DEAMINATION OF SERINE

Full text

DEAMINATION OF SERINE
II. D-SERINE DEHYDRASE, A VITAMIN Bs ENZYME FROM
ESCHERICHIA COLP
BY
DAVID E. METZLER
AND
ESMOND E. SNELL
(From the Biochemical Institute
and
the Department
of
Chemistry, The University
of
Texas, and the Clayton Foundation
for
Research, Austin, Tezaa)
(Received for publication, April 14, 1952)
Non-enzymatic deamination of serine and cysteine is catalyzed by py-
ridoxal and certain metal salts at 100” (2). This finding suggested that
pyridoxal phosphate might be involved in the enzymatic deamination of
these amino acids. Vitamin B, has already been implicated in the de-
sulfhydration of cysteine by rat liver (3) and of cysteine and homocysteine
by bacteria (4). Several similarities of cysteine desulfhydrase and of ser-
ine dehydrase have been reported (5, 6). These findings supported the
supposition that vitamin Be might be involved in serine dehydration, in
spite of the recent report that adenosine-5-phosphate and glutathione are
the only demonstrable cofactors of serine and threonine dehydrases (de-
aminases) from Escherichiu c& (7).
The preparation in cell-free form of a pyridoxal phosphate (PLP) re-
quiring n-serine dehydrase from cells of E. coli is described below. This
enzyme is readily separated from the serine dehydrase of Wood and Gun-
salus (7), which appears to be an L-serine dehydrase.
EXPERIMENTAL
Preparation
of
Cells-Both the Crookes strain of E. c&i (ATCC 8739)
and a vitamin Be-requiring mutant of E. cobi1 were used. These were
grown on either of two media. Medium A contained 20 gm. of Difco
tryptone, 10 gm. of Difco yeast extract, and 5 gm. of KzHPOd per liter;
its initial pH was 7.4. Medium B contained 7 gm. of KZHPOI, 3 gm. of
KH2POI, 0.5 gm. of sodium citrate trihydrate, 0.1 gm. of MgSOa.7Hz0,
0.4 gm. of (NH&SO*, 0.1 gm. of Fe(NH&(SO&.6Hz0, 10 gm. of glucose,
3.4 gm. of the amino acid mixture described by Sauberlich and Baumann
(S), 10 mg. each of adenine, guanine, and uracil, 0.4 mg. each of riboflavin,
calcium pantothenate, and nicotinic acid, 0.2 mg. each of thiamine hydro-
chloride and p-aminobenzoic acid, 10 y of folic acid, and 2 y of biotin per
liter; its initial pH was 7.0.
* A preliminary report has appeared (1).
1 We are indebted to Dr. Bernard D. Davis for supplying this organism, mutant
M154-59L.
363
by guest, on July 16, 2011www.jbc.orgDownloaded from

