Communication

The Chemical Labeling of Glutamate Decarboxylase in Viuo*

(Received for publication, July 1, 1980, and in revised form, November IO, 1980)

Robert R. Rando From the Department of Pharmacology, Haruard Medical School, Boston, Massachusetts 02115

Mouse brain glutamate decarboxylase(s) was specif- ically titrated in vivo and in crude brain homogenates by a combination of gabaculine and [a-'H]acetylenic y- aminobutyric acid. This specific titration is based on the differential spectra of action of these two mecha- nism-based enzyme inactivators. The specificity of the titration in vitro was demonstrated by showing that the time course of radioactivity incorporation exactly paralleled the time course for glutamate decarboxylase inactivation. Furthermore, pretreatment of the crude homogenate with aminooxyacetic acid and a-methyl- trans-3-dehydroglutamate, two inactivators of gluta- mate decarboxylase which function by entirely differ- ent mechanisms, decreased count i n c o ~ o r a t ~ o n ([n- 3H]acetylenic y-aminobutyric acid) to the same extent as the activity was decreased. Injection of [a-3H]acety- lenic y-aminobutyric acid intraperitoneally after gaba- culine injection led to incorporation of 0.46 nmol of inactivatorfmouse brain, when approximately 70% of the enzyme was inactivated. This means that there is approximately 0.66 nmol of glutamate decarboxylase/ 0.5 g of mouse brain, assuming the stoichiometry of inactivator bound to enzyme is one. This value is simi- lar to the one (0.646 nmol) obtained from a calculation based on the enz-yme purification data (Wu, J.-Y. (1974) in y-Arninobutyric Acid in Neroous System Function (Roberts, E., Chase, E. N., and Tower, D. B., eds) pp. 7- 55, Raven Press, New York).

L-Glutamate decarboxylase(s) (EC 4.1.1.5) is a pyridoxal phosphate-linked enzyme which is responsible for the biosyn- thesis of the major inhibitory neurotransmitter y-aminobu- tyric acid (1). This enzyme appears to be specifically located in neurons which secrete y-aminobutyric acid and can there- fore be t.aken as a marker for them. In fact, peroxidase-coupled anti-glutamate decarboxylase antibodies have been used to localize GABA'-ergic neurons in the central nervous system (2). Since this enzyme is of such great importance, a chemical method that would enable the specific and irreversible titra- tion of it ia wiuo would be of considerable interest. Not, only could such a titrant be used histochemically to easily locate GABA-ergic neurons in t.he central nervous system, but it could be used to quantitatively monitor the absolute levels of glutamate decarboxylase in vivo as a function of neurophar-

* This work was supported by United States Public Health Services Grant NS 11550 from the National Institutes of Health. The costs of publication of this art,icle were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduer- tisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviation used is: GABA, y-aminobutyric acid.

macological and behavioral manipulation. In principle it should be possible to specifically label the

enzymets) which d e c ~ b o x y ~ a t e L-glutamate in vivo with highly selective irreversible enzyme inhibitors. Mechanism- based enzyme inactivators, which require the target enzyme to catalyze its own destruction from chemically unreactive substrates, are potential candidates for these studies (3, 4). However, even enzyme inactivators of this type are likely not to show absolute specificity due to their being substrates for more than one enzyme. This can be clearly shown with 4- amino-I-hexynoic acid GABA ( 5 ) . This compound, although originally designed as a specific inactivator of y-aminobutyric

