The Localization of Tightly Bound Cardiolipin in Cytochrome Oxidase*

One to two molecules of tightly bound cardiolipin are associated with resolved fractions of cytochrome oxidase containing subunits I to I11 or I to IV. Large scale isolation of subunits I to IV indicates the presence of approximately 0.5 molecule of cardiolipin per molecule of subunit I. Lipoprotein staining of sodium dodecyl sulfate/urea/acrylamide gels of cytochrome oxidase support the findings that subunit I is a lipoprotein. The resistance of this tightly bound cardiolipin to organic solvent extraction suggests a specific association of some tenacity with the protein. In contrast to the more general requirement of phospholipid and detergents for cytochrome oxidase (EC 1.9.3.1) activity this enzyme has been shown to have an absolute catalytic requirement for cardiolipin Past studies have indicated the relatively high proportion of cardiolipin in the isolated beef heart cytochrome oxidase (4), the presence of residual cardiolipin following lipid depletion of the enzyme (5, 6), and the resistance of this tightly bound cardiolipin to extraction with chloroform/methanol (2:1, v/v) (7) or cleavage by phos-pholipase A (8). Tightly bound cardiolipin in cytochrome oxidase is not removed by many solvents that extract loosely bound lipids but can be extracted with alkaline chloroform/ methanol (8) or acetone/ethanol mixtures (9), as well as by exchange with the detergent Triton X-100 (2, 3). The conclu-sion


METHODS
Preparation of cytochrome oxidase and sodium dodecyl sulfate/ urea/gel electrophoresis were as previously described (10). The biuret method (11) was used to determine protein with bovine serum albumin as the standard. Protein samples were solubilized with deoxycholate, and where necessary, sonication or addition of alkali (KOH) was used to compel solubilization. Phospholipid content was calculated GM-12847 of the National Institute of General Medical Sciences. The * This research was supported in part by Program Project Grant costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact, from an analysis of total phosphorus (12) and by assuming 4% phosphorus content in phospholipids. Identification of extracted phospholipids was made by bidimensional thin layer chromatography on 0.25mm-thick Silica Gel 60 on glass plates according to Awasthi et al. (7).
Chromatographed extracts were visualized by charring with sulfuric acid or spraying with the phosphorus-specific spray "Phospray" (Supelco, Inc.).
The following procedures were used for the staining of sodium dodecyl sulfate/urea gels with lipoprotein stains, adapted from histochemical methods (13). Approximately 200 pg of cytochrome oxidase protein was electrophoresed per gel. Staining for protein was made with 0.2% Coomassie blue R-250 as previously described (10). For lipoprotein staining, gels were fmt fixed in a solution of 10% acetone containing 10% trichloroacetic acid for 4 h a t room temperature, immediately following electrophoresis. Nile blue A (Fisher Scientific Co.) was made 0.2% (w/v) in methanol/water/acetic acid (5:4: 1) and gels were stained in this solution for 1 h a t 50°C. Sudan black B (National Aniline Division) was made 0.2% (w/v) in acetone/water/ acetic acid (5:4:1) and gels were stained in this solution for 30 min a t 70°C. Staining for lipoprotein by the acid hematein technique (13) was performed as follows. Following electrophoresis, gels were rinsed in 50 mM Tris-HCI, pH 7, and placed in a dichromate/calcium solution (5% w/v dichromate, 1% w/v CaCI,) for 1 h at 37°C. After rinsing thoroughly with distilled water, gels were treated for 5 h a t 37OC with an acid hematein solution, prepared by boiling 50 mg of hematoxylin (National Aniline Division) and 1 ml of 1% sodium iodate in 48 ml of double-distilled water, cooled, and acidified with 1 ml of glacial acetic acid. Gels were removed and soaked in a borax/ferricyanide solution for 3 h a t 60°C (0.25 g of sodium borate and 0.25 g of potassium ferricyanide in 100 ml of distilled water). All gels stained with lipoprotein stains, or with Coomassie blue, were destained for various times in a methanol/water/acetic acid (5:5:1) solution.
Subunits I, 11,111, and IV of cytochrome oxidase were purified after dissociation and elution in 2% sodium dodecyl sulfate on a Bio-Gel P-60 column (5 X 180 cm), essentially according to Steffens and Buse (14). Approximately 300 mg of cytochrome oxidase protein was loaded onto the column. T o 0.8 volume of the collected aqueous protein peak fractions in sodium dodecyl sulfate was added 2.0 volumes of methanol and 1.0 volume of chloroform. The resulting homogeneous solutions were stirred at room temperature for 1 h and centrifuged to remove the precipitated subunit fractions. The protein precipitate was washed once by suspension in and centrifugation from 100 ml of methanol/ chloroform/water (2:1:0.8), resuspended in 100 ml of distilled water, and evaporated under vacuum to remove traces of methanol and chloroform before lyophilization of protein samples.
An acid-insoluble fraction of cytochrome oxidase (subunits I to 111) and acid-soluble fraction (subunits IV to VII) were obtained by resolution in an acidic butanol/methanol mixture (10). An insoluble fraction of cytochrome oxidase (subunits I to IV) and soluble fraction (subunits V to VII) were obtained by resolution in 1% acetic acid. This method of resolution has been fully described elsewhere (15) and involves extraction of the washed and dialyzed particulate cytochrome oxidase in 1% acetic acid, the smaller subunits of the enzyme (V to VII) being selectively solubilized in this medium. Resolved fractions of cytochrome oxidase were thoroughly washed by homogenization in distilled water before lyophilization.
Purified protein subunits or resolved fractions of cytochrome oxidase were subjected to a series of organic solvent extractions. Firstly, approximately 10 mg of lyophilized protein was added to stoppered glass tubes and vortexed continually for 1 h at room temperature with 1 ml of acetone. After removal of the acetone, protein was similarly extracted in 1 ml of chloroform/methanol (2:l) and then again in 1 ml of chloroform/methanol/NH40H 28% (100:501.5), the phosphorus 9967 content of the protein residue being measured a t each stage of the extractions. After the final extraction in alkaline chloroform/methanol, 1 ml of ethanol was added to this organic mixture to precipitate the small amount of protein that was invariably solubilized in this alkaline mixture.

