Mitochondrial Targeted Cyclophilin D Protects Cells from Cell Death by Peptidyl Prolyl Isomerization*

Cyclophilin D (CyPD) is thought to sensitize opening of the mitochondrial permeability transition pore (mPTP) based on the findings that cyclosporin A (CsA), a pseudo-CyPD substrate, hyperpolarizes the mitochondrial membrane potential ( (cid:1)(cid:2) ) and inhibits apoptosis. We provide evidence that contrasts with this model. Using live cell imaging and two photon microscopy, we report that overexpression of CyPD desensitizes HEK293 and rat glioma C6 cells to apoptotic stimuli. By site-directed mutagenesis of CyPD that compromises peptidyl-prolyl cis-trans isomerase (PPIase) activity, we demonstrate that the mechanism involved in this protective effect requires PPIase activity. Furthermore, we show that, under resting conditions, (cid:1)(cid:2) is hyperpolar-ized in CyPD wild type-overexpressing cells but not in cells overexpressing mutant forms of CyPD that lack PPIase activity. Finally, in glutathione S -transferase (GST) pull-down assays, we demonstrate that CyPD binding to the adenine nucleotide translocator (ANT), which is considered to be the core component of the mPTP, is not affected by the loss of PPIase activity. Collectively, our data suggest that CyPD should be viewed as a cell survival-signaling molecule and indicate a protective role of CyPD against apoptosis that is mediated by one or more targets other than the ANT. In programmed cell death (apopto-sis) between Bam HI and Not I sites from pGEM-HNb-CyPD vector. The CyPD-enhanced yellow fluorescent protein (EYFP) fusion fragment was engineered by a two-stage fusion PCR strategy. First, full-length coding regions of CyPD and EYFP were amplified in two separate reactions. The 3 (cid:3) primer for CyPD had 21 bp of the 5 (cid:3) of EYFP, and the 5 (cid:3) primer for EYFP contained 20 bp of the 3 (cid:3) of CyPD. These complementary sequences were annealed in the second PCR stage to generate the CyPD plus EYFP fusion fragment. The CyPD stop codon was removed from the PCR primers. A Bam HI site was attached to the forward primer of CyPD, and an Xba I site was added to the reverse primer of EYFP. The fusion PCR fragment was then subcloned into pcDNA3.1/zeo( (cid:4) ) be- tween Bam HI and Xba I sites. The CyPD and EYFP bicistronic expression vector was generated by digesting pGEM-HNb-CyPD vector with Sma I and Not I. The CyPD fragment was subsequently subcloned into pIRES-EYFP (CLONTECH, Palo Alto, CA) between Eco RV and Not I sites.The rat ANT1 cDNA (gift of Dr. Y. Shinohara, University of To- kushima, Japan) was subcloned into the pGEM-HNb vector by PCR with a forward primer, 5 (cid:3) -actgcccgggatgggggatcaggctttgagc-3 (cid:3) , and a reverse primer, 5 (cid:3) -cggaattcttacacatattttttgatctcatcatac-3 (cid:3) . An Sma I site was attached to the 5 (cid:3) -end of the forward primer, and an Eco RI site was attached to the 5 (cid:3) -end of the reverse primer. The PCR fragment was subcloned into pGEM-HNb between Sma I and 0.1 (cid:1) g/ (cid:1) l and translated for 45 min in a rabbit reticulocyte lysate translation supplemented with L -[ S]methionine Binding of in vitro translated ANT1 to GST-CyPD fusion proteins was performed in phosphate-buffered saline, pH 7.2, at room temperature for 1 h, followed by three washes in equilibration buffer containing 0.5 M NaCl and 1% Triton X-100. Proteins bound to glutathione-Sepharose were resolved by SDS-PAGE followed by autoradiography.

In multicellular organisms, programmed cell death (apoptosis) is an essential process of normal development, tissue maintenance, and aging (1,2). Abnormal regulation of apoptosis results in multiple human diseases, including cancer, AIDS (3), and many neurological disorders (2,4). Mitochondria provide a key regulatory role in cell death by releasing apoptogenic proteins into the cytosol that initiate the caspase cascade (5)(6)(7)(8). Release of these factors is thought to occur via a tightly regulated increase in the permeability of the inner mitochondrial membrane, due to the opening of the mitochondrial permeability transition pore (mPTP) 1 (9 -12). The primary components of the mPTP include the voltage-dependent anion channel (VDAC) in the outer membrane, ANT in the inner membrane, and CyPD within the mitochondrial matrix (10,13). Bcl-2 family members can regulate the release of apoptotic factors from mitochondria through direct interaction with VDAC and regulation of ANT activity during apoptosis (14 -16).
