Transcriptional downregulation of ATM by EGF is defective in ... - Nature

Oncogene (2001) 20, 4281 ± 4290 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

ORIGINAL PAPERS

Transcriptional downregulation of ATM by EGF is defective in ataxia-telangiectasia cells expressing mutant protein Katherine E Keating1,4, Nuri Gueven2,4, Dianne Watters3, H Peter Rodemann2 and Martin F Lavin*,1,3 1

The Queensland Cancer Fund Research Laboratory, The Queensland Institute of Medical Research, PO Royal Brisbane Hospital, Brisbane, Qld. 4029, Australia; 2Section for Radiobiology and Molecular Environmental Research, RoÈntgenweg 11, 72076 TuÈbingen, Germany; 3The Department of Surgery, University of Queensland, PO Royal Brisbane Hospital, Brisbane, Qld. 4029, Australia

There is evidence that ATM plays a wider role in intracellular signalling in addition to DNA damage recognition and cell cycle control. In this report we show that activation of the EGF receptor is defective in ataxia-telangiectasia (A-T) cells and that sustained stimulation of cells with EGF downregulates ATM protein in control cells but not in A-T cells expressing mutant protein. Concomitant with the downregulation of ATM, DNA-binding activity of the transcription factor Sp1 decreased in controls after EGF treatment but increased from a lower basal level in A-T cells to that in untreated control cells. Mutation in two Sp1 consensus sequences in the ATM promoter reduced markedly the capacity of the promoter to support luciferase activity in a reporter assay. Overexpression of anti-sense ATM cDNA in control cells decreased the basal level of Sp1, which in turn was increased by subsequent treatment of cells with EGF, similar to that observed in A-T cells. On the other hand full-length ATM cDNA increased the basal level of Sp1 binding in A-T cells, and in response to EGF Sp1 binding decreased, con®rming that this is an ATM-dependent process. Contrary to that observed in control cells there was no radiation-induced change in ATM protein in EGF-treated A-T cells and likewise no alteration in Sp1 binding activity. The results demonstrate that EGF-induced downregulation of ATM (mutant) protein in A-T cells is defective and this appears to be due to less ecient EGFR activation and abnormal Sp1 regulation. Oncogene (2001) 20, 4281 ± 4290. Keywords: ataxia-telangiectasia ATM; transcriptional downregulation; EGF; Sp1 transcriptional factor

*Correspondence: MF Lavin 4 These two authors contributed equally to the work Received 27 November 2000; revised 6 April 2001; accepted 9 April 2001

Introduction The human genetic disorder ataxia-telangiectasia (A-T) is characterized by a complex phenotype including neurodegeneration, immunode®ciency and a predisposition to develop tumours primarily leukaemias and lymphomas (Sedgwick and Boder, 1991). Positional cloning led to the identi®cation of the gene (ATM) defective in this syndrome (Savitsky et al., 1995). The feature most widely studied in A-T is increased sensitivity to ionizing radiation (Goto et al., 1967; Taylor et al., 1975; Lavin and Shiloh, 1997). It appears likely that ATM is a sensor of DNA damage that leads to the repair of double strand breaks or a form of damage that is converted into a break in DNA (Cornforth and Bedford, 1985; Foray et al., 1997). Recognition of a speci®c form of DNA damage activates ATM by an unknown mechanism to trigger a number of cell cycle checkpoints (Beamish and Lavin, 1994; Nagaswa and Little, 1983). The best described of these is the G1/S checkpoint where a defect in the stabilization of p53 was initially described in A-T cells (Kastan et al., 1992). This defect was subsequently shown to extend to the p53 eector genes p21/WAF1, GADD45, MDM2, cyclin E ± cdk2 kinase and downstream substrates (Canman et al., 1994; Dulic et al., 1994; Khanna et al., 1995). The mechanism of activation/stabilization of p53 in response to DNA damage is complex involving both phosphorylation and acetylation at several sites (Sakaguchi et al., 1998; Shieh et al., 1997). Evidence for a direct involvement of ATM with p53 was provided by the demonstration that these two molecules interact directly (Watters et al., 1997) and ATM phosphorylates p53 on serine 15 in response to DNA damage (Banin et al., 1998; Canman et al., 1998; Khanna et al., 1998). These data suggest that ATM is the protein kinase responsible for the rapid phosphorylation and activation of p53 in response to ionizing radiation damage to DNA. ATR, (ATM ± and rad3 ± related protein) another member of the phosphatidylinositol 3-kinase (PI3kinase) family also phosphorylates p53 on serine 15 in response to DNA damage but this is signi®cantly delayed compared to ATM-mediated phosphorylation

Transcriptional regulation of ATM KE Keating et al

4282

Oncogene

(Tibbetts et al., 1999). A number of other substrates have been reported for ATM kinase in response to DNA damage. These include c-Abl (Khanna et al., 1997; Baskaran et al., 1997); BRCA1 (Cortez et al., 1999; Gatei et al., 2000); Nibrin (NBS1) (Gatei et al., 2000; Lim et al., 2000; Wu et al., 2000; Zhao et al., 2000); mdm2 (Khosravi et al., 1999) and chk2 (Chaturvedi et al., 1999; Matsuoka et al., 1998). These proteins participate in DNA damage recognition, signal transduction and cell cycle checkpoint activation demonstrating that ATM plays a central/upstream role in the ecient induction of a major cell cycle signalling network. Compartmentalization of ATM, predominantly in the nucleus of proliferating cells, is supportive of a role in DNA damage recognition and cell cycle control. ATM has also been localized to cytoplasmic vesicles (Brown et al., 1997; Lim et al., 1998; Watters et al., 1997), is a cytoplasmic protein in human and mouse brain and its absence leads to increased lysosomal numbers (Oka and Takashima, 1998; Barlow et al., 2000). Several lines of evidence have implicated ATM in extra-nuclear initiated signal transduction including defective signalling through the T-cell and B-cell receptors in A-T cells (Khanna et al., 1997; Kondo et al., 1993; a defect in mitogen-mediated signalling in AT lymphocytes (O'Connor and Scott-Linthicum, 1980); only partial protection by cytokines against apoptosis in A-T cells (Lynn and Wang, 1997); a greater demand for growth factors in A-T ®broblast survival (Shiloh et al., 1982); poor growth capacity of embryonic ®broblasts from atm7/7 mice (Xu and Baltimore, 1996), defective regulation of K+ channel activity (Rhodes et al., 1998) and more recently Yang and Kastan (2000) have reported that ATM is involved in an insulinsignalling pathway that controls initiation of protein synthesis. These observations might well be explained by the coupling of ATM function to signalling through one or more receptors on the cell surface. This could also provide an explanation for part of the defective response to radiation in A-T cells (Lavin and Shiloh, 1997), since it has been demonstrated that agents such as ionizing radiation are capable of initiating a response outside the nucleus by providing either gain of function or by abrogating the function of receptor tyrosine kinases (Knebel et al., 1996; Kwok and Sutherland, 1991, 1992). The epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase that can be activated by a variety of agents including oxidants that inhibit receptor tyrosine dephosphorylation in a thiol-sensitive and reversible manner (Knebel et al., 1996). EGF has been reported to enhance the radiosensitivity of some cell types (Balaban et al., 1996; Knebel et al., 1996; Kwok and Sutherland, 1991; Nelson, et al., 1989) and increase the radioresistance of other cells (Woltman et al., 1994). We have recently shown that EGF increases radiosensitivity in both human ®broblasts and B lymphoblastoid cells by downregulating ATM and DNA-dependent protein kinase (DNA-PK) at the mRNA and protein levels (Gueven et al., 2001). Under these conditions the DNA

