|Home | About | Journals | Submit | Contact Us | Français|
The MRE11 complex (MRE11, RAD50 and NBS1) and the ataxia-telangiectasia mutated (ATM) kinase function in the same DNA damage response pathway to effect cell cycle checkpoint activation and apoptosis1–3. The functional interaction between the MRE11 complex and ATM has been proposed to require a conserved C-terminal domain of NBS1 for recruitment of ATM to sites of DNA damage4,5. Human Nijmegen breakage syndrome (NBS) cells and those derived from multiple mouse models of NBS express a hypomorphic NBS1 allele that exhibits impaired ATM activity despite having an intact C-terminal domain3,6–11. This indicates that the NBS1 C terminus is not sufficient for ATM function. We derived Nbs1ΔC/ΔC mice in which the C-terminal ATM interaction domain is deleted. Nbs1ΔC/ΔC cells exhibit intra-S-phase checkpoint defects, but are otherwise indistinguishable from wild-type cells with respect to other checkpoint functions, ionizing radiation sensitivity and chromosome stability. However, multiple tissues of Nbs1ΔC/ΔC mice showed a severe apoptotic defect, comparable to that of ATM- or CHK2-deficient animals. Analysis of p53 transcriptional targets and ATM substrates showed that, in contrast to the phenotype of Chk2−/− mice, NBS1ΔC does not impair the induction of proapoptotic genes. Instead, the defects observed in Nbs1ΔC/ΔC result from impaired phosphorylation of ATM targets including SMC1 and the proapoptotic factor, BID.
To address the role of the conserved C-terminal domain of NBS1, homologous recombination was used to delete exon 15 of the Nbs1 (also known as Nbn) gene (Supplementary Fig. 1). Splicing from exon 14 to 16 in the ensuing allele, hereafter designated Nbs1ΔC, results in a nonsense mutation. The messenger RNA transcribed from the targeted allele encodes an NBS1 protein that terminates after a non-native isoleucine and lacks the C-terminal 24 amino acids, which include the ATM binding domain (Fig. 1a, b). Western blotting confirmed that a smaller NBS1 protein was produced (Fig. 1c) and immunoprecipitation with NBS1 antisera demonstrated that the MRE11 complex was intact and present at normal levels in Nbs1ΔC/ΔC cells (Fig. 1d). In contrast to cells from NBS patients and Nbs1ΔB/ΔB mice6,9, the MRE11 complex exhibited normal nuclear localization and ionizing-radiation-induced foci (IRIF)-formation in Nbs1ΔC/ΔC cells (Supplementary Fig. 2a). Nbs1ΔC/ΔC mice were viable and born in normal mendelian ratios, and they did not exhibit overt developmental defects.
Unlike Atm−/−cells, Nbs1ΔC/ΔC mouse embryo fibroblasts (MEFs) did not senesce prematurely, did not exhibit increased spontaneous chromosome aberrations and were not sensitive to γ-irradiation (Supplementary Fig. 2b, c, and data not shown)12–14. Atm−/− mice uniformly develop thymic lymphoma from 2 to 8 months of age12,14. Whereas 90% of Atm−/−mice in our colony present with lymphoma by 8 months, none has been observed in Nbs1ΔC/ΔC mice of the same age (Supplementary Fig. 3a).
Atm−/− mice and ATM-deficient cells from ataxia telangiectasia patients are defective in the activation of DNA-damage-dependent checkpoints at the G1/S and G2/M transitions, and within S phase3,12–14. Cells established from NBS patients, and from mice that model the Nbs1657Δ5 allele, have normal G1/S checkpoints15,16, but are defective in the imposition of intra-S-phase and G2/M DNA-damage-dependent checkpoints9,11,17. Neither the G1/S nor the G2/M DNA-damage-dependent cell cycle checkpoints were altered in early passage Nbs1ΔC/ΔC MEFs, indicating that these ATM-dependent checkpoints did not require the NBS1 C terminus (Fig. 2a, b). In contrast, Nbs1ΔC/ΔC cells exhibited an intra-S-phase checkpoint defect comparable to Nbs1ΔB/ΔB, suppressing DNA synthesis after 10 Gy of ionizing radiation by 36% compared with 51% in wild type (Fig. 2c).
