Production and properties of threonine aldolase immobilisates

Article abstract of DOI:10.1016/j.molcatb.2014.01.021

Dichiral β-hydroxy-α-amino acids are a highly valuable class of compounds from which pharmaceutically active intermediates for the synthesis of e.g. β-sympathomimetic drugs can be obtained. In lieu of laborious multi-step "classical" organic synthesis, biocatalysis using threonine aldolases (TAs) opens up a way to synthesise β-hydroxy-α-amino acids in one step. Although enzyme kinetics, stereospecificity as well as substrate specificity were and are matters of investigation, there is a lack of investigations addressing enzyme stability, which is crucial if the reaction is thought to be transferred to an economical scale. Hence methods to immobilise the l-low specificity threonine aldolase of Escherichia coli (l-TA) were studied. After extensive screening the entrapment of the enzyme into a porous network of orthosilicate appeared to be the most promising method.

Full text of DOI:10.1016/j.molcatb.2014.01.021

Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Production and properties of threonine aldolase immobilisates
Sandra Kurjatschij , Michael Katzberga,b, Martin Bertaua,∗
a
a
Technische Universität Bergakademie Freiberg, Institute of Industrial Chemistry, Leipziger Straße 29, 09599 Freiberg, Germany
ThyssenKrupp Uhde GmbH, Am Haupttor, Bau 3668, 06237 Leuna, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 8 February 2014
Dichiral ␤-hydroxy-␣-amino acids are a highly valuable class of compounds from which pharmaceutically
active intermediates for the synthesis of e.g. ␤-sympathomimetic drugs can be obtained. In lieu of labori-
ous multi-step “classical” organic synthesis, biocatalysis using threonine aldolases (TAs) opens up a way
to synthesise ␤-hydroxy-␣-amino acids in one step. Although enzyme kinetics, stereospecificity as well
as substrate specificity were and are matters of investigation, there is a lack of investigations addressing
enzyme stability, which is crucial if the reaction is thought to be transferred to an economical scale.
Hence methods to immobilise the l-low specificity threonine aldolase of Escherichia coli (l-TA) were
studied. After extensive screening the entrapment of the enzyme into a porous network of orthosilicate
appeared to be the most promising method.
Keywords:
Enzyme immobilisation
Threonine aldolase
Biocatalysis
Amino alcohols
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
important as improving stereoselectivity and yield. However in
case of TAs investigations on improving stability of those enzymes
are rather scarce, which prompted us to study ways to improve
the stability of TAs.
The most common strategy to improve stability is immobilisa-
tion [11,12] which can also be used to make an enzyme separable
from the reaction mixture and hence reusable, given the activity
of the enzyme is retained. Furthermore immobilisation enables to
run a process continuously as the immobilised enzyme can be used
as a fixed bed.
Threonine aldolases (TAs) and the related serine hydrox-
ymethyltransferases (SHMT) [1,2] are a group of enzymes that
are able to catalyse formation and cleavage of the C2 C3-bond in
␤
logical role of TAs is to cleave threonine yielding acetaldehyde and
glycine (Scheme 1), it is the reverse reaction that is attracting inter-
est of synthetic organic chemists, since it would allow to synthesise
stereopure ␤-hydroxy-␣-amino acids and it derived molecules like
-hydroxy-␣-amino acids like threonine [3,4]. While the physio-
1
,2-aminoalcohols from simple starting materials (an aldehyde and
But although numerous immobilisation methods have been
developed so far [13], no universal principles for selecting the
appropriate method for a given enzyme have yet been identified,
making selection of the most appropriate immobilisation method
for a particular enzyme still a matter of empirical work [14]. Con-
sequently different immobilisation methods needed to be studied
in order to identify suitable methods for stabilising threonine
aldolases (TAs).
an amino acid) [5–7].
By using this biocatalytic approach highly valuable chiral build-
ing blocks needed to synthesise pharmaceutically active substances
like antibiotics as well as ␤-sympathomimetic drugs (Fig. 1) would
be accessible in a more straightforward and environmentally
benign way than the currently used laborious multi-step synthesis
making use of “classical” organic chemistry [8].
However this straightforward and promising biocatalytic
approach still suffers from drawbacks such as unsatisfactory
diastereoselectivity of TAs and the achievable product yield. As a
consequence those issues are matter of current investigations [10].
