-r-AMINOBUTYRIC iiCID IN BRAIN: ITS FORMATION FROM

GLUTAMIC ACID*
BY EUGENE ROBERTS AND SAM FRANKEL
(From the Division of Cancer Research, Washington University School of Medicine,
St. Louis)

(Received for publication, June 3, 1950)

Relatively large quantities of an unidentified ninhydrin-reactive mate-

rial were found in numerous two-dimensional paper chromatograms of pro-
tein-free extracts of fresh mouse, rat, rabbit, guinea pig, human, and frog
brains. At most, only traces of this material were found in a large num-
ber of extracts of many other normal and neoplastic tissues and in urine
and blood. In the present experiments and in the following paper (l),

Downloaded from www.jbc.org by guest, on June 7, 2010

the unknown compound in brain extract is conclusively identified as y-
aminobutyric acid. It is shown that this compound can be formed from
glutamic acid by homogenates, washed residues, and acetone powders of
brain. A preliminary report of some of these findings has been made (2).
EXPERIMENTAL

Chromatography.-One-dimensional and two-dimensional chromatograms

were made by the descending method as outlined by Consden et al. (3) and
extended by Dent (4, 5).
Separation of Unknown Material-Mice were killed by dislocation of the
cervical vertebrae, and the brains were removed immediately and frozen
in dry ice. A pooled sample weighing 7.8 gm. was homogenized in 100
ml. of 75 per cent alcohol. After centrifugation, the extract was evap-
orated almost to dryness, taken up in 2.6 ml. of water, and the aqueous
extract was centrifuged to remove insoluble material. This extract was
then employed for chromatography.
In Fig. 1 is shown a tracing of a chromatogram obtained from an aliquot
of extract corresponding to 75 mg. of fresh mouse brain. Intense spots
were given by aspartic acid, glutamic acid, glycine, glutamine, taurine,
alanine, and the unknown material. Cystine, serine, and p-alanine were
present in smaller amounts, and traces of valine, threonine, and glutathione
were noted. Two substances which disappeared on acid hydrolysis were
located to the left of aspartic and glutamic acids on the chromatograms.
Among the substances which have been studied chromatographically (5),
the unknown material corresponded most closely in its behavior to histi-
*Aided by grants from the Charles F. Kettering Foundation and the United
States Public Health Service.
55
$6 Y-AMINOBUTYFUC ACID IN BRAIK

dine, methionine sulfoxide, and y-aminobutyric acid. Its position on the

chromatogram was especially favorable for separation from most of the
other detectable constituents by one-dimensional chromatography with
phenol, thus making possible a more adequate comparison of its proper-
ties with those of the aforementioned substances.

Downloaded from www.jbc.org by guest, on June 7, 2010

IO-+

FIG. 1 FIG. 2
FIN. 1. Tracing of a chromatogram of a peroxide-treated aliquot of brain extract
corresponding to 75 mg. of fresh tissue. The letters identify the spots and the num-
bers represent visual estimates of relative intensities of color, a rating of 1 being
taken as the color given by 1 y of amino acid. The spots delineated by a broken
line reprcscnt traces. AZ, cystine (cystcic acid); L, asps&c acid; K, glutamic acid;
J, swine; I, glycine; N, taurine; Q, glutamine; H, alanine; T, ,9-alanine; 8, threonine;
F, valine; X, unknown substance; R, glutathionc; P and 0, unidentified materials
which disappew on acid hydrolysis.
FIG. 2. Chromatography of unknown (X) and r-aminobutyric acid (G) in (I) col-
lidinc-lutidine, (2) phenol, (3) butanol.

Ten spots of 10 Al., each corresponding to 30 mg. of fresh tissue, were

placed along the narrow edge of each of four sheets of Whatman No. 1
filter paper and run in water-saturated phenol for 24 hours. The two end
strips of each sheet were cut out and sprayed with a 0.1 per cent solution
of ninhydrin in butanol. Six bands appeared in all instances. The un-
known material was located in the band furthest from the starting point.
The most likely contaminant would be the small quantity of valine shown
in Fig. 1. With the two end strips as reference guides, the strips of paper
 E. ROBERTS AND S. ??RANKEL 57

containing the unknown material were cut out of the four chromatograms.
ill1 ninhydrin-reactive material was removed from the strips by allowing
water to flow down t.hem into beakers. Xo color was given when these
strips were sprayed subsequently with ninhydrin after drying. The eluate
was evaporated to dryness and taken up in 2 ml. of distilled water. Sub-
sequent chroma,tography of aliquots of t’he eluate before and after hydrol-
ysis with 6 N HCl showed the presence of a large quantity of the unknown
substance, which was stable to hydrolysis, and traces of vsline and a pep-
tide material.
IdentiJicaLion of Unknown Compound As y-Aminobutyric Acid-Micro-
biological assay for hist.idine’ and a comparison of the chemical and chro-
matographic propert.ies with those of histidine and mcthionine sulfoxide

