Structural and mechanistic studies on anthocyanidin synthase catalysed oxidation of flavanone substrates: The effect of C-2 stereochemistry on product selectivity and mechanism

Article abstract of DOI:10.1039/b507153d

During the biosynthesis of the tricyclic flavonoid natural products in plants, oxidative modifications to the central C-ring are catalysed by Fe(II) and 2-oxoglutarate dependent oxygenases. The reactions catalysed by three of these enzymes; flavone synthase I, flavonol synthase and anthocyanidin synthase (ANS), are formally desaturations. In comparison, flavanone 3β-hydroxylase catalyses hydroxylation at the C-3 pro-R position of 2S-naringenin. Incubation of ANS with the unnatural substrate (±)-naringenin results in predominantly C-3 hydroxylation to give cis-dihydrokaempferol as the major product; trans-dihydrokaempferol and the desaturation product, apigenin are also observed. Labelling studies have demonstrated that some of the formal desaturation reactions catalysed by ANS proceed via initial C-3 hydroxylation followed by dehydration at the active site. We describe analyses of the reaction of ANS with 2S- and 2R-naringenin substrates, including the anaerobic crystal structure of an ANS - Fe-2-oxoglutarate-naringenin complex. Together the results reveal that for the 'natural' C-2 stereochemistry of 2S-naringenin, C-3 hydroxylation predominates (>9:1) over desaturation, probably due to the inaccessibility of the C-2 hydrogen to the iron centre. For the 2R-naringenin substrate, desaturation is significantly increased relative to C-3 hydroxylation (ca. 1:1); this is probably a result of both the C-3 pro-S and C-2 hydrogen atoms being accessible to the reactive oxidising intermediate in this substrate. In contrast to the hydroxylation-elimination desaturation mechanism for some ANS substrates, the results imply that the ANS catalysed desaturation of 2R-naringenin to form apigenin proceeds with a syn-arrangement of eliminated hydrogen atoms and not via an oxygenated (gem-diol) flavonoid intermediate. Thus, by utilising flavonoid substrates with different C-2 stereochemistries, the balance between C-3 hydroxylation or C-2, C-3 desaturation mechanisms can be altered. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b507153d

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Structural and mechanistic studies on anthocyanidin synthase
catalysed oxidation of flavanone substrates: the effect of C-2
stereochemistry on product selectivity and mechanism
Richard W. D. Welford,† Ian J. Clifton, Jonathan J. Turnbull, Stuart C. Wilson
and Christopher J. Schofield*
Chemical Research Laboratory, Department of Chemistry and Oxford Centre for Molecular
Sciences, Mansfield Rd, Oxford, UK OX1 3TA. E-mail: christopher.schofield@chem.ox.ac.uk;
Fax: +44 (0)1865 275674; Tel: +44 (0)1865 275625
Received 20th May 2005, Accepted 19th July 2005
First published as an Advance Article on the web 1st August 2005
During the biosynthesis of the tricyclic flavonoid natural products in plants, oxidative modifications to the central
C-ring are catalysed by Fe(II) and 2-oxoglutarate dependent oxygenases. The reactions catalysed by three of these
enzymes; flavone synthase I, flavonol synthase and anthocyanidin synthase (ANS), are formally desaturations. In
comparison, flavanone 3b-hydroxylase catalyses hydroxylation at the C-3 pro-R position of 2S-naringenin.
Incubation of ANS with the unnatural substrate (± )-naringenin results in predominantly C-3 hydroxylation to give
cis-dihydrokaempferol as the major product; trans-dihydrokaempferol and the desaturation product, apigenin are
also observed. Labelling studies have demonstrated that some of the formal desaturation reactions catalysed
by ANS proceed via initial C-3 hydroxylation followed by dehydration at the active site. We describe analyses of the
reaction of ANS with 2S- and 2R-naringenin substrates, including the anaerobic crystal structure of an
ANS–Fe–2-oxoglutarate–naringenin complex. Together the results reveal that for the ‘natural’ C-2 stereochemistry of
2S-naringenin, C-3 hydroxylation predominates (>9 : 1) over desaturation, probably due to the inaccessibility of the
C-2 hydrogen to the iron centre. For the 2R-naringenin substrate, desaturation is significantly increased relative to
C-3 hydroxylation (ca. 1 : 1); this is probably a result of both the C-3 pro-S and C-2 hydrogen atoms being accessible
to the reactive oxidising intermediate in this substrate. In contrast to the hydroxylation–elimination desaturation
mechanism for some ANS substrates, the results imply that the ANS catalysed desaturation of 2R-naringenin to form
apigenin proceeds with a syn-arrangement of eliminated hydrogen atoms and not via an oxygenated (gem-diol)
flavonoid intermediate. Thus, by utilising flavonoid substrates with different C-2 stereochemistries, the balance
between C-3 hydroxylation or C-2, C-3 desaturation mechanisms can be altered.
Introduction
During flavonoid biosynthesis, flavone synthase I catalyses
desaturation of 2S-naringenin 3 to give apigenin 4 (Fig. 1A).
It is thought that this reaction proceeds by abstraction of the
The flavonoids are polyphenolic natural products found in plants
and some micro-organisms. They play roles in pigmentation,
protection against UV photodamage, can act as signalling
molecules, and have been reported to possess a variety of bio-
medicinal properties, including anti-malarial, anti-oxidant and
anti-tumour activities.
One of the key sites for structural diversification of flavonoids
is the central C-ring. Modifications to this ring are brought about
C-3 pro-R and C-2 hydrogen atoms that are in a syn-facial
4,15
arrangement.
Flavanone 3b-hydroxylase catalyses hydroxy-
lation of the pro-R hydrogen at C-3 of 2S-naringenin 3 pro-
ducing 2R,3R-trans-dihydrokaempferol 5 (2R,3R-trans-DHK).
Flavonol synthase catalyses a formal desaturation of 2R,3R-
trans-dihydroflavonols (e.g. 5) to form flavonols. However, the
results from incubations of flavonol synthase and ANS with un-
natural flavanone (e.g. 2S-naringenin 3) substrates have provided
evidence that both these enzymes catalyse C-3 hydroxylation of,
at least, some dihydroflavonol (e.g. 5) substrates, followed by
1
2,3
by oxygenases and reductases (Fig. 1A). Four of the oxyge-
nases known to catalyse modification of the C-ring in plants
are Fe(II) and 2-oxoglutarate dependent oxygenases: flavone
4
5
6,7
synthase I, flavanone 3b-hydroxylase, flavonol synthase and
6,7,16
dehydration, rather than direct desaturation (Fig. 2A).
8,9
anthocyanidin synthase (ANS) (Fig. 1A).
The non-haem Fe(II), 2-oxoglutarate dependent oxygenase
family of enzymes utilise a 2-oxoglutarate 1 (2OG) co-substrate
17
There is uncertainty about the exact in vivo role of ANS.
ANS was originally thought to catalyse conversion of 2R,3S,4S-
8
cis-leucoanthocyanidins (e.g. 6) to anthocyanidins (e.g. 7).
1
0,11
to effect a 2 electron oxidation of their substrate (Fig. 1B).
However, in vitro incubations of purified ANS from Ara-
bidopsis thaliana (A. thaliana) with 2R,3S,4S-cis-leucocyanidin
Most commonly, substrate oxidation involves hydroxylation,
though some members of the family catalyse other oxidative
reactions such as desaturation or ring expansion. The 2OG 1
co-substrate is decarboxylated to give succinate 2 containing
one oxygen atom derived from dioxygen; the other dioxygen
atom is typically incorporated into the product in the case of
hydroxylation reactions. The oxidising intermediate effecting
6
(2R,3S,4S-cis-LCD) produces cyanidin 7 as only a minor
product, with the major product, quercetin 8 arising from 4
electron oxidation, a process that requires two catalytic cycles
17
(
Fig. 2B). ANS can also catalyse the oxidation of other
flavonoid substrates, often producing a mixture of products.
