ATM activation and its recruitment to damaged DNA require binding to the C terminus of Nbs1

Abstract

ATM has a central role in controlling the cellular responses to DNA damage. It and other phosphoinositide 3-kinase-related kinases (PIKKs) have giant helical HEAT repeat domains in their amino-terminal regions. The functions of these domains in PIKKs are not well understood. ATM activation in response to DNA damage appears to be regulated by the Mre11-Rad50-Nbs1 (MRN) complex, although the exact functional relationship between the MRN complex and ATM is uncertain. Here we show that two pairs of HEAT repeats in fission yeast ATM (Tel1) interact with an FXF/Y motif at the C terminus of Nbs1. This interaction resembles nucleoporin FXFG motif binding to HEAT repeats in importin-beta. Budding yeast Nbs1 (Xrs2) appears to have two FXF/Y motifs that interact with Tel1 (ATM). In Xenopus egg extracts, the C terminus of Nbs1 recruits ATM to damaged DNA, where it is subsequently autophosphorylated. This interaction is essential for ATM activation. A C-terminal 147-amino-acid fragment of Nbs1 that has the Mre11- and ATM-binding domains can restore ATM activation in an Nbs1-depleted extract. We conclude that an interaction between specific HEAT repeats in ATM and the C-terminal FXF/Y domain of Nbs1 is essential for ATM activation. We propose that conformational changes in the MRN complex that occur upon binding to damaged DNA are transmitted through the FXF/Y-HEAT interface to activate ATM. This interaction also retains active ATM at sites of DNA damage.

FIG. 1.

Conserved motifs at the C terminus of Nbs1 mediate binding to Tel1/ATM. (A) Two-hybrid assays of S. pombe Tel1 and Mre11 (Rad32) with full-length Nbs1, a C-terminal truncation of 60 amino acids (Nbs1-ΔC60), or the C-terminal 50 amino acids of Nbs1 (Nbs1-C50). Growth on SC-HTLA medium indicates a positive two-hybrid interaction. (B) A mutation in the Mre11-binding domain of Nbs1 abolishes the two-hybrid interaction with Mre11 without diminishing the interaction between Nbs1 and Tel1. The K522A mutation was created in the Mre11 (Rad32)-binding motif of full-length Nbs1. (C) Alignment of C termini of Nbs1 proteins. The conserved Mre11-binding region, acidic patch, and FXF/Y motif are shown. Mutations created in fission yeast, Xenopus, and human Nbs1 proteins are shown. (D) Two-hybrid assays involving S. pombe Nbs1 and Tel1, human Nbs1, and ATM. (E) Coprecipitation of Nbs1 with Tel1 in S. pombe. TAP-Tel1 expressed from the nmt1 promoter at the tel1 genomic locus while 13myc-tagged Nbs1 was expressed from its own promoter at the endogenous locus. TAP-Tel1 was precipitated with protein A Sepharose, and Nbs1-13myc was detected with anti-myc antibodies.

FIG. 2.

C-terminal mutations in Nbs1 ablate Tel1 function without impairing the function of Nbs1 in survival of DNA damage. (A) Detection of phospho-histone H2A in chromosome spreads of fission yeast in response to 100 Gy IR. Rad3 and Tel1 both phosphorylate H2A, and therefore rad3Δ and tel1Δ single mutants have phospho-H2A foci, whereas the rad3Δ tel1Δ double mutant has no phospho-H2A foci. In a rad3Δ background, mutations in the Tel1-binding motif of Nbs1 abolish induction of phospho-H2A foci. (B) Detection of YFP-Crb2 foci in live cells. Tel1-binding motif mutations in Nbs1 in a rad3Δ background abolish induction of YFP-Crb2 foci in response to 30 Gy IR. (C) IR and MMS survival assays show that mutations in the Tel1-binding domain of Nbs1 do not increase the sensitivity to DNA damage.

FIG. 3.

