Analysis of DNA polymerase ν function in meiotic recombination, immunoglobulin class-switching, and DNA damage tolerance

DNA polymerase ν (pol ν), encoded by the POLN gene, is an A-family DNA polymerase in vertebrates and some other animal lineages. Here we report an in-depth analysis of pol ν–defective mice and human cells. POLN is very weakly expressed in most tissues, with the highest relative expression in testis. We constructed multiple mouse models for Poln disruption and detected no anatomic abnormalities, alterations in lifespan, or changed causes of mortality. Mice with inactive Poln are fertile and have normal testis morphology. However, pol ν–disrupted mice have a modestly reduced crossover frequency at a meiotic recombination hot spot harboring insertion/deletion polymorphisms. These polymorphisms are suggested to generate a looped-out primer and a hairpin structure during recombination, substrates on which pol ν can operate. Pol ν-defective mice had no alteration in DNA end-joining during immunoglobulin class-switching, in contrast to animals defective in the related DNA polymerase θ (pol θ). We examined the response to DNA crosslinking agents, as purified pol ν has some ability to bypass major groove peptide adducts and residues of DNA crosslink repair. Inactivation of Poln in mouse embryonic fibroblasts did not alter cellular sensitivity to mitomycin C, cisplatin, or aldehydes. Depletion of POLN from human cells with shRNA or siRNA did not change cellular sensitivity to mitomycin C or alter the frequency of mitomycin C-induced radial chromosomes. Our results suggest a function of pol ν in meiotic homologous recombination in processing specific substrates. The restricted and more recent evolutionary appearance of pol ν (in comparison to pol θ) supports such a specialized role.


Introduction
In mammalian cells, a diverse group of DNA polymerases carry out genomic DNA replication and genome maintenance. These include the core enzymes for semi-conservative DNA replication (pols α, δ, ε and telomerase), base excision repair (pol β), mitochondrial DNA replication and repair (pol γ and Primpol), non-homologous end-joining and immunological diversity (pols λ, μ, θ and terminal-deoxynucleotidyl transferase), and DNA damage tolerance by translesion synthesis (η, ι, κ, z, and Rev1). Some of these enzymes have roles in more than one pathway of DNA processing [1,2]. The consequences of genetic disruption of an individual DNA polymerase vary widely, ranging from embryonic lethality to no discernable phenotype.
Pol ν, encoded by the POLN gene in mammalian cells, is not assigned in the preceding list, and it is the only identified polymerase for which an analysis of knockout animals remains to be described. The catalytic domain of pol ν belongs to the A-family of DNA polymerases, and is related to the pol domain of pol θ (encoded by mammalian POLQ / Drosophila Mus308) [3][4][5][6]. Pol θ participates in a pathway of DNA double-strand break repair by alternative end joining [7,8], and consequently defects in Mus308 or POLQ confer hypersensitivity to various DNA damaging agents [8][9][10]. The function of pol ν is currently uncertain, but several roles have been suggested. For example, it was reported that siRNA-mediated knockdown of POLN sensitizes human cells to DNA crosslinking agents [11,12].
Mammalian genomes encode three Mus308 homologs: POLN, POLQ, and DNA helicase HELQ [3,13,14]. Pol θ possesses both a helicase-like and a DNA polymerase domain while pol ν has only a DNA polymerase domain and HELQ has only a helicase domain. Our phylogenetic analysis of the distribution of these three genes is shown in Fig 1. POLQ genes are found throughout most of the eukaryotic lineage, in both animals and in plants, and are inferred to be present in the last eukaryotic common ancestor, 1500 million years ago. Even the unicellular green algae Ostreococcus (the smallest free-living eukaryote) contains POLQ. HELQ is found in most eukaryotic branches, but not in plants. HELQ orthologs are also present in both euryarchaea and crenarchaea [15]. In contrast, POLN genes are restricted to a much more limited range of genomes, in some metazoan groups. POLN appeared during genomic evolution much more recently than POLQ. The POLN gene arose, either by gene duplication or horizontal transfer, in the last common ancestor of the metazoa, around 600 million years ago. POLN genes are distinguished by characteristic sequence insertions in the polymerase domain (S1 Fig), with the location of insert 3 distinct from that of POLQ [5]. Genomic DNA sequence databases for the indicated model organisms, accessed from the NCBI Genome portal, were searched using Blastp for each genus, and then individually verified by alignment. For illustrative purposes, the occurrence of each gene was mapped onto the phylogenetic distribution of the universal stress protein (USP) genes [19]. The presence of POLN in the indicated organism is shown by a light blue colored bar, of POLQ by a light green bar and HELQ by a light orange bar. All identifications were crosschecked and confirmed by accessing Gene Tree view in Ensembl and Ensembl Metazoa (metazoa.ensembl.org). POLN Human pol ν (900 amino acid residues) has a strong strand-displacement activity and low fidelity, with exceptionally efficient insertion of T opposite template G in the steady state [4,16,17]. Purified pol ν can bypass the major groove DNA lesion 5S-thymine glycol (Tg) by inserting A, and bypasses major groove DNA-peptide and some DNA-DNA crosslinks [4,18], but does not bypass other DNA modifications including an abasic (AP) site, a cisplatininduced intrastrand d [GpG] crosslink, a cyclobutane pyrimidine dimer, a 6-4 photoproduct, or minor groove DNA-peptide or DNA-DNA crosslinks [4,18]. A unique cavity in the polymerase domain allows pol ν to generate and accommodate a looped-out primer strand [6]. However, it is currently unknown how these unique biochemical properties are implemented in vivo.
To help illuminate pol ν function, we examined POLN gene expression patterns in mice, and phenotypes of mice and cells where POLN was disrupted. The results suggest a specific function in germ cells, but do not support a role for pol ν in tolerance of DNA crosslinks.

