Flow synthesis of phenylserine using threonine aldolase immobilized on Eupergit support

Article abstract of DOI:10.3762/bjoc.9.254

Threonine aldolase (TA) from Thermotoga maritima was immobilized on an Eupergit support by both a direct and an indirect method. The incubation time for the direct immobilization method was optimized for the highest amount of enzyme on the support. By introducing the immobilized TA in a packed-bed microreactor, a flow synthesis of phenylserine was developed, and the effects of temperature and residence time were studied in particular. Calculations of the Damkoehler number revealed that no mass transfer limitations are given in the micro-interstices of the packed bed. The yield does not exceed 40% and can be rationalized by the natural equilibrium as well as product inhibition which was experimentally proven. The flow synthesis with the immobilized enzyme was compared with the corresponding transformation conducted with the free enzyme. The product yield was further improved by operating under slug flow conditions which is related to the very short residence time distribution. In all cases 20% diastereomeric excess (de) and 99% enantiomeric excess (ee) were observed. A continuous run of the reactant solution was carried out for 10 hours in order to check enzyme stability at higher temperature. Stable operation was achieved at 20 minute residence time. Finally, the productivity of the reactor was calculated, extrapolated to parallel run units, and compared with data collected previously.

Full text of DOI:10.3762/bjoc.9.254

Flow synthesis of phenylserine using threonine
aldolase immobilized on Eupergit support
Jagdish D. Tibhe, Hui Fu, Timothy Noël, Qi Wang, Jan Meuldijk
and Volker Hessel*§
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2013, 9, 2168–2179.
Micro Flow Chemistry and Process Technology, Department of
Chemical Engineering and Chemistry, Eindhoven University of
Technology, P.O. Box 513, 5600 MB Eindhoven, the Netherlands
doi:10.3762/bjoc.9.254
Received: 13 June 2013
Accepted: 24 September 2013
Published: 22 October 2013
Email:
Volker Hessel* - v.hessel@tue.nl
This article is part of the Thematic Series "Chemistry in flow systems III".
Guest Editor: A. Kirschning
* Corresponding author
§ Tel.: +31(0)402472973
Keywords:
© 2013 Tibhe et al; licensee Beilstein-Institut.
License and terms: see end of document.
Eupergit; flow chemistry; immobilized enzyme; threonine aldolase
Abstract
Threonine aldolase (TA) from Thermotoga maritima was immobilized on an Eupergit support by both a direct and an indirect
method. The incubation time for the direct immobilization method was optimized for the highest amount of enzyme on the support.
By introducing the immobilized TA in a packed-bed microreactor, a flow synthesis of phenylserine was developed, and the effects
of temperature and residence time were studied in particular. Calculations of the Damköhler number revealed that no mass transfer
limitations are given in the micro-interstices of the packed bed. The yield does not exceed 40% and can be rationalized by the
natural equilibrium as well as product inhibition which was experimentally proven. The flow synthesis with the immobilized
enzyme was compared with the corresponding transformation conducted with the free enzyme. The product yield was further
improved by operating under slug flow conditions which is related to the very short residence time distribution. In all cases 20%
diastereomeric excess (de) and 99% enantiomeric excess (ee) were observed. A continuous run of the reactant solution was carried
out for 10 hours in order to check enzyme stability at higher temperature. Stable operation was achieved at 20 minute residence
time. Finally, the productivity of the reactor was calculated, extrapolated to parallel run units, and compared with data collected
previously.
Introduction
Enzymes are bio-based catalysts having some distinct prop- complex synthesis in a green and clean manner [1]. Recently,
erties like high activity, high stereo-, regio- and chemoselec- enzymes have received much attention for making processes
tivity and high substrate specificity, which allow their use in a more economically and ecologically beneficial, as they facili-
2168