Cyclophilins are a group of PPIases with highly conserved protein sequences (17). These proteins are thought to be important for protein folding (18). CyPD is a mitochondrial-targeted PPIase (19). Although its specific physiological function remains largely unknown, CyPD is considered critical for the opening of the mPTP (20,21). This view is based on the observations that CsA, a potent inhibitor of mitochondrial-mediated apoptosis (9), blocks the mPTP at concentrations similar to those needed to inhibit the enzymatic activity of CyPD. This suggested that the PPIase activity was a necessary step in the opening of the mPTP (22,23). Cyclophilins bind CsA via a conserved hydrophobic pocket, which is critical for PPIase activity (24 -26). Thus, early models of CsA inhibition of the permeability transition suggested that CsA acted as a pseudosubstrate of CyPD that prevented it from interacting with the mPTP. However, the addition of CsA was shown not to disrupt the binding of CyPD and ANT (27). In addition, Scorrano et al. (28) reported that CsA inhibited mPTP opening, whereas diethylpyrocarbonate (DPC) induced mPTP opening. Because both CsA and DPC inhibit PPIase activity, the authors argued that CyPD must regulate the mPTP independently of this enzymatic activity.
In this manuscript, we examine the role of CyPD in the regulation of cell death. Overexpression of CyPD was used to ascertain its physiological function in response to oxidative stress and staurosporin-induced cell death. Two-photon imaging was employed to reduce photo-bleaching of fluorophores and photo-toxicity in live cells, which was important for the prolonged imaging periods in our experiments. The collapse of ⌬⌿ and the loss of membrane integrity were monitored to assess the initiation of cell death (29,30). In contrast to the current view of CyPD acting as a key accessory protein that sensitizes the mPTP to opening, we demonstrate that overexpression of CyPD delays both the onset of ⌬⌿ collapse under oxidative stress and the loss of membrane integrity induced by staurosporin. The protective effect of CyPD was dependent on its PPIase activity as demonstrated by site-directed mutagenesis. Furthermore, under resting conditions, the ⌬⌿ was significantly higher in cells overexpressing CyPD. Taken together, these data suggest that CyPD is an important factor with functional consequences in the regulation of programmed cell death.

Construction of Expression Vectors-The coding fragment of rat
CyPD cDNA (gift of Dr. Halestrap, University of Bristol, UK) was amplified using a forward primer, 5Ј-cgggatccatgctagctctgcgctgcgg-3Ј, containing a BamHI site and a reverse primer, 5Ј-agtctctagaggctgtgacttagctcaactg-3Ј, containing an XbaI site. The PCR product was subcloned into the Xenopus expression vector pGEM-HNb between the BamHI and XbaI sites (31). The CyPD coding region was subsequently subcloned into pcDNA3.1/zeo(ϩ) (Invitrogen, Carlsbad, CA) between the BamHI and XbaI sites. All restriction enzymes are from Life Technologies (Rockville, MD). CyPD R96G and H167Q mutants were generated by QuikChange site-directed mutagenesis (Stratagene, La Jolla, CA) using plasmid pGEM-HNb-CyPD as template. The forward primer for the R96G mutant was 5Ј-ccttccacggggtcatcccagccttcatgtgcc-3Ј and the reverse complement primer was 5Ј-ggcacatgaaggctgggatgaccccgtggaagg-3Ј. The forward primer for the H167Q was 5Ј-ggctggatggcaagcaagttgtgtttggccatg-3Ј and the reverse complement primer was 5Ј-catggccaaacacaacttgcttgccatccagcc-3Ј. Both mutants were subsequently subcloned into pcDNA3.1/zeo(ϩ) between the BamHI and XbaI sites.
To generate the GST-CyPD fusion vector, the CyPD fragment was subcloned into pGEX-4T-2 (Amersham Biosciences, Piscataway, NJ) between BamHI and NotI sites from pGEM-HNb-CyPD vector. The CyPD-enhanced yellow fluorescent protein (EYFP) fusion fragment was engineered by a two-stage fusion PCR strategy. First, full-length coding regions of CyPD and EYFP were amplified in two separate reactions. The 3Ј primer for CyPD had 21 bp of the 5Ј of EYFP, and the 5Ј primer for EYFP contained 20 bp of the 3Ј of CyPD. These complementary sequences were annealed in the second PCR stage to generate the CyPD plus EYFP fusion fragment. The CyPD stop codon was removed from the PCR primers. A BamHI site was attached to the forward primer of CyPD, and an XbaI site was added to the reverse primer of EYFP. The fusion PCR fragment was then subcloned into pcDNA3.1/zeo(ϩ) between BamHI and XbaI sites. The CyPD and EYFP bicistronic expression vector was generated by digesting pGEM-HNb-CyPD vector with SmaI and NotI. The CyPD fragment was subsequently subcloned into pIRES-EYFP (CLONTECH, Palo Alto, CA) between EcoRV and NotI sites.
The rat ANT1 cDNA (gift of Dr. Y. Shinohara, University of Tokushima, Japan) was subcloned into the pGEM-HNb vector by PCR with a forward primer, 5Ј-actgcccgggatgggggatcaggctttgagc-3Ј, and a reverse primer, 5Ј-cggaattcttacacatattttttgatctcatcatac-3Ј. An SmaI site was attached to the 5Ј-end of the forward primer, and an EcoRI site was attached to the 5Ј-end of the reverse primer. The PCR fragment was subcloned into pGEM-HNb between SmaI and EcoRI sites.
