Glutamate Transport into Synaptic Vesicles

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 267, No. 22, Issue of August 5, PP ,1992 Printed in U. S. A. Glutamate Transport into Synaptic Vesicles ROLES OF MEMBRANE POTENTIAL, ph GRADIENT, AND INTRAVESICULAR ph* (Received for publication, March 11, 1992) Joel S. TabbSS, Phillip E. KishS, Rebecca Van Dykell, and Tetsufumi UedaSII ** From the $Mental Health Research Institute, the TDepartment of Internal Medicine, and the IIDepartments of Pharmacology and Psychiatry, The University of Michigan, Ann Arbor, Michigan Glutamate, the major excitatory neurotransmitter in icin. We conclude from these experiments that optimal the mammalian central nervous system, is transported ATP-dependent glutamate uptake requires a large A* into bovine synaptic vesicles in a manner that is ATP and a small ApH. Chloride stimulates glutamate uptake dependent and requires a vesicular electrochemical by increasing the vesicle ApH, and we suggest that the proton gradient. We studied the electrical and chemical resulting decrease in intravesicular ph at low C1- conelements of this driving force and evaluated the effects centrations plays an important role in stimulating veof chloride on transport. Increasing concentrations of sicular glutamate uptake. C1- were found to increase the steady-state ATP-dependent vesicular ph gradient (ApH) and were found to concomitantly decrease the vesicular membrane potential (A*). Low millimolar chloride concentrations, Synaptic vesicles play important roles in the storage of which cause 3-6-fold stimulation of vesicular glutaneurotransmitters in the nerve terminal and their release into mate uptake, caused small but measurable increases in the synaptic cleft (Kelly, 1988; Nicholls and Attwell, 1990). ApH and decreases in A*, when compared to control vesicles in the absence of chloride. Nigericin in potas- Specific transport systems have been identified in synaptic sium buffers was used to alter the relative proportions vesicle membranes for the major amino acid neurotransmitof ApH and A*. Compared to controls, at all chloride ters: glutamate (Naito and Ueda, 1983, 1985; Maycox et al., concentrations tested, nigericin virtually abolished 1988; Sidon and Sihra, 1989), y-aminobutyric acid (Hell et ApH and increased the vesicle interior positive A*. al., 1988; Fyske and Fonnum, 1988; Kish and Ueda, 1989), Concomitantly, nigericin increased ATP-dependent and glycine (Kish et al., 1989; Christensen et al., 1990), as glutamate uptake in 0-1 mm chloride but decreased well as acetylcholine (Anderson et al., 1982) and catecholglutamate uptake in 4 mm (45%), 20 mm (80%), and amines (Toll and Howard, 1978; Floor et al., 1990). While 140 mm (75%) C1- (where ApH in the absence of niger- these transport systems have been described in detail as to icin was large). These findings suggest that either A*, their kinetic characteristics and substrate specificity, none of ApH, or a combination can drive glutamate uptake, but these transport proteins have been isolated and little is known to different degrees. In the presence of 4 mm C1-, where uptake is optimal, both A* and ApH contribute to the driving force for uptake. When the extravesicular ph was increased from 7.4 to 8.0, more C1- was required to stimulate vesicular glutamate uptake. In the absence of C1-, as extravesicular ph was lowered to 6.8, uptake was over %fold greater than it was at ph 7.4. As extravesicular ph was reduced from 8.0 toward 6.8, less C1- was required for maximal stimulation. Decreasing the extravesicu- lar ph from 8.0 to 6.8 in the absence of C1- significantly increased glutamate uptake activity, even though pro- Agents that inhibit vacuolar-type ATPases (V-type ATPases) ton-pumping ATPase activity actually decreased about also inhibit glutamate uptake, indicating that a proton-pump- 45% under identical conditions. In the absence of chlo- ing ATPase is responsible for establishing a A~H+ that enerride, nigericin increased glutamate uptake at all the ph values tested except ph 8.0. Glutamate uptake at ph 6.8 in the presence of nigericin was over 6-fold greater than uptake at ph 7.4 in the absence of niger- * This study was supported by National Institutes of Health Grant NS and Alcohol, Drug Abuse and Mental Health Administration Post-doctoral Fellowship MH (to J. S. T.). 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: Dept. of Physiology, Dartmouth Medical School, Hanover, NH ** To whom correspondence and reprint requests should be addressed: Mental Health Research Institute, The University of Michigan, 205 Zina Pitcher Place, Ann Arbor, MI about their mechanisms of action. The uptake of glutamate, the major excitatory neurotransmitter in the mammalian brain, into synaptic vesicles is temperature- and ATP-dependent (Naito and Ueda, 1983), quite specific for L-glutamate, has a relatively high K, (1-2 mm) and is stimulated 3-6-fold by low millimolar concentrations of C1- (Naito and Ueda, 1985). The uptake of glutamate into synaptic vesicles is inhibited by proton ionophores, suggesting that glutamate uptake is driven by an electrochemical proton gradient (Ap,+) across the vesicular membrane. gizes the glutamate transport system (Naito and Ueda, 1985; Maycox et al., 1988; Cidon and Sihra, 1989; Shioi et al., 1989). This has been further demonstrated by the fact that glutamate can accumulate in vesicles in the absence of ATP when A* is artificially supplied, either by using valinomycin and high external concentrations of K or Rb (Shioi et al., 1990), or by reconstituting the vesicular glutamate transporter with bacteriorhodopsin and using light to drive protons into pro- teoliposomes (Maycox et al., 1990). While it is generally accepted that A ~H+ is the actual driving The abbreviations used are: A/.LH+, electrochemical proton gradient; FCCP, carbonyl cyanide p-(trifluoromethoxy)-phenylhydrazone; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; NEM, N-ethylmaleimide; A*, membrane potential; V-type ATPase, vacuolar-type ATPase.

