A Rapid Selection Procedure for Simple Commercial Implementation of ω-Transaminase Reactions

Article abstract of DOI:10.1021/acs.oprd.5b00159

A stepwise selection procedure is presented to quickly evaluate whether a given ω-transaminase reaction is suitable for a so-called "simple" scale-up for fast industrial implementation. Here "simple" is defined as a system without the need for extensive process development or specialized equipment. The procedure may be used when investment in intensive process development cannot be justified or when rapid execution is paramount, for applications such as small singular batches. The three-step evaluation procedure consists of: (1) thermodynamic assessment, (2) biocatalyst activity screening, and (3) determination of product inhibition. The method is exemplified with experimental work focused on two products: 1-(4-bromophenyl)ethylamine and (S)-(+)-3-amino-1-Boc-piperidine, synthesized from their corresponding pro-chiral ketones each with two alternative amine donors, propan-2-amine, and 1-phenylethylamine. Each step of the method has a threshold value, which must be surpassed to allow "simple" implementation, helping select suitable combinations of substrates, enzymes, and donors. One reaction pair, 1-Boc-3-piperidone with propan-2-amine, met the criteria of the three-step selection procedure and was subsequently run at 25 mL scale synthesizing (S)-(+)-3-amino-1-Boc-piperidine at concentrations up to 75 g/L. However, the highest product yield (70%) was obtained at a lower substrate concentration of 50 g/L.

Full text of DOI:10.1021/acs.oprd.5b00159

Article
pubs.acs.org/OPRD
A Rapid Selection Procedure for Simple Commercial Implementation
of ω‑Transaminase Reactions
Maria T. Gundersen,† Par Tufvesson,† Emma J. Rackham,‡ Richard C. Lloyd,‡ and John M. Woodley*,†
̈
†Department of Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Lyngby, Denmark
‡Dr. Reddy’s Chirotech Technology Centre, 410 Cambridge Science Park, Milton Road, CB4 0PE, Cambridge, U.K.
ABSTRACT: A stepwise selection procedure is presented to quickly evaluate whether a given ω-transaminase reaction is
suitable for a so-called “simple” scale-up for fast industrial implementation. Here “simple” is defined as a system without the need
for extensive process development or specialized equipment. The procedure may be used when investment in intensive process
development cannot be justified or when rapid execution is paramount, for applications such as small singular batches. The three-
step evaluation procedure consists of: (1) thermodynamic assessment, (2) biocatalyst activity screening, and (3) determination of
product inhibition. The method is exemplified with experimental work focused on two products: 1-(4-bromophenyl)ethylamine
and (S)-(+)-3-amino-1-Boc-piperidine, synthesized from their corresponding pro-chiral ketones each with two alternative amine
donors, propan-2-amine, and 1-phenylethylamine. Each step of the method has a threshold value, which must be surpassed to
allow “simple” implementation, helping select suitable combinations of substrates, enzymes, and donors. One reaction pair, 1-
Boc-3-piperidone with propan-2-amine, met the criteria of the three-step selection procedure and was subsequently run at 25 mL
scale synthesizing (S)-(+)-3-amino-1-Boc-piperidine at concentrations up to 75 g/L. However, the highest product yield (70%)
was obtained at a lower substrate concentration of 50 g/L.
Scheme 1. Examples of Potential ω-TA Reactions Using (A)
a Synthetic Route and (B) a Resolution Route
1. INTRODUCTION
Over the past decade, biocatalysis has become an established
and useful complement to conventional chemical catalysis for
the synthesis of fine chemicals. Most often, biocatalytic
methods have been selected due to exceptional selectivity
(regio- and/or enantioselectivity).1 In fact, the majority of
industrially applied biocatalytic reactions today yield optically
pure chiral products that are used in the fine chemical industry
as building blocks for agrochemicals and pharmaceuticals.2 In
particular, biocatalytic transamination chemistry has been
identified as one of the key emerging areas for the
pharmaceutical industry1,3 as a means of producing optically
pure chiral amines. This paper focuses on the biocatalytic
synthesis (and resolution) of chiral amines of high optical
purity using ω-transaminase (ω-TA) (E.C. 2.6.1.18), which is a
type of amino transferase. ω-TA was chosen as a catalyst for
this work due to its outstanding stereoselectivity and broad
ketone substrate repertoire. Two ω-TA-catalyzed paths are
available toward optically pure chiral amines, using either
asymmetric synthesis or kinetic resolution. Although the latter
is challenged by a maximum 50% yield,4−6 both are considered
as potential options for the “simple” scale-up.
ω-TA catalyzes an amino transfer reaction, illustrated in
Scheme 1. Briefly, in the synthetic direction (Scheme 1A) the
amino donor (an amine), and the amino acceptor (a prochiral
ketone), here referred to simply as the “donor” and “acceptor”,
respectively, react with the enzyme in a sequential fashion
producing the desired target chiral amine product and a
coproduct. Detailed descriptions of the sequential ping-pong bi-
bi enzymatic reaction mechanism can be found elsewhere.7,8 In
the resolution reaction (Scheme 1B) the same reaction takes
place, but now the amino donor is added as a racemic mixture.
