Cleavage of GSK-3β by calpain counteracts the inhibitory effect of Ser9 phosphorylation on GSK-3β activity induced by H2O2

Authors

  • Ye Feng,

    1. Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
    • These authors contributed equally to this work.
  • Yiyuan Xia,

    1. Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
    • These authors contributed equally to this work.
  • Guang Yu,

    1. Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
  • Xiji Shu,

    1. Department of Pathology & Pathophysiology, School of Medicine, Jianghan University, Wuhan, China
    Search for more papers by this author
  • Haoliang Ge,

    1. Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
  • Kuan Zeng,

    1. Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
  • Jianzhi Wang,

    1. Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
  • Xiaochuan Wang

    Corresponding author
    • Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author

Address correspondence and reprint requests to Xiaochuan Wang, Department of Pathophysiology, Key Laboratory of Neurological Disease of National Education Ministry, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, China. E-mail: wxch@mails.tjmu.edu.cn

Abstract

Glycogen synthase kinase-3 beta (GSK-3β) dysfunction may play an essential role in the pathogenesis of psychiatric, metabolic, neurodegenerative diseases, in which oxidative stress exists concurrently. Some studies have shown that GSK-3β activity is up-regulated under oxidative stress. This study evaluated how oxidative stress regulates GSK-3β activity in human embryonic kidney 293 (HEK293)/Tau cells treated with hydrogen peroxide (H2O2). Here, we show that H2O2 induced an obvious increase of GSK-3β activity. Surprisingly, H2O2 dramatically increased phosphorylation of GSK-3β at Ser9, an inactive form of GSK-3β,while there were no changes of phosphorylation of GSK-3β at Tyr216. Moreover, H2O2 led to a transient [Ca2+]i elevation, and simultaneously increased the truncation of GSK-3β into two fragments of 40 kDa and 30 kDa, whereas inhibition of calpain decreased the truncation and recovered the activity of GSK-3β. Furthermore, tau was hyperphosphorylated at Ser396, Ser404, and Thr231, three most common GSK-3β targeted sites after 100 μM H2O2 administration in HEK293/Tau cells, whereas inhibition of calpain blocked the tau phosphorylation. In addition, we found that there were no obvious changes of Cyclin-dependent kinase 5 (CDK5) expression (responsible for tau phosphorylation) and of p35 cleavage, the regulatory subunit of CDK5 in H2O2-treated HEK293/Tau cells. In conclusion, Ca2+-dependent calpain activation leads to GSK-3β truncation, which counteracts the inhibitory effect of Ser9 phosphorylation, up-regulates GSK-3β activity, and phosphorylates tau in H2O2-treated HEK293/Tau cells.

Abbreviations used
AD

Alzheimer's disease

[Ca2+]i

intracellular calcium concentration

GSK-3β

glycogen synthase kinase-3 beta

H2O2

hydrogen peroxide

MDA

malondialdehyde

SOD

superoxide dismutase

Glycogen synthase kinase-3 beta (GSK-3β) – also known as human tau protein kinase (Ishiguro et al. 1992) – is a multifunctional serine/threonine protein kinase that was originally identified as a regulator of glycogen metabolism. Previous studies have shown that GSK-3β is also involved in embryonic development, microtubule dynamics, protein synthesis, cell cycle, cell death, oncogenesis as well as inflammation (Embi et al. 1980; Frame and Cohen 2001; Grimes and Jope 2001; Doble and Woodgett 2003). Therefore, it is now recognized as a critical component and central figure in numerous cellular signaling pathways with a plethora of cellular targets (Jope 2004; Jope and Johnson 2004). GSK-3 dysfunction may also be associated with certain psychiatric diseases such as bipolar mood disorders and schizophrenia (Jope and Roh 2006), metabolic diseases such as type II diabetes (Planel et al. 2002; Henriksen and Dokken 2006), and neurodegenerative diseases such as Parkinson's disease and Alzheimer's disease (AD). Taking into account GSK-3β contribution to AD, it is up-regulated and associated with Alzheimer neurofibrillary degeneration of abnormally hyperphosphorylated tau, also promotes β-amyloidosis through increased amyloidogenic processing of β-amyloid precursor protein (Dugo et al. 2007).

