Synthesis of D- and L-Phenylalanine Derivatives by Phenylalanine Ammonia Lyases: A Multienzymatic Cascade Process

Article abstract of DOI:10.1002/anie.201410670

The synthesis of substituted D-phenylalanines in high yield and excellent optical purity, starting from inexpensive cinnamic acids, has been achieved with a novel one-pot approach by coupling phenylalanine ammonia lyase (PAL) amination with a chemoenzymatic deracemization (based on stereoselective oxidation and nonselective reduction). A simple high-throughput solid-phase screening method has also been developed to identify PALs with higher rates of formation of non-natural D-phenylalanines. The best variants were exploited in the chemoenzymatic cascade, thus increasing the yield and ee value of the D-configured product. Furthermore, the system was extended to the preparation of those L-phenylalanines which are obtained with a low ee value using PAL amination.

Full text of DOI:10.1002/anie.201410670

Angewandte
Chemie
DOI: 10.1002/anie.201410670
Biocatalysis
Synthesis of d- and l-Phenylalanine Derivatives by Phenylalanine
Ammonia Lyases: A Multienzymatic Cascade Process**
Fabio Parmeggiani, Sarah L. Lovelock, Nicholas J. Weise, Syed T. Ahmed, and
Nicholas J. Turner*
Abstract: The synthesis of substituted d-phenylalanines in
high yield and excellent optical purity, starting from inex-
pensive cinnamic acids, has been achieved with a novel one-pot
approach by coupling phenylalanine ammonia lyase (PAL)
amination with a chemoenzymatic deracemization (based on
stereoselective oxidation and nonselective reduction). A simple
high-throughput solid-phase screening method has also been
developed to identify PALs with higher rates of formation of
non-natural d-phenylalanines. The best variants were exploited
in the chemoenzymatic cascade, thus increasing the yield and
ee value of the d-configured product. Furthermore, the system
was extended to the preparation of those l-phenylalanines
which are obtained with a low ee value using PAL amination.
d-phenylalanines would represent an atom-economic and
low-cost approach compared to alternative methods.[3]
Nonetheless, to date no evidence of wild-type PAL
enzymes which are able to produce d-phenylalanines with
high selectivity have been found. We recently reported[4] that
PAL-catalyzed amination of the cinnamic acids 1 can lead to
the formation of significant levels of the enantiomers d-2
together with the expected l-2 products (Scheme 1, left). In
T
he asymmetric hydroamination of activated alkenes, cata-
lyzed by ammonia lyases, represents a very attractive
approach to the synthesis of enantiomerically enriched
chiral amino acids. In particular, phenylalanine ammonia
lyases (PALs, EC 4.3.1.24), identified in various organisms,
mainly plants and yeasts, catalyze the interconversion of
cinnamic acids and l-phenylalanines.[1] The reaction is
synthetically very valuable since it does not require expensive
cofactors or recycling systems, and its practical applications
have already been demonstrated on small and large scale.[2]
However, the use of PALs for the production of non-
natural d-phenylalanines would increase their value further,
since the latter are important building blocks for many natural
products (e.g. macrolide antibiotics) and active pharmaceut-
ical ingredients [e.g. nateglinide and PPACK (d-phenylalanyl-
l-prolyl-l-arginine chloromethyl ketone)], either as key
chiral structural fragments, or in peptides to confer resistance
to enzymatic hydrolysis. The PAL-mediated synthesis of
Scheme 1. Amination/deracemization cascade concept.