Incubation of ANS with leucocyanidin of the ‘unnatural’ stere-
ochemistry (2R,3S,4R-trans-LCD, the C-4 epimer of 2R,3S,4S-
IV
hydroxylation is thought to be a ferryl species (Fe =O) (Fig. 1B).
This intermediate has recently been characterised spectroscopi-
cis-LCD 6) leads to a mixture of both 2 and 4 electron oxidation
12–14
cally in the reaction cycle of the enzyme taurine dioxygenase.
9,16,17
products;
quercetin 8 (30%) and the thermodynamically
less stable cis-diastereoisomer of the dihydroflavonol, dihydro-
quercetin (DHQ) (55% of products) predominate, with the latter
being produced as a single enantiomer. ANS also catalyses
†
Current address: UC Berkeley, Department of Chemistry, Berkeley,
CA 94720, USA.
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6
3 1 1 7

View Article Online
Fig. 1 A Current proposals for the flavonoid biosynthetic pathway. Flavonoid nomenclature is indicated on the flavanone (2S-naringenin 3)
structure. Under each structure a flavonoid class name is given and underneath that, the compound name for different hydroxylation patterns is
given. Subsequent to 2S-naringenin 3, ‘B’ is used to represent the B-ring and the A-ring is not shown. This representation is also used in subsequent
figures. For the steps after formation of the leucoanthocyanidins (e.g. 6), the dotted lines represent the originally proposed pathway, while the solid
17
lines indicate a new proposal based on recent in vitro evidence. CHS = chalcone synthase, CHI = chalcone isomerase, FNS I/II = flavone synthase
I or II, FHT = flavanone 3b-hydroxylase, FLS = flavonol synthase, DFR = dihydroflavonol reductase, ANS = anthocyanidin synthase, FGT =
IV
flavonol glycosyl transferase. B Hydroxylation reaction of a Fe(II), 2OG dependent oxygenase; the Fe =O intermediate is shown boxed.
Fig. 2 A Outline mechanism for C-3 hydroxylation and dehydration mechanism for desaturation of 2R,3R-trans-DHQ 9 by ANS or flavonol
23
synthase to form quercetin 8. FLS = flavonol synthase, ANS = anthocyanidin synthase. B Proposed mechanism for formation of quercetin 8 from
17,23
2
R,3S,4S-cis-LCD 6 via an initial C-3 oxidation step—details are given in Turnbull et al. ; a mechanism proceeding via oxidation at C-4 cannot be
ruled out.
3
1 1 8
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6

View Article Online
Table 1 Percentage of total of products formed by ANS incubations with racemic naringenin 3/13 and racemic eriodictyol 14 as analysed by
HPLC. Note that cis-10 and trans-DHK 5/11 were not resolved under the conditions used, see Table 3 for the ratio of cis and trans. DHK =
dihydrokaempferol, DHQ = dihydroquercetin
Substrate
trans-Dihydroflavonol (%)
cis-Dihydroflavonol (%)
Flavone (%)
Flavonol (%)
(
(
± )-Naringenin 3/13
± )-Eriodictyol 14
60 (DHK)
29 (DHQ)
33 (Apigenin)
10 (Luteolin)
7 (Kaempferol)
<1 (Quercetin)
61 (DHQ)
desaturation of 2R,3R-trans-DHQ 9; a reaction catalysed in
vivo by flavonol synthase (Fig. 1A). The crystal structures
of three ANS complexes have been solved: ANS–2OG, with
the unnatural 2R,3R-trans-DHQ 9 substrate ANS–Fe–2OG–
14 was not an efficient substrate of ANS (with ca. 30% of the
turnover of (± )-naringenin 3/13 being observed under standard
conditions). The HPLC analyses indicated that in both cases,
the hydroxylated dihydroflavonol product(s) predominate (e.g.
DHK 5/10 from naringenin), with some desaturation product,
flavone (e.g. apigenin 4 from naringenin) also being observed
(Table 1). The flavone product made up a greater proportion of
the overall product in assays with (± )-naringenin 3/13 than with
(± )-eriodictyol 14. These observations imply that the balance
between destauration and hydroxylation is in part determined
by the hydroxylation pattern on the flavonoid B-ring.
9
DHQ
2
and following exposure of the 2R,3R-trans-DHQ 9
complex crystals to dioxygen, a ANS–Fe–succinate–quercetin–
18
DHQ product complex.
In vitro, both ANS and flavonol synthase have been reported
to catalyse oxidation of 2S-naringenin 3, the natural substrate
of flavone synthase I and flavanone 3b-hydroxylase. We have
communicated that incubation of ANS with the unnatural sub-
strate (± )-naringenin 3/13 gives the C-3 hydroxylation product
cis-DHK 10 as a single enantiomer as the major species (ee
Enantiomerically enriched naringenin substrates (2S-3 and
2R-13) were prepared by the enzymatic cleavage of the appropri-
ate 7-O-rhamnose of 2S- and 2R-naringin using commercially
available naringinase from Penicillium decumbens. The chiral
purity of the naringenin products (3 and 13) could not be
measured directly either by chiral-NMR (the lanthanide chiral
16
>
70%), with lower levels of trans-DHK 5/11 (of undetermined
enantiomeric purity) and kaempferol 12 also being observed.
Flavonol synthase from A. thaliana has also been shown to react
6
with both enantiomers of the unnatural substrate naringenin.
16,17
It converts the ‘natural’ C-2 enantiomer 2S-naringenin 3 to
shift method developed detects dihydroflavonol enantiomers
2
R,3S-cis-DHK 10, 2S,3S-trans-DHK 11 and kaempferol 12.
at the C-3 position) or chiral-HPLC (which was unable to
resolve the enantiomers under the conditions tested with a
ChiraSpherꢀR NT column). However, subsequent HPLC analy-
sis revealed significantly differing product distributions obtained
from incubations of ANS with the racemic 3/13 and the chirally
pure (3 and 13) naringenin substrates (Table 2). These results
infer that the chiral integrity of the naringin had largely been
unaffected under the conditions required for cleavage of its
glycosidic bond.
Similarly flavonol synthase from Citrus unshiu has been shown
to convert 2S-naringenin 3 to kaempferol 12, in a process that
was thought to be due to two catalytic cycles with the first
converting 2S-naringenin 3 to 2R,3R-trans-DHK 5, with this
product then being desaturated enzymatically to kaempferol 12.
Flavonol synthase from both Citrus unshiu and A. thaliana has
been shown to convert 2R-naringenin 13 to 2S,3S-trans-DHK
7
7
6
11, resulting from hydroxylation of the C-3 pro-S hydrogen atom.
1
4
14
Members of the family of non-haem Fe(II), 2OG dependent
Analyses measuring the production of CO
2
from 1-[ C]-
oxygenases are known to catalyse different types of oxidation
reactions with different substrates. For example, clavaminic acid
synthase catalyses three reactions comprising hydroxylation,
oxidative ring expansion and desaturation processes in the
2OG co-substrate has indicated that both 2S-3 and 2R-
23
naringenin 13 are substrates of ANS. The similarity in the
1
4
specific activity values as measured by CO formation using
2
either 2R-13 or 2S-naringenin 3 enantiomers was in marked
contrast to those observed on incubations of ANS with racemic
and chirally pure LCD and DHQ. In the cases of both LCD
and DHQ, ANS was much more active with the chirally
pure substrates of the natural stereochemistry (2R,3S,4S-6
(LCD), and 2R,3R-9 (DHQ) respectively). Thus the presence
of a C-3 hydroxyl group, as in leucoanthocyanidins (6) and
dihydroflavonols (9), enables ANS to be selective for substrates
with a particular C-2 stereochemistry.
