The catalytic potential of Coptis japonica NCS2 revealed - Development and utilisation of a fluorescamine-based assay ETI

Article abstract of DOI:10.1002/adsc.201200641

The versatility and potential of a norcoclaurine synthase (NCS) from Coptis japonica NCS2 has been investigated, together with the development and application of a novel fluorescence-based high-throughput assay using nearly forty amines/aldehydes. The stereocontrol exerted by CjNCS2 on selected non-natural substrates has been determined, where the tetrahydroisoquinolines (THIAs) were formed as the (1S)-isomer in >95% ee, as observed with the natural product norcoclaurine. Docking calculations involving THIA mechanism intermediates, utilising the reported Thalictrum flavum NCS X-ray crystallographic structure, were carried out and combined with the CjNCS2 screening results to further understand the mode of action of NCS. These findings suggested that in addition to the key active-site residues K122 and E110, D141 is also mechanistically essential for the enzymatic transformation. The exceptional tolerance of NCS towards aldehyde substrates is furthermore supported by our proposed mechanism in which the aldehydes protrude out of the enzymatic pocket. Copyright

Full text of DOI:10.1002/adsc.201200641

FULL PAPERS
DOI: 10.1002/adsc.201200641
The Catalytic Potential of Coptis japonica NCS2 Revealed –
Development and Utilisation of a Fluorescamine-Based Assay
Thomas Pesnot,a Markus C. Gershater,b John M. Ward,b and Helen C. Hailesa,*
a
Department of Chemistry, University College London, Christopher Ingold Laboratories, 20 Gordon Street, London
WC1H 0AJ, U.K.
Fax : (+44)-(0)207-679-7463; e-mail: h.c.hailes@ucl.ac.uk
Research Department of Structural & Molecular Biology, ISMB, University College London, Gower Street, London
b
WC1E 6BT, U.K.
Received: July 20, 2012; Published online: November 4, 2012
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201200641.
Abstract: The versatility and potential of a norco- bined with the CjNCS2 screening results to further
claurine synthase (NCS) from Coptis japonica NCS2 understand the mode of action of NCS. These find-
has been investigated, together with the develop- ings suggested that in addition to the key active-site
ment and application of a novel fluorescence-based residues K122 and E110, D141 is also mechanistically
high-throughput assay using nearly forty amines/al- essential for the enzymatic transformation. The ex-
dehydes. The stereocontrol exerted by CjNCS2 on ceptional tolerance of NCS towards aldehyde sub-
selected non-natural substrates has been determined, strates is furthermore supported by our proposed
where the tetrahydroisoquinolines (THIAs) were mechanism in which the aldehydes protrude out of
formed as the (1S)-isomer in >95% ee, as observed the enzymatic pocket.
with the natural product norcoclaurine. Docking cal-
culations involving THIA mechanism intermediates, Keywords: alkaloids; biotransformations; norcoclaur-
utilising the reported Thalictrum flavum NCS X-ray ine synthase; Pictet–Spengler reaction; tetrahydroi-
crystallographic structure, were carried out and com- soquinolines
Introduction
for their production, syntheses of THIAs remain
either step-intensive or lack stereocontrol to generate
The benzylisoquinoline alkaloids (BIAs) are a large single isomers.[9,11,12]
and structurally diverse group of secondary metabo-
lites that have a key role in chemical plant defense.[1] a recombinant monoamine oxidase and nanoscale
Many BIA natural products have been studied for
their pharmacological properties including magnoflor-
ine (anti-HIV),[2] the antibiotic berberine,[3] papaver-
ine (coronary vasodilatation),[4] and the well-known
analgesic morphine (Figure 1).[5] A variety of synthetic
BIAs including dihydro- and tetrahydroisoquinoline
alkaloids (DHIA and THIA, respectively) have also
been identified as bioactive molecules against a wide
array of diseases.[6–8] This myriad of biological activi-
ties has highlighted the THIA scaffold as a versatile
pharmacophore, and the BIA natural products and
structural analogues as sought after pharmaceutical
candidates.
Interesting recent reports have described the use of
The chemical synthesis of chiral BIAs is often chal-
lenging as it requires the regioselective functionalisa-
tion of phenols and amines, and the formation of
cyclic structures in a stereocontrolled manner.[8–10] De-
Figure 1. Examples of biologically active BIA natural prod-
spite the development of elegant chemical strategies ucts.
Adv. Synth. Catal. 2012, 354, 2997 – 3008
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2997

