Neurochemical Research, VoL 17, No. 4, 1992, pp. 339-343

Is There a High Molecular Weight Glutamic Acid Decarboxylase? M. P6rez-de la Mora 1, A. B. Rizo-Silva ~, and J. M6ndez-Franco I (Accepted September 18, 1991)

Gel-filtration chromatography from crude extracts of mouse brain indicates the presence of a high molecular weight (HMW) (more than 750 kd) and a low molecular weight (LMW) (87.5 Kd) glutamic acid decarboxylase (GAD) when they are concentrated prior chromatography either by precipitation with salts or ethanol. Kinetically both peaks of GAD-activity exhibit an almost identical Km for L-glutamate. Only LMW-GAD appears if the concentration step is carried out by ultrafiltration or-if the extract is chromatographied without the concentrating step. HMW-GAD disappears from the chromatographic profiles if ethanolic extracts of GAD are treated with Triton X-100 before the chromatographic procedure. The sodium sulfate precipitation of a previously separated LMW-GAD gives rise to the reappearance of a HMW-GAD peak. Apparently HMWGAD does not exist as a different molecular entity; indeed it may be an artefactual aggregation of LMW-GAD. KEY WORDS: Gel filtration; glutamic acid decarboxylase; multiple forms; molecular weight; GABA.

INTRODUCTION

also reported in several species of mammals (12-15). Although a considerable effort has been devoted to the study LMW-GAD, practically no work has been made to establish the actual existence and properties of HMWGAD. The aim of this work is to establish if indeed a high-molecular weight GAD actually exists in brain as a distinct molecular entity.

It is now well established that -/-aminobutyric acid (GABA) plays a fundamental role in the mammalian central nervous system (1,2). GABA neurons inervate all regions of the brain (3-5) and participate in a wide variety of functional manifestations. GABA is synthesized from glutamic acid in a single-reaction, step by the action of glutamic acid decarboxylase (EC 4.1.1.15) (GAD) (6). Similarly to several other enzymes involved in the biosynthesis of neurotransmitters this enzyme seems to be heterogeneous (79). Thus, both a high- and low-molecular weight (ttMW, LMW) forms have been described in the mouse and rat brains (4,10,11) and a number of low molecular weight species differing in kinetic properties, isoelectric points, electrophoretic mobility and hydrophobicity have been

EXPERIMENTAL PROCEDURE For the preparation of crude extracts of GAD several albino mice (local strain) were killed by decapitation and their brains quickly removed and weighed. The brains were homogenized in 5 vol of icecold water in a Sorvall Omni-Mixer. The homogenate was centrifuged for 30 min at 100 000 x g in a Beckman ultracentrifuge and the supernatant adjusted to pH 7.0 by the addition of a small aliquot of 1.0 M K-phosphate buffer pH 7.0. Concentrated solutions of pyridoxal phosphate (PLP) and aminoethylisothiouronium (AET) were added to the extracts to give a final concentration of 1.0 mM PLP and 10 mM AET. High-speed supernatants (GAD specific activity 3.7 - 4.3 nmol/

i Departamento de Neurociencias, Instituto de Fisiologia Celular, Universidad Nacional Aut6noma de M6xico. Apdo. Postal 70-600, 04510 M6xico, D.F. M6xico.

