A Simple and Sensitive Procedure for Measuring Isotope Fluxes through Ion-specific Channels in Heterogenous Populations of Membrane Vesicles*

In this paper, we describe a simple and highly sen- sitive manual assay for isotope fluxes through ion-conducting pathways, particularly cation-specific channels, in heterogenous populations of small membrane vesicles. We measure uptake of tracer of the ion of interest, against a large chemical gradient of the same ion. As a result of the imposed chemical gradient, a transient electrical diffusion potential is set up across the membranes of those vesicles which are highly permeable to the ion of interest. The isotope tends to equilibrate with the diffusion potential and is therefore concentrated selectively and transiently into those ve- sicle containing the channels. Furthermore, when performed in this way, the time course of tracer equilibration occurs over several minutes, rather than the sub- second range expected for tracer equilibration into channel-containing vesicles in the absence of an opposing chemical gradient of the permeant ion. The use of the procedure is demonstrated for three Na-conducting channels: gramicidin D incorporated into phospholipid vesicles, amiloride-blockable Na channels in toad bladder microsomes, and veratridine- activated tetrodotoxin-blockable Na channels in rat brain synaptic membranes. For all three cases, it proved simple to measure a specific 22Na uptake, in a minute time range, using very low concentrations of the channel-containing vesicles. By comparison with isotope flux measurements performed without an opposing Na gradient, the power

In this paper, we describe a simple and highly sensitive manual assay for isotope fluxes through ionconducting pathways, particularly cation-specific channels, in heterogenous populations of small membrane vesicles. We measure uptake of tracer of the ion of interest, against a large chemical gradient of the same ion. As a result of the imposed chemical gradient, a transient electrical diffusion potential is set up across the membranes of those vesicles which are highly permeable to the ion of interest. The isotope tends to equilibrate with the diffusion potential and is therefore concentrated selectively and transiently into those vesicle containing the channels. Furthermore, when performed in this way, the time course of tracer equilibration occurs over several minutes, rather than the subsecond range expected for tracer equilibration into channel-containing vesicles in the absence of an opposing chemical gradient of the permeant ion.
The use of the procedure is demonstrated for three Na-conducting channels: gramicidin D incorporated into phospholipid vesicles, amiloride-blockable Na channels in toad bladder microsomes, and veratridineactivated tetrodotoxin-blockable Na channels in rat brain synaptic membranes. For all three cases, it proved simple to measure a specific 22Na uptake, in a minute time range, using very low concentrations of the channel-containing vesicles. By comparison with isotope flux measurements performed without an opposing Na gradient, the power of the present assay derives from both the very large gain in sensitivity and the convenient time course.
Functional properties of ion-conducting channels can be studied by electrophysiological techniques on native cells or phospholipid bilayers. For isolation of channel proteins and biochemical characterization, it is essential to assess the functional state in a well defined subcellular system, such as in membrane vesicles. But measurement of ion fluxes in small * This research was supported in part by National Institutes of Health Grant AM 31089-82 ("Structural Analysis of Na/K ATPase'') to I. S. Edelman. The 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.