Thomas Pesnot et al.
FULL PAPERS
reoselectivities. To develop these properties, high-
throughput assays are invaluable to allow the rapid
characterisation of native and mutant enzyme activi-
ties against a range of substrates. In addition, chiral
assays as well as methods to confirm absolute configu-
rations are vital. When developing enzyme variants
with improved properties such as solvent tolerance
again these assay tools are essential. Here we describe
the versatility and potential of an NCS from Coptis ja-
ponica, the development and application of a novel
fluorescence-based high-throughput assay, the abso-
Scheme 1. NCS-mediated biosynthesis of (S)-norcoclaurine
3.
Pd(0) particles in an enantioselective deracemisation lute stereochemistry of the norcoclaurine formed, and
reaction, or use of a berberine bridge enzyme and ki- for the first time detail the stereocontrol that NCSs
netic resolution to THIAs.[13,14] An alternative syn- can exert on non-natural substrates. In addition, dock-
thetic approach is the use of biosynthetic enzymes at ing calculations have been performed and combined
key stages. The first committed step in the plant bio- with screening results to further understand the mode
synthesis of BIAs is mediated by norcoclaurine syn- of action of NCS.
thase (NCS).[10,15] The enzyme catalyses the stereose-
lective coupling of dopamine 1 and 4-hydroxyphenyl-
acetaldehyde (4-HPAA) 2 to generate (S)-norcoclaur- Results and Discussion
ine (S)-3, the key biosynthetic precursor to BIA natu-
ral products (Scheme 1). The enzyme NCS therefore A previous study investigating norcoclaurine synthase
has significant potential as a tool for the stereoselec- in Coptis japonica revealed NCS activity from cul-
tive production of libraries of THIA and BIA com- tured C. japonica cells when using substrates 1 and 2.
pounds.
Recombinant NCSs were generated in E. coli and
Early NCS studies focused on identifying the natu- designated as C. japonica NCS1 (CjNCS1) and C. ja-
ral precursors involved in BIA alkaloid biosynthesis. ponica PR10A (CjPR10A) since the protein was ini-
The subsequent use of cultured plant cells enabled tially classified as pathogenesis-related 10 (PR10).[25]
NCSs to be purified and characterised in more detail, Both enzymes were reported to produce (S)-3, al-
particularly from Thalictrum flavum.[16,17] More re- though the LC-MS elution time of the products dif-
cently, the cloning and over-expression of T. flavum fered, which was surprisingly suggested to be due to
NCS (TfNCS) led to high levels of expression and a different mechanism of norcoclaurine production.[25]
then elegant mechanistic, X-ray crystallographic and In subsequent work, recombinant CjPR10A was se-
NMR studies, as well as investigations into substrate lected for incorporation into alkaloid pathways, rather
tolerance with a small number of analogues.[18–22] This than CjNCS1, to give (S)-reticuline.[26,27] More recent-
was followed by a scale-up of the reaction using ly, CjNCS1 was further studied and found not to form
TfNCS for the generation of (S)-norcoclaurine.[23] norcoclaurine 3.[28] As part of a study to use little ex-
However, in some of these studies the enzymatic reac- plored NCSs in biocatalytic studies with non-natural
tions were performed under phosphate buffer reac- substrates CjPR10A, denoted here as CjNCS2 (acces-
tion conditions, favouring the undesirable chemical sion number A2A1A1), was obtained as a synthetic
background reaction which can complicate interpreta- gene, subcloned into pET29a and expressed in E. coli.
tion of the results.[12,16–18,20,22,23] In parallel with our cur- After purification of the his-tagged protein 0.3 mg/mL
rent studies, TfNCS has been recently used to gener- of protein were isolated (4.4 U/mg of protein). When
ate a range of THIAs to expand its application in bio- used in a bioconversion (5 mM scale) with 1 and 2,
catalytic transformations.[24] Notably, in TfNCS studies the formation of norcoclaurine 3 was confirmed by
to date, limited attention has been focused on the ste- HPLC and comparison to racemic 3 generated by
reoselectivity of the reaction. The CD spectra in some a biomimetic reaction.[12]
reports has indicated that one isomer was formed
In order to evaluate the catalytic versatility of
preferentially but here ees were not determined.[20,22] CjNCS2 and validate it as tool for the synthesis of
However, in the scale-up study using TfNCS with novel THIAs several assays were considered, includ-
1 and 2 an ee for norcoclaurine 3 of 93% was report- ing HPLC and GC. However, to enhance throughput,
ed.[23] In the recent publication to generate non-natu- particularly for future library screening, fluorescent or
ral THIAs using TfNCS no stereoselectivities were re- colorimetric screens were considered. Fluorescamine
ported.[24]
4, is a water soluble spirolactone that exclusively gen-
For biocatalytic applications enzymes need to be erates fluorescent pyrrolidinones upon reaction with
readily expressed, stable under the conditions used, primary amines.[29–33] This chemoselectivity for pri-
accept a broad range of substrates, and give high ste- mary amines (e.g., dopamine 1) to generate fluores-
2998
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2997 – 3008

The Catalytic Potential of Coptis japonica NCS2 Revealed
Figure 2. Fluorescamine-based assay results and control re-
action. Reactions with dopamine 1 were performed at 0.4
and 0.1 mgmLÀ1 CjNCS2. Non-productive NCS control re-
actions were performed with tyramine
5 at 0.4 and
Scheme 2. Fluorescamine assay yielding high fluorescence
with primary amines 1 and 5, but not 3.
0.1 mgmLÀ1 CjNCS2. Error bars are standard deviations
from triplicate experiments.
cent species over secondary amines (e.g., norcoclaur- cence was noted. A significant fluorescence signal was
ine 3) that give non-fluorescent furanones was utilised observed due to CjNCS2 C-terminal amine and side
to monitor NCS activity by quantifying the depletion chain reactions with 4. Indeed, incubation with
of dopamine 1 (Scheme 2). Fluorescence intensities 0.4 mg/mL of CjNCS2 yielded more than 50% of
resulting from the reaction of 4 with dopamine 1 and background fluorescence, compared to a 25% level
norcoclaurine 3 were measured and the signal was with 0.1 mg/mL of the protein (Figure 2). This led to
1000-fold greater for the pyrrolidinone, than for the the selection of CjNCS2 at concentrations of 0.1 mg/
furanone (Scheme 2; see the Supporting Information). mL for use in the assay as it gave a rapid rate of con-
Fluorescamine auto-fluorescence was also observed sumption of dopamine, but also limited the back-
but at a negligible level. Purified CjNCS2 in HEPES ground fluorescence signal. Overall, these results
buffer was reacted with dopamine 1 and increasing demonstrated that a fluorescamine-based assay could
concentrations of 4-HPAA 2 at 378C over 1 h. The be used to quantitate the consumption of dopamine
addition of fluorescamine 4 after 1 h generated fluo- in an NCS-mediated reaction.
rescence by reaction with the residual dopamine 1. A
After validating and establishing procedures for the
drop in fluorescence intensity was observed which fluorescamine-based assay it was used for screening li-
was dependent on the concentration of 4-HPAA used, braries of amine and aldehyde substrate analogues.
showing
a
typical dose-response correlation While several substrates were obtained from commer-
(Scheme 2, Figure 2). It should be noted that in the cial sources, many compounds such as arylacetalde-
absence of NCS, but in a suitable medium (phosphate hydes and variously decorated phenethylamines were
buffer), dopamine can readily react with 4-HPAA to commercially unavailable so were synthesised. The
yield norcoclaurine.[12] However, the use of HEPES aryl and heteroaromatic acetaldehydes were readily
buffer minimised the non-enzymatic reaction (<1%). prepared from either the corresponding terminal al-
Several control reactions were performed to con- kenes via ozonolysis or the corresponding alcohols via
firm that the consumption of dopamine in the assay a Parikh–Doering oxidation.[34] This enabled access to
was due to a CjNCS2-mediated reaction. Tyramine 5 a variety of electron-rich, electron-deficient, carbocy-
(a dopamine analogue inert towards NCSs) instead of clic and polyfunctionalised aldehydes (6–32)
dopamine was used with CjNCS2 and a high fluores- (Table 1). Commercially available amines (1, 33–37)
cence signal observed irrespective of the concentra- (Table 2) were used, together with 3-substituted phen-
tion of 4-HPAA, indicating no reaction had occurred
ACHTUNGTRENeNGNU thylamines (38–42) prepared from the corresponding
(Scheme 2, Figure 2). When no enzyme was used in phenylacetonitriles via standard reduction reactions.
the assay an elevated and constant fluorescence signal Using the assay procedure established above, the
was observed, whereas in the absence of 4, no fluores- CjNCS2-mediated conversion of dopamine was ini-
Adv. Synth. Catal. 2012, 354, 2997 – 3008
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2999

Thomas Pesnot et al.
FULL PAPERS
3000
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2997 – 3008