There are two forms of mammalian GSK-3, GSK-3α, and GSK-3β, and they are encoded by separate genes (Woodgett 1990). These two isoforms share high homology in their catalytic domains, but significantly differ in their N- and C-terminal parts that could play a role in the regulation of GSK3 activities. Despite their similar biochemical and substrate properties, GSK-3α and GSK-3β are not always functionally identical (Goode et al. 1992; Wang et al. 1994a; Hoeflich et al. 2000; Phiel et al. 2003; Su et al. 2004; Liang and Chuang 2007). Besides the difference in N and C-termini of the two isoforms, differential expression of GSK-3α and GSK-3β in different tissues and their subcellular localizations have also been demonstrated. Dynamic regulation of GSK-3 kinase activity through modulation of serine/tyrosine phosphorylation in neurons at different developmental or maturation stages is thus possibly mediated via diverse spatial or temporal distribution of these two GSK-3 isoforms. Yet, their precise roles in different tissues are not fully understood.

GSK-3β, the smaller of the two proteins, is a single polypeptide of 482 amino acids with a molecular weight of 47-kDa, and contains a central protein kinase catalytic domain (Woodgett 1991). GSK-3β is subject to multiple regulatory mechanisms. Among those regulatory pathways, phosphorylation at tyrosine-216 in GSK-3β enhances the enzymatic activity of GSK-3β, while phosphorylation at serine-9 in GSK-3β decreases active site availability (Jope et al. 2007). Therefore, phosphorylation of GSK-3β endows itself with dual specificity, capable of autophosphorylating its Ser, Thr, and Tyr residues and inducing its net inhibition by a three to fivefold reduction in enzyme activity (Wang et al. 1994b; Planel et al. 2002), probably via autophosphorylation of Ser9. Besides, Wnt stimulation acts on GSK-3β in a multiprotein complex that includes axin, adenomatous polyposis coli-associated protein, and β-catenin (Harwood 2001). GSK-3β phosphorylates all three of these proteins and the phosphorylation of β-catenin leads to its degradation. When Wnt proteins bind to the frizzled receptor, the dishevelled protein is activated, which in turn inhibits GSK-3 activity by disrupting this multi-protein complex. As a consequence, inhibition of GSK-3 allows the accumulation of β-catenin which translocates into the nucleus in conjunction with transcription factors to control gene expression. Recent studies showed a novel mechanism of GSK-3 regulation in which N-terminal cleavage of GSK-3 by calpain, a calcium-dependent, non-lysosomal cysteine protease expressed ubiquitously in mammals and many other organisms, leads to the loss of its inhibitory domain, producing two fragments with kinase activity (Goñi-Oliver et al. 2007, 2011). Strong oxidative stresses usually induce calcium overload, and in turn resulting in calpain activation. Taken together, regulation of GSK-3β activity differs from different conditions.

Previously, we reported that calyculin A, a strong serine/threonine protein phosphatase inhibitors, while inhibits protein phosphatase 2 (PP2A) activity, also induces oxidative stress and an increase in GSK-3β activity; melatonin, vitamin E, and berberine all can exert antioxidative effects and rescue GSK-3β activity in human embryonic kidney 293 (HEK293)/tau cells and N2a cells (Yu et al. 2011). However, the exact mechanism of the regulation of GSK-3β activity under oxidative stress is still not brought forth. In this study, we employed H2O2 to create an oxidative stress condition, and found that the GSK-3β activity was increased in H2O2-treated HEK293/Tau cells. Surprisingly, H2O2 dramatically increased phosphorylation of GSK-3β at Ser9, an inactive form of GSK-3β, while there were no obvious changes of phosphorylation of GSK-3β at Tyr216. Intracellular Ca2+ [Ca2+]i of HEK293/Tau cells was increased after exposure to H2O2, and western blotting results showed GSK-3β was cleaved into two fragments. The specific calpain inhibitor IV rescued the truncation of GSK-3β and the GSK-3β activity was inhibited accordingly. Taken together, our data suggest that H2O2 increases total GSK-3β activity through calpain activation caused by calcium influx, which counteracts the inhibitory effect of phosphorylation of GSK-3β at serine9 induced by H2O2.