But although product yield and stereoselectivity are important per-
formance criteria of a biocatalytic process, it often is the stability
of the enzyme determining whether a biocatalytic route becomes
economically viable or not. Hence investigating methods and ways
to maximise stability of an enzyme under process conditions is as
Among known TAs the low-specificity TA of Escherichia coli (l-
TA) appeared to be the most promising one to be studied for immo-
bilisation as it is well studied and reported to catalyse synthesis of
amino alcohols and hydroxy amino acids with good stereoselectiv-
ity (values obtained by Baer et al. for o-chlorophenylserine: conver-
sion > 95%, enantioselectivity > 99%, diastereoselectivity with d.r.
(threo/erythro) = 80:20) making it one of the promising candidates
to become commercially interesting [15].
2. Results and discussion
∗ Corresponding author. Tel.: +49 3731 39 2384; fax: +49 3731 39 2324.
E-mail address: martin.bertau@chemie.tu-freiberg.de (M. Bertau).
The low-specificity l-threonine aldolase of E. coli (l-TA) is a
homotetrameric enzyme [16] for which inactivation (16% loss in
1
381-1177/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2014.01.021

4
S. Kurjatschij et al. / Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
OH
O
O
O
TA
+
H C
3
OH
OH
H3C
H
NH2
NH2
Scheme 1. Threonine aldolase catalysed cleavage of threonine to give glycine and acetaldehyde.
O
Ar
OH
OH
O
N
H
HN
S
O
(
+)-L-733,060
O
Cl
Cl
receptor antagonist
Thiamphenicol
1
R
antibiotic
1
OH
O
R
OH
O
2
R
1
N
H
R
O
OH
NH2
oxazolidinones
NH2
antibiotics
rare amino-acids
histone deacetylase inhibitors
OH
NH
CH3
OH
O
HO
O
KUL7211
sympathomimetic
-
Fig. 1. Pharmaceutically active products based on dichiral intermediates having a 1,2-aminoalcohol moiety.
activity in 6 weeks) was observed even if the enzyme was stored
at pH = 7.0. From the resulting enzyme activity in the supernatant
which was >93% in all cases, it was concluded that in fact the poor
adsorption behaviour of l-TA was responsible for the low activity
found for l-TA immobilised by adsorption.
◦
at 4 C indicating that immobilisation is clearly required in order to
◦
maintain enzyme activity under reaction conditions (T ∼ 30 C) as
well as over normal periods of storage.
In general there are four basic methods for enzyme immo-
bilisation: (1) adsorption onto a carrier, (2) entrapment into a
gel or porous media, (3) covalent attachment onto a carrier and
These results show that binding the enzyme via ionic interaction
is not efficient in case of l-TA as the overall charge of the enzyme
at process conditions of pH = 8.0 does not appear to be sufficient
to permanently link the enzyme to the carrier. Hence alternative
immobilisation methods were investigated.
(4) crosslinking of enzyme crystals and aggregates (CLEA, CLEC)
17,18].
[
2.1. Adsorption onto carriers
Table 1
Screening of immobilisation procedures for l-threonine aldolase of Escherichia coli
and obtained yields.
Adsorption makes use of the ionic interactions between the
enzymes surface and a carrier substrate. However investigations on
Enzyme
Technique
Inclusion
Support
Residual
activity (%)
adsorption of l-TA onto different materials (CaCO , Al O , Alginit)
3
2
3
[
19,20] having ion exchange properties showed that this method
Alginate beads
Orthosilicates
Superabsorber Favor®
Amberlite XAD-2
Al2O3
5–12
is not a promising way to obtain l-TA immobilisates, as adsorption
yields did not exceed 1% of the original activity (Table 1). Only in the
case of the styrene-divinylbenzene based commercial ion exchange
resin Amberlite XAD 2 [21] a significant amount (Table 1) of enzyme
remained on the carrier.
In order to make sure that poor adsorption properties instead
of other reasons caused the low enzyme activity after fixation onto
the adsorbent, the l-TA was eluted with sodium phosphate buffer
≤ 30
–
<6
<1
<1
<1
<1
<1
5
l-TA E. coli
Adsorption
CaCO3
Cellulose
Alginit
Eupergit®
Chitosan
C
Covalent
attachment

S. Kurjatschij et al. / Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
5
Scheme 2. Direct covalent attachment of enzyme onto Eupergit® C.
2
.2. Covalent attachment onto a carrier
towards reactive groups and that the essential amino groups in
the active site experience modification resulting in enzyme inac-
tivation. Therefore no further experiments were conducted with
respect to a potential deactivating physical interaction.
The use of carriers bearing reactive groups on its surface capa-
ble of reacting with e.g. enzyme amino groups enables the covalent
attachment of an enzyme onto a carrier. In comparison to immo-
bilisation products obtained from equilibrium dependent ionic
interactions, covalent attachment of enzymes onto a carrier yields
stable immobilised enzymes with preserved activity [22].
