Entamoeba histolytica Dmc1 Catalyzes Homologous DNA Pairing and Strand Exchange That Is Stimulated by Calcium and Hop2-Mnd1

Meiosis depends on homologous recombination (HR) in most sexually reproducing organisms. Efficient meiotic HR requires the activity of the meiosis-specific recombinase, Dmc1. Previous work shows Dmc1 is expressed in Entamoeba histolytica, a eukaryotic parasite responsible for amoebiasis throughout the world, suggesting this organism undergoes meiosis. Here, we demonstrate Dmc1 protein is expressed in E. histolytica. We show that purified ehDmc1 forms presynaptic filaments and catalyzes ATP-dependent homologous DNA pairing and DNA strand exchange over at least several thousand base pairs. The DNA pairing and strand exchange activities are enhanced by the presence of calcium and the meiosis-specific recombination accessory factor, Hop2-Mnd1. In combination, calcium and Hop2-Mnd1 dramatically increase the rate of DNA strand exchange activity of ehDmc1. The biochemical system described herein provides a basis on which to better understand the role of ehDmc1 and other HR proteins in E. histolytica.


Introduction
In most eukaryotic organisms meiosis is essential for reproduction and comprises one round of DNA replication followed by two rounds of cell division to produce haploid gametes. To initiate meiosis, the Spo11 enzyme generates DNA double-strand breaks (DSBs) throughout the genome [1]. The newly formed DSBs are repaired through a DNA repair pathway known as homologous recombination (HR) [2], which forms physical connections between homologous chromosomes. The linkage between homologous chromosomes helps to ensure proper segregation of paired chromosomes at meiotic prophase I. Once the DSB forms, both ends of the DSB are processed to produce 3' ssDNA tails that, in turn, become the nucleation site for the Rad51 and Dmc1 recombinases-two Escherichia coli RecA orthologs [3][4][5][6][7]. After binding the ssDNA tail, Rad51 and Dmc1 form right-handed helical nucleoprotein filaments known as presynaptic filaments [8][9][10][11]. The presynaptic filament searches for homology by invading the homologous chromosome. When homology is located, Rad51 and Dmc1 catalyze homologous DNA pairing and displace the complementary strand. The resulting structure is known as a displacement loop, or D-loop. Formation of the D-loop is followed by DNA strand exchange. Most organisms express Rad51 ubiquitously, while Dmc1 is expressed only during meiosis. Normal meiosis relies on both Rad51 and Dmc1 [12]. During meiosis, Saccharomyces cerevisiae Rad51 serves as an accessory cofactor promoting Dmc1-mediated recombination [13]. Deletion of DMC1 in S. cerevisiae results in meiotic arrest in prophase I [14], while DMC1-/-knockout mice remain viable but sterile [15]. These results indicate a conserved role for Dmc1 in meiotic recombination [14,15].
Entamoeba histolytica is a protozoan parasite that causes amoebiasis, which can manifest as amebic dysentery and liver abscesses in more than 50 million humans a year-with approximately 70,000 annual deaths worldwide [24][25][26]. E. histolytica has a two-stage life cycle. In the first stage, amoeboid trophozoites proliferate in the colon and may cause disease. As a result of unknown cues, the trophozoite enters the second stage of the life cycle, encystation, which is characterized by genome duplications and formation of tetra-nucleated, environmentally-stable cysts. The cyst-stage facilitates spread to new hosts in contaminated food and water. After ingestion by the host, tetra-nucleated metacystic amoebae emerge in the small intestine (excystation) and undergo several divisions to yield eight trophic amoebae [27][28][29]. Very little is known about the encystation and excystation processes in E. histolytica as no axenic encystation condition exists [30]. Therefore, Entamoeba invadens, a reptilian parasite that encysts in vitro, has been used as a model.
In many organisms, meiosis often results in transmission of gametes or zygotes in order to locate new environments or in the case of parasites, a new host [31]. Several pathogens (Leishmania [32], Trypanosoma brucei [33][34][35], and Giardia lamblia [36]) undergo meiosis, while E. histolytica is thought to be asexual. Several lines of evidence suggest that meiosis may occur in E. histolytica. First, E. histolytica possesses most of the RAD52 epistasis group of DNA repair genes (RAD50, RAD51, RAD52, RAD54/RDH54, RAD55, RAD57, RAD59, MRE11 AND XRS2) involved with HR, which are highly conserved among eukaryotes [37,38]. These genes are differentially expressed in response to DNA damage [37]. Second, Singh et al. [39] employed an in vivo PCR-based method, using inverted repeats, to demonstrate that HR occurs in E. histolytica and E. invadens. Third, trophozoites of E. histolytica and E. invadens [40,41] have one nucleus, while cysts have four nuclei, which suggests meiosis occurs during encystation [42]. Fourth, genes known to be involved with meiotic HR, including Dmc1, were identified in E. histolytica [42]. Lastly, these meiotic genes were expressed [39,42] during developmental transitions and the formation of the tetra-nucleated cysts, providing support for the idea that meiosis occurs in E. histolytica.
In the current study, we report for the first time the purification of the ehDmc1 recombinase and the biochemical characterization of its recombination activities. We demonstrate that ehDmc1 forms presynaptic filaments. These filaments are competent to search for homology in duplex DNA to promote ATP-dependent homologous DNA pairing and DNA strand exchange over at least several thousand base pairs. We show that calcium and mHop2-Mnd1 separately enhance ehDmc1 D-loop formation and DNA strand exchange. Importantly, the rate of these recombination activities was significantly increased when both calcium and mHop2-Mnd1 were present. We present evidence that the small molecule, 4,4 0 -diisothiocyanostilbene-2,2 0 -disulfonic acid (DIDS), inhibits D-loop formation by ehDmc1. Taken together, our data reveal that ehDmc1 is a functional recombinase providing support for the idea that meiosis occurs in E. histolytica.