Nbs1 interacts with specific HEAT repeats 17-18 and 21-22 in Tel1. (A) Schematic of two-hybrid assays with fission yeast Nbs1 and Tel1. (B) Two-hybrid analyses of the regions of Tel1 that interact with Nbs1. The regions of Tel1 that are encoded by the two-hybrid plasmids are described in Materials and Methods. Nbs1-CT indicates the C-terminal 50 amino acids of Nbs1.

FIG. 4.

Characterization of antibodies. (A) Anti-C147 antibodies from rabbits, which were affinity purified with GST-C147 that contained the C-terminal 147 amino acids of Xenopus NBS1, were immunoblotted with Xenopus egg extract. In addition to Nbs1, a larger-molecular-weight protein (indicated by an asterisk) was also recognized by anti-C147. (B) Anti-C50 antibodies from rabbits, which were affinity purified with GST-C50 that contained the C-terminal 50 amino acids of Xenopus NBS1, were immunoblotted with Xenopus egg extract. (C) Egg extract was treated with control IgG or anti-C147 to immunodeplete Nbs1. Following depletion, extracts were immunoblotted with anti-C147. Depletion of Nbs1 from the extract using anti-C147 removed Nbs1 but not the larger protein shown in panel A. (D) Anti-ATM antibodies were generated in rabbits using GST-XATMC containing the C-terminal 300 amino acids of Xenopus ATM. These antibodies were affinity purified with His-GFP-XATMC and immunoblotted with Xenopus egg extract. (E) Shown is an immunoblot of Xenopus egg extract and HeLa cell lysate with anti-Mre11 antibodies purchased from Oncogene Research Products.

FIG. 5.

The acidic patch and FXY motif are essential elements of an ATM-binding domain at the C terminus of Xenopus NBS1. (A) NBS1 coprecipitates with ATM in Xenopus egg extracts. ATM or control immunoprecipitates from Xenopus egg extracts were washed four times with buffer 1 (without detergent) or buffer 2 (with 0.1% NP-40), followed by immunoblotting for ATM and NBS1. (B) Mre11 coprecipitates with Nbs1. Control, anti-C147, and anti-C50 immunoprecipitates of NBS1 from Xenopus egg extracts were washed four times with buffer 1 or buffer 2, as described for panel A, followed by immunoblotting for NBS1 and Mre11. (C) The C-terminal 50 amino acids of Xenopus NBS1 bind ATM but not Mre11, whereas the C-terminal 147 amino acids of NBS1 bind to both ATM and Mre11. Glutathione beads coupled with 20 μg of GST, GST-C147, or GST-C50 recombinant proteins were incubated with 40 μl of Xenopus egg extracts for 2 h at 4°C. The beads were isolated and washed four times with buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, 0.1% NP-40). Proteins associated with the beads were immunoblotted for ATM and Mre11. (D) The acidic patch and FXY motif are essential elements of an ATM-binding domain at the C terminus of Xenopus NBS1. The procedure performed was the same as that described for panel C except that glutathione beads were coupled with 20 μg of GST, GST-C50, GST-C50m1, or GST-C50m2 recombinant proteins.

FIG. 6.

