The Polymerase Activity of Mammalian DNA Pol ζ Is Specifically Required for Cell and Embryonic Viability

Abstract

DNA polymerase ζ (pol ζ) is exceptionally important for maintaining genome stability. Inactivation of the Rev3l gene encoding the polymerase catalytic subunit causes a high frequency of chromosomal breaks, followed by lethality in mouse embryos and in primary cells. Yet it is not known whether the DNA polymerase activity of pol ζ is specifically essential, as the large REV3L protein also serves as a multiprotein scaffold for translesion DNA synthesis via multiple conserved structural domains. We report that Rev3l cDNA rescues the genomic instability and DNA damage sensitivity of Rev3l-null immortalized mouse fibroblast cell lines. A cDNA harboring mutations of conserved catalytic aspartate residues in the polymerase domain of REV3L could not rescue these phenotypes. To investigate the role of REV3L DNA polymerase activity in vivo, a Rev3l knock-in mouse was constructed with this polymerase-inactivating alteration. No homozygous mutant mice were produced, with lethality occurring during embryogenesis. Primary fibroblasts from mutant embryos showed growth defects, elevated DNA double-strand breaks and cisplatin sensitivity similar to Rev3l-null fibroblasts. We tested whether the severe Rev3l-/- phenotypes could be rescued by deletion of DNA polymerase η, as has been reported with chicken DT40 cells. However, Rev3l-/- Polh-/- mice were inviable, and derived primary fibroblasts were as sensitive to DNA damage as Rev3l-/- Polh+/+ fibroblasts. Therefore, the functions of REV3L in maintaining cell viability, embryonic viability and genomic stability are directly dependent on its polymerase activity, and cannot be ameliorated by an additional deletion of pol η. These results validate and encourage the approach of targeting the DNA polymerase activity of pol ζ to sensitize tumors to DNA damaging agents.

Fig 1. Expression of human REV3L complements Rev3l-deficient MEFs.

(A) Top, the human REV3L gene was cloned into a pOZ vector for expression in mammalian cells with an N-terminal FLAG-HA epitope tag. The vector also expresses the interleukin 2 receptor (IL2R) gene via an internal ribosomal entry site (IRES). Below, domains in mammalian REV3L protein. Indicated here are the N-terminal domain, positively-charged domain, two REV7-binding domains, the KIAA2022 homology domain (see S1 Fig), and the C-terminal Fe-S cluster for interaction with other subunits. Vertical bars in the polymerase domain represent highly conserved motifs. The location of the D2781A/D2783A active site mutations (ASM) is shown. (B) Expression of REV3L in MEF cell lines. A set of primers and a Taqman probe were used that recognizes both human and mouse Rev3l, but does not amplify knockout transcript. Functional mouse Rev3l is expressed in Rev3l^+/+ but not Rev3l^-/Δ cells. Wild-type or ASM recombinant REV3L mRNA was expressed in immortalized Rev3l^-/Δ MEFs at about half of the endogenous level. (C) Doubling time (in hr) of MEFs harboring empty vector (EV) Rev3l^+/Δ, Rev3l^-/Δ, and Rev3l^-/Δ MEFs expressing wild-type or ASM recombinant REV3L. (D) Survival of MEFs harboring empty vector (EV): Rev3l^+/Δ (maroon), Rev3l^-/Δ (orange); and Rev3l^-/Δ MEFs expressing wild-type (green) or ASM (blue) recombinant REV3L. ATP content was measured 48 hr after addition of cisplatin. (E) Survival of these cell lines 48 hr after ultraviolet C (UVC) radiation as measured by ATP content. Data represent mean ± SEM.

Fig 2. Spontaneous DNA double-strand break formation is reduced in Rev3l complemented MEFs.

(A) DAPI staining of empty vector (EV)-expressing Rev3l^+/Δ, Rev3l^-/Δ, as well as Rev3l^-/Δ MEFs expressing wild-type or ASM recombinant REV3L; arrows indicate micronuclei, with an enlarged example in the inset. (B) Merged immunofluorescence staining of the same MEFs as in (A) with DAPI (blue), 53BP1 (red) and γ-H2AX (green); foci indicate areas of DNA double-strand breaks. (C) Quantification of percent of nuclei that have associated micronuclei. (D) Quantification of cells with fewer than 3, 3 to 5, 6 to 10 or greater than 10 53BP1 foci (as measured using CellProfiler). The bars are color-coded exactly as in Part C to indicate the genotype of the MEFs (*) p < 0.01. Data represent mean ± SEM.

