Isolation of Plasma Membrane Vesicles from Rabbit Skeletal Muscle and Their Use in Ion Transport Studies*

A method has been developed for the isolation of sealed plasma membrane vesicles from rabbit white skeletal muscle. The final preparation was highly pu- rified ds indicated by enrichment of plasma membrane marker enzymes (ie. ouabain-sensitive (Na+,K+)-ATP- ase, adenylate cyclase, and acetylcholinesterase). The absence of sarcoplasmic reticulum and mitochondria as contaminants was indicated by the low specific activity of marker enzymes, ie. Ca2+-ATPase, succinate- cytochrome c reductase, and monoamine oxidase. Thin section and negative staining electron microscopy con- firmed the absence of sarcoplasmic reticulum and mitochondrial contamination. The plasma membrane preparation consisted largely of sealed vesicles as observed by electron microscopy and as also demonstrated by latency of enzymic activities, which were unmasked by preincubation with de- tergent (sodium dodecyl sulfate). Membrane sidedness was estimated from latency of ouabain-sensitive (Na+,K+)-ATPase activity and acetylcholinesterase activity. The latency studies suggest that most of the vesicles are oriented inside out with respect to the orientation of the sarcolemma membrane in the muscle fiber. The inside-out plasma membrane vesicles actively accumulated sodium ions upon addition of ATP. The sodium ions were concentrated greater than %fold inside the vesicles and were released upon addition of the ionophore monensin. The sodium ions were taken up in the presence of K+ or NK' but not of choline. Uptake was inhibited by low concentrations of vanadate or digitoxin. The Na+ uptake was concomitant with Rb'

Isolation of Plasma Membrane Vesicles from Rabbit Skeletal Muscle and Their Use in Ion Transport Studies* (Received for publication, July 7, 1982) Steven SeilerS and Sidney Fleischer From the Department of Molecular Biology, Vanderbilt University, Nashville, Tennessee 37235 A method has been developed for the isolation of sealed plasma membrane vesicles from rabbit white skeletal muscle. The final preparation was highly purified ds indicated by enrichment of plasma membrane marker enzymes (ie. ouabain-sensitive (Na+,K+)-ATPase, adenylate cyclase, and acetylcholinesterase). The absence of sarcoplasmic reticulum and mitochondria as contaminants was indicated by the low specific activity of marker enzymes, ie. Ca2+-ATPase, succinatecytochrome c reductase, and monoamine oxidase. Thin section and negative staining electron microscopy confirmed the absence of sarcoplasmic reticulum and mitochondrial contamination.
The plasma membrane preparation consisted largely of sealed vesicles as observed by electron microscopy and as also demonstrated by latency of enzymic activities, which were unmasked by preincubation with detergent (sodium dodecyl sulfate). Membrane sidedness was estimated from latency of ouabain-sensitive (Na+,K+)-ATPase activity and acetylcholinesterase activity. The latency studies suggest that most of the vesicles are oriented inside out with respect to the orientation of the sarcolemma membrane in the muscle fiber.
The inside-out plasma membrane vesicles actively accumulated sodium ions upon addition of ATP. The sodium ions were concentrated greater than %fold inside the vesicles and were released upon addition of the ionophore monensin. The sodium ions were taken up in the presence of K+ or NK' but not of choline. Uptake was inhibited by low concentrations of vanadate or digitoxin. The Na+ uptake was concomitant with Rb' efflux. Therefore, the sodium ion transport and the resulting gradients formed appear to have been generated by the ouabain-sensitive (Na+,K+)-ATPase. Batrachotoxin, which opens Na+ channels in excitable tissues, prevents most of the Na' uptake, suggesting the presence of toxin-activated Na+ channels in these plasma membrane vesicles.
Skeletal muscle plasma membrane generates ionic gradients which are essential for excitable function. Maintenance of these ionic gradients is provided by vectorial transport of Na' outward and K+ inward across the plasma membrane. The * This work was supported by Grant AM14632 from the National Institutes of Health, by a grant from the Muscular Dystrophy Association of America, and by a National Institutes of Health Biomedical Research Support grant and University Research Council grant (Vanderbilt University). 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, Department of Physiology, Vanderbilt School of Medicine, NashviUe, Tennessee 37235. transport of these ions is mediated by the (Na',K+)-pump also referred to as the (Na',K')-ATPase. The resulting downhill, inwardly directed, Na' electrochemical gradient is also utilized to drive uptake of amino acids (1) as well as the countertransport of Ca2+ via a Na'/Ca2' exchange mechanism (reviewed in Ref. 2).
The availability of a purified plasma membrane vesicle system makes possible the study of plasma membrane composition and function uncomplicated by the presence of other organelles. The enrichment of fractions in plasma membrane from skeletal muscle has been carried out in a number of laboratories using a variety of approaches (3-17). This is the first report of the isolation of skeletal muscle plasmalemma as vesicles which measures vesicle integrity and membrane sidedness and uses such vesicles for transport studies. A preliminary report of this work has already appeared.'

EXPERIMENTAL PROCEDURES
Materials-Radioactive ["C]tryptamine bis-succinate, "Na, 86Rb, and '%a were obtained from New England Nuclear. Sodium dodecyl sulfate that was used for sample preincubations before assay was a specially purified grade obtained from BDH Chemicals Ltd. (Poole, England). SDS2 that was used to stop reactions was obtained from Fisher Scientific (Pittsburgh, PA). Imidazole (A grade), HEPES (ULTROL grade), and the ionophores A23187 and monensin were obtained from Calbiochem-Behring. Dextran T-10 was obtained from w/w), filtered through a 0.45-pm Millipore filter (Bedford, MA), and Pharmacia (Uppsala, Sweden) and was made as a stock solution (30% from EM Laboratories (Elmsford, NY) or Schwarz/Mann. Ouabain, stored frozen until used. Density gradient grade sucrose was obtained acetylthiocholine, valinomycin, digitoxin, phosphocreatine (di-Tris salt), rabbit muscle creatine phosphokinase, and rabbit muscle adenylate kinase were obtained from Sigma. Glutaraldehyde was obtained from Polysciences, Inc. (Warrington, PA).
