Production and characterization of polyclonal and monoclonal antibodies to rat brain l -glutamate decar☐ylase

Brain Research, 373 (1986) 1-14 Elsevier

1

BRE 11662

Research Reports Production and Characterization of Polyclonal and Monoclonal Antibodies to Rat Brain L-Glutamate Decarboxylase J.-Y. W U 1'2'3, L.A. DENNER 1, S.C. WEI 3, C.-T. LIN2.3, G.-X. SONG 2.3, Y.F. XU 3, J.W. LIU 3 and H.S. LIN2,3

1Program in Neuroscience and 2Department of Cell Biology, Baylor College of Medicine, Houston, TX 77030and 3Department of Physiology, The Milton S. Hershey Medical Center, The Pennsylvania State University, Hershey, PA 17033 (U.S.A.) (Accepted September 10th, 1985)

Key words: glutamate decarboxylase (GAD) - - GAD monoclonal antibody - - GAD polyclonal antibody - ~,-aminobutyric acid (GABA)

Specific monoclonal and polyclonal antibodies to rat brain glutamate decarboxylase (GAD) were produced and characterized. Polyclonal antibodies against GAD were raised in rabbits by injecting a total of 70-210/zg of purified GAD i.m. The specificity of antiGAD serum was established from a variety of tests including Ouchterlony immunodiffusion, immunoelectrophoresis, immunoprecipitation, dot immunoassay, ELISA tests and Western immunoblottings. In immunodiffusion and immunoelectrophoresis tests using partially purified GAD preparations and anti-GAD serum a single, sharp precipitin line corresponding to GAD activity was obtained. Quantitative immunoprecipitation of GAD activity was achieved using anti-GAD IgG and Staphylococcus aureus. Specificity of the antiserum was further indicated from a dot immunoassay and ELISA tests in which the intensity of the reaction product was proportional to the amount of GAD protein present. In the Western immunoblotting experiments using partially purified GAD preparations only two protein bands corresponding to the position of the two subunits of GAD were stained by anti-GAD IgG, further supporting the specificity of polyclonal antibodies against GAD. In addition to polyclonal antibodies, several specific GAD-antibodies-producing clones were also obtained by the hybridoma technique. The specificity of monoclonal antibodies against GAD were established from the following criteria: positive on ELISA test using homogenous GAD as antigen; formation of GAD-anti-GAD IgG complex as indicated from gel filtration chromatography and sodium dodecyl sulfate-polyacrylamidegel electrophoresis; and specific recognition of GAD subunit in a partially purified GAD preparation in Western immunoblotting test. Monoclonal antibodies were further characterized by immunohistochemical localization of known GABAergic neurons and their processes in the cerebellum and retina. INTRODUCTION G A B A has been established as a m a j o r inhibitory neurotransmitter in both v e r t e b r a t e and invertebrate nervous systems19,29,35,43. L - G l u t a m a t e decarboxylase ( G A D ) , the biosynthetic enzyme for y-aminobutyric acid ( G A B A ) , has been shown to be a specific m a r k e r for G A B A e r g i c neurons and their processes 30,31,43,45. Since the purification of G A D from mouse brain to h o m o g e n e i t y was achieved and specific antibodies against brain G A D b e c a m e available 4°,42,46, much progress has been m a d e in the identification of G A B A e r g i c neurons and their projections by immunocytochemical techniques (for review see refs. 30, 43, 45). Unfortunately, the supply of the

well-characterized monospecific antibodies against mouse brain G A D has been depleted. H e n c e we decided to u n d e r t a k e the purification of G A D from rat brain and to use the purified G A D p r e p a r a t i o n s for the production of specific polyclonal antibodies. In addition, we also decided to take the advantages of the recent advances in h y b r i d o m a techniques to prepare monoclonal antibody against G A D . T h e purification of G A D to homogeneity, the criteria of purity, the subunit structure and the kinetic properties of the purified G A D have been described elsewhereS-lo. In this communication, we describe the p r o c e d u r e for the production of both polyclonal and monoclonal antibodies against G A D . In addition, the following characterizations are included: immunodiffusion, im-

Correspondence: J.Y. Wu, Department of Physiology, The Milton S. Hershey Medical Center, The Pennsylvania State University, P.O. Box 850, Hershey, PA 17033, U.S.A. 0006-8993/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

munoelectrophoresis, Western immunoblotting, gel filtration and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), immunoprecipitation, enzyme inhibition, dot immunoassay, enzymelinked immunoadsorbent assay (ELISA) and immunocytochemical localization. A portion of this work has been presented in a preliminary form8,9.

GAD was assayed by a radiometric method as previously described40. 42. Disposable culture tubes, 15 x 85 mm, containing 10/A L-[1-14C]glutamic acid (5 aCi/ml, 40 mM sodium glutamate in 0.1 M potassium phosphate, pH 7.0) were capped with rubber stoppers holding center wells containing 0.2 mt hyamine hydroxide. The reaction was started by injecting 100 /A enzyme solution containing 1 mM 2-aminoethylisothiouronium bromide hydrobromide (AET), 0.2 mM pyridoxal-5'-phosphate (PLP) and 1 mM E D T A into the tube. After a 30-min incubation at 37 °C in a shaking water bath, 0.2 ml of 0.5 N H2SO 4 was injected through the stopper to terminate the reaction. Tubes were shaken for an additional 60 min at 37 °C to allow a complete evolution and absorption of 14CO2 in hyamine. Center wells were then transferred to vials containing 5 ml toluene, 0.3% PPO, and 0.01% POPOP, and 14CO2was determined in a liquid scintillation counter with ca. 85% counting efficiency.

tric focusing (pH 4-7) in which a single protein band with comigrating GAD activity was obtained 8-1~. The purity of GAD preparations was further established from SDS-PAGE which showed two protein bands corresponding to two GAD subunits of 80,000 and 40,000 daltons 10. The purified GAD preparations were then used as antigen for the production of polyclonal antibodies. Briefly, 4 rabbits were given 7 biweekly subscapular injections of 10 or 30/~g of purified GAD each in Freund's complete adjuvant followed by intermittent monthly boosters of 10 to 30~g as previously described 42. Animals were bled from the ear vein one week after the seventh injection. Immunoglobulin fractions were prepared as previously describedn, 42. Serum was first fractionated with 50% ammonium sulphate and the pellet thus obtained was dissolved in and dialysed against 20 mM Tris, 20 mM sodium chloride, pH 7.8 (buffer A) before it was loaded on a DE52 (Whatman) column which had been equilibrated with buffer A. Typically, less than 4 mg of protein was loaded for each ml of resin. After washing with one bed vol. of the equilibration buffer, IgG-enriched fractions were eluted with 5 column vols. of a linear gradient which was made of equal volumes of buffer A and 40 mM Tris, 400 mM sodium chloride, pH 7.8. Fractions containing the highest absorbance at 280 nm were pooled and concentrated on an Amicon stirred cell with PM30 membranes. The final concentration of the buffer was adjusted to 15 mM sodium phosphate, 140 mM sodium chloride, pH 7.4 (PBS). All sera and IgG were stored at -80 °C.

