Elucidation of active site residues of Arabidopsis thaliana flavonol synthase provides a molecular platform for engineering flavonols

Article abstract of DOI:10.1016/j.phytochem.2007.07.006

Arabidopsis thaliana flavonol synthase (aFLS) catalyzes the production of quercetin, which is known to possess multiple medicinal properties. aFLS is classified as a 2-oxoglutarate dependent dioxygenase as it requires ferrous iron and 2-oxoglutarate for catalysis. In this study, the putative residues for binding ferrous iron (H221, D223 and H277), 2-oxoglutarate (R287 and S289) and dihydroquercetin (H132, F134, K202, F293 and E295) were identified via computational analyses. To verify the proposed roles of the identified residues, 15 aFLS mutants were constructed and their activities were examined via a spectroscopic assay designed in this study. Mutations at H221, D223, H277 and R287 completely abolished enzymes activities, supporting their importance in binding ferrous iron and 2-oxoglutarate. However, mutations at the proposed substrate binding residues affected the enzyme catalysis differently such that the activities of K202 and F293 mutants drastically decreased to approximately 10% of the wild-type whereas the H132F mutant exhibited approximately 20% higher activity than the wild-type. Kinetic analyses established an improved substrate binding affinity in H132F mutant (K_m: 0.027 ± 0.0028 mM) compared to wild-type (K_m: 0.059 ± 0.0063 mM). These observations support the notion that aFLS can be selectively mutated to improve the catalytic activity of the enzyme for quercetin production.

Full text of DOI:10.1016/j.phytochem.2007.07.006

Available online at www.sciencedirect.com
PHYTOCHEMISTRY
Phytochemistry 69 (2008) 66–75
www.elsevier.com/locate/phytochem
Elucidation of active site residues of Arabidopsis thaliana
flavonol synthase provides a molecular platform
for engineering flavonols
a
a
a,
*
Chun Song Chua , Daniela Biermann a,b, Kian Sim Goo , Tiow-Suan Sim
a
Department of Microbiology, Yong Loo Lin School of Medicine, National University of Singapore, 5 Science Drive 2, Singapore 117597, Singapore
b
Department of Pharmaceutical Biology, Philipps-Universita¨t Marburg, Deutschhausstraße 17A, 35037 Marburg, Germany
Received 5 January 2007; received in revised form 4 June 2007; accepted 3 July 2007
Available online 24 August 2007
Abstract
Arabidopsis thaliana flavonol synthase (aFLS) catalyzes the production of quercetin, which is known to possess multiple medicinal
properties. aFLS is classified as a 2-oxoglutarate dependent dioxygenase as it requires ferrous iron and 2-oxoglutarate for catalysis.
In this study, the putative residues for binding ferrous iron (H221, D223 and H277), 2-oxoglutarate (R287 and S289) and dihydroqu-
ercetin (H132, F134, K202, F293 and E295) were identified via computational analyses. To verify the proposed roles of the identified
residues, 15 aFLS mutants were constructed and their activities were examined via a spectroscopic assay designed in this study. Muta-
tions at H221, D223, H277 and R287 completely abolished enzymes activities, supporting their importance in binding ferrous iron and 2-
oxoglutarate. However, mutations at the proposed substrate binding residues affected the enzyme catalysis differently such that the activ-
ities of K202 and F293 mutants drastically decreased to approximately 10% of the wild-type whereas the H132F mutant exhibited
approximately 20% higher activity than the wild-type. Kinetic analyses established an improved substrate binding affinity in H132F
mutant (Km: 0.027 0.0028 mM) compared to wild-type (Km: 0.059 0.0063 mM). These observations support the notion that aFLS
can be selectively mutated to improve the catalytic activity of the enzyme for quercetin production.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: Arabidopsis thaliana; Cruciferae; Flavonoids; Quercetin; 2-Oxoglutarate dependent dioxygenase; Flavonol synthase; Site-directed mutagenesis
1. Introduction
and Lamb, 2005; Havsteen, 1983, 2002; Ormrod et al.,
1995; Pietta, 2000).
Flavonoids constitute a large group of polyphenolic sec-
ondary metabolites in plants. Approximately 8000 different
flavonoid compounds have been reported to date and they
have been ascribed as important for growth and normal
physiology of plants (Pietta, 2000). Besides providing col-
oration to plants, flavonoids are important for protection
against UV-B radiation and pathogenic microorganisms.
The antimicrobial activities of flavonoids in plants have
raised interest and further investigations have shed light
on their medicinal properties (Cushnie et al., 2003; Cushnie
The biosynthesis pathway to flavonoids has been well
established and flavonol synthase (FLS) is one of the
enzymes which produces quercetin 6 (Fig. 1; Harborne
and Williams, 2000). This compound possesses a wide
range of therapeutic properties, e.g. cardiovascular protec-
tion, anti-cancer, anti-inflammatory and anti-oxidant
activities (Narayana et al., 2001). The medicinal values of
quercetin 6 warrant the molecular and biochemical charac-
terization of the enzyme responsible for its production. As
a 2-oxoglutarate dependent dioxygenase (2-ODD), FLS
requires 2-oxoglutarate (2-OG) and ferrous iron for
catalytic function. So far, crystallographic studies of 2-
ODDs have established a characteristic topology of a
*
Corresponding author. Tel.: +65 65163280; fax: +65 67766872.