$ vesicles is difficult. First, the typical conductance of channels is in the order of 10"o-10-'2 mho/channel (1-6). Thus, one would expect that the equilibration time of a tracer added to a suspension of vesicles will be in the order of seconds (if not less) and therefore inaccessible to manual techniques. Furthermore, the density of ion channels in cell membranes can be as low as 0.5/pm2 (3, 4), and thus on isolation of native membrane vesicles with a radius of less than 0.4 pm, one can expect a ratio of less than one channel/cell membrane vesicle. Also, membrane vesicles are usually derived from heterogeneous populations of cells and are always contamined with vesicles from internal organelles such as mitochondria, endoplasmic reticulum, etc. Thus, the volume of interest, i.e. that bounded by the membranes containing the channels of interest, can often be expected to be a relatively small fraction of the total vesicle volume. One is therefore faced with the problem of measuring an extremely fast flux into a small fraction of the vesicle population and distinguishing it from a nonspecific flux into the bulk of the vesicle volume. For purification of channel proteins and their assay, one invariably wishes to reconstitute the protein into artificial phospholipid vesicles. In a number of cases where channel proteins have been isolated, binding of a specific ligand such as tetrodotoxin (7) or saxitoxin (8,9) has been used as the criterion for preservation of function during purification, but not all channels show high affinity ligand binding. A transport assay would of course be preferable. However, the small diameter of phospholipid vesicles in the various reconstitution procedures (300-1000 A) and the requirement for an excess of lipid compared to protein should lead to a very small fraction of the vesicle population containing channels and makes the problems of measurement even more severe than in cell membrane vesicles. Assay of isotope uptake or release from reconstituted vesicles has proved to be possible only in favorable cases in which the channel protein is available in large quantities, e.g. acetylcholine receptors (10) and the tetrodotoxin receptor (11,12). As expected, the tracer equilibration time with these systems is so fast as to preclude continuous monitoring of the time course with simple manual methods.
In this paper, we describe a simple, sensitive, and convenient flux assay for selective ionic channels in membrane vesicles and demonstrate its use to measure "Na fluxes through gramicidin incorporated into phospholipid vesicles, amiloridesensitive Na channels in membranes isolated from toad bladder, and tetrodotoxin-sensitive Na channels in brain synaptic membranes. This assay was conceived partly on the basis of an interpretation offered by Glynn and Warner (13) for the transient accumulation in human red cells of 42K flowing through Ca-activated K channels: the Gardos effect.

Vesicle Preparations
Liposomes-80 mg of crude soybean lecithin were suspended in 2 ml of media containing 25 mM imidazole (pH 7), 5 mM EDTA, and 150 mM NaCl. The lipid suspension was vortexed for 5 min and then sonicated in a Branson 12 bath sonicator until it became translucent ( -30 rnin).
Toad Bladder Microsomes-Toads (Bujo marinus, Dominican Republic origin) were obtained from National Reagents (Bridgeport, CT) and kept partially submerged in tap water. The animals were double pithed and perfused through the ventricle with about 500 ml of Ringer's solution. The bladders were then excised and immersed in ice-cold homogenizing buffer containing 55 mM NaCl, 87.5 mM sucrose, 12.5 mM imidazole (pH 71, 2.5 mM EDTA, and 0.1 mM amiloride. All the subsequent operations were carried out at 0 "C. The preparation of microsomes is based on a procedure developed by Palmer and Edelman.' The epithelial cells were scraped from the supporting tissue with a glass microscope slide and washed twice in the above homogenization buffer by centrifugation a t 800 X g for 10 min. Washed cells obtained from four to six hemibladders were suspended in 2 ml of homogenization buffer and disrupted by a single 5-s burst with a Polytron tissue grinder (Brinkmann Instruments) at setting 6. The homogenate was centrifuged at 800 X g for 10 min, and the pellet (mainly nuclei and unbroken cells) was discarded. The supernatant was centrifuged at 8,000 X g for 5 min, yielding a cloudy supernatant, a loose white-yellow pellet, and a hard brown pellet. The loose layer was separated from the tight pellet by gentle shaking and was combined with the supernatant. The combined fraction was centrifuged for 60 min at 27,000 X g, yielding a clear supernatant and pellet. This pellet, "the microsomal fraction," was suspended in the homogenizing buffer to a final concentration of 0.5-1.5 mg of protein/ ml and used for transport measurements within 4 h.
