Transmembrane movement of phosphatidylcholine in mitochondrial outer membrane vesicles.

One of the steps in the import of phosphatidylcholine (PC) in mitochondria is transmembrane movement across the outer membrane. This process was investigated in vitro using isolated mitochondrial outer membrane vesicles (OMV) from rat liver. 14C-Labeled PC was introduced into the OMV from small unilamellar vesicles by a PC-specific transfer protein (PCTP). The membrane topology of the newly introduced PC was determined from its accessibility to phospholipase A. Under conditions where the OMV stay intact, externally added phospholipase A is able to hydrolyze up to 50% of both the introduced [14C]PC and the endogenous PC. Pool size calculations showed that close to 100% of the PC in the OMV can be exchanged by PCTP. A back-exchange experiment revealed that the introduction of the labeled PC is reversible. The results demonstrate that newly introduced PC molecules readily equilibrate over both leaflets of the OMV membrane. The kinetics of the PCTP-mediated exchange process indicate that the t of the transmembrane movement at 30°C is 2 min or less.

One of the steps in the import of phosphatidylcholine (PC) in mitochondria is transmembrane movement across the outer membrane. This process was investigated in vitro using isolated mitochondrial outer membrane vesicles (OMV) from rat liver. 14 C-Labeled PC was introduced into the OMV from small unilamellar vesicles by a PC-specific transfer protein (PCTP). The membrane topology of the newly introduced PC was determined from its accessibility to phospholipase A 2 . Under conditions where the OMV stay intact, externally added phospholipase A 2 is able to hydrolyze up to 50% of both the introduced [ 14 C]PC and the endogenous PC. Pool size calculations showed that close to 100% of the PC in the OMV can be exchanged by PCTP. A back-exchange experiment revealed that the introduction of the labeled PC is reversible. The results demonstrate that newly introduced PC molecules readily equilibrate over both leaflets of the OMV membrane. The kinetics of the PCTP-mediated exchange process indicate that the t1 ⁄2 of the transmembrane movement at 30°C is 2 min or less.
Phosphatidylcholine (PC) 1 is the major constituent phospholipid in both the inner and the outer membrane of mammalian mitochondria (1). The terminal steps of the de novo synthesis of PC (and most other phospholipids) occur primarily on the cytoplasmic surface of the endoplasmic reticulum (2). Consequently, the biogenesis of mitochondria requires efficient import of PC from the ER. Several mechanisms for the intracellular transport of phospholipids have been proposed (for a recent review see Ref. 3). There is growing evidence that the import of phosphatidylserine into mitochondria occurs via direct interorganelle contact between mitochondria and the so-called mitochondria-associated membrane (4 -7). This ERlike membrane contains several phosholipid biosynthetic enzymes including some involved in the synthesis of PC (7,8). The mechanism of PC import into mitochondria is not known.
Additional mechanisms, so far unknown, are required for the sorting of the newly imported phospholipids within the mitochondrion, in order to maintain the specific phospholipid composition and transbilayer phospholipid distribution of both mitochondrial membranes. In vivo labeling studies in hepatocytes employing [ 14 C]choline have shown that radiolabeled PC appears rapidly in the mitochondrial outer membrane, while transfer to the inner membrane is slow (9). This suggests that newly synthesized PC is transferred from the ER to the inner membrane via the outer membrane. In isolated mitochondria, movement of labeled endogenous PC from the outer to the inner membrane has been demonstrated (10). Thus, the outer membrane is the first barrier to be taken by a newly imported PC molecule on its way to the inner membrane or to the inner leaflet of the outer membrane.