2.3. Entrapment into a matrix
From the observations that neither adsorption nor covalent
attachment of l-TA onto a carrier resulted in immobilisates with
satisfactory activity, it becomes clear that successful immobili-
sation requires methods not modifying catalytically important
functional groups of the enzyme or resulting in full or partial
denaturation and thus retaining enzyme activity. An immobilisa-
tion technique from which neither effects on the active site nor
negative effects on protein integrity were to be expected, is entrap-
ment of the enzyme into a matrix among which alginate [30] and
orthosilicate gels [31] were studied.
Alginate encapsulation provided immobilisation yields in the
range of 10% (Table 1). However entrapment into orthosilicates
proved to be the immobilisation procedure of choice as up to 30% of
residual enzyme activity were retained after immobilisation. Thus
the l-TA immobilisates produced with this technique showed the
highest activity of all immobilisates investigated in this study.
In case of orthosilicate gels, the three-dimensional network
around the entrapped enzymes is formed by acid-catalysed hydrol-
ysis of alkyl orthosilicates, namely tetraethyl/tetramethyl orthosili-
cate (TEOS/TMOS), and triethoxy/trimethoxymethyl orthosilicate
(TRIEMOS/TRIMMOS), resulting in silanols which in turn undergo
condensation polymerisation (Scheme 3).
SEM studies revealed that the resulting silicates exhibit a highly
porous structure as shown in Fig. 2 depicting TEOS entrapped l-
TA. The latter image proves that there are sufficient number of
pores providing enough space enabling unhindered transfer of the
substrate and product to and from the enzyme, respectively.
But although an immobilisation method for E. coli l-TA could be
identified, the observed immobilisation yield of 15–27% shows that
the method still needs optimisation.
One reason for the loss in enzyme activity in the course of TEOS
immobilisation is the fact that hydrolysis of TEOS not only yields
the silicate gel matrix encapsulating the enzyme but also an alco-
hol (in case of TEOS, ethanol) being a potential enzyme denaturant.
From the reaction equation (Fig. 3) it becomes evident that per mol
of alkyl orthosilicate 4 mol of the respective alcohol are formed. As
a consequence the resulting concentration of the alcohol formed
reaches a level high enough to impair enzyme activity. In case of
immobilisation making use of TEOS, the concentration of ethanol
formed during the polycondensation reaction was determined to
be 14.5 mol/l, which is three times higher than the concentration
at which ethanol caused inactivation of l-TA could be observed
As a starting point immobilisation of l-TA using Eupergit®
C
was tested. The material is commercially available and well known
for efficient covalent enzyme immobilisation [23,24] However,
Eupergit C proved not to be suitable for l-TA from E. coli as immo-
®
bilisation yields (based on activity) did not exceed 1%. The observed
immobilisation yields are contrary to results obtained in compar-
ative experiments with l-TA from T. maritima for which activity
based immobilisations yields reached 47% applying the same con-
ditions (Scheme 2). The latter is in good accordance with the values
reported by Fu et al. (52% immobilisation yield) [25]. The rea-
sons for the observed differences in immobilisation yields may be
attributable to catalytically important surficial amino groups that
are present and accessible for reaction in the E. coli l-TA but not in
the T. maritima l-TA. However further research is needed to fully
understand the observations made.
As a consequence alternative carriers needed to be sought. An
interesting alternative to Eupergit® C is chitosan, a non-toxic and
inexpensive biopolymer [26]. The polyaminosaccharide is accessi-
ble to functionalisation with glutaraldehyde thus placing aldehyde
moieties on the surface with the aim of establishing covalent
bonds with surficial amino groups (␧-amino- or N-terminal-amino
groups) of l-TA [27].
To achieve immobilisation of l-TA two strategies were pursued:
(
i) linkage of l-TA to chitosan pre-treated with glutaraldehyde and
(
ii) cross-linking the enzyme with non-pre-treated chitosan in solu-
tion to which the aldehyde is added. In addition, direct cross-linking
of l-TA using only glutaraldehyde (CLEA, CLEC approach) was also
tested. In fact none of the strategies proved to be satisfactory. While
linking the enzyme to chitosan resulted in approx. 5% residual
activity, irrespective of the method applied, enzyme cross-linking
experiments were found to lead to full enzyme deactivation. The
observed maximal immobilisation yield (Table 1) implies that glu-
taraldehyde, fixed to chitosan or not, affects catalytically important
amino groups crucial for enzyme activity [28].
