The Presenilin-2 Loop Peptide Perturbs Intracellular Ca2+ Homeostasis and Accelerates Apoptosis*

In cells undergoing apoptosis, a 22-amino-acid presenilin-2-loop peptide (PS2-LP, amino acids 308–329 in presenilin-2) is generated through cleavage of the carboxyl-terminal fragment of presenilin-2 by caspase-3. The impact of PS2-LP on the progression of apoptosis, however, is not known. Here we show that PS2-LP is a potent inducer of the mitochondrial-dependent cell death pathway when transduced as a fusion protein with HIV-TAT. Biochemical and functional studies demonstrate that TAT-PS2-LP can interact with the inositol 1,4,5-trisphosphate receptor and activate Ca2+ release from the endoplasmic reticulum. These results indicate that PS2-LP-mediated alteration of intracellular Ca2+ homeostasis may be linked to the acceleration of apoptosis. Therefore, targeting the function of PS2-LP could provide a useful therapeutic tool for the treatment of cancer and degenerative diseases.

Eleven years ago, presenilin-1 (PS1) 2 and -2 (PS2) were discovered through familial and molecular genetic studies that implicated their involvement in the etiology of Alzheimer disease (1). PS1 and PS2 proteins are composed of 467 and 448 amino acids, respectively, and share ϳ70% identity (2). They are both membrane proteins with eight predicted transmembrane domains and a large hydrophilic loop of ϳ120 amino acids between the sixth and seventh transmembrane domains. PS1 and PS2 appear to be ubiquitously expressed throughout all tissues (1,3) and are the essential components of the ␥-secretase complex that are involved in the processing of Notch receptor and amyloid-␤ precursor protein (4). Mutations in PS-1 and PS-2 can lead to increased formation of amyloid-␤ precursor protein-42, a highly neurotoxic amyloid oligomer (5).
A characteristic feature of PS1 and PS2 is their proteolysis by an endogenous presenilinase, generating amino-terminal (NTF) and carboxyl-terminal fragments (CTF) within cells (6,7). Many studies have shown that both NTF and CTF are involved in regulating proliferation and death of cells (8,9). It is also known that in cells undergoing apo-ptosis, activation of caspase-3 leads to a specific cleavage of CTF, yielding a shorter form of CTF plus a 22-amino-acid hydrophilic peptide (amino acids 308 -329 in PS2) (10 -12). The caspase-3-mediated cleavage of PS-2 has been linked to the execution process of apoptosis. For example, phosphorylation of PS2 at a caspase-recognition site (S330) that inhibited the cleavage of CTF by caspase-3 reduced the apoptotic activity of CTF (13). It is not known, however, how the 22-amino-acid presenilin-2-loop peptide (PS2-LP) generated during the initial stage of apoptosis contributes to the execution of apoptosis.
Mutations in PS have been linked to alterations of intracellular Ca 2ϩ ([Ca 2ϩ ] i ) homeostasis (14 -16). Disturbances in [Ca 2ϩ ] i homeostasis can lead to the modification of a multitude of Ca 2ϩ -dependent phenomena in cells and initiate cell death (17). In this study, we tested the hypothesis that PS2-LP, as an apoptosis by-product, can act as an enhancer of apoptosis by perturbing [Ca 2ϩ ] i homeostasis. We found that recombinant PS2-LP coupled with a TAT membrane penetrating peptide, can enter cells and induce apoptotic cell death. The proapoptotic effect of PS2-LP appears to correlate with its action on the intracellular Ca ϩ2 release machinery at the endoplasmic reticulum (ER). Our data suggests that perturbation of [Ca 2ϩ ] i homeostasis can act as an amplifying factor contributing to the efficient execution of apoptosis.

