Electrophoretic Analysis of the Mitochondrial Outer Membrane

May 30, 2008 - Electrophoretic Analysis of the Mitochondrial Outer Membrane Rupture Induced by Permeability Transition. Hans Zischka* ... Citation dat...
7 downloads 13 Views 2MB Size
Download PDF
Anal. Chem. 2008, 80, 5051–5058

Electrophoretic Analysis of the Mitochondrial Outer Membrane Rupture Induced by Permeability Transition Hans Zischka,*,† Nathanael Larochette,‡,§,| Florian Hoffmann,⊥ Daniela Hamo ¨ ller,† † † # Nora Ja¨gemann, Josef Lichtmannegger, Luise Jennen, Josef Mu¨ller-Ho¨cker,∇ Frigga Roggel,∇ Martin Go¨ttlicher,†,[ Angelika M. Vollmar,⊥ and Guido Kroemer‡,§,| Institute of Toxicology, Helmholtz Center Munich, German Research Center for Environmental Health, 85764 Oberschleissheim, Germany, INSERM, U848, Institut Gustave Roussy, University Paris-Sud, Paris-11, 94805 Villejuif, France, Department of Pharmacy, Ludwig-Maximilians University Munich, 81377 Munich, Germany, Institute of Pathology, Helmholtz Center Munich, German Research Center for Environmental Health, 85764 Oberschleissheim, Germany, and Institute of Pathology, Ludwig-Maximilians University Munich, 80337 Munich, Germany, and Institute of Toxicology, Technische Universität München, 80802 Munich, Germany A pathological increase of the permeability of the mitochondrial membranes may culminate in the irreversible rupture of the mitochondrial outer membrane. Such a permeability transition is lethal because it results in the release of death-inducing molecules from mitochondria and/or metabolic failure. Current methods to assess this outer membrane damage are mostly indirect or scarcely representative of the overall mitochondrial population. Here we present an analytical and preparative approach using free flow electrophoresis to directly distinguish rat liver mitochondria that have undergone the permeability transition from unaffected organelles or from organelles that are damaged to a minor degree. Mitochondrial populations, which considerably differ in outer membrane integrity or cytochrome c content, were separated by this means. We further show that the relative abundance of each population depends on the dose of the permeability transition inducer and the duration of the treatment time. Finally, we have employed this approach to investigate the impairment of mitochondria that were isolated from livers subjected to ischemia/reperfusion damage. The permeabilization of mitochondrial membranes is a critical event of cell death pathways culminating in either apoptosis or necrosis.1–3 One peculiar mode of membrane permeabilization is * To whom correspondence should be addressed. Phone: ++49 89 3187 2663. Fax: ++49 89 3187 3449. E-mail: [email protected]. † Institute of Toxicology, Helmholtz Center Munich. ‡ INSERM, U848. § Institut Gustave Roussy. | University Paris-Sud. ⊥ Department of Pharmacy, Ludwig-Maximilians University Munich. # Institute of Pathology, Helmholtz Center Munich. ∇ Institute of Pathology, Ludwig-Maximilians University Munich. [ Institute of Toxicology, Technische Universität München. (1) Bernardi, P.; Krauskopf, A.; Basso, E.; Petronilli, V.; Blachly-Dyson, E.; Di Lisa, F.; Forte, M. A. FEBS J. 2006, 273, 2077–2099. (2) Grimm, S.; Brdiczka, D. Apoptosis 2007, 12, 841–855. (3) Kroemer, G.; Galluzzi, L.; Brenner, C. Physiol. Rev. 2007, 87, 99–163. 10.1021/ac800173r CCC: $40.75  2008 American Chemical Society Published on Web 05/30/2008

