- Email: [email protected]
Glutamate in rat brain cortex synaptic vesicles: influence of the vesicle isolation procedure
Brain Research, 386 (1986) 405-408 Elsevier
405
BRE21872
Glutamate in rat brain cortex synaptic vesicles: influence of the vesicle isolation procedure NORA RIVEROS, JENNY FIEDLER, N. LAGOS, C. MUlqOZ and F. ORREGO
Department of Physiology and Biophysics, Facultyof Medicine, Universidadde Chile, Santiago (Chile) (Accepted 22 July 1986)
Key words: Glutamate - - Synaptic vesicle - - Brain cortex
Rat brain cortex synaptic vesicles have been isolated by 3 different procedures. The one of Hata et al. (J. Neurochem., 27 (1976) 139) gave synaptic vesicles with a high glutamate content, but also, as judged by [3H]ouabain binding and electron microscopy, with considerable contamination by plasma membrane vesicles. This did not allow a precise estimation of the glutamate content of each synaptic vesicle. The second procedure used (Life Sci., 21 (1977) 1075), in which the tissue is homogenized with an all glass homogenizer, yielded vesicles of higher purity, but with no glutamate. A slightly modified Kadota and Kadota procedure (J. Cell Biol., 58 (1973) 135) gave synaptic vesicles of a very high purity that were filtered on a Sepharose 4B column, and there, the synaptic vesicle fraction of highest purity was estimated to contain 3640 glutamate molecules in each glutamatergic vesicle. This is equivalent to an intravesicular concentration of 0.21 M, that is, at least 10 times higher than the glutamate concentration in the rat brain cortex.
The neuroexcitatory action of glutamate in the mammalian brain was established more than 30 years ago 21. Since then several groups have looked for the presence of glutamate in mammalian brain synaptic vesicles 3'9'11'12'15, since such a finding would strongly suggest a neurotransmitter role for it ~4. Although very small amounts of glutamate and other amino acids have been consistently found in synaptic vesicles, their relative amounts are similar to those seen in the soluble cytoplasm 3, and their subcellular distribution is similar to that of the cytoplasmic markers K or lactate dehydrogenease 12, thus suggesting that their presence in vesicles was due to contamination. The only exception to this is taurine, that is present in higher amounts 3'9. In a search for endogenous ligands for kainic acid (KA) receptors, we found that synaptic vesicles prepared by a relatively rapid procedure contained substantial amounts of glutamate, which was the only ligand present for such receptors ~6. These findings, that are important for establishing a transmitter role for glutamate, were supported by later studies of Naito and Ueda 13, who showed that highly purified synaptic vesicles contain
an ATP-dependent glutamate uptake system. In the present work we have compared the influence that 3 different synaptic vesicle isolation procedures have on their glutamate content. This has allowed us, using a highly purified synaptic vesicle fraction, to estimate the number of glutamate molecules present in a single brain cortex synaptic vesicle. In the initial experiments glutamate was measured by the [3H]kainic acid radioreceptor assay in which [3H]KA of 4 - 5 . 6 Ci/mmol was used, exactly as described 16. When synaptic vesicles prepared by the procedure of Hata et al. 6 were passed through a Sepharose 4B gel filtration column, most of the endogenous glutamate appeared in the particulate fraction, where almost all of the protein was also present (Fig. 1A), while very much less endogenous glutamate was seen in the fractions that corresponded to free amino acids, where the added internal standard of [3H]glutamate appeared. This exogenously added [3H]glutamate did not bind to the particulate fraction. The particulate fraction also contained endogenous acetylcholine and [3H]noradrenaline when the synaptic vesicles were prelabeled by incubating rat brain cor-
Correspondence: F. Orrego, Department of Physiology and Biophysics, Faculty of Medicine, Universidad de Chile, Casilla 70055, Santiago, Chile.