The C-terminal region of NBS1 is required for ATM activation following addition of linear plasmid DNA or damaged sperm chromatin to the Xenopus egg extract. (A) Addition of linear double-stranded DNA induces ATM and NBS1 phosphorylation. Xenopus egg extract was incubated for 15 min with no DNA (lane 1) or 1 ng/μl of pBluescript circular plasmid DNA that was incubated with buffer (lane 2), AluI DNA endonuclease (lane 3), or heat-inactivated AluI (lane 4). Extracts were then immunoblotted for pS1981-ATM, total ATM, and NBS1. The gel mobility shift of NBS1 in lane 3 was due to phosphorylation, which is partially dependent on ATM. (B) Immunodepletion of Nbs1 from the extract using the anti-C147 antibodies prevents ATM autophosphorylation induced by linear plasmid DNA. Mock-depleted extracts (lanes 2 and 4) or extracts depleted of NBS1 with anti-C147 antibodies (lanes 1 and 3) were incubated with 1 ng/μl or 20 ng/μl of linearized pBluescript for 15 min. Extracts were immunoblotted for pS1981-ATM, total ATM, and NBS1. Quantification of the immunoblot showed that Nbs1 immunodepletion removed 57% of the Mre11 and 28% of the ATM in the samples that contained 1 ng/μl DNA. Similar values were measured in the 20-ng/μl sample (60% and 24%, respectively). (C) Addition of the anti-C147 and anti-C50 antibodies to the Xenopus egg extract blocks ATM autophosphorylation induced by linear plasmid DNA. Extracts containing 300 ng/μl of control IgG or affinity-purified anti-C147 or anti-C50 NBS1 antibodies were incubated with 1 ng/μl of linearized pBluescript for 15 min. Extracts were immunoblotted for pS1981-ATM and total ATM. (D) The inhibitory effects of the anti-C50 and anti-C147 antibodies on ATM autophosphorylation induced by linear plasmid DNA are abrogated by preincubating the antibodies with GST-C50. Anti-C147 or anti-C50 NBS1 antibodies (250 ng/μl) were preincubated with 2 μg/μl of GST or GST-C50 for 20 min at room temperature before being added to extracts containing 1 ng/μl of linearized pBluescript. Following 15 min of incubation, extracts were immunoblotted for pS1981-ATM and total ATM. (E) EcoRI-treated sperm chromatin induces ATM and NBS1 phosphorylation. Xenopus egg extract was incubated for 15 min with no DNA (lane 1) or 2,500 demembranated sperm nuclei/μl that were incubated with buffer (lane 2), EcoRI DNA endonuclease (lane 3), or heat-inactivated EcoRI (lane 4). Extracts were then immunoblotted for pS1981-ATM, total ATM, and NBS1. (F) Damaged chromatin fails to induce ATM autophosphorylation in the NBS1-depleted samples. Mock-depleted extracts (lanes 1 and 3) or extracts depleted of NBS1 with anti-C147 antibodies (lanes 2 and 4) were incubated with 2,500 EcoRI-treated demembranated sperm nuclei/μl for 15 min or 30 min. Extracts were immunoblotted for pS1981-ATM, total ATM, and NBS1. Quantification of the immunoblot showed that Nbs1 immunodepletion removed 47% of the Mre11 and 20% of the ATM in the 15-min sample. The corresponding numbers for the 30-min sample were 40% and 21%, respectively. (G) Anti-C50 and anti-C147 antibodies prevent ATM autophosphorylation induced by damaged chromatin. Xenopus egg extracts containing 250 ng/μl of control IgG or affinity-purified anti-C147 or anti-C50 NBS1 antibodies were incubated with 2,500 EcoRI-treated demembranated sperm nuclei/μl for 15 min. Extracts were immunoblotted for pS1981-ATM and total ATM. (H) The inhibitory effects of the anti-C50 and anti-C147 antibodies on ATM autophosphorylation induced by damaged chromatin are abrogated by preincubating the antibodies with GST-C50. Anti-C147 or anti-C50 NBS1 antibodies (250 ng/μl) were preincubated with 2 μg/μl of GST or GST-C50 for 20 min at room temperature before being added to extracts containing 2,500 EcoRI-treated demembranated sperm nuclei/μl. Following 15 min of incubation, extracts were immunoblotted for pS1981-ATM and total ATM.

FIG. 7.

