![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 16, Issue 8, 3705-3718, August 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* The Institute of Molecular and Cell Biology, Proteos, Singapore 138673, Singapore;
National Cancer Centre, Singapore 169610, Singapore;
Howard Hughes Medical Institute and Department of Biochemistry and Molecular Biophysics, College for Physicians and Surgeons, Columbia University, New York, NY 10032; and
Howard Hughes Medical Institute and Department of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA 98109
Submitted April 11, 2005;
Accepted May 24, 2005
Monitoring Editor: Tony Hunter
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
; Droge, 2002). They are produced upon exposure to 
-ray or UV light, during oxidative phosphorylation in mitochondria, or during inflammation by macrophages, and cause damage to DNA, proteins, and lipids, leading to cell death in the form of necrosis (high dosage) or apoptosis (low dosage). A connection between oxidative stress and aging has been established in Drosophila and Caenorhabditis elegans (Finkel and Holbrook, 2000). DAF-16 in C. elegans, encoding the forkhead related transcription factor (FOXO3a/FKHRL1 in mammals), promotes longevity in response to reduced insulin/IGF-1 signaling (Kenyon et al., 1993). An analogous example in mammals is the p66shcA knockout mice. Deletion of p66shcA from mouse genome by gene targeting renders the mice resistant to oxidative stress, and a 30% increase in average lifespan of the mice was observed. Mouse embryonic fibroblasts (MEFs) isolated from these mice are also resistant to the cytotoxic effects of H2O2 (Migliaccio et al., 1999). Oxidative stress leads to phosphorylation of p66shcA on Ser36. p66shcA carrying a S36A mutation, when reintroduced back into p66shcA null MEFs, could not correct the resistance of these cells to oxidative stress, suggesting that Ser36 phosphorylation is critical for the proapoptotic activity of p66shcA in response to oxidative stress (Migliaccio et al., 1999). Recent studies have suggested a link between p66shcA and FOXO3a. In p66shcA-deficient MEFs, H2O2-induced phosphorylation/inactivation of FOXO3a is inhibited, correlated with increased activity of FOXO3a and up-regulation of target genes involved in free radical scavenging and oxidative stress resistance. Hence, the proapoptotic activity of p66shcA in response to oxidative stress might be mediated by FOXO3a (Nemoto and Finkel, 2002). Furthermore, phosphorylation of Ser36 on p66shcA was found to be critical for FOXO3a phosphorylation (Nemoto and Finkel, 2002). Hence, Ser36 phosphorylation is essential for p66shcA function in oxidative stress responses, in which Akt/FOXO3a play critical roles. However, the kinase responsible for this phosphorylation has not been identified.
p66shcA, p52shcA, and p46shcA constitute the ShcA family (reviewed in Luzi et al., 2000; Ravichandran, 2001). All three isoforms contain a phospho-tyrosine binding domain (PTB), a collagen homology domain 1 (CH1), and a Src homology 2 domain (SH2), whereas p66shcA has a unique collagen homology domain 2 (CH2) at the N' terminus in which Ser36 is located (Luzi et al., 2000). Working as adaptor proteins, p52 and p46 transmit signals from receptor tyrosine kinases (RTK) to the RAS-MAPK pathway (reviewed in Luzi et al., 2000). Their SH2 and PTB domains are responsible for binding to the phosphorylated tyrosine residues of activated RTK, which subsequently phosphorylates the CH1 domain on tyrosine residues Y239/240 and/or Y317. Phosphorylated ShcA proteins form a complex with Grb2, which constitutively interacts with Sos (Rozakis-Adcock et al., 1992). Recruitment of Grb2-Sos complex to the plasma membrane by p46/52 shcA leads to activation of Ras-ERKs pathway, and this promotes cell proliferation/transformation and cell survival (Ullrich and Schlessinger, 1990; Schlessinger, 2000). p66shcA is also phosphorylated after growth factor stimulation, but it competes with p52shcA for Grb2 binding (Migliaccio et al., 1997). Overexpression of p66shcA markedly inhibits activation of ERK after epidermal growth factor (EGF) stimulation in Chinese hamster ovary cells (Okada et al., 1997). Thus p66shcA and p52/46shcA have distinct functions in response to growth factor stimulation. Although p66shcA plays an important role in cell response to oxidative stress, the roles for p52shcA and p42shcA have not been explored.
In addition to their toxic and proapoptotic roles, ROS have been recently recognized as molecules with important physiological functions including acting as an oxygen sensor in the regulation of erythropoietin production, maintenance of redox homeostasis, and enhancement of cell signaling and amplification of cell responses to growth factors and antigens (Reth, 2002; Mikkelsen and Wardman, 2003). In some cases, ROS appear to function as second messengers in cell signaling, being produced in response to external stimuli and subsequently used for signal amplification (reviewed in Droge, 2002; Finkel, 2003). ROS can directly activate signaling molecules via modification, or indirectly activate signaling cascades by inhibiting protein tyrosine phosphatases. Yet the molecular mechanisms by which ROS function as a mitogen are still elusive. In an attempt to identify the kinase responsible for ShcA66 Ser36 phosphorylation upon oxidative stress, we found that 1) ERKs were responsible for H2O2-induced Ser36 phosphorylation of p66shcA; 2) ERKs acted upstream of p66shcA-Akt-FOXO3a pathway, which was delineated from studies on p66shcA-/- cells (Nemoto and Finkel, 2002); 3) one major target of ERK-ShcA66-FOXO3a pathway was p27, whose down-regulation by H2O2 required Ser36 phosphorylation of p66shcA; 4) p27 participated in cell response to oxidative stress; 5) H2O2-induced ERK activation was facilitated by p52/46shcA but hindered by p66shcA, and that H2O2 promoted the interaction between p66shcA and p52/46shcA. This suggests the existence of a negative feedback loop for regulation of ERK activation. The interaction between ShcA proteins may provide a novel mechanism by which ShcA proteins exert their diverse functions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Cell Culture
MEFs were prepared following a standard protocol and these primary MEFs, NIH3T3, and immortalized Jnk1-/-Jnk2-/-, ShcA+/+, ShcA-/-, Src+/+, and Src-/- MEFs were cultured in DMEM (Life Technologies, Rockville, MD) containing 10% fetal calf serum (HyClone, Logan, UT) supplemented with glutamine and penicillin/streptomycin, and COS7 cells were cultured in DMEM containing 10% bovine calf serum (HyClone) supplemented with glutamine and penicillin/streptomycin. To generate ShcA-/- cells expressing p66shcA or p66shcA(S36A), ShcA-/- MEFs were transfected with constructs expressing these proteins or the empty vector, respectively, selected against puromycin (expressed by these constructs) for 2 wk. The expression of p66shcA or p66shcA(S36A) was analyzed by Western blot in comparison to ShcA+/+ MEFs.