406
Vo 3H iOO
GLU
I
o G: Fz 0 o
50
LL 0
I E c © co eo
c~ z
'L
i
Vo
0
o
34) - -
c3
o
I
{13
-1
too
o z
I o
2
I I
I
50
I
o
o
I0
15
20
25
FRACTION
Fig. 1. Gel filtration on Sepharose 4B of Synaptic Vesicles. In A, synaptic vesicles were prepared by the method of Hata et al. 6, and in B, by that of Seidler et al. is. The columns were 13 × 1.1 cm, their flow rate was set at 0.25 ml/min, and fractions of 0.6 ml were collected. A [3H]L-glutamicacid internal standard (50,000 cpm in A, and 10,900 cpm in B) were added to the synaptic vesicles just prior to their application to the column. These amounts of glutamate are far below the sensitivity limit of the glutamate radioreceptor assay. Endogenous glutamate was detected as an inhibition of [3H]kainic acid (4-5.6 Ci/ mmol) binding to its receptors (left ordinate). For this assay 3-6 column fractions were combined, as depicted by the width of each bar in the figure, heated at 90 °C for 7 min, cooled, and centrifuged at 100,000 g for 30 min to sediment denatured proteins. The supernatants were then lyophilized. Just before the receptor assay, they were dissolved in a smaller volume of water. Absorbance at 280 nm and radioactivity of the internal standard were measured in each fraction. In A, elution was with 10 mM Tris HCI pH 7.4, 50 mM sucrose, and in B, with 130 mM potassium phosphate pH 7.4 buffer. Both A and B represent averages of two independent experiments. tex slices with [3H]L-noradrenaline for 30 min 19, and then isolating their synaptic vesicles by processing the slices together with the non-radioactive brain cortices as usually done in this procedure (not shown). This indicated that synaptic vesicles were indeed present in the particulate fraction. It was also found,
however, that this fraction had an appreciable contamination by plasma membrane fragments, as judged by [3H]ouabain binding (1310 _+ 128 fmol bound/mg protein, n = 12, when the fraction was incubated with 100 nM [3H]ouabain (New England Nuclear, 20.9 Ci/mmol) following the procedure of Hansen 5. Highly purified plasma membrane of mammalian kidney or skeletal muscle transverse tubules bind, under the same conditions, about 150 pmol of ouabain/mg protein (M. de la Fuente and C. Hidalgo, personal communication). This contamination by plasma membrane was also seen in electron micrographs of this fraction (not shown), and was already reported by Hata et al. 6. To test whether the endogenous glutamate present in the particulate fraction was truly inside synaptic vesicles or, alternatively, represented soluble glutamate that had been trapped inside resealed m e m b r a n e fragments formed during tissue homogenization or subsequently, the tissue was homogenized in the usual medium to which 2/~Ci of [3H]glutamate had been added, and synaptic vesicles obtained and then filtered on Sepharose 4B. Only 0.1% of the added radioactivity appeared in the particulate peak. In the case of endogenous glutamate 1.06% of the total glutamate present in the tissue was recovered in the vesicular fraction (Fig. 1A). This indicated that, at most, 10% of the glutamate present in particles isolated by filtration on Sepharose 4B, could be due to glutamate trapped inside resealed plasma membrane vesicles. We next studied the synaptic vesicle isolation procedure of Seidler et al.tS, in which axon terminals are not disrupted by hypo-osmotic shock, but by rigorous grinding with an all glass homogenizer. The purity of these vesicles following gel filtration on Sepharose 4B, as judged by [3H]ouabain binding (122.6 +_ 10 fmol of ouabain bound/mg protein) was higher than with the previous procedure, however, virtually no glutamate appeared in the vesicle fraction, and all of it was recovered in the free amino acid fraction (Fig. 1B). We attribute this to synaptic vesicle damage during the rigorous homogenization procedure. We finally studied synaptic vesicle isolation by a slight modification of the procedure of Kadota and Kadota 4,s, in which the synaptosomes are submitted to hypo-osmotic conditions only for a brief period, and which yields a highly purified synaptic vesicle fraction. Following gel filtration on Sepharose 4B of
407 this vesicle preparation, we found that negligible [3H]ouabain was bound to it (37.2 + 2.6 fmol/mg protein). This was also seen in an extensive electron microscopical study of the same preparation, where non-synaptic vesicle membrane profiles were quite rare (not shown). In this latter part of the study, the radioreceptor assay for glutamate was performed with [3H]kainic acid of high specific activity [3H]vinylidene, 60 Ci/mmol KA, New England Nuclear). This allowed a much more sensitive detection of glutamate. Gel filtration on Sepharose 4B of these purified synaptic vesicles (Fig. 2) showed that a considerable amount of glutamate (27.74% of the total eluted) was associated with the vesicles, and that 72.26% of the glutamate eluted in the free amino acid fraction. In addition, a third minor peak of 'glutamate-like' activity was found in fraction 22. This fraction is much retarded relative to the free amino acid fraction, and may correspond to a peptide that interacts with the Sepharose granule matrix, and because of this elutes
~ 2O
60c
aoG J2
.