In the present study the process of PC movement across the outer mitochondrial membrane is investigated in an in vitro system consisting of mitochondrial outer membrane vesicles (OMV) isolated from rat liver mitochondria. These OMV are sealed and have a right-side-out topology with an almost symmetric distribution of PC over the inner and outer membrane leaflet (11). Phosphatidylcholine-specific transfer protein (PCTP) was used to introduce radiolabeled PC into the OMV from donor vesicles. The use of PCTP also allowed the determination of the pool size of exchangeable PC in the OMV. The transmembrane distribution of the introduced PC was assessed from its accessibility to externally added phospholipase A 2 (PLA 2 ). It is concluded that newly introduced [ 14  were purchased from Avanti (Birmingham, AL). PCTP was purified from bovine liver and stored at Ϫ20°C at a concentration of ϳ0.5 mg/ml in 20 mM Tris-HCl, pH 7.2, 100 mM NaCl, containing 50% glycerol (12). Porcine pancreatic phospholipase A 2 (ppPLA 2 ) was purified as in Ref. 13. Bee venom phospholipase A 2 (bvPLA 2 ), bovine serum albumin (essentially fatty acid-free), trypsin, and soybean trypsin inhibitor were obtained from Sigma. All other chemicals were analytical grade.
Isolation of Mitochondrial OMV-Mitochondria were isolated from the livers of male Wistar rats by a combination of differential centrifugation and isotonic Percoll gradient centrifugation as described (14). OMV were obtained using a modified version of the method recently introduced for mitochondria from Neurospora crassa (15).
Briefly, the mitochondria were suspended in hypotonic buffer (5 mM KP i , pH 7.2, 5 mM EDTA) at a protein concentration of 5 mg/ml and incubated on ice for 20 min under stirring. The outer membrane was detached from the remaining mitoplasts by 20 strokes in a Potter-Elvehjem homogenizer with a tightly fitting Teflon pestle. The homogenate (15 ml) was layered on a sucrose step gradient consisting of 9 ml of 0.25 M sucrose and 12 ml of 1.1 M sucrose in 10 mM MOPS, pH 7.2, 2.5 * This work was carried out with financial support from the Human Capital and Mobility program of the European Union. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ Present address: Institut de Biologie Physico-Chimique "Biophysique Cellulaire," URA-CNRS 526, 13 1 The abbreviations used are: PC, phosphatidylcholine; DOPC, dioleoylphosphatidylcholine; DOPA, dioleoylphosphatidic acid; CE, cholesteryl oleoyl ether; MOPS, 3-(N-morpholino)propanesulfonic acid; PLA 2 , phospholipase A 2 ; ppPLA 2 , porcine pancreatic PLA 2 ; bvPLA 2 , bee venom PLA 2 ; OMV, outer membrane vesicle(s); PCTP, PC-specific transfer protein; SUV, small unilamellar vesicle(s); ER, endoplasmic reticulum; HPTLC, high performance thin layer chromatography. mM EDTA (EM buffer). After ultracentrifugation for 1 h at 141,000 ϫ g (SW28 rotor, Beckman), the outer membrane fraction was collected from the 0.25-1.1 M sucrose interface, its sucrose concentration was adjusted to 1.1 M, and it was loaded on the bottom of a flotation sucrose gradient (4 ml/tube). A layer of 7 ml 1.065 M sucrose in EM buffer was put on top, and the tube was filled up with EM buffer. The purified OMV were collected from the 0 -1.065 M sucrose interface after ultracentrifugation for 16 h at 141,000 ϫ g (SW41 rotor), diluted in EM buffer, and pelleted by centrifugation for 1 h at 300,000 ϫ g (Ti60 rotor). The pellet was resuspended in 220 mM mannitol, 70 mM sucrose, 2 mM Hepes, pH 7.4 (MSH buffer) at a protein concentration of 1-3 mg/ml. Aliquots of the OMV were frozen in liquid nitrogen and stored at Ϫ80°C.
Typically, the procedure yielded 1 mg of OMV/100 mg of mitochondria based on protein. Based on the total activity of the outer membrane marker enzyme monoamine oxidase, 25% of the outer membranes are recovered in the OMV preparation. The purity of the OMV was assessed by marker enzyme analyses and will be documented in detail elsewhere. 2 Typically, the specific activity of monoamine oxidase was enriched 26 times, while that of succinate cytochrome c reductase (mitochondrial inner membrane marker) was depleted by a factor of 30, compared with the specific activities determined for the mitochondria. The specific activities of monoamine oxidase, the microsomal markers glucose-6-phosphatase and NADPH-cytochrome c reductase, the lysosomal marker acid phosphatase, and the plasma membrane marker 5Ј-nucleotidase in the OMV fraction relative to those of the nuclei-free rat liver homogenate were 69.8, 0.45, 0.55, 2.8, and 1.3, respectively.