361
DEAMINATION OF SERINE. II
The bacteria were grown either in liquid media without aeration or in
Roux bottles on media solidified with 2 per cent agar. In either case the
cells were harvested after 14 to 16 hours, washed once or twice with 0.9
per cent NaCl, and suspended in sufficient 0.1
M
phosphate buffer (pH 7.0
or 7.8) to give a suspension containing 10 to 100 mg. of dry cells per ml.
The buffer was usually 3 X lO+
M
in glutathione (GSH). Most sus-
pensions were vacuum-dried over potassium hydroxide. Cell-free extracts
were prepared by freezing and thawing twice, then autolyzing suspensions
of dried cells as described by Wood and Gunsalus (7). The suspending
liquid was phosphate buffer at either pH 7.0 or 7.8 containing 6 X 1w3
M
GSH.
Assay of Serine Dehydrase-Optically pure
D-
and L-serine2 and commer-
cially available nn-serine and
L-
and m-threonine were employed. A
suitable aliquot of the cell suspension or extract to be assayed was placed
in a tube containing 0.3 ml. of the desired 0.5
M
buffer and water to make
the final volume after addition of activators and substrate 1.5 ml, In
early work the mixture was made 0.001
M
in MgS04, since reports in the
literature indicated a possible need for Mg++. This has not been found to
have any activating effect on the preparations we tested. Activators and
other additions were usually incubated with the enzyme for 10 minutes at
37.5” before addition of the substrate. 0.3 ml. of 0.1
M
m-serine or thre-
onine was added and the samples incubated aerobically for 10 or 30 min-
utes. Since only limited amounts of
D-
and L-serine were available, these
were used at lower concentrations, even though the reaction rate was not
linear with time under these conditions. The reaction was stopped with
0.5 ml. of 25 per cent trichloroacetic acid, the precipitate centrifuged, and
the supernatant solution analyzed for keto acid by the direct method of
Friedemann and Haugen (9). This procedure was standardized against
sodium pyruvate and against the ketobutyrate formed by deamination
of a known amount of L-threonine by dried E. coli cells. 98 per cent of the
theoretical amount of ammonia was produced in this reaction. Dupli-
cate assays usually agreed within 5 per cent or better. The data given are
average values of the duplicates, corrected for the small amount of carbonyl
compounds present in the cells and for the pyridoxal phosphate added.
While cell-free extracts show approximately linear kinetics with respect
to time and enzyme concentration (Fig. l), some dried cell preparations
deviate markedly from this behavior. For this reason, no formal unit
of enzyme activity has been employed. For comparative purposes, the
amounts of keto acid produced in 10 or 30 minutes per mg. of dry cells or
their equivalent have been tabulated along with the actual weights of
cells used.
* Dr. Jesse P. Greenstein generously supplied these isomers.
by guest, on July 16, 2011www.jbc.orgDownloaded from

D. E. METZLER AND E. E. SNELL
3%
Ammonia determinations, chromatography, and spectrophotometric
characterization of keto acid 2,4-dinitrophenylhydrazones were carried
out as described in Paper I (2). The source and purity of pyridoxal
and pyridoxamine phosphates have been described (10).
Results
Activation of Serine Dehydrase by Pyridoxal Phosphate-Pyruvate pro-
duction by cells of the vitamin Be-requiring mutant grown without added
vitamin Ba in Medium B is increased 50 to 100 per cent by the addition of
8
’ 0.6
8
%
w
t 0.4
3
E I
0 50 100 150 200
MG DRIED CELLS (AS EXTRACT) x TIME IN MINUTES
FIQ. 1. Pyruvate production from nn-serine as a function of time and enzyme
concentration. A cell-free extract of dried
E.
coli cells (Crookes strain, grown in
Medium A) was the enzyme source. Tris(hydroxymethyl)aminomethane buffer,
pH 8.2, was used. Pyridoxal phosphate added was 5 X 10-O M. Incubation at (
l
)
7.5 minutes, (0) 15 minutes, (A) 30 minutes, and (0) 60 minutes.
PLP (Table I). Pyruvate production by cells grown in the presence of a
high level of pyridoxamine is increased only 10 to 20 per cent by such ad-
ditions. The activity of freshly prepared cell-free extracts from vitamin
Be-deficient cells is increased 6 times by PLP, whereas similar extracts
from cells grown on high vitamin BE are activated only slightly. An ex-
tract from vitamin Be-deficient cells purified S-fold on an organic solids
basis by ammonium sulfate precipitation showed a g-fold increase of ac-
tivity when PLP was added. These results demonstrate that the serine
dehydrase of these extracts is a pyridoxal phosphate enzyme, and that
resolution was not affected by the treatments applied to those cells grown
with excess pyridoxamine.
Dried cells of a wild type E. coli grown on Medium A also show increased
pyruvate production from nn-serine when PLP is added (Table II). The
by guest, on July 16, 2011www.jbc.orgDownloaded from