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
tate downstream processing requiring less separation steps, hydrolysis and esterification [27], oxidation and reduction [28],
providing the best product quality with high purity at low C–C bond formation [29], and polymerization [30].
energy consumption [2]. Still, the industrial applications of
enzymes are hampered by many factors, like the lack of opera- Recent advancements in the field of enzymatic microreactors
tional stability and the difficulty of enzyme recovery. Recently, include the use of alginate/protamine/silica hybrid capsules with
Wang et al. have shown that the activity and recyclability of the ultrathin membranes [31], monolithic enzyme microreactors
CalB enzyme can be enhanced by the use of a polymersome [29,32], and biodegradable enzymatic microreactors based on
Pickering emulsion. As such, a biphasic system could be used surface-adhered physical hydrogels of PVA [33]. Babich et al.
by loading the enzyme in the aqueous phase and organic demonstrated the possibility of gram scale synthesis of phos-
reagents in the polymersome [3]. The above mentioned draw- phorylated compounds using phosphatase immobilized on
backs may also be largely avoided by the use of an immobi- Immobeads [34]. Buchegger et al. used a microfluidic mixer to
lized enzyme [4]. Immobilization has also been shown to study the pre-steady state development of an enzymatic bioreac-
enhance the stability of the enzyme [5]. A number of different tion and found that the dynamics of a biochemical reaction can
immobilization methods were reviewed by Sheldon [6].
be studied in a few seconds [35].
Despite the many advantages to use enzymes, their activity is Although many promising routes have been developed for the
often low, creating a demand for process intensification. In the synthesis of chiral α-amino alcohols, these often depend on the
last decade, different methods of process intensification have use of toxic and expensive chiral ligands coupled to metal
been proposed and tested for both fine-chemical and bulk- complexes [36]. Enzymes overcome these drawbacks as they
chemical processing [7-9]. It makes sense to test process inten- are not toxic and they can be obtained easily from microorgan-
sification for biochemical processing as well. Microreactors, as isms. Threonine aldolases (TA) are a class of enzymes which is
a preferred process intensification tool, have gained consider- PLP (pyridoxal-5’-phosphate) dependent and can catalyze the
able importance due to their many advantages over conven- aldol reaction between glycine and a variety of aromatic and ali-
tional batch reactors, including rapid heat and mass transfer, phatic aldehydes [37-39]. The same enzyme can also catalyze
high surface area-to-volume ratios for dispersed media, and the reverse reaction, i.e. the cleavage of threonine into glycine
short processing times [10]. Indeed, there is indication that and acetaldehyde [40-42]. Fesko et al. conducted kinetic and
biocatalysis under flow conditions provided by a microreactor thermodynamic studies using the phenylserine synthesis from
can result in better process control by external numbering-up in glycine and benzaldehyde as a model reaction [43]. The same
which each subunit of reaction can be examined separately with group also investigated the effect of ring-sided substituents of
enhanced productivity [11].
benzaldehyde on the product yield, revealing that TA accepts
aromatic compounds having electron withdrawing groups as a
Microreactors have been used in many fields of chemistry such substrate [44]. Aldolases are known to be hampered by thermo-
as analytical systems [12], multiphase reaction systems [13-15], dynamic and kinetic limitations, such as low diastereoselectiv-
cross coupling reactions [16] and in chemical synthesis of phar- ity and product yield. To overcome this drawback of aldolases,
maceutical products [17]. Furthermore, the use of novel process dynamic kinetic asymmetric transformation has been carried out
windows to enhance the chemical production has also been in which a bi-enzymatic process was performed to achieve a
reviewed [18-22]. From the industrial perspective, Hessel et al. high yield of the product by shifting the reaction equilibrium
have analyzed the patent situation in the field of microreaction [45].
technology [23]. In extension to such fine-chemical and phar-
maceutical applications being investigated for almost 20 years, Eupergit oxirane acrylic beads provide a rapid and simple
the quite recent introduction of microfluidic devices in support for immobilization and have been used to immobilize
bioprocess intensification and biocatalysis has been reviewed various enzymes for a number of reactions [46]. Immobiliza-
[24]. Asanomi et al. have also summarized the use of microflu- tion on Eupergit (a porous material) can be achieved without the
idic devices in biocatalysis and compared them with conven- need for any additional reagents, as the epoxy groups on
tional batch reactors [25]. The advantages of enzymatic Eupergit can react directly with the nucleophilic groups of the
microreactors have been demonstrated both in process develop- enzyme by forming strong covalent linkages like amino,
ment and for the production scale [26]. Similarly these reactors hydroxy or mercapto functional groups. Eupergit has a high
can provide high throughput opportunities, reduced reaction density of epoxy groups on the surface (oxirane density
time with high conversion efficiency and high productivity per 300 μmol/g dry beads [47]), increasing the possibility of multi-
unit reaction volume for biocatalysis in fine chemistry [27]. point attachment of the enzyme. This multipoint attachment
These advantages have been demonstrated for reactions such as provides enhanced conformational stability, which translates to
2169