particular, for cinnamic acids with an electron-deficient
aromatic ring, a considerable reduction in the ee value of l-
2 was observed over time. Herein, we exploit this activity and
describe a system for the synthesis of optically enriched d-
amino acids. For this purpose we designed a cascade process
involving the deracemization of 2 (i.e., an enantioselective
conversion of l-2 into a prochiral intermediate, along with
a nonspecific complementary reaction which regenerates rac-
2, thus resulting in the accumulation of d-2).[5] Specifically, we
selected a system involving the enzymatic oxidation of l-2
into the corresponding imino acid 3 and the nonselective
chemical reduction of the latter (Scheme 1, right). Such
a reduction can be carried out with several reagents, for
example, borohydrides, borane complexes, or catalytic trans-
fer hydrogenation. The borane–ammonia complex is the
mildest, cheapest, and most efficient option.[6] Its compati-
bility with proteins and its stability in water at high pH values
and high ammonia concentrations made it ideal for our
purpose. For the oxidation step, the first choice would be an l-
amino acid oxidase (LAAO), although no enzyme of this class
with broad substrate specificity has been successfully
expressed in a prokaryotic host in an active form.[7] However,
the same reaction is also performed by l-amino acid
deaminases (LAADs),[8] which are membrane-bound pro-
teins associated with the respiratory electron transport chain,
and use oxygen as a co-substrate but produce water instead of
H2O2. We selected the LAAD from the enterobacterium
[*] Dr. F. Parmeggiani, Dr. S. L. Lovelock, N. J. Weise, S. T. Ahmed,
Prof. N. J. Turner
Manchester Institute of Biotechnology and School of Chemistry
University of Manchester
131 Princess Street, M1 7DN, Manchester (UK)
E-mail: nicholas.turner@manchester.ac.uk
[**] This work was funded by the Biotechnology and Biological Sciences
Research Council (BBSRC) and Glaxo-SmithKline (GSK) under the
Strategic Longer and Larger (sLoLa) grant initiative (ref. BB/
K00199X/1). N.J.W. was supported by the European Union’s 7th
Framework program FP7/2007-2013 under grant agreement no.
289646 (KYROBIO). We thank the Royal Society for a Wolfson
Research Merit Award (N.J.T.). Dr. Emma Jones (GSK) is gratefully
acknowledged for fruitful discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201410670.
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Proteus mirabilis,[9] which has been successfully expressed in
E. coli, is active on a broad spectrum of substrates, and has
already been exploited as a whole-cell biocatalyst on prep-
arative scale.[10]
The activity of LAAD in the deracemization of the model
substrate p-nitrophenylalanine (rac-2a) was tested under
different conditions (Table 1), at high pH and in the presence
variabilis[11] is particularly suited for biocatalysis for three
reasons: 1) the expression level is higher than nonbacterial
PALs, 2) the enzyme possesses a broader substrate specific-
ity,[2f] and 3) the X-ray crystal structure has been solved with
the mobile loops in the active closed conformation, thus
facilitating mutagenesis studies.
The coupling of PAL amination and LAAD/NH3:BH3
deracemization (Scheme 1) resulted in 71% conversion of
p-nitrocinnamic acid (1a) into d-2a with 96% ee, and would
thus be useful for preparative applications compared to
alternative methods. However, perfect enantioselectivity
could not be obtained, and we therefore decided to try to
improve the performance of the PAL enzyme by directed
evolution.
Table 1: Deracemization of rac-2a into d-2a with LAAD.[a]
Buffer system
pH
NH3:BH3
[equiv]
ee [%][b]
2a [%][b]
KPi 100 mm
KPi 100 mm
NH4OH 100 mm
NH4OH 1m
NH4OH 3m
NH4OH 5m
NH4OH 5m
NH4OH 5m
7.0
8.0
9.0
9.0
9.0
9.0
9.6
9.6
30
30
30
30
30
30
30
50
>99
>99
>99
96
92
93
84
82
82
80
81
79
78
87
Initially we developed
a high-throughput screening
method for the identification of PAL variants with increased
selectivity for the formation of the compounds d-2. The assay
(Scheme 2) is based on the detection of hydrogen peroxide
formed by the reaction of the d-amino acid with a d-amino
acid oxidase (DAAO, from Trigonopsis variabilis)[12] coex-
pressed together with the PAL variant, thus yielding the
corresponding imino acid 3. In the presence of horseradish
peroxidase (HRP), H2O2 is reduced to water, and colorless
3,3’-diaminobenzidine (DAB) is oxidized to a polymeric
brown dye which accumulates in the colony.[13] The intensity
of the color is proportional to the concentration of peroxide,
and related in turn to the amount of d-2 formed by PAL.