Initially, both the sodium and Tris salts of adenosine 5"triphosphate (both substantially vanadium-free) were obtained from Sigma Chemical. In later studies, Tris/ATP was made from the NazATP Tris (AG 50W-X8,100-200 mesh, from Bio-Rad). The ATP was kept (Sigma) by passing it over a cation exchange resin equilibrated with as a stock solution of 0.2-0.3 M, pH 7.0-7.2, the concentration of which was determined spectrophotometrically (molar extinction coefficient at 259 nm was 1.54 X l@ liters/mol.cm).
Batrachotoxin was the kind gift of Dr. John Daly, National Institutes of Health (Bethesda, MD). It was made up as a stock solution of 5 mM in absolute ethanol and diluted into the assay media to the desired concentration.
SDS-polyacrylamide gel electrophoresis was performed using a modification of the method of LeStourgeon and Beyer (28).
Inorganic phosphate released was measured using the method of Ottolenghi (20) with the modifications that reagent IV contained 12% (w/v) ascorbate, 20% (w/v) trichloroacetic acid, 1% (w/v) molybdate; 0.5 ml of this reagent instead of 1 ml was used.
Acetylcholinesterase-The membrane fractions were pretreated with SDS before assaying acetylcholinesterase activity as described for the (Na',K+)-ATPase assay. Acetylcholinesterase was measured spectrophotometrically at 23 "C using acetylthiocholine as substrate according to the method of Steck and Kant (21). The time-dependent change in absorbance at 412 nm (extinction coefficient of 1.36 X lo4 liters/mol. cm) was used to calculate the acetylcholinesterase activity.
CuZ+-ATPase-The Ca'+-stimulated ATPase was measured in the presence of the ionophore A23187. The addition of ionophore made the activity linear with respect to time and protein concentration. Ca2+-stimulated ATPase was measured by release of inorganic phosphate in the presence ( 5 0 p~) and absence of CaZ' (1 m~ Tris/ EGTA). The 0.5-ml final reaction volume contained 100 m~ KCI, 3 m~ MgCl,, 3 mM Tris-ATP, 30 m~ imidazole, pH 7.2, 5 pg/ml of ionophore A23187,2-15 pg of protein, and either 50 p~ CaCl, or 1 mM Tris/EGTA. The reaction was carried out at 25 "C for 5 min and then terminated by the addition of 0.5 ml of 5% (w/v) SDS, 10 mM Na,EDTA and placement on ice. Inorganic phosphate was determined as described earlier for the (Na+,K+)-ATPase activity.
Monoamine Oxiduse-Monoamine oxidase activity was measured by a modification of the method of Wurtman and Axelrod (23). In order to make the activity linear with respect to protein concentration, the reaction medium (300 pl final volume) contained 50 m~ KP04, pH 7.4,2 m~ EDTA, 0.4 m g / d of bovine serum albumin, 3 p~ ['*C] tryptamine bis-succinate (3 X lo5 dpm/ml), and air as the gas phase.
The reaction was started by the addition of sample and incubated for 20 min at 37 "C before termination by addition of 200 pi of 2 N HCI and placement on ice.
Succinate-cytochrome c Reductase-This was measured at 32 "C as described previously (24), and adenylate cyclase activity was measured at 37 "C by the method of Jakobs et al. (25).
ATP-dependent Zon Transport-Ion transport was measured using the Millipore filtration method. A protocol similar to that of Lau et al. (29) was used initially. However, we found that Na+ uptake was increased with higher concentrations of Na' and KC and in the presence of an ATP-regenerating system (creatine phosphokinase and creatine phosphate) (30, 31). A longer pre-equilibration (3 h on ice and 30 min at 37 "C) was used to obtain ionic equilibration. Unless otherwise stated, the samples were preincubated for 3 h in a medium containing 10 mM NaCl (2-3 X lo6 cpm/ml of "Na), 100 mM KCl, 5 m~ MgC12,5 units/ml of creatine phosphokinase, 10-15 m~ creatine phosphate ( d i -T r i s salt), 40 m~ imidazole/HEPES, pH 7.3, and a protein concentration of 0.04 mg/ml. The sample was then incubated 30 min at 37 "C, and, after ion equilibration, an aliquot was withdrawn (zero time point) to determine the amount of Naf in the vesicles prior to the addition of ATP. Active transport was initiated by the addition of 5 mM Tris/ATP (pH 7.0-7.2). Aliquots (0.5 ml containing 20-50 pg of protein) were withdrawn and rapidly filtered on 0.22-pm Millipore filter discs (GSWP) (Bedford, MA). The filters were immediately washed three times (2 ml/wash) with 0.5 M NaCl, 0.5 M NazS04 or the reaction medium minus ATP and radioactive tracer. Similar results were obtained with each of these media. The filters were blotted on absorbent tissue (Kimwipes, Kimberly Clark Corp., Neenah, WI) to remove excess adherent filtrate. The filters were then placed in glass vials and 2 ml of ethylene glycol monoethyl ether (Cellosolve) were added. After the filters dissolved, counting scintillant (10 ml of ACS, Beckman Instruments) was added and the radioactivity was measured.
Rubidium (=Rb), which can substitute for potassium in activation of the (Na+,K+)-ATPase (32), was used as a tracer without the addition of unlabeled Rb for potassium transport experiments. The "Na and =Rb tracer studies were carried out using identical samples and assay media except for the isotope. A 10-fold higher radioactivity for Rb (20 X lo6 cpm/ml) was used. An extensive wash of the fiters was employed to remove the extravesicular radioactivity. Experimental details are described in legend to Fig. 10.
Procedure for Zsolutwn of Plasma Membrane Vesicles-Female New Zealand White rabbits (1-1.8-kg weight) were killed by cervical dislocation. All subsequent steps were carried out at 0-4 "C unless stated otherwise. The white muscles of the hind legs were separated from red muscle, large blood vessels, nerves, tendon, and fascia. Sharp dissection was performed on an ice-packed glass tray. About 40 g of trimmed muscle were obtained from a single rabbit leg. The muscle was then ground in a meat grinder (General Model H meat grinder), the meat being forced through a plate with 2-mm holes.