E-mail address: micsimts@nus.edu.sg (T.-S. Sim).
0031-9422/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2007.07.006

C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
67
double-stranded beta helix (DSBH) in this protein family
(Clifton et al., 2006). The DSBH consists of two sets of
four beta sheets that are arranged in an anti-parallel fash-
ion, forming a sandwich-like structure. This rigid structure
provides structural support for housing two conserved
motifs, Hx(D/E)xnH and RxS, which are essential for bind-
ing ferrous iron and 2-OG, respectively. Notably, the
HxDxnH is the more commonly found motif in the 2-
ODDs described to date (Clifton et al., 2006). This motif
is known to be crucial for anchoring the ferrous iron
required as a platform for dioxygen activation and subse-
quent catalysis (Koehntop et al., 2005; Lange and Que,
1998). While the use of a conserved mechanism involves
generation of a reactive ferryl species for catalysis, studies
have indicated that the 2-ODDs can carry out a variety
of oxidative reactions such as hydroxylation, ring cycliza-
tion and fragmentation, epimerization and desaturation
(Clifton et al., 2006). Some of these reactions are observed
in four other 2-ODDs identified in the flavonoid biosynthe-
sis pathway; flavone synthase type I (FSI), flavanone 3-
hydroxylase (F3H), anthocyanidin synthase (ANS) and fla-
vonol 6-hydroxylase (Anzellotti and Ibrahim, 2000, 2004).
FLS was first identified from a parsley cell culture by
Britsch et al. (1981) and subsequently genes encoding this
protein were cloned from other species such as Petunia hyb-
rida (Forkmann et al., 1986), Arabidopsis thaliana (Pelletier
et al., 1997) and Citrus unshiu (Moriguchi et al., 2002).
Wellmann et al. (2002) reported the role of C. unshiu in cat-
alyzing the desaturation of 2R,3R-trans-dihydroflavonols
3, 4 to flavonols 5, 6 (Fig. 1). In the same study, the con-
served ferrous iron and 2-OG binding residues were
exposed by alignment with nucleotide sequences of other
2-ODDs. The two conserved residues, G68 and G261, were
also identified via sequence alignment and their structural
roles in the enzyme function were verified via site-directed
mutagenesis (SDM). In another study, the bifunctionality
of C. unshiu FLS was reported as it catalyzed both the con-
version of 2R,3R-trans-dihydroflavonols 3, 4 to flavonols 5,
6, as well as the hydroxylation of flavanones (Lukacin
et al., 2003). Interestingly, when FLS was incubated with
2S-naringenin 1, a natural substrate of F3H, high yields
of kaempferol 5 with traces of the intermediate 2R,3R-
trans-dihydrokaempferol 3 were detected. These observa-
tions unveiled the possibility of exploiting the promiscuity
of enzymes of the flavonoid biosynthesis pathway in pro-
ducing novel compounds.
Hydroxylation and desaturation activities described in
FLS were similarly reported for ANS, which shares
R
OH
C
B
HO
O
B
C
A
A
OH
O
Flavanones
R = H - 2S-Naringenin 1
R = OH - 2S-Eriodictyol
2
R
R
F3H
OH
OH
HO
O
O
HO
O
O
FLS
OH
OH
OH
OH
Dihydroflavonols
R = H - 2R,3R-trans-Dihydrokaempferol
R = OH - 2R,3R-trans-Dihydroquercetin
Flavonols
R = H
R = OH - Quercetin
- Kaempferol 5
3
4
6
R
R
DFR
OH
OH
HO
O
HO
O
ANS
OH
OH
OH OH
OH
Leucoanthocyanidins
Anthocyanidins
R = H - Pelargonidin
R = OH - Cyanidin
R = H
- 2R,3S,4S-cis-Leucopelargonidin
7
8
9
10
R = OH - 2R,3S,4S-cis-Leucocyanidin
Fig. 1. Part of the flavonoid biosynthesis pathway. FLS catalyzes desaturation of dihydroflavonols to flavonols. F3H, flavanone 3-hydroxylase; FLS,
flavonol synthase; DFR, dihydroflavonol reductase; ANS, anthocyanidin synthase.

68
C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
approximately 50–60% sequence similarity to FLS at the
polypeptide level. ANS was proposed to catalyze the
in vivo oxidation of 2R,3S,4S-cis-leucoanthocyanidins 7, 8
to anthocyanidins 9, 10. Surprisingly, in vitro incubation
of ANS with 2R,3S,4S-cis-leucocyanidin 8 or 2R,3S,4R-
trans-leucocyanidin resulted in quercetin 6 being the major
product formed instead of cyanidin 10 (Turnbull et al.,
2003). Desaturation of 2R,3R-trans-dihydroquercetin 4
(DHQ) to quercetin 6 in ANS was also reported (Turnbull
et al., 2000), and this has supported the crystallographic
study performed on A. thaliana ANS (aANS) where the
enzyme was crystallized with 2R,3R-trans-DHQ 4 as sub-
strate (Wilmouth et al., 2002). Apart from the FLS-like cat-
alytic properties, ANS was also found to be able to react
with naringenin 1 (Turnbull et al., 2004). The high amino
acid homology between these two enzymes, ANS and
FLS, suggests a likelihood of similar three dimensional
structures to support the similar catalytic reactions
observed. Hence, the availability of the crystal structure
of aANS with known active site residues would be useful
for the characterization of other 2-ODDs in the flavonoid
pathway.
