The application of glutamic acid α-decarboxylase for the valorization of glutamic acid

Article abstract of DOI:10.1039/b913741f

Glutamic acid is an important constituent of waste streams from biofuels production. It is an interesting starting material for the synthesis of nitrogen containing bulk chemicals, thereby decreasing the dependency on fossil fuels. On the pathway from glutamic acid to a range of molecules, the decarboxylation of glutamic acid to γ-aminobutyric acid (GABA) is an important reaction. This reaction, catalyzed by the enzyme glutamic acid α-decarboxylase (GAD) was studied on a gram scale. In this study, GAD was immobilized on Eupergit and in calcium alginate and its operational stability was determined in a buffer free system, using various reactor configurations. Immobilization was shown to increase the GAD stability. The conditions for the highest GABA production per gram of enzyme were determined by extrapolation of enzyme stability data. At 30 °C in a fed batch process this results in an average volumetric productivity of 35 kg m^-3 hr^-1. The cost of using GAD immobilized in calcium alginate was estimated as €5 per metric ton of product. Furthermore it was shown that the cofactor pyridoxal-5′-phosphate (PLP) could be regenerated by the addition of a small amount of α-ketoglutaric acid to the reactor. In conclusion the application of immobilized GAD in a fed batch reactor was shown to be a scalable process for the industrial production of GABA from glutamic acid. The Royal Society of Chemistry 2009.

Full text of DOI:10.1039/b913741f

Cutting-edge research for a greener sustainable future
www.rsc.org/greenchem
Volume 11 | Number 10 | October 2009 | Pages 1485–1704
ISSN 1463-9262
Lammens et al.
Application of glutamic acid
α-decarboxylase
Pietropaolo et al.
Selective transfer hydrogenolysis of
glycerol
Braga et al.
Cross-coupling of diaryl diselenides
Ninomiya et al.
Synthesis of α-acyloxyacrylate esters

PAPER
www.rsc.org/greenchem | Green Chemistry
The application of glutamic acid a-decarboxylase for the valorization
of glutamic acid†
a
b
c
a
a
Tijs M. Lammens, Daniela De Biase, Maurice C. R. Franssen,* Elinor L. Scott and Johan P. M. Sanders
Received 29th June 2009, Accepted 9th July 2009
First published as an Advance Article on the web 24th July 2009
DOI: 10.1039/b913741f
Glutamic acid is an important constituent of waste streams from biofuels production. It is an
interesting starting material for the synthesis of nitrogen containing bulk chemicals, thereby
decreasing the dependency on fossil fuels. On the pathway from glutamic acid to a range of
molecules, the decarboxylation of glutamic acid to g-aminobutyric acid (GABA) is an important
reaction. This reaction, catalyzed by the enzyme glutamic acid a-decarboxylase (GAD) was
studied on a gram scale. In this study, GAD was immobilized on Eupergit and in calcium alginate
and its operational stability was determined in a buffer free system, using various reactor
configurations. Immobilization was shown to increase the GAD stability. The conditions for the
highest GABA production per gram of enzyme were determined by extrapolation of enzyme
◦
stability data. At 30 C in a fed batch process this results in an average volumetric productivity of
5 kg m- hr . The cost of using GAD immobilized in calcium alginate was estimated as €5 per
3
-1
3
metric ton of product. Furthermore it was shown that the cofactor pyridoxal-5¢-phosphate (PLP)
could be regenerated by the addition of a small amount of a-ketoglutaric acid to the reactor. In
conclusion the application of immobilized GAD in a fed batch reactor was shown to be a scalable
process for the industrial production of GABA from glutamic acid.