HPLC analyses revealed significantly differing product distri-
butions in incubations of ANS with the racemic (3/13), 2S-
3 and 2R-naringenin 13 substrates (Table 2, Figs. 3 and 4).
For incubation with (± )-naringenin 3/13 the major product
observed was DHK. With the ‘natural’ C-2 stereochemistry of
2S-naringenin 3, the major product (∼74% of isolated products
as analysed by HPLC) was also DHK (5/10/11), with only
minor amounts of the desaturated product, apigenin 4 (∼1%),
being formed. However, with the ‘unnatural’ stereochemistry
of 2R-naringenin 13, apigenin 4 was formed in significant
quantities (∼44%). Note in the racemic mixture 3/13, the S-
isomer 3 is likely to be selectively oxidised, hence rationalising
19–21
biosynthesis of clavulanic acid.
Further, the structurally
related Fe(II) dependent oxidase isopenicillin N-synthase has a
diverse substrate and reaction selectivity. With different unnat-
ural substrates it has been observed to catalyse hydroxylation,
22
ring expansion, desaturation and epoxidation reactions. Un-
derstanding the factors that determine which type of oxidative
reaction a non-haem Fe(II) oxygenase carries out is a challenge
that may bring about a better understanding of these catalysts.
This information may also be used in the design of successful
synthetic catalysts that mimic enzymes.
Here we describe the in vitro products of assays of ANS with
R-13 and 2S-naringenin 3, a crystal structure of an ANS–
2
Fe–2OG–naringenin complex and the activity of mutant forms
of ANS. The results demonstrate that in the case of naringenin
substrates, the product selectivity of ANS is dependent upon the
C-2 stereochemistry and that the balance between desaturation
and hydroxylation relates to the accessibility of the C-2 hydrogen
atom to the reactive oxidising intermediate.
Results
Table 2 Ratios of products formed from ANS incubations with racemic
Assays
3
/13, 2S-3 and 2R-13 naringenin as analysed by HPLC
Initially, the product ratios from incubations of ANS (A.
thaliana) with racemic mixtures of unnatural flavanone sub-
strates, (± )-eriodictyol 14 and (± )-naringenin 3/13, which have
differing B-ring hydroxylation patterns, were analysed by HPLC
Substrate
DHK
Apigenin
Kaempferol
(
± )-Naringenin 3/13
60
74
50
33
1
44
7
25
6
2
2
S-Naringenin 3
R-Naringenin 13
(
eriodictyol 14 has the same hydroxylation pattern as naringenin
ꢀ
3
but with an additional 3 hydroxyl group). (± )-Eriodictyol
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6
3 1 1 9

View Article Online
Fig. 3 The reaction of ANS with the 2S-naringenin 3 substrate showing pathways for formation of cis-DHK 10, trans-DHK 5/11 and kaempferol
1
2 products.
Fig. 4 The reaction of ANS with the 2R-naringenin 13 substrate showing pathways for formation of cis-DHK 10, trans-DHK 5/11, apigenin 4 and
kaempferol 12 products.
the similar results obtained for the racemic and S-isomer. These
results imply that a 2R-C-2 stereochemistry coupled to the
presence of a syn hydrogen at C-3 is required for a direct
desaturation of flavanone (e.g. 3) substrates by ANS.
After isolation of the DHK (5/10/11) produced from in vitro
assays of ANS incubated with racemic 3/13, 2S-naringenin 3
and 2R-naringenin 13, chiral HPLC (using a ChiraSpherꢀR NT
chiral column) was used to separate the trans-5/11 and cis-
DHK 10 diastereomers (Table 3). Resolution of the enantiomers
of both trans-5/11 and cis-DHK 10 was not possible under
the conditions tested. HPLC analyses of incubation mixtures
were carried out immediately after quenching, so as to minimise
conversion of the cis-isomers 10 to the more thermodynamically
stable trans-isomers 5/11. The thermodynamic ratio of trans-11 :
cis-10 DHK, is ca. 9 : 1, with isomerisation occurring via C-
ꢀ
16,24
4 hydroxyl assisted opening of the C-ring.
For incubations
of (± )-naringenin 3/13 and 2S-naringenin 3 with ANS, cis-
DHK 10 was produced in excess to the thermodynamically
more stable trans diastereomer (5 and 11) (ca. 4 : 1 and 5.2
: 1, respectively). For the 2R-naringenin 13 substrate of ANS,
the diastereoselectivity of DHK formation was much reduced
with similar amounts of both cis (10 and/or enantiomer) and
trans-DHK 5/11 products being produced.
The flavonoid products from an incubation of ANS with (± )-
naringenin 3/13 were separated from the co-factor, co-products
1
7,23
and enzymes in the assay mixture by solid phase extraction
Table 3 Ratios of trans- and cis-DHK production of ANS incubations
with racemic 3/13, 2S-3 and 2R-13 naringenin as determined by HPLC
1
and analysed by H NMR with minimal exposure to solvent and
elevated temperature (Fig. 5). The H NMR analyses confirmed
1
Substrate
trans-DHK
cis-DHK
the assignment of cis-DHK 10, trans-DHK 5/11 and apigenin
4
products, though insufficient material of the minor product
(
2
2
± )-Naringenin 3/13
1
1
1.1
4.0
5.2
1
kaempferol 12 was present to obtain a definitive spectrum. The
ratio of cis-10 to trans-DHK 5/11 observed (4 : 1 by integration
of the C-2 proton resonance) was similar to that observed by
S-Naringenin 3
R-Naringenin 13
3
1 2 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6

View Article Online
1
Fig. 5 H NMR spectrum of cis 10 and trans-dihydrokaempferol 5/11 purified from an incubation of ANS with (± )-naringenin 3/13.
HPLC. Using NMR spectroscopy, the lanthanide chiral shift
reagent europium tris[3-heptafluoropropylhydroxymethylene)-
the trans-DHK resonance was too small to determine whether
any splitting occurred on addition of the chiral shift reagent;
some or all of the trans-DHK observed may thus be formed
by isomerisation of cis-DHK 10. In comparison for a standard
of (± )-cis/trans-DHQ (9 and stereoisomers) the C-5 hydroxyl
proton resonance of both diastereomers was split on addition of
(
+)-camphorate (Eu(hfc)
3
) can be used to distinguish the
enantiomers of dihydroflavonols, e.g. DHQ (9 and enantiomer)
16
(
Fig. 6). Addition of increasing amounts of Eu(hfc)
cause any splitting of the C-5 hydroxyl proton resonance at
11.75, indicating that cis-DHK 10 was formed, at least
3
did not
d
H
between 0.25 and 1 equivalent of Eu(hfc) (Fig. 6).