Through reaction therefore, one isomer is left unreacted, which
becomes the desired optically pure product.
The amino moiety alone is transferred between the two
starting substrates, and therefore in the synthetic direction, the
molecular structure of the chiral product will be determined by
the structure of the acceptor molecule. This means that the
donor molecule can be freely chosen, since it neither affects the
target product structure nor the stereoselectivity. In principle
therefore, a plethora of possible donors could be chosen,
although in the scientific and patent literature only a handful of
amine donors have been reported. The authors have recently
proposed a novel quantum mechanical method to determine
the free energy of compounds and hence the thermodynamic
feasibility of using novel amino donors for this reaction type,
irrespective of kinetic considerations.9 This along with a wider
implementation of this technology in the future is likely to lead
to a broader range of different amino donors.
Received: May 21, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
a
Table 1. Compounds Used in This Study
a
Acceptor ketones: 4-bromoacetophenone (1), 1-Boc-3-piperidone (2); amino donors: propan-2-amine (3), 1-phenylethylamine (chiral) (4); target
chiral amine products: 1-(4-bromophenyl)ethylamine (chiral) (5), 3-amino-1-boc-piperidine (chiral) (6); coproducts: acetone (7) and
acetophenone (8).
Despite the interest in such reactions, they are often
demanding to implement on an industrial scale due to frequent
thermodynamic and kinetic challenges.10 While many technical
solutions are available to overcome these challenges, the
proposed solutions are frequently complex and often require
significant process development time. Indeed, for some
applications, a fast and simple process development is not
only desirable but may be essential for commercial success. In
these cases, it will be more important to rapidly develop a
simple process, than to obtain an economically optimal process.
Such situations include pharmaceutical synthesis in the early
phases of clinical testing and other cases such as small singular
batches, where investment in extensive process development
cannot be justified. Using this logic, and from a knowledge of
the properties of a given ω-transaminase-catalyzed reaction and
the available enzymes to catalyze the reaction, we reasoned that
it should be possible to categorize a particular reaction as
“complex” (requiring extensive development) or “simple” (with
easy implementation and scale-up). We therefore suggest that
an evaluation method allowing the identification and selection
of “simple” reactions (and eliminating the “complex” ones)
would prove a valuable tool for process chemists.
The scope of this manuscript is therefore to present a
stepwise decision-making procedure to quickly identify if a
“simple” scale-up is feasible for a given reaction. Hence,
solutions such as biocatalyst modification by protein engineer-
ing,10 amino donor recycling,11,12 and equilibrium shifting
methods13,14 have not been considered here. The three-step
decision-making procedure involves an evaluation of: (1)
thermodynamics, (2) biocatalyst activity, and (3) product
inhibition. Each step is evaluated against a threshold value,
which must be met in order to identify a given case as suitable
for “simple” implementation.
2. RESULTS
In order to exemplify the method, experimental data on two
chiral target products were evaluated, 1-(4-bromophenyl)-
ethylamine (5) and (S)-(+)-3-amino-1-Boc-piperidine (6).
These compounds were selected because both products are
commercially attractive and additionally biocatalytic trans-
aminations to synthesize both 515 and 611,16 have been
reported previously. In these reactions optical purity was
necessary and evaluated, but a specific stereoisomer was not
required. The prochiral ketone substrates, the amino acceptors,
4-bromoacetophenone (1) and 1-Boc-3-piperidone (2), corre-
sponding to the products above, were reacted with two possible
donor molecules propan-2-amine (3) and 1-phenylethylamine
(4) (Table 1). Both amino donors have frequently been used in
a wide variety of biocatalytic transamination. Between them,
they represent different classes of donor. For instance, donor 3
serves as an inexpensive achiral donor. In contrast amino donor
4 is a more costly chiral compound which has also been
reported to be inhibitory17 with downstream processing
complications due to separation issues when the product
shares structural similarity. However, donor 4 also offers a
significant thermodynamic advantage, since the carbonyl
coproduct, acetophenone (8) formation is highly favorable
(Table 1). Academically, a more common amino donor that has
often been reported is the use of alanine (or pyruvate for the
resolution reaction).4,18 We have previously shown that the
thermodynamics using this donor very strongly favors the
reverse resolution reaction,19 and therefore we have not
considered this further in this work.
2.1. Method Development. In order to enable rapid
evaluation, the three selection criteria are each assigned a
threshold value, which must be met to enable implementation
of a simple scale-up. The proposed procedure is outlined in
Figure 1. In the figure, full lines indicate that the reaction has
met (green lines) or failed (red lines) the individual criteria.