Preparation of Donor Vesicles-A dry lipid film was prepared, which consisted of 1.935 mol of DOPC including 4 Ci of [ 14 C]DOPC, 0.1 mol of DOPA, and 37 Ci of [ 3 H]CE (1 nmol) as nonexchangeable marker. The lipid film was hydrated in 1 ml of MSH buffer. The resulting phospholipid suspension was subjected to 10 cycles of 30-s ultrasonication with 30-s intervals at 0°C, using a Branson B12 sonifier equipped with a microtip and operated at 55 W output. The sonicated phospholipid suspension was centrifuged for 30 min at 315,000 ϫ g in a Beckman TL-100 ultracentrifuge. The supernatant containing the small unilamellar vesicles (SUV) was stored in aliquots at Ϫ20°C until use. Storage did not adversely affect the recovery of the SUV in the supernatant after a 20-min centrifugation at 185,000 ϫ g as used in the forthcoming.
PCTP-catalyzed Exchange of PC-OMV (at a concentration of 0.25 mg of protein/ml) were incubated with radiolabeled donor SUV (0.15 mM PC) and PCTP (18 g/ml) in MSH buffer containing 0.05% (w/v) bovine serum albumin at 30°C for 20 min unless otherwise indicated. Subsequently, the OMV were separated from the SUV by ultracentrifugation for 20 min at 185,000 ϫ g (TLA100.2 rotor). To determine the extent of the PC exchange, 14 C and 3 H were counted in the supernatant and in the resuspended pellet by a Packard 1500 Tricarb Liquid Scintillation Analyzer.
Quantification of the transfer of [ 14 C]DOPC to the OMV involved a correction for the SUV contamination in the OMV pellet using the 3 H-labeled nonexchangeable marker. This contamination typically amounted to 10% of the total amount of SUV added. In the calculation it was assumed that the 14 C-specific radioactivity of the co-pelleted SUV is the same as that determined for the SUV in the supernatant. In order to calculate the obtained specific radioactivity of PC in the OMV, the phospholipid phosphorus/protein ratio and the PC content of the OMV were determined. OMV aliquots corresponding to 150 g of protein were extracted according to Bligh and Dyer (16), and the phosphorus content of the organic phase was quantitated (17) to yield the phospholipid phosphorus/protein ratio. Phospholipid analysis of the OMV extracts by two-dimensional HPTLC was performed as described (14). Spots were visualized by iodine vapor and scraped off, and phospholipid phosphorus was quantitated (17), yielding the PC content of the OMV. Apart from the considerations above, the calculation of the pool size of exchangeable PC in the OMV ((specific radioactivity of OMV/specific radioactivity of SUV outer leaflet) ϫ 100%) was based on the notion that 65% of the PC in the donor SUV is accessible to PCTP (18 -20).
In back-exchange experiments, OMV with [ 14 C]DOPC introduced as described above were incubated at 0.25 mg/ml with unlabeled SUV consisting of DOPC/DOPA (95:5, mol/mol) at a concentration of 1.5 mM based on phosphorus, in the presence of 18 g/ml PCTP at 30°C for the periods of time indicated. OMV were recollected by centrifugation at 185,000 ϫ g as above. Radioactivity in the pellet and the supernatant was counted, and the percentage of [ 14 C]DOPC extracted from the OMV was calculated, taking into account the above corrections, while assuming that 10% of the SUV with associated 14 C label co-pellets with the OMV.
