Whole-Cell Biocatalysts for Stereoselective C-H Amination Reactions

Article abstract of DOI:10.1002/anie.201510028

Enantiomerically pure chiral amines are ubiquitous chemical building blocks in bioactive pharmaceutical products and their synthesis from simple starting materials is of great interest. One of the most attractive strategies is the stereoselective installation of a chiral amine through C-H amination, which is a challenging chemical transformation. Herein we report the application of a multienzyme cascade, generated in a single bacterial whole-cell system, which is able to catalyze stereoselective benzylic aminations with ee values of 97.5 %. The cascade uses four heterologously expressed recombinant enzymes with cofactors provided by the host cell and isopropyl amine added as the amine donor. The cascade presents the first example of the successful de novo design of a single whole-cell biocatalyst for formal stereoselective C-H amination.

Full text of DOI:10.1002/anie.201510028

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201510028
German Edition: DOI: 10.1002/ange.201510028
Biocatalysis
Hot Paper
À
Whole-Cell Biocatalysts for Stereoselective C H Amination Reactions
Peter Both, Hanna Busch, Paul P. Kelly, Francesco G. Mutti, Nicholas J. Turner, and
Sabine L. Flitsch*
Abstract: Enantiomerically pure chiral amines are ubiquitous
chemical building blocks in bioactive pharmaceutical products
and their synthesis from simple starting materials is of great
interest. One of the most attractive strategies is the stereoselec-
ases (ADH) with complementary stereoselectivity. Two
À
ADHs were deemed necessary, since C H activation is not
always stereoselective and thus we would be able to maximize
product yields. Finally, a transaminase (ATA) would be used
for generating the amine with the desired stereoselectivity. To
our knowledge, such a system represents the first example of
a four-enzyme cascade and poses the question whether these
enzymes are compatible in a single whole-cell system, given
the complex demand for cofactors, transport of substrate and
product through the bacterial cell wall, and metabolic stability
and potential toxicity of intermediates and products.
À
tive installation of a chiral amine through C H amination,
which is a challenging chemical transformation. Herein we
report the application of a multienzyme cascade, generated in
a single bacterial whole-cell system, which is able to catalyze
stereoselective benzylic aminations with ee values of 97.5%.
The cascade uses four heterologously expressed recombinant
enzymes with cofactors provided by the host cell and isopropyl
amine added as the amine donor. The cascade presents the first
example of the successful de novo design of a single whole-cell
The target enzymes for the cascade were carefully chosen
based upon their known complementarity with respect to
substrate recognition. The first enzyme is an engineered
À
biocatalyst for formal stereoselective C H amination.
À
S
tereoselective C H amination is a very
attractive strategy for the conversion of
simple low-cost chemical starting materials
to high-value chiral amine building blocks,
and has therefore attracted much interest in
organic chemistry. The most successful chem-
ical strategies reported so far have involved
intramolecular and transition-metal-cata-
lyzed reactions[1] with a number of elegant
methods using both chemical[2] and enzy-
matic[3,4] catalysts. We were interested in
constructing de novo biosynthetic multien-
zyme cascades that are guided in design and
built by retrosynthetic considerations,[5]
which led us to the design of a single whole-
cell system that can catalyze the stereoselec-
À
tive benzylic C H amination of simple non-
functionalized organic compounds using
molecular oxygen and isopropyl amine
(Figure 1).
Figure 1. E. coli BL21(DE3) cells harboring a monooxygenase [Y96F], R- and S-selective
alcohol dehydrogenases [LbRADH and ReSADH], and an w-transaminase [ATA117] capable
of converting 4-substituted ethylbenzenes 1a–e into alcohols 2a–e, ketones 3a–e, and
finally amines 4a–e.
