Accumulation of Phosphorylated β-Catenin Enhances ROS-Induced Cell Death in Presenilin-Deficient Cells

Presenilin (PS) is involved in many cellular events under physiological and pathological conditions. Previous reports have revealed that PS deficiency results in hyperproliferation and resistance to apoptotic cell death. In the present study, we investigated the effects of PS on β-catenin and cell mortality during serum deprivation. Under these conditions, PS1/PS2 double-knockout MEFs showed aberrant accumulation of phospho-β-catenin, higher ROS generation, and notable cell death. Inhibition of β-catenin phosphorylation by LiCl reversed ROS generation and cell death in PS deficient cells. In addition, the K19/49R mutant form of β-catenin, which undergoes normal phosphorylation but not ubiquitination, induced cytotoxicity, while the phosphorylation deficient S37A β-catenin mutant failed to induce cytotoxicity. These results indicate that aberrant accumulation of phospho-β-catenin underlies ROS-mediated cell death in the absence of PS. We propose that the regulation of β-catenin is useful for identifying therapeutic targets of hyperproliferative diseases and other degenerative conditions.

Presenilin (PS) is a multipass membrane protein that regulates many physiological and pathological processes. Mutations in two PS genes, PS1 and PS2, cause familial Alzheimer disease (FAD) and elevate levels of the longer form of amyloid b-peptide (Ab), which is derived from amyloid precursor protein (APP) [8,9]. PS is the catalytic subunit of the c-secretase complex that cleaves many type 1 membrane proteins. Besides its activity in the c-secretase complex, PS has effects on various intracellular activities, such as Wnt/b-catenin signaling, phosphatidylinositol 3-kinase/Akt and MEK/ERK signaling, calcium homeostasis, and apoptosis [10][11][12][13][14]. PS1-null mice are embryonic lethal and exhibit skeletal and central nervous system (CNS) defects [15]. Studies of mouse embryonic fibroblasts (MEFs) from PS1-null mice and skin of adult PS1 conditional knockout mice, however, show that PS1 deficiency elicits tumor-like phenomena at least in part through stabilization of b-catenin and epidermal growth factor receptor (EGFR) [10,16,17]. In addition, PS deficiency and inhibition of csecretase activity cause resistance to apoptotic stimuli [14,18].
In the present study, we identified a novel function of phosphorylated b-catenin in the absence of PS. Phosphorylated b-catenin accumulated abnormally in PS-deficient MEFs under conditions of serum deprivation, provoking reactive oxygen species (ROS) generation and ROS-induced cell death.

PS dKO MEFs are vulnerable to serum deprivationinduced cell death
Although we confirmed protein expression of PS1 and nicastrin in crude extracts of PS WT and human PS1-rescued PS dKO (hPS1) MEFs, PS was not detected in PS dKO MEFs, as previously reported [19]. Immature form of nicastrin was detected in PS dKO cells as expected. Exogenous human PS1, however, recovered the PS1-CTF band and maturation of nicastrin ( Figure 1A). To examine the role of PS in cell survival, conditions of serum deprivation were applied to each cell type. By LDH release assay, PS dKO MEFs were more vulnerable to serum deprivation than PS WT cells (p,0.001, Figure 1B). Restoration of hPS1 in PS dKO MEF decreased this vulnerability and restored the rate of cell death to PS WT levels (p,0.001, Figure 1B).

Serum deprivation accentuates ROS generation in PS dKO MEFs
To determine the mechanism that underlies serum deprivationinduced cell death in PS dKO MEFs, we examined ROS generation in these MEFs by fluorescence microscopy using the DCFDA fluorophore. After 36 hr in DMEM, PS dKO MEFs generated high levels of ROS, as evidenced by damaged morphology with DCFDA staining (Figure 2A). Unlike PS dKO MEFs, PS WT and hPS1 MEFs had healthier cell morphologies, exhibiting lower DCFDA signals ( Figure 2A). We also assessed the effects of trolox, a water-soluble derivative of vitamin E, as an antioxidant and LiCl, a widely used GSK-3 inhibitor, on stress from serum deprivation. Trolox and LiCl treatment protected PS dKO MEFs against serum deprivation-induced cell death (p,0.001, Figure 2B) and inhibited ROS generation (Figure 2A).

Phospho-b-catenin accumulates in PS dKO MEFs under serum deprivation
We examined b-catenin protein levels in PS dKO MEFs because PS interacts with b-catenin [20]. By Western blot, we observed higher levels of b-catenin in PS dKO MEFs, but bcatenin levels in hPS1 MEFs and PS WT MEFs were comparable in basal condition ( Figure 3A). Based on our results that LiCl treatment protected PS dKO MEFs against serum deprivation induced-ROS generation and cell death (Figure 2A and 2B, respectively), we investigated the effect of GSK-3b on b-catenin phosphorylation. By immunofluorescence using a specific antibody for phospho-b-catenin (S33, S37, and T41), we detected robust phospho-b-catenin signals in PS dKO MEFs under conditions of serum deprivation, but PS WT MEFs in either growth or serumdeprived media and PS dKO MEFs in growth media showed little fluorescence ( Figure 3B). In PS dKO MEFs, GSK-3b inhibition with LiCl reduced phospho-b-catenin levels (S33, S37, T41), but trolox did not ( Figure 3C), as confirmed by Western blot ( Figure 3D).

