A Transposon-Derived DNA Polymerase from Entamoeba histolytica Displays Intrinsic Strand Displacement, Processivity and Lesion Bypass

Entamoeba histolytica encodes four family B2 DNA polymerases that vary in amino acid length from 813 to 1279. These DNA polymerases contain a N-terminal domain with no homology to other proteins and a C-terminal domain with high amino acid identity to archetypical family B2 DNA polymerases. A phylogenetic analysis indicates that these family B2 DNA polymerases are grouped with DNA polymerases from transposable elements dubbed Polintons or Mavericks. In this work, we report the cloning and biochemical characterization of the smallest family B2 DNA polymerase from E. histolytica. To facilitate its characterization we subcloned its 660 amino acids C-terminal region that comprises the complete exonuclease and DNA polymerization domains, dubbed throughout this work as EhDNApolB2. We found that EhDNApolB2 displays remarkable strand displacement, processivity and efficiently bypasses the DNA lesions: 8-oxo guanosine and abasic site. Family B2 DNA polymerases from T. vaginalis, G. lambia and E. histolytica contain a Terminal Region Protein 2 (TPR2) motif twice the length of the TPR2 from φ29 DNA polymerase. Deletion studies demonstrate that as in φ29 DNA polymerase, the TPR2 motif of EhDNApolB2 is solely responsible of strand displacement and processivity. Interestingly the TPR2 of EhDNApolB2 is also responsible for efficient abasic site bypass. These data suggests that the 21 extra amino acids of the TPR2 motif may shape the active site of EhDNApolB2 to efficiently incorporate and extended opposite an abasic site. Herein we demonstrate that an open reading frame derived from Politons-Mavericks in parasitic protozoa encode a functional enzyme and our findings support the notion that the introduction of novel motifs in DNA polymerases can confer specialized properties to a conserved scaffold.


Introduction
The genome of Entamoeba histolytica contains replicative DNA polymerases a, d and e; lesion repair polymerases Rev1 and Rev3, and a family A DNA polymerase able of thymine glycol bypass [1,2,3]. Protozoan parasites Trichomonas vaginalis, Giardia lambia and E. histolytica encode a great variety of transposable elements (TEs) [4]. Among these TEs, a novel class of DNA transposons dubbed Polintons or Mavericks are elements of 15 to 20 kb that encode a family B2 DNA polymerase, a retroviral integrase, a protease and a putative ATPase [5,6]. It is suggested that Politons-Mavericks maybe related to double-stranded DNA viruses and have a direct influence in the evolution of these parasites [7]. For instance, it is estimated that 5% of the genome of T. vaginalis consists of multiple copies of Polintons-Mavericks [5,6]. DNA polymerases from Polinton-Mavericks have to efficiently replicate these long repetitive DNA elements. However, to date no studies on the biochemical characterization of proteins involved in the replication process of Politons-Mavericks have been carried out in any organism. In principle, family B2 DNA polymerases from Politons-Mavericks must be highly proccesive in order to be able to replicative over 20 kbs [5,6,7]. Family B2 DNA polymerases are modular proteins that contain a polymerization and a 39-59 exonuclease domain and two extra elements dubbed Terminal Protein Regions (TPR) 1 and 2. The polymerization is divided in 3 subdomains: thumb, fingers and palm. The structural arrangement of these subdomains forms a cupped right hand in which a double stranded DNA is positioned for nucleotide addition [8,9].
Nature has found two structural solutions for DNA polymerases to incorporate thousands of nucleotides without falling off a template strand. One is the use of processivity factors, like torodial shape proteins or factors that encircle or increment the surface/ area between a DNA polymerase and double stranded DNA substrate, such as PCNA, b-clamp, thioredoxin, UL44, and the b subunit of DNA polymerase c [10,11,12,13]. The second solution is to confer intrinsic processivity to replicative DNA polymerases by the addition of novel domains, as it occurs in T5 and Q29 DNA polymerases [14,15]. Q29 DNA polymerase is the archetypical family B2 DNA polymerase and its TPR2 is responsible for processivity and strand displacement [16,17]. TPR2 structurally resembles the promoter specificity loop of single subunit RNA polymerases, suggesting that nature has used the two beta strand extended loop for processivity and promoter selectivity and that the presence of this loop may have occurred before the specialization of single subunit DNA and RNA polymerases [18,19]. Family B2 DNA polymerases are present in bacteriophages, yeast, cnidarians and parasitic protozoa [20]. However, the only family B2 DNA polymerases characterized to date are those from phages.
A recent report corroborates that E. histolytica contains four family B2 DNA polymerases [21], however the report only characterized its cellular localization and in vivo expression. Herein we report the biochemical characterization of a family B2 DNA polymerase from E. histolytica, this polymerase in comparison to Q29 DNA polymerase contains 21 amino acids extra in its TPR2. We found that this extended TPR2 is responsible for processive polymerization, strand displacement and abasic site lesion bypass.