EXPERIMENTAL PROCEDURES
Materials-Synthetic PS-2 loop peptide (PS2-LP) and anti-PS2-LP polyclonal antibody were generated by the ProteinTech Group (Chicago, IL). Nickel-nitrilotriacetic acid columns were purchased from Qiagen, and Slide-A-Lyzer (MW 3.5 kDa) and the FITC labeling kit were purchased from Pierce. G-25 Sephadex gravity columns were obtained from Amersham Biosciences. The pET-28b-TAT vector (V2.1) was kindly provided by Dr. Steven F. Dowdy (18 -20). Anti-cytochrome c monoclonal antibody and anti-His-tag, anti-␤-actin monoclonal antibodies were purchased from BD Biosciences. Anti-manganese superoxide dismutase was purchased from Stressgen Biotechnologies (San Diego, CA). Anti-InsP 3 receptor rabbit polyclonal antibody was purchased from Calbiochem. Goat polyclonal antibody for SERCA2 was purchased from Santa Cruz Biotechnology, Inc. Cleaved caspase-3 (Asp-175) (5A1) monoclonal antibody was purchased from Cell Signaling Technology (Beverly, MA). EnzChek caspase-3 assay kit was purchased from Molecular Probes (Eugene, OR). Propidium iodide (PI) and thapsigargin were purchased from Sigma. All cell culture reagents were ACS grade and were purchased from Invitrogen.

Plasmids and Protein Purification-Oligonucleotides
Positive recombinant clones were sequenced prior to transformation into the E. coli strain BL21(DE3) pLysS. Bacterial cultures were grown overnight, and protein expression was induced by isopropyl 1-thio-␤-D-galactopyranoside treatment for 4 -6 h followed by sonication in a buffer solution containing 100 mM NaH 2 PO 4 , 10 mM, Tris-HCl, 10 mM imidazole, 7 M urea, pH 8.0 (Buffer B-7M). The His-tagged fusion proteins were purified using Ni 2ϩ -nitrilotriacetic acid-agarose affinity column, through sequential wash with B-7 M buffer and C-7 M buffer containing 100 mM NaH 2 PO 4 , 10 mM imidazole, 7 M urea, pH 6.3, followed by the B-7 M buffer and elution with a buffer containing 100 mM NaH 2 PO 4 , 10 mM Tris-HCl, 250 mM imidazole, 8 M urea, pH 4.5. The elution step was followed by dialysis against phosphate-buffered saline using the Slide-A-Lyser dialysis cassette. The TAT-fusion proteins were then desalted on a PD-10 column into phosphate-buffered saline, flashfrozen in 10% glycerol and stored at Ϫ80°C. FITC-labeled TAT fusion proteins were generated with fluorescein labeling according to standard protocols.
Caspase-3 Activity Assay-Enzymatic activity for caspase-3 was assayed using the EnzChek caspase-3 kit following the manufacturers protocol (23,24). Briefly, cells treated with 10 M TAT-PS2-SP or TAT-PS2-LP for 3 h were harvested and resuspended in 50 l of 1ϫ cell lysis buffer, followed by three cycles of freeze and thaw. 50 l of the supernatant from each sample was incubated with 50 l of 2ϫ substrate solution at room temperature for 30 min. Caspase-3-mediated cleavage of the substrate leads to increase in fluorescence, determined by the fluorescence measurement at excitation wavelength of 342 nm and emission wavelength of 441 nm. A standard curve using known amounts of 7-amino-4-methylcoumarin was used to convert fluorescent values to specific catalytic activity.
Mitochondria Isolation-Mitochondria were isolated from NRP-154 cells following the protocol of Unkila et al. (25). Briefly, after treatment with TAT-PS2-LP or TAT-PS2-SP, NRP-154 cells were washed three times with ice-cold phosphate-buffered saline. The cell pellet was sus-pended in a solution containing 250 mM sucrose, 20 mM HEPES, 1 mM dithiothreitol, 10 mM KCl, 1 mM EDTA, 1 mM EGTA, 1.5 mM MgCl 2 , and a mixture of protease inhibitors at 4°C. Cells were then homogenized with 20 strokes in a 22-gauge syringe and centrifuged at 800 ϫ g for 10 min to remove intact cells and nuclei. Mitochondria were pelleted by centrifugation at 15,000 ϫ g for 10 min at 4°C, and resuspended in RIPA buffer (150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 20 mM Tris-HCl, pH 7.5). For each sample, 30 g of total proteins from the cytosolic and mitochondrial fraction were used for SDS-PAGE and Western blot with anti-cytochrome c antibody, also with anti-␤-actin and anti-manganese superoxide dismutase as loading controls for cytosol and mitochondria, respectively (26).
