Phosphorylation of the Bloom's syndrome helicase and its role in recovery from S-phase arrest

Abstract

Bloom's syndrome (BS) is a human genetic disorder associated with cancer predisposition. The BS gene product, BLM, is a member of the RecQ helicase family, which is required for the maintenance of genome stability in all organisms. In budding and fission yeasts, loss of RecQ helicase function confers sensitivity to inhibitors of DNA replication, such as hydroxyurea (HU), by failure to execute normal cell cycle progression following recovery from such an S-phase arrest. We have examined the role of the human BLM protein in recovery from S-phase arrest mediated by HU and have probed whether the stress-activated ATR kinase, which functions in checkpoint signaling during S-phase arrest, plays a role in the regulation of BLM function. We show that, consistent with a role for BLM in protection of human cells against the toxicity associated with arrest of DNA replication, BS cells are hypersensitive to HU. BLM physically associates with ATR (ataxia telangiectasia and rad3(+) related) protein and is phosphorylated on two residues in the N-terminal domain, Thr-99 and Thr-122, by this kinase. Moreover, BS cells ectopically expressing a BLM protein containing phosphorylation-resistant T99A/T122A substitutions fail to adequately recover from an HU-induced replication blockade, and the cells subsequently arrest at a caffeine-sensitive G(2)/M checkpoint. These abnormalities are not associated with a failure of the BLM-T99A/T122A protein to localize to replication foci or to colocalize either with ATR itself or with other proteins that are required for response to DNA damage, such as phosphorylated histone H2AX and RAD51. Our data indicate that RecQ helicases play a conserved role in recovery from perturbations in DNA replication and are consistent with a model in which RecQ helicases act to restore productive DNA replication following S-phase arrest and hence prevent subsequent genomic instability.

FIG. 1.

(A) Western blotting of nuclear extracts to confirm expression of BLM in GMO8505 transfectants. Lane 1, GMO8505 cells; lanes 2 to 4, transfectants of GMO8505 cells containing the pcDNA3 vector only (PSNG1, PSNG13, and PSNV4); lanes 5 to 7, cells expressing the BLM cDNA in pcDNA3 (PSNB2, PSNF5, and PSNP2). β-Tubulin was used as a loading control. (B) BS cells are hypersensitive to HU. Clonogenic survival analyses were conducted on three stable transfectants of GMO8505 cells containing either the pcDNA3 vector only (PSNG1, PSNG13, and PSNV4), which are indicated by the open symbols, or cells expressing the BLM cDNA in pcDNA3 (PSNB2, PSNF5, and PSNP2), which are indicated by the filled symbols. Cells were exposed to HU for 48 h before washing and being left for 3 weeks to allow colonies to develop. Analyses were performed in triplicate. Error bars represent standard errors of the means.

FIG. 2.

(A) Validation of anti-ATR antibodies. Immunoprecipitations were carried out by using two independent anti-ATR antibodies raised against the C-terminal region (ATR-C) or the N-terminal region (ATR-N) of ATR. Preimmune antisera were used as negative controls. The immunoprecipitates were Western blotted with an ATR antibody that recognizes the N-terminal domain of ATR (a generous gift of S. P. Jackson, University of Cambridge). ATR was found in the precipitate (IP) when either ATR-C or ATR-N was employed but was found only in the supernatant (SUP) when the preimmune serum was employed. (B) Immunoprecipitations were conducted as described for panel A, except that the Western blotting antibody employed was the commercial anti-ATR FRP1 sc1887 antibody (Santa Cruz Biotechnology). (C) ATR can be coimmunoprecipitated with BLM in cells exposed to replicational stress. Immunoprecipitations were carried out using IHIC33 anti-BLM antibody, or an immunoglobulin G control, on extracts from cells exposed to no treatment (−), HU, or aphidicolin (+), as indicated above the lanes. Theimmunoprecipitates were Western blotted for the presence of ATR by using the sheep anti-ATR-N antibodies. (D) BLM can be coimmunoprecipitated with ATR following replicational stress. Immunoprecipitations using the sheep anti-ATR-N antibodies were carried out for normal human lymphoblastoid cells expressing BLM (+/+) or lymphoblastoid BS cells (−/−) and Western blotted for BLM by using the IHIC33 antibody. (E) BLM and ATR colocalize to nuclear foci. Indirect immunofluorescence of BLM (red) and ATR (green) in untreated cells expressing BLM (PSNSF5) are shown. In the merged image, colocalization is shown as yellow. The DNA (blue) panel shows the position of the nucleus, as judged by staining with Hoechst. (F) BLM is not required for the formation of ATR foci. Immunofluorescence analysis was performed as described in panel E, except that the PSNG13 vector-only control BS cells lacking BLM expression were used. (G) The percentage of cells with ATR foci increases in response to HU treatment, and the degree of colocalization of BLM and ATR also increases in response to HU treatment. Cells were incubated with 5 mM HU and fixed for immunofluorescence analyses at the times indicated. The percentage of cells with ATR-, BLM-, and ATR-BLM-colocalizing foci were determined microscopically. A minimum of 100 cells of each type were assessed at each time point.