binding activity of Sp1 was also decreased and a link was established between these processes. We have extended these studies to A-T cells and demonstrate that EGF signalling and transcriptional downregulation of ATM mutant protein is defective in these cells.

Results Activation of EGFR Engagement of the EGF receptor by EGF and other ligands initiates signal transduction pathways giving rise to mitogenesis and can also alter the radiosensitivity status of cells (Aaronson, 1991; Carpenter, 1987; Kwok and Sutherland, 1991). Since we have previously shown that EGF sensitizes control but not A-T cells to ionizing radiation we examined the initial step in this pathway by comparing EGFR activation in A-T and control ®broblasts using an antibody that recognizes activated EGFR. The results, outlined in Figure 1A demonstrate that EGFR is rapidly activated (within 10 min) by incubation of control (NFF) ®broblasts with EGF (50 ng/ml). Immunoblotting with an anti-EGFR antibody also detects multiple bands after EGF treatment representing dierentially phosphorylated forms (Figure 1A). A second control ®broblast line (HDF) showed a similar pattern of activation (results not shown). On the other hand there was only a minimal change in activation of EGFR in A-T ®broblasts (AT5BI) (Figure 1A) and this was con®rmed by immunoblotting with anti-EGFR antibody. EGFR activation in a second A-T ®broblast (GM02052C) was also de®cient compared to the control and this was also con®rmed by immunoblotting with anti-EGFR antibody (Figure 1B). It is evident that the amount of EGFR is somewhat reduced in both A-T ®broblast lines and stimulation with EGF leads to a marked decrease in EGFR protein compared to controls where the amount remains largely unchanged (Figure 1A,B, middle panels). A third A-T ®broblast line (GM03395) was also defective in EGFR signalling (results not shown). The mutation in AT5B1 (7281del6) would be predicted to give rise to near fulllength protein (Savitsky et al., 1995), while in GM02052C a homozygous mutation results in a truncated unstable protein (Gilad et al., 1996) and the mutation in GM03395 is not known. None of these A-T ®broblast lines showed activation of ATM kinase in response to radiation (results not shown). Effect of EGF on ATM mutant and protein expression We have previously shown that EGF downregulates the amount of ATM protein in both ®broblasts and lymphoblastoid cells at the transcriptional level (Gueven et al., 2001). To check this in A-T cells we selected ATIABR cells (homozygous for an inframe mutation (7636del9)) that express near full-length ATM mutant protein detectable by immunoblotting (Watters et al., 1997). The results in Figure 2A reveal


4283

Figure 1 Dierential activation of the EGFR in control and A-T cells. (a) EGFR activation levels in NFF (control) and AT5BI ®broblasts, treated with 50 ng/ml EGF for 10 and 30 min, was determined by immunoblotting 100 mg of whole cell lysate with the EGFR antibody E12120 (upper panel) which detects the activated form of the receptor. The EGFR antibody E1005 (middle panel) was used to determine the total amount of EGFR protein. Protein loading was determined using the a actin antibody A4700. (b) Activation of EGFR in control and GM02052C, A-T ®broblasts. Immunoblotting was performed as described in (a)

Figure 2 Eect of EGF incubation on the amount of ATM protein in A-T and control cells (a) Eect of EGF (50 ng/ml) incubation on ATM protein in AT1ABR cells. Immunoblotting was carried out with 100 mg of extract to determine ATM levels using the anti-ATM antibody, ATM4BA. Since DNA-PKcs was present in greater abundance than ATM, one sixth the amount of protein was loaded for DNA-PKcs blotting. Actin was used as a loading control. (b) The eect of EGF (50 ng/ml) treatment (16 h) on ATM levels in dierent A-T cell lines (AT1ABR, GM1525, AT3LA, AT9ABR, AT17ABR and AT5ABR) and a control (C3ABR) lymphoblastoid line. Protein loading in each lane was determined by Poinceau S staining (not shown)

that EGF fails to alter the amount of ATM in AT1ABR cells up to 24 h incubation. Under these conditions the amount of DNA-PKcs also remained unchanged. In three other A-T cell lines expressing ATM mutant protein GM1525, AT17ABR (unusually stable protein) and AT5ABR (very unstable protein), EGF failed to alter the level of protein (Figure 2B). In two other A-T cell lines, AT3LA and AT9ABR, with mutations predicted to give rise to truncated proteins (Watters et al., 1997), there was no eect of EGF as expected. Under these conditions ATM protein is reduced in a control cell line C3ABR (Figure 2B).