Protein assay

Ouchtertony double immunodiffusion

Protein was assayed by the protein-dye binding method described previously3. Standards containing 1-10 pg of bovine serum albumin were included in each determination. Ten /~g gave an absorbance change of ca. 0.21 OD units at 595 nm.

Double immunodiffusion was performed according to the procedure previously described42 with minor modifications. Immunodiffusion gels were made of 1% SeaKem ME (FMC) agarose in PBS containing 0.02% sodium azide, 1 mM AET, 0.2 mM PLP, pH 7.4. One and 10 #1 of serum and 10 ~1 of a concentrated, partially purified GAD preparation (ca. 5% pure, 5 mg/ml) was placed in the wells. Plates were kept covered in a humidified atmosphere at 4 °C for 24-72 h. Precipitin lines were either visualized directly or after extensive washing and brief staining with 0,05% Coomassie brilliant blue G-250 in 40% methanol, and 7% acetic acid followed by destaining in 7% acetic acid with several changes.

MATERIALS AND METHODS

GAD assay

Polyclonal antibody production GAD was purified ca. 1300-fold to apparent homogeneity by a combination of column chromatographies on D E A E cellulose, hydroxylapatite and gel filtration and preparative polyacrylamide get electrophoresisSA 0. The purified GAD preparations were homogenous as judged from non-denaturing regular (5%) and gradient (3.6-25% and 6-710% ) polyacrylamide gel electrophoresis and narrow-range isoelec-

Immunoelectrophoresis Immunoelectrophoresis was performed in gels containing 1% SeaKem ME agarose, 25 mM Tris, 192 mM glycine, 1 mM AET, 0.2 mM PLP and 0.01% sodium azide, pH 8.4. Ca. 5/A of partially purified GAD preparation (about 5% pure, 5 mg/ml) was applied and electrophoresis performed for 1.5 h at 4 °C with 10 volts/cm constant voltage. The electrophoresis running buffer contained 25 mM Tris, 192 mM glycine, I mM AET, 0.2 mM PLP and 0.5% 2-mercaptoethanol, pH 8.4. One lane was cut from the gel and stained for protein as described for immunodiffusion gels. A parallel lane (1 cm wide, ca. I mm thick) was cut in 0.5 cm lengths and assayed for GAD activity. Slices were macerated in disposable culture tubes containing 125/zl of 200 mM potassium phosphate, 1 mM AET, 0.2 mM PLP and 1 mM EDTA, pH 6.0. After adding 12.5/~1 L-[1-14C]glutamic acid (5 /zCi/ml, 40 mM sodium glutamate), GAD activity was measured. Narrow troughs (ca. 1 mm wide, 10 cm long) were then cut between the remaining lanes in the direction of electrophoresis. Preimmune or anti-GAD serum (100 #l) was placed in the troughs and immunodiffusion carried out for 12-36 h. Gels were then washed extensively with large volumes and several changes of PBS for 1-3 days at room temperature with constant gentle agitation. Precipitin lines were either visualized directly or stained for protein as for immunodiffusion gels. Immunoprecipitation and enzyme inhibition GAD-anti-GAD complexes were precipitated using the heat-killed, formalin-fixed Staphylococcus aureus Cowan Type I (SAC; Bethesda Res. Lab.) procedure previously described 17. Unless stated otherwise, all dilutions were made with, and pellets resuspended in, PBS containing 1 mM AET, 0.2 mM PLP, 1 mM EDTA, pH 7.0. Partially purified GAD preparation (80/~l, about 10% pure, 2 mg/ml) was incubated for 16 h at 4 °C with an equal volume of either anti-GAD serum, preimmune serum, or IgG fractions (10 mg/ml) at dilutions of 8, 40 and 160-fold. Fifteen ~l of SAC was added to each microfuge tube and samples were incubated at room temperature with agitation for 30 min, followed by brief centrifugation. Supernatants and pellets were then assayed for GAD activity. The following controls were performed: no anti-GAD serum, no SAC, same amount

of preimmune serum or IgG instead of anti-GAD serum or anti-GAD IgG, or neither anti-GAD serum nor SAC. Total GAD activity in the supernatant and pellet in the presence of anti-GAD serum or antiGAD IgG compared to preimmune serum or normal IgG indicated the degree of inhibition of enzyme activity by the antiserum.

Dot immunoassay Dot immunoassay was performed according to the procedure previously described 15 with minor modifications. Nitrocellulose (Bio-Rad) discs were made with a paper hole punch to fit snugly into the bottom of standard 96-well microtiter plates. Unless otherwise mentioned, all procedures were performed at room temperature with constant gentle agitation. GAD protein (5-500 ng) from either purified or partially purified (10% pure) GAD preparations were spotted on the discs in 2/~1 vols. and incubated at 37 °C for 4 h. The amount of GAD protein in the partially purified preparations was estimated from % of purity which was calculated from the specific activity based on that of purified GAD preparation, 2.4 units per mg protein as 100%. All subsequent dilutions and washes were with PBS containing 0.05% Tween-20, 0.1% bovine serum albumin, pH 7.4 (buffer B). Discs were washed 3 x for 5 min each before the addition of 100 gl/well of preimmune serum or antiserum (1:100 dilution). After 2 h incubation, discs were washed 3 x and incubated with 100gl of peroxidaseconjugated goat anti-rabbit IgG (Miles-Yeda) diluted 1:200 for one additional hour. Discs were then washed 3 additional times before the addition of the substrate (0.05% 3,3'-diaminobenzidine tetrahydrochloride in 50 mM Tris, pH 7.6 and 0.01% hydrogen peroxide). The reaction was terminated after 2 min by aspiration of the substrate followed by several rapid washes. Enzyme-linked immunoadsorbent assay (ELISA ) Aliquots of 50/~l of purified or partially purified (about 10% pure) GAD preparations containing 50-500 ng of GAD were added to microtiter wells and incubated with constant agitation at room temperature overnight 45. Wells were washed 3 x with 200/~1 of buffer B. Fifty/zl of either 1:100 diluted antiGAD or preimmune serum, culture medium, or ascites fluid was added and incubated for 2 h. The wells

were washed 3 x. After an additional 2 h incubation with peroxidase-conjugated goat anti-rabbit IgG (for polyclonal antibodies from rabbits) or peroxidaseconjugated rabbit anti-mouse IgG (Miles-Yeda) (for monoclonal antibodies from mice) diluted 1:200 with buffer B, wells were washed 3 more times and treated with 100 ~1 of substrate (0.05% 2,2'-azino-di-(3-ethyi-benzothiazolin)-suifonate; ABTS, 0.1 M citric acid, pH 4.2, 0.02% hydrogen peroxide) for 15 min at room temperature. The reaction was terminated by aspirating the substrate and adding 40 ktl 0.0006% sodium azide.