Phylogenetic analysis and structural modeling of EhDNApolB2
Putative family B2 DNA polymerase were searched in Pathema database (http://pathema.jcvi.org/Pathema/) using the amino acid sequence of Q29 DNA polymerase (Table S1). Initial amino acid alignment was carried out with ClustalW and manually corrected. Phylogenetic reconstruction of the family B2 DNA polymerase sequences was obtained using the maximum likelihood method with LG+I+G as substitution model with gamma = 1.66 and 1,000 bootstrap replicates on phyML 3.0 program at the web server (http://www.lirmm.fr/,gascuel) [22] (Table S2). The homology model of EhDNApolB2 was constructed using the Molecular Operating Environment (MOE) program with the crystal structure of the complex Q29 DNA polymerase/primertemplate DNA as template (PDB ID: 2PZS) [19].

Cloning, Protein expression and purification
Full-length ORF located in loci EHI_018010 and a N-terminal 153 amino acids deletion were PCR amplified using oligonucleotides directed from the Pathema database (Table S3) and subcloned into the pCOLD I vector (Takara). As full-length EHI_018010 is poorly expressed in E. coli, through this work we focused on the N-terminal deletion that we dubbed EhDNApolB2 ( Figure 1). The pCOLDI-EhDNApolB2 construct was transformed into E. coli BL21 DE3-Rosseta II. Transformants were inoculated into 50 ml of LB supplemented with 100 mg/ml of ampicillin and 35 mg/ml of chloramphenicol and used to inoculate 1 liter of LB. This culture was grown at 37uC until it reached an OD600 of 0.6 and induced with 0.5 mM IPTG for 16 hours at 16uC. The cell pellet was harvested by centrifugation at 4uC. Bacterial lysis was carried out by the freeze-thawing method; briefly the pellet was resuspended in 40 ml of 50 mM potassium phosphate pH 8, 300 mM NaCl, 1 mM PMSF, 0.5 mM DTT and 0.5 mg/ml of lysozyme, incubated on ice for 30 minutes and freeze-thaw two times. The resuspended cell culture was sonicated and centrifuged at 17,000 rpm for 30 minutes at 4uC. Recombinant EhDNApolB2 was purified by Immobilized Metal Affinity Chromatography (IMAC) using a 1 ml pre-packed column. The eluate was dialyzed in 50 mM potassium phosphate pH 7.0, 1 mM DTT, 100 mM NaCl and 2 mM EDTA. To further purify EhDNApolB2, the dialyzed protein was loaded onto a heparin column and eluted with a NaCl gradient (50 to 1000 mM). EhDNApolB2 eluted between 400 to 700 mM of NaCl. The fractions were dialyzed in 50 mM potassium phosphate pH 7.0, 1 mM DTT, 100 mM NaCl, 0.5 mM EDTA, 50% glycerol and stored at 220uC. Purity was verified on a 10% SDS-PAGE stained with Coomassie Brilliant Blue R-250.

Deletion and site directed mutagenesis
Exonuclease deficient EhDNApolB2 was constructed by mutating residue Asp345 to alanine using the Quick-Change protocol (Stratagene) accordingly to the manufacturer instructions. DTPR2 DNA polymerase mutant was obtained by deletional PCR mutagenesis using Phusion DNA polymerase (Finnzyme) with primers designed to flanking the ends of the TPR2 region. The complete oligonucleotide sequences used is described in Table  S3. Exonuclease deficient and deletion polymerases were confirmed by automated DNA sequencing.

Polymerization and degradation reactions
Oligonucleotides were used to generate double stranded polymerization substrates as previously described. For a complete list of oligonucleotides used as substrates please refer to Table S4. The nucleotide sequence of the 45 mer template strand is 59-cct tgg cac tag cgc agg gcc agt tag gtg ggc agg tgg gct gcg-39 and 24 mer primer sequence is: 59-cgc agc cca cct gcc cac cta act-39 [3]. Several rounds of buffer optimization revealed that the optimal primer extension buffer for EhDNApolB2 consists of 50 mM Tris-HCl pH 7.5, 2.5 mM MgCl2, 1.5 mM DTT, 0.2 mg/ml BSA (data not shown). Polymerization reactions were carried out using a final concentration of 1 nM primer-template and 20 nM EhDNApolB2 at 37uC. Proofreading exonuclease activity was performed using a set of double stranded matched and mismatched substrates as indicated in the figure legend. The reactions were carried out to 25uC and stopped to indicate time with stop buffer. Single stranded exonuclease activity was performed by a time course experiment using a 59 radiolabeled single stranded DNA (Table S4) at 37uC. Reactions were stopped with a buffer containing 95% formamide, 50 mM EDTA, 0.01% bromophenol blue.
Reaction mixtures were run on a 15% polyacrylamide gel and 8 M urea. Denaturing polyacrylamide gels were dried and analyzed by phosphorimagery on a Molecular Dynamics Phos-phorImager.