All vectors were automatically sequenced by the University of Texas Health Science Center at San Antonio sequencing core facility to demonstrate correct engineering. The oligonucleotides used were purchased from Operon Technologies (Alameda, CA).
Cell Culture and the Generation of Stable Cell Lines-Cells were maintained at 37°C in Dulbecco's modified Eagle's medium/F-12 (Invitrogen, Rockville, MD) supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 200 g/ml penicillin, and 100 g/ml streptomycin in a humidified atmosphere of 5% CO 2 /95% air. Cells were subcultured 1:10 by incubating in 0.05% Trypsin, 0.53 mM EDTA (Invitrogen, Rockville, MD) for 5 min when they were 70% confluent. DNA was transfected into cells by LipofectAMINE reagent (Invitrogen, Rockville, MD) according to the instructions of the manufacturer. After 72 h of transfection, HEK293 cells were trypsinized and transferred from a six-well dish to a 100-mm dish with medium containing 200 g/ml Zeocin (Invitrogen, Carlsbad, CA). Surviving cells were grown for 2-3 weeks, and symmetrical colonies were marked under a microscope. A grease boundary was drawn around marked colonies. A solution of 5 l of 0.05% Trypsin and 0.53 mM EDTA was briefly applied to the colony, and the cell suspension was transferred to 96-well plates. 24 colonies were picked for each construct to examine their protein expression level.
Image Acquisition and Analysis-Cells transiently expressing CyPD-EYFP were transferred to a 35-mm dish with a coverslip bottom 24 h after transfection and allowed to re-attach for another 24 h. Five minutes prior to imaging, the growth medium was replaced with recording buffer (120 mM NaCl, 4.5 mM KCl, 1 mM CaCl 2 , 2 mM MgCl 2 , 10 mM HEPES, pH 7.4) containing 50 nM tetramethyl rhodamine ethyl ester (TMRE, Molecular Probes, Eugene, OR) at room temperature. Fluorescent images were acquired using an Olympus Fluoview 300 confocal microscope with a 40ϫ oil objective (numerical aperture (N.A.) ϭ 1.35) at zoom 5. CyPD-EYFP was excited at 488 nm and fluorescence was detected between 510 and 550 nm. TMRE was excited at 568 nm, and fluorescence was detected beyond a 585-nm long pass barrier filter.
A Focht Chamber System 2 (FCS2, Bioptechs Inc., Butler, PA) was used for time-lapse two-photon imaging. Cells were transferred to the FCS2 coverslips 24 h prior to each experiment and maintained in the cell culture incubator. The chamber system was subsequently assembled and mounted on an Olympus Fluoview 300 confocal microscope customized for two-photon excitation and external photo-multiplier detection. Chamber temperature was maintained at 37°C during imaging, and cells were constantly perfused with phenol red-free Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum, 25 mM HEPES, pH 7.4, 2 mM L-glutamine, 200 g/ml penicillin, 100 g/ml streptomycin, 50 nM TMRE, and 100 M tert-butyl hydroperoxide (t-BuOOH) at 15 ml/h. TMRE was excited at 820 nm using a Ti-sapphire Coherent Mira 900 laser pumped with a 5-watt Verdi laser (Coherent Inc., Santa Clara, CA). Laser intensity was attenuated with a neutraldensity filter wheel such that no significant photo-bleaching of TMRE was observed during a 24-h recording period. Signal was collected through an Olympus 60ϫ oil objective (N.A. ϭ 1.4) at zoom 1.
For CyPD and EYFP bicistronic expression, the FCS2 chamber was mounted on a Zeiss 510 confocal microscope adapted for two-photon imaging. EYFP was excited at 488 nm, and emission was collected through a 505-to 550-nm band pass barrier filter. PI was excited at 543 nm and emission was collected through a 560-nm long pass filter. TMRE was excited at 820 nm, and emission was collected through a 565-to 615-nm barrier filter. Confocal and two-photon imaging was performed in two tracks alternating in the line-by-line mode. Differential interference contrast (DIC) images were collected simultaneously with either the PI or TMRE track. Signal was collected through a Zeiss 63ϫ oil objective (N.A. ϭ 1.4) at zoom 1. Once an EYFP-positive cell was identified, 488-nm excitation was turned off, and only the TMRE/PI and DIC signals were acquired in the single-track setting. Image analysis was performed using the public domain program ImageJ (National Institutes of Health, available at rsb.info.nih.gov/ij/) and the software ANALYZE (Mayo Foundation, Rochester, MN).