Likewise, dashed lines indicate an alternative strategy by
B
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
it is more accurate to describe the constant as “apparent”,
Keqapp. In principle to save time as an alternative to
experimental measurement, in silico methods could be used
to estimate such values, although the accuracy is perhaps
questionable. Here the Keqapp for the two chiral amine products
5 and 6 were measured experimentally using the two donors 3
app
and 4, as described above. The Keq for the four reactions
(Table 2) varied by a factor of 104 from the most challenging
ap
Table 2. Experimental Values for Keq
donors
acceptors
3
4
a
1
2
0.025
32
0.5
450
a
Data previously reported in ref 13.
Figure 1. Decision making procedure for a simple scale-up. Green
lines marked with a check mark or red lines marked with an X indicate
if a given criterion is met or not met, respectively. Dashed lines and
boxes indicate options for reassessment if a criterion is not met. Each
criteria has cut off values for simple implementation. 1. The
thermodynamic criteria is meet when Keq is less than 0.02 (resolution
reactions) or greater than 1 (synthetic direction). 2. The activity
criterion requires a specific activity greater than 0.05 g/g/h. 3. The
inhibition criteria is met at less than 50% activity loss, with 5% of
target concentration product present. Possible remediation options, if
a given criterion for a simple scale-up is not met, can be to consider an
alternate amino donor or test an alternative biocatalyst (dashed lines).
pair, 1 and 3, at 0.025 to the most favorable pair, 2 and 4, which
app
had a Keq
of 450, in the synthetic direction. Thus,
thermodynamics is indeed highly variable between the four
selected reaction pairs. After applying the threshold criteria one
of the two products, 5, was eliminated from further
investigation. This may indicate that highly conjugated aryl
compounds are not suitable for simple scale up and should be
assisted by other process technologies and strategies. For
example, it has been reported that one of the compounds we
have used as a donor here 4, could also synthesized and
successfully scaled in combination with in situ product removal,
alleviating both the thermodynamic and inhibitory strains.14
None of the reaction pairs evaluated here was found suitable for
the resolution reaction, although alanine, the amine donor
often found most suited for resolutions reactions was not tested
as discussed previously.19
Clearly it is possible to carry forward more than one amine
donor to the subsequent evaluation steps, although this is not
helpful for the procedure, which aims to focus effort on those
cases with the biggest chance of simple scale-up success. In this
case, due to the low cost and high water solubility, amine donor
3 was selected for further evaluation.
2.1.2. Biocatalyst Activity Screening. No matter how
favorable the thermodynamics, without sufficient activity the
reaction will not be completed in a reasonable time, and issues
like enzyme inactivation may arise. Hence, the next step of the
procedure is to find a suitable biocatalyst with sufficient activity.
Candidates for biocatalyst screening can be obtained from
commercial screening kits or in-house enzymes. For the
“simple” scale-up, strategies such as protein engineering are
not considered. Low activity of an enzyme preparation will
negatively impact downstream processing, by adding extra
proteinaceous material which impedes product recovery.
Therefore, the maximum biocatalyst loading was set to 10%
v/v irrespective of the biocatalyst formulation. Additionally,
product concentration should be in the range of ≥50 g/L6,10 to
assist downstream product recovery. Finally, due to biocatalyst
stability concerns, we reasoned it necessary to complete the
reaction within 96 h. On this basis, we calculated a minimal
biocatalyst specific activity (sometimes termed “biocatalyst
productivity”), as a threshold value for the “simple” of 0.05 g/
g/h (g product/g biocatalyst/hour).
adjusting one of the variable reaction components (the amino
donor or the biocatalyst). The threshold values for each
criterion are also indicated in the legend of Figure 1, the
justification for which is given in the following section.
2.1.1. Thermodynamic Assessment. Unfavorable thermody-
namics presents one of the main barriers to the implementation
of the transaminase−catalyzed reactions on an industrial scale.4
The thermodynamic equilibrium constant (Keq) of the reaction
is important since it determines the maximum reaction yield for
a given concentration of substrates. Thus, we reasoned it is one
of the most important parameters for determining the optimal
process configuration.6,13,20 For this reason, we suggest the first
step in the procedure should be to determine if a candidate
reaction has a suitable thermodynamic equilibrium constant to
make a “simple” scale-up feasible.
In the synthetic mode, thermodynamic feasibility is here
defined as a Keq above 1.0, since lower values of Keq would
require a high excess (more than 20-fold) of the amino donor
to obtain sufficient reaction yields (95% or higher), for eventual
industrial implementation. Use of such an excess makes the
reaction costly and practically difficult to carry out at high
substrate concentrations. In a similar way, we reasoned that for
reactions with a low Keq, a kinetic resolution would be a better
choice for the reaction. On the other hand, the resolution
reaction requires more stringent conversion requirements since
the separation of the amine product from the unreacted half of
the racemic donor starting material is of course quite
challenging. Hence we have chosen a Keq threshold of 0.02 in
the resolution direction, meaning only values lower than this
are suitable for a simple scale-up.