Previous results using quantitative PCR and Northern blotting have demonstrated that the EGF-induced decrease in ATM is due to downregulation of ATM mRNA (Gueven et al., 2001). We also employed RT ± PCR here to show that there was no alteration in the amount of mRNA after EGF treatment of AT1ABR cells (results not shown). Sp1-binding activity after EGF treatment ATM and NPAT (nuclear protein at A-T locus) are arranged in a head-to-head con®guration in both the Oncogene


4284

human and mouse genomes (Byrd et al., 1996; Imai et al., 1996). The bidirectional promoter (approximately 0.5 kb in size) contains two Sp1 transcription factor, consensus binding sequences in the ATM gene direction. Accordingly, we determined whether Sp1 DNA binding activity was altered in A-T cells in response to EGF. Gel-shift analysis with an oligonucleotide-binding consensus sequence for Sp1 revealed the presence of a single, well-de®ned, retarded band, the basal level of which was low in AT1ABR extracts (Figure 3A). A transient, further decrease at 3 h posttreatment with EGF was observed but there was a marked increase in binding at later times after treatment (Figure 3A). Under these conditions the amount of Sp1 protein did not change in the ATIABR cells (Figure 3B). The extent of the binding in EGF-treated AT1ABR cells was comparable to that in the untreated C3ABR cells (Figure 3C). To check the universality of this, dierent A-T lines were incubated with EGF for 16 h prior to measuring Sp1-binding activity revealing that ®ve of six A-T cell lines showed a signi®cant increase in activity and the other sample (AT5ABR) a smaller increase (Figure 3C). This was the case irrespective of whether the A-T cell lines expressed ATM protein or not. The decrease in Sp1 binding activity in response to EGF in the control cell line C3ABR was also observed previously in other control cell lines and the binding was speci®c for Sp1 since it was supershifted with antibody to Sp1 (Gueven et al., 2001). Further evidence to support Sp1 involvement was obtained by mutating the Sp1 consensus sequences in the ATM promoter in a reporter assay. Replacement of GGGCGGGG with GAACGGGG in the ®rst putative Sp1 site (7440 ± 441) resulted in a reduction of luciferase activity to 0.1% of basal wild-type activity. The same replacement of GG with AA in the second Sp1 consensus site (7425 ± 426) led to a similar reduction in luciferase activity (results not shown). Effect of ATM cDNA sense and anti-sense expression To further con®rm that the dierence in response to EGF in A-T and control cells was ATM-dependent we employed full-length ATM cDNA and anti-sense constructs under the control of an inducible metallothionein promoter, transfected into either A-T or control cells. In control cells stably transfected with anti-sense ATM and induced with CdCl2, ATM protein decreases markedly and the cells exhibit a phenotype similar to that for A-T (Zhang et al., 1998). Expression of ATM anti-sense pMAT2 (+ CdCl2) in C3ABR cells led to a decrease in basal Sp1 binding reminiscent of that seen in A-T cell extracts (Figure 4). Furthermore, EGF treatment of C3ABR cells expressing pMAT2 led to a marked increase in Sp1-DNAbinding as observed previously in A-T cells, which contrasts with the results seen in uninduced pMAT2transfected cells treated with EGF alone where a decrease occurred as expected (Figure 4). On the other hand expression of full-length ATM cDNA increased

Oncogene

Figure 3 Eect of EGF on nuclear protein binding to the Sp1binding consensus sequence in AT1ABR cells. (a) Nuclear extracts (20 mg), prepared at dierent times after EGF addition were incubated with 32P-labelled double-stranded oligonucleotide (22 mer) containing a single Sp1 binding site. Free and retarded fragments were separated on native 5% acrylamide gels. (b) Eect of EGF on the amount of Sp1 protein in AT1ABR cells with time of incubation. Actin was used as a loading control. (c) Dierent A-T cell lines (AT1ABR, GM1525, AT3LA, AT9ABR, AT17ABR and AT5ABR) were untreated or treated with EGF for 16 h (7 and+respectively) prior to preparation of extracts for binding to the Sp1 consensus sequence as described in a. C3ABR was used as a control

the basal level of Sp1 binding in AT1ABR cells but when these cells were also treated with EGF, Sp1 binding decreased as observed previously in control


Figure 4 Eect of sense and antisense ATM cDNA transfection on Sp1 binding in EGF treated A-T and control cells. Extracts were prepared from C3ABR cells stably transfected with an antisense ATM cDNA construct (pMAT2) and from AT1ABR cells transfected with full-length ATM cDNA (pMAT1) (Zhang et al., 1997, 1998). Cells were treated with CdCl2 for 8 h to induce transcript (anti-sense or sense) followed by 16 h incubation with EGF (50 ng/ml). Sp1 binding was determined as described in the legend to Figure 3

AT1ABR cells, EGF increased basal Sp1 binding to a level comparable to that in the untreated control and there was no further change with GM-CSF (Figure 5A) and the amount of mutant ATM did not change (Figure 5B). As described in Figure 2, EGF reduces ATM protein in controls but not in A-T cells. Exposure of control cells to radiation, subsequent to EGF treatment, led to an increase in both ATM protein and ATM kinase activity (in vitro) which was somewhat less than that in untreated, irradiated cells at 30 min post-irradiation (Figure 6A, compare lanes 1 and 3 and lanes 2 and 4). However by 3 h postirradiation both ATM protein and ATM kinase activity were comparable to values in cells not exposed to EGF (Figure 6A, lanes 5 and 6). When ATM kinase activity was determined in vivo using antibody against phosphorylated serine 15 of p53 and increase was also observed over the same time course. On the other hand only a low level of non-inducible ATM kinase was observed in AT1ABR cells under the same conditions and ATM protein was unchanged in amount (Figure 6B). A delayed increase in serine 15 phosphorylation was obtained as observed previously in A-T cells.

4285

A

cells (Figure 4). Uninduced, pMAT1 transfected AT1ABR cells behaved the same as the parental line (AT1ABR) revealing a lower basal level of Sp1 binding which increased in response to EGF (Figure 4). Finally, transient transfection of A-T cells with a reporter construct driven by the ATM promoter as described in Gueven et al. (2001) revealed an increase in promoter activity in response to EGF compatible with the increase in Sp1 binding activity. Luciferase activity was 1.0+0.26 for untreated cells and 1.31+0.169 (P=0.09) for EGF-treated cells. This was the opposite to that observed in control cells where luciferase activity decreased in response to EGF treatment (Gueven et al., 2001).