As a starting pool of variants to test this screening
platform, we employed a mixture of 48 separate NNK site-
saturation mutagenesis libraries (960 variants) targeting 48
residues in close proximity to the binding site (see Fig-
ure S1).[14] The screening was performed as follows (Figure 1):
1) ultracompetent E. coli BL21(DE3) cells were transformed
with two plasmids, one harbouring the PAL variant and the
95
>99
[a] Reaction conditions: 5 mm rac-2a, 15 mgmLÀ1 LAAD E. coli wet cells,
378C, 220 rpm, 7 h. [b] Determined by HPLC analysis.
of high concentrations of NH3, thus confirming its compat-
ibility with the optimal conditions for PAL aminations. The
addition of at least 40 equivalents of borane allowed complete
deracemization under these conditions. The formation of
small amounts of the keto acid 4 (see Scheme 1) by sponta-
neous hydrolysis of 3 led to a slight reduction in overall yield.
The use of the alternative reducing agents mentioned above
did not lead to any improvement in conversion (see Table S1
in the Supporting Information).
Then we tested the performance of wild-type PAL in the
coupled system. The PAL from the cyanobacteria Anabaena
Scheme 2. Reactions involved in the solid-phase screening for d-PAL activity.
Figure 1. Protocol for the solid-phase screening for d-PAL activity.
2
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
other the DAAO, and plated onto a nylon
membrane laid on an agar plate; 2) col-
onies were grown overnight at 308C;
3) the membrane was transferred to
a fresh plate containing IPTG and incu-
bated overnight at 208C; 4) partial lysis
of the cells was achieved by freeze–thaw
cycles in liquid nitrogen; 5) the mem-
brane was then treated with a HRP
solution to remove any peroxide present;
6) the membrane was transferred to
a filter paper soaked with a solution of
substrate, NH3, HRP, and DAB, and
incubated at room temperature.
p-Nitrocinnamic acid (1a) was
selected as a model substrate, although
similar behavior was observed with other
d-forming substrates (1b–e): a qualitative
estimation of the activity of the variant in
each colony could be easily carried out by
visual inspection (see inset in Scheme 1
and Figures S2 and S3). No color change
was observed when employing the sub-
strates 1 f–j, which are known to afford
Scheme 3. Enantiocomplementary amination/deracemization cascades to afford either optically
pure d- or l-2.
exclusively the l-enantiomer, even after a prolonged incuba-
tion with the wild-type enzyme.[4] The screening was then
performed on approximately 5000 colonies and the most-
active clones were picked, sequenced, and assayed in
analytical-scale biotransformations to measure conversion
and product ee values by HPLC. All positive clones were
found to produce d-2a at a faster or similar rate compared to
the wild-type enzyme (see Table S2), and several variants
appeared to be considerably faster. To validate the assay, 35
clones randomly picked from the colorless colonies showed
low to no formation of d-2a when analysed by HPLC (see
Table S3).
To avoid bias resulting from varying levels of protein
expression, the most active variants were purified and
analytical-scale biotransformations were performed using
normalized protein concentrations (see Figure S4). The best
mutants thus identified are listed in Table 2. In spite of their
enhanced d-PAL activity, none of the variants led to a change
in the final position of the equilibrium, thus giving a practi-
cally racemic product and no useful enrichment of d-2a, as
previously observed with the wild-type enzyme.[4]
However, we anticipated that these d-PAL variants could
be exploited in our amination/deracemization cascade
system. By employing the best mutant identified from the
screening (H359Y) the ee value of d-2a increased to greater
than 99% and the conversion to 78% under the same reaction
conditions, thus further demonstrating the faster production
of d-2a with respect to wild-type PAL (other variants
behaved similarly; see Table S4).
The amination/deracemization cascade (Scheme 3a) was
then tested with a range of electron-deficient cinnamic acids
(1b–e), thus affording d-2b–e in good yields and excellent
optical purities (Table 3). Since the deracemization also
works efficiently when starting from pure l-2, the same
approach was extended successfully to the cinnamic acid
substrates 1 f–j, thus leading to optically pure d-2 f–j. A
representative reaction profile for 1 f is shown in Figure 2.
Even with the cascade system, some of the cinnamic acid
remained unconverted as a consequence of the reversible
nature of the PAL reaction. However, the amount of
unconverted 1 is lower than in the case of PAL alone, thus
Table 2: PAL variants with the highest specific d-formation activity.[a]
Variant
Specific d-PAL
Relative activity
compared to WT
activity [UmgÀ1
]
WT
3.23
11.35
10.76
9.83
9.10
7.95
–
H359Y
H359K
S456P
A88Q
S263A
3.52
3.34
3.05
2.82
2.47
[a] Reaction conditions: 5 mm 1a, 5m NH4OH, 25 mgmLÀ1 purified PAL,
pH 9.6, 378C, 220 rpm, 4 h.