In this work, the concentration-based equilibrium constant
was experimentally determined using a previously described
method.20 Since the value is obtained for comparative purposes,
practical (rather than standard) conditions were used, meaning
For this case study a small screen with four enzymes was
conducted, using the reactant pair 2 and 3 selected from the
previous section. In this screen four selected enzymes were
tested, two of which were known to be (R)-selective and two
C
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(S)-selective. We reasoned that for this case study the particular
stereoselectivity of the enzyme did not influence the overall
procedure. Additionally, since this screen was conducted with
an achiral amine donor and the pro-chiral ketone, the selectivity
of the enzyme would not affect the reactivity with these
substrates. The screen showed a large variation between the
least and most reactive candidates (Table 3). Details of the
the threshold for the enzyme is defined as minimum specific
activity, which for the “simple scale-up” was set at 0.05 g/g/h.
Biocatalyst recycle was not considered for the simple scale-up.
Finally a determination of product inhibition is carried out.
This is a frequent hurdle for process intensification of ω-TA’s,14
due to the high product concentrations (50 g/L) required to
simplify the product recovery.2 In contrast to the high
concentration intensity of commercial processes, enzymes are
designed to work under physiological (dilute) conditions. This
frequently leads to process intensification challenges with
biocatalytic reactions. For example transaminases display a
ping-pong bi-bi reaction mechanism, with two sequential half
reactions,8 and this type of reaction mechanism is often plagued
by inhibition from competitive dead-end complexes of products
bound to the apo-enzyme or the incorrect form of the holo-
enzyme. Hence, understanding the inhibition profile of a
potential product is vital in evaluating the possibility of a simple
scale-up. As such, we advocate that, if severe inhibitory effects
are observed with low product concentrations, it implies a high
risk of inhibition under process scale concentrations.
Table 3. Specific Activities Obtained with Reactant Pair 2
and 3, with Selected Enzymes
enzyme
selectivity
specific activity (g/g/h)
Ars-ωTA
Tar0
S
0.048
0.003
0.012
0.054
R
R
S
Tar1
ATA 47
individual enzymes (ATA-47, Tar0, Tar1, and Ars-ωTA) are
given in the experimental section of the paper. Tar0 was found
to give a specific activity of 0.003 g/g/h, whereas the best
candidate (ATA-47) gave a 20-fold higher value of 0.054 g/g/h.
ATA-47 was therefore carried to the next step. Likewise the
enzyme Ars-ωTA had a high specific activity of 0.048 g/g/h,
close to the threshold value.
2.1.3. Determination of Product Inhibition. The final step
of the procedure considers product inhibition of the enzyme,
which due to the requirement for high product concentrations
(50 g/L) in industrial processes,2 is a frequent hurdle for
process intensification of enzyme reactions in general, and ω-
TAs in particular.14 Hence, we set the threshold value here at a
50% reduction in reaction rate in the presence of 2.5 g/L
product, under the assay conditions used here (see
Experimental Section). Here only product inhibition is
assessed, since substrate inhibition can relatively easily be
overcome by substrate feeding.
The three-step evaluation method has been successfully
applied to a case study, and one reaction pair with one
biocatalyst was deemed suitable for “simple” scale-up.
2.2. Intensification and Scale-Up. In the previous
sections, the selection procedure for a simple scale-up toward
the synthesis of 6 identified acceptor 2 with donor 3 (Scheme
2) using ATA-47 as suitable. In the event ATA-47 was
a
Scheme 2. Synthetic Transaminase Reaction Carried Out
In order to experimentally test for product inhibition, the
initial reaction rate of ATA-47 was measured using 100 mM 3
and 10 mM 2, in the presence of various concentrations of the
product 6, up to 10 mM. Importantly, the substrate
concentrations were chosen to avoid limiting the reaction by
thermodynamic constraints. Inhibition was observed with 10
mM product and amounted to a 10% initial rate reduction,
compared to initial conversion rates in the absence of product.
Initial conversion rates were assumed when less than 10% of
limiting starting material was converted.
2.1.4. Discussion. First, with respect to thermodynamics, the
procedure enables the elimination of unfavorable cases. Clearly
each donor or acceptor molecule has an associated free energy
which contributes to the net thermodynamics of a given
reaction. In this way for instance a comparison of the
equilibrium constants of two reactions (with different accept-
ors, but using the same donor) can be used to interpret the
effect of changing acceptors. In an analogous way, one could
determine the Keq for a given acceptor with one donor and
extrapolate the Keq to other donors with the same acceptor,
given one knows the difference in ΔG between the reactions, as
discussed elsewhere.19
a
Compounds: 1-Boc-3-piperidone (2), propan-2-amine (3), (S)-
(+)-3-amino-1-Boc-piperidine (S)-(+)-6), acetone (7).
substituted by ArS-ωTA since the difference in activity was
negligable and the latter enzyme has been reported to have
excellent stereoselectivity.21,22
2.2.1. Reaction Optimization: pH and Donor Loading.
Prior to scale-up, a small optimization study was undertaken to
evaluate if reaction rates could be enhanced by simple
optimization within the biocatalyst stability range. A range of
pH and donor loadings was explored in an attempt to improve
kinetics, with both short (0.5−2 h) and long (18 h) reaction
times; the latter time point was chosen to investigate enzyme
stability under the given conditions.
