|Home | About | Journals | Submit | Contact Us | Français|
DNA double-strand breaks (DSBs) are dangerous lesions that can lead to potentially oncogenic genomic rearrangements or cell death. The two major pathways for repair of DSBs are nonhomologous end joining (NHEJ) and homologous recombination (HR). NHEJ is an intrinsically error-prone pathway while HR results in accurate repair. To understand the origin of genomic instability in human cells it is important to know the contribution of each DSB repair pathway. Studies of rodent cells and human cancer cell lines have shown that the choice between NHEJ or HR pathways depends on cell cycle stage. Surprisingly, cell cycle regulation of DSB repair has not been examined in normal human cells with intact cell cycle checkpoints. Here we measured the efficiency of NHEJ and HR at different cell cycle stages in hTERT-immortalized diploid human fibroblasts. We utilized cells with chromosomally-integrated fluorescent reporter cassettes, in which a unique DSB is introduced by a rare-cutting endonuclease. We show that NHEJ is active throughout the cell cycle, and its activity increases as cells progress from G1 to G2/M (G1<S<G2/M). HR is nearly absent in G1, most active in the S phase, and declines in G2/M. Thus, in G2/M NHEJ is elevated, while HR is on decline. This is in contrast to a general belief that NHEJ is most active in G1, while HR is active in S, G2 and M. The overall efficiency of NHEJ was higher than HR at all cell cycle stages. We conclude that human somatic cells utilize error-prone NHEJ as the major DSB repair pathway at all cell cycle stages, while HR is used, primarily, in the S phase.
DNA double-strand breaks (DSBs), in which both strands in the double helix are severed, are the most dangerous type of DNA lesion. If left unrepaired, or repaired incorrectly, DSBs may result in massive loss of genetic information, genomic rearrangements, or cell death. DSBs are repaired by two major mechanisms: non-homologous end joining (NHEJ) and homologous recombination (HR) 1. The two pathways differ in their fidelity and template requirements. NHEJ modifies the broken DNA ends, and ligates them together with little or no homology, generating deletions or insertions 2. In contrast, HR uses an undamaged DNA template on the sister chromatid or homologous chromosome to repair the break, leading to the reconstitution of the original sequence 3. Thus, the choice of DSB repair pathway determines the fidelity of repair, which in turn may influence the rates of aging and tumorigenesis 4–6.
Both DSB repair pathways play important roles in mammalian DSB repair 7, 8. The exact mechanism by which the choice between the two DSB repair pathways is made remains unclear. In part, the choice of repair pathway is determined by cell cycle stage. The dependence of DSB repair on the cell cycle was first shown by analyzing the sensitivity of chicken DT40 cells, deficient in NHEJ or HR factors, to ionizing radiation 9. NHEJ mutants were highly sensitive in G1 and early S, while HR mutants were sensitive only in S/G2. Studies in hamster CHO cell lines containing mutations in DSB repair genes showed that NHEJ-defective cells have reduced repair at all cell cycle stages, while HR-defective cells have a minor defect in G1, and greater impairment in S/G2/M 10, 11. Thus, NHEJ is a major DSB repair pathway in G1 stage, while both NHEJ and HR may be active in S/G2/M. The studies of mutant cell lines do not allow for comparison of the contributions of each pathway, because various mutations inactivate NHEJ and HR to a different extent. Furthermore, the effect of cell cycle stage on DSB repair has not been analyzed in primary human cells, which maintain intact cell cycle checkpoints.
We recently developed sensitive fluorescent reporter assays to examine NHEJ and HR in hTERT-immortalized diploid human cells. These cells retain all characteristics of untransformed primary cells 12, including intact cell cycle checkpoints. We show that NHEJ is active throughout the cell cycle, but is the highest in G2/M. HR is nearly absent in G1, most active in the S phase, and is low in G2/M. Thus, in normal human fibroblasts NHEJ is the major DSB repair pathway, while HR primarily repairs DNA breaks that occur during replication.
The efficiency of NHEJ and HR during cell cycle has not been examined in normal human cells. The use of normal cells is important because normal cells maintain an intact cell cycle control apparatus, which may be involved in regulating DNA repair. To examine the contribution of NHEJ and HR in DSB repair during the cell cycle we used hTERT-immortalized normal human fibroblasts (HCA2-hTERT) containing chromosomally integrated GFP-based NHEJ and HR reporter constructs. The construction of NHEJ and HR reporter cell lines has been described previously 13, 14. Here we used two NHEJ (I9a and S13a) and two HR (H15c and H32c) cell lines for analysis of repair during the cell cycle.