Translesion synthesis
DNA amplifications were carried out using 1 nM DNA and variable polymerase concentration (20 nM EhDNApolB2 or 40 nM DTPR2) and 100 mM each dNTP and aliquots were removed as a function of time, added to a stop buffer. Subsequently, samples were run on a 15% polyacrylamide 8 M urea gel and analyzed using a PhosphorImager and Quantity-One software.

Strand-displacement
Strand displacement was carried out with a template of 45 oligonucleotides and a c-P 32 ATP labeled 21mer primer. A third oligonucleotide of 24, 21 and 18 nucleotides was hybridized to create gaps of 1, 3 and 6 nt. Reactions were carried out with 1 nM primer-template, 20 nM of EhDNApolB2 and 40 nM of DTPR2 at 37uC. Reaction mixtures were run on a 15% polyacrylamide gel, 8 M urea.

Processivity Assay
Processivity assays were carried out using single stranded M13mp18 hybridized with a forward primer of 17 nucleotides labeled with c-P 32 ATP. The template was present at 1 nM at the polymerases at 20 y 40 nM. Aliquots were taken at 10, 20 and 40 minutes and stopped with equal amounts of 90% formamide, 50 mM EDTA, 0.1% bromophenol blue. Samples were run on a 6% polyacrylamide gel 8 M urea. The dried gel was visualized and analyzed using a PhosphorImager and Quantity-One software.

Results and Discussion
Identification of Polinton-Maverick derived family B2 DNA polymerases in E. histolytica We performed a BLAST search in the genome of E. histolytica and found 4 family B2 DNA polymerases, although with different amino acid lengths that a previous report which classified them as organellar DNA polymerases [21] ( Figure 1A). The BLAST search indicates that the closest orthologs to the family B2 DNA polymerases of E. histolytica are DNA polymerases related to a TE dubbed Polinton-Maverick present in Entamoeba invadens. Our phylogenetic analysis corroborates an initial observation that the four family B2 DNA polymerase of E. histolytica are closely related to Polinton-Maverick DNA polymerases from G. lambia and T. vaginalis [1,4,6] ( Figure S1A). In this analysis the four family B2 DNA polymerases from E. histolytica are grouped into well-defined branches with a bootstrap value of 1000 for the nearest branch. Family B2 DNA polymerases from linear protein-primed replicated plasmids and bacteriophages are located in different branches of this phylogenetic tree. Thus, our phylogenetic tree analysis strongly suggests that the four family B2 DNA polymerases from E. histolytica are related to TEs. Furthermore, we were able to identify the conserved exonuclease and polymerization motifs of family B and the TPR1 and TPR2 extensions of family B2 DNA polymerases in those DNA polymerases ( Figure 1A and Figure S1B) [23,24,25,26,27]. The conservation of the critical amino acids involved in catalysis for the polymerization and exonucleolytic domains indicates that all four family B2 DNA polymerase in E. histolytica may display polymerase and exonuclease activities ( Figure 1A). The non-conserved N-terminal segment of Polinton-Maverick derived family B2 DNA polymerases of E. histolytica maybe related to a protein segment that functions as a terminal protein as is observed in family B2 DNA polymerases from protein-primed replicated plasmids (For a recent review [28]). We were unable to find the retroviral-like integrase, adenoviral-like protease and ATPase in the genome of E. histolytica associated with Polintons-Mavericks. However, the genome of E. invadens contains a 16,504 bp Polinton-Maverick that contains these canonical proteins [6] and the associated family B2 DNA polymerase shares approximately 85% amino acid identity in polymerization domain of family B2 DNA polymerases from E. histolytica. It is possible that our failure in finding integrase, adenoviral-like protease and a putative ATPase orthologous in E. histolytica maybe to an error in the genome assembly or due to gene lost.
As the amino acid length of the four family B2 DNA polymerases of E. histolytica varies between 813 to 1239 amino acid and the main divergences are located at the N-terminal of these polymerases, we decided to biochemically characterize the ORF in loci EHI_018010 because of its reduced amino length and similitude to the well characterized Q29 DNA polymerase. This polymerase is dubbed in this work ''full-length EhDNApolB2'' ( Figure S1B). We cloned and sequenced two independent clones of full-length EhDNApolB2. After sequencing them we corrected the identity of several residues located in the exonuclease domain, among them those present in motif Exo III. An amino acid sequence alignment of full-length EhDNApolB2 in comparison to Q29 DNA polymerase indicates that both proteins share 38% amino acid identity in their exonuclease and polymerization domains and that the main difference appears at the length of the TPR2 motif, which is 21 amino acids longer in EhDNApolB2 ( Figure S2). A structural model of the 39-59 exonuclease and polymerase domains of full-length EhDNApolB2 depicts these 21 extra amino acids as two beta strands adjacent to the finger subdomain. In this structural model the TPR2 motif completely encircles the double stranded DNA ( Figure 1B).