3
16
predominantly, as a single enantiomer (Fig. 7). The intensity of
To investigate the orientation of naringenin in the ANS active
site, co-crystallisation of ANS with (± )-naringenin 3/13 was
carried out (Figs. 8–10). In the ANS–Fe–2OG–DHQ–MES
crystal structure, two molecules of DHQ 9 were observed to
be complexed to the active site. The DHQ molecule in place to
be oxidised (i.e. that closest to the iron) was the 2R,3R-trans-9
stereoisomer; its C-3 hydrogen was in a position where it could
potentially be oxidised, with relatively little movement, by an
IV
iron centred ferryl species (Fe =O), while the C-2 hydrogen was
18
more distant and pointed away from the iron centre. This ob-
servation is consistent with a mechanism involving oxygenation
at C-3 during desaturation of 2R,3R-trans-DHQ 9 by ANS. A
more direct desaturation (either concerted or non-concerted)
process appeared much less likely. In addition, a second molecule
of DHQ substrate was present in the active site of this crystal
structure (assigned as the 2S,3S-trans enantiomer), along with
a molecule of the 2-[N-morpholino]ethanesulfonic acid (MES)
buffer present in the crystallisation solution.
Fig. 6 Effect of the lanthanide chiral shift reagent Eu(hfc)
3
-(+)
on the NMR spectrum of a thermodynamic mixture of (± )-trans-
dihydroquercetin (9 and enantiomer) and (± )-cis-dihydroquercetin.
Note that if pure (+)-trans-dihydroquercetin 9 was added to the mixture,
only one of the peaks observed in the lower spectrum increased in
magnitude.
The overall arrangement of the active site residues in
the ANS–Fe(II)–2OG–naringenin structure is very similar
to that observed in the previously described structures
of ANS–Fe(II)–2OG–DHQ –MES and ANS–Fe(II)–succinate–
2
1
8
˚
quercetin (RMS of 0.42 A to 1gp5). A single molecule of 2S-
naringenin 3 is complexed in the active site in a similar manner
to DHQ-1 9 in the ANS–Fe(II)–2OG–DHQ structure. A 2S-
naringenin 3 substrate molecule was modelled into the electron
˚
density, although the lower resolution (2.3 A) of the ANS–
Fig. 7 Effect of lanthanide chiral shift reagent Eu(hfc)
NMR spectrum of cis-10 and trans-dihydrokaempferol 5/11 purified
from an incubation of (± )-naringenin 3/13.
3
-(+) on the
Fe(II)–2OG–naringenin meant that (partial) occupancy by the
2R-naringenin 13 could not be ruled out.
2
Fig. 8 Overall view of ANS–Fe(II)–2OG–DHQ –MES. Left: PDB structure 1gp5 containing 2OG (pink), two molecules of MES (purple) and two
molecules of dihydroquercetin (DHQ). DHQ-1 and DHQ-2 are 2R,3R-trans (buff) and 2S,3S-trans (grey) stereochemistries respectively. Right: the
structure described here, with 2OG (pink) and one molecule of 2S-naringenin (buff).
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6
3 1 2 1

View Article Online
Fig. 9 Stereo diagram of the ANS active site with 2mF
o
c
-DF electron density map plotted at a contour level of 1r.
Fig. 10 The active site of ANS depicted using the same colour scheme as Fig. 8. Left: the 2R,3R-trans-DHQ 9 structure. Right: the 2S-naringenin
3
structure.
There are some significant changes between the active sites
of the ANS–Fe(II)–2OG–DHQ –MES and ANS–Fe(II)–2OG–
2
naringenin–MES structures. In the 2S-naringenin 3 structure,
there was no density in the active site corresponding to a second
ꢀ
molecule of naringenin equivalent to ‘DHQ-2 in the ANS–
Fe(II)–2OG–DHQ –MES structure. The lower concentration
2
of substrate present in the co-crystallisation of ANS with
naringenin 3/13 compared to DHQ (9 and enantiomer), (final
concentrations of 2.5 mM and 10 mM respectively) may in part,
have resulted in this; however, the 2.5 mM solution of naringenin
3/13 used still represents an eight-fold excess of naringenin 3/13
relative to ANS.
In the ANS–Fe(II)–2OG–DHQ –MES structure, the 3 -
ꢀ
2
hydroxyl group of the second DHQ (enantiomer of 9) molecule is
in position to hydrogen bond with the C-5 hydroxyl group of the
A-ring of DHQ-1 9 and the main chain carbonyl of Thr233. The
Fig. 11 The effect of different concentrations of flavonoid substrate
on the rate of decarboxylation of the 2OG 1 co-substrate. Note that the
rate is non-zero in the absence of flavonoid substrate due to the substrate
uncoupled reaction of 2OG. Black circles with a (± )-trans-DHQ (9 and
enantiomer) substrate; open circles with a (± )-naringenin 3/13 substrate.
ꢀ
absence of a 3 -hydroxyl group in the naringenin 3/13 substrate
may not allow efficient binding in an analogous manner.
1
4
An assay monitoring the decarboxylation of a 1-[ C]-2OG
co-substrate was used to investigate the effect of varying the
initial concentration of different flavonoid substrates on the rate
of ANS reaction (Fig. 11). For both (± )-naringenin 3/13 and
ꢀ
ꢀ
ꢀ
substrates with 3 - or 3 ,5 -hydroxyl groups. It is possible that
binding of a second substrate (flavonoid) represents a method
for obtaining a degree of selectivity related to the hydroxylation
pattern of the B-ring (note that the B-ring hydroxylation pattern
affects catalytic efficiency, e.g. naringenin 3/13 versus eriodictyol
14 (Table 1)).
(
± )-trans-DHQ (9 and enantiomer) substrates the initial rate
increased to a maximum at approximately 800 lM. After this
the rate began to decrease. This demonstrates that substrate
inhibition occurs for both (± )-naringenin 3/13 and (± )-trans-
DHQ 9, despite the observation that two molecules of flavonoid
were observed in the crystallographic analysis of the active site
only in the latter case. The precise mechanism(s) of substrate
inhibition is uncertain. However, it seems reasonable to propose
that binding at the second substrate binding site, observed
with DHQ (9 and enantiomer), will inhibit product release and
hence reduces the efficiency of enzymatic turnover, at least for
ꢀ
The absence of a 3 -hydroxyl group in the 2S-naringenin 3
substrate, compared to DHQ 9, also causes an alteration in the
hydrogen bonding between the B-ring of 2S-naringenin 3 with
Tyr142 relative to the contacts between the B-ring of 2R,3R-
trans-DHQ 9 and Tyr142. In the naringenin structure, there
ꢀ
appears to be single hydrogen bond between Tyr142 and the 4 -
hydroxyl group of 2S-naringenin 3, whereas DHQ-1 9 uses both
ꢀ
ꢀ
its 3 - and 4 -hydroxyl to hydrogen bond to Tyr142. Further,
3
1 2 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6

View Article Online
the position of the side chain of Y142 in the 2S-naringenin 3
structure is clearly defined and is rotated by ca. 14 relative to the
DHQ 9 structure, bringing it closer to the naringenin substrate
Table 5 Relative activity of enzymes in comparison with wild types
for assays with the (± )-trans-DHQ (9 and enantiomer) substrate as
determined by HPLC monitoring of quercetin 8 production. Figures
are normalised to wild type enzymes (ANS or flavonol synthase). For
flavonol synthase, figures in parentheses are normalised to those for
ANS
◦
(
if the A- and C-rings of 2R-naringenin 13 occupied the same
position where 2S-naringenin 3 is modelled, there would be a
˚
distance of only 2 A from the B-ring hydroxyl to Ce1 of Y142).
The conformation of the side chain of Gln117 projects towards
Tyr142, but it is too distant (3.7 A) for a direct hydrogen bond to
Enzyme
Activity relative to wild type enzyme (%)
˚
its phenolic oxygen. There is also an apparent rotation of V229,
which is close to both the 2OG 1 and substrate binding sites.