Mutagenesis
Point mutations of ShcA were generated using a site-directed mutagenesis Kit from Invitrogen (Carlsbad, CA), following the manufacturer's protocol. Truncations of ShcA were synthesized by PCR and subsequently cloned into the expression vector pcDNA3.1. All the expression constructs were sequenced to verify the mutations.
Immunoprecipitation, Western Blot, and Immunohistochemistry
Cells were subcultured the day before and then treated with H2O2 for various durations of time. To test the effects of the kinase inhibitors, specific inhibitors were added to the cell cultures 1 h before the addition of H2O2. Cells were washed with phosphate-buffered saline (PBS) and lysed in TNEN buffer containing 50 mM Tris, pH 7.5, 100 mM KCl, 1 mM EDTA, 0.5% NP-40, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM sodium orthorvanadate, 10 mM NaF, 1 mM beta-glycerol phosphate, and 10 µg/ml each of aprotonin and leupeptin (Li et al., 2002). Protein concentrations were determined by Bio-Rad protein quantitation assays (Richmond, CA). Equal amounts of protein (20 µg) were fractionated by electrophoresis on SDS-PAGE gels, transferred onto a nitrocellulose membrane (Millipore, Billerica, MA), probed with primary and secondary antibodies, and visualized using an ECL kit (Amersham, Piscataway, NJ). For immunoprecipitation (IP), antibodies against HA, myc, ShcA, or ERKs were added to the cell lysate (400 µg) for overnight incubation, followed by addition of protein A plus G agarose beads for two more hours. The beads were washed with TNEN buffer three times and once with PBS. The immunoprecipitated proteins were released from the beads by boiling in 1x sample buffer for 5 min and subsequently analyzing by Western blot. Total cell lysate (20 µg) was run next to the IP samples to assess the portions of proteins that were coprecipitated.
COS7 cells grown on coverslips were transfected with lipofectamine (Life Technologies, Gaithersburg, MD). After 24 h, the cells were permeabilized with methanol at -20°C for 3 min, washed three times with PBS containing 10% bovine calf serum, and incubated with the primary antibodies for 1 h at room temperature. 9E10 was used to detect p66shcA (myc-tagged), and polyclonal anti-ERK antibodies were used to detect ERKs. After interaction with the primary antibody, the slides were washed three times with PBS containing 10% bovine calf serum and then incubated with secondary antimouse antibodies conjugated with Texas Red (Sigma, St. Louis, MO) and anti-rabbit antibodies conjugated with FITC for 1 h at room temperature. The slides were then rinsed with PBS and examined under a confocal microscope.
GST Fusion Protein Preparation
ShcA 46, ShcA52, ShcA66, and ShcA66(S36A) were each cloned into the pGEX vector that was then transformed into BL21 cells. The bacteria were cultured in 40 ml LB with 100 µg/ml ampicillin at 37°C overnight with vigorous shaking. On the second day, the cultures were used to inoculate 800 ml LB medium with 100 µg/ml ampicillin and shaken at 37°C until OD600 reached 
1.0. IPTG was added to the culture to a final concentration of 0.15 mM. The cells were further cultured at 30°C for 6 h, harvested, resuspended in MTPBS buffer (150 mM NaCl, 16 mM Na2HPO4, 5 mM NaH2PO4, 1 mg/ml lysozyme, 1% Trition X-100, 0.05 mM PMSF, 1 mM dithiothreitol [DTT]), and sonicated 30 s for five times. Insoluble debris was removed by centrifugation at 3000 xg for 30 min at 4°C. The cell lysates were combined with 50% slurry of glutathione-agarose resin and then incubated for 30 min at 4°C. The resin was washed with PBS five times. Bound GST fusion protein was eluted using elution buffer (10 mM reduced glutathione, 50 mM Tris-HCl, pH 8.0).
In Vitro Kinase Assay for ERK
ERK1 (HA-tagged) was expressed in COS7 cells, and one set of cells was treated with 0.5 mM H2O2 for 10 min. The other set was left untreated. ERK1 was immunoprecipitated from either set of cells using anti-HA antibody and protein A-agarose beads. Purified GST-ShcA66 or GST-ShcA66(S36A) protein (3 µg each) was mixed with protein A beads bound-ERK in 30 µl kinase assay buffer (25 mM Tris-HCl, pH 7.4, 5 mM MgCl2, 0.5 mM EGTA, 1 mM DTT, 20 µM ATP, 20 µCi [-32P]ATP, 0.1 mM sodium orthovanadate) and incubated at room temperature for 20 min. The reaction was terminated by addition of 5x SDS-PAGE sample buffer and boiling at 100°C for 5 min. Proteins were analyzed by SDS-PAGE, and the phosphorylated form of GST-ShcA66 or GST-ShcA66(S36A) was detected by auto-radiography.
RT-PCR
To determine the mRNA levels of p27, we isolated total RNA from cells using Trizol. Reverse transcription reactions were carried out with 5 µg of total RNA following the standard protocol supplied with the reverse transcriptase (Roche, Indianapolis, IN). The resulting cDNA was used for PCR and actin was used as a control. The primers used are: 5'-GTCAAACGTGAGAGTGTCTAAC-3';5'-GTTTACGTCTGGCGTCGAAGGC-3'. The primers for actin were as follows: forward, 5'-AGATGTGGATCAGCAAGCAG-3'; reverse, 5'-GCGCAAGTTAGGTTTTGTCA-3'.