DNA Substrates
ϕX174 viral (+) strand (ssDNA) and ϕX174 replicative form I (dsDNA) were purchased from New England Biolabs. The ϕX174 replicative form I was digested by ApaLI (New England Biolabs) to generate linearized dsDNA. The supercoiled pBluescript DNA was purified from E. coli, according to manufacturer instructions using a commercially available Giga kit (Qiagen). All oligonucleotides were purchased from Integrated DNA Technologies. The oligonucleotides used in each assay were gel purified using denaturing polyacrylamide gel electrophoresis as described [43]. For the DNA binding assay, the H3 oligonucleotide (ssDNA) was 5'-radiolabeled with [ 32 P-γ]-ATP using T4 polynucleotide kinase [43]. To obtain the double-stranded substrate (dsDNA), 32 P-H3 was annealed to the unlabeled complementary H3c oligonucleotide. In the strand exchange assay, OL83-1 was used as the single-stranded substrate, and the double-stranded duplex DNA was generated by radiolabeling OL83-1, as described above, and annealing it with the unlabeled complementary oligonucleotide, OL83-2. The oligonucleotide OL90 was radiolabeled as described above and was used in the D-loop assay and nuclease protection assay. All previously mentioned oligonucleotides sequences can be found in Table 1.

Isolation and Modification of the Genes Encoding E. histolytica DMC1 and RAD51
Neither the E. histolytica DMC1 gene (XM_651488) nor the RAD51 (XM_648984) gene have apparent introns. Thus, it was possible to isolate DMC1 and RAD51 cDNA directly from genomic DNA by nested-PCR. Total genomic DNA was isolated from trophozoites using the Wizard Genomic DNA Purification Kit (Promega). The first round of PCR used genomic DNA as a template, and Primers 1 and 3 (Table 1) for DMC1 and Primers 4 and 5 (Table 1) for RAD51 isolation. These primers allowed for the incorporation of nucleotides encoding three histidines, a flexible linker of two glycine residues between the tag and the DMC1 or the RAD51 coding sequence, and a 3' BamHI site. A second round of PCR using the first-round PCR products as templates, and Primers 2 and 3 (Table 1) for DMC1 and Primers 2 and 5 (Table 1) for RAD51, facilitated the integration of nucleotides encoding three additional N-terminal histidines (to form the final N-terminal 6 histidine tag) and a 5' NdeI restriction site. The resulting PCR products for both DMC1 and RAD51 were digested with NdeI and BamHI and ligated separately into pET-11a (Novagen). The fidelity of PCR and correctness of the DNA constructs were confirmed by DNA sequencing.

Expression and Purification of ehDmc1 and ehRad51
The pET-ehDmc1-(HIS) 6 expression plasmid was introduced into the BL21 DE3 Rosetta (Novagen) strain of E. coli. The cells were grown at 37°C until an OD 600 of~1.0, and protein expression was induced by the addition of 0.4 mM IPTG at 16°C for 20 hr. The cells were harvested by centrifugation. The cell paste (40 g) was resuspended in Buffer A (50 mM Tris-HCl pH 7.5, 1 mM EDTA, 10% sucrose, 0.01% Igepal, 1 mM β-mercaptoethanol, 0.1 mg/ml lysozyme, 1 mM benzamidine, 1 mM PMSF, and protease inhibitors at a final concentration of 5 μg/ml: aprotinin, chymostatin, leupeptin, and pepstatin A) containing 250 mM KCl. All subsequent steps were performed at 4°C. The resuspended cells were lysed by sonication, and the extract was clarified by ultracentrifugation at 40,000 rpm (Beckman Ti-45 rotor) for 90 min. The clarified lysate was diluted 1:3 in Buffer B (20 mM KH 2 PO 4 pH 7.5, 10% glycerol, 1 mM EDTA, 1 mM DTT) and loaded onto a 40 ml Q Sepharose column (GE Healthcare). The column was washed with Buffer B containing 100 mM KCl followed by a linear gradient of Buffer B containing 0-800 mM KCl. The peak fractions (~250 mM KCl) containing ehDmc1 were pooled and incubated with 2 ml of Ni-NTA Sepharose (GE Healthcare). The bound protein was eluted with 6 ml of Buffer B containing 500 mM imidazole and 300 mM KCl. The eluate was diluted 1:4 in Buffer B and loaded onto a 1 ml Source 15S column (GE Healthcare). The column was washed with Buffer B containing 100 mM KCl. A linear gradient of Buffer B containing 0-800 mM KCl was applied to the column. The peak fractions containing ehDmc1 (~300 mM KCl) were pooled and loaded onto a 1 ml Source 15Q column (GE Healthcare). Following a wash, the column was subjected to a linear gradient of Buffer B containing 0-700 mM KCl. Peak fractions (~350 mM KCl) containing ehDmc1 were pooled, concentrated, and stored in small aliquots at -80°C. The yield of purified ehDmc1 was~1 mg. Three independent preparations yielded the same results in the biochemical experiments. ehRad51 was expressed from pET-ehRad51-(HIS) 6 and purified using the same protocol described for ehDmc1, except that the Q Sepharose column was omitted.

Expression and Purification of mHop2-Mnd1
The mHop2-Mnd1 expression plasmid was a kind gift from Dr. Daniel Camerini-Otero (National Institute of Health, Bethesda, MD). The mHop2-Mnd1 protein complex was purified as described [20].

Expression and Purification of hRPA
The expression and purification of the human single strand DNA binding protein, replication protein A (RPA) was performed as described [9,45].

Western Analysis
E. histolytica cells (50 ml,~4 x10 7 cells) were pelleted and re-suspended in 10 ml of Buffer A, followed by two freeze-thaw cycles. Glass beads (1 ml; 0.5 mm, BioSpec Products) were added to the resuspended cells followed by vortexing for 3 min. The lysate was centrifuged, and the supernatant was transferred to a new tube. The clarified supernatant was then mixed with 0.25 ml of SP Sepharose (GE Healthcare) followed by gentle rocking at 4°C for 2 hr. The beads were then washed with 0.5 ml of Buffer B containing 100 mM KCl. The bound proteins were eluted with 1 ml of Buffer B containing 500 mM KCl. The eluate was TCA-DOC precipitated, and the precipitation product (15 μl), along with purified ehDmc1 (1 μg) and ehRad51 (1 μg), was subjected to 12% SDS-PAGE followed by transfer to nitrocellulose membrane (Whatman). Rabbit antibody against scRad51 was a kind gift from Dr. Patrick Sung (Yale University, New Haven, CT) and was used at 1:2000. For the secondary antibody, commercially available HRP-conjugated anti-rabbit IgG (Sigma Aldrich) was used (1:5000), and the membrane was developed using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific).