Rat Brain Synaptic Membrarzes-Brain membranes were obtained from synaptosomal lysates. Synaptosomes were prepared from rat brain homogenates by a procedure similar to that described by Krueger et al. (14). All operations were carried out at 0-4 "C. The synaptosome pellet was homogenized with four strokes at 600 rpm in a glass Teflon homogenizer (0.15-mm clearance) in 5 mM Tris-HC1 (pH 8.0) (about 40 mlhrain). Solid ultrapure sucrose was then added to the suspension to bring it to 15% w/v. Twenty ml of this suspension were layered on a 3-ml cushion of 32.5% sucrose and centrifuged at 120,000 X g for 90 min in an SW 28 Beckman rotor. The interface between the 15% and the 32.5% sucrose was collected, diluted in 2 volumes of the Na buffer described in the legend to Fig. 4, and centrifuged at 100,000 X g for 30 min. The pellets contain the membranes used in the experiments. All solutions also contained 0.1 mM phenylmethanesulfonyl fluoride, 1 mM iodoacetamide, and 1 p~ pepstatin A. For flux measurements, the membranes were sonicated for 20 s in a bath sonicator prior to being used.

Transport Assay
Two different procedures were used to remove Na ions from the external media of vesicle suspensions prior to the transport assay. For experiments with liposomes, a volume of 200 pl was centrifuged through a Sephadex G-50-40 column pre-equilibrated with 150 mM Tris-C1, 25 mM imidazole (pH 7), and 5 mM EDTA, as described previously (15). This step exchanged the external NaCl by an equal amount of Tris-C1 without changing the total volume. The eluted vesicle suspension was mixed with 800 p1 of the above solution to which gramicidin had been added. The transport assay was initiated 30 s later by adding 10 pl of 22NaCI (2 pCi) to the suspension.
For experiments with toad bladder microsomes and brain synaptic membranes, the external Na was removed by passing the vesicles through a cation-exchange column (Dowex 50-X8 Tris form). Volumes of 100-200 p1 of the vesicle suspension were applied to small Dowex columns (see below) and eluted with 1-1.5 ml of 175 mM sucrose. This step exchanged the external cations by Tris and diluted the suspension 5-15-fold. Various reagents (arniloride, veratridine, and terodotoxin) were added, as required, to the vesicle suspension. The assay was initiated 30 s later by adding 10-15 pl of "NaC1 (2-3 pCi).
The vesicle suspensions were incubated with the isotope in the ' L. G. Palmer and I. S. Edelman, manuscript in preparation. different conditions and for the times indicated in the legends to the Figs. 1-4. In order to separate the vesicles from the medium, 100-pl aliquots of the vesicle suspension were applied to 5-6-cm columns of Dowex 50-X8 (Tris form) poured in Pasteur pipettes, and the vesicles were eluted into counting vials by addition of 1.5 ml of ice-cold sucrose solution (175 mM) (cf. Ref. 16). Prior to use, the columns were washed with 1-2 ml of 175 mM sucrose containing 25 mg/ml of bovine serum albumin and stored a t 0 "C. The amount of 22Na trapped within the vesicles (eluted on the Dowex columns) was estimated by scintillation counting. The 22Na content is expressed everywhere as a fraction of the initial total radioactivity in the vesicle reaction medium. Transport assays were performed at 25 or 0 "C for experiments using cell membrane vesicles or liposomes, respectively.

RESULTS AND DISCUSSION
The Principle of the Measurement-The principle of the assay is as follows. The vesicles are prepared to contain a relatively high concentration of NaCI. Shortly before the assay, the external Na is replaced by a relatively impermeant ion such as Tris. As a consequence of the Na gradient, an electrical diffusion potential will be set up, the magnitude of which will be determined by the relative permeabilities of Na, C1, and Tris through the membrane. Only in those vesicles containing the Na channels is the Na permeability likely to be much greater than the C1 and Tris permeabilities, and hence a Na diffusion potential of maximal size and interior negative will be formed. If an isotope that permeates through the channel (in our case "Na) is added to the exterior solution, it will tend to equilibrate with the membrane potential without itself significantly affecting the potential. It will therefore accumulate selectively into that fraction of the vesicle population containing the channels. In time, the Na gradient will dissipate, as will the interior negative membrane potential, and so **Na will leave the vesicles. It will be shown that by arranging the flux assay in this way, the measurement of Z' Na uptake is highly sensitive due to its accumulation, the time course of the selective "Na uptake is convenient (in the minute range), and one can distinguish permeability properties of the channels of interest from nonselective Na permeabilities.