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
long term operational stability. Fu et al. have already investi- higher activity retention. For the direct method, our results
gated the thermal stability of TA on different supports [48] and demonstrate a high activity to those reported in the literature.
found that the enzyme stability increases after immobilization There, only 52.4% enzyme retention and activity retention of
on Eupergit support.
20.7% were achieved while for the indirect method 37%
enzyme loading with 31% of activity retention was observed
In continuation of our previous work [48], this paper includes [49].
further investigation of the thermal stability of TA on Eupergit
by a so-called direct and indirect method. Also, a flow syn-
Table 1: Comparison of direct and indirect method.
thesis of Eupergit-immobilized TA in a packed bed microre-
actor was established and compared to the use of free enzymes
in the reaction of glycine and benzaldehyde. The reaction
investigated is the synthesis of phenylserine starting from
benzaldehyde and glycine (Scheme 1). The segmented flow
experiments were carried out as an alternative to the free
enzyme in a single phase flow to maximize the yield. A con-
tinuous activity check was performed for determining the
Method
Enzyme retention (%)
Activity retention (%)
Direct
99
78
52
89
Indirect
Comparison of thermal stability of
stability of the immobilized enzyme over a longer period of TA immobilized by two different methods
time. The productivity of the flow systems reported here was For TA immobilized by using the direct method, the increase in
also determined and compared with the performance published stability of the enzyme is likely due to a multipoint attachment.
in previous literature.
Around 80% of activity retention was observed after 5 hours.
The multipoint attachment may provide a rigid structure for the
enzyme and increased stability. In case of the indirect method
the enzyme almost behaves like a free enzyme, which is due to
the introduction of spacer between the enzyme and the support.
This enhances the conformational freedom of the enzyme. The
behavior of the curve for the three cases free, indirect immobi-
lized and direct immobilized TA is shown in Figure 1.
Results and Discussion
Comparison of direct and indirect enzyme im-
mobilization
For the direct immobilization method, 99% enzyme retention
and 52% activity retention were observed, while in the indirect
method the values were 78% and 89%, respectively (Table 1).
In case of direct immobilization, the enzyme and support are
both directly in contact with each other which resulted in higher
values for enzyme retention. For the indirect method, however,
the epoxy groups were converted into aldehyde groups. These
groups can only react with nucleophiles such as amino groups.
During treatment of Eupergit with ethylenediamine, a coupling
of two adjacent epoxy groups on the support can occur which
would result in a lower amount of enzyme retention. The multi-
point attachment in case of direct immobilization may block the
active site of the enzyme and, as a consequence, this would
result in a lower degree of activity retention. In case of the indi-
rect method the enzyme was separated from the support by a
spacer element. This was formed by the reaction of ethylenedi-
amine and glutaraldehyde. This reaction reduces the probability
of blocking the active site of the enzyme which would provide
Optimization of immobilization time
The optimization of immobilization is discussed here only for
the direct method. In order to make the process more feasible
and cost effective for industrial applications, we reduced the
time of immobilization of TA on Eupergit [50]. The amount of
immobilized enzyme increases until it reaches a plateau after
about 24 hours (Figure 2). At this point 99% of enzyme immo-
bilization was achieved which is a substantial improvement
with respect to 72 hours originally employed.
The phenylserine synthesis in batch using
free enzyme
Batch reactions were carried out to estimate the yield of phenyl-
serine formation and reaction time which can principally be
Scheme 1: Phenylserine synthesis.
2170

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
Figure 1: Activity loss of TA immobilized by two different methods and
as a free enzyme at 80 °C. Reproduced with permission from Elsevier
[48].
Figure 3: Batch reaction using free enzyme. Reaction conditions:
Reaction volume (10 mL), TA (2.7 mg, specific activity – 0.135 U/mg),
temperature (70 °C).
Product inhibition study
The product inhibition study was carried out in order to under-
stand whether there is an effect of product formed during the
reaction which might block the active site of the enzyme. To
achieve this we carried out three reactions for 40 minutes at
70 °C. The amount of product which was added before the reac-
tion is 20, 40 and 60 mol % (Figure 4). It has been observed that
when the amount of product increases the yield starts to
decrease which indicates that the product inhibition effect does
exist. This is due to the very high concentration of product as
compared to the amount of enzyme used for the reaction.
Figure 2: Degree of immobilization versus incubation time.
achieved. Flow processing, and in particular its productivity
highly depends on achieving a complete transformation in a
very short time, ideally within a few minutes or less. After
20 minutes, a 40% yield of phenylserine was achieved and no
further yield increase was observed (Figure 3). TA-based reac-
tions are known to be of equilibrium-type which limits the
achievable yield to about 40% in the given case. Another reason
could principally be deactivation of the enzyme. Yet, this is not
so likely, since Figure 1 shows that even after 5 hours about
50% of the enzyme still shows activity. Thus, a self-inhibitory
effect of the product on the TA at a given concentration can be
made responsible or alternatively the reactant glycine which is
present in high concentration could cause inhibition [51].
Figure 4: Product inhibition study.
2171