Cell Cycle and Cell Viability Analysis
Cells were serum-starved overnight, treated with increasing concentrations of H2O2 for 24 h, trypsinized and fixed in 70% ethanol, and washed and analyzed by flow cytometry after PI staining. To measure cell death rates, cells were cultured in 96-well plates at 1 x 104/well, serum-starved for 24 h, and then treated with different concentrations of H2O2 for 4 h. Cell proliferation reagent WST-1 (Roche) was added to each well, and the cells were incubated for 1 more hour at 37°C. The absorbance was measured against a background control by microplates (ELISA) reader at 430 nm. The reference wavelength is 650 nm. Trypan blue exclusion assays were used to confirm the results obtained from this quantitative assay.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
; Le et al., 2001; Foschi et al., 2001). In these studies, a retarded migration of p66shcA on a Western blot was used as an indicator for Ser/Thr phosphorylation. It was discovered that H2O2-induced p66shcA phosphorylation occurred on Ser36 (Migliaccio et al., 1999). Because p66shcA can be phosphorylated at multiple sites: Ser36, Ser138, and tyrosine residues on the CH1 domain (Faisal et al., 2002), migration retardation may not necessarily represent phosphorylation at Ser36 (Supplementary Data S1). Instead we used a recently developed monoclonal antibody that specifically recognizes Ser36 phosphorylated p66shcA throughout our studies (Figure 1A). A dosage course study showed that as little as 0.2 mM of H2O2 could induce Ser36 phosphorylation in serum-starved cells (Figure 1B). Ser36 phosphorylation reached the maximal level at 0.5 mM and then declined with higher doses of H2O2. A time course study revealed that Ser36 phosphorylation (close to the maximal level) occurred at 10 min after stimulation with 0.5 mM H2O2 (Figure 1C).
|
), and ERK has been implicated in p66shcA phosphorylation in response to endothelin or taxol (Yang and Horwitz, 2000; Foschi et al., 2001). As expected, JNKs were activated upon H2O2 treatment in NIH3T3 cells (unpublished data). Yet pretreatment with 20 µM of JNK specific inhibitor (L- or D-form) for 1 h did not affect H2O2-induced Ser36 phosphorylation (unpublished data). The JNK inhibitor, at 20 µM, is sufficient to inhibit JNK activity (Le et al., 2001). Furthermore, immortalized MEFs derived from mice deficient for both JNK1 and JNK2 was compared with NIH3T3 cells and we found that phosphorylation of p66shcA on Ser36 was not significantly inhibited in the absence of JNK1/2 (Figure 2A). Because JNK3 is not expressed in MEFs (Tournier et al., 2000), we conclude that activation of JNKs is not an absolute requirement for H2O2-induced phosphorylation of Ser36. We observed that the knockout cells showed a slight increase in the basal level of phosphorylated p66shcA. This may reflect a compensation mechanism occurred to cells deficient for all three JNKs, e.g., inactivation of phosphatases involved in terminating JNK signaling. Similarly, inhibition of p38 MAPK activation by SB25083 showed no effect on Ser36 phosphorylation (unpublished data).
|
We next tested whether ERKs, which were also activated by H2O2 (Figure 2B), played a role in phosphorylation of p66shcA on Ser36 in response to oxidative stress. We used a widely used inhibitor U0126 to inhibit MEK1, a kinase that directly phosphorylates and activates ERK1/2. We found that Ser36 phosphorylation of p66shcA was inhibited in a dosage-dependent manner when cells were pretreated with U0126 (Figure 2C). ERK1/2 activation was indeed suppressed by the inhibitor in a dosage-dependent manner, suggesting that ERK1/2 activation was required for Ser36 phosphorylation.
To further confirm this conclusion, a constitutively active form of MEK1 was coexpressed with p66shcA and phosphorylation of Ser36 was tested on immunoprecipitated p66shcA (Mansour et al., 1994). It was found that activated MEK1 led to Ser36 phosphorylation even without H2O2 treatment and to much higher levels (Figure 2D), suggesting that ERK activation was sufficient for Ser36 phosphorylation.
To demonstrate that p66shcA is a substrate for ERKs, an in vitro kinase assay was carried out. p66shcA was fused to GST and the fusion protein was expressed and purified. A mutant GST-p66shcA, Ser36 to Ala (S36A), was also generated by site-directed mutagenesis, expressed in bacteria, and purified. ERK1 was expressed in COS7 cells and immunoprecipitated. The kinase assay was carried out using p66shcA or p66shcA (S36A) as a substrate. It was observed that activated ERK1 but not activated MEK1, which were isolated from H2O2-treated cells, was able to phosphorylate p66shcA (Figure 2E). Phosphorylation of p66shcA (S36A) was considerably less, indicating that Ser36 was a major site for ERK1. Furthermore, Western blot analysis with the specific anti-phospho-p66shcA antibodies confirmed that the Ser36 was indeed phosphorylated in the in vitro assay (Figure 2F). In the same settings, ERK5, another member of the MAPK family that can be activated by oxidative stress (Abe et al., 1996), failed to phosphorylate p66shcA at Ser36 (Figure 2F). The results indicate that p66shcA is a bona fide substrate for ERK1/2, but not for ERK5. In the p66shcA protein sequence, Ser36 is followed by Pro37, constituting a MAP kinase recognition motif (Kolch, 2000; Sharrocks et al., 2000).