the so-called “mitochondrial permeability transition” (PT),4 a process that can be inhibited by submicromolar concentrations of the cyclophilin D (Cyp D) inhibitor cyclosporine A (CsA)5 and that involves the creation of a pore/channel-like structure.6,7 As a result of the colloid-osmotically driven influx of water into the mitochondrial matrix through the inner mitochondrial membrane4 (“swelling”), mitochondria undergo a transition from the “aggregated” to the “orthodox” state,4 culminating in the distension and later rupture of the outer membrane and hence mitochondrial outer membrane permeabilization (MOMP).8–10 It is generally accepted that the PT is an “all-or-nothing” event,8 which can be triggered by a plethora of distinct stimuli.11 However, the molecular constituents of the PT pore have remained elusive,1,2,12,13 and it is still unknown whether mitochondria that have undergone PT are all equally damaged or whether different stages of damage exist. This question is of particular importance with respect to the ongoing discussion about the implication of PT in distinct cell death scenarios.1–3,11,14–16 Substantial mitochondrial damage may cause ATP depletion and necrosis. Partial damage or damage in which some mitochondria maintain their membrane potential and produce ATP may drive apoptotic cell death. Minor damage that only affects few mitochondria can simply provoke the removal of the damaged organelles by (4) Hunter, D. R.; Haworth, R. A.; Southard, J. H. J. Biol. Chem. 1976, 251, 5069–5077. (5) Crompton, M.; Ellinger, H.; Costi, A. Biochem. J. 1988, 255, 357–360. (6) Sorgato, M. C.; Keller, B. U.; Stuhmer, W. Nature 1987, 330, 498–500. (7) Szabo, I.; Zoratti, M. J. Bioenerg. Biomembr. 1992, 24, 111–117. (8) Petronilli, V.; Cola, C.; Massari, S.; Colonna, R.; Bernardi, P. J. Biol. Chem. 1993, 268, 21939–21945. (9) Vander Heiden, M. G.; Chandel, N. S.; Williamson, E. K.; Schumacker, P. T.; Thompson, C. B. Cell 1997, 91, 627–637. (10) Brenner, C.; Grimm, S. Oncogene 2006, 25, 4744–4756. (11) Green, D. R.; Kroemer, G. Science 2004, 305, 626–629. (12) Saris, N. E.; Carafoli, E. Biochemistry (Moscow, Russ. Fed.) 2005, 70, 187– 194. (13) Baines, C. P.; Kaiser, R. A.; Sheiko, T.; Craigen, W. J.; Molkentin, J. D. Nat. Cell Biol. 2007, 9, 550–555. (14) Halestrap, A. P. Biochem. Soc. Trans. 2006, 34, 232–237. (15) Kinnally, K. W.; Antonsson, B. Apoptosis 2007, 12, 857–868. (16) Tsujimoto, Y.; Shimizu, S. Apoptosis 2007, 12, 835–840.

Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

5051

autophagy.14,17 Thus, accurate methods of PT detection might furnish important information on the fate of damaged cells.1,18 Here, we report on a novel approach to assess (PT-induced) MOMP. This technology is based on the difference in surface charge between the inner and outer mitochondrial membranes of isolated rat liver mitochondria. Heidrich et al. have reported in 1970 that inner membrane-matrix vesicles (IMV) generated from rat liver mitochondria display a stronger deflection in an electrical field toward the anode as compared to the intact organelles.19 With the use of zone electrophoresis in a free flow electrophoresis device (ZE-FFE), we and others have recently reported that ZE-FFE may be an adequate tool for the preparative analysis of mitochondria isolated from yeast20,21 or plants.22 In ZE-FFE, charged particles are injected into a laminar buffer stream, deflected by a perpendicular electrical field, and collected at the end of the separation chamber. Here, we applied ZE-FFE to rat liver mitochondria and confirmed that ZE-FFE detects differences in their surface charge between the inner and the enveloping outer membrane. We therefore reasoned that ZE-FFE could detect whether mitochondria have undergone a PT-induced MOMP, due to the resulting exposure of more negative inner membrane charges. Upon PT, damaged mitochondrial populations can be separated and collected by ZE-FFE, and the relative abundance of these populations depends on the intensity and the duration of PT induction. We show that this method can be employed to analyze mitochondrial damage after liver insult by ischemia/reperfusion, a clinically relevant tissue-damaging process, which has been reported to cause PT-induced MOMP.14,23,24 EXPERIMENTAL PROCEDURES Isolation of Rat/Mouse Liver Mitochondria. Mitochondria were isolated by differential centrifugation according to standard protocols. Briefly, freshly removed liver tissue was homogenized with a glass Teflon homogenizer in isolation buffer (10 mM triethanolamine (TEA), 10 mM acetic acid (HAc), 280 mM sucrose, 0.2 mM EGTA, pH 7.4 with KOH). Homogenates were cleared from debris and nuclei by two times centrifugation at 750g (10 min at 4 °C), and mitochondria were pelleted at 9000g (10 min at 4 °C). Organelles were washed three times (once at 9000g and two times at 15 000g, 10 min, 4 °C) and resuspended in FFE separation buffer (10 mM TEA, 10 mM HAc, 280 mM sucrose, pH 7.4 with KOH). Ischemia/Reperfusion Treatment. Six week old male Sprague-Dawley rats (Charles-River Laboratories, Sulzfeld, Germany) were anesthetized by ip injection of 0.005 mg/kg fentanyl (Janssen-Cilag, Neuss, Germany) and 2.0 mg/kg midazolam (Ratiopharm, Ulm, Germany). Anaesthesia was maintained by 1.5% isoflurane (Abbott, Wiesbaden, Germany) using a vaporizer with carbogen (5% CO2/95% O2, Air-Liquide, Duesseldorf, Germany); body temperature was maintained with a warming lamp. The (17) Ott, M.; Gogvadze, V.; Orrenius, S.; Zhivotovsky, B. Apoptosis 2007, 12, 913–922. (18) Bernardi, P.; Scorrano, L.; Colonna, R.; Petronilli, V.; Di Lisa, F. Eur. J. Biochem. 1999, 264, 687–701. (19) Heidrich, H. G.; Stahn, R.; Hannig, K. J. Cell Biol. 1970, 46, 137–150. (20) Zischka, H. Proteomics 2003, 3, 906–916. (21) Zischka, H. Mol. Cell. Proteomics 2006, 5, 2185–2200. (22) Eubel, H. Plant J. 2007, 52, 583–594. (23) Saris, N. E.; Eriksson, K. O. Acta Anaesthesiol. Scand. Suppl. 1995, 107, 171–176. (24) Weiss, J. N.; Korge, P.; Honda, H. M.; Ping, P. Circ. Res. 2003, 93, 292– 301.