The Catalytic Potential of Coptis japonica NCS2 Revealed
Table 2. Amines used in the fluorescence assay and the con-
versions observed.
(Table 1; + + +) after 1 h with electron-donating
groups (2, 12, 17) and electron-withdrawing moieties
(15, 16). The conversions were higher than for phenyl-
acetaldehyde 19, perhaps reflecting aldehyde solubili-
ties.
Compound number Amine
Conversion[a,b,c] [%]
1
+ + + (81)
meta-Substituted arylacetaldehydes were also ac-
cepted, though less readily, with conversions towards
50% (Table 1; + +) (9, 10). The combination of meta-
and para-substitution lead to reduced levels of con-
version in the range 10–40% (Table 1; trace, +) (6, 7,
14, 18). Previous studies using NCS from cultured C.
japonica cells reported that 18 and 19 produced the
corresponding THIAs (no data shown), and more re-
cently CjNCS was successfully used with substrate 18
in a pathway to (S)-reticuline, consistent with our re-
sults using CjNCS2.[25,27] The pentafluorophenyl ana-
logue 13 gave low conversions with CjNCS2 (Table 1;
trace). Our data using arylacetaldehydes suggested
that unfavourable steric interactions may occur in the
active site when going from 4-, to 3- to further substi-
tuted substrates. Interestingly, in recent work with
TfNCS, several 2-, 3-, 4-, and 3,4-arylacetaldehydes
were used with 1 and conversions determined using
an HPLC dopamine consumption-based assay.[24] All
conversions were in the range 51–71%, although a 3-h
reaction time was used, perhaps enabling less differ-
entiation of substrate tolerances.[24] With CjNCS2 the
heteroaromatic acetaldehydes 8 and 11 gave moder-
ate conversion levels (Table 1; +), highlighting the po-
tential of using this enzyme with a range of function-
alised aldehydes. Crucially, under the reaction condi-
tions used non-enzymatic conversions were low with
the aryl and heteroaromatic acetaldehydes (2, 6–19).
Benzaldehydes (21, 22, 26, 27) and 2-naphthalde-
hyde (30) were poor substrates for CjNCS2, with low
or no enzymatic activities compared to the non-enzy-
matic reaction, which was still evident even in
HEPES buffer (see the Supporting Information).
Short aliphatic aldehydes showed more varied activi-
ties. Formaldehyde 24 is a highly reactive substrate
that is fully converted in the absence of enzyme and
33
34
35
36
0
0
0
0
37
+ (33)
38
39
40
41
+ + (77)
0
0
0
0
42
[a]
In each well was 20 mL amine (0.5 mM in H2O), 20 mL 2
(1 mM in CH3CN) and 160 mL of purified CjNCS2
(140 mU in pH 7.4 HEPES). The fluorescence micro-
plates were incubated for 1 h at 378C and reactions
quenched with 50 mL of 4 (2 mM in CH3CN). Reactions
were performed in triplicate.
Conversions represented by + + + (>80%), + + (40–
80%), + (25–40%), trace (10–25%), 0 (<10%) and % in
parentheses; see the Supporting Information, sd Æ5%.
Amine 42 was consumed in the assay; a secondary LC-
MS assay confirmed no THIA product was formed.
[b]
[c]
tially explored with aldehydes 2 and 6–32 (Table 1). was not explored further. CjNCS2 was inactive to-
In order to detect any non-enzymatic reactions, con- wards ethanal 20 and cyclopentane-carboxaldehyde
trols with buffer replacing the CjNCS2 enzyme solu- 23, however isobutyraldehyde 28 was accepted at de-
tion were performed alongside each enzymatic reac- tectable levels. For TfNCS, no products were reported
tion. Experiments in the absence of dopamine 1 gave when using formaldehyde, ethanal 20 or benzaldehyde
the background fluorescence level, and in the absence 21, consistent with our data, however competing non-
of 2 corresponded to a 0% conversion (no consump- enzymatic conversions were not noted.[24] The longer
tion of dopamine). These controls enabled compara- chain aliphatic aldehyde heptanal 25 and hydrocin-
bility between plates (see the Supporting Informa-
ACHTUNGTRENnNUNG amaldehyde 32 proved to be good substrates for
tion). Data from the assays indicated the tolerance of CjNCS2 with enzymatic conversions of approximately
CjNCS2 to aldehydes and % conversions were calcu- 60% (Table 1; + +). The more sterically constrained
lated (Table 1). The assay demonstrated that all aryl cinnamaldehyde 29 and substitution of the aldehyde
and heteroaromatic acetaldehydes studied (2, 6–19) moiety in 2 for a ketone (31) precluded both the en-
were substrates for CjNCS2. Substitution at the para- zymatic and non-enzymatic reactions. These results
position on the aromatic ring was particularly well tol- overall demonstrated that the fluorescent assay was
erated, with almost complete enzymatic conversion able to provide data on the differential acceptance of
Adv. Synth. Catal. 2012, 354, 2997 – 3008
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3001