E295. The latter four residues are also conserved in FLS
of other species (Fig. 2), suggesting their possible structural
or functional roles in the enzyme.
2.2. Choice of mutations
The functional significance of the identified residues was
investigated via SDM and a total of 15 mutants were con-
structed. In most 2-ODDs, the ferrous iron and 2-OG are
bound by the HxDxnH and RxS motifs, with a few excep-
tions known, i.e. (1) a HxExnH motif in Streptomyces cla-
vuligerus clavaminate synthase, (2) a RxT motif in A.
thaliana gibberellin C20 oxidase and (3) a KxnT motif for
binding 2-OG in human factor-inhibiting hypoxia-induc-
ible factor (Elkins et al., 2003; Lukacin et al., 2000; Zhang
et al., 2002). To ascertain the relevance of the putative fer-
rous iron and 2-OG binding residues in aFLS, they were
replaced with alternative amino acids found at the corre-
sponding positions of other 2-ODDs. The strictly con-
served histidine residues were substituted with amino
acids with ring structures. Hence, the mutants constructed
were H221W, D223E, H277F, R287K and S289T.
In this paper, the active sites of A. thaliana FLS (aFLS)
were first evaluated based on the crystallographic structure
of aANS. Computational analyses facilitated the identifica-
tion of plausible cosubstrate, cofactor and substrate bind-
ing residues in aFLS. The relevance of these residues as
binding ligands was then examined by SDM and kinetic
analyses of the mutants generated.
Similarly, mutations were carried out on the proposed
substrate binding residues. Each residue was mutated to
amino acids with similar and dissimilar properties to the
original residue. The dissimilar mutations constructed
were, F134A, K202M, F293A, E295L and the similar
mutations were F134L, K202R, F293L and E295Q. The
residue H132 was mutated to a tyrosine (similar to aANS
residue Y142) and phenylalanine to examine the role of
any potential hydrogen bond formed between this residue
and the substrate, DHQ 4.
2. Results and discussion
In addition, the secondary structures of the aFLS
mutants were predicted using PSIPRED (Jones, 1999)
and the results were similar to that of the wild-type (WT)
(data not shown). This suggested minimal aberration in
the global structure of the mutant enzymes.
2.1. Computational analyses
Most enzymes in the 2-ODD protein family are known
to possess the HxDxnH and RxS motifs for binding ferrous
iron and the 2-OG, respectively. Protein sequence align-
ment of FLS from various species with that of aANS indi-
cated the conserved residues in aFLS were H221, D223,
H277, R287 and S289 (Fig. 2). To expose the less conserved
substrate binding residues, structural modeling of aFLS
was carried out. The modeled aFLS was found to be struc-
turally similar to the crystal structure of aANS (Protein
Data Bank accession number: 1GP5) (Wilmouth et al.,
2002). Since the reported structure of aANS (1GP5) was
complexed with DHQ 4, which is the natural substrate of
aFLS, the resolved crystal structure of aANS provided a
useful platform for predicting the substrate binding resi-
dues in aFLS.
Using the Swiss-PdbViewer (SPdbV) program (Guex
et al., 1999) for superimposition, the two structures were
structurally aligned with a RMS value of 0.32 A. Based
on the DHQ 4 binding ligands in aANS, the residues
Y142, F144, K213, F304 and E306 (Wilmouth et al.,
2002) were selected and used to map out the corresponding
residues in aFLS which are H132, F134, K202, F293 and
2.3. Heterologous expression of soluble aFLS WT and
mutant fusion proteins
The WT and mutant aFLS enzymes were expressed in
Escherichia coli BL21 (DE3) as GST–aFLS fusion proteins.
GST tag was incorporated into the enzymes for rapid puri-
fication of the fusion proteins via affinity binding to gluta-
thione sepharose beads. Sodium dodecyl sulfate
polyacrylamide gel electrophoresis (SDS–PAGE) analyses
showed that the fusion proteins were solubly expressed
and their levels of expression did not differ significantly
from the WT. The single band which is approximately
64 kDa corresponded to the combined molecular weight
of the calculated aFLS (38 kDa) and the GST protein
(26 kDa) (Fig. 3). The purified fusion proteins were at least
90% pure as estimated by the Quantity Oneꢁ software (Bio-
Rad).