~, a z
~ 8
o a:
? ~-P
0
6
re
PO
~5
20
FRACTION
Fig. 2. Glutamate in purified synaptic vesicle fractions. Synaptic vesicles were purified by the procedure of Kadota and Kadota 8, except that after 2 min of hypo-osmotic shock, the preparation was made 165 mM in KCI and 10 mM Tris citrate pH 6.5, by adding 10% volume of concentrated solutions 4. Synaptic vesicles (0.75 ml) derived from about 5.6 g of rat brain cortex were applied to a 10 x 1 cm Sepharose 4B column equilibrated with 0.25 M sucrose, 25 mM Tris citrate pH 7.1. Each fraction was 0.74 ml. Protein was measured by the procedure of Bradford 2, and endogenous glutamate by a receptor assay2° in which [3H]kainic acid of high specific activity (60 Ci/mmol) was used. This allowed measurement of glutamate in single fractions. Other conditions were as in Fig. 1. The figure represents a single experiment, from which the data for estimating the glutamate content per synaptic vesicle were obtained. Two other very similar experiments were done. The [3H]glutamic acid internal standard co-elutes with the larger endogenous glutamate peak centered on fraction 15.
after smaller molecules. This substance, which we have not yet characterized, may be enzymatically degraded during the receptor assay, releasing free glutamate which is then detected. An example of this has been described by US 17. The synaptic vesicles were extracted and their soluble contents fractionated by high voltage electrophoresis as described 16. Once again, only a single component that corresponded to glutamate was detected (not shown). We next proceeded to estimate the number of glutamate molecules per synaptic vesicle (SV) as follows: the radius of a small spherical SV is 250 A (ref. 22), its surface area: 4m-2, and the total area of the lipid bilayer: 8zrr2 = 1,571,250 A 2. The area occupied by a single phospholipid (P-lipid) molecule at the surface pressure present in mammalian membranes (ca. 33 mN'm-1), and above the membrane transition temperature is 55 A 2 (ref. 7). Therefore, each SV has 28,568 P-lipid molecules. The weighted (for relative abundance in SVs) (ref. 22) average molecular weight of SV P-lipids is 820.75, therefore, the weight in grams of 28,568 P-lipid molecules is 28,568 times 820.75 g/N (Avogadro's number = 6.023 by 1023) = 3.89 by 10-17 g. The P-lipid/protein ratio in highly purified synaptic vesicles is 1.44 (ref. 22), therefore, the protein content of each SV is 2.7 by 10-17 g. The peak SV fraction 6 in Fig. 2 contained 15.31 nmol glutamate, that represent 9.22 x 1015 molecules. It also contained 240 #g protein. If the SV purity were 95% in terms of protein, the SV protein present in the fraction would be 228 #g, that represents 8.44 × 1012 SV (228 × 10 -6 g/2.7 x 10 -17 g). As they contained 9.22 x 1015 glutamate molecules, the number of glutamate molecules/SV is 9.22 x 1015/8.44 x 1012 = 1092. This, to our knowledge, is the highest content yet reported for any transmitter present in mammalian brain synaptic vesicles. But brain cortex SVs are heterogeneous regarding their transmitters. Therefore, if one assumes that 30% of all the SVs present operate with glutamate as a transmitter, this would raise the number of glutamate molecules/SV to 3640, that is quite close to the estimated quantum size of 6000 ACh molecules in cholinergic motor nerve endings 1°. We estimate the internal volume of the small spherical vesicles present in central excitatory synapses to be 2.87 x 107 A 3. If each contained 3640 glutamate molecules, their intravesicular concentration would be 0.21 M. The glutamate con-
408 tent of rat brain cortex is 11.83 f~mol/g wet wt. 1, that
centration in synaptic vesicles, together with the
when corrected for extracellular space and intracellular water, gives an intracellular concentration of about 21.1 mM. This indicates that glutamate in syn-
presence of well defined postsynaptic glutamate receptors and physiological effects 21, provides firm
aptic vesicles is concentrated at least 10-fold relative to the whole cortex. As the latter value also includes the glutamate present in synaptic vesicles, the actual concentration ratio of vesicular to extravesicular glutamate is, possibly, much higher than 10. In the above calculations only two major assump-
proof that glutamate is a bona fide central neurotransmitter H. We are grateful to Patricio Cancino for technical assistance, to Drs. Cecilia Hidalgo and M. de la Fuente and to S. Villanueva for discussion. Sup-
glutamate occurred during the SV isolation. This lat-
ported by project D I B B 1590 of the Universidad de Chile, and 1018 of F o n d o Nacional de Ciencias. The doctoral thesis of Nora Riveros was supported by P N U D U N E S C O project CHI 81/001. F.O. is
ter is clearly u n w a r r a n t e d , and the real glutamate content is probably higher than our present estimate. In any event, the presence of this high glutamate con-
tion for a fellowship, during which part of this work was done.