Western immunoblotting test The Western blot transfer technique of Towbin et al. 38 was followed with some modifications. Briefly, 20/~g of partially purified G A D preparation (about 10% pure) was applied to a 10% SDS slab gel and electrophoresed for 4 h at 20 mA. Proteins were then transferred from SDS gel to nitrocellulose sheet by electrophoretic transfer at 25 V for 16 h. The nitrocellulose sheet was then treated with 3% bovine serum albumin at room temperature for 2 h. After rinsing in 0.01 M phosphate-saline buffer containing 0.2% bovine serum albumin 0.05% Tween 20, the nitrocellulose strips were immersed in 0.02 mg/ml of either polyclonal or monoclonal anti-GAD IgG for 2 h at room temperature. For the control experiments, the same amount of IgG isolated from preimmune serum was used instead of anti-GAD IgG. After brief rinsing in the same buffer, the strips were further incubated with peroxidase-labeled goat IgG anti-rabbit lgG (for polyclonal anti-GAD) or peroxidase-labeled rabbit IgG anti-mouse IgG (for monoclonal antiGAD) at 1:100 dilution for 1 h, followed by a brief incubation with peroxidase substrate. Protein markers included fl-galactosidase, 116,500 dalton; phosphorylase b, 94,000 dalton; transferrin, 90,000 dalton; bovine serum albumin, 67,000 dalton; ovalbumin, 43,000 dalton; and carbonic anhydrase, 31,000 dalton.

Monoclonal antibody production Monoclonal antibodies were prepared by procedures previously described12,1s,45 with some modifications. BALB/C mice were immunized by weekly

intraperitoneal injections of 100 ug of partially purified GAD preparation (ca. 50% pure) emulsified in Freund's complete adjuvant for 6 weeks. Blood was collected from the tail vein and serum analyzed for antibodies production by immunodiffusion. When an immunoprecipitin band was seen, spleen cells were harvested, incubated with 0.17 M ammonium chloride at 4 °C for 10 rain to lyse erythrocytes, and fused with a myeloma cell line (P3x63Ag3) at a ratio of 10:1 in 50% polyethyleneglycol 1000 for 4 min. Cells were evenly suspended and distributed into 96-wetl microtiter plates (one drop/well). The next day an additional drop of 2X HAT medium (hypoxanthine 1.36 mg/100 ml, aminopterin 0.036 mg/100 ml, thymidine 0.387 mg/100 ml) was added to each welt. After one week, cells were fed with one drop of HT medium (hypoxanthine, thymidine; no aminopterin) to select for the growth of hybrids. About 17 days after cell fusion, antibody-producing hybrids were detected by an ELISA using pure GAD. Positive clones were subcloned by limiting dilution, rescreened with the pure antigen and grown to confluency in 250-ml bottles. Clones yielding positive responses in an additional ELISA were then injected into the peritoneal cavity of immunosuppressed mice. Ascites fluid was collected about 7-10 days later, retested in the ELISA, and stored at -80 °C.

Demonstration of GAD-anti-GAD IgG complex by Sephadex G-200 column chromatography and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) About 3.55 mg of partially purified GAD preparation (ca. 10% pure) was incubated with 4.52 mg of monoclonal anti-GAD IgG (monoclone 12-24) for 16 h prior to the application to a Sephadex G-200 column, 2.5 x 60 cm. The colurnn was eluted with 50 mM phosphate buffer, pH 7.2 containing 1 mM AET, 0.2 mM PLP and 1 mM EDTA. About 2 ml per fraction was collected. For the control experiments the conditions were the same as described above with the exception that monoclonal anti-GAD IgG was replaced by the same amount of normal mouse IgG. Fractions containing GAD activity from Sephadex G-200 column were analyzed in 10% SDS-PAGE as previously described 10. Phosphorylase b, bovine serum albumin, ovalbumin and carbonic anhydrase were used as molecular weight markers.

lmmunocytochemistry The procedure used for immunocytochemical studies was similar to those described previously2t-23,45. Male rats were perfused through the left ventricle with 500 ml of 4% paraformaldehyde, 0.1% glutaraldehyde, 8.5% sucrose, 0.01% calcium chloride and 0.1 M phosphate, pH 7.4. Brains and retinae were removed. The cerebella were dissected and cut into 0.6 x 0.4 x 0.2 cm blocks. Tissue blocks were then immersed in 2% paraformaldehyde in the same buffer overnight, rinsed with phosphate buffer and immersed in buffered 0.03 M hydroxylamine for 30 min to block excess aldehyde groups. Fifty-gm sections were obtained with a vibratome and immersed for 10 min in normal goat serum diluted 1:100. The sections were divided into 4 groups and incubated for 16 h at 4 °C with the following solutions diluted 1:400 with phosphate buffer: (1) rabbit preimmune serum, (2) rabbit anti-GAD serum, (3) control mouse ascites fluid against myeloma cells, or (4) ascites fluid against rat GAD. Sections were washed twice for 10 min each with phosphate buffer and then incubated for 1 h with peroxidase-conjugated goat anti-rabbit IgG (for polyclonal antibodies) or peroxidase-conjugated rabbit anti-mouse IgG (for monoclonal antibodies) diluted 1:200. After washing with phosphate buffer, sections were briefly refixed in 2.5% glutaraldehyde for 10 min, washed and incubated for 3-5 min with 0.05% 3,3'-diaminobenzidine tetrahydrochloride in 0.05 M Tris, pH 7.6, containing 0.01% hydrogen peroxide. After washing, sections were further fixed in 0.05% OsO4 for 2 min, washed with distilled water, mounted with glycerin, sealed with fingernail enamel, and examined under the light microscope. RESULTS

Immunodiffusion Ouchterlony double immunodiffusion tests were performed with partially purified rat GAD and various rabbit sera as shown in Fig. 1 where precipitin bands were visualized directly without protein staining. The center well.contained 10 ~1 of partially purified GAD preparation (about 5% pure, containing 50/~g total protein). Ten gl (well 1) or one #1 (well 2) of serum from an animal that had received 7 injections of 30/~g each gave a sharp, single precipitin

5

Fig. 1. Immunodiffusiontest with partially purified GAD and anti-GAD serum. The center well contained 10/A of partially purified GAD (about 5% pure, containing 50/~gtotal protein). Wells 1 and 2 contained 10 and 1/tl, respectively,of serum from a rabbit that had received a total of 210 ~g of purified GAD. Wells 3 and 4 contained 10 and 1~1, respectively,of serum fro~ a rabbit that had received a total of 70 #g of purified GAD. Wells 5 and 6 contained 10/Aof preimmune serum from the respective rabbits.

band within 24 h. No precipitin band was detected with preimmune serum (wells 5 and 6) or with sera from a rabbit which had received an equal number of injections of 10/~g each (wells 3 and 4). Rabbits receiving 10/~g per injection later also developed antibodies against GAD after two additional injections of 10~g each (data not shown).

lmmunoelectrophoresis Immunoelectrophoresis performed with partially purified (5% pure) rat GAD and rabbit antiserum is shown in Fig. 2. The distribution of GAD activity in the agarose gel is indicated at the top. A single peak of activity was centered around 1.8 cm from the origin. The protein profile of the sample loaded on the gel shown at the bottom of Fig. 2 indicated the complex protein pattern of the crude GAD preparation. The protein pattern after electrophoresis, immunodiffusion, extensive washing and Coomassie staining is shown in the center of Fig. 2 in which a single, smooth immunoprecipitin arc corresponding to the position of GAD activity was seen. No precipitin band was seen when preimmune serum was substituted for anti-GAD serum (not shown).

t~

b6 ×5

b4 "3

m i B

m

O

'-2 E 1

e~ o

I

0

I

!