2 Vesicles Synaptic into Transport Glutamate force for glutamate uptake, it has not been fully established which of the two components of ApH+, the membrane potential (A*) or the ph gradient (ApH), or both, are involved in driving glutamate uptake. It is also unclear how C1- stimulates glutamate accumulation into vesicles. In the chromaffin granule, a well-characterized storage vesicle that is isolated from adrenal chromaffin cells, both A9 and ApH can drive catecholamine accumulation (Holz, 1978; Schuldiner et al., 1978; Johnson and Scarpa, 1979), and a combination of both provides optimal uptake conditions. High concentrations of C1- (about 75 mm) stimulate catecholamine uptake by increasing the ApH portion of the ApH+ that is required for optimal uptake (Johnson et al., 1979). In synaptic vesicles the situation is less clear. In contrast with chromaffin granules, 4 mm, and not 75 mm C1-, provides maximal stimulation of glutamate uptake (Naito and Ueda, 1985). Maycox et al. (1988) demonstrated that at 4 mmc1- synaptic vesicles acidify only slightly. They also showed that at higher C1- concentrations ApH increased significantly, but glutamate uptake decreased. Maycox et al. concluded that glutamate uptake was dependent on the A* and that ApH did not appear to play an important role in vesicular glutamate accumulation. Several hypotheses exist to explain the mechanism of how C1- might stimulate the ATP-dependent accumulation of glutamate within synaptic vesicles. One hypothesis suggests that C1- uptake into synaptic vesicles might be directly coupled to glutamate uptake. Another hypothesis suggests that C1- might bind directly to a regulatory site on the glutamate transport protein. A third theory suggests that C1- might enter the vesicle and act to establish a ApH, which consequently stimulates glutamate uptake into the vesicle. The focus of this report is to attempt to understand mechanism the by which C1- stimulates glutamate transport in synaptic vesicles. In this study, we carried out detailed analyses of A\k and ApH formation in synaptic vesicles, in the presence of various C1- concentrations. By manipulating the intravesicular ph in the absence and the presence of nigericin, we were able to demonstrate that 4 mm C1- caused significant and measurable changes in both A\k and ApH and that, depending upon medium conditions, either A* or ApH could drive glutamate uptake. We also show that changes in the intravesicular ph, and not just the ApH, were responsible for the stimulation of glutamate uptake that occurred at low C1- concentrations. Portions of this report have been previously published in abstract form (Tabb and Ueda, 1991). EXPERIMENTAL PROCEDURES Materials-All chemicals were purchased from Sigma, with the exception of oxonol V and acridine orange, which were purchased from Molecular Probes, Inc. (Eugene, OR), and [14C]methylamine and ~-[2,3-~H]glutamate, which were purchased from Amersham Corp. Carbonyl cyanide p-(trifluoromethoxy)-phenylhydrazone (FCCP) was stored as a 10 mm solution in either ethanol or dimethyl sulfoxide, and nigericin was stored as 0.1 mm stock solution in ethanol. Bovine Synaptic Vesicle Preparation-Bovine synaptic vesicles were prepared in bulk using a modified version of the procedure described by Kish and Ueda (1989). Typically, six bovine brains were obtained from a local slaughterhouse. On ice, the meninges were removed from each brain, the cerebral cortex dissected, and excess white matter removed (yield, approximately 1.2 kg of cortex). About 300 g of cortex were briefly blended in 600 ml of 0.32 M sucrose, 1 mm NaHC03, 1 mm magnesium acetate, 0.5 mm calcium acetate, 0.2 mm phenylmethylsulfonyl fluoride (solution A) in a Waring blender with 3 X 7 s bursts; solution A was used in all steps unless noted otherwise. The homogenate was diluted to 2 liters in solution A and was rehomogenized in a 300-ml tight-fitting Teflon-glass homogenizer (Kontes, Vineland, NJ) with two strokes, at 1,900 rpm. Each batch was diluted to 3 liters and was centrifuged in a Sorvall GSA rotor at 2,500 rpm (1,000 gmaj for 10 min. The supernatant (Sl) was saved, and the pellets (PI) from various batches were pooled, diluted to 4 liters, rehomogenized, and recentrifuged at 2,500 rpm (1,000 gmax) for 10 min. This supernatant (Sl) was pooled with the previous S1, and was centrifuged in a GSA rotor at 13,000 rpm (27,300 gmaj for 15 min. This pellet (P2) was saved, and the supernatant was discarded. The P2 was resuspended in solution A (final volume, 1.6 liters). This suspension was diluted with an equal volume of 1.28 M sucrose, to make the final concentration of sucrose 0.8 M, and was then centrifuged at 13,000 rpm (27,300 g,j in a GSA rotor for 45 min. The floating myelin bands and supernatant were aspirated, and the pellets (synaptosomes) were saved. The synaptosomes were resuspended in 2 liters of ice-cold lysing buffer (6 mm Tris-maleate, ph 8.1), diluted to 8 liters with lysing buffer, mechanically stirred at 4 "C for 45 min, and were then centrifuged at 19,000 rpm (43,500 g,) in a Sorvall SS-34 rotor for 15 min. The supernatant was then concentrated from 8 liters to 800 ml in an Amicon spiral ultraconcentrator, equipped with an Sly30 cartridge (30,000 molecular weight cut-off). The retentate was then centrifuged at 43,000 rpm (200,000 g,,,) in a Beckman 45Ti rotor ultracentrifuge rotor for 70 min. The pellets (crude synaptic vesicles) were resuspended in 20 ml of lysing buffer and were layered over six discontinuous sucrose gradients (12 m10.4 M, 6 m10.6 M, 6 mlo.8 M), and were then centrifuged at 28,000 rpm (141,000 g,,,) in a Beckman SW28 rotor for 2 h. The lysing buffer and 0.4 M sucrose layers (but not the 0.4 M, 0.6 M sucrose interface; it contains plasma membrane contaminants) were removed, diluted with lysing buffer, and were centrifuged at 47,000 rpm (200,000 gmax) in a Beckman Ti50 rotor for 60 min. The pellets were saved and were either resuspended in solution B (0.32 M sucrose, 1 mm NaHC03. 