˚
A recent study done by Welford et al. (2005) reported
difficulty in determining the activities of untagged K213A

C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
69
C.unshiu FLS
E.grandiflorum FLS
F.ananassa FLS
G.max FLS
M.domestica FLS
P.hybrida FLS
S.tuberosum FLS
V.vinifera FLS
aFLS
LPQEEKEVYSRPADAKDVQGYGTKLQKEVEG----KKSWVDHLFHRVWPPSSINYRFWPK 150
LPQEEKELIAKPEGSQSIEGYGTRLQKEVDG----KKGWVDHLFHKIWPPSSINYQFWPK 150
LPQEEKEVYAKDPNSKSVEGYGTFLQKEL-EG---KKGWVDHLFHMIWPPSAINYRFWPK 149
LPQEEKELIAKPAGSDSIEGYGTKLQKEVNG----KKGWVDHLFHIVWPPSSINYSFWPQ 149
LPQEEKEAYAKPPDSGSIEGYGTKLFKEISEGDTTKKGWVDNLFNKIWPPSVVNYQFWPK 152
VPQEEKELIAKTPGSNDIEGYGTSLQKEVEG----KKGWVDHLFHKIWPPSAVNYRYWPK 163
VPQEEKELIAKKPGAQSLEGYGTSLQKEIEG----KKGWVDHLFHKIWPPSAINYRYWPK 167
LPQVEKELYAKPPDSKSIEGYGTRLQKEEEG----KRAWVDHLFHKIWPPSAINYQFWPK 150
LPSSEKESVAKPEDSKDIEGYGTKLQKDPEG----KKAWVDHLFHRIWPPSCVNYRFWPK 150
LSVEEKEKYANDQATGKIQGYGSKLANNASG----QLEWEDYFFHLAYPEEKRDLSIWPK 160
aANS
C.unshiu FLS
E.grandiflorum FLS
F.ananassa FLS
G.max FLS
M.domestica FLS
P.hybrida FLS
S.tuberosum FLS
V.vinifera FLS
aFLS
NPPSYRAVNEEYAKYMREVVDKLFTYLSLGLGVEGGVLKEAAGG-DDIEYMLKINYYPPC 209
NPPAYREANEEYAKRLQLVVDNLFKYLSLGLDLEPNSFKDGAGG-DDLVYLMKINYYPPC 209
NPASYREANEEYAKNLHKVVEKLFKLLSLGLGLEAQELKKAVGG-DDLVYLLKINYYPPC 208
NPPSYREVNEEYCKHLRGVVDKLFKSMSVGLGLEENELKEGANE-DDMHYLLKINYYPPC 208
NPPSYREANEEYAKHLHNVVEKLFRLLSLGLGLEGQELKKAAGG-DNLEYLLKINYYPPC 211
NPPSYREANEEYGKRMREVVDRIFKSLSLGLGLEGHEMIEAAGG-DEIVYLLKINYYPPC 222
NPPSYREANEEYAKWLRKVADGIFRSLSLGLGLEGHEMMEAAGS-EDIVYMLKINYYPPC 226
NPPSYRDANEAYAKCLRGVADKLLSRLSVGLGLGEKELRESVGR-DELTYLLKINYYPPC 209
NPPEYREVNEEYAVHVKKLSETLLGILSDGLGLKRDALKEGLGG-EMAEYMMKINYYPPC 209
TPSDYIEATSEYAKCLRLLATKVFKALSVGLGLEPDRLEKEVGGLEELLLQMKINYYPKC 220
aANS
C.unshiu FLS
E.grandiflorum FLS
F.ananassa FLS
G.max FLS
M.domestica FLS
P.hybrida FLS
S.tuberosum FLS
V.vinifera FLS
Afls
PRPDLALGVVAHTDLSALTVLVPNEVPGLQVFKDDRWIDAKYIPNALVIHIGDQIEILSN 269
PRPDLALGVVAHTDMSAITVLVPNEVQGLQVYKDGHWYDCKYIPNALIVHIGDQVEIMSN 269
PRPDLALGVVAHTDMSALTILVPNEVQGLQACRDGQWYDVKYIPNALVIHIGDQMEIMSN 268
PCPDLVLGVPPHTDMSYLTILVPNEVQGLQACRDGHWYDVKYVPNALVIHIGDQMEILSN 268
PRPDLALGVVAHTDMSTVTILVPNDVQGLQACKDGRWYDVKYIPNALVIHIGDQMEIMSN 271
PRPDLALGVVAHTDMSYITILVPNEVQGLQVFKDGHWYDVKYIPNALIVHIGDQVEILSN 282
PRPDLALGVVAHTDMSYITLLVPNEVQ---VFKDGHWYDVNYIPNAIIVHIGDQVEILSN 283
PRPDLALGVVSHTDMSAITILVPNHVQGLQLFRDDHCFDVKYIPNALVIHIGDQLEILSN 269
PRPDLALGVPAHTDLSGITLLVPNEVPGLQVFKDDHWFDAEYIPSAVIVHIGDQILRLSN 269
PQPELALGVEAHTDVSALTFILHNMVPGLQLFYEGKWVTAKCVPDSIVMHIGDTLEILSN 280
aANS
C.unshiu FLS
E.grandiflorum FLS
F.ananassa FLS
G.max FLS
M.domestica FLS
P.hybrida FLS
S.tuberosum FLS
V.vinifera FLS
aFLS
GKYKAVLHRTTVNKDKTRMSWPVFLEPPADTVVG-PLPQLVD-DENPPKYKAKKFKDYSY 327
GKYKSVYHRTTVNKEKTRMSWPVFLEPPPDHEVG-PIPKLVN-EENPAKFKTKKYKDYAY 327
GKYRAVLHRTTVSKDKTRISWPVFLEPPADQVIG-PHPKLVDDKENPPKYKTKKYSEYVY 327
GKYKAVFHRTTVNKDETRMSWPVFIEPKKEQEVG-PHPKLVN-QDNPPKYKTKKYKDYAY 326
GKYTSVLHRTTVNKDKTRISWPVFLEPPADHVVG-PHPQLVN-AVNQPKYKTKKYGDYVY 329
GKYKSVYHRTTVNKDKTRMSWPVFLEPPSEHEVG-PIPKLLS-EANPPKFKTKKYKDYVY 340
GKYKSVYHRTTVNKYKTRMSWPVFLEPSSEHEVG-PIPNLIN-EANPPKFKTKKYKDYVY 341
GKYKSVLHRTTVKKDMTRMSWPVFLEPPPELAIG-PLPKPTS-EDNPPKYKKKRYCDYVY 327
GRYKNVLHRTTVDKEKTRMSWPVFLEPPREKIVG-PLPELTG-DDNPPKFKPFAFKDYSY 327
GKYKSILHRGLVNKEKVRISWAVFCEPPKDKIVLKPLPEMVS-VESPAKFPPRTFAQHIE 339
aANS
Fig. 2. Protein sequence alignment of FLS from various species with aANS. The sequence alignment indicated the amino acid residues involved in iron
binding (yellow) and 2-oxoglutarate binding (green). Five potential substrate binding residues (red) were identified via superimposition of aANS and the
modeled aFLS tertiary structures. The five selected residues are highly conserved in FLS of various species. The predicted secondary structure of aFLS is
shown on top of the sequence alignment depicting arrangement of b-sheets (arrows) and a-helices (cylinder) in aFLS.