Accumulation of phospho-b-catenin induces cytotoxicity in H4 neuroglioma cells
To examine whether accumulated phospho-b-catenin affects cell viability, we introduced several b-catenin constructs into H4 neuroglioma cells. After transient transfection with each construct, we measured protein expression from these constructs in H4 cells. To induce phospho-b-catenin accumulation, the K19/49R mutant construct was transfected, because this mutant fails to undergo ubiquitination and degradation due to their mutated ubiquitination site [7]. By Western blot, levels of the phosphorylated form of b-catenin (on residues 33, 37, and 41) were high in K19/49R b-catenin mutant transfectants ( Figure 4A), suggesting that phospho-b-catenin accumulated in the cells. In contrast, the S37A b-catenin mutant induced the highest levels of total bcatenin but showed low reactivity by immunoblotting with antiphospho-b-catenin ( Figure 4A). Endogenous b-catenin and b-actin in each transfectant cell type were expressed similarly. Using these constructs, we performed the LDH release assay to measure cell viability under conditions of serum deprivation. K19/49R bcatenin mutant-transfected H4 cells had increased rates of cell death compared with WT b-catenin-transfected cells (p,0.005, Figure 4B); S37A b-catenin-transfected H4 cells, however, had greater viability compared with WT b-catenin transfectants (p,0.05, Figure 4B). To determine the effects of phospho-bcatenin on ROS generation, we stained H4 cells with DCFDA. Similar to the results of the LDH release assay, most K19/49R bcatenin transfectants were brightly stained, although only few cells remained ( Figure 4C). WT b-catenin and S37A b-catenin transfectants did not stain as intensely as the K19/49R transfectants ( Figure 4C), suggesting that accumulation of phospho-b-catenin induces ROS generation.

Discussion
PS performs many functions in such physiological and pathological processes as development, calcium signaling, apoptosis, and Alzheimer disease. Previous studies demonstrated that bcatenin accumulation occurs in the absence of PS, which is accompanied by increased proliferation [16,17]; these results were recapitulated in our experiments in which PS dKO MEFs experienced an accumulation in b-catenin ( Figure 3A) and increased proliferation (data not shown) in normal growth media. In contrast, several reports have shown that b-catenin suppresses growth and induces cytotoxicity [21][22][23]. Here, we examined cell mortality and b-catenin expression under conditions of serum deprivation in the absence of PS. Because PS dKO MEFs died easily on removal of serum compared with PS WT MEFs, we determined whether PS was associated with this phenomenon using hPS1-supplied MEFs. Reconstitution of human PS1 in PS dKO MEFs reversed the vulnerability to serum deprivationinduced cell death, suggesting that PS1 is protective under such circumstances.
To identify the underlying mechanism of these phenomena, intracellular ROS levels were measured, because ROS are formed during serum deprivation-induced cell death [24], although there has been no report that links between b-catenin and ROS generation. Many cells were stained brightly with DCFDA in PS dKO MEFs during serum deprivation, but PS-containing MEFs and LiCl-and trolox-treated PS dKO MEFs experienced reduced fluorescent intensity and cell death. When PS dKO MEFs were deprived of serum, they accumulated more phospho-b-catenin, generated more ROS, and died at higher rates than PS WT MEFs or hPS1 MEFs. Treatment with the antioxidant trolox inhibited ROS generation and serum deprivation-induced cell death in PS dKO MEFs (Figure 2A & B), but it could not block phospho-bcatenin levels ( Figure 3C & D). Treatment with LiCl, a potent inhibitor that blocks the phosphorylation of b-catenin by GSK-3b, reduced phospho-b-catenin levels, ROS generation, and serum deprivation-induced cell death in PS dKO MEFs (Figures 2A & B and Figures 3C & D). These results suggest that accumulation of phospho-b-catenin causes the generation of ROS and cell death in serum-deprived PS dKO MEFs. To assess the cytotoxic effect of GSK-3b-induced phospho-b-catenin, we transiently transfected H4 neuroglioma cells with wild-type and mutant b-catenin constructs. Because b-catenin undergoes sequential processing, including phosphorylation, ubiquitination, and degradation, we designed an experiment with various b-catenin mutants, in which we substituted serine for alanine at position 37. This mutation inhibited GSK-3b-mediated phosphorylation of b-catenin, curtailing subsequent ubiquitination and proteosomal degradation. Substitution of lysine with arginine at positions 19 and 49 inhibited ubiquitination of b-catenin and degradation as well. Western blot analysis with anti-phospho-b-catenin (S33, 37, T41) demonstrated that the K19/49R b-catenin mutant had the highest immunoreactivity, as expected. The S37A b-catenin mutant was robustly detected by an anti-b-catenin antibody that recognized the total protein but had low immunoreactivity against the GSK-3bmediated phospho-b-catenin antibody. This mutant was well labeled by anti-phospho-b-catenin (T41, S45), as expected (data not shown). In contrast, because exogenous WT b-catenin also was regulated by endogenous machinery, little exogenous WT bcatenin remained in the b-catenin mutants. In the LDH release assay, K19/49R b-catenin mutant-transfected H4 cells had higher rates of cell death and ROS generation than transfectants of other mutants. S37A b-catenin-transfected H4 cells experienced less death and ROS generation than wild-type b-catenin transfectants, despite higher total levels of b-catenin in S37A b-catenin mutant transfectants than in K19/49R b-catenin transfectants and wildtype cells (Figure 4). These results imply that increases in b-catenin are not essential in generating ROS or inducing cell death under conditions of serum deprivation. Only accumulation of phosphob-catenin is responsible for ROS generation and the ensuing cell death. This is the first report that demonstrates that accumulation of phospho-b-catenin generates ROS and leads to cell death in a PS-dependent manner. It is to be clarified why phospho-b-catenin is more toxic. It maybe phosphorylation changes the conformation of b-catenin such that it may interact with proteins in a different way or inhibits the E3 ligase machinery. Or, it may not be the phosphorylation per se, but the activity of GSK-3, b-TrCP E3 ligase, or proteasome [25][26][27]. We checked whether c-secretase inhibitor can mimic the increased vulnerability to serum withdrawal using PS expressing cells. Since c-secretase inhibitor did not affect vulnerability of the cells, we can rule out intramembrane proteolysis, and b-catenin is not regulated by csecretase inhibitor under serum deprivation condition. One other point is why serum deprivation preferentially increases b-catenin phosphorylation in the absence of PS. Under normal basal conditions, we previously showed that PS facilitates b-catenin phosphorylation via a scaffolding mechanism (bringing b-catenin and GSK-3 in complex with PS). At the same time, generalized GSK-3 activity (as detected by phospho-GSK3) is increased in PS dKO cells. Serum withdrawal should further enhance general GSK-3 activity in all cells because of lack of growth factor, etc. and loss of PI3K/Akt signaling [28,29]. But perhaps under conditions of prolonged serum withdrawal (36 hrs), GSK-3 activity, localization, etc may be altered (i.e. directed at tau for example) such that PS1 still sequesters b-catenin but cannot bring together GSK-3. If so, then the loss of PS may indirectly promote GSK-3 mediated phosphorylation of b-catenin by neither sequestering b-catenin nor GSK-3. These points are needed to be clarified by further study.
Taken together, b-catenin facilitates cell proliferation under normal growth conditions, but when serum is limited, phosphorylated b-catenin accumulates in the absence of PS and induces ROS generation and cell death.