Thomas Pesnot et al.
FULL PAPERS
Table 3. Biotransformations to generate THIAs 3, 43–45.
a range of aldehyde analogues by CjNCS2. It simulta-
neously provided information on the non-enzymatic
reaction, which must be assessed for each new sub-
strate. In general arylacetaldehydes were good to ex-
cellent substrates with dopamine for CjNCS2 and
some aliphatic aldehydes could be used as substrates
by the enzyme.
Amine Aldehyde THIA product Yield[a] [%] ee [%]
The substrate tolerance of CjNCS2 towards a selec-
tion of amines was also investigated with 4-HPAA 2
(Table 2; see the Supporting Information). Out of the
amines (1, 33–42) tested, CjNCS2 displayed signifi-
cant activity towards compounds 1, 37 and 38. The
common feature uniting these three analogues is
a meta-phenol group. The lack of a phenol at this po-
sition (33, 34) or substitution with an electron-donat-
ing amine, or electron-withdrawing halide group (39–
41) resulted in no product formation. In previous
work aniline 39 gave THIA products in a biomimetic
phosphate-mediated reaction.[12] Here the lack of ac-
ceptance of 39 compared to 1 by CjNCS2 may be due
to the higher reactivity of the phenolate intermediate
generated via deprotonation in the active site.[22] The
presence of substituents on the ethylamine moiety re-
duced amine consumption with CjNCS2 from approx-
imately 80% with 38 down to 35% with metaraminol
(37), probably due to unfavourable steric or hydro-
gen-bonding interactions of 37 in the active site. A
complete loss of fluorescence was observed with 3-ni-
trophenethylamine 42, in both the enzymatic and
non-enzymatic reaction. This reaction is mechanisti-
cally unlikely due to the strong electron-withdrawing
effect exerted by the nitro group. In light of this
1
2
99
>95 (S)
1
1
15
25
2
77
>95 (S)
>95 (S)
>95 (S)
56[b]
38
73
[a]
Reactions were carried out using CjNCS2 (7.5 mL; 10 U
in pH 7.4 HEPES), 40 mmol of 1 and 60 mmol of 2, 15 or
25 (to give 3, 43, 44, respectively) or 20 mmol of 38 and
30 mmol of 2 (to give 45) at 378C for 3 h.
The ortho-addition product 44a was also formed in 11%
yield.
[b]
result, enzymatic and non-enzymatic reactions for the formations. It was therefore crucial to determine the
amine and aldehyde substrates were performed and optical purity of not just norcoclaurine 3, but also
reactions analysed by LC-MS (see the Supporting In- compounds 43–45, and additionally to establish the
formation). For 42, neither reaction led to THIA absolute configurations which are essential for the
product formation, highlighting an example of a false bioactivity of THIAs.[3,35]
positive result in the fluorescamine-based assay.
Previous studies using TfNCS have either indicated
To study the application of CjNCS2 as a biocatalyst that one isomer of 3 was formed preferentially using
the biotransformations were then performed on CD (no quantitative stereochemical data), phosphate
a larger scale (5 mM amine, 10 U CjNCS2). Five sub- buffers were used enabling the non-enzymatic reac-
strates were selected from the hits identified in the tion to occur, or stereoselectivities of products other
fluorescamine-based assay. Dopamine 1 was reacted than 3 were not reported.[20,22–24] When cultured C. ja-
with 4-HPAA 2, 4-bromophenylacetaldehyde 15 and ponica cells were used, the results indicated that (S)-3
heptanal 25, and 2 was also reacted with 3-hydroxy- was the major isomer resulting from the bioconver-
phenethylamine 38.
sion, but no ee was reported.[20] The synthesis of 45
All four enzymatic reactions yielded THIA prod- has also been reported using TfNCS, although no
ucts (3, 43–45) in good to excellent isolated yields, yield or selectivity were given and reaction times
comparable to conversion yields calculated from the were over 24 h in phosphate buffer, which is known
fluorescamine-based assay (Table 1 and Table 2). Neg- to catalyse the non-enzyme-mediated coupling of
ligible conversions were observed in enzyme-free con- these amines and aldehydes.[20]
trol reactions, demonstrating that these were exclu-
To facilitate the routine determination of product
sively produced via CjNCS2-mediated reactions. The stereoselectivities from NCS-mediated reactions, and
reaction between 1 and heptanal 25 also generated absolute configurations, Mosherꢁs derivatisation of
1
the ortho-addition product 44a in 11% isolated yield the product followed by analysis of the H NMR spec-
(Table 3). One of the essential advantages biocatalysts tra was performed.[37,38] Initially the stereoselectivity
have is their ability to catalyse stereoselective trans- of norcoclaurine 3 produced using CjNCS2 was inves-
3002
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2997 – 3008

The Catalytic Potential of Coptis japonica NCS2 Revealed
Figure 3. 1H NMR spectra (in ppm) for 1-H (5.60–5.95 ppm) of CjNCS2-3 (top), rac-3 (middle) and (1S)-3 (>80% ee)
(bottom), all derivatised with (R)-MTPA chloride.
tigated. Racemic 3 and (S)-3 (>80% ee by chiral ly at 5.70 ppm [Figure 3, CjNCS2-3 (top)], indicating
HPLC of N-Boc-3 and Mosherꢁs derivatisation) stand- that CjNCS2 catalysed the stereoselective synthesis of
ards were synthesised using a phosphate-mediated re- (S)-norcoclaurine in >95% ee, as observed with
action, and by modification of a published procedure TfNCS.[23]
using a chiral auxiliary Bischler–Napieralski cyclisa-
With the conserved environment at 1-H in 3 and
tion and reduction strategy (see the Supporting Infor- 43–45, the same approach was used for the other
mation).[11,12] These were reacted with (R)-MTPA THIAs. Compounds 43, 44 and 45 were derivatised
1
1
chloride. Analysis of the H NMR spectrum of the with (R)-MTPA chloride and analysis of the H NMR
rac-3 derivative highlighted differences between the spectra revealed that for all three analogues, a unique
(1S,2’S) and (1R,2’S) diastereoisomers, particularly at 1-H signal was observed between 5.5 ppm and
1-H, which was observed at 5.70 ppm and 5.89 ppm 5.7 ppm with no other diastereoisomer detected at
[Figure 3, rac-3 (middle)]. Analysis of (S)-3 (>80% a higher chemical shift. This indicated that again the
ee) indicated that the major 1-H signal at 5.70 ppm (1S)-enantiomer of 43–45 was formed in high stereo-
[Figure 3, (1S)-3 (bottom)] corresponded to the selectivities of >95% ee.
(1S,2’S)-isomer, and that at 5.89 ppm to the (1R,2’S)-
While the fluorescamine-based assay demonstrated
isomer. The derivatisation was repeated with that recombinant NCSs such as CjNCS2 are attractive
CjNCS2-3, and the 1-H signal was observed exclusive- catalysts for the stereoselective production of THIAs,
Adv. Synth. Catal. 2012, 354, 2997 – 3008
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3003