339 0364-3190/92/0400-0339506.50/09 1992PlenumPublishingCorporation

340 min/mg protein) were concentrated prior to gel filtration chromatography by precipitation with either sodium sulfate or ethanol. In some experiments supernatants were concentrated by ultrafiltration in Amicon stirred cells equipped with PM30 membranes. When sodium sulfate was used, the enzyme was precipitated by adding the salt (22.5 g/100 ml) under stirring to supernatants kept at room temperature. The mixture was centrifuged for 20 min at room temperature at 20 000 x g; the resulting pellet was resuspended in a 3.0 - 3.5 ml ice-cold 20 mM K-phosphate buffer pH 7.0; 1.0 mM PLP; 10 mM AET (St buffer). When ethanol was used, the high-speed supematants submerged in a bath of acetone which was cooled by gradual addition of dry ice, were taken at I~ and under stirring absolute ethanol was added drop-wise to give a 10% (v/v) concentration. The temperature was then lowered to -3~ and ethanol was further added to attain a 13% final concentration. The mixture was stirred for an additional 5 min at - 6~ and then centrifuged at the same temperature. The corresponding pellets were processed as indicated for the sodium sulfate precipitation experiments. In some experiments ethanolic GAD extracts prepared as above were incubated for 20 min (0-4~ under gentle stirring with 0.25% (V/V) Triton X-100. Chromatographic Procedures. LMW- and HMW-GAD activities were resolved by gel filtration chromatography on LKB columns (1.6 x 42 cm) packed with Ultrogel AcA34 (exclusion limit 750,000). Concentrated GAD extracts (2.15 ml) previously centrifuged at 105,000 g (30 min) were used. The column was equilibrated and run with St buffer at 20 ml/h and 1.6 ml fractions were collected. GAD activity and protein concentration were measured in each third fraction. The void volume of the column was determined with 2.15 ml dextran blue (1.0 mg/ml St buffer) chromatographied under the same conditions. In some experiments GAD extracts were chromatographied on DEAE-cellulose D-52. A LKB column (1.3 x 8.2 cm) equilibrated in St buffer was loaded with 15.0 ml GAD extract. The column was washed with St buffer and subsequently eluted with 200 and 500 mM K-phosphate pH 7.0, 1.0 mM PLP 10 mM AET. The column was run at 20 ml/h and 1.6 ml fractions were collected. Optical density at 280 nm was continuously monitored and GAD activity was assayed in selected fractions. Km Dete;mination. Km values from HMW and LMW-GAD were obtained from double reciprocal plots (16) using seven different final concentrations of glutamate (0.6, 1.9, 3.7, 5.5, 11, 20 and 33 raM). The specific activity of L-glutamate was 37.8 ixCi/mmol. Determination of GAD Molecular Weight. An Ultrogel AcA 34 column (1.6 x 44 cm) was utilized. The column was calibrated with: ovoalbumin (MW 43 000), bovine serum albumin (MW 67 000), rabbit muscle aldolase (MW 158,000) and bovine liver catalase (MW 232 000). Dextran blue 2,000 was used for establish the void volume. Markers were dissolved in 2.15 ml St buffer and chromatographied as described under chromatographic procedures. Fractions were monitored continuously at 280 rim, and the elution volumes were recorded at the peaks of the corresponding protein profiles. GAD molecular weight was extrapolated from a graph showing (elution volumes - void volume) vs log of the markers MWs. Analytical Measurements. GAD-activity was measured by triplicate according to Albers and Brady (17) with some modifications. To a glass micro-tube (0.3 x 25 mm) containing 8 p.l of 87.5 mM (1-14C) L-glutamate (specific activity 37.8 ~Ci/mmol), 0.5 mM PLP and 50 mM K-phosphate pH 7.0 was added 62 ul GAD. The micro-tube was joined to another microtube containing 50 Ixl Hyamine hydroxide through a 12 cm latex tube. Both micro-tubes were incubated for 20 rain in a water bath set at 37~ At the end of the incubation period, the reaction

P6rez-de la Mora, Silva, and M6ndez-Franco was stopped by the injection of 75 izl 3.8 M H2SO4. In order to allow for a complete evolution and absorption of 14CO2 in Hyamine, the measuring device in a nearly extended position was incubated for an additional 60 rain at 40~ in an electric oven. Finally the Hyamine micro-tube was dropped into a vial containing 10 ml of a scintillation cocktail (4.0 g 2,5-diphenyloxazole (PPO) and 0.1 g (2(4-methyl-5phenyl-oxazolyl) benzene (DMPOPOP) per 1 1 toluene), and its content pumped-out with the aid of a glass rod. Radioactivity was determined in a Tri-Carb Packard Scintillation counter with ca. 80% efficiency. Protein was measured according to Bradford (18) using bovine serum albumin as standard. Materials. [1-14C]Dc-glutamic acid (23 mCi/mmol) and Hyamine hydroxide were obtained from Amersham Laboratories. Ultrogel AcA 34 was from LKB Produkter. DEAE-Cellulose D52 was from Whatman Bio Systems Ltd. Dextran blue 2,000 and protein markers came from Pharmacia Biotechnology Products. PLP, AET, Triton X-100, PPO, DMPOPOP, bovine semm albumin and Coomassie brilliant blue G 250 were purchased from Sigma Chemical Co. All other chemicals used were of the best purity locally available.