Gramicidin Channels in Phospholipid Vesicles-The experiments in this section demonstrate the use of the transport assay for the case of a well characterized channel-forming ionophore, gramicidin (17). Fig. 1A shows the time course of '*Na uptake into soybean phospholipid vesicles prepared to contain 150 mM NaC1, in the presence and absence of gramicidin and a large Na gradient. In the presence of a Na gradient and the ionophore, a large amount of "Na was taken up, and after reaching a maximal level of about 1% of the total added radioactivity, after about 8 min, the '*Na content of the vesicles declined slowly. If we assume that the average molecular weight of the phospholipids is 1000 and there are about 3000 phospholipid molecules/sonicated vesicle (18), then the molar concentration of vesicles is about 2-3 X 10"j M. Assuming optimally that the gramicidin molecules are all incorporated into the vesicles and one requires two gramicidin molecules to produce an active channel (19), then at the concentration of ionophore used, 5 X lo-@ M, one could expect that not more than about one in every 100 vesicles contains the channel. The ratio of internal to external volume in our Assay of Isotope Fluxes through Channels in Vesicles   FIG. 1. "Na uptake through gramicidin channels in liposomes. Sonicated phospholipid vesicles (40 mg/ml) were prepared in 150 mM NaCl solution, and the external Na was substituted by Tris as described under "Methods." A, the eluted vesicles (200 pl) were diluted 5-fold in one of the following solutions: 150 mM Tris-C1, 25 mM imidazole (pH 7), 5 mM EDTA, 5 X lo-* M gramicidin (O), same solution but without gramicidin (A), same solution using a solution that contains NaCl instead of Tris-C1 (0). About 30 s after the dilution, 10 pl of 22NaC1 (2 pCi) were added to each suspension. The tracer uptake at 0 "C was measured as described under "Methods." B, the initial stages of this assay were as in part A , using a diluting solution containing 150 mM Tris-C1 and gramicidin (0). At t = 4.5 min (indicated by the arrow), the suspension was divided into two portions of 600 and 100 p l , respectively. The larger volume of suspension was mixed with 60 p1 of a solution containing 150 mM KCI, 25 mM imidazole (pH = 7), and 5 mM EDTA. 110-p1 aliquots of this suspension were removed to the Dowex columns at the times indicated (0). The smaller portion (100 p l ) served as a control. This portion was mixed with 10 pl of Tris-imidazole buffer and was applied to a Dowex column at t = 10 min (a).
conditions is roughly 0.01. The fact that at the maximum about 1% of the total radioactivity was taken up suggests that the "Na had been accumulated to roughly 100-fold excess in those vesicles that did contain the ionophore. This calculation takes into account the finding (20) that gramicidin molecules do not migrate from vesicle to vesicle. Other experiments (not shown) indicated that it was easy to detect differences in isotope uptake with gramicidin concentration only %fold higher or lower than 5 X M, a finding which supports the conclusion that the molar concentration of channels is much lower than that of the vesicles. The slow decline in internal "' Na content (Fig. 1A) was due presumably to dissipation of the Na gradient limited by the slow permeation of Tris ions into the vesicles or net loss of NaC1. A rapid loss of *'Na could be induced (Fig. 1 B ) by addition to the exterior medium of K ions, which permeate readily through gramicidin (17) and should thus bring about immediate depolarization of the membrane potential (for an analysis of the effects of the gradient on the time course, see Miniprint'). In the absence of a Na gradient or the ionophore, the uptake of isotope was very small and monotonic (Fig. 1A). This "Na uptake is essentially that which occurs into the ionophore-free lipid vesicles. These results show that, compared to the '!Na uptake without a gradient, the imposition of the Na gradient produced a t least a 50-fold increase in "Na uptake and this occurred over minutes.
