c-Jun N-terminal Kinase Specifically Phosphorylates p66ShcA at Serine 36 in Response to Ultraviolet Irradiation*

Mice lacking expression of the p66 isoform of the ShcA adaptor protein (p66ShcA) are less susceptible to oxidative stress and have an extended life span. Specifically, phosphorylation of p66ShcA at serine 36 is critical for the cell death response elicited by oxidative damage. We sought to identify the kinase(s) responsible for this phosphorylation. Utilizing the SH-SY5Y human neuroblastoma cell model, it is demonstrated that p66ShcA is phosphorylated on serine/threonine residues in response to UV irradiation. Both c-Jun N-terminal kinases (JNKs) and p38 mitogen-activated protein kinases are activated by UV irradiation, and we show that both are capable of phosphorylating serine 36 of p66ShcA in vitro. However, treatment of cells with a multiple lineage kinase inhibitor, CEP-1347, that blocks UV-induced JNK activation, but not p38, phosphatidylinositol 3-kinase, or MEK1 inhibitors, prevented p66ShcAphosphorylation in SH-SY5Y cells. Consistent with this finding, transfected activated JNK1, but not the kinase-dead JNK1, leads to phosphorylation of serine 36 of p66ShcA in Chinese hamster ovary cells. In conclusion, JNKs are the kinases that phosphorylate serine 36 of p66ShcA in response to UV irradiation in SH-SY5Y cells, and blocking p66ShcA phosphorylation by intervening in the JNK pathway may prevent cellular damage due to light-induced oxidative stress.

The accumulation of oxidative cellular damage has been noted in a number of aging and neurodegenerative diseases, such as Alzheimer's, Parkinson's, and amyotrophic lateral sclerosis (1)(2)(3)(4). The molecular mechanisms leading to oxidative cell death have not been not fully elucidated; however, recent studies suggest that phosphorylation cascades may be involved in events triggered by free radical overproduction. For example, members of the mitogen-activated protein kinase (MAPK) 1 family, including p38 and JNKs, are activated upon exposure of cells to DNA-damaging agents, hydrogen peroxide, UV irradiation, and hypoxia (5)(6)(7)(8)(9). However, it is not clear what events downstream of kinase activation contribute to the ultimate damage induced by oxidative stress. In this regard, it is of interest that phosphorylation of the p66 ShcA adaptor protein has been shown to play an important role in signaling events leading to cell death in response to oxidative damage (10).
The mammalian ShcA adaptor protein has three isoforms: p46, p52, and p66 (11). All Shc isoforms contain a common structure with a C-terminal SH2 (Src homology 2) domain, a proline-and glycine-rich region, collagen-homologous region 1 (CH1), and an N-terminal PTB (phosphotyrosine binding) domain (12). In addition, there is a unique collagen homologous region 2 (CH2) domain at the N terminus of the p66 ShcA isoform. It is well established that the Shc proteins are phosphorylated at tyrosine residues in response to stimulation by a variety of growth factors and cytokines (12)(13)(14)(15)16). The p46 and p52 isoforms transmit signals from receptor tyrosine kinases to the Ras/MAPK pathway by forming a stable complex involving Grb2 and a Ras exchange factor, SOS (Son of Sevenless) (12,(17)(18)(19)(20)(21). However, p66 ShcA appears functionally different from the p46 and p52 isoforms. Unlike the other isoforms, p66 does not transform mouse fibroblasts, nor does it induce MAPK activation, although p66 ShcA is transiently phosphorylated at tyrosine residues in response to growth factor stimulation (22,23).
The p66 protein is phosphorylated at serine/threonine residues in response to insulin, Taxol, UV, or H 2 O 2 treatment (10,24,25). It has recently been demonstrated that, in response to UV and H 2 O 2 treatment, p66 ShcA is phosphorylated mainly at serine 36 in the N-terminal CH2 domain (10). p66 ShcA -null mouse fibroblast cells have enhanced cellular resistance to oxidative damage (10). Wild type p66 ShcA , but not a phosphorylation-defective mutant, can restore the normal stress response in the p66 ShcA -null mouse fibroblast cells. Moreover, mice lacking expression of p66 ShcA are less susceptible to chemical induced oxidative damage by paraquat and have an extended life span (10). Therefore, p66 ShcA , particularly its phosphorylation at serine 36, appears to be important for the cell death response upon oxidative stress, and the prevention of this phosphorylation may have therapeutic impact on diseases that are associated with oxidative damage.