PPIase Activity Assay-PPIase activity was estimated according to Kofron et al. (32). Briefly, purified GST-CyPD or GST (10 nM final concentration) was pre-equilibrated with buffer (50 mM HEPES, 100 mM NaCl, pH 8.0). Immediately before assaying activity, chymotrypsin (60 mg/ml in 1 mM HCl, final concentration 6 mg/ml, Calbiochem, La Jolla, CA) was added. The PPIase substrate I (Calbiochem) was dissolved in the solvent trifluoroethanol with LiCl (470 mM) to a 3 mM stock concentration. The PPIase substrate I was added to the reaction to give a final concentration of 75 nM. Absorption at 380 nm was measured in a BioSpec-1601 Spectrophotometer (Shimadzu, Columbia, MD). Reactions with no PPIase were used as background controls.
Western Blot Analysis-Western blot procedures were essentially performed as previously described (33). Briefly, cells were harvested by trypsinization and washed three times with phosphate-buffered saline. Subsequently, cells were solubilized in 20 mM Tris-HCl, pH 7.6, 140 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 1 M pepstatin A, 20 M leupeptin, 0.2 mM 4-(2-aminoethyl)benzenesulfonylfluoride hydrochloride at 4°C for 30 min. The cell lysate was centrifuged at 900 ϫ g for 5 min to separate the nuclei. The concentration of total protein was determined by the BCA assay kit (Pierce, Rockford, IL). For SDS-PAGE, 10 g of total protein was loaded per lane. CyPD was detected with polyclonal rabbit anti-CyPD antibody (Pocono Rabbit Farm & Laboratory, Inc., Canadensis, PA). Horseradish peroxidase-conjugated secondary antibody (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) was used and visualized by chemiluminescence (PerkinElmer Life Sciences, Boston, MA). Goat polyclonal anti-actin antibody sc-1615 and donkey anti-goat IgG sc-2304 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were used to show the equal loading of proteins.
GST Fusion Protein Purification-BL21(DE3) bacteria transformed with pGEX-4T-2-CyPD construct were grown until a 0.6 OD at 600 nm was reached. Isopropylthio-␤-D-galactoside (Invitrogen, Rockville, MD) was added to a final concentration of 1 mM to induce GST-CyPD expression for 4 h at 37°C. Bacteria were harvested and lysed by sonication in 1ϫ phosphate-buffered saline containing 1 mM phenylmethylsulfonyl fluoride, 100 mM EDTA. Bacterial lysate was centrifuged at 22,000 ϫ g, and the supernatant was collected. Binding of GST or GST fusion protein to glutathione-Sepharose 4B (Amersham Biosciences) was performed at 4°C for 1 h followed by three washes with equilibration buffer (0.5 M Tris-HCl, pH 8.0, 4 mM EDTA, 0.1% 2-mercaptoethanol, 5% glycerol, 1 mM phenylmethylsulfonyl fluoride). Elution of bound protein was performed in equilibration buffer with 15 mM glutathione.
In Vitro Translation of ANT1 and Binding to GST Fusion Protein-Preparation of synthetic mRNAs and in vitro translations were performed as previously described (34,35). Briefly, mRNAs were diluted to 0.1 g/l and translated for 45 min in a rabbit reticulocyte lysate translation system (Promega, Madison, WI) supplemented with L-[ 35 S]methionine (PerkinElmer Life Sciences, Boston, MA). Binding of in vitro translated ANT1 to GST-CyPD fusion proteins was performed in phosphate-buffered saline, pH 7.2, at room temperature for 1 h, followed by three washes in equilibration buffer containing 0.5 M NaCl and 1% Triton X-100. Proteins bound to glutathione-Sepharose were resolved by SDS-PAGE followed by autoradiography.