EhDNApolB2 is an active DNA polymerase
Protein expression of full-length EhDNApolB2 resulted in a poorly expressed heterologous protein with yield of less than 0.1 mg per liter of bacterial culture (data not shown). In order to circumvent this problem we decided to create a construct in which the first 153 amino acids were eliminated. These 153 amino acids have no homology with any know protein in the GenBank. The deleted protein resembles in length to Q29 DNA polymerase, which is the archetypical family B2 DNA polymerase ( Figure  S1B). The deleted protein was cloned in a pCOLD I vector and purified by IMAC and heparin chromatography with typical yields of 2 mg per liter of cell culture. After these two purification steps, the protein is more than 95% pure as assessed in a denaturating SDS-PAGE and present a molecular weight of 78 kDa ( Figure 2A). We refer to protein as EhDNApolB2 throughout this work. In order to corroborate the putative 39-59 exonuclease and polymerization activities of EhDNApolB2 we measured its enzymatic activities for double stranded DNA exonucleolytic degradation and primer-template polymerization. As observed in Figure 2B, in the absence of dNTPs EhDNApolB2 gives rise to degradation products that increase in direct relationship with its concentration, indicating that EhDNApolB2 is an active 39-59 exonuclease ( Figure 2B, lanes 1 to 5). In the presence of dNTPs, EhDNApolB2 is able to elongate a 21mer primer to a fulllength 45mer product in a concentration-dependent fashion ( Figure 2B, lanes 6 to 10). The exonucleolytic products observed in lanes 6 to 10 indicate that, as is observed for other DNA polymerases, the exonucleolytic and polymerization activities of EhDNApolB2 are in competition, and that the presence of dNTPs shifts the reaction towards the polymerization mode. The 39-59 exonuclease domain of DNA polymerases like Klenow Fragment, T7 DNA polymerase or Q29 DNA polymerase is responsible of mismatch proofreading an contributes to overall polymerase fidelity [29,30]. To investigate the role of the 39-59 exonuclease of EhDNApolB2 in proofreading we performed a time course experiment using a primer-template substrate with a paired and a mispaired 39OH at 25uC ( Figure 2C). As observed in Figure 2B, the reaction product of EhDNApolB2 after an incubation of 8 minutes with the paired substrate results in approximately 50% of the 24mer hydrolyzed to exonucleolytic products, whereas in the mispaired substrate it has been completely hydrolyzed to smaller 22mer and 21mer products ( Figure 2C, lanes 5 and 10). At 25uC the exonuclease product for the paired nucleotide is limited to 21 nt ( Figure 2B, lane 1 to 5) in contrast to the lower migrating product observed in the exonucleolytic degradation at 37uC ( Figure 2B, lanes 1 to 5).
The differential in activity accordingly to the temperature correlates with the shuttle between exonuclease and polymerase active sites; at lower temperature the primer translocates to the polymerase active site before it subsequent hydrolysis. The exonucleolytic degradation differential observed for paired and mispaired substrates is in agreement with the idea that the frayed end of the mispaired primer translocates to the exonuclease active site of EhDNApolB2 more effectively than a paired end. This preference for a mispaired versus a paired primer-template indicate that the 39-59 exonuclease domain of EhDNApolB2 contributes to the overall fidelity of this polymerase as is the case for other polymerases [31,32,33]. These polymerization and proofreading activities of EhDNApolB2 are in agreement to the presence of 4 invariant amino acids containing carboxylic acid groups (Asp163 and Glu 165, Asp221, and Asp345) in the 39-59 exonuclease motifs ExoI, ExoII, and ExoIII of the exonuclease domain, and 3 invariant aspartic acids and a tyrosine residue (Asp430, Tyr 582, Asp673 and Asp675) located in motifs A, C, and B of the polymerization domain of EhDNApolB2 with respect to family B2 DNA polymerases ( Figure S2).
