Eur. J. Biochem. 209, 267-273 (1992) ( i ~ FEBS 1992 1ON-Nonyl acridine orange interacts with cardiolipin and allows the quantification of this phospholipid in isolated mitochondria Jean-Michel Pk I I T , Abderrahman MAFTAH, Marie-Helene RATINAUD and Raymond JULIEN lnstitut de Riotechnologie, Faculte des Sciences, Limoges, France (Rcceived April 30/July 10, 1992) ~ EJB 92 0604 The acridine orange derivative, l0N-nonyl acridine orange, is an appropriate marker of the inner mitochondrial membrane in whole cells. We use membrane model systems to demonstrate that 10N- nonyl acridine orange binds to negatively charged phospholipids (cardiolipin, phosphatidylinositol and phosphatidylserine). The stoichiometry has been found to be 2 mol 10N-nonyl acridine orange/ mol cardiolipin and 1 mol dye/mol phosphatidylscrine or phosphatidylinositol, while, with zwitterionic phospholipids, significant binding could not be detected. The affinity constants were 2 x 1 O6 M ~ for cardiolipin-I ON-nonyl-acridine-orange association and only 7 x lo4 M ~ ' for that of phosphatidylserine and phosphatidylinositol association. The high affinity of the dye for cardiolipin may be explained by two essential interactions firstly an electrostatic interaction betwcen the quatern- ary ammonium of nonyl acridine orange and the ionized phosphate residues of cardiolipin and secondly, hydrophobic interactions between adjacent chromophores. A linear relationship was dem- onstrated between the cardiolipin content of model membranes and the incorporated dye. Conse- quently, a convenient and rapid method for cardiolipin quantification in membranes was established and applied to the cardiolipin-containing organelle, the mitochondrion. The 10N-nonyl acridine orange (NAO), which is specifi- cally incorporated into the inner mitochondrial membrane [l], plays a prominent role in the study of mitochondria in whole cells [2, 31. It enables monitoring of mitochondria in different situations, such as the cell cycle [4] and cell ageing [5] and also discrimination between different subpopulations of a heterogenous cell population, according to thcir mitochon- drial contents [6, 71. Nevertheless, up to now the membrane molccular species which are specifically recognized by NAO, have not been determined. The large number of inner-mitochondrial-membrane enzymes implicated in oxidative phosphorylation, differing in their conformation and biological properties require a similar lipid environment for their activity. Cardiolipin, one of the three major phospholipids present in the inner membrane [8 - 121, has been reported to be essential for the activity of the ADP/ATP carrier protein [I 31, for the phosphate transport protein [14] and for various other enzyme complexes [15,16]. This phospholipid has also been reported to be associated with the F1-FO ATPase [17]. Consequently, the NAO inhibition of the inner-mitochondrial-membrane enzymes [ 181 may be due to the interaction between the positively charged dye and cardiolipin, the main acidic phospholipid present in the inner mitochondrial membrane. The purpose ofthis work was to establish, with reference to different model membranes, that lON-nonyl acridine orange intcracts with acidic phospholipids and, more particularly, with cardiolipin. The absorbance spectra of NAO incubated Correspondence to J.-M. Petit, Institut de Biotechnologie, 123 Avcnue Albert Thomas. F-87060 Limoges ckdex, France Ahhreviationx AO. Acridine orange NAO, 1 ON-nonyl acridine orange. with liposomes and measurements of the degree of saturation allowed us to determine the specificity and the stoichiometry of NAO interactions with acidic phospholipids. In order to discriminate between the hydrophobic contribution and the charge effect, zwitterionic phospholipids were also tested, Fi- nally, the high affinity of NAO for cardiolipin was used to quantify this phospholipid in isolated mitochondria. MATERIALS AND METHODS Chemicals Cardiolipin from bovine heart, phosphatidylserine from bovine brain, phosphatidylinositol from bovine liver, dipal- mitoylphosphatidylcholine, dimyristoylphosphatidylcholinc and dipalmitoylphosphatidylethanolamine were purchased from Sigma Chemical Co. 10N-nonyl acridine orange was a gift from H. W. Zimmermann (Institut fur Physikalische Chemie der Universitat Freiburg, FRG), and acridine orange (AO) from Polysciences. As NAO is sensitive to light, stock solution was prepared just prior to use. All othcr chemicals were of analytical grade deionized double-distilled water was used throughout. Absorbance values were measured with a Perkin-Elmer Lambda-2 spectrophotometer. Values were re- ported as averages of at least triplicate measurements. Preparation of liposomes Lipid mixtures, described in Table 1, were prepared in chloroform and then evaporated to dryness under a stream of nitrogen. After addition of 220 mM mannitol, 70 mM sucrose, 0.1 mM EDTA, 10 mM Hepes and enough KOH to raise the