Downloaded from www.jbc.org by guest, on June 7, 2010

showed that the unknown was neither of &se compounds.
A known sample of -y-aminobutyric acid2 (15 y) and the unknown com-
pound (52 ~1.) were then chromatographed separat,ely and in mixture in
three different. water-saturat,ed solvent systems: n-butanol, phenol, and
collidine-lutidine. The results are shown in Fig. 2. Although the R,
values (3) arc widely different in the solvent systems employed, the main
constituent of the unknown. mst,erial and the r-aminobutyric acid traveled
at the same rate in each case and gave completel:r superimposable bands
when mixed. It was therefore concluded that the unknown compound is
y-aminobutyric acid. This conclusion was confirmed by application of
the isotope derivat,ive method to the eluted material (1).
Quantitative Determination of y-Aminobutyric Acid-Advantage was
taken of the fact that this amino acid could be separated from all but
traces of two of the ninhydrin-reactive materials found in brain by one-
dimensional chromatography in phenol (see the previous discussion). By
using smaller concentrations of brain extract the quantities of the inter-
fering substances were reduced below a level at which positive reactions
with ninhydrin were given. Under these conditions the color given by
the y-aminobutyric acid could be used for quantitative determination.
Spots corresponding to known amounts of r-aminobutyric acid and ac-
curately measured amounts of unknown samples were placed along the
narrow edge of sheets of Whatman Xo. 1 filter paper and run in water-
saturated phenol for 24 hours. Aft,er thorough removal of the phenol by
directing a fan at the sheet,sfor 18 to 24 hours, the sheets were sprayed
on both sideswith a 0.1 per cent solution of ninhydrin in butanol. Maxi-
mal color development took place in 24 to 36 hours at room temperature or
in 30 minutes at 93”. The developed spots were cut out, care being taken
1The microbiologicaldeterminationswere kindly performedby Dr. G. B. Rama-
satma. The organismLeuconostoc mesenteroides P-60 was employed in the assay.
* y-Aminobutyric acid waspurchasedfrom the Department of Chemistry, Univor-
eity of Illinois. It wasrecrystallized from alcohol.
58 ‘y-AMINOBIM’YRIC ACID IN BRAIN

to include the same total area of paper in all of the known and unknown
samples in any set of determinations. Suitably chosen paper blanks were
always included. The pieces of paper containing the spots were cut into
small strips with minimal handling and were put into test-tubes. 5 ml. of
water distilled in glass were then added and the tubes were shaken vigor-
ously for aminute or two. Q uan t i t at ive elutions of the color were achieved
by this method. The paper fiber was then centrifuged and the color was
read at 570 rnp in the Beckman spectrophotometer. The optical density
was proportional to the concentration in the range studied (1 to 15 7).
The color was stable for at least 4 hours at room temperature when the
tubes were kept out of strong light. Although the standard curves ob-
tained at different times were in good agreement, a series of standards was