Poln knockout mice have normal development and lifespan
To explore possible biological functions of pol ν, we established two different Poln knockout mouse models. In one model, the second exon was deleted in mice of C57BL/6J;129Sv background (Fig 2 and S2 Fig). This exon encodes the ATG initiation codon and the first 46 amino acids, including an N-terminal protein domain that is highly conserved in vertebrates [21]. We avoided targeting the first exon, because POLN shares this exon with HAUS3, an essential gene encoded within the first intron of POLN [21]. In another model, a zinc finger nuclease (ZFN)mediated 4 or 13 bp frameshift was introduced in the DNA polymerase domain of pol ν (FVB/ NCrl background) (Fig 3A and S3 Fig). ZFN-mediated Poln mutant mice are missing critical parts of the DNA polymerase domain including the highly conserved motifs 5 and 6 (Fig 3B). A universally conserved Asp residue in motif 5 of A-family DNA polymerases (D804 in pol ν) that coordinates bivalent metal ions for interaction with an incoming nucleotide, is lost [22]. In all of the engineered mice studied here, a long noncoding RNA gene that overlaps with Poln exons 3-6 remains intact.
Both homozygous Poln-deficient mouse models were viable and fertile. Comparing them to their wild-type littermates, no reproductive or developmental differences were observed. Heterozygous Poln Ex2 mutant mouse crosses yielded pups with normal litter sizes and within expected proportions of genotypes and genders ( Table 1). Homozygous Poln knockout mouse crosses for the ΔEx2, del4 and del13 constructs all yielded normal litter sizes and with normal sized pups at birth. A phenotype analysis, which included a complete gross necropsy, hematology and serum chemistry, full histopathology on all tissues, and survey radiographs was completed with adult ZFN-mediated Poln knockout mice. A comparison between wild-type mice and Poln-deficient mice did not reveal any obvious differences in size, body weight, organ  Table 2), or skeletal appearance. Additionally, no significant differences between genotypes were observed in hematology or serum chemistry analysis ( Table 2).
Poln +/+ , Poln +/ΔEx2 , and Poln ΔEx2/ΔEx2 mice were monitored for over 2 years to assess survival, general health, and tumor incidence. Loss of functional Poln did not affect overall survival compared to wild-type and heterozygous littermate controls (Fig 4). There were no significant differences in tumor prevalence, multiplicity or distribution of tumor types (Tables  3 and 4). Non-neoplastic lesions were of types and incidence commonly found in aging laboratory mice, and there appeared to be no pattern of susceptibility related to genotype ( Table 5) or sex. All mice had a variety of age-related lesions that were considered incidental and not included in non-neoplastic diagnoses. These included atrophies of testicles, ovaries, uterus, and thymus, mild multifocal lymphoid infiltrates in various organs, and mild glomerulopathy. Similarly, no obvious alterations in lifespan, general health or cause of death were noted in Poln del4 and del13 mice.
Poln has high relative expression in spermatocytes, but is not essential for spermatogenesis and meiotic recombination We previously reported that the POLN transcript is most prominent in adult testis from the human, mouse, and zebrafish [3,21]. To quantify absolute transcript levels, we compared expression of Poln, Polq, and the replicative polymerase Pold1 in different mouse tissues. Poln expression was highest in testis by an order of magnitude, much higher than Polq and Pold1 in this organ. However, Poln transcript was almost undetectable in other tissues, in contrast to Polq and Pold1 (Fig 5A). To investigate when Poln starts expressing during testis development, we isolated testes from young male mice of various ages. Poln expression gradually increased and was higher than Polq and Pold1 after 16-days, when the population of spermatocytes in the pachytene stage increases [23]. In contrast, the amounts of mRNA of Polq and Pold1 were not markedly increased during testis development (Fig 5B). Fractions enriched in spermatids and pachytene spermatocytes were prepared by centrifugal elutriation, and Poln expression was higher in the pachytene-enriched fraction (Fig 5C). These data suggested a possible meiosis-related function of pol ν in testis.
Testis morphology was normal in Poln knockout mice (Fig 6A), and testes size measurements, histology and breeding experiments described above did not suggest a marked perturbation of spermato-and spermiogenesis. To test for an overt role of pol ν in meiotic recombination, chromosome spreads were stained with SYCP3 (a component of the synaptonemal complex) and MLH1 (a crossover marker) to assess progression through meiotic prophase I (Fig 6B). MLH1 foci numbers per nucleus were not statistically different between pol ν proficient and Poln del4/del4 mice (Fig 6C). Poln ΔEx2/ΔEx2 spermatocytes had one less MLH1 focus per nucleus than controls (p = 0.03) (Fig 6D). The results suggest that while pol ν is not essential for crossing over or meiosis, it may be required for efficient crossing over at specific loci, at least in some genetic backgrounds (see legend to S4 Fig) triangles and loxP sites by blue triangles. First, the neomycin positive selection cassette (neo) was excised by breeding with C57BL/6J Flp deleter mice. A subsequent cross with Cre-expressing mice led to excision of the Poln exon 2 and surrounding intron sequence. Heterozygous Poln knockout mice (Poln +/ΔEx2 ) were then used for breeding. (B) Example of Southern blot analysis of (left) the recombined locus (neo+), (middle) the Flpmediated excised locus and (right) the constitutive Poln excised locus. Genomic DNA of the tested animals was compared with C57BL/6J wild-type genomic DNA (WT). AflII-digested DNA was blotted on a nylon membrane and hybridized with the 5' probe. Restriction fragment sizes were: recombined neo+ locus 10.5 kb, Flp-mediated excised locus 8.8 kb, constitutive Poln excised locus 7.7 kb and the wild-type locus 6.0 kb. Genomic DNA was further analyzed extensively and confirmed by specific PCR assays and complete DNA sequencing as described in the Materials and Methods. https://doi.org/10.1371/journal.pgen.1006818.g002 Generation of polν defective mammalian cells Possible function of pol ν at a meiotic recombination hotspot We examined the potential role of pol ν in meiotic homologous recombination-associated DNA synthesis. Crossovers, arising by recombination at one meiotic hotspot A3 [24,25], were Generation of polν defective mammalian cells isolated from sperm of Poln ΔEx2/ΔEx2 C57BL/6J x DBA/2J and Poln del4/del4 FVB/NCrl x DBA/2J F1 hybrid mice, and control animals. The crossover frequency at A3 in FVB/NCrl x DBA/2J is half that of C57BL/6J x DBA/2J (Fig 6E and 6F). The A3 locus in FVB/NCrl lacks a high affinity PRDM9 binding site (S4 Fig) and, therefore it likely does not receive meiotic DSBs to initiate recombination [26]. As such, the FVB/NCrl x DBA/2J F1 hybrid receives meiotic DSBs on only the DBA/2J allele, rather than at both the C57BL/6J and DBA/2J alleles as in the C57BL/6J x DBA/2J F1 hybrid. Consistent with this model, the A3 locus shows reciprocal crossover asymmetry, an altered distribution of crossovers depending upon their orientation, which is a hallmark of meiotic hotspots with biased DSB formation on only one of the two parental alleles (S4 Fig). A total of 155, 157, 185, 180 crossovers from Poln del4/del4 , Poln +/del4 , Poln ΔEx2/ΔEx2 , and Poln +/ΔEx2 males were isolated and mapped. Consistent with reduced MLH1 foci, the Poisson corrected crossover frequency in male Poln ΔEx2/ΔEx2 mice was 80% of that in heterozygous littermates (p = 0.04) (Fig 6F). The crossover frequency in Poln del4/del4 mutants was not significantly reduced compared to controls (p = 0.09, Fig 6E), but combining both datasets showed a statistically significant (p = 0.01) reduction in Poln homozygous mutants to about 80% of the control frequency (S5 Fig).
We further examined the so-called 'crossover refractory zone' (CRZ) at the A3 hotspot in Poln mutant mice. Crossover formation in this 258 bp region of the A3 hotspot is strongly inhibited, even though DSBs are formed there abundantly [24,27]. The region contains secondary structure-forming sequence that may inhibit heteroduplex formation [24]. The combined Poln mutant data (S5 Fig) showed a statistically significant (p = 0.04) 50% reduction in crossover frequency at this region (10 of 340 crossovers in knockouts, 21 out of 337 crossovers in controls).