Typically, the preparation began with 200 g of ground leg muscle obtained from three rabbits. The ground muscle (50 g at a time) was blended in 250 ml of wash buffer (0.75 M KCl, 5 mM imidazole, pH 7.4) using a Waring blender at half-maximal speed for 5 s. The blendate was centrifuged for 20 min at 9,000 rpm in a J A 10 rotor (Beckman model 5-21 centrifuge) and the supernatant was discarded. The pellet was resuspended with a stirring rod. The wash procedure was repeated twice, the first time in wash buffer and the second time in 250 ml of 0.3 M sucrose, 5 mM imidazole, pH 7.4. The pellet was then homogenized in 250 ml of 0.3 M sucrose, 5 mM imidazole, pH 7.4, using a Waring blender at full speed for 1 min. The homogenate was centrifuged in a JA 10 rotor at 9,000 rpm for 20 min and the pellet was discarded. The supernatant was filtered through several layers of cheesecloth to remove free floating fat and the filtrate was centrifuged at 30,000 rpm for 2 h in a Beckman type 35 rotor. The microsomal pellet was resuspended in 40-60 ml of ice-cold 0.6 M KCl, 4 mM imidazole, pH 7.4, with several strokes of a loose fitting Dounce homogenizer and maintained on ice for 1-3 h. The microsomes (approximately 125 mg of protein in 10-12 ml/centrifuge tube) were layered onto a sucrose density step gradient consisting of 17% (w/w) sucrose (10 ml) and 23% sucrose (10 ml), buffered with 5 mM imidazole, pH 7.4, and 50% sucrose (7 ml) as a cushion. The gradients were centrifuged in a Beckman SW 27 rotor at 20,000 rpm for 12-16 h.
PMs was at the 17/23% sucrose interface (Fig. L4). Most of the microsomal protein migrated to higher density and consisted predominantly of sarcoplasmic reticulum and mitochondria.
PMs was usually purified further and enriched with respect to sealed vesicles by centrifugation on a discontinuous dextran T-10 gradient. The sucrose gradient fraction was diluted with an equal volume of 5 m~ imidazole, pH 7.4, and the vesicles were recovered by sedimentation at 35,000 rpm for 2 h. The pellets were resuspended in 5 mM imidazole, pH 7.4. Typically, 2-3 mg of partially purified plasma membrane were layered on top of a discontinuous dextran T-10 gradient (buffered with 5 m~ imidazole, pH 7.4). The two-step gradient, containing 8.7 and 18% dextran T-10, was centrifuged in a Beckman SW 56 rotor or SW 41 rotor for 16-20 h at 35,000 rpm. The fraction at the 5 m~ imidazole and 8.7% dextran T-10 interface (Fig.  1B) contained the highly purified plasma membrane vesicles (PMd).
The membrane fraction was diluted with an equal volume of 5 mM imidazole, pH 7.4, and recovered by centrifugation at 35,000 rpm for 2 h. The pellets were resuspended in 0.3 M sucrose, 5 mM imidazole, pH 7.4, to a protein concentration of 3-10 mg/ml and were quick frozen in liquid nitrogen and stored at -65 "C until used.

Isolation and Characterization of the Plasma Membrane
Vesicles from Skeletal Muscles-A method has been described for the isolation of plasma membrane vesicles from skeletal muscle (cf "Methods"). The ground muscle first received a limited blending and washing in 0.75 M KCl. The low speed sediment was then homogenized in isotonic buffered sucrose (0.3 M) and a microsomal fraction was obtained by centrifugation. PMs was obtained by isopycnic centrifugation

Muscle Phsmale
Purification of plasmalemma fractions on sucrose and dextran T-10 density gradients. A, sucrose gradient. Separation of the muscle microsomes was performed using a discontinuous sucrose gradient. The sample, approximately 9 ml, was layered onto the gradient containing 17% (IO ml), 23% (10 ml), and 50% (10 ml) sucrose and was centrifuged overnight in a SW 27 rotor. The plasma membrane fraction (PMs) was in Band 2 at the 17/23% interface. At the sample/l7% sucrose interface, Band 1 contained mostly aggregated contractile protein and leaky plasma membrane, as visualized by thin section electron microscopy. This fraction contained little in the way of plasma membrane diagnostic activities (Na',K')-ATPase was 8.0 pmol/mg.h (cf Table I for comparison). Band 3 contained "light" SR, transverse tubule, and plasma membrane. Band 4 contained about 80-85% of the microsomal protein and consisted mainly of sarcoplasmic reticulum vesicles and mitochondria as indicated by the high amounts of Ca"-ATPase and succinate-cytochrome c reductase activities in this fraction. B, dextran gradient. PMs was further purified on a discontinuous dextran gradient. The sample was layered onto a step gradient containing 8.7% (2 ml) and 18% (1 ml) dextran T-

10.
An SW 56 centrifugation tube shows the separation of the fraction into subpopulations. The plasma membrane vesicles (PMd) (Band 1 ) were obtained from the sample/8.5% dextran T-10 interface. Band 2 contains "light" SR and outer membrane of mitochondria and leaky plasma membrane vesicles. using a sucrose density step gradient (Fig. lA, Bund 2). PMs was further purified on an isopycnic, discontinuous, dextran T-10 gradient (Fig. lB, Bund 1). The yield was approximately 3 mg for PMs and 0.5-1.0 mg for PMd from 50 g of ground muscle. A typical preparation made use of three rabbits (200 g of ground muscle) which yielded 12 mg of sucrose gradient or 2-4 mg of the dextran gradient-purified plasma membrane vesicles, respectively. The preparation of PMd took 2 days due to the time required for the two isopycnic gradients.