kDa
95
~ 64kDa
54
Fig. 3. SDS–PAGE analysis of WT and mutant aFLS fusion proteins obtained after purification. Lane M shows the protein standard marker (Bio-Rad)
used for estimation of the protein size. All of the aFLS mutants were solubly expressed and the fusion proteins were at least 90% pure as estimated by the
Quantity Oneꢁ software (Bio-Rad).

70
C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
Table 1
(aANS) and K202A (aFLS) mutants due to insoluble pro-
tein expression. In this study, insolubility was not encoun-
tered in the expression of the mutant fusion proteins,
including the K202 mutants.
Specific activities and relative activities of the WT and mutant aFLSs
FLS
Specific activity
Relative activity
(%)a
(nmol min–1 mg–1
)
WT
26.9 5.3
100
2.4. Enzyme assay to determine aFLS activity
Substrate binding sites
H132Y
22.2 4.0
33.4 5.7
3.9 0.4
14.5 2.5
1.1 0.2
3.3 0.3
2.2 0.2
2.3 0.7
1.8 0.5
13.0 2.5
83
124
15
54
4
12
8
9
H132F
F134A
F134L
K202M
K202R
F293A
F293L
E295L
E295Q
A continuous spectroscopic enzyme assay was set up in
this study for efficient determination of the amount of
quercetin 6 produced by the enzyme. In the previous FLS
assay described by Wellmann et al. (2002), catalase was
included as a component of the enzyme assay as it has been
reported to provide stimulating effects for several 2-ODDs
(Blanchard et al., 1982; Britsch and Grisebach, 1986;
Kondo et al., 1981). However, the mechanism and function
of catalase in the enzyme reaction mixture remain obscure
and its use resulted in formation of a cloudy suspension.
Hence, product extraction procedures were necessary prior
to spectroscopic measurement. In this study, catalase was
excluded without seriously compromising tangible mea-
surements of FLS activities. Accepting a slight underrating
in the activity of aFLS by omitting catalase (data not
shown) enabled an easy and direct spectroscopic measure-
ment of the product formed.
7
48
Iron binding site
H221W
D223E
NDb
ND
ND
H277F
2-OG binding site
R287K
S289T
a
ND
12.9 1.6
48
Relative activity of each mutant enzyme is expressed as the percentage
of the specific activity of the mutant enzyme relative to that of the WT
enzyme at 100%.
b
ND- not detectable.
After a full spectral scan of all the components and
products of the enzyme assay, a wavelength of 375 nm
was found to be optimum for quantitating the amount of
quercetin 6 formed in the enzyme reaction. The reagents
used in the enzyme assay and succinate, the by-product
formed during catalysis, had negligible absorbance reading
at this particular wavelength (data not shown). The high
performance liquid chromatography (HPLC) results fur-
ther verified that the absorbance reading obtained at
375 nm was solely contributed by quercetin 6, as the reten-
tion time of the converted product (single peak) coincided
with that of the quercetin 6 standard (36 min) (data not
shown). Hence, the spectroscopic enzyme assay was carried
out in a 96-well format with the absorbance of the product
measured at 375 nm. Subsequent kinetic studies of the
enzymes were performed using the same assay.
other 2-ODDs at similar sites (Sim and Sim, 1999; Doan
et al., 2000). The abolished activity of D223E mutant
clearly showed that the aspartate in the aFLS HxDxnH
motif is non-interchangeable with a glutamate. The unique
structure of clavaminate synthase that requires an amino
acid with a slightly longer side-chain for securing the fer-
rous iron may account for the presence of a glutamate in
the iron binding motif, HxExnH (Zhang et al., 2002). How-
ever, in 2-ODDs with an HxDxnH motif, substitution to a
glutamate residue would disrupt the precise arrangement of
the 2-His-1-carboxylate facial triad and may render ineffec-
tive iron binding and catalysis.
Similar abolishment in the activity of R287K mutant
demonstrated that this residue is crucial in aFLS. Crystal-
lographic studies carried out on 2-ODDs showed that the
conserved arginine residue would orientate the 2-OG
through hydrogen bond for priming the ferrous iron (Wil-
mouth et al., 2002; Zhang et al., 2000). Hence, when substi-
tuted to a lysine with a shorter side, the iron priming ability
of 2-OG would be reduced due to the lack of hydrogen
bond support and this may lead to the loss of the enzyme
activity.