ANS and flavonol synthase (FLS) mutants were designed
based on potential involvement of amino acid side chains in
acid–base catalysis during desaturation (ANS-K128A), binding
of flavonoid substrates (ANS Y142H, ANS-K341H, K341N,
FLS-H131Y and FLS-K329N) and binding of the MES
molecule (ANS-E230Q, ANS-N131D, ANS-N131A), the latter
of which had been postulated to be replacing L-ascorbic acid in
the active site. ANS residue Y142 and the flavonol synthase
residue H132 were targeted on the basis of crystallographic
analyses, as the identity of the residues at this position represents
a potential difference between the ANS and predicted flavonol
synthase active sites. It was thought that these residues may be
involved in determining substrate selectivity, so mutants in which
the residues at position-142 for ANS (equivalent to position-128
for flavonol synthase) were interchanged were made (i.e. ANS-
Y142H and FLS-H132Y).
ANS wt
100.00
67.00
70.00
48.00
71.00
82.00
66.00
100 (224)
63 (140)
28 (64)
ANS Y142H
ANS K341N
ANS E230Q
ANS K128A
ANS N131D
ANS N131A
FLS wt
FLS H132Y
FLS K329N
dependent on L-ascorbic acid. Incubation with (± )-naringenin
3/13 demonstrated that all the mutants converted it to DHK
5/10/11, apigenin 4 and kaempferol 12.
Discussion
The work presented here is consistent with, and supports
the proposal that ANS catalyses the oxidation of flavanone
substrates (e.g. naringenin 3/13) initially by removal of the
C-3 pro-S hydrogen. At least for naringenin 3/13 and related
substrates, the type of oxidation reaction catalysed, in particular
hydroxylation versus desaturation, is dependent on the C-
2 stereochemistry of the substrate. It seems likely that this
influences the stereoelectronic relationship between the other
reactive oxidising intermediate(s) and the naringenin substrate
(3 or 13).
The six ANS mutants (Y142H, E230Q, K128A, N131D,
N131A, K341N) and two flavonol synthase mutants (H132Y
and K329N) were prepared and purified (to >90% purity by
SDS-PAGE analysis) as reported for the wild type enzymes,
(
ANS-E230Q bound less tightly to the Reactive Green column
used for purification, but was obtained in a similar purity to
the other mutants). Two other mutants, ANS-K213A and FLS-
K202A, were produced in a predominantly insoluble form in
Escherichia coli BL21(DE3) cells and were not pursued. ESI-MS
analyses of the mutant proteins were all within 10Da (Table 4) of
the calculated values and circular dichroism analyses indicated
that they had similar folds to native ANS and flavonol synthase
For some 2OG dependent oxygenases, including taurine
dioxygenase and deacetoxycephalosporin C synthase, kinetic
isotope studies are consistent with a mechanism in which a
ferryl intermediate abstracts a hydrogen atom from the prime
25,26
(
data not shown).
substrate.
This can be followed by substrate hydroxylation,
The mutants were assayed by using HPLC to monitor for
desaturation or rearrangement. Desaturation can be envisaged
to occur either by initial oxygenation followed by dehydration
or via a more direct process not involving oxygenation (Fig. 12).
In accord with an analogous radical rebound mechanism
proposed for the ANS catalysed hydroxylation of naringenin
3/13 substrates, all the reactions may occur via formation of a
common C-3 flavanone radical intermediate by removal of the
C-3 pro-S hydrogen atom (Figs. 3 and 4).
For 2S-naringenin 3, a single stereoisomer of cis-DHK 10 is
produced as the major product (>70% of total detected flavanoid
products), presumably 2R,3S-cis-DHK 10 (assuming retention
of C-2 stereochemistry), formed by hydroxylation of the C-3
pro-S position (Fig. 3). This mechanism is supported by the
ANS–Fe–2OG–naringenin–MES crystal structure (see below).
In contrast with 2S-naringenin 3, in the case of 2R-
naringenin 13 there is a significant decrease in the product
(dia)stereoselectivity of hydroxylation (i.e. a ca. 1 : 1 ratio of
cis : trans DHK was observed) coupled with a large increase in
the relative amount of the direct desaturation product, apigenin
conversion of (± )-trans-DHQ (9 and enantiomer) to quercetin 8.
All of the mutants displayed significant activity under standard
assay conditions (Table 5). The K128A mutant was less active
than wild type ANS (ca. 70% with (± )-trans-DHQ substrate)
and when incubated with (± )-naringenin 3/13, the relative
amounts of apigenin 4 to dihydrokaempferol 5/10/11 produced
remained reasonably consistent with that of the wild type.
These results indicate that Lys128 is not essential for catalysis
and specifically that it does not act as a general base during
desaturation, at least under in vitro conditions. As mutation
of Y142H in ANS did not increase the activity of ANS to
that of flavonol synthase (Table 5) it can be concluded that
the presence of the corresponding residue H132 in flavonol
synthase is not solely responsible for the increased efficiency
of flavonol synthase with 2R,3R-trans-DHQ 9 substrate. As for
the wild type enzyme, optimal activity of all the mutants was
Table 4 Masses of wild type and mutant enzymes
4
(Fig. 4).
Enzyme
Predicted mass/Da
Experimental mass/Da
ANS is proposed to catalyse the formal desaturation of
R,3R-trans-DHQ 9 to give quercetin 8 via an oxygenating
2
ANS wt
40 396
40 343
40 382
40 395
40 339
40 397
40 353
38 281
38 307
38 267
40 404 ± 14
40 348 ± 11
40 378 ± 12
40 402 ± 15
40 345 ± 6
40 404 ± 8
40 358 ± 6
38 284 ± 6
38 313 ± 7
38 272 ± 5
mechanism involving C-3 hydroxylation to give an enzyme
bound geminal C-3,3-diol 15 intermediate, followed by dehy-
ANS Y142H
ANS K341N
ANS E230Q
ANS K128A
ANS N131D
ANS N131A
FLS wt
23
dration (Fig. 2A). ANS catalysed desaturation of some LCDs
(e.g. 2R,3S,4S-cis-LCD 6) may occur via an analogous process
(Figs. 2B and 12), though for these substrates a mechanism
proceeding via C-4 hydroxylation cannot be ruled out. However,
it is unlikely that the apigenin 4 observed in the assay of 2R-
naringenin 13 arises by an oxygenating hydroxylation followed
by dehydration mechanism. This is because ANS is unable to
FLS H132Y
FLS K329N
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6
3 1 2 3

View Article Online
Fig. 12 Outline of possible non-oxygenating versus oxygenating mechanisms for desaturation. In the case of ANS, the oxygenation mechanism is
ꢀꢀ
proposed to occur when R = OH.
dehydrate dihydroflavonols (i.e. DHK 5 or DHQ 9) to flavones
fate for an intermediate involved in formation of the C-3 pro-S
hydroxylated product.
(
i.e. apigenin 4 and luteolin respectively), even when the former
are formed in the active site, e.g. cis-DHQ is formed in the ANS
active site in the reaction of 2R,3S,4R-trans-LCD substrates.