Thomas Pesnot et al.
FULL PAPERS
it also highlighted some catalytic limits of the enzyme
regarding substrate acceptance. In order to develop
more versatile NCS catalysts, for instance via site-di-
rected mutagenesis, a thorough understanding of the
topology and mechanism of action of NCS are essen-
tial. To date, there have been few reports of the cata-
lytic mode of action of NCS. The tertiary structure of
TfNCS has recently been investigated by Rçsch et al.
via CD and NMR titration experiments and Boffi
et al. using X-ray crystallography.[19,21,22,39] In the crys-
tallographic studies, a stacked configuration was ob-
served between the non-productive aldehyde used
and 1, with the aldehyde carbonyl hydrogen-bonded
to the mechanistically important residue Lys122.[22]
However, the stacked configuration reported would
suggest the formation of (R)-norcoclaurine, unless
there is substantial reorientation within the active
site. This may be because the 4-HPAA non-produc-
tive analogue 4-hydroxybenzaldehyde used, has
a more electron-withdrawing aryl ring than 2 and may
therefore bind to NCS in an unproductive orientation.
Glu110 and Tyr108 were also highlighted as potential-
ly important residues mechanistically. The NMR ex-
periments used methyl(4-hydroxyphenyl) acetate and
(4-hydroxyphenyl)ethanol as HPAA 2 substitutes, and
suggested that Glu75 may be a catalytically important
residue.[39] With further mechanistic insights required,
computational docking experiments were performed
(with autodock4.2 and autodock vina[40]) using the re-
ported crystal structure of TfNCS (PDB 2VQ5).[22]
TfNCS has a high 77% sequence homology to
CjNCS2 (see the Supporting Information), and there-
fore provides a suitable model for the investigation of
CjNCS2. The aminol productive intermediates 46a–d
(Figure 4, a) which lead to the THIAs 3 and 43–45
were docked: analysis of these docking calculations
highlighted two putative binding modes. A first bind-
ing mode (binding mode A, Figure 4, b and c) was the
lowest in energy, and characterised by the aminol
adopting an L-shape conformation to fit the enzymat-
ic cavity with the catechol group buried in the cavity
and the R1 aldehyde moiety occupying the entrance
to the active site. In the second putative binding
Figure 4. a) Structure of aminol intermediates 46a–d. b)
Lowest energy docking conformations of binding mode A
and aminols 46a–d in red (46a and d), blue, yellow, respec-
tively (2VQ5 in grey ribbon and water channel is highlight-
ed with red spheres). c) TfNCS1 active site with compound
46a represented as an electronic density surface around
a green core (binding mode A). Amino acid residues located
within 5 ꢂ to the putative aminol intermediates are high-
lighted in grey with catalytically active residues highlighted
mode (binding mode B, not shown), aminols 46 also in cyan.
adopted an L-shape topology but with a flipped con-
formation in which the aldehyde group R1 is buried in
the enzyme cavity and the catechol occupies the en- extension the R1 end of the corresponding aminol in-
trance to the active site. These two new binding termediate, occupies a large space in the active site
modes are in contrast to the stacked X-ray crystallo- with limited steric constraints. In this respect, the
graphic configuration.[22]
stacked model proposed by Boffi and co-workers is
The fluorescence assay results showed that CjNCS2 disfavoured.[22] Similarly, the putative binding mode B
tolerates aldehydes that are significantly more steri- with the R1 of the aminol deeply buried in the enzy-
cally challenging than 4-HPAA 2, the natural sub- matic cavity is less likely. In the proposed binding
strate (e.g., 2-indolecarboxaldehyde 8 and 3-methoxy- mode A, the end of the R1 substituent occupies the
phenylacetaldedyde 10), or possess longer alkyl entrance to the active site. As a result, bulkier sub-
chains (e.g., heptanal 25 and hydrocinnamaldehyde strates and those possessing longer alkyl chains could
32). This suggests that the aldehyde substrate, and by more readily be tolerated by the enzyme, as they
3004
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2997 – 3008

The Catalytic Potential of Coptis japonica NCS2 Revealed
would protrude out of the enzymatic cavity towards
the solvent.
The fluorescence screen also highlighted a differen-
tial in CjNCS2 tolerance towards arylacetaldehyde
substituents with 4-substituted arylacetaldehydes (2,
12, 15–17) being excellent substrates, while 3-arylace-
taldehydes and heteroaromatic acetaldehydes (8–11)
were moderate to good substrates, and 3,4-disubstitut-
ed arylacetaldehydes (6, 7, 14, 18) and pentafluoro-
ACHTUNGTRENNUNGphenylacetaldehyde 13 were generally poor sub-
strates. This reactivity is in accordance with binding
mode A which results in the aldehyde 4-aryl substitu-
ents directed towards the solvent, 3-aryl substituents
slightly oriented towards active site residues and poly-
substituted arylacetaldehydes facing major steric chal-
lenges with hydrophobic residues (Phe80, Phe99 and
Tyr108).
The poor substrate tolerances observed with the
benzaldehydes 21, 22, 26, 27 could, according to bind-
ing mode A, be due to the aromatic ring being too
deeply buried in the enzymatic cavity resulting in un-
favourable steric interactions with residues from
within the enzymatic pocket. In addition 23 was not
accepted and 28 was a very poor substrate, both pos-
sessing substitution at the a-position, giving rise to
unfavourable steric interactions. Poor substrate ac-
ceptance of a-substituted aldehydes has been noted
for TfNCS.[24] Binding mode A furthermore supports
the observed poor substrate tolerance towards the
amines, where metaraminol 37 was a poorer substrate
for CjNCS2 than dopamine 1 or 3-hydroxyphenethyl-
amine 38. In binding mode A, the phenethylamine
group in 46 docks into a sterically hindered binding
pocket. Analysis of CjNCS2 substrate tolerances from
our fluorescent screen therefore strongly support
binding mode A. Similar substrate tolerance studies
recently reported for TfNCS and other mechanistic
studies are also are in accordance with binding mode
A.[20,24,25]
Figure 5. Proposed NCS enzyme mechanism with 1 and 15.
a) Docking of imine intermediate. b) Docking of intermedi-
ate prior to rearomatisation.
The combined analysis of docking (theoretical) and nucleophilicity of the amine functionality by deproto-
assay (empirical) results has lead to the formulation nation of the ammonium cation (predominant at neu-
of a putative mechanism for the catalytic activity of tral pH). The aminol intermediate resulting from con-
NCS, building upon earlier studies (Figure 5). Initially, densation of the amine with the aldehyde subsequent-
binding of the phenethylamine substrate to the cata- ly undergoes dehydration to form an iminium ion, sta-
lytic pocket via hydrogen bonding could occur be- bilised by an electrostatic interaction with Asp141 or
tween the 3-hydroxy moiety and Lys122 (~2.5 ꢂ). An- potentially Tyr108, highlighted as an important resi-
choring of the amine substrate could be further en- due in previous work.[22] The putative dehydration
hanced by hydrogen bonding of the amine functional- would then be promoted by the evacuation of a water
ity with Asp141 (~2.4 ꢂ). Subsequent entry by the al- molecule via a water channel (Figure 4, b, Figure 5).
dehyde substrate into the enzymatic cavity, could be The formation of the isoquinoline bond could then be
held in place via hydrogen bonding with Glu110 (~ assisted by residues Lys122 and Asp141, in a concerted
2.8 ꢂ). Then Glu110 and Asp141 residues assist the manner: abstraction of the 3-hydroxy proton on 1 by
nucleophilic attack of the primary amine onto the al- Lys122 would generate a reactive phenolate anion
dehyde. Due to their proximity to one another, that can cyclise onto the iminium cation (stabilised by
Glu110 and Asp141 could complex to both amine and Asp141) and assist the electrophilic addition reaction
aldehyde functions and orientate them into a produc- to form the 1,2-hydroquinoline ring (Figure 5, a). The
tive conformation. Asp141 would also enhance the stereoselectivity associated with the cyclisation step
Adv. Synth. Catal. 2012, 354, 2997 – 3008
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3005