366
DEAMINATION OF SERINE. II
effect is most pronounced in the older cell-free extracts. This partial
resolution of the holoenzyme is probably a result of repeated thawings of
the extracts, which were stored in the frozen state; prolonged storage does
not in itself cause any resolution. In line with the observations of others
TABLE
I
Pyruvate Production from DL-Serine by Vitamin B6-Requiring Mutant of E. coli.
E. coli M154-59L was grown on Medium B with no added vitamin Bs (low B6) or
with 20 mg. per liter of pyridoxamine dihydrochloride (high Be). Fresh cells were
treated with toluene to prevent keto acid metabolism (12). All dried cells and
cell-free extracts were prepared in phosphate buffer, pH 7.0, and incubated in the
same buffer.
Preparation No.
I. Fresh cells from liquid medium
II. Dried cells from liquid medium
III. Dried cells from agar surface
Cell-free extract from III
“ “ “ II
Fractionated extract1 from III
Medium
Low Bs
High Be
Low Be
High Be
Low Be
Low Be
High B6
Low Be
t
Additions*
w.
2.4 None
2.4 PLP
2.4 None
2.4 PLP
2.4 None
2.4 PLP
2.4 None
2.4 PLP
1.5 None
1.5 PLP
2.4 None
2.4 PLP
2.4 PLP + AMP
2.4 None
2.4 PLP
5.3 None
5.3 PLP
I
Y pn mg.
cells
0.063
0.112
0.084
0.100
0.061
0.117
0.096
0.099
0.107
0.263
0.023
0.133
0.140
0.064
0.068
0.0058
0.051
* Pyridoxal phosphate (PLP), 5 X 10-4 br; adenosine-5-phosphate (AMP), 5
x lo-‘M.
t Chromatographic experiments established that the keto acid formed was pyru-
vate.
$ The fraction used was soluble in 1.72 M and insoluble in 1.98
M
ammonium sul-
fate at room temperature. The product was dialyzed free of ammonia.
(ll), we found that cells grown on the carbohydrate-free Medium A at pH
7.4 have a much higher deaminase activity than those grown on the carbo-
hydrate-containing Medium B at pH 7.0. The former cells, especially
when dried at pH 7.8, also deaminate threonine and are activated by
adenylic acid and glutathione, as reported by Wood and Gunsalus (7).
by guest, on July 16, 2011www.jbc.orgDownloaded from

D. E. METZLER AKD E. E. SNELL
367
In contrast, most of the cell-free extracts show little or no activation by these
compounds and are only slightly active in the deamination of threonine.
Ik:$ect of Various Buffers on Serine Dehydrase Activity-A striking dif-
TABLE II
Pyruvate Production
from
m-Serine by Wild Type E.
coli
All enzyme
preparations were
from Crookes
strain
of
E. co2i
grown on Medium A.
Incubations were carried out in phosphate buffer, pH 7.8.
Preparation Additions*
X. Dried cells, pH 7.8
Cell-free extract from X, pH 7.8
Same extract 6 days old
0.60
None
0.60
PLP
0.60
AMP + GSH
0.40 None
0.40 PLP
0.40 AMP + GSH
1.20 None
1.20 PLP
fly #$*rn”.
0.45
0.55
1.57
0.29
0.55
0.37
0.14
0.59
* Pyridoxal phosphate, 5 X lo-’ M; adenosine-5-phosphate, 5 X lo-’ M; gluta-
thione, 1 X 10e3 M.
TABLE III
Effect
of
Different Buffers
on Serine
Dehydration by Eztracta
of
E. coli
All tubes contained cell-free extract, Preparation IX, equivalent to 0.23 mg. of
dry cells of
E.
coli, Crookes strain, grown on Medium A.
Buffer. pR 7.8
Phosphate
Tris(hydroxymethyl)aminomethane
Triethanolamine
None
PLP
PLP
None
PLP
* 5 X 10eB M pyridoxal phosphate was used.
ference in serine dehydrase activity occurs when phosphate buffer is re-
placed by tris(hydroxymethyl)aminomethane buffer (Tris buffer, Table
III). In the absence of added PLP, the activity in the latter buffer is only
10 per cent of that in its presence; the activity in phosphate buffer (without
PLP) is much higher. A possible explanation for the high apparent reso-
by guest, on July 16, 2011www.jbc.orgDownloaded from