5052

Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

abdomen was opened by midline laparotomy, and the portal triad was prepared. The arterial and portal blood flow to the left lateral and median lobe of the liver was interrupted by applying an atraumatic clip, resulting in a 70% liver ischemia. After 90 min of ischemia, the blood supply was restored by removal of the clip and the reperfusion period was initiated. Animals were sacrificed after 3 h of reperfusion by bleeding. Control animals were sacrificed directly after the midline laparotomy. The livers were rinsed free from blood by perfusing the organs with phosphatebuffered saline (PBS). Mitochondria from the left lateral and median lobe of the liver were isolated by differential centrifugation as above with the modification that mitochondria were pelleted from the 750g supernatant at 15 000g (10 min at 4 °C) and washed once at 15 000g. Mitochondria were stored in liquid nitrogen according to Fleischer25 until further use. For ZE-FFE experiments stored mitochondria were thawed, adapted to FFE separation buffer by washing, and subsequently analyzed. For electron scan microscopy tissue samples were fixed in 3% glutaraldehyde for 2 h and then kept in 1% glutaraldehyde until further examination. All animals received humane care in compliance with the “Principles of Laboratory Animal Care”. Studies were registered and approved by the government authorities. In Vitro Treatments of Isolated Rat Liver Mitochondria. IMVs were prepared from mitochondria by osmotic swellingshrinking as described by Heidrich et al.19 PT-induced swelling of freshly isolated rat liver mitochondria suspensions upon Ca2+ addition was routinely measured by light scattering at 540 nm in a microplate absorbance reader (µ-Quant, Bio-Tek, Bad Friedrichshall, Germany) over a period of 30 min at room temperature (RT). The final assay volume was 200 µL, containing mitochondria at 0.5 mg/mL in “standard swelling buffer” (10 mM MOPS Tris, pH 7.4, 200 mM sucrose, 5 mM succinate, 1 mM Pi (inorganic phosphate), 10 µM EGTA, and 2 µM rotenone). CsA (5 µM) was added 5 min before Ca2+. Intactness of the isolated mitochondria was routinely checked by standard respiratory measurements, and only mitochondrial preparations that showed Ca2+-induced PT which could be inhibited by CsA were used for ZE-FFE experiments. To study the PT by ZE-FFE 10 mg of mitochondria were incubated in 20 mL of “FFE-compatible swelling buffer” (10 mM TEA, 10 mM HAc, 280 mM sucrose, pH 7.4 with KOH, 5 mM succinate, 1 mM Pi, 10 µM EGTA, and 2 µM rotenone) at RT. CsA/Ca2+-inhibited mitochondria were treated with CsA (0.6-4 µM) for 5 min, and Ca2+ (25 nmol/mg mitochondria) was subsequently added with 10 min further incubation. FCCPuncoupled mitochondria were incubated in FFE-compatible swelling buffer without succinate and treated with FCCP (2 µM) for 5 min and either pelleted (15 000g, 10 min, 4 °C) and analyzed by ZE-FFE or treated with Ca2+ (25 nmol/mg mitochondria) and incubated for an additional 10 min. Ca2+-treated mitochondria were incubated with different Ca2+ doses or treatment times as indicated in the text. Treated mitochondria were softly homogenized to ensure proper mixing, pelleted (15 000g, 10 min, 4 °C), resuspended, and analyzed by ZE-FFE. Untreated reference mitochondria were incubated in FFE-compatible swelling buffer alone, pelleted after 5 min, and analyzed by ZE-FFE. (25) Fleischer, S. Methods Enzymol. 1979, 55, 28–32.