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
Synthesis of phenylserine by using
immobilized TA in a microreactor
Effect of temperature
The effect of the reaction temperature on the yield of phenyl-
serine was investigated at a residence time of 10 minutes (flow
rate = 25 μL/min). As shown in Figure 5, the yield increased up
to a temperature of 70 °C reaching a value of maximal 25%
which is in line with the limitations based on the equilibrium
discussed before. Above 70 °C, a decrease in yield was
observed (Figure 5). It can be assumed that above 70 °C the
reaction rate is even higher, however enzyme deactivation
becomes more substantial. Therefore, we chose 70 °C as the
optimum temperature for the following experiments.
Figure 6: Effect of flow rates on yield for immobilized enzymes in a
packed bed microreactor (70 °C). Reaction conditions: Reaction
volume (0.250 mL), TA (1.1 mg, specific activity – 0.052 U/mg),
temperature (70 °C).
(ε) = 0.5 estimated, density (ρ) = 1024 kg/m3, viscosity of
fluid (μ) = 0.00134 kg/ms and thermal conductivity (K) =
0.58 (W/m·K).
The Reynolds number (Re) for a fluid flow through a bed of
approximately spherical particles of diameter D can be calcu-
lated, according to Equation 1 in which the void fraction is ε
and the superficial velocity is V.
(1)
Figure 5: Effect of temperature on product yield. Reaction conditions:
Reaction volume (0.250 mL), TA (1.1 mg, specific activity –
0.052 U/mg), flow rate (25 μL/min). Samples were taken after
30 minutes at the indicated temperature.
The Reynolds number was 0.1 and the Schmidt (Sc) number
was calculated to be 1.309. Next, the Sherwood number (Sh)
was calculated using the following equation,
Effect of flow rates
When operating the packed bed microreactor with immobilized
TA at 70 °C, the yield of phenylserine can be further increased
to 30% at longer residence times (Figure 6). This result resem-
(2)
bles the batch performance reported above. The reason can be The Sherwood number was found to be 1.093 which provided a
the already supposed product inhibition of the enzyme or a mass transfer coefficient (k) of 17 × 10−6. The reaction rate
specific flow effect. At higher flow rates and shorter residence vmax was calculated using experimental data and the value was
times the product yield decreases. In order to check whether 1.67 × 10−6 mol/m2s. This allowed us to determine the
mass transfer limitations are involved, we estimated the mass Damköhler number (Da) using following equation,
transfer coefficient for transport of the substrate to the particles.
(3)
Mass transfer calculations
The calculation of the Damköhler number was carried out for
the reaction with immobilized TA under the same conditions The Damköhler number turned out to be 0.98 × 10−3, which is
discussed before in the section above. The following values much smaller than the threshold criterion 1. Therefore, no mass
were used to calculate the mass transfer coefficient, porosity transfer limitation can be encountered in the present case.
2172

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
Free enzymes in flow using a capillary microreactor
profound transversal recirculation patterns within the liquid
Some flow experiments were performed in Teflon tubing at slugs which constitute a permanent highly mixed fluid system.
70 °C with the free enzyme and unraveled an almost linear Single-phase flow operations do not show similarly strong
increase of product yield with increasing residence times. Under passive mixing. However, as no mass transfer limitations exist
these homogeneous conditions, very short diffusion distance for the immobilized enzyme which also should apply here, the
between the reactants and the enzyme is guaranteed. Yet, the mixing issue can be regarded of less relevance than the resi-
yield does not increase above 15% for a residence time of dence time control issue.
20 minutes which is about half compared to heterogeneous
conditions. This could point towards a higher enzyme activity Indeed, the slug flow derived yields are higher compared to the
and/or to higher enzyme loadings in the immobilized state.
single-phase flow process. The product yield for the slug flow
process with a residence time of 20 min, is lower compared to
For a residence time of 60 min (Figure 7) a yield of 32% was the flow operation with the immobilized enzyme. For a resi-
achieved, which is not substantially higher than found for the dence time of 40 min the product yield is 34%. Thus, it is
best yield using the immobilized enzyme under microreactor slightly higher than the best yield obtained for a flow reactor
operation conditions. Studies under even longer residence times using the immobilized enzyme as packed bed. When comparing
could not be conducted, since the corresponding flow rates this result with the free enzyme reaction using a capillary
could not be accommodated with the pumps used.
microreactor, the yield increases from 16 to 25% (20 min resi-
dence time) and only a 5% increase of the yield for 40 min resi-
dence time. Obviously, slug flow operation does have a slight
effect on yields. When the amount of enzyme was doubled for a
single-phase flow experiment using a Teflon tube microreactor,
the product yield improved from 16% to 25% (20 min resi-
dence time). A longer residence time of 40 min further
increases the yield to 30% which did not change when
employing a residence time of 60 min. Again, either deactiva-
tion of TA or product inhibition can be made responsible as
discussed above.
Long term enzyme stability
A continuous experiment was performed to study the long term
stability of immobilized enzyme at 70 °C. The stability of
Eupergit-bound TA (direct method) has already been discussed.
The studies revealed that at 80 °C there is 20% loss of enzyme
activity within 4 hours. It was anticipated that at lower tempera-
tures the lifetime of the immobilized enzyme would be
extended and would allow long term operation of the flow
system. Indeed, the product yield was almost constant at around
30–35% even after 10 hours of continuous operation (Figure 8).
Figure 7: Effect of residence time on yield for free enzymes in a Teflon
tube microreactor at 70 °C. Reaction conditions: For free enzyme in
flow: Reaction volume (0.500 mL), TA (0.135 mg, specific activity –
0.135 U/mg). Slug flow: Reaction volume (0.500 mL), TA (0.135 mg,
specific activity – 0.135 U/mg). Double amount single phase: Reaction
volume (0.500 mL), TA (0.27 mg, specific activity – 0.135 U/mg).
Next, segmented flow experiments were carried out by injecting Here, the inhibitory effect by the product might be reduced due
argon in the tube microreactor which generated a slug flow with to the large amount of enzyme (1.1 mg) and the small reactor
a flow rate of 12.5 µL/min and 6 µL/min, respectively, which volume of about 250 µL. The continuous removal of the prod-
corresponded to residence times of 20 and 40 minutes. We uct from the reaction system may reduce this effect. This signi-
aimed to achieve a much narrower residence time distribution fies the stable operation of the immobilized enzyme over a
under slug flow conditions as opposed to the laminar parabolic longer period of time which is useful to reduce the cost of the
flow profile of the single-phase flow system. The latter can lead process by reusing the catalyst.
to a considerable share of reactants experiencing a (much)
shorter residence time than the given (averaged) residence time Conclusion
based on the nominal flow rates. Under slug flow conditions, all To the best of our knowledge, we report the first use of immobi-
reactants will see almost the same residence time due to reduced lized threonine aldolase in a microreactor for the flow synthesis
axial dispersion allowing the reactor to operate as an ideal plug of phenylserine. So far, we achieved a maximum of about 30%
flow reactor. Additionally, slug flow is known to create yield under the typical flow rates using TA bound to Eupergit.
2173