I

l

I

1

2

3

4

cm from origin

Fig. 2. Immunoelectrophoresis of partially purified GAD and anti-GAD serum. Ca. 5 ~1 of partially purified GAD preparation (about 5% pure, 5 mg/ml) was applied and electrophoresis was performed as described in Materials and Methods. After eleetrophoresis, one lane was cut into 0.5 cm per slice and assayed for GAD activity as shown at the top. GAD activity was expressed as cpm/slice/30 min. The protein pattern of the gel after electrophoresis, immunodiffusion and extensive washing in PBS is shown in the middle. The protein profile of the sample applied to the gel is shown at the bottom.

lmmunoprecipitation and enzyme inhibition When the amount of GAD and SAC was kept constant, increasing amounts of antiserum decreased GAD activity in the supernatant (filled circles) with a concomitant increase in GAD activity in the pellet (open circles). The data in Fig. 3 are corrected for the effects of preimmune serum, which caused about 25% inhibition of GAD activity at the highest amount tested. Hence, no higher concentration of antiserum was tested. At this concentration (1:8 dilution), about 40% of GAD activity could be recovered as G A D - a n t i - G A D complex in the pellet. No significant inhibition of GAD activity was observed by normal IgG at a concentration of 0.625 mg/ml (square). GAD activity was recovered quantitatively as immunoprecipitate by anti-GAD IgG at the same concentration of normal IgG (supernatant: filled circles; pellet: open circles) (Fig. 4). No immunoprecipitate was obtained with normal IgG. Dot immunoassay Two different GAD preparations, one pure and another one about t0% pure, were used in dot immunoassay (Fig. 5). Nitrocellulose discs in the top row were spotted with (left to right) 500, 160, 50, 16 and 5 ng of pure GAD. About 10 × as much protein as pure GAD from 10% pure GAD preparation were spotted on the middle row. As little as 16 ng of GAD could be detected. The intensity of reaction product

oeo E0

100,

0

# 60 O

~ 4o
•~ 5 0 o

I 0.~ 2.5 Volume Antiserum, ~1

I 10

< o <

",,

° I

Fig. 3. Immunoprecipitation of GAD with anti-GAD serum. Eighty pl of partially purified GAD (about 10% pure, 2 mg/ml) was incubated for 16 h at 4 °C with an equal volume of either anti-GAD serum or p r c i m m u n e s e r u m at8-, 40- and 160-fold dilution. The immune complexes were then precipitated with SAC as described in Materials and Methods. GAD activity after immunoprecipitation was measured in the supernatant (filled circles) and the pellet (open circles). GAD activity was expressed as % of control which was the GAD activity in the presence of preimmune serum. The amount of serum used was expressed as the volume of the undiluted s e r u m in pl.

0

O.lli6

v ~

tie, la

i

I

2.5

t0

Fig. 4. Immunoprecipitation of GAD with anti-GAD IgG. The conditions were the same as those described in Fig. 3 with the exception that serum was replaced by IgG preparations. The amount of anti-GAD IgG was expressed as the volume of the undiluted anti-GAD IgG (I0 rag/rot) in/~1. GAD activity after immunoprecipitation was m©a~tLred in the supernatant (filled circles) and tl~e pellet (open cir~s). Total GAD activity of the incubation mixture is the sum of the activity in the supernatant and pellet (squares).

ELISA test of polyclonal antibody

Fig. 5. Dot immunoassay with GAD and anti-GAD serum. Five to 500 ng of protein from purified GAD preparation (top row) or 10 x that amount of protein from 10% pure GAD preparation (middle row) were spotted on nitrocellulose discs. Five thousand and 500 ng ~f 10% pure GAD were spotted on the bottom row as control. After treatment with either anti-GAD serum (1:100; top and middle rows) or preimmune (bottom row) serum (1:100), immuno complexes were detected with peroxidase-conjugated goat anti-rabbit IgG (1:200) using diaminobenzidine and I-I202 as the substrates.

Polyclonal antibodies against G A D were further tested by E L I S A test (Fig. 6). Wells were coated with various amounts of G A D protein either from purified preparation (top row) or partially purified preparation (10% pure, middle bottom rows). Fifty ng of G A I ) could be detected by this method. Again, the intensity of the reaction product is roughly proportional to the amount of antigen present and appears to be independent of the purity of the G A D preparations as shown in Fig. 6 where a similar pattern was obtained either with pure (top row) or partially purified (middle row) G A D preparations. The specificity of a n t i - G A D serum was further indicated by the lack of reaction between partially purified G A D preparations and preimmune serum (bottom row).

is shown to be roughly proportional to the amount of antigen spotted regardless of the purity of G A D preparations. The specificity of this reaction was further indicated in the bottom row where controls yielded no reaction product. The controls contained 500 and 50 ng of protein of 10% pure G A D preparation and were treated with preimmune serum instead of anti-GAD serum (bottom row).

Fig. 6. Enzyme-linked immunoadsorbent assay with anti-GAD serum. Microtiter wells were coated with 50-500 ng of GAD protein from either pure (top row) or partially purified preparations (about 10% pure; middle and bottom rows). After treatment with anti-GAD serum (1:100; top and middle rows) or preimmune serum (1:100; bottom row), immune complexes were detected with peroxidase-conjugated goat anti-rabbit IgG (1:200) using ABTS and H20 2 as the substrates.

Fig. 7. Western transfer tests with polyclonal and monoclonal GAD antibodies. About 20/tg of partially purified GAD preparation (10% pure) was applied to 10% SDS slab gel and electrophoretically transferred to the nitrocellulose sheet. Lane A: stained with polyclonal anti-GAD IgG showing two protein bands corresponding to the position of GAD subunits, namely, 80,000 dalton and 40,000 dalton. Lane B: stained with monoclonal anti-GAD IgG (12-24) showing a protein band corresponding to the position of the GAD, 80,000-dalton a-subunit. The numbers on the left represent the positions of standard molecular weight markers in kdaltons.

Western immunoblotting test When the partially purified G A D (about 10% pure) preparation was analyzed by SDS-PAGE, many protein bands were stained with Amido Black (data not shown). A similar protein pattern was obtained after the gel was transferred to nitrocellulose sheets. However, only two protein bands corresponding to the same position as G A D subunits - - a molecular weight of 80,000 and 40,000 daltons were stained with polyclonal anti-GAD IgG and only the 80,000 dalton subunit was stained with monoclonal anti-GAD IgG (12-24; Fig. 7). No protein band was

stained in the control experiments where polyclonal anti-GAD IgG and monoclonal anti-GAD I g G were replaced by the same amount of normal rabbit and mouse IgG respectively.

