Conductometric method for the rapid characterization of the substrate specificity of amine-transaminases

Article abstract of DOI:10.1021/ac9028483

Amine-transaminases (ATAs, ω-transaminases, ω-TA) are PLP-dependent enzymes that catalyze amino group transfer reactions. In contrast to the widespread and wellknown amino acid-transaminases, ATAs are able to convert substrates lacking an a-carboxylic functional group. They have gained increased attention because of their potential for the asymmetric synthesis of optically active amines, which are frequently used as building blocks for the preparation of numerous pharmaceuticals. Having already introduced a fast kinetic assay based on the conversion of the model substrate α-methylbenzylamine for the characterization of the amino acceptor specificity, we now report on a kinetic conductivity assay for investigating the amino donor specificity of a given ATA. The course of an ATA-catalyzed reaction can be followed conductometrically since the conducting substrates, a positively charged amine and a negatively charged keto acid, are converted to nonconducting products, a noncharged ketone and a zwitterionic amino acid. The decrease of conductivity for the investigated reaction systems were determined to be 33-52 μS mM^-1. In contrast to other ATA-assays previously described, with this approach all transamination reactions between any amine and any keto add can be monitored without the need for an additional enzyme or staining solutions. The assay was used for the characterization of a ATA from Rhodobacter sphaeroides, and the data obtained were in excellent agreement with gas chromatography analysis.

Full text of DOI:10.1021/ac9028483

Anal. Chem. 2010, 82, 2082–2086
Conductometric Method for the Rapid
Characterization of the Substrate Specificity of
Amine-Transaminases
Sebastian Scha¨ tzle,† Matthias Ho¨ hne,† Karen Robins,‡ and Uwe T. Bornscheuer*,†
Institute of Biochemistry, Department of Biotechnology and Enzyme Catalysis, Greifswald University,
Felix Hausdorff-Strasse 4, 17487 Greifswald, Germany, and Lonza AG, Valais Works, Visp, Switzerland
Amine-transaminases (ATAs, ω-transaminases, ω-TA) are
PLP-dependent enzymes that catalyze amino group trans-
fer reactions. In contrast to the widespread and well-
known amino acid-transaminases, ATAs are able to con-
vert substrates lacking an r-carboxylic functional group.
They have gained increased attention because of their
potential for the asymmetric synthesis of optically active
amines, which are frequently used as building blocks for
the preparation of numerous pharmaceuticals. Having
already introduced a fast kinetic assay based on the
conversion of the model substrate r-methylbenzylamine
for the characterization of the amino acceptor specificity,
we now report on a kinetic conductivity assay for inves-
tigating the amino donor specificity of a given ATA. The
course of an ATA-catalyzed reaction can be followed
conductometrically since the conducting substrates, a
positively charged amine and a negatively charged keto
acid, are converted to nonconducting products, a non-
charged ketone and a zwitterionic amino acid. The de-
crease of conductivity for the investigated reaction systems
were determined to be 33-52 µS mM-1. In contrast to
other ATA-assays previously described, with this ap-
proach all transamination reactions between any amine
and any keto acid can be monitored without the need
for an additional enzyme or staining solutions. The
assay was used for the characterization of a ATA from
Rhodobacter sphaeroides, and the data obtained were
in excellent agreement with gas chromatography
analysis.
Figure 1. Transaminase reaction and principle of the conductivity
assay. A positively charged amine and a negatively charged keto acid
are converted to a zwitterionic amino acid and a noncharged ketone,
thus the conductivity of the reaction solution decreases.
are capable of being performed in a high-throughput screening
(HTS) format. Examples includes solid-phase bound assays related
to immunoassay,1 IR thermography,2 circular dichroism (CD)
spectroscopy,3,4 fluorescence image analysis,5 and UV-vis-based
methods.6,7
Amine-transaminases (ATA, ω-transaminases, ω-TA) gained
increased attention in research in the last years because of their
potential for the synthesis of optically active amines, which are
frequently used as building blocks for the preparation of numerous
physiologically active compounds. However, for the screening of
the substrate specificities and enantioselectivities of ATA only a
limited number of methods have been reported, due to the special
type of reaction catalyzed (Figure 1): An amine and a carbonyl
substrate bearing a ketone or aldehyde group are converted into
products by exchanging their amino and carbonyl functional
groups, without the stochiometric consumption of a cofactor. In
case of other enzymatic reactions like e.g. hydrolysis of esters or
amides, the substrates (ester/amide) and products (alcohol/amine
The identification of novel biocatalysts and the optimization
of existing enzymes are key steps for developing highly efficient
processes. Thus, protein engineering and the metagenome ap-
proach have emerged as powerful tools that are used frequently
to achieve these aims. While working with large enzyme libraries
which are generated during these strategies, the fast, effective,
and reliable identification of enzymes with the desired properties
is often the bottleneck since this is the most time-consuming step.
