ATM-mediated Transcriptional Elevation of Prion in Response to Copper-induced Oxidative Stress*

Increasing evidence suggests that the cellular prion protein (PrPC) plays a protective role in response to oxidative stress, but the molecular mechanism is unclear. Here, we demonstrate that murine neuro-2a and human HeLa cells rapidly respond to an increase of intracellular copper concentration by up-regulating ataxia-telangiectasia mutated (ATM)-mediated transcription of PrPC. Copper stimulation activates ATM by phosphorylation at Ser-1981, which leads to phosphorylation of p53 at Ser-15 and the initiation of the mitogen-activated protein kinase kinase/extracellular-related kinases/extracellular-related kinases (MEK/ERK)/Sp1 pathway. As results, Sp1 and p53 bind to the PrP promoter, leading to increase PrPC expression. Elevated PrPC correlates with reduction of intracellular copper concentration and suppression of Cu(II)-induced accumulation of reactive oxygen species and cell death. Depletion of PrPC, ATM, p53, and/or Sp1 further demonstrates that ATM is a key regulatory protein to promote activation of p53 and Sp1 leading to PrPC elevation, which is required to reduce Cu(II) toxic effects and may play an important role in modulation of intracellular copper concentration.

Conformational conversion of cellular prion protein (PrP C ) into its scrapie form, PrP Sc , is believed to cause transmissible spongiform encephalopathies, including bovine spongiform encephalopathy (BSE) 2 and human Creutzfeld-Jacob disease (1). Although research in prion diseases over past two decades has made significant progress, the normal function of PrP C remains elusive. PrP C , encoded by Prnp gene, is a glycoprotein that expresses in almost all tissues, including neurons and glial cells in the central nervous system (2,3). PrP C specifically localizes on the cell membrane (4), implicating a potential role in cellular sensing. In genomic knock-out studies, PrP-deficient mice or cattle develop and behave normally, implying that the function of PrP C is not essential under normal physiological conditions (5)(6)(7). However, PrP C -deficient mice are more sensitive to oxidative stress (8). In addition, mouse neural CF10 cells derived from PrP knock-out mice were more sensitive to reactive oxygen species (ROS) than cells from wild type mice induced by manganese, Mn(II) (9). These observations suggest that PrP C may play a protective role in cellular resistance to oxidative stress.
Many lines of evidence suggest that copper plays an important role in prion biology. First, the N-terminal octarepeat domain and its neighboring sequence of PrP C are capable of binding copper ions, Cu(II) (10 -16). Second, binding of copper to PrP C may contribute to formation of the PrP protease resistance, a hallmark of PrP Sc (17)(18)(19). Third, the binding of copper to PrP C is pH-dependent, which favors the idea that PrP C might aid in transporting copper across the plasma membrane (14). Although PrP C is believed to involve the metabolism of copper, the molecular mechanism of the interaction between copper and PrP C is not well understood.
Copper is an essential trace element associated with a variety of physiological functions. However, high concentrations of copper are toxic to the cells, which induce cellular oxidative stress leading to ROS accumulation. In this work we study the functional relationship between the Cu(II)-induced oxidative stress and the protective effect of PrP C . Our goal was to understand the molecular mechanisms underlying the potential response of PrP C to oxidative stress induced by Cu(II). We demonstrate for the first time that endogenous PrP C rapidly reacts to Cu(II). Specifically, the Cu(II)-induced elevation of PrP C is modulated through transcriptional up-regulation mediated via the ataxia-telangiectasia mutated (ATM). The elevated PrP C protects against copper-induced oxidative stresses and cell death and plays an active role in modulation of intracellular copper concentration.

Cell Culture, Small Interfering RNA (siRNA) Transfection, and Cu(II) Treatment
Murine neuro-2a (N2a) and human HeLa cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Invitrogen) at 37°C and 5% CO 2 . To establish a stable PrP-knockdown N2a cell line, N2a cells were transfected with the short hairpin RNA to the mouse PrP (Open Biosystems, Huntsville, AL) and then selected by the addition of 1 g/ml puromycin (Alexis Biochemicals, San Diego, CA) into the growth media. To transiently knock down the gene expressions of interest, we used siRNA pre-designed specifically to against Prnp (Ambion, Austin, TX), ATM, p53, and/or Sp1 along with the non-silencing control siRNA (Qiagen, Valencia, CA). Each siRNA at a concentration of 100 nM was transfected into ϳ5 ϫ 10 5 N2a or HeLa cells using 15 l of Oligofectamine following the manufacturer's instructions (Invitrogen). Cells were incubated with or without 100 M of CuCl 2 or MgCl 2 for different time periods and then harvested for further experiments.

Determination of ROS
N2a and PrP-knockdown N2a cells in 96-well plates were incubated with 50 M 2Ј,7Ј-dichlorodihydrofluorescein diacetate (DCFH-DA) (Invitrogen) for 45 min to allow DCFH-DA to diffuse into cells. After washing with phosphate-buffered saline (PBS), cells were then treated with 100 M CuCl 2 for 0, 10, 30, 60, 120, or 240 min. DCFH-DA is hydrolyzed by intracellular esterase to yield DCFH that is oxidized by H 2 O 2 or low molecular weight peroxides in cells to produce the highly fluorescent compound, 2Ј,7Ј-dichlorofluorescein (DCF). The DCF fluorescence intensity is determined using a Multiable Counter (Model Wallac 1420, PerkinElmer Life Sciences) with an excitation wavelength of 485 nm and an emission wavelength of 530 nm.

Measurement of Cell Viability
The cell viability was measured by a 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT)-based cell growth determination kit (Sigma-Aldrich). N2a or PrP-knockdown N2a cells were incubated with 100 M CuCl 2 and collected at the indicated time points. After removal of medium, cells were aseptically added with MTT solution in an amount equal to 10% of the culture volume and incubated for 4 h. MTT formazan crystals were dissolved by the addition of MTT solvent in an amount equal to the original culture volume and incubated for 1 h. Absorbance was spectrophotometrically measured at a wavelength of 570 nm with subtraction of background absorbance measured at 690 nm.