Introduction
One can, for example, think of isolating L-glutamic acid from
dried distiller’s grains with solubles (DDGS), a cheap co-product
of bioethanol production from maize or wheat. DDGS from
Ever increasing oil prices, environmental concerns and the
insecurity of oil supply force the need to look for alter-
natives to fossil fuels. This is not only the case for trans-
portation fuels, but also for raw materials for the chemical
industry. Most polymers still depend on fossil resources for
their building blocks. In the case of nitrogen containing
molecules, production usually involves reactions of molecules
such as propylene with ammonia in very energy-intensive
3
maize contains on average 25 to 30 wt% crude protein, which
4–7
typically consists of 20% L-glutamic acid. This means that
there is a very high potential for using L-glutamic acid as a
raw material for chemicals. Moreover, the structure of glutamic
acid with its amine and carboxylate functionalities and the small
carbon backbone resembles many industrial intermediates, so a
variety of chemicals could be made from it in a relatively small
amount of steps.
1
processes.
A major challenge is to create new, economically feasible
routes from plant materials to industrial chemicals. Amino
acids could play an important role in the making of nitrogen-
containing bulk chemicals, because amino acids already contain
nitrogen. That makes them a good starting point to by-pass the
use of ammonia and thus to save a lot of energy and investment
costs. With the current worldwide increase in biofuels produc-
tion, it may be expected that in the near future it will be possible
to cost-effectively isolate amino acids from protein rest streams.
In a potential pathway from L-glutamic acid to industrial
chemicals, g-aminobutyric acid (GABA) could be an intermedi-
ate from which various other chemicals can be made. An example
of a next step could be the simple lactamization of GABA to
8
2
-pyrrolidone, which is an industrial solvent and also the most
9
important precursor for the monomer N-vinylpyrrolidone. In
this way, the use of biomass for the production of chemicals can
be integrated in the current industrial infrastructure.
2
The reaction of L-glutamic acid to GABA can be done
enzymatically, with the enzyme glutamic acid a-decarboxylase
(
GAD), a pyridoxal 5¢-phosphate (PLP)-dependent enzyme
a
Valorisation of Plant Production Chains, Wageningen University and
10
widely distributed among living organisms. Recently a study
of this reaction using immobilized whole cells of Lactobacillus
brevis was published. However, using whole cells for this
reaction has disadvantages such as the decomposition of
product by GABA transaminase in the presence of pyruvate or
a-ketoglutarate, or the possible decrease in activity due to lack
of nutrients for the cells. Until now, to the best of our knowl-
edge, it has not been shown that this process could be profitable
Research Centre, Wageningen, The Netherlands.
E-mail: tijs.lammens@wur.nl
11
b
Dipartimento di Scienze Biochimiche “A. Rossi Fanelli”, Sapienza
Universit a` di Roma, Roma, Italy
c
Laboratory of Organic Chemistry, Wageningen University & Research
Centre, Dreijenplein 8, 6703 HB, Wageningen, The Netherlands. E-mail:
maurice.franssen@wur.nl; Fax: +31 317 484914; Tel: +31 317 482976
12
11
†
Electronic supplementary information (ESI) available: Fed batch mass
balance. See DOI: 10.1039/b913741f
1
562 | Green Chem., 2009, 11, 1562–1567
This journal is © The Royal Society of Chemistry 2009

◦
for the manufacture of other than food or pharmaceutical grade
products.
stored as such at 4 C. The filtrate was tested for GAD presence
with an activity assay.
In this paper we develop an enzyme-based process for the
scalable production of GABA from L-glutamic acid, applicable
in the chemical industry. Purified GAD from E. coli was immo-
bilized in two different manners (covalent binding to Eupergit
and gel entrapment in calcium alginate) and the performance
of the immobilized enzyme was studied in a buffer-free reaction
system, to eliminate problems such as incompatibility with the
co-enzyme PLP (the case with phosphate) or inhibition (the case
GAD immobilization on Eupergit 250 C
In a typical procedure, dry Eupergit beads (500 mg, 0.25 mm
bead diameter, 100 nm pore diameter) were suspended in
potassium phosphate buffer (5 ml, 1 M at pH 7.0) containing
0.1 mg/ml GAD and left overnight at room temperature on
an overhead rotating disc. The suspension was filtered and
the beads washed with sodium acetate (0.01 M, pH 4.6), re-
suspended in sodium acetate (7 ml, 0.01 M, pH 4.6) and left for
13
with acetate). Other advantages of a buffer-free system would
be lowering process operation costs by using fewer chemicals
and simplifying product recovery.