In addition, also physical interaction between enzyme and
the carriers surface may have contributed to the observed low
immobilisation yields. In some cases this phenomenon could be
avoided or its extent reduced by increasing the length of the spacer
group connecting carrier and enzyme [29]. On the other hand the
immobilisation tests with glutaraldehyde and Eupergit® C (glyci-
dol) clearly indicate that the enzyme is not sufficiently tolerant
RO
OR
HO
OH
condensation
poymerisation
+
4 H O
2
Si
Si
+ 4 ROH
SiO + 2 H O
2 2
RO
OR
HO
OH
Scheme 3. Schematic illustration of the hydrolysis and condensation reaction of alkyl orthosilicates.

6
S. Kurjatschij et al. / Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
method are depicted in Fig. 3. The presence of organic solvents in
the reaction mixture used for immobilisation even is beneficial as
it enables controlling pore size distribution in the resulting porous
silicate matrix [33]. Studies within the working group identified
n-hexane, cyclohexane, n-octane and n-decane as most promising
candidates [34].
Since organic solvents are known enzyme denaturants, only
enzyme compatible ones, i.e. with no or little effect on enzyme
activity, are suitable for this approach. Hence various organic
solvents have been screened. After 18 h of incubation with the
respective solvent at water/solvent ratios typically used for silicate
gel formation, residual activities were determined as given in Fig. 3.
After it had turned out that the presence of an organic solvent not
only had no detrimental effect on enzyme activity (with the excep-
tion of 25% ethanol in water), but it also became evident that 8 out
of 13 tested solvent/water mixtures displayed a considerable (up to
4
1%) promoting effect on the observed enzyme activity. The under-
lying reasons for these observations are not yet fully understood
and hence the matter of forthcoming investigations.
Fig. 2. SEM image of TEOS immobilised l-TA.
As can be seen from Fig. 3 there are opposing effects in
dependence of the water/solvent ratio. While n-decane promotes
threonine cleavage with increasing amounts of solvent present, the
opposite is observed for n-hexane, for which the aldolase catal-
ysed reaction already profits from a rather low fraction of the
organic solvent. On the other hand water miscible solvents such
as ethanol remain ineffective if concentrations were kept below
5% (i.e. 0.9 mol/l in the case of ethanol). At higher values the l-
TA is inactivated. Hence minimising the concentration of alcohols
formed during alkyl orthosilicate hydrolysis through extraction is
mandatory, regardless of the nature of the alcohol formed.
Consequently the solvent to be chosen should have an excellent
extraction capacity for the alcohol formed while not impairing pore
size distribution. BET measurements proved that none of the sol-
vents investigated (for the solvents tested see Fig. 3) affected pore
sizes significantly. Hence n-hexane was selected as it is very effec-
tive in extracting water soluble alcohols from an aqueous reaction
solution.
(
4.2 mol/l, Fig. 3). Experiments with non-immobilised enzyme con-
firmed these findings. After exposing l-TA to the same amount of
ethanol present during immobilisation a residual activity of 24%
was obtained compared to 15–27% found for the TEOS immobil-
isate.
Consequently ethanol needs to be removed from the reaction
mixture for which purpose two approaches were found to be viable:
(
1) ethanol removal by vacuum evaporation [32] and (2) ethanol
extraction with organic solvents.
Attempts to remove the alcohol by vacuum evaporation resulted
in increased polycondensation rates in the course of preparing
the sol solution. The almost instantaneous solidification prevented
the enzyme from being fully entrapped into the gel causing lower
immobilisation yields.
However the method could still be improved by switching
from TEOS to TMOS (tetramethyl orthosilicate). Using the latter,
methanol instead of ethanol is released during immobilisation
towards which l-TA is more stable (Fig. 3). The rather disadvanta-
geous ecopotential of methanol was considered negligible in view
of the amounts released in the course of the production of a cat-
alyst material, in particular in view of the higher stability of l-TA
towards methanol.
Optimisation efforts led to a maximum enzyme loading of 50 U/g
carrier depending upon the crude enzyme extract.
The TMOS immobilisate obtained with the optimised immobili-
sation method showed no observable loss in activity upon storage
◦
for ten weeks at 4 C, whereas the non-immobilised enzyme lost
A more straightforward alternative to evaporation of the alcohol
formed during alkylorthosilicate hydrolysis under reduced pres-
sure is extraction of the by-product using a water immiscible
organic solvent. The results of experiments making use of this
up to 16% of its activity already after 6 weeks.