Recruitment of ATM to DSBs precedes its autophosphorylation, and ATM autophosphorylation in Nbs1-depleted extracts can be restored by the C-terminal 147 amino acids of Nbs1. (A) The anti-C147 antibodies prevent ATM binding to the linear plasmid DNA. Extracts preincubated with 300 ng/μl of control IgG or affinity-purified anti-C147 were incubated with magnetic avidin beads coupled with a one-end-biotinylated 2-kb DNA fragment derived from pBluescript. Extracts were incubated with extracts at room temperature for 5 min. Proteins associated with beads were immunoblotted for S1981-ATM (unphosphorylated ATM), pS1981-ATM, total ATM, Ku70, and NBS1. (B) S1981 phosphorylation is not a prerequisite for the interaction of ATM with linear plasmid DNA. Reactions were similar to panel A except samples were not preincubated with antibodies and samples were taken at 5-min intervals following addition of DNA beads. (C) The anti-C147 antibodies prevent ATM binding to damaged sperm chromatin. Extracts preincubated with 250 ng/μl of control IgG or affinity-purified anti-C147 were incubated with 2,500 EcoRI-treated demembranated sperm nuclei/μl. Extracts were incubated at room temperature for 5 min. Proteins associated with chromatin were immunoblotted for S1981-ATM (unphosphorylated ATM), pS1981-ATM, total ATM, NBS1, and Ku70. (D) A recombinant fragment of Nbs1 that contains the Mre11- and ATM-binding domains restores ATM autophosphorylation to an Nbs1-depleted extract. Anti-C147 antibodies were used to deplete Nbs1 from a Xenopus egg extract. Buffer, 100 ng/μl GST, GST-C147, GST-C97, or GST-C50 was added to the extract together with linearized plasmid DNA. After 15 min of incubation, extracts were then immunoblotted for pS1981-ATM and total ATM. (E) Linearized plasmid DNA and damaged sperm chromatin induce ATM autophosphorylation in an Nbs1-depleted extract supplemented with a recombinant fragment of Nbs1 that contains the Mre11- and ATM-binding domains. GST-C147 (100 ng/μl) was incubated with buffer, 1 ng/μl of linearized pBluescript, or 2,500 sperm chromatin treated with EcoRI/μl. After 15 min of incubation, extracts were then immunoblotted for pS1981-ATM and total ATM. (F) The C147 fragment of Nbs1 restores ATM binding to linearized plasmid DNA and damaged sperm chromatin in an Nbs1-depleted extract. The Nbs1-depleted extract supplemented with 100 ng/μl of GST or GST-C147 was incubated with magnetic avidin beads coupled with one-end-biotinylated 2-kb fragments derived from pBluescript or sperm chromatin treated with EcoRI. After 2 min of incubation, DNA beads and chromatin were isolated and proteins associated with DNA beads or damaged chromatin were immunoblotted for S1981-ATM, pS1981-ATM, total ATM, and Ku70. Similar results were observed after a 10-min incubation except that the S1981-ATM signal was very weak (Z. You and T. Hunter, unpublished data).

FIG. 8.

Model of ATM activation. (A) Location of FXF/Y motifs preceded by the acidic patch in Nbs1 proteins from Schizosaccharomyces pombe (Sp), Saccharomyces cerevisiae (Sc), Homo sapiens (Hs), Xenopus laevis (Xl), and Drosophila melanogaster (Dm). The HEAT interaction region of S. cerevisiae nucleoporin Nsp1 is also shown. The lower panel shows a schematic representation of the Nbs1 proteins and the positions of FHA domains, BRCT domains, Mre11-interaction motif (MIM), and FXF/Y motifs. (B) Structure of an FXFG peptide interacting with HEAT repeats 5 and 6 of importin-β (6). (C) Proposed mechanism of ATM activation by DSBs. In step 1, the inactive ATM/MRN complex is recruited to sites of DNA damage. In step 2, ATP binding and hydrolysis induce conformational changes within the DNA-bound MRN/ATM complex that disengage the ATM dimer. ATM dimer disengagement leads to ATM autophosphorylation and phosphorylation of downstream effectors, such as H2AX. In step 3, the amplification of checkpoint signal by recruitment, activation, and release of ATM and by recruitment of additional MRN/ATM complexes via interactions with phospho-H2AX and MDC1 is shown. See the text for a more detailed description.