Overexpression of CyPD Protects Cells from Oxidative
Stress-CyPD is thought to be an accessory protein that controls the opening of the mPTP in response to apoptosis-inducing factors (36). Because CsA is known to bind CyPD and, thereby, reduce the probability of mPTP opening, we hypothesized that overexpression of CyPD would increase sensitivity of cells to oxidative stress. Our approach to test this hypothesis was 2-fold. First, we tested whether cells overexpressing CyPD exhibited the correct targeting of this protein to the mitochondria. For this purpose, we created a fusion protein consisting of CyPD tagged with EYFP at its COOH-terminal and then transiently transfected this construct into HEK293 cells. A cell expressing CyPD-EYFP is presented in Fig. 1A. When cells expressing this fusion protein were also labeled with the mitochondrial marker TMRE (Molecular Probes, Eugene, OR), we found that CyPD-EYFP was correctly targeted to the mitochondria (Fig. 1A). The second step in our experimental design was to create stable cell lines overexpressing wild type CyPD (nontagged). A Western blot analysis of CyPD expression in one of these cells lines is presented in Fig. 1B. The physiological effects of CyPD overexpression were tested in this stable cell line and compared with a control cell line, expressing vector alone. Oxidative stress was induced by exposure to t-BuOOH and measured by monitoring the ⌬⌿ (37). Briefly, cells were seeded on glass coverslips and allowed to adhere for 24 h. Plated cells were initially labeled with TMRE (50 nM) in a culture incubator for 5 min. The glass coverslip was then sealed in a closed perfusion chamber, and the cells were imaged on a two-photon confocal microscope for 12 h. Two-photon imaging was used to simultaneously excite the membrane potential indicator TMRE and to produce a DIC image. An overlay of both the DIC image and the TMRE-labeled mitochondria is presented in Fig. 1C. Throughout the course of the experiment, a complete z section (six images in 1-m steps) of the cells was taken every 5 min. The cells were then monitored for ⌬⌿ collapse, which was measured by the loss of TMRE fluorescence. Simultaneous imaging with DIC optics insured that a loss of fluorescence was not due to a change in the focal plane or to the loss of the cell. Spatial-temporal stacks were created with only a single image from each z section (Fig. 1D). The chosen image from each z section was selected on the basis of brightest TMRE fluorescence, which generally coincided with the middle image. These data reveal that cells overexpressing CyPD maintain their mitochondrial membrane potentials in the presence of t-BuOOH for significantly longer periods of time compared with cells expressing vector alone (p Ͻ 0.0005, t test). A histogram of the average time of ⌬⌿ collapse is presented in Fig. 1E. The average time until the ⌬⌿ collapsed for CyPD-overexpressing cells was 9.9 Ϯ 0.3 h (n ϭ 69 cells, pooled from four separate experiments). The average time until ⌬⌿ collapsed was 6.2 Ϯ 0.1 h in control cells expressing vector alone (n ϭ 85 cells, pooled from five separate experiments). We conclude from these data that overexpression of CyPD significantly prolongs the time until ⌬⌿ collapses. Thus, rather than sensitizing cells to oxidative stress, CyPD appears to significantly protect the cells.
CsA Inhibits the Collapse of ⌬⌿ in Cells Exposed to Oxidative Stress-CsA is a well-established inhibitor of mPTP opening (38). Because the action of CsA on mPTP is thought to be mediated by CsA binding to CyPD, we examined how CsA exposure affected the collapse of ⌬⌿ in response to oxidative stress. HEK293 cells were plated on glass coverslips, labeled with TMRE, sealed in the perfusion chamber, and imaged as described above. In addition, cells were pre-treated with CsA (10 M) for 30 min and then continuously perfused with Dulbecco's modified Eagle's medium containing CsA (10 M), TMRE (50 nM), and t-BuOOH (100 M). Under these experimental conditions, we found that the average time for ⌬⌿ to collapse in CsA treated cells was 20.5 Ϯ 0.1 h (n ϭ 52 cells, pooled from two experiments) (Fig. 2). In comparison, the average time to ⌬⌿ collapse in control cells (vehicle only) exposed to t-BuOOH was 6.2 Ϯ 0.1 (n ϭ 54, pooled from two experiments). Because it is known that CsA also has non-mitochondrial targets that can affect cell viability (39), we also tested how the immunosuppressant FK506 affected the collapse of ⌬⌿ in response to oxidative stress. FK506 is an immunosuppressant that, like CsA, prevents activation of calcineurin. However, FK506 has no effect on mitochondria (40). Pre-treatment of cells with FK506 (1 M, 30 min) had no protective effect on the collapse of ⌬⌿ in cells exposed to t-BuOOH (Fig. 2). The average time to ⌬⌿ collapse was 6.0 Ϯ 0.1 h (n ϭ 58 cells, pooled from two experiments). We conclude from these data that CsA inhibits the collapse of ⌬⌿ when cells are exposed to oxidative stress. However, the effects of CsA are unlikely to be mediated by calcineurin inhibition, because treatment of cells with FK506 did not affect the collapse of ⌬⌿.
The Protective Effect of CyPD on ⌬⌿ is Independent of Celltype-To determine whether the protective effect of CyPD on ⌬⌿ collapse was dependent on the cell-type, we re-examined the effects of CyPD overexpression in another cell line. Rat C6 glioma cells were transiently transfected with a bicistronic vector for independent expression of CyPD and EYFP (see "Materials and Methods"). Cells were allowed to express the proteins for 48 h, and positively transfected cells were identified by EYFP fluorescence (Fig. 3A, left panels). Plated cells were then labeled with TMRE, treated with t-BuOOH, and imaged as described above. Consistent with our findings in the HEK293 cells, we found that C6 glioma cells overexpressing CyPD maintained their ⌬⌿ for significantly longer periods than non-transfected cells when exposed to t-BuOOH (Fig. 3A, upper panels). C6 glioma cells transfected with the bicistronic vector containing EYFP alone exhibited no significant difference from control, non-transfected cells (Fig. 3A, lower panels).  (Fig. 3B). We conclude from these data that, under oxidative stress, overexpression of CyPD prolongs the time until ⌬⌿ collapses in both HEK293 and rat C6 glioma cells, suggesting that it might be a generalized mechanism operating in many cell types.

FIG. 3. Prolongation of the time until ⌬⌿ collapses under oxidative stress is independent of cell type.