Phospholipase A 2 Treatment-OMV at a protein concentration of 0.6 mg/ml in MSH buffer were incubated for the indicated periods with ppPLA 2 and bvPLA 2 at a concentration of 0.4 units/ml each, in the presence of 0.1 mM Ca 2ϩ at room temperature. This mixture of phospholipases A 2 is optimal for determining the phospholipid topology in OMV, as stable levels of phospholipid degradation are obtained rapidly (11). PLA 2 activity was inhibited by the addition of 0.75 mM EGTA. The OMV were pelleted (20 min at 185,000 ϫ g) and subjected to phospholipid extraction. The specific activities of ppPLA 2 (1.25 ϫ 10 3 units/mg) and bvPLA 2 (4.2 ϫ 10 3 units/mg) were determined using egg yolk lipoproteins as substrate (13).
Intactness of the OMV-After treatment with PLA 2 , aliquots of the OMV suspension corresponding to 40 g of protein were incubated with 1.7 mg of trypsin/mg of OMV for 20 min on ice, prior to the addition of a 2.5-fold excess (w/w) of soybean trypsin inhibitor over trypsin (11). The remaining activity of the enzyme adenylate kinase, determined spectrophotometrically at 37°C (14,21), served as a measure for the intactness of the OMV.
Analysis of PC Degradation-Phospholipids were extracted according to the method of Rose and Oklander (22) from samples corresponding to 100 -150 g of OMV. To determine the degradation of the endogenous PC by PLA 2 , phospholipids were separated by two-dimensional HPTLC, visualized by iodine vapor, and quantitated as described (14).
To assess the degradation of [ 14 C]DOPC, the lipid extract was analyzed by one-dimensional HPTLC on silica gel 60 (Merck, FRG), using chloroform, methanol, 25% ammonia, water (90:54:5.5:5.5 (v/v/v/v)) as eluent. Subsequently, the radioactive spots on the TLC plate were quantified by a Berthold Automatic TLC linear analyzer (Wildbad, FRG). As the hydrolysis of [ 14 C]DOPC to lyso-[ 14 C]PC and [ 14 C]oleate by PLA 2 did not affect the recovery of the radiolabel from the extraction procedure (see Fig. 1A), it was possible to quantitate the extents of [ 14 C]DOPC degradation from the HPTLC elution profiles. No hydrolysis of PC was detected in control experiments without PLA 2 .
Other Methods-Protein concentrations were determined by the BCA assay (Pierce) with 0.1% (w/v) SDS added and bovine serum albumin as standard. Concentrations of SUV were determined by phosphate analysis (17).

Localization of [ 14 C]DOPC in OMV by Treatment with PLA 2 -14 C-Labeled DOPC was introduced into isolated mitochondrial OMV from donor vesicles by the action of the PCTP.
This protein catalyzes a one-to-one PC exchange resulting in a replacement of the endogenous PC in the acceptor membrane with PC from the donor vesicles, i.e. without changing the PC content of the membranes (23)(24)(25). [ 14 C]DOPC was efficiently transferred from donor SUV to acceptor OMV under the conditions used, an incubation for 20 min at 30°C with equal amounts of PC present in the populations of donor and acceptor vesicles (PC donor /PC acceptor molar ratio approximately 1), and in the presence of PCTP. Typically, around 4 nmol of [ 14 C]DOPC were transferred per mg of OMV protein.
After removal of the donor SUV by centrifugation, the OMV were incubated with a mixture of phospholipases A 2 from bee venom and porcine pancreas to determine the membrane topology of the labeled PC molecules. Fig. 1A shows the HPTLC analysis of OMV phospholipid extracts from a typical experiment. Approximately 40% of the [ 14 C]DOPC is hydrolyzed within 10 min by the action of PLA 2 (Fig. 1, A and B). Degradation of the remainder proceeds slowly and is accompanied by a gradual loss of the intactness of the OMV as indicated by the increasing accessibility of adenylate kinase present in the OMV lumen to externally added trypsin (Fig. 1C). In the presence of an excess of PLA 2 (Fig. 1A), up to 94% of the [ 14 C]DOPC is degraded in 10 min. This is accompanied by a loss of barrier function of the OMV membrane as evidenced by the proteolysis of over 80% of adenylate kinase by trypsin (data not shown). Under similar conditions with an excess of PLA 2 present virtually all endogenous PC is hydrolyzed (11).