Therefore a large variety of fast methods for determining enzyme
activity have been developed for many biocatalysts, which often
(1) Taran, F.; Gauchet, C.; Mohar, B.; Meunier, S.; Valleix, A.; Renard, P. Y.;
et al. Angew. Chem., Int. Ed. 2002, 41, 124–127
(2) Reetz, M. T.; Becker, M. H.; Ku¨hling, K. M.; Holzwarth, A. Angew. Chem.,
Int. Ed. 1998, 37, 2647–2650
(3) Ding, K.; Ishii, A.; Mikami, K. Angew. Chem., Int. Ed. 1999, 38, 497–501
(4) Reetz, M. T.; Ku¨hling, K. M.; Hinrichs, H.; Deege, A. Chirality 2000, 12,
479–482
(5) Soleihac, J.-M.; Cornille, F.; Martin, L.; Lenoir, C. Anal. Biochem. 1996,
241, 120–127
(6) Mor´ıs-Varas, F.; Shah, A.; Aikens, J.; Nadkarni, N. P.; Rozzell, J. D.;
Demirjian, D. C. Bioorg. Med. Chem. 1999, 7, 2183–2188
(7) Baumann, M.; Hauer, B. H.; Bornscheuer, U. T. Tetrahedron: Asymmetry
2000, 11, 4781–4790
.
.
.
.
.
* To whom correspondence should be addressed: Fax: (+49) 3834-86-80066.
E-mail: uwe.bornscheuer@uni-greifswald.de.
† Greifswald University.
.
‡ Lonza AG.
.
2082 Analytical Chemistry, Vol. 82, No. 5, March 1, 2010
10.1021/ac9028483  2010 American Chemical Society
Published on Web 02/11/2010

Table 1. Calibration of the Conductivity Assaya
substrate/product
1 & 6
2 & 6
3 & 6
4 & 6
5 & 6
3 & 7
3 & 8
3 & 9
3 & 10
conductivity [µs mM-1 cm-1
]
36.6
39.0
44.0
34.6
32.9
37.1
40.0
49.5
52.1
a The values reflect the change in conductivity due to the conversion of 1 mM substrates to products (see Figure 4 for structures of substrates
and products 1-10). The resolution of the applied conductometer was 0.1 µS.
and acid) can be discriminated very easily,8 or during reactions
involving a cofactor like NAD(P)H or side product like H2O2,
these compounds can be detected specifically.9,10 For ATA, this
is not the case, and there are only a few special methods to
discriminate the similar substrates and products without affecting
enzyme activity.11
Buffer Preparation. All buffers had a concentration of 20 mM
and were adjusted to pH 7.5. Tris was adjusted with HCl, the three
buffers containing BES, CHES, and HEPES were adjusted with
Bis-Tris (bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane),
and EPPS and tricine were adjusted with 1,8-diazabicyclo[5.4.0]-
undec-7-en.
In addition to ketones and aldehydes, ATA also convert the
R-keto carboxylates pyruvate and glyoxylate. The methods for
assaying aminotransferase activity published so far comprise (i)
measurement of the generated amino acid (alanine, glycine) via
detection of the corresponding copper complex11 and (ii) a pH
indicator based multienzyme cascade assay.12 Obviously, there
are several drawbacks: the copper staining solution in method
(i) inhibits the enzyme, and nonspecific color formation with
several buffers and crude cell extracts are observed. Thus, only
end point measurements are possible, and appropriate measure-
ments of the blank are necessary. Method (ii) works in the
asymmetric synthesis mode (Figure 1, reaction from right to left
for bottom partners), and thus no information about enantiopref-
erence or -selectivity can be obtained.
Alternatively, we recently published an UV-spectrophotometric
assay based on the conversion of the widely used model substrate
R-methylbenzylamine.13 The product from this reaction, acetophe-
none, can be detected spectrophotometrically at 245 nm with high
sensitivity (ε ) 12 mM-1 cm-1), since the other reactants
showed only a low absorbance. Besides the standard substrate
pyruvate, all low-absorbing ketones, aldehydes, or R-keto
carboxylates could be used as cosubstrates. Thus, the assay
is a fast and easy method for determining transaminase activity.