Downloaded from www.jbc.org by guest, on June 7, 2010

run with each set of determinations in order to eliminate the influence of
possible differences in paper and solvent on the development of the color.
A similar procedure was employed in which the unknown and standards
were run on two-dimensional chromatograms in phenol and collidine-luti-
dine. Good checks were obtained when brain extracts were analyzed by
both the one-dimensional and two-dimensional methods. Three experi-
mental samples of brain homogenates gave total values of 149, 202, and
278 y of y-aminobutyric.acid by the one-dimensional procedure, while
values of 156, 200, and 295 y, respectively, were obtained for the same
samples by two-dimensional chromatography. This indicates the ade-
quacy of the one-dimensional method under the conditions of the experi-
ment. Recoveries of y-aminobutyric acid which had been added to brain
homogenates ranged from 92 to 100 per cent. The presence of glutamic
acid and aspartic acid in concentrations 200 times as great as those of the
r-aminobutyric acid did not affect the recovery.
The above methods have been applied successfully to the determination
of several other amino acids in pure solutions and in protein hydrolysates.
y-Aminobutyric Acid Present in Brain Chiefly in Free Form-The aver-
age value of a number of determinations of y-aminobutyric acid in alco-
holic extracts of mouse brain was 55 mg. per 100 gm. of fresh weight of
tissue. Closely similar values were found in the supernatant solutions
after heat coagulation of brain homogenates, in trichloroacetic acid ex-
tracts, and in acid hydrolysates of whole brain. It was also found that
hydrolysis with 6 N HCl did not increase the content of y-aminobutyric
acid in the protein-free extracts. These results show that the y-amino-
butyric acid is present in brain chiefly in the free form.
In$uence of Glutamic Acid on Content of y-Aminobutyric Acid during In-
cubation of Brain Homogenates, Washed Residues, or Acetone Powders-
Homogenates of freshly excised mouse brain containing 250 mg. of tissue
per ml. were prepared in ice-cold 0.05 M phosphate buffer, pH 7.42. In
 K. ROBERW AND S. FRANKEL 59

several experiments the homogenates were used directly. In others the

homogenates were centrifuged in the cold at 3500 r.p.m., the supernatants
discarded, and the residues washed with varying volumes of phosphate
buffer one or more times. The residues were then suspended in a volume
of buffer such that the final volume of the suspension was the same as that
of the original homogenate. The acetone powders were prepared by sus-
pending fresh brain in 100 to 500 volumes of acetone at - 15” in a Waring
blendor, filtering rapidly with suction, resuspending in the same volume of
fresh acetone, filtering, and finally drying in vacua over PZOS and paraffin
at 4”. Suitable amounts of the powders were then suspended in phos-
phate buffer with the aid of a ground glass homogenizer prior to use.
Immediately after preparation of the suspensions, 2 ml. aliquots were

Downloaded from www.jbc.org by guest, on June 7, 2010

pipetted into tubes containing 8.45 ml. of 95 per cent alcohol and 0.67 ml.
of distilled water to serve as blanks. 2 ml. portions of the preparations
were also added to 0.67 ml. of water or to 0.67 ml. of a solution containing
a suitable concentration of glutamic acid, previously adjusted to pH 7.4.
These latter mixtures were then incubated for 2 hours at 38” in an atmos-
phere of O2 in a Dubnoff incubator. The reactions were stopped by the
addition of 8.45 ml. of 95 per cent alcohol and the mixtures stirred thor-
oughly and centrifuged. The clear supernatants were poured off and ali-
quots were used for analysis. Whenever aliquots greater than 0.3 ml.
were used, the two-dimensional chromatographic procedure was employed.
The data for representative experiments are summarized in Table I. In
all of the preparations employed, the increase of r-aminobutyric acid was
accelerated by the presence of added glutamic acid. Placing the prepara-
tions in boiling water for 1 minute destroyed the activity. In several in-
stances, there were increases in the content of y-aminobutyric acid in the
absence of added glutamic acid. This occurred to the largest extent in
the crude homogenates. Even in the most extensiveiy washed prepara-
tions there were still easily detectable amounts of those amino acids which
were present in the fresh tissue in the free form to the largest extent.
From this it would be expected that constituents of brain, other than those
which react with ninhydrin, would also adhere to the washed residues or
acetone powders. For this reason it was not possible to conclude from
these experiments that glutamic acid was a precursor of y-aminobutyric
acid. The possibility was not ruled out that glutamic acid was indirectly
stimulating the formation of r-aminobutyric acid from some other sub-
stance. The following experiments were undertaken to determine whether
or not the carbon of glutamic acid would be found in y-aminobutyric acid
after incubation with a preparation which was actively forming the latter
compound.
Formation of y-Aminobutyric Acid from Glutamic Acid by Brain Acetone
GO ‘y-AMINOBUTYRIC ACID IN BRAIX

Powder-Glutamic acid which was uniformly labeled with CL4was kindly

furnished to us by Dr. Konrad Bloch. This amino acid had been isolated
from a hydrolysate of algae grown in C140n. The glutamic acid had a
specific activity of 27,900 c.p.m. per mg. when counted as the amino acid
with a windowless proportional counter connected to a scaling unit. The
sampleswere deposited from dilute aqueous solution on cups 10.7 sq. cm.
in area. A semple containing 1 mg. of the labeled amino acid to which

