Cyclophilin A as negative regulator of apoptosis by sequestering
11111Claude Bonfils , Nicole Bec , Christian Larroque , Maguy Del Rio , Céline Gongora ,
2 1 Martine Pugnière and Pierre Martineau .
1 IRCM, INSERM U896, UM1, CRLC Val d’Aurelle-Paul Lamarque, Montpellier,
2 CNRS, UMR5236, UM1-UM2, Faculté de Pharmacie, Montpellier, France.
Corresponding author: Claude Bonfils
IRCM, INSERM U896, CRLC Val d’Aurelle-Paul Lamarque, 208 Rue des Apothicaires, 34298 Montpellier cedex 5, France. Tel : +33 (0) 467 61 85 36
ax : +33 (0) 467 61 37 87 F
The release of cytochrome c from the mitochondrial intermembrane space is a decisive event in programmed cell death. Once in the cytoplasm, cytochrome c is involved in the formation of the macromolecular complex termed apoptosome, which activates procaspase-9 which in turn activates downstream procaspase-3. There are increasing evidences indicating that cyclophilin A is highly expressed in many tumours and cell lines where it exerts an anti-apoptotic function. In brain tissue, which over-expresses constitutively cyclophilin A, we evidenced mixed dimers composed of cyclophilin A and cytochrome c. In a cell free system we observed that pure cyclophilin A inhibited cytochrome c-dependent procaspase-3 activation. Moreover, we evidenced cyclophilin A-cytochrome c complexes within the cytoplasm of HCT116 cells following staurosporine-induced apoptosis. Our results strongly support that, in tumour cells, cyclophilin A is able to inhibit procaspase-3 activation by sequestering cytochrome c.
Apoptosis, brain, tumour, cyclophilin A, cytochrome c, procaspase-3,
Cyt c: cytochrome c
Cyp A: cyclophilin A
PEBP: phosphatidylethanolamine-binding protein
CsA: cyclosporin A
AMC: aminomethyl coumarin
DEVD-AMC: Acetyl-L-aspartyl-L-glutamyl-L-valyl-L-aspartic acid ;- (4-methyl-coumaryl-
EDAC: N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride
Signalling for apoptosis occurs through multiple pathways that are initiated either by triggering events within the cell (DNA damage, cytoskeleton disruption) or from outside the cell by ligation of death receptors . Cells respond by an active process resulting in characteristic morphological changes (DNA fragmentation and membrane blebbing). The degradation of nuclear DNA is due to caspase-dependent as well as caspase-independent DNAse activation . The proteolytic cleavages occurring during apoptosis are mediated at least in part by a group of proteases termed caspases, which are synthesized as inactive proenzymes. Two principal mechanisms, the mitochondrial and the death receptor pathways, are able to trigger caspase activation. The intrinsic apoptotic pathway involves the permeabilization of the outer mitochondrial membrane and the release of cytochrome c (Cyt c). This hemoprotein participates to the formation of a multimeric complex termed apoptosome, which activates procaspase-9, which in turn activates downstream procaspase-3.
Cell death or survival is determined by a balance between pro- and anti-apoptotic molecules, which are under the control of extra- as well as intracellular signals. In addition, apoptosis is modulated by a set of proteins, which are not specific of programmed cell death. For instance, in differentiated and non-dividing cells such as neurons and muscular cells, molecular chaperones are critical for maintaining cellular homeostasis. The heat shock protein family (HSP), known to control protein folding by chaperone activity, was shown to down
regulate apotosis . Moreover, in tumour cells there are increasing evidences indicating that cyclophilin A (Cyp A), a cyclosporin A-binding protein with chaperone function, can participate to cell death regulation. It has been reported that Cyp A is overexpressed in hepatocellular carcinoma , in small and non-small cell lung cancer [5,6] in pancreatic denocarcinoma , in endometrial carcinoma  as well as in esophageal cancer . It was a
shown that stable RNA interference-mediated suppression of Cyp A diminishes non-small-cell lung tumour growth in vivo . Moreover it was observed that over-expression of Cyp A, in human prostate cancer cell line, prevented hypoxia- and cisplatin-induced apoptosis, whereas small interfering RNA-based Cyp A knockdown reversed this effect . A biological mechanism accounting for the anti-apoptotic effect of Cyp A was recently reported . The authors evidenced, in interleukin-6-dependent multiple myeloma cells, that cyclophilins A and B support the anti-apoptotic action of signal transducer and activator of transcription 3 (Stat3).