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
factor of being restricted to a maximum of 40% yield. Here,
removal of the product from the enzyme thereby readjusting the
equilibrium is the method of choice. One opportunity to achieve
this could be the decarboxylation of phenylserine as depicted in
Scheme 2 which is currently been investigated in our labora-
tories.
Taking into account the experimentally derived flow perfor-
mance mentioned above, the calculated TA-enzyme based
microreactor productivity was found to be 5.4 mg/h. From an
industrial perspective on microreactors, it is reasonable to
assume that an external numbering-up of 10 (respectively,
54 mg/h) along with scaling-out of those reactors the produc-
tion of 200 mg/h in total can be achieved. This value is not too
far away from the production need for commercial high-value
products in pharmacy at a price of 500 €/g and higher. The
comparison between theoretical and experimental values for
productivity has been shown in Table 2.
Figure 8: Long term enzyme stability at 70 °C. Reaction conditions:
Reaction volume (0.250 mL), TA (1.1 mg, specific activity –
0.052 U/mg), flow rate (12.5 μL/min), temperature (70 °C).
Both the theoretical and experimental productivities are based
on a quite low (experimentally used) amount of active enzymes;
simply for reasons of availability. In [48] we calculated also
productivities of other enzymatic microreactors (with other
enzymes, having no availability limitation) and this result in
much higher productivities.
The yield was about 35% for the free enzyme using a tubular
microreactor. Here slug flow performed slightly better than
single-phase flow operation. Several explanations have been
provided which rationalizes the findings reported above. These
include the residence time distribution present for the single-
phase flow operation, different degrees of enzyme deactivation,
and different degrees of enzyme activity – each being different
for the three flow operations investigated. This gives room for
future optimization. Above all, however, the existence of an
equilibrium between starting materials and products is the major
In this paper, we made a first step towards a critical view on the
commercial potential for high-priced pharmaceutical products
using enzymatic microreactor technology.
Scheme 2: Synthesis of chiral α-aminoalcohol by telescoping aldolase reaction with decarboxylation.
Table 2: Comparison of productivity between the theoretical calculations and experimental observations.
Calculations
Specific activity of enzyme (U/mg)
Amount of Eupergit inside reactor (mg)
Productivity values (g/h)
Theoreticala
Experimental
0.188
0.135
133
125
0.280
0.0054
asee [48].
2174

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
Experimental
Materials
The enzyme L-allo-threonine aldolase (L-low-TA) (EC
4.1.2.48) having concentration of 3 mg/mL and activity of
0.135 U/mg has a strong preference for L-allo-threonine from
Thermotoga maritima and was kindly donated by the Junior
Research Group ‘‘Industrial Biotechnology’’ (University
of Leipzig, Germany). L-threonine, β-nicotinamide adenine
dinucleotide disodium salt (NADH) were purchased from
AppliChem GmbH (Darmstadt, Germany). Glycine, benzalde-
hyde, pyridoxal-5’-phosphate (PLP), Eupergit CM, 2-mercap-
toethanol, 25% glutaraldehyde solution in water, ethylenedi-
amine and other reagents were all purchased from Sigma
Aldrich (Zwijndrecht, NL) and used as received. For immobili-
Figure 9: Direct immobilization.
zation of TA on Eupergit 1 M phosphate buffer solution was 2.5 mL of the TA enzyme solution (4.365 mg of TA) in a test
used, while for the phenylserine synthesis a 50 mM solution tube at room temperature. The test tube was then kept in an
was used.
orbital shaker for a predetermined time. Five different contact
times were determined (6, 12, 18, 24 and 72 h).
TA immobilization
The immobilization was carried out by a direct and an indirect Enzyme reactions
method. For the direct method, the enzyme was bonded directly Batch reactions
to the surface of the support, while the indirect method a spacer Batch reactions were carried out in 20 mL test tubes at 70 °C,
is used to provide enhanced enzyme mobility. The experi- while stirring. The test tube was charged with 750 mg glycine
mental procedure for direct immobilization (Figure 9) and indi- (1 M), 106 mg benzaldehyde (0.1 M), 100 μL 5 mM PLP solu-
rect immobilization (Figure 10) as well as enzyme retention, tion, 2 mL DMSO, 0.9 mL (2.7 mg) TA solution (activity =
activity retention and determination of enzyme concentration 0.407 U/mL) and 7 mL of a 50 mM phosphate buffer solution.
are well explained in [48].
The samples were collected at three different reaction times (20,
40 and 60 min). In each case 1 mL of sample was collected and
the reaction was terminated by adding a 30% trichloroacetic
Immobilization at different incubation times
At first the immobilization was carried out using the procedure acid solution. Then, all samples were extracted with 2 mL of
given in literature [51]. 500 mg of Eupergit was treated with internal standard solution (1,3-dimethoxybenzene in ethyl
Figure 10: Indirect immobilization.
2175