Trypan Blue Assay
The trypan blue exclusion assay was used for the quantification of dead cells. Monolayers of N2a or the PrP-knockdown N2a cells were incubated with or without 100 M of CuCl 2 for an indicated period of time. Cells were trypsinized and then collected by centrifugation at 1,500 rpm for 10 min. After resuspension with PBS, cells were stained with 0.2% trypan blue (Sigma-Aldrich), and dying cells were identified by trypan bluepositive staining. Average and S.D. of copper-induced cell death were calculated based on three independent experiments. Student's t test was employed to determine significant differences between the control and different test groups.

Prnp Transcription
PrP mRNA Stability Assay-N2a cells were incubated with 100 M CuCl 2 for an indicated period of time and then treated with or without 5 g/ml actinomycin D for 2 h. Cells were harvested, and the total RNA was extracted by RNAqueous-4RCR kit (Ambion). Two micrograms of the total RNA from each sample was used for reverse transcription (RT)-PCR by SuperScript One-Step RT-PCR with Platinum Tag (Invitrogen) to synthesize the cDNA of the mouse PrP, and ␤-actin was used as a control. The PCR products were separated on a 1% 1ϫ Tris acetate EDTA (Tris-acetate EDTA)-agarose gel and analyzed on a PhosphorImager (GE Healthcare). Densitometric analyses were used to normalize PrP cDNA to ␤-actin control cDNA and resulted in relative -fold increase in Prnp expression in N2a cells after treatment with Cu(II).

Real-time RT-PCR
A one-step real-time RT-PCR in which the RT reaction took place in the same tube was performed to determine Prnp gene transcription using LightCycler Probe Design software (Roche Applied Science). The PCR primers were designed as follows: mouse PrP C , forward 5Ј-CCA AGC TTA TGG CGA ACC TTG GCT GCT-3Ј and reverse 5Ј-CCG AAT TCT CCC ACT ATC AGG AAG AT-3; mouse ␤-actin, forward 5Ј-GCC CTA GAC TTC GAG C-3Ј and reverse 5Ј-CGC ATG TCA ACG TCA C-3. N2a cells were treated with 100 M CuCl 2 and collected at the indicated time points. Cells were harvested, and total RNA was extracted using RNAqueous-4RCR kit (Ambion) following the manufacturer's instructions. Equal amounts of total RNA (2 g) from each sample were subjected to real-time RT-PCR using the iQ5 Multicolor Real-Time PCR detection system with the iScript one-step RT-PCR kit with SYBR Green (Bio-Rad) to quantify the possible elevation of the PrP mRNA after exposure of cells to Cu(II). The endogenous reference gene, ␤-actin, was used as a control. Assuming that the amplification efficiencies of the target gene (Prnp) and the control gene (␤-actin) were very similar, ⌬⌬C T (⌬⌬threshold cycle) was calculated using the following formulae (Qiagen): ⌬C T(control) ϭ ⌬C T(the target gene in the start sample) Ϫ ⌬C T(the reference gene in the start sample) ; ⌬C T(test) ϭ ⌬C T(the target gene in the test sample) Ϫ ⌬C T(the reference gene the test sample) ; ⌬⌬C T ϭ ⌬C T(control) Ϫ ⌬C T(test) . Then, normalized target gene expression in the test sample ϭ 2 Ϫ⌬⌬CT .

Western Blot Analyses
Cells were homogenized in a lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100) containing a protease inhibitor mixture (Roche Applied Science) on ice for 30 min. After centrifugation at 15,000 rpm for 10 min, the supernatant was collected for experiments. For nuclear extracts, harvested cells were gently resuspended in lysis buffer A (10 mM Hepes, pH 8, 1.5 mM MgCl 2 , 10 mM KCl, 0.3 M sucrose, 1 mM dithiothreitol, 0.5% Nonidet P-40, containing a protease inhibitor mixture (Roche Applied Science) and then incubated on ice for 15 min. After centrifugation at 800 rpm for 5 min, the cell pellets were re-suspended in lysis buffer B, 10 mM Hepes, pH 8, 25% glycerol, 0.42 M NaCl, 1.5 mM EDTA, containing a protease inhibitor mixture (Roche Applied Science) and then rocked at 4°C for 2 h. After centrifugation at 15,000 rpm for 10 min, the supernatants were stored as the nuclear extracts at Ϫ80°C until use. After determination of protein concentrations by BCA protein assay (Pierce), the equivalent of 30 g of total protein was loaded onto SDS-PAGE gels (Bio-Rad) and analyzed by Western blotting with the appropriate primary antibodies, i.e. 1:500 dilution of anti-PrP monoclonal antibody (mAb) SAF-32 (Cayman Chemical, Ann Arbor, MI), 1:1,000 dilution of antibodies against ATM, phosphorylated ATM-Ser-1981, p53, phosphorylated p53-Ser-15, MEK1, and phosphorylated ERK1/2 (Cell Signaling, Beverly, MA), 1:500 dilution of anti-Sp1 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), and a 1:10,000 dilution of anti-␤-actin antibody (Sigma-Aldrich). The appropriate horseradish peroxidase-conjugated secondary antibodies (Bio-Rad) were used. The blots were visualized by using ECL kit (Amersham Biosciences) and exposed to Fuji Super RX x-ray film (Fuji Photo Film, Valhalla, NY) (20).