Ca 2ϩ Measurements-50 -70% confluent NRP-154 cells were loaded with 2 M fura-2-AM for 45 min at 37°C, in a balanced salt solution (BSS, 140 mM NaCl, 2.8 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 10 mM HEPES, pH 7.2). Following wash out of fura-2-AM from the culture medium, cells were resuspended in BSS buffer and incubated with 10 M TAT-PS2-SP, TAT-PS1-LP, or TAT-PS2-LP for 30 min at 37°C. To measure the ATP-induced Ca 2ϩ release from the ER, cells were then resuspended in BSS buffer containing 0 Ca 2ϩ and 0.5 mM EGTA and transferred to a cuvette system attached to the PTI fluorometer (Photon Technology International, Monmouth Junction, NJ). 0.5 mM ATP was added to the cell suspension solution within 45 s after resuspension in the Ca 2ϩ -free BSS solution. The ATP-induced Ca 2ϩ release was assayed through ratiometric excitation of fura-2 at excitation wavelength of 340 and 380 nm, according to our published protocols (27).
Western Blot and Co-immunoprecipitation-After treatment with TAT-PS2-LP or TAT-PS2-SP, NRP-154 cells were harvested and lysed in ice-cold RIPA buffer (150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 20 mM Tris-HCl, pH 7.5) in the presence of a mixture of protease inhibitors. The cell lysates were centrifuged at 12,000 rpm for 10 min at 4°C, and the supernatants were mixed with Laemmli sample buffer. 20 g of protein were separated on a 4 -12% SDS-PAGE. The proteins were transferred to a polyvinylidene difluoride membrane and blotted with primary antibody and secondary horseradish peroxidase antibody. Peroxidase activity was developed with ECL.
Co-immunoprecipitation assays of TAT-PS2-SP, TAT-PS2-LP, InsP 3 R and SERCA2 were performed as follows. Lysates of NRP-154 cells treated with TAT-PS2-SP or TAT-PS2-LP was resuspended in 0.5 ml modified RIPA buffer plus protease inhibitors. 500 g of whole cell lysate was incubated overnight with 5 g of polyclonal anti-His-tag, anti-InsP 3 R, or anti-SERCA2 antibody. As a negative control, 5 g of preimmune rabbit IgG was used. The immunocomplexes were collected on protein G-Sepharose beads by incubating for 2 h and were washed four times with RIPA buffer. The beads were resuspended in 30 l of 2ϫ concentrated Laemmli electrophoresis buffer, separated by SDS-PAGE, and transferred onto polyvinylidene difluoride membranes for the detection of His-tagged TAT-PS2-SP and TAT-PS2-LP, InsP 3 R, and SERCA2.
Confocal Fluorescence Imaging-Detection of TAT-PS2-LP or TAT-PS2-SP inside NRP-154 cells was performed using fluorescence imaging of FITC that is conjugated with the various membrane-penetrating peptides. For assays of cell death, cells were treated with PI (1 g/ml) and visualized by confocal fluorescence microscopy using a Zeiss 510-META system.
Electrophysiology-Spodoptera frugiperda (Sf9) cells were grown and maintained as describe (28). Nuclei (5-10-m diameter) were isolated and selected for electrophysiology based on unique morphology. For patch clamp measurements of InsP 3 R channel activity, the pipette solution contained 140 mM KCl, 0.5 mM Na 2 ATP, 10 mM HEPES, pH 7.3, 1 M [Ca 2ϩ ], and 33 nM InsP 3 . All solutions were carefully buffered to desired free [Ca 2ϩ ] (29), which was confirmed by fluorometry. Single channel data were acquired as described (29). Segments of current traces exhibiting a single InsP 3 R channel were used for determination of open probability (P o ), and dwell time analyses by QuB software (30).
Statistical Analysis-Values are mean Ϯ S.E. Significance was determined by Student's t test. A value of p Ͻ 0.01 was used as criterion for statistical significance.

Expression and Purification of TAT Fusion Proteins and Delivery of
TAT-PS2-LP into NRP-154 Cells-Previous studies have shown that PS2-LP could be generated via double cleavage of PS-2 by presenilinase and caspase-3 (Fig. 1A) (6,7,11). To test the cellular function of PS2-LP, we attempted to express the PS2-LP peptide by transfection of cDNA plasmids into cultured cells. However, multiple trials have failed to result in detectable expression of PS2-LP in these cells, e.g. NRP-154, HeLa, or HEK 293 cells. This is not surprising, because small peptide molecules are frequently vulnerable to degradation by cellular quality control mechanisms.