tions have been made. O n e is that 30% of the SVs present carry glutamate. The other is that no loss of
1 Benjamin, A.M. and Quastel, J.H., Locations of amino acids in brain slices from the rat, Biochem., 128 (1972) 631-646. 2 Bradford, M.M., A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem., 72 (1976) 248-254. 3 De Belleroche, J.S. and Bradford, H.F., Amino acids in synaptic vesicles from mammalian cerebral cortex: a reappraisal, J. Neurochem., 21 (1973)441-451. 4 De Lorenzo, R.J. and Freedman, S.D., Calcium dependent neurotransmitter release and protein phosphorylation in synaptic vesicles, Biochem. Biophys. Res. Commun, 80 (1978) 183-192. 5 Hansen, O., Relationship between g-strophantin binding capacity and ATPase activity in plasma membrane fragments from ox brain, Biochim. Biophys. Acta, 233 (1971) 122-132. 6 Hata, F., Kuo, C.H., Matsuda, T. and Yoshida, H., Factors required for Ca-sensitive acetylcholine release from crude synaptic vesicles, J. Neurochem., 27 (1976) 139-144. 7 Jain, M.K., Non-random lateral organization in bilayers and biomembranes. In R.C. Aloia (Ed.), Concepts of Membrane Structure, Membrane Fluidity in Biology, Vol. 1. Academic Press, New York, 1983, pp. 1-37. 8 Kadota, K. and Kadota, T., Isolation of coated vesicles, plain synaptic vesicles, and flocculent material from a crude synaptosome fraction of guina pig whole brain, J. Cell Biol., 58 (1973) 135-151. 9 Kontro, P., Marnela, K.-M. and Oja, S.S., Free amino acids in the synaptosome and synaptic vesicle fractions of different bovine brain areas, Brain Research, 184 (1980) 129-141. 10 Kuffler, S.W. and Yoshikami, D., The number of transmitter molecules in a quantum: an estimate from iontophoretic application of acetylcholine at the neuromuscular synapse, J. Physiol. (London), 251 (1975) 465-482. 11 L~ihdesm~iki, P., Karppinen, A., Saarni, H. and Winter, R.. Amino acids in the synaptic vesicle fraction from calf brain: content, uptake and metabolism. Brain Research. 138 (1977) 295-308.
most grateful to the John Simon Memorial Founda-
12 Mangan, J.L. and Whittaker, V.P., The distribution of free amino acids in subcellular fractions of guinea-pig brain, Biochem. J., 98 (1966) 128-137. 13 Naito, S. and Ueda, T., Adenosine triphosphate-dependent uptake of glutamate into protein 1-associated synaptic vesicles, J. Biol. Chem., 258 (1983) 696-699. 14 Orrego, F., Criteria for the identification of central neurotransmitters and their application to studies with some nerve tissue preparations in vitro, Neuroscience, 4 (1979) 1037-1057. 15 Rassin, D.K., Amino acids as putative transmitters: failure to bind to synaptic vesicles of guinea pig cerebral cortex, J. Neurochem., 19 (1972) 139-148. 16 Riveros, N. and Orrego. F., A search in rat brain cortex synaptic vesicles for endogenous ligands for kainic acid receptors, Brain Research, 236 (1982) 492-496. 17 Riveros, N. and Orrego, F., A study of possible excitatory effects of N-acetylaspartyl glutamate in different in vivo and in vitro brain preparations. Brain Research, 299 (1984) 393-396. 18 Seidler, F.J., Kirksey, D.F., Lau, C., Whitmore, W.L. and Slotkin, T.A., Uptake of (-)(3H)-norepinephrine by storage vesicles prepared from whole rat brain: properties of the uptake system and its inhibition by drugs, Life Sci., 2l (1977) 1075-1086. 19 Vargas, O. and Orrego, F., Elevated extracellular potassium as a stimulus for releasing (3H)-norepinephrine and (~4C)a-aminoisobutyrate from neocortical slices. Specificity and calcium dependency of the process, J. Neurochem., 26 (1976) 31-34. 20 Vincent, S.R. and McGeer, E.G., Kainic acid binding to membranes of striata! neurons, Life Sci., 24 (1979) 265-270. 21 Watkins, I.C. and Evans, R.H., Excitatory amino acid transmitters, Ann. Rev. Pharmacol. Toxicol., 21 (1981) 165-204. 22 Whittaker, V.P., The synaptic vesicles. In A. Lajtha (Ed.), Handbook of Neurochemistry, 2nd edn., Vol. 7, Plenum, New York, 1984, pp. 41-69.