Electrophoretic Mobility Shift Assays (EMSA)
The linear DNA fragments containing DNA sequences in the mouse Prnp promoter region (Ϫ1928/ϩ52) (21) and the human PRNP promoter region (Ϫ2560/ϩ125) (22) were used as target DNAs. The nucleotide sequences of the probe sense strands are as follows: the putative Sp1 binding sites (Ϫ65/Ϫ35) 5Ј-ATC ACG CCC CGC CCC TCG CCC AGC CTA GCT CC-3Ј and the putative p53 binding site (Ϫ1832/Ϫ1810) 5Ј-CTT GTC AAG ACT AGT TTG CCT CG-3Ј in the mouse Prnp promoter; the putative Sp1 binding site (Ϫ71/Ϫ47) 5Ј-CTC GGC CGG CCG CCC GCC GGG GGC A-3Ј and the putative p53 binding site (Ϫ757/Ϫ732) 5Ј-TTT CCC CAG GGC ATG CCT GGT TTA C-3Ј in the human PRNP promoter. The competing oligonucleotides include the Sp1 consensus 5Ј-ATT CGA TCG GGG CGG GGC GAG C-3Ј (23) and the p53 consensus 5Ј-AGG CAT GTC TAG GCA TG-3Ј (24). Target DNAs were labeled with biotin by using a biotin 3Ј-end DNA labeling kit (Pierce). Two complement DNA sequences of the biotin-labeled target DNAs or unlabeled (cold) consensuses were annealed to become double strand oligonucleotides. N2a or HeLa cells were mock-transfected or transfected with control siRNA or siRNA to ATM for 48 h and then incubated with or without 100 M CuCl 2 for 30 min. Cells were harvested in PBS by centrifugation. The pelleted cells were lysed in the lysis buffer A containing 10 mM HEPES, pH 8, 1.5 mM MgCl 2 , 10 mM KCl, 0.3 M sucrose, 1 mM dithiothreitol, 0.5% Nonidet P-40, and the protease inhibitor mixture (Roche Applied Science). The nuclei were pelleted, and nuclear proteins were extracted in lysis buffer B containing 10 mM HEPES, pH 8, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 1 mM dithiothreitol, and protease inhibitor mixture (Roche Applied Science). After measurement of the protein concentration by BCA protein assay (Pierce), the nuclear protein extracts were ready for EMSA. LightShift chemiluminescent EMSA kit (Pierce) was used in the DNA-protein binding experiments. The biotin-labeled double strand target DNA was incubated with nuclear extracts from N2a or HeLa cells in solution containing 2.5% glycerol, 5 mM MgCl 2 , 1 mM EDTA, 0.05% Nonidet P-40, and 50 ng/l poly(dI⅐dC) at room temperature for 20 min and then loaded onto 5% polyacrylamide gel in 0.5ϫ Tris-buffered EDTA for electrophoresis. The binding reaction was then transferred to a Nylon membrane (Pierce). After cross-linking the transferred DNA to the membrane by ultraviolet light, the biotinlabeled DNA signals were detected by sequential incubations of the membrane with stabilized streptavidin-horseradish peroxidase conjugate and lumino/enhancer solution (Pierce) and then exposure of the membrane to x-ray film. The binding of the predicted transcription factor to the specific target DNA formed a slower-migrated band than the original size of the starting target DNA. If an up-shifted band is seen in EMSA, it is an indication that the testing transcription factors, Sp1 or p53, may bind to the specific DNA sequence in the Prnp promoter. Specific binding of p53 or Sp1 to the oligonucleotides tested can be further confirmed by the competition experiment using 100fold molar excess of the unlabeled consensus oligonucleotides, which should eliminate the observed protein/DNA binding complex observed in EMSA.

Determination of Intracellular Copper Concentration by Inductively Coupled Plasma Mass Spectrometry
ICP-MS was used to determine the intracellular copper concentration. The ThermoFinnigan Element 2 (Waltham, MA) single collector ICP-MS device is a high resolution double focusing instrument operated in low resolution (m/⌬m 300). Cells were treated with 100 M CuCl 2 and collected at 0, 10, 30, 60, 120, and 240 min after treatment. After washing twice with PBS, the number of cells was counted. Samples containing equal number of cells were centrifuged and dried at 60°C for 3 h. The dried cell material was dissolved in 200 l of 70% nitric acid and then diluted to 2% nitric acid and spiked with gallium. 71 Ga was chosen as an internal standard because its m/z is similar to that of copper, and it causes no major spectroscopic interference. A 10 l aliquot of a 10-ppm gallium standard was added to the each sample. A copper standard was prepared using 10 l aliquots of standard solutions of in 1 ml of 2% nitric acid. All solutions were prepared with Milli-Q 18.2 M water. The nitric acid blank, the copper standard, and each of the samples were introduced to the ICP via a 200 l/min self-aspirating nebulizer in a cyclonic spray chamber (Glass Expansion). Between samples, the nitric acid blank was used to rinse the nebulizer. Results for each sample were calculated using the integrated average background-subtracted peak intensities from 20 consecutive scans. The concentration of copper of each sample was calculated with its protein concentration.