ERK1 Formed a Complex with p66shcA
Having shown that in response to H2O2 treatment, ERK1/2 phosphorylated p66shcA on Ser36, we next questioned whether ERKs physically interact with p66shcA. To test whether ERK1 and p66shcA interact, we first used COS7 cells to do co-IP assays. COS7 cells were transfected with constructs expressing either HA-tagged ERK1, or ERK1 and myc-tagged p66shcA. Ectopically expressed p66shcA, along with its associated proteins, was precipitated with anti-myc antibodies, and fractionated onto an 8% SDS-PAGE gel. ERK1 was detected with anti-HA antibodies. We found that ERK1 (20% of expressed ERKs) coprecipitated with p66shcA, but was not pulled down when p66shcA was not expressed (Figure 3A). Treatment of COS7 cells with H2O2 for 10 min did not increase the formation of ShcA-ERK1 complex. Reciprocally, COS7 cells were transfected with DNA expressing p66shcA, or DNA expressing p66shcA and ERK1. ERK1, along with its associated proteins, was precipitated from cell lysate with anti-HA antibodies and fractionated on 8% SDS-PAGE gel. p66shcA was detected with anti-myc antibodies by Western blot. p66shcA (15% of the expressed p66shcA) was pulled down from the lysate by ERK1, but was not pulled down with anti-HA antibodies when ERK1 was not coexpressed (Figure 3B). Treatment of COS7 cells with H2O2 for 10 min slightly increased formation of ShcA-ERK1 complex. Together these results indicate that p66shcA interacts with ERK1 and formation of a complex may facilitate p66shcA phosphorylation by ERK1. Furthermore, co-IP assays confirmed the interaction between endogenous p66shcA and ERKs (Figure 3C), as p66shcA was precipitated by anti-ERK antibodies conjugated to agarose beads. Similarly, ERK1/2 could be precipitated with anti-p66shcA antibodies (Figure 3D). The fact that H2O2 was not found to increase ERK-p66ShcA complex formation consistently could be due to technical difficulties, especially with proteins like p66shcA that is expressed at very low levels and have many interacting partners (see below). Alternatively, it could be that the increase in complex formation is mediated by phosphorylation, which is removed quickly in the signaling cascade. For example, tyrosine phosphorylation of ShcA proteins reached the maximal level after 5 min of stimulation with H2O2 and started to decline after 10 min (see below).
|
Alignment of ShcA against the consensus MAPK docking domains revealed that ShcA proteins do have a docking site in the SH2 domain (479 LQGEPWFHGKLSRREAEALLQL 500; residues in bold are highly conserved). Although no canonical ERK binding site (Phe-Xaa-Phe-Pro) was found, the CH2 domain of p66shcA did contain a Phe-Phe-Promotif (24 aa downstream of Ser36). Together with Ser36-Pro37, it may constitute an ERK docking site (reviewed in Kolch, 2000; Sharrocks et al., 2000). To determine whether this putative docking site was required for the interaction between ERK and p66shcA, we generated two point mutations in the docking sites of p66shcA (Arg491Arg492 to AlaAla and ArgArg to GluGlu) and tested their influence on ERK-p66shcA interaction in co-IP experiments. It was found that these two mutants could still bind to ERK in a manner similar to that of the wild-type p66shcA (Figure 3E). Furthermore, we found that neither p52 nor p46shcA, which contains the docking site, was able to interact with ERK in the co-IP experiments, suggesting that this docking site is not important in this interaction (Figure 3F). Rather, these results indicate that the CH2 domain of p66shcA plays an essential role in this interaction.
|
Inhibition of ERKs Diminished FOXO3a Phosphorylation
p66shcA plays a proapoptotic role in response to oxidative stress, by modulating FOXO3a activity, a member of forkhead-related transcription factors (Nemoto and Finkel, 2002). The activity of FOXO3a is dependent on its Thr32 phosphorylation status: the nonphosphorylated form can enter the nucleus and is transcriptionally active, whereas the phosphorylated form remains in the cytoplasm. Furthermore, phosphorylation of FOXO proteins can also affect their DNA-binding activities and transactivation activities (Van Der Heide et al., 2004). We did observe Akt activation, FOXO3a phosphorylation at Thr32, and a significant translocation of FOXO3a from the nucleus to the cytoplasm in response to H2O2 (Figure 4A and Supplementary Data S2). It was also found that the majority of the FOXO3a proteins were already in the cytoplasm, likely due to the fact that the cells were serum-starved (Supplementary Data S2). Nonetheless, these results indicate that H2O2 treatment could inactivate FOXO3a via cytoplasm sequestration. Thr32 phosphorylation is a function of Akt kinase and requires Ser36 phosphorylation of p66shcA in response to oxidative stress, because H2O2-induced Thr32 phosphorylation of FOXO3a is dramatically reduced p66shcA-deficient cells (Nemoto and Finkel, 2002). Because ERK1/2 were upstream of p66shcA and responsible for Ser36 phosphorylation, we predicted that interference with ERK1/2 would also affect the Akt activity and FOXO3a phosphorylation. As expected, H2O2-induced activation of Akt1 was inhibited by U0126 in a dosage-dependent manner, and so was phosphorylation of FOXO3a on Thr32 (Figure 4B). These results further support the conclusion that ERK1/2 act as positive regulators of p66shcA in response to oxidative stress. Comparison of Akt activation and p66shcA phosphorylation (Figure 1B vs. 4A) revealed that p66shcA phosphorylation, but not Akt activation, declined in the presence of a higher concentration of H2O2. Activation of Akt at high concentrations of H2O2 may be mediated by a pathway independent of p66shcA, e.g., phosphoinositide-dependent kinase 1 (PDK1). It has been reported that PDK1 could be activated by H2O2 in fibroblasts via tyrosine phosphorylation, especially at higher concentrations (>1.0 mM; Prasad et al., 2000). Furthermore, inhibition of Akt activation by wortmannin abrogated the FOXO3a phosphorylation (unpublished data), confirming the observation that Akt is the major kinase responsible for FOXO phosphorylation in response to oxidative stress.
We also analyzed another member of the forkhead family, FOXO1, in response to H2O2 treatment. We found that phosphorylation of FOXO1 on Thr28 (corresponding to Thr32 of FOXO3a, which can be recognized by the same antibody) was barely detectable in presence of H2O2, suggesting that Thr28 phosphorylation of FOXO1 may not be involved in oxidative stress response (unpublished data). On the contrary, phosphorylation of Ser256 on FOXO1, also a function of Akt, was strongly stimulated by H2O2. Inhibition of ERK also diminished Ser256 phosphorylation in dosage-dependent manner (unpublished data).