1 mm dithiothreitol) at about 5 mg/ml and stored in liquid Nz or were stored as pellets at -80 "C. Vesicles stored under either condition were stable for at least 1 month. Assay for [3H]Glutamate and ['4C]Methylamine Accumulation into Synaptic Vesicles-The uptake of [3H]glutamate or ['4C]methylamine into synaptic vesicles was determined using a slight modification of the filtration procedure described by Kish and Ueda (1989). Standard uptake assay conditions used potassium gluconate instead of sucrose to maintain iso-osmotic conditions and contained 0.14 M potassium gluconate, 20 mm HEPES (ph 7.4), 4 mm MgSO,,50-100pgof vesicle protein, and 0 or 2 mm Tris-ATP (from a 100 mm stock, ph 7.2). Vesicles were incubated in potassium gluconate buffer at 4 "C for at least 1 h before uptake was measured. Uptake was initiated by the addition of a mixture (20 +I) of unlabeled and [3H]glutamate (or [14C]methylamine) and ATP; the final volume was 100 +l. In experiments where the external ph was manipulated, care was taken to ensure that all buffer components and the resulting final solution were at the desired ph. In all uptake experiments, the final concentration of either ['Hlglutamate or ['*C]methylamine was 50 +M. ATPdependent uptake activities of ["Hlglutamate and [14C]methylamine were calculated as the amount of glutamate or methylamine taken up in the presence of ATP minus the amount taken up in the absence of ATP and were expressed as picomoles/milligram protein. All uptake experiments were carried out at 30 "C. Measurements of A* and ApH by Fluorescence Quenching-The A* across the synaptic vesicle membrane was monitored by measuring ATP-dependent quenching of the fluorescent dye oxonol V (Scherman and Henry, 1980; Van Dyke, 1988), while the ApH was monitored by measuring ATP-dependent quenching of acridine orange (Maycox et al., 1988 Cidon and Sihra, 1989). Synaptic vesicles ( pg) and standard assay medium containing varying concentrations of potassium gluconate and KC1 (total volume, 2 ml) were prewarmed to 30 "C in a 3-ml disposable cuvette with a magnetic stirring bar. Oxonol V was used at a final concentration of 1.3 +M, and acridine orange was used at 2.5 p ~ Fluorescence. changes were measured and analyzed using a SPEX DM3000CM spectrofluorometer equipped with a temperature-controlled cuvette holder and a magnetic stirrer. Excitation and emission wavelengths for oxonol V were 617 and 643 nm, respectively; those for acridine orange were 492 and 537 nm, respectively. All fluorescence experiments were carried out at 30 "C. After a stable base line was obtained, recordings were started and, within about 60 s, Tris-ATP was added to a final concentration of 2 mm (20 p1 of 200 mm Tris-ATP stock solution). Additions of FCCP or nigericin are labeled on the respective figures. FCCP was added to a final concentration of 10 p~ (2 pl of 10 mm ethanolic FCCP stock solution), and nigericin was added to a final concentration of 0.1 pm (2 p1 of 0.1 mm ethanolic nigericin stock

3 15414 Glutamate Transport into Synaptic Vesicles solution); ethanol itself had no effect on either oxonol V or acridine orange fluorescence. Studies were repeated in triplicate. Fluorescence changes were quantified using % fractional quench = [F(I,- F(ATP~]/F~ X 100, where F, is the initial fluorescence intensity, and F(ATP) and F(I) are the fluorescence intensities after the addition of Tris-ATP and dissipating ionophore, respectively. ATPase Actiuity-V-type ATPase activity was measured by monitoring the N-ethylmaleimide (NEM)-sensitive hydrolysis of ATP using a coupled enzymatic system similar to that described by Norby (1988). Synaptic vesicles (100 pg) were added to 1 ml of buffer (prewarmed to 30 c) containing 20 mm HEPES, 140 mm potassium gluconate, 7 mm MgS04,l mm phosphoenolpyruvate, 0.4 mm NADH, 10 units of pyruvate kinase, and 30 units of lactate dehydrogenase. When used, NEM was added to a concentration of 0.5 mm. ATP hydrolysis was initiated by the addition of 5 mm Tris-ATP (final concentration), and the reaction was monitored by measuring the reduction in absorbance at 340 nm. The ph values of all solutions were adjusted to the desired ph, and the final buffer was also checked to ensure that the ph was correct. RESULTS AND DISCUSSION Effect of Cl- on A\k and ApH Formation in Synaptic Vesicles-Chloride concentrations of 1-10 mm stimulate the ATPdependent uptake of glutamate into isolated bovine synaptic vesicles 3-6-fold, and higher concentrations of C1- have been found to attenuate this stimulation (Naito and Ueda, 1985). Maycox et al. (1988) examined the effects of 4 and 150 mm C1- on A\k and ApH formation in synaptic vesicles and found that changes in glutamate uptake at these concentrations of C1- did not correlate with the vesicular changes in A\k and ApH. Maycox and co-workers also reported that uptake increased greatly at 4 mmc1- but that ApH increased only slightly. At 150 mmc1-, uptake was less than it was in the absence of C1-, but the changes in ApH and A\k were large; ApH was increased greatly, and