Additionally, the amino acceptor specificity of a given ATA can
be characterized very quickly.
Calibration. For calibrating the assay, different standard
solutions with decreasing substrate and increasing product
concentrations of 0-10 mM were prepared, and the conductivity
and pH were measured (Table 1). No change of the pH could be
detected. It is noteworthy that the accuracy of the assay was
significantly higher starting from 10 mM substrates than starting
from 5 mM substrates. Furthermore, it was important to use all
reactants simultaneously, including the nonconducting ketone and
alanine, for the calibration (see results). For analyzing the
influence of crude extract on the conductivity measurements,
standard curves with different amounts (0-18% (v/v)) of crude
extract (OD600 ≈ 10) of E. coli BL21 without ATA were
measured in parallel. For the preparation of the crude extract,
cells were washed twice and disrupted by sonication in the
buffer used for conductivity measurements.
Kinetic Measurements. The reactions were carried out at
room temperature (RT). Kinetic measurements were performed
in reaction mixtures containing 10 mM substrates in buffer (20
mM, pH 7.5), 10% dimethyl sulfoxide, and an appropriate amount
of enzyme (purified enzyme or crude extract). Using the platinum
(L 5 mm) electrode, the reaction can be carried out in a well of
a microtiter plate, and a reaction volume of 200 µL is sufficient.
The course of the reaction was followed by measuring the
conductivity of the solution continuously. Each measurement was
repeated at least three times. For blank measurement, the reaction
was carried out with cell extract lacking an amine-transaminase
or with reaction mixtures containing only one of the substrates,
but no significant change in conductivity could be detected. The
specific activity was expressed as units per milligram protein. One
unit of activity was defined as the amount of enzyme that produced
1 µmol ketone product per minute.
As the acetophenone assay is limited to R-methylbenzylamine
as amino donor, we now developed a method for a fast and easy
characterization of the amine donor specificity of a given ATA.
With this conductivity based approach every amine donor and
acceptor can be applied as substrate, as long as one of the
substrates is an amino or keto acid, respectively.
EXPERIMENTAL SECTION
Validation of the Assay with Gas Chromatography. During
kinetic measurement, aliquots were taken at certain points for
validation of the conductivity assay by gas chromatography. The
reaction was stopped by adding 1/10 of the volume of 1 M HCl.
The ketones formed during the reaction were extracted with one
volume of ethyl acetate and determined by gas chromatography
(Hewlett-Packard 5890 Series II Plus) using a forte BP-21 column
(SGE, Griesheim, Germany) with benzaldehyde as internal
standard.
All conductivity mesasurements were performed using either
a Qcond 2200 (VWR, Germany) or a CX-401 (Elmetron, Poland)
multimeter with a graphite (L 12 mm) or a platinum (L 5 mm)
electrode, respectively.
(8) Krebsfa¨nger, N.; Schierholz, K.; Bornscheuer, U. T. J. Biotechnol. 1998,
60, 105–111
(9) Hinman, L. M.; Blass, J. P. J. Biol. Chem. 1981, 256, 6583–6586
(10) Holt, A.; Palcic, M. M. Nature Prot. 2006, 1, 2499–2505
(11) Hwang, B.-Y.; Kim, B.-G. Enzyme Microb. Technol. 2004, 34, 429–436
(12) Truppo, M. D.; Rozzell, J. D.; Moore, J. C.; Turner, N. J. Org. Biomol. Chem.
2009, 7, 395–398
(13) Scha¨tzle, S.; Ho¨hne, M.; Redestad, E.; Robins, K.; Bornscheuer, U. T. Anal.
Chem. 2009, 81, 8244–8248
.
.
.
.
Characterization of the ATA from Rhodobacter sphaeroi-
des. The ATA from Rhodobacter sphaeroides 2.4.1 (DSM 158) was
recombinantly expressed in E. coli BL21 (DE3) and purified via
.
.
Analytical Chemistry, Vol. 82, No. 5, March 1, 2010 2083

given amine substrate can be used separately in the assay, so that
also information about enantioselectivity can be obtained, and (iii)
the amine substrate usually is much better soluble than the
corresponding ketone. Thus, even with higher substrate concen-
trations a homogeneous assay solution is obtained rather than a
biphasic system, which otherwise might interfere with the
measurements.