The rate dependency on pH was tested between pH 7 and 9,
with 40 mM acceptor and 500 mM donor (Figure 2). Other
studies have found up to 40% variation in yield in this pH range
for similar reactions.23 Here the fastest reaction rates were
identified at pH 9 for all time points. The greatest difference
was found in the 18 h reaction times, where average reaction
rates are 45% faster at pH 9 compared with pH 7, indicating
that this is the best pH, within the pH range tested, with respect
to kinetics, and that the enzyme is more stable under these
conditions. Since the pKa of the amine donor 3 is 10.6,24
meaning a higher pH would render a higher fraction of the
substrate uncharged and thus reactive, in principle operating at
a higher pH would therefore be beneficial from the perspective
Second, the biocatalyst activity is assessed, since low activity
will have drawbacks in the form of low space-time yields and
may prevent the reaction from going to completion due to
enzyme deactivation. One solution would be to apply high
biocatalyst concentration, but these may negatively impact
downstream processing by hindering product recovery. Thus,
D
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 4. Reaction profile over 96 h with initial substrate
concentrations of 25, 50, and 75 g/L.
Figure 2. Specific rates measured at four reaction time points at four
different reactions, each carried out at different pH values (pH 7.0, pH
7.5, pH 8, and pH 9 from left to right at each time point, respectively).
investigated the stability of this compound in water, either for
biocatalysis11 or chemical catalysis. In the 25 and 50 g/L
reactions final conversions of acceptor 2 to chiral amine target 6
of 70% were observed. Figure 5 shows that the initial reaction
rates are similar at all substrate concentrations tested, indicating
that the reaction is not kinetically controlled (above KM).
of the reaction rate. Nevertheless, in this study we limited the
pH range to keep the study simple and manageable, consistent
with the philosophy of this work, and therefore did not test the
reaction at higher pH values than 9.
Furthermore, the same method was used to determine
optimal donor loading. Donor concentrations could potentially
be limiting, dependent upon KM12. Clearly an excess
concentration of the donor (over acceptor) could be used
which might also drive the equilibrium.23,25 This was tested
experimentally but at all concentrations tested, the rate was
unaffected by donor concentration (Figure 3), suggesting a KM
beneath 100 mM. For subsequent experiments 1 M 3 was used.
Figure 5. Initial product formation for the first 12 h of the reaction
with initial substrate concentrations of 25, 50, and 75 g/L.
2.2.3. Product Identification. Finally, the reaction was run at
25 mL scale for 96 h to isolate product. At 50 g/L substrate
concentration the final reaction composition was analyzed to
contain 89% 6 and 11% 2 (with an isolated product yield of
around 70%). This composition is in excellent agreement with
that found in the 50 g/L 1 mL scintillation vial experiment,
which gave 91% target chiral amine 6 and 9% ketone 3, after 96
h.
Figure 3. Specific rates found at four time intervals at four donor
concentrations (0.1, 0.5, 1, and 2 M from left to right at each time
point, respectively) used to investigate the optimal donor loading for
the reaction.
2.2.2. Reaction Intensification. As indicated above a viable
scale-up depends on reaction intensification (i.e., the synthesis
of high product concentrations).10 This is important in the
simple scale-up because too low a concentration will add
volume to the reaction and thus complicate the process. The
reaction of 2 and 3 using Ars-ωTA was therefore intensified by
increasing the substrate concentration up to 75 g/L. Three
reactions were done in scintillation vials at concentrations of
25, 50, and 75 g/L. The reactions proceeded smoothly (Figure
4) at both 25 g/L and 50 g/L but not at 75 g/L, the latter most
likely due to mass transfer limitations from low solubility and
decomposition of the starting material in aqueous conditions.
The latter was further investigated and confirmed (data not
shown). To the best of our knowledge no other study has
3. CONCLUSION
A simple stepwise procedure has been described, to facilitate
the selection of suitable substrate-donor-enzyme combinations
to allow so-called “simple” scale-up. Each step in the procedure
has a threshold value which must be met to allow simple
implementation. We believe that this method will prove useful
both to select good candidates for this technology and to
eliminate those that may require further development. A simple
case study was used to illustrate the power of the procedure,
sequentially eliminating unsuitable substrates, donors, and
enzymes. Beyond this case study, we furthermore suggest that
analogous procedures could be used for the evaluation of other
“simple” biocatalytic processes.
E
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
4. EXPERIMENTAL SECTION
Article
NTA), Ars-ωTA was eluted with 500 mM imidazole. The
purified enzyme solution was exchanged into phosphate buffer
(0.1 M, pH 7) using an Amicon Ultra 15 Centrifugal filter (10
k) unit and SDS-PAGE analysis performed to confirm that the
Ars-ωTA had been purified (>85%) and concentrated (35 mg/
mL).