A\k was significantly reduced. Cidon and Sihra (1989) obtained similar results using 10 mm KC1 as lowc1- and 100 mm KC1 as high C1-. Both groups concluded similarly that ApH did not play a role in glutamate uptake into synaptic vesicles, since there was no correlation between the effect of C1- on ApH and its effect on glutamate uptake and, therefore, that glutamate uptake into synaptic vesicles was driven solely by A*. We suggest that although C1- has different effects on vesicular ApH formation and ATP-dependent glutamate uptake, ApH changes may still play an important role in regulating glutamate transport. To determine how ApH may affect glutamate uptake into synaptic vesicles, we examined the effect of various concentrations ofc1- on A\k and ApH formation in synaptic vesicles by measuring the quenching of the fluorescent dyes oxonol V (for A*) and acridine orange (for ApH). Representative fluorescence measurements of oxonol V and acridine orange quenching are shown in Fig. 1, and quantitation of these measurements is presented in Fig. 2. When Tris-ATP (2 mm) was added to the synaptic vesicles, A\k rapidly formed and, consequently, the fluorescence of the oxonol V was rapidly quenched (Fig. 1A). Upon the addition of 0.1 PM nigericin, a K+/H+ exchanging ionophore, the fluorescence was quenched further, indicating an additional increase in A\k, as would be expected if nigericin essentially substitutes K+ for intravesicular H+ while the proton pump is active. The addition of FCCP, an electrogenic proton ionophore, dissipated A\k as expected, returning the fluorescence to base line. As the C1- concentration in the assay mixture was increased from 0 to 140 mm (replacing the impermeant anion gluconate), the steady-state level of A\k was steadily reduced (Fig. 1A). Oxonol V was quenched so rapidly that the initial rate of A\k A 1 + ATP + Nlgerlcln + FCCP I Tlme (sec) + Nlgerlcln ATP 4 mm CI 140 mm CI * Time (sec) FIG. 1. Effect of C1- on membrane potential and ph gradient formation in bovine synaptic vesicles. Membrane potential (A@) formation was monitored by measuring oxonol V fluorescence quenching, and ph gradient (ApH) formation was monitored by measuring acridine orange fluorescence quenching. Changes in A@ formation (A) or ApH formation (B) were monitored in the presence of various concentrations of C1-. Final concentrations of 2 mm Tris- ATP, 0.1 p~ nigericin, and 25 PM FCCP were added from concentrated stock solutions as described under Experimental Procedures at the times shown by the urrows J ph Gradlent Potentlal [Chloride] (mm) FIG. 2. Quantitation of the effect of C1- on A* and ApH formation in synaptic vesicles. ATP-dependent A* (closed circles) and ApH (open circles) changes in synaptic vesicles were monitored in the presence of various concentrations of C1-, and the fractional fluorescence quench at steady-state was calculated. Data represent the average of three separate fluorescent scans. formation could not be accurately determined using this technique. At low c1- levels, for example 1 or 4 mm, A\k was slightly, but measurably, decreased as compared to A\k meas- ured in the absence of C1-. As the concentration of C1- was increased further, A\k further decreased to a minimum value at about 50 mm C1- (Fig. 2). At mmc1-,a*, as measured by oxonol V quenching, was reduced by about 40% when compared to the A\k observed in the absence of C1-. Changes in ApH were monitored under similar conditions. When Tris-ATP was added to vesicles, a ApH was generated,

4 Vesicles Synaptic into Transport Glutamate as assessed by concomitant acridine orange fluorescence quenching (Fig. 1B). Addition of nigericin dissipated ApH (in contrast to its effect on A*, which it increased), and, hence, returned fluorescence to its basal level. As can be seen in Fig. 1B, as the medium C1- concentration increased (assay conditions that decreased A*), both the initial rate and steadystate level of ApH increased. Measurable changes in steadystate levels of ApH were seen with C1- concentrations as low as 1 mm (Fig. 2). In contrast to A*, ApH formation continued to increase at medium Cl- concentrations above 50 mmc1- and did not plateau even at 140 mmc1- (Fig. 2). At 140 mm C1-, the magnitude of ApH, as measured by acridine orange quenching, had increased over &fold. These results show that at 4 mmc1-, the concentration that causes maximal stimulation of glutamate uptake, small, but measurable and significant changes occur in both ApH and A\k formation in the synaptic vesicle. Simultaneous Measurement of ATP-dependent ['HIGlutamate and ['4C]Methylamine Accumulation in Synaptic Vesicles-The accumulation of trace concentrations of methylamine has been used extensively to demonstrate acidification of subcellular compartments (Johnson and Scarpa, 1976; Loh et al., 1984; Van Dyke, 1988). To further demonstrate that measurable ph gradients form in synaptic vesicles under conditions that cause maximal stimulation of glutamate uptake, we simultaneously measured the ATP-dependent uptake of [3H]glutamate and ['*C]methylamine in a double-labeled experiment (Fig. 3). Maximal glutamate uptake occurred at 4 mm Cl-, with a 3.6-fold stimulation of uptake above control in the absence of C1-. These results are quite similar to those published previously (Naito and Ueda, 1985). At 4 mmc1-, methylamine accumulation was increased about 1.6-fold above control. As C1- concentrations were increased above 4 mm, glutamate uptake decreased steadily, while methylamine accumulation increased. At 