ZE-FFE Analysis of Mitochondrial Samples. The mitochondrial samples suspended in FFE separation buffer were analyzed in a free flow electrophoresis apparatus (FFE-Weber GmbH, Planegg, Germany) under the following conditions: (i) circuit electrolyte solution was 100 mM HAc, 100 mM TEA adjusted to pH 7.4 with KOH; electrolyte stabilizing solution was 100 mM HAc, 100 mM TEA, 0.28 M sucrose, pH 7.4; FFE separation buffer was 10 mM TEA, 10 mM HAc, 280 mM sucrose, pH 7.4; (ii) all buffers were cooled on ice during separation to avoid thermal gradients; (iii) mitochondrial samples were applied at the cathodal side into ZE-FFE with a sample flow rate of 1-2 mL/h; (iv) electrophoresis was performed in horizontal mode at 5 °C with a flow rate of 330 mL/h and a voltage of 750 V; (v) fractions were collected in 96-well plates, and the distribution of resolved particles was monitored at a wavelength of 260 nm with a µ-Quant microplate reader (Bio-Tek instruments, Bad Friedrichshall, Germany); relative abundances of the separated mitochondrial populations (Figure 3) were calculated from their respective OD 260 nm (optical density) peak values taken from the ZE-FFE separation spectra; (vi) ZE-FFE-separated mitochondrial fractions were concentrated by centrifugation at 4 °C (10 min at 16 000g); (vii) routine tests of the ZE-FFE apparatus were conducted as recently published;21 (viii) comparisons of mitochondrial populations, e.g., dose-dependent Ca2+-treated mitochondria, were done consecutively on the same day, under identical conditions regarding voltage, buffer composition, buffer and sample velocity, and temperature. From each sample, several separation profiles were recorded to ensure time stability and exclude drifts in separation conditions. Electron Microscopy. ZE-FFE-separated mitochondrial fractions were immediately pelleted, fixed in 3% glutaraldehyde, postfixed with 1% osmium tetroxide, dehydrated with ethanol, and embedded in Epon. Ultrathin sections were negative stained with uranyl acetate and lead citrate and then analyzed on a Zeiss EM 10 CR electron microscope. Glutaraldehyde-fixed liver tissues were treated similarly but dehydrated with acetone and analyzed on a Philips EM 420 electron microscope. Miscellaneous. Quantification of mitochondrial protein amounts was done by the Bradford assay. Immunoblotting analysis of ZE-FFE-separated mitochondrial protein extracts was carried out according to Towbin et al.26 Equal protein amounts were applied per lane, and blots were routinely stained with Ponceau red to control for proper transfer. For protein detection polyclonal antisera against VDAC (voltage-dependent anion-selective channel protein, Acris) and monoclonal antibodies against Cyt c (cytochrome c), TIM 23 (mitochondrial import inner membrane translocase subunit TIM 23), HSP 60 (heat shock protein (60 kD), BD), and Cyp D (Acris) were used. Glycogen detection was done by color development upon mixing 200 µL of the separated ZE-FFE fractions with 50 µL of Lugol’s iodine solution (Fluka, Germany).27 RESULTS AND DISCUSSION Discrimination of Intact Rat Liver Mitochondria and Inner Membrane Vesicles by ZE-FFE. Incubation of purified mitochondria in low- and high-osmolar sucrose solutions ruptures the (26) Towbin, H.; Staehelin, T.; Gordon, J. Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 4350–4354. (27) Govons, S.; Vinopal, R.; Ingraham, J.; Preiss, J. J. Bacteriol. 1969, 97, 970– 972.

Figure 1. Discrimination of liver mitochondria and inner membrane vesicles by ZE-FFE. (A) ZE-FFE separation profiles of untreated rat/ mouse liver reference mitochondria showed one major peak (R). Several peaks which are deflected more towards the anode were observed for IMVs (one representative separation out of six independent experiments is shown). (B) Electron micrographs of ZE-FFEpurified IMVs and rat liver reference mitochondria.

mitochondrial outer membrane and generates IMVs,19 which expose a more negative surface charge than untreated mitochondria with an intact outer membrane.19 Electrophoresis (ZE-FFE) of iso-osmotic “reference” mitochondria resulted in a single major peak that we termed the R peak (Figure 1A), comprising mainly intact mitochondria as confirmed by electron microscopy (Figure 1B). In contrast, osmotically treated mitochondria deflected more toward the anode and demonstrated IMV morphology (shown for the most anodal peak in Figure 1B). Beside the major anodal IMV peak some smaller IMV peaks were detected in ZE-FFE (Figure 1A), and first data indicate that they differ in their amount of residual outer membrane (not shown). The mechanisms accounting for the generation of these discrete IMV peaks remain to be studied. Regardless, IMV exhibit an anodal deflection, demonstrating that rupture and/or (partial) removal of the outer mitochondrial membrane can be monitored by ZE-FFE. Preparative Analysis of Rat Liver Mitochondria Undergoing PT by ZE-FFE. Isolated mitochondria undergo PT upon the addition of Ca2+ in the presence of phosphate in vitro, and this process can be inhibited by CsA.5 PT is typically monitored as a Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