The characteristics of the plasmalemma fractions are summarized in Table I. The sucrose gradient considerably enriched the fraction in plasma membrane from mixed microsomes ( Table I). The (Na',K')-ATPase was 13-fold higher than that of the microsomes, while SR and mitochondrial contamination were decreased by &fold and 5-fold, respectively, as judged by Ca*'-ATPase and succinate-cytochrome c reductase activities.
The plasma membrane fraction obtained from the dextran gradient fraction, although decreased in yield (Table I), contained still higher ouabain-sensitive (Na',K')-ATPase and acetylcholinesterase (1.4-and 1.8-fold higher than sucrose preparation, respectively) ( Table I). The dextran purification procedure further decreased the specific activity of succinatecytochrome c reductase, monoamine oxidase, and Ca2'-ATPase activities ( Table I), suggesting that the dextran T-10 gradient step also separated the sealed (see below) plasma membrane vesicles from intact mitochondria, outer mitochondrial membranes, and SR, respectively. Based on the amounts of contaminant marker activities, we estimated that the dextran plasma membrane fraction contained less than 1% sar-mmal Vesicles coplasmic reticulum and less than 1% mitochondria contamination (Table I).
Acetylcholinesterase co-isolated with the ouabain-sensitive (Na',K')-ATPase activity (Table I) in accord with the observations of others (35) that the highest specific activity of acetylcholinesterase was in sarcolemma-enriched fractions. The dextran plasma membrane fraction also exhibited a high specific adenylate cyclase activity (330 pmol/mg. min).
The purified skeletal muscle plasma membrane was examined by thin section electron microscopy. The fraction contained sealed membrane vesicles of irregular shape and varying in "diameter" from approximately 0.2-0.5 pm. Only a few of the vesicles had a flattened appearance characteristic of transverse tubules (Fig. 2 A ) . Transverse tubule-like vesicles constituted only a minor percentage of the total vesicle population.
Thin section electron microscopy confirmed the absence of mitochondria and sarcoplasmic reticulum. The latter has a characteristic membrane asymmetry which can be detected by tannic acid enhancement (27). Negative staining did not reveal vesicles with 90-8, surface particles which are diagnostic TABLE I Characteristics of fractions in the purification of skeletal muscle plasma membranes PMs and PMd refer to plasma membrane fractions obtained from sucrose gradient and dextran T-10 gradient, respectively. The numbers in parentheses represent the number of preparations analyzed.
Ouabain-sensitive ATPase as obtained by subtracting the values obtained in the presence of ouabain from the total ATPase (above). e Ten-min time point, measuring uptake capacity at 25 "C in a medium containing 5 m~ NaCI, 50 mM KC1, 5 mM MgC12, 30 mM imidazole, 1 mM EGTA, and 5 m~ Tris/ATP.
Purified sarcoplasmic reticulum (36) has Ca2'-ATPase activity of 3.0 pmol/mg.min measured under the conditions described. This value was used to estimate the upper limit of SR contamination. The microsome fraction contained 65-70% sarcoplasmic reticulum, the sucrose gradient fraction contained 1055, and the dextran T-lO-purified fraction contained less than 1% sarcoplasmic reticulum contamination.
e No available data exist on the specific activity of monoamine oxidase activity of purified skeletal muscle outer mitochondrial membrane. The pig heart outer mitochondrial membrane has an activity of 12.8 nmol/mg.min using benzylamine as substrate (37).
'Purified bovine heart mitochondria have a succinate-cytochrome c reductase rate of approximately 800-900 nmol/mg.min (38). Using this value, the microsomal fraction and PMs contained approximately 5% and 1% mitochondria, respectively. The mitochondrial contamination was further reduced to 0.3% in the PMd plasma membrane fraction. of submitochondrial vesicles (Fig. 2). We did not observe any muscle filaments or collagen in the purified fractions of plasmalemma, either PMs or PMd.
The protein profiles of sucrose-and dextran-purified sarcolemma were compared by SDS-polyacrylamide gel electrophoresis (Fig. 3). A variety of polypeptides of different molecular weights were found in the plasma membrane-enriched fractions. PMd has a significantly different relative intensity of bands compared with PMs. Most notably, the 100,000-and 40,000-dalton bands were decreased and the 65,000-and 30,000-dalton components are increased. Few of the polypeptides have been identified with regard to their functional activities. This gel pattern is distinctly different and more complex than that of purified sarcoplasmic reticulum (36).
The phospholipid content in PMd was twice the protein content on a weight basis, based on the phosphorus to protein ratio (Table I), and the amount was increased over that of PMs. The cholesterol to phospholipid molar ratio was 0.39." Estimate of Sidedness-The estimate of sidedness is based on two assumptions. 1) Acetylcholinesterase is localized only S. Seiler, L. Jones, and S. Fleischer, manuscript in preparation.
on the outer face of the plasmalemma and is exposed in right side-out vesicles; and 2) ouabain-sensitive (Na',K+)-ATPase cannot be measured in sealed vesicles. Ouabain-sensitive (Na',K')-ATPase activity should not be measurable in sealed plasma membrane vesicles regardless of orientation of the vesicles since ATP and ouabain bind on opposite membrane faces of this transmembrane pump (40, 41). In sealed right side-out vesicles, the ouabain but not the ATP binding would be accessible so that ATPase activity would not be measured (Fig. 4), whereas the opposite situation would pertain for sealed inside-out vesicles. (Na',K')-ATPase activity would be expressed but would not be inhibited by the ouabain.
The latent activities can be expressed by preincubation of the plasma membrane fraction with SDS (Fig. 5), which makes the vesicles leaky so that both the ouabain-sensitive (Na',K')-ATPase and acetylcholinesterase activities are expressed. The amount of detergent required in order to obtain maximal enzyme activity was the same for both the (Na+,K+)-ATPase and the acetylcholinesterase (Fig. 5), consistent with a common action for detergent activation, i.e. the sealed vesicles become leaky. The enzymic activity declined somewhat when excess detergent was used in the preincubation  9 4 , o O O ; bovine serum albumin, 68, oOO, ovalbumin, 43,000; carbonic anhydrase, 30,000, soybean trypsin inhibitor, 21,000, and lysozyme, 14,000. medium. Therefore, it was desirable to optimize detergent concentration. The optimal SDS concentration for preincubation depended on the purity of the sample being assayed. It was 0.2 and 0.4 mg of SDS/ml for 1 mg/ml of membrane protein for the microsomes and the plasma membranes purified a t either gradient stage, respectively. The increased optimal SDS concentrations paralleled the increased lipid content of the membrane.