ELISA test of monoclonat antibody In the initial screening, among 960 wells about 420 wells showed a positive reaction in ELISA test using purifed G A D as antigen (data not shown). Fortyeight of these revealed strong reactions and were further cloned by limiting dilution. Following screening and subcloning, 4 of the single clones 12-1, 12-t0, 12-24, 12-33, were used for ascites fluid production. An ELISA test using pure G A D and ascites fluid diluted 1:200 is shown in Fig. 8 (positive controls, 10A and 10B; negative controls, l l A and liB). Ascites fluids used in wells 7A, 7B, 7 F - 7 H , 8A, 8E, 8F, 8H, 9B, 9G, 9H, 10C, 10E, l l F , 12A, 12D-12F and 12H contained strong antibodies against GAD.

Demonstration of GAD-anti-GAD complex by gel filtration and SDS-PA GE

IP

Fig. 8. Enzyme-linked immunoadsorbent assay of monoclonal antibodies to GAD. Two hundred ng of purified GAD was used as antigen in each well. The details of condition of incubation with primary antibodies, peroxidase-conjugated second antibodies and peroxidase reaction were described under Materials and Methods. Mouse anti-GAD serum was used as positive control (10A and 10B). Ascites fluid from mice receiving myeloma cells was used as negative control (llA and llB). Wells in rows A and B, C and D, E and F, andG and H contained ascites fluid derived from clone 12-1, 12-10, 12-24 and 12-33, respectively. Wells, e.g. 7A, 7B, 7F-H, 8A, 8E, 8F, 8H, 9B, 9G, 9H, 10C, 10E, llF, 12A, 12D-F and 12H showing deep peroxidase reaction product, contained monoclonal antibodies against GAD.

Monoclonal antibodies against G A D were further characterized by Sephadex G-200 gel filtration column chromatography and SDS-PAGE. When the incubation mixture containing G A D and monoclonal anti-GAD IgG 12-24 was applied to Sephadex G-200 column, the elution position of G A D was found to shift from fraction 90 to fraction 82 suggesting the formation of a higher molecular weight complex, presumably G A D - a n t i - G A D IgG (Fig. 9). Similar results were obtained with monoclonal a n t i - G A D IgG 12-1, 12-10 and 12,331 When the peak fraction and one fraction on each side of the G A D activity peak from the Sephadex G-200 column were examined by SDS-PAGE, it was found that the fractions deriving from G A D and monoclonal anti-GAD IgG mixture contained much higher amounts of IgG than the comparable fractions deriving from G A D and normal mouse IgG mixture (Fig. 10) further suggesting that the complex is G A D - a n t i - G A D IgG complex.

Immunocytochemical localization In addition to the ELISA test, the demonstration of G A D - a n t i - G A D complex by gel filtration column chromatography and SDS-PAGE, and Western immunoblotting test, monoclonal antibody against

~°

I

E

• o~O t

i

40

t

•,/ i

°~e

~o I

I

I

I

I

I

i

I

I

• o• o 0 "~ i

i

/ ; ....-..a. °'i'-*l "°, 50

"~:

60

/

1-o .

. . . . . .

70

O0

90

I00

l

X

exhibited small punctate reaction product similar to those stained with the polyclonai antibodies (Fig. l l C ) . In addition to the punctate staining a r o u n d Purkinje cell bodies, n u m e r o u s G A D - p o s i t i v e stainings were seen in the m o l e c u l a r as well as granual layer corresponding to the innervations of stellate cells on the dendritic processes of Purkinje cells in the molecular layer and the Golgi cells and their processes in the granular layer, respectively. In the retina, both polyclonal and monoclonal antibodies against G A D also showed similar staining pattern, namely, the staining was mainly concentrated in the inner plexiform layer (IPL) and to a lesser extent in the inner nuclear layer (INL) (Fig. l i E and F). Note that a clear multistriated staining p a t t e r n is seen in the IPL. Control sections which were treated with p r e i m m u n e

"

°-~o..;,o o TIO

120

Fraction Number

Fig. 9. Demonstration of GAD-anti-GAD IgG complex by Se-' phadex G-200 gel filtration column chromatography. A: 3.55 mg of partially purified GAD (about 10% pure) was incubated with 4.52 mg of monoclonal anti-GAD IgG (monoclone 12-24) for 16 h prior to the application to a Sephadex G-200 column, 2.5 x 60 cm. The peak fraction was fraction 82. About 2 ml/ fraction was collected. B: the conditions were the same as A except that monoclonal anti-GAD IgG was replaced by normal mouse IgG and the amount of GAD and IgG was changed to 2.72 and 3.46 mg, respectively. The ratio of GAD to IgG was the same in A and B. The peak fraction was fraction 90. A shift of elution position of GAD towards higher molecular weight (from fraction 90 to 82) is obtained in A.

G A D was further characterized immunohistochemically in the areas where G A B A e r g i c neurons and their processes have been well established, e.g. cerebellum and retina. Polyclonal antibodies against G A D were included in a parallel e x p e r i m e n t as a positive control. In the cerebellum, discrete, punctate deposits of reaction product were seen surrounding Purkinje cell bodies (P) in sections treated with antiG A D serum (Fig. l l A ) . Sections treated with preimmune serum or control ascites fluid derived from myeloma cell which was negative in the E L I S A screening showed no staining or only diffuse background staining (Fig. l l B and D). Sections treated with G A D monoclonal antibodies from ascites fluid

Fig. 10. Analysis of GAD-containing fractions from Sephadex G-200 column on SDS-PAGE. Approximately the same amount of protein, 10/Agfrom each fraction, was applied to the gel. In addition, monoclonal anti-GAD IgG, 50/~g, and standard molecular weight markers, 40/~g each, were included in the gel. Lanes 1-3: fractions 88, 82 and 76 from Sephadex G-200 column (Fig. 9A) containing GAD and anti-GAD IgG (monoclone 12-33). Lane 4: Molecular weight markers; phosphorylase b (94,000); bovine serum albumin (67,000); ovalbumin (43,000); carbonic anhydrase (30,000). Lane 5: partially purified GAD sample used in the experiment. Lane 6: monoclohal anti-GAD IgG (monoclone 12-24). Lanes 7-9: fractions 96, 90 and 86 from Sephadex G-200 column containing GAD and normal mouse IgG (Fig. 9B). Lanes 10-12: fractions 90, 82 and 76 from Sephadex G-200 column containing GAD and anti-GAD IgG (monoclone 12-24).