Cell viability assay
Dead and membrane-ruptured cells were quantitatively measured by lactate dehydrogenase (LDH) release. Aliquots (50 ml) of cell culture medium were mixed with 0.42 mM b-NADH in potassium phosphate buffer, pH 7.4 (125 ml; sigma) and 20 mM sodium pyruvate (25 ml; Sigma) in a microplate. After being mixed with sodium pyruvate, the samples were measured immediately 7 times at 340 nm for 5 min. Kinetic measurements were monitored using a microplate spectrophotometer (PowerWave XS; BIO-TEK instruments Inc., Winooski, VT). We used 0.5% Triton X-100-treated samples as positive controls.

Western blot analysis
Harvested cell pellets were resuspended in RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, and 50 mM Tris-HCl, pH 7.4) containing protease inhibitors (Sigma) and incubated on ice for 20 min. After centrifugation at 15,000 g for 20 min, the supernatant was collected. Crude membrane fractions of the cultured cells were prepared according to previous reports, with minor modifications [32,33]. Protein concentrations were determined by BCA assay (Sigma), and equal amounts of protein were loaded on appropriate gels. The separated samples were transferred to a PVDF membrane and incubated with antibodies against the target proteins. Protein bands were visualized by enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech, Buckinghamshire, England) using a bioimaging analyzer (LAS-3000; Fuji, Tokyo, Japan).

ROS measurement
Levels of hydrogen peroxide were determined using dichlorofluorescein diacetate (DCFDA; Invitrogen, Carlsbad, CA). Treated cells were incubated with 10 mM DCFDA for 30 min and washed with PBS. Fluorescent signals were captured using a fluorescence microscope (Olympus, Tokyo, Japan).

Immunofluorescence
Treated cells were fixed with 2% paraformaldehyde (PFA) for 10 min and washed with PBS. Fixed cells were permeabilized with PBST buffer (0.5% Triton X-100 in PBS) and incubated with blocking solution (2% goat serum and 2% horse serum in PBST) for 1 hr. Cells were incubated with primary antibodies in blocking solution overnight at 4uC. Labeled cells were washed with PBST and relabeled with Cy-3-conjugated secondary antibody. After cells were counterstained with DAPI (Sigma), fluorescent signals were visualized by fluorescence microscopy (Olympus).

Statistical analysis
All data were expressed as mean6SEM. Significant differences were calculated using one-way ANOVA with Tukey's post-test. A P value of less than 0.05 denoted a statistically significant difference.