5053

decrease in light scattering caused by mitochondrial swelling28 (Figure 2A). Due to its limited extensibility, the outer membrane ruptures (MOMP) upon matrix swelling,9 exposing parts of the inner membrane surface and possibly generating more negatively charged organelles. Consequently, we assessed whether a PTinduced MOMP can be monitored by ZE-FFE. Isolated rat liver mitochondria were energized with succinate, challenged for 20 min at RT with Ca2+ (25 nmol/mg mitochondria), and subsequently analyzed by ZE-FFE (Figure 2B). Energized mitochondria from the same preparation without Ca2+ addition were chosen to assess the ZE-FFE deflection of undamaged “reference” mitochondria (R). Energized mitochondria pretreated with CsA before Ca2+ addition (“CsA/Ca2+ mitochondria”) were used as PT-inhibited control. In order to assess the effect of passive Ca2+ binding on the outer membrane surfaces in ZE-FFE, mitochondria were pretreated with the protonophore FCCP before Ca2+ addition (“FCCP-uncoupled mitochondria ± Ca2+”) since FCCP dissipates the mitochondrial inner membrane potential (∆Ψm) and thus abolishes the electrophoretic Ca2+ uptake into the mitochondrial matrix. Reference mitochondria eluted in one major R peak (Figure 2B). Ca2+-treated mitochondria, however, deflected more toward the anode into several peaks, which were labeled M1, M2, and M3 from the cathode to the anode, respectively (Figure 2B). As compared to the R peak, an intensity decrease was noted for the M1 peak and a peak broadening of around two fractions toward the anode (i.e., anodal shoulder of M1, Figure 2B). Unlike the Ca2+-treated mitochondria, CsA/Ca2+ and FCCP-uncoupled mitochondria ± Ca2+ eluted in one peak which was highly similar in deflection and height to that of the reference mitochondria (Figure 2, parts B and C, respectively). Electron microscopy of the M1 peak revealed that it contained morphologically heterogeneous mitochondria, comprising a minority of apparently intact mitochondria and a majority of mitochondria with dilated matrixes and disorganized cristae (cp. Figure 2D and Figure 1B). M1 mitochondria were still surrounded by an outer membrane which was, however, damaged to different extents (Figure 2D, see also the immunoblotting analyses in Figure 3B). M2 and M3 mitochondria exhibited IMV morphology (Figure 2D). Slightly higher amounts of residual outer membrane were detected in M2 than in M3 (arrows in Figure 2D, see also Figure 3B). In contrast to Ca2+-treated mitochondria, CsA/Ca2+ and FCCP-uncoupled mitochondria ± Ca2+ had intact outer membranes and electron-dense matrixes (Figure 2D), yet slightly altered cristae structures were observed in the latter as compared to reference mitochondria (cp. Figure 1B and Figure 2D). Taken together, these results demonstrate that the mitochondrial populations M1-M3 result from Ca2+-induced PT, as they are not observed upon PT inhibition by CsA. Furthermore, mitochondria which have undergone a full-blown PT-induced MOMP (i.e., M2 and M3) can be clearly separated by ZE-FFE from intact mitochondria or mitochondria damaged to a minor degree (i.e., M1). The anodal deflection of M1-M3 cannot be attributed to a dissipation of the ∆Ψm nor to the passive binding of Ca2+ to the outer surfaces of mitochondrial membranes (because FCCP treatment ± Ca2+ of mitochondria does not cause (28) Beatrice, M. C.; Palmer, J. W.; Pfeiffer, D. R. J. Biol. Chem. 1980, 255, 8663–8671.

