Repair of a Specific Double-Strand Break Generated Within a Mammalian Chromosome by Yeast Endonuclease L-Scel
Total Page:16
File Type:pdf, Size:1020Kb
.,. 1994 Oxford University Press Nucleic Acids Research, 1994, Vol. 22, No. 25 5649-5657 Repair of a specific double-strand break generated within a mammalian chromosome by yeast endonuclease l-Scel Tamas Lukacsovich, Di Yang and Alan S.Waldman* Department of Biological Sciences, University of South Carolina, Columbia, SC 29208, USA Received August 29, 1994; Revised and Accepted November 4, 1994 ABSTRACT We established a mouse Ltk- cell line that contains yeast (8), underscoring the importance of mechanisms for efficient within its genome a herpes simplex virus thymidine repair of DSBs. kinase gene (tk) that had been disrupted by the Eukaryotic cells potentially have several mechanisms for insertion of the recognition sequence for yeast repairing a DSB. One possible repair pathway is through endonuclease l-Scel. The artificially introduced 18 bp recombination between the broken sequence and an intact l-Scel recognition sequence was likely a unique homologous sequence. It is well documented that DSBs can sequence in the genome of the mouse cell line. To stimulate extrachromosomal homologous recombination in yeast assess whether an induced double-strand break (DSB) (reviewed in 9) and mammalian cells (reviewed in 10). Several in the genomic tk gene would be repaired preferentially reports on recombination in manunalian cells suggest that a DSB by gene targeting or non-homologous recombination, repair mechanism similar to that originally described for we electroporated the mouse cell line with recombination in yeast (11) also operates in extrachromosomal endonuclease l-Scel alone, one of two different gene recombination of DNA sequences transfected into mammalian targeting constructs alone, or with l-Scel in conjunction cells (10), as well as in gene targeting (12-14). DSBs and single- with each of the two targeting constructs. Each strand breaks have been shown to stimulate homologous targeting construct was, in principle, capable of recombination within the yeast genome (9), but there has been correcting the defective genomic tk sequence via no report directly assessing what effect(s) a break introduced into homologous recombination. tk+ colonies were a mammalian genome may have on homologous recombination. recovered following electroporation of cells with l-Scel Some mammalian cell mutants that are defective in the repair in the presence or absence of a targeting construct. of DSBs have been shown to display elevated levels of sister Through the detection of small deletions at the l-Scel chromatid exchanges (SCE) when treated with DNA damaging recognition sequence in the mouse genome, we agents (reviewed in 15), although the precise relationship between present evidence that a specific DSB can be introduced SCE and homologous recombination remains unclear. Treatment into the genome of a living mammalian cell by yeast of mouse cells with a chemical inhibitor of DNA break repair endonuclease l-Scel. We further report that a DSB in has been shown to induce a 4-fold stimulation of the genome of a mouse Ltk- cell is repaired intrachromosomal homologous recombination (16). However, in preferentially by non-homologous end-joining rather these latter studies it was not clear if chromosomal breaks served than by targeted homologous recombination with an as recombinogenic lesions, or whether recombination was exogenous donor sequence. The potential utility of this stimulated by an indirect mechanism. Understanding how DSBs system is discussed. may influence homologous recombination in mammalian chromosomes can contribute to our understanding of naturally occurring genetic rearrangements, as well as have potential INTRODUCTION application in designing strategies for improving the efficiency The genome of a mammalian cell is likely to experience a of gene targeting. spectrum of DNA damage every day. One type of DNA lesion Another potential pathway for the repair of a DSB in a that virtually all organisms must contend with is the formation mammalian cell is through DNA end-joining, or non-homologous of double-strand breaks (DSBs). It has been estimated that a recombination. It has been shown that mammalian cells can mammalian genome suffers about eight DSBs per day (1). efficiently ligate together virtually any two non-cognate ends of Unrepaired DSBs can be lethal or contribute to the induction of DNA (reviewed in 17). The molecular mechanisms for DNA abnormal chromosomal rearrangements in eukaryotes (2-7). It end-joining in mammalian cells have not yet been elucidated; has recently been demonstrated that even a single unrepaired DSB several investigators have recently initiated biochemical in a dispensible plasmid can serve as a signal for cell death in characterizations of end-joining activities isolated from *To whom correspondence should be addressed 5650 Nucleic Acids Research, 1994, Vol. 22, No. 25 mammalian cells (18-22). Efficient end-joining of DNA might gentamicin sulfate. To select for tk+ cells, medium be responsible for the efficiency with