A, DIC images of rat C6 glioma cells (gray) overlaid with fluorescent signals representing overexpression of CyPD (green) and ⌬⌿ (red), during exposure to 100 M t-BuOOH (upper panels). CyPD and EYFP were transiently expressed using a bicistronic expression vector. CyPD-positive cells were identified as EYFP-positive cells and followed by DIC images throughout the recording. Cells were imaged for EYFP and TMRE before each experiment. Note that after 2 h, ⌬⌿ collapse did not occur only in the cell expressing CyPD and EYFP. Cells transfected with the bicistronic vector containing EYFP alone, and subsequently imaged during t-BuOOH exposure, are presented in the lower panels. Note that after 2 h, ⌬⌿ also collapsed in the cell expressing EYFP alone. Scale bar, 10 m. B, histogram of the average time when ⌬⌿ collapsed under oxidative stress for control non-transfected cells (Cntrl), CyPD-and EYFPtransfected cells (CyPD), and EYFP-transfected cells (EYFP). The asterisk denotes a statistical significant difference (p Ͻ 0.0005, single factor ANOVA).

Overexpression of CyPD Protects Cells from Staurosporininduced Cell
Death-To determine whether CyPD protects cells against other apoptotic stimuli, the effects of CyPD overexpression were examined in staurosporin-treated cells. To induce cell death, rat C6 glioma cells were treated with 1 M staurosporin for 6 h. Apoptosis was confirmed by staining the cells with Annexin V-FITC (Molecular Probes), which detects phosphatidylserine in the outer leaflet of the plasma membrane. Phosphatidylserine flips to the outer leaflet of the plasma membrane preceding the loss of membrane integrity and is an early indication of apoptosis (41). Co-staining with propidium iodide (PI, Molecular Probes) identified those cells with compromised plasma membrane integrity and that had progressed through apoptosis (Fig. 4A). These data demonstrate that staurosporin induces apoptosis in these cells. To test whether CyPD confers protection under these conditions, rat C6 glioma cells were transiently transfected with the bicistronic vector for independent expression of CyPD and EYFP.
Plated cells were treated with staurosporin in the presence of PI and imaged as described above following 48 h of expression. Positively transfected cells were identified by EYFP fluorescence. A spatio-temporal stack of PI staining during the course of the experiment demonstrated that C6 glioma cells overexpressing CyPD maintained their plasma membrane integrity for a significantly longer period than non-transfected cells when exposed to staurosporin (Fig. 4B, upper panels). C6 glioma cells expressing EYFP alone were not significantly different from control, non-transfected cells (Fig. 4B, lower panels). The average time when PI-positive staining occurred in the presence of staurosporin was 7.4 Ϯ 0.1 h in control, non-transfected cells (n ϭ 8 experiments) and 7.6 Ϯ 0.2 h for EYFP alone expressing cells (n ϭ 4 experiments). In contrast, this value was increased to 10.1 Ϯ 0.5 h (n ϭ 4 experiments) in CyPD and EYFP expressing cells (Fig. 4C). We conclude from these data that overexpression of CyPD protects rat C6 glioma cells from staurosporin-induced cell death.
PPIase Activity Is Required for the Protective Effect of CyPD-PPIase enzymatic activity in cyclophilins catalyzes the rotation of prolyl peptide bonds in target proteins (42). The crystal and NMR structures of cyclophilin A in complex with CsA revealed the presence of a hydrophobic pocket that subsequently was found to play a critical role both in substrate binding and PPIase activity (24 -26). A point mutation within this pocket, H126Q, completely abolished substrate binding and consequently, enzymatic activity (43). In contrast, a mutation near the mouth of the pocket, R55G, abolished PPIase activity without affecting substrate binding. We engineered the corresponding mutants in CyPD to test whether PPIase activity was required for the protective effect of CyPD on ⌬⌿ (Fig.  5A). CyPD wild type and mutants, R96G and H167Q, were created as GST fusion proteins, purified from bacteria, and assayed for PPIase activity (Fig. 5B). We found the mutagenesis at both residues abolished PPIase activity from 100% in CyPD to 3.0 Ϯ 1.9% in mutant R96G and 3.8 Ϯ 3.0% in mutant H167Q (n ϭ 3 experiments). Finally, GST alone did not display PPIase activity (3.4 Ϯ 1.5%) (Fig. 5C). These experiments demonstrate that the CyPD mutants, R96G and H167Q, do no exhibit detectable PPIase activity above background levels (GST), similar to the corresponding mutants in CyPA that abrogate PPIase activity. These data corroborate models of the structural requirements for substrate binding and PPIase enzymatic activity in the cyclophilins (24 -26).