This allows a simple measurement of the reaction progress, if
some crucial requirements are fulfilled: First, to maximize
sensitivity the buffer system should have low background con-
ductivity. The pH of the buffer has to be kept in a pH range from
4.0-8.0, where the net charge of alanine is ≈0. Furthermore, for
practical reasons or for high-throughput screening purposes, the
assay should not be sensitive toward variations in crude extract
concentrations.
Table 2. Summary of the Buffer Systems Investigatedc
conductivity
rel
buffera
pKa
[µS cm-1
]
activity [%]b
PB
7.2
7.3
7.7
8.0
8.1
8.3
9.4
-
2760
445
141
246
1600
240
46
100
120
24
76
71
98
67
0
BES
HEPES
EPPS
Tris
tricine
CHES
A. dest
3
a PB: phosphate buffer; BES: N,N-bis(2-hydroxyethyl)taurine; HEPES:
4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid; EPPS: 4-(2-hydroxy-
ethyl)-1-piperazinepropanesulfonic acid; Tris: 2-amino-2-(hydroxym-
ethyl)-1,3-propanediol; tricine: N-[tris(hydroxymethyl)methyl]glycine;
CHES: 2-(cyclohexylamino)ethanesulfonic acid. b Activities in the
different buffer systems were determined using the acetophenone
assay.13 c All buffers had a concentration of 20 mM and were adjusted
to pH 7.5 (see the Experimental Section).
Standard buffer systems like phosphate buffer or Tris-HCl
could not be applied due to their very high conductivity, whereby
the relative change in conductivity caused by the enzymatic
reaction is rather small. Thus, Good’s buffers were investigated
regarding the conductivity of 20 mM solutions at pH 7.5 and effects
on aminotransferase activity (Table 2). Due to their zwitterionic
nature, solutions of these compounds show a much smaller
conductivity if the pH is kept within 2 units near the pI of the
respective compound (isoelectric buffers).
The first system investigated was BES-Bis-Tris. It showed a
very good performance with purified enzyme, but unfortunately
activities of crude extract measured in this buffer did not match
the results obtained by gas chromatography. Although the CHES-
Bis-Tris system showed excellent performance with both purified
enzyme and crude extract, due to its inhibitory effect (67% enzyme
activity compared to phosphate buffer) and its rather low buffer
capacity at pH 7.5, we explored an alternative buffer system. The
inhibitory effect was the same for the EPPS buffer (76% rel activity)
and even greater for the HEPES-Bis-tris system (24% rel activity).
Finally, the tricine buffer showed both a good buffer capacity at
pH 7.5 and a very good agreement of determined activities of
purified enzyme and crude extract by measuring the conductivity
compared to GC analysis (Figure 2). For this reason the tricine
buffer was used for all subsequent experiments. The inhibitory
effect of the EPPS and CHES buffer may vary for other transami-
affinity chromatography (IMAC sepharose) as reported previ-
ously.13
Protein concentrations were determined using the BCA assay
kit (Uptima, Montlucon, France).
To determine the substrate specificity of the enzyme, 10 mM
of different substrates and an appropriate amount of purified
enzyme (0.1-2 mg mL-1) were mixed with 1 mL of 20 mM
tricine buffer pH 7.5. Dimethyl sulfoxide (10% (v/v)) was added
to the mixture to dissolve hydrophobic substrates. Solutions
of oxaloacetate and R-ketoglutarate were prepared freshly prior
to measurements. The reaction was carried out at RT and
followed by measuring the conductivity continuously. For
validation by GC, aliquots were taken at certain points, and
the concentration of the ketone products 1b-5b was analyzed
as described above.
RESULTS AND DISCUSSION
In the course of an ATA-catalyzed reaction the conductivity of
the reaction medium changes since charged reactants (amine and
keto acid) are converted to noncharged species (the ketone and
a zwitterionic amino acid) (Figure 1). For all experiments, the
transamination was carried out in kinetic resolution mode. The
rationale behind this approach is as follows: (i) the reaction
equilibrium favors product formation, (ii) both enantiomers of a
Figure 2. Validation of the assay by following the reaction of 10 mM (S)-R-methylbenzylamine and pyruvate using both conductivity measurements
(black symbols/bar) and detection of the acetophenone formed by gas chromatography (white symbols/bar). The calculated activities differed
only by 5% for purified enzyme (left, 0 0.25 mg/mL, 3 0.5 mg/mL, O 1 mg/mL) and crude extract (10% (v/v)) (right, the curve displays the
conductivity measurement, the bars show the calculated acetophenone concentrations).