4.2.2. Reaction Conditions. Sections 2.1.1 and 2.1.3. Each
reaction, performed in duplicate, contained: 1 g L−1 ATA-47, 2
mM PLP, 5% DMSO, 0.1 M tris-HCl buffer pH 7.5, and up to
10 mM pro-chiral ketone acceptor 2 or 1.5 mM pro-chiral
ketone acceptor 1 together with 10 mM amino donor 4 or 100
mM 3, and was run for up to 48 h in 4 mL reaction vessels, at
30 °C in a thermos-shaker. Keq values were determined by
measuring conversion at varying concentrations of substrates
and products according to a previously described protocol.20
Inhibition studies were made by measuring initial rates (less
than 10% of the limiting substrate consumed), in the presence
of increasing product.
Section 2.2.1. All samples were carried out in 0.5 mL
reactions in a 96 well plate format. Short reactions of 30 min, 1
h, and 2 h were run with 20 g/L lyophilized cells of Ars-ωTA;
the 18 h reactions were run with 2 g/L lyophilized cells. All
reactions were run with 10 g/L acceptor 2 in 0.1 M Tris-HCl
buffer. The pH optimum was tested at pH’s 7, 7.5, 8, and 9,
with 0.5 M donor 3. Donor optimization was tested with
concentrations of donor 3 of 0.1, 0.5, 1, and 2 M, carried out at
pH 7.5. All experiments were carried out in triplicate.
Section 2.2.2. Reactions were done at 1 mL scale in a 96 well
plate, with 11 identical reactions per substrate concentration.
The reaction contained 0.1 M Tris-HCl buffer pH 9.0 with 0.4
g/L purified Ars-ωTA, 0.5 M donor 3, 0.1 g/L PLP, 5−10%
DMSO, and 25, 50, or 75 g/L pro-chiral acceptor 2. The
reaction was agitated 250 rpm at 25 °C. Samples were taken at
regular time points throughout the experiment.
Section 2.2.3. Reactions were performed at 25 mL scale in
an Easymax vessel, which was stirred at 400 rpm, maintained at
25 °C, with substrate concentrations of 25 and 50 g/L.
Otherwise the composition in the reactor was identical to that
described in section 2.2.2.
Section: 2.1.2. Experiments were carried out by resuspend-
ing 500 mg wet cells in 4.75 mL of 500 mM 3 hydrochloride,
50 mM potassium phosphate buffer pH 7.0, and shaking in an
orbital shaker at 250 rpm, maintained at 30 °C, for 30 min. 0.25
mL of a 200 g/L solution of pro-chiral ketone acceptor 2 in
DMSO was added and the reaction returned to the shaker for
18 h. Reactions were analyzed by GC. The activity for ATA-47
was extrapolated from rates measured in the experiments
carried out as described in sections 2.1.1 and 2.1.3.
4.2.3. Product Isolation. Section 2.2.3. The pH of the
reaction was adjusted to 13 with 5 M NaOH and extracted with
MTBE (3 × 20 mL). The combined organic extracts were
filtered through Celite to remove emulsion and dried (MgSO4),
filtered, and concentrated in vacuo. Being volatile, the excess
amine donor 3 was removed with the organic solvents during
concentration.
4.2.4. Work up of Samples for Analysis. Sections 2.1.1 and
2.1.3. The analytical samples were prepared as follows; 0.1 mL
of sample was added to 0.4 mL of 1 M NaOH with 10 mM
dibenzyl ether as external standard. The compounds were
extracted with 0.3 mL of MTBE, and the organic layer was
dried with anhydrous MgSO4, which was removed by
centrifugation.
4.1. Materials. Three plasmids encoding the enzymes, Tar0,
Tar1, and Ars-ωTA, were kindly provided by Professor NJ
Turner (University of Manchester, Manchester, UK). Tar0
encoded the ω-transaminase from Arthrobacter sp. KNK168
(Sequence 2 from US 7169592) inserted between the Nde I
and Xho I (with C-terminal His tag) site of pET21a (Accession
number ABN35871). Tar1 encoded ω-transaminase from
Arthrobacter sp. KNK168 (Sequence 110 from US 8293507,
Tar1) inserted between the Nde 1 and Xho 1 (with N-terminal
His tag) site of the pET16b (Accession number AFX11601).
Ars-ωTA encoded mutated ω-transaminase from Arthrobacter
citreus (Sequence 16 from US 7172885,) inserted between the
Nde 1 and Xho 1 (with C-terminal His tag) sites of the pET21a
(Accession number ABN37907). The commercial enzyme
ATA-47 (30902-2; activity 0.41 U/mg; batch LH1-01-02) was
purchased from c-LEcta GmbH (Leipzig, Germany).