RESULTS When a crude brain extract was concentrated either by sodium sulfate or ethanol and chromatographied on Ultrogel AcA 34, two main different peaks of GADactivity were evident (Figure 1): the first was excluded from the column (HMW-GAD) and the second (LMWGAD) was eluted within its included volume. According to the exclusion volume, the molecular weight for the HMW GAD was above 750,000, whereas the LMWGAD was found to exhibit a molecular weight of 87,500 (graph not shown). Similar results were obtained when high-speed supernatant were concentrated with 60% saturation ammonium sulfate (data not shown). In contrast, only the LMW-GAD peak was obtained if the GAD extract was concentrated by ultrafiltration on Amicon PM 30 membranes or chromatographied without any concentration step (Figure 2). Considering that the HMW-GAD peak could result from aggregation of the LMW-GAD during the concentration of the supematants, we incubated in the cold for 20 min an ethanolic GAD extract with 0.25% Triton X100 in an attempt to disaggregate the HMW-GAD. Thereafter, the extract was loaded into a DEAE cellulose column in order to eliminate any GAD which could be trapped into detergent miscelles. GAD was eluted from the ion-exchange column with 200 mM phosphate buffer (Figure 3A) and the fractions with the highest GAD activity were pooled and concentrated by ultrafiltration on Amicon PM 30 membranes. Under these conditions, only the LMW-GAD peak could be identified after gel filtration chromatography of the ultrafiltrate (Figure 3B). In order to further ascertain if LMW-GAD could aggregate

Are There Different MW GADs?

6,0

600

4,0

I

341

v.

~

3.0-

A

~

~

- 4.50

2,0-

I

-.300

E 2,O

E c

900

A

Vo

I,O-

500

.~_ E

-i50

E

-~

J

x,

c

E

0

"//

0

'B >-

I--

O < I

,,=s 0

9,0.

E

-900

6.o 6,0-

-600

3,0-

-300

O-

t 20

30 FRACTION

i 40

0

50

No

Fig. 1. Chromatographic profiles of a GAD extract cortcentrated by precipitation with either sodium sulfate (A) or ethanol (B). UltrogeJ AcA 34 Columns were loaded respectively with 21 mg protein (specific activity 4.5 nmol/min/mg protein) and 25 nag protein (specific activity 6.7 nmol/min/mg protein) and chromatographied as described under Experimental Porcedure. V~ indicates the point at which dextran blue gave maximal absorbance.

to produce HMW-GAD, the "light" GAD form was obtained by chromatography in Ultrogel AcA 34, and subsequently concentrated by ultrafiltration to give a protein concentration of 4.3 mg/ml. LMW-GAD, once concentrated, was precipitated by sodium sulfate as described above, and fractionated by gel filtration chromatography. Both the HMW-GAD and the LMW-GAD forms were eluted from the column (Figure 4). This indicates that LMW-GAD had indeed been aggregated upon precipitation. To find out whether or not HMW-GAD and LMW-GAD had similar properties we determined their Km values for L-glutamate. Km values were 2.3 • 0.35 and 2.4 -4_-_0.5 (n=5) for the HMW-GAD and LMWGAD forms respectively.

DISCUSSION In agreement with previous reports (4,10,11), gelfiltration chromatograpy of GAD extracts concentrated by sodium sulfate precipitation resulted in the appearance of two peaks of enzyme activity (Figure 1). The

,< I ,..-., < O

iooo

0 (2._

3,0

0

500

'L'

~ 20

i 30

i

FRACTION

40

i

0

50

N~

Fig. 2. Ultrage; AcA 34 chromatography of G A D extracts. (A) 2.I5 ml of a high-speed supematant from the homogenate containing 16 mg protein (specific activity 3.7 nmol/min/mg protein) were introduced into the column and chromatograhied. (B) A GAD extract was concentrated by ultrafiltration on Amicon PM 30 membranes and introduced into the column (38 mg protein, specific actMty 4.1 nmol/min/ mg protein). For the chromatographic details see Experimental Procedure and Figure 1.