405
BRE21872
Glutamate in rat brain cortex synaptic vesicles: influence of the vesicle isolation procedure NORA RIVEROS, JENNY FIEDLER, N. LAGOS, C. MUlqOZ and F. ORREGO
Department of Physiology and Biophysics, Facultyof Medicine, Universidadde Chile, Santiago (Chile) (Accepted 22 July 1986)
Key words: Glutamate - - Synaptic vesicle - - Brain cortex
Rat brain cortex synaptic vesicles have been isolated by 3 different procedures. The one of Hata et al. (J. Neurochem., 27 (1976) 139) gave synaptic vesicles with a high glutamate content, but also, as judged by [3H]ouabain binding and electron microscopy, with considerable contamination by plasma membrane vesicles. This did not allow a precise estimation of the glutamate content of each synaptic vesicle. The second procedure used (Life Sci., 21 (1977) 1075), in which the tissue is homogenized with an all glass homogenizer, yielded vesicles of higher purity, but with no glutamate. A slightly modified Kadota and Kadota procedure (J. Cell Biol., 58 (1973) 135) gave synaptic vesicles of a very high purity that were filtered on a Sepharose 4B column, and there, the synaptic vesicle fraction of highest purity was estimated to contain 3640 glutamate molecules in each glutamatergic vesicle. This is equivalent to an intravesicular concentration of 0.21 M, that is, at least 10 times higher than the glutamate concentration in the rat brain cortex.
The neuroexcitatory action of glutamate in the mammalian brain was established more than 30 years ago 21. Since then several groups have looked for the presence of glutamate in mammalian brain synaptic vesicles 3'9'11'12'15, since such a finding would strongly suggest a neurotransmitter role for it ~4. Although very small amounts of glutamate and other amino acids have been consistently found in synaptic vesicles, their relative amounts are similar to those seen in the soluble cytoplasm 3, and their subcellular distribution is similar to that of the cytoplasmic markers K or lactate dehydrogenease 12, thus suggesting that their presence in vesicles was due to contamination. The only exception to this is taurine, that is present in higher amounts 3'9. In a search for endogenous ligands for kainic acid (KA) receptors, we found that synaptic vesicles prepared by a relatively rapid procedure contained substantial amounts of glutamate, which was the only ligand present for such receptors ~6. These findings, that are important for establishing a transmitter role for glutamate, were supported by later studies of Naito and Ueda 13, who showed that highly purified synaptic vesicles contain
an ATP-dependent glutamate uptake system. In the present work we have compared the influence that 3 different synaptic vesicle isolation procedures have on their glutamate content. This has allowed us, using a highly purified synaptic vesicle fraction, to estimate the number of glutamate molecules present in a single brain cortex synaptic vesicle. In the initial experiments glutamate was measured by the [3H]kainic acid radioreceptor assay in which [3H]KA of 4 - 5 . 6 Ci/mmol was used, exactly as described 16. When synaptic vesicles prepared by the procedure of Hata et al. 6 were passed through a Sepharose 4B gel filtration column, most of the endogenous glutamate appeared in the particulate fraction, where almost all of the protein was also present (Fig. 1A), while very much less endogenous glutamate was seen in the fractions that corresponded to free amino acids, where the added internal standard of [3H]glutamate appeared. This exogenously added [3H]glutamate did not bind to the particulate fraction. The particulate fraction also contained endogenous acetylcholine and [3H]noradrenaline when the synaptic vesicles were prelabeled by incubating rat brain cor-
Correspondence: F. Orrego, Department of Physiology and Biophysics, Faculty of Medicine, Universidad de Chile, Casilla 70055, Santiago, Chile.