268 Table 1. Phospholipid composition of liposomes. Phospholipids Mixture 1 2 3 4 % total lipid phosphorus PdtCho 54 40 47 47 PdtEth 46 33 40 40 Cardiolipin - - - PdtScr - - 13 - 13 Pdtlns - ~- - 27 pH to 7.4 (buffer A), liposomes were formed by sonication for 5 min at 4'C with a probe Branson Sonifier B-30 (power setting 30 W). In order to eliminate very small liposomes which could not be pelleted by centrifugation, the vesicles were washed once by centrifugation at I50000 x g for 60 min. The pellet was resuspended in the same buffer and the stock lipid suspensions were stored at 4 'C under nitrogen. An ali- quot was removed for the determination of the phospholipid composition of each liposome preparation. Their relative com- positions are given in Table 1. NAO binding to liposomes Increasing amounts of NAO (10-200 pM) from a freshly prepared 5-mM stock solution in ethanol, were added to lipo- somes and the final volume was adjusted to 1 ml with buffer A. The lipid concentrations in each sample, determined by P, analysis, were 240 pM for mixture 1, 325 FM for mixture 2, and 275 pM for mixtures 3 and 4 (the percentage of phospholipids in each mixture is given in Table 1). The suspen- sion was mixed and immediately centrifuged at 150000 x g for 60 min. An aliquot of 100 pl supernatant was taken for P, analysis to control that the supernatant was free of phospholipids. The unbound dye in the supernatant was mea- sured by its absorbance at 495 nm with a Perkin-Elmer type- Lambda-2 double-beam spectrophotometer. The concen- tration of unbound NAO was calculated from a calibration curve run with NAO solutions of known content (0- 10 pM) in buffer A, and bound NAO was calculated as total minus free. Preparation of mitochondria Mitochondria were isolated from the liver of adult male Wistar rats in 2 mM Tris/HCl, pH 7.4, containing 250 mM sucrose (buffer B) according to Hogeboom et al. [19]. The liver was submerged in ice-cold buffer B. The chopped tissue was homogenized using a glass Teflon motorized Potter- Elvejhem homogenizer. Large cell debris and nuclei were pel- leted by centrifuging the homogenate for 10 min at 600 x g. The supernatant was centrifuged for 15 min at 7000 x g. The mitochondria1 pellet was washed twice in buffer A by Centrifugation at 8000 x g for 15 min. The entire procedure was carried out at 4��C. The protein content was determined by a Biuret procedure using bovine serum albumin as a stan- dard [20]. NAO binding to mitochondria To 100 p1 freshly isolated mitochondria (10 mg protein/ ml), increasing amounts of NAO (10-300 pM) were added 0.3 1 0 100 200 NAO added (nmol) Fig. 1. NAO binding to liposomes. PdtCholPdtEth (mixture 1, I), PdtCho/PdtEth/cardiolipin (mixture 2, O), PdtCho/PdtEth/PdtScr (mixture 3, V) and PdtCho/PdtEth/PdtIns (mixture 4, M). The phospholipid compositions are described in Table 1. Increasing amounts of NAO were added to liposomes, and the final volume of the assay was adjusted to 1 ml with 220 mM mannitol, 70 mM sucrose, 0.1 mM EDTA, 10 mM Hepes and enough KOH to raise the pfI to 7.4. The phospholipid concentration in the sample was 240 pM for mixture 1, 325 pM for mixture 2 and 275 pM for mixtures 3 and 4. The liposome suspensions were immediately centrifuged (60 min, 150 000 x g), and free dye determined by the absorbance at 495 nm. Bound dye was calculated as total minus free. No lipids were detected by Pi analysis of I00 pl supernatant. and the final volume was adjusted to 1 ml with buffer A. The samples were incubated for 2 min, then centrifuged at 30000 x g for 5 min. The absence of phospholipids in the supernatant was controlled by Pi analysis on 100 pl super- natant. Bound NAO was determined as described above. Lipid extraction and analysis Total phospholipids were extracted from freshly isolated mitochondria (20 mg protein/ml) according to Folch et al. [21]. The mitochondrial suspension was homogenized with 20 volumes chloroform/methanol (2 : 1). After filtration, the residue was re-extracted twice, each time with ten volumes of solvent. The non-lipid contaminants wcre removcd by Folch partition. The extract was evaporated to dryness and the lipids were dissolved in chloroform. Phospholipid species were sep- arated on silica-gel plates (Merck) using chloroform/ methanol/water (65: 25:4, by vol.) as the developing system. The spots were visualized by I2 vapor and their lipid phos- phorus contents were determined according to Ames [22]. Hepatocyte preparation and staining Hepatocytes were prepared as previously described [23, 241. Rat liver was perfused with a solution of collagenase (2 mg/ml) for 10 min at 30"C, then dilacerated in a Krebs- Ringcr solution. Cells were washed twice by Centrifugation at 50 x g for 2 min. 2 x lo6 cells/ml were stained with 0.5 pM NAO or 0.5 @I A 0 for 15 min at 25'C, then examined by fluorescence microscopy (BHT, Olympus). Photographs were taken using 400 Asa color film (Fujichrom).