The percentage of the sealed vesicles was estimated by measuring the percentage of latent activity. Approximately 71% of the ouabain-sensitive (Na',K')-ATPase of the sucrose gradient-enriched plasma membranes was latent, and 85% was latent in the dextran fraction (Table 11). Therefore, the dextran gradient purification removed some 14% of unsealed vesicles. Acetylcholinesterase was used as an index of right side-out and leaky vesicles (Fig. 4). Approximately 3 4 % of the acetylcholinesterase was measured in PMd without detergent treatment and, subtracting 15% leaky vesicles, we estimate that 19% were right side-out. Most of the vesicles (66%) were sealed and inside out ( Table 11).
The sidedness of the purified plasmalemmal vesicles was also estimated from the ouabain inactivation of the (Na+,K+)-ATPase activity, per se, assayed in the presence of the ionophore monensin (Fig. 7). Monensin renders the vesicles permeable to Na' and K' , thereby eliminating the electrochemical gradient which would otherwise build up and retard the (Na',K')-ATPase activity in a sealed vesicle. When monensin-treated vesicles were pretreated with detergent, the enhanced activity (23%) estimated the sealed right side-out vesicles (Fig. 6, C versus E ) . The increase in the ouabainsensitive ATPase activity by detergent pretreatment (Fig. 6, D uersus F ) (50%) is a measure of the sealed inside-out plasma membrane vesicles. By difference from loo%, approximately 27% of the vesicles were estimated to be leaky by this procedure.
The two methods to measure membrane sidedness ( Fig. 6 and Table I1 and Fig. 5) gave similar results albeit with some consistent differences.
Na' Transport Studies-Both types of plasma membrane preparations (PMs and PMd) concentrated Na' upon energization with ATP (Fig. 7). The amount of Na' taken up reached a plateau within approximately 12 min. The dextran- and acetylcholinesterase (0) is presented as a function of SDS concentration in the preincubation medium. The PMs (1 m g / d of protein) was pretreated as described with varying concentrations of SDS before diluting into the assay mixture. This particular sample (PMs) had a maximal ouabain-sensitive (Na+,K')-ATPase activity of 37.8 pmol of Pi/mg. h and acetylcholinesterase activity of 153 nmol/ mg min. This assay forms the basis for estimating membrane sidedness (Table I1 and text). Values for both PMs and PMd are summarized in Table 11. ZO, Inside-out; RO, right side-out; L, leaky. Acetylcholinesterase (nmol/mg.min) 73.0 (2) 153 (2) 53 (2) 96.0 * Per cent latent activity = (total activity (SDS-pretreated)activity (no SDS))/(total activity (SDS-pretreated)).
e The latency of ouabain-sensitive (Na+,K')-ATPase is a measure of sealed vesicles both right side-out and inside-out (85%). The difference The latency of acetylcholinesterase is a measure of inside-out vesicles (66%) and when subtracted from the percentage of sealed vesicles from 100% gives the percentage of leaky vesicles (15%).
(85%) gives the percentage of right side-out vesicles (19%). ATPase activity is given under different conditions. A, in the presence of 120 n" NaCl, 20 m~ KCl, 30 m~ imidazole/Cl, pH 7.5, 4 m~ MgC12,0.5 m~ Tris/EGTA, 5 m~ NaN3, and 4 m~ Na2ATP at 37 "C. B, as in A but with 1 m~ ouabain, C, as in A but with 1 p~ monensin. 0, as in C but with 1 m~ ouabain. E, as in A after pretreatment with 0.4 mg/ml of SDS. F, as in E with 1 m~ ouabain. Abbreviations are as in Fig. 5. purified fraction exhibited both a faster uptake rate and a larger capacity than PMs, consistent with the increased purity and increased amount of sealed vesicles (Tables I and 11). A concentration gradient was generated which was %fold greater inside. This is a minimum estimate and could be closer to 10fold when correction is made for approximately 20% right side-out vesicles.
The addition of the Na+ ionophore monensin (Fig. 7) caused rapid release of Na+ from the vesicles, confirming that Na+ was taken up against a concentration gradient. At the concentration used, monensin had little effect on the (Na+,K+)-ATPase activity after detergent pretreatment and, therefore, did not appear to inhibit the Na' pump activity.
The effect of various other ions on the Na+ uptake was determined to further characterize the Na+-pumping activity. NH,' and K+ ions facilitated the Na+ uptake, whereas choline did not (Fig. 8). Ammonium ion stimulated the Na+ uptake more effectively than K+. Replacing the potassium in the uptake medium with choline inhibited most of the Na+ uptake (Table 1 1 1 and Fig. 8). However, there was a small amount of Na+ uptake (~20%) in the absence of K+ as has already been described for the (Na+,K+)-pump (49). These observations are consistent with the known characteristics of the (Na+,K+)pump (32,48). Substituting 5 m~ NaN3 for 5 mM NaCl in the assay medium did not change the kinetics of Na+ uptake and final capacity, confirming that mitochondrial ATPase was not involved.
Low concentrations of digitoxin (10 PM), a lipid-soluble cardiac glycoside (44), completely inhibited the sodium uptake (Fig. 8). Ouabain, an impermeant cardiac glycoside, had no effect on the Na+ uptake rate (Table 111), consistent with the view that the inhibitory site of ouabain is on the interior of the vesicle.