first one was excluded from the column and had, according to the exclusion limit of Ultrogel AcA 34 a molecular weight above 750,000. The second peak of GADactivity was recovered within the included volume of the Ultrogel AcA 34 column, and exhibited a molecular weight of 87 500 daltons; this is close to the value reported by Wu et ai. (19), i.e., 86 000. Moreover a similar chromatographic profile was found when GAD extracts concentrated by precipitation with ethanol were passed through an Ultrogel AcA 34 column (Figure 1). However, only the LMW-GAD peak showed up if in the GAD-extracts the concentrating step was omitted or they are concentrated by ultrafiltratiou on Amicon PM 30 membranes (Figure 2). Since salt and organic solvent precipitations induce a high degree of protein aggregation due, among other effects, to dehydration (20), it is conceivable that the HMW-GAD does not exist intracellularly, but that rather it is an artefact produced by aggregation of LMW-GAD, as has been already suggested by Denner et al. (11). The fact that HMW-GAD does not appear when GAD ex-

342

P~rez-de la Mora, Silva, and M~ndez-Franco

30

I

2O m~

2OO m~a

5OO mM

~.o

o

0,8

!ilo

5 < I

c:~

AI

I

8,0

- 200

4,0

-I00

0

0

i~

-500

c

".~

_

/

2O0

E 4.0-

z bJ ~

-IOQ

2.o

< i

_

o

o-t

o

2o

3o

50

4O

FRACTION

No

Fig. 4. Gel filtration rechromatography of LMW-GAD concentrated by precipitation with sodium sulfate. 2.3 ml (16 mg protein) LMWGAD (11.5 nmol/min/mg protein) concentrated succesivley by ultrafiltration and sodium sulfate precipitation were used. For chromatographic and other details see the text and Figure 1.

0 20

50

FRACTION

40

50

No

Fig. 3. DEAE cellulose and gel filtration chromatography of GAD extracts. (A) An ethanolic GAD extract treated with 0.25% Triton X100 was applied to a DEAE cellulose D-52 column and after washing with St buffer GAD was eluted stepwise with 200 a n d 500 mM Kphosphate St buffer (see arrows). Fractions 18-21 containing the highest GAD activity were pooled, concentrated on Amicon PM 30 membranes and centrifuged at 100 000 g for 30 min. (B) 2.1 ml (11.3 mg protein) from the above GAD preparationg (14 nmoles/min/mg protein) were introduced into a Ultrogel AcA 34 column and chromatographies. See the text and Figure 1 for details of both chromatographies.

tracts are not subject to aggregating steps, or when concentrated by ultrafiltration (which does not increase the interactions among proteins) supports this view. Furthermore, treatment of ethanolic extracts of GAD with Triton X-100 chromatography on Ultragel AcA 34 diminishes both the amount of protein present in the void volume of the column, and prevents the appearance of the HMW-GAD peak. The possibility that in these last experiments HMW-GAD had not been eluted from the ion-exchange column, or was lost in the subsequent ultrafiltration step seems unlikely, since no GAD activity was eluted even with 500 mM phosphate and not more than 15% of the enzyme activity was lost during the ultrafiltration step. More likely HMW-GAD was disaggregated by the detergent. Finally the fact that it is possible to reconstitute a HMW-GAD peak from LMWGAD (Figure 4) when this last form of the enzyme is concentrated by precipitation with NazSO 4 argues very strongly against the existence an intracellular "heavy" GAD form, and supports the suposition that the HMWGAD activity peak represents LMW-GAD trapped in a protein aggregate formed during the concentration step(s).

In addition the following findings indicate that both forms correspond to the same enzyme: 1) identical Km values for L-glutamate; 2) lack of differences in pH optimum; 3) similar immunodifusion pattern; and 4) similar inhibition by antibodies raised against LMW-GAD (this work and 10). In conclusion, although the results of this paper do not argue against the possibility that intracellularly LMWGAD could aggregate under certain physiological conditions, i.e., Ca 2§ concentration (21) they are consistent with the view that HMW-GAD does not exist as a distinct molecular entity, but that it only represents an artefactual form of the enzyme produced during its precipitation with either salts or organic solvents.

ACKNOWLEDGMENTS We thank Dr. A. G6mez-Puyou for his critical review and comments and Mrs. Mafia Teresa Torres-Peralta for her excellent secretarial skills. This work was partially supported by the grant D-III903548 from Consejo Nacional de Ciencia y Tecnologh (CONACyT).