Deionized water (18 Ω) was used for all experiments. All
chemicals where purchased from chemical vendors at reagent
grade or higher and used without modification. GC and NMR
solvents were of analytical grade, and products used for the
enzyme expression were of biological grade.
4.2. Methods. 4.2.1. Enzyme Expression. Section 2.2.1.
The plasmid that encodes Ars-ωTA with its C-terminal hexa-
histidine tag, was transformed into E. coli BL21 (DE3)
(Novagen from Merck KGaA, Darmstadt, Germany), using
standard procedures,26 and maintained with 100 μg/mL
ampicillin. Briefly, Ars-ωTA was expressed in autoinducing
medium as follows: a 1% glycerol stock inoculum was used to
inoculate 400 mL of ZYP-5052 medium.27 The culture was
incubated at 37 °C, shaking at 250 rpm for 24 h, in a Sartorius
Stedim CERTOMAT BS-1. The culture was centrifuged at
8000 rpm (Beckman Coulter Avanti J-26S XP centrifuge, JLA
8.1 rotor) at 4 °C for 30 min, the supernatant were decanted
and the cell pellet stored at −20 °C. The average yield was 9 g
pellet mass per liter of culture.
Sections 2.1.2, 2.2.2, and 2.2.3. A sample of 1 mL of E. coli
BL21 (DE3) expressing Ars-ωTA enzyme was inoculated in 50
mL vegetable peptone broth with 20 g/L glucose and 15 μg/
mL kanamycin and cultivated for 6−7 h at 37 °C, 250 rpm in a
rotary shaker incubator. This preculture was used to seed two 1
L fermentation vessels. Both fermenters were run with the same
fermentation procedure, which consisted of culturing the cells
in a sugar free semidefined base medium, controlled at pH 7.2,
30 °C, 20% dissolved oxygen, and feeding at a predefined linear
rate with base medium containing 400 g/L glucose from the
point of inoculation. The culture was induced when 100 OD600
was reached by the addition of 0.5 mM IPTG final
concentration and reduction in culture temperature to 25 °C.
Following 24 h elapsed fermentation time the feed rate was
reduced to a predefined constant rate until cell harvest at 41 h.
The final cell population reached 212 and 192 OD600, 80 and
81 g/L dry cell weight, respectively. A portion of 1 L of
fermentation broth from each reactor was harvested by
centrifugation at 6000 g for 35 min; the supernatant was
discarded, and the pellet was frozen at −80 °C.
Enzyme Purification. E. coli cells (25 g) expressing Ars-
ωTA were added to 250 mL 0.1 M phosphate buffer (pH 7)
and sonication at 2 °C. The lysate was concentrated by
ammonium sulfate to 50−60% ammonium sulfate fraction. The
precipitate was resuspended in 25 mL 0.1 M phosphate buffer
with 30 mM imidazole, and purified using a His-Trap (Ni-
F
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(10) Tufvesson, P.; Lima-Ramos, J.; Nordblad, M.; Woodley, J. M.
Section 2.2.1. The sample was mixed for indicated time at
700 rpm, 30 °C. Samples were sacrificed by addition of 0.5 mL
MeCN and spun down. 0.2 mL of the supernatant was
transferred to a new plate with 0.8 mL of MeCN and MgSO4.
Finally 0.5 mL was transferred to an analysis plate and
derivatized with 15 μL of Et3N and 10 μL of Ac2O, preceding
analysis.
Sections 2.1.2 and 2.2.2. A 1 mL reaction was mixed
thoroughly with 9.0 mL MeCN containing 4.5 mg/mL dibenzyl
ether. 1.0 mL of this mixture was put in a GC vial and
derivatized with 30 μL of Et3N/20 μL of Ac2O prior to GC
analysis (Chiraldex Dex-CB column 25 m × 0.25 mm × 0.25
μm, oven temp 170 °C for 15 min, Carrier He @ 20 psi,
injector/detector 200 °C). Quenched 50 g/L and 75 g/L
reactions were further diluted 1:1 with MeCN prior to
derivatization and analysis.
Org. Process Res. Dev. 2011, 15, 266−274.
(11) Hohne, M.; Kuhl, S.; Robins, K.; Bornscheuer, U. T.
̈
̈
ChemBioChem 2008, 9, 363−365.
(12) Hohne, M.; Bornscheuer, U. T. ChemCatChem 2009, 1, 42−51.
̈
(13) Tufvesson, P.; Bach, C.; Woodley, J. M. Biotechnol. Bioeng. 2014,
111, 309−319.
(14) Truppo, M. D.; Rozzell, J. D.; Turner, N. J. Org. Process Res. Dev.
2010, 14, 234−237.
(15) Shin, J.; Kim, B. J. Org. Chem. 2002, 67, 2848−2853.
(16) Hohne, M.; Robins, K.; Bornscheuer, U. T. Adv. Synth. Catal.
̈
2008, 350, 807−812.