The position of 2S-naringenin 3 in the active site relative to the
iron in the ANS–Fe–2OG–naringenin–MES crystal structure
supports the conclusions of the solution studies. The C-3 pro-S
hydrogen is the most accessible to the iron centre and appears
the most likely position for oxidation of both enantiomers of
naringenin 3/13 by ANS. The structure of 2S-naringenin 3
bound to the active site shows that the C-2 hydrogen atom of 2S-
flavanones (e.g. 3) is likely to be inaccessible to oxidation. This
is consistent with the low levels of apigenin 4 formed with this
substrate and the proposal that apigenin 4 formation in vitro re-
quires a syn-facial relationship between the C-2 hydrogen and the
C-3 pro-S hydrogen, which is only possible with 2R-flavanones
(e.g. 13). Modelling 2R-naringenin 13 into the active site of ANS
implied that the C-2 hydrogen is accessible for oxidation for this
substrate (data not shown). However, the greater distance of the
assumed position of the ferryl oxygen to the C-2 position of 2R-
17
Thus, ANS catalysed C-2, C-3 desaturation of 2R-naringenin 13
to give apigenin 4 is likely to occur by removal of two hydrogen
atoms by a reactive oxidising species, possibly without the direct
aid of a catalytic base from an amino acid residue (Fig. 4). The
mutagenesis studies are consistent with this proposal, as they
rule out the critical involvement of residues including K128 of
ANS, that could act as a general base.
For the ‘direct’ (compared to the hydroxylation–elimination
reactions proposed for dihydroflavonols) desaturation reactions
catalysed by non-haem Fe(II), 2OG dependent oxygenases,
it has been suggested that following abstraction of the first
III
hydrogen atom, the iron–oxygen derived species (probably Fe –
OH) removes a second hydrogen atom from the substrate to
27
generate awater molecule. This process mayeither be concerted
with desaturation, or proceeds via a radical species that collapses
to form the product, after rearrangement in the case of some
enzymes, e.g. deacetoxycephalosporin C synthase.
˚
naringenin 13 compared to that of the C-3 position (ca. 6 A and
˚
5 A, respectively) suggests that a concerted syn-facial C-2, C-3
mechanism is unlikely. As discussed above it seems more likely
that a C-3 radical is initially formed by removal of the C-3 pro-S
hydrogen of 2R-naringenin 13 prior to subsequent removal of
its C-2 hydrogen and formation of apigenin 4 (Fig. 4).
The proposed C-3 flavanone radical intermediate formed by
removal of the C-3 pro-S hydrogen atom of 2R-naringenin 13 is
a plausible intermediate in the formation of both 2S,3S-trans-
DHK 11, by a-face hydroxylation and apigenin 4, by removal of
the syn-facial C-2 hydrogen (Fig. 4). In comparison, a syn-facial
desaturation mechanism involving the C-3 pro-S hydrogen of
In the ANS–Fe–2OG–DHQ crystal structure, two molecules
2
of DHQ (9 and enantiomer) were observed to be bound to
the active site; exposure of the crystals to dioxygen resulted
in conversion of the DHQ 9 molecule closest to the iron
centre to quercetin 8, demonstrating that this complex is still
catalytically active. However, it was not clear how the quercetin
8 product could exit the active site, without movement of
the second molecule of DHQ (enantiomer of 9). Although
other explanations are possible, the observation of inhibition of
ANS activity by high concentrations of flavonoid substrates is
consistent with the catalysis being hindered by the formation
of non-productive ANS–flavonoid complexes in which two
molecules of flavonoid are bound at the active site.
2S-naringenin 3 cannot occur, consistent with the observation
that only very low amounts of apigenin 4 are produced in
incubations of ANS with 2S-naringenin 3. These observations,
together with prior work, demonstrate that C-3 oxidation on
the a-face (equivalent to the pro-S C-3 hydrogen of naringenin
3
/13) is the predominant mechanism in ANS catalysis, but that
in the case of the 2R-naringenin 13 substrate, b-face oxidation
(
hydroxylation) can also occur (Fig. 4).
Interestingly, the product stereoselectivity of hydroxylation at
the C-3 position is reduced for incubations of 2R-naringenin
3 with ANS relative to that with 2S-naringenin 3, possibly
1
In vitro assay results along with sequence similarities imply
that flavone synthase I and flavanone 3b-hydroxylase (Fig. 1A)
are predominantly b-face oxygenases (i.e. preferentially catalyse
due to an alteration in the spatial arrangement between the
ferryl (or equivalent) intermediate (Fig. 13) or an alternative
Fig. 13 Scheme outlining the possible position of the ferryl intermediate relative to the C-2 and C-3 hydrogen atoms of both 2S-3 and 2R-naringenin
1
3 substrates. Note: the relative accessibility of the C-3 hydrogen in the latter case.
3
1 2 4
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6

View Article Online
oxidation at the position equivalent to the C-3 pro-R hydrogen
of naringenin 3/13), while ANS and flavonol synthase are
predominantly a-face oxygenases (the C-3 pro-S hydrogen atom
of naringenin 3/13). In support of this proposal, the results
presented here indicate that ANS can catalyse a syn-facial a-face
desaturation. In comparison, the desaturation of 2S-naringenin
Reverse 5ꢀ-GTGAGCTAGCACACCGAGTGCTAGCTC-
AG-3ꢀ
ANS K128A
23
ꢀ
Forward 5 -CAAGGCTATGGAAGTGCATTGGCTAACA-
ꢀ
ACGCGAG-3
ꢀ
Reverse 5 -CCACTCGCGTTGTTAGCCAATGCACTTCC-
3
by flavone synthase I (Fig. 1A) is thought to occur by a syn-
ATCGCC-3ꢀ
4,15
facial b-face mechanism.
The results also reveal that ANS
ANS N131D:
ꢀ
can catalyse b-face oxidation, albeit as a minor pathway with
the unnatural R-naringenin isomer-13, which gives 2S,3R-cis-
DHK 10 as one of its products (Fig. 4).
Forward 5 -CAAGGCTATGGAAGTAAATTGGCTGACA-
ꢀ
ACGCGAG-3
ꢀ
Reverse 5 -CCACTCGCGTTGTCAGCCAATTTACTTGC-
There is current interest in the generation of synthetic catalysts
ATAGCC-3ꢀ
10
based on the active sites of non-haem oxygenases. Like
the non-haem Fe(II), 2OG dependent oxygenases clavaminic
acid synthase and deacetoxy/deacetylcephalosporin C synthase,
ANS is able to catalyse more than one type of oxidative reaction,
indicating the versatility of these catalysts. Studies on the 2OG
oxygenase clavaminic acid synthase have shown that the balance
between hydroxylation and desaturation can be affected by the
ANS N131A:
ꢀ
Forward 5 -CAAGGCTATGGAAGTAAATTGGCTGCCA-
ꢀ
ACGCGAG-3
ꢀ
Reverse 5 -CCACTCGCGTTGGCAGCCAATTTACTTCC-
ATAGCC-3ꢀ
ANS K341N:
ꢀ
Forward 5 -CAACATATTGAGCATGCGTTGTTTGGGA-
21
ꢀ
nature of the substrate side chain. In the case of ANS, the
balance between hydroxylation and desaturation appears to
be more subtle, being dependent on the stereochemistry at a
single position. By choice of substrate stereochemistry it may
be possible to use the same catalyst for both hydroxylation
and desaturation reactions. The results demonstrate that ANS
can catalyse formal desaturation by more than one type of
mechanism: oxygenating or non-oxygenating. Thus, although
the oxygenating mechanism is probably unusual, and proposed
to proceed via a gem-diol intermediate then elimination of water,
it should not be assumed that there is a unique desaturation
mechanism for iron-dependent oxidising enzymes.