Poln deficiency does not perturb transcription in testis
Whole genome transcriptome analysis by RNAseq was performed to determine whether Poln disruption influences gene expression in testis from Poln ΔEx2/ΔEx2 mice. Using rRNA-depleted total RNA, we obtained an average of 65 million reads per sample (range 53 to 73 million) with an average mapping rate of 90% to the reference mouse genome. Poln disruption did not perturb overall transcription (Fig 7A). Only a few genes had significantly altered expression, at a false discovery rate (FDR) 0.05. Expression of only four of these genes was increased more than 2-fold (S1 Table), comprising two that encode proteins (Lgi2 and 9330182L06Rik), and two specifying noncoding RNAs (BB283400 and Speer4cos). Expression of only two genes was reduced more than 2-fold, Egfem1 and Poln itself. Because these genes have no close physical linkage and their functions are unknown, we have not pursued them further at present. As expected, expression of the targeted exon 2 of Poln was not detected in Poln ΔEx2/ΔEx2 mice. All other Poln exons showed greatly reduced or absent expression, likely related to the disruption of normal splicing. Haus3 expression was not influenced by deletion of the targeted Poln exon (Fig 7B).

Poln deficiency does not influence radiation-induced DNA damage levels in spermatids
We asked whether pol ν influences levels of ionizing radiation-induced DNA damage in testis. This could be the case if pol ν is directly involved in DSB repair (by analogy with the related pol θ) or if pol ν is necessary for translesion DNA synthesis of radiation-induced lesions such as thymine glycol (Tg). Pol ν is proficient in bypass of Tg, which blocks the progression of replicative DNA polymerases [4]. A dose dependent increase in γ-H2AX foci (a surrogate for DNA damage including DSB) was observed in wild-type, Poln ΔEx2/ΔEx2 , and Polq -/mice ( Fig  8A). The number of γ-H2AX foci in the absence of irradiation was higher in Polq -/mice than in wild-type or Poln ΔEx2/ΔEx2 mice. This is consistent with the higher basal level of γ-H2AX foci in pol θ-suppressed human cells [28]. Following irradiation of mice with 1 or 2 Gy, the average γ-H2AX foci numbers in round spermatids of testis sections was measured [29]. After 5 hr, the number of foci was reduced to a similar level in all mice (Fig 8B), irrespective of Poln and Polq status. As a control experiment, blood was taken from killed mice after irradiation and micronuclei were measured in reticulocytes. Frequencies of micronuclei were increased in Polq -/mice after IR, as previously reported [10,30,31]. Poln status did not influence the level of micronuclei in reticulocytes (Fig 8C). We also examined Poln ΔEx2/ΔEx2 Polq -/double knockout mice and found that Poln deletion did not further increase the frequency of micronuclei above that in Polq -/reticulocytes (Fig 8C).

Poln disruption does not influence mouse immunoglobulin classswitching
Because pol θ can participate in end-joining of DNA breaks produced during immunoglobulin class-switch recombination (CSR) [8], we tested whether pol ν is also involved in CSR. Naïve B cells isolated from the spleens of wild-type and Poln ΔEx2/ΔEx2 mice were stimulated for IgM to IgG class switching, and then the fraction of IgG1-positive B cells was measured by flow cytometry. Parallel B-cell cultures were incubated with NU7026, an inhibitor of DNA-PKcs that increases the proportion of CSR with >1 bp insertion at the junction [8,32]. We found that stimulated B cells from Poln-proficient and deficient mice had similar overall frequencies of IgG1 (6.1%). Inhibition of DNA-PKcs increased the frequency of CSR in both genotypes by 2-fold (11.7% in wild-type, 12.5% in Poln ΔEx2/ΔEx2 ) (S2 Table), as observed previously with wild-type and Polq -/mice [8]. The Sμ-Sγ1 junction was then sequenced from 100 clones of each group of IgG1-positive B cells. Insertions of >1 bp at Sμ-Sγ1 junctions were pol θ-dependent [8], but pol ν-independent (Fig 9A and 9B). We also performed qPCR analysis of transcript expression in splenic B cells for Polq, Poln, Helq, Haus3, and Pold1. Interestingly, among those genes only Polq expression was increased after B-cell activation by lipopolysaccharide and interleukin 4 treatment (Fig 9C).