140 mmc1-, glutamate uptake was about 50% of the uptake measured in the absence of C1-, while methylamine accumulation was about 3.5-fold greater than it was in the absence of C1-. Effect of Nigericin on A* and ApH Formation-Since nigericin catalyzes an exchange of protons for K', it has been used extensively to equilibrate the intravesicular ph with the extravesicular ph and hence to dissipate the ApH in various subcellular organelles (Russell, 1984; Cidon and Sihra, 1989). Fig. 4A shows that 0.1 pm nigericin virtually eliminated ATP [Chloride] (mm) FIG. 3. ATP-dependent uptake of [3H]glutamate and ['"C] methylamine in synaptic vesicles. Synaptic vesicles were incubated with both 50 pm [3H]glutamate(open circles) and 50 ~ L M ["C] methylamine (closed circles), and the uptake of both labels was measured for 5 min. ATP-dependent uptake was calculated as the uptake in the presence of ATP minus the uptake in the absence of ATP. All data represent the average of triplicate determinations, and the error is the standard deviation [Chlorlde] (mh) / H -Nlgericln +Nigerkln SO 140 [Chloride] mm FIG. 4. Effect of C1- and nigericin on the ATP-dependent ApH and A\t formation in synaptic vesicles. A, ATP-dependent methylamine accumulation was measured in synaptic vesicles incubated with various C1- concentrations in the absence (solid bars) or presence (hatched bars) of 0.1 pm nigericin for 5 min. Data represent the average of triplicate determinations, and the error is the standard deviation. B, ATP-dependent oxonol V quenching was measured in synaptic vesicles incubated with various concentrations of C1- in the absence (solid bars) or presence (hatched bars) of 0.1 p~ nigericin. Data represent the average of three separate determinations. dependent methylamine accumulation (and therefore ApH formation) in synaptic vesicles, at all C1- concentrations tested. Under similar conditions, nigericin was found to increase ATP-dependent A\k formation (as measured by oxonol V fluorescence quenching) at all C1- concentrations tested (Fig. 4B). These results demonstrate that, in synaptic vesicles, low nigericin concentrations in the presence of potassium dissipated the vesicular ApH and concomitantly increased A*. Since nigericin acts to equilibrate intravesicular and extravesicular protons, and since methylamine accumulates into compartments in a manner that is proportional to the proton gradient, it would be expected that, in the presence of nigericin, the internal and external concentrations of methylamine would be the same. Therefore, if we use the amount of methylamine accumulated within vesicles in the absence of nigericin as a measure of the intravesicular methylamine concentration, and use that in the presence of nigericin as a measure of extravesicular methylamine concentration, we can calculate, from the data in Fig.4A, the ratios of internal methylamine concentrations uersus external methylamine concentrations. This will allow us to determine the ApHs in vesicles in the presence of ATP and various concentrations of c1-. Table I shows that, in the absence of ATP, and therefore in the absence of proton pumping, some methylamine accumulated within the vesicle; these data indicate that there exists a ApH of about 0.4 ph units. In the absence of ATP, increasing the c1- concentration to 140 mm increased the ApH to about 0.5 ph units. In the absence of C1-, the addition of ATP increased ApH from 0.4 to about 0.6 units. The further

5 15416 Synaptic into Glutamate Transport TABLE I Effect of Cl- on methylamine accumulation ratios and ApH in synaptic vesicles. The ratios of the intravesicular methylamine concentrations versus the extravesicular methylamine concentrations and the ApH values were calculated from the data presented in Fig [MA];, intravesicular methylamine concentration; [MA],,, extravesicular methylamine concentration. The ratio [MA]J[MA], was calculated from the methylamine accumulation which was measured in the absence and presence, respectively, of 0.1 PM nigericin. ApH = log([ma],/[ma],). [MAlz/[MAlo APH -ATP -ATP +ATP +ATP ATP-dependent mm addition of C1- increased methylamine accumulation, and at 140 mmc1- the vesicles maintained a ApH of about 1.1 ph units, of which about 0.6 ph units were ATP-dependent. These results demonstrate that these synaptic vesicles have a pre-existing ApH of about 0.4 ph units that is present even in the absence of ATP. Whether thi signifies that the vesicles were not fully equilibrated with the buffer prior to measuring the ApH or that an endogenous ApH exists within these synaptic vesicles is not known. These findings also demonstrate that a C1- gradient, in the absence of ATP, caused synaptic vesicles to acidify slightly, probably due to electrical coupling. When ATP was added, a larger ApH formed and the subsequent addition ofc1- further stimulated vesicle acidification. While the calculations presented here may not reflect the actual vesicular ApH found in vivo, due to the nonphysiological nature of the assay conditions, the trends in ApH changes that occur with manipulations in the external Cl- concentration should reflect the trends that would occur in vivo. Effect of Nigericin on Glutamate Uptake into Synaptic Ves- icles-naito and Ueda (1985) and Fyske et al. (1989) have previously shown that nigericin, in the presence of potassium, inhibited glutamate uptake at 4 mmc1-. To determine what effect reduced ApH would have on how C1- stimulates glutamate uptake, we examined the effect of nigericin on uptake at several