AGGAACAAGAAG-3
ꢀ
Reverse 5 -CTTGTTCCTTCCCAAACAACGCATGCTCA-
ATATGTTGAGC-3ꢀ
ANS K341A:
ꢀ
Forward 5 -CAACATATTGAGCATGCGTTGTTTGGGA-
ꢀ
AGGAACAAGAAG-3
ꢀ
Reverse 5 -CTTGTTCCTTCCCAAACAACGCATGCTCA-
ATATGTTGAGC-3ꢀ
FLS K329N:
ꢀ
Forward 5 -GGATTACAGTTACCGCAACCTCAATAAA-
ꢀ
CTTCCTCTGG-3
ꢀ
Reverse 5 -GAGGAAGTTTATTGAGGTTGCGGTAACT-
GTAATCCTTG-3ꢀ
FLS K329A:
ꢀ
Forward 5 -GGATTACAGTTACCGCGCGCTCAATAAA-
Materials and methods
ꢀ
CTTCCTCTGG-3
ꢀ
ANS and flavonol synthase (A. thaliana) were over-expressed
and purified according to published procedures.
protocols, equipment and reagent sources were as reported.
Reverse 5 -GAGGAAGTTTATTGAGCGCGCGGTAACT-
23,28
GTAATCCTTG-3ꢀ
General
23
The Stratagene QuikChange Site-Directed mutagenesis kit was
used to create mutant pET-24a(+)-ans and pET-3a(+)-fls genes.
The mutants were purified using the same procedure as for wild
type ANS and flavonol synthase, except for ANS-E230Q which
bound less tightly to the Reactive Green column. The mutations
were confirmed by automated DNA sequencing carried out
by the DNA Sequencing Service, Department of Biochemistry,
University of Oxford.
Assay procedures
For the assays investigating the effect of different concentrations
of (± )-naringenin 3/13 and (± )-trans-DHQ (9 and enantiomer)
on the initial rate of 2OG 1 decarboxylation, the method used
1
4
was that of monitoring the decarboxylation of 1-[ C]-2OG as
described previously, except that assays were quenched after
23
1 min.
HPLC assays were carried out and analysed as described
The primers used for mutagenesis were as follows:
ANS Y142H:
23
previously, except for those to separate cis-DHK 10 and trans-
DHK 5/11 where the conditions were as follows. Chiral HPLC
analysis—after SPE separation of the assay components, 200 lL
ꢀ
Forward 5 -GTGGACAATTGGAATGGGAAGATCACTT-
ꢀ
CTTTCATCTTGCGTATC-3
ꢀ
−1
Reverse 5 -CAGGATACGCAAGATGAAAGAAGTGATC-
of the 50% MeOH fraction were loaded at 0.8 mL min onto
ꢀ
TTCCCATTCCAATTGTC-3
a ChiraSpherꢀR NT 250 × 4.6 column pre-equilibrated with
FLS H132Y:
40% buffer A and 60% buffer B. The column was then run
ꢀ
−1
Forward 5 -GCTTGGGTCGATTATCTCTTCCATCGAA-
isocratically at 0.8 mL min under similar solvent conditions.
TCTGG-3ꢀ
The percentage yields were calculated by integration of
peaks on HPLC traces. For assays of (± )-trans-DHQ (9 and
enantiomer) and (± )-naringenin 3/13 substrates, the elutant
was monitored between 200 and 400 nm. For assays of LCD
(e.g. 6) substrates, the elutant was monitored between 200 and
600 nm. The HPLC system was calibrated for different flavonoid
standards by 15 injections of known amounts of a compound
giving an absorbance of <1.5 at the wavelength of interest. Using
the equation area = bx, where x is the number of nmoles of
compound and b is a conversion factor, the values calculated
for b were cyanidin 7 (198 70828 at 520 nm), (± )-trans-DHQ (9
and enantiomer) (180 32188 at 290 nm), quercetin 8 (172 92176
at 370 nm) and (± )-naringenin 3/13 (173 60805 at 290 nm). A
standard of DHK 5/10/11 was not available in sufficient quan-
tities for a calibration, so the same conversion factor was used
as for naringenin 3/13 (which has a very similar chromophore).
ꢀ
Reverse 5 -GTGGCCAGATTCGATGGAAGAGATAATC-
GACCC-3ꢀ
ANS K213A:
ꢀ
Forward 5 -GAGCTTCTTCTACAAATGGCGATAAATT-
ꢀ
ACTATCCAAAATGTCCTCAGC-3
ꢀ
Reverse 5 -GAGGACATTTTGGATAGTAATTTATCGCC-
ꢀ
ATTTGTAGAAGAAGCTCTTC-3
FLS K202A:
ꢀ
Forward 5 -GGCGGAGTATATGATGGCGATTAACTAT-
ꢀ
TATCCGCCGTGTCC-3
ꢀ
Reverse 5 -CACGGCGGATAATAGTTAATCGCCATCAT-
ꢀ
ATACTCCGCCATC-3
ANS E230Q:
Forward 5ꢀ-GCTAGCACTCGGTGTGCAAGCTCACAC-
CG-3ꢀ
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6
3 1 2 5

View Article Online
Table 6 Data collection and refinement statistics for the ANS–Fe(II)–
concentration of 10% (v/v) MeOH), pH 6.5. The crystals
were immersed in mother liquor supplemented with 10% (v/v)
2
OG–naringenin structure (PDB id code: 2brt)
ethylene glycol and flash cooled in liquid N . Diffraction data
were collected at 100 K using beamline 14.2 (SRS, Daresbury
Laboratory, UK) with an ADSC Quantum4 CCD detector
2
Space group
1 1 1
P2 2 2
2.2 (2.31–2.2)
56.8
62.3
102.5
109 275
18 607
0.082 (0.606)
98.0 (99.5)
15.8 (3.1)
0.211
Resolution/A˚
Unit cell/A˚
a
b
c
(
Table 6). The data were processed using MOSFLM and SCALA
29
(CCP4 suite ).
Total reflections
Unique reflections
Acknowledgements
a
R
merge (%)
Completeness (%)
We thank the Biotechnology and Biological Sciences Research
Council (Biomolecular Sciences Committee) and the Engineer-
ing and Physical Sciences Research Council for funding. We
thank Dr J. L. Firmin (John Innes Centre, Norwich, UK) for
providing the purified 2S- and 2R-naringin.
I/r
I
R
R
free
0.268
Mean B-values/ A˚ 2
Main chain
Side chain
Substrates
Iron
47.3
48.0
52.3
31.7
47.8
References
Solvent
1
2
3
4
5
6
7
8
9
B. A. Bohm, Introduction to Flavonoids, Harwood Academic Pub-
lishers, Amsterdam, 1998.
G. Forkmann and W. Heller, Comprehensive Natural Products
Chemistry, Elsevier Science, Amsterdam, 1999, pp. 713–748.
K. Springob, J. Nakajima, M. Yamazaki and K. Saito, Nat. Prod.
Rep., 2003, 20, 288–303.
RMS deviations from ideal
Bonds/A˚
0.025
1.89
128
◦
Angles/
No. solvent molecules
a
R
merge = R
j
R
h
|Ih,j − ꢁI
h
ꢂ|/R
j
R
h
ꢁI ꢂ × 100 (outer resolution shell statistics
h
S. Martens, G. Forkmann, L. Britsch, F. Wellmann, U. Matern and
R. Lukacin, FEBS Lett., 2003, 544, 93–8.
shown in parentheses).
R. Lukacin, C. Urbanke, I. Groning and U. Matern, FEBS Lett.,
2
000, 467, 353–8.
Substrate synthesis
A. G. Prescott, N. P. Stamford, G. Wheeler and J. L. Firmin,
Phytochemistry, 2002, 60, 589–93.
2
R-13 and 2S-naringenin 3 were produced by naringinase
R. Lukacin, F. Wellmann, L. Britsch, S. Martens and U. Matern,
Phytochemistry, 2003, 62, 287–92.