Thomas Pesnot et al.
FULL PAPERS
would be attributed to the topology of the active and can be used to generate novel THIAs in high ste-
pocket orientating the iminium intermediate for the reoselectivities.
formation of the (S)-enantiomer. Re-aromatisation of
Examination of the fluorescence assay results com-
the 1,2-hydroquinoline into the tetrahydroisoquino- bined with docking calculations also leads to the for-
line product is mediated by the abstraction of 8a-H mulation of an alternative catalytic mechanism for
(Figure 5, b). A possible candidate for this step is NCS, with the aldehyde protruding out of the enzy-
Glu110 in agreement with previous work,[22] which ac- matic pocket, explaining the versatility towards alde-
cording to docking calculations is 2.2 ꢂ away from hyde substrates. Our putative mechanism revealed
8a-H (see the Supporting Information), and could three key enzymatic residues (E110, K122 and D141)
also be assisted by Tyr108. Following rearomatisation, that are essential for the catalytic activity of NCS.
the tetrahydroisoquinoline product leaves the active This, in turn, may help engineer NCS mutants with
site. This proposed mode of action is consistent with enhanced activities and substrate tolerances, and pro-
data from previous studies: the essential role of vide an essential tool for the regio- and stereocon-
Lys122 has been demonstrated by the generation of trolled synthesis of THIAs.
a site-specific mutant K122A which was inactive and
mutants E110A and Y108F demonstrated lower activ-
ities.[22]
Experimental Section
The putative role of amino acid residues bearing
carboxylates such as Asp141 and Glu110 for the enzy- Heterologous Expression and Purification of CjNCS2
matic catalysis of Pictet–Spengler reactions is further
supported by studies on strictosidine synthase, which
CjNCS2 (also known as CjPR10 A, accession A2A1A1) was
obtained as a synthetic gene with the addition of six C ter-
catalyses the condensation of the aldehyde secologa-
minal histidine residues from DNA 2.0. The synthetic con-
nin with tryptamine.[41] Here residue Glu309 was iden-
struct was subcloned into pET29a (Novagen) using NdeI
tified as essential for the Pictet–Spengler condensa-
tion. Not only is it involved in holding together the
amine and aldehyde residues, but it is also believed to
enhance the nucleophilicity of tryptamine and facili-
tate the final rearomatisation step by the abstraction
of proton 2-H.[41,42] Our putative mechanism which
complies with previous empirical results, suggests that
the catalytic activity observed with NCS is based on
the key amino acid residues Glu110, Lys122 and
Asp141. However, our docking calculations highlight-
ed many other residues that are likely to be catalyti-
cally inactive, but involved in sculpting the topology
of the active site (Figure 5).
and HindIII, and the resulting construct was used to trans-
form Escherichia coli BL21 (DE3) pLysS (Promega). Terrif-
ic broth with 50 mg/mL kanamycin was inoculated with 4%
overnight culture (grown in terrific broth, 378C, 250 rpm)
and was incubated for 2 h at 378C, shaking at 250 rpm,
before lowering the temperature to 308C for 1 h. The cul-
tures were then induced with 0.5 mM IPTG and grown for
a
further 3 h. Cells were harvested by centrifugation
(10000ꢃg, 10 min) and lysed using sonication (20ꢃ15 s at
10 m amplitude) before being purified by nickel affinity
chromatography. Purified protein was then desalted using
a PD10 desalting column (GE Healthcare).
General Procedure for the CjNCS2-Mediated
Synthesis of THIAs
Conclusions
To a mixture of the amine in water (1 equiv.) and aldehyde
in acetonitrile (1.5 equiv.) was added purified CjNCS2
(7.5 mL, 10 U) in HEPES buffer (0.1M, pH 7.4). The reac-
tion mixture was stirred at 378C for 3 h, and subsequently
quenched by the addition of HCl (1M; 1 mL). The crude
mixtures were filtered to remove enzyme precipitate and
purified by semi-preparative HPLC according to either gra-
dient 1 or gradient 2. Semi-preparative HPLC were per-
A novel fluorescence assay based on the chemoselec-
tive reactivity of fluorescamine for primary amines
has been developed. The assay allowed us to map
CjNCS2 tolerance for both amine and aldehyde sub-
strates. A limited set of dopamine analogues were tol-
erated by the enzyme, and highlighted the require-
ment of a meta-hydroxy moiety in dopamine, consis- formed on a Varian Prostar instrument equipped with an au-
tosampler, a UV-visible detector and a DiscoveryBIO wide
Pore C18–10 Supelco column (25ꢃ2.12 cm). Elutions were
monitored at 280 nm and carried out according to either of
the following gradients. Gradient 1: 5% to 40% of acetoni-
trile/water (0.1% TFA) and gradient 2: 5% to 90% of aceto-
nitrile/water (0.1% TFA). Fractions containing the desired
product were combined and concentrated under reduced
pressure. Iterative co-evaporation of the residue with metha-
nol (3ꢃ5 mL) allowed the complete removal of any residual
TFA and yielded the product as pale yellow to colourless
glassy solids.
tent with previous reports. In contrast, CjNCS2
showed surprising versatility towards aldehyde sub-
strates. Variously decorated arylacetaldehydes were
excellent substrates for CjNCS2, which can also ac-
commodate some aliphatic aldehydes. These results
were validated by four successful biotransformations
with selected substrates, and stereoselectivities estab-
lished (>95% ee). Overall these results demonstrated
that CjNCS2 has significant potential as a biocatalyst,
3006
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2997 – 3008