REFERENCES 1. Krogsgaard-Larsen, P., Scheel-Kriiger, J., and Kofod, H. (eds.) 1979. GABA Neurotransmitters, Alfred Benzon Symposium XII". Munksgaard, Copenhagen. 2. Fonnum, F. 1987. Biochemistry, anatomy and pharmacology of GABA neurons. Pages 173-182, in Miltzer, H. Y. (ed) Psychopharmacology: The Third Generation of Progress, New York, Raven Press. 3. Roberts, E. 1978. Immunocytochemical visualization of GABA neurons. Pages 95-102, in Lipton, M. A., DiMascio, A., and

A r e There Different M W G A D s ?

4.

5.

6. 7. 8. 9. 10. 11. 12.

Killam, K. F. (eds.) Psychopharmacology: A Generation of Progress, Raven Press, New York. P6rez de la Mora, M., Possani, L.D., Tapia, R., Ter~in, L., Palacios, R., Fuxe, K., HOkfelt, T., and Ljungdahl,/~. 1981. Demonstration of central ",/-aminobutyrate-containingnerve terminals by means of antibodies against glutamate decarboxylase. Neuroscience 6:875-895. Mugnaini, E., and Oertel, W. H. 1985. An atlas of the distribution of GABAergic neurons and terminals in the rat CNS as revealed by GAD immunocytochemistry. Pages 436--608, in Bj6rklund, A., H6kfelt, T. (eds). Handbook of Chemical Neuroanatomy, Elsevier, Amsterdam. Roberts, E., and Simonsen, D. G. 1963. Some properties of Lglutamic acid decarboxylase in mouse brain. Biochem. Pharmacol. 12:113-134. Benishin, C. G., and Carroll, P. T. 1983. Multiple forms of choline-O-acetyltransferase in mouse and rat brain: solubilization and characterization. J. Neurochem. 41:1030-1039. Sze, P. Y., Alderson, R. F., and Hedrick, B. J. 1983. Two forms of striatal tyrosine hydroxylase from DEAE-cellulose chromatography. Brain Res. 268:129-137. Hersh, L. B., Wainer, B. H., and Andrews, L. P. 1984. Multiple isoelectric and molecular weight variants of choline acetyltransferase. Artifact or real? J. Biol. Chem. 259:1253-1258. Wu, J. -Y., Wong, E., Saito, K., Roberts, E., and Schousboe, A. 1976. Properties of L-glutamate decarboxylase from brains of adult and newborne mice. J. Neurochem. 27:653-659. Denner, L. A., Wei, S. C., Lin, H. S., Lin, C. -T., and Wu, J. -Y. 1987. Brain L-glutamate decarboxylase: purification and subunit structure. Proc. Natl. Acad. Sci. USA 84:668-672. Spink, D. C., Wu, S. J., and Martin, D. L. 1983. Multiple forms

343

13. 14. 15.

16. 17. 18. 19. 20.

21.

of glutamate decarboxylasein porcine brain. J. Neurochem. 40:113119. Denner, L. A., and Wu, J. -Y. 1985. Two forms of rat brain glutamic acid decarboxylase differ in their dependence on free pyridoxal phosphate. J. Neurochem. 44:957-965. Spink, D. C., Porter, T. G., Wu, S. J., and Martin, D. L. 1985. Characterization of three kinetically distinct forms of glutamate decarboxylase from pig brain. Biochem. J. 231:695-703. Legay, F., Henry, S., and Tappaz, M. 1987. Evidence for two distinct forms of native glutamic acid decarboxylase in rat brain soluble extract: an immunoblottingstudy. J. Neurochem. 48:10221026. Dawes, E. A. 1963. Quantitative Problems in Biochemistry. Pages 51-74, in F. and S. Livingstone Ltd. Edimburgh and London. Albers, R. W., and Brady, R. O. 1959. The distribution of glutamate decarboxylase in the nervous system of the rhesus monkey. J. Biol. Chem. 234:926-928. Bradford, M. 1976. Methods for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248--254. Wu, J. -Y., Matsuda, T., and Roberts, E. 1973. Purification and characterization of glutamate decarboxylase from mouse brain. J. Biol. Chem. 218:3029-3034. Green, A. A., and Hughes, W. L. 1955. Protein fractionation on the basis of solubility in aquous solutions of salts and organic solvents. Pages 67-90, in Colowick, S.P., and Kaplan, N.O. (eds.) Vol. 1. Methods in Enzymology, Academic Press, New York. Fonnum, F. 1968. The distribution of glutamate decarboxylase and aspartate transaminase in subcellular fractions of rat guineapig brain. Biochem. J. 106:401-412.