Figure 2. Composition profile and ee values for the conversion of 1 f
into d-2 f (ee values are positive for l-2 f and negative for d-2 f).
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Table 3: Synthesis of d- and l-2a–j from 1a–j.[a]
Keywords: amino acids · biocatalysis · enzymes ·
lyases · synthetic methods
.
PAL(H359Y)/LAAD cas-
cade
PAL(WT)/DAAO cas- PAL(WT) only
cade
Subs.
R
Conv.
ee d-2
Conv.
ee l-2
Conv. ee l-2
[%][b]
[%][b]
[%][b]
[%][b]
[%][b] [%][b]
[1] N. J. Turner, Curr. Opin. Chem. Biol. 2011, 15, 234 –
240; M. M. Heberling, B. Wu, S. Bartsch, D. B.
Janssen, Curr. Opin. Chem. Biol. 2013, 17, 250 – 260.
1a
1b
1c
1d
1e
1 f
1g
1h
1i
p-NO2
m-NO2
o-NO2
p-CF3
p-CN
o-F
m-F
p-F
3,5-F2
m-Cl
79
74
80
64
77
77
74
62
78
63
>99
98
98
98
99
>99
98
>99
>99
99
82
76
83
66
79
>99
>99
>99
>99
>99
n.r.
n.r.
n.r.
n.r.
n.r.
84
79
91
65
72
84
67
53
57
86
rac
89
rac
76
70
>99
>99
>99
97
´
[2] a) A. Gloge, J. Zon, A. Kçvꢀri, L. Poppe, J. Rꢁtey,
Chem. Eur. J. 2000, 6, 3386 – 3390; b) C. Paisz, A.
Katona, J. Rꢁtey, Chem. Eur. J. 2006, 12, 2739 – 2744;
c) B. de Lange, D. J. Hyett, P. J. D. Maas, D. Mink,
F. B. J. van Assema, N. Sereinig, A. H. M. de Vries,
J. G. de Vries, ChemCatChem 2011, 3, 289 – 292;
d) S. Bartsch, U. T. Bornscheuer, Angew. Chem. Int.
Ed. 2009, 48, 3362 – 3365; Angew. Chem. 2009, 121,
3412 – 3415; e) S. Bartsch, U. T. Bornscheuer, Pro-
tein Eng. Des. Sel. 2010, 23, 929 – 933; f) S. L. Love-
lock, N. J. Turner, Bioorg. Med. Chem. 2014, 22,
n.r.
n.r.
n.r.
n.r.
n.r.
1j
>99
[a] Reaction conditions: 5 mm subs., 5m NH4OH, 25 mgmLÀ1 PAL E. coli wet cells,
35 mgmLÀ1 LAAD or DAAO E. coli wet cells, 40 equiv NH3:BH3, pH 9.6, 378C,
220 rpm. [b] Determined by HPLC analysis (yields of isolated products are given in
the Supporting Information). n.r.=not required, because of the high ee value
products from the PAL-mediated reaction alone.
´
5555 – 5557; g) B. Wu, W. Szymanski, G. G.
Wybenga, M. W. Heberling, S. Bartsch, S. de Wilde-
man, G. J. Poelarends, B. L. Feringa, B. W. Dijkstra,
D. B. Janssen, Angew. Chem. Int. Ed. 2012, 51, 482 –
486; Angew. Chem. 2012, 124, 497 – 501; h) W.
Szymanski, B. Wu, B. Weiner, S. de Wildeman,
B. L. Feringa, D. B. Janssen, J. Org. Chem. 2009, 74, 9152 – 9157.
[3] A. S. Bommarius, M. Schwarm, K. Drauz, Chimia 2001, 55, 50 –
59.
showing at least a partial beneficial effect of the cascade
coupling of amination and deracemization (Figure 2 and
Table 3).
[4] S. L. Lovelock, R. C. Lloyd, N. J. Turner, Angew. Chem. Int. Ed.
2014, 53, 4652 – 4656; Angew. Chem. 2014, 126, 4740 – 4744.