Poln deficient cells have a normal response to DNA damaging agents
Cells with inactivation of pol θ or its ortholog HELQ show increased sensitivity to DNA damaging agents, such as bleomycin (pol θ) [8] and mitomycin C or cisplatin (HELQ) [33][34][35]. To determine whether inactivation of pol ν also sensitizes cells to DNA damage, we examined the relative sensitivity of Poln ΔEx2/ΔEx2 MEFs to mitomycin C (MMC), cisplatin, bleomycin, hydroxyurea, olaparib, formaldehyde and acetaldehyde. Aldehyde sensitivity was tested because of the observation that purified pol ν can bypass some DNA-peptide crosslinks [18]. However, we detected no hypersensitivity of immortalized MEFs to any of the tested DNA damaging agents (Fig 10). We also examined primary Poln del4/del4 MEFs and found no hypersensitivity to MMC, olaparib, etoposide or 5-fluorouracil (S6 Fig). Rev3L null MEFs were used as positive controls in the same experiment, and they showed sensitivity to MMC as reported [36]. We generated Poln ΔEx2/ΔEx2 Polq −/− MEFs by crossing Poln ΔEx2/ΔEx2 and Polq −/− mice, isolating double mutant MEFs from the progeny and immortalizing with SV40 Tag. Polq −/− MEFs were more sensitive to bleomycin, as reported [8]. Poln ΔEx2/ΔEx2 MEFs were not more sensitive to bleomycin and Poln ablation did not further increase the bleomycin sensitivity of These results contrast with previous reports that siRNA directed against POLN causes an increased sensitivity to MMC and cisplatin [11,12]. Therefore, we performed siRNA-mediated knockdown in human cells using the previously reported targeting sequence [11]. In that study, the efficiency of siRNA-mediated knockdown was assessed using an antibody raised against pol ν. We have developed~80 Pol ν monoclonal antibodies and two polyclonal antibodies against pol ν -77 [37], and identified several of them that can immunoblot and immunoprecipitate overexpressed pol ν in human cells (e.g. Mab#40 in S8 Fig). However, we have not been able to detect endogenous pol ν with these or other antibodies, and so we turned to a different approach to monitor the efficiency of siRNA-mediated POLN knockdown. Pol ν was expressed in a doxycycline-inducible manner in 293T-Rex cells. We confirmed complete knockdown of tagged protein expression, showing that the POLN-specific siRNA construct Generation of polν defective mammalian cells very effectively inactivates POLN mRNA (Fig 11A, top panel). We also performed quantitative PCR assays for POLN and HAUS3 as described [21] and detected significant reduction of doxycycline-induced POLN transcript (P = 0.01) as well as endogenous POLN transcript (P = 0.03) by the POLN-specific siRNA (S9 Fig). Note that only incomplete transcripts of POLN have been detected in 293T cells [21]. HAUS3 expression was not affected by this siRNA (S9 Fig). Nevertheless, neither pol ν-depletion (siN) nor pol ν-overexpression (293T-REx(POLN) + Dox, siC) had any influence on mitomycin C sensitivity in human cells (Fig 11B and S8  Fig). It was also proposed that pol ν depletion caused an increased formation of radial chromosomes in human cells treated with MMC [11]. We therefore analyzed metaphase spreads of 293T-REx depleted of pol ν, and 293FT cells depleted of pol ν, FANCA, or FANCD2 (Fig 11A).
Suppression of FANCA and FANCD2 resulted in an increased frequency of radial chromosomal formation after treatment with mitomycin C as expected for these Fanconi anemia gene products. However, pol ν depletion did not increase the frequency of radial chromosome formation (Fig 11C and 11D).

Discussion
Viability and lifespan of mice lacking pol ν The investigations presented here answer many long-standing questions about pol ν function and expression. An analysis of mice with genetic disruptions of Poln has not been reported previously. Because Poln is expressed only in some metazoan lineages, many widely-used model organisms are not relevant for biological analysis of pol ν. We found that Poln is not essential for mouse embryonic development or for viability. Three independently established and targeted mouse models that disrupt Poln were fully viable, in C57BL/6J;129Sv and FVB/ NCrl backgrounds. The Poln deficient mice have a normal lifespan and lack abnormalities for many investigated anatomic features. Consistent with this, the International Mouse Phenotyping Consortium (mousephenotype.org) has recently listed mice with a homozygous knockout of exon 6. No abnormalities in adults or embryos were detected in a battery of anatomical, physiological and behavioral tests. No expression of Poln was detected in embryos, also consistent with our reported observations [21].

Evidence for a specialized function of pol ν in testis
Our observations point to a function of pol ν in the testis. We discovered that Poln is uniquely upregulated during mouse testicular development and that it is enriched in spermatocytes. This may be a conserved feature in organisms where POLN occurs. For example, we found that POLN expression was essentially undetectable in somatic adult cells or embryonic cells from the zebrafish, but was present in adult zebrafish testis [21]. Further, the phylogenetic analysis described here shows that POLN genes exist only in metazoan animals that use sexual reproduction and deploy sperm. This is consistent with a specialized function for pol ν in discrete spermatocyte-forming organs. The evolutionary distribution of POLN is a contrast to the wide distribution of POLQ genes in animals and plants, including many unicellular species and some that reproduce asexually. The DNA polymerase activity of pol ν appears to be critical for its function, as all organisms encoding the enzyme retain the six conserved DNA polymerase motifs with the residues necessary for catalytic activity. Expression of recombinant protein from the human, mouse and zebrafish POLN cDNAs all produce active DNA polymerase [21]. Importantly, it is uncertain whether the pol ν protein is expressed significantly in mammalian somatic cells or tissues. Most POLN transcripts in somatic human cells are inactive, alternatively spliced variants [3]. The protein is not detectable with the antibodies currently available, although mass spectrometry experiments detect a few peptides representing pol ν in various mouse, human, and rat tissues including brain, heart, kidney, liver, lung, pancreas, spleen, and testis [38]. The low level expression of mostly inactive transcripts is consistent with the lack of phenotypes in somatically derived cells, described here.
Pol ν may be important at specific recombination hot spots where DNA forms unique DNA secondary structures which provide challenges for other DNA polymerases. We detected a modest but significant reduction in meiotic recombination frequencies at the A3 hot spot in Poln deficient mice. The frequency was lower in the crossover refractory zone of A3 in Poln deficient mice. This zone harbors insertion/deletion polymorphisms and possibly forms a hairpin structure inhibiting replicative DNA polymerases [24]. The strand displacement activity of pol ν [4] may be effective in helping to synthesize the secondary structure-forming sequence. In fact, the crystal structure of pol ν reveals a cavity in the polymerase domain, which could accommodate a looped-out primer strand [6]. Such a looped-out primer could be formed during annealing of insertion/deletion polymorphisms during meiotic recombination.
Further, we previously reported proteins associated with pol ν when it was overexpressed in human cells. Mass spectrometry analysis showed associations of pol ν with homologous recombination factors including BRCA1, FANCJ, BRCA2, and PALB2 [21]. This is consistent with the participation of pol ν in a specialized homologous recombination reaction.
Pol ν-defective mice and human cells have a normal response to DNA crosslinking agents Our results show that mammalian cells have normal responses to DNA crosslinking agent exposure following pol ν elimination or depletion. This contrasts with the reported increased sensitivity to the DNA crosslinking agents MMC and cisplatin in human cells after an siRNAmediated knockdown of POLN [11,12], Our conclusions are based on several independent knockouts of Poln function in the mouse, and on controlled siRNA and shRNA-depletion of POLN in human cells. These cells showed normal survival responses to crosslinks, and no evidence of increased frequency of MMC-induced radial chromosomes. We did not detect cellular crosslink-sensitivity after suppression of POLN. In studying pol ν, it is also crucially important to insure, as we have done here, that knockdown of POLN does not interfere with the expression of the essential HAUS3 gene encoded in intron 1 of vertebrate POLN. Further, we emphasize that antibodies cannot be used currently to monitor endogenous knockdown of POLN. Finally, we have described that there is little evidence for expression of pol ν in somatic cells, which would explain the lack of phenotypes in non-germline cells when the gene is disrupted. Current evidence points towards a specific germ cell function.
Although pol ν shares homology with the DNA polymerase domain of Drosophila Mus308, which is involved in resistance to DNA crosslinking agents [9], it appears that a different Mus308 homolog is more involved in crosslink sensitivity in mammalian cells. HELQ, which shows similarity to the Mus308 helicase domain, is involved in DNA crosslink tolerance [34,33,35]. Our results are more in line with the lack of sensitivity to DNA crosslinking agents that was reported for POLN −/− chicken DT40 cells [39,40]. The ability of pol ν to perform limited bypass of a major groove DNA-peptide adduct (N 6 -A-peptide) and a DNA-DNA (N 6 -A-N 6 -A) crosslink residue [18] may be more broadly related to properties of pol ν that allow it to bypass blocks such as the secondary DNA structures that exist within meiotic recombination hotspots.