406
Vo 3H iOO
GLU
I
o G: Fz 0 o
50
LL 0
I E c © co eo
c~ z
'L
i
Vo
0
o
34) - -
c3
o
I
{13
-1
too
o z
I o
2
I I
I
50
I
o
o
I0
15
20
25
FRACTION
Fig. 1. Gel filtration on Sepharose 4B of Synaptic Vesicles. In A, synaptic vesicles were prepared by the method of Hata et al. 6, and in B, by that of Seidler et al. is. The columns were 13 × 1.1 cm, their flow rate was set at 0.25 ml/min, and fractions of 0.6 ml were collected. A [3H]L-glutamicacid internal standard (50,000 cpm in A, and 10,900 cpm in B) were added to the synaptic vesicles just prior to their application to the column. These amounts of glutamate are far below the sensitivity limit of the glutamate radioreceptor assay. Endogenous glutamate was detected as an inhibition of [3H]kainic acid (4-5.6 Ci/ mmol) binding to its receptors (left ordinate). For this assay 3-6 column fractions were combined, as depicted by the width of each bar in the figure, heated at 90 °C for 7 min, cooled, and centrifuged at 100,000 g for 30 min to sediment denatured proteins. The supernatants were then lyophilized. Just before the receptor assay, they were dissolved in a smaller volume of water. Absorbance at 280 nm and radioactivity of the internal standard were measured in each fraction. In A, elution was with 10 mM Tris HCI pH 7.4, 50 mM sucrose, and in B, with 130 mM potassium phosphate pH 7.4 buffer. Both A and B represent averages of two independent experiments. tex slices with [3H]L-noradrenaline for 30 min 19, and then isolating their synaptic vesicles by processing the slices together with the non-radioactive brain cortices as usually done in this procedure (not shown). This indicated that synaptic vesicles were indeed present in the particulate fraction. It was also found,
however, that this fraction had an appreciable contamination by plasma membrane fragments, as judged by [3H]ouabain binding (1310 _+ 128 fmol bound/mg protein, n = 12, when the fraction was incubated with 100 nM [3H]ouabain (New England Nuclear, 20.9 Ci/mmol) following the procedure of Hansen 5. Highly purified plasma membrane of mammalian kidney or skeletal muscle transverse tubules bind, under the same conditions, about 150 pmol of ouabain/mg protein (M. de la Fuente and C. Hidalgo, personal communication). This contamination by plasma membrane was also seen in electron micrographs of this fraction (not shown), and was already reported by Hata et al. 6. To test whether the endogenous glutamate present in the particulate fraction was truly inside synaptic vesicles or, alternatively, represented soluble glutamate that had been trapped inside resealed m e m b r a n e fragments formed during tissue homogenization or subsequently, the tissue was homogenized in the usual medium to which 2/~Ci of [3H]glutamate had been added, and synaptic vesicles obtained and then filtered on Sepharose 4B. Only 0.1% of the added radioactivity appeared in the particulate peak. In the case of endogenous glutamate 1.06% of the total glutamate present in the tissue was recovered in the vesicular fraction (Fig. 1A). This indicated that, at most, 10% of the glutamate present in particles isolated by filtration on Sepharose 4B, could be due to glutamate trapped inside resealed plasma membrane vesicles. We next studied the synaptic vesicle isolation procedure of Seidler et al.tS, in which axon terminals are not disrupted by hypo-osmotic shock, but by rigorous grinding with an all glass homogenizer. The purity of these vesicles following gel filtration on Sepharose 4B, as judged by [3H]ouabain binding (122.6 +_ 10 fmol of ouabain bound/mg protein) was higher than with the previous procedure, however, virtually no glutamate appeared in the vesicle fraction, and all of it was recovered in the free amino acid fraction (Fig. 1B). We attribute this to synaptic vesicle damage during the rigorous homogenization procedure. We finally studied synaptic vesicle isolation by a slight modification of the procedure of Kadota and Kadota 4,s, in which the synaptosomes are submitted to hypo-osmotic conditions only for a brief period, and which yields a highly purified synaptic vesicle fraction. Following gel filtration on Sepharose 4B of
407 this vesicle preparation, we found that negligible [3H]ouabain was bound to it (37.2 + 2.6 fmol/mg protein). This was also seen in an extensive electron microscopical study of the same preparation, where non-synaptic vesicle membrane profiles were quite rare (not shown). In this latter part of the study, the radioreceptor assay for glutamate was performed with [3H]kainic acid of high specific activity [3H]vinylidene, 60 Ci/mmol KA, New England Nuclear). This allowed a much more sensitive detection of glutamate. Gel filtration on Sepharose 4B of these purified synaptic vesicles (Fig. 2) showed that a considerable amount of glutamate (27.74% of the total eluted) was associated with the vesicles, and that 72.26% of the glutamate eluted in the free amino acid fraction. In addition, a third minor peak of 'glutamate-like' activity was found in fraction 22. This fraction is much retarded relative to the free amino acid fraction, and may correspond to a peptide that interacts with the Sepharose granule matrix, and because of this elutes
~ 2O
60c
aoG J2
.