Fig 3. Knock-in mice and MEFs expressing active site mutant Rev3l have knockout phenotypes.

(A) Diagram of the mouse Rev3l ASM knock-in allele, with the wild-type (WT) locus shown at the top. Green rectangles indicate Rev3l coding sequences and the gray line represents chromosomal sequence. In the middle diagram, FRT sites are represented by double red triangles, loxP sites by blue triangles and lox511 sites by yellow triangles. The targeted exon 27 (starred) carries D2773A and D2775A point mutations and is inserted in an inverted orientation between wild-type exons 26 and 27. Splicing donor and acceptor sites flanking wild-type and ASM exon 27 are kept intact. The knock-in was produced by a Cre-dependent genetic switch. First, the neomycin positive selection cassette (neo) was excised by breeding with C57BL/6 Flp deleter mice. A subsequent cross with Cre-expressing mice led to excision of the wild-type exon 27 and inversion of ASM mutant exon 27 into the functional orientation. In the constitutive ASM knock-in locus shown in the lower diagram, the D2773A/D2775A Rev3l gene is expressed under the control of the endogenous Rev3l promoter and wild-type Rev3l exon 27 is absent from the locus. Heterozygous ASM knock-in mice (Rev3L^+/M) were then used for breeding. (B) Example of Southern blot analysis of (left) the inducible knock-in locus (neo⁺) and (right) the constitutive ASM knock-in locus. Genomic DNA of the tested animals was compared with C57BL/6 wild-type genomic DNA (WT). EcoRV digested DNA was blotted on a nylon membrane and hybridized with the external 3’ probe with the position shown at the top of part A. Restriction fragments of 15 kb, 11.5 kb and 9.5 kb were observed for the wild-type, inducible knock-in locus (neo⁺) and constitutive ASM knock-in locus, respectively. Genomic DNA was further analyzed extensively and confirmed by specific PCR assays and complete DNA sequencing as described in the Materials and Methods. (C) Genotypes of mouse pups produced by breeding parental Rev3l^+/M mice. (D) Growth of Rev3l^+/Δ and Rev3l^M/Δ cells. These cells were produced by addition of AdCre to Rev3l^M/lox or Rev3l^+/lox MEFs, deleting the floxed allele of Rev3l. (E) Survival of Rev3l^+/lox, Rev3l^M/lox, Rev3l^+/Δ and Rev3l^M/Δ primary MEFs 120 hr after addition of cisplatin, as measured by ATP content. (F) The MEFs as in (E) were stained with DAPI, and for 53BP1 and γ-H2AX by immunofluorescence as in Fig 2 to detect foci of DNA double-strand breaks. The quantification shows the percentage of cells with >2 53BP1 and γ-H2AX foci in Rev3l^+/lox, Rev3l^M/lox, Rev3l^+/Δ and Rev3l^M/Δ primary MEFs 9 days after AdCre treatment. (*) p < 0.01. Data represent mean ± SEM.

Fig 4. Deletion of Polh does not ameliorate phenotypes caused by knockout of Rev3l.

(A) Genotypes of mouse pups produced by breeding parental Rev3l^-/lox Polh^-/—mice. (B) MEFs with the indicated genotypes were stained with DAPI, and for 53BP1 and γ-H2AX by immunofluorescence as in Fig 2 to detect foci of DNA double-strand breaks. The quantification shows the percentage of cells with >2 53BP1 and γ-H2AX foci in Rev3l^-/lox Polh^+/+, Rev3l^-/lox Polh^-/-, Rev3l^-/Δ Polh^+/+ and Rev3l^-/Δ Polh^-/- MEFs 9 days after AdCre treatment. (C) Survival of primary MEFs as in part B, 120 hr after addition of cisplatin, as measured by ATP content. For panels B& C, Data represent mean ± SEM.