As an alternative approach to test the in vivo function of PS2-LP, we took advantage of the well established HIV-mediated method for delivery of synthetic peptides into cells (19,20). Extensive studies have shown that an 11-amino-acid TAT peptide "YGRKKRRQRRR" can enable peptides of various sizes to penetrate the cell membrane. Recently, Dowdy and co-workers (18,19) developed the bacterial expression vector pHis-TAT for production of recombinant TAT fusion proteins and has kindly provided us with this plasmid. Using the pHis-TAT plasmid, we have generated the following three constructs, TAT-PS2-LP, TAT-PS1-LP, and TAT-PS2-PS. The TAT-PS1-LP and TAT-PS2-SP constructs were used for comparative studies and for control purpose (Fig. 1B).
We used the BL21 E. coli strain for expression of these recombinant fusion proteins. Through the use of Ni 2ϩ -affinity columns, we found that abundant amount of TAT-PS1-LP, TAT-PS2-LP, and TAT-PS2-SP proteins could be purified to homogeneity (Fig. 1C). Based on gel electrophoresis and Western blot, the molecular size of the peptide appeared to be ϳ7 kDa, as expected from the predicted molecular size for TAT-PS1-LP, TAT-PS2-LP, or TAT-PS2-SP.
To analyze the ability of these TAT fusion proteins to penetrate into NRP-154 cells, two independent assays were employed. First, we used immunoblot to detect the presence of TAT-PS2-LP and TAP-PS2-SP inside NRP-154 cells after a brief incubation in the culture medium. For this purpose, we have generated a polyclonal antibody against the presenilin-2 loop peptide. This antibody could recognize the synthetic PS2-LP, CTF, and the full-length PS2 (not shown). As shown in Fig. 1D, TAT-PS2-LP could efficiently enter NRP-154 cells by a 30-min incubation with either 1 or 10 M TAT-PS2-LP. Under these conditions, the cells retained TAT-PS2-LP expression for 3 h after the fusion protein was washed out of the medium.
Second, we used fluorescence microscopy to detect the presence of FITC-conjugated TAT-PS2-LP inside individual NRP-154 cells (Fig. 1E). Clearly, green fluorescent cells could be observed following a 30-min incubation with 5 M TAT-PS2-LP, revealing the effective membrane penetrating capability of the TAT-PS2-LP peptide into NRP-154 cells.
TAT-PS2-LP Triggers Apoptosis in NRP-154 Cells-Continuous monitoring of FITC-labeled NRP-154 cells demonstrated a potent cytotoxic effect of TAT-PS2-LP ( Fig. 2A). At 30 min after incubation of 5 M TAT-PS-LP with NRP-154 cells, most of the green fluorescent cells are healthy, as illustrated by the negative staining with PI. By 9 h, the majority of green fluorescent cells were PI-positive, illustrating that they were dead or dying. The pronounced cell death phenomenon was not observed in cells incubated with TAT-PS2-SP at either 30-min or 9-h incubation times.
Quantitative analysis showed that greater than 90% of FITC-labeled TAT-PS2-LP cells were dead at 9 h, whereas less than 10% of FITC-labeled TAT-PS2-SP cells suffered such fate (Fig. 2B). These results indicate that TAP-PS2-LP has a potent cytotoxic activity in NRP-154 cells.

TAT-PS2-LP Causes Release of Cytochrome c from Mitochondria and
Activation of Caspase-3-To distinguish the nature of TAT-PS2-LPinduced cell death, e.g. apoptosis or necrosis, we performed the following two sets of experiments. First, we examined whether treatment of cells with TAT-PS2-LP could lead to activation of caspase-3, a marker for apoptosis. As shown in Fig. 3A, activation of caspase-3 was observed in NRP-154 cells treated with TAT-PS2-LP (10 M) but not with TAT-PS2-SP. In addition, 3 h after washing out the TAT-PS2-LP, the degree of caspase-3 activation was higher compared with that at 30 min, consistent with the cytotoxic effect of TAT-PS2-LP.