~, a z
~ 8
o a:
? ~-P
0
6
re
PO
~5
20
FRACTION
Fig. 2. Glutamate in purified synaptic vesicle fractions. Synaptic vesicles were purified by the procedure of Kadota and Kadota 8, except that after 2 min of hypo-osmotic shock, the preparation was made 165 mM in KCI and 10 mM Tris citrate pH 6.5, by adding 10% volume of concentrated solutions 4. Synaptic vesicles (0.75 ml) derived from about 5.6 g of rat brain cortex were applied to a 10 x 1 cm Sepharose 4B column equilibrated with 0.25 M sucrose, 25 mM Tris citrate pH 7.1. Each fraction was 0.74 ml. Protein was measured by the procedure of Bradford 2, and endogenous glutamate by a receptor assay2° in which [3H]kainic acid of high specific activity (60 Ci/mmol) was used. This allowed measurement of glutamate in single fractions. Other conditions were as in Fig. 1. The figure represents a single experiment, from which the data for estimating the glutamate content per synaptic vesicle were obtained. Two other very similar experiments were done. The [3H]glutamic acid internal standard co-elutes with the larger endogenous glutamate peak centered on fraction 15.
after smaller molecules. This substance, which we have not yet characterized, may be enzymatically degraded during the receptor assay, releasing free glutamate which is then detected. An example of this has been described by US 17. The synaptic vesicles were extracted and their soluble contents fractionated by high voltage electrophoresis as described 16. Once again, only a single component that corresponded to glutamate was detected (not shown). We next proceeded to estimate the number of glutamate molecules per synaptic vesicle (SV) as follows: the radius of a small spherical SV is 250 A (ref. 22), its surface area: 4m-2, and the total area of the lipid bilayer: 8zrr2 = 1,571,250 A 2. The area occupied by a single phospholipid (P-lipid) molecule at the surface pressure present in mammalian membranes (ca. 33 mN'm-1), and above the membrane transition temperature is 55 A 2 (ref. 7). Therefore, each SV has 28,568 P-lipid molecules. The weighted (for relative abundance in SVs) (ref. 22) average molecular weight of SV P-lipids is 820.75, therefore, the weight in grams of 28,568 P-lipid molecules is 28,568 times 820.75 g/N (Avogadro's number = 6.023 by 1023) = 3.89 by 10-17 g. The P-lipid/protein ratio in highly purified synaptic vesicles is 1.44 (ref. 22), therefore, the protein content of each SV is 2.7 by 10-17 g. The peak SV fraction 6 in Fig. 2 contained 15.31 nmol glutamate, that represent 9.22 x 1015 molecules. It also contained 240 #g protein. If the SV purity were 95% in terms of protein, the SV protein present in the fraction would be 228 #g, that represents 8.44 × 1012 SV (228 × 10 -6 g/2.7 x 10 -17 g). As they contained 9.22 x 1015 glutamate molecules, the number of glutamate molecules/SV is 9.22 x 1015/8.44 x 1012 = 1092. This, to our knowledge, is the highest content yet reported for any transmitter present in mammalian brain synaptic vesicles. But brain cortex SVs are heterogeneous regarding their transmitters. Therefore, if one assumes that 30% of all the SVs present operate with glutamate as a transmitter, this would raise the number of glutamate molecules/SV to 3640, that is quite close to the estimated quantum size of 6000 ACh molecules in cholinergic motor nerve endings 1°. We estimate the internal volume of the small spherical vesicles present in central excitatory synapses to be 2.87 x 107 A 3. If each contained 3640 glutamate molecules, their intravesicular concentration would be 0.21 M. The glutamate con-
408 tent of rat brain cortex is 11.83 f~mol/g wet wt. 1, that
centration in synaptic vesicles, together with the
when corrected for extracellular space and intracellular water, gives an intracellular concentration of about 21.1 mM. This indicates that glutamate in syn-
presence of well defined postsynaptic glutamate receptors and physiological effects 21, provides firm
aptic vesicles is concentrated at least 10-fold relative to the whole cortex. As the latter value also includes the glutamate present in synaptic vesicles, the actual concentration ratio of vesicular to extravesicular glutamate is, possibly, much higher than 10. In the above calculations only two major assump-
proof that glutamate is a bona fide central neurotransmitter H. We are grateful to Patricio Cancino for technical assistance, to Drs. Cecilia Hidalgo and M. de la Fuente and to S. Villanueva for discussion. Sup-
glutamate occurred during the SV isolation. This lat-
ported by project D I B B 1590 of the Universidad de Chile, and 1018 of F o n d o Nacional de Ciencias. The doctoral thesis of Nora Riveros was supported by P N U D U N E S C O project CHI 81/001. F.O. is
ter is clearly u n w a r r a n t e d , and the real glutamate content is probably higher than our present estimate. In any event, the presence of this high glutamate con-
tion for a fellowship, during which part of this work was done.