In order to corroborate that EhDNApolB2 belongs to the family B of DNA polymerases and to further ensure that the observed polymerase and exonuclease activities are intrinsic of EhDNA-polB2 and not due to a co-purified DNA polymerase from E. coli we tested the effect of aphidicolin, a specific inhibitor of family B DNA polymerases, on EhDNApolB2. Aphidicolin inhibits Q29 DNA polymerase, DNA polymerase a and family B DNA polymerases from archaea like Aeropyrum pernix or Pyrococcus furiosus [34,35,36]. For instance, at a concentration of 10 mM dNTPs, the amount of polymerized substrate by Q29 DNA polymerase is reduced by half in the presence of 400 mM of aphidicolin [34]. In  Figure S3 we shown that in the presence of 25 mM dNTPs, the amount of primer extension of M13 single stranded DNA annealed to a 17mer by EhDNApolB2 is reduced by half at a concentration of 242 mM of aphidicolin. Thus, EhDNApolB2 is sensible to aphidicolin inhibition in a similar level than Q29 DNA polymerase. As in other eukaryotes, aphidicolin inhibits the growth and DNA synthesis of E. histolytica indicating the presence of family B2 DNA polymerases in this parasite, as E. histolytica contain canonical family DNA polymerase (a, d, and e) involved in nuclear DNA replication [37].
EhDNApolB2 efficiently bypasses 8-oxo guanosine and abasic site lesion E. histolytica is exposed to oxidative damage during macrophage attack and genes that cope with free radicals are over-expressed in those conditions [38,39]. As a nuclear family A DNA polymerase from E. histolytica is able to efficiently bypass thymine glycol [3] we determined the lesion bypass of recombinant EhDNApolB2. To determine translesion synthesis of EhDNApolB2, we assayed primer extension using a set of primer-templates in which a specific DNA lesion is to be used as a template immediately after the end of the primer. Thus, the first nucleotide to be incorporated into the primer is incorporated opposite the lesion. In a primer extension experiment in which the first base to be used as a template is a control cytosine, a time course experiment shows that at the shortest time (2.5 min) only 4% of the substrate has been used and at the longest time (20 minutes) 22% of the product is fully extended (Figure 3, lanes 1 to 5). Lesion bypass was studied using 8-oxoguanosine, abasic site, 5R, 6S and 5S, 6R thymine glycol, cyclobutane prymidine dimer (CPD) and 6-4 product (6-4 PP) (Figure 3, lanes 6 to 33). We found that EhDNApolB2 efficiently bypasses 8-oxoguanosine and abasic sites, but is unable to bypass thymine glycol, cyclobutane prymidine dimer and 6-4 product (Figure 3). EhDNApolB2 efficiently incorporates and extends from 8-oxoguanosine, after an incubation of 20 minutes 20% of the primer-template is utilized ( Figure 3, lanes 6 to 10). 8-oxoguanosine only posses a moderate block to replicate DNA polymerase and it is readily bypassed by orthologous DNA polymerases like Q29 DNA polymerase [40,41]. Interestingly primer extension of an abasic site containing template is of 12% at the shortest incubation time and 25% complete at the longer extension time, indicating that in this experiment the abasic site is utilized with similar efficiency that an undamaged base (Figure 3, lanes 11 to 15). The only other known DNA polymerase able to incorporate and extend opposite an abasic site with high efficiency is DNA polymerase h [42]. Yfamily DNA polymerases can incorporate opposite an abasic site but are only moderately active to extend after nucleotide incorporation [43].
Thymine glycol posses a strong block for family B polymerase that only incorporate one nucleotide opposite to this lesion and are unable of bypass. EhDNApolB2 bypasses 5R, 6S thymine glycol and 5S, 6R thymine glycol lesions with very low efficiency. The amount of full-length product after 20 minutes is 2 and 3% for 5R, 6S thymine glycol and 5S, 6R thymine glycol respectively. As in other DNA polymerases EhDNApolB2 incorporates only one nucleotide in 5R, 6S thymine glycol and 5S, 6R thymine glycol substrates [44]. The percentage of single nucleotide addition is 38 and 43% respectively (Figure 3 lanes 16 to 24). However, EhDNApolB2 only incorporates one nucleotide opposite thymine glycol, and is not able to extend after single nucleotide incorporation. This is in contrast to DNA polymerase n and a family A DNA polymerase from E. histolytica which efficiently bypass thymine glycol [3,42]. RB69 DNA polymerase incorporates one nucleotide opposite thymine glycol, but it does not elongate from it indicating that family B DNA polymerase are unable to bypass thymine glycol [45]. Structural studies indicate that the extra methyl group of thymine glycol displaces the incoming template base into a non catalytically competent conformation for further elongation [45]. The low percentage of thymine glycol by EhDNApolB2 may indicate a more flexible active site in comparison to other family B DNA polymerases. As expected EhDNApolB2 is unable to bypass CPD and 6-4 photoproduct (Figure 3 lanes 25 to 33). To date, no replicative DNA polymerase is able to bypass those UV-generated lesions and only specialized DNA polymerases are able to efficiently insert or elongate opposite these lesions [46,47,48].

