Mutation at the polymerase active site of mouse DNA polymerase delta increases genomic instability and accelerates tumorigenesis

Abstract

Mammalian DNA polymerase delta (Pol delta) is believed to replicate a large portion of the genome and to synthesize DNA in DNA repair and genetic recombination pathways. The effects of mutation in the polymerase domain of this essential enzyme are unknown. Here, we generated mice harboring an L604G or L604K substitution in highly conserved motif A in the polymerase active site of Pol delta. Homozygous Pold1(L604G/L604G) and Pold1(L604K/L604K) mice died in utero. However, heterozygous animals were viable and displayed no overall increase in disease incidence, indicative of efficient compensation for the defective mutant polymerase. The life spans of wild-type and heterozygous Pold1(+/L604G) mice did not differ, while that of Pold1(+/L604K) mice was reduced by 18%. Cultured embryonic fibroblasts from the heterozygous strains exhibited comparable increases in both spontaneous mutation rate and chromosome aberrations. We observed no significant increase in cancer incidence; however, Pold1(+/L604K) mice bearing histologically diagnosed tumors died at a younger median age than wild-type mice. Our results indicate that heterozygous mutation at L604 in the polymerase active site of DNA polymerase delta reduces life span, increases genomic instability, and accelerates tumorigenesis in an allele-specific manner, novel findings that have implications for human cancer.

FIG. 1.

Primary sequence and structural conservation of the Pol δ active site. (A) Domain structure of murine Pol δ. The exonuclease proofreading domain is located in the amino-terminal region, the polymerase domain occupies the central region, and a putative zinc coordinating domain is located near the carboxy terminus. (B) Sequence alignment of motif A. Primary sequences were aligned by CLUSTALW (http://align.genome.jp/). (C, part I) Conservation of the three-dimensional structure of motif A among DNA polymerases. Coordinates of crystal structures were downloaded from the Protein Data Bank (www.rcsb.org) and manipulated in a suite of programs available in the Molecular Operating Environment package at http://www.chemcomp.com/. The polymerase domains shown are from Thermus aquaticus (KlenTaq, 3KTQ) Bacillus stearothermophilus (Bst, 1XWL), E. coli (Klenow, 1KFD), bacteriophage T7 (T7, 1T7P), Thermococcus gorgonarius (Tgo, 1TGO), “Desulfurococcus tok” (D.Tok, 1QQC), “Pyrococcus kodakaraensis” (Kod, 1GCX), and bacteriophage RB69 (RB69, 1IG9). Motifs A, B, and C were aligned with the ternary structure of the RB69 DNA polymerase. The essential catalytic aspartic acids in motifs A and C that coordinate metal ions are shown. The orthologous motif A residues corresponding to mouse Leu604 (see panel B) are shown making contact with the incoming dNTP. (C, part II) Molecular modeling of motif A mutations L415G and L415K on the ternary structure of the RB69 DNA polymerase. L415G and L415K are orthologous to murine Pol δ L604G and L604K, respectively. Molecular modeling was performed by using Engh and Huber structural parameters and the software described above (9).

FIG. 2.

Generation and verification of PoldI^+/L604G and PoldI^+/L604K mice. (A) The structure of the murine Pold1 locus is shown above the targeting vector. Gray numbered boxes represent exons. A neomycin selection cassette was inserted into intron 11. TK1 and TK2 represent thymidine kinase genes. Targeting vectors carrying the motif A mutation L604G or L604K were generated by site-specific mutagenesis of the corresponding codon in exon 15. (B) Southern blotting of PCR-screened Pold1-targeted ES clones. Digestion with HindIII yields 9.3-kb and 8-kb bands, while digestion with HindIII and SphI yields 8-kb and 6.2-kb bands. The numbers at the top of refer to different ES clones. (C) DNA sequence verification of Pold1-targeted heterozygous clones. (D) PCR confirmation of heterozygous clones. The top ∼3-kb band is the targeted allele, and the bottom 1.5-kb band is the wild type (WT). (E) Expression of Pol δ in whole-cell extracts of liver. Aliquots of extracts containing 5 μg to 60 μg of total protein were applied to gels, and Western blots were probed with anti-Pol δ and antitubulin antibodies.

FIG. 3.