Cu(II) Triggers Rapid Elevation of PrP C -
The initial goal of this study was to evaluate the potential response of PrP C to the Cu(II) stimulation. As shown in Fig. 1A, endogenous PrP C was normally seen in N2a cells as di-, mono-, and non-glycosylated forms by Western analyses (Fig. 1A, a and e, lanes 1). To compare PrP C protein productions before and after incubation with Cu(II), the PrP C protein levels ( Fig. 1A, a) were measured and normalized based on ␤-actin levels ( Fig. 1A, b). Quantitative comparison of PrP C levels over time in Cu(II)-treated N2a cells are summarized in Fig. 1B (solid columns). Compared with the PrP C level in untreated N2a cells (Fig. 1A, a1), PrP C levels in Cu(II)-treated N2a cells increased rapidly to 2.7 Ϯ 0.1-and 10.4 Ϯ 3.2-fold above the base line 30 and 60 min after the treatment (Fig. 1, A, a lanes 3 and 4, and B, filled columns). Gradual declines of PrP C levels were seen followed with prolonged incubation, but the PrP C level remained significantly higher than the base line with a 4.7 Ϯ 0.1-fold increase at 120 min and a 3.6 Ϯ 0.17-fold increase at 240 min of incubation ( Fig.  1A, a, lanes 5 and 6, and B, filled columns). Because cations such as Cu(II) induce ROS in cells, but Mg(II) does not (25), we were interested in seeing whether Mg(II) can also trigger similar levels of PrP C elevation. The PrP C protein levels were essentially unchanged before and after incubation with 100 M Mg(II) over the same time period as Cu(II) treatment ( Fig. 1, A, e, lanes 1-6, and B, open columns). These data suggest that PrP C specifically responds to Cu(II), but it does not respond to Mg(II) (p Ͼ 0.05). To verify the observed protein elevations were indeed PrP C , the Prnp gene transcription was depleted by transfection with short hairpin RNA. Immunoblotting analysis showed that the Prnp gene depletion was successful, as the PrP C expression was essentially undetectable under the conditions as described in Fig. 1A, a.
PrP Elevation Is Triggered through Increased Gene Transcription-To further investigate whether Cu(II)-induced acute elevation of PrP C occurs at the transcriptional level, we carried out RT-PCR and densitometric analyses to compare the Prnp mRNA levels with the control ␤-actin mRNAs in N2a cells after exposure them to Cu(II). With these analyses, we were able to determine the relative -fold increase of Prnp expression to the reference gene ␤-actin mRNA. Results of these analyses showed that the control ␤-actin mRNA expression levels as expected were nearly constant. In contrast, the PrP mRNA level in N2a cells increased from 1.53 Ϯ 0.3-fold at 10 min to 4.46 Ϯ 0.7-fold at 30 min after incubation with Cu(II). A gradual decline of the PrP mRNA levels were observed after 30 min ( Fig.  2A, a, and B, filled columns).
To distinguish whether the gradual decease of the PrP mRNA levels are due to mRNA stability or reduced ongoing transcription triggered by Cu(II), the same cells were treated at the indicated time points with actinomycin D (Act D), which stops gene transcription (26,27), and all Act D-treated cells were collected at end of the experiments, i.e. 240 min. The   (II) for 0, 10, 30, 60, 120, or 240 min. Cells were harvested, and total RNA was extracted. 2 g of total RNA of each sample was subjected to real-time RT-PCR to determine the threshold cycle (C T ) (see "Results") for calculating the starting mRNA level of PrP C and ␤-actin. The Prnp mRNA was normalized against ␤-actin mRNA (see "Experimental Procedures") and expressed as -fold change compared with the mRNA level in untreated control cells. Bars represent the mean Ϯ S.D. of three independent experiments. Student's t test was employed to determine significant differences between the control and experimental test groups (**, p Ͻ 0.01).
mRNA levels with or without Act D treatment was then compared side-by-side (Fig. 2A), and their relative levels to ␤-actin were summarized in Fig. 2B. The addition of Act D did not cause significant changes in the levels of the PrP mRNA in N2a cells after incubation with Cu(II) at different time points (p Ͼ 0.05) ( Fig. 2A, c, and B, open columns), suggesting that the elevated PrP mRNA levels observed over the experimental period was caused by an initial Cu(II)-mediated increase of Prnp gene transcription and not maintained by constant gene transcription. Thus, the PrP mRNA is relatively stable over a period of 240 min. To have a more accurate quantification of the PrP mRNA, we further performed real-time RT-PCR to quantify the amount of the Prnp mRNA in reference to the ␤-actin gene expression. The ␤-actin mRNA expression levels as measured by threshold cycle (C T ) were nearly constant over the experimental period, i.e. within the normal cycle variation of 1.25 (p Ͼ 0.05). In contrast, a significant increase in the Prnp mRNA was observed in N2a cells under the same conditions. As shown in Fig. 2C, Prnp mRNA levels in N2a cells increased from 2.5 Ϯ 0.7-fold at 10 min to maximum level of 8.8 Ϯ 1.9-fold at 30 min after the initial addition of Cu(II). A gradual decline was also observed thereafter, suggesting that PrP C expression is a dynamic process resulting from rapid cellular responses to Cu(II). However, at all time points the level of the Prnp mRNA in cells incubated with Cu(II) was significantly higher than that in the untreated cell control (p Ͻ 0.01), indicating that the rapid PrP C elevation induced by Cu(II) indeed occurs at the transcriptional level. Thus, our data demonstrate that production of PrP C can be rapidly elevated by Cu(II) through increased gene transcription, and this transcriptional modulation lasted up to 240 min without significant decay.
Elevated PrP C Protects against Cu(II)-induced ROS Accumulation and Cell Death-High levels of Cu(II) causes cellular oxidative stress leading to ROS accumulation (28); therefore, our finding that Cu(II) rapidly increases PrP C expression ( Fig.  1) may indicate that the elevated PrP C could be the specific cellular response to antagonize copper cytotoxicity. To test the potential effect of elevated PrP C on Cu(II)-induced oxidative stress or cell death, we monitored ROS accumulation, cellular viability, and cell death in cells that were treated with Cu(II) the same way as described in Fig. 1A. DCFH-DA was used as a specific probe to detect ROS in cells as described previously (29,30). DCFH-DA is a stable compound that readily diffuses into cells where it is hydrolyzed by intracellular esterase, resulting in loss of its diacetate group, thus yielding DCFH. Cells were washed with PBS, incubated with 100 M Cu(II), and then collected at different time points. Cu(II)-induced H 2 O 2 and low molecular weight peroxides can oxidize DCFH to produce the highly fluorescent compound, DCF (31,32), which is indicative of ROS production. Briefly, N2a and PrP-knockdown N2a cell cultures were first incubated with 50 M DCFH-DA for 45 min. The intensity of the DCF fluorescence was determined. After exposure of N2a cells to Cu(II), the DCF fluorescence rapidly increased (2.1 Ϯ 0.3-fold) within 10 min (Fig. 3A, open circle) in comparison with the untreated N2a cells (Fig. 3A, filled triangles). A similar level of DCF increase was also observed in the PrP-knockdown N2a cells (Fig. 3A, open circle). However, instead of a gradual decrease of the DCF fluorescence as shown in the wild type cells (Fig. 3A), a small but steady increase of DCF was detected in the PrP-knocking down cells over a period of 240 min (Fig. 3A). Overall, Cu(II) induced consistently higher levels of ROS in the PrP-knockdown N2a cells than in the wild type N2a cells (p Ͻ 0.01). Because the only difference between these two cell lines is the presence or absence of PrP C , this observation suggests that Cu(II)-induced PrP C elevation (Figs. 1 and 2) may prevent ROS accumulation. It is notable there is a clear discrepancy between the DCF that peaks at 10 min upon Cu(II) treatment (Fig. 3A, open circle) and the effect of Cu(II) on PrP C reactivity that peaks 30 min after Cu(II) exposure (Fig. 1A,  a) in N2a cells. This time lag might be necessary for Cu(II)induced ROS to trigger PrP C expression. Consistently, the DCF level declines at 30 min after Cu(II) incubation (Fig. 3A, open circle), whereas the PrP C level continues to increase (Fig. 1A, a), suggesting a possible cause-effect relationship between DCF and the PrP C protein levels.
Accumulation of ROS is toxic to cells (29). To test the potential of the elevated PrP C to prevent cytotoxicity induced by Cu(II), we measured the effects of Cu(II) on cellular viability by a MTT colorimetric assay. After cells were incubated with Cu(II), cellular viability was calculated by MTT-based cell growth curves. In the wild type N2a cells, Cu(II) reduced cellular viability to ϳ80% over a period of 240 min (Fig. 3B, open circle) in comparison with untreated cells (Fig. 3B, filled triangles). Additional and significant (p Ͻ 0.05) reduction of cellular viability (60 -68%) was observed when PrP C was depleted (Fig.  3B, open circle). Depletion of PrP C by itself had no effect on cell survival (Fig. 3B, open triangles). Trypan blue staining was also used to further determine whether PrP C prevents cell death induced by Cu(II). After incubation with Cu(II), N2a cells had background levels of cell death (from 14.6 Ϯ 1.7% at 30 min to 18.7 Ϯ 1.7% at 240 min) (Fig. 3C, open circle). A significant (p Ͻ 0.05) increase in cell death was observed in PrP-depletion N2a cells (from 25.6 Ϯ 1.8% at 30 min to 33.0 Ϯ 2.8% at 240 min) after incubation with Cu(II) (Fig. 3C, open circle). Together, these data suggest that elevated PrP C induced by Cu(II) protects against ROS accumulation and cell death in N2a cells.