Perspective
In many organisms, it is a common observation that knocking out a single DNA repair protein generates little or no DNA damage or growth-related phenotype. Double mutations, or disruption of backup pathways, are frequently necessary to reveal significant phenotypes. For example, no phenotype of pol ι disruption has been detected as a single mutant. In contrast, pol η /pol ι double-deficient mice show an altered spectrum of UV radiation-induced tumors [41,42]. We constructed pol ν /pol θ double-defective mice, and as described here, this double mutation did not exacerbate cellular sensitivity to DNA damaging agents beyond what we observed with pol θ defective animals. Further screening of double mutants seems a possible approach to unveil the biological function of pol ν in vertebrates. The present work provides a major resource and foundation that paves the way for further exploration of the function of this unique human enzyme.

Ethics statement
Research mice were handled according to the policies of the MD Anderson Cancer Center Institutional Animal Care and Use Committee, under approved protocol number 00001119-RN00.
Targeting the Poln ΔEx2 allele disrupted from exon 2 The strategy for cre-loxP based knockout and targeting vector construction was designed and performed by genOway (Lyon, France). The genomic region of interest containing the murine Poln locus was isolated by PCR from genOway's 129Sv BAC library. PCR fragments were subcloned into the pCR4-TOPO vector (Invitrogen). The following primers were used. For the short homology arm: 5'-CATAACAGGTCAGAGTCACAAAACAGATATGC-3'and 5'-TAT CTCACAGACATCAAAACCTACACATGCC-3'; for proximal long homology arm: 5'-CCAG GTAATTTAGATGTGTGAACCAGATGC-3' and 5'-ACAAACTTTCCAGAACAAGGACA ATGACC-3'; for distal long homology arm: 5'-GGGACAGAATAGAAACAAAATGACAAA ATAGACC-3' and 5'-GCTTATGCAAGTAGAGATTCAAAGTTGATGTAAGG-3'. The resulting sequenced clones (containing intron 1 to intron 2) were used to construct the targeting vector. A region including exon 2 was flanked in the adjacent introns by a Neo cassette (LoxP transcription, comparing Poln +/+ and Poln ΔEx2/ΔEx2 testes. Positive values indicate genes with higher transcription in the absence of pol ν, negative values indicate lower transcription. Black dots indicate genes with no significant differential transcription; red dots indicate differentially expressed genes (p < 0.05). (B) Genome browser screenshots of the Poln /Haus3 locus showing RNA-seq data for four wild-type and four Poln disrupted mice. The red box highlights exon 2 knocked out in the Poln mutant mice. This exon contains the initiation codon and a conserved N-terminal domain [21]. The globular DNA polymerase / pseudoexonuclease domain [6] is encoded by exons 4-25. A long non-coding RNA gene (4930557 J02Rik) is specified by the strand complementary to Poln exons 3-6.
https://doi.org/10.1371/journal.pgen.1006818.g007 (B) DNA repair kinetics. At defined time-points after irradiation with 2 Gy, round spermatids were analyzed by counting foci from three mice ± SE; *significant difference (p < 0.01) compared with wild-type control. (C) Spontaneous and radiation-induced micronucleus frequencies were measured in reticulocytes. Data are presented as means from three mice ± SE; **** significant difference (p < 0.0001) compared with wild-type control. site-FRT site-PGK promoter-Neo cDNA-FRT site) and by a distal loxP site (Fig 2A). This allowed generation of constitutive and conditional knockout lines which had deleted the 145 bp Poln exon 2 containing the translational initiation codon, 760 bp of the upstream intron and 63 bp of the downstream intron. In the knockout, the deleted region is replaced by a 50 bp fragment containing the distal loxP site, simultaneously disrupting an endogenous AflII restriction enzyme site.  PCR analysis, one primer pair was designed to amplify sequences spanning the 5' homology region. This primer pair designed to specifically amplify the targeted locus was: sense, 5' GAA AAGCCTCGAAGATATGGGCACC-3', anti-sense (Neo cassette) 5'-GCCTCCCCTACCC GGTAGAATTAGATC-3'. Targeting was confirmed by Southern blot analysis using internal (B) Cells were exposed to the indicated doses of mitomycin C for 48 hr. Viability was determined by measuring ATP content as described in Materials and Methods. The mean of three separately plated and treated plates is shown, with SD indicated by error bars. (C) Mitomycin C -induced radial chromosomes in the cells shown in A. The cells were exposed to 40 ng/mL mitomycin C for 48 hr. 100 metaphases per sample were analyzed. (D) Arrows indicate radial chromosomes in POLN-depleted 293FT cells exposed to mitomycin C. Scale bar: 10μm. and external probes on both 3' and 5' ends. Two clones were identified as correctly targeted at the Poln locus.

Zinc-finger nuclease (ZFN) targeting
Sixteen ZFN pairs targeting Poln were designed and assembled by PCR subcloning into the pZFN plasmid by Sigma. All pairs were tested for efficiency of generating double strand breaks using the Surveyor Mutation assay in cultured mouse Neuro2A cells. The selected ZFN pair, targeting exon 21, was used for subsequent microinjections into FVB/NCrl embryos. The ZFN target site (cleavage site in lowercase) is: 5'-GGCCTCTCCCGAGGAtctgtGCACAGGACCAG-CAG-3'. Validation used forward primer: 5'-CCCTGGGAATACTTGGGACT and reverse primer: 5'-ACTACCAGGCAGGACAGGTG.
Embryos were obtained from superovulated Charles River FVB/NCrl female mice. ZFN pairs were microinjected and transferred into pseudopregnant foster females. Thirty-six pups were born. The pups were sampled for genotyping at approximately 3 weeks of age. Four founders (11% efficiency) have been identified. Three of the four founders resulted in early stop codons. Founder #22 has a 4 base pair deletion (FVB/NCrl.Cg-Poln em1Rwd designated as Poln-del4 allele). Founder #26 has a 13 base pair deletion (FVB/NCrl.Cg-Poln em2Rwd designated as Poln del13 allele) and #27 is chimeric with a 4 and 13 base pair deletion. An ApaL1 restriction site is present in the ZFN targeted site and the sequence is lost in the 4-bp and 13-bp-deleted allele. This is used to validate the genotyping result (S3 Fig).

Mouse aging study
Thirty mice (15 males and 15 females) of each genotype (Poln ΔEx2/ΔEx2 , Poln +/ΔEx2 , and Poln +/+ ) were monitored over time for tumor incidence and survival. Mice were housed in an SPF modified barrier facility under standard light cycle and temperature conditions. Sentinal animals were free of multiple murine infectious and parasitic agents including Helicobacter and mouse parvoviruses. Mice were fed irradiated standard rodent diet (Harlan Laboratories Irradiated Teklad 22/5 Rodent Diet #8940) ad libitum and provided with reverse osmosis purified water via an automated system. Mice were monitored from approximately 3 weeks of age until they reached a moribund state. Mice were monitored daily for overall health status. Additionally, body condition score [43] and body weight were assessed weekly and recorded. Mice were removed from the study if any of the following criteria were met: Loss of greater than 20% body weight in one week, body condition score of 1 or 2 [43], respiratory or ambulatory difficulties, or non-healing dermatitis.
Mice removed from study were killed by carbon dioxide inhalation followed by cervical dislocation before complete necropsies were performed. Body weight and weights of the following organs were obtained: spleen, kidneys, liver, adrenals, testes, heart, thymus, and brain. All tissues were examined grossly, fixed for 24-48 hr in neutral buffered formalin, and stored in 70% ethanol. For all mice, H&E-stained slides of the following tissues were prepared by standard techniques and examined by a board-certified veterinary pathologist: kidneys, liver, heart, lungs, thymus, spleen, testes or ovaries, brain, all gross lesions. For all suspected lesions of histiocytic sarcoma or lymphoma, immunohistochemistry was performed to identify T cells (CD3), B-cell (CD45R), and macrophages (F4/80 or MAC). Additional immunohistochemistry was performed as required for specific diagnosis.

Serum chemistry
Physiological serum parameters were measured using the VetScan VS2 Analyzer with the Comprehensive Diagnostic Profile reagent rotor (Abaxis).

Isolation of the mouse Poln cDNA
We prepared a cDNA library was prepared from 129Sv mouse testis. The full-length open reading frame (ORF) was amplified with primers (5'-caccaaaatggaaaattatgaggcatgtg and 5'-caggagatcctgggctcagccgactacaaagacgatgacgacaagtaggaattcatatat), and cloned into pENTR/ D-TOPO vectors and then recombined into pDEST17 vectors (Invitrogen).

Quantitative PCR (qPCR) assay
Total RNA was extracted using TRIzol (ThermoFisher Scientific), and RNA integrity assessed using the Agilent 2100 bioanalyzer (Agilent Technologies, Inc.). Total RNA (1 μg) was used as template to synthesize cDNA with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). qPCR was then performed on the ABI 7900HT Fast Real-Time PCR System (Applied Biosystems). Custom assays for mouse Poln, Polq, and Pold1 were designed using FileBuilder 3.1 software (Applied Biosystems) and ordered from Applied Biosystems. TaqMan primer and probe sets for each gene are shown in Table 6. The absolute quantity (AQ) of transcripts for Poln and Polq was determined using the generated standard curves and Applied Biosystems' Sequence Detection Software version 2.2.2 (ABI). Standard curves for each gene were determined using the plasmids pDEST17 carrying cDNA coding full length of Poln, and Polq clone (MGC:189905, IMAGE:9088092). For relative quantification, TaqMan primers and the probe set for Gapdh were purchased from Applied Biosystems. Triplicate qPCR reactions each containing cDNA representing 40 ng of reverse-transcribed total RNA were then assayed for transcript quantity with Gapdh serving as endogenous controls to normalize input RNA levels.

Preparation of spermatogenic cells
Cell suspensions were prepared from 25 adult mice (~100 days old) by previously published methods [44,45]. Seminiferous tubules were isolated by incubating the decapsulated testes with collagenase (0.5 mg/mL) and DNase I (200 μg/mL) in enriched DMEM/F12 (Invitrogen), to which 0.1 mM non-essential amino acids (Invitrogen), 1 mM sodium pyruvate (Invitrogen), and 5 mM sodium lactate (Sigma) were added. This decapsulated testis tissue was shaken for 15 min at 35˚C in a water bath, until it was mostly dispersed into tubules. The dispersed tubules were allowed to settle and, after removal of the supernatant, were resuspended in 32 mL of DMEM/F12 solution. In each of four 50 mL tubes, 8 mL of this suspension was layered onto 40 mL of 5% Percoll solution (Sigma), and the tubules were allowed to settle until most of the larger tubules and clumps were at the bottom. The supernatants were removed and the settled tubules were washed with DMEM/F12 solution and then further digested with trypsin (1 mg/mL) and DNase I (200 μg/mL) in enriched DMEM/F12 for 20 min at 35˚C with shaking. Fetal bovine serum was added to 10%, and the cells were dispersed by pipetting. Total cell suspensions were separated by centrifugal elutriation (JE-6B rotor, Beckman) to obtain fractions enriched in spermatids (flow rate: 15.6 mL/min, rotor speed: 2250 rpm) and pachytene primary spermatocytes (37 mL/min, 2250 rpm). The purified pachytene fraction was obtained by plating the Percoll-enriched fraction on DSA (Sigma) coated dishes, to which Sertoli cells bind strongly, and the pachytene spermatocytes were recovered in the unbound cell fraction. The Sertoli cells recovered from the elutriator were purified by plating them on the DSA-coated dishes, removing the unbound and loosely bound cells, and directly extracting the bound Sertoli cells with RLT lysis buffer (QIAGEN). The purity of each fraction was initially determined by cell smears stained with periodic acid Schiff-hematoxylin.

Testis morphology
Testes were removed from killed mice and fixed in neutral-buffered formalin for 24 hr before being processed for histologic sectioning (4 μm) and stained with hematoxylin and eosin (H&E). The samples were analyzed using a BX41 Olympus microscope with 10X objective.

Generation of F1 hybrid mice
In order to detect recombinant molecules at the A3 locus in sperm, F1 hybrid animals carrying heterozygous A3 alleles were generated. The Poln ΔEx2 allele in the C57BL/6J background and the Table 6. TaqMan primer and probe sets for mouse.

TaqMan primer
TaqMan probe

Spermatocyte spreads and immunofluorescence
Spermatocyte spreads were prepared according to previously published protocol [46]. Briefly, slides rinsed with PBS were blocked for 30 min at room temperature in Gelatin Block Solution (GBS: 0.2% v/v fish gelatin (Sigma G-7765), 0.2% v/v IgG-free BSA (Jackson ImmunoResearch 001-000-162), 0.05% w/v Tween-20). Next, covered slides were incubated overnight at 4˚C with 100 μl of primary antibody diluted in GBS. Slides were then washed with GBS on a rotating platform shaker (one quick rinse followed by 5, 10 and 15 min rinses). Next, covered slides were incubated for 45 min at 37˚C with 100 μl of secondary antibody diluted ( DNA isolation and determination of amplifiable DNA concentration DNA samples were isolated from sperm from epididymides of adult (2-5 m.o.) F1 hybrid animals, as previously described [24,25]. Two animals were analyzed for each allele, per genotype. Briefly, a standard phenol/ chloroform/ isoamyl alcohol DNA isolation protocol was followed by ethanol precipitation. An aliquot of each DNA sample was used to quantify DNA concentration by UV absorbance and by comparison to a dilution series by agarose gel electrophoresis, using high quality sperm DNA of defined concentration. The number of amplifiable DNA molecules/pg was determined by performing 12-24 PCR reactions per sample seeded with 12 pg per reaction (equivalent to 2 amplifiable molecules/well).

Crossover assay
The crossover assay for A3 was as described previously [24,25]. Briefly, recombinant molecules were identified after two rounds of nested PCR by 0.8% agarose gel electrophoresis and positive reactions (putative crossovers) were then PCR-amplified, together with positive and negative controls. Products in a 96-well format were dot-blotted onto nylon hybridization membrane (Roche, 11417240001) and genotyped by Southern blotting with allele-specific oligonucleotide (ASO) probes. Somatic DNA from spleen or liver for each assayed animal was used as a negative control at total DNA inputs equivalent to or higher than sperm DNA. No crossovers were detected in somatic controls (frequency < 1.05 x 10 −5 ).

Genotyping with ASO probes
A detailed version of the Southern blotting protocol for the A3 hotspot has been described [24,25]. Briefly, ASO probes were radiolabeled using T4 polynucleotide kinase and hybridized with nylon membranes containing PCR products from the crossover assay. After hybridization, blots were exposed for 4-5 hr on phosphorimager screens and scanned using the Typhoon FLA 9500 phosphorimager (GE Healthcare). Scans were then scored and positive signals used to generate crossover breakpoint maps. Each ASO is an 18-bp oligonucleotide designed to specifically hybridize with one of the two parental genotypes. At the A3 locus, most of the FVB/NCrl sequence is identical to the C57BL/6J strain.
Induction and repair of radiation-induced γ-H2AX foci in round spermatids Wild-type and Poln ΔEx2/ΔEx2 mice were generated from Poln heterozygous knockout crosses, and Polq -/mice were generated from Polq heterozygous knockout crosses. 8-12 week-old mice received whole-body irradiation with 1 Gy or 2 Gy at 2 Gy/min, 160 kV peak energy (Rad Source 2000 irradiator, Suwanee, GA). For DSB induction, three wild-type mice per dose were analyzed at 30 min post-irradiation. For the DSB repair kinetics three mice per time-point were analyzed at 0.5, 5, 24, and 48 hr after irradiation with 2 Gy. In each experiment three mock-irradiated mice served as controls. After animals were sacrificed, testes were immediately removed and placed in fixative. Formalin-fixed tissues were embedded in paraffin and sectioned at an average thickness of 4 μm. Tissue sections were incubated with primary antibody against γ-H2AX (Bethyl) followed by Alexa Fluor-488-conjugated secondary antibody (Invitrogen). Finally, sections were mounted in VECTASHIELD mounting medium with 4 0 ,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). For quantitative analysis, radiation-induced foci were counted by eye using a Leica DMI 6000 microscope equipped with a 63X oil objective. Tissue sections were incubated with primary antibody against γ-H2AX. To evaluate potential differences in DSB repair kinetics, the two-way ANOVA test was performed for each dose and repair-time. The criterion for statistical significance was p 0.05.

RNA-seq analysis
For RNA-seq analysis, four biological replicates were prepared for Poln +/+ and Poln ΔEx2/ΔEx2 mice. RNA was purified from testis of 68-day old male mice with RNeasy kit (QIAGEN) with on-column DNase treatment. The libraries were prepared using the Illumina TruSeq stranded total RNA kit according to the manufacturer's protocol, except that the PCR amplification cycle was reduced to 10. The libraries were sequenced on HiSeq 2000 (Illumina), generating 53-73 million pairs of 75 bp reads per sample. Each pair of reads represents a cDNA fragment from the library. The reads were mapped to mouse genome (mm10) by TopHat (Version 2.0.7) [47]. By reads, the overall mapping rate is 90-96%. 83-93% fragments have both ends mapped to mouse genome. The number of fragments in each known gene from RefSeq database 48 (downloaded from UCSC Genome Browser on March 6, 2013) was enumerated using htseq-count from HTSeq package (version 0.5.3p9) [48]. Genes with less than 10 fragments in all the samples were removed before differential expression analysis. The differential expression between conditions was statistically accessed by R/Bioconductor package edgeR (version 3.0.8) [49]. Genes with false discovery rate 0.05 and fold change !2 were called significant. Gene clustering and heatmap were done by Cluster 3.0 and TreeView. GO analysis was performed using Ingenuity Pathway Analysis (IPA) software. The transcriptional profiling plot (Fig 7A) was made using R software.