B

Effect of radiation and GM-CSF on ATM mutant protein We have previously shown that EGF-reduced ATM protein and kinase activity in control cells are rapidly restored to constitutive levels by exposure to either ionizing radiation or GM-CSF (Gueven et al., 2001). A parallel increase in Sp1 DNA binding activity in response to these agents appears to be responsible for the increase in ATM. Here again we demonstrated that Sp1 binding activity was reduced by EGF in two control cell lines and rapidly restored when GM-CSF was subsequently added (Figure 5A). Concomitant with this the amount of ATM protein decreased after EGF treatment and increased to normal levels when GM-CSF was added (Figure 5B). In the case of

Figure 5 Eect of GM-CSF on ATM protein (a) Sp1 binding activity was determined as described in the legend to Figure 3. Controls (C37ABR and C2ABR) and A-T (AT1ABR) cells were treated with EGF for 16 h or with EGF followed by GM-CSF for 2 h or with GM-CSF only. (b) Changes in the amount of ATM protein in response to EGF and GM-CSF treatment. Cells were treated as described in (a). Immunoblotting was carried out for ATM and Sp1 protein and actin was used as a loading control Oncogene


4286

Figure 6 Eect of radiation on ATM kinase activity in EGF-treated control (C3ABR) and A-T (ATIABR) cells. ATM kinase was determined by immunoprecipitation and utilization of p531 ± 40 as a substrate. In vivo ATM kinase activity was determined by immunoblotting with antibody, speci®c for phosphorylated serine 15 on p53. Immunoblotting was used to determine ATM and p53 proteins and actin was used as a loading control

Discussion We have shown previously that incubation of human cells with EGF reduces the expression of ATM at both the mRNA and protein levels and sensitizes these cells to radiation (Gueven et al., 2001). Here we show that this response is defective in A-T cells due to abnormal signalling through EGFR. EGF treatment failed to decrease the amount of ATM protein in an A-T cell line, AT1ABR, which is characterized by an inframe deletion of three amino acids, producing near fulllength inactive ATM protein and it also did not change the response to radiation (Watters et al., 1997; Gueven et al., 2001). Three other A-T cell lines, expressing mutant ATM, were also refractive to downregulation of ATM by EGF treatment, suggesting that this is a general defect in A-T cells. This is the ®rst example of downregulation of ATM protein in response to a physiological stimulus (Gueven et al., 2001). In control proliferating cells, where ATM is present in normal levels, neither ionizing radiation nor radiomimetic chemicals change the amount or distribution of ATM in the cell (Brown et al., 1997; Watters et al., 1997). On the other hand ATM can be upregulated (from low basal levels) in peripheral blood mononuclear cells stimulated to divide with phytohaemagglutinin (PHA) (Fukao et al., 1999). In that case an increase in ATM protein and ATM kinase activity correlated with increased cell proliferation. In this study the nature of the defect in A-T cells can be traced to the inecient activation of the EGFR in ®broblasts exposed to EGF. The basal amount of EGFR is somewhat reduced in AOncogene

T ®broblasts compared to control but not to an extent that would explain the very poor activation response in the presence of EGF observed in all three A-T lines investigated. Instead it appears that shortly after EGF engagement, EGFR becomes unstable or is rapidly downregulated since the amount of protein decreases markedly. This may be another manifestation of a more general signalling defect in A-T cells. In this context Peretz et al. (2001) have shown that in the absence of ATM, insulin-like growth factor-1 receptor is downregulated. One of the earliest observations supporting a more general role for ATM was the failure to record a cytoplasm to nucleus signalling after internalization of PHA in A-T lymphocytes (O'Connor and Scott-Linthicum, 1980). Kondo et al. (1993) demonstrated that Ca2+ mobilization, in response to both PHA and CD3 cross-linking, was defective in lymphocytes from A-T patients. Defective signal transduction and Ca2+ mobilization has also been described after B cell receptor ligation in A-T lymphoblastoid cells (Khanna et al., 1997). More recently Rhodes et al. (1998), have shown that the ability of A-T ®broblasts to depolarize in response to increasing concentration of extracellular K+ is signi®cantly reduced compared to controls and that the outward recti®er K+ currents are largely absent in these cells. These data support a role for ATM in the regulation of K+ channel activity and membrane potential. Other evidence for defective signalling in A-T includes greater demand for growth factors in A-T ®broblasts (Elmore and Swift, 1976; Shiloh et al., 1982), poor growth capacity of ®broblasts from atm7/7


mice (Xu and Baltimore, 1996) and only partial protection against cell death in peripheral blood mononuclear cells from A-T patients by serum and added cytokines (Lynn and Wang, 1997). A more direct role for ATM in general signalling has been demonstrated by Yang and Kastan (2000) who showed that insulin-induced phosphorylation of eiF-4E-binding protein is mediated by ATM. The results outlined here demonstrate that signalling through EGFR modulates the activity and amount of ATM protein and this is defective in A-T cells. In order to explain defective signalling through several dierent receptors it suggests that there exists some step common to all these processes mediated by ATM or dierent targets for ATM kinase activity, each inherently associated with an early stage of receptormediated signalling. Alternatively, the generalized signalling defect might arise indirectly as a consequence of absence of or mutation in ATM. Loss of ATM is characterized by increased sensitivity to ionizing radiation which mediates its damaging eects through reactive oxygen species. Rotman and Shiloh (1998) have hypothesized that a number of the features in A-T patients and Atm7/7 mice including neurodegeneration arise as a consequence of oxidative stress. In support of this Barlow et al. (1999) have demonstrated that organs which develop pathologic changes in Atm7/7 mice are targets for oxidative damage and Purkinje cells in the cerebellum are exquisitively sensitive. In addition Watters et al. (1999) have reported increased lipid peroxidation and decreased catalase activity in A-T cells. A constitutively high level of reactive oxygen species in A-T cells might also lead to membrane damage and defects in universal receptor signalling. Approximately 0.5 kb separates the 5' end of the ATM gene from the 5' end of another gene NPAT, indicating that they share a bidirectional promoter (Byrd et al., 1996; Saito et al., 1998). This promoter contains several CCAAT boxes and Sp1 consensus sites but no TATA box, resembling a promoter for housekeeping genes. We have shown previously that Sp1 is a candidate for common regulation of these two genes and its downregulation or inactivation might account for the eect of EGF in reducing the amount of ATM (Gueven et al., 2001). A causal relationship was established in this regulation when we showed that GM-CSF could rapidly restore Sp1 DNA binding activity in EGF treated cells and in parallel restore the amount of ATM protein to constitutive levels (Gueven et al., 2001). Borellini and Glazer (1993) also showed that GM-CSF leads to a marked increase in Sp1 binding activity in proliferating cells. Evidence for a role for EGF in reducing the amount of Sp1 in a rat pituitary cell line (GH4) has been reported previously (Mortensen et al., 1997). We have provided additional data in this report supporting a role for Sp1 in the regulation of ATM. Mutation at either of the Sp1 consensus binding sequences in the ATM promoter reduced markedly the capacity of the promoter to express luciferase activity in a reporter assay. In the case of A-T cells, the basal level of Sp1 binding activity