which cultured mammalian supplemented with hypoxanthine/aminopterin/thymidine (HAT) cells randomly integrate transfected DNA molecules into their was prepared using 100 x HAT supplement (GIBCO). Cells were genomes (so-called 'illegitimate recombination') (17). maintained at 37°C in a humidified atmosphere of 5% Co2. To contribute to our understanding of how mammalian cells process DSBs, we wanted to develop an experimental system that Plasmid constructions would allow us to introduce a specific DSB within the genome All plasmids used are based on vector pJS-l, which is a derivative of a living cell. We were particularly interested in determining of pSV2neo (25) with restriction site modifications as previously if a genomic DSB might be repaired efficiently by targeted described (26). Plasmid pTKl (8.3 kb) contains the wild-type homologous recombination with a transfected DNA sequence (and HSV-l tk gene on a 2.5 kb BamHI fragment inserted into the therefore be ofpotential use in the development of a methodology unique BamHI site of the vector pJS-l. Into the Sacd site at for enhancing gene targeting efficiency) or if, instead, a genomic position 960 of the tk coding region (numbering according to 27) DSB would be repaired preferentially by non-homologous joining was inserted the following double-stranded oligonucleotide: of the broken chromosomal ends. I In our experimental system, we wanted to be able to introduce 5'-T AGGGAT AACAGGGT AATAGCT-3' a single break at a defined genetic locus. Wholesale genomic 3'-TCGAA T C CCT AT T GT C CC AT T A -5' breakage, as occurs in methods involving the introduction of t bacterial restriction enzymes into mammalian cells (see 23 for This oligonucleotide contains the recognition sequence for example), may induce cellular responses to global damage and endonuclease I-SceI (24), shown in bold. This oligonucleotide this may obscure the effect of any single specific DSB. (Massive also terminates in sticky ends compatible with ends produced by genomic breakage is also clearly undesirable in gene targeting SacI cleavage. Endonuclease I-SceI cleaves to the 3' side of the strategies.) To accomplish our goals, we made use of the nucleotides indicated by arrows, leaving sticky ends with 3' endonuclease I-SceI which is encoded by a mobile type I intron single-strand extensions. Insertion of the oligonucleotide at the from the mitochondrial genome of Saccharomyces cerevisiae SacI site of the tk gene inactivated the tk gene by introducing (24). The recognition sequence for I-SceI cleavage is 18 bp in a frame-shift (a net gain of 22 nucleotides was caused by the length, with some tolerance for single bp degeneracies (24). insertion, when the duplicated 4 bp terminal sticky ends are taken Arithmetical calculations reveal that the I-SceI recognition site into account). The plasmid containing the tk gene disrupted by is not likely to occur randomly in a mammalian genome. the I-SceI recognition site was named pTK1-Sce (Fig. 1). Plasmid We designed a DNA construct containing a herpes simplex pTARG is identical to pTKl, except that the tk gene is rendered virus type one (HSV-1) thymidine kinase gene (tk) disrupted by non-functional due to a 104 bp inversion ofthe sequences between the insertion of an 18 bp oligonucleotide containing the I-SceI the two EcoRV sites at positions 744 and 848 of the tk coding recognition site. This disrupted tk gene was introduced into the region (Fig. 1). Plasmid pTARG has been described previously genome of mouse Ltk- (thymidine kinase deficient) fibroblasts, (28). A similar plasmid, pAL2 (Fig. 1), contains a defective establishing a unique, or nearly unique, I-Scel target sequence internal fragment of the tk gene inserted at the HindmH site of within the mouse genome. We reasoned that subsequent the vector. This tk fragment maps between the Hincd and SmaI introduction of endonuclease I-SceI into the cell could, in sites at positions 442 and 1621 of the tk gene. Plasmid pAL2 principle, lead to the reconstruction of a functional tk gene via (7.0 kb) has been described previously (29). the induction of a DSB at the I-Scel recognition site followed by appropriate resection and joining of DNA ends to restore a Assay for transient tk expression coding sequence. Alternatively, if an appropriate targeting To establish conditions for efficient electroporation of mouse construct were introduced in conjunction with I-SceI, a DSB Ltk- cells, we introduced pTKl (described above) into mouse induced by I-SceI might be repaired via homologous Ltk- cells under a variety of conditions and assayed the recombination with the targeting construct. electroporated cells for transient tk gene expression. The assay In this work, we show that an I-SceI recognition site embedded for a functional tk gene provided a minimal estimate for the within the genome of mouse fibroblasts can be cleaved by percentage of cells that had taken up DNA. Following electroporating endonuclease I-Scel into the cells. This is the first electroporation, cells were plated into 35 mm dishes at a density reported use of an endonuclease to induce a specific DSB at a of 1.5 x l10 cells per dish.