Fidelity of translesion DNA synthesis by EhDNApolB2
EhDNApolB2 contains an active 39-59 exonuclease domain that in seconds degrades a labeled P-32 primer if a primer-templated duplex is annealed to an abasic site or thymine glycol in the absence of dNTPs at 37uC (data not shown). In order to test the fidelity of lesion bypass by EhDNApolB2 opposite to these lesions we constructed an exonuclease deficient polymerase Asp345Ala mutant that eliminates one of the essential carboxilates of motif ExoIII. We tested the fidelity of lesion bypass using as templates 8oxo guanosine, abasic site, and thymine glycol. To avoid sequence context we used an undamaged 45mer template with identical sequence to the template containing 8-oxo guanosine and abasic site ( Table S4). Using single dNTPs in independent reaction mixtures we observed that EhDNApolB2 incorporates dATP opposite a template thymine ( Figure 4A, lane 4), misincorporates dTTP ( Figure 4A, lane 5) and is unable to incorporate dGTP and dCTP opposite thymine ( Figure 4A, lanes 2 and 3). EhDNApolB2 preferentially incorporates dATP opposite 8-oxo guanosine ( Figure 4A, lane 9) and misincorporates dTTP ( Figure 4A, lane 10). As expected, EhDNApolB2 is unable to incorporate dGTP (Figure 4, lane 7), but unexpectedly EhDNApolB2 does not incorporate dCTP opposite 8-oxo guanosine ( Figure 4A, lane 8). 8-oxo guanosine is a dual coding lesion in which the syn conformation of 8-oxo guanosine mimics a template thymine allowing dATP incorporation opposite the lesion [49,50]. Several DNA polymerases like DNA polymerase I of Bacillus stearothermophilus incorporate dATP with high preference opposite from 8-oxo guanosine [50] by allowing the syn conformation of 8-oxo guanosine and DNA polymerase i selectively incorporates dCTP opposite 8-oxo guanosine by presenting a narrow active site that does not allows the syn conformation of 8-oxo guanosine [51]. Thus, it is possible that EhDNApolB2 presents a specific interaction with 8-oxo guanosine that favors the syn over the anti conformation and thus favors dATP incorporation. DNA polymerases incorporate dATP or dGTP opposite an abasic site following the ''A rule'' [52]. As expect EhDNApolB2 preferentially incorporates dATP opposite an abasic site ( Figure 4A, lane 14). The migration of the fulllength primer extension product is in the same position in the abasic site, thymine and 8-oxoguanosine templates ( Figure 4A,  lanes 1, 6 and 11) indicating that EhDNApolB2 selectively incorporates dATP opposite and abasic site and that the mechanism of bypass does not involve skipping this lesion as is the case for DNA polymerase b [53]. This observation is corroborated by the migration of the 30mer band corresponding to dATP incorporation in thymine, 8-oxoguanosine and abasic site ( Figure 4A, lanes 4, 9, and 14) Exonuclease deficient EhDNApolB2 displays an increased bypass opposite thymine glycol isomers ( Figure 4B, lanes 1, 6  and 11) in comparison to wild type ( Figure 4B, lanes 1 and 6) as is observed for an exonuclease deficient RB69 DNA polymerase [45]. EhDNApolB2 preferentially incorporates dATP and dGTP opposite thymine glycol ( Figure 4B, lanes 2, 5, 7 and 10). Although more detailed kinetic experiments are needed to understand lesion bypass fidelity of EhDNApolB2 the preliminary data indicates that EhDNApolB2 misincorporates thymine opposite a template thymine, abasic site, and 8-oxo guanosine and misincorporates dGTP opposite thymine glycol. This observation is consistent with that fact that family A polymerases involved in lesion bypass also incorporate with low fidelity [54,55,56].
A mutant DTPR2 is active but with hampered polymerase and exonuclease activities EhDNApolB2 contains a TPR2 motif 21 amino acid longer than the one present in Q29 DNA polymerase ( Figure 5A). Mutagenesis experiments have corroborated that TPR2 is involved in processivity and strand displacement in Q29 DNA polymerase [14]. As the TPR2 motif of EhDNApolB2 is twice the size of the orthologous Q29 DNA polymerase we hypothesized that this domain may have the same or novel functions. To determine the putative involvement of TPR2 in EhDNApolB2 lesion bypass we constructed a deletion mutant that eliminates 43 amino acids of TPR2 from EhDNApolB2. This mutants is readily purified using the same purification than wild-type EhDNApolB2 ( Figure 5B,  lanes 1 and 2). We measured the polymerization and exonuclease activities for EhDNApolB2 and DTPR2 using a fixed concentration of both polymerase in a time course reaction with Figure 6. Processivity of EhDNApolB2 in comparison to Q29 DNA polymerase, and its dependence on TPR2. The processivity of wild-type EhDNApolB2 was measured in comparison to Q29 DNA polymerase and DTPR2. Reactions were carried out with 20 nM of the indicated polymerase and 1 nM of c-P 32 17mer primer annealed to circular M13mp18 ssDNA. Aliquots were taken at 10, 20 and 40 min and loaded onto a 6% denaturing polyacrylamide gel. The arrows in the right correspond to the full-length M13 DNA amplification and abortive products. doi:10.1371/journal.pone.0049964.g006 and with out added dNTPs ( Figure 4S). As observed wild-type EhDNApolB2 is able to efficiently elongate a primer template in a time dependent manner, however exonucleolytic degradation bands are also observed and accumulated over time ( Figure 4S, lanes 2 to 5). In the other hand, the DTPR2 mutant is halted after the incorporation of three nucleotides suggesting a processivity problem at this position, perhaps triggered by a favored misinsertion ( Figure 4S, lanes 6 to 9).
The 39-59 exonuclease activity of the DTPR2 mutant in the presence of nucleotides is drastically reduced ( Figure 4S, lanes 6 to 9) suggesting that the extended TPR2 insertion has a role in the conformational changes that occurs during editing and polymerization modes. In order to discern the impact of DTPR2 during exonucleolytic degradation, the same experiment but without added dNTPs was performed. As shown before EhDNApolB2 contains a strong exonuclease activity and is able to efficiently degrade a 24mer to a 4mer in 20 minutes. On the other hand DTPR2 only degrades a 24mer to a 15mer after 20 minutes. ( Figure 4S, lanes 1 to 9).
To further evaluate the role of TPR2 in proofreading and 39-59 exonuclease activity we performed a time course experiment using a mispaired primer-template and a single stranded labeled oligonucelotide ( Figure 5C). During an incubation period from 2 to 10 minutes only 30% of the mispaired 24mer is degraded to products of 22 to 19 nucleotides by DTPR2 ( Figure 5C, lanes 1 to 5). This is in contrast to the almost complete exonucleolytic 24mer degradation to products from 22 to 4 nucleotides by wild type EhDNApolB2 ( Figure 5C, lanes 11 to 15). Suggesting a putative role of EhDNApolB2's TPR2 in exonucleolytic degradation of mispaired primers. Q29DNApolymerase's DTPR2 also presents a decay in exonucleolytic degradation in comparison to wild-type enzyme, however this domain is not involved in coordinating exonuclease and polymerization activities [14,57]. In contrast mutation in intrinsic processive elements like the thioredoxin binding loop of T7 DNA polymerase diminish the extend of exonuclease activity indicating that other polymerases processive elements couple exonuclease and polymerization activities [58]. DTPR2 and wild-type EhDNApolB2 present a similar extended of exonucleolytic degradation at a single stranded DNA oligonucleotide. At the longest incubation time (10 minutes) approximately 60% of the substrate has been degraded. Interestingly DTPR2 degrades to 4-8mers whereas wild-type EhDNA-