ATP Hydrolysis Assay
Purified ehDmc1 (1 μM) was incubated with increasing amounts of [ 32 P-γ]-ATP in 10 μl reaction mixture with Buffer C (20 mM Tris-HCl pH 7.5, 2.4 mM MgCl 2 , 50 mM KCl, 1 mM DTT) at 37°C. After 60 min of incubation, 1.5 μl aliquots were removed, and the reactions were stopped by the addition of an equal volume of 0.5 M EDTA. The products were then subjected to thin-layer chromatography (TLC) using polyethyleneimine-cellulose (PEI) plates (Sigma Aldrich). The amount of ATP hydrolysis was determined using a phosphorimager (Typhoon FLA 7000, GE Healthcare). The ATP hydrolysis assay to determine saturating concentrations of DNA was processed as described above, except ehDmc1 was incubated with the saturating concentration of [ 32 P-γ]-ATP at (1. were incubated with dsDNA composed of 32 P-H3 annealed to H3c using the same experimental conditions as above. A control reaction for both substrates was deproteinized via treatment with SDS (0.5%) and Proteinase K (0.5 μg/ml). The reaction products were subjected to 12% native polyacrylamide gel electrophoresis. The gels were dried on Whatman cellulose chromatography paper (Sigma-Aldrich), and the results were analyzed using a phosphorimager. The DIDS electrophoretic mobility shift assay (EMSA) was processed the same as above, except for the addition of increasing amounts of DIDS (2.5 μM, 5 μM, 7.5 μM, 10 μM, 33.3 μM, 66.7 μM, and 100 μM) with ehDmc1 (5.2 μM) and 32 P-radiolabeled H3 substrate (0.05 pmol) at 37°C for 20 min. The reaction products were analyzed as described above.

D-loop Assay
32 P-radiolabeled OL90 oligonucleotide (3.5 μM nucleotides) was incubated in the presence or absence of ehDmc1 (1.5 μM) in Buffer E (25 mM Tris-HCl, 2 mM ATP, 2.4 mM MgCl 2 ) containing 0.1 mg/ml BSA, 1 mM DTT, and the ATP regeneration system consisting of 16 mM creatine phosphate and 36 μg/ml creatine kinase in a final reaction volume 12.5 μl for 10 min at 37°C. The reaction was initiated by the addition of supercoiled pBluescript (35 μM base pairs). At indicated time points, the reactions were deproteinized with the addition of SDS (0.5%) and Proteinase K (0.5 μg/ml). The reaction products were separated using 0.9% agarose gel electrophoresis, dried on DE81 anion exchange paper (GE Healthcare), and analyzed with a phosphorimager. Where indicated, ADP and the ATP analogues, ATP-γ-S and AMP-PNP, were used with ehDmc1 in lieu of ATP in the reaction. In the order of addition D-loop assay, the order in which ssDNA and dsDNA were added to the reaction was altered for two reactions. Specifically, (a) ehDmc1 was incubated at 37°C with dsDNA (pBluescript) for 10 min prior to the addition of ssDNA ( 32 P-OL90) and (b) ehDmc1 was incubated at 37°C with both ssDNA ( 32 P-OL90) and dsDNA (pBluescript) for 10 min. The time course D-loop reactions with calcium and mHop2-Mnd1 were processed as described above except ssDNA ( 32 P-OL90) was incubated at 37°C in the presence or absence of ehDmc1 (1.5 μM) for 5 min prior to the addition of CaCl 2 (0.32 mM) and/or mHop2-Mnd1 (0.16 μM). After an additional 5 min incubation at 37°C, the reaction was initiated by the addition of dsDNA (pBluescript). The reactions were stopped at the indicated time points, as described above. The DIDS (4,4 0diisothiocyanostilbene-2,2 0 -disulfonic acid) D-loop assay was processed the same as above, except for the addition of increasing amounts of DIDS (2.5 μM, 5 μM, 7.5 μM, and 10 μM) after 2 min of ehDmc1 incubation at 37°C with ssDNA ( 32 P-OL90). After 8 min of incubating, the reaction was initiated with the addition of pBluescript followed by a 12 min incubation. The reactions were deproteinized and the products were analyzed as described above.
Nuclease Protection Assay 32 P-radiolabeled OL90 (3 μM nucleotides) was incubated with ehDmc1 (1.5 μM) for 1, 2, 5, and 10 min at 37°C in 10 μl total reaction volume in Buffer D. DNase I (2 units; Promega) was added to the reaction followed by an additional 15 min incubation at 37°C. The reaction was ehDmc1 Catalyzes D-Loop Formation and DNA Strand Exchange deproteinized by treatment with SDS (0.5%) and Proteinase K (0.5 μg/ml) at 37°C for 15 min. The products were subjected to 12% native polyacrylamide electrophoresis, and the gels were analyzed with a phosphorimager. When different ATP analogues and ADP were used, the reactions were performed under the same experimental conditions as above, except ATP-γ-S, ADP, and AMP-PNP (all at 2.5 mM) were substituted for ATP. The DIDS nuclease protection assay was processed the same as above, except for the incubation of increasing concentrations of DIDS (10 μM, 33.3 μM, 66.7 μM, and 100 μM) with ehDmc1 (1.5 μM) and 32 P-OL90. The reactions were deproteinized, and the products were analyzed as described above.

Pull-down Assay
The mHop2-Mnd1 complex (3.5 mg) and BSA (5 mg) were immobilized on 0.5 ml of Affi-gel 15 matrix (BioRad) as per the manufacturer instructions. The mHop2-Mnd1-Affi-gel beads or BSA-Affi-gel beads (17.5 μl) were incubated with ehDmc1 (7 μg) in Buffer B with 50 mM KCl. The reactions were incubated at 4°C with mixing for 45 min. After incubating, the supernatant was removed, and the beads were washed three times with Buffer B. An equal volume of SDS loading dye (160 mM Tris-HCl pH 6.8, 60% glycerol, 4% SDS (w/v)) was added to the beads to elute the bound protein. Aliquots of the supernatant, wash, and bead fractions were separated using 12% SDS-PAGE. The gel was visualized by staining with Coomassie blue.