5054

Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

any anodal deflection). Since M1-M3 demonstrated different levels of PT-induced MOMP, it appears plausible that their variable degree of anodal deflection reflects different levels of exposure of negative surface charges of the inner membrane, similar to the results observed with IMVs (see below). Ca2+-Induced PT Causes Variable Degrees of Mitochondrial Damage in a Dose- and Time-Dependent Fashion. PTassociated mitochondrial swelling is typically assessed by a decrease in the OD at 540 nm and occurs above a certain threshold of Ca2+ (Figure 2A). Depending on the Ca2+ dose, different kinetics of swelling are observed. The higher the Ca2+ dose, the more rapid is the reduction of the OD 540 (Figure 2A). Such optical density measurements result from the summation of the response of individual mitochondria to Ca2+. We employed ZE-FFE to monitor whether PT induction by Ca2+ might result in the generation of heterogeneously damaged mitochondria. Energized mitochondria were incubated for 20 min at RT with Ca2+ doses ranging from 12.5 to 50 nmol Ca2+/mg mitochondria and subsequently analyzed by ZE-FFE. At a dose of 12.5 nmol Ca2+/mg mitochondria, one major peak was detectable in ZEFFE (i.e., M1, Figure 3A). At 25 nmol Ca2+/mg, a reduction of M1 together with a broadening toward the anode (i.e., M1 anodal shoulder, see Figure 2B) as well as the emergence of M2 and M3 was noted (Figure 3A). Further depletion of M1 coinciding with an increase of M2 and M3 was found at 37.5 nmol Ca2+/mg and at 50 nmol Ca2+/mg (Figure 3A). It should be noted that the sensitivities of mitochondria toward Ca2+-induced PT differed somewhat between the individual preparations. The given Ca2+ doses should thus not be misconceived as absolute values but rather as the doses and sensitivities typically observed in the majority of these experiments. A major advantage of ZE-FFE is its preparative character, enabling downstream analyses of the separated mitochondrial subpopulations. Immunoblotting analysis, for example, of the organelles collected from these peaks (Figure 3B) demonstrated significant differences in the amount of outer membrane (VDAC) and intermembrane space (Cyt c) proteins but showed relatively constant amounts of inner membrane (TIM 23) and matrix proteins (HSP 60 and Cyp D). With respect to the reference mitochondria (R), the more anodal the subfraction was collected the more substantial was the depletion in VDAC (Figure 3B, cp. M3 and M2 at 50 nmol Ca2+/mg mitochondria). No Cyt c was detected in mitochondria collected from the M3 and M2 mitochondria, while Cyt c was still present in the anodal shoulder of M1, though at reduced levels as compared to M1 (Figure 3B). To study the time-dependent alteration of mitochondrial subpopulations upon PT induction by ZE-FFE, energized mitochondria were incubated with 20 nmol Ca2+/mg mitochondria for 5-60 min and subsequently analyzed (Figure 3C). A decrease of M1 and the emergence of M2 mitochondria were already noted after 5 min of PT induction. At 10 min M3 mitochondria became clearly detectable (Figure 3C). A further continuous depletion of the M1 mitochondria was observed from 20 to 60 min coinciding with an increase of M2 and M3 mitochondria (Figure 3C). Next, we assessed how the separated mitochondrial subpopulations contribute to the changes in OD 540 that are determined in spectrometric swelling assays (Figure 3D). Similar OD 540 values were found for reference mitochondria (R) and Ca2+-treated

Figure 2. ZE-FFE analysis of rat liver mitochondria undergoing PT. (A) PT assessment by optical density measurements (“swelling”) of mitochondrial suspensions at different Ca2+ doses. Swelling occurred with different kinetics and could be inhibited by CsA. Each data curve represents the average of four individual measurements. (B and C) PT assessment by ZE-FFE. Untreated reference mitochondria showed one major peak (R). Mitochondria treated with Ca2+ to induce PT consistently deflected more towards the anode (>20 independent experiments). In the displayed example, mitochondria were incubated at 20 nmol Ca2+/mg for 20 min, and subsequent ZE-FFE analysis revealed several peaks (M1-M3, anode +; cathode -; one representative separation out of nine independent experiments is shown). CsA/Ca2+-treated mitochondria and FCCP-uncoupled mitochondria ( Ca2+ eluted in one peak which was similar in deflection and height to the reference mitochondria (representative separations out of seven and three independent experiments, respectively, are shown). (D) Electron micrographs of ZE-FFEseparated mitochondrial populations. The arrows indicate residual outer membrane in M2 mitochondria. Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

5055

Figure 3. PT causes variable degrees of mitochondrial damage. (A) Mitochondria treated with different Ca2+ doses show differences in the relative abundances of the resulting subpopulations (incubation time 20 min). Increasing Ca2+ doses cause a decrease of M1 and a concomitant increase of M2 and M3 mitochondria (one representative analysis out of six independent experiments is shown). (B) Immunoblotting analysis of ZE-FFE-separated mitochondrial populations from (A) showed significant differences in their amount of detectable outer membrane VDAC but relatively constant amounts of inner membrane TIM 23 and matrix proteins HSP 60 and Cyp D, respectively. No Cyt c was detected in mitochondria collected from the M3 and M2 mitochondria, while Cyt c was still present, although at reduced levels, in mitochondria collected from the anodal shoulder of M1 as compared to M1. Protein load per lane was 5 µg. (C) Mitochondria treated with 25 nmol Ca2+/mg mitochondria show time-dependent differences in the relative abundances of the resulting subpopulations. Prolonged treatment causes a decrease of M1 and a concomitant increase of M2 and M3 mitochondria (one representative analysis out of two independent experiments each with at least four different time points is shown). (D) Absorbance measurements at 540 nm of ZE-FFE-separated mitochondrial subpopulations normalized to 100 µg of mitochondrial protein of each population (data represent average values of three independent mitochondrial preparations and ZE-FFE separations).