TABLE I
Injlxence of Glutamic Acid on Formation of y-Aminobutyric Acid in Brain
Preparalions
Incubation conditions described in the text. In the case of the homogenates

Downloaded from www.jbc.org by guest, on June 7, 2010

and washed residues an amount of tissue equivalent to 500 mg. of original fresh
weight of tissue was employed. 50 mg. of acetone powder were used.
--- - -----.- - _._-. _
Type of preparation
T Experiment No. Final molarity of Change in content of
y-aminobutyric
added glutamic acid acid per tube
-j- -
-I- 7
Homogenate 0 49
0.1250 111
0 50
0.1250 100
Washed residue 0 -5
0.0608 4
0.0039 35
0.0078 47
0 22
0.0078 69
0.0313 85
0.1256 179
0.1250% 3
Acetone powder 0 8
0.1250 369
I 0.1250” 8
0.0367 40
0.1250 144
L i -. ---
* Hrnt -inactivated.

5 y of y-aminobutyric acid had been added was subjected to two-dimen-

sional chromatography. At the same time a chromatogram was run con-
taining 5 y of y-aminobutyric acid, and one containing a seriesof samples
of knonq concentrations was run to serve as a standard curve. There
were no ninhydrin-reactive constituents other than glutamic and y-amino-
butyric acids on the paper containing the radioactive glutamic acid. On
elution and estimation of the quantity of r-aminohut(yric acid, 5 y were
 E. ROBERTS AND ti. BXANKPL 61

recovered. The eluate contained no radioactivity whatsoever. This

shows that the sample of glutamic acid employed contained no r-amino-
butyric acid or other ninhydrin-reactive impurities, that no y-aminobutyric
acid was formed from the glutamic acid during the chromatographic pro-
cedures, and that no radioactive contaminants were present which mi-
grated to the same position as the y-aminobutyric acid on the chromato-
gram. The recovery from the paper containing y-aminobutyric acid alone
was 4.7 y, a value within the range of experimental error.
An experiment was then performed in which 40 mg. of fresh acetone
powder from mouse brain were incubated with 146,000 counts of glutamic
acid (5.25 mg.) in a final volume of 1 ml. under the conditions previously
described; 3.16 ml. of 95 per cent alcohol were then added. Samples of

Downloaded from www.jbc.org by guest, on June 7, 2010

TABLE II
Radioactivity Found in Eluates of Ninhydrin-Positive Areas frost Two-Dimensional
Chromalograms ajfer Incubation of @Labeled Glutamic Acid with Acetone
Powder of Mouse Brain
146,OfXl counts of glutamic acid were added originally to th incubation mixture.
I
Amino acid *rota1c0unts
insamp~e
--_-_.-_ -. -.~.~- ..- ._ _. -- --. ~..
s.J%n.
Glutamic acid. 127,800
y-Aminobutyric acid 354
Glutamine. ‘20
Glycine............ ,... ,....._.,...i 0
Leucines............ ,,.,... _...___...i 0
Valine.... ,...,...,,,...... ,. . . .. 0
Alanine..... ,.._.......,..... ,. ., 0
Taurine. . . . . . . . . . . . .._.._.........._.. ., ,. 0
_._ ._._ . ..___~._ .___-~.--. _...~ ..-.

the clear supernatant (0.75 ml.) were subjected to two-dimensional chro-

matography. It was found that t’he total content’ of y-aminobutyric acid
after incubation exceeded by 40 y that found in the zero time control.
The spots corresponding to glutamic acid and to all of the detectable
constituent,s which were clearly separated from the large glutamic acid
spot were eluted with suitable volumes of water, plated, and count,ed. In
this manner it was possible to obtain results for glutamic acid, y-amino-
butyric acid, glutamine, glycine, the leucines, valine! alanine, and taurine.
The counts are shown in Table II.
Of the constituents other than glutamic acid, by far the largest, act.ivity
was found in the y-aminobut,yric acid. Glutamine was the only ot,her con-
stituent having a significant numbrr of counts. The act,ivit,y rceovcrcd in
the glutamic a&1 spot (127,801)a.p.m.) was somewhaBt8 higher than would
62 y-AMINOBUTYRIC ACID IN BRAIN

have been expected if 1 mole of COZ had been liberated per mole of glu-
tamic acid present (117,000 c.p.m.). The reaction with ninhydrin prob-
ably does not go to completion on paper when such large quantities of
amino acid are present.
The above experiment shows conclusively that the carbon chain of glu-
tamic acid can be converted to that of y-aminobutyric acid. It is not
possible at the present time to deal with the stoichiometry of the reaction
because of the relatively low specific activity of the glutamic acid and the
weak activity of the enzyme preparation employed. It appears highly
probable, however, that the formation of r-aminobutyric acid from glu-
tamic acid in brain takes place by a! decarboxylation.
Toxicity of y-Aminobutyric Acid-Two mice, weighing approximately