Portions of this paper (including Figs. 5-7) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 83M-1146, cite the authors, and include a check or money order for $2.80 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.
It is of interest that the transient accumulation of tracer in the gramicidin-containing vesicles is reminiscent of the classical "counter transport" phenomenon observed for carrier kinetic mechanisms. Although the phenomenon of isotope accumulation against a chemical gradient has been taken as diagnostic of a "carrier" mechanism (21), the result with gramicidin, a well characterized channel-forming ionophore, shows that in the experimental conditions chosen, the distinction between carrier and "channel" kinetic mechanisms is not possible.
Amiloride-sensitiue N u Channel.-A crude preparation of membranes was prepared from epithelial cells scraped from toad bladders. The apical surface of the epithelial cells contains the amiloride-sensitive Na-specific channel (3, 22). The membrane preparation used consists of a mixture of vesicles derived from apical and basolateral membranes of the cells and is also heavily contaminated with fragmented mitochondria and probably other internal organelles. In the experiment of Fig. 2, the cells were homogenized and the membranes prepared in a medium containing 55 mM NaCl and 100 PM amiloride. Shortly before the assay, the external Na was replaced with Tris as described under "Methods." Amiloride is a weak base (pK 8.7), and it is essentially all in the protonated form at the experimental pH of 7.0. External amiloride will therefore be removed on the Dowex column. Then "Na uptake was monitored in the conditions described in the legend to Fig. 2. As seen in Fig. ZA, when unlabeled NaCl was added to the medium and the Na gradient was thus abolished, a small, slow, and monotonic uptake of "Na was observed. Inhibition by externally added amiloride was not detectable. Conversely, when only "Na was added to the medium and the transmembrane Na gradient was maximal, a very large uptake was found and it declined slowly from the maximum. Amiloride added to the external medium inhibited both the initial rate and the maximal extent of "Na uptake by about 65%. In five different preparations, the average inhibition by amiloride was 69 5 5% (mean & S.E.). Since amiloride acts from the outside of epithelial cells, it is likely that the amiloride-sensitive flux is occurring in vesicles oriented right side out with respect to the cellular orientation. The purpose of adding amiloride to the membrane preparation media was that it might, when incorporated inside, block Na Assay of Isotope Fluxes through Channels in Vesicles fluxes through channels in any membranes oriented inside out with respect to the cellular orientation. The uptake of *'Na which persists in the presence of amiloride added to the assay medium indicates the presence of other Na-or cationspecific permeation pathways, possibly in a different population of vesicles. The fact that most of the tracer is retained even 120 min after imposition of the NaCl gradient implies that the relevant vesicle space is either highly impermeable to C1 or cannot contract to a great extent (see "Analysis of "Na Flow" in Miniprint). On the other hand, abolishing the Na gradient by addition of NaCl to the medium at the peak of *'Na accumulation led to an immediate loss of **Na from the vesicles both in the absence and presence of amiloride (Fig. 2 B ) . This phenomenon is quite similar to that described for gramicidin in Fig. 1B. The sensitivity of the 22Na flux to externally added amiloride, in an experiment like that in Fig.  2, was examined in Fig. 3, which shows the initial phase of 22Na uptake at different concentrations of amiloride. The inhibitor is effective at very low concentrations and similar to those required to inhibit the Na flux in the intact bladder (22, 23). We conclude that much if not all of the amilorideblockable flux measured in the vesicles is passing through the same Na channels characterized previously in intact bladder

Recently, Chase and
Al-Awqati (25) have been able to detect an amiloride-blockable "Na efflux in an apical membrane-enriched fraction isolated from toad bladder cells. The "Na flux was measured using rapid flow equipment and occurred over a subsecond time scale. These properties of the "Na flux are expected for conditions without a Na gradient. The necessity of using the rapid flow equipment makes such measurements less convenient than the manual assay described in this paper. Labelle and Valentine (26) have reported a much slower Na flux in toad bladder microsomes. This flux was inhibited by amiloride but only at concentrations higher than 0.6 mM. The slow time course and relative insensitivity to amiloride make it doubtful that the apical Na channels were being observed.