368
DE.kMI~ATION OF SERIKE. II
lution in Tris buffer is that the primary amine group of this buffer combines
loosely with PLP, perhaps to form a Schiff base. Triethanolamine buffer
cannot react ih this way and shows effects intermediate between phosphate
and Tris buffers (Table III), suggesting that this explanation may be par-
tially correct.
The dependence of activity of partially purified serine dehydrase upon
pH is shown in Fig. 2. The optimum is similar to that reported (11) for
intact, resting cells of E. coli.
Optical Specijicity of Sol&e Enzyme-The cell-free extracts attack
D-
serine rapidly and convert it completely to pyruvate and ammonia.
L-
5 6 7 6 9
,,H AT 25’
FIG. 2. pH dependence of pyruvate production by the salt-fractionated extract
of
E.
coli. Pyruvate production at 30 minutes per mg. of original cell weight is
plotted. 0, acetate and phosphate buffers; A, tris(hydroxymethyl)aminomethane
buffers.
Serine is attacked at a rate less than one-tenth as great (Fig. 3). Thus,
the enzyme is a n-serine dehydrase and differs from that described by Wood
and Gunsalus (7), which appears to be L-serine (or threonine) dehydrase.
Further clarification of the relationship of these two enzymes is provided
by the data in Table IV. Whereas keto acid production by dried cells from
L-serine and m-threonine is greatly activated by addition of AMP, pyru-
vate production from n-serine is unaffected. In contrast, dehydration of
n-serine but not of L-serine or threonine, is greatly increased by PLP.
Under our conditions D-serine dehydrase is more stable than the
L-
serine dehydrase, since most cell-free extracts are almost free of the latter
enzyme. Whether the small amount of deamination of L-serine by such
extracts (Fig. 3) is due to contaminating traces of the L-serine dehydrase
is not yet known.
by guest, on July 16, 2011www.jbc.orgDownloaded from

D. E. METZLER AND E. E. SNELL
369
In recent experiments, extraction of 70 to 95 per cent of the D-serine
dehydrase from dried cells was effected by merely suspending them for
1 hour at 5’ in 0.1
M
phosphate buffer. The extract contained very little
L-serine dehydrase, while a large part was retained in the residue (Table V).
A more detailed study of the properties of the D-serine dehydrase from
E’. coli is planned after its further purification. Experiments to date give
the following additional information. Pyruvate and ammonia are pro-
D-SERINE
L-SERINE
0 IO 20 30 60 120
TIME, MINUTES
FIG. 3. Comparative pyruvate production from D- and L-serine by a cell-free
extract of
E.
coli (Crookes strain). Substrates were 2 X 1OP M; pyridoxal phosphate
6 X 1O-B M. The extract of 2 mg. of dry cells was present per ml. of phosphate buffer,
pH 7.8. Determinations on the 120 minute samples showed that 98 per cent of the
theoretical ammonia was evolved in the deamination of n-serine. Chromatography
and spectral characterization of the 2,4-dinitrophenylhydrazone of the product
showed that pyruvate was produced. No detectable amounts of hydroxypyruvate
or other keto acids were formed.
duced in equimolecular amounts, and pyruvate is the only keto acid
formed (Fig. 3). Pyruvate production under nitrogen is as rapid as in
air. ?rTo stimulation by either AMP or GSH has been observed; however,
the latter was always added during preparation of dried cells and hence
was present in small amounts. Approximately 50 per cent saturat,ion of
the enzyme with coenzyme and substrate respectively occurs with 1 X lO+
M
PLP and 3
X
1O-4
M
n-serine. With a crude cell-free preparation, pyri-
doxamine phosphate (PMP) activates the enzyme fully at, a concentration
of 1
X
lO+
M,
hut conversion of PMP to PT,P during preincubation of
enzyme with activators (e.g. by transamination with small amounts of
keto acid present in such extracts) has not been excluded.
by guest, on July 16, 2011www.jbc.orgDownloaded from