2.5. Effects of mutations on the aFLS activity
To test the effects of the directed mutations on enzyme
activity, the specific activities of the mutant aFLS enzymes
were assayed using DHQ 4 as substrate and compared with
the WT enzyme (Table 1). The specific activity of the
enzyme is defined as the number of moles of quercetin pro-
duced per mg of enzyme in 1 min at 37 ꢂC under the pre-
scribed condition.
On the other hand, the S289T mutant retained approx-
imately 50% (12.9 nmol minꢀ1 mgꢀ1) of WT activity
(26.9 nmol minꢀ1 mgꢀ1). This is similar to results of a
study on F3H, where the S290T mutant retained approxi-
mately 20% of WT activity (Lukacin et al., 2000). Con-
versely, in C. unshiu FLS mutant, S289L, its activity was
completely abolished (unpublished). In retrospect, these
2.5.1. Cosubstrate and cofactor binding residues
Mutations at the proposed iron binding residues,
H221W, D223E and H277F knocked out the activities of
the mutants supporting the widely accepted concept that
these residues are crucial for enzyme catalysis. This is also
consistent with results of mutational studies performed on

C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
71
Table 2
results support the requirement of the serine residue in FLS
The kinetic properties of WT and mutant aFLSs
and F3H for sustaining high enzyme activity.
FLS
WT
Km (mM)a
kcat (sꢀ1 a
)
kcat/Km (Mꢀ1 sꢀ1
)
2.5.2. Substrate binding residues
0.059 0.0063
0.056 0.0106
948.2
Mutations at the proposed substrate binding residues
generally led to a significant decrease in the activities of
the mutants except for H132F (Table 1). Most notably,
the dissimilar mutations (F134A, K202M, F293A and
E295L) resulted in enzymes with less than 20% of WT
activity. Upon mutation to similar residues, F134L and
E295Q mutants showed recovery of activities and retained
approximately 50% of WT activity. However, the activities
of K202R and F293L were consistently low as only 10% of
WT activity was retained. The lack of recovery in the activ-
ities of K202 and F293 mutants suggest that these residues
are important for substrate conversion.
Computational analyses further revealed that the resi-
due K202 is capable of forming a hydrogen bond with
the C ring of DHQ 4. Hence, the lack of activities in
K202 mutants may be attributable to the loss of hydrogen
bond. Although arginine is structurally similar to lysine,
the increased distance between the guanidinium group of
arginine and the C ring of DHQ 4 may weaken the existing
hydrogen bond. Therefore, the activities of K202R
remained significantly low despite replacement with an
amino acid of similar properties. On the other hand, the
residue F293 was postulated to be involved in forming p
stacking with the phenyl A ring of DHQ 4 as described
by Wilmouth et al. (2002). As p stacking involves interac-
tion between the electron clouds of two phenyl rings, the
interaction force may be extensive. Hence, it may represent
the major force responsible for binding the DHQ 4 in an
optimal spatial position for desaturation to occur. The loss
of the phenyl ring structures in the F293 mutants could
possibly explain the diminished activities observed in the
mutants.
H132Y
H132F
F134L
E295Q
S289T
a
0.031 0.0040
0.027 0.0028
0.129 0.0150
0.070 0.0028
0.320 0.0287
0.030 0.0022
0.034 0.0037
0.019 0.0008
0.015 0.0008
0.024 0.0011
966.5
1247.9
151.2
208.3
75.3
The results were obtained from at least three sets of independent
experiments with duplicates.
(0.059 mM). This indicates slight improvement in the sub-
strate binding affinities of the H132 mutants and provides a
basis for the improved specific activities observed in the
H132F mutant. However, for the S289T and F134L
mutants, their substrate binding affinities were negatively
affected, as reflected by their increased Km values. All the
aFLS mutants assayed were found to exhibit an approxi-
mately 2–4-fold lower kcat than WT (0.056 sꢀ1) including
the H132 mutants with decreased Km values; this implies
a decreased turnover number in the mutant enzymes.
2.7. Conclusion
The pronounced reduction in activities of aFLS mutants
exemplified the significance of the substituted amino acid
residues for enzyme catalysis and hence, further corrobo-
rated their respective recognition as ferrous iron, 2-OG,
or DHQ 4 binding residues. The trends observed in the
kinetic analyses of the selected mutants further supported
the results of the enzyme assay and two observations were
made: (1) mutations at these sites consistently led to a
decrease in turnover number of the enzyme and (2) H132
is an attractive site for modification to engineer aFLS with
improved activity. As these selected residues were found to
be highly conserved in the FLSs of other species (Fig. 2), it
is reasonable to postulate that corresponding residues in
other FLSs are potentially active site residues as well. Of
note, the substrate binding residues of aFLS in this study
was assigned based on incubation of aFLS with DHQ 4.
Hence, further investigation is necessary to determine the
roles of these residues in handling flavanones (e.g. eri-
odictyol and pinocembrin) as a substrate for the hydroxyl-
ation reaction.