In response to ionizing radiation, ATM phosphorylates SMC1; this event is required for imposition of the intra-S-phase checkpoint18. Consistent with the defect observed, SMC1 phosphorylation on ionizing radiation exposure was reduced in Nbs1ΔC/ΔC cells (Fig. 2d). This did not seem to reflect impaired ATM activation because ATM autophosphorylation (at Ser 1981), an index of ATM activation19, was unaffected in Nbs1ΔC/ΔC (Fig. 2e). These data suggest that the NBS1 C-terminal domain governs the access of activated ATM to SMC1, and that impairing this event contributes to the checkpoint defect of Nbs1ΔC/ΔC cells.
In contrast to the relatively minor impact on cell cycle checkpoint functions, Nbs1ΔC exerted a profound influence on apoptosis. Rad50S/S mice, which express the hypermorphic Rad50S allele, exhibit ATM-dependent apoptotic attrition of haematopoietic cells, resulting in death from anaemia at 2–3 months of age2,20. Rad50S/S mice thus provide a uniquely sensitive context to assess ATM function. The onset of age-dependent anaemia in Rad50S/S mice was markedly reduced by the presence of even a single Nbs1ΔC allele2. Rad50S/S Nbs1+/ΔC and Rad50S/S Nbs1ΔC/ΔC mice did not exhibit pathology at 8 months, an age at which 97.5% of Rad50S/S mice succumbed to anaemia (Supplementary Fig. 3b)2,20. Flow cytometry confirmed that the attrition of B-, T- and myeloid-cell lineages was mitigated in Rad50S/S Nbs1ΔC/ΔC mice (Supplementary Fig. 4). Rad50S/S mice also exhibit apoptosis in the semeniferous tubules20 and the gut epithelium (Fig. 3a). Apoptosis in Rad50S/S Nbs1ΔC/ΔC testes and gut was substantially mitigated, demonstrating that the effect of Nbs1ΔC on apoptosis was not confined to haematopoietic cells (Fig. 3a, and Supplementary Figs 5 and 6).
Having established that Nbs1ΔC impaired apoptotic cellular attrition induced by the Rad50S allele, we examined the induction of apoptosis by ionizing radiation. Nbs1ΔC/ΔC mice were irradiated and thymi were examined by immunohistochemical staining for cleaved caspase-3. Similar to Atm−/−, thymi from Nbs1ΔC/ΔC mice showed reduced caspase staining after 10 Gy, indicating an attenuated apoptotic response to ionizing radiation in vivo (Fig. 3b).
To obtain a more quantitative view of the apoptotic defect, ionizing-radiation-induced apoptosis was assessed in Nbs1ΔC/ΔC thymocytes ex vivo. Cultured thymocytes were γ-irradiated with 0.5, 1, 2, 4, 6 and 8 Gy, and annexinV-positive cells, indicative of apoptosis, were scored by flow cytometry. At each ionizing radiation dose, apoptosis of Nbs1ΔC/ΔC thymocytes was reduced (Fig. 3c); the reduction was comparable in magnitude to Atm−/− or Chk2−/− (also known as Chek2−/−) thymocytes at 5 Gy (Fig. 3d). This analysis also revealed that the distribution of CD4, CD8 and double-positive thymocytes in Nbs1ΔC/ΔC was indistinguishable from wild type (Supplementary Fig. 4b); hence, Nbs1ΔC/ΔC does not phenocopy Atm−/− in which thymic differentiation is impaired14.
If the apoptotic function of ATM were dependent on the NBS1 C terminus, Nbs1ΔC/ΔC would be epistatic to Atm deficiency with respect to its apoptotic defect4,5. To test this, we interbred Nbs1+/ΔC and Atm+/− mice. Homozygous double mutants were viable and born at the expected mendelian ratios (data not shown). Thymocyte apoptosis in double mutants was comparable to Atm−/− or Nbs1ΔC/ΔC, consistent with the interpretation that the apoptotic functions of ATM are largely dependent on the C-terminal domain of NBS1 (Fig. 3d). To determine if CHK2 functioned in the same signalling pathway as ATM and NBS1 in the thymus, we generated Atm−/− Chk2−/− double-mutant mice. Apoptosis in response to ionizing radiation was as deficient as in cells lacking p53 (Fig. 3d). Hence, p53-dependent apoptosis is regulated in parallel, with CHK2 on one arm and the MRE11 complex and ATM on the other.
p53’s influence on apoptosis in the thymus is mediated in part through transcriptional regulation of proapoptotic genes. This aspect of p53 function is dependent on CHK2, and only partially impaired by ATM deficiency21,22. To address the mechanistic basis of the Nbs1ΔC/ΔC apoptotic defect, changes in the levels of Bax and Puma (also known as Bbc3) mRNA were assessed at 8 h post 5 Gy of ionizing radiation using quantitative PCR. The levels of Bax and Puma mRNA were similar in both Nbs1ΔC/ΔC and wild-type thymocytes (Fig. 4a). In contrast, cells lacking CHK2 or p53 were almost completely deficient in their induction (Fig. 4a). These data support the view that MRE11-complex-dependent apoptotic induction is largely CHK2-independent, consistent with previous data indicating that NBS1 and CHK2 exert parallel influences on the intra-S-phase checkpoint23.
