DNA polymerase beta null mouse embryonic fibroblasts harbor a homozygous null mutation in DNA polymerase iota

Dear Editor

Prior to the explosion of gene targeting technology, considerable biochemical and cell-based experiments suggested a role for DNA polymerase β (pol β) in DNA repair. Since its initial discovery in 1971 [1], the enzyme that was to eventually be designated pol β [2, 3] was unique in its enzymatic properties [4–6] as compared to the other newly characterized mammalian DNA polymerases alpha (pol α), gamma (pol γ) and delta (polδ) [1, 7]. Of the four eukaryotic DNA polymerases identified by 1977, pol β soon earned the moniker of ‘the’ DNA repair polymerase [8–11]. These early studies defined a role for pol β in repair using isolated nuclei or nuclear extracts, monitoring the incorporation of radioactive nucleosides following ultraviolet (UV) light-induced DNA damage [8–11]. A more definitive role for pol β in excision repair in vitro was demonstrated in 1983, in which complete removal of pyrimidine dimers in DNA was conducted by purified enzyme preparations of pol β, T4 denV (UV endonuclease) and HeLa AP endonuclease II [12]. Although it was subsequently shown that pol α can also carry-out gap-filling DNA synthesis in a base excision repair (BER) reaction [13], the evidence continued to mount in support of pol β acting as ‘the’ DNA repair polymerase in the nucleus, while pol γ participated in replication and repair in mitochondria [1]. Studies continued to identify a role for pol β in the repair of damage induced by many different DNA damaging agents, including bleomycin [14, 15], UV-radiation [16], benzo[a]pyrene [17], methylmethane sulfonate [18], ionizing radiation [19], G-T mis-pairs [20] and uracil [21–24]. Several groups reported complete BER in vitro with pol β and additional purified proteins [22, 23, 25]. Although it was demonstrated in heterologous systems (E. coli and Saccharomyces cerevisiae) that pol β can conduct DNA replication and repair in vivo [26, 27], it was not until a mouse gene knockout was made that the specificity of the repair conducted by pol β was defined [28].

Characterization of the pol β knockout mouse [29, 30] and mouse embryonic fibroblasts (MEFs) deficient in pol β [28] clearly demonstrated a requirement for pol β in repair of alkylation and oxidative DNA damage [28, 31] and provided a valuable resource to explore additional functions of pol β [32, 33], to evaluate the impact of pol β on mutagenesis [34–38] and mechanisms of genotoxin-induced cell death [39–47], to investigate alternate or compensatory repair pathways in the absence of pol β [48–53] and to address structure-function relationships or protein partners of pol β in vivo [34, 54, 55]; among other studies too numerous to mention herein. The most definitive and reproducible endpoint that has been used to evaluate pol β participation in repair in vivo is survival following DNA damage such as exposure to alkylating agents [28, 55]. Surprisingly, it is the 5′dRP lyase function of pol β [56] that appears to be essential and sufficient for alkylating-agent resistance [55].

In the absence of pol β (in MEFs), cells are unable to efficiently repair the highly toxic 5′dRP moiety and therefore are hypersensitive to different types of alkylating agents such as methylmethane sulfonate, N-methyl-N-nitrosourea and N-methyl-N′-nitro-N-nitrosoguanidine [28, 36, 42, 44, 55], the thymidine analog 5-hydroxymethyl-2′ -deoxyuridine [40] as well as the therapeutic agent temozolomide [40, 44]. Recently, it was also demonstrated that in addition to pol β, DNA polymerase λ (pol λ) and DNA polymerase ι (pol ι), as well as the mitochondrial polymerase pol γ have 5′dRP lyase activity and can function (in vitro) in the removal of the 5′ dRP group that arises during BER [53, 57–61]. This would suggest that pol ι and/or pol λ may participate in BER, possibly as a back-up to pol β or may function with unique lesion specificity, in-line with the proposition that the initiating lesion directs the repair sub-pathway [51, 62]. Although MEFs deficient in either pol ι or pol λ are not hypersensitive to alkylating agents [44], the ability of either pol ι and/or pol λ to complement pol β in vivo has not been evaluated. Using MEFs isolated from pol λ knockout mice [63], a back-up role for pol λ was observed following oxidative damage, although the requirement for the 5′dRP activity of pol λ has not been determined [53, 59]. To date, a targeted deletion or inactivation of the mouse POLI gene has not been reported. However, it was observed that mice derived from the 129 strain harbor a nonsense mutation in exon 2, rendering these mice deficient in pol ι expression [64].