H-C"C-CH-CH.-CHL-CO~H I

NHr 1

2

acid transaminase, is also a potent inactivator of glutamate decarboxylase and ornithine S-aminotransferase (6, 7). Thus this inhibitor in and of itself is not likely to be useful for labeling studies in uiuo. However, it could be rendered useful in a differential labeling protocol if a second reagent were available that inactivated y-aminobutyric acid transaminase, ornithine ~-aminotransfer~e, and other enzymes possibly in- activated by 1. It turns out that such a reagent is readily available in gabaculine 2. This naturally occurring neurotoxin does not inactivate glutamate decarboxylase but is a potent inactivator of y-aminobutyric acid transaminase and ornithine S-aminotransferase by well described mechanisms (7-9). Thus this compound shares a spectrum of action comparable to that of acetylenic GARA save for the fact that it cannot inactivate glutamate decarboxylase. In this report we show that. something approaching absolute specificity in the in riirlo labeling of mouse brain glutamate decarboxylase can be real- ized using gabaculine and [ff-.~H]acetylenic GABA in combi- nation. The experiments are performed by pretreating wibh unlabeled gabaculine 2 followed by t~he addition of [a-.'H]- acetylenic GABA.

MATERIALS AND METHODS

Gabaculine was synthesized by the published procedure, as was [a-"Hlacetylenic GABA (5,8,9). a-Methyl-trans-3-dehydroglutamate was prepared by a route similar to that described in the literature (10). Aminooxyace~ic acid was purchased from Sigma. Glutamate decarboxylase activity was assayed by measuring the amount of ['4C]COr released from [I-'*C]glutamate ( I ) . The crude mouse brain homogenate was prepared by the method of Wu ( I ) . Mice were sacrificed by decapitation and their brains were removed. One to two ml of chilled 5 mM potassium phosphate (pH = 7.5) containing 0.2 mM pyridoxal phosphate and 5 mM a-aminoethylisothiouronium bro- mide were added per mouse brain and the mixture was homogenized with a Teflon-glass homogenizer. The suspension was treated with an additional 1.5 ml of the above buffer/mouse brain at 5°C for 1 h for the purpose of extraction. The suspension was centrifuged a t 10,ooO X g for 15 min according to the published purification scheme (I). The crude supernatant was used in the labeling experiments. Radio- activity incorporation from fa-"HJacetylenic GABA was measured by

1111

1112 Glutamate Decarboxylase Labeling

trichloroacetic acid precipitation of the labeled protein followed by extensive washing (4x1, dissolution in Protosol (New England Nu- clear), and counting in Econoflor (New England Nuclear).

RESULTS AND DISCUSSION

In the initial experiments described below, the rate of glutamate decarboxylase inactivation, after gabaculine treat- ment by [a-'Hlacetylenic GABA, was determined as a func- tion of radioactivity incorporation into protein in order to estimate the specificity of the inactivator toward glutamate decarboxylase in a crude mixture. In the experiments shown in Fig. 1, a crude extract of mouse brain, containing approxi- mately 0.1% glutamate decarboxylase (total protein), was pre- pared and subjected to inactivation by [a-'Hlacetylenic GABA. The rate of inactivation of glutamate decarboxylase was determined at a [a-'3H]acetylenic GABA concentration of 1 mM. At the same time radioactivity incorporation into protein was also determined with and without gabaculine (IO-" M) pretreatment. The gabaculine pretreatment corn- pletely abolishes transaminase activity within 0.5 h. As can be seen in Fig. 1, the incorporation of radioactivity into the protein precisely parallels the time course for the inactivation of glutamate decarboxylase in the gabamline pretreated case. This strongly suggests that labeling of protein in this crude

Control rnlnus inhibitor

40

a

1500

k!