Reticulocyte micronuclei
Approximately 100 μL of blood obtained from individual mice by cardiac puncture were collected into tubes containing 350 μL of heparin solution and fixed in ultra-cold methanol according to the protocol in the Mouse MicroFlowBasic Kit (Litron Laboratories). The fixed samples were stored at −80˚C until the flow cytometry analysis was performed. Methanolfixed blood samples were washed and labeled with anti-CD71-FITC, anti-CD61-PE and PI for high speed flow cytometry using CellQuest software, v5.2 (Becton Dickinson, San Jose, CA). For each sample, 2 × 10 4 CD71-positive reticulocytes were analyzed for the presence of micronucleated reticulocytes. Flow cytometers were calibrated by staining Plasmodium bergheiinfected rodent blood (malaria biostandards) in parallel with test samples on each day of analysis. Statistical analysis was performed using the Student's t-test or one-way ANOVA followed by Tukey's test.
B cell culture and CSR analysis B cells were isolated from mouse spleens, purified by negative selection with anti-CD43 depletion (Miltenyi) and stimulated with IL-4 and Lipopolysaccharide and IL-4 (Sigma) for 72 hr.
Where indicated, cultures were incubated with DNA-PKcs inhibitor 20 μM NU7026 (Tocris) dissolved in DMSO, or mock-treated. Three mice were analyzed in triplicate and cell count numbers and viability were similar for all groups. The culture, flow-cytometric analysis for CSR analysis and junction analysis has been described [32,50]. Sμ-Sγ1 CSR junctions were amplified by PCR using the following conditions for 25 cycles at 95˚C (30 s), 55˚C (30 s), 68˚C (180 s) using the primers (FWD 5 0 -AATGGATACCTCAGTGGTTTTTAATGGTGGGTTT A-3 0 ; REV 5 0 CAATTAGCTCCTGCTCTTCTGTGG-3 0 ) and Pfu Turbo (Stratagene). To the PCR reaction, 5 U of Taq polymerase (Promega) was added and incubated at 72˚C for 10 min. The resulting product was TOPO TA cloned and transformed into Top10 E. coli cells (Life Technologies) and plasmids were purified and sent for sequencing using M13 FWD and REV primers in addition to the amplification primers for sequencing. 100 clones for each group were analyzed for mutations, deletions, insertions, and sequence overlaps at the junction and both 30 nt upstream and downstream of the junction.

Generation of stable cell lines
T-REx 293 cells (Invitrogen) were transfected with the pcDNA5/FRT/TO TOPO TA construct. Full-length pol ν with six His residues at the N-terminus and a FLAG tag at the C-terminus [4,21] was inserted into the plasmid vector. Stable clones were selected with hygromycin and blasticidin. Expression of the Flag-tagged pol ν was induced with 0.1 μg/mL doxycycline (SIGMA) for 24 hr. Basal repression and doxycycline-induced expression of pol ν were confirmed by immunoblotting.

shRNA transfection
The following sequence was cloned into pSIF-H1-copGFP vector to target POLN: 5'-GATCC TCTTTGGCGAGTTAGAGCTGTACTTCCTGTCAGATGCAGCTTTAACTTGCCAAAG AGGATTTTTT-3', underlined sequence indicates the target sequence (System Biosciences). Lentiviral particles were generated by transfection of three plasmids (the expression plasmid, e.g., pSIF-H1-copGFP-hPOLN.1; plus pFIV-34N and pVSV-G) into 293FT cells using FuGene 6. Culture media from transfected cells was collected 48 hr after transfection to isolate the viral particles, passed through 0.45 μm filters, used immediately or stored at −80˚C in single-use aliquots. Lentiviral transduction was completed as follows: Briefly, 6.0 × 10 4 cells were seeded into 6-well plate and incubated for 24-30 hr at 5% CO 2 at 37˚C. Cells were transduced for 18 hr with shRNA-expression lentiviral stocks at 32˚C and cultured for 72 hr at 37˚C as described [51]. Stable cell lines were selected by detecting GFP expression (a co-expressed marker gene).

Survival experiments
For the ATPlite assay (PerkinElmer), 1,250 cells were plated per well into white 96 well plates and incubated twenty-four hr prior to inducing DNA damage. The cells were incubated with DNA damage inducing agents for the indicated time. After incubation, the cells were immediately lysed and assayed for ATPlite luminescence as described in the manufacturer's instructions. For the clonogenic assay, 1.0 x 10 5 cells were plated in 60 mm culture plates and incubated for 24 hr prior to DNA damage induction. Groups of plates were exposed to indicated doses of mitomycin C for 1 hr. After making a dilution series for each group, cells were returned to the incubator until colonies could be detected in the samples (7 to 14 days), and then were fixed, stained, and scored for survival.

Chromosome analysis
Cells were treated with 40 ng/mL of mitomycin C for 48 hr. At forty-four hr after mitomycin C treatment, cells were treated with 0.03μg/mL colcemid solution (Sigma) for 4 hr. The cells were then trypsinized and exposed to 0.075M KCl for 15 min at 37˚C, and were fixed in 3:1 methanol:glacial acetic acid. The cells were spread on glass slides, Giemsa stained and metaphases were analyzed using a BX41 Olympus microscope, with 60X or 100X oil objectives. Photographs were taken with the 60X oil objective on a Spot Idea 5 color digital camera. 100 metaphases per sample were analyzed to identify cell population with radial chromosome.   The binding of a meiosis-specific zinc finger protein, PRDM9, dictates the location of most meiotic DSBs in the mouse. There is a PRDM9 binding motif at the center of the A3 locus [26]. The consensus PRDM9 binding motif derived by Brick et al. [52] is also indicated. The binding motif sequence differs between strain backgrounds, which affects the affinity of PRDM9 binding [26] and likely the frequency of meiotic DSBs. The yellow bar and yellow shading represent the predicted PRDM9 binding site. Red arrowheads indicate polymorphisms implicated in differential PRDM9 binding. The PRDM9 binding motif in FVB/ NCrl is predicted to have lower affinity than the motif in the A/J background. It was previously shown that in A/J x DBA/2J F1 hybrids, the A3 hotspot has a strong reciprocal crossover asymmetry [24], similar to that observed in F x D. Taken together, the reduced recombination frequency in F x D as compared to B x D mice can be explained by reduced DSBs on the FVB/ NCrl A3 allele. The TaqMan primers spanned across adjacent exons of the human gene as described [21]. siC served as a negative control. HAUS3 was analyzed simultaneously. Note that full-length POLN is not appreciably expressed in 293T cells, although partial transcripts representing portions of the mRNA can be detected [21]. To evaluate the extent of the reduction, a paired t-test was performed.