Materials and methods

Cell culture

HEK293 cell lines stably transfected with the longest human tau cDNA (HEK293/Tau) were cultured in 90% Dulbecco's modified Eagle's medium (DMEM) in the presence of 1 μg G418 (Invitrogen Corporation, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum (Invitrogen Corporation) at 37°C in 5% CO2 and 95% air in a humidified incubator. The medium was changed every other day, and cells were plated at an appropriate density according to each experimental scale. Prior to experimentation, cells were grown to approximately 80% confluence on six-well culture plates for cell lysates preparation or on 96-well culture plates for measurement of cell viability.

H2O2 administration and measurement of cell viability by cell counting kit-8

Cells were grown on 96-well culture plates at a density of 5000 cells in 100 μL, and then the cells were exposed to various concentrations (0, 50, 100, 250, 500, or 1000 μM) of H2O2 for 2 h at 37°C. Then, 10 μL/well CCK-8 (Dojindo Laboratories, Kumamoto, Japan) solution was added and the cells were incubated for 3–4 h in the dark, and the absorbance at 450 nm was read. Assays were repeated independently for five times.

Measurement of superoxide dismutase (SOD) and malondialdehyde (MDA)

Cells were cultured in six-well plate, after exposure to 100 μM H2O2 for 2 h, then cells were washed down from the wells with phosphate-buffered saline (pH 7.4) and lysed with buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM sodium chloride, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 0.02% sodium azide, 100 μg/mL phenylmethylsulfonyl fluoride, and 10 μg/mL protease inhibitors (leupeptin, aprotinin, and pepstatin) for at least 15 min on ice. Then, cells were centrifuged at 10 000 g for 15 min at 4°C. Total SOD activity in the supernatant was determined with a commercial kit from Jiancheng Bioengineering Institute (Central Road, Nanjing, China), based on its ability to inhibit the oxidation of hydroxylamine by the xanthine–xanthine oxidase system. The mauve product (nitrite) produced by the oxidation of hydroxylamine has an absorbance at 550 nm. One unit of SOD activity was defined as the amount that reduced the absorbance at 550 nm by 50%. Malondialdehyde, a metabolite of lipid peroxides, was used as an indicator of lipid peroxidation. MDA was determined with a commercial kit from Jiancheng Bioengineering Institute by measuring the malondialdehyde formed by the thiobarbituric acid reaction. N-butyl alcohol was used for extraction during this process. 4 mL N-butyl alcohol was added into each tube of mixed solution prepared according to the instruction, and then we centrifuged the new mixed solution at 3000 g for 10 min, put it at 4°C overnight, and detected the supernatant at excitation wavelength of 485 nm and emission wavelength of 528 nm the next day.

GSK-3β activity assay

Activity of GSK-3 was measured using GENMED GSK-3β Activity Assay Kit (GENMED Scientifics INC., Arlington, MA, USA), based on its ability to phosphorylate its targeting sequence GPHRSTPESRAAV in the presence of ATP and GSK-3α inhibitor Aloisine A, and then through a series of reactions involving pyruvate kinase and lactate dehydrogenase; the phosphorylated products in the system transform NADH (reduced nicotinamide adenine dinucleotide) to NAD (nicotinamide adenine dinucleotide), accordingly, detect the absorbance at 340 nm to evaluate the specific activity of GSK-3β.

Cell lysates preparation

Cells were rinsed twice in ice-cold phosphate-buffered saline (PBS, pH 7.4) and lysed with buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM sodium chloride, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 0.02% sodium azide, 100 μg/mL phenylmethylsulfonyl fluoride, and 10 μg/mL protease inhibitors (leupeptin, aprotinin, and pepstatin) followed by sonication for 15 min on ice. After centrifugation at 10 000 g for 5 min at 4°C, supernatants were fetched out and added to one third volume of 4× sample buffer [200 mM Tris-HCl, pH 6.8, 8% SDS, 40% glycerol, 10% β-mercaptoethanol (ME), and 0.05% bromophenol blue], boiled for 10 min, and then stored at −20°C or used immediately. The concentration of cell lysates was measured by BCA Protein Assay kit (Thermo Scientific, Pierce protein biology products, Rockford, IL, USA).