8 900 2

5

1200 2

600 9 8

300

0

hi/ NU TES

FIG. 1. Time course for the in uitm labeling of glutamate decarboxylase. The crude mouse brain extract was pretreated with 0.1 p~ unlabeled acetylenic GABA for 15 s at 35°C as a precaution to block any possible nonspecific binding. A t this concentration and incubation period the inhibitor has no effect on glutamate decarbox- ylase (GAD) activity. Samples of the supernatant were treated with either M gabaculine for 30 min, IO-3 M unlabeled acetylenic GABA for 45 min, or left untreated (all a t 35°C). The samples were then treated with M [ff-~H~acetylenic GABA (specific activity = 2.1 mCi/mmol) a t 35°C and at indicated times samples were removed for assaying remaining glutamate decarboxylase activity as well as count incorporation. GIutamate decarboxylase activity was assayed by measuring the amount of ['4C]C02 released from [1-14C]glutamate (1). For count incorporation, 2.5 ml of extract were added to 7.5 ml of 10% trichloroacetic acid and the suspension was centrifuged and washed with the trichloroacetic acid solution four times. The final precipitate was dissolved in Protosol (New EngIand Nuclear) over- night and counted in Econoflor (New England Nuclear). In the count incorporation curve, the extract of one mouse brain was used to generate each point. A duplicate of each point was run and the counts incorporated into the unlabeled acetylenic GABA-treated controls were subtracted. In the absence of gabaculine, 2850 cpm were maxi- mally incorporated into the trichloroacetic acid-precipi~ble radioac- tivity after treatment with [u-'H]acetylenic GABA. This additional incorporation is due to the labeling of GABA transaminase and ornithine &aminotransferase by [a-'H]acetylenic GABA in the ab- sence of gabaculine. 0, enzyme activity in the presence of [a-'H]- acetylenic GABA; 0, control enzyme activity; A, radioactivity incor- poration (gabaculine + [a-'H]acetyienic GABA).

mixture occurs as a consequence of glutamate decarboxylase inactivation; otherwise it would be highly fortuitous for the two time courses to be identical. In addition, roughly double the amount of counts are incorporated into the preparation which did not contain gabaculine. This higher labeling is due to the incorporation of [a-3H]acetylenic GABA into the trans- aminases and possibly other enzymes. High concentrations of gabaculine do not decrease the amount of putative glutamate decarboxylase labeling, and pretreatment of the extract with

M unlabeled acetylenic GABA for 1 h abolishes all count incorporation.

TO further explore the selectivity of the labeling of gluta- mate decarboxylase by the [a-3H]acetylenic GABA/gabacu- line combination, radioactivity incorporation was studied in combination with pretreatment with known irreversible glu- tamate decarboxylase inactivators. The gabaculine-modified crude extract was treated with different concentrations of aminooxyacetic acid, a compound that reacts with pyridoxal phosphate, and with a-methyl-trans-3-dehy~oglutamate, a known mechanism-based inactivator of glutamate decarbox- ylase, at a single concentration but for varying times. These two inhibitors inactivate glutamate decarboxylase by entirely different mechanisms. Aminooxyacetic acid is a dose-depend- ent, relatively nonspecific inactivator which reacts with active site pyridoxal phosphate (1). a-MethyI-trans-3-dehydroglu- tamate, on the other hand, is a time-dependent specific inac- tivator of the enzyme which alkylates an active site residue (11). By inactivating the enzyme to different extents with the two inactivators and then titrating the remaining enzyme with [a-"Hlacetylenic GABA, the selectivity of count incorporation can be estimated. Linearity between residual glutamate de- carboxylase activity and radioactivity incorporation would only result if acetylenic GABA is specific for the glutamate decarboxylase under the conditions of the experiment. As far as can be determined, the only property that aminooxyacetic acid and a-methyl-trans-3-dehydroglutamate share is their ability to inactivate glutamate decarboxylase. The results of these experiments are shown in Fig. 2. Indeed, the curve is linear, a result consistent with the anticipated selectivity of the [a-'Hlacetylenic GABA/gabaculine combination. Had other discrete enzymes been inactivated by acetylenic GABA under these conditions they would have had to show sensitiv- ities toward aminooxyacetic acid and a-methyl-trans-3-dehy- ~ ~ g i u t a r n a t e similar to that of glutamate decarboxylase itself. However, the possibility that an enzyme is labeled which is present in much lower levels than glutamate decarboxylase cannot be ruled out by these experiments. To explore this possibility, peptide mapping will be required using [u- 3H]acetylenic GABA of high specific activity.