We sought to identify the kinase(s) responsible for phosphorylating serine 36 of p66 ShcA . We report here that JNK1, JNK2, and JNK3 phosphorylate p66 ShcA in vitro and further demonstrate that blocking JNK activation prevents p66 ShcA phosphorylation in UV-irradiated SH-SY5Y human neuroblastoma cells.

EXPERIMENTAL PROCEDURES
Reagents-The kinase inhibitors SB203580, LY294002, and PD98059 were purchased from BIOMOL Research Laboratories (Plymouth Meeting, PA). CEP-1347, also known as KT7515, is a semi-synthetic derivative of K-252a and was provided by Kyowa-Hakko (Tokyo) (26). All stocks of inhibitors were dissolved in dimethyl sulfoxide (Me 2 SO) and stored at Ϫ20°C in amber glass vials. Further dilutions were made in appropriate solutions containing 0.05% bovine serum albumin (Fraction V, Protease-free). Pefabloc and bovine serum albu-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. min were purchased from Roche Molecular Biochemicals. Dulbecco's modified Eagle's medium, Neural Basal media, and B-27 serum-free supplement were purchased from Life Technologies, Inc. Aprotinin and protease inhibitor mixture were purchased from Sigma. Recombinant mixed lineage kinases (MLKs) and JNK3␣1 were made as described elsewhere (27,28). Recombinant active MEKK1, MKK4, MKK7, JNK1␣1, JNK2␣2, p38, ERK1, and immunoprecipitating p66 ShcA antibody were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Western blotting p66 ShcA antibody was purchased from Transduction Laboratories (Lexington, KY). The Grb2 antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Phospho-JNK and JNK antibody were purchased from Cell Signaling Technology (Beverly, MA). Myc antibody was purchased from Babco (Richmond, CA). IGF-1 was purchased from Calbiochem (San Diego, CA).
Cell Culture-The human neuroblastoma cell line SH-SY5Y, kindly provided by Dr. June Biedler (Memorial Sloan-Kettering Cancer Center, Rye, NY) and the Chinese hamster ovary cells (CHO; ATCC CCL-61) from the American Type Culture Collection (Manassas, VA) were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum at 37°C in 10% CO 2 . Cells were detached for passage by adding 0.25% trypsin. For all experiments in the presence of compounds the cells were maintained in Neural Basal media supplemented with B27 serum-free nutrient.
UV Treatment of SH-SY5Y Cells-In experiments utilizing MAPK inhibitors, cells were treated with vehicle (0.2% dimethyl sulfoxide), 15 M SB203580, 0.5 M CEP-1347, 10 M LY294002, or 50 M PD98059 for 1 h at 37°C. Cells in a minimal amount of medium were irradiated in a UV Stratalinker (Stratagene, La Jolla, CA) for varying times as indicated. The medium was then restored, and the cells were returned to the incubator for varying amounts of time before further analysis.
Immunoblots and Immunoprecipitations-Cell extracts were prepared by suspending cells in lysis buffer containing 10 mM Tris, 50 mM NaCl, 1% Triton X-100, 2 mM sodium vanadate, 0.2 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF, Roche Molecular Biochemicals), and proteinase inhibitor mixture (Sigma). Immunoblots and immunoprecipitations were performed according to the procedure of Harlow and Lane (29). Cell lysates or immunoprecipitates were separated on precast 10% Tris/glycine SDS-PAGE gels (Invitrogen, San Diego, CA), transferred to polyvinylidene difluoride membranes (Millipore, Chicago, IL), and probed with antibodies as indicated in buffer containing 10 mM Tris-HCl (pH 7.2), 150 mM NaCl, 0.1% Triton X-100, and 3% bovine serum albumin. After incubating with alkaline phosphataseconjugated secondary antibody, blots were developed by the ECF reagent (Amersham Biosciences, Inc.) and analyzed on the STORM Phos-phorImager system (Molecular Dynamics, San Francisco, CA). In the instances where the degree of the p66 ShcA mobility shift was determined, the distance between the peak volume of the p66 ShcA band relative to the reference p52 band was determined for each sample on the PhosphorImager. PP2A1 Treatment-In the alkaline phosphatase experiment, immunoprecipitated samples were treated with PP2A1 (Upstate Biotechnology, Inc.) as described elsewhere (25). In brief, immunoprecipitates were incubated with PP2A1 for 30 min at 30°C. The reaction was stopped by adding sample buffer and boiling.