In the course of investigating the loading capacity of different
sol/enzyme combinations protein leaching of the immobilisates
was examined (Fig. 4). For this purpose the immobilisate was
Fig. 3. Relative activity [%] of native enzyme after 18 h of incubation in different organic solvent/water mixtures with organic solvent contents of 5 and 25%.

S. Kurjatschij et al. / Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
7
Table 2
Activity of the supernatant
Comparison of biocatalytic conversions with native enzyme and immobilised
enzyme.
6
4
2
0
0
0
0
Compound l-TA
Conversion
%)
d.r.
ee (threo)
(
(threo/erythro)
3a
Native enzyme
50
12
7
70
20
63:37
46:54
67:33
80:20
77:23
60:40
>99%
>99%
>99%
>99%
>99%
>99%
3
3
3
3
3
a
a
b
b
b
Native enzyme
Immobilised enzyme
Native enzyme
Native enzyme
Immobilised enzyme 17
2
.4. Biocatalytic conversion
1
2
3
4
5
6
7
8
9
10 11 12
After the successful immobilisation of l-TA its applicability in
conditions of immobilisation
biocatalytic reactions yielding dichiral compounds in one step
needed to be tested. A reaction known to be catalysed by l-TA is
the reaction of benzaldehyde (1a) and o-chlorobenzaldehyde (1b)
yielding phenylserine and o-chlorophenylserine respectively [15]
experiment TEOS (v/v) TRIEMOS (v/v) TMOS (v/v)
enzyme solution
10
1
2
3
4
5
6
7
8
9
0,5
1
0
0,5
1
0
0,5
1
0
0,5
1
0
0,5
0
0
0,5
0
0
0,5
0
0
0,5
0
0
0
0
1
0
0
1
0
0
1
0
0
1
10
10
1
(
Scheme 4). A comparison of the conversion as well as the diaste-
reomeric and enantiomeric purity of the products obtained from
reactions using native and immobilised l-TA is shown in Table 2.
From Table 2 it can be seen that for both aldehyde starting mate-
rials the conversion obtained with immobilised enzyme is lower
than the one observed with native enzyme over the same period of
time. On the other hand diastereoselectivity is enhanced for 1a and
it is lower for 1b. This behaviour is owed to the nature of the sur-
face of the silicatic entrapment material. Through intramolecular
hydrogen bonds the inner surface of the immobilisate is rather
hydrophobic. Consequently both – the aldehydes and the aldol con-
densation products – will display a tendency of being adsorbed to
the silicatic matrix. Thus, the observed rate of conversion is appar-
ently lower when compared to the rate observed in reactions using
non-immobilised l-TA since a part of the product is not in solution
but adsorbed to the silicatic matrix.
Consequently the tendency of the substrate to adsorb to the
silicatic matrix will determine the rate of its transfer to the
enzyme. This interpretation is supported by experiments investi-
gating adsorption of hydrophobic molecules such as benzaldehyde
1a) onto the silicate matrix. For 1a it was found that it is extracted
from the aqueous solution and adsorbed to a large extent (up to
8%) by the silicatic matrix. The same effect could be observed in
1
1
3
3
3
1
1
1
0
1
2
5
5
5
Fig. 4. Residual enzyme activities (relative activity based on native enzyme activity)
in supernatants of different immobilisates.
suspended in a buffer/glycerol/PLP solution and the supernatant
was tested for aldolase activity (Fig. 4). Again immobilisates pro-
duced by using TMOS for immobilisation showed the best results.
Besides enzyme leaching mechanical stability of an immobil-
isate is an important criterion especially if the biocatalytic reaction
catalysed by the immobilised enzyme is done in a stirred tank. In
order to evaluate the mechanical stability of the porous immobil-
isate particle sizes were investigated using a Beckmann Coulter
particle analyser. Fig. 5 shows a slight increase in particle size
from d50 = 155.8 ␮m to d50 = 207.9 ␮m within a period of 14 h. This
increase is explained by the swelling behaviour of the silicate.
(
6
experiments using zeolites which are well known to form intra-
molecular hydrogen bonds rendering the inner surfaces of these
Fig. 5. Exemplified particle size distribution of an immobilisate (Fig. 4, experiment 5) before (blue bars) and after 14 h. (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of the article.)

8
S. Kurjatschij et al. / Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
O
O
OH
*
O
TA
PLP, pH 8, rt
OH
OH
R
R
+
NH2
NH2
100 mM
1 M
H
1a
H
3a L-threo /L-erythro
2
R =
R =
o-Cl 1b
o-Cl 3b L-threo / L-erythro
Scheme 4. Biotransformation of glycine with benzaldehyde (1a) and o-chlorobenzaldehyde (1b).
silicates hydrophobic. In particular 1a was adsorbed almost quan-
titatively (≥95%) by zeolites [35,36].