Cu(II)-induced PrP C Correlates ATM-mediated p53 and MEK/ERK/Sp1
Activations-Because Cu(II)-induced PrP C elevation starts at the transcriptional level (Fig. 2), it is possible that responsive binding of certain transcription factors to the PrP promoter may initiate the Prnp gene transcription for more PrP C production. Although no direct binding of transcriptional factors to the Prnp promoter has been described previously, prior sequencing analyses predict a putative p53 binding site and three putative Sp1 binding sites on the human PRNP promoter (22). Similarly, by analyzing the DNA sequences of the mouse Prnp promoter, we also identified a putative p53 binding site that is highly homologous to the human p53 binding sequence between nucleotides Ϫ1823/Ϫ1814. Three putative Sp1 binding sites in the mouse Prnp promoter have been described previously (21). Binding of Sp1 to a gene promoter can be activated by MEK/ERK (33), and both MEK/ERK and p53 pathways can be activated by ATM (34,35). To first test whether ATM is involved in the Cu(II)-induced PrP C elevation, we measured the potential activation of ATM by phosphorylation at Ser-1981. As we showed in Fig. 1A, treatment of N2a cells with Cu(II) caused the rapid increase of PrP C (Fig. 4A, a,  lanes 1-3). This rapid increase of PrP C was accompanied with the strong phosphorylation of ATM at Ser-1981, indicating activation of ATM (Fig. 4A, c, lanes 1-3). Similarly, the ATM substrates p53 and MEK were also activated as shown by Ser-15 phosphorylation of p53 (Fig. 4A, e, lanes 1-3) and increased MEK protein level (Fig. 4A, f, lanes 1-3), which in turn promotes ERK1/2 activation by phosphorylations (Fig. 4A, g lanes  1-3). Consequently, protein levels of Sp1 were also increased (Fig. 4A, h, lanes 1-3). These results indicate that Cu(II) indeed induces activation of ATM, which may lead to activations of its downstream stress-responsive molecules, including p53 and MEK/ERK1/2-mediated Sp1 activation. To further test whether ATM is the key regulatory protein for p53 and Sp1 activations leading to PrP C elevation, we measured the levels of the same proteins in cells with depleted ATM by siRNA. Introduction of a control siRNA did not affect protein elevations induced by Cu(II) (Fig. 4A, lanes 4 -6). In contrast, depletion of ATM completely eliminated all of the responsive protein elevations (Fig. 4A, a-h, lanes 7-9) with the exception of the ␤-actin protein controls (Fig. 4A, lanes 7-9). To detail pathways of the Cu(II)-induced PrP C elevation, we further knocked down gene expression of p53 and/or Sp1 using specific siRNA. Depletion of p53 (Fig. 4A, d, lanes 10 -12, and e, lanes 10 -12) did not affect activations of ATM (Fig. 4A, c, lanes 10 -12) and MEK/ ERK/Sp1 pathways (Fig. 4A, f, lanes 10 -12, g, lanes 10 -12, and  h, lanes 10 -12) but resulted in a week elevation of PrP C (Fig. 4A,  a, lanes 10 -12). Similarly, depletion of Sp1 (Fig. 4A, h, lanes  13-15) did not affect activation of ATM (Fig. 4A, c, lanes  13-15) and p53 (Fig. 4A, e, lanes 13-15) but also resulted in attenuated elevation of PrP C . Significantly, double depletions of p53 (Fig. 4A, d, lanes 16 -18) and Sp1 (Fig. 4A, h, lanes 16 -18) completely abolished elevation of PrP C (Fig. 4A, a, lanes 16 -18) even though they did not affect activation of ATM (Fig. 4A, c, lanes 16 -18), MEK1 (Fig. 4A, f, lanes 16 -18), and ERK1/2 ( Fig.  4A, g, lanes 16 -18). Collectively, these data suggested that Cu(II)-induced PrP elevation in N2a cells is most likely through ATM-mediated activation of two independent p53 and Sp1 pathways.
To test whether these Cu(II)-induced effects are neuronal specific or mouse cell-specific, we carried out the same set of experiments using a non-neuronal human HeLa cell line. Essentially the same effects of Cu(II) on HeLa cells were also observed (Fig. 4B). Thus, these results clearly indicate that in both murine neuronal N2a and human non-neuronal HeLa cells, ATM is a key molecule that regulates the responsive expression of PrP C to Cu(II) through ATM-dependent p53 and MEK/ERK/Sp1 pathways.