In this paper we show that Cyp A could control cell survival at a different level. Brain tissue expresses constitutively Cyp A in high amount . We evidenced in neuron subcellular extracts stable heterodimers composed of Cyp A and Cyt c. We measured the inhibiting effect of purified Cyp A on Cyt c-dependent procaspase-3 activation in vitro. We
determined the Kd of the Cyt c-Cyp A complex and we evidenced the formation of Cyt c-Cyp
A heterodimers within the cytoplasm of apoptotic cells. Our experimental results strongly support that the anti-apoptotic effect of Cyp A may be due to the ability of Cyp A to bind free cytoplasmic Cyt c.
2 MATERIALS AND METHODS
Mouse monoclonal clone 7H8 2C12 anti-Cyt c antibodies were from Bioscience (Interchim, France). Rabbit polyclonal antibodies against Cyp A were from Chemicon-
tate (Southampton, UK). Anti-phosphatidylethanolamine-binding protein (PEBP) IgG Ups
were kindly provided by Dr Françoise Schoentgen from CNRS (Orléans, France).
Electrophoresis and Western blots.
Proteins were resolved alternatively in denaturing and semi-denaturing 15% polyacrylamide gels. The first method was conducted essentially as described by Laemmli  in a discontinuous buffer system. The second electrophoretic technique was performed in a continuous buffer system (25 mM Tris, 192 mM glycine, pH 8.6, 0.1% SDS). In this case the samples were treated for 3 min at 100?C in a non-reducing solubilization solution containing 5% SDS.
Proteins resolved in either semi- or in fully-denaturing gels were blotted on a PVDF membrane (Immobilon P from Millipore, Bedford, MA) with a Trans-Blot SD semi-dry apparatus from BioRad (Hercules, CA). Membranes were probed for 2 hours with either anti-Cyt c (1/5000), anti-Cyp A (1/1000), anti-PEBP (1/1000) antibodies. Then the blots were incubated with the appropriate secondary antibody (1/2000) conjugated with peroxidase.
Measurement of caspase-3 activity.
Caspase activity was assayed by measuring the cleavage of the fluorophore aminomethyl coumarin (AMC) from DEVD-AMC (Acetyl-L-aspartyl-L-glutamyl-L-valyl-L-aspartic acid ;- (4-methyl-coumaryl-7-amide)) substrate . We prepared S-100 cytosolic fractions from HeLa cells according to the protocol described by Liu et al. . The cells
were suspended in HK buffer (20 mM Hepes/KOH, pH 7.5, 10 mM MgCl, 1 mM EDTA, 1 2
mM EGTA) supplemented with protease inhibitors (0.1 mM PMSF, 5 ，g/ml pepstatin A, 10
，g/ml leupeptin, 2 ，g/ml aprotinin and 25 ，g/ml ALLN). The suspension was introduced in a
Mini-Bomb cell disruption chamber (Kontes, Vineland, NJ). A 15 bar nitrogen pressure was applied to the chamber for 30 min. Cell debris were eliminated by centrifugation at 15 000 g for 10 min. Then the solution was centrifuged at 100 000 g for 60 min. The S-100 supernatant was stored at –80?C.
The Cyt c concentration was determined with a MC2 (Safas, Monaco)
spectrophotometer from the sodium dithionite reduced spectrum, using an absorbance
414-600-1-1coefficient ?？ 101 mM cm.