Fig. 11. Immunocytochemical staining of rat cerebellum and retina using polyclonal and monoclonal antibodies against GAD. A: the rat cerebellar section was stained with polyclonal antibodies. The punctate reaction product of anti-GAD is found surrounding Purkinje cell bodies (large arrowheads) in the molecular and granular layers (small arrowheads); x 1200. B: a control section taken from the rat cerebellum and stained with preimmune rabbit serum shows no reaction product; x 1200. C: the rat cerebellar section was stained with one of the monoclonal antibody (12-24). The punctate reaction product is also seen surrounding the Purkinje cell bodies (large arrowheads) and in the molecular and granular layers (small arrowheads); x940. D: a control section taken from the rat cerebellum and stained with control ascites shows no reaction product; x 1200. E: the rat retinal section was stained with one of the monocional antibodies (12-24). Reaction product is shown mainly in the inner plexiform layer (IPL) and some in the inner nuclear layer (INL). Note that a clear muitistriated staining pattern is seen in the IPL (arrowheads). The outer nuclear layer (ONL) is unstained; x870. F: the rat retinal section was stained with polyclonal antibodies. Reaction product is also seen in the IPL and INL but not in the ONL. Note that in the IPL the multistriated staining pattern is also found (arrowheads); x750. G: a control section taken from the rat retina and stained with preimmune rabbit serum shows no reaction product; x900.

11 serum or control ascites fluid showed no specific staining (Fig. l l G ) . This observation is in good agreement with the earlier observation made in this laboratory21,22 as well as others 1,39 and is compatible with the large body of evidence that some amacrine cells in the mammalian retina are GABAergic 4,39,44. DISCUSSION In this communication, we described an approach which differed greatly from those described by Oertel et al. 28 for the production of polyclonal antibodies against rat brain GAD. The approach we used is similar to our earlier works of mouse brain GAD40.42, namely, to purify brain GAD to apparent homogeneity and use the purified GAD preparations as antigen for the production of anti-GAD serum while Oertel et al. 28 used partially purified GAD preparations to prepare 'trapping' antiserum, followed by the isolation of G A D - a n t i - G A D precipitate using the 'trapping' antiserum and finally used the precipitate as antigen for the production of anti-GAD serum. Although Oertel et al. 28 claimed that they had obtained specific polyclonal antibodies against rat brain GAD, no convincing evidence was actually presented to support their claim. The main criteria that they had was the precipitating characteristics of antiserum. However, the precipitating properties of antiserum simply indicate that the anti-serum does contain antibodies against GAD but it does not prove its specificity. In fact, from their own results26-2s it seems likely that their anti-GAD serum may also contain antibodies against GABA-transaminase (GABA-T). This notion is based on the following observations. First of all, the purity of the GAD preparations used as antigen for the generation of the so called 'trapping' antiserum and for the subsequent immunoprecipitation experiments to yield the immune complex for the production of the 'specific' antiserum is rather low, probably less than 30% and 15% pure, respectively28. Hence, several immunoprecipitin lines were obtained as it is expected 2s. Secondly, they relied on the specificity of y-acetylenic G A B A to determine which immunoprecipitin line to use for the generation of 'specific' antiserum. However, y-acetylenic GABA is at least 3 x more potent to GABA-T than GAD16. Jung et al.16 showed that the amount of protein bound with ~[SH]acetylenic GABA in the

brain homogenate could be accounted for by the amount of GABA-T present. Therefore, the immunoprecipitin lines that Oertel et al. 2~ identified with ~'-acetylenic GABA might also contain G A B A - T anti-GABA-T complex in addition to GAD-antiGAD complex. Furthermore, it is also likely that the precipitin lines contain other proteins in addition to antigen-antibody complex because of non-specific absorption occurring during precipitation process. Thirdly, Oertel et al. 26 reported that when partially purified GAD preparation labeled with [2-3H]-y-acetylenic GABA was analyzed by immunoelectrophoresis and two-dimensional electrophoresis, two radioactive protein bands corresponding to mol. wts. of 54,000 + 1,500 and 58,000 + 1,500 were obtained. Although the subunit structure of rat brain GAD is still not fully characterized, none of the reports using purified GAD preparations agrees with their reSUITS2,10,24. The above results reported by Oertel et al. 26 regarding the subunit structure of GAD are rather similar to our previous report for GABA-T in which we have shown that GABA-T consists of two non-identical subunits with mol. wts. of 53,000 and 58,000 dalton34, 40. Finally, the immunohistochemical staining of GAD reported by Oertel et al. 27 also is different from those reported by Saito et al. 32 and McLaughlin et a1.25 using the well-characterized antimouse GAD serum and the present studies using antibodies against the purified rat brain GAD, in that the former showed strong staining in the cell body, e.g. Purkinje cells in the cerebellum, while the latter showed strong staining in the terminal and weak or no staining in the cell body. It is interesting to note that anti-GABA-T serum was also reported to give strong staining in the cell body 1,7. From the above brief discussion, it seems reasonable to conclude that anti-serum produced by Oertel et al. 2s using partially purified GAD preparations may not be specific for GAD and probably also contain antibodies against GABA-T. Since the specificity of anti-GAD serum produced by our procedures depend solely on the purity of GAD preparations used as antigen, we have characterized the purified GAD preparations extensively prior to their use. The purified GAD preparations have been shown previously to be homogenous as judged by a variety of analytical techniques such as electrophoresis on 5% regular polyacrylamide gels,

12 3.6-25% and 6-10% gradient gels and isoelectric focusing in pH 4-7 agarose gels. In all 4 non-denaturing gel systems, GAD preparations migrated as a single protein band and the position of the protein band coincided with the GAD activitys-t~/. Furthermore, in the denaturing gel, SDS-PAGE, GAD dissociated into two non-identical subunits with mol. wts. of 40,000 and 80,000 dalton, which is in excellent agreement with a mol. wt. of 120,000 dalton for the native enzyme 10. The use of~tg instead of mg quantities of protein as antigen, as used in the conventional method of immunizing animals for the production of antibody, is essential because of the scarcity of the purified enzymes from the nervous tissue. Furthermore, the chance of producing antibodies against trace impurities, which might be associated with the highly purified preparations and escape detection by sensitive physical and chemical techniques, is much less because of the small amount ~ g quantities) of antigen used in the immunization of animals. This technique has been successfully used in our laboratory for the production of polyclonal antibodies to a variety of antigens such as mouse brain GAD and GABA-transaminase33,40, 42, choline acetyltransferase (CHAT) from Torpedo 5.6, GAD 36 and ChAT 37 from catfish brain, GAD and cysteine sulfinic acid decarboxylase from bovine brain 41, neurofilament protein 20 and coated vesicle protein 13. The specificity of the polyclonal antibodies described here was established based on the following criteria. First of all, a single sharp precipitin band was seen in Ouchterlony double immunodiffusion using a partially purified GAD preparation. Secondly, a single, sharp, symmetrical precipitin arc was obtained with comigrating GAD activity in immunoelectrophoresis using a crude GAD preparation. In both cases the presence of only one protein band corresponding to the enzyme activity further attests to the purity of the antigen and the specificity of the immunologic response. Thirdly, the specific formation of G A D - a n t i - G A D immune complexes between antiGAD serum or anti-GAD IgG and GAD was demonstrated in the SAC precipitation experiments. The quantitative recovery of GAD activity in the pellet is additional evidence for the formation of GAD-antiGAD complexes. Fourthly, in the Western immunoblotting test, anti-GAD recognized only GAD subunits, the 40,000- and 80,000-dalton subunits in a rela-