Downloaded from www.jbc.org by guest, on June 7, 2010

20 gm., were given a single intraperitoneal injection of 100 mg. of y-amino-
butyric acid in 0.5 ml. of water, and two other mice were given 0.25 ml.
of the same solution. There were no signs of toxicity and the mice ap-
peared to be completely normal thereafter. This indicates that the acute
toxicity of y-aminobutyric acid is low.
DISCUSSION

An extensive survey of the free amino acids in many normal and malig-
nant mouse tissues, a part of which has been reported (6, 7), revealed the
presence of considerable quantities of a ninhydrin-reactive substance in ex-
tracts of brain which was present, at most, in traces in the other tissues
examined. This substance was also detected in the brains of other species.
The experiments reported in this paper and in the following one (1) estab-
lished with certainty that this compound is y-aminobutyric acid. Experi-
ments with homogenates, washed residues, and acetone powders of mouse
brain invariably showed increases in the content of y-aminobutyric acid
upon incubation with glutamic acid. With isotopically labeled glutamic
acid, it was found that glutamic acid could serve as a precursor of y-amino-
butyric acid. Recent unpublished experiments in our laboratory with
homogenates and sterile autolysates of liver, muscle, and tumors have
shown increases of a substance which corresponds closely in its properties
to y-aminobutyric acid after incubation for varying periods of time. It
may thus be that the formation of y-aminobutyric is of wide occurrence
in mammalian tissues. The uniquely high concentrations of this substance
in brain may possibly be a result of a greater rate of formation than in
other tissues, a slower rate of removal, or a combination of both. A study
is being made of the metabolism of y-aminobutyric acid.
y-Aminobutyric acid has recently been reported to be present in bac-
teria (8), plants (9), human blood and urine (5), adult ox and embryo calf
muscle (lo), and yeast (11). In bacteria and plants, this compound can
 E. ROBERTS AND S. FRANKEL 63

arise as a result of the action of glutamic acid decarboxylase (12, 13).

The presence of this enzyme has not yet been demonstrated in any of the
other sources in which y-aminobutyric acid has been reported. The ex-
periments reported in this paper indicate that such an enzyme may be
present in brain tissue.

SUMMARY

1. y-Aminobutyric acid has been shown to be a constituent of brain in

which it exists preponderantly in the free form.
2. Experiments with various crude enzyme preparations from mouse
brain showed a net increase of y-aminobutyric acid, which was accelerated
by the addition of glutamic acid.

Downloaded from www.jbc.org by guest, on June 7, 2010

3. It was shown by the use of isotopically labeled glutamic acid that
the latter amino acid can serve as a precursor for the y-aminobutyric acid.
4. It is suggested that the formation of this compound in brain takes
place by the Q! decarboxylation of glutamic acid.
BIBLIOGRAPHY

1. Udenfriend, S., J. Biol. Chem., 187, 65 (1950).

2. Roberts, E., and Frankel, S., Federation Proc., 9, 219 (1950).
3. Consden, R., Gordon, A. H., and Martin, A. J. P., Biochem. J., 38,224 (1944).
4. Dent, C. E., Biochem. J., 41, 240 (1947).
5. Dent, C. E., Biochem. J., 43, 169 (1948).
6. Roberts, E., and Frankel, S., Cancer Res., 9, 645 (1949).
7. Roberts, E., Frankel, S., and Harman, P. J., Proc. Sot. Exp. Biol. and Med., 74,
383 (1959).
8. Work, E., Biochim. et biophys. acta, 3, 400 (1949).
9. Steward, F. C., Thompson, J. F.;and Dent, C. E., Science, 110, 439 (1949).
10. Gordon, A. H., Biochem. J., 100, 99 (1949).
Il. Reed, L. J., J. Biol. Chem., 183, 451 (1949).
12. Gale, E. F., Advances in Enzymol., 6, 1 (1946).
13. &hales, O., Mims, V., and Schales, S. S., Arch. Biochem., 10, 455 (1946).