4.2. Instrumentation
The adsorption properties of benzaldehyde on the inner sur-
face of the immobilisate are also suited to explain the changes
in diastereoselectivities observed in comparison with the native
enzyme. It is obviously the residence time in the system and the
time allowed for equilibration which eventually determines prod-
uct diastereopurity. Thus the enzyme displays only an insufficient
capability to discriminate between threo and erythro product for-
mation. Eventually it is the repeated course of product formation
and cleavage which after all leads to the diastereomer distribution
observed in the equilibrated state.
Analytical HPLC was carried out with a Dionex Ultimate
3000 system. The separation was achieved on
a
Chirex®
3126 D-penicillamine chiral column (250 mm × 4.6 mm) from
®
Phenomenex .
GC–MS separation and analysis were performed on a Varian
431-GC Gas Chromatograph and a 220-MS IT Mass Spectrometer
system.
Photometric measurements were conducted on a Jasco V-630
spectrophotometer.
Particle size distribution and the swelling capacity of the immo-
bilisates were investigated using a Beckmann Coulter equipped
with a LS 13320 Laser Diffraction, a universal liquid module (ULM)
and a sonication control unit.
The SEM images were produced by a Tescan REM/EDT coupled
with EDX INCAx-sight from Oxford Instruments and a Cressington
Sputter Coater 108anfo.
On the hydrophobic surfaces there is only a retardation effect
observable leading to prolonged residence times. There is no indi-
cation for a process taking effect on product stereopurity (3a).
Moreover, longer residence times provide sufficient time for the
reaction equilibrium to establish, and it is this residence time in the
immediate proximity of the enzyme which in this way takes effect
on diastereomer distribution. Consequently the situation is differ-
ent for 1b, where the compound exhibits a higher dipole moment,
is more polar than 1a, adsorbs less effectively and consequently
shows higher conversion within the same time frame, yet accom-
panied by lower diastereoselectivity (d.r. = 60:40).
In order to circumvent these intricacies caused by hydrophobic-
ity of the inner surface of the entrapment material, attempts were
made to modify the surface by inserting a methyl group through
reacting trialkylmethoxyorthosilicates with TMOS instead of the
pure orthosilicate. However, these immobilisates showed a poorer
solidification behaviour and a higher leaching potential (Fig. 4) due
to less efficient formation of three-dimensional silicate network.
For these reasons this option was not further pursued.
4.3. Activity assays
The native enzyme activity was determined by a 2 step coupled
assay. At first threonine is cleaved by the threonine aldolase to give
acetaldehyde and glycine. Parallel to this the evolving acetalde-
hyde is reduced to ethanol by a yeast alcohol dehydrogenase with
simultaneous oxidation of NADH to NAD*. The resulting decrease
of absorbance was recorded spectrophotometrically at 340 nm.
Experiments were carried out in a standard cuvette with a total
◦
solution volume of 1 ml at 30 C. The solution consisted of 100 ␮l
1
1
3
M buffer pH = 8.0, 20 ␮l 2.5 mM PLP, 100 ␮l 0.5 M threonine, 20 ␮l
0 mM NADH (H O), 10 ␮l threonine aldolase crude extract, 10 ␮l
2
U/␮l ADH and 720 ␮l H O. One unit of TA is defined to catalyse the
2
formation of 1 ␮mol of acetaldehyde (corresponds to the oxidation
of NADH) per min at 30 C.
3
. Conclusion
◦
The activity of immobilised enzyme was obtained based on the
TA catalysed retro-aldol reaction of phenylserine giving glycine and
benzaldehyde. The formed aldehyde was extracted with chloro-
form from the aqueous solution and quantified by means of GC–MS
measurements. In this case one unit of TA is defined to catalyse
Threonine aldolases can efficiently be immobilised by entrap-
ment into silicatic matrices. The reaction behaviour of these
immobilisates is determined by the hydrophobicity of their inner
surfaces. The more hydrophobic the latter, the slower the biocat-
alytic aldol condensation will proceed. Further activities will be
directed towards optimising diastereoselectivity of the l-TA catal-
ysed reaction.