Serum Stimulation Led to p66shcA Phosphorylation on Ser36 in an ERK-dependent Manner
Because serum stimulation also leads to ERK activation, we tested whether it induces p66shcA phosphorylation on Ser36. NIH3T3 cells were serum-starved for 20 h and thereafter stimulated with 10% serum for 10 min. Endogenous p66shcA was immunoprecipitated, and its phosphorylation on Ser36 was analyzed by Western blot with specific antibodies. We found that serum greatly induced Ser36 phosphorylation, which was blocked by ERK inhibition (Figure 4C), suggesting that ERKs are also responsible for serum-induced Ser36 phosphorylation. As expected, Akt activation was also observed and so was the phosphorylation of FOXO3a. However, inhibition of ERK activation showed no effect on Akt activation or FOXO3a phosphorylation, suggesting that the ERK-p66shcA-Akt pathway is not a main pathway in serum-induced Akt activation (Figure 4D). The results excluded the possibility that U0126 may somehow inhibit Akt activation in an unspecific manner.
H2O2 Treatment Down-regulated FOXO Target Gene p27
FOXO3a controls cell cycle, cell death, and oxidative stress response through transactivating different sets of genes (Tran et al., 2003). Expression profiles of a couple of known FOXO3a target genes such as MnSOD and catalase were tested in response to H2O2 treatment (Kops et al., 2002). In our experimental settings, MnSOD and catalase levels did not show any significant change after H2O2 exposure. Instead, H2O2 exposure at 0.2–0.5 mM from 2 to 20 h was able to reduce the protein levels of p27kip1 (Figure 5A and Supplementary Data S3), a cyclin-dependent kinase inhibitor and a target gene for FOXO3a (Dijkers et al., 2000; Medema et al., 2000; Stahl et al., 2002), whereas p21 and cyclin D was not affected (unpublished data). p27 has FOXO3a-binding sites in its promoter region, and its expression requires activation of FOXO3a (Tran et al., 2003). The reduction of p27 on H2O2 exposure could be a consequence of inactivation of FOXO3a through the p66shcA-Akt pathway because it was accompanied by a translocation of FOXO3a from the nucleus to the cytoplasm (Supplementary Data S2). More interestingly, inhibition of ERK activation, which also compromised Akt activation and inactivation of FOXO3a, was able to suppress the down-regulation of p27 (Figure 5A). RT-PCR assays confirmed that H2O2 treatment down-regulated p27 mRNA levels in several cell types tested, including NIH3T3 and MEFs, and that inhibition of ERKs abolished the down-regulation (Figure 5B). These results indicate that H2O2 treatment also leads to a decrease in the levels of a CDK inhibitor and might facilitate the cellular events involving p27 and that ERK activation negatively regulates p27 expression at the mRNA levels. More importantly, MEFs lacking p66shcA did not show this down-regulation of p27 (Figure 5C), which was restored by reconstitution of p66shcA in the knockout cells (Figure 5D). But the MEFs expressing p66shcA (S36A) behaved like the knockout mutant (Figure 5D). These results indicate that p27 down-regulation is also dependent on the phosphorylation of Ser36 in p66shcA, similar to the activation of Akt and phosphorylation of FOXO3a in response to H2O2 (Nemoto and Finkel, 2002).
|
; Mikkelsen and Wardman, 2003) prompted us to test whether H2O2 could support cell proliferation in our experimental settings. Serum-starved cells were treated with increasing concentrations of H2O2 for 20 h (or 2 h), and their cell cycle profiles were analyzed by FACS analysis. Only a marginal increase was observed in the percentage of cells in G2/M phase (unpublished data). Bromodeoxyuridine (BrdU) incorporation assays did not reveal an increase in the BrdU-positive cells in the presence of H2O2, suggesting that the mitogenic effect is minimal. One interpretation for failure to detect a mitogenic effect for H2O2 is because H2O2 also has a proapoptotic effect. Therefore, the cells have to balance life and death under this condition, and the ultimate cellular response to H2O2 is apoptosis in the experimental settings. In fact, p27 is suitable for such a role because it not only regulates cell cycle progression but also apoptosis (Coqueret, 2003). p27 can be proapoptotic or antiapoptotic, depending on the cell types used and the factors used for apoptosis induction (Katayose et al., 1997; Hiromura et al., 1999; Dijkers et al., 2000). To study the role for p27 in oxidative stress–induced apoptosis, MEFs were prepared from p27-deficient and control wild-type mice. These primary MEFs were challenged with different concentrations of H2O2 for 5 h, and the cell death rates and cell survival rates were measured by trypan blue exclusion method and a cell proliferation/survival kit, respectively. It was found that p27-/- MEFs survived better than control wild-type cells in both assays (Figure 5E, unpublished data for trypan blue exclusion method). These results indicate that p27 possesses a proapoptotic function in oxidative stress response. Therefore, down-regulation of p27 by H2O2 might normally protect the cells against oxidative stress. A similar role for p27 and FOXO3a has been found in cytokine-mediated survival of hematopoietic cells (Dijkers et al., 2000).
Tyrosine Phosphorylation of p52/46shcA and Their Role in H2O2-induced ERK Activation
It is known that ShcA proteins regulate ERK activation in response to growth factors. Having shown that ERK could phosphorylate p66shcA in the presence of H2O2, we intended to study whether ShcA proteins have an influence on ERK activation in cell response to oxidative stress. During the studies of Ser36 phosphorylation, it was found that H2O2 also induced tyrosine phosphorylation of p66shcA, as well as p52 and p46shcA (Figure 6A). For all three isoforms, tyrosine phosphorylation was detectable at 0.2 mM and reached the maximal level at 1.0 mM H2O2. A time course study revealed that tyrosine phosphorylation of p66shcA occurred at 5 min after treatment with 0.5 mM H2O2 (Figure 6B), indicating that tyrosine phosphorylation precedes Ser36 phosphorylation (Figure 1C). Hence, H2O2 treatment leads to successive phosphorylation of all three ShcA proteins on tyrosine residues and of p66shcA on Ser36 in murine fibroblasts.