tively crude GAD preparation which contains many other proteins further indicating the high degree of specificity of anti-GAD serum. Fifthly, in both ELISA and dot immunoassays, the intensity of reaction product appears to be roughly proportional to the amount of GAD present regardless whether the preparation was pure or only partially purified, thus further suggesting that the antiserum specifically recognizes GAD. Although the hybridoma technique does not require a pure antigen preparation in order to obtain a specific monoclonal antibody, a specific screening probe is essential. In the present study, the purified GAD preparations were used in the ELISA tests for screening anti-GAD producing clones. The specificity of monoclonal antibodies was established by further ELISA tests using pure GAD as antigen for single clones and for ascites fluid obtained from mice which had been injected with anti-GAD producing single clones. The direct evidence of GAD-antiGAD complex formation comes from the gel filtration and SDS-PAGE experiments. The former shows that GAD activity shifted towards a higher molecular weight position when GAD was incubated with monoclonal anti-GAD IgG suggesting a complex formation. The SDS-PAGE analysis further indicated that the complex was indeed enriched in IgG. Perhaps the most convincing evidence regarding the specificity of GAD monoclonal antibody is the Western immunoblotting test, in which the monoclonal antibody reacts only with the 80,000-dalton GAD subunit in a crude GAD preparation. It is interesting to note that GAD polyclonal antibodies recognize both the 80,000- and 40,000-dalton subunits while monoclonal antibody only recognizes the 80,000-dalton subunit. This is similar to benzodiazepine receptors in that some monoclonal antibodies against the receptor also just recognize one subunit, the 50,000dalton ct-subunit, but not the 55,000-dalton fl-subunit TM. Further characterization of polyclonal and monoclonal antibodies comes from the immunocytochemical demonstration of G A D in rat cerebellum and retina. In the cerebellum, discrete, punctate reaction products were found to be associated with Purkinje cell bodies and dendritic processes. In the retina, GAD-positive staining is most concentrated in the IPL. The distribution of GAD-staining neurons and

13 their processes agrees well with the known distribution of G A B A e r g i c neurons in the cerebellum and retina4,2t,22,25,32,39,44 further supporting the specificity

plication for immunohistochemical studies has been demonstrated.

of the polyclonal as well as monoclonal a n t i - G A D described here. In summary, we have r e p o r t e d the p r e p a r a t i o n and characterization of new polyclonal and monoclonal antibodies against rat G A D . The specificity of the antibodies has been extensively tested and their ap-

ACKNOWLEDGEMENTS

REFERENCES

13 Garbern, J.-Y. and Wu, J.-Y., Purification and characterization of clathrin from bovine brain, J. Neurochem., 36 (1981) 602-612. 14 H~iring, P., Stahli, C., Schoch, P., Takacs, B., Staehelin, T. and M6hler, H., Monoclonal antibodies reveal structural homogeneity of y-aminobutyric acid/benzodiazepine receptors in different brain areas, Proc. Natl. Acad. Sci. U.S.A., 82 (1985) 4837-4841. 15 Hawkes, R., Niday, E. and Gordon, J., A dot-immunobinding assay for monocional and other antibodies, Anal. Biochem., 119 (1982) 142-147. 16 Jung, M.J., Lippert, B., Metcalf, B.W., Schechter, P.J., B6hlen, P. and Sjoerdsma, A., The effect of 4-aminohex-5ynoic acid (y-acetylenic GABA, 7-ethynyl GABA) a catalytic inhibitor of GABA transaminase, on brain GABA metabolism in vivo, J. Neurochem., 28 (1977) 717-723. 17 Kessler, S., Rapid isolation of antigens from cells with a Staphylococcal protein A-antibody adsorbent: parameters of the interaction of antibody-antigen complexes with protein A, J. lmmunol., 115 (1975) 1617-1624. t8 Kfhler, G. and Milstein, C., Continuous cultures of fused cells secreting antibody of predefined specificity, Nature (London), 256 (1975) 495-497. 19 Kravitz, E.A., Acetylcholine, y-aminobutyric acid and glutamic acid: physiological and chemical studies related to their roles as neurotransmitter agents. In G.C. Quarton, T. Melnechuk and F.O. Schmidt (Eds.), The Nearosciences, Rockefeller University Press, New York, 1967, pp. 433-444. 20 Lasek, R.J. and Wu, J.-Y., Immunochemical analysis of the proteins comprising myxicola (10 rim) neurofilaments, Soc. Neurosci. Abstr., 2 (1976) 40. 21 Lin, C.-T., Li, H.-Z. and Wu, J.-Y., Immunocytochemical localization of L-glutamate decarboxylase, gamma-aminobutyric acid transaminase, cysteine sulfinic acid decarboxylase, aspartate aminotransferase and somatostatin in rat retina, Brain Research, 270 (1983) 273-283. 22 Lin, C.-T., Song, G.-X. and Wu, J.-Y., Ultrastructural demonstration of L--glutamate decarboxylase and cysteinesulfinic acid decarboxylase in rat retina by immunocytochemistry, Brain Research, 331 (1985) 71-80. 23 Lin, C.-T., Song, G.-X. and Wu, J.-Y., Is taurine a neurotransmitter in rabbit retina?, Brain Research, 337 (1985) 293-298. 24 Maitre, M., Blinderman, J.-M., Ossola, L. and Mandel, P., Comparison of the structures of L-glutamate decarboxylase from human and rat brains, Biochem. Biophys. Res. Commun., 85 (1978) 885-890.

1 Barber, R. and Sito, K., Light microscopic visualization of GAD and GABAtF in immunocytochemical preparations of rodent CNS. In E. Roberts, T.N. Chase and D.B. Tower (Eds.), GABA in Nervous System Function, Raven, New York, 1976, pp. 113-132. 2 Blinderman, J.-M., Maitre, M., Ossola, L. and Mandel, P., Purification and some properties of L-glutamate decarboxylase from human brain, Eur. J. Biochem., 86 (1978) 143-152. 3 Bradford, M.M., A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Biochemistry, 72 (1976) 248-254. 4 Brandon, C., Lam, D.M.K. and Wu, J.-Y., The y-aminobutyric acid system in rabbit retina: localization by immunocytochemistry and autoradiography, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 3557-3561. 5 Brandon, C. and Wu, J.-Y., Electrophoretic and immunochemical characterization of choline acetyltransferase from Torpedo, Soc. Neurosci. Abstr., 3 (1977) 404. 6 Brandon, C. and Wu, J.-Y., Purification and properties of choline acetyltransferase from Torpedo californica, J. Neurochem., 30 (1978) 791-797. 7 Chan-Palay, V., Wu, J.-Y. and Palay, S.L., Immunocytochemical localization of GABA-transaminase at cellular and ultrastructural levels, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 2067-2071. 8 Denner, L.A., Lin, C.-T., Song, G.-X. and Wu, J.-Y., Multiple forms of rat brain L-glutamate decarboxylase: purification and immunochemical characterization, Fed. Proc. Fed. Am. Soc. Exp. Biol., 42 (1983) 2008. 9 Denner, L.A., Lin, C.-T., Song, G.-X. and Wu. J.-Y., Characterization, purification and immunochemical studies of multiple forms of rat brain L-glutamate decarboxylase, Soc. Neurosci. Abstr., 9 (1983) 1478. 10 Denner, L.A., Wei, S.C., Lin, H.S. and Wu, J.-Y., Purification and characterization of L-glutamate decarboxylase from whole rat brain, Brain Research, submitted. 11 Fahey, J.L., Chromatographic separation of immunoglobulins. In C.A. Williams and M.W. Chase (Eds.), Methods in Immunology and Immunochemistry, Vol. L Academic Press, New York, 1967, pp. 321-332. 12 Galfre, G., Howe, S.C., Milstein, C., Butcher, G.W. and Howard, J.C., Antibodies to major histocompatibilityantigens produced by hybrid cell lines, Nature (London), 266 (1977) 550-552.