Vanadate, at low concentrations, inhibited the Na+ uptake ( Figs. 9 and 11). The concentration required for half-maximal inhibition was between 100 and 200 n~, similar to that reported by Karlish and Pick (47) for reconstituted (Na+,K+)-ATPase isolated from pig kidney outer medulla. Vanadate has been shown to bind to the site of ATP hydrolysis of the (Na+,K+)-ATPase (45,46) that is on the outer face of insideout vesicles (Fig. 4). Rubidium Efflux Studies-By pre-equilibrating the vesicles with =Rb, which can substitute for K+, we confiied that Na+ inward pumping was coupled to Rb' efflux as expected for an operating (Na+,K+)-pump (50, 51). Not all of the =Rb was pumped out of these vesicles as sodium was pumped in (Fig. lo), perhaps reflecting a higher permeability of the vesicle to Rb' or K+ under these conditions as well as the presence of some right-side-out vesicles which do not pump ions under the conditions of the assay.
Toxin-activated Nu+ Channels-The plasma membrane of electrically excitable tissues contains Na+ channels that increase membrane sodium permeability when depolarized or when incubated with certain plant or animal toxins (52). Batrachotoxin, at low concentrations, causes depolarization of excitable membranes by opening Na+ channels (53).
In order to investigate the presence of Na+ channels in our isolated plasma membrane vesicles, we incubated them with batrachotoxin (10 p~) , which prevented most of the ATPdependent Na+ accumulation by the plasma membrane vesicles (Fig. 11). Batrachotoxin (10 p~) had no effect on the ouabain-sensitive (Na+,K+)-ATPase activity measured after pretreatment with SDS, suggesting that batrachotoxin did not . 7 (left). Active transport of sodium into plasma membrane vesicles. The active accumulation of "Na was measured using the Millipore filter technique. PMs (A) (100 pg/ml) and PMd (0) (60 g/ml) were preincubated for 30 min at 25 "C in an uptake medium containing 5 m MgClZ, 30 m imidazole, pH 7.2, 1 m EGTA, 50 m KC1, and 5 m NaCl (2 X lo6 cpm/ml). The uptake reaction was started at zero time by the addition of 5 m Tris/ATP. At the indicated time, 0.5-ml aliquots of sample were withdrawn from the assay tube and filtered on 0.22-pm Millipore filters. The filters were washed twice with 2-ml washes of 0.5 M NaCl and prepared for scintillation counting. Arrow, the addition of monensin ( M ) (1 p~) . At the concentration used, monensin had no effect on the detergenttreated (Na+,K+)-ATPase activity.  . 10 (left). Sodium uptake occurs concomitantly with potassium (=Rb) efflux. PMd was preincubated in the cold for 2 h in an uptake medium containing 5 m NaC1,50 nm KC1,5 m MgClz, 10 nm creatine phosphate, 5 units/ml of myokinase with 15 mM (NH&S04), 10 units/ml of creatine phosphokinase, 30 m 4-mOrphOlinepropanesulfonic acid, pH 7.2, with 2 X IO6 cpm/ml of "Na (A) or 2 X lo7 cpm/ml of =Rb (0). The sample was then temperatureequilibrated at 25 "C for 30 min. After ATP was added, aliquots (0.5 m l ) were taken and filtered. The Nters were washed four times with 2-ml washes of 0.5 M NazS0, prior to preparation for counting. FIG. 11 (right). Inhibition of ATP-dependent sodium uptake by batrachotoxin and vanadate. The sample was preincubated 3 h on ice in a medium containing 10 m NaC1,100 m K (85 m C1, 15 m 4-morpholinepropanesulfonic acid salts), 1 m EGTA, 5 mM MgC12, 40 pg/ml of sample (PMd), 30 m 4-morpholinepropanesulfonic acid, pH 7.2, *'Na (2 X IO6 cpm/ml), 15 m Tris/creatine phosphate, 10 units/ml of creatine phosphokinase, and 5 units/ml of myokinase with 15 m (NH4)zSO~. The sample was then temperature equilibrated for 1 h at 37 "C with the indicated effector. Na+ uptake p~ batrachotoxin and 0.3% ethanol (W) and 10 p~ vanadate (0) was was measured as described in Fig. 7. Na+ uptake in the presence of 10 measured. The control (A) also contained 0.3% ethanol. inhibit the Na+ pump; this is consistent with the observations of others (55). This experiment suggests that most of the vesicles had Na+ channels which were activated by batrachotoxin since nearly all the Na' accumulation was prevented ice and then at 37 "C for 30 min. The Na+ uptake was measured as described in Fig. 7. FIG. 9 (right). Inhibition of ATP-dependent sodium uptake rate by vk-adate. Initial Na uptake rat,& (obtained from 30-i time points) were measured in an uptake medium consisting of 10 mM NaCl (2 X IO6 cpm/ml of "Na), 100 m KCl, 20 nm HEPES/ imidazole, pH 7.3, 1 m EGTA, 5 m MgClz, and 40 pg/ml of protein (PMd) with the indicated concentrations of Na3V04. The sample was preincubated on ice for 3 h and then for 30 min with the indicated concentration of N&V04. Thirty seconds after the ATP was added, 0.5-ml aliquots were rapidly withdrawn, filtered, and washed twice with 0.5 M NaC1,5 m Tris-Cl, pH 7.4. The amount of Na+ pumped in the presence of ATP minus the amount of Na' passively equilibrated was used to calculate the net uptake rate. Each point is the average of two determinations. The apparent K, under these conditions is approximately 150 m.

TABLE I11
The effects of cardiac glycosides on sodium uptake rate Initial Na uptake rates (obtained from 30-s time points) were measured on two different dextran gradient preparations (A and B) in an uptake medium consisting of 10 m NaCl (2 X IO6 cpm/ml of 22Na), 100 mM KC1,20 nm HEPES/imidazole, pH 7.3, 1 mhf EGTA, 5 m MgC12, 40 pg/ml of protein with effectors as given in the table.