The reporter cassette for detecting NHEJ 15 contains a GFP gene with an artificially engineered 3 kb intron from the Pem1 gene (GFP-Pem1). The Pem1 intron contains an adenoviral exon flanked by 18 bp recognition sequences for I-SceI endonuclease (Figure 1A). I-SceI is used to generate site-specific DSBs in vivo. In the NHEJ-I9a clone (Figure 1A) I-SceI sites are in inverted orientation, which generate incompatible ends (Figure 1C), and in NHEJ-S13a (Figure 1A) I-SceI sites are in a direct orientation, which generate compatible DNA ends (Figure 1D). An un-rearranged NHEJ cassette is GFP negative since the adenoviral exon disrupts the GFP ORF. Upon induction of DSBs by the expression of I-SceI 16, the adenoviral exon is removed and NHEJ restores function of the GFP gene. This reporter can detect a wide spectrum of NHEJ events since the intron can tolerate deletions and insertions. The HR reporter (Figure 1B) is built using the same GFP-Pem1 gene as the NHEJ reporter 14. In the HR reporter, the first exon of GFP-Pem1 contains a 22 bp deletion combined with the insertion of three restriction sites, I-SceI-HindIII-I-SceI, which are used for inducing DSBs. The deletion ensures that GFP cannot be reconstituted by an NHEJ event. The two I-SceI sites are in an inverted orientation, so that I-SceI digestion leaves incompatible ends (Figure 1C). The first copy of GFP-Pem1 is followed by a promoter-less/ATG-less first exon and intron of GFP-Pem1. The intact construct is GFP-negative. Upon induction of a DSB by I-SceI digestion the functional GFP gene will be reconstituted by intramolecular or intermolecular gene conversion between the two mutated copies of the first GFP-Pem1 exon. Since the second copy of the GFP gene lacks a promoter, the first ATG codon, and the second exon, crossing over or single-strand annealing will not restore the GFP activity. This design allows for the exclusive detection of gene conversion, which is the predominant HR pathway in mammalian cells 17. The H15c and H32c cell lines used in this study are two independent integrants of the HR reporter. Reconstitution of the functional GFP gene by both NHEJ or HR has been confirmed by sequencing reporter cassettes from GFP-positive clones 13.
To study DSB repair during cell cycle we first determined the treatments that arrest HCA2-hTERT cells at various cell cycle stages. Normal cells are sensitive to contact inhibition, and arrest in G1 stage at confluence (Figure 2A). Treatment with a DNA polymerse α inhibitor, aphidicolin 18, arrested HCA2-hTERT in S stage (Figure 2A). Colchicine, which prevents microtubule polymerization 19, arrested the cells in G2/M stage (Figure 2A). Cell cycle distribution in the treated cells was verified by propidium iodide staining and flow cytometry every day for 7 days following treatment. Drug-treated cells entered cell cycle arrest 2 days after treatment, and remained arrested for at least 7 days. No cell death was observed in either confluent or drug-treated cells. Confluent cells were in G1 stage after reaching confluence, and could be maintained in G1 indefinitely with regular media changes. Figure 2A shows cell cycle distribution on day 7 after reaching confluence and day 4 after drug treatment. This is the time point when DSB repair was taking place (discussed below).
To analyze NHEJ and HR G1 arrested cells were co-transfected with 5 µg of I-SceI-expressing plasmid 20 to induce DSBs and 0.1 µg DsRed plasmid to normalize for the transfection efficiency. G1-arrested cells were transfected on day 6 after cells reached confluence, and drug-treated cells were transfected on day 3 after drug treatment. In HCA2 fibroblasts I-SceI expression reaches a maximum during the first 24 hours after transfection, and then progressively declines 14. Therefore, the majority of DSBs were induced on day 7 of confluence and day 4 after drug treatment. Repair of I-SceI-induced breaks results in the appearance of GFP+ cells. To quantify NHEJ and HR events cells were analyzed by flow cytometry 4 days post-transfection to allow for maximum GFP expression in drug-treated cells. Our pilot experiments showed no decline in GFP or DsRed signal during the first 4 days after transfection. GFP and DsRed fluorescence was analyzed using the green-versus-red plot (Figure 2B) as described previously 15. To normalize for the efficiency of transfection, the ratio of GFP+ to DsRed+ cells was used as a measure of NHEJ and HR efficiency.
In G1 stage NHEJ was active while HR was almost completely repressed (Figure 2C, D). As the HR reporter contains a duplication of the GFP gene, HR can potentially occur intrachromosomally. The near absence of HR in G1 indicates that HR does not occur when only an intrachromosomal template is available.
In S phase the frequency of NHEJ was increased by approximately 1.5 to 3 fold relative to the G1 stage cells (Figure 2C). The frequency of HR in S increased 20 to 27 fold relative to G1 phase (Figure 2D). Despite the sharp increase in HR frequency in S phase it remained lower than S-phase NHEJ.
Interestingly, in G2/M NHEJ frequency was further elevated 1.5 to 4 fold relative to the S phase (Figure 2C). NHEJ of both compatible and incompatible ends was elevated in G2/M, but the rise in compatible-end NHEJ (line S13a) was more dramatic. The frequency of HR declined in G2/M by 5 fold relative to HR in S phase (Figure 1D). The increase of NHEJ and decline of HR in G2/M is an unexpected finding, since HR is believed to be efficient in G2/M.
In summary, our results indicate that NHEJ is the major DSB repair pathway at all cell cycle stages, and is most active in G2/M phase. HR operates predominantly in S-phase, but even in S phase HR may be less efficient than NHEJ.