Bindings of Sp1 and p53 to Mouse (Prnp) and Human (PRNP) Promoters Are Enhanced by Cu(II) in an ATM-dependent
Manner-Data described in the above sections suggested that the bindings of transcription factors, i.e. Sp1 and p53, might be involved in up-regulation of the PrP promoter. To confirm these possible gene regulating mechanisms, we used an EMSA to examine the binding of Sp1 or p53 to the mouse Prnp promoter in N2a cells or to the human PRNP promoter in HeLa cells. Incubation of the biotin-labeled target mouse DNA fragment containing the putative Sp1 binding sites (Ϫ65/Ϫ35) (21) with nuclear extracts from N2a cells generated a specific DNA/ protein binding band on the gel (Fig. 5A, a, lane 2). The addition of a 100-fold excess of non-biotin-labeled "cold" Sp1 consensus oligonucleotides abolished this band, suggesting a specific binding between the nuclear protein and the putative Prnp Sp1 binding sequence (Fig. 5A, a, lane 3). Similarly, the biotin-labeled putative p53 binding sequence (Ϫ1835/Ϫ1810) also reacted with the N2a nuclear extract and produced a band shift (Fig. 5A, b, lane 2). This binding was completely blocked by the addition of cold p53 consensus sequence (Fig. 5A, b, lane 3). These results indicated that Sp1 and p53 do bind to the mouse Prnp promoter under the normal physiological conditions. Similar bindings of p53 and Sp1 to human PRNP promoter were also observed by using the same assay in HeLa cells (Fig. 5B). Together, our data show that both Sp1 and p53 bind to the mouse and human PrP promoters under normal physiological conditions.
Next, we tested whether Cu(II) or ATM could affect the bindings of p53 or Sp1 to the PrP promoters. After mock transfection or transfection with control siRNA or siRNA to ATM for 48 h, N2a cells were exposed to the same concentration of Cu(II), i.e. 100 M, for 30 min. Nuclear proteins were then extracted, and the same amount of nuclear extracts and biotinlabeled oligonucleotides as described above were used for EMSA. As shown in Fig. 5A, a, lanes 4 and 5, and b, lanes 4 and 5, exposure of N2a cells to Cu(II) significantly increased the binding of Sp1 or p53 to the Prnp promoter as DNA/protein binding bands appeared to be much stronger than the ones without CuI(II) treatment. Introduction of the control siRNA did not interfere with the binding of Sp1 or p53 to the Prnp promoter ( Fig. 5A, a, lanes 6 and 7, b, lanes 6 and 7). Significantly, however, depletion of ATM by siRNA completely demolished formation of the DNA/protein binding bands (Fig.  5A, a, lanes 8 and 9, b, lanes 8 and 9), confirming that ATM is the key molecule in the up-regulation of the Prnp promoter after the Cu(II) stimulation. Consistently, similar effects of Cu(II) or ATM were also observed in HeLa cells (Fig. 5B, lanes  4 -9). Noticeably, depletion of ATM did not completely abolish the bindings of p53 or Sp1 to the PRNP promoter, suggesting other cellular factors may also involve in the p53 or Sp1 binding in HeLa (Fig. 5B). Together these data support the idea that Cu(II) triggers ATM-dependent activations of p53 and Sp1, which bind to the PrP promoter leading to the increase of the PrP gene transcription.
PrP C Delays Cu(II)-induced Activations of ATM and p53-The data described above have shown that Cu(II) induces activation of ATM and its downstream p53 and MEK/ERK1/2 signaling pathways, leading to increased production of PrP C . Interestingly, PrP C appears to prohibit ATM and p53 activations. These effects can be shown by depletion of PrP C . We incubated N2a and PrP-knockdown N2a cells with 100 M Cu(II) for a period of 240 min and observed the timing and intensity of the ATM and p53 activations in response to the Cu(II) stimulation. The same responsive patterns of PrP C , ATM, and p53 were observed in the wild type N2a cells (Fig. 6A,  a-c, and B, filled columns). Interestingly, relatively higher protein levels of ATM and p53 were consistently detected in the PrP-depleted cells over the entire testing period than that in the wild type cells (Fig. 6A, e-g, and B, open columns). In addition, both ATM and p53 were elevated at least 20 min earlier in the PrP-depleted cells than that in the wild type cells (Fig. 6A, f and  g versus b and c, and B). Collectively, these results suggest that Cu(II) triggers activation of ATM resulting in the increase of the PrP C expression. The elevated PrP C may in turn counteract ATM-mediated p53 phosphorylation (Fig. 6A, b and c, lanes 4), which constitutes a negative and regulatory feedback mechanism to balance the production of PrP C and ATM activation.
PrP C Modulates Intracellular Copper Concentration-PrP C is a membrane protein that binds copper (10,12). Because there is an active and antagonistic interaction between PrP C and Cu(II), we hypothesized that the elevated PrP C may regulate intracellular copper concentration. To test this hypothesis, we employed ICP-MS to measure copper concentration in cells before and after cellular exposure to Cu(II). Three types of samples were prepared; they are N2a cells without treatment, N2a cells transfected with the control siRNA, and N2a cells transfected with the specific siRNA to the mouse PrP. After incubation of cells with 100 M Cu(II), the same elevated response of PrP C to the Cu(II) was observed as shown in Fig. 1A by Western blotting analyses (Fig. 7A). Measurement of the copper concentration inside cells by ICP-MS indicated that normal intracellular copper concentrations were in a range between 45.7 Ϯ 1.1 and 47.8 Ϯ 1.1 ng/mg of protein as shown in the N2a cells without treatment or transfected with control siRNA (Fig. 7B, filled triangles or filled squares). A slightly lower copper level (38.3 Ϯ 1.0 ng/mg of protein) was detected in PrP-knockdown N2a cells (Fig. 7B, open circle) (p Ͻ 0.05). As soon as the addition of Cu(II), the intracellular copper concentration was rapidly increased to the level of 194.0 Ϯ 4.2 ng/mg of protein (4.1fold), in 10 min (Fig. 7B, filled triangles). This initial burst of copper concentration was, however, followed by a rapid decline of the copper concentration back close to the baseline at 66.7 Ϯ 1.4 and 55.5 Ϯ 1.7 ng/mg of protein 30 and 60 min after the addition of Cu(II) (Fig. 7B, filled triangles). Similar patterns of intracellular copper contents were also observed in N2a cells transfected with the control siRNA (Fig. 7B, filled squares). A different pattern of the intracellular copper concentration was seen in the PrP-depleted N2a cells. After exposure to Cu(II), a slow but stable increase of intracellular copper concentration was observed over a period of 240 min (Fig. 7B, open circle). Most importantly, the burst of high intracellular concentration of copper was only observed in the wild type PrP C cells, implicating possible and active regulation of PrP C on intracellular concentration of copper. Collectively, these data suggest that PrP C may play an active role in regulating intracellular copper concentration.