◦
the formation of 1 ␮mol of benzaldehyde per min at 30 C. The
reaction mixture contained 20 ␮l 2.5 mM PLP, 100 ␮l 1 M buffer
pH = 8.0, immobilised enzyme, 980 ␮l 0.05 M phenylserine (dis-
solved in buffer pH = 8.0). The aqueous phase was extracted with
4
. Experimental
8
00 ␮l chloroform for 1 min and centrifuged for 1 min in order to
separate both phases. 1 ␮l of the organic layer was then injected
into the GC system and separated on a Varian column VF-35ms
4.1. Materials
(
1
30 m × 0.25 mm × 0.39 ␮m). The flow of the carrier gas Helium was
All chemicals and solvents were of appropriate purity and were
obtained from commercial sources or synthesised in our laborato-
ries.
◦
.0 ml/min and the temperature of the injector was 250 C. The tem-
◦
◦
◦
◦
perature programme 60 C (0 min); 60–90 C with 3.0 C/min; 90 C
5 min) gave a total of 15 min and the following retention time for
(
l-TA in the form of crude extracts of E. coli cells overexpressing
l-TA was kindly provided by the research group of Prof. Hummel
benzaldehyde 9.2 min. The analytes were ionised via EI and the MS
acquisition method recorded mass to charge values in the range of
(
University of Düsseldorf) and evocatal [16].
4
0–150.

S. Kurjatschij et al. / Journal of Molecular Catalysis B: Enzymatic 103 (2014) 3–9
9
4.4. Immobilisation procedures
[2] G.-H. Zhao, H. Li, W. Liu, W.-G. Zhang, F. Zhang, Q. Liu, Q.-C. Jiao, Amino Acids
0 (2011) 215–220.
4
[
[
3] A.K. Samland, G.A. Sprenger, Appl. Microbiol. Biotechnol. 71 (2006) 253–264.
4] M. Brovetto, D. Gamenara, P. Saenz Méndez, G.A. Seoane, Chem. Rev. 111 (2011)
4346–4403.
The immobilisation procedure is composed of two steps, the
preparation of the different silica sol stock solutions and the addi-
tion of crude enzyme extract and subsequent gelation. The reaction
mixture comprises of 0.02 mol of the correspondent silicate (4.5 ml
TEOS, 3.95 ml TMOS, 3.98 ml TRIEMOS, 2.85 ml TRIMMOS), 0.4 ml
[
[
5] T. Kimura, V.P. Vassilev, G.-J. Shen, C.-H. Wong, J. Am. Chem. Soc. 119 (1997)
11734–11742.
6] N. Dückers, K. Baer, S. Simon, H. Gröger, W. Hummel, Appl. Microbiol. Bio-
technol. 88 (2010) 409–424.
ultrapure H O and 1 ml 0.01 M HCl [37]. The initial two phase
[7] K. Baer, N. Dückers, W. Hummel, H. Gröger, ChemCatChem 2 (2010) 939–942.
[8] V.P. Vassilev, T. Uchiyama, T. Kajimoto, C.-H. Wong, Tetrahedron Lett. 36 (1995)
2
◦
system is stirred vigorously at 40 C until a homogenous sol is
5063–5064.
obtained. According to the different experiments sol stock solution
and enzyme solution are mixed together until solidification occurs.
[10] J.-Q. Liu, T. Dairi, N. Itoh, M. Kataoka, S. Shimizu, H. Yamada, J. Mol. Catal. B:
Enzym. 10 (2000) 107–115.
[11] C. Mateo, J.M. Palomo, G. Fernandez-Lorente, J.M. Guisan, R. Fernandez-
Lafuente, Enzyme Microb. Technol. 40 (2007) 1451–1463.
12] R.A. Sheldon, S. van Pelt, Chem. Soc. Rev. 42 (2013) 6223.
[13] Ulf Hanefeld, Chem. Soc. Rev. 38 (2009) 453–468.
◦
The gel is stored for 12 h at 4 C for ageing purposes. In order to mod-
ify the porous structure the sol and enzyme solution are added to
an organic solvent (n-hexane, cyclohexane, n-octane, decane) and
mixed well until solidification.
[
[
[
14] A. Liese, L. Hilterhaus, Chem. Soc. Rev. 42 (2013) 6236.
15] K. Baer, N. Dückers, T. Rosenbaum, C. Leggewie, S. Simon, M. Kraußer, S. Oßwald,
W. Hummel, H. Gröger, Tetrahedron: Asymmetry 22 (2011) 925–928.
16] J.-Q. Liu, T. Dairi, N. Itoh, M. Kataoka, S. Shimizu, H. Yamada, Eur. J. Biochem.
[
4
.5. Biotransformation
255 (1998) 220–226.
[
17] R.A. Sheldon, Org. Process Res. Dev. 15 (2011) 213–223.
The reaction mixture consisted of 1.375 ml of crude protein
[18] C.L. Kielkopf, S.K. Burley, Biochemistry 41 (2002) 11711–11720.