As for the H132Y mutant, the enzyme retained approx-
imately 83% of WT activity, whereas H132F demonstrated
higher activity than WT by approximately 20%. The results
clearly indicated that any hydrogen bond formed between
either the tyrosine or histidine residue with DHQ 4 was
non-essential. Instead, the hydrophobic environment pro-
vided by the phenylalanine side chain was preferred for
desaturation of DHQ 4. Therefore, substitution with other
suitable hydrophobic amino acids can be explored for engi-
neering aFLS with improved activities.
2.6. Kinetic analyses of WT and selected aFLS mutants
3. Experimental
The kinetic parameters of WT, S289T, H132F, H132Y,
F134L and E295Q mutants were investigated to further
examine their roles in enzyme catalysis (Table 2). As for
the other mutants, their specific activities were significantly
lower than WT, and hence their kinetic properties were not
pursued.
3.1. Computational analyses
Nucleotide and protein sequences of aFLS were
retrieved from the GenBank Database, with accession
numbers NM_120951 and NP_196481, respectively. Simi-
larly, the FLS sequences of various species were obtained:
C. unshiu (AB011796), Eustoma grandiflorum (AB078965),
The Km of H132F (0.027 mM) and H132Y (0.031 mM)
mutants were approximately 2-fold lower than WT

72
C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
Fragaria ananassa (DQ834905), Glycine max (AB246668),
Malus domestica (AY965343), P. hybrida (Z22543), Sola-
num tuberosum (X92178) and Vitis vinifera (AB086055).
The protein sequence of aANS (1GP5) was retrieved from
the Protein Data Bank (Wilmouth et al., 2002). Subse-
quently, multiple sequence alignment was performed using
CLUSTAL W Multiple Alignment Program (Thompson
et al., 1994) to identify the conserved regions in the nucle-
otide sequences of FLSs in various species.
The secondary structure of FLS was predicted using the
PSIPRED program (http://bioinf.cs.ucl.ac.uk/psipred/)
(Jones, 1999). The tertiary structure of aFLS was predicted
using SWISS-MODEL program (http://swissmodel.exp-
asy.org/) (Schwede et al., 2003), based on comparative
modeling. The ‘‘first approach mode’’ was selected for
the modeling process and the submitted sequence (aFLS)
was modeled using templates which share at least 25%
sequence identity.
aFLS gene and the results were analyzed using ABI
PRISMe 3100 Genetic Analyzer.
3.4. Subcloning of aFLS
The vector containing the aFLS gene was subsequently
double-digested with BamHI and SalI and subcloned into
the expression vector, pGK (Loke and Sim, 2000). The
resultant plasmid was designated as pGK-aFLS. The
recombinant plasmids were transformed into E. coli BL21
(DE3) cells for high level expression of soluble GST–aFLS
fusion protein. Similarly, sequencing was performed to
identify and authenticate recombinant clones containing
the gene of interest.
3.5. Heterologous expression and purification of GST–aFLS
fusion protein
All protein structures were viewed using SPdbV pro-
gram (http://www.expasy.org/spdbv/) (Guex et al., 1999),
and this program also allowed superimposition of crystal
structures. The protein structures analyzed in detail in this
study were aANS (1GP5) and aFLS (predicted).
Recombinant cultures were grown in 50 mL Luria Ber-
tani broth supplemented with 50 lg mLꢀ1 kanamycin
(Sigma) for expression studies. The cultures were grown
at 37 ꢂC with continuous agitation at 250 rpm until an
OD600 of 0.5–0.7 was reached. They were subsequently
induced with 1 mM isopropyl-b-D-thiogalactopyranoside
(Sigma) and incubated at room temperature for 15 h with
shaking at 250 rpm.
3.2. Reverse transcription of aFLS
Total RNA from the roots of A. thaliana was generously
provided by Dr. Liu Xianan from Colorado State Univer-
sity. The mRNAs were purified from the total RNA
through the Oligotex mRNA Mini Kit (Qiagen) in accor-
dance with the protocol provided. After purification, the
mRNAs were reverse-transcribed to obtain cDNA frag-
ments using Sensiscript Reverse Transcriptase Kit (Qia-
gen). Subsequently, polymerase chain reaction (PCR) was
performed to amplify the gene of interest, aFLS, using
the forward (OL1231) and reverse (OL1232) primers with
BamHI and SalI restriction sites incorporated, respectively
(Table 3). The PCR protocol consisted of 1 cycle of 1 min
at 95 ꢂC followed by 35 cycles of 30 s at 95 ꢂC, 30 s at
60 ꢂC and 3 min at 72 ꢂC, prior to the final extension step
of 5 min at 72 ꢂC.
Following induction, the cultures were harvested by cen-
trifugation at 7000 rpm for 5 min at 4 ꢂC. For each prepa-
ration, the supernatant was discarded and the pelleted cells
were resuspended in cold phosphate buffered saline. The
cells were subsequently lysed via sonication with a 1/800
probe at an intensity of 10 lm for 8 cycles (8 s pulse with
10 s rest in each cycle). The crude extract obtained after
sonication was centrifuged at 12,000 rpm for 15 min and
the soluble protein fraction (supernatant) was subsequently
purified using the Microspin GST Purification Module
(Amersham Pharmacia Biotech). Following the manufac-
turer’s guidelines, each protein of interest was eluted in glu-
tathione elution buffer (GEB). All soluble protein fractions
and purified enzymes were subsequently analyzed using
10% SDS–PAGE. The images of the gel were captured
using Bio-Rad Molecular Imager Gel Doc XR, equipped
with the Quantity Oneꢁ software. The program was also
utilized to estimate the amount and purity of the expressed
proteins. The purified protein concentration was measured
by Bio-Rad protein assay using bovine serum albumin
(Sigma) as the protein standard.