Oligonucleotide DNA Strand Exchange Assay
The unlabeled OL83-1 oligonucleotide (10 μM nucleotides) was incubated with ehDmc1 (6 μM) for 10 min at 37°C in a 10 μl reaction of Buffer D containing 0.1 mg/ml BSA, 1 mM DTT, and the previously described ATP regeneration system, in the absence or presence of CaCl 2 (2 mM) and/or mHop2-Mnd1 (0.6 μM). This incubation was followed by the addition of duplex DNA ( 32 P-OL83-1/OL83-2; 5 μM base pairs) and spermidine (4 mM final). At the indicated time points, the reactions were stopped by treatment with SDS (0.5% final) and Proteinase K (0.5 μg/ml) at 37°C for 15 min. The products were subjected to 12% native polyacrylamide gel electrophoresis. The gels were analyzed using a phosphorimager.

Plasmid Length DNA Strand Exchange Assay
Purified ehDmc1 (12.5 μM) was mixed with ϕX174 virion ssDNA (30 μM nucleotides) in Buffer E containing the ATP regeneration system, described above, for 10 min at 37°C followed by the addition of RPA (3.8 μM) and KCl (150 mM final). After an additional incubation for 8 min at 37°C, the linearized double-stranded ϕX174 DNA (30 μM base pairs) and spermidine (4 mM final) were added to the reaction. The reactions were stopped at the indicated time points with the addition of SDS (0.5% final) and Proteinase K (0.5 μg/ml) followed by a 20 min incubation at 37°C. The deproteinized samples were separated on 0.9% agarose gel and stained with ethidium bromide. The gels were analyzed using Image Lab software (BioRad).

Sequence Alignment and Phylogenetic Classification
Reference Dmc1 amino acid sequences were downloaded from the reference protein database at GenBank (http://www.ncbi.nlm.nih.gov/) and aligned using MUSCLE [46] with 20 maximum iterations. A neighbor-joining tree was prepared using the Jukes-Cantor [47] genetic distance model and edited in Geneious v6.1.6 (www.geneious.com).

Statistical Analysis
For each assay described, at least three experiments were performed and the error bars represent standard error of the mean.

Cloning, Expression, and Purification of ehDmc1
The presence of the DMC1 meiotic recombinase gene in the genome of E. histolytica suggested that E. histolytica may undergo meiotic recombination. To determine if the putative ehDMC1 gene encoded a functional recombinase, the DMC1 gene encoding a 334 amino acid protein was amplified from E. histolytica genomic DNA and a six-histidine epitope tag was fused to the 5' end of the gene. The gene was sequenced to confirm the absence of any PCR-generated mutations and inserted into the pET11 bacterial expression plasmid. The resultant pET11-ehDmc1-(HIS) 6 expression plasmid was introduced into bacteria and the cells were induced to produce ehDmc1-(HIS) 6 protein. We purified ehDmc1 to greater than 95% homogeneity ( Fig 1A) using a combination of nickel affinity chromatography and ion exchange chromatography.
In most eukaryotic organisms, Dmc1 is expressed exclusively during meiosis. One exception is the upregulation of Dmc1 in response to radiation-induced mitotic catastrophe (MC) in human cell lines [48] creating a pseudomeiotic state. Considering E. histolytica is polyploid, we asked whether ehDmc1 protein was expressed in E. histolytica. Unfortunately, antibodies specific to ehDmc1 are not available commercially. This led us to use a strategy previously reported by Kant et al. (2005) that used heterologous antibodies raised against Saccharomyces cerevisiae Dmc1 to confirm the identity of the Dmc1 protein from the rice plant, Oryza sativa [49]. Previous experience with scRad51 antibodies revealed a cross-reaction with both hRAD51 and hDMC1, likely due to high sequence conservation (data not shown). Based on the crossreactivity of the scRad51 antibodies with hRAD51 and hDMC1, we reasoned the high sequence conservation (61%) between ehDmc1 and hDMC1 may be sufficient for the scRad51 antibodies to detect ehDmc1. To test this idea, we first asked whether the scRad51 antibodies recognized highly purified recombinant ehRad51 and ehDmc1. As shown in Fig 1B, scRad51 antibodies recognized purified recombinant ehRad51 (~40.3 kDa, lane 1) and ehDmc1 (~37.1 kDa, lane 2) confirming that the scRad51 antibodies cross-reacted with the ehRad51 and ehDmc1 protein.
Using the scRad51 antibodies, we performed Western analysis on partially purified lysate from E. histolytica. Our results show two bands were detected in lysate from E. histolytica (Fig 1B, lane 3) that correspond to the same molecular weights of the purified ehRad51 and ehDmc1 (Fig 1B). This suggested that E. histolytica expresses both recombinase proteins at the same time. These results are in agreement with previous reports that ehRad51 and ehDmc1 mRNA transcripts are present during normal cell culture [37,42].
ehDmc1 Hydrolyzes ATP Analysis of the primary sequence of ehDmc1 revealed conserved Walker A and B motifs (S1 Fig) [50] suggesting the potential for ATP hydrolysis activity. In support of this, both hDMC1 and scDmc1 proteins harbor a Walker A motif and possess ATP hydrolysis activity [9,[51][52][53].
To determine if ehDmc1 possessed ATP hydrolysis activity, we performed classic Michaelis-Menten analysis of the ATP hydrolysis activity of ehDmc1. The ehDmc1 protein was incubated with increasing concentrations of [ 32 P-γ]-ATP followed by TLC. From three independent experiments, a relatively weak ATP hydrolysis activity (k cat = 0.53 min -1 , K m = 480 +/-50 μM) was detected (Fig 1C). Previous reports indicated the ATP hydrolysis activity of both hDMC1 and scDmc1 is stimulated by the presence of DNA [9,52,53]. We first determined saturating concentrations of ssDNA and dsDNA to be used in the ATP hydrolysis assay (Fig 1D, 90 μM nucleotides and 45 μM base pairs, respectively). We asked if ehDmc1 ATP hydrolysis activity was stimulated in the presence of DNA by incubating the ehDmc1 protein with saturating concentrations of [ 32 P-γ]-ATP in the absence or presence of saturating amounts of ssDNA or dsDNA prior to TLC. Similar to both hDMC1 and scDmc1, ehDmc1 ATPase activity was stimulated by both dsDNA and ssDNA, with the greatest stimulation occurring in the presence of ssDNA (Fig 1E). The turnover rate (k cat ) for ehDmc1 ATP hydrolysis was 0.53 min -1 which is similar to that reported for hDMC1 (k cat = 0.6 min -1 , [9,43] and scDmc1 (k cat = 0.7 min -1 , [52]). It is interesting to note that the turnover rate for ehDmc1, hDMC1 and scDmc1 are all over 7-fold greater than hRAD51 (k cat = 0.07 min -1 , [54]) and 35-fold greater than ecRecA (0.015 min -1 , [55][56][57]). The catalytic efficiency (k cat /K m ) for ehDmc1 (18 s -1 M -1 ) is 1.6 fold greater than hRAD51 (11 s -1 M -1 , [54]).