Western blotting

Total cell lysates of HEK293/Tau cells were harvested from each treatment. Primary rabbit monoclonal antibody p35/p25, mouse polyclonal antibody phospho-GSK-3β (Tyr 216), mouse monoclonal antibody DM1A (alpha Tubulin antibody), and rabbit polyclonal antibodies including phospho-GSK-3β (Ser9), phospho-AKT (Thr308), phospho-AKT (Ser473), and Fyn were obtained from Cell Signaling Technology Inc (Beverly, MA, USA); rabbit polyclonal antibody GSK-3β (Ab-216, #21301) was purchased from Signalway Antibody LLC (College Park, MD, USA); rabbit polyclonal antibodies including phospho-tau (Ser396), phospho-tau (Ser404), and phospho-tau (Thr231) were obtained from Biosource (Camarillo, CA, USA); and mouse monoclonal antibody Cyclin-dependent kinase 5 (CDK5) was obtained from Abcam (Cambridge, MA, USA). The dilution of all the primary antibodies mentioned above used in this study was 1 : 1000. The specific fluorescent secondary antibodies for Odessy imaging system use were purchased from LI-COR IRDye (Lincoln, NE, USA), and the dilution ratio was 1 : 10 000. The blots were scanned by Odessy and computer image analysis software (Image-Pro Plus; Media Cybernetics, Silver Spring, MD, USA) was used for quantitative analysis of the blots.

Measurement of intracellular calcium concentration

Fura-2-acetoxymethyl ester (Fura-2-AM) is a membrane-permeable derivative of the ratiometric calcium indicator Fura-2 used in biochemistry to measure intracellular calcium concentrations [Ca2+]i by fluorescence. In this study, cells were grown to approximately 70% confluence on round cover glasses in Petri dish, then cells were washed three times with standard extra cellular fluid (pH 7.4) containing 140 mM NaCl, 5.4 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 33 mM glucose, and 20 mM HEPES. Cells were then incubated with final concentration of 5 μM Fura-2-AM (Dojindo Laboratories) dissolved in dimethyl sulphoxide (DMSO) in the dark for 30 min, and washed cells five times with standard extra cellular fluid mentioned above in the dark, and then using I× 71 confocal microscopy (Olympus, Tokyo, Japan) and analysis software Invivo-IPA to measure the intensity of fluorescence of 32 randomly chosen cells which showed clear cell configuration before and after 100 μM H2O2 treatment, record the curves of changes and calculate the ratio of fluorescence intensity at absorbance 340 nm and 380 nm (Hu et al. 2002; Zhu et al. 2008).

Cell treatment with calpain inhibitor IV

Cells were cultured in 90% DMEM in the presence of 1 μg G418 (Invitrogen Corporation) supplemented with 10% fetal bovine serum (Invitrogen Corporation). The cells were divided into three groups, the blank control cells, cells exposed to 100 μM H2O2 for 2, 3, 5, 15, 30, and 120 min, and cells treated with both 100 μMH2O2 and 10 μM calpain inhibitor IV (Calbiochem, Darmstadt, Germany) simultaneously at the above time points at 37°C in 5% CO2 and 95% air in a humidified incubator, and then prepared cell lysates for western blotting analysis.

Statistical analysis

Data were expressed as mean ± SD and analyzed using SPSS 18.0 statistical software (SPSS Inc, Chicago, IL, USA). Statistical significance was determined by one-way anova procedure followed by Student–Newman–Keuls post hoc test with 95% confidence and Student's two-tailed t-test.

Results

H2O2 induces oxidative stress and increases GSK-3β activity

To investigate the regulation of GSK-3β activity under oxidative stress, we employed H2O2 to create the condition in this study. First, the HEK293/tau cells were exposed to various concentrations (0, 50, 100, 250, 500, or 1000 μM) of H2O2 for 2 h at 37°C. Assuming the absorbance of control cells was 100%. Compared with control group, viabilities of H2O2-treated cells measured by CCK8 kit were 90% for 50 μM, 75% for 100 μM, 47% for 250 μM, 27% for 500 μM, and 23% for 1000 μM (Fig. 1a). Considering the optimal living status of cells for experimentation, we chose 100 μM H2O2 as the final concentration of drug administration in the following study. Then, we measured the level of malondialdehyde as indicator of lipid peroxidation and activity of SOD. We found that incubation of HEK293/tau cells with 100 μM H2O2 resulted in oxidative stress, characterized by a significant increased level of malondialdehyde and a decreased activity of SOD compared with the control group (Fig. 1b and c).