[
19] J. Li, Z. Jiang, H. Wu, L. Long, Y. Jiang, L. Zhang, Compos. Sci. Technol. 69 (2009)
39–544.
20] R. Reshmi, G. Sanjay, S. Sugunan, Catal. Commun. 7 (2006) 460–465.
extract (≥10 U/ml) or water respectively (immobilisates), glycine
5
2
(3 mmol), aldehyde 1a (1b) (0.3 mmol), 1.375 ml glycerol, 300 ␮l
[
buffer (1 M, pH = 8.0) and pyridoxal-5-phosphate (50 ␮M). The
different reaction mixtures were stirred at 25 C for different
[21] M. Basri, K. Ampon, W.M.Z. Wan Yunus, C.N.A. Razak, A.B. Salleh, J. Am. Oil
Chem. Soc. 72 (1995) 407–411.
◦
[
22] F. López-Gallego, L. Betancor, A. Hidalgo, N. Alonso, G. Fernandez-Lorente, J.M.
Guisan, R. Fernandez-Lafuente, Enzyme Microb. Technol. 37 (2005) 750–756.
periods of time. Conversion, diastereo- and enantioselectivity were
determined by analytical HPLC. Therefore samples (100 ␮l) were
withdrawn from the reaction mixture and the enzymatic reaction
was stopped via acidic denaturation. This was followed by a dilu-
tion of 1:10 with a water/methanol mixture to give the final ratio of
[23] M.H.A. Janssen, L.M. van Langen, S.R.M. Pereira, F. van Rantwijk, R.A. Sheldon,
Biotechnol. Bioeng. 78 (2002) 425–432.
[
[
24] M. Katchalski-Katzir, D.M. Kraemer, J. Mol. Catal. B: Enzym. 10 (2000) 157–176.
25] H. Fu, I. Dencic, J. Tibhe, C.A. Sanchez Pedraza, Q. Wang, T. Noel, J. Meuldijk, M.
de Croon, V. Hessel, N. Weizenmann, et al., 22nd International Symposium on
Chemical Reaction Engineering (ISCRE 22), vol. 207/208, 2012, pp. 564–576.
26] B. Krajewska, Enzyme Microb. Technol. 35 (2004) 126–139.
2
5% methanol and 75% water. The solvent consisted of 75% 0.2 mM
[
[
CuSO ·5H O and 25% methanol. Additional measurement parame-
4
2
27] D.S. Rodrigues, A.A. Mendes, W.S. Adriano, L.R. Gon c¸ alves, R.L. Giordano, J. Mol.
Catal. B: Enzym. 51 (2008) 100–109.
◦
ters were flow rate: 1 ml/min, column oven temperature: 40 C and
detection: variable wavelength detector at 254 nm.
[28] C. Mateo, J.M. Palomo, L.M. van Langen, F. van Rantwijk, R.A. Sheldon, Bio-
technol. Bioeng. 86 (2004) 273–276.
[
29] L. Cao, Carrier-Bound Immobilized Enzymes. Principles, Applications and
Design, Wiley-VCH, Weinheim, 2005.
Acknowledgment
[30] A. Blandino, M. Macias, D. Cantero, Process Biochem. 36 (2001) 601–606.
[
[
[
31] A. Pierre, Biocatal. Biotransform. 22 (2004) 145–170.
32] M.L. Ferrer, F. del Monte, D. Levy, Chem. Mater. 14 (2002) 3619–3621.
33] A.J. Bush, R. Beyer, R. Trautman, C.J. Barbé, J.R. Bartlett, J. Sol-Gel Sci. Technol.
32 (2004) 85–90.
Support by the German Federal Environmental Foundation
(
Deutsche Bundesstiftung Umwelt, DBU) (Project # 13217-32) is
[
[
34] F. Simon, Diplomarbeit, Technische Universität Bergakademie Freiberg,
Freiberg, 2008.
35] S. Kropp, Diplomarbeit, Technische Universität Bergakademie Freiberg,
Freiberg, 2013.
gratefully acknowledged.
References
[36] E. Hain, Diplomarbeit, Technische Universität Bergakademie Freiberg, Freiberg,
011.
37] J.R. Premkumar, E. Sagi, R. Rozen, S. Belkin, A.D. Modestov, O. Lev, Chem. Mater.
4 (2002) 2676–2686.
2
[
[1] M.L. Gutierrez, X. Garrabou, E. Agosta, S. Servi, T. Parella, J. Joglar, P. Clapés,
1
Chem. Eur. J. 14 (2008) 4647–4656.