DNA Binding Activities of ehDmc1
The observation that the ATPase activity of ehDmc1 was stimulated by the presence of DNA suggested that ehDmc1 binds DNA. To demonstrate that ehDmc1 bound DNA, an EMSA was performed. A 32 P-radiolabeled 80-mer oligonucleotide was used as a ssDNA substrate or was annealed to an unlabeled complementary 80-mer oligonucleotide to create a dsDNA substrate. Increasing concentrations of the ehDmc1 protein were incubated with ssDNA (Fig 2A and 2B) or dsDNA (Fig 2C and 2D) substrates, and the reactions were resolved on a polyacrylamide gel. Fig 2 shows ehDmc1 bound to ssDNA with approximately 5-fold greater affinity than to dsDNA. These results show that ehDmc1 had a strong preference for ssDNA over dsDNA, similar to hDMC1 [43].

Presynaptic Filament Formation by ehDmc1
Formation of a nucleoprotein filament is critical for the recombination activities of Dmc1 [9,51,58]. Dmc1 forms stacked octameric protein rings on ssDNA in the absence of ATP [9,51,59]. These protein rings are not the active form of Dmc1-ssDNA. In the presence of ATP, Dmc1 forms an active right-handed nucleoprotein filament on ssDNA [9]. The difference between the Dmc1 nucleoprotein filament and stacked octameric rings of Dmc1 on ssDNA can be visualized by using a nuclease protection assay [43,60,61]. To determine if ehDmc1 formed presynaptic filaments on ssDNA, we used this nuclease protection assay [43,60,61]. In this assay, if ehDmc1 forms a nucleoprotein filament, the ssDNA will be protected from nucleolytic digestion by DNase I by the ehDmc1 nucleoprotein filament. hRAD51, scRad51, hDMC1 and scDmc1 require ATP binding in order to form a presynaptic filament [8,9,62,63]. Therefore, 32 P-radiolabeled ssDNA was incubated with ehDmc1 in the presence of ATP to allow for presynaptic filament formation. After a brief incubation, DNase I was added to the reaction. The time course of the DNase I digestion showed that ehDmc1 formed a presynaptic filament that protected the ssDNA from DNase I digestion (Fig 3A) The filament formed within one minute and lasted throughout the course of the assay, suggesting ehDmc1 is capable of forming a rapid and stable presynaptic filament. Next, we wished to determine the nucleotide dependence of ehDmc1 presynaptic filament formation. Using the same nuclease protection assay as described above, we incubated ehDmc1 with 32 P-radiolabeled ssDNA in the absence of nucleotide or in the presence of ATP, ADP, the slowly hydrolyzable analog ATP-γ-S, or the non-hydrolyzable ATP analog AMP-PNP. DNase I was added to the reaction after a brief incubation. When either ATP or AMP-PNP, ( Fig  3B, lanes 1 and 4, respectively) was incubated with ehDmc1, the ssDNA was strongly protected from DNase I digestion. Incubation of ehDmc1 with, ATP-γ-S, or ADP ( Fig 3B, lanes 2 and 3, respectively) resulted in greatly reduced protection of the ssDNA from DNase I digestion. In the absence of ATP, ehDmc1 was unable to protect the ssDNA from DNase I (Fig 3B, lane 6). Taken together, the results suggest that ehDmc1 requires ATP binding but not hydrolysis to form presynaptic filaments.

ehDmc1 Catalyzes DNA Homologous Pairing
The ability of ehDmc1 to form a stable presynaptic filament suggested that ehDmc1 may be able to commence homologous DNA pairing. To determine if the ehDmc1 presynaptic filaments mediated DNA homologous pairing, an assay was used that monitors the formation of a D-loop [43]. In this assay, ehDmc1 was incubated with a 32 P-radiolabeled oligonucleotide (ssDNA) to allow presynaptic filament formation. Upon addition of this nucleoprotein complex to supercoiled duplex DNA harboring a region of complementary sequence, the 32 P-radiolabeled oligonucleotide was assimilated into a supercoiled duplex DNA through base-pairing with the complementary sequence within the duplex DNA. As a result, the homologous sequence was displaced forming a D-loop (Fig 4A). Guided by previous work on hDMC1 [9], we determined the preferred order of addition for the ssDNA and supercoiled dsDNA. Our results show that ehDmc1 strongly prefers to form a presynaptic filament on ssDNA prior to the addition of duplex DNA (Fig 4B, lane 2). Co-addition of ssDNA and dsDNA resulted in greatly reduced D-loop formation (Fig 4B, lane 4). No D-loop was detected when ehDmc1 was incubated on supercoiled dsDNA prior to the addition of ssDNA (Fig 4B, lane 3). This preference of forming a presynaptic filament on ssDNA to form D-loop is in agreement with that of hDMC1 [9]. To determine if ehDmc1 required ATP hydrolysis for D-loop formation, ATP was substituted with AMP-PNP, ATP-γ-S, or ADP. D-loop formation by ehDmc1 was seen in the presence of ATP (Fig 4C, lanes 2-4). In the presence of AMP-PNP, the D-loop formation by ehDmc1 was significantly weaker than that seen in the presence of ATP (Fig 4C, compare lanes  4 and 7). This result is similar to that seen with hDMC1 [43]. Contrastingly, scDmc1-mediated D-loop formation was reported to be greater in the presence of AMP-PNP than with ATP [52]. Neither ATP-γ-S nor ADP supported ehDmc1-mediated D-loop formation (Fig 4C, lanes 5  and 6, respectively). As expected, no D-loop formation was seen in the absence of ATP (Fig 4C,  lane 8). These results suggest ehDmc1, like scDmc1 and hDMC1, requires ATP binding but not hydrolysis to form D-loop [9,52].