was lower in 5/6 cell lines, compared to that observed in control cells, but the amount of Sp1 protein was unchanged and the amount of binding activity after EGF treatment increased in these A-T cell lines. GMCSF was not capable of increasing the low basal level of Sp1 binding in A-T cells nor did it alter the amount of ATM protein. An explanation for reduced levels of Sp1 binding activity in AT1ABR cells could be the presence of an inhibitor of this transcription factor. Indeed Chen et al. (1994) have described a heat-labile and protease-sensitive Sp1 negative regulator that speci®cally inhibits Sp1 binding to a c-jun Sp1 site. In addition recombinant retinoblastoma (Rb) protein stimulates Sp1 mediated transcription by liberating it from inhibitor (Chen et al., 1994), and Rb upregulates the Werner's syndrome gene promoter which is dependent on a region containing two Sp1 binding sites (Yamabe et al., 1998). Of particular signi®cance to this observation is our previous ®nding that Rb is largely present in a hyperphosphorylated form in A-T cells (Khanna et al., 1995). Based on the data of Chen et al. (1994), it would be expected that there would be less stimulation of the DNA binding activity of Sp1 in cells with a disproportionate amount of the hyperphosphorylated form of Rb and thus lower binding in A-T cell extracts as seen in this study. But why does Sp1 binding increase after EGF treatment? It is evident that the amount of Sp1 binding is dependent on ATM. Anti-sense to ATM decreased the basal level of Sp1binding activity and in the presence of EGF an increase in binding activity was obtained, similar to the pattern in A-T cells. Expression of ATM cDNA in A-T cells had the opposite eect, enhancing the basal activity and in the presence of EGF lowering Sp1 binding. It is not clear how ATM maintains the level of Sp1 activity but it is possible that Sp1 is a substrate for ATM or is downstream in a pathway regulated by ATM. What also is unclear is why, in the presence of increased Sp1 binding activity after incubation with EGF, mutant ATM protein does not increase in amount. In essence two important contributions to understanding the physiological role of ATM have been described here: ATM is downregulated by EGF in control cells and this signalling pathway is defective in A-T. The actual cause of the defect in EGFR activation and signalling in A-T remains unresolved but appears to be related to a more general signalling defect. Since we can correct the signalling defect downstream with full-length ATM cDNA it is likely to be a consequence of the A-T phenotype. The ability of EGF to change the level of ATM adds to a number of observations demonstrating that ATM is controlled in a complex fashion. ATM is activated by ionizing radiation and other agents (Banin et al., 1998; Canman et al., 1998); its distribution in the cell appears to be in¯uenced by the dierentiation state, being predominantly a cytoplasmic protein in human and mouse brain (Oka and Takashima, 1998; Barlow et al., 2000) and in oocytes (Barlow et al., 1998); it increases up to 10-fold in amount in response to mitogenic stimuli to

4287

Oncogene


4288

Materials and methods

(25 mM HEPES pH 8, 0.25 M sucrose, 1 mM EGTA, 5 mM MgSO4, 50 mM NaF, 1 mM DTT, 1 mM PMSF) and disrupted by at least four freeze-thaw-cycles. Insoluble material was removed by centrifugation at 13 000 g for 20 min. Protein concentration was determined using a BioRad DC protein assay kit according to the manufacturer's recommendations. All steps were carried out at room temperature. Protein samples were separated on 5 or 10% denaturing gels and blotted on nitrocellulose membranes. After blocking in 4% milk powder/0.1% Tween-20/PBS for at least 1 h the blot was incubated for 1 h at room temperature or overnight at 48C, with the relevant primary antibody. The blot was washed three times in 0.1% Tween 20/PBS and incubated with peroxidase-conjugated secondary antibody for 1 h. Following several washing steps, the blot was developed using a Chemoluminescence kit (Dupont).

Cells and chemicals

Gel retardation assay

Fibroblasts: normal human skin ®broblasts (NFF and HDF) were from healthy donors expressing ATM protein; A-T ®broblasts AT5BI (GM05823C), GM02052C and GM03395 were obtained from Coriell Cell Repositories. Lymphoblastoid cells: control, C3ABR, C37ABR and C2ABR and A-T cells, AT1ABR, AT5ABR, AT9ABR, AT17ABR, GM1525 and AT3LA have been described previously (Watters et al., 1997). All cells were grown at 378C in a humidi®ed atmosphere of 5% CO2 and 95% air in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) (15% FCS for GM02052C, AT5BI and GM03395). Irradiation of cells was performed at room temperature using a 137Cs source delivering gamma rays at a dose rate of 2.8 Gy/min. EGF (R&D Systems) was added at a concentration of 50 ng/ml to the culture medium of log-phase cultures or serum-reduced ®broblasts for various times as indicated. Antibodies used: polyclonal EGFR antibody (E1005), p53 monoclonal antibody (pAb 1801) and polyclonal Sp1 antibody (Sp1pEp2) were from Santa Cruz; monoclonal EGFR activated form (E12120) was from Transduction Laboratories; monoclonal a-actin (A4700) from Sigma; polyclonal anti-Ku and DNAPKcs antibodies were obtained from Susan Lees-Miller, University of Calgary, Canada; anti-phosphorylated p53 serine 15 was obtained from New England Biolabs and monoclonal ATM (CT1) and polyclonal ATM raised in sheep (ATM5BA) antibodies were generated in this laboratory.