Caspase-3 activity was measured according to the conditions described by Hampton et
al. . The activation of procaspase was conducted at 37?C. A sample of S-100 supernatant (50 ，g of protein) was mixed with 1 mM dATP and 400 nM of pure Cyt c, in some experiments bovine albumin, Cyp A or PEBP were added in the medium. The final volume was adjusted to 70 ，l with buffer HK. At the end of the incubation, DEVD-AMC (50 ，M)
was added to the medium. The AMC fluorescence was monitored with a luminescence spectrometer Perkin Elmer (Buckinghamshire, England) ？ 354 nm and ？ 460 excitationemission
nm. The measure was standardised with a solution of 1 ，M AMC.
Determination of Kd values.
Surface plasmon resonance analysis: The kinetic parameters were
determined with a Biacore 3000 instrument (GE Healthcare, Biacore AB Uppsala, Sweden)
Equilibrium chromatography method: We used the method of Hummel and
Dreyer  with the technical improvements indicated by Berger and Girault .
Induction of apoptosis in HCT 116 cells:
HCT116 colon adenocarcinoma cells were grown in complete medium, i.e., RPMI 1640 supplemented with 10% calf fetal serum and 2 mM L-glutamine at 37 ?C under a humidified atmosphere with 5% CO. When the cells were at 80% confluence apoptosis was 2
induced by treatment with 0.3 ，M staurosporine for 24 hours. The cells were disrupted under nitrogen pressure in a Mini-Bomb chamber as indicated above. The supernatant obtained by centrifugation for 30 min at 18 000 g was stored at –80?C.
Cross-linking Cyt c-protein complexes.
20 ，l of 0.1 M N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride
(EDAC) were added to 100 ，l of HCT116 cells supernatant (4 mg/ml). The reaction was
developed for 1 to 3 hours at room temperature and stopped with 20 ，l of 0.1 M ethanolamine.
The covalent protein complexes were separated in 15% SDS-PAGE in denaturing conditions and blotted on a PVDF membrane.
3 1 Effect of Cyp A on Cyt c-dependent procaspase-3 activation
In previous experiments we evidenced the ability of Cyt c to link small proteins, which were not directly involved in the apoptotic program, while we studied proteins of brain cell cytosol. These results are detailed in supplementary file 1. Briefly, from 115 000 g brain-cell supernatant we isolated a protein fraction, termed fraction R, containing two heterodimers of Cyt c exhibiting molecular masses of 30 and 35 kDa (figure 1A). These molecular complexes were disrupted in the presence of a disulfide reducing agent (mercaptoethanol). We later identified the Cyt c-associated proteins as being respectively Cyp A and PEBP isoform-1 (figure 1B). We investigated the ability of fraction R to trigger caspase-3 activity in vitro, by
comparison to pure Cyt c (data not shown). We observed that fraction R was less active than Cyt c to promote the reaction, the activity being about 30 per cent of that obtained with Cyt c. This lack of activity could result of the fact that part of the Cyt c in fraction R is immobilized in molecular complexes and had no access to the caspase activation cascade. However, fraction R was a mixture of at least 15 proteins and we could not exclude the presence of inhibitors of the reaction. In consequence we decided to investigate caspase activation in the presence of purified proteins.
We purified Cyp A and PEBP from fraction R by liquid chromatography (see
supplementary file 2). We measured procaspase-3 activation triggered by Cyt c in vitro in a
cell free system. The influence of Cyp A on the reaction is indicated in figure 2A. Cyp A inhibits the reaction at low concentration. It is active in the range of 0.3 to 0.6 ，M. The mean
value is roughly stoichiometric with the amount of Cyt c present in the reaction (0.4 ，M).
Cyp A is the intracellular receptor of cyclosporin A (CsA) and exerts a peptidyl-prolyl isomerase function (PPIase). The catalysis of prolyl isomerization by Cyp A is strongly inhibited by low amounts of CsA . We tested the effect of increasing amounts of CsA on Cyp A inhibition of Cyt c-dependent procascaspase-3 activation. From figure 2B, CsA is not able to suppress the inhibiting effect of Cyp A, indicating that the ability of Cyp A to bind Cyt c is independent of its PPIase activity. Cyp A, in addition to its PPIase activity, is known to exert a chaperone function within cells , its ability to link Cyt c could result of this latter function.