RESULTS AND DISCUSSION
When gels that had been heavily loaded with cytochrome oxidase (200 pg of protein/gel) were stained for lipoprotein, all three of the lipoprotein stains used stained that band corresponding in position to subunit I (Fig. 1). In addition to staining for subunit I, Nile blue and Sudan black dyes stained for subunit IV, while acid hematein stained for subunits 111 and IV. No lipoprotein stain was found at the gel origin or toward the lower end of the gel where protein-free lipid might be expected to migrate (16). If the dyes used in these studies were specifically staining for lipoprotein, then the association of protein and lipid must clearly be of some tenacity, considering the conditions of gel electrophoresis and the procedures used for fixing, staining, and destaining of the gels. However, these stains are not considered to be of absolute specificity for lipoprotein (13) and it was possible that staining could have resulted from other causes, for example, the high hydrophobicity of the larger subunits of cytochrome oxidase (17).
In an attempt to identify the localization of tightly bound cardiolipin in cytochrome oxidase, resolved fractions of this enzyme were subjected to organic solvent extraction. Densitometric gel profiles of the resolved fractions used in these studies are shown in Fig. 2. The results of organic solvent extraction on these fractions are tabulated in Table I. The phosphorus content of those fractions containing the smaller subunits of cytochrome oxidase (either IV to VI1 or V to VII) was readily extracted in acetone, and, following extraction in chloroform/methanol (2:1), the phosphorus content of such fractions was undetectable. In contrast, following extraction with acetone and chloroform/methanol (2:1), a small but significant phosphorus content remained associated with those fractions containing the larger subunits of cytochrome oxidase (I to I11 or I to IV). Only after extraction with alkaline chloroform/methanol was this residual phosphorus content completely removed (Table I). Thin layer chromatographic analysis of this alkaline extract confirmed the exclusive presence of cardiolipin (cf. Fig. 4).
A large scale isolation procedure on Bio-Gel P-60 was used to obtain essentially purified subunits I, 11, 111, and IV of   (15). Gels were scanned a t 550 nm after staining in Coomassie blue. Roman numerals refer to subunits I to VI1 by the nomenclature of Downer et al. (18).