Processive DNA polymerization by EhDNApolB2 depends on TPR2
Family B DNA polymerases interact with PCNA to increase their processivity. DNA polymerases a and d from E. histolytica contain canonical PCNA binding motifs and are expected to be highly processive, as PCNA from E. histolytica (EhPCNA) assembles as a trimeric toroid [59]. EhDNApolB2 does not contain a canonical PIP box binding motif and full-length EhDNApolB2 is not stimulated by EhPCNA (data not shown). Q29 DNA polymerase is able to incorporate more than 70 kb of DNA in a single DNA binding event [60] and mutagenesis studies have demonstrated that the TPR2 motif of QDNA polymerase is involved in processivity [14]. To asses the intrinsic processivity of EhDNApolB2 we used a single stranded M13mp18 substrate annealed to a 17mer with equimolar amounts of control Q29 DNA polymerase and DTPR2 ( Figure 6). After 40 minutes, full length M13mp18 is synthesized by Q29 DNA and EhDNApolB2 polymerases and no abortive/distributive products are observed ( Figure 6, lanes 1 to 3 and 5 to 7) in comparison to a control primer template without added polymerase ( Figure 6, lane 4). As observed after 10 minutes, EhDNApolB2 completely extends a M13mp18 substrate and after a period of 20 to 40 minutes is able to perform a second round of synthesis over the same substrate displacing the newly synthesized DNA. Thus, EhDNApolB2 is more processive than Q29 DNA polymerase ( Figure 6, lanes 1 to 3). As expected DTPR2 synthesized DNA in a distributive/ abortive fashion, demonstrating that the TPR2 motif is crucial for processivity in this DNA polymerase ( Figure 6, lanes 8 to 10).