The time course of the degradation of [ 14 C]DOPC parallels that of the hydrolysis of endogenous PC under the same conditions (Fig. 1B). From the amount of [ 3 H]CE present in the lipid extracts (Fig. 1A) it is estimated that approximately 10% of the total 14 C label present originates from co-pelleted SUV. This contamination may account for the slightly different degrees of protection from PLA 2 observed for labeled and endogenous PC after 10 min of incubation (Fig. 1B). When the time of the incubation with PCTP is reduced to 3 min, still 60 Ϯ 5% of the 14 C label is recovered as intact PC after 10 min of treatment with PLA 2 (data not shown). The data suggest that the newly introduced PC rapidly adopts a transbilayer distribution similar to that of the endogenous PC, with approximately 50% of the PC content localized in the inner leaflet and 50% in the outer leaflet in agreement with Ref. 11.
The Exchangeable Pool of PC in the OMV-In order to calculate the specific radioactivity of the PC pool in the OMV after exchange by PCTP, the PC content of the OMV was determined. The OMV have a phospholipid phosphorus/protein ratio of 1220 Ϯ 150 nmol/mg (Ϯ S.D., n ϭ 6) and a PC content of 54.7 Ϯ 3.1% (Ϯ S.D., n ϭ 11) (compare Refs. 1 and 14).
The existence of a fast equilibration of newly introduced [ 14 C]DOPC across the OMV membrane is supported by the results depicted in Fig. 2. The specific radioactivity of the PC pool in the OMV was determined after different times of exchange by PCTP and compared with that of the outer leaflet of the donor SUV, as PCTP only has access to the outer leaflet of phospholipid vesicles (19). After 5 min of exchange, the specific radioactivity of both PC pools reaches a similar equilibrium value (Fig. 2), demonstrating that the entire PC pool of the OMV has exchanged with the outer leaflet of the SUV. The exchangeable pool of PC in the OMV was determined in several independent experiments, yielding an average value of 91 Ϯ 12% (Ϯ S.D., n ϭ 22).
After introduction into the OMV membrane, [ 14 C]DOPC is completely extractable in a back-exchange experiment as shown in Fig. 3. OMV with [ 14 C]DOPC incorporated were incubated with an excess of unlabeled SUV (PC SUV /PC OMV molar ratio of 10) and PCTP. Under these conditions all of the [ 14 C]DOPC is in exchange with the PC in the outer leaflet of the SUV (Fig. 3), demonstrating the reversibility of the transmembrane movement. DISCUSSION The present study demonstrates that [ 14 C]DOPC introduced into the outer leaflet of isolated mitochondrial outer membrane vesicles by PCTP is rapidly equilibrated over the two membrane leaflets. This conclusion is based on two lines of experimental evidence. (i) Similar to the extent of hydrolysis of endogenous PC, only 50% of the introduced [ 14 C]DOPC can be degraded by externally added PLA 2 , while the OMV stay intact, indicating that the other 50% has flipped to the inner leaflet. (ii) The pool size of exchangeable PC in the OMV is close to 100%. This result, combined with the reversibility of the exchange process (Fig. 3), indicates that the probe molecule [ 14 C]DOPC behaves like endogenous PC. The monophasic kinetics observed for the PCTP-catalyzed PC exchange process (Fig. 2) do not allow distinction between the rates of the actual exchange and the PC transmembrane movement. Consequently, the rate constant of the transmembrane movement of PC (flip-flop) is close to or higher than the rate constant of the overall exchange process, which implies that at 30°C the t1 ⁄2 for PC transmembrane movement in OMV is 2 min or less. Direct access of PCTP to the inner leaflet of the OMV membrane by permeation could in principle also explain the results obtained in the present study. However, this possibility is considered remote in view of the molecular size of PCTP (25 kDa (26)) relative to that of cytochrome c (12 kDa), for which the outer membrane is impermeable (27). Accordingly, labeled PC is not recovered in the inner membrane fraction after its introduction into intact mitochondria by PCTP (20).