◦
another 6 hours on the rotating disc at 4 C, to remove any non-
covalently bound enzyme. Then the beads were filtered again
◦
and stored as such at 4 C, giving 0.24 mg GAD per 1 g of wet
Materials and methods
beads (theoretical value). Both filtrates were tested for GAD
presence with an activity assay and a Bradford protein assay.
Materials
GAD activity assay
Sodium alginate and Eupergit 250 C were both obtained from
Sigma. All other chemicals were obtained from Sigma or Fluka.
Their quality was analytical grade or higher and they were used
as received.
The enzyme activity assay was performed in water without
any additional buffer present. The setup was a Metrohm 718
stat titrino with a titration vessel equipped with a thermostatic
jacket. Titration was performed with an aqueous solution of
HCl (0.1 M).
The purified GAD isoform GadB was isolated from E. coli
14
and purified as published before.
In a typical experiment, 10 ml of an aqueous solution of
L-glutamic acid (0.08 M, 0.118 g) and PLP (0.5 mM, 1.33 mg)
in water was brought to pH 4.6 with NaOH at 40 C. Then
Analytical methods
◦
For HPLC analysis, glutamic acid and GABA were derivatized
15
0.02 mg GAD (or an equivalent of 0.02 mg GAD in the case of
immobilized GAD) was added and the titration curve recorded.
GAD specific activity was determined as a function of the slope
of the titration curve over 10 minutes time and is defined as
with phenylisothiocyanate (PITC). A 50 ml sample of reaction
mixture or of a standard solution of glutamic acid or GABA
was dried under vacuum. The residue was dissolved in 20 ml
ethanol-water-triethylamine (2:2:1 by volume) and dried under
vacuum. Then the residue was re-dissolved in 30 ml ethanol-
water-triethylamine-PITC (7:1:1:1 by volume), allowed to react
for 20 minutes and subsequently left overnight to dry under
vacuum at room temperature. The remaining dry residue was
dissolved in 0.8 ml of mobile phase, consisting of an aqueous
solution of 6.57 g sodium acetate, 0.4 ml triethylamine, 0.6 ml
acetic acid and 125 ml acetonitrile, made up to 1000 ml with
-
1
+
-1
-1
U mg , equal to mmol H added min mg .
GAD stability assay
The enzyme stability was determined by measuring the activity
for a prolonged period of time under steady-state conditions,
with a continuously operated stirred tank reactor (CSTR)
setup. For this a Gilson Minipuls 2 pump with 0.25 mm (ID)
PVC tubing was connected to the reaction vessel, continuously
pumping a solution of L-glutamic acid and PLP (0.1 M and
16
water. The final pH was 5.9.
Isocratic HPLC separation was performed on a Waters appa-
ratus with a model 600 controller, a model 717 plus autosampler
with 100 ml injection loop, a model 2487 dual wavelength
absorbance detector and a temperature control module. The
column used was a HP Hypersil BDS C-18 (250 ¥ 4.0 mm, 5 mm
particle size). UV detection was done at 254 nm. The flow rate
0.5 mM, respectively, pH 4.6) in at 0.06 ml/min and pumping
reaction mixture out at the same rate. A Biozym polyethylene
filter with 16 mm pore diameter was used to keep the Eupergit
beads in the reactor. Titration was done with an aqueous
solution of HCl (0.1 M).
◦
was 0.6 ml/min at 30 C. Retention times of PITC-glutamic acid
and PITC-GABA were 5.2 and 11.6 minutes, respectively.