The DNA-binding activity of transcription factor Sp1 was determined by modi®cation of a method previously described (Dittmann et al., 1998). Brie¯y, 107 cells were washed in icecold PBS, resuspended in 400 ml of buer A (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.2 mM PMSF) and incubated for 15 min on ice. After addition of 25 ml of 10% NP-40 the cells were subsequently vortexed for 10 s and the nuclei pelleted by centrifugation (1 min, 13 000 g at 48C). The nuclear pellet was resuspended in 40 ml of buer C (20 mM HEPES pH 7.9, 400 mM KCl, 1 mM EDTA, 1 mM EGTA, 10% glycerol, 1 mM DTT, 0.2 mM PMSF) and rotated for 30 ± 60 min at 48C to elute nuclear proteins. After additional centrifugation for 15 min at 48C the supernatant containing the nuclear proteins was aliquoted, frozen in liquid nitrogen and stored at 7708C until used in binding reactions. Approximately 35 fmol of 32 P-radiolabelled double-strand Sp1-consensus sequence (5'ATTCGATCGGGGCGGGGCGAGC-3'; Promega) (labelled using T4 polynucleotide kinase) was incubated with 10 mg of nuclear extract in the presence of 1 mg of herring sperm DNA and 20 mg of BSA in binding buer (10 mM HEPES pH 7.9, 50 mM KCl, 0.5 mM EDTA, 10% glycerol, 1% NP-40, 5 mM DTT, 0.2 mM PMSF) in a 20 ml reaction. Binding was for 25 min at room temperature before DNA-protein complexes were separated on a native 5% acrylamide gel. Exposure time to X-ray ®lm was usually 30 ± 45 min.

EGFR activation

Luciferase assays

Logarithmically growing ®broblasts were incubated in RPMI medium containing reduced serum (2%) for 24 h to enhance the expression of EGFR. Cells were treated with 50 ng/ml EGF and then harvested by scraping in PBS containing phosphatase inhibitors (10 mM EDTA, 50 mM NaF, 10 mM Na3VO4). Cells were lysed according to Bandyopadhyay et al. (1998) and cellular protein was estimated using the BioRad DC protein assay kit. Cell lysates containing 100 mg of protein were resolved on 5% SDS ± PAGE, transferred to membranes and probed with the antibodies E1005 and E12120 to detect EGFR and activated EGFR respectively. Equal protein loading was established by probing samples run on 10% gels with an antibody to actin.

The ATM promoter was ampli®ed from human genomic DNA as described previously (Gueven et al., 2001). The 615 bp ampli®cation product was cloned into the pGL3-basic luciferase reporter vector (Promega) using a SmaI/HindIII digest. Mutations were introduced into the two Sp1 consensus sequences by replacing underlined GG sequences with AA (GGGCGGGG, 7440 ± 441 and GGGCGGGG, 7425 ± 426) using a QuickChange in vitro mutagenesis kit (Stratagene). These constructs were co-transfected with the pRL-CMV vector (Promega) at a ratio of 2 : 1 into lymphoblastoid cells by electroporation (280 V, 960 mF, one pulse). Thirty-six hours after electroporation the cells were incubated with 50 ng/ml EGF for a further 16 h before cell extracts were prepared. Luciferase activity was measured using the Dual Luciferase Assay (Promega, E1910). Brie¯y, 100 ml, ®re¯y luciferase substrate (LARII) and 20 ml of cell extract were mixed and the reaction was immediately measured for 10 s. Then 100 ml of renilla luciferase substrate including an inhibitor for ®re¯y luciferase (Stop & Glow) was

peripheral blood mononuclear cells (Fukao et al., 1999) and it is upregulated dramatically in the myoepithelium of sclerosing adenosis of the breast (Clarke et al., 1998). The present results provide further evidence that ATM responds to multiple intracellular and extracellular initiated stimuli, in recognising damage in DNA and in controlling and perhaps responding to the proliferation status of the cell. Elucidation of this EGF signalling pathway for transcriptional control of ATM protein will assist in our understanding of a more general role of ATM in intracellular signalling.

Immunoblotting EGF-treatment was carried out on log-phase cells for dierent times as indicated. After washing in PBS, 56106 cells were resuspended in 40 ml of nuclear extraction buer Oncogene


added and light emission was detected for another 10 s interval. The ratio of both measurements (pGL/pRL) results in the relative luciferase activity, avoiding variabilities in transfection. The absolute values of luminescence measured (RLU) was generally in the range of 56104 ± 56106 for basal activity. The untreated control was set to 1.

noglobulin G and protein A/G-sepharose beads. ATM was immunoprecipitated with anti-ATM antibody (ATM5BA) and kinase activity determined as described. ATM kinase activity was also determined in vivo using an antibody that detects serine 15 phosphorylated p53 (Banin et al., 1998; Canman et al., 1998; Khanna et al., 1998).

4289

ATM-kinase assay ATM-kinase activity was determined using the method described by Canman et al. (1998). For whole cell lysate isolation, cells were lysed on ice in TGN buer (50 mM Tris pH 7.5, 50 mM b-glycerophosphate, 150 mM NaCl, 10% glycerol, 1% Tween-20, 1 mM NaF, 1 mM Na3 VO4, 1 mM PMSF, 2 mg/ml pepstatin, 5 mg/ml leupeptin, 10 mg/ml aprotinin, 1 mM DTT). After centrifugation at 13 000 g for 1 min, 2 mg of extract was precleared with mouse immu-

Acknowledgments We wish to thank the Australian National Health and Medical Research Council, the Queensland Cancer Fund, the Deutsche Forschungsgemeinschaft (grant number 488/ 1 ± 1) and the A-T Children's Foundation for support. Thanks to Aine Farrell for technical support and Ann Knight and Tracey Laing for typing the manuscript.