Solvent extraction of lipid associated with resolved fractions of
cytochrome oxidase Resolved fractions of cytochrome oxidase (see Fig. 2) were subjected to a series of extractions (1 h a t room temperature with vortexing) in acetone, chloroform/methanol (2:l). and finally chloroform/methanol/28% NH,OH (100:501.5). After each extraction, the phosphorus content of the protein fractions was determined. Results are averaged from determinations on three different preparations of resolved fractions. cytochrome oxidase (Fig. 3). Subunit I isolated by this procedure was essentially free of other subunits, while subunit I1 was about 20% contaminated with subunit I (calculated from their relative gel peak areas). Following removal of sodium dodecyl sulfate with precipitation of protein from methanol/ chloroform/water (2:1:0.8), subunits were lyophilized and then analyzed for phosphorus content (Table 11). Subunit I was found to contain approximately 0.9 pg of phosphorus/mg of protein and subunits I1 and 111 approximately 0.4 and 0.2 pg of phosphorus/mg of protein, respectively. The phosphorus associated with these subunits was resistant to extraction with acetone or chloroform/methanol but was removed in alkaline chlorofom/methanol ( Table 11). The alkaline extracts from subunits I, 11, and 111 were analyzed by thin layer chromatography that confirmed cardiolipin associated with these subunits (Fig. 4). The cardiolipin extracted from subunit I chromatographed identically to the beef heart cardiolipin standard used, as did that tightly bound cardiolipin extracted under alkaline conditions from resolved fractions of the larger subunits of cytochrome oxidase. Assuming a molecular weight of about 40,000 for subunit I (18), approximately 0.5 mol of cardiolipin/mol of subunit I remain associated following purification of this subunit. This calculation is based on the assumption that the phosphorus content measured in subunit I is wholly accounted for by the cardiolipin content of this subunit. This was confirmed to be the case by chromatographing an alkaline extract of a known amount of subunit I protein and analyzing the phosphorus content of the chromatogram (by scraping off the cardiolipin spot); this amount agreed within 5% with the value for the phosphorus extractable from the same sample weight of lyophilized subunit I. Studies in our laboratory have confirmed the absolute catalytic requirement for cardiolipin in cytochrome oxidase (2). A minimum of some 2 molecules of cardiolipin per complex of cytochrome oxidase are necessary for the full potential activity A B C D

After purification and lyophilization of subunits I to I V isolated on
Bio-Gel P-60 (see Fig. 3). approximately 10 mg of protein of each were analyzed for phosphorus content, before and after extraction in organic solvents (   All of the procedures used in these studies in an attempt to determine the localization of cardiolipin in cytochrome oxidase might normally be expected to dissociate or extract loosely bound phospholipids. For this reason, we feel it is unlikely that the association of tightly bound cardiolipin with the larger subunits of cytochrome oxidase (and in particular subunit I) is merely a happenstance. There have been numerous reports on the isolation and purification of cytochrome oxidase subunits, primarily from the standpoint of determining amino acid sequences or of identifying bound prosthetic groups (heme and copper) but apparently the lipoprotein nature of some of these subunits had been overlooked. The present studies suggest a major portion of the tightly bound cardiolipin is associated with subunit I, although we cannot rule out the possibility of association with other subunits (most likely subunits I1 and 111); in the unresolved enzyme, cardiolipin might be associated between subunits I and IV. In our studies on the resolution of cytochrome oxidase (15), we have shown that a subunit I.IV complex, containing all the heme and copper of the original enzyme in a spectrally unaltered state, may represent the limit of resolution of cytochrome oxidase that remains functionally indistinguishable from the isolated soluble holoenzyme. These studies emphasized the importance of subunit IV in retaining the integrity of this complex; removal of subunit IV from the I IV complex resulted in loss of heme and copper, changes in their spectral characteristics, and loss of enzymic activity. It is significant therefore, that tightly bound catalytic cardiolipin (2) should be found in this fraction (see Table I). The finding of some 0.5 molecule of cardiolipin associated with purified subunit I may mean that a major portion of the 2 molecules of catalytic cardiolipin bound to the enzyme complex are simply lost during the isolation procedure, or that 1 molecule of cardiolipin is more labile in its association with the complex and is more readily extracted upon resolution of the enzyme (e.g. upon resolution of subunit IV from the I-IV complex).
Given the catalytic requirement for cardiolipin in cytochrome oxidase (2), its mode of association with the mitochondrially synthesized subunits of this enzyme (20) and its role therein will clearly be of great importance in finally understanding the mechanism of action of cytochrome oxidase.