Chromosome instability in cultured primary embryonic fibroblasts. (A) Total numbers of spontaneous chromosome aberrations observed in wild-type fibroblasts and in Pold1^+/L604G and Pold1^+/L604K heterozygous fibroblasts. C-banding metaphase spreads of three different Pold1^+/+ cell lines, four different Pold1^+/L604G lines, and three different Pold1^+/L604K lines (passage 2) were used for analysis. A minimum of 1,000 chromosomes was analyzed for each line. P values were calculated by unpaired Student t test. (B) Commonly observed aberrations in heterozygous lines. Arrows indicate chromatid breaks (parts 1 to 3) and an acentric fragment (part 4).

FIG. 4.

DNA damage-induced checkpoint response is proficient in Pold1^+/L604K and Pold1^+/L604G MEFs. Ser139 phosphorylation of histone variant H2AX was analyzed as a marker for checkpoint response in MEFs derived from wild-type, Pold1^+/L604G, and Pold1^+/L604K mice. (A) Cell cycle distribution of γ-H2AX foci. The distribution of singlet gated Pold1^+/L604K cells is shown, with and without treatment with replication inhibitors. Negative control, untreated cells stained with an unrelated nuclear antibody; spontaneous, untreated cells stained for γ-H2AX; 0.4 μg/ml aphidicolin (Aph), cells treated for 8 h with drug and stained for γ-H2AX; 0.4 mM hydroxyurea (HU), cells treated for 8 h with drug and stained for γ-H2AX. Cells above the light green line were used for the analysis in panel B. (B) Fraction of cells positive for γ-H2AX in untreated cells and cells treated with replication inhibitors. For each sample, ∼32,000 cells were sorted and ∼20,000 cells were used for analysis. (C) Indirect immunofluorescence of γ-H2AX foci in nuclei of Pold1^+/L604G and Pold1^+/L604K MEFs. CT, control showing staining with irrelevant antibody; UT, cells not treated with replication inhibitor showing spontaneous focus formation; T, cells treated with 0.4 μg/ml aphidicolin for 8 h. Images were collected at ×63 magnification.

FIG. 5.

Survival of heterozygous Pol δ L604K and L604G mice. Kaplan-Meier survival curves for combined F₁ and F₂ generation PoldI^+/L604G mice and PoldI^+/L604K mice are shown. Sixty-one PoldI^+/+ (24 male, 37 female), 64 PoldI^+/L604G (30 male, 34 female), and 87 PoldI^+/L604K (44 male, 43 female) animals were used. The life span of PoldI^+/L604K mice was significantly shorter than that of the combined PoldI^+/+ and PoldI^+/L604G mice (P < 0.0001, log rank and Wilcoxon tests). The median survival time of the +/L604K heterozygotes was 18 months, relative to 22 months for the wild-type and +/L604G animals.

FIG. 6.

Time to death with histologically diagnosed tumors. The risk of dying with a tumor as a function of time was analyzed by Cox regression analysis. In the Pold1^+/L604K mice, the HRs were elevated for all tumors combined, for lymphoreticular tumors alone, and for all tumors excluding lymphoreticular and hepatocellular tumors (other tumors) but not for hepatocellular tumors. The latter tumors may be associated with Helicobacter infection. HRs for the four groups of tumors were not statistically significantly elevated in Pold1^+/L604G mice, and probability curves are not shown. WT, wild type.

FIG. 7.

No loss of heterozygosity at the Pold1 locus in tumors from Pold1^+/L604G and Pold1^+/L604K mice. Mouse strain 129X1/SvJ- and C57BL/6-specific SNPs were genotyped at the Pold1 locus in 12 histologically graded tumors from Pold1^+/L604K and Pold1^+/L604G mice; representative results for six tumors are shown. Tumors contained ∼90% to 95% malignant cells. Genomic DNA was extracted and amplified with primers flanking exon 13 and intron 16 and sequenced. Sequence chromatograms are shown, and coding is by the complementary strand. Genotyped SNPs were as follows: 1, rs3720949 encoding Asn528, G in strain 129 and A in strain Bl6; 2, rs3720481 encoding Thr536, G in strain 129 and C in strain Bl6; 3, rs3719256 encoding position Gly/Ser575, C in strain 129 and T in strain Bl6. The 12 tumors analyzed included a follicular lymphoma, histiocytic sarcoma, hepatocellular adenoma, and hepatocellular carcinoma from Pold1^+/L604G mice and a follicular lymphoma, mesenteric lymphoma, fibrosarcoma, uterine sarcoma, and hepatocellular carcinoma from Pold1^+/L604K mice; these were also used for analysis of RT-PCR products (see Results). A histiocytic sarcoma and lymphocytic sarcoma from Pold1^+/L604G mice and a lymphocytic sarcoma from a Pold1^+/L604K mouse were analyzed in addition.