Western blot analyses using the polyclonal anti-PS2-LP antibody showed that prolonged incubation of cells with TAT-PS2-LP led to maintenance of the integrity of the peptide inside the cell, because the apparent molecular size of the TAT-PS2-LP protein did not change, and there were no degradation products detected. For quantitative assay of the caspase-3 enzymatic activity, we used the EnzChek caspase-3 kit purchased from Molecular Probes. As shown in Fig. 3B, pretreatment of cells with TAT-PS2-LP significantly increased the caspase-3 activity, an effect that was not observed with treatment of cells with TAT-PS2-SP.
Second, we assayed whether TAT-PS2-LP could trigger the release of cytochrome c from mitochondria, another marker for cells undergoing apoptosis. As shown in Fig. 3C, treatment of NRP-154 cells with TAT-PS2-LP induced significant release of cytochrome c from mitochondria into the cytosol. In contrast, TAT-PS2-SP had negligible effect on cytochrome c release as compared with the control. Furthermore, no release of other mitochondrial proteins such as manganese superoxide dismutase (26) was detectable in the same samples (Fig. 3C). Together, our data demonstrate that TAT-PS2-LP triggers mitochondria-dependent cascade of cell death in NRP-154 cells.
TAT-PS2-LP Reduces Intracellular Ca 2ϩ Content in NRP-154 Cells-Previous studies (14 -16) have shown that PS mutations are associated  Notice that 3 h after washout of TAT-PS2-LP, the degree of caspase-3 activation is higher compared with that at 30 min, even though the intracellular content of TAT-PS2-LP declined with time. Western blot with anti-PS2-LP was performed with a polyclonal antibody against the synthetic PS2-LP peptide. B, quantitative assay for caspase-3 enzymatic activity was performed using the EnzChek kit (see "Experimental Procedures"). Incubation of cells with TAT-PS2-LP (10 M, 3 h) led to significant activation of caspase-3, an effect that was not observed with TAT-PS2-SP. C, treatment of NRP-154 cells with TAT-PS2-LP induces significant release of cytochrome c from mitochondria into the cytosol. Compared with the control, TAT-PS2-SP is less effective in triggering cytochrome c release. ␤-Actin and manganese superoxide dismutase were used as loading control for cytosol and mitochondrial pellet, respectively. The Western blots were representative of three other experiments.
with alterations of intracellular Ca 2ϩ homeostasis. Changes in intracellular Ca 2ϩ homeostasis have also been linked to the initiation and amplification of apoptosis (31)(32)(33). To test whether PS2-LP-mediated apoptosis is related to changes in intracellular Ca 2ϩ stores, we measured the amount of Ca 2ϩ inside the ER, by activating the InsP 3 -receptor-mediated Ca 2ϩ release pathway using 500 M ATP in the absence of extracellular Ca 2ϩ .
Pretreatment of NRP-154 cells with 10 M TAT-PS2-LP (30 min) significantly reduced the amount of Ca 2ϩ released from the ER by ATP, an effect that was not seen with the TAT-PS2-SP protein (Fig. 4A). Moreover, compared with TAT-PS2-LP, pretreatment of cells with TAT-PS1-LP did not alter the ER Ca 2ϩ storage (Fig. 4B). Because the PS1-LP and PS2-LP do not show conservation in a primary amino acid sequence, this result suggests that PS2-LP may have a specific effect on intracellular Ca 2ϩ homeostasis. The reduction of the ER Ca 2ϩ store may reflect two possibilities, either by altering Ca 2ϩ uptake or Ca 2ϩ release across the ER membrane.
TAT-PS2-LP Interacts with the InsP 3 R and Regulates Its Channel Activity-To understand the mechanisms underlying the TAT-PS2-LP-mediated changes in [Ca 2ϩ ] i homeostasis, we first examined the possible interactions between TAT-PS2-LP and InsP 3 R or SERCA2, through co-IP assays. As shown in Fig. 5A, the IP of TAT peptides with anti-His antibody could pull down InsP 3 R from cells treated with TAT-PS2-LP but not from cells treated with TAT-PS2-SP (left panel), and IP with anti-InsP 3 R antibody could pull down TAT-PS2-LP but not TAT-PS2-SP (right panel). This interaction appears to be specific, because we did not observe any interaction between TAT-PS2-LP and SERCA2 (lower panel). These results indicate that TAT-PS2-LP can interact with InsP 3 R, but not with the SERCA2, raising the possibility that PS2-LP may affect the activity of InsP 3 R.