Figure 1.

Hydrogen peroxide (H2O2) induces oxidative stress and increases glycogen synthase kinase-3 beta (GSK-3β) activity in human embryonic kidney 293 (HEK293)/tau cells. (a) Relative cell viabilities. Assuming the absorbance of control cells was 100%. Compared with control group, viabilities of cells treated with different concentrations of H2O2 for 2 h were 90% for 50 μM, 75% for 100 μM, 47% for 250 μM, 27% for 500 μM, and 23% for 1000 μM. (b, c) Measurement of malondialdehyde (MDA) and superoxide dismutase (SOD). Compared with control group, cells treated with 100 μM H2O2 for 2 h showed increased level of MDA and decreased activity of SOD. (d) GSK-3β activity measurement. Compared with control cells, the activity of GSK-3β in 100 μM H2O2-treated cells was increased by more than twofold. The data were from three to four separate experiments. *p < 0.05, **p < 0.01 versus (-) H2O2 control.

H2O2 was once reported to inhibit the activity of GSK-3β in a phosphatidylinositol-3-kinase (PI3K)-dependent manner (Blair et al. 1999). However, recent study suggested that H2O2 caused dephosphorylation of Ser9 of GSK-3β (an activated form) also via PI3K/Akt signaling pathway in PC12-D2R cells (Nair and Olanow 2008). Obviously, the exact regulatory mechanism of GSK-3 β activity is still vagued. We here measured GSK-3β activity in HEK293/Tau cells after treatment with 100 μM H2O2 for 2 h. The result showed that the activity of GSK-3β was significantly increased by more than twofold as compared with that in control cells (Fig. 1d). Taken together, these findings suggest that GSK-3β activity is up-regulated under oxidative stress induced by 100 μM H2O2.

H2O2 increases phospho-Ser9 of GSK-3β via Akt activation

Phosphorylation of GSK-3β is regarded as the main regulatory mechanism on GSK-3β activity in physiological and some pathological conditions. Phospho-GSK-3β at Tyr216 is an activated form of GSK-3β while phospho-GSK-3β at Ser9 is an inactivated form of GSK-3β (Jope et al. 2007). The activation of the PI3K/Akt pathway seems to be predominant in down-regulating GSK-3β activity via Ser9 phosphorylation, and the activation of Akt is associated with increased phosphorylation of its Ser473 and Thr308 (Alessi et al. 1996). Furthermore, Fyn, a protein kinase of the Src tyrosine kinase family expressed in the brain and T lymphocytes, may directly phosphorylate Tyr216 of GSK-3β and thereby activate GSK-3β. In this study, western blotting showed that H2O2 treatment did not change the expression of phospho- Tyr216 GSK-3β, Fyn, and DM1A as a loading control (Fig. 2a, c, f, and g), but markedly increased the expression of phosphorylated-GSK-3β at Ser9, phospho-Thr308 Akt and phospho-Ser473 Akt (Fig. 2a, b, d, and e). These results demonstrate that H2O2 increases phospho-Ser9 of GSK-3β via Akt activation, but there were no obvious changes of phosphorylation of GSK-3β on Tyr216. In other words, H2O2 regulates GSK-3β via its phosphorylation at Ser9 rather than Tyr216, and from our GSK-3β activity assay, we can conclude that some other mechanisms may play essential roles in modulating GSK-3β activity which counteract the inhibitory effects of phosphorylation of GSK-3β at Ser9 induced by H2O2 in HEK293/Tau cells.

Figure 2.

Hydrogen peroxide (H2O2) increases phosphorylation of glycogen synthase kinase-3 beta (GSK-3β) at Ser9. (a) Western blotting showed the expression of phospho-Ser9 GSK-3β, phospho-Tyr216 GSK-3β, phospho-Thr308 Akt, phospho-Ser473 Akt, Fyn, and DM1A against alpha-tubulin as a loading control, in human embryonic kidney 293 (HEK293)/Tau cells after treatment with 100 μM H2O2 for 2 h. (b–g) Quantitative analyses of the blots showed that H2O2 markedly increased phospho-Ser9 GSK-3β, phospho-Thr308 Akt, phospho-Ser473 Akt, but did not change phospho-GSK-3β-Tyr216, Fyn, and DM1A. The data were presented as means ± SD of three independent experiments. *p < 0.05, **p < 0.01 versus (-) H2O2 control.