FIG. 3.

(A) Schematic representation of the two potential PI-3 kinase recognition sites in the N-terminal region of BLM. The sites at Thr-99 and Thr-122 in each case contain a downstream glutamine residue, as required for recognition by PI-3 kinases. The putative PCNA binding motif QQRVKDFF⁹⁰ (conserved residues underlined) is indicated. (B) Phosphorylation of GST-BLM recombinant fragments representing the N-terminal domain between residues 1 and 212. ATM was immunoprecipitated from cells expressing WT ATM (+/+) or from cells expressing mutated ATM (−/−). The immunoprecipitated material was used in phosphorylation reactions containing the GST-BLM-1-212 fragment containing either WT sequence, Thr-99 mutated to alanine (T99A), or Thr-122 mutated to alanine (T122A), as indicated above the lanes. Note the lack of phosphorylation of the T99A protein. (C) ATR kinase assays. ATR was immunoprecipitated from HeLa cell extracts and used to phosphorylate the GST-BLM derivatives depicted in panel B, a doubly substituted derivative (T99A/T122A), and p53 (to act as a positive control) or GST (to act as a negative control). Note the lack of phosphorylation of the T99A/T122A BLM protein. (D) The kinase responsible for phosphorylation of BLM in the immunoprecipitates is ATR. ATR was immunoprecipitated using an anti-Flag antibody from cells expressing inducible WT ATR or a KD version of ATR, and the immunoprecipitate was used to phosphorylate the same proteins as depicted in panel C. Note the considerably reduced phosphorylation of the T99A/T122A BLM derivative.

FIG. 4.

Formation of BLM foci and colocalization of BLM foci with RAD51, phosphorylated histone H2AX, ATR, PCNA, and topoisomerase IIIα is not dependent upon phosphorylation of BLM on residues Thr-99 or Thr-122. Immunofluorescence analyses were conducted as described in Fig. 2 for cells either untreated (lower 3 rows of panels) or treated with 5 mM HU for 5 h prior to fixation. Colocalization in the merged image is indicated by yellow coloration.

FIG. 5.

Cells expressing the ATR phosphorylation site mutants arrest in G₂ following release from replication arrest. (A) Western blotting of crude nuclear extracts to show equivalent BLM expression in control PSNF5 cells (lane 1) and four independent clones of GMO8505 cells expressing BLM-T99A/T122A is shown; clone C2.1 (lane 3), clone C20 (lane 4), clone C1.2 (lane 5) and clone C1.4 (lane 6) are depicted. Lane 2 indicates the untransfected GMO8505 cells for validation of the antibody. β-Tubulin was used as loading control. (B) Exponentially growing cultures of either PSNF5 cells expressing WT BLM or clone 1.4 cells expressing BLM-T99A/T122A were exposed to HU (2 mM) for 16 h and then released into drug-free medium. The DNA content was analyzed by flow cytometry at the times indicated. (C) Analysis of a caffeine-sensitive G₂/M checkpoint in HU-treated cells expressing BLM-T99A/T122A. Cells were treated with HU as before and then released in the presence or absence of 4 mM caffeine. Cells were harvested at the time points indicated and analyzed by flow cytometry for expression of phosphorylated histone H3 to determine the percentage of cells in mitosis (left panel) and for DNA content (right panel). Cells expressing phosphorylated histone H3, and with a 2 N DNA content, are circled and labeled with an M for the 24-h time point. The percentage of cells expressing phosphohistone H3 is shown graphically. Results are expressed as the means of three experiments. Error bars indicate standard errors of the means. (D) Expression of phospho-Chk1. Whole-cell extracts were prepared from cells expressing T99A/T122A-BLM at the times indicated after release from HU block. The extracts were analyzed by Western blotting for total Chk1, phospho-Chk1 (Ser 345), and β-tubulin levels (shown by representative blots). The bar graph indicates quantification of the blotting data (normalized for β-tubulin levels) and represents the means of two determinations.

FIG. 6.

Model for the role of ATR-mediated phosphorylation of BLM in recovery from replicative stress. Two parallel pathways exist. Pathway 1 (left) is activated in cells expressing BLM protein (+BLM), while pathway 2 (right) is operational in BS cells lacking BLM protein. In pathway 1, cells expressing WT BLM (BLM-WT) display ATR-dependent phosphorylation of Thr-99 and Thr-122 (circled P), and these cells recover from replicative stress via a pathway that avoids SCEs. In cells expressing BLM-T99A/T122A, ATR cannot phosphorylate BLM (circled P with a cross) but are committed to the BLM-dependent pathway, resulting in aberrant recovery associated with G₂ checkpoint arrest.