K. Saito, M. Kobayashi, Z. Gong, Y. Tanaka and M. Yamazaki,
Plant J., 1999, 17, 181–9.
J. J. Turnbull, W. J. Sobey, R. T. Aplin, A. Hassan, J. L. Firmin, C. J.
Schofield and A. G. Prescott, Chem. Commun., 2000, 2473–4.
mediated cleavage of the glycosidic bond of 2R- and 2S-naringin
23
as described.
1
H NMR scale naringenin incubations
Assays were carried out as 18 × 0.5 ml aliquots for 45 minutes,
1
1
0 M. Costas, M. P. Mehn, M. P. Jensen and L. Que, Jr., Chem. Rev.,
004, 104, 939–86.
then combined and loaded onto an SPE column (Strata C-18E,
2
5
00 mg). The cofactors were eluted using 8 ml 5% MeOH (v/v);
DHK 5/10/11 with 4 ml 40% MeOH (v/v); and naringenin 3/13
apigenin 4 with 60% MeOH (v/v). The H NMR (500 MHz)
1 R. P. Hausinger, Crit. Rev. Biochem. Mol. Biol., 2004, 39, 21–68.
12 J. C. Price, E. W. Barr, B. Tirupati, J. M. Bollinger, Jr. and C. Krebs,
1
Biochemistry, 2003, 42, 7497–508.
3 D. A. Proshlyakov, T. F. Henshaw, G. R. Monterosso, M. J. Ryle and
R. P. Hausinger, J. Am. Chem. Soc., 2004, 126, 1022–3.
1
1
spectra were assigned by comparison to standards (including
doping experiments) and couplings confirmed by COSY spectra.
4 P. J. Riggs-Gelasco, J. C. Price, R. B. Guyer, J. H. Brehm, E. W.
Barr, J. M. Bollinger, Jr. and C. Krebs, J. Am. Chem. Soc., 2004, 126,
1
1
For chiral H NMR analyses, after the initial H spectra had been
recorded, Eu(hfc)
3
-(+) was added portion wise (2 × 85 nmol,
8
108–9.
1
1
× 170 nmol) from a 40 mM solution with a H NMR spectrum
15 L. Britsch, Arch. Biochem. Biophys., 1990, 282, 152–60.
16 R. W. D. Welford, J. J. Turnbull, T. D. W. Claridge, A. G. Prescott
and C. J. Schofield, Chem. Commun., 2001, 1828–9.
recorded after each addition. Selected data as follows:
Naringenin 3/13; d (500 MHz; CD CN): 2.8 (dd, 1H, J 3.0
H
3
1
1
1
7 J. J. Turnbull, M. J. Nagle, J. F. Seibel, R. W. D. Welford, G. H. Grant
and C. J. Schofield, Bioorg. Med. Chem. Lett., 2003, 13, 3853–7.
8 R. C. Wilmouth, J. J. Turnbull, R. W. D. Welford, I. J. Clifton, A. G.
Prescott and C. J. Schofield, Structure, 2002, 10, 93–103.
9 E. N. Marsh, M. D. Chang and C. A. Townsend, Biochemistry, 1992,
31, 12648–57.
1
1
(
[
7.0, H-3), 3.1 (dd, 1H, J 13.0 17.0, H-3), 5.3 (dd, 1H, J 3.0
3.0, H-2), 5.8 (2d, 2H, J 2.1, H-6 H-8), 6.75 (d, 2H, J 8.5), 7.3
d, 2H, J 8.5); kmax: 291.4 nm; m/z (ESI-MS negative ion mode):
−
M − H] 271.0.
Apigenin 4; d
H
(500 MHz; CD
3
CN): 6.3 (d, 1H, J 2.0), 6.55
(
d, 1H, J 2.0), 6.65 (s, 1H, H-3), 7.0 (d, 2H, J 9.0), 8.0 (d, 2H,
20 Z. H. Zhang, J. S. Ren, D. K. Stammers, J. E. Baldwin, K. Harlos
and C. J. Schofield, Nat. Struct. Biol., 2000, 7, 127–33.
21 M. D. Lloyd, K. D. Merritt, V. Lee, T. J. Sewell, B. Wha-Son, J. E.
Baldwin, C. J. Schofield, S. W. Elson, K. H. Baggaley and N. H.
Nicholson, Tetrahedron, 1999, 55, 10201–20.
J 9.0); kmax: 337.9 nm; m/z (ESI-MS negative ion mode): [M −
−
H] 269.0.
trans-DHK 5/11; d
H-3), 5.0 (d, 1H, J 11.5 H-2); kmax: 292.4 nm; m/z (ESI-MS
H
(500 MHz; CD CN): 4.6 (d, 1H, J 11.5,
3
2
2 J. E. Baldwin and M. Bradley, Chem. Rev., 1990, 90, 1079–88.
−
negative ion mode): [M − H] 287.0.
23 J. J. Turnbull, J. Nakajima, R. W. D. Welford, M. Yamazaki, K. Saito
cis-DHK 10; d
H
(500 MHz): 4.3 (d, 1H, J 3.0, H-3), 5.45 (d,
and C. J. Schofield, J. Biol. Chem., 2004, 279, 1206–16.
2
2
4 E. Kiehlmann and E. P. M. Li, J. Nat. Prod., 1995, 58, 450–5.
5 J. C. Price, E. W. Barr, T. E. Glass, C. Krebs and J. M. Bollinger, Jr.,
J. Am. Chem. Soc., 2003, 125, 13008–9.
1
7
1
H, J 3.0, H-2), 6.0 (2d, 2H, J 2.2, H-6 8), 6.85 (d, 2H, J 8.5),
(500 MHz):
.35 (d, 2H, J 8.5); + 380 nmol Eu(hfc)
3
-(+); d
H
1.75 (s, 1H); kmax: 292.4 nm; m/z (ESI-MS negative ion mode):
2
6 J. E. Baldwin, R. M. Adlington, N. P. Crouch, J. W. Keeping, S. W.
Leppard, J. Pitlik, C. J. Schofield, W. J. Sobey and M. E. Wood,
J. Chem. Soc., Chem. Commun., 1991, 768–70.
−
[
M − H] 287.0.
Eriodictyol 14 assay; trans-DHQ; m/z (ESI-MS negative ion
−
mode): [M − H] 303.0; kmax: 290.2 nm. cis-DHQ; m/z (ESI-MS
27 M. D. Lloyd, H. J. Lee, K. Harlos, Z. H. Zhang, J. E. Baldwin,
C. J. Schofield, J. M. Charnock, C. D. Garner, T. Hara, A. C. T. van
Scheltinga, K. Valegard, J. A. C. Viklund, J. Hajdu, I. Andersson, A.
Danielsson and R. Bhikhabhai, J. Mol. Biol., 1999, 287, 943–60.
−
negative ion mode): [M − H] 303.0; kmax: 290.2 nm.
Crystals of (± )-naringenin 3/13 complexed with ANS–Fe(II)–
2
OG were grown from a solution of 18% (w/v) PEG 2000
2
8 J. J. Turnbull, A. G. Prescott, C. J. Schofield and R. C. Wilmouth,
Acta Crystallogr., Sect. D: Biol. Crystallogr., 2001, 57, 425–7.
monomethylether, 50 mM MES, 200 mM ammonium acetate,
mM FeSO , 10 mM potassium 2OG, 10 mM sodium ascorbate
2
4
29 Collaborative Computational Project Number 4, Acta Crystallogr.,
Sect. D: Biol. Crystallogr., 1994, 50, 760–763.
and 2.5 mM (± )-naringenin 3/13 (in MeOH to give a final
3
1 2 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 1 7 – 3 1 2 6