ehDmc1-mediated D-loop Formation is Stimulated by mHop2-Mnd1 and Ca 2+
Previous reports indicated the D-loop forming activity of scDmc1 [63,71,72] and hDMC1 [51] is enhanced by calcium. Using the preferred order of addition and nucleotide to support ehDmc1-mediated homologous DNA pairing, we used the D-loop assay to determine if calcium stimulated ehDmc1 D-loop formation. Calcium mediated an approximate 2-fold stimulation of D-loop formation catalyzed by ehDmc1 (Fig 6, compare lanes 2-4 with 6-8). Our observation that ehDmc1 interacts with mHop2-Mnd1 and previous reports demonstrating mHop2-Mnd1 enhances both spDmc1 [23] and hDMC1 D-loop formation [20,67] led us to ask if mHop2-Mnd1 stimulated ehDmc1 D-loop formation. When mHop2-Mnd1 was added to the ehDmc1 D-loop reaction, there was an approximate 3.5-fold increase in D-loop formation (Fig 6, lanes 10-12). Since both calcium and mHop2-Mnd1 stimulated D-loop formation, we investigated whether co-addition of calcium with mHop2-Mnd1 would further enhance ehDmc1 D-loop formation. As shown in Fig 6 (lanes 14-16), addition of calcium did not an aliquot was removed and deproteinized. The reaction products were separated on a 12% native polyacrylamide gel followed by analysis with a phosphorimager. The mean percent protection of the ssDNA from DNase I digestion of three independent experiments was graphed. Error bars represent SEM. Lane   The ability of ehDmc1 to form D-loops led us to determine whether ehDmc1 was capable of DNA strand exchange in a 3-strand assay that utilizes ϕX174 DNA viral (+) ssDNA and linearized ϕX174 dsDNA that are 5.4 kilobase pairs in length [9]. In this assay, ehDmc1 was preincubated with the ϕX174 DNA viral (+) ssDNA to allow for presynaptic filament formation. This was followed by the addition of hRPA, a single strand DNA binding protein that stabilizes the ssDNA by preventing the formation of secondary structure. The addition of linear ϕX174 dsDNA initiated the reaction leading to joint molecule formation and strand exchange forming a nicked circular product ( Fig 7A). As shown in Fig 7B, ehDmc1 catalyzed ATP-dependent DNA strand exchange over 5.4 kilobase pairs. This DNA strand exchange was similar to the activity shown for Dmc1 proteins from other species [9,18,73]. Although the molecular effect of spermidine on the ability of other recombinases [9,74] to commence DNA strand exchange is not known, we found ehDmc1 DNA strand exchange was dependent upon the presence of spermidine (data not shown).

Oligonucleotide DNA Strand Exchange by ehDmc1
Since ehDmc1 weakly catalyzed DNA strand exchange on plasmid length DNA substrates, we switched to an oligonucleotide-based DNA strand exchange assay [43,45] to further investigate the influence of calcium and mHop2-Mnd1 on ehDmc1. In this assay, an oligonucleotide ssDNA substrate was incubated with ehDmc1 followed by the addition of radiolabeled dsDNA to initiate the reaction. DNA strand exchange occurs when the radiolabeled strand of the dsDNA substrate is replaced by the unlabeled homologous ssDNA within the ehDmc1 presynaptic filament (Fig 8A). Our results demonstrate that ehDmc1 is adept at DNA strand exchange  (Fig 8). The presence of calcium in the reaction increased not only the rate of DNA strand exchange but also the amount of strand exchange product (~38%) compared to the absence of calcium (~16%). The addition of mHop2-Mnd1 alone or in combination with calcium greatly increased the rate of ehDmc1-mediated DNA strand exchange ( Fig  8B and 8C). The time scale of the reactions containing mHop2-Mnd1 separately or in combination with calcium did not allow us to determine if calcium enhanced the ehDmc1/ mHop2-Mnd1-mediated DNA strand exchange. Therefore, we performed the same DNA strand exchange assay using shorter time points. The presence of calcium greatly increased the rate of the DNA strand exchange activity of ehDmc1-mHop2-Mnd1 (Fig 8D and 8E) with the reaction reaching a plateau in less than 15 seconds. There was over a 240-fold increase in the rate of the strand exchange reaction mediated by ehDmc1 in the presence of both calcium and mHop2-Mnd1. Taken together, these results indicate that calcium and mHop2-Mnd1 enhance ehDmc1-mediated DNA strand exchange.

Phylogenetic Relationship of ehDmc1 to Other Pathogens, Yeast, and Higher Eukaryotes
The similarity of the response by ehDmc1 and hDMC1 to calcium and mHop2-Mnd1 in Dloop formation and DNA strand exchange prompted us to evaluate the relationship of ehDmc1 to hDMC1, scDmc1, spDmc1, and Dmc1 from other pathogens by sequence alignment (S1 Fig). The phylogenetic tree (S2 Fig) shows that ehDmc1 is more closely related to other pathogens than to the yeasts and human. However, ehDmc1 shares 201 identical sites with the