|
; Blake et al., 2000). We then made point mutations at Y239/240/F, Y317/F, and Y239/240/317F by site-directed mutagenesis in p52shcA. NIH3T3 cells were transfected with DNA expressing the mutant p52shcA and subsequently treated with H2O2 for 10 min. Tyrosine phosphorylation of these proteins was analyzed by Western blot using anti-phospho-tyrosine antibodies (Figure 6C). Neither Y239/240/F nor Y317/F mutations significantly affected the phosphorylation, whereas triple mutations completely abolished tyrosine phosphorylation, indicating that all three sites were phosphorylated upon H2O2 treatment. We then swapped the triple mutations into p66shcA and found that they were the only tyrosine residues phosphorylated by H2O2 treatment (Figure 6D). We showed, for the first time, that H2O2 treatment led to phosphorylation of all three tyrosine residues in all three ShcA isoforms. Although Src kinase and PDGFR could phosphorylate these residues, we found that pretreated NIH3T3 with inhibitors for Src, PDGFR, or a combination of the two did not significantly affected H2O2-induced tyrosine phosphorylation of ShcA proteins (unpublished data). Furthermore, Src-/- MEF cells showed normal tyrosine phosphorylation of all three ShcA isoforms even in the presence of PDGFR inhibitor (unpublished data), suggesting that neither Src nor PDGFR is required for ShcA tyrosine phosphorylation or that there exist functionally redundant tyrosine kinases in the cell.
Tyrosine phosphorylation has been involved in growth factor induced ERK activation. Like growth factors, H2O2 can activate ERK1/2 in a dosage-dependent manner (Rao, 1996). Similarly, we found that in NIH3T3 cells, 0.25 mM of H2O2 could significantly activate ERK1/2 and this was accompanied by increased complex formation between ShcA and Grb2 (unpublished data). It has been reported that Src plays important roles in H2O2-induced ERK activation (Aikawa et al., 1997). We tested this conclusion using the Src-/- and control MEFs and found that Src was not required for ERK activation (unpublished data), suggesting that redundant kinases may indeed exist in the cell. The results are in agreement with our finding that Src deficiency did not alter ShcA tyrosine phosphorylation.
To determine whether tyrosine phosphorylation of ShcA plays a role in ERK activation, NIH3T3 cells were transfected with HA-tagged ERK1, in combination with p52shcA or p52shcA (Y239/240/317F), serum-starved for 20 h, and treated with 0.5 mM H2O2 for 10 min. ERK1 was immunoprecipitated using polyclonal anti-HA antibodies, and activated ERK1 was detected with anti-phosphorylated ERK antibodies by Western blot. Expression of normal p52shcA was able to enhance activation of ERK1 in response to H2O2 (Figure 6, E and F). On the other hand, expression of the mutant p52shcA (Y239/240/371F) failed to do so (Figure 6, E and F), suggesting that tyrosine phosphorylation of ShcA is necessary for activation of ERKs in response to H2O2. The triple mutations (Y239/240/371F) should not affect the folding of ShcA proteins, because p66shcA carrying the triple mutations was similarly phosphorylated at Ser36 upon H2O2 treatment and that p66shcA carrying these mutations still possessed its ability to physically interact with its partner p52shcA (Supplementary Data S3 and S4). To further confirm that ShcA proteins are involved in ERK activation in response to H2O2, ShcA-/- fibroblasts were challenged with different concentrations of H2O2, and ERK activation was assessed by Western blot analysis. It was found that ShcA-/- fibroblasts showed compromised ERK activation (Figure 6G), although the basal level of activated ERKs was slightly higher in the ShcA-deficient cells. A similar defect in ERK activation in response to low concentrations of growth factors has been reported in ShcA-/- fibroblasts (Lai and Pawson, 2000). Taken together, our data indicate that H2O2 treatment results in tyrosine phosphorylation of p52/46shcA, which in turn facilitates H2O2-induced ERK activation.
Phosphorylation of Ser36 Played a Negative Role in H2O2-induced ERK Activation
Having shown that ERK interacted with p66shcA and phosphorylated p66shcA on Ser36 and that p52/46shcA facilitated H2O2-induced ERK activation, we wanted to determine whether Ser36 phosphorylation of p66shcA played a role in H2O2-induced ERK activation. NIH3T3 cells were cotransfected with ERK1 and p66shcA or p66shcA (S36A), serum-starved for 20 h, and stimulated with H2O2 for 10 min. ERK1 was precipitated with anti-HA antibodies conjugated to agarose beads, and its activation was determined by Western blot analysis. Unlike p52/46shcA, which strongly facilitate ERK activation, p66shcA was found to suppress activation of ERK (Figure 6H), consistent with the findings that p66shcA plays an inhibitory role in ERK activation stimulated by EGF (Migliaccio et al., 1997; Okada et al., 1997). Surprisingly, mutant p66shcA (S36A) was found to facilitate ERK activation, although only to a mild extent (Figure 6H), suggesting that the normal function of p66shcA is to negatively regulate ERK activation and that phosphorylation of Ser36 may play an important role. To confirm this conclusion, we overexpressed p66shcA, or p66shcA(S36A) in COS7 cells, which were subjected to stimulation with 0.5 mM H2O2 for different periods of time and the activation of endogenous ERKs were detected by Western blot. It was found that p66shcA inhibited H2O2-induced ERK activation (Figure 6I), whereas expression of S36A mutant led to a slightly enhanced activation of ERK1/2.
Complex Formation among ShcA Proteins
Having shown that H2O2-induced ERK activation could be facilitated by p52/46shcA but hampered by p66shcA, we wanted to study the molecular basis for their different actions. p66shcA contains an extra 110 amino acid CH2 domain compared with p52shcA and is a minor isoform in fibroblasts. We first used pulldown assays to test whether they exist as complexes in the cell. A GST-p52shcA fusion protein was made and used in pulldown experiments against p46shcA expressed in COS7 cells. Although GST-p52shcA was able to precipitate p46shcA, GST itself failed to do so (Figure 7A). Similarly, GST-p46shcA was made and used in pulldown experiments against p52shcA. COS7 cells were transfected with p52shcA, which expressed both p52shcA and p46shcA due to two different start codons being used. We found that GST-p46 was able to precipitate both p52shcA and p46shcA, suggesting that p52-p46 and p46-p46 dimers may exist (Figure 7B). Similar experiments were carried out using GST-p46shcA fusion protein to precipitate CH1, PTB, or SH2 domains alone, and it was found that PTB and CH1 domains were able to bind p46shcA (unpublished data), suggesting that those two domains are involved in complex formation between ShcA proteins. Because p66shcA contains PTB and CH1 domains, we believe that p66shcA could also form complexes with p52shcA or p46shcA.
|
Interaction between p46 and p52 was confirmed by co-IP experiments. We found that p52shcA was able to immunoprecipitate coexpressed p46shcA, whereas p46 could not be precipitated when p52 was not coexpressed (Figure 7C). H2O2 treatment did not significantly promote p52-p46 interaction. Reciprocally, p46shcA was able to precipitate both p52shcA and p46shcA (expressed from the same construct; Figure 7D), and H2O2 treatment did not significantly enhance the complex formation. These results suggest that ShcA proteins do interact in vivo.