References Aaronson SA. (1991). Science, 254, 1146 ± 1153. Bandyopadhyay D, Mandal M, Adam L and Mendelsohn J. (1998). J. Biol. Chem., 273, 1568 ± 1573. Balaban N, Moni J, Shannon M, Dang L, Murphy E and Goldkorn T. (1996). Biochem. Biophys Acta., 1314, 147 ± 156. Banin S, Moyal L, Shieh SY, Taya Y, Anderson CW, Chessa L, Smorodinsky NI, Prives C, Reiss Y, Shiloh Y and Ziv Y. (1998). Science, 281, 1647 ± 1677. Barlow C, Liyanage M, Coens PB, Tarsounas M, Nagashima K, Brown K, Rottinghaus S, Jackson SP, Tagle D, Ried T and Wynshaw-Boris A. (1998). Development, 125, 4007 ± 4017. Barlow C, Dennery PA, Shigenaga MK, Smith MA, Morrow JD, Jackson Roberts II L, Wynshaw-Boris A and Levine RL. (1999). Proc. Natl Acad. Sci. USA, 96, 9915 ± 9919. Barlow C, Ribaut-Barassin C, Zwingman TA, Pope AJ, Brown KD, Owens JW, Larsen D, Harrington EA, Haeberie AM, Mariani J, Eckhaus M, Herrup K, Bailly Y and Wynshaw-Boris A. (2000). Proc. Nat. Acad. Sci. USA, 97, 871 ± 876. Baskaran R, Chiang GG, Mysliwiec T, Kruh GD and Wang JY. (1997). J. Biol. Chem., 272, 18905 ± 18909. Beamish H and Lavin MF. (1994). Int. J. Radiat. Biol., 65, 175 ± 184. Borellini F and Glazer RI. (1993). J. Biol. Chem., 268, 7923 ± 7928. Brown KD, Ziv Y, Sadanandan SN, Chessa L, Collins FS, Shiloh Y and Tagle DA. (1997). Proc. Natl. Acad. Sci. USA, 94, 1840 ± 1845. Byrd PJ, McConville DM, Cooper P, Parkhill J, Stankovic T, McGuire GM, Thick JA and Taylor AMR. (1996). Hum. Mol. Genet., 5, 145 ± 149. Canman CE, Lim DS, Cimprich KA, Taka Y, Tamai K, Sakaguchi K, Appella E, Kastan MB and Siliciano JD. (1998). Science, 281, 1677 ± 1679. Canman CE, Wol AC, Chen CY, Fornace Jr AJ and Kastan MB. (1994). Cancer Res., 54, 5054 ± 5058. Carpenter G. (1987). Annu. Rev. Biochem., 56, 881 ± 914. Chaturvedi P, Eng WK, Zhu Y, Mattern MR, Mishra R, Hurle MR, Zhang X, Annan RS, Lu O, Faucette LF, Scott GF, Li X, Carr SA, Johnson RK, Winkler JD and Zhou BB. (1999). Oncogene, 18, 4047 ± 4054. Chen LI, Nishinaka T, Kwan K, Kitabayashi I, Yokoyama K, Fu YH, Grunwald S and Chiu R. (1994). Mol. Cell Biol., 14, 4380 ± 4389.

Clarke RA, Kairouz R, Watters D, Lavin MF, Kearsley JH and Lee CS. (1998). J. Clin. Pathol. Mol. Pathol., 51, 224 ± 226. Cornforth MW and Bedford JS. (1985). Science, 227, 1589 ± 1591. Cortez D, Wang Y, Qin J and Elledge SJ. (1999). Science, 286, 1162 ± 1166. Dittmann KH, Gueven N, Mayer C and Rodemann HP. (1998). Int. J. Radiat. Biol., 74, 225 ± 230. Dulic V, Kaufmann WK, Wilson SJ, Tisty TD, Lees E, Wade Harper J, Elledge SJ and Reed SI. (1994). Cell, 76, 1013 ± 1023. Elmore E and Swift M. (1976). J. Cell Physiol., 89, 429 ± 431. Foray N, Priestley A, Alsbeih G, Badie D, Capulas EP, Arlett CF and Malaise EP. (1997). Int. J. Radiat. Biol., 72, 271 ± 283. Fukao T, Kaneko H, Birrell G, Gatei M, Tashata H, Kasahara K, Watters D, Khanna KK, Kondo N and Lavin MF. (1999). Blood, 94, 1998 ± 2006. Gatei M, Young D, Cerosaletti KM, Desai-Mehta A, Spring K, Kozlov S, Lavin MF, Gatti RA, Concannon P and Khanna K. (2000). Nature Genet., 25, 115 ± 119. Gilad S, Khosravi R, Shkedy D, Uziel T, Ziv Y, Savitsky K, Rotman G, Smith S, Chessa L, Jorgensen TJ, Harnik R, Frydman M, Sanal O, Portnoi S, Goldwicz Z, Jaspers NGJ, Gatti RA, Lenoir G, Lavin MF, Tatsuni K, Wegner RD, Shiloh Y and Bar-Shira A. (1996). Hum. Mol. Genet., 5, 433 ± 439. Gatei M, Scott SP, Filippovitch I, Soronika N, Lavin MF, Weber B and Khanna KK. (2000). Cancer Res., 60, 3299 ± 3304. Goto SP, Amirmokri E and Liebner EJ. (1967). Am. J. Dis. Child., 114, 617 ± 625. Gueven N, Keating KE, Chen P, Fukao F, Khanna KK, Watters D, Rodemann PH and Lavin MF. (2001). J. Biol. Chem. 276, 8884 ± 8891. Imai T, Yamauchi M, Seki N, Sugawara T, Saito T, Matsuda T, Ito H, Nagase T, Nomura N and Hori TA. (1996). Genome Res., 6, 439 ± 447. Kastan MB, Zhan Q, el-Deiry WS, Carrier F, Jacks T, Walsh WV, Plunkett BS, Vogelstein B and Fornace Jr AJ. (1992). Cell, 71, 587 ± 597. Khanna KK, Beamish H, Yan J, Hobson K, Williams R, Dunn I and Lavin MF. (1995). Oncogene, 11, 609 ± 618.