[5] N. J. Turner, Curr. Opin. Chem. Biol. 2004, 8, 114 – 119; T. M.
Beard, N. J. Turner, Chem. Commun. 2002, 246 – 247; S. Servi, D.
Tessaro, G. Pedrocchi-Fantoni, Coord. Chem. Rev. 2008, 252,
715 – 726; M. Rachwalski, N. Vermue, P. J. T. Rutjes, Chem. Soc.
Rev. 2013, 42, 9268 – 9282.
[6] F. Alexandre, D. P. Pantaleone, P. P. Taylor, I. G. Fotheringham,
D. J. Ager, N. J. Turner, Tetrahedron Lett. 2002, 43, 707 – 710.
[7] L. Pollegioni, P. Motta, G. Molla, Appl. Microbiol. Biotechnol.
2013, 97, 9323 – 9341.
[8] G. S. Hossain, J. Li, H. Shin, G. Du, L. Liu, J. Chen, Appl.
Microbiol. Biotechnol. 2014, 98, 1507 – 1515; D. P. Pantaleone,
A. M. Geller, P. P. Taylor, J. Mol. Catal. B 2001, 11, 795 – 803.
[9] G. Massad, H. Zhao, H. L. T. Mobley, J. Bacteriol. 1995, 177,
5878 – 5883.
[10] R. L. Hanson, B. L. Davis, S. L. Goldberg, R. M. Johnston, W. L.
Parker, T. P. Tully, M. A. Montana, R. N. Patel, Org. Process Res.
Dev. 2008, 12, 1119 – 1129.
Furthermore, we were able to adapt this strategy (Scheme
3b) to the formation of optically pure l-2a–e from 1a–e by
employing E. coli producing DAAO instead of LAAD. This
approach achieves enantiocomplementarity with the same
panel of biocatalysts, thus providing access to substituted l-
phenylalanines which result in low ee values in direct PAL
aminations (Table 3).
In conclusion, an efficient high-throughput solid-phase
screening assay was successfully developed and used to
identify variants with enhanced activity towards d-phenyl-
alanines, thus providing a starting platform for further
directed evolution studies. The best variant was exploited in
a cascade process to convert a range of cinnamic acids into
optically pure d-phenylalanine derivatives. The cascade was
also adapted to enhance the enantioselectivity for l-phenyl-
alanines starting from the same cinnamic acids. This approach
(employed in both the screening and the preparative appli-
cation), represents a useful multienzymatic one-pot system
involving PALs and thus widens the range of applications of
these valuable biocatalysts in preparative biotransformations.
[11] M. C. Moffitt, G. V. Louie, M. E. Bowman, J. Pence, J. P. Noel,
B. S. Moore, Biochemistry 2007, 46, 1004 – 1012.
[12] L. Lin, H. R. Chien, W. Wang, T. Hwang, H. Fu, W. Hsu, Enzyme
Microb. Technol. 2000, 27, 482 – 491.
[13] a) M. Alexeeva, A. Enright, M. J. Dawson, M. Mahmoudian,
N. J. Turner, Angew. Chem. Int. Ed. 2002, 41, 3177 – 3180; Angew.
Chem. 2002, 114, 3309 – 3312; b) S. C. Willies, J. L. White, N. J.
Turner, Tetrahedron 2012, 68, 7564 – 7567.
Received: November 2, 2014
Revised: January 8, 2015
Published online: && &&, &&&&
[14] The library was supplied by Codexis, Inc. (US).
4
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Biocatalysis
PAL around: The cascade coupling of
phenylalanine ammonia lyase (PAL) with
l-amino acid deaminase (LAAD) or d-
amino acid oxidase (DAAO), led to the
conversion of cheap and easily accessible
cinnamic acids into both non-natural d-
and l-phenylalanines by a chemoenzy-
matic process. PAL variants possessing
higher d-activity were selected by using
a novel solid-phase screening assay and
employed to increase yield and ee values.
F. Parmeggiani, S. L. Lovelock,
N. J. Weise, S. T. Ahmed,
N. J. Turner*
&&& — &&&
Synthesis of d- and l-Phenylalanine
Derivatives by Phenylalanine Ammonia
Lyases: A Multienzymatic Cascade
Process
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!