We next explored the functional consequences of the interaction of TAT-PS2-LP with the InsP 3 R by performing patch clamp studies of endogenous InsP 3 R channels in their native membrane environment (28,29,34) (Fig. 5B) (Fig. 5B). The number of channels activated by InsP 3 (N A ) was also enhanced in the presence of TAT-PS2-LP (Fig. 5B). The product N A P o , a measure of the total InsP 3 -mediated flux, was enhanced by nearly 3-fold and over 4-fold by 0.5 or 2.5 M TAT-PS2-LP, respectively (Fig. 5B).
These data demonstrate that the interaction of TAT-PS2-LP with the InsP 3 R leads to a marked activation of channel gating in the presence of subsaturating concentrations of InsP 3 . These results are consistent with our findings that TAT-PS2-LP decreases the ER Ca 2ϩ store.

DISCUSSION
In the present study, we have used the TAT protein-penetrating transduction system to investigate the possible involvement of PS2-LP in cell apoptosis and [Ca 2ϩ ] i homeostasis. The advantage of this system results from the fact that it is technically challenging to use DNA transfection to introduce small peptides into living cells (35,36), as these peptides are often unstable and susceptible to cellular degradation. TAT-mediated protein transduction occurs in a rapid, concentration-dependent fashion that is independent of receptors and transporters. This technology has been used to introduce proteins ranging in size from 15 to 120 kDa into a wide variety of human and murine cells (36). To determine whether this method could be used to deliver PS2-LP into cultured cells, we generated the recombinant TAT-PS2-LP fusion protein, plus the necessary control constructs. We found that transduction of TAT-PS2-LP, but not TAT-PS2-SP, into NRP-154 cells induced the release of cytochrome c from mitochondria, increased the activity of caspase-3, and triggered cell apoptosis. We also observed that TAT-PS2-LP-treated cells exhibited significantly reduced Ca 2ϩ storage inside the ER, but TAT-PS1-LP and TAT-PS2-SP did not exhibit such effect. TAT-PS2-LP interacts with the InsP 3 R and increases its Ca 2ϩ release channel activity. These results provide the first physiological evidence for PS2-LP in regulating intracellular Ca 2ϩ homeostasis and cell apoptosis.
A number of previous studies have explored the role of PS2 and its derivatives in cell apoptosis. For example, PS2 has been shown to function as a proapoptotic effector (13,(37)(38)(39)(40)(41). Overexpression of the CTF of PS-2 has been shown to increase amyloid-␤ precursor protein recovery and to decrease cell viability, by augmenting caspase-3 activity (8).
Our results showed that PS2-LP, as a by-product of caspase-3 cleavage of CTF, could act as a potent amplifier of apoptosis.
Previous studies have suggested that PS1 and PS2 also play different roles in cell death. Specifically, overexpression of PS2 in PC12 cells led to increased cell death in response to a variety of apoptotic stimuli (42,43). The expression of a PS2 mutation associated with Alzheimer disease led to increased levels of apoptosis at both basal and stimulated conditions (41). On the other hand, the role of PS1 in cell death is less clear. Over- expression of L286V PS1 mutant but not wild-type PS1 resulted in increased susceptibility to cell death induced by trophic factor withdrawal and Ab-mediated neurotoxicity (44), whereas a separate study showed that overexpression of A246E mutant in PS1 did not enhance apoptosis (45). These studies are consistent with our data indicating that PS2-LP, but not PS1-LP, could perturb intracellular Ca 2ϩ homeostasis and accelerate cell death.
In summary, our data demonstrate that PS2-LP, as a by-product of apoptosis, can provide a potent feed-forward mechanism to accelerate the apoptosis process. Because altered intracellular Ca 2ϩ homeostasis is associated with the cellular function of PS2-LP, we suspect that this small peptide molecule may play important roles in the overall cell growth and death processes. Targeting PS2-LP may thus offer an attractive therapeutic approach for cancer-related human diseases.