DISCUSSION
In this report we have shown that production of the cellular prion protein is responsive to oxidative stress induced by cop- per. This is to the best of our knowledge the first description of a rapid and interactive response of PrP C to cellular stress. In addition, we demonstrate that activation of ATM is the key molecular event that drives the elevated production of PrP C . Specifically, Cu(II)-induced ATM activation promotes MEK/ ERK/Sp1 and p53 signaling pathways, which promotes bindings of Sp1 and p53 to Prnp or PRNP promoter, resulting in up-regulated transcription of Prnp or PRNP. The elevated PrP C appears to function as a protective cellular mechanism to prevent Cu(II)-induced ROS accumulation and cell death, possibly involving in re-adjustment of intracellular copper concentration. A negative feedback mechanism also appears to be in place as PrP C down-regulates ATM-mediated p53 and ATM phosphorylations.
It should be mentioned that several aspects of the presented paradigm have been described previously. For example, inducible expression of Prnp by cooper has been reported (36). A role for PrP C in the cellular resistance to oxidative stress is known, and MEK1 has been shown to be a positive activator of the PRNP promoter that inhibits the AKT pathway (37,38). However, the contribution of the present study is that we have presented direct evidence for a comprehensive paradigm to show an active interaction between the copper stimulation and acute cellular protective response of PrP C . In particular, the effect of copper is shown for the first time meditated through ROS that results in activation of ATM, which in turn activates the MEK/ERK/Sp1 and p53 signaling cascades and leads to the activation of transcriptional activation of PrnP. Another cautionary note is that although we showed strong correlation between the copper treatment and ROS (Fig. 3A), additional studies are needed to confirm the Cu(II)-mediated ROS is the driving force for the observed PrP C elevation.
Even though the conformational conversion of PrP C to PrP SC is believed to be responsible for many of the prion diseases including Creutzfeld-Jacob disease and mad cow diseases, the normal physiologic function of PrP C remains elusive. One of the most suspected roles of PrP C is that it may play a protection role against cytotoxicity induced by metals such as copper (39,40). However, the molecular mechanism underlying this protection effect was unknown. Here we provide direct evidence to show that PrP C respond specifically to the copper stimulation (Fig. 1A), and the elevated PrP C reduces ROS accumulation and cell death (Fig. 3).
Our subsequent biochemical analyses show that ATM is a key regulatory factor in the response of PrP C to Cu(II) stimulation. This finding is significant because ATM, a nuclear serine-threonine kinase, is well known as a central molecular mediator for cellular responses to DNA damages (35,41). A possible role of ATM in regulation of cellular oxidative stress has not been very well understood. We demonstrate specifically that cellular oxidative stress induced by Cu(II) rapidly activates ATM as shown by phosphorylation at its Ser-1981 residue (Figs. 4 and 6). ATM activation in turn triggers activations of the MEK/ERK/Sp1 and p53 signaling cascades (Fig. 4). In the cellular signaling transduction network, MEK/ERK/Sp1 and p53 are known downstream effector molecules of ATM (33,35). Our molecular analyses using the EMSA further reveals that Cu(II)-induced PrP C elevation requires up-regulation of the PrP promoter by binding of the two well known transcriptional activators, i.e. Sp1 and p53, to the putative binding sites on the PrP promoter. It should be mentioned that binding of p53 or Sp1 to the PrP promoters have not be described previously even though putative binding sites were predicted by early reports (21,22) or revealed by our PrP promoter sequence analysis (Fig. 5). However, prior studies have suggested a possible regulatory relationship between p53 and PrP C (42,43). Therefore, Cu(II) triggers an ATM-dependent cellular mechanism that  d, and f). B, quantification of the intracellular copper contents. N2a cells were mock-transfected (filled triangles) or transfected with control siRNA (filled squares) or siRNA to the mouse PrP C (open circle) for 48 h and then incubated with 100 M Cu(II) for indicated period of time. After washing twice with PBS, the number of cells was counted. Samples containing equal number of cells were centrifuged and dried at 60°C for 3 h. The dried cell materials were dissolved in 200 l of 70% nitric acid and then diluted to 2% nitric acid and spiked with gallium. A 10-l aliquot of a 10 ppm 71 Ga standard was added to the each sample as an internal standard. A copper standard was prepared using 10-l aliquots of standard solutions of in 1 ml of 2% nitric acid. Each of the samples was introduced to the ICP (ThermoFinnigan Element 2) via a 200-l/min self-aspirating nebulizer in a cyclonic spray chamber (Glass Expansion). Results for each sample are calculated using the integrated average background-subtracted peak intensities from 20 consecutive scans. The concentration of copper of each sample was calculated with its protein concentration. results in cellular protection against the cellular oxidative stress and cell death by PrP C .
Interestingly, in response to DNA damage, ATM-mediated activation of p53 is typically shown to increase cell death through caspase-mediated apoptosis (44,45). In contrary to these early findings, we demonstrate that ATM-mediated activation of p53 actually confers protection against Cu(II)-induced cell death through PrP C . Although it is not clear at present how activated ATM can play such a dual role in either promoting or suppressing cell death. Our data described here may suggest a possible balance between the activation of ATM, which could lead to apoptosis, and the elevation of PrP C , which could prevent cell death and apoptosis (4). For example, depletion of ATM or p53/Sp1 abolishes cellular response of PrP C to Cu(II) stimulation (Fig. 4). These data indicated that ATMmediated activation of p53 and Sp1 pathways elevate PrP C . Conversely, depletion of PrP C results in higher levels of ATM and p53 activation (Fig. 6), which suggests a negative regulatory effect of PrP C on Cu(II)-induced activation of ATM and its downstream signaling molecules. Thus, our data support a central position of ATM in the protective response of PrP C to Cu(II) cytotoxicity as well as in a possible balance between ATM activation and the PrP C elevation to avoid ATM-mediated cellular death.
Even though we now know PrP C plays an active and protective role against Cu(II)-induced cellular oxidative stress through an ATM-dependent mechanism, how PrP C confers such a protective effect against the Cu(II)-induced cytotoxic effect is still unclear. Our data suggest that PrP C may minimize cellular oxidative stress by active regulation of intracellular copper concentration. By using the ICP-MS analysis, we demonstrated that the intracellular copper concentration in the wild type N2a cells is somewhat higher than that of the PrPdepleted N2a cells (Fig. 7). This result is similar to previous reports showing lower copper concentration in the PrP-knockout mouse brain tissues (10) and in the PrP-knock-out mouse neural CF10 cells (9). It is notable that the difference of the copper contents between the wild type and PrP-depleted N2a cells is relatively small. This could potentially be explained by incomplete depletion of PrP C by siRNA. A stronger argument for an active role of PrP C in the regulation of intracellular copper concentration could come from the inability of cells to stop gradual increase of intracellular copper without PrP C (Fig. 7B). With PrP C , intracellular copper increased with an acute burst within 10 min after the addition of a high level of Cu(II) but followed by a quick and strong correction back close to the normal cellular copper levels. In contrast, relatively slow and possibly passive accumulation of intracellular copper was seen in cells without PrP C . Such an acute response of intracellular copper to PrP C arguably indicates PrP C plays an active role in the copper uptake and regulation. One possible and logical way to confirm the role of PrP C in regulating intracellular cooper is to overexpress PrP and measure potential prevention of influx of intracellular copper upon Cu(II) stimulation. However, recent studies including ours (46) demonstrated that endogenous PrP C could behave quite differently from the exogenously produced PrP. In fact, opposite effects could sometimes be observed between these two types of PrP c . Thus, because of the molecular differences between the endogenous and exogenous PrPc are unknown at present, data generated by overexpression of PrPc through a plasmid would be difficult to interpret.
Other evidence that could support a role of PrP C in regulation of intracellular copper concentration is that PrP C is a copper-binding protein (13,14,47,48). Furthermore, because PrP C localizes on the cell membrane (4), it is possible that PrP C could transport copper across the cell membrane (16, 49 -51). Indeed, it has been suggested that PrP C could bind Cu(II) in extracellular medium at the neutral pH and then release it in acidic endosomes (16). Because copper chaperones can only bind and deliver the monovalent Cu(I) to target proteins in the cytoplasm (52,53), the endocytosed divalent Cu(II) ions need to be reduced to Cu(I). The N-terminal octa-repeat FIGURE 8. Model for responsive elevation of PrP C to Cu(II). Exposure of cells to the high concentration of copper causes accumulation of ROS, which leads to activation of ATM by phosphorylation at Ser-1981. The activated ATM activates p53 by phosphorylation at Ser-15 and the MEK/ERK/Sp1 pathway, which leads to bindings of p53 and Sp1 to the mouse Prnp or human PRNP promoter resulting in up-regulation the PrP C expression. The elevated PrP C reduces the intracellular copper content and in turn reduces the ATM activation. Through such signaling transduction as ATM-mediated cascade events, cells respond to copper cytotoxic effects by elevating PrP C , which protects against copper-mediated cellular damages. FEBRUARY 13, 2009 • VOLUME 284 • NUMBER 7 region of PrP C has been suggested to be able to reduce Cu(II) to Cu(I) (54,55).