The experiments described above show that in whole brain homogenates, glutamate decarboxylase can be specifically ti- trated with the gabac~line/[a-~H]acetylenic GABA combina- tion. It should also be possible to accomplish the same result in uivo. To these ends mice were treated with unlabeled acetylenic GABA followed by {a-3H]acetylenic GABA, by gabaculine followed by [~-'H]acetyIenic GABA, and by [a-3H]acetylenic GABA. The gabaculine concentration chosen completely abolishes brain y-aminobutyric transami- nase levels under the conditions of the experiment and the [a-3H]acetylenic GABA inactivates approximately 70% of the brain glutamate decarboxylase. The results of these experi- ments are shown in Table I. As indicated, gabaculine markedly decreases the amount of incorporated radioactivity, and fur- thermore the level of incorporation is quite consistent with the in vitro studies and with the level of expected glutamate decarboxylase based on the enzyme purifkation data (1). Each average 0.5 g of mouse brain is expected to contain 0.65 nmol

Glutamate Decarboxylase Labeling 1113

of glutamate decarboxylase, and the titration data indicate that the 0.66 (0.46/0.7) nmol is present, assuming a stoichi- ometry of inhibitor incorporation of one. These experiments are consistent with the notion that the brain enzymeis) has been specifically titrated in vivo.

The experiments described show that something approach- ing absolute specificity in the titration of an enzyme can be achieved using a combination of inactivators. Gabaculine and

0 AOAA Treated

A aMDG Treated

I 500 1000 I500

/ A aMDG Treated

I I - 500 1000 I500

CPM /NCURPORATD FIG. 2. Titration of glutamate decarboxylase in the presence

of glutamate decarboxylase inactivators. Crude mouse brain enzyme was prepared as above. The extract was treated with lo-' M gabaculine to inactivate the transaminases. The extract was then divided into aliquots. Aliquots which were treated with aminooxy- acetic acid (AOAA) did not contain pyridoxal phosphate. Samples were then treated either with aminooxyacetic acid at concentrations of 0.025, 1.4, 2.5, 10, and 40 PM or with a-methyl-trans-dehydroglu- tamate (aMDG) (0.5 mM). The samples containing aminooxyacetic acid were incubated for 0.5 h at 35°C. The remaining enzymatic activity was determined at each concentration and the remainder of each sample was treated with 1 mM [a-'H]acetylenic GABA for 1 h and the counts incorporated into protein were determined as before. Corrections were made for the small reactivation (10-20%) of the aminooxyacetic acid-treated samples under the conditions of extended incubation. This reactivation was assessed by measuring the slight recovery of enzymatic activity found with the aminooxyacetic acid- treated enzyme carried through all of the above steps. In the a- methyl-trans-dehydroglutamate case, the samples were treated with the compound at 0.5 mM for varying periods of time at 35°C and the remaining enzymatic activity was determined. The samples were quickly cooled to 5OC to slow down the inactivation process. The samples were then filtered through a Sephadex G-10 column to remove the a-methyl-tr~s-dehydroglutamate. Approximately 75% of the control enzymatic activity was obtained in the void volume. The glutamate decarboxylasft (GAD) collected in the void volume was titrated with 1 mM [a-"Hfacetylenic GABA and the incorporated counts were ascertained as before. Resufts were corrected for the slightly different protein concentrations in the void volumes. The trichloroacetic acid-precipitable count incorporation is plotted as a function of the enzyme activity remaining after treatment with the two inhibitors. Each point represents the extracts from a single mouse brain. Duplicates were run for all points. 0, the aminooxyacetic acid- treated enzyme; A, a-methyl-trans-dehydoglutamate-treated en- zyme.