370
DEAMINATION OF SERINE. II
TABLE
IV
Keto Acid Production by Dried Cells from Various Substrates
0.5 mg. of cells of E. coli, Crookes strain, dried at pH 7.8, per tube. Incubation
at pH 7.8, phosphate buffer, for 10 minutes.
Substrate* I Additionst Keto acid production
I I
PM per mg. cells
n-Serine.
“
“
“
L-Serine
I‘
.......
.......
.......
“
‘I
nL-Serine.
“ . .
“
“ . .
nL-Threonine.
“
.........
0.38
.........
PLP 0.62
.........
AMP 0.38
.........
PLP + AMP 0.62
.........
None 0.048
....... .........
PLP 0.056
........ .........
AMP 1.00
....... .........
PLP + AMP 1.03
.........
None 0.38
.........
PLP 0.59
....... .........
AMP 1.25
....... .........
PLP + AMP 1.48
....... .........
None 0.14
.........
PLP 0.19
“
....... .........
AMP 2.08
“
....... .........
PLP + AMP 2.12
* 0.3 ml. of 0.01
M D-
or L-serine or 0.02 Y nL-serine or threonine per tube.
t 5 X 10-O
M
pyridoxal phosphate and 5 X lo-’
M
adenosine-5-phosphate.
TABLE
V
Partial Separation
of
D-
and L-Serine Dehydrases
Dried cells of E. coli, Crookes strain, grown on Medium A were treated as shown.
Assays were carried out in phosphate buffer, pH 7.8.
Preparation Pyruvate production in 10 min.
-- n-S&m? 1 L-Serine*
Dried cells.. .I
0.40 0.56 1.73
Extract At.. ’
0.36 0.40 0.10
“ Bt.. . j 0.40 0.50 0.20
Washed residue from A$. j 0.33 0.11 / 0.639
* 0.3 ml. of 0.01
M
n-serine with 5 X 1O-6
M
pyridoxal or 0.3 ml. of 0.01
M
L-serine
with 5 X lo-’
M
adenosine-5-phosphate and 1 X 10-*
M
glutathione.
t Cells suspended in 0.1
M
phosphate buffer, pH 7.8, containing 3 X 10eJ
M
ade-
nosine-5-phosphate and 6 X 10-a
M
glutathione. Extract A contained 20 mg.,
Extract B 50 mg., of dried cells per ml. Residue centrifuged after standing 1 hour
at 5”.
$ Residue washed once with 2) times the original volume of buffer.
5 Variable results were obtained, depending on the assay conditions.
by guest, on July 16, 2011www.jbc.orgDownloaded from

D. E. METZLER AND E. E. SNELL
371
Pyridoxal, pyridoxamine, 5-desoxypyridoxal, and 3-methoxypyridoxal
were without effect in activating the enzyme at concentrations of 2.5 X 10m6
M.
In frozen crude extracts, the enzyme is stable for at least 3 months.
The enzyme is inhibited almost completely by 1V
M
copper or zinc sulfates
and is inactivated by boiling for 2 minutes.
DISCUSSION
The data presented show that E. coli contains a n-serine dehydrase
which is activated by pyridoxal phosphate. Recently Yanofsky (12) and
Reissig (13) have reported that extracts of Neurospora contain a serine-
deaminating enzyme which is stimulated by pyridoxal phosphate but not
by AMP or GSH. No evidence that the L-serine dehydrase of Wood and
Gunsalus (7) is a vitamin Bs enzyme has yet been obtained, but it would
seem strange if entirely different mechanisms for this reaction were em-
ployed for the
D
and
L
isomers. It seems likely that L-serine dehydrase
is also a vitamin Be enzyme, but that the coenzyme is bound much more
tightly. Reissig (13) reports that L-threonine dehydrase of Neurospora is a
pyridoxal phosphate-requiring enzyme.
Serine is deaminated in at least two ways by biological systems: oxi-
datively to hydroxypyruvate (14), or with no net oxidation to yield py-
ruvate. The latter reaction corresponds to the non-enzymatic pyridoxal-
catalyzed reaction (2), and is that carried out by the
D-
and L-serine
dehydrases studied here. It is predominant in bacterial cells (11, 15).
Chargaff and Sprinson (16) have suggested that in this reaction the en-
zyme catalyzes the dehydration of serine to a-aminoacrylic acid, which is
then spontaneously converted to pyruvate and ammonia. The non-en-
zymatic dehydration by pyridoxal and the enzymatic dehydration by
pyridoxal phosphate proteins are readily fitted into this mechanism if
one assumes formation of a Schiff base between serine and pyridoxal or
pyridoxal phosphate, as suggested for the non-enzymatic reaction (2).
Whether the enzymatic reaction also requires a metal ion, as does the non-
enzymatic reaction, remains to be determined. A metal ion might stabilize
the intermediate Schiff base, as suggested for the non-enzymatic reaction.
If a metal is not involved, the intermediate would still be stabilized by
by guest, on July 16, 2011www.jbc.orgDownloaded from