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
acetate). The enantiomeric excess (ee) and the diasteriomeric glycine, 106 mg (0.1 M) of benzaldehyde, 100 μL of 5 mM PLP
excess (de) of phenylserine were determined by HPLC analysis solution, 2 mL of DMSO, and 7.9 mL of 50 mM phosphate
of the aqueous phase, while conversion of benzaldehyde was buffer solution. For each experiment a fresh batch of immobi-
determined from analysis of the organic layer using gas chro- lized TA was used. For each reaction three samples were
matography. Since no byproducts could be detected, the degree collected and the average was calculated.
of conversion reflected the yield of phenylserine formation.
Analysis of the reaction
Flow reactions
Determination of conversion (yield)
A Teflon tubing with an inner diameter of 500 μm and a volume The conversion of benzaldehyde was determined by GC–FID
of 500 μL was used to perform the flow experiments. The (Varian 430-GC) having a CP-Sil 5 CB column of 30 meters
tubing was completely immersed into the thermostat bath at (0.25 mm ID with film thickness of 1 micron), and a flame
70 °C. Amounts of reactants employed were kept similar as for ionization detector. The carrier gas was helium (1 mL/min, split
the batch experiments. The solution was taken up in a single ratio 100). The oven temperature was maintained at 100 °C for
syringe and pumped at different flow rates (25 μL/min, 1 minute and then increased to 250 °C at a rate of 25 °C/min.
12.5 μL/min, 6 μL/min). The samples were collected at the The injector temperature was fixed at 250 °C and the detector
outlet and immediately hydrolyzed by adding a 30% solution of temperature at 250 °C.
trichloroacetic acid. For each reaction three samples were
collected. Then a sample was taken from the syringe after Determination of enantiomeric excess (ee) and dia-
12 hours and conversion was determined to be 7%. The setup stereomeric excess (de)
for the reaction is shown in Figure 11.
Determination of the enantiomeric excess (ee) and the dia-
stereomeric excess (de) was achieved by HPLC using a chiral
column (Chirex 3126 (D)-penicillamine Column 250 × 4.6 mm)
Immobilized enzyme reaction using micro flow
A simple glass tube with an inner diameter of 3.5 mm and a under the following conditions: 75% 2 mM CuSO4 solution +
length of 5 cm was used as housing for the TA immobilized 25% methanol at a flow rate of 1 mL/min. Detection of prod-
Eupergit support. A neck was constructed at the outlet for ucts was achieved at a wavelength of 254 nm. The column
holding the glass wool which prevented wash out of the temperature was kept at 40 °C throughout the analysis. The sep-
immobilized support from the reactor. In each experiment the aration of all the four isomers is depicted in Figure 13. The
same amount of Eupergit (125 mg corresponding to dry wet of correct elution of the different stereoisomers was elucidated by
Eupergit) was used as fixed-bed material. The length of the injecting pure sample of each compound and comparison with
Eupergit bed was determined to be 4.7 cm +/− 0.2 cm, while the literature [43-45].
volume of the filled reactor was around 250 μL. The reactor
was kept in a vertical position during all experiments In all experiments, 20% of diastereomeric excess of phenyl-
(Figure 12a and b). During operation the reactor was incased in serine was observed and a >99% enantiomeric excess was
a thermostat. The reaction mixture contained 750 mg (1 M) of obtained for both diastereomers.
Figure 11: Flow reaction set-up using free enzyme.
2176

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
Figure 12: Experimental setup for packed be microreactor.
Figure 13: Analysis of the four isomers of phenylserine on a chiral column.
References
Acknowledgements
1. Koeller, K. M.; Wong, C. H. Nature 2001, 409, 232–240.
The authors are thankful to the ProAChim project in the frame
of ERA-NET Industrial Biotechnology (ERA IB, No. 035581)
for providing funding and would also like to thank Dr. Nicole
Weizenmann from Junior Research Group ‘‘Industrial Biotech-
nology’’, Institute of Biochemistry, University of Leipzig,
Perlickstraße 5, D-04103 Leipzig, Germany for detailed discus-
sions on experiments.
doi:10.1038/35051706
2. Hilterhaus, L.; Thum, O.; Liese, A. Org. Process Res. Dev. 2008, 12,
618–625. doi:10.1021/op800070q
3. Wang, Z.; van Oers, M. C. M.; Rutjes, F. P. J. T.; van Hest, J. C. M.
Angew. Chem., Int. Ed. 2012, 51, 10746–10750.
doi:10.1002/anie.201206555
2177