Here we report the first direct comparison of NHEJ and HR in normal human cells. We employ sensitive fluorescent reporter assays that allow for a direct comparison of the efficiencies of NHEJ and HR events upon induction of chromosomal DSBs with a rare-cutting endonuclease. Fluorescent assays score DSB repair events in thousands of cells, and are highly quantitative.
Our data shows that NHEJ is active throughout the cell cycle. Our distinct finding is that NHEJ efficiency is G1<S<G2/M. In earlier studies the roles of NHEJ and HR during the cell cycle were analyzed by assaying the sensitivity of synchronized NHEJ or HR mutant cells to ionizing radiation 9–11, 21–23. Most of the NHEJ-deficient cell lines are hypersensitive in G1 phase, leading to conclusion that NHEJ acts primarily in G1. NHEJ has been proposed to act also in S-phase 24, 25, or in all cell cycle stages 10, 11. It was also shown that fidelity of NHEJ is higher in G2 than G1 26. We show that the efficiency of NHEJ in G2/M is 4–6 fold higher than in G1. Thus, NHEJ plays the major role in DSB repair in human cells and its activity is increased in S and G2/M stages despite availability of HR. Our results also suggest that G2/M phase of the cell cycle is characterized by the highest efficiency of DNA break repair.
We show that HR is extremely low in G1, is most efficient in S, and is decreased in G2/M. The cell survival studies uniformly agree that HR is not utilized in G1 stage 9–11, 27. Ectopic and allelic sequences may potentially be used as templates for repair, but in organisms with large repetitive genomes this may generate chromosomal rearrangements and a loss of heterozygosity. Indeed, studies in both plants and animals show that homologous and ectopic sequences are used at very low frequencies 17, 28, 29. Theoretically, HR may occur in S, G2 and M since sister chromatids are present. Our results indicate that HR is used more frequently in S than in G2/M. It was shown that HR is important for repairing lesions resulting from replication block 10, 30, 31. Thus, a possible explanation for decreased HR in G2/M is that a primary function of HR in mammalian cells is to repair DNA damage associated with DNA replication, and HR is active when DNA replication machinery is present.
Since our assay directly measures NHEJ and HR using the same type of reporter we can evaluate the contribution of NHEJ and HR at each cell cycle stage. The comparison between NHEJ and HR is summarized in a model (Figure 3) in which NHEJ contributes to DSB repair at all cell cycle phases, but is most active in G2/M, while HR contributes primarily in S phase, and modestly in G2/M. Importantly, even in the S phase when HR is the most active, NHEJ is more efficient than HR. Our data is consistent with NHEJ being the major DSB repair pathway during all cell cycle stages in human somatic cells. The reason for preferential use of NHEJ by human cells is likely to be the repetitive nature of the human genome 5. If an incorrect template were to be used for repair, an HR event could result in a gross genomic rearrangement. Organisms with highly repetitive genomes may therefore favor NHEJ, as a small deletion associated with an NHEJ event is less deleterious than an aberrant recombination event. This strategy will protect young organisms from genomic rearrangements; however, over time mutations introduced by NHEJ will accumulate in the genome and contribute to a functional decline associated with aging.
HCA2 are human neonatal foreskin fibroblasts isolated by the laboratory of O.M. Pereira-Smith (UTHSCSA, Houston, TX). Construction of NHEJ-I9a, NHEJ-S13a, HR-H15c, and HR-H32c reporter cell lines is described in 14. Cells were cultured at 37° in a 5% CO2, 3% O2 incubator, in EMEM media supplemented with 15% fetal bovine serum, 100 units/ml penicillin and 100 μµ/ml streptomycin.
In order to arrest the cells at various stages of the cell cycle, several growth conditions, and drug treatments were used. For G1 arrest, cells were plated on 100 mm plates at a concentration of 5×105 and allowed to grow to confluence and kept there for 6 days before transfection. Following transfection, the confluent cells were replated on 60 mm plates to maintain a state of contact inhibition and therefore G1 arrest. S-phase arrest was induced by the addition of 1 µg/ml aphidicolin, and G2/M arrest was induced by the addition of 0.1 mg/ml colchicine.
The transfections were performed using an Amaxa Nucleofector; program U-20. In each transfection 2×106 cells were transfected with 0.1 µg of a DsRed expressing plasmid, and 5 µg of an I-SceI expressing plasmid.
For the analysis of NHEJ and HR cells were harvested, resuspended in ~1ml 1×PBS, put on ice, and run on a FACS Calibur instrument. GFP and DsRed fluorescence was analyzed using the red-versus-green plot as described previously 15. Cell cycle distribution was analyzed by PI staining. Data was analyzed using CellQuest software.
We thank Dara Brown, A’Shantee O’Steen, and Anna Sokolov for help with construction of HR reporter cell lines. This work was supported by grants from US National Institute of Health AG027237 (V.G.), American Federation for Aging Research (V.G.), the Komen Foundation (V.G.), and Ellison Medical Foundation (V.G. and A.S.).