372
DEAMINATION OF SERINE. II
hydrogen bonding (17), as shown in the accompanying diagram. The
immediate product of enzymatic activity would be the Schiff base of amino-
acrylic acid with pyridoxal phosphate, which could then hydrolyze to t,he
unstable free aminoacrylic acid. It is possible that in other vitamin Be-
catalyzed reactions involving serine this Schiff base of aminoacrylic acid,
which would be somewhat stabilized by the long conjugated bond system,
could react with other substances by addition across the double bond;
e.g., with indole to yield tryptophan (12) and with homocysteine to give
cystathionine (18).
Purified preparations of n-amino acid oxidase attack n-serine slowly (19)
or not at all (20). The natural occurrence of a special enzyme, D-serine
dehydrase, that rapidly converts this amino acid to pyruvate thus assumes
added interest as another indication of the participation of n-amino acids
in metabolism.
We are grateful to Dr. Chozo Mitoma for performing several experi-
ments establishing the optical specificity of the enzyme.
SUMMARY
1. Vitamin Ba-deficient cells of a mutant strain of Escherichia coli con-
tain a serine dehydrase which requires pyridoxal phosphate as coenzyme.
The enzyme is readily obtained in cell-free extracts and is partially re-
solved when such extracts of a wild type E. coli are aged.
2. The enzyme is specific for n-serine and can be separated readily
from an L-serine dehydrase that is also present in dried cells.
3. Unlike L-serine dehydrase, n-serine dehydrase is not activated by
adenosine-5-phosphate. No evidence that pyridoxal phosphate is a com-
ponent of the enzyme deaminating L-serine and L-threonine was obtained,
although its essential r81e in this process is considered likely.
4. A possible mechanism for serine dehydration is discussed briefly.
BIBLIOGRAPHY
1. Metzler, D. E., and Snell, E. E.,, Federation Proc., 11, 258 (1952).
2. Metzler, D. E., and Snell, E. E., J. Biol. Chem., 198, 353 (1952).
3. Braunshtein, A. E., and Azarkh, R. M., Doklady Akad. Nauk S. S. S. R., 71, 93
(1950); Chem. Abstr., 44, 7900 (1950).
4. Kallio, R. E., J. Biol. Chem., 192, 371 (1951).
5. Binkley, F., J. Biol. Chem., 160, 261 (1943).
6. Smythe, C. V., Ann. New York Acad. SC., 46, 425 (1944).
7. Wood, W. A., and Gunsalus, I. C., J. Biol. Chem., 181, 171 (1949).
8. Sauberlich, H. E., and Baumann, C. A., J. Biol. Chem., 176, 165 (1948).
9. Friedemann, T. E., and Haugen, G. E., J. Biol. Chem., 147, 415 (1943).
10. Metzler, D. E., and Snell, E. E., J. Am. Chem. Sot., 74, 979 (1952).
Il. Gale, E. F., and Stephenson, M., Biochem. J., 32, 392 (1938).
by guest, on July 16, 2011www.jbc.orgDownloaded from

D. E. METZLER AXD E. E. SNELL
373
12. Yanofsky, C., J. Biol. Chem., 194, 279 (1952).
13. Reissig, J. L., Arch. Biochem. and Biophys., 36, 234 (1952).
14. Sprinson, D. B., and Chargaff, E., J. Biol. Chem., 194, 411 (1946).
15. Chargaff, E., and Sprinson, D. B., J. Biol. Chem., 161, 273 (1943).
16. Chargaff, E., and Sprinson, D. B., J. Biol. Chem., 148, 249 (1943).
17. McIntire, F. C., J. Am. Chem. Sot., 69, 1377 (1947).
18. Binkley, F., Christensen, G. M., and Jensen, W. N., J. Biol. Chem., 194, 109
(1952).
19. Klein, J. R., and Handler, P., J. Biol. Chem., 139, 103 (1941).
20. Horowitz, S. H., J. Biol. Chem., 164, 141 (1944).
by guest, on July 16, 2011www.jbc.orgDownloaded from