Oncogene


4290

Oncogene

Khanna KK, Yan J, Watters D, Hobson K, Beamish H, Spring K, Shiloh Y, Gatti RA and Lavin MF. (1997). J. Biol. Chem., 271, 29335 ± 29341. Khanna KK, Keating KE, Kozlov S, Scott S, Gatei M, Hobson K, Taya Y, Gabrielli B, Chan D, Lees-Miller SP and Lavin MF. (1998). Nature Genet., 20, 398 ± 400. Khosravi R, Maya R, Gottlieb T, Oren M, Shiloh Y and Shkedy D. (1999). Proc. Natl. Acad. Sci. USA, 96, 14973 ± 14977. Knebel A, Rahmsdorf HJ, Ullrich A and Herrlich P. (1996). EMBO J., 15, 5314 ± 5325. Kondo N, Inoue R, Nishimura S, Kasahara K, Kameyama Y, Miwa Y, Lorenzo PR and Orii T. (1993). Scand. J. Immunol., 38, 45 ± 48. Kwok TT and Sutherland RM. (1991). Int. J. Cancer, 19, 73 ± 76. Kwok TT and Sutherland RM. (1992). Int. J. Oncol. Biol. Phys., 22, 525 ± 527. Lavin M and Shiloh Y. (1997). Annu. Rev. Immunol., 15, 177 ± 202. Lim DS, Kirsch DG, Canman CE, Ahn JH, Ziv Y, Newman LS, Darnell RB, Shiloh Y and Kastan MB. (1998). Proc. Natl. Acad. Sci. USA, 95, 10146 ± 10151. Lim DS, Kim ST, Xu B, Maser RS, Lin J, Petrini JH and Kastan MB. (2000). Nature, 404, 613 ± 617. Lynn W and Wang P. (1997). Neuroimmunomod., 4, 277 ± 284. Matsuoka S, Huang M and Elledge SJ. (1998). Science, 282, 1893 ± 1897. Mortensen ER, Marks PA, Shiotani A and Merchant JL. (1997). J. Biol. Chem., 272, 16540 ± 16547. Nagaswa H and Little JB. (1983). Mutat. Res., 109, 297 ± 308. Nelson J, Mcgivern M, Walker B, Bailie JR and Murphy RF. (1989). Eur. J. Cancer Clin. Onco., 12, 1851 ± 1855. O'Connor RD and Scott-Linthicum D. (1980). Clin. Immun. Immunopathol., 15, 66 ± 75. Oka A and Takashima S. (1998). Neuroscience Lett., 252, 195 ± 198. Peretz S, Jensen R, Baserga R and Glazer PM. (2001). Proc. Natl. Acad. Sci. USA, 87, 1676 ± 1681. Rhodes N, DSouza T, Foster CD, Ziv Y, Kirsch DG, Shiloh Y, Kastan MB, Reinhart PH and Gilmer TM. (1998). Genes Dev., 12, 3686 ± 3692. Rotman G and Shiloh Y. (1998). Hum. Mol. Genet., 7, 1555 ± 1563. Saito T, Matsuda Y, Ishii H, Watanabe F, Mori M, Hayashi A, Araki R, Fujimori A, Fukumura R, Morimyo M, Tatsumi K, Hori T-A and Abe M. (1998). Mammal. Genet., 9, 769 ± 772. Sakaguchi K, Herrera JE, Saito S, Miki T, Bustin M, Vassilev A, Anderson CW and Appella E. (1998). Genes Dev., 15, 2831 ± 2841.

Savitsky K, Bar-Shira A, Gilad S, Rotman G, Ziv Y, Vanagaite L, Tagle OA, Smith S, Uziel T, Sfez S, Ashkenazi M, Pecker I, Harnik R, Patanjali SR, Simmons A, Frydman M, Sartiel A, Gatti RA, Chessa L, Sanal O, Lavin MF, Jaspers NGL, Malcolm A, Taylor R, Arlett CF, Miki T, Weissman SM, Lovett M, Collins FS and Shiloh Y. (1995). Science, 268, 1749 ± 1753. Sedgwick RP and Boder E. (1991). Ataxia-telangiectasia (208900; 208910; 208920). In: Hereditary Neuropathies and Spinocerebellar Atrophies (JBM Vianney De Jong ed). Elsevier Science Publishers, Amsterdam, pp. 347 ± 423. Shieh SY, Ikeda M, Taya Y and Prives C. (1997). Cell, 91, 325 ± 334. Shiloh Y, Tabor E and Becker Y. (1982). Carcinogenesis, 3, 815 ± 820. Taylor AM, Harnden DG, Arlett CF, Harcourt SA, Lehmann AR, Stevens S and Bridges BA. (1975). Nature, 258, 427 ± 429. Tibbetts RS, Brumbaugh KM, Williams JM, Sarkaria JN, Cliby WA, Shieh S-Y, Taya Y, Prives C and Abraham RT. (1999). Genes Dev., 13, 152 ± 157. Watters D, Khanna KK, Beamish H, Birrell G, Spring K, Kedar P, Gatei M, Stenzel D, Hobson K, Kozlov S, Farrell A, Ramsay J, Gatti R and Lavin MF. (1997). Oncogene, 14, 1911 ± 1921. Watters D, Kedar P, Spring K, Bjorkman J, Chen P, Gatei M, Birrell G, Garrone B, Srinivasa P, Crane DI and Lavin MF. (1999). J. Biol. Chem., 274, 34277 ± 34282. Woltman R, Yahalom J, Maxy R, Pinto J and Fuks Z. (1994). Int. J. Radiat. Oncol. Biol. Phys., 30, 91 ± 98. Wu X, Ranganathan V, Weisman DS, Heine WF, Ciccone DN, O'Neill TB, Crick KE, Pierce KA, Lane WS, Rathbun G, Livingston DM and Weaver DT. (2000). Nature., 405, 477 ± 482. Xu Y and Baltimore D. (1996). Genes Dev., 10, 2401 ± 2410. Yamabe Y, Shimamoto A, Goto M, Yokota J, Sugawara M and Furuichi Y. (1998). Mol. Cell Biol., 18, 6191 ± 6200. Yang DO and Kastan MB. (2000). Nat. Cell. Biol., 2, 893 ± 898. Zhao S, Weng YC, Yuan SS, Lin YT, Hsu HC, Lin SC, Gerbino E, Song MH, Zdzienicka MZ, Gatti RA, Shay JW, Ziv Y, Shiloh Y and Lee EY. (2000). Nature, 405, 473 ± 477. Zhang N, Chen P, Khanna KK, Scott S, Gatei M, Kozlov S, Watters D, Spring K, Yen T and Lavin MF. (1997). Proc. Natl. Acad. Sci. USA, 94, 8021 ± 8026. Zhang N, Chen P, Gatei M, Scott S, Khanna KK and Lavin MF. (1998). Oncogene, 17, 811 ± 818.