The pol β knockout mouse [29] and hence the derived MEFs [28] were developed by gene-targeting using 129-derived embryonic stem cells. As such, there is the possibility that the pol β deficient cells may harbor a null mutation in pol ι in addition to the targeted null mutation in pol β. We have therefore evaluated all of the pol β deficient MEFs we have available for the presence of the nonsense mutation in the POLI gene. The PCR-based genotyping screen of codon 27 differentiates a wild-type allele (codon = TCG) from a null allele (codon = TAG) as this codon is part of a TaqI restriction enzyme site in the wild-type allele (TaqI recognition sequence = TCGA) but the TaqI site is destroyed in the null allele (TAGA) [64]. As shown in Figure 1 and described in Table I, one of the pol β null MEF cell lines (19tsA, clone 2B2), initially described in 1996 [28] and available from ATCC, also harbors a null mutation in pol ι. We have notified ATCC to change the description for this cell line and will point out, as indicated in Table I that a second matched pair of MEFs are available from ATCC each of which are WT for pol ι. The alkylation sensitivity of pol β deficient cells is not in question - several additional reports have confirmed this observation using the matched pairs listed in Table I that are WT for pol ι, including the 36.3 and 38 4 matched pair [36, 55] and the 92TAg and 88TAg matched pair [42, 44] as well as pol β deficiency mediated by RNA interference [44, 65] and many other studies from other laboratories. It remains to be determined if these double null cells have an increased sensitivity to alkylating agents and most importantly, if complementation with mouse pol ι effects a change as well.

FIGURE 1.

PCR genotyping of mouse POLI codon 27 in pol β null and matched wild-type cell pairs. Genomic DNA isolated from each cell line was purified and characterized for the pol ι codon 27 Wt and null allele as described [64]. Briefly, amplification was performed using flanking PCR primers that amplify an 88 bp fragment surrounding codon 27 in a volume of 25 μl, with 250 μM of each dNTP, 10 pmoles of each primer, and 1.25 units of AmpliTaq Gold DNA polymerase. After denaturation at 95°C for 10 min, the reaction mixture is subjected to 35 PCR cycles as follows: denaturation at 95°C for 0.5 min, annealing at 57°C for 0.5 min and primer extension for 1 min at 72°C, followed by a single primer extension step at 72°C for 10 min. The resulting fragment is then purified using a Qiagen PCR spin-column and the DNA is incubated in the presence of TaqI (NEB) at 65°C for 2 hours. The final products were then analyzed by electrophoresis using a 4% NuSieve 3:1 agarose gel in TBE. The amplicon (88 bp) containing a WT allele is digested to two fragments of 39 and 49 bp whereas the null allele (mut) remains as the undigested 88 bp fragment. The cell lines and corresponding genomic DNA used for lanes 1–10 are listed in Table I. M = DNA size markers; WT = DNA derived from C57BL/6 mice harboring a WT POLI codon 27; mut = DNA derived from c129 mice harboring a mutant POLI codon 27.

TABLE I.

Cell lines screened for the POLI codon 27 mutation.

Lane #	Cell Line	ATCC Number	Overall Genotype regarding pol β and pol ι	Reference
1	16tsA (clone 1B5)	CRL-2307	WT	[28]
2	19tsA (clone 2B2)	CRL-2308	pol β null and pol ι mutant (null)	[28]
3	36.3	N/A*	WT	[36, 55]
4	38 Δ 4	N/A	pol β null	[36, 55]
5	50TAg	N/A	pol β null	-
6	53TAg	N/A	WT	-
7	88TAg	CRL-2820	pol β null	[42, 44]
8	92TAg	CRL-2816	WT	[42, 44]
9	369TAg	N/A	pol ι mutant (null)	[44]
10	370TAg	N/A	pol ι heterozygous (+/−)	-

^*N/A = Not available from ATCC.

This same procedure can be utilized to check other MEF cell lines derived from mouse 129 strains. Note that in general, this should not be a significant problem, as most knockout lines are bred for several generations to the C57BL/6 strain before MEFs are developed. As indicated in Table I, only the 19tsA cell line harbors the 129 strain-derived POLI gene mutation. This cell line was developed from crossing an F1 generation and was not found in MEFs that were derived from later generations. In addition, as it is almost standard practice to breed newly developed knockout lines for a few generations to C57BL/6 before initiation of detailed experimental analysis, we do not foresee this as impacting any of the published whole animal studies with regard to the pol β +/− mice.