M1 mitochondria (Figure 3D). In contrast, M2 and M3 mitochondria showed significantly lower OD 540 values (Figure 3D). Thus, the OD decrease that is determined in “classical” swelling assays appears to be related to the emergence of M2 and M3 mitochondria. A decrease in absorbance translates into a transition from R/M1 to M2/M3 mitochondria in ZE-FFE. Of note, a slightly higher OD 540 nm value was observed for ZE-FFE-purified CsA/ Ca2+-inhibited mitochondria compared to reference mitochondria (Figure 3D), which is in agreement with reported observations of an absorbance increase of CsA-treated mitochondrial suspensions upon addition of PT inducers.29–31 Taken together, these data show that ZE-FFE is suitable for the assessment of mitochondrial damage, as it is inflicted by Ca2+induced PT in a dose- and time-dependent fashion. PT-Induced MOMP in Ischemia/Reperfusion Analyzed by ZE-FFE. PT has been considered as an in vitro artifact for a long time,1 and many unsolved questions surround the role of the mitochondrial PT in vivo. It is generally accepted that mitochondrial membrane permeabilization represents a committing step in cell death scenarios.11 It is a matter of controversy, however, whether PT is a cause or a consequence of cell death,15 under which conditions and to which extent PT occurs in vivo,1 and (29) Petronilli, V.; Cola, C.; Bernardi, P. J. Biol. Chem. 1993, 268, 1011–1016. (30) Ricchelli, F.; Gobbo, S.; Moreno, G.; Salet, C. Biochemistry 1999, 38, 9295– 9300. (31) Li, M. Toxicology 2003, 194, 19–33.

5056

Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

which threshold of PT is associated with cell death in distinct cell types.32 However, there is some consensus that PT occurs and contributes to massive cell death during the reperfusion phase of tissues which have undergone a long period of ischemia, e.g., due to surgery, thrombosis, or stroke.1,14,23 We observed pronounced mitochondrial alterations in electron micrographs from fixed liver tissue upon ischemia/reperfusion (I/R) (Figure 4A). Mitochondria were markedly rounded, appeared swollen, and had lost their typical cristae organization (Figure 4A, I/R). Moreover, mitochondria presented partial outer membrane ruptures and losses (Figure 4A, I/R, arrows). Untreated reference liver tissues, in contrast, had ellipsoid-shaped, tubular cristae containing mitochondria (Figure 4A, R). It should be noted, however, that I/R did not induce the appearance of vesicles with IMV morphology that would resemble M2 and M3 mitochondria (Figure 2D), suggesting that a full-blown PT did not occur under these conditions. Next, we compared isolated I/R mitochondria and untreated reference mitochondria by ZE-FFE (Figure 4B). The ZE-FFE separation profiles of reference samples yielded a major peak containing mitochondria and a smaller cathodal glycogen-containing peak. This glycogen peak was absent from I/R mitochondria, (32) Rodriguez-Enriquez, S.; He, L.; Lemasters, J. J. Int. J. Biochem. Cell Biol. 2004, 36, 2463–2472.

Figure 4. PT-induced MOMP in ischemia/reperfusion (I/R). (A) Electron micrographs of normal liver tissue (R) and after I/R. Disintegrating cristae, configurational matrix transitions, and outer membrane ruptures (arrows) were noted in mitochondria after I/R. (B) ZE-FFE separation of mitochondria isolated from normal (R) and I/R-treated rat liver. A cathodal peak containing glycogen as identified by positive reaction with Lugol’s solution was noted in reference samples which was absent in I/R-treated samples. In comparison to reference mitochondria, I/R mitochondria shifted two fractions towards the anode (anode +; cathode -; one representative separation out of six ZE-FFE experiments from two independent preparations, each, is shown). (C) Immunoblotting analysis of mitochondria from the ZE-FFE comparison. Depletions of outer membrane VDAC and intermembrane space Cyt c were found in I/R-treated mitochondria in comparison to untreated reference mitochondria. Relatively constant amounts of the inner membrane TIM 23 and the matrix proteins HSP 60 and Cyp D were detected. Protein load per lane was 15 µg.

presumably as a result of glycogen catabolism. A quantitative anodal shift was seen for I/R mitochondria as compared to reference mitochondria (Figure 4B). Immunoblotting analysis of ZE-FFE-purified mitochondrial fractions demonstrated marked depletions of outer membrane VDAC and Cyt c in I/R mitochondria versus untreated reference mitochondria (Figure 4C). A mild depletion was observed for the inner membrane protein TIM 23. Rather little variations were noted for the matrix proteins HSP 60 and Cyp D (Figure 4C). Thus, electron microscopy, ZE-FFE, and the subsequent immunoblotting analysis consistently demonstrated outer mitochondrial membrane damage upon I/R. It further appears that I/R induced a less pronounced mitochondrial damage with respect to outer membrane ruptures and losses as compared to the observed mitochondrial alterations from the in vitro experiments. A critical issue is whether the mitochondrial damage observed in the liver tissue is adequately mirrored by the ZE-FFE analysis of the isolated organelles. The I/R treatment may have severely impaired mitochondria or may have “sensitized” them resulting in either sample losses or inflicted damage during the isolation procedure, meaning an under- or overestimation of mitochondrial damage. On the one hand, we did not observe a massive