Another method for monitoring the GAD activity at a high
substrate concentration for a long period of time was by titration
with 0.08 M L-glutamic acid in water. In this case, GAD specific
-
1
-1
-1
GAD immobilization in calcium alginate
activity (U mg ) was determined as mmol L-Glu min mg
added. When performing the reaction with GAD in calcium
alginate, CaCl (0.04 M) was added to the reaction mixture in
In a typical procedure, 1 mg/ml GAD solution was added
to a 2% (w/v) solution of sodium alginate in sodium acetate
buffer (0.1 M, pH 4.6), to give 25 mg/ml GAD. This solution
was added dropwise with a syringe with a 0.4 mm (ID)
needle to a continuously stirred aqueous solution of calcium
chloride (0.2 M), resulting in the formation of off-white calcium
alginate beads containing 0.05 mg GAD per 1 g of wet beads
2
both cases to keep the beads from dissolving in the course of
time.
Computational fitting
Fitting of experimental data to model equations was performed
with a Kevin Raner Software package called WinCurveFit,
version 1.1.8, 2002.
(
GAD:alginate = 1:800 by weight). The beads were allowed to
◦
harden overnight at 4 C, after which they were filtrated and
This journal is © The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1562–1567 | 1563

Results and discussion
GAD on the two different carriers. However, it is unsuitable to
determine the stability of native enzyme, because the enzyme
will not be retained by the polyethylene filter. The activity of
GAD on Eupergit and alginate was measured for 24 hours
continuously. HPLC analysis of reaction samples showed that
GAD immobilization
Two different ways to immobilize GAD were chosen, namely
gel entrapment in calcium alginate and covalent binding to
Eupergit 250C epoxide beads, in order to determine whether
immobilization has an influence on the enzyme performance.
GAD entrapment in calcium alginate yielded beads of ap-
proximately 1.5 mm diameter. The filtrate showed no activity
for conversion of glutamic acid into GABA, so apparently the
immobilization yield was 100%.
GAD binding to Eupergit was performed at pH 7. Bradford
analysis of the supernatants showed no residual protein after 18
hrs of reaction and there was no activity for the conversion of
L-glutamic acid (at pH 4.6), suggesting the immobilization yield
was 100%.
◦
in the CSTR setup at 40 C and pH 4.6, a steady state with a
constant concentration level of both glutamic acid and GABA
was reached after four hours. The activity measured as a function
of time was for both GAD in calcium alginate and GAD on
Eupergit the same as determined with a fed batch setup, which
was chosen as the preferred method to assess the stability of the
native GAD and compare it with immobilized GAD. In this case,
titration was done with 80 mM glutamic acid in water instead of
with hydrochloric acid, in order to replenish the substrate and
in doing so keeping the substrate concentration well above K .
m
A comparison of the enzyme stability of native GAD, GAD
in calcium alginate and GAD on Eupergit during 24 hours
of turnover conditions is shown in figure 1. This shows that
immobilizing GAD significantly improves the enzyme stability.
Enzyme activity assay and kinetics
◦
◦
Similar curves were obtained at 40 C and 50 C (data not
shown).
In many micro-organisms, the GAD system is assumed to
control the acidity of the cytosol, via the decarboxylation of
glutamic acid to GABA, thereby incorporating a proton in the
17
molecule. Thus when performing this reaction without added
buffer, the pH of the solution will rise in the course of the
reaction. It was found that when comparing the titration curve
with the GABA formation as measured by HPLC, these two
graphs match (data not shown), indicating that both methods are
equally suitable for determining the enzyme activity. Titration
is the preferred method, because it provides a simple (online)
activity assay.
With this assay Michaelis-Menten kinetics of the native and
immobilized enzyme were studied. For this, the initial enzyme
activity was measured as a function of the glutamic acid
concentration and the data points were fitted to the Michaelis-
Fig. 1 The influence of immobilization on GAD stability in a fed batch
setup. Individual titration curves with native GAD, GAD in calcium
alginate and GAD on Eupergit were recorded for 24 hours at pH 4.6
and 30 C. The residual activity is defined as the specific activity at a
given time divided by the initial specific activity.
Menten equation. The obtained values for the apparent K
m
in the case of native GAD, GAD in calcium alginate and
GAD on Eupergit were 2.0, 6.3 and 1.6 mM, respectively.