As mentioned above, p66shcA and p52/46shcA play distinct roles in ERK activation induced by growth factors or H2O2 (Figure 6 and Migliaccio et al., 1997). To test whether p66shcA interact with p52 or p46, we transfected COS7 cells with myc-tagged p66 and HA-tagged p52. One set was stimulated with H2O2, whereas the other set was left untreated. p66 and associated proteins were precipitated with anti-myc antibodies and fractionated onto an 8% SDS-PAGE gel. p52 was detected with anti-HA antibody on a Western blot (Figure 8A). Dimers of p66-p52 were found even when cells were not treated, and H2O2 treatment modestly enhanced this dimerization. Similar results were obtained for p66-p46 interaction (Figure 8B). Because p66 and p52/46 have different functions in ERK activation, increased dimer formation between p66shcA and p52shcA or p46shcA in response to H2O2 may help the cells to attenuate signal transmission. This is consistent with our findings that tyrosine phosphorylation of p52 and p46 precedes Ser36 phosphorylation of p66.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
), our results demonstrated for the first time that p66shcA (Ser36) is a substrate of ERKs. p66shcA has a perfect MAPK recognition motif (Ser36-Pro37) in the CH2 domain. We demonstrated that p66shcA and ERK do interact in vitro and in vivo. In vitro kinase assays confirmed that H2O2-activated ERKs phosphorylated p66shcA with Ser36 being a major site. Furthermore, ERK activation is necessary and sufficient for H2O2-induced Ser36 phosphorylation. This conclusion is further supported by the findings that inhibition of ERK activation also impeded p66shcA downstream signaling molecules such as Akt and downstream transcription factor FOXO3a in a signaling pathway triggered by H2O2 (Nemoto and Finkel, 2002). Even though serum feeding also resulted in Ser36 phosphorylation in an ERK-dependent manner, inhibition of ERKs, and therefore inhibition of p66shcA phosphorylation, did not affect Akt activation or FOXO3a phosphorylation induced by serum stimulation, suggesting that the p66shcA-Akt-FOXO3a pathway is specific to H2O2, but not to growth factors. One possible explanation is that serum-induced Akt activation is mainly through RTK-PI-3 kinase pathway, whereas H2O2-induced Akt activation is mainly through p66shcA.
What cellular events does the p66shcA-Akt-FOXO pathway regulate in response to oxidative stress? Reactive oxygen species have been documented to have a proapoptotic role and under certain conditions, a mitogenic role. Meanwhile, ROS can also activate the antioxidant defense system to protect the cells. Accumulating evidence suggests that some of these effects of ROS are mediated by the FOXO transcription factors, which are known to control cell cycle, cell death, and stress detoxification by regulating transcription of different sets of genes under various conditions (Tran et al., 2003). For example, H2O2 has been reported to down-regulate genes involved in both H2O2 scavenging and oxidative stress resistance, e.g., catalase and MnSOD, by inactivating FOXO3a through the p66shcA-Akt pathway (Nemoto and Finkel, 2002; Trinei et al., 2002). In the absence of p66shcA, this pathway is disrupted and as a consequence, expression of these antioxidant proteins is maintained and p66shcA-deficient cells exhibit resistance to oxidative stress (Migliaccio et al., 1999; Nemoto and Finkel, 2002; Trinei et al., 2002). In this regard, the action of the p66shcA-Akt-FOXO pathway is to transduce proapoptotic signals. In contrast, FOXO proteins also control expression of genes involved in apoptotic induction such as Fas ligand and the Bcl2 family member Bim. Activation of Akt and phosphorylation of FOXO would lead to down-regulation of these two proteins, facilitating cell survival. Furthermore, activation of Akt, a well-studied mediator of survival, can lead to inactivation of caspase 9 and the proapoptotic Bcl2 family member Bad, in a transcription-independent manner. It appears that the Akt-FOXO pathway also transduces antiapoptotic signals. In addition, oxidative stress also activates many other signaling pathways, including the protein kinase C (PKC) 
-Nrf2 pathway, which up-regulates the levels of antioxidant proteins such as peroxiredoxin I, promoting cell survival (Li et al., 2002, 2004). Therefore, all these signals would be integrated to influence the decision making as to cell death or survival in response to oxidative stress.