C1- concentrations (Fig. 5). We found that the effects of nigericin on uptake varied significantly with the amount ofc1- present. In the absence ofc1-, rather than inhibiting glutamate uptake, we found that nigericin increased ATP-dependent glutamate uptake approximately 45% when compared to control uptake. Glutamate uptake in the presence of 1 mmc1- was also stimulated by nigericin (over 10%). At higher concentrations ofc1-, the effect was reversed, and uptake was reduced by the presence of nigericin. Glutamate uptake was reduced approximately 45% at 4 mm C1-, 80% at 20 mm Cl-, 90% at 50 mm Cl-, and about 75% at 140 mm Cl-. These results indicate that ApH is clearly important in driving vesicular glutamate uptake and that, when AP predominates, nigericin increases glutamate uptake, whereas, when ApH predominates, nigericin decreases glutamate uptake. This experiment suggests that, under certain conditions, either of the two components of A~H+ provides a dominant driving force for glutamate uptake. At 0 or 1 mm C1-, uptake was stimulated by the addition of nigericin, a condition that eliminated ApH and enhanced AP. At these low C1- concen- trations, AP must be the major driving force for uptake. This is in contrast to the situation that occurred at 50 or 140 mm Vesicles 1WO [Chloride] (mm) FIG. 5. Effect of C1- and nigericin on the ATP-dependent glutamate uptake in synaptic vesicles. ATP-dependent glutamate uptake was measured in synaptic vesicles incubated with various Cl- concentrations in the absence (solid bars) or presence (hatched bars) of 0.1 PM nigericin. Uptake was measured for 10 min. Data represent the average of triplicate determinations, and the error is the standard deviation. 600, [Chiorlde] (mm) FIG. 6. Effect of extravesicular ph and C1- on glutamate uptake in synaptic vesicles. Synaptic vesicles were incubated at ph 6.3 (triangles), 6.8 (squares), 7.4 (circles), and 8.0 (diamonds), in the presence of mm C1-. The ATP-dependent uptake of glutamate was measured for 10 min. Control uptake was defined as the ATP-dependent glutamate uptake measured at ph 7.4 in the absence of C1-. Data represent the average of triplicate determinations, and the error is the standard deviation. C1-. Under these conditions, uptake was reduced more than 75% by nigericin. Since A9 was enhanced by the addition of nigericin, but uptake decreased, ApH must be the major driving force at these high C1- concentrations. At 4 mmc1-, uptake was reduced by the presence of nigericin but by less than 50%. Therefore, at 4 mm C1-, both A* and ApH appear to be important driving forces for uptake. Effect of Varying External ph on Glutamate Uptake-We have shown that, depending upon conditions, either A9 or ApH plays a major role in glutamate uptake, and that, at 4 mmc1-, uptake is driven by both AP and ApH. If low c1- concentrations stimulate uptake by increasing the vesicular ApH, it is unclear whether the ApH or the lower intravesicular ph that would result from an increase in the ApH is responsible for the increase in uptake activity. To help determine whether ApH plays an important role in determining glutamate uptake activity, we attempted to independently alter ApH by varying buffer ph and assessed the effects on ATP-dependent glutamate uptake in the presence of various C1- concentrations as we varied the extravesicular ph from 6.3 to 8.0 (Fig. 6). For this experiment, control glutamate uptake was defined as ATP-dependent uptake, at

6 ph 7.4, in the absence ofc1-. When the external ph was increased to ph 8.0, the entire C1- stimulation curve was shifted to higher C1- concentrations, when compared to ph 7.4. At ph 8.0, mm C1- (instead of 4-10 mm, at ph 7.4) were required to maximally stimulate uptake, and attenuation of this stimulation occurred at 50 mm C1- instead of 20 mm. When the external ph was lowered to ph 6.8, the opposite trend occurred. Glutamate uptake was maximal at 1-4 mm C1-, and uptake in the absence of C1- was already more than 3-fold greater than control uptake. Uptake was attenuated by C1- in concentrations higher than 10 mm. At ph 6.3, C1- had a minimal effect on uptake. In the absence of C1-, glutamate uptake was about 1.5-fold greater than control uptake and, as the C1- concentration was increased, uptake gradually de- creased. The resultant C1- curve looks as if it had been shifted too far left, and all that remained was the attenuating effect of high C1- concentrations. One possible explanation for the finding that changes in extravesicular ph affect the degree to which C1- stimulates glutamate uptake may be that the V-type ATPase and not the glutamate transport system is being stimulated by a decrease in the external ph. To test this hypothesis, we monitored the V-type ATPase activity in synaptic vesicles as we manipulated the extravesicular ph. Since the fluorescence of oxonol V varies with ph, we decided to use the NEMsensitive ATPase activity as a measure of vesicular V-type ATPase activity (Cidon and Sihra, 1989). Fig. 7 shows that as the buffer ph decreased from 8.0 to 6.3, the V-type ATPase activity also decreased substantially. At ph 6.8, where uptake was maximally stimulated in the absence of chloride, ATPase activity was only 53% of the activity at ph 7.4 and 48% of that at ph 8.0. The lack of correlation between the effect of ph on V-type ATPase activity and its effect on C1 induced stimulation of glutamate uptake strongly suggests that the effect of ph on glutamate uptake is not mediated by increasing proton-pumping activity. Effect