So immobilization of GAD in calcium alginate increases the
◦
apparent K value. An explanation for this difference could be
m
diffusion limitation. Since the alginate beads are much larger in
size, glutamic acid needs to diffuse further to reach the GAD
located in the center of the beads, a process which is favored by
a higher concentration of glutamic acid. The obtained apparent
Reaction conditions
The effect of pH on the enzyme activity was determined
for the immobilized as well as the native GAD. The native
enzyme showed a strong decrease of the activity above pH 5,
which is similar to data reported before. The immobilized
enzyme showed in both cases this profile too (data not shown),
indicating that the immobilization had no influence on the
pH dependency of the enzyme activity. The pH dependency
of the enzyme activity is known to be associated to a protein
K
m
value of native GAD in water is a little higher than the K
m
13
found by Shukuya in pyridine buffer (0.8 mM) and lower than
13
18
the K
decrease in K
thought to be the partial displacement of substrate from the
m
found in acetate buffer (6.7 mM). The reason for a
m
compared to the one found in acetate buffer is
13
enzyme surface by acetic acid.
19
conformational change, so the immobilization had no effect
on the conformational mobility of the enzyme. 24 Hour fed
batch experiments were also performed at different pHs in the
range of 3.6 to 4.6 (data not shown). No differences were found
in the enzyme stability.
Enzyme stability
The GAD activity as a function of time was determined under
semi steady-state conditions, keeping the substrate concentra-
tion well above K . This was done in order to eliminate any
m
concentration effects from the measured enzyme activity and
thus to determine the stability in time under turnover conditions.
Two different ways to do this were chosen. The first was a CSTR
setup, which is a convenient way to compare the stability of
The influence of temperature on the enzyme performance
(activity and stability) was also studied. A higher temperature
gives a higher enzyme activity, but also a faster decrease of the
activity, as can be seen from figure 2, which shows the specific
1
564 | Green Chem., 2009, 11, 1562–1567
This journal is © The Royal Society of Chemistry 2009

approach can be the regeneration of PLP from PMP by GAD
24
transaminase activity, which is known from human GAD. This
is shown in figure 3. When a small amount of a-ketoglutaric acid
is added to the system, GAD catalyzes the transamination of
a-ketoglutaric acid to glutamic acid, thereby converting PMP
into PLP.
Fig. 2 The influence of temperature on the long-term GAD activity
in a fed batch setup. Individual titration curves with GAD in calcium
◦
alginate were recorded at 30, 40 and 50 C for 24 hours at pH 4.6.
enzyme activity as a function of time with GAD immobilized in
calcium alginate. Similar curves were obtained with native GAD
and GAD on Eupergit (data not shown).
GAD deactivation
Fig. 3 The regeneration of PLP with a-ketoglutaric acid.
The cause of the decrease of GAD activity with time is known to
be related to a side reaction that takes place once every 300,000
turnovers, which is an abortive transamination reaction.
20,21
Figure 4 shows that if neither PLP nor a-ketoglutaric acid is
added, that leads to a fast deactivation of GAD. If no PLP is
added, it will only be present in a stoichiometric amount, for in
active GAD from E. coli PLP is bound covalently to the active
site. This stoichiometric amount of PLP will be transaminated
over time, leading to a fast decrease of the enzyme activity. The
addition of 5 mM a-ketoglutaric acid, however, has the same
influence on the enzyme activity as the addition of extra PLP
to the system, probably because the PLP is now regenerated.
Addition of less (0.5 mM) and more (50 mM) a-ketoglutaric
acid was also tested. The first showed no effect, which is likely
due to unfavorable kinetics, and the second showed a stable but
lower GAD activity, probably due to competitive inhibition of
glutamic acid by a-ketoglutaric acid. The fact that replacing
PLP with a-ketoglutaric acid is a possibility for this process is
especially interesting from an economic point-of-view, because
a-ketoglutaric acid is much less expensive than PLP. Moreover,
the transamination reaction regenerates glutamic acid which can
then be recycled.
In this case, GAD replaces a carbonyl group of PLP with
an amine group from a GABA molecule, thereby producing
pyridoxamine-5-phosphate (PMP) and succinic semialdehyde.