We discovered that p27kip1, a well-studied target gene of FOXO3a, was down-regulated by H2O2, in correlation to inactivation of FOXO3a. Furthermore, H2O2-induced Ser36 phosphorylation, Akt activation, as well as down-regulation of p27 could be suppressed by inhibition of ERK activation. This is the first time that a connection between oxidative stress and p27 expression was established. What is the biological significance of reduced expression of p27 in response to oxidative stress? p27 is a CDK inhibitor and is a negative regulator of cell cycle progression. The levels of p27 are decreased in many human cancers and growth factor–stimulated cell proliferation is associated with down-regulation of p27. In addition, previous studies showed that ectopic expression of activated FOXO3a led to cell cycle arrest in which induction of p27 through FOXO3a appeared to play an important role (Medema et al., 2000). We propose that H2O2-induced down-regulation of p27 may facilitate cell cycle entry, consistent with the finding that ROS can have a mitogenic role under certain conditions (Li et al., 1997; Reth, 2002) and the recent finding that p66shcA is actually a negative regulator of cell cycle progression (Pacini et al., 2004). Yet, H2O2 alone was not sufficient to promote cells to enter cell cycle, because we could not detect a marked increase in proliferating cells under stimulation of a low dose of H2O2. However, we could not exclude the possibility that H2O2-induced p27 reduction may facilitate cell cycle progression under proper conditions, for example, when apoptosis is blocked or other mitogenic signals are provided. Recent studies also suggest that p27 may participate in apoptosis. Overexpression of p27 induces apoptosis in several cell types and p27 deficiency protects cells from stress-induced apoptosis, indicating a proapoptotic role for p27 (Dijkers et al., 2000). In contrast, there are also reports suggesting an antiapoptotic role for p27 (reviewed in Coqueret, 2003). It appears that p27 can have different functions in apoptosis, dependent on the cell types and stress types (Coqueret, 2003). Studies of p27-/- MEFs in comparison to wild-type control cells revealed that p27 has a proapoptotic role in response to oxidative stress, as p27-/- MEFs showed improved survival against H2O2. Therefore, H2O2-induced down-regulation of p27 may provide a protective mechanism in oxidative stress response. The modest effect of p27 supports the concept that the end result of oxidative stress is determined by a coordinated action of various pathways. Taken together, we found that H2O2 led to down-regulation of p27, which may participate in cell response to oxidative stress.
Our studies provided evidence that p52/46shcA are also involved in oxidative stress response. They participated in H2O2-induced activation of ERKs. It has long been known that H2O2 activates ERK but the molecular mechanisms are less clear. We found that oxidative stress led to phosphorylation of ShcA proteins on all three tyrosine residues (239/240/371), the complex formation between ShcA and Grb2, and activation of ERK. Mutant p52shcA (Y239/240/371 to Phe) inhibited H2O2-induced ERK activation, and cells deficient for ShcA showed diminished ERK activation. Thus, H2O2-induced activation of ERKs follows the same pathway triggered by growth factors. The question remains as to what are the tyrosine kinases activated by H2O2 and capable of phosphorylating ShcA proteins at all three residues. Although previous studies have shown that H2O2 treatment activates growth factor receptors such as EGFR and PDGFR, probably by suppressing protein tyrosine phosphatases (Knebel et al., 1996; Kamata and Hirata, 1999), PDGFR was found unnecessary for ShcA phosphorylation or ERK activation. Other mechanisms have also been suggested by which H2O2 activates ERK in different settings (Guyton et al., 1996). Studies in cardiomyocyte showed that H2O2 directly targeted G
i and Go, which led to dissociation of G
and subsequent activation of ERK (Nishida et al., 2000). In these cells, activation of ERK required Src activation (Nishida et al., 2000), but we found that deficiency of Src did not affect phosphorylation of ShcA on tyrosine residues, Akt activation, or ERK1/2 activation induced by H2O2 in fibroblasts. The discrepancy could be due to the ways by which Src is inhibited. In previous studies, either an inhibitor or a dominant negative form of Src was used, whereas in our studies, Src-/- MEFs were used. It is possible that dominant negative Src may interfere with the function of Src homologues such as Yes and Lyn.
Our findings also suggest that p66shcA played a negative role in H2O2-induced ERK activation, in which Ser36 phosphorylation is critical, because mutant p66shcA (S36A) showed some stimulatory effect. A negative role for p66shcA ERK activation has been previously reported in response to growth factor (Migliaccio et al., 1997; Okada et al., 1997). Hence, all the three ShcA isoforms participate in p66shcA phosphorylation: p52/46shcA facilitates ERK activation; activated ERK in turn phosphorylates p66shcA, leading to FOXO3a phosphorylation. Therefore, in this event, they functioned in a cooperative way. Later, the Ser36-phosphorylated p66shcA may block ERK activation as a feedback regulation. In this regard, ShcA proteins clearly have distinct functions. What is the molecular basis for the isoform specific effects? How do ShcA isoforms achieve their functions? We believe that our studies may provide some hints to the above questions. We found that ShcA isoforms interact with each other and may exist as dimers or multimers in the cell. H2O2 treatment promotes the association between p66shcA and p52shcA or p46shcA, but not between p52 and p46. Because p52shcA and p46shcA are more abundant than p66shcA in fibroblasts, they may exist as dimers in the cell and facilitate ERK activation in response to H2O2. The switch to p66-p52 or p66-p46 may act to interfere with p52/46shcA function and thereby attenuate ERK activation. In accordance with this, it was found that ectopic expression of p66shcA in ShcA-/- MEFs did not inhibit ERK activation (Hu and Li, unpublished data), suggesting that the inhibitory effect of p66shcA on ERK activation was via p52/46ShcA. Alternatively, it is possible that the interaction among ShcA isoforms may help to assemble upstream and downstream components, e.g., protein phosphatase 2A, MAPKAP kinase2, SHIP, Grb2 (all interact with ShcA), into an organized signaling complex, enhancing the efficiency of signaling propagation (Lamkin et al., 1997; Ugi et al., 2002; Yannoni et al., 2004). In this regard, the complex may act as an anchoring scaffold to facilitate signal transmission (Burack and Shaw, 2000). Certainly, the exact function of the complex formation among ShcA proteins warrants further investigation.
The major finding of this study is that phosphorylation of p66shcA on Ser36 serves two purposes in cell response to oxidative stress: transmitting signals to modulate p27 expression to influence cell response to oxidative stress; and inhibiting ERK activation through a negative feedback mechanism. In response to H2O2, all three ShcA proteins participate in transmitting signals to inactivate transcription factor FOXO3a, in two steps: H2O2 first induces phosphorylation of tyrosine residues of ShcA proteins, facilitating ERK activation; and activated ERKs in turn phosphorylate p66shcA on Ser36, leading to phosphorylation of FOXO3a by Akt and down-regulation of p27. p66shcA may inhibit ERK activation by forming complexes with p52shcA or p46shcA, the stimulatory ShcA proteins for ERK activation.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METH |
|---|
) led to massive phosphorylation of p66shcA on Ser36. The mutant form of MEK1