Concerning PEBP we saw that it was able to inhibit procaspase-3 activation (data not shown). However, it was less efficient than Cyp A since it requires a tenfold excess (from 3 to 6 ，M) to inhibit the effect of Cyt c.
3 2 Determination of the Kd of the molecular complexes with Cyt c.
The binding affinity of Cyt c for Cyp A was measured by surface plasmon resonance analysis (see supplementary file 3). The sensorgrams led to calculate a Kd value of 16.0?1.3
2-1-1，M, the association and dissociation rate constants being respectively 1.05?0.08 10 M s
-3 -1and 1.67?0.02 10s. We cannot visualize the Cyt c binding on immobilized PEBP. This probably reflects an adverse effect of the chemical immobilization of PEBP. To circumvent this problem, we used an equilibrium chromatographic technique to determine the Kd values
values obtained are 17.3?4.8 ，M and 9.7?3.8 ，M of the proteins in solution. The Kd
respectively for the complexes Cyp A-Cyt c and PEBP-Cyt c. For Cyp A-Cyt c interaction the affinities obtained by the two methods are in a same range
We determined the concentrations of the two proteins in pig brain, Cos and Hela cell cytosols by densitometry of blots standardized with samples of pure Cyp A or PEBP. We found that CypA was abundant in the cell lines. Its concentration expressed in ，M was 17.6 ?
2.4 and 14.5 ? 0.9 respectively in HeLa and Cos cells while it was about half in pig brain (8.8 ? 2.3). It was reported by others  that there is a two to three fold increase of Cyp A amount when comparing tumour versus normal tissues. If we assume that the concentration of Cyp A measured in the cytosol of the cell lines, is indicative of the concentration present within tumour tissues, we can conclude that Cyp A is able to combine about half of the cytosolic Cyt c since the Kd value of the complex of the two proteins is in the same range (15 to 20 ，M).
In contrast we deduced that the ability of PEBP to associate Cyt c in the cell lines was very low since we found 3.1 ? 0.2 and 0.5 ? 0.1 ，M of PEBP respectively in HeLa and Cos
cells cytosols . However, PEBP is highly concentrated brain cytosol (10.8 ? 2.1 ，M). Its
ability to combine free Cyt c in this tissue is not negligible since the Kd of the complex of the two proteins is very similar.
3 3 Detection of Cyp A/Cyt c heterodimers in apoptotic HCT116 cells.
3 3 1 Expression of Cyp A in colorectal cancer.
The protein amount of Cyp A within various tumours, was investigated by Koletsky et
al. . These authors reported that Cyp A concentration in colon adenocarcinomas was twofold to threefold greater than that found in adjacent normal tissue. In our laboratory, we obtained data corroborating these results from a previous study on gene expressed in advanced colorectal cancer (see supplementary file 4).
3 3 2 Detection of Cyt c/Cyp A protein complexes in apoptotic cells.
Cell death was induced in HCT116 cell cultures by treatment with staurosporine. The induction of apoptosis by this drug was controlled both by flow cytometry following double labelling of the cells and by measuring Cyt c efflux from mitochondria (Figure 3A). A sample of the cytosol of apoptotic HCT116 cells was cross-linked with the carbodiimide reagent EDAC. The cross-linked polypeptides were then resolved in 15% polyacrylamide gel and blotted. As indicated in figure 3B, the blot shows that one protein band cross-reacts both with anti-Cyt c and anti-Cyp A antibodies. It is likely a dimer of Cyp A and Cyt c since it exhibits an apparent molecular mass of 32 kDa. Concerning PEBP we could not evidence mixed molecular complexes with Cyt c on the blots.
In this study we evidenced that both PEBP and Cyp A possess the property to link Cyt c with Kd values in the range of 10 to 20 ，M. This affinity for a key component of the
apoptotic machinery could lead to the inhibition of apoptosis in cells overexpressing these proteins.
The PEBP protein family is a highly conserved group of proteins found in a great variety of organisms from plants to mammals . PEBP-1, a 23 kDa basic protein, was