PMP dissociates from the active site, leaving GAD in its apo-
22
form, which is conformationally less stable. Meeley and Martin
showed that the inactivation of GAD from hog brain is not first
order, but is best described as the sum of two exponential decay
23
processes. The reason for this biphasic inactivation could be in
the multiple PLP binding sites that GAD contains, which might
inactivate at different rates. In the case of GAD from E. coli the
measured deactivation curves of GAD show a good fit to this
model, at different temperatures and for both the immobilized
and the native GAD. The results for the fit of the curves in figure
2
are shown in table 1. The rate constant of the first term (a)
shows a clear temperature dependency, with an increasing rate of
deactivation at increasing temperature. The second rate constant
(
b) shows no clear temperature dependency; however, the errors
in the determined values for b are high, making it difficult to
draw any conclusions from the second term of deactivation.
Cofactor regeneration
It was shown before that the stability of GAD can be greatly
18,21
increased by adding PLP to the reaction mixture.
Another
Table 1 The determined coefficients and rate constants for a fit of the
obtained data for GAD in calcium alginate with a double exponential
-at
-bt
decay function: residual activity =A
0
e
+ B e . Given errors are the
0
deviations between two measurements
-1
Coefficient (% of initial rate)
Rate constant (hr )
T
A
0
B
0
a
b
Fig. 4 The influence of a-ketoglutaric acid on the GAD stability.
Individual titration curves were recorded with native GAD at pH 4.6
and 40 C, in the presence of 0.5 mM PLP, in the presence of 5 mM a-
ketoglutaric acid and in the presence of neither PLP, nor a-ketoglutaric
acid.
◦
3
4
5
0 C 71 ± 0.2
36 ± 6.5 0.014 ± 0.002
19 ± 1.7 0.037 ± 0.008
20 ± 1.2 0.058 ± 0.016
0.74 ± 0.19
1.13 ± 0.45
0.40 ± 0.01
◦
◦
0 C 85 ± 3.8
◦
0 C 84 ± 0.6
This journal is © The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1562–1567 | 1565

Reactor design and productivity
For the immobilization of an enzyme in calcium alginate,
an enzyme to alginate weight ratio of 0.5 is achievable with
The long-term experiments with immobilized GAD described
in this article have been performed in two different ways of
operation, in a CSTR and in a fed batch reactor. From the
obtained data it is possible to derive what would be the best
mode of operation and to calculate if this process would be
economically viable in an industrial application. In order to
do this, a mass balance of the reactor was made and from
this the amount of GABA that can be produced per gram of
GAD was calculated, as was the volumetric productivity of the
reactor in that case.† GAD immobilized in calcium alginate was
chosen as the model catalyst, as it has been shown to perform
better than GAD immobilized on Eupergit. Alginate is also
less expensive than Eupergit. The chosen operating conditions
26
lipase. Assuming that would also be the case with GAD, 300
gram alginate is needed for the immobilization of 150 gram
GAD. The alginate concentration is 2% (w/v), so 300 gram
alginate is equal to 15 L of beads. If the catalyst loading is 10%,
3
then the total reactor volume would be 0.15 m . The average
volumetric productivity of this reactor over eight days based on
-
3
-1
the above assumptions would then be 35 kg GABA m hr .
The same calculation can be done for performing the process
◦
at 50 C. Then the reaction time to 5% remaining enzyme
activity is 50 hours and by then 1 gram GAD has formed 29 kg
GABA. The average volumetric productivity would then become
-
3
-1
as high as 111 kg GABA m hr .
A cascade of three CSTR’s which each 80% conversion
for these calculations were calcium alginate at pH 4.6 and
(overall giving 99% conversion) showed to be less effective than
◦
3
0 C. These conditions were chosen because an extrapolation of
a fed batch process. This is because of the low glutamic acid
concentration in the second and third reactor (4 and 1 mM,
respectively). These concentrations are below the determined
the reactor productivity based on the derived inactivation data
Table 1) shows that at 30 degrees the total expected GABA
yield per gram GAD is the highest, although it takes more time
figure 5).