DNA, Plasmids, and Transfection-Mouse p66 ShcA cDNA was provided by Dr. Jeffrey Pessin (University of Iowa, Iowa City, IA (30)). Serine 36 was mutagenized to alanine using the QuikChange kit (Stratagene). The JNK1 kinase-dead (JNK1KD) plasmid was constructed by substituting an EcoRV-EcoRI fragment of the pcDNA3-HA-JNK1 with a similar fragment that contained alanine 55 instead of lysine. DNAs were transfected into CHO cells by LipofectAMINE Plus reagents as described by the manufacturer (Life Technologies, Inc.).
Fusion Protein Generation and Purification-For generating calmodulin-binding protein (CBP)-p66 ShcA fusion proteins, either wild type or mutant S36A p66 ShcA cDNA was cloned into pCal-n (Stratagene) to generate CBP-p66 ShcA fusions. Expression plasmids were transformed into BL21(DE3) Escherichia coli strain (Novagen, Madison, WI) and grown to mid-log phase. Protein expression was induced by 1 mM isopropyl-1-thio-␤-D-galactopyranoside at 30°C for 3 h. Bacterial lysates were purified on a calmodulin affinity column according to the manufacturer's protocol (Stratagene).

UV Irradiation Results in Serine/Theronine Phosphorylation of the 66-kDa ShcA Isoform in SH-SY5Y Cells-P66
ShcA is serine/threonine-phosphorylated in response to growth factors, such as epidermal growth factor and insulin, and to cellular stress induced by UV irradiation, hydrogen peroxide, or Taxol treatment (30,24,10,25). To determine whether p66 ShcA was also phosphorylated on serine/threonine residues in response to cellular stress in a neuronal environment, the human neuroblastoma SH-SY5Y cell line was UV-irradiated, and electrophoretic mobility was assessed. A retarded electrophoretic mobility shift of p66 ShcA was observed in response to UV irradiation after 3 h of post-treatment incubation (Fig. 1, A and  B, lanes 1 and 2). In contrast, the mobilities of the p52 and p46 isoforms were not affected. The decreased mobility of p66 ShcA is characteristic of post-translational modifications, such as protein phosphorylation. Indeed, this decreased mobility was abolished by serine/threonine protein phosphatase 2A1 treatment (Fig. 1A, lanes 4 and 5). These data support the observation that p66 ShcA is phosphorylated at serine/threonine residue(s) after UV irradiation in SH-SY5Y cells, similar to previous results obtained in other cell systems (10,24,25,30).
Inhibition of JNK Activation Prevents Phosphorylation of p66 ShcA in SH-SY5Y Cells upon UV Irradiation-MAPKs, most notably JNKs and p38, have been shown to play an important role in response to oxidative damage in neuronal environments (31)(32)(33)(34). To investigate whether the MAPKs were responsible for p66 ShcA phosphorylation, several known kinase inhibitors were used to determine whether they could block p66 ShcA phosphorylation in response to UV irradiation. SH-SY5Y cells were treated with CEP-1347, an MLK inhibitor, SB203580, a p38 inhibitor, LY294002, a phosphatidylinositol 3-kinase inhibitor, or PD98059, a MEK1 inhibitor, followed by UV irradiation. The SB203580, LY294002, and PD98059 compounds were effective in blocking UV-induced phosphorylation of their respective downstream targets, ATF-2, AKT, and ERK1/2 (data not shown); however, they had no effect on the UV-induced p66 ShcA  9 and 10) and UV-irradiated (even-numbered lanes). Cell lysates were Western-blotted with the Shc antibody. The positions of the p66, p52, and p46 isoforms are indicated on the right. The degree of p66 ShcA mobility shift as indicated by the separation value was measured as the distance between the p52 and p66 Shc isoforms. These experiments were repeated twice, and similar results were obtained. mobility shift (Fig. 1B). In contrast, at a concentration at which CEP-1347 blocked JNK phosphorylation (data not shown), p66 ShcA mobility shift was also inhibited to the same degree as protein phosphatase 2A1 treatment (Fig. 1, A and B). As previously reported, both MLK3 and JNK activation are blocked in the presence of CEP-1347 in SH-SY5Y cells after UV irradiation (27). These results suggested that MLKs or kinase(s) regulated by MLKs might be responsible for p66 ShcA phosphorylation upon UV irradiation.
Phosphorylation of p66 ShcA at Serine 36 Is Regulated by the JNK Pathway-The major site of serine phosphorylation of p66 ShcA after hydrogen peroxide treatment of mouse embryonic fibroblasts is serine 36, and overexpression of a phospho-deficient serine 36 mutant of p66 ShcA in p66 ShcA -null fibroblasts attenuates hydrogen peroxide induced cell death (10). To determine whether the MLK/JNK pathway affected p66 ShcA phosphorylation as suggested by use of the kinase inhibitors in the whole cell studies described above (Fig. 1B) and in an effort to identify the kinase(s) responsible for its phosphorylation, kinases in the MLK/JNK pathway were co-expressed with either wild type (wt p66 ShcA ) or a serine 36 to alanine (S36A) mutant of p66 ShcA and electrophoretic mobility was assessed. The wild type and mutant S36A cDNAs were tagged with a myc epitope and expressed in CHO cells along with MLK3 and JNK1. As described previously (27), overexpression of MLK3 led to phosphorylation/activation of JNK ( Fig. 2A, lower panel). As shown in Fig. 2A, in the absence of an upstream activator of JNK1, co-expression of JNK1 with either wt p66 ShcA or S36A p66 ShcA did not alter the mobility of the Shc proteins. In contrast, co-expression of MLK3, JNK1 and wt p66 ShcA , but not S36A p66 ShcA protein, induced a decreased electrophoretic mobility of p66 ShcA (Fig. 2A). Interestingly, the decreased mobility of wt p66 ShcA was not observed when MLK3 and kinase-dead JNK1 (JNK1KD) were co-expressed (Fig. 2B). It was concluded from these results that the Shc kinase was not MLK3 or kinases downstream of MLK3 and upstream of JNK. These data suggested that JNK, when activated either by UV treatment or MLK3 expression, was able to phosphorylate p66 ShcA in whole cells.
Both JNKs and p38 Can Phosphorylate Serine 36 of p66 ShcA in Vitro-Whole cell analysis indicated that the JNKs might be responsible for p66 ShcA phosphorylation. An in vitro kinase assay was used to test candidate kinases directly. To generate recombinant wt p66 ShcA and S36A p66 ShcA proteins, wild type or mutant p66 ShcA cDNAs were fused to the CBP sequence. Shc proteins were expressed as CBP fusion proteins and purified on a calmodulin affinity column. The recombinant Shc proteins were then used as substrates in a kinase assay to determine whether proteins in the MAPK pathway could phosphorylate Shc specifically at serine 36 in vitro.
As shown in the Fig. 3, A and B, all kinases tested were capable of phosphorylating both themselves and MBP, indicating that all of the kinases tested were active in the kinase reactions. ERK1 (Fig. 3B) and MEKK1 (data not shown) were not able to phosphorylate Shc proteins, whereas MLK1, MLK2, MLK3, MKK4, and MKK7 appeared to phosphorylate wt p66 ShcA with no selectivity at serine 36 (Fig. 3C). In contrast, only ϳ10% of radioactivity was incorporated into S36A p66 ShcA relative to wt p66 ShcA when JNK1 and JNK2 were incubated with Shc proteins (Fig. 3, B and C). Similarly, JNK3 and p38 also demonstrated some selectivity to the serine 36 site but to a lesser extent then JNK1 and JNK2. Furthermore, there was a decreased electrophoretic mobility of p66 ShcA protein detected  1, 3, and 5 in panel A and lanes 1-4 in panel B), mutant S36A p66 ShcA (A; lanes 2, 4, and 6 in panel A), JNK1 (lanes 3-6 in panel A and lanes  2 and 3 in panel B), JNK1 kinase-dead (JNK1KD, lane 4 in panel B), and MLK3 ( lanes 5 and 6 in panel A and lanes 3 and 4 in panel B).  2, 5, 8, 11, and 14) or mutant p66 ShcA (S36A; lanes 3, 6, 9, 12, and 15) were incubated with ERK1 (lanes 1-3), p38 (lanes 4 -6), JNK1 (lanes 7-9), JNK2 (lanes 10 -12), or JNK3 (lanes 13-15) together with [␥-32 P]ATP in an in vitro kinase assay. The position of p66 ShcA is indicated on the right. The auto-phosphorylated kinases are indicated by a filled dot to the right. C, quantitation of the 32 P incorporation into p66 ShcA proteins in the kinase assay. The percentage of 32 P incorporation is expressed as the fraction of 32 P incorporated into mutant S36A p66 ShcA (filled bar) relative to wt p66 ShcA (shaded bar), which was set at 100%. The different kinases are indicated at the x axis. Data are the average Ϯ range from two independent experiments. A significant difference between phosphorylation of wt Shc and S36A-Shc was observed in samples containing JNK1, JNK2, and p38 as determined by one-tailed Student's t test statistics (p Ͻ 0.04, 0.01, and 0.07, respectively).
UV Treatment Does Not Result in an Increased Grb2 Binding to p66 ShcA -The p46 and p52 Shc proteins are known to be tyrosine-phosphorylated and form a complex with Grb2 and SOS protein upon growth factor stimulation (12,(17)(18)(19)(20)(21). This complex formation couples receptors to the Ras pathway and is critical for cell proliferation and survival. Although p66 ShcA is transiently phosphorylated at tyrosines upon growth factor stimulation, binding of Grb2 to p66 ShcA does not lead to activation of Ras, but rather evidence suggests that p66 ShcA may act as a sink for Grb2 and thereby effectively block growth factor-mediated responses (23,30).
To test whether serine phosphorylation of p66 ShcA by UV irradiation altered its binding to Grb2, co-immunoprecipitation experiments were performed. As a control, SH-SY5Y cells were serum-starved and stimulated with insulin-like growth factor-1 (IGF-1) (24). As expected, the amount of p66 ShcA associated with Grb2, after normalizing to Grb2 protein in each anti-Grb2 immunoprecipitation, was 3.8-fold higher after IGF-1 stimulation (Fig. 4, lanes 2 and 3). A greater association of Grb2 with p66 ShcA was observed in the presence of serum as compared with serum-starved cells (Fig. 4, lanes 2 and 4). UV treatment in the presence of serum, however, resulted in a decreased amount of p66 ShcA associated with Grb2 (Fig. 4,  lanes 4 and 5). Furthermore, we observed no mobility shift in the p66 ShcA associated with Grb2 either before or after UV treatment, suggesting that a non-serine phosphorylated form of p66 ShcA may be associated with Grb2 (Fig. 4, lanes 4 and 5). Similar results were obtained when Shc proteins were immunoprecipitated and Western-blotted with Grb2 antibody (data not shown). These data support the conclusion that, unlike growth factor stimulation, serine-phosphorylated p66 ShcA does not result in increased binding to Grb2 during UV irradiation. DISCUSSION The p66 ShcA protein appears to be a key protein in mediating cellular responses upon oxidative stress. P66 ShcA knock-out mice are less susceptible to the pesticide paraquat and also live 30% longer than their normal counterparts (10). Of particular interest is the demonstration that phosphorylation of p66 ShcA at serine 36 is important for the normal cell death response due to oxidative damage. Blocking phosphorylation of serine 36 prevents hydrogen peroxide-induced cell death in fibroblasts (10). In consideration of these observations, it was of interest to determine the kinase(s) that mediated serine 36 phosphorylation of p66 ShcA in response to UV irradiation.
In a UV-irradiated SH-SY5Y model, it is first shown that an inhibitor of the MLK pathway, CEP-1347, prevents p66 ShcA phosphorylation (Fig. 1). This inhibition appears to be specific because p38, phosphatidylinositol 3-kinase, and MEK1 inhibitors have no effect. CEP-1347 has been previously demonstrated to inhibit members of the MLK family, which are upstream activators of MKK4/7 leading to JNK activation. Thus, our initial results implicated members of the MLK family or kinases downstream of these proteins as potential Shc kinases. Phosphorylation of p66 ShcA at serine 36 is blocked in cells overexpressing MLK3 and a kinase-dead JNK1 (Fig. 2) suggesting that JNK1 is an Shc kinase. In support of these findings, three JNK isoforms can phosphorylate p66 ShcA , predominantly at serine 36, as determined in an isolated kinase assay (Fig. 3). It is concluded from both the whole cell and in vitro analyses that we have identified JNKs as candidate Shc kinases mediating phosphorylation at serine 36 of p66 ShcA .
The JNKs (also called stress-activated protein kinases (SAPKs)) are an integral part of the signal transduction pathways that transmit extracellular signals of stress and injury to a variety of cellular responses including cell death (5)(6)(7). JNK is activated upon DNA and oxidative damage due to UV irradiation, hypoxia, or apoptotic agents such as Taxol (35)(36)(37). By both molecular and pharmacological approaches, it has been demonstrated that blocking JNK activation can prevent apoptosis after UV irradiation and oxidative stress (9, 38 -40). One substrate of JNKs is the transcription factor c-Jun. In primary embryonic superior cervical ganglia and hippocampal neurons, activation of c-Jun after trophic withdrawal or potassium depolarization can lead to neuronal death, whereas blocking c-Jun either by neutralizing antibody, antisense oligonucleotides, or expression of dominant-negative proteins prevents cell death (41)(42)(43)(44). In support of the cell culture results, the hippocampal neurons from either JNK3 knock-out or phospho-null c-Jun mice are protected from kainate-induced cell death (45,46). Although it is clear that blocking c-Jun is sufficient to prevent certain types of cell death, the discovery that p66 ShcA is also a substrate of the JNKs suggests that perhaps more than one pathway is triggered by activation of JNKs leading to cell death. Whether p66 ShcA phosphorylation and c-Jun activation represent two unique or redundant events leading to cell death remains to be elucidated.
It has been reported that half of p66 ShcA protein is phosphorylated at serine/threonine residues and associates with Grb2 protein upon epidermal growth factor stimulation (30). The serine/threonine-phosphorylated form of p66 competes with the p52 and p46 Shc protein for the limited pool of Grb2 and thus dominantly inhibits epidermal growth factor receptor downstream signaling events. We tested whether phosphorylated p66 ShcA protein acts as a sink for Grb2 binding and thus may dominantly interfere or attenuate the survival pathway. Rather than an increase in p66 ShcA -associated Grb2 binding, UV irradiation resulted in a decreased amount of p66 ShcA protein associated with Grb2. These results are consistent with the lack of Grb2 binding to Taxol-induced serine-phosphorylated p66 ShcA (25). Although these studies do not distinguish the specific serine phosphorylated sites induced by the various stimuli, overall the results suggest that serine phosphorylation may lead to distinct downstream events dependent perhaps upon the site of phosphorylation and the upstream kinases involved.
Both JNKs and p38 are activated by UV irradiation and are capable of phosphorylating p66 ShcA in vitro (Fig. 3). However, only JNKs appear to phosphorylate p66 ShcA in SH-SY5Y cells, because inhibition of JNKs by CEP-1347 prevented p66 ShcA phosphorylation, whereas inhibition of p38 kinase by SB25038 had no effect after UV irradiation (Fig. 1). Nevertheless, it is possible that the choice of kinase employed to phosphorylate p66 ShcA may depend upon the cellular context.
The genetic data from the p66 ShcA knock-out mice suggest that serine phosphorylation of this protein may have profound effects on the extent of oxidative damage either directly inflicted or incurred with aging (10). Our data link the phosphorylation of p66 ShcA to the JNKs, a family well established as involved in oxidative stress responses. It will be of further interest to determine the downstream effects of p66 ShcA phosphorylation and the extent to which this protein contributes to JNK-mediated cell death.