DIDS Inhibits ehDmc1 Homologous DNA Pairing
To date, no inhibitors for any Dmc1 protein have been identified. However, several compounds have been reported to inhibit hRAD51 [75]. We tested one of these compounds, 4,4 0 -diisothiocyanostilbene-2,2 0 -disulfonic acid (DIDS) to determine if it would inhibit ehDmc1-mediated functions [76,77]. Previous work suggested DIDS may inhibit DNA binding to the hRAD51 recombinase [76]. We wished to determine if DIDS directly inhibited DNA binding by ehDmc1. To do this, we used an EMSA where ehDmc1 was incubated with ssDNA, ATP and increasing concentrations of DIDS. As shown in Fig 9A, DIDS inhibited ssDNA binding by ehDmc1. The concentration of DIDS that was required to completely inhibit ssDNA binding by ehDmc1 was~6 fold higher (66 μM) than reported for hRAD51 (10 μM, [76]). Because ATP was present in the EMSA, we wished to determine whether DIDS directly inhibited ssDNA binding by ehDmc1 or whether DIDS indirectly inhibited DNA binding by interfering with ATP binding that is necessary for presynaptic filament formation. We addressed these possibilities using the ATP hydrolysis assay. Incubation of ehDmc1 in the presence of ssDNA with the amount of DIDS that inhibited DNA binding resulted in attenuation of the stimulatory effect ssDNA has on ehDmc1 ATP hydrolysis activity (Fig 9B). In the absence of ssDNA, DIDS had no effect on the ATP hydrolysis activity of ehDmc1 (Fig 9B). These results suggest DIDS directly interferes with ssDNA binding and not with binding of ATP. We used the nuclease  protection assay to monitor the stability of the ehDmc1 nucleoprotein filament in the presence of DIDS. ehDmc1 was incubated in the presence or absence of ssDNA and/or increasing concentrations of DIDS. Our results show that DIDS inhibits presynaptic filament formation ( Fig  9C). Notably, the concentration of DIDS that inhibits presynaptic filament formation is the same concentration that inhibits ssDNA binding by ehDmc1. Lastly, we asked whether the inhibition of presynaptic filament formation by DIDS would compromise the ability of ehDmc1 to catalyze D-loop formation. The results show that increasing concentrations of DIDS inhibited D-loop formation by ehDmc1 (Fig 9D). The IC 50 of DIDS on ehDmc1-mediated homologous pairing is 4.5 +/-0.23 μM. Our results demonstrate that ehDmc1 is a potential target for small molecule inhibitors in a manner similar to hRAD51 [75,76,[78][79][80][81][82][83].