This w o r k was s u p p o r t e d by Grants NS-20922, NS20978 and EY-05397 from the National Institutes of Health, U . S . A . The authors thank W . M . H u a n g and D i a n e Evans for their skillful technical assistance and Pat Gering for her excellent secretarial assistance.

14 25 McLaughlin, B.J., Wood, J.G., Saito, K., Barber, R., Vaughn, J.E., Roberts, E. and Wu, J.-Y., The fine structural localization of glutamate decarboxylase in synaptic terminals of rodent cerebellum, Brain Research, 76 (1974) 377-391. 26 Oertel, W.H., Schmechel, D.E., Daly, J.W., Tappaz, M . L and Kopin, l.J., Localization of glutamate decarboxylase on line-immunoelectrophoresis and two-dimensional electrophoresis by use of the radioactive suicide substrate [2-3H]-7-acetylenic GABA, Life Sci., 27 (1980) 21332141. 27 Oertel, W.H., Schmechel, D.E., Mugnaini, E., Tappaz, M.L. and Kopin, I.J., Immunocytochemical localization of glutamate decarboxylase in rat cerebellum with a new antiserum, Neuroscience, 6 (1981) 2715-2735. 28 Oertel, W.H., Schmechel, D.E., Tappaz, M.L. and Kopin, I.J., Production of a specific antiserum to rat brain glutamic acid decarboxylase by injection of an antigen-antibody complex, Neuroscience, 6 (1981) 2689-2700. 29 Roberts, E., GABA in nervous system function - - an overview. In D.B. Tower (Ed.), The Nervous System, Vol 1., The Basic Neurosciences, Raven Press, New York, 1975, pp. 541-552. 30 Roberts, E., New directions in GABA research. I. Immunocytochemical studies of GABA neurons. In P. K0rgsgaard-Larson, J. Scheel-Kruger and H. Koford (Eds.), GABA-Neurotransmitters: Pharmacochemical, Biochemical and Pharmacological Aspects, Munksgaard, Copenhagen, 1979, pp. 28-45. 31 Roberts, E. and Kuriyama, K., Biochemical-physiological correlations in studies of the 7-aminobutyric acid system, Brain Research, 8 (1968) 1-35. 32 Saito, K., Barber, R., Wu, L-Y., Matsuda, T., Roberts, E. and Vaughn, J.E., Immunohistochemical localization of glutamic acid decarboxylase in rat cerebellum, Proc. Natl. Acad. Sci. U.S.A., 71 (1974) 269-273. 33 Saito, K., Schousboe, A., Wu, J.-Y. and Roberts, E., Some immunochemical properties and species specificity of GABA-a-ketoglutarate transaminase from mouse brain, Brain Research, 65 (1974) 287-296. 34 Schousboe, A., Wu, J.-Y. and Roberts, E., Subunit structure and kinetic properties of 4-aminobutyrate-2-ketoglutarate transaminase purified from mouse brain, J. Neurochem., 23 (1974) 1189-1195. 35 Snyder, S.H., G A B A in nervous system function - - an overview. In D.B. Tower (Ed.), The Nervous System, Vol.

I., The Basic Neurosciences, Raven, New York, 1975, pp. 355-361. 36 Su, Y.Y.T., Wu, J.-Y. and Lam, D.M.K., Purification of L-glutamic acid decarboxylase from catfish brain, J. Neurochem., 33 (1979) 169-179. 37 Su, Y.Y.T., Wu, J.-Y. and Lam, D.M.K., Purification and some properties of choline acetyltransferase from catfish brain, J. Neurochem., 34 (1980) 438-445. 38 Towbin, H., Staehelin, T. and Gordon, J., Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications, Proe. Natl. Acad. Sci. U.S.A., 76 (1979) 4350-4354. 39 Vaughn, J.E., Famiglietti, E.V., Barber, R.P., Saito, K., Roberts, E. and Ribak, C.E., GABAergic amacrine cells in rat retina: immunocytochemical identification and synaptic connectivity, J. Comp. Neurol., 197 (1981)113-127. 40 Wu, J.-Y., Purification, characterization and kinetic studies of GAD and GABA-T from mouse brain. In E, Roberts, T.N. Chase and D.B. Tower (Eds.), GABA in Nervous System Function, Raven, New York, 1976, pp. 7-60. 41 Wu, J.-Y., Purification and characterization of cysteic acid and cysteine sulfinic acid decarboxylase and t,-glutamate decarboxylase from bovine brain, Proc. Natl. Acad. Sci. U.S.A., 79 (1982) 4270-4274. 42 Wu, J.-Y., Preparation of glutamic acid decarboxylase as immunogen for immunocytochemistry. In A.C. Cuello (Ed.), Neuroimmunocytochemistry, IBRO Handbook Series: Methods in the Neurosciences, Wiley, Sussex; 1983, pp. 159-191. 43 Wu, J.-Y., lmmunocytochemical identification of GABAergic neurons and pathways. In L. Hertz, E. Kvamme, E.G. McGeer and A. Schousboe (Eds.), Glutamine, Glutamate and G A B A in the Central Nervous System, Liss, New York, 1983, pp. 161-176. 44 Wu, J.-Y., Brandon, C., Su, Y.Y.T. and Lam, D.M.K., Immunocytochemical and autoradiographic localization of GABA system in the vertebrate retina, Mol. Cell, Biochem., 39 (1981) 229-238. 45 Wu, J.-Y., Lin, C.-T., Brandon, C., Chan, D.-S., MOhter, H. and Richards, J.G., Regulation and immunocytochemical characterization of GAD. In S. Palay and V. Chan-Palay (Eds.), Cytochemical Methods in Neuroanatomy, Liss, New York, 1982, pp. 279-296. 46 Wu, J.-Y., Matsuda, T. and Roberts, E., Purification and characterization of glutamate decarboxylase from mouse brain, J. Biol. Chem., 245 (1973) 3029-3034.