H2O2 induces [Ca2+]i elevation and cleavage of GSK-3β in HEK293/Tau cells

It is evident that GSK-3β activity can be increased by rises in intracellular Ca2+ (Hartigan and Johnson 1999; Cedazo-Minguez et al. 2003). In this study, we used Fura-2AM to measure [Ca2+]i by fluorescence. We found that 100 μM H2O2 triggered [Ca2+]i increase from basal level of 102.3 ± 9.2 nM to 556.7 ± 12.1 nM of peak over 32.2 ± 8.7 s, then [Ca2+]i decreased from the peak level to 101.1 ± 12.2 nM over 61.2 ± 11.3 s and stayed at this plateau during our experimental period (n = 32 cells, as shown in Fig. 3a). These data suggest that H2O2 increases [Ca2+]i in HEK293/Tau cells, and also provide further indication that some calcium-dependent substances may upregulate GSK-3β activity. Simultaneously, when we detected the total GSK-3β, we were surprised to find that GSK-3β was cleaved into two fragments of 40 kDa and 30 kDa (Fig. 3b).

Figure 3.

Hydrogen peroxide (H2O2) induced [Ca2+]i elevation and cleavage of glycogen synthase kinase-3 beta (GSK-3β). (a) Fura-2-AM labeled cells detected by I× 71 confocal microscopy showed that [Ca2+]i of 100 μM H2O2-treated human embryonic kidney 293 (HEK293)/Tau cells was increased from basal level of 102.3 ± 9.2 nM to 556.7 ± 12.1 nM of peak over 32.2 ± 8.7 s, then [Ca2+]i decreased from the peak level to 101.1 ± 12.2 nM over 61.2 ± 11.3 s and stayed at the plateau (n = 32 cells). (b) Western blots showed that GSK-3β was cleaved into two fragments of 40 kDa and 30 kDa.

Calpain truncates GSK-3β and up-regulates its activity

In 2007, Paloma Goni-Oliver et al. reported that N-terminal cleavage of GSK-3 by calpain leads to the loss of its inhibitory domain, and calcium influx in cultured cortical neurons induces GSK-3 proteolysis through calpain activation (Goñi-Oliver et al. 2007, 2011). To further investigate the effect of calcium influx and calpain activation on GSK-3β truncation and activity in H2O2-treated HEK293/Tau cells, we detected the activities and cleaved fragments of GSK-3β in 2, 3, 5, 15, 30, and 120 min, respectively, after 100 μM H2O2 with or without 10 μM specific calpain inhibitor IV treatment. Using western blot analysis, we found that H2O2 markedly increased the truncation of GSK-3β into two fragments of 40 kDa and 30 kDa while total GSK-3β had no difference at each time point mentioned above (Fig. 4a and b), whereas inhibition of calpain decreased the truncation (Fig. 4c). These data show that calpain truncated GSK-3β into two fragments after 100 μM H2O2 treatment in HEK293/Tau cells.

Figure 4.

Hydrogen peroxide (H2O2) induced-calpain activation caused truncation of glycogen synthase kinase-3 beta (GSK-3β) and up-regulated its activity. (a, b, c) Cleaved fragments of GSK-3β were detected 2, 3, 5, 15, 30, and 120 min after 100 μM H2O2 with or without 10 μM specific calpain inhibitor IV treatment. H2O2 markedly increased the truncation of GSK-3β into two fragments of 40 kDa and 30 kDa while total GSK-3β level was not different at each time point. Calpain inhibitor IV decreased the truncation. (d) GSK-3β activity assay showed that H2O2 led to an increased of GSK-3β activity, and calpain inhibitor IV recovered GSK-3β activity to control level. (e) Calpain inhibitor IV did not influence the transient [Ca2+]i elevation and recovered GSK-3β activity. The data were presented as means ± SD of three independent experiments. **p < 0.01 versus (-) H2O2 control.