Based upon previous work by us and
others[4,6–13] we designed the target cascade to
include four enzymes as shown in Figure 1. We envisaged
chimeric self-sufficient P450 monooxygenase (Y96F)[9] con-
sisting of the P450camY96F catalytic domain fused to
a reductase domain from Rhodococcus sp. (RhFRed). The
enzyme produces scalemic mixtures of alcohols with a general
preference for the R-enantiomer.[10] Two alcohol dehydrogen-
ases with complementary stereoselectivity were selected for
the next step. LbRADH from Lactobacillus brevis is known
to oxidize R-alcohols using NADP+,[11] whereas ReSADH
from Rhodococcus erythropolis oxidizes S-alcohols using
NAD+ as the cofactor.[8,12] The final step for the stereoselec-
tive formation of the chiral amine is catalyzed by w-trans-
aminase ATA117 from Arthrobacter sp. requiring an amine
donor and pyridoxal 5’-phosphate (PLP).[13]
À
initial C H activation by a self-sufficient cytochrome P450
monooxygenase to generate the benzylic alcohol, which could
then be oxidized to the ketone using two alcohol dehydrogen-
[*] Dr. P. Both, H. Busch, Dr. P. P. Kelly, Dr. F. G. Mutti, Prof. N. J. Turner,
Prof. S. L. Flitsch
School of Chemistry, Manchester Institute of Biotechnology
The University of Manchester
131 Princess Street, Manchester M1 7DN (UK)
E-mail: Sabine.Flitsch@manchester.ac.uk
Supporting information and ORCID(s) from the author(s) for this
article are available on the WWW under http://dx.doi.org/10.1002/
anie.201510028.
Angew. Chem. Int. Ed. 2016, 55, 1511 –1513
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1511

Angewandte
Communications
Chemie
Figure 2. E. coli BL21(DE3) cell cotransformed with two plasmids: one
harboring a monooxygenase [Y96F], the other harboring R- and S-
selective alcohol dehydrogenases [LbRADH and ReSADH], and an w-
transaminase [ATA117].
Figure 3. Gas chromatogram of the optimized whole-cell biotransfor-
mation using 300 mgmLÀ1 cells and 1 mm of substrate 1a in 50 mm
sodium phosphate buffer pH 7.4 after 24 h and 48 h with the addition
of isopropylamine to a final concentration of 200 mm at 24 h (incu-
bated at 208C and 220 rpm).
The pathway was introduced into Escherichia coli
(Figure 1) using a two-plasmid system as shown in Figure 2.
In order to avoid possible problems with inefficient tran-
scription and/or translation we opted for a monocistronic
design. This ensures independent regulation of each gene via
its own operon (Figure S1). For the majority of the cascade we
constructed a new plasmid, pPB01, which was designed to be
void of regulatory elements adjacent to the EcoRI and PstI
restriction sites. The plasmid was derived from the pTrcHis
plasmid (Lifetechnologies) and is composed of the pBR322
ori, the gene for ampicillin resistance, and the gene for the lac
repressor, and thus it can be used with any operons containing
the lac operator for expression regulation. pPB01 allows the
insertion of operons in the BioBrick format for fast cloning
strategies.[14] Using this method the final construct pPB01/
ReSADH + LbRADH + ATA117 bearing three new genes
was generated (Figure S2) and inserted into E. coli together
with plasmid pCDF/Y96F bearing the P450 gene. This two-
plasmid strategy was designed to ensure better control over
P450 gene expression, which can be toxic to bacterial cells.
Analysis of the E. coli BL21(DE3) cells harboring both
plasmids after induction with isopropyl b-d-1-thiogalactopyr-
anoside (IPTG) by protein gel electrophoresis and western
blotting of the soluble fraction of cell lysates showed bands at
molecular weights expected for enzymes of the cascade
(Figure S3). The whole-cell system was then tested for
conversion of substrate 1a to 4a thus measuring activity of
the cascade encoded by plasmids pCDF/Y96F and pPB01/
ReSADH + LbSADH + ATA117, respectively. It was gratify-
ing to see that product 4a could clearly be detected in all
samples using GC analysis and comparison to authentic,
commercially available standard amine 4a (Figure 3 and
Figure S9).
aminolevulinic acid (ALA) upon induction (for more details
see Section 1D in the Supporting Information).
The use of isopropylamine (IPA) rather than alanine as
the amine donor resulted in higher conversions with this
whole-cell system (Table 1) and hence IPA was used in
subsequent optimization experiments. However, IPA was
found to inhibit the first P450-catalyzed step (loss of activity),
which could be overcome by adding 1a (substrates were
added from 500 mm stock solutions in DMSO) and amine in
Table 1: Conversion of 1a–e and 2a into corresponding amines 4a–
e using the cascade of enzymes Y96F, LbADH, ReADH, and ATA117. I
Substrate
Amine donor
Conversion to
amines 4a–e [%][a]
2a
1a
1a
1b
1c
1d
1e
IPA
l-Ala
IPA
IPA
IPA
IPA
IPA
45
0
26
17
14
15
5
[a] Conversion determined by GC after 48 h.