Having established that transcriptional regulation was unaffected in Nbs1ΔC/ΔC mice, we examined ATM substrates. The levels and phosphorylation status of the ATM substrates p53, CHK2 and the apoptotic effector BID were examined after ionizing radiation treatment. The ATM-dependent phosphorylation of BID was markedly reduced in Nbs1ΔC/ΔC cells (Fig. 4b). This finding is particularly compelling in light of the fact that Nbs1ΔC/ΔC phenocopies BID deficiency with respect to apoptotic and intra-S-phase checkpoint defects24,25. In contrast, no defects in the ATM-dependent and MRE11-complex-dependent hyperphosphorylation of CHK2 was observed in Nbs1ΔC/ΔC cells (Fig. 4c). Similarly, the phosphorylation and stabilization of p53 after ionizing radiation, which is defective in Atm−/− and Chk2−/− cells21,22,26, was normal in Nbs1ΔC/ΔC (Fig. 4b). These data support a model wherein the MRE11 complex, through the C terminus of NBS1, facilitates access of ATM to substrates that include effectors of apoptosis, and, in which the MRE11 complex and ATM act in parallel to CHK2. An implicit prediction of this model is met: the apoptotic defects of ATM- and CHK2-deficiency are additive (Fig. 3d).
Loss of the NBS1 C terminus exerts a relatively circumscribed effect: NBS1ΔC does not impair p53 phosphorylation, stabilization or transcriptional responses but reduces the ability of ATM to phosphorylate SMC1 as well as the proapoptotic BID protein. The phenotypic similarities between Nbs1ΔC/ΔC and Bid−/− are consistent with the view that BID is among the major ATM-dependent apoptotic effectors impaired in Nbs1ΔC/ΔC (refs 24, 25). The precise role of BID phosphorylation in apoptosis remains unclear, but our data are consistent with the view that dynamic modification of BID influences apoptosis.
The findings presented support the functional significance of the NBS1 C terminus for ATM activity in vivo; however, the specificity of the Nbs1ΔC/ΔC phenotype clearly demonstrates that ATM recruitment is not mediated solely by the NBS1 C terminus. In vitro analyses using purified proteins27, as well as the phenotypic differences between Nbs1ΔB/ΔB (lacking the N terminus)9 and Nbs1ΔC/ΔC mice, illustrate that ATM makes multiple contacts with members of the MRE11 complex. We propose that ATM signalling (and presumably recruitment) may be mediated by distinct molecular determinants within the MRE11 complex, as well as in other DNA damage sensors and response mediators, and that different outcomes of the ATM–MRE11-complex DNA damage response may be governed by distinct molecular interactions.
Ionizing radiation sensitivity, analysis of chromosomal aberrations, G1/S, G2/M, and intra-S-phase checkpoint assays were performed as described28.
Immunoblotting and immunoprecipitations were carried out as described previously2,29. For analysis of NBS1 localization, MEFs were fixed in 4% formaldehyde and permeabilized with 0.25% TX-100. For IRIF analysis cells were fixed in 1:1 methanol:acetone 8 h post treatment with 10 Gy of ionizing radiation. Images were captured on a Zeiss Axiovert and imaged with a CCD camera using Volocity (Improvision) and cropped in Photoshop (Adobe).
We thank N. Copeland, N. Jenkins, and C. Adelman for assistance with recombineering and ES cell culture, J. Theunissen for assistance with checkpoint and apoptotic analysis, G. Oltz and E. Rhuley for AC1 ES cells, Y. Shiloh for anti-ATM (MAT3) antibodies, and Petrini laboratory members for helpful suggestions. T.H.S. was supported by an NRSA fellowship and this work was supported by NIH grants awarded to J.H.P. and the Joel and Joan Smilow Initiative.
Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature.
Author Contributions T.H.S. and J.H.P. conceived the experiments and wrote the paper. T.H.S., M.M., S.S.C., and H.H. performed the experiments.
Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests.