p66shc Inhibits Pro-survival Epidermal Growth Factor Receptor/ERK Signaling during Severe Oxidative Stress in Mouse Renal Proximal Tubule Cells*

The fully executed epidermal growth factor receptor (EGFR)/Ras/MEK/ERK pathway serves a pro-survival role in renal epithelia under moderate oxidative stress. We and others have demonstrated that during severe oxidative stress, however, the activated EGFR is disconnected from ERK activation in cultured renal proximal tubule cells and also in renal proximal tubules after ischemia/reperfusion injury, resulting in necrotic death. Studies have shown that the tyrosine-phosphorylated p46/52 isoforms of the ShcA family of adaptor proteins connect the activated EGFR to activation of Ras and ERK, whereas the p66shc isoform can inhibit this p46/52shc function. Here, we determined that severe oxidative stress (after a brief period of activation) terminates activation of the Ras/MEK/ERK pathway, which coincides with ERK/JNK-dependent Ser36 phosphorylation of p66shc. Isoform-specific knockdown of p66shc or mutation of Ser36 to Ala, but not to Asp, attenuated severe oxidative stress-mediated ERK inhibition and cell death in vitro. Also, severe oxidative stress (unlike ligand stimulation and moderate oxidative stress, both of which support survival) increased binding of p66shc to the activated EGFR and Grb2. This binding dissociated the SOS1 adaptor protein from the EGFR-recruited signaling complex, leading to termination of Ras/MEK/ERK activation. Notably, Ser36 phosphorylation of p66shc and its increased binding to the EGFR also occurred in the kidney after ischemia/reperfusion injury in vivo. At the same time, SOS1 binding to the EGFR declined, similar to the in vitro findings. Thus, the mechanism we propose in vitro offers a means to ameliorate oxidative stress-induced cell injury by either inhibiting Ser36 phosphorylation of p66shc or knocking down p66shc expression in vivo.

The canonical Ras/MEK 2 /ERK pathway is activated upon ligand or, in some instances, H 2 O 2 stimulation of the epidermal growth factor receptor (EGFR) (1,2). Both ligand stimulation and H 2 O 2 induce phosphorylation of the EGFR, and the consequent binding of the Shc-Grb2-SOS1 complex to the activated EGFR results in activation of the Ras/Raf/MEK/ERK module (1, 3, 40 -42). However, toxic stress conditions uncouple the activated EGFR from ERK (4). The Shc (Src homology 2/␣-collagen-related protein) adaptor proteins have been shown to bind to a variety of receptors, including growth factor receptors such as the EGFR, and to couple those receptors to activation of Ras (5). In mammals, three shc family genes have been identified (shcA, shcB, and shcC), among which the shcA gene is the most ubiquitous (5). The shcA gene encodes three different proteins through differential usage of transcription/translation initiation sites and splicing (6). The p66 shc isoform of shcA is uniformly expressed in most cells except some hematopoietic cells and contains an extra N-terminal CH-like region with Ser 36 , which is absent in the p46/52 isoforms (5).
It has been demonstrated that the tyrosine-phosphorylated p46 shcA and p52 shcA isoforms couple the activated EGFR to Ras/ ERK activation during growth factor stimulation (5) or oxidative stress (1). Stress conditions (UV light and H 2 O 2 ) may also lead to Ser 36 phosphorylation of the 66-kDa isoform (7), which functions as a dominant-negative regulator of p46/52 shc and terminates Ras/ERK activation (8 -10). Interestingly, p66 shcϪ/Ϫ mice have an extended life span and show reduced sensitivity to oxidative stress (11). Thus, p66 shc is involved in signal transduction pathways that regulate cellular responses to oxidative stress and life span, and its absence increases resistance to oxidant injury and increases survival (12). Stress-activated p66 shc has also been implicated in non-mitochondrial events that facilitate cell death through negative regulation of the Ras/ERK pathway (13).
Activation of ERK is important for survival in the kidney in vivo and in renal epithelial cells in vitro during oxidative stress (14 -17). During the in vivo model of oxidative stress (ischemia/ reperfusion (I/R) injury), excess amounts of reactive oxygen species and free radicals (18,19), including H 2 O 2 , are formed and have been postulated to play a crucial role in the pathogenesis of renal injury (19). During I/R injury, the EGFR is activated in the proximal tubules of the kidney, but ERK activation is absent, and the proximal tubules undergo necrotic death (20,21). We made a similar observation in vitro: a high dose of H 2 O 2 causes necrotic (oncotic) death of renal proximal tubule cells with concomitant activation of the EGFR but not ERK, and ectopic activation of endogenous ERK rescues the cells from injury (14,15). On the other hand, during moderate stress (moderate dose of H 2 O 2 ) in vitro, renal proximal tubule cells survive through activation of the EGFR and ERK (14,15). These observations suggest that the activated EGFR could serve a prodeath function (22) in addition to its more widely accepted role of enhancing regeneration of the injured segments of the kidney (20,23,43,44). These dual roles of the EGFR have been described previously, as reactive oxygen species-dependent activation of the EGFR leads to cell death in renal proximal tubule cells exposed to cisplatin (24), and functional inactivation of the EGFR in renal proximal tubular cells reduces tubular-interstitial lesions after renal injury (25). Because the expression of p66 shc in the kidney has been demonstrated (26), we postulate here that activated p66 shc inhibits the survival signaling pathway by disconnecting the activated EGFR from Ras/ ERK activation depending on the extent of oxidative stress.
Accordingly, the aim of this study was to test the hypothesis that serine phosphorylation of p66 shc during severe oxidative stress in renal proximal tubule cells inactivates ERK and leads to cell death. We sought means to manipulate either expression or Ser 36 phosphorylation of p66 shc to restore ERK activation and survival.
Cell Lines and Animals-The immortalized mouse proximal tubule line TKPTS was used as described (14,15). Oxidative stress was induced by treatment of semiconfluent cells with 0.5 or 1 mM H 2 O 2 for various time points. For in vivo experiments, 129Sv mice were used, and I/R injury was induced as described (14,15).
Protein Isolation, Western Blotting, Immunoprecipitation, and Ras Activation-Kidneys were removed and homogenized in radioimmune precipitation assay buffer as described (14,15). Similarly, monolayers of TKPTS cells were lysed in radioimmune precipitation assay buffer. Protein content was determined using a Bio-Rad protein determination assay. 100 g of proteins from cell or tissue lysates were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Bio-Rad). The filters were hybridized with the appropriate primary antibodies followed by an horseradish peroxidase-conjugated secondary antibody. The bands were visualized using the ECL method (Amersham Biosciences) and quantified by densitometry (UN-SCAN-IT, Silk Scientific, Inc., Orem, UT). For immunoprecipitation, 400 g of total cell lysates were incubated with the appropriate primary antibody overnight at 4°C using the Catch and Release Version 2.0 reversible immunoprecipitation system (Millipore). Immunoprecipitated proteins were resolved by SDS-PAGE as described above. In immunoprecipitation studies, negative control IgGs (rat or sheep) were also used to determine cross-reactivity. We found that these control antibodies did not cross-react upon Western blotting (data not shown). Activated ras was determined using an activation kit (Millipore) following the instructions of the manufacturer.
Elk1-Luciferase Reporter Transactivation Assay-The pFR-Luc (reporter) and pAF2-Elk1 (fusion transactivator) plasmids were purchased from Stratagene (La Jolla, CA). Plasmids were transiently transfected into TKPTS cells using GenePORTER 2 reagent (Genlantis, Inc., San Diego, CA) together with a ␤-galactosidase plasmid (Invitrogen) in 6-well plates as described (27). Luciferase activity was determined using a luciferase assay kit (Promega, Madison, WI) as suggested by the manufacturer under control conditions and 6 h after treatment with either epidermal growth factor (EGF) or H 2 O 2 (0.5 or 1.0 mM). This time point was chosen because cells show no obvious damage at this time. The relative luciferase activity was measured and normalized to the amount of activity detected for a cotransfected ␤-galactosidase plasmid.
Transient Transfection of shc Plasmids-The following plasmids were used: wild-type p66 shc plasmid, p66 shc small interfering RNA (siRNA)-expressing (pTERsi66shc) plasmid, the p66 shc (S36A) mutant plasmid (in which Ser 36 was mutated to Ala), and the p66 shc (S36D) phosphomimetic mutant plasmid (in which Ser 36 was mutated to Asp). These plasmids were transiently transfected into TKPTS cells using GenePORTER 2 reagent in 6-well plates. Treatment protocols are described in the legends of the appropriate figures.
Establishment of a p66 shc Knockdown Cell Line-TKPTS cells were stably transfected in a T-25 culture flask with either 6 g of pTERsi66shc plasmid or the appropriate vector using Gene-PORTER 2 reagent. After 48 h, the cells were split into 100-mm Petri dishes in the presence of 200 g/ml Zeocin. 7-10 days later, the surviving colonies were picked up by Scienceware sterile cloning disks and serially propagated in 24-, 12-, and 6-well plates. The extent of p66 shc knockdown was determined by Western blotting. For additional experiments, we used the control T18C and p66 shc knockdown Tsi66-21 clones.
Statistical Analysis-Statistical differences between the treated and control groups were determined by Student's paired t test. Differences between means were considered significant if p Ͻ 0.05. All analyses were performed using the SigmaStat 3.5 software package.

Activation of the EGFR-mediated Ras/MEK/ERK Pathway Is
Attenuated during Severe Oxidative Stress-TKPTS cells survive 0.5 mM H 2 O 2 treatment via growth arrest, whereas they undergo necrotic (oncotic) death after treatment with 1 mM H 2 O 2 (14,15). Previously, we determined that both 0.5 and 1 mM H 2 O 2 tyrosine-phosphorylate the EGFR, similar to EGF treatment (15), but that ERK activation is absent after treatment of the cells with 1 mM H 2 O 2 , whereas 0.5 mM H 2 O 2 -induced ERK phosphorylation is persistent for a longer time (14), similar to treatment with EGF (data not shown). Here, we wanted to determine whether the activated EGFR is connected to downstream elements of signaling such as Ras, MEK, and ERK. Accordingly, TKPTS cells were treated with EGF or H 2 O 2 (0.5 mM or 1 mM) for various time points, and cell lysates were collected. Ras activation was determined by a kinase assay together with phosphorylation of MEK1 and ERK1/2 ( Fig. 1). Surprisingly, we found that 30 min after treatment, the Ras/ MEK/ERK pathway was activated regardless of the concentration of H 2 O 2 , similar to treatment with EGF. We also determined that the observed ERK activation at this time point required both the EGFR and ras, as pretreatment of the cells with AG1478 (an EGFR inhibitor) or infection with a dominant-negative ras adenovirus (28) inhibited ERK1/2 phosphorylation (data not shown). By contrast, activation of ras, as well as phosphorylation of MEK1 and ERK1/2, was greatly attenuated after 60 min of treatment with 1 mM H 2 O 2 , whereas EGF or 0.5 mM H 2 O 2 treatment still sustained ras/ MEK1/ERK activation.

Status of Shc Phosphorylation during Oxidative Stress in Vitro-
The Shc adaptor proteins can be tyrosine-phosphorylated by various agents (11,29,45), whereas stress conditions such as H 2 O 2 and UV irradiation can phosphorylate p66 shc at Ser 36 (11). The tyrosine-or serine-phosphorylated Shc isoforms play opposite roles in EGFR-mediated ERK activation (8,30). TKPTS cells were treated with either 0.5 or 1 mM H 2 O 2 , and cell lysates were immunoprecipitated with an anti-Shc antibody. After SDS-PAGE and transfer, the blots were immunoblotted with an anti-phospho-Shc (Tyr 239 /Tyr 240 ), anti-phospho-p66 shc (Ser 36 ), or anti-Shc antibody (Fig. 2). The p52 shc and p66 shc isoforms were tyrosine-phosphorylated by 0.5 and 1 mM H 2 O 2 . Ser 36 phosphorylation of p66 shc was observed only after treatment with 1 mM H 2 O 2 . We also determined whether ERK or JNK plays a role in Ser 36 phosphorylation of p66 shc as   (31,32). Accordingly, TKPTS cells were pretreated with the MEK/ERK inhibitor U0126 or infected with a dominant-negative JNK adenovirus prior to treatment with 1 mM H 2 O 2 , and Ser 36 phosphorylation of p66 shc was determined (Fig. 2B). Under this condition, both ERK and JNK are activated (phosphorylated) and can be inhibited by U0126 or dominantnegative JNK, respectively (14). Surprisingly, inhibition of both ERK and JNK also attenuated Ser 36 phosphorylation of p66 shc , but not tyrosine phosphorylation of p66 shc or p52 shc . We next determined whether Ser 36 -phosphorylated p66 shc is linked to ERK inhibition and cell death.
Involvement of p66 shc and Its Ser 36 Phosphorylation in Inhibition of ERK Function during Severe Oxidative Stress-We determined ERK activation/function using the Elk1-luciferase reporter transactivation system. Elk1 is a downstream target of ERK, and as such, its activation status reflects the activation status of ERK (33). Accordingly, TKPTS cells were transiently transfected with the Elk1-luciferase system together with a ␤-galactosidase plasmid and the plasmids indicated in Fig. 3A.
As shown, a constitutively active MEK plasmid dramatically increased luciferase activity (8-fold) compared with the untreated control. EGF or 0.5 mM H 2 O 2 treatment also increased luciferase activity, although to a lesser extent. By contrast, 1 mM H 2 O 2 significantly suppressed Elk1-mediated luciferase activity. Notably, cotransfection of a p66 shc siRNA-expressing plasmid or the p66 shc (S36A) mutant (34) restored Elk1-luciferase activity after treatment with 1 mM H 2 O 2 . In contrast, the phosphomimetic S36D mutant did not rescue 1 mM H 2 O 2 -mediated inhibition of Elk1-luciferase activity.
We also used the Tsi66-21 cell line, which was derived from TKPTS cells by stably transfecting the p66 shc siRNA plasmid, or its control vector-expressing counterpart (T18C). Fig. 3B shows significant knockdown of p66 shc expression in Tsi66-21 cells. Treatment of this p66 shc knockdown cell line with a high dose of H 2 O 2 significantly increased phosphorylation of ERK, whereas the same treatment of the control cell line failed to demonstrate increased ERK phosphorylation (Fig. 3C). In addition, tyrosine phosphorylation of the p52 shc isoform was longer lasting in the p66 shc knockdown cell line than in the control cell line after treatment with H 2 O 2 (data not shown). These results support the notion that Ser 36 -phosphorylated p66 shc indeed attenuates ERK activation during severe oxidative stress.
Involvement of p66 shc and Ser 36 Phosphorylation in Cell Death during Severe Oxidative Stress-Previously, we showed that in the absence of ERK activation, TKPTS cells undergo necrotic death during severe oxidative stress but that ectopic activation of endogenous ERK rescues cells from that death (14). Thus, isoform-specific knockdown of endogenous p66 shc or expression of its mutant (S36A) that restores ERK activation (Fig. 3) should ameliorate cell death under severe oxidative stress. Transient transfection of TKPTS cells by a p66 shc siRNA-expressing vector (34) significantly increased (by 2.5fold) the number of surviving cells 24 h after treatment with 1 mM H 2 O 2 as determined by trypan blue staining (Fig. 4A). In addition, transient transfection of the S36A mutant, but not the phosphomimetic mutant (S36D), also increased survival (by 2-fold). Similarly, survival of Tsi66-21 cells was significantly higher 6 or 24 h after treatment with a high dose of H 2 O 2 (Fig.  4B) compared with that of its control counterpart. These results prove that through Ser 36 phosphorylation, p66 shc is necessary for ERK inactivation (Fig. 3) and the consequent cell death during severe oxidative stress.
p66 shc Disrupts the EGFR-p52 shc -Grb2-SOS1 Complex during Severe Oxidative Stress-In the next step, we wanted to determine how p66 shc uncouples the activated EGFR from ERK activation during severe oxidative stress. Accordingly, TKPTS cells were treated with EGF or H 2 O 2 (0.5 or 1 mM) for different time points. Cell lysates were obtained, and protein complexes were immunoprecipitated with an anti-EGFR antibody, followed by immunoblotting with an anti-Shc, anti-Grb2, anti-SOS1, or anti-EGFR antibody. As shown in Fig. 5A, 30 min after treatment, the activated EGFR bound p46/52 shc , Grb2, and SOS1 adaptor proteins regardless of the type of treatment. 60 min after treatment with 1 mM H 2 O 2 , however, binding of the p66 shc isoform to the EGFR was also increased, with a concomitant decrease in SOS1 binding (Fig. 5B). By contrast, both EGF and 0.5 mM H 2 O 2 retained the EGFR-p46/52 shc -Grb2-SOS1  MARCH 7, 2008 • VOLUME 283 • NUMBER 10 complex formation without evidence of p66 shc binding. Furthermore, immunoprecipitation of proteins with an anti-Grb2 antibody from these 1 mM H 2 O 2 -treated cells showed that Grb2 increasingly bound p66 shc , but not p52 shc , and consequently less SOS1 (Fig. 6A). Also, we demonstrated that transient overexpression of wild-type or mutant (S36A or S36D) p66 shc increased Grb2 binding to p66 shc in 1 mM H 2 O 2 -treated cells (Fig. 6B). However, SOS1 bound to Grb2 only in the presence of the S36A mutant (Fig. 6B). These results demonstrate that Ser 36 phosphorylation of p66 shc is important for disruption of the Grb2-SOS1 complex. Further evidence for p66 shc -mediated inhibition of effective complex formation is shown in the experiments summarized in Fig. 6C, where the p66 shc knockdown cell line Tsi66-21 showed diminished p66 shc but increased SOS1 binding to the EGFR compared with the control cell line (T18C) after treatment with a high dose of H 2 O 2 .

p66 shc Inhibits Renal Cell Survival during Oxidative Stress
These results suggest that p66 shc -bound Grb2 has decreased affinity for SOS1 during severe oxidative stress. Knockdown of p66 shc expression decreased p66 shc binding to the EGFR and restored SOS1 binding to the EGFR-p46/52 shc complex (Fig.  6C). The results are summarized in Fig. 7. Accordingly, EGF, 0.5 mM H 2 O 2 , or even early 1 mM H 2 O 2 treatment facilitates formation of the EGFR-p52 shc -Grb2-SOS1 complex, which leads to ras/ERK activation (Fig. 7A). In contrast, at a later time point (Fig. 7B), 1 mM H 2 O 2 treatment induces binding of p66 shc to the EGFR, resulting in Grb2 binding through p66 shc and the consequent dissociation of the p52 shc -Grb2-SOS1 complex. Under this condition, ras and consequently ERK are not activated.

I/R Injury in the Kidney Increases Phosphorylation of Shc Proteins and Their
Binding to the EGFR-First, protein lysates from kidneys of 129Sv mice that underwent I/R injury were immunoprecipitated with an anti-Shc antibody, followed by immunoblotting with an anti-phospho-p66 shc (Ser 36 ), antiphospho-Shc (Tyr 239 /Tyr 240 ), or anti-Shc antibody. As shown in Fig. 8A, Ser 36 phosphorylation of p66 shc was significantly increased 30 min and 24 h after reperfusion. Tyrosine phosphorylation of the p52 shc isoform that preceded p66 shc Ser 36 phosphorylation also occurred in the kidney. Probing with an anti-EGFR antibody showed that Shc proteins increasingly bound to the EGFR in the reperfusion phase. Immunoprecipitation of proteins with an anti-EGFR antibody, followed by immunoblotting with an anti-Shc or anti-phospho-p66 shc (Ser 36 ) antibody, revealed that the EGFR increasingly bound Ser 36 -phosphorylated p66 shc during reperfusion (Fig. 8, A and B). In addition, there was a decline in the initial SOS1 binding to the EGFR complex 30 min after reperfusion that coincided with the increased binding of Ser 36 -phosphorylated p66 shc to the EGFR (Fig. 8, A and B). Whether EGFR-bound and Ser 36 -phosphorylated p66 shc is localized to the proximal tubules that undergo necrotic death during I/R injury needs further studies. It is important to note that the EGFR is expressed mostly in renal proximal tubules (35). These observations are very similar to those shown in vitro after severe oxidative stress (Fig. 5B), suggesting a similar mechanism that inhibits the pro-survival ERK activation in vivo.

DISCUSSION
Increasing evidence shows a role for the p66 shc adaptor protein in mediating oxidative stress-related injury in a variety of cell types and under a variety of non-physiological conditions such as obesity, diabetes, inflammation, etc. (7,9). Other studies have also shown that Ser 36 phosphorylation of p66 shc is necessary for this effect (11,36): Ser 36 -phosphorylated p66 shc can act as a dominant-negative factor to antagonize the function of tyrosine-phosphorylated p46/52 shc , i.e. inhibits coupling of the activated EGFR to downstream effectors such as Ras/MEK/ERK (8,30). Lack of ERK activation could facilitate cell death during oxidative stress in renal tubular cells (14,15,37). This may be especially relevant under conditions in which phosphorylation of p66 shc at Ser 36 (11) prevents activation of ERK and increases cell death (8,30).
Here, we have demonstrated that phosphorylation of ShcA proteins by H 2 O 2 depends on the dose of the oxidant in renal proximal tubule cells in vitro: a moderate dose of H 2 O 2 tyrosine-phosphorylates ShcA proteins (Fig. 2), similar to EGF treatment (data not shown). On the other hand, a high dose of H 2 O 2 phosphorylates p66 shcA at Ser 36 , in addition to tyrosine phosphorylation of p52 shcA and p66 shcA (Fig. 2A). This Ser 36 phosphorylation of p66 shc is MEK/ERK-and JNKdependent (Fig. 2B), as shown previously (31,32). Whether Ser 36 phosphorylation of p66 shc is a dosedependent or all-or-none event will require further studies at the single-cell level rather than in a mixture of cells.
These results also suggest that the observed inhibition of the Ras/MEK/ERK pathway under severe oxidative stress (Fig. 1), the consequent death of these cells (14,15), and the Ser 36 phosphorylation of p66 shc ( Fig. 2A) are probably related and that the observed early ERK activation (Fig. 1), together with activation of JNK (14), serves as a negative feedback mechanism to shut down the ERK-dependent survival pathway during severe oxidative stress.
Indeed, siRNA-mediated knockdown of the p66 shc isoform restored ERK function both in a transient reporter transactivation assay (Fig. 3A) and in a p66 shc knockdown TKPTS cell line after treatment with a high dose of H 2 O 2 (Fig. 3C). Similarly, knockdown of p66 shc increased resistance to high dose H 2 O 2induced cell death (Fig. 4). Our results also show that substitution of Ser 36 with Ala blunted high dose H 2 O 2 -induced inhibition of ERK function (Fig. 3A) and cell death (Fig. 4), whereas the phosphomimetic S36D mutant did not.
Furthermore, we sought to determine the mechanism by which Ser 36 -phosphorylated p66 shc uncouples the activated EGFR from ras/MEK/ERK activation (Fig. 1) during severe oxidative stress. The activated EGFR recruits adaptor proteins such as Shc, Grb2, and SOS1 to activate ras (5). In this process, the tyrosine-phosphorylated p46/52 shc isoforms play an important role, whereas p66 shc could antagonize this process (30). It has been reported that p66 shc and p52 shc compete for the available Grb2 (8). Accordingly, we determined the assembly of the EGFR-Shc-Grb2-SOS1 complex by immunoprecipitation after   (30 min) activates the EGFR, which in turn recruits p52 shc , Grb2, and SOS1 to activate ras and ultimately ERK. Under these conditions, cells survive. B, at a later time point (60 min), 1 mM H 2 O 2 activates p66 shc , which binds the EGFR in addition to p52 shc . p66 shc , but not p52 shc , binds Grb2 but is unable to recruit SOS1; ras and ERK are not activated, and cells die. At this time point, both EGF and 0.5 mM H 2 O 2 sustain the activation of the Ras/ERK pathway through formation of the EGFR-p52 shc -Grb2-SOS1 complex (A) and the consequent survival. treatment with 0.5 or 1 mM H 2 O 2 and compared it with a fully executed assembled unit using EGF treatment. At an early time point, all agents induced the assembly of the EGFR-p46/52 shc -Grb2-SOS1 complex (Fig. 5A), with its resultant activation of ras/MEK/ERK (Fig. 1). Interestingly, at a later time point, 1 mM H 2 O 2 treatment increased binding of p66 shc to the EGFR, with a concomitant decrease in SOS1 binding (Fig. 5B). Not surprisingly, 1 mM H 2 O 2 treatment decreased ras/MEK/ERK activation under this condition (Fig. 1). On the other hand, EGF and 0.5 mM H 2 O 2 treatment sustained the EGFR-p46/52 shc -Grb2-SOS1 complex (Fig. 5B) at this time point and the consequent activation of ras/MEK/ERK (Fig. 1). Further co-immunoprecipitation studies revealed that Grb2 increasingly bound to p66 shc but bound less to SOS1 (Fig. 6A). Overexpression of wild-type p66 shc and its phosphomimetic mutant (S36D) increased binding of Grb2 to p66 shc but decreased binding to SOS1 (Fig. 6B) after treatment with a high dose of H 2 O 2 . In contrast, in the presence of the S36A mutant, SOS1 binding to Grb2 was restored during severe oxidative stress (Fig. 6B); this observation further supports the role of the Ser 36 phosphorylation of p66 shc in inhibition of ERK activation (Fig. 3) and cell death (Fig. 4). Khanday et al. (38) demonstrated that the N-terminal (CH2) domain of p66 shc competes with the C-terminal region of SOS1 for the SH3 (Src homology 3) domain of Grb2, resulting in dissociation of the Grb2-SOS1 complex upon activation of p66 shc during oxidative stress. This mechanism also implies reduced ras activation, as the Grb2-SOS1 complex formation is impaired. Okada et al. (8) showed that serine/threonine-phosphorylated p66 shc is associated with Grb2 and competes for Grb2 binding with p52 shc , resulting in inhibition of EGFR function. Migliaccio et al. (30) demonstrated that the p66 shc -Grb2 complex does not activate ERK. Furthermore, the CH2 domain of p66 shc contains Ser 36 , the phosphorylation of which might affect p66 shc -Grb2 binding and the consequent dissociation of the Grb2-SOS1 complex, as our results suggest (Fig. 6B). We hypothesize that during severe oxidative stress, EGFR-bound p66 shc binds at least part of the available Grb2 and decreases the available Grb2-SOS1 complex, which would connect the activated EGFR to Ras/ERK activation.
We have also shown that in the p66 shc knockdown cell line, the EGFR bound significantly less p66 shc but more SOS1 compared with the control cell line during severe oxidative stress (Fig. 6C). Under these circumstances, the knockdown line also showed activation of ERK, whereas the control cell line did not (Fig. 3C).
Taken together, our data are consistent with the notion that severe oxidative stress (1 mM H 2 O 2 ) disrupts the EGFR-p46/ 52 shc -Grb2-SOS1 complex through binding the available Grb2 to activated p66 shc , as observed by others (8), resulting in the subsequent disruption of EGFR signaling to ERK. This process requires Ser 36 phosphorylation of p66 shc , as inhibition of Ser 36 phosphorylation by the S36A mutant, but not the phosphomimetic S36D mutant, restored ERK function (Fig. 3) and increased survival (Fig. 4).
Notably, we found that the EGFR increasingly bound Ser 36phosphorylated p66 shc in addition to tyrosine-phosphorylated p46/52 shc in the mouse kidney after I/R injury (Fig. 8), similar to the results found in proximal tubule cells in vitro (Fig. 5B). Parallel with the increased p66 shc binding, association of SOS1 with the EGFR was diminished in vivo (Fig. 8B), as observed under severe oxidative stress in vitro (Fig. 5B). Several studies have demonstrated increased expression and activation of the EGFR after I/R injury in the kidney (20,21), but its role has not been defined. The EGFR, which is present in the proximal tubules (39), is activated by I/R but fails to activate ERK under such conditions. Ser 36 -phosphorylated p66 shc and its association with the activated EGFR would seem to be a reasonable explanation for the failure of ERK to be activated in I/R injury. Further studies using p66 shcϪ/Ϫ mice (11) will clarify this phenomenon. This mechanism may also offer a means to ameliorate oxidative stress-induced cell injury by either inhibiting Ser 36 phosphorylation of p66 shc or knocking down p66 shc expression in vivo. Acknowledgment-We thank Dr. Judit Megyesi for valuable help in the in vivo experiments.