As described above, we first tested whether cells overexpressing the PPIase mutants exhibited the correct targeting to the mitochondria. Fusion proteins consisting of CyPD R96G and H167Q were tagged with EYFP at their respective COOH terminals and then transiently transfected into HEK293 cells. Cells expressing these EYFP-tagged CyPD mutants are presented in Fig. 6A. Both mutants were correctly targeted to the mitochondria (Fig. 6A). Stable cell lines expressing the CyPD mutants were then created and selected for protein expression levels that were comparable to the CyPD wild type cell line (Fig. 6C). To test whether PPIase activity was required for the action of CyPD, plated cell cultures were prepared, labeled, and then imaged while exposed to t-BuOOH (100 M). Spatio-temporal stacks of a single field of cells are presented for each cell type for the duration of the recording (Fig. 6B). We found that, in response to oxidative stress, mutagenesis of CyPD abolished the protective effect observed with wild type CyPD as measured by the latency to ⌬⌿ collapse. Specifically, the average time until ⌬⌿ collapse was 6.3 Ϯ 0.1 h for R96G mutant cells (n ϭ 73 cells, from four experiments) and 6.2 Ϯ 0.1 h for H167Q mutant cells (n ϭ 85 cells, from five experiments) (Fig. 6D). These values were not significantly different from cells trans-fected with vector alone 6.2 Ϯ 0.1 (Fig. 6D, cf. Fig. 1E). We conclude from these data that the protective effect of CyPD during oxidative stress requires PPIase activity.
Overexpression of CyPD Increases Resting ⌬⌿-⌬⌿ is significantly affected by the open probability of the mPTP, and, at the same time, the mPTP open probability is tightly regulated by ⌬⌿ (44 -46). The lower the ⌬⌿, the higher the probability that the mPTP opens and the more likely it becomes that this positive feedback loop collapses ⌬⌿. Given the interdependence of mPTP and ⌬⌿, we examined whether overexpression of CyPD wild type and its mutants affected ⌬⌿ under resting conditions. Plated cells were obtained from the stable cell lines expressing CyPD wild type, R96G, and H167Q used above (Fig.  6C). The cells were labeled with TMRE and imaged at higher magnification (Fig. 7A). ⌬⌿ was estimated as the log ratio of mitochondrial (F mito ) to cytosolic (F cyt ) fluorescent intensity according to the Nernst equation: ⌬⌿ ϳ log 10 (F mito /F cyt ) (47). F mito was estimated by selecting the highest pixel value within a 5 ϫ 5 pixel region surrounding a single mitochondrion. In a given cell, the fluorescence intensity of the five brightest mitochondria was averaged. F cyt was also averaged from five measurements for each individual cell. The single-cell log ratios were then pooled to provide the average log 10 (F mito /F cyt ) for a particular cell line. We found that cells overexpressing CyPD wild type protein exhibited significantly higher log ratios of 1.96 Ϯ 0.02 (n ϭ 148 cells, p Ͻ 0.0005), as compared with 1.84 Ϯ 0.03 for mutant R96G (n ϭ 103 cells); 1.83 Ϯ 0.02 for mutant H167Q (n ϭ 157 cells) and 1.81 Ϯ 0.02 for control cells transfected with vector alone (n ϭ 166 cells) (Fig. 7B). These data demonstrate that the overexpression of CyPD increases the resting ⌬⌿ and provides a potential mechanism for the protective effect of CyPD.
CyPD Binding to ANT Is Not Affected by Mutations in the PPIase Binding Pocket-The binding of CyPD to ANT is considered important for mPTP regulation (48). CsA does not disrupt the binding of CyPD and ANT (27), and pharmacolog- ical inhibitors of PPIase activity have both excitatory and inhibitory effects on mPTP opening (28). These reports suggest that the site of interaction between CyPD and ANT does not involve the PPIase substrate-binding pocket. The hydrophobic residue H167 of CyPD is critical for PPIase substrate binding and subsequent enzymatic activity (43). If CyPD interacts with ANT via the hydrophobic pocket, then the CyPD H167Q mutant should not be able to bind ANT. To test this hypothesis, GST fusion proteins of CyPD wild type and mutants, R96G and H167Q, were purified from bacteria and bound onto glutathione-Sepharose beads. In vitro translated ANT, labeled with [ 35 S]methionine, was mixed with the GST-CyPD beads, washed, and subsequently resolved on SDS-PAGE followed by autoradiography. We found that ANT binding to CyPD was not significantly affected by mutations in the PPIase binding pocket. Densitometry readings showed no statistically significant differences between ANT binding to CyPD wild type (100%), in comparison to mutant H167Q (103 Ϯ 14%; n ϭ 3) and mutant R96G (104 Ϯ 11%; n ϭ 3) (Fig. 8). GST alone exhibited ANT binding of 6 Ϯ 5% (n ϭ 3). We conclude from these data that the interaction of ANT with CyPD does not occur through the hydrophobic pocket.