GSK-3β activity assay showed that H2O2 led to an increased of GSK-3β activity during the whole experimental process (Fig. 4d), whereas calpain inhibitor IV did not influence the transient [Ca2+]i elevation (Fig. 4e) but recovered GSK-3β activity (Fig. 4d). These results suggest that H2O2 activates GSK-3β via calpain activation not directly by [Ca2+]i elevation.

Taken together, Ca2+-dependant calpain activation truncates GSK-3β into two fragments of 40 kDa and 30 kDa, which counteracts the inhibitory effect of Ser9 phosphorylation and up-regulates GSK-3β activity in H2O2-treated HEK293/Tau cells.

Tau is phosphorylated in H2O2-treated HEK293/Tau cells

GSK-3β is known as human tau protein kinase (TPK I) in Alzheimer's disease (Ishiguro et al. 1992). To test the effect of truncated GSK-3β on tau phosphorylation, we here also detected the phosphorylation of tau at Ser396, Ser404, and Thr231, three most common GSK-3β targeted sites, in HEK293/Tau cells treated with 100 μM H2O2 only or simultaneously treated with both 100 μM H2O2 and 10 μM calpain inhibitor IV. Western blotting showed that tau was plainly phosphorylated at Ser396, Ser404, and Thr231 after 100 μM H2O2 administration, whereas inhibition of calpain blocked the tau phosphorylation (Fig. 5a).

Figure 5.

Hydrogen peroxide (H2O2) induced-glycogen synthase kinase-3 beta (GSK-3β) truncation led to tau hyper-phosphorylation. (a) Western blotting showed that tau was obviously phosphorylated at Ser396, Ser404, and Thr231 after 100 μM H2O2 administration, whereas inhibition of calpain blocked the tau phosphorylation. (b) CDK5, another main tau protein kinase, and its regulatory subunit p35/p25 fragments did not change, while GSK-3β was truncated in H2O2-treated human embryonic kidney 293 (HEK293)/Tau cells.

CDK5 is another main tau protein kinase, whose regulatory subunit p35 can be also proteolytically cleaved by calpain, generating a p25 form and further activating CDK5 (Kusakawa et al. 2000). In this study, we detected whether p35 is cleaved or CDK5 is responsible for tau phosphorylation in H2O2-treated HEK293/Tau cells. We found that there were no obvious changes of CDK5 expression and p35/p25 fragments after H2O2 treatment while GSK-3β is truncated in HEK293/Tau cells (Fig. 5b). These results demonstrated that 100 μM H2O2 induced calpain specific GSK-3β activation, while no effects were exerted on CDK5 activity regulation. Thus, we may conclude that aberrant tau phosphorylation in 100 μM H2O2-treated HEK293/Tau cells is because of calpain induced GSK-3β activation rather than CDK5 activation.

Discussion

GSK-3β is subjected to multiple regulatory mechanisms. Among those modulatory pathways, phosphorylation of GSK-3β is the most widely studied one. The activity of GSK-3β can be inhibited by phosphorylation of Ser-9 which is located in its regulatory N-terminal domains and be activated by phosphorylation of Tyr216, but the role of Tyr216 phosphorylation in the regulation of GSK-3β activity may be much smaller than the role of Ser9 phosphorylation (Takahashi-Yanaga et al. 2004). Previous studies reported that GSK-3β activity is up-regulated under oxidative stress (Yu et al. 2011). In this study, we investigated the mechanism of GSK-3β regulation under oxidative stress. Employing 100 μM H2O2 to create the oxidative stress condition, we found that GSK-3β activity in H2O2-treated HEK293/tau cells was significantly increased by more than twofold as compared with that in control cells. Moreover, compared with the control group, there were no obvious changes of phosphorylation of GSK-3β at Tyr216 or the expression level of Fyn after H2O2 administration, which suggests that regulation of GSK-3β activity via phosphorylation at Tyr216 may not contribute much in oxidative stress. Furthermore, unexpectedly, H2O2 markedly increases phospho-Ser9 of GSK-3β via Akt activation. An insulin-induced stimulation of the PI3K/Akt pathway may collaborate in the down-regulation of GSK-3β activity and the phosphorylated Ser-9 binds as a competitive pseudo-substrate to the primed-binding site, inhibiting the binding of the protein and its phosphorylation (Dajani et al. 2001; Frame and Cohen 2001; ter Haar et al. 2001). These data suggest that some other mechanisms may play essential roles in modulating GSK-3β activity which counteract the inhibitory effect of phosphorylation of GSK-3β at Ser9 induced by H2O2 in HEK293/Tau cells.