The Catalytic Potential of Coptis japonica NCS2 Revealed
Synthesis of 3, 43–45 using CjNCS2
27.0, 30.0, 32.7, 33.0, 38.7, 53.4, 112.4, 116.0, 120.4, 123.3,
144.8, 146.4; HR-MS (ESI+): m/z=250.1812, calcd. for
C15H24NO2 [MH+]: 250.1807.
Determination of the absolute configuration is described in
the Supporting Information via Mosherꢁs derivatisation.
(1S)-1-(4-Hydroxybenzyl)-1,2,3,4-tetrahydroisoquinoline-
6,7-diol [(S)-norcoclaurine; (S)-3]: Compound (S)-3 was pre-
pared according to the general procedure from dopamine
1 (40 mmol in 400 mL water) and 4-hydroxyphenylacetalde-
hyde 2 (60 mmol in 120 mL acetonitrile). The product was
purified by semi-preparative HPLC using gradient 1 (reten-
tion time: 11.8 min), to give (S)-3 as a colourless glassy
solid; yield: 10.7 mg (99%). Spectroscopic characterisation
data was identical to that previously described and for (S)-3
(see the Supporting Information).[12] [a]D25: À25.0 (c 1.0,
MeOH) {lit. [a]2D5: À24.7 (MeOH)}.[23]
(1S)-1-(4-Hydroxybenzyl)-1,2,3,4-tetrahydroisoquinoline-
6-ol [(S)-45]: Compound (S)-45 was prepared according to
the general procedure from 2-(3-hydroxyphenyl)ethylamine
38 (20 mmol in 400 mL water) and 4-hydroxyphenylacetalde-
hyde 2 (30 mmol in 120 mL acetonitrile). The product was
purified by semi-preparative HPLC using gradient 2 (reten-
tion time: 11.1 min) to give (S)-45 as a colourless glassy
solid; yield: 4.8 mg (73%). IR (neat): nmax =3475, 3008,
2962, 1668, 1611, 1530 cmÀ1
;
1H NMR (600 MHz;
CD3OD):[20] d=2.95–3.11 (m, 3H, CHCHH and 4-H2), 3.28–
3.30 (m, 1H, 3-HH), 3.40 (dd, J=14.5 and 5.8 Hz, 1H,
CHCHH), 3.48 (app. quintet, J=6.3 Hz, 1H, 3-HH), 4.64
(dd, J=8.9 and 5.8 Hz, 1H, 1-H), 6.65 (d, J=2.5 Hz, 1H, 5-
H), 6.69 (dd, J=8.5 and 2.5 Hz, 1H, 7-H), 6.81 (d, J=
8.5 Hz, 2H, 3’-H and 5’-H), 7.03 (d, J=8.5 Hz, 1H, 8-H),
7.13 (d, J=8.5 Hz, 2H, 2’-H and 6’-H); 13C NMR (150 MHz;
CD3OD): d=26.4, 40.4, 40.6, 58.0, 115.6, 116.0, 116.9, 123.6,
126.9, 129.2, 131.7, 133.9, 158.3, 158.5; HR-MS (ESI+): m/
z=256.1342, calcd. for C16H18NO2 [MH+]: 256.1338.
(1S)-1-(4-Bromobenzyl)-1,2,3,4-tetrahydroisoquinoline-
6,7-diol [(S)-43]: Compound (S)-43 was prepared according
to the general procedure from dopamine 1 (40 mmol in
400 mL water) and 4-bromophenylacetaldehyde 15 (60 mmol
in 120 mL acetonitrile). The product was purified by semi-
preparative HPLC using gradient
2 (retention time:
13.1 min) to give (S)-43 as a colourless glassy solid; yield:
10.3 mg (77%). IR (neat): nmax =2950 br, 1670, 1598,
1
1532 cmÀ1; H NMR (600 MHz; CD3OD): d = 2.90–3.03 (m,
2H, 4-H2), 3.08 (dd, J=14.4 and 8.5 Hz, 1H, CHCHH), 3.28
(app. quint, J=6.4 Hz, 1H, 3-HH), 3.42 (dd, J=14.4 and
6.0 Hz, 1H, CHCHH), 3.47 (app. quint, J=6.4 Hz, 1H, 3-
HH), 4.65 (dd, J=8.2 and 6.3 Hz, 1H, 1-H), 6.54 (s, 1H, 8-
H), 6.63 (s, 1H, 5-H), 7.25 (d, J=8.4 Hz, 2H, 2’-H and 6’-
H), 7.55 (d, J=8.4 Hz, 2H, 3’-H and 5’-H); 13C NMR
(150 MHz; CD3OD): d=25.6, 40.6, 40.8, 57.4, 114.2, 116.2,
122.6, 123.3, 123.6, 132.6, 133.3, 136.0, 145.8, 147.0; MS
Acknowledgements
We gratefully acknowledge the Biotechnology and Biological
Sciences Research Council (BBSRC) (BB/G014426/1) for
funding T.P. and M.C.G.
(ES+): m/z=336 [MHACTHNUTRGENNGU ACHTUGNTRENNUGN
(81Br)+, 48%], 334 [MH(79Br)+, 50],
References
319 (50), 317 (52), 273 (100); HR-MS (ESI+): m/z=
334.0433, calcd. for C16H1779BrNO2 [MH+]: 334.0443.
(1S)-1-Hexyl-1,2,3,4-tetrahydroisoquinoline-6,7-diol [(S)-
44] and 1-hexyl-1,2,3,4-tetrahydroisoquinoline-7,8-diol (44a):
Compounds 44 and 44a were prepared according to the gen-
eral procedure from dopamine 1 (40 mmol in 400 mL water)
and heptanal 25 (60 mmol in 120 mL acetonitrile). The prod-
uct was purified by semi-preparative HPLC using gradient 2
to give (S)-44 (yield: 5.6 mg, 56%) (retention time:
14.1 min) and 44a (yield: 1.1 mg, 11%) (retention time:
13.8 min).
[1] D. K. Liscombe, B. P. MacLeod, N. Loukanina, O. I.
Nandi, P. J. Facchini, Phytochemistry 2005, 66, 1374–
1393.
[2] M. A. Rashid, K. R. Gustafson, Y. Kashman, J. H. Car-
dellina, J. B. McMahon, M. R. Boyd, Nat. Prod. Lett.
1995, 6, 153–156.
[3] K. Iwasa, M. Kamigauchi, M. Ueki, M. Taniguchi, Eur.
J. Med. Chem. 1996, 31, 469–478.
[4] Y. J. Kim, H. K. Hong, H. S. Lee, S. H. Moh, J. C. Park,
S. H. Jo, H. Choe, J. Cardiovasc. Pharmacol. 2008, 52,
485–493.
[5] A. J. Goodman, B. Le Bourdonnec, R. E. Dolle, Chem-
MedChem 2007, 2, 1552–1570.
[6] K. Iwasa, M. Moriyasu, Y. Tachibana, H.-S. Kim, Y.
Wataya, W. Wiegrebe, K. F. Bastow, L. M. Cosentino,
M. Kozuka, K.-H. Lee, Bioorg. Med. Chem. 2001, 9,
2871–2884.
[7] R. Gitto, L. De Luca, S. Ferro, S. Agnello, E. Russo, G.
De Sarro, A. Chimirri, Chem. Pharm. Bull. 2010, 58,
1602–1605.
Major regioisomer (S)-44: IR (neat): nmax =3008 br, 1671,
1613 cmÀ1
;
1H NMR (600 MHz; CD3OD): d=0.93 (t, J=
5.9 Hz, 3H, (CH2)4CH3), 1.32–1.53 (m, 8H, (CH2)4CH3),
1.87 (m, 1H, CHCHH), 2.03 (m, 1H, CHCHH), 2.86–3.02
(m, 2H, 4-H2), 3.29–3.33 (m, 1H, 3-HH), 3.50 (app. quintet,
J=6.1 Hz, 1H, 3-HH), 4.34 (dd, J=7.9 and 5.0 Hz, 1H, 1-
H), 6.61 (s, 1H, 5-H), 6.65 (s, 1H, 8-H); 13C NMR
(150 MHz; CD3OD): d=14.4, 23.6, 25.7, 26.4, 30.2, 32.7,
35.2, 41.0, 56.7, 113.8, 116.2, 123.6, 124.2, 145.9, 146.7; MS
(ES+): m/z=250 (MH+, 100%), 233 ([MHÀOH]+, 72);
HR-MS (ESI+): m/z=250.1812, calcd. for C15H24NO2
[MH+]: 250.1807.
[8] J. D. Scott, R. M. Williams, Chem. Rev. 2002, 102, 1669–
1730.
Minor ortho-regioisomer 44a: 1H NMR (600 MHz;
CD3OD): d=0.93 (t, J=5.9 Hz, 3H, (CH2)4CH3), 1.30–1.62
(m, 8H, (CH2)4CH3), 1.83 (m, 1H, CHCHH), 2.12 (m, 1H,
CHCHH), 2.92–3.04 (m, 2H, 4-H2), 3.29–3.33 (m, 1H, 3-
HH), 3.47 (m, 1H, 3-HH), 4.61 (dd, J=9.5 and 3.4 Hz, 1H,
1-H), 6.55 (d, J=8.2 Hz, 1H, 5-H), 6.74 (d, J=8.2 Hz, 1H,
6-H); 13C NMR (150 MHz; CD3OD): d=14.4, 23.6, 25.5,
[9] M. Chrzanowska, M. D. Rozwadowska, Chem. Rev.
2004, 104, 3341–3370.
[10] J. Stçckigt, A. P. Antonchick, F. Wu, H. Waldmann,
Angew. Chem. 2011, 123, 8692–8719; Angew. Chem.
Int. Ed. 2011, 50, 8538–8564.
[11] A. L. Zein, O. O. Dakhill, L. N. Dawe, P. E. Georghiou,
Tetrahedron Lett. 2010, 51, 177–180.
Adv. Synth. Catal. 2012, 354, 2997 – 3008
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3007