tions have been made. O n e is that 30% of the SVs present carry glutamate. The other is that no loss of
1 Benjamin, A.M. and Quastel, J.H., Locations of amino acids in brain slices from the rat, Biochem., 128 (1972) 631-646. 2 Bradford, M.M., A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem., 72 (1976) 248-254. 3 De Belleroche, J.S. and Bradford, H.F., Amino acids in synaptic vesicles from mammalian cerebral cortex: a reappraisal, J. Neurochem., 21 (1973)441-451. 4 De Lorenzo, R.J. and Freedman, S.D., Calcium dependent neurotransmitter release and protein phosphorylation in synaptic vesicles, Biochem. Biophys. Res. Commun, 80 (1978) 183-192. 5 Hansen, O., Relationship between g-strophantin binding capacity and ATPase activity in plasma membrane fragments from ox brain, Biochim. Biophys. Acta, 233 (1971) 122-132. 6 Hata, F., Kuo, C.H., Matsuda, T. and Yoshida, H., Factors required for Ca-sensitive acetylcholine release from crude synaptic vesicles, J. Neurochem., 27 (1976) 139-144. 7 Jain, M.K., Non-random lateral organization in bilayers and biomembranes. In R.C. Aloia (Ed.), Concepts of Membrane Structure, Membrane Fluidity in Biology, Vol. 1. Academic Press, New York, 1983, pp. 1-37. 8 Kadota, K. and Kadota, T., Isolation of coated vesicles, plain synaptic vesicles, and flocculent material from a crude synaptosome fraction of guina pig whole brain, J. Cell Biol., 58 (1973) 135-151. 9 Kontro, P., Marnela, K.-M. and Oja, S.S., Free amino acids in the synaptosome and synaptic vesicle fractions of different bovine brain areas, Brain Research, 184 (1980) 129-141. 10 Kuffler, S.W. and Yoshikami, D., The number of transmitter molecules in a quantum: an estimate from iontophoretic application of acetylcholine at the neuromuscular synapse, J. Physiol. (London), 251 (1975) 465-482. 11 L~ihdesm~iki, P., Karppinen, A., Saarni, H. and Winter, R.. Amino acids in the synaptic vesicle fraction from calf brain: content, uptake and metabolism. Brain Research. 138 (1977) 295-308.
most grateful to the John Simon Memorial Founda-
12 Mangan, J.L. and Whittaker, V.P., The distribution of free amino acids in subcellular fractions of guinea-pig brain, Biochem. J., 98 (1966) 128-137. 13 Naito, S. and Ueda, T., Adenosine triphosphate-dependent uptake of glutamate into protein 1-associated synaptic vesicles, J. Biol. Chem., 258 (1983) 696-699. 14 Orrego, F., Criteria for the identification of central neurotransmitters and their application to studies with some nerve tissue preparations in vitro, Neuroscience, 4 (1979) 1037-1057. 15 Rassin, D.K., Amino acids as putative transmitters: failure to bind to synaptic vesicles of guinea pig cerebral cortex, J. Neurochem., 19 (1972) 139-148. 16 Riveros, N. and Orrego. F., A search in rat brain cortex synaptic vesicles for endogenous ligands for kainic acid receptors, Brain Research, 236 (1982) 492-496. 17 Riveros, N. and Orrego, F., A study of possible excitatory effects of N-acetylaspartyl glutamate in different in vivo and in vitro brain preparations. Brain Research, 299 (1984) 393-396. 18 Seidler, F.J., Kirksey, D.F., Lau, C., Whitmore, W.L. and Slotkin, T.A., Uptake of (-)(3H)-norepinephrine by storage vesicles prepared from whole rat brain: properties of the uptake system and its inhibition by drugs, Life Sci., 2l (1977) 1075-1086. 19 Vargas, O. and Orrego, F., Elevated extracellular potassium as a stimulus for releasing (3H)-norepinephrine and (~4C)a-aminoisobutyrate from neocortical slices. Specificity and calcium dependency of the process, J. Neurochem., 26 (1976) 31-34. 20 Vincent, S.R. and McGeer, E.G., Kainic acid binding to membranes of striata! neurons, Life Sci., 24 (1979) 265-270. 21 Watkins, I.C. and Evans, R.H., Excitatory amino acid transmitters, Ann. Rev. Pharmacol. Toxicol., 21 (1981) 165-204. 22 Whittaker, V.P., The synaptic vesicles. In A. Lajtha (Ed.), Handbook of Neurochemistry, 2nd edn., Vol. 7, Plenum, New York, 1984, pp. 41-69.