TABLE I Titration of mouse brain glufamate decurhoxytuse in uzuo

A group of four mice (25-30 g) were injected i n ~ a ~ r i t o n e a l l y with 100 mg/kg of gabaculine/mouse. A second group of four were injected with 200 mg/kg of uniabeled acetylenic GABAimouse and a third group of four were injected with saline. After 4 h all three groups of mice were injected intrapentonea~y with 200 mg/kg of [a-"Hlacety- lenic GABA (specific activity = 2.1 mCi/mmol). After an additional 4 h the mice were killed and their brains were removed and homog- enized in 5 ml of 5 mM potassium phosphate (pH = 7.5). It was determined that under these conditions approximateiy 70% of the mouse brain glutamate decarboxylase was inactivated when only [a- 'H]acetylenic GABA was added. The membrane fragments were spun down at lo00 X g. The supernatants were precipitated with 5% trichloroacetic acid. These precipitates were taken up in 5 ml of 5% trichloroacetic acid, dispersed by sonication, and centrifuged. This process was repeated four times. Finally the precipitates were dis- solved in Protosol overnight and counted in Econoflor. The average disintegrations per min found for the mice treated according to controls described are given. Each figure is the average of three determinations.

~

Incorporation per 0.5 g of mouse brain

~

&m nmol [a-"H]Acetylenic GABA 5110 1.12

+ Acetylenic GABA 300 0.066 + Gabaculine 2 100 0.46

acetylenic GABA are highly specific enzyme inactivators in their own right, being of a class of mechanism-based irrevers- ible enzyme inhibitors. However, as with any inactivator of this type, they can, in principle, inactivate any enzyme that can utilize them as a substrate. They will, of course, not chemically react with biomolecules which cannot activate them. Gabaculine is a potent inactivator of y-aminobut~ic acid transaminase, less so of ornithine S-aminotransferase, and possibly other y- or S-aminotransferases. Acetylenic GABA inactivates these same enzymes but, in addition, it is also a potent inactivator of glutamate decarboxylase, since the en- zyme can use the inhibitor as a substrate. Most importantly, glutamate decarboxylase cannot process gabaculine and hence is not inactivated by it. This is probably clue to the fact that the C-H bond, which must be enzymatically cleaved prior to inhibition, is much more acidic in acetylenic GABA than in gabaculine. This difference in the inhibitory spectrum of the two inactivators can thus be exploited to gain the necessary specificity of action.

A method of specifically titrating glutamate decarboxylase in vivo opens up the possibility of many different and inter- esting experiments. It will now be possible to quantitatively measure the neuronal levels of glutamate decarboxylase in vivo without destroying the tissue. In addition, using [a-3H]acetylenic GABA of high specific activity (10 Ci/mmol), it should be possible to map GABA-ergic neurons in the central nervous system by autoradiography with a much higher degree of resolution than currently possible.

REFERENCES 1. Wu. J.-Y. (1974) in y-Aminohutyric Acid in. Nerlious System

Function (Roberts, E., Chase, E. N. & Tower, D. B., eds) pp. 7- 55, Raven Press, New York

2. Saito, K., Barber, R., Wu, J.-Y., Matsuda, T., Roberts, B. & Vaughn, J . E. (1974) Proc. Nutl. Acad. Sei. Lr. S. A. 71, 269- 273

3. Rando, R. R. (1974) Science 185, 320-324 4. Abeles, R. H. & Maycock, A. L. (1976) Accts. Chem. Res. 9, 313-

5. Jung, M. J., Lippert, B., Metcalf, B. W., Schechter, P. J., Bohlem,

6. Jung, M. J., Metcalf, B. W., Lippert, B. & Casara, P. (1978)

319

P. & Sjoerdsma, A. (1977) J. Neurochem. 28, 717-723

1114 Glutamate Decarboxylase Labeling

Biochemistry 17,2628-2632 Commun. 76, 1276-1281 7. Jung, M. M. & Seiler, N. (1978) J. B i d . Chem. 253, 7431- IO. Bey, P. & Vevert, J. P. (1980) J. Org. Chem. 45, 3249-

7434 3253 8. Hando, R. R. (1977) Biochemistry 16,4604-4610 11. Chrystal, E., Bey, P. & Rando, R. R. (1979) J . Neurochem. 33, 9. Rando, R. R. & Bangerter, F. W. (1977) Biochem. Biophys. Res. 1501-1507