Thomas Pesnot et al.
FULL PAPERS
[12] T. Pesnot, M. C. Gershater, J. M. Ward, H. C. Hailes,
Chem. Commun. 2011, 47, 3242–3244.
[13] J. M. Foulkes, K. J. Malone, V. S. Coker, N. J. Turner,
[27] A. Nakagawa, H. Minami, J.-S. Kim, T. Koyanagi, T.
Katayama, F. Sato, H. Kumagai, Nat. Commun. 2011, 2,
334.
J. R. Lloyd, ACS Catal. 2011, 1, 1589–1594.
[28] E.-J. Lee, P. Facchini, Plant Cell 2010, 22, 3489–3503.
[29] S. Udenfriend, S. Stein, P. Bçhlen, W. Dairman, W.
Leimgruber, M. Weigele, Science 1972, 178, 871–872.
[30] P. Bçhlen, S. Stein, J. Stone, S. Udenfriend, Anal. Bio-
chem. 1975, 67, 438–445.
[31] W. L. Baker, Analyst 1997, 122, 447–453.
[32] M. L. Read, T. Etrych, K. Ulbrich, L. W. Seymour,
FEBS Lett. 1999, 461, 96–100.
[14] J. H. Schrittwieser, V. Resch, J. H. Sattler, W.-D. Lien-
hart, K. Durchschein, A. Winkler, K. Gruber, P. Ma-
cheroux, W. Kroutil, Angew. Chem. 2011, 123, 1100–
1103; Angew. Chem. Int. Ed. 2011, 50, 1068–1071.
[15] P. J. Facchini, B. St-Pierre, Curr. Opin. Plant Biol. 2005,
8, 657–666.
[16] N. Samanani, P. J. Facchini, Planta 2001, 213, 898–906.
[17] N. Samanani, P. J. Facchini, J. Biol. Chem. 2002, 277,
33878–33883.
[18] N. Samanani, D. K. Liscombe, P. J. Facchini, Plant J.
2004, 40, 302–313.
[19] H. Berkner, J. Engelhorn, D. K. Liscombe, K. Schweim-
er, B. M. Wçhrl, P. J. Facchini, P. Rçsch, I. Matecko,
Protein Expression Purif. 2007, 56, 197–204.
[20] L. Y. P. Luk, S. Bunn, D. K. Liscombe, P. J. Facchini,
M. E. Tanner, Biochemistry 2007, 46, 10153–10161.
[21] A. Pasquo, A. Bonamore, S. Franceschini, A. Macone,
A. Boffi, A. Ilari, Acta Crystallogr. Sect. F: Struct. Biol.
Cryst. Commun. 2008, 64, 281–283.
[22] A. Ilari, S. Franceschini, A. Bonamore, F. Arenghi, B.
Botta, A. Macone, A. Pasquo, L. Bellucci, A. Boffi, J.
Biol. Chem. 2009, 284, 897–904.
[33] C. F. Zhang, L. Fan, P. Yang, Chinese Chem. Lett. 2007,
18, 97–98.
[34] T. Hirose, T. Sunazuka, T. Zhi-Ming, M. Handa, R.
Uchida, K. Shiomi, Y. Harigaya, S. Omura, Heterocy-
cles 2000, 53, 777–784.
ˇ
ˇ ´
ˇ
[35] J. Rꢅcny, F. Gualtieri, S. Tucek, Physiol. Res. 2002, 51,
131–137.
[36] B. E. Maryanoff, D. F. McComsey, J. C. Craig, Chirality
1998, 10, 169–172.
[37] L. F. Tietze, Y. Zhou, E. Tçpken, Eur. J. Org. Chem.
2000, 2247–2252.
[38] T. R. Hoye, M. K. Renner, J. Org. Chem. 1996, 61,
2056–2064.
[39] H. Berkner, K. Schweimer, I. Matecko, P. Rçsch, Bio-
chem. J. 2008, 413, 281–290.
[40] O. Trott, A. J. Olson, J. Comput. Chem. 2010, 31, 455–
461.
[41] X. Ma, S. Panjikar, J. Koepke, E. Loris, J. Stçckigt,
Plant Cell 2006, 18, 907–920.
[42] J. J. Maresh, L.-A. Giddings, A. Friedrich, E. A. Loris,
S. Panjikar, B. L. Trout, J. Stçckigt, B. Peters, S. E.
OꢁConnor, J. Am. Chem. Soc. 2008, 130, 710–723.
[23] A. Bonamore, I. Rovardi, F. Gasparrini, P. Baiocco, M.
Barba, C. Molinaro, B. Botta, A. Boffi, A. Macone,
Green Chem. 2010, 12, 1623–1627.
[24] B. M. Ruff, S. Brꢄse, S. E. OꢁConnor, Tetrahedron Lett.
2012, 53, 1071–1074.
[25] H. Minami, E. Dubouzet, K. Iwasa, F. Sato, J. Biol.
Chem. 2007, 282, 6274–6282.
[26] H. Minami, J.-S. Kim, N. Ikezawa, T. Takemura, T. Ka-
tayama, H. Kumagai, F. Sato, Proc. Natl. Acad. Sci.
USA 2008, 105, 7393–7398.
3008
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2997 – 3008