Discussion
Here, we report that ehDmc1 protein is expressed in E. histolytica. We cloned and devised a purification procedure for the ehDmc1 recombinase. We demonstrated purified ehDmc1 possesses ATP hydrolysis activity that was stimulated preferentially by ssDNA. ehDmc1 binds ssDNA with a higher affinity than with dsDNA and forms presynaptic filaments on ssDNA in an ATP-dependent manner. Additionally, ehDmc1 catalyzed robust ATP-dependent D-loop formation and DNA strand exchange, which is stimulated by the presence of calcium. We showed that mHop2-Mnd1 interacts with ehDmc1 to enhance the D-loop formation and DNA strand exchange activity of ehDmc1. Our results demonstrated that ehDmc1 catalyzes DNA strand exchange over several thousand base pairs. Based on phylogenetic analysis, we found that ehDmc1 is more similar to higher plants and pathogens. Our biochemical and phylogenetic analysis show that ehDmc1 is more similar to hDMC1 than scDmc1. Lastly, we demonstrated that a small molecule, DIDS, is an effective inhibitor of ehDmc1 homologous DNA pairing. Our data provide strong evidence that ehDmc1 is catalytically active and may be a potential target for therapeutic treatment for E. histolytica infection.
The co-expression of both Rad51 and Dmc1 recombinases is typically seen only in meiosis [14,[84][85][86]. We were surprised to find two proteins bands recognized by our scRad51 antibodies that correspond to the molecular weights of ehDmc1 and ehRad51 in E. histolytica normal cell culture. Importantly, these two bands migrate similarly in SDS-PAGE with purified recombinant ehDmc1 and ehRad51. We are careful to note we relied on the cross-reactivity of scRad51 antibodies to detect expression of ehDmc1 and ehRad51. Previously, two bands (41 and 46 kDa) of ehRad51 were observed by Lopez-Casamichana et al. [37]. These molecular weights differ from the predicted molecular weight of ehRad51 (ehRad51 is 40.3 kDa containing 366 amino acids and ehDmc1 is 37.1 kDa containing 334 amino acids). The difference in molecular weight of the two protein bands observed in this study and that previously reported [37] were both~4 kDa. Therefore, like the antibodies used in this study, the antibodies used by Lopez-Casamichana et al. [37] may have also cross-reacted with ehDmc1. As a result, Lopez-Casamichana may have actually detected both ehDmc1 and ehRad51.
Our results demonstrate that ehDmc1 has similar biochemical characteristics as those previously described for Dmc1 recombinases from other organisms [9,18,23,51,52]. The ATPase is stimulated by both ssDNA and dsDNA with ssDNA providing the greater level of stimulation. The preference of binding ssDNA is in agreement with the preferred order of addition in the D-loop formation assay where the recombinase binds to ssDNA prior to the addition of dsDNA. The formation of a stable presynaptic filament is dependent upon the binding and not  DIDS inhibits presynaptic filament formation by ehDmc1. A. ehDmc1 was incubated with 32 P-radiolabeled OL90 ssDNA in the presence and absence of increasing amounts of DIDS at 37°C for 20 min. Products were separated on 12% polyacrylamide gels and analyzed with a phosphorimager. B. ehDmc1 was incubated with saturating amounts of [ 32 P-γ]-ATP in the presence and absence of ϕX174 ssDNA and/or DIDS (66.6 μM). The reactions were stopped at the indicated times, subjected to TLC, and analyzed using a phosphorimager. C. ehDmc1 was incubated with 32 P-radiolabeled OL90 in the presence and absence of increasing amounts of DIDS followed by exposure to DNase for 15 min at 37°C. The reactions were stopped, separated on 12% polyacrylamide gels, and analyzed with a phosphorimager. D. ehDmc1 was incubated with 32 P-radiolabeled OL90 ssDNA for 2 min prior to the addition of DIDS hydrolysis of ATP as the non-hydrolyzable ATP analog, AMP-PNP, supported presynaptic filament formation to the same extent as ATP. The observation that AMP-PNP did not support the same level of D-loop formation is similar to that reported for hDMC1 [9] but unlike scDmc1 [72]. The lack of D-loop formation by ehDmc1 in the presence of ATP-γ-S is similar to what is reported for hDMC1 [9]. It is not clear why ATP-γ-S fails to support D-loop formation or DNA strand exchange by ehDmc1 and hDMC1. However, ATP-γ-S weakly supports DNA strand exchange and D-loop formation by hRAD51 [87]. This difference could reflect slight differences within the ATP binding pocket that permit hRAD51 to accommodate ATPγ-S. It is possible, that ehDmc1 has low affinity for ATP-γ-S. In support of this idea, hRAD51-ssDNA nucleoprotein filaments formed in the presence of ATP-γ-S are less extended than ecRecA nucleoprotein filaments [62]. Furthermore, the hRAD51-ssDNA nucleoprotein filaments formed in the presence of ATP-γ-S resemble the inactive ecRecA nucleoprotein filaments formed in the absence of a nucleotide cofactor [88].
The ATP-dependent DNA strand exchange using plasmid length DNA substrates was rather weak when compared hDMC1 [9], hRAD51 [87], and scRad51 [74]. We suggest the full potential of ehDmc1 may not be realized due to the use of human single-strand DNA binding protein, replication protein A (RPA) in the 3-strand DNA strand exchange assay.
Our results show that mHop2-Mnd1 physically interacts with ehDmc1. We find this result intriguing given the identity of ehDmc1 to mDmc1 is 61% and to hDMC1 61%. Furthermore, mHop2 has only 32% identity with ehHop2 and mMnd1 has 41% identity with ehMnd1. We interpret these results to suggest that the interaction surfaces between Dmc1 and Hop2-Mnd1 proteins are well conserved. In agreement with this idea, Ploquin et al. (2007) demonstrated the spDmc1 interacted with mHop2-Mnd1. Here, the 32% identity between mHop2 and spHop2 is the same as that between mHop2 and ehHop2 while the 33% identity between mMnd1 and spMnd1 is lower than the identity between 41% mMnd1 and ehMnd1. Despite this, both spDmc1 and ehDmc1 interact with mHop2-Mnd1.
We show that calcium and mHop2-Mnd1 enhance both homologous DNA pairing and DNA strand exchange by ehDmc1. When both calcium and mHop2-Mnd1 were present, we failed to see any further enhancement of ehDmc1-mediated D-loop formation than seen with either calcium or mHop2-Mnd1 alone. This may be due to the selection of time points for the D-loop formation experiments. To resolve this issue, we switched to a slower oligonucleotidebased DNA strand exchange assay [43]. When we initially used the oligonucleotide-based DNA strand exchange assay, the results were similar to those seen in the ehDmc1-mediated Dloop formation assay where no apparent difference between incubation with mHop2-Mnd1 alone or in combination with calcium was observed. However, the use of shorter time points revealed that calcium synergistically worked with mHop2-Mnd1 to dramatically increase the rate of the ehDmc1-mediated DNA strand exchange reaction over 240-fold. Importantly, we show that calcium is not required for mHop2-Mnd1-mediated stimulation of ehDmc1 as reported for scHop2-Mnd1 and scDmc1 [71]. These results suggest that the activation of ehDmc1 by mHop2-Mnd1 is mechanistically different than the activation by calcium. In support of this idea, calcium was shown to inhibit the ATP hydrolysis activity of hDMC1 and promote a more stable hDMC1-ADP-ssDNA complex [51], while Hop2-Mnd1 is reported to stabilize the hDMC1 presynaptic filament and facilitate the capture of dsDNA [67]. Our results demonstrating that ehDmc1 interacts with, and is stimulated by mHop2-Mnd1 suggests ehHop2-Mnd1 will likely enhance the recombination activities of ehDmc1. It will be important to demonstrate these putative interactions upon availability of purified ehHop2-Mnd1 protein complex.
Currently, meiosis is only proposed to occur during encystation in E. histolytica [42]. Given that HR was demonstrated in E. histolytica [39], ehDmc1 is expressed in E. histolytica [42], our results demonstrating the ehDmc1 protein is likely present in E. histolytica in cell culture, and our biochemical analysis demonstrating ehDmc1 is an active recombinase, provides additional support for the possibility that meiosis occurs in E. histolytica.
An alternate explanation is that polyploidy in E. histolytica may be an adaptation for survival. This is seen in human cancer cells that are known to be aneuploid, tetraploid, and polyploid [91,92]. Radiation treatment of cancer cells induces MC [79] leading to aberrant mitosis. As a result, radiation treated cancer cells become polyploid and aneuploid. Polyploid cells are more resistant to radiation treatment than diploid cells [93]. While polyploidy may be temporarily beneficial to cancer cells, the genome is highly unstable and often triggers DNA checkpoint cell cycle arrest. These cells may attempt to escape cell death through depolyploidization using HR, but most often, the cells undergo apoptosis [94]. However, there are instances of meiotic genes, including Dmc1, being aberrantly upregulated in cancer cells in response to radiation-induced MC [48]. These cells were shown to depolyploidize to become smaller mononucleated cells that survive the radiation treatment and produce progeny [48]. This response to radiation treatment in human cancer cells is similar to the response of Cryptococcus neoformans to fluconazole where the cell becomes aneuploid for specific chromosomes [95]. Once the fluconazole is removed, the cells depolyploidize to their original chromosome copy number. It is possible that E. histolytica does not undergo meiosis, but uses ehDmc1-mediated HR in a manner similar to these radiation-induced MC surviving cancer cells to maintain the pseudomeiotic polyploid state of the E. histolytica cell in response to unknown cues that induce encystation. These examples suggest the formation of a polyploid or aneuploid state may be a conserved survival tactic utilized by eukaryotic organisms.
We present the first report of a small molecule inhibitor for any Dmc1 recombinase. Our demonstration that DIDS is a small molecule inhibitor of ehDmc1 may provide a helpful tool to reveal if ehDmc1 has a role in encystation or as a potential drug therapeutic. The biochemical system described herein should provide a basis on which to better understand the role of ehDmc1 and other HR proteins in E. histolytica.