(
apparent K of the enzyme in calcium alginate and therefore
m
(
the reaction rates are relatively low.
Cost analysis
To estimate the price of the needed enzyme per metric ton of
produced GABA, one can assume that a typical price for a crude
enzyme applied in the chemical industry would be €100 per kg of
pure enzyme equivalent (private communication from industry).
Then the cost of enzyme would be €3 per ton GABA. Using
crude enzyme and immobilizing that with alginate would cost
0
.3 kg alginate at €6 per kg is €1.80 per ton GABA extra. The
cost of using immobilized GAD in this process would therefore
be around €5 euro per ton of produced GABA. That means the
price will be in the same range as for example amylase (used for
27
the hydrolysis of starch), which costs €2–3 per ton of product.
Fig. 5 The expected productivity of GAD in calcium alginate at
different temperatures, based on the inactivation data shown in
table 1.
Furthermore, because of the high volumetric productivity of
this bioreactor (which is related to the high enzyme activity), the
operational costs are not expected to form an obstacle. Also the
costs of PLP and a-ketoglutaric acid are not expected to pose a
problem, because the amounts used in the process will be very
limited and most of it can be recycled.
This reaction can conveniently be performed in a fed batch
process, as was shown above. In order to keep the volume of
the reactor constant, the reaction could be performed by the
addition of solid glutamic acid instead of an aqueous solution
of glutamic acid. In this way, the substrate concentration would
remain constant as the product concentration increases, result-
ing in an accumulation of product. Therefore, it is important
that product inhibition of GAD by GABA does not take
place. This was examined up to a GABA concentration of
The major obstacle for the application of this process is the
current cost of glutamic acid, which is too high because it is still
28
produced by a fermentative process. The isolation of glutamic
acid from protein rest streams is essential to make this process
cost-effective, and is therefore a challenge that is under current
investigation.
5
mol/L. No decrease in enzyme activity was detected, as was
expected because the presence of two free carboxylic acid groups
in a molecule is required for effective inhibition of GAD.
Conclusion
25
Figure 5 showed the expected productivity of the reactor with
GAD in calcium alginate at 30 C. After eight days only 5%
Here we have shown that it is possible to develop a practical and
economically feasible process for the industrial bulk production
of GABA, based on glutamic acid derived from plant materials.
GABA can then be a precursor for different nitrogen-containing
materials in such a way that it can be implemented in the current
infrastructure of the chemical industry, thereby reducing the
dependency on fossil fuels.
◦
of the initial GAD activity remains. After this time 1 gram of
GAD has formed 34 kg GABA. For a production of 1000 kg
GABA in these eight days, 29.4 gram GAD of high purity would
be needed. The activity of the purified GAD is 5.2 times higher
14
than that of the crude extract as reported by De Biase et al.
Should a crude extract of GAD be used (which is likely to be
used in industry), 150 gram would be needed.
By way of immobilization, the enzyme (GAD) was made
more stable and easily separable from the reaction mixture.
1
566 | Green Chem., 2009, 11, 1562–1567
This journal is © The Royal Society of Chemistry 2009

Furthermore, a reactor setup with a pH stat was tested on a gram
scale with the immobilized enzyme, showing that it is possible to
perform the reaction without any buffer present. The subsequent
reactor design, based on the acquired results and assuming no
major scale-up problems, showed that a fed batch process would
be the preferred method. A major advantage of this is that
the process can be run up to a high product concentration,
making the product recovery easier. Finally, and important for
an industrial application, it was shown that the enzyme cost will
not be an obstacle for this process.
8 A. Blade-Font, Tetrahedron Lett., 1980, 21, 2443.
9
A. L. Harreus, 2-Pyrrolidone, In: Ullmann’s Encyclopedia of Indus-
trial Chemistry, Online edn., 2000, Wiley InterScience.
1
0 H. Ueno, J. Mol. Catal. B: Enzym., 2000, 10, 67.
11 J. Huang, L. Mei, H. Wu and D. Lin, World J. Microbiol. Biotechnol.,
2007, 23, 865–871.