mitochondrial destruction either on the electron micrographs from the tissues or in ZE-FFE, as this should have resulted in unspecific separation profiles with very broad peaks ranging more than 10 fractions, as was recently demonstrated with yeast mitochondria,21 making a severe underestimation of the mitochondrial damage unlikely. On the other hand, and in order to deal with the inherent difficulties of introducing damage upon mitochondrial isolation, we have chosen a “mild” isolation procedure avoiding numerous washing and resuspension steps. Indeed, we found a good agreement between mitochondrial damage observed in the originating tissue and in the herefrom isolated organelles subjected to ZE-FFE analysis of the treated tissues. Clearly, this preliminary analysis needs further attention in future experiments to arrive at quantitative conclusions. Assessment of MOMP by ZE-FFE. In the present study we have analyzed rat liver mitochondria that were either damaged in vitro osmotically or by Ca2+-induced PT or in vivo by I/R treatments. As evidenced by electron microscopy such mitochondria undergo MOMP. In ZE-FFE, these “treated” mitochondria consistently demonstrated a higher electrophoretic mobility as compared to untreated reference mitochondria, i.e., a stronger deflection toward the anode. Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

5057

According to “FFE theory”, the electrophoretic deflection depends on the charge and (in reciprocal manner) on the size of the separated particles.33 Due to their increased diameter (Figure 2D) the electrophoretic mobility of mitochondrial populations M2 and M3 should have decreased in comparison to reference mitochondria and consequently a more cathodal deflection should have occurred. Since we have observed the opposite, a more anodal deflection, mitochondrial size has likewise a minor impact on ZE-FFE of liver mitochondria. This is in agreement with Heidrich et al.,19 who reported the altered deflection to be caused by a more negative surface charge of the inner compared to the outer mitochondrial membrane. This finding was further substantiated by ultrastructural investigations from Hackenbrock and Miller34 and from Parmley et al.35 Both studies demonstrated the presence of patches of negative charges, presumably acidic glycoproteins, at the outer surface of the inner membrane of rat liver mitochondria. It thus appears that MOMP is linked to a (partial) removal of the outer membrane which leads to the exposure of these inner membrane anionic sites. Consequently, and in agreement with our observations, increasing anodal deflection reflects a more pronounced stripping of the outer membrane. CONCLUSIONS This study advocates the use of ZE-FFE for the discrimination and characterization of mitochondria exhibiting increasing levels of MOMP. The coexistence of different stages of mitochondrial (33) Roman, M. C.; Brown, P. R. Anal. Chem. 1994, 66, 86A–94A (34) Hackenbrock, C. R.; Miller, K. J. J. Cell Biol. 1975, 65, 615–630. (35) Parmley, R. T.; Spicer, S. S.; Poon, K.; Wright, J. J. Histochem. Cytochem. 1976, 24, 1159–1168.

5058

Analytical Chemistry, Vol. 80, No. 13, July 1, 2008

damage (M1, M2, and M3) was detected. Although mitochondria accumulated in a major stage of PT (M2), ZE-FFE allowed for the detection of an intermediate stage (M1), suggesting that PTinduced MOMP may progress through discrete steps. It will be interesting to see whether further mitochondrial subpopulations emerge, e.g., when PT is induced by other triggers than Ca2+. In contrast to the in vitro experiments, mitochondria isolated from I/R-treated livers showed a less pronounced damage. It is tempting to speculate that the presence of cytosol (with its considerable ion buffer capacities and colloid osmotic pressure) and/or geometric constraints (induced by the cytoskeleton and the presence of the mitochondrial network) prevent mitochondria from being completely spoiled of their outer membrane as this can occur in vitro. Irrespective of these theoretical considerations, it appears that ZE-FFE enables the detection of mitochondrial damage by PT, a direct comparison of the extent of mitochondrial damage upon PT, as well as a comparison between mitochondria from PT inflicting in vivo situations and in vitro experiments. ACKNOWLEDGMENT The authors thank Dr. E. E. Rojo for critical reading of the manuscript. H.Z. thanks Dr. A. Wendel for helpful discussions and Dr. M. Ueffing for his support. This work was funded by EMBO (ASTF No. 129-06 to H.Z.), by a special Grant of the Ligue National contre le Cancer (e´quipe labellise´e), as well as a Grant from EU (ApoSys to G.K.).

Received for review January 24, 2008. Accepted April 15, 2008. AC800173R