of Varying the Intravesiculur ph on Glutamate Uptake-As the external ph was lowered, ATP-dependent glutamate uptake was stimulated by decreasing concentrations of C1-. This suggests the possibility that intravesicular ph, and not simply ApH, might be involved in stimulating gluta- mate uptake. To further investigate this hypothesis we varied the intravesicular ph and then monitored its effect on glutamate uptake. We have shown that, using both acridine orange fluorescent quenching (Fig. 1B) and methylamine accumula- Glutamate Transport into Synaptic Vesicles tion (Fig. 4A), at ph 7.4, nigericin collapsed the vesicular ApH. To determine whether nigericin also collapsed ApH across a ph range, we measured ATP-dependent methylamine accumulation at various buffer ph values in the absence or presence of nigericin. The results presented in Fig. 8 show that, in the absence of ionophore, ApH increased as the ph was decreased from ph 8.0 to 6.8, at which point it leveled off as the external ph was further decreased to 6.3. In the presence of nigericin, there was no ATP-dependent methylamine uptake; at ph values below 8.0, there was even a net efflux of methylamine from the vesicle. This demonstrates that, at all ph values tested, nigericin effectively dissipated ApH and, therefore, equilibrated the intravesicular and extravesicular ph. To determine the effect of adjusting the internal ph on ATP-dependent glutamate uptake, we incubated vesicles at various external ph values in the absence and presence of nigericin (Fig. 9). In the absence of both C1- and nigericin, glutamate uptake increased as the external ph was decreased from 8.0, peaking at ph 6.8, and then decreased slightly as the ph was further lowered to 6.3. In the absence of C1- but in the presence of nigericin, glutamate uptake exhibited a similar trend; uptake increased as the ph was decreased from +Nigericln \ I Extravesicular ph FIG. 8. Effect of extravesicular ph and nigericin on ATPdependent [ 4C]methylamine accumulation into synaptic vesicles. Synaptic vesicles were incubated with 50 pm [14C]methylamine in the absence of C1-, and ATP-dependent methylamine accumulation was measured as described in the legend to Fig. 3. Uptake was measured in the absence (open circles) or presence (filled circles) of 0.1 pm nigericin for 5 min. I I Extravesicular ph FIG. 7. Effect of extravesicular ph on V-type ATPase activity. V-type ATPase activity was measured in the absence of C1- as described under Experimental Procedures. The external ph was varied from 6.3 to 8.0, and ATP hydrolysis was monitored for 5 min at 30 C. V-type ATPase activity represents the activity observed in the absence of NEM minus the activity obtained in the presence of 0.5 mm NEM Extravesicular ph FIG. 9. Effect of nigericin on glutamate uptake into synaptic vesicles incubated at various extravesicular ph values. Vesicles were incubated for 10 min at various ph values from 6.3 to 8.0, in the absence (open circles) or presence (filled circles) of 0.1 p~ nigericin, each in the absence of C1-. Control uptake was defined as the ATP-dependent glutamate uptake measured at ph 7.4 in the absence of nigericin. Data represent the average of triplicate determinations, and the error is the standard deviation.

7 15418 Synaptic into Transport Glutamate Vesicles 8.0 to 6.8 and then decreased as the ph approached ph 6.3. It is important to note that, at all the ph values tested except The fact that reducing the intravesicular ph to 6.8 stimulated uptake suggests that there might be amino acid residues 8.0, uptake was significantly higher in the presence of nigeri- on the glutamate transport protein that are highly sensitive cin than in its absence. At ph 6.8, with nigericin, uptake was to changes in the internal ph. It is possible that, when some approximately 6-fold greater than it was at ph 7.4, without of these residues become protonated, the transport protein nigericin. This finding demonstrates that glutamate can be becomes more active and the rate of glutamate uptake is stimulated up to 6-fold, in the absence of C1-, by reducing the intravesicular ph. There are several ways to explain the results presented above. Wt have demonstrated that changing the extravesicular ph does not stimulate the ATPase responsible for driving glutamate uptake. It is also possible that glutamate is being protonated within the vesicle at the lower intravesicular ph values and that the glutamic acid (protonated glutamate) is increased, while, when other residues become protonated, the transporter becomes less active and uptake is decreased. This could explain why glutamate uptake activity does not correlate with ApH changes at C1- concentrations in excess of 10 mm. High C1- concentrations might decrease the intravesicular ph beyond the ph where this particular transport protein is most active. In conclusion, the results presented here demonstrate that, being trapped within the vesicle. This is unlikely, since the in glutamate transporting synaptic vesicles, measurable pka of the y-carboxyl group is approximately 4.3 and, therefore, even at ph 6.3, only 1% of the glutamate would be converted to glutamic acid. Another possible interpretation is that it is the reduction in intravesicular ph and not simply changes do occur in both A* and ApH in the presence of low concentrations of C1-, and that A*, ApH, or both, can drive glutamate uptake, depending on the C1- concentration. Some of the other halides, bromide and iodide, can act in a similar the ApH that is responsible