One millimolar KPO, was added to facilitate cardiac glycoside binding. The reaction was started with the addition of 5 m~ Tris/ATP. Thirty s after the ATP was added, a 0.5-ml aliquot was rapidy withdrawn, filtered, and washed twice with the uptake medium minus ATP and "Na tracer. The amount of Na+ pumped in the presence of ATP was corrected for the amount of Na' passively equilibrated. The data represent the mean * the standard deviation. The numbers in parentheses denote the number of determinations. by batrachotoxin (Fig. 11). Veratridine (100 /AM) also prevented Na+ accumulation (data not shown), but was not as effective as batrachotoxin ( 5 4 ) . DISCUSSION A method is described for isolating sealed plasmalemma vesicles from rabbit skeletal muscle. The preparation is highly enriched in plasma membrane marker enzymes, containing the highest specific activity of ouabain-sensitive (Na+,K+)-ATPase, adenylate cyclase, and acetylcholinesterase activities (Table I) reported for a skeletal muscle plasmalemma preparation (Table IV). It is practically devoid of mitochondria and sarcoplasmic reticulum as indicated by low activities of diagnostic enzymes (Table I). The vesicles are largely impermeable to substrates and ions, making them especially suitable for ion transport studies. The sidedness has been characterized and found to be mostly inside-out. These vesicles are capable of vectorial countertransport of Na' and K' as reported here, as well as Na/Ca exchange? The vesicles contain Na' channels as demonstrated by the ability of batrachotoxin to prevent Na' accumulation. The dextran gradient was used to separate sealed from leaky vesicles. Dextran does not penetrate into the sealed vesicles so that their buoyant density is lower compared with the leaky membranes (56). The purification step removes contamination by other membranes as well (Table I). The plasma membrane fraction is practically devoid of the major sources of membrane contaminants, mitochondria and sarcoplasmic reticulum. Polyacrylamide gel electrophoresis does not reveal major bands referable to actin and myosin, and electron microscopy does not reveal muscle filaments or collagen.
The plasmalemma preparation makes use of young rabbits (1.8 kg or less). When larger rabbits were used (2.0 kg or greater), double membrane vesicles contaminated the plasmalemma at the sucrose gradient step (57) which could not be removed by centrifugation on a dextran T-10 gradient. Preparations from even larger rabbits contained multilamellar vesicles, possibly originating from peripheral nerve myelin and from which the double walled vesicles seem to have been derived. The problem of the double membrane vesicles in the purified preparation was practically eliminated with the use of 2-kg or smaller rabbits and resulted also in much higher specific activity of plasma membrane marker enzymes (57).
The purified plasmalemma preparation appears to be stable to storage in the frozen state. There is no loss of (Na+,K+)-ATPase activity observed with freezing and thawing of either (PMs or PMd) purified plasmalemma preparation. The ion gradient generating capability can also be preserved for months. However, at the microsome stage of purification, approximately one-half of the (Na',K')-ATPase activity was inactivated by freezing and thawing. This is the first report of isolated skeletal muscle plasmalemma vesicles which has been characterized with regard to membrane sidedness. We estimate that 50-66% of the vesicles (PMd) are sealed and inside out, whereas 19-25% of the vesicles are sealed and right side-out. The range expresses more the differences obtained between the two methods used to estimate sidedness (Table I1 and Fig. 6) than variation from preparation to preparation. The method using the ouabain sensitivity of the (Na+,K+)-ATPase in the presence of monensin tends to overestimate the percentage of leaky vesicles while indicating a decreased amount of inside-out vesicles. It would appear that the monensin treatment makes the vesicles somewhat more leaky.
Two approaches were used to estimate membrane sidedness in our studies. Both approaches use the latency of ouabainsensitive (Na+,K+)-ATPase for the quantitation of sealed uersus leaky vesicles. Sealed vesicles do not display this activity since both faces of the membrane must be accessible in order to measure this activity. The fiit procedure combines this assay with the latency of acetylcholinesterase to measure the amount of inside-out vesicles ( Fig. 5 and Table 11). The second procedure makes use of only the ouabain-sensitive (Na+,K+)-ATPase assay by itself. The percentage of right side-out vesicles is obtained as the difference in ATPase rate with and without detergent pretreatment (Fig. 6). Both procedures make use of SDS pretreatment to release the latency in order to express total activity. Monensin is an ionophore which, when added to the assay medium, dissipates the Na' and K' gradients generated by the (Na+,K+)-pump of the vesicles (Fig. 7). Therefore, initial rates of (Na+,K+)-ATPase in inside-out vesicles can be obtained. ATPase activity is enhanced because Na+, which would otherwise compete with K' for binding sites on the inner face, does not concentrate. Also, K' , which is required on the inner face for optimal (Na',K')-ATPase activity, can leak in.
The measurement of ATPase activity includes "basal" ATPase, i.e. MgATPase which is not sensitive to ouabain (Fig. 6). The latter must be subtracted from the ouabainsensitive (Na+,K+)-ATPase in order to estimate inside-out vesicles. To minimize the basal ATPase, EGTA and azide were used in the assay medium to reduce ATPase from sarcoplasmic reticulum and mitochondria, respectively. Inactivation of basal ATPase by detergent treatment would result in an inflated estimate of the amount of inside-out vesicles. Therefore, the use of detergent to optimize activity must be carefully calibrated.
There is little experience in the use of acetylcholinesterase to estimate membrane sidedness in muscle. Most of the acetylcholinesterase activity from muscle is associated with the extracellular surface of the plasmalemma as determined by histochemistry (42), although a small portion is located intra cellularly (64, 65). Some acetylcholinesterase appears to be associated with the basal lamina (66, 67); we have no indications that the latter isolates with the plasmalemma preparation described here. There is heterogeneity in the distribution of this enzyme in skeletal muscle sarcolemma. Approximately one-third is localized at the neuromuscular junction (64, 66,  68). Therefore, caution must be exercised since a small fraction of the vesicle population could skew the interpretation for the entire population. Pragmatically, the use of acetylcholinesterase to assess membrane sidedness is valid for this preparation. It gives simiiar results to the ouabain-dependent (Na+,K+)-ATPase method which makes use of monensin. Furthermore, a third estimate of membrane sidedness used in our laboratory, the latency of [3H]ouabain binding sites: leads to the same conclusion that most of the vesicles are sealed and oriented inside out.