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
4. Krajewska, B. Enzyme Microb. Technol. 2004, 35, 126–139.
doi:10.1016/j.enzmictec.2003.12.013
32.Szymańska, K.; Pudło, W.; Mrowiec-Białoń, J.; Czardybon, A.;
Kocurek, J.; Jarzębski, A. B. Microporous Mesoporous Mater. 2013,
170, 75–82. doi:10.1016/j.micromeso.2012.11.037
5. Cabral, J. M. S.; Kennedy, J. F. Immobilization techniques for altering
thermal stability of enzymes. In Thermostability of enzymes;
Gupta, M. N., Ed.; Springer Verlag: Berlin, 1993; pp 163–179.
6. Sheldon, R. A. Adv. Synth. Catal. 2007, 349, 1289–1307.
doi:10.1002/adsc.200700082
33.Fejerskov, B.; Smith, A. A. A.; Jensen, B. E. B.; Hussmann, T.;
Zelikin, A. N. Langmuir 2013, 29, 344–354. doi:10.1021/la3040903
34.Babich, L.; Hartog, A. F.; van der Horst, M. A.; Wever, R.
Chem.–Eur. J. 2012, 18, 6604–6609. doi:10.1002/chem.201200101
35.Buchegger, W.; Haller, A.; van den Driesche, S.; Kraft, M.; Lendl, B.;
Vellekoop, M. Biomicrofluidics 2012, 6, 012803. doi:10.1063/1.3665717
36.Bodkin, J. A.; McLeod, M. D. J. Chem. Soc., Perkin Trans. 1 2002,
2733–2746. doi:10.1039/B111276G
7. Stankiewicz, A.; Moulijn, J. Chem. Eng. Prog. 2000, 96, 22–34.
8. Lutze, P.; Gani, R.; Woodley, J. M. Chem. Eng. Process. 2010, 49,
547–558. doi:10.1016/j.cep.2010.05.002
9. Freund, H.; Sundmacher, K. Chem. Eng. Process. 2008, 47,
2051–2060. doi:10.1016/j.cep.2008.07.011
37.Kimura, T.; Vassilev, V. P.; Shen, G.-J.; Wong, C. H.
J. Am. Chem. Soc. 1997, 119, 11734–11742. doi:10.1021/ja9720422
38.Miura, T.; Kajimoto, T. Chirality 2001, 13, 577–580.
doi:10.1002/chir.1180
10.Ehrfeld, W.; Hessel, V.; Loewe, H. Microreactors; Wiley-VCH:
Weinheim , 2000.
11.Urban, P. L.; Goodall, D. M.; Bruce, N. C. Biotechnol. Adv. 2006, 24,
42–57. doi:10.1016/j.biotechadv.2005.06.001
39.Vassilev, V. P.; Uchiyama, T.; Kajimoto, T.; Wong, C.-H.
Tetrahedron Lett. 1995, 36, 4081–4084.
12.Peterson, D. S. Lab Chip 2005, 5, 132–139. doi:10.1039/B405311G
13.Doku, G. N.; Verboom, W.; Reinhoudt, D. N.; van den Berg, A.
Tetrahedron 2005, 61, 2733–2742. doi:10.1016/j.tet.2005.01.028
14.Shang, M.; Noël, T.; Wang, Q.; Hessel, V. Chem. Eng. Technol. 2013,
36, 1001–1009. doi:10.1002/ceat.201200703
doi:10.1016/0040-4039(95)00720-W
40.Bell, S. C.; Turner, J. M. Biochem. Soc. Trans. 1973, 1, 678–681.
doi:10.1042/bst0010678
41.Dainty, R. H. Biochem. J. 1970, 117, 585–592.
15.Stouten, S. C.; Wang, Q.; Noël, T.; Hessel, V. Tetrahedron Lett. 2013,
54, 2194–2198. doi:10.1016/j.tetlet.2013.02.064
42.John, R. A. Biochim. Biophys. Acta 1995, 1248, 81–96.
43.Fesko, K.; Reisinger, C.; Steinreiber, J.; Weber, H.; Schürmann, M.;
Griengl, H. J. Mol. Catal. B: Enzym. 2008, 52–53, 19–26.
doi:10.1016/j.molcatb.2007.10.010
16.Noël, T.; Buchwald, S. L. Chem. Soc. Rev. 2011, 40, 5010–5029.
doi:10.1039/C1CS15075H
17.Watts, P.; Haswell, S. J. Chemical Synthesis in Microreactors. In
Microengineering in Biotechnology; Hughes, M. P.; Hoettges, K. F.,
Eds.; Methods in Molecular Biology, Vol. 583; Springer Verlag, 2008;
pp 109–120. doi:10.1007/978-1-60327-106-6_4