Oxidative stress is associated with calcium overload, and in turn resulting in calpain activation (Goll et al. 2003; Peng and Jou 2010). Activation of calpain in vitro truncates GSK-3 into two fragments and thus loses its inhibitory domain (Goñi-Oliver et al. 2007). Our study showed that H2O2 led to a transient [Ca2+]i elevation, and simultaneously increased the truncation of GSK-3β into two fragments of 40 kDa and 30 kDa and up-regulated the activity of GSK-3β, whereas calpain inhibitor IV did not influence the transient [Ca2+]i elevation but decreased the truncation and recovered the activity of GSK-3β. These results suggest that H2O2 activates GSK-3β via calpain activation not directly by [Ca2+]i elevation.

Truncation of GSK-3β by calpain should decrease total phosphorylation of GSK-3β at Ser9 because of the loss of Ser9 with the N-terminal cleavage. However, there was a significant increase in phospho-Ser9 of GSK-3β via Akt activation in H2O2-treated HEK293/Tau cells. To explain the net increase in GSK-3β activity under oxidative stress condition, we think that the activated truncations counteract the inhibitory effect of Ser9 phosphorylation and up-regulates GSK-3β activity in H2O2-treated HEK293/Tau cells. In other words, it should be very careful to evaluate GSK-3β activity when phosphorylation of GSK-3β at Ser9 is increased under oxidative stress condition.

Oxidative stress is closely associated with the development of AD (Feng and Wang 2012), which is characterized by two lesions: intracellular neurofibrillary tangles made from hyperphosphorylated tau and extracellular senile plaques containing β-amyloid peptide. GSK-3β, a tau protein kinase, is up-regulated and associated with neurofibrillary pathology at all Braak stages in AD (Pei et al. 1999). This study shows that H2O2 induced-hyperphosphorylation of tau at Ser396, Ser404, and Thr231 resulted from activated truncations of GSK-3β. CDK5 was not responsible to tau phosphorylation because calpain exerted no effect on CDK5 activity regulation in H2O2-treated HEK293/Tau cells. It implies that GSK-3β, especially truncated GSK-3β lost the inhibitory domain, plays a crucial protein kinase on tau phosphorylation in AD patients accompanied by oxidative stress.

Prior study described that 1 mM H2O2 led to dephosphorylation of tau in rat brain primary neuronal cultures because of activation of PP2A (Davis et al. 1997). Moreover, retinoblastoma protein (pRb)-associated PP2A activity increased in a dose-dependent fashion following the treatment of H2O2 ranging from 200 to 800 μM (Cicchillitti et al. 2003). In this study, we suppose that 100 μM H2O2 is not enough to strengthen tau-association with PP2A, and the increase in GSK-3β activity overrides PP2A-dependent dephosphorylation of tau, finally results in tau phosphorylation.

In summary, our data prove the importance of N-terminal end in regulating GSK-3β activity and we also provide evidence that truncated GSK-3β by transient increased intracellular calcium concentration-induced calpain activation mainly alters its activity in H2O2-treated HEK293/Tau441 cells, and this truncation process counteracts the inhibitory effect of increased phosphorylation of GSK-3β at serine9 via Akt activation and serves as a dominant mechanism to up-regulate GSK-3β activity. Under this circumstance, GSK-3β rather than CKD5 caused aberrant tau phosphorylation in H2O2-treated HEK293/Tau cells.

Acknowledgements

We thank Dr. Liping Zhu from Key Laboratory of Pulmonary Diseases of Ministry of Health of China for discussion on [Ca2+]i analysis. This study was supported in part by grants from National Natural Science Foundation of China (81271405, 31071226). The authors have no commercial affiliations or conflict of interests. Author contributions: YF performed most of the experiments and writing. GY and HG carried out some of cell culture; YX and KZ carried out some of western blotting; XS directed measurement of intracellular calcium concentration; JZW directed most of the experiments; XW designed and oversaw the research project, analyses, and writing.

Ancillary