DISCUSSION
The mPTP is generally depicted as a core protein complex composed of VDAC, ANT, and CyPD (10,13). The importance of CyPD in the mPTP regulation was initially recognized by pharmacological studies involving the immunosuppressant CsA. CsA was shown to inhibit the mPTP and PPIase activity at similar concentrations (22,23). Biochemical data supported a tight interaction between CyPD and ANT (27,48). Thus, CsA was envisioned as a pseudosubstrate for the PPIase enzyme that removed CyPD binding from ANT, thereby inhibiting mPTP opening. However, Crompton and colleagues (27) recently reported that CsA does not displace CyPD from ANT, suggesting that the binding site important for PPIase activity does not directly involve in the ANT binding. Furthermore, Scorrano and co-workers showed that the PPIase inhibitor DPC induced mPTP opening, suggesting that CyPD affects the mPTP open-close transition independently of its PPIase activity (28). Lemasters and He (49) have recently presented evidence suggesting that the mPTP may be formed by aggregation of misfolded membrane proteins, which cluster to form a large conductance transmembrane pore. Moreover, CyPD was hypothesized to bind to these clusters in their model, possibly to inhibit pore formation by refolding damaged proteins to their native state. Taken together, these data indicate that the specific role of CyPD in mPTP function is different than previously assumed.
We have used overexpression studies to investigate the role of CyPD in the regulation of the mPTP. Based on previous models of CyPD interaction with the ANT, we postulated that overexpression of CyPD would sensitize mitochondria to oxidative stress. In fact, we found that CyPD overexpression results in a protective effect, increasing significantly the time it takes for the ⌬⌿ to collapse under oxidative stress. CyPD overexpression also protected cells against staurosporin-induced cell death. Point mutations in CyPD that eliminated PPIase activity did not have similar protective effects, clearly demonstrating that CyPD requires functional PPIase activity. Interestingly, suppression of cyclophilin A, the cytosolic isoform of the family of PPIases that bind cyclosporin A, has been reported in markedly sensitize cells to oxidative stress (50). An involvement of PPIase activity in oxygen defense has been reported for cyclophilin 18 in rabbit blastocysts (51). The thiol-specific antioxidant protein Aop1 has also been reported to bind human cyclophilin 18 (52). Taken together, these reports and our data indicate that cyclophilins are important components of our oxygen defense system.
The traditional view of the mPTP regulation is that CyPD binds to the ANT sensitizing the pore to open by stimuli such as Ca 2ϩ , pH, and reactive oxygen species (45,53,54). Our overexpression studies, on the other hand, indicate that CyPD can desensitize and close the mPTP. This view of mPTP function is consistent with recent work reported by Bauer and colleagues (55). These authors showed that transient overexpression of ANT1 induced apoptosis in HEK293 cells and that this effect could be inhibited by co-expression of CyPD. The authors suggested that CyPD might mask an interaction domain of ANT1 with an unknown repressor. This view is inconsistent with our data, because overexpression of CyPD, in the absence of ANT1 overexpression, would be predicted to increase mPTP opening by blocking repressor binding to the ANT. The alternative explanation proposed by the authors was that overexpression of ANT1 may titrate out endogenous inhibitors of mPTP, thereby inducing apoptosis. In this case, Bauer and colleagues suggested that CyPD would act as the endogenous inhibitor of the mPTP. However, Bauer and colleagues did not investigate the protective effects of CyPD expression on apoptosis, because ANT1 expression itself was used as their apoptotic stimulus. Our data are consistent with the second conceptual alternative (i.e. that CyPD acts as a inhibitor of the mPTP) and further demonstrate that CyPD protects cells against apoptotic stimuli.
Several possible mechanisms could account for the protective effect of CyPD against cell death. CyPD has been shown to directly bind ANT, a primary component of the mPTP, which could inhibit mPTP opening. This mechanism appears unlikely, because mutation of the PPIase binding pocket of CyPD removes the protective effect but does not affect ANT binding. Published reports are also consistent with our finding that ANT binding is not through the PPIase binding pocket of CyPD (27). Although the protective effect of CyPD appears unlikely to be mediated solely through ANT binding, it is possible that one or more other binding targets of CyPD are recruited to the ANT, which in turn directly inhibits mPTP opening. In fact, He and Lemasters (49) have recently proposed that CyPD and an unidentified chaperone-like protein bind to clusters of damaged proteins to inhibit mPTP opening. Our data are consistent with this new paradigm of mPTP pore structure and regulation. The protective effect of CyPD could also be mediated by an unknown target, which indirectly inhibits mPTP opening. For ). There is no statistical difference between CW, CR, and CH (single factor ANOVA) example, any protein that can functionally hyperpolarize ⌬⌿ could clearly account for the protective effect of CyPD against apoptotic stimuli given the ⌬⌿ dependence of mPTP opening. In support of this mechanism of action, mitochondrial respiration in yeast has been reported to decrease when the CyPD homolog CPR3 is mutated (56). Furthermore, our data demonstrate that overexpression of CyPD hyperpolarizes ⌬⌿, whereas overexpression of the two PPIase mutants that do not provide oxidant protection do not increase ⌬⌿. Irrespective of the precise underlying mechanism of action, our data underscore the importance of CyPD in the regulation of mPTP and programmed cell death and suggest CyPD can be viewed as a cell survival molecule.