44.Steinreiber, J.; Fesko, K.; Reisinger, C.; Schürmann, M.;
van Assema, F.; Wolberg, M.; Mink, D.; Griengl, H. Tetrahedron 2007,
63, 918–926. doi:10.1016/j.tet.2006.11.035
45.Steinreiber, J.; Schurmann, M.; Wolberg, M.; van Assema, F.;
Reisinger, C.; Fesko, K.; Mink, D.; Griengl, H. Angew. Chem., Int. Ed.
2007, 46, 1624–1626. doi:10.1002/anie.200604142
46.Boller, T.; Meier, C.; Menzler, S. Org. Process Res. Dev. 2002, 6,
509–519. doi:10.1021/op015506w
18.Stouten, S. C.; Noël, T.; Wang, Q.; Hessel, V. Aust. J. Chem. 2013, 66,
121–130. doi:10.1071/CH12465
19.Hessel, V.; Kralisch, D.; Kockmann, N.; Noël, T.; Wang, Q.
ChemSusChem 2013, 6, 746–789. doi:10.1002/cssc.201200766
20.Hessel, V.; Vural-Gürsel, I.; Wang, Q.; Noël, T.; Lang, J.
Chem. Eng. Technol. 2012, 35, 1184–1204.
47.Sukyai, P.; Rezić, T.; Lorenz, C.; Mueangtoom, K.; Lorenz, W.;
Haltrich, D.; Ludwig, R. J. Biotechnol. 2008, 135, 281–290.
doi:10.1016/j.jbiotec.2008.04.002
doi:10.1002/ceat.201200038
21.Hessel, V.; Wang, Q. Chim. Oggi 2011, 29 (3), 81–84.
22.Hessel, V.; Wang, Q. Chim. Oggi 2011, 29 (5), 70–74.
23.Dencic, I.; Hessel, V.; de Croon, M. H. J. M.; Meuldijk, J.;
van der Doelen, C. W. J.; Koch, K. ChemSusChem 2012, 5, 232–245.
doi:10.1002/cssc.201100389
48.Fu, H.; Dencic, I.; Tibhe, J.; Sanchez Pedraza, C. A.; Wang, Q.;
Noël, T.; Meuldijk, J.; de Croon, M.; Hessel, V.; Weizenmann, N.;
Oeser, T.; Kinkeade, T.; Hyatt, D.; Van Roy, S.; Dejonghe, W.; Diels, L.
Chem. Eng. J. 2012, 207–208, 564–576. doi:10.1016/j.cej.2012.07.017
49.Knezevic, Z.; Milosavic, N.; Bezbradica, D.; Jakovljevic, Z.;
Prodanovic, R. Biochem. Eng. J. 2006, 30, 269–278.
doi:10.1016/j.bej.2006.05.009
24.Marques, M. P. C.; Fernandes, P. Molecules 2011, 16, 8368–8401.
doi:10.3390/molecules16108368
25.Asanomi, Y.; Yamaguchi, H.; Miyazaki, M.; Maeda, H. Molecules 2011,
16, 6041–6059. doi:10.3390/molecules16076041
50.Dencic, I.; Meuldijk, J.; de Croon, M.; Hessel, V. J. Flow Chem. 2011,
1, 13–23. doi:10.1556/jfchem.2011.00005
26.Bolivar, J. M.; Wiesbauer, J.; Nidetzky, B. Trends Biotechnol. 2011, 29,
333–342. doi:10.1016/j.tibtech.2011.03.005
51.Janssen, M.; van Langen, L. M.; Pereira, S. R. M.; van Rantwijk, F.;
Sheldon, R. A. Biotechnol. Bioeng. 2002, 78, 425–432.
doi:10.1002/bit.10208
27.Matsuura, S.-i.; Ishii, R.; Itoh, T.; Hamakawa, S.; Tsunoda, T.;
Hanaoka, T.; Mizukami, F. Chem. Eng. J. 2011, 167, 744–749.
doi:10.1016/j.cej.2010.10.042
28.Tudorache, M.; Mahalu, D.; Teodorescu, C.; Stan, R.; Bala, C.;
Parvulescu, V. I. J. Mol. Catal. B: Enzym. 2011, 69, 133–139.
doi:10.1016/j.molcatb.2011.01.007
29.Dräger, G.; Kiss, C.; Kunz, U.; Kirschning, A. Org. Biomol. Chem. 2007,
5, 3657–3664. doi:10.1039/B712804E
30.Kundu, S.; Bhangale, A. S.; Wallace, W. E.; Flynn, K. M.;
Guttman, C. M.; Gross, R. A.; Beers, K. L. J. Am. Chem. Soc. 2011,
133, 6006–6011. doi:10.1021/ja111346c
31.Wang, J.-Y.; Yu, H.-R.; Xie, R.; Ju, X.-J.; Yu, Y.-L.; Chu, L.-Y.;
Zhang, Z. AIChE J. 2012, 59, 380–389. doi:10.1002/aic.13834
2178

Beilstein J. Org. Chem. 2013, 9, 2168–2179.
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.9.254
2179