1
1
1
1
2 B. J. Shelp, A. W. Bown and M. D. McLean, Trends Plant Sci., 1999,
, 446–452.
4
3 R. Shukuya and G. W. Schwert, J. Biol. Chem., 1960, 235, 1649–
1652.
4 D. De Biase, A. Tramonti, R. A. John and F. Bossa, Protein
Expression and Purif., 1996, 8, 430.
5 S. Gunawan, N. Y. Walton and D. M. Treiman, J. Chromatogr. A,
1
990, 503, 177.
Acknowledgements
16 V. Rossetti and A. Lombard, J. Chromatogr. B, 1996, 681, 63.
1
1
1
2
2
2
7 G. Capitani, D. De Biase, C. Aurizi, H. Gut, F. Bossa and M. G.
Gr u¨ tter, EMBO J., 2003, 22, 4027–4037.
We wish to thank professor Kenji Soda (Kansai University,
Japan) for his fruitful suggestions concerning the reactivation
of apo-GAD and Kees van Kekem (WUR, The Netherlands)
for providing HPLC apparatus and assistance. Further, we are
grateful to NWO-Aspect for funding of this work.
8 A. Tramonti, D. De Biase, A. Giartosio, F. Bossa and R. A. John,
J. Biol. Chem., 1998, 273, 1939–1945.
9 H. Gut, E. Pennacchietti, R. A. John, F. Bossa, G. Capitani, D. De
Biase and M.G. Gr u¨ tter, EMBO J., 2006, 25, 2643–2651.
0 T. G. Porter, D. C. Spink, S. B. Martin and D. L. Martin, Biochem. J.,
1
985, 231, 705–712.
1 V. P. Almazov, I. V. Morozov, F. A. Savin and B. S. Sukhareva, J. Mol.
Biol. U.S.S.R., 1985, 19, 359–370.
Notes and References
2 P. L. Grant, J. M. Basford and R. A. John, Biochem. J., 1987, 241,
1
2
F. Veatch, Chem. Eng. Prog., 1960, 56, 65–67.
6
99–704.
E. Scott, F. Peter and J. Sanders, Appl. Microbiol. Biotechnol., 2007,
2
2
3 M. P. Meeley and D. L. Martin, Cell. Mol. Neurobiol., 1983, 3, 39–54.
7
5, 751–762.
4 T. G. Porter, S. B. Martin and D. L. Martin, J. Neurochem., 1986, 47,
3
4
5
6
J. Sanders, E. Scott, R. Weusthuis and H. Mooibroek, Macromol.
Biosci., 2007, 7, 105–117.
4
68–471.
2
2
5 M. L. Fonda, Biochemistry, 1972, 11, 1304–1309.
R.L. Belyea, K. D. Rausch and M.E. Tumbleson, Bioresour. Technol.,
6 K. Won, S. Kim, K. Kim, H. W. Park and S. Moon, Process Biochem.,
2
004, 94, 293–298.
2
005, 40, 2149–2154.
M. J. Spiehs, M. H. Whitney and G.C. Shurson, J. Anim. Sci., 2002,
2
7 P. S. J. Cheetham, in Applied Biocatalysis, ed. J. M. S. Cabral, D. Best,
8
0, 2639–2645.
st
L. Boross and J. Tramper, Harwood Academic Publishers, Chur, 1
Y. Kim, N. S. Mosier, R. Hendrickson, T. Ezeji, H. Blaschek, B. Dien,
edn., 1994, ch. 11, pp. 431–469.
M. Cotta, B. Dale and M. R. Ladisch, Bioresour. Technol., 2008, 99,
2
8 T. Kawakita1, C. Sano, S. Shioya, M. Takehara and S. Yamaguchi,
Monosodium Glutamate, In: Ullmann’s Encyclopedia of Industrial
Chemistry, Online edn., 2000, Wiley InterScience.
5
165–5176.
7
J. L. Wolfson and G. Shearer, Agronomy J., 1981, 73, 611–613.
This journal is © The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1562–1567 | 1567