2084 Analytical Chemistry, Vol. 82, No. 5, March 1, 2010

and thus the net charge of an analyte may be affected by other
compounds present in the solution. On the other hand, more
importantly, ions or even neutral molecules from the background
electrolyte may affect the mobility and thus the molar conductivity
of an analyte significantly by intermolecular electrostatic or
hydrophobic interactions. This explains why the conductivity of
a mixture comprised of different ions usually differs to some extent
from the sum of the contributions of the single ions to overall
conductivity. For this reason, all participating reactants and even
the neutral species had to be included in the calibration of the
assay. If otherwise the change of the conductivity per milimolar
reaction conversion [µS/mM] is calculated from molar conductivi-
ties of the single compounds, the obtained value would be higher
(49.4 µS/mM conversion, O in Figure 3) compared with the
measurement of the complete mixtures (44.1 µS/mM conversion,
b in Figure 3).
Figure 3. Influence of different concentrations of crude extract (b
0%, 1 5%, 9 10%, [ 14%, 2 18% (v/v)) on the change of conductivity
during the transaminase reaction. While measuring the standard
curves all reactants have to be present, as if measured separately,
the calculated rate ∆µS mM-1 (O) is significantly higher. Contrary to
other buffer systems investigated (and at substrate concentrations
of 5 mM), no significant influence of different crude extract concentra-
tions were observed.
Application of the Assay in the Characterization of an ATA
from Rhodobacter sphaeroides. To verify the method described
above, an ATA identified in the genome of Rhodobacter sphaeroides
2.4.1 was characterized using this assay. The enzyme was
expressed recombinantly in E. coli BL21 and purified as de-
scribed.13 Investigation of the amino donor profile revealed that
besides the model substrate R-methylbenzylamine also benzyl-
amine is a very well accepted substrate (Figure 4). Aliphatic and
arylaliphatic compounds were converted only slowly, especially
the cyclic aliphatic amine 1-N-Boc-3-aminopiperidine. The results
obtained for the amino acceptor profile are in good agreement
with previously measured data using the acetophenone assay.13
Pyruvate is the best converted acceptor, glyoxylate and succinic
semialdehyde are converted moderately, and the keto-dicarboxylic
acids oxalacetate and R-ketoglutarate are converted only very
slowly.
The conductivity assay is an excellent complementary method
to the acetophenone assay, since the usually more interesting
amino donor specificity of an ATA can be investigated. With both
assays in hand, the complete characterization of the substrate
specificity can be performed very rapidly. Furthermore, the
conductivity assay allows a fast screening of ATA variants for the
reaction of any desired amine with a keto acid. One limitation of
the conductivity assay is that an appropriate low conductivity buffer
nases, so these buffers should be considered for further
experiments.
In practice and especially for screening of many samples, crude
cell extract is used as enzyme preparation instead of purified
proteins. Thus an important question was whether the presence
of different amounts of crude extract has an impact on the
conductivity measurement. As expected, the background conduc-
tivity of the reaction medium increased proportional to the
concentration of cell extract used. Next, calibration experiments
were performed where different conversions were simulated by
varying the concentrations of all reactants according to a real
reaction. Fortunately, the addition of different amounts of crude
extract did not seem to affect the change of conductivity
significantly (Figure 3). This was surprising, because in theory
the molar conductivity of a given compound depends on the
concentration of all electrolytes present in the reaction system
for mainly two reasons: on the one hand, the protonation state
Figure 4. Substrate specificity of Rhodobacter sphaeroides ATA was determined using the conductivity assay (black) as well as gas
chromatography (gray). All measurements were repeated three times.
Analytical Chemistry, Vol. 82, No. 5, March 1, 2010 2085

in the pH-range of 4-8 has to be used. For determining other
enzymatic properties like the pH- or temperature optimum of an
ATA, the acetophenone assay is more flexible since every low
absorbing buffer at any pH may be used and absorbance is
virtually not dependent on temperature in contrast to conductivity,
so that multiple calibrations can be avoided.
of the amino donor substrate profile of ATA. This was achieved
without the need for any additional enzymes or staining solutions.
With appropriate equipment, we envision an application of this
assay as a high-throughput method for screening enzyme libraries.
Received for review December 15, 2009. Accepted
January 30, 2010.
CONCLUSIONS
In this contribution we described a simple and sensitive
conductivity assay that can be used for the fast characterization
AC9028483
2086 Analytical Chemistry, Vol. 82, No. 5, March 1, 2010