several genes encoding organellar DNA polymerases of the family B has also been reported in the   genomes of other organisms. Agrocybe aegerita has two pol B sequences, the Aa-pol B gene that is potentially functional and a disrupted Aa-polB P1 gene. A. chaxingu has two pol B sequences that contain disrupted ORFs, which could encode nonfunctional enzymes [38]. ORFs encoded by Ehodp genes in E. histolytica conserve the catalytic domains, suggesting that they are functional. By RT-PCR we found that all four Ehodp genes were transcriptionally active and showed different expression levels in asynchronic cultures (Figure 4). In situ PCR experiments showed that Ehodp1 gene was located both in the nuclei and cytoplasmic structures that could be either EhkOs or cryptons. The finding of Ehodp1 gene both in the nucleus and these cytoplasmic structures ( Figure 5) suggests a possible interaction between them as suggested by Solis et al. [20]. Furthermore, EhkOs have the nuclear transcription factors EhTBP [39], Ehp53 [40], and EhCBP [41], and the Ehtbp gene was also found in the nucleus and EhkOs [39] as we determined for Ehodp1 gene in the present work.
The genes encoding organellar or viral DNA polymerases have also been found in transposable elements named Mavericks [42]. Mavericks have an average size of 15-20 kb and are present in eukaryotic genomes including the T. vaginalis genome [42]. When we aligned the protein sequence of EhODP1 with the amino acid sequence of the DNA polymerase encoded in the T. vaginalis Mav Tv1.1 Maverick element, we found that they have 29% identity and 36% similarity (data not shown). However, there are no reports about the presence of Maverick-like elements in E. histolytica, although its genome is rich in transposable elements (nonLong Terminal Repeats) of LINES and SINES classes [43]. Further studies will define whether E. histolytica has or has not Mavericks.
Additionally, our immunolocalization experiments detected EhODP1 polypeptide in fixed trophozoites. EhODP1 protein was located in cytoplasmic DNAcontaining structures but not in nuclei, and in an EhkOenriched fraction, that was not tested for the presence of crypton organelles. For this reason we cannot discard crosscontamination between both DNA-containing organelles. Its presence in these organelles and the conservation of the DNA polymerase catalytic domain in the protein, suggest that EhODP1 could be involved in their DNA replication. We want to note here that due to the similarity in amino acid sequences between EhODP1 and EhODP4 (Figure 3), the anti-rEhODP1-168 antibodies could also recognize the Ehodp4 gene product, which was annotated in the E. histolytica genome database while performing the present work. However, we decided to include EhODP4 here to have a broad panorama of the E. histolytica family B DNA polymerases related to mitochondrial plasmids.

Conclusions
We reported here the presence of a family of four active Ehodp genes in the E. histolytica genome, encoding putative organellar DNA polymerases of family B. EhODP1, EhODP2, EhODP3, and EhODP4 conserve the 3 -5 exonuclease II and 5 -3 polymerization domains and show high similarity to DNA polymerases present in fungi mitochondrial plasmids. EhODP1 protein was detected in EhkOs suggesting that it could be involved in EhkO DNA replication. Interestingly, the Ehodp1 gene was located in nuclei and cytoplasmic DNA-containing structures, indicating a close relationship between these organelles.