(17) Al-Haque, N.; Santacoloma, P. A.; Neto, W.; Tufvesson, P.;
Gani, R.; Woodley, J. M. Biotechnol. Prog. 2012, 28, 1186−1196.
(18) Kroutil, W.; Fischereder, E.; Fuchs, C. S.; Lechner, H.; Mutti, F.
G.; Pressnitz, D.; Rajagopalan, A.; Sattler, J. H.; Simon, R. C.; Siirola,
E. Org. Process Res. Dev. 2013, 17, 751−759.
(19) Gundersen, M. T.; Abu, R.; Schurmann, M.; Woodley, J. M.
̈
Section 2.2.3. 100 μL of the reaction was removed and
diluted with MeCN (900 μL). This mixture was derivatized
with 30 μL of Et3N/20 μL of Ac2O and analyzed by GC.
4.2.5. Analytical. Sections 2.1.1 and 2.1.3. All analytical
work was carried out with gas chromatography, with a
PerkinElmer (Santa Clara, CA, USA) Clarus 500 apparatus,
with PerkinElmer Elite-5 column. 1 μL was injected with a 30:1
split ratio and ran with a constant flow rate of 1.6 mL min−1
helium with a temperature gradient from at 120 to 230 °C.
Sections 2.1.2, 2.2.1, 2.2.2, and 2.2.3. Chiraldex Dex-CB
column 25 m × 0.25 mm × 0.25 μm, oven temp. 170 °C for 15
min, carrier He @ 20 psi, injector/detector 200 °C.
Tetrahedron: Asymmetry 2015, 26, 567−570.
(20) Tufvesson, P.; Jensen, J. S.; Kroutil, W.; Woodley, J. M.
Biotechnol. Bioeng. 2012, 109, 2159−2162.
(21) Mutti, F. G.; Kroutil, W. Adv. Synth. Catal. 2012, 354, 3409−
3413.
(22) Koszelewski, D.; Goritzer, M.; Clay, D.; Seisser, B.; Kroutil, W.
ChemCatChem 2010, 2, 73−77.
(23) Koszelewski, D.; Lavandera, I.; Clay, D.; Rozzell, J. D.; Kroutil,
W. Adv. Synth. Catal. 2008, 350, 2761−2766.
(24) Hall, H. J. J. Am. Chem. Soc. 1957, 256, 1955−1958.
(25) Shin, J. S.; Kim, B. G. Biotechnol. Bioeng. 1997, 55, 348−358.
(26) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular cloning: a
laboratory manual; Cold Spring Harbor Laboratory: Cold Spring
Harbor, NY, 1989.
4.3. 1H NMR of Isolated Product. N-Boc-3-amino-
1
piperidine H NMR (400 MHz, CD3OD) 3.96 (1H, m), 3.82
(27) Studier, F. W. Protein Expression Purif. 2005, 41, 207−234.
(1H, m), 2.78 (1H, br), 2.67 (1H, m), 2.60 (1H, br s), 1.91
(1H, m), 1.67 (1H, m), 1.44 (9H, s), 1.40 (1H, m), 1.26 (1H,
m).
AUTHOR INFORMATION
■
Corresponding Author
*Fax: (+45) 4525 2885. E-mail: jw@kt.dtu.dk.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The support of AMBIOCAS financed through the European
Union Seventh Framework Programme (Grant agreement no.
245144) is acknowledged.
REFERENCES
■
(1) Pollard, D. J.; Woodley, J. M. Trends Biotechnol. 2007, 25, 66−73.
(2) Straathof, A. J. J.; Panke, S.; Schmid, A. Curr. Opin. Biotechnol.
2002, 13, 548−556.
(3) Clouthier, C. M.; Pelletier, J. N. Chem. Soc. Rev. 2012, 41, 1585−
1605.
(4) Koszelewski, D.; Tauber, K.; Faber, K.; Kroutil, W. Trends
Biotechnol. 2010, 28, 324−332.
(5) Malik, M. S.; Park, E.-S.; Shin, J.-S. Appl. Microbiol. Biotechnol.
2012, 94, 1163−1171.
(6) Tufvesson, P.; Lima-Ramos, J.; Jensen, J. S.; Al-Haque, N.; Neto,
W.; Woodley, J. M. Biotechnol. Bioeng. 2011, 108, 1479−1493.
(7) Steffen-Munsberg, F.; Vickers, C.; Thontowi, A.; Schatzle, S.;
̈
Meinhardt, T.; Svedendahl Humble, M.; Land, H.; Berglund, P.;
Bornscheuer, U. T.; Hohne, M. ChemCatChem 2013, 5, 154−157.
̈
(8) Henson, C.; Cleland, W. Biochemistry 1964, 3, 338−345.
(9) Meier, R. J.; Gundersen, M. T.; Woodley, J. M.; Schurmann, M.
̈
ChemCatChem 2015, 7, 2594−2597.
G
DOI: 10.1021/acs.oprd.5b00159
Org. Process Res. Dev. XXXX, XXX, XXX−XXX