Copper-induced Transcriptional Activation of PrP C
Copper, as a transition metal, serves as an essential role in a variety of normal physiological processes. However, a higher than normal intracellular copper concentration is cytotoxic because it promotes formation of ROS such as H 2 O 2 , O 2 Ϫ , and OH via Haber-Weiss and Fenton reactions and inactivates enzymatic and protein activities via interactions with special and chemical functional groups (56). Metalloproteins such as copper superoxide dismutase (SOD1) is also known to regulate intracellular copper concentration (57)(58)(59). Interestingly, early studies have suggested that PrP C can either enhance copper incorporation into superoxide dismutase or may actually have superoxide dismutase-like activity (47,49). Altogether, PrP C could in principle serve as a copper-binding protein on the cell membranes to control the intracellular copper concentration especially in response to such cellular stresses as that induced by copper. Obviously, further and detailed investigation is warranted to confirm this possibility. We propose a model (Fig. 8) to summarize the possible events that may happen after cellular exposure to Cu(II). The high concentration of Cu(II) causes cellular oxidative stress as indicated by the accumulation of ROS (Fig. 3), which in turn triggers the activation of ATM through Ser-1981 phosphorylation (Figs. 4 and 6). The activated ATM initiates the MEK/ERK/ Sp1 and p53 signaling pathways (Fig. 4) and enhances binding of Sp1 and p53 to the PrP promoter (Fig. 5). Up-regulation of Prnp or PRNP gene transcription by p53 and Sp1 results in the elevated production of PrP C (Figs. 2, 4, 6, and 7). The elevated PrP C counteracts Cu(II)-induced ROS and cell death (Fig. 3), probably through suppression of ATM activation and its downstream signaling molecules, such as p53 and Sp1 (Fig. 6), as well as reduction of the intracellular copper content (Fig. 7). Because our experiments were carried out in both mouse neuronal cells and non-neuronal human cells, the protection of PrP C against the Cu(II)-induced oxidative stress could be a general phenomena through similar signaling transduction pathways in both types of cells. In summary, findings described in this study should contribute to our understanding toward normal physiologic function of PrP C , interaction of copper with PrP C , and their role in prion biology.