Tetrodotoxin-sensitiue N u Channels-The third system that we have utilized to demonstrate the efficacy of the assay is the veratridine-activated, tetrodotoxin-inhibitable Na channel in rat brain synaptic membranes (Fig. 4). For this experiment, sonicated synaptosomal membrane fragments, equilibrated with 150 mM NaCl, were prepared as described under "Methods." Fig. 4 shows the time course of "Na uptake into synaptic membranes and the effects of veratridine, tetrodotoxin, and a transmembrane Na gradient.
It is expected that in the presence of veratridine these channels will be open, while tetrodotoxin should block them. The criterion of a successful measurement will therefore be inhibition by tetrodotoxin of a flux observed in the presence of veratridine. As seen in Fig. 4, when such measurements are made in synaptic membranes without a Na gradient, the isotope uptake is small and inhibition by tetrodotoxin is only just detectable. However, with a transmembrane Na gradient, "Na uptake into the veratridine-activated synaptic membranes was large and showed the by now expected biphasic kinetics. Here, tetrodotoxin inhibited the 22Na uptake by about 60-70%, to the same level as that observed in synaptic membranes which were not preincubated with veratridine.
Stimulation of the Na Flux by veratridine could be observed at concentrations as low as lo-' M. The concentration dependence of the flux followed a simple hyperbolic saturation curve with half-maximal activation at 1 X M veratridine. Tetrodotoxin inhibited the stimulation of the flux produced by 1 X lo-* M veratridine. The concentration of tetrodotoxin giving a half-maximal inhibition of the initial rate, i.e. Is,,, was about 3 x lo-' M. These affinities are similar to those observed in excitable cells (27). The 2'Na uptake in the absence of both veratridine and tetrodotoxin presumably reflects the presence in the vesicle preparation of a Na conductance which is neither activated by veratridine nor inhibited by tetrodotoxin.
shown in Fig. 4 1s far better than that obtalned in previous The resolution of the tetrodotoxin-sensitive "Na flux studies using synaptic vesicles without an opposing Na gradient (11,12,28). Measurements of 22Na uptake into phospholipid vesicles reconstituted with the partially purified Na channel protein have failed to show any stimulation by veratridine, although in two reports evidence has been produced that active Na channels were incorporated into the vesicles (29, 30). By contrast, it has been shown4 that in vesicles reconstituted with partially purified Na channel protein, veratridine produces a 400-500% increase in "Na flux when assayed in the presence of an opposing Na gradient. This flux is sensitive to tetrodotoxin and occurs in a time scale of minutes.
Significance and Potential Uses of the Assay-Comparison of the experiments described in this paper with previous work on membrane vesicles from both toad bladder and rat brain synaptic membranes emphasizes the convenient time course and tremendous gain in sensitivity of the flux measurements when performed according to the present procedure.
The success of this assay depends on the existence of a large differential membrane permeability between the ion of interest and the other ions present. It is therefore applicable to all transport systems involving net conductance of the ion of interest, but especially to the case of channel mechanisms. The method is particularly useful for work with heterogenous membrane systems because the accumulation of the isotope into vesicles containing the channels of interest greatly magnifies the flux into those vesicles and separates functionally the different classes of vesicles. Assay of channels in a heterogenous population of vesicles is the first step towards purification of channel proteins and biochemical characterization. Reconstitution of functionally active channel protein into artificial phospholipid vesicles will also make use of the advantages of the assay. The experiment of Fig. 1 clearly demonstrates that a successful reconstitution can be detected even if the channel of interest is incorporated into a small fraction of the lipid vesicles (less than l%), and the assay provides also a sensitive measure of the concentration of the channels. By assaying channel fluxes in different vesicle fractions isolated from structurally complex tissues such as muscle, it should also be possible to use the procedure to localize channels to particular membranes.