TPR2 is involved in strand displacement
To corroborate the potentially strong strand displacement of EhDNApolB2 we prepared a set of four primer-template constructs in which a DNA polymerase should be able to fill a gap of 1, 3 and 6 nucleotides before displacing a duplex DNA and a control primer-template in which no strand-displacement is needed. After an incubation of 20 minutes, EhDNApolB2 efficiently displaces duplex DNA with gaps of 1, 3 and 6 nts with an efficiency of 21%, 27% and 34% in comparison to the control with out a 39 annealed oligonucleotide that is extended with an efficiency of 32% ( Figure 7A, lanes 1 to 17). In contrast the DTPR2 mutant only synthesizes full-length 45mer when a 39 duplex barrier is not present ( Figure 7B, lanes 1 to 5). In the presence of 1 nt gap, the DTPR2 mutant is halted at 25 nts and 27 nts and only 5% of the substrate is completely extended to the 45mer product. If the gap is of 3 nts, the mutant is halted at 27 nt and 28 nt and only 4% of the substrate is extended and if the gap is of 6 nts the polymerase is halted at 30 nt and only 2% of the substrate is fully extended ( Figure 7B, lanes 6 to 17).

DTPR2 confers lesion bypass opposite an abasic site
Two structural solutions exists to improve the efficiency of DNA polymerases involved in translesion DNA synthesis: one solution is the presence of a wide active site in which a bulky lesion like a thymidine dimer can be easily accommodated, the other solution is the presence of extra insertions, like insertion 2 of DNA polymerase h with respect to other family A DNA polymerases [61,62]. In order to elucidate if the extra-length of TPR2 may have a role in lesion bypass, we carried out a time course DNA lesion bypass opposite an undamaged template, 8-oxoguanosine and an abasic site was assayed side by side with EhDNApolB2 and DTPR2. As previously demonstrated, EhDNApolB2 efficiently bypasses 8-oxoguanosine and abasic site (Figure 8, lanes 1 to  14). Interestingly the DTPR2 mutant efficiently bypasses 8-oxo guanosine but only incorporates 1 nt opposite an abasic site ( Figure 8, lanes 15 to 28). In this experiment, the primer extension efficiencies having as a template cytosine, 8-oxo guanosine and abasic are 28%, 33% and 27% after 20 minutes of incubation (Figure 8, lanes 1 to 14). This is similar to extension efficiencies of DTPR2 in which 21% and 28% of the primer-template extension is observed using cytosine and 8oxoguanosine, but in clear contrast to the extension opposite an abasic site, in which no fully 45mer product is observed. In this case 22% of the substrate is elongated only one nucleotide indicated by an asterisk.
EhDNApolB2 efficiently extends from a primer in which an oligonucleotide containing a 39OH AMP or CMP overlap with an abasic site (Figure 9, lanes 1 to 5 and 11 to 15). In contrast Figure 9. TPR2 is responsible of lesion bypass extension opposite an abasic site. Lesion bypass of wtEhDNApolB2 (lanes 1 to 5 and 11 to 15) and DTPR2 (lanes 6 to 10 and 16 to 20) extending from a primer containing a 39OH purine or a pyrimidine opposite an abasic site. A primer containing a 39OH dAMP (lanes 1 to 10) or dCMP (11 to 20) opposite an abasic site was subject to a time course primer extension reaction from 2.5 to 20 minutes using equal amounts of DNA polymerases and 100 mM dNTPs. The reaction products were run onto a 15% denaturing polyacrylamide gel. doi:10.1371/journal.pone.0049964.g009 DTPR2 is not able to extend this template that mimics the situation in which a nucleotide is incorporated opposite an abasic site. (Figure 9, lanes 6 to 10 and 16 to 20).
Thus, as in other DNA polymerases, EhDNApolB2 incorporates opposite and abasic site and TPR2 is the key element to pass this lesion. A recent report indicates that PCNA confers lesion bypass capabilities to DNA polymerase d opposite an abasic site indicating an intrinsic ability of family B DNA polymerases to bypass this lesion [63]. The presence of the extra 21 amino acids in the TPR2 insert opens the possibility to speculate if this extra amino acids distort the active site to allow that an abasic site can be efficiently used as a non instructive template or if this TPR2 insertion contributes to an increased binding affinity of EhDNA-polB2 that permits the polymerase extension from an abasic site.