The plasma membrane preparation described in this report contains mostly inside-out plasma membrane vesicles and is especially suitable for Na+ transport studies energized by ATP. This is the first such study with skeletal muscle plasma membrane. Na+ gradients can be generated which are 10-fold, comparable with that in the muscle fiber. It is clear that the pumping is achieved by the (Na+,K+)-pump of the skeletal muscle plasma membrane since 1) it is energized by ATP and there is concomitant Rb+ efflux with Na' uptake; 2) the characteristics of non-Na+ ions required for activation are similar to that reported for the (Na+,K+)-pump from other systems; 3) the pumping is inhibited by cardiac glycosides and vanadate comparable with the (Na+,K+)-pumps which have been studied; and 4) batrachotoxin and veratridine obviate Na+ accumulation, indicating that an excitable membrane containing Na+ channels is involved.
Lau and co-workers (29) have isolated a transverse tubule preparation from skeletal muscle which is mainly inside-out and is capable of energized Na+ uptake. The plasma membrane vesicle preparation described in this report is mainly inside-out in orientation, but the indications are that it derives mainly from the surface and not from a transverse tubule. A transverse tubule has a morphologically distinctive appearance, i.e. flattened tubules, 0.1 X 0.04 pm, with electron opaque "caps" at opposite ends Thin section electron microscopy of the plasmalemma preparation reveals more rounded although irregular shaped vesicles averaging 0.2-0.5 pm in diameter. It did not reveal appreciable transverse tubules (Fig.  2) although the latter were present in other fractions from the sucrose gradient (Fig. L4, Band 3)

.'j
The plasma membrane preparation (PMd) described here has severalfold higher (Na+,K+)-ATPase activity than has been obtained for isolated transverse tubules (Table IV). This is consistent with ouabain binding studies on intact tissue in which a much higher density of [3H]ouabain binding sites was measured on the plasmalemma than in the transverse tubule (62). It is interesting to note that the Ca2+-ATPase activity in our plasma membrane-enriched fraction is very low compared with that reported for transverse tubule (60, 61, 63).
The time required to reach maximal Na+ uptake capacity for these vesicles occurs within 12 min and is more than 5-fold faster than the transverse tubules isolated from triads (29). The Na+ flux is faster in PMd than transverse tubule, possibly reflecting the larger amount of (Na+,K+)-ATPase in PMd as compared with the transverse tubule preparation (Table IV). Progress in the study of heart plasmalemma sidedness and transport, thus far, has outpaced studies with skeletal muscle (30, 70, 71, 74, 75).
The best characterized cardiac sarcolemma preparation by Jones et al. (22) is largely right side-out. This preparation does not concentrate Na+ and would not be expected to since it is predominantly right side-out? Nonetheless, it has a 2-fold higher specific activity of ouabain-sensitive (Na+,K+)-ATPase and adenylate cyclase than does our skeletal muscle preparation. This difference between skeletal and cardiac muscle plasma membranes probably reflects different characteristics of the two types of muscle tissue.
Heart plasmalemma preparations have been reported which P. Volpe, S. Seiler, and S. Fleischer, unpublished data. are capable of concentrating Na+ (31, 69). Grosse et al. (31) have prepared cardiac vesicles which appear to be largely inside-out and these are capable of pumping Na+ with similar capacity to skeletal muscle PMd, although the rate of Na+ pumping and the concentration gradient which is generated are not given. An interesting aspect of their study is the interaction of the (Na',K+)-ATPase with creatine phosphokinase. creatine phosphokinase was found associated with their cardiac plasma membrane preparation (30), and an added ATP-regenerating system stimulated both the Na+ uptake and (Na+,K+)-ATPase. We also found that an ATPregenerating system stimulated Na+ uptake. There are major difficulties in the isolation of a defined muscle plasma membrane from skeletal muscle. The muscle fibers are entrapped in an extensive collagen network, making them resistant to disruption. The plasma membrane is heterogeneous, and the surface membrane represents only a small portion of the plasma membrane which invaginates, giving rise to transverse tubule (71). In frog skeletal muscle, Peachey (71) has estimated that there is 7 times more transverse tubule than surface membrane and there is 5-6 times as much sarcoplasmic reticulum as transverse tubule. Only recently have transverse tubule preparations been described (60, 61, 63), and their diagnostic characteristics are not known with any degree of certainty. There is no marker enzyme available to discriminate between transverse tubule and surface membrane. The characterization of membrane sidedness is not trivial. The aspect of membrane sidedness is a dimension which has not previously been considered for skeletal muscle plasma membrane isolation. Another problem is that contractile filaments, which comprise most of the muscle mass, serve to "glue" the components together, making membrane separation more difficult and the yield low.

5R. Mitchell
Early skeletal muscle plasma membrane isolation techniques yielded a "sarcolemma" or outer sheath preparation. The sarcolemma surrounds the muscle fiber (4, 5, 72) and consists of three distinct layers, collagen, glycocalyx, or basal lamina and plasma membrane (70). The sarcolemma preparation was prepared from muscle by homogenization and subsequent extraction with 0.4 M LiBr and later with 1.0 M KC1 which solubilizes the contractile proteins and releases internal organelles (4). Another sarcolemma preparation utilized homogenization in Ca2+-containing solutions and incubation at 37 OC, at which temperature the contractile proteins become more soluble and endogenous proteolysis likely takes place. Plasma membranes can be released from sarcolemma preparations by additional shear, and purification can be achieved with the use of sucrose density gradient centrifugation (8,9,11,73). Yet another approach to plasma membrane isolation utilized direct and vigorous homogenization of the muscle tissue to obtain microsomes. The microsomes were then fractionated on a sucrose density gradient (15, 16).
The isolation of highly purified plasma membrane vesicles from rabbit skeletal muscle described here makes use of a limited blending in high salt, followed by homogenization in buffered sucrose. Two gradients are used sequentially for purification, including a second dextran gradient which removes leaky vesicles. The preparation appears to consist mainly of surface membrane rather than transverse tubule. The surface membrane fraction is largely sealed and oriented inside out. The vesicles have been shown to effectively generate a sodium ion gradient, the magnitude of which is comparable with that in the muscle fiber.