It may seem paradoxical that on the one hand almost 100% of the PC pool of the OMV is accessible to PCTP, while on the other hand only half of the introduced [ 14 C]DOPC can be degraded by external PLA 2 in intact OMV (Fig. 1). A possible explanation would be that the PCTP-catalyzed exchange process, or PCTP itself, induces the rapid transmembrane movement of PC in OMV. This possibility cannot be excluded; however, so far there are no indications that PCTP would have this ability in other biomembranes (23,24,28). Furthermore, PCTP does not induce phospholipid transbilayer movement in protein-free phospholipid vesicles (17,18). The alternative explanation, i.e. that the lyso-PC molecules produced by PLA 2 do not participate in the transmembrane equilibration of PC, is considered more likely.
The microsomal contamination present in the preparations of OMV requires consideration in view of the rapid rate of PC flip-flop occurring in rat liver microsomes (t1 ⁄2 ϳ 2-3 min at 37°C (17)) and the proposed role of a specialized ER fraction (mitochondria-associated membrane) in phospholipid transport to mitochondria (4,8). Based on the specific activity of the microsomal marker enzyme glucose-6-phosphatase in OMV preparations and in microsomes isolated according to Ref. 14, the microsomal contamination of the OMV does not exceed 9.4% based on protein. 2 Given the phospholipid:protein ratio of microsomes (1), it can be calculated that this percentage cor-responds to a contamination of at most 4% on a phospholipid basis. Consequently, microsomal PC hardly contributes to the total PC pool. Vance (8) has shown that the specific activity of glucose-6-phosphatase is higher in mitochondria-associated membrane than in microsomes, whereas that of NADPH-cytochrome c reductase is lower. There are no indications that the microsomal fraction contaminating the OMV has characteristics of mitochondria-associated membrane, as the specific activities of both enzymes in the OMV relative to those in the rat liver homogenate are similar (0.45 and 0.55, respectively). 2 In vitro studies on the import of PC by rat liver mitochondria have shown that labeled PC is confined to the outer membrane after its introduction by PCTP from phospholipid vesicles (11,20). These results raised the question of which of the putative consecutive steps in intramitochondrial PC transport is limiting, the movement across the outer membrane or the transfer from outer to inner membrane. The present study strongly argues that the first step is not limiting. Calculations of the pool size of exchangeable PC in the outer membrane of intact mitochondria (20) suggest that the occurrence of transmembrane equilibration in the outer membrane depends on the acyl chain composition and the type of label of the PC molecule. In the case of [ 14 C]DOPC the size of the exchangeable PC pool in the outer membrane of intact mitochondria was found to be 56%, indicating that the label stays in the outer leaflet (11). Combined with the present results this suggests that the inner membrane might have a regulatory role in the transmembrane movement of PC across the outer membrane.
In a study on the topology of phospholipids in mitochondrial OMV from Saccharomyces cerevisiae, the transmembrane movement of PC in the outer membrane was reported to proceed relatively slowly (t1 ⁄2 ϳ 50 min (29)). These authors used a different, non-PC-specific transfer protein and a different donor-acceptor system, which complicates direct comparison with the present data.
In vivo labeling in rat hepatocytes has shown that 15 min after the addition of [ 14 C]choline, the specific radioactivity in the mitochondrial outer membrane is approaching equilibrium with that in the microsomes, which is consistent with a rapid transmembrane movement of PC in the outer membrane (9).
The existence of rapid PC transmembrane movement across the mitochondrial outer membrane makes a lot of sense from the point of view of biology. Since mitochondria are highly dynamic organelles that should have easy access to newly synthesized membrane components, it does not come as a surprise that the outer membrane has a mechanism for rapid shuttling of phospholipids. Comparable fast rates of energyindependent phospholipid transmembrane movement have been reported in membranes that are sites of phospholipid synthesis, the ER membrane (18, 30) and the inner membrane of Escherichia coli (31).