for stimulating ATP-dependent manner on glutamate uptake (data not shown), and these glutamate uptake into synaptic vesicles. We suggest that this is the most plausible explanation for the results presented above. The data presented in this report allow us to describe the synaptic vesicle glutamate uptake system in considerably more detail than was previously available. ATP is hydrolyzed by a V-type ATPase that pumps protons into the vesicle and creates a ArH+. In the absence of permeant anions, the majorhalides probably act by entering the synaptic vesicle through a halide-specific transport system. Once inside the vesicle, C1- acts to increase the magnitude of the vesicular ApH. We propose that it is the lowering of the intravesicular ph, and not necessarily only the actual ApH, that stimulates glutamate uptake into synaptic vesicles. It should be noted that r the results presented here do not rule out the possibility that C1- acts through additional mechanisms, such as direct bindity of the ApH+ is comprised of A*, positive inside, while a ing to the translocator, to stimulate glutamate uptake. Further small ApH also exists. Under these conditions, glutamate understanding of the mechanism of glutamate accumulation uptake is driven largely by A*, since dissipating the small within synaptic vesicles and the effects of C1- on glutamate existing ApH and increasing A* further actually stimulates uptake will require purification and molecular characterizauptake. tion of the major components of this active transport system. When C1- is present in the extravesicular medium, it enters the vesicle through an ATP-dependent, halide-selective trans- Acknowledgments-We are grateful to Dr. Carolyn Bovenkerk and Francis Lee for their tireless help in preparing synaptic vesicles and port system. Intravesicular C1- may neutralize the charge of to Mary Roth for her assistance in preparing this manuscript. the pumped protons, facilitating further transport of protons, leading to an increased intravesicular proton concentration. REFERENCES This results in a decrease in the intravesicular ph (hence an Anderson, D. C., King, S. C., and Parsons, S. M. (1982) Bioehemistry21, increase in ApH) with concurrent decrease in A*. At high Christensen, H, Fyske, E. M., and Fonnum, F. (1990) J. Neurochem. 64,1142- C1- concentrations (above 50 mm), ApH is large and A* is 1147 Cidon, S., and Sihra, T. S. (1989) J. Biol. Chem. 264, smaller than in the absence of C1-. Under these conditions, Floor, E., Leventhal, P. S., and Schaeffer, S. F. (1988) J. Neurochem. 55,1663- glutamate uptake is driven largely by ApH, since increasing 1670 Fyske, E. M., and Fonnum, F. (1988) J. Neurochem. 50, A* at the expense of ApH significantly inhibits uptake. At Fyske, E. M., Christensen, H., and Fonnum, F. (1989) J. Neurochem. 52,946- C1- concentrations of 4-10 mm, A9 provides the major driving 951 Hell, J. W., Maycox, P. R., Stadler, H., and Jahn, R. (1988) EMBOJ. 7,3023- force for glutamate uptake, although ApH also plays an im portant role, since increasing A* and decreasing ApH by the Hell, J. W., Maycox, P. R., and Jahn, R. (1990) J. Biol. Chem. 265, Holz, R. W. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, addition of low concentrations of nigericin (such as 0.1 PM) Johnson, R. G., and Scarpa, A. (1976) J. Biol. Chem. 251, have intermediate effects; uptake activity is reduced about Johnson, R. G., and Scarpa, A. (1979) J. Biol. Chem. 254, Johnson, R. G., Pfister, D., Carty, S. E., and Scarpa, A. (1979) J. Biol. Chem. 50% under this condition. 254, Kelly, R. B. (1988) Neuron 1, In a manner similar to that which exists for the catechol- Kish, P. E., and Ueda, T. (1989) Methods Enzymol. 174, amine carrier in the chromaffin granule (Johnson et al., 1979) Kish, P. E., Fischer-Bovenkerk, C., and Ueda, T. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, or the y-aminobutyric acid transporter in synaptic vesicles Loh, Y. P., Tam, W. W. H., and Russell, J. T. (1984) J. Bid. Chem. 259,8238- (Hell et al, 1990), the glutamate-specific translocator can 8245 Maycox, P. R., Deckwerth, T., Hell, J. W., and Jahn, R. (1988) J. Biol. Chem. harness either A*, ApH or a combination of the two, to drive 263, glutamate uptake. In chromaffin granules, catecholamine up- Mayeox, P. R., Deckwerth, T., and Jahn, R. (1990) EMBO J. 9, Naito. S.. and Ueda. T. (1983) J. Biol. Chem. 258, take is maximal in 75 mmc1- and, in y-aminobutyric acid Naito; S.; and Ueda; T. (1985) J. Neurochem. 44, vesicles, mm C1- is optimal for uptake. This is in Nicholls. D. G.. and Attwell. D. (1990) Trends Pharmncol Norby, J. G. (1988) Methods Enzymol. 156, contrast to glutamatergic vesicles, where 4 mm c1- is optimal Russell, J. T. (1984) J. Biol. Chem. 259, for uptake. In these glutamatergic vesicles, A* appears to be Scherman, D., and Henry, J. P. (1980) Biochim. Biophys. Acta 599, Schuldiner. S.. Fiskes. H.. and Kanner. B. I. (1978) Proc. Natl. Acad. Sei. U. S. a more effective driving force than ApH although, under A. 75, certain conditions, ApH can drive glutamate uptake as well. Shioi J. Naito S. and Ueda T. (1989) Biochem. J. 258, Shioi' J.: Naito' S.: and Ueda: T. (1990) Biochem. J. 267,63-68 At 4 mm C1-, the intravesicular ph is slightly acidic, and it is Toll, i., and Hbward, B. D. (1978) Biochemistry 17, this acidic internal ph that appears to create the optimal Tabb, J. S., and Ueda, T. (1991) Thmf IBRO World Congress of Neurosci. (Abstr. P43.14) conditions for glutamate uptake. Van Dyke, R. W. (1988) J. Biol. Chem. 263,