consecutive steps to the reaction mixture: in the first step,
substrate 1a was added and the reaction mixture incubated
for 24 h at 208C in sealed reaction vials with sufficient
headspace to supply oxygen but minimize loss of volatile
substrate. In the next step 200 mm IPA was added to the
biotransformation, which was incubated for another 24 h at
208C, as reported by Koszelewski and co-workers.[13e] We also
examined whether additional PLP cofactor was required;
however, no improvement in yield was observed consistent
with the observation that E. coli produces high levels of
endogenous PLP.[15]
For optimized biotransformations, the effects of different
components such as cell mass (Figure S5), amine donor
concentration (Figure S6), and substrate concentration (Fig-
ure S7) were investigated with the best conversion obtained
shown in Figure 3 and Table 1. When racemic 1-(4-fluoro-
The effect of a number of parameters on the conversion of
1a to 4a was investigated in order to optimize the reaction
conditions. A number of variables for cell growth were
investigated and the following found to produce the highest
conversions: 16 h expression of protein in BL21(DE3) cells at
208C after induction with 0.4 mm IPTG at OD600 = 0.6 in M9
minimal media supplemented with 2 mm MgSO4, 0.2 mm
CaCl2, and 3 mm FeCl3, and with the addition of 0.5 mm 5-
1512
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1511 –1513

Angewandte
Communications
Chemie
phenyl)ethanol (2a) was added as a substrate instead of 1a,
composed of financial contribution from the European
conversion to the corresponding amine was 45% while there
was 18% of unconverted keto intermediate 3a with no traces
of alcohol 2a (Figure S9). Conversions for the whole cascade
starting from 1a were lower (26%), suggesting that the
reaction performed by Y96F might be the limiting step of the
cascade. A number of attempts were made to improve the
activity, such as supplementing higher concentrations of ALA
or co-expression of glucose dehydrogenase II from Bacillus
megaterium in order to facilitate cofactor recycling; however,
none of these resulted in higher yields.
The transaminase ATA117 is well known to be highly R-
selective and formation of (R)-4a with an ee value of 97.5%
was established when the product of the cascade was analyzed
by GC chromatography using a chiral stationary phase
(Figure 4). The substrate scope of the whole-cell enzyme
cascade was further explored using a small panel of substrates
(1b–1e) resulting in overall conversions of 5–26% from ethyl
benzene derivatives 1b–1e to amines 4b–4e, respectively
(Table 1).
In summary, we have developed a new biocatalytic
cascade, based upon the co-expression of four genes within
a single bacterial cell, for the conversion of ethylbenzenes 1a–
e to enantiopure (R)-1-phenylethanamines 4a–e, respectively,
with conversions of up to 26%. Under the present reaction
conditions, no additional cofactor except for the amine donor
IPA and molecular oxygen was required. None of the
substrates, cofactors, intermediates, or products appeared to
have significant adverse effects on the transformation. The
flexible generic nature of the design of the cascade will allow
for the easy substitution of individual enzymes with homo-
logues or mutants and thereby provide a cassette-based
modular approach for the design and construction of alter-
Unionꢀs Seventh Framework Programme (FP7/2007–2013)
and EFPIA companiesꢀ in kind contribution. We thank Dr.
Andrew Collis (GSK) for useful discussions on plasmid and
operon design. N.J.T. and S.L.F. are grateful to the Royal
Society for Wolfson Research Merit Awards.
Keywords: aminations · biocatalysis · chiral amines ·
enzyme cascades · whole-cell biotransformation
How to cite: Angew. Chem. Int. Ed. 2016, 55, 1511–1513
Angew. Chem. 2016, 128, 1533–1536
[1] M.-L. Louillat, F. W. Patureau, Chem. Soc. Rev. 2014, 43, 901.
[2] R. A. Green, J. F. Hartwig, Angew. Chem. Int. Ed. 2015, 54, 3768;
Angew. Chem. 2015, 127, 3839.
[3] a) C. C. Farwell, R. K. Zhang, J. A. McIntosh, T. K. Hyster, F. H.
Arnold, ACS Cent. Sci. 2015, 1, 89; b) R. Singh, J. N. Kolev, P. A.
Sutera, R. Fasan, ACS Catal. 2015, 5, 1685.
[4] M. Schrewe, N. Ladkau, B. Bühler, A. Schmid, Adv. Synth. Catal.
2013, 355, 1693.
[5] N. J. Turner, E. OꢀReilly, Nat. Chem. Biol. 2013, 9, 285.
[6] K. Tauber, M. Fusch, J. H. Sattler, J. Pitzer, D. Pressnitz, D.
Koszelewski, K. Faber, J. Pfeffer, T. Haas, W. Kroutil, Chem. Eur.
J. 2013, 19, 4030.
[7] S. Klatte, V. F. Wendisch, Bioorg. Med. Chem. 2014, 22, 5578.
[8] C. A. Müller, A. Denning, T. Welters, T. Winkler, A. J. Ruff, W.
Hummel, H. Grçger, U. Schwaneberg, J. Biotechnol. 2014, 191,
196.
[9] a) A. Robin, V. Kçhler, A. Jones, A. Ali, P. P. Kelly, E. OꢀReilly,
N. J. Turner, S. L. Flitsch, Beilstein J. Org. Chem. 2011, 7, 1494;
b) E. OꢀReilly, V. Kçhler, N. J. Turner, S. L. Flitsch, Chem.
Commun. 2011, 47, 2490; c) E. J. Basom, J. W. Spearman, M. C.
Thielges, J. Phys. Chem. B 2015, 119, 6620.
[10] P. P. Kelly, A. Eichler, S. Herter, D. C. Kranz, N. J. Turner, S. L.
Flitsch, Beilstein J. Org. Chem. 2015, 11, 1713.
[11] a) K. Niefind, J. Müller, B. Riebel, W. Hummel, D. Schomburg, J.
Mol. Biol. 2003, 327, 317; b) M. Reuter, A. Prokoph, J. Kasprzak,
K. Becker, K. Baronian, R. Bode, G. Kunze, H. Vorbrodt, Appl.
Microbiol. Biotechnol. 2015, 99, 4723; c) V. Erdmann, U.
Mackfeld, D. Rother, A. Jakoblinnert, J. Biotechnol. 2014, 191,
106; d) B. Li et al., Sci. Rep. 2014, 4, 6750; e) F. G. Mutti, T.
Knaus, N. S. Scrutton, M. Breuer, N. J. Turner, Science 2015, 349,
1525.
À
native cascades for the enantioselective C H amination of
other substrates.
[12] a) K. Abokitse, W. Hummel, Appl. Microbiol. Biotechnol. 2003,
62, 380; b) C. A. Müller, B. Akkapurathu, T. Winkler, S. Staudt,
W. Hummel, H. Grçgen, U. Schwaneberg, Adv. Synth. Catal.
2013, 355, 1787.
[13] a) M. D. Truppo, N. J. Turner, J. D. Rozzell, Chem. Commun.
2009, 16, 2127; b) M. D. Truppo, N. J. Turner, Org. Biomol.
Chem. 2010, 8, 1280; c) E. OꢀReilly, C. Iglesias, D. Ghislieri, J.
Hopwood, J. L. Galman, R. C. Lloyd, N. J. Turner, Angew.
Chem. Int. Ed. 2014, 53, 2447; Angew. Chem. 2014, 126, 2479;
d) A. P. Green, N. J. Turner, E. OꢀReilly, Angew. Chem. Int. Ed.
2014, 53, 10714; Angew. Chem. 2014, 126, 10890; e) D. Kosze-
lewski, I. Lavandera, D. Clay, G. M. Guebitz, D. Rozzell, W.
Kroutil, Angew. Chem. Int. Ed. 2008, 47, 9337; Angew. Chem.
2008, 120, 9477.
Figure 4. Chiral GC of biotransformation products leading to 4a after
acetylation in comparison with authentic standards.
[14] a) G. Røkke, E. Korvald, J. Pahr, O. Oyås, R. Lale, Methods Mol.
Biol. 2014, 1116, 1; b) BioBrick is a trademark of The BioBricks
Foundation (http://www.biobricks.org).
[15] J. Kim, J. P. Kershner, Y. Novikov, R. K. Shoemaker, S. D.
Copley, Mol. Syst. Biol. 2010, 6, 436.
Acknowledgements
The research leading to these results has received funding
from the Innovative Medicine Initiative Joint Undertaking
under grant agreement no 115360, resources of which are
Received: October 27, 2015
Published online: December 21, 2015
Angew. Chem. Int. Ed. 2016, 55, 1511 –1513
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1513