3.3. Cloning of aFLS
The amplified gene products were separated via agarose
gel electrophoresis using 0.8% (w/v) agarose gel. The band
containing the fragment of interest, aFLS gene (ꢁ1 kb),
was excised and purified using the MinElutee Gel Extrac-
tion Kit (Qiagen) and subsequently ligated into pCR-Blun-
tII-TOPO Vector using the Zero Blunte TOPOe PCR
Cloning Kit (Invitrogen). The plasmids of the positive
recombinant clones were then extracted using Wizardꢁ
Plus SV Minipreps DNA Purification System from Pro-
mega. Full sequencing of both strands of the aFLS gene
was carried out using the ABI PRISMe BigDyee Termi-
nator Cycle Sequencing Ready Reaction Kit from Applied
Biosystems to verify the recombinant vectors harboring the
3.6. Site-directed mutagenesis
The recombinant expression vector, pGK-aFLS, was
used as a template to construct aFLS mutants using Quik-
Changee Site-directed Mutagenesis Kit from Stratagene
(Loke and Sim, 2001). The primers required for each of
the mutations were designed as shown in Table 3. Sequenc-
ing in both strands of the DNA templates was employed to

C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
73
Table 3
The list of primers used in this study
Purpose
Primer
Sequence (50–30)
Cloning of aFLS gene into pGK vector
OL1231
BamHI
CCAAAAGGATCCATGGAGGTCGAAAGAGTCCAAGAC
SalI
OL1232
CGTATTGAGTCGACTCAATCCAGAGGAAGTTTATTGAG
Mutagenesis
Substrate binding site
H132Y
OL1245
OL1246
GGGTCGATTATCTCTTCCATCG
GGAAGAGATAATCGACCCAAGC
H132F
F134A
F134L
K202M
K202R
F293A
F293L
E295L
E295Q
OL1270
OL1271
GGGTCGATTTTCTCTTCCATCG
GGAAGAGAAAATCGACCCAAGC
OL1247
OL1248
GATCATCTCGCCCATCGAATCTGG
GATTCGATGGGCGAGATGATCGAC
OL1272
OL1273
GATCATCTCCTCCATCGAATCTGG
GATTCGATGGAGGAGATGATCGAC
OL1274
OL1275
GCGGAGTATATGATGATGATTAACTATTATCCGCCG
GGCGGATAATAGTTAATCATCATCATATACTCCGCC
OL1276
OL1277
GTATATGATGAGGATTAACTATTATCCGCCG
GATAATAGTTAATCCTCATCATATACTCCGC
OL1251
OL1252
GGCCGGTTGCCTTGGAGCCTCCC
GGCTCCAAGGCAACCGGCCACG
OL1278
OL1279
GGCCGGTTCTCTTGGAGCCTCCC
GGCTCCAAGAGAACCGGCCACG
OL1253
OL1254
GGTTTTCTTGCTGCCTCCCCGTG
CGGGGAGGCAGCAAGAAAACCGG
OL1280
OL1281
GGTTTTCTTGCAGCCTCCCCGTG
CGGGGAGGCTGCAAGAAAACCGG
Iron (II) binding site
H221W
OL1282
OL1283
GTACCGGCTTGGACAGATCTCAGTGG
CTGAGATCTGTCCAAGCCGGTACAC
D223E
H277F
OL1284
OL1285
GCTCATACAGAGCTCAGTGGAATC
TCCACTGAGCTCTGTATGAGCCGG
OL1286
OL1287
AATGTGTTGTTCAGGACGACGGTGG
CCGTCGTCCTGAACAACACATTTTTATACC
2-OG binding site
R287K
OL1288
OL1289
GAGAAGACGAAGATGTCGTGGCCG
CCACGACATCTTCGTCTTCTCTTTATCC
S289T
OL1290
OL1291
GACGAGGATGACGTGGCCGGTTTTC
AACCGGCCACGTCATCCTCGTCTTC
verify the authenticity of the desired aFLS mutants after
SDM.
Each assay reaction mixture contained 100 lM of DHQ,
111 mM NaOAc, 2.5 mM NaAsc, 42 lM iron (II) sulfate,
approximately 12.5 lg of enzyme and 83 lM 2-OG (Well-
mann et al., 2002). The reaction mixture was added in
the order described above into a 96-well plate. Subse-
quently, the reaction mixture was incubated at 37 ꢂC for
7 min prior to measuring the absorbance of the product
(quercetin 6) at 375 nm.
The same assay was carried out to determine the kinetic
properties of the enzyme. Each enzyme was tested using
eight different substrate concentrations (between 20 lM
to 1000 lM of DHQ) and the absorbance was measured
per minute for 15 min. The Km and kcat of the enzymes
were determined using Hanes–Woolf plot.
3.7. Assay for aFLS activity
The full spectral scan, from wavelengths 230 nm to
1000 nm (at an increment of 1 nm), was performed using
the Safire, Tecan Instrument (Austria) on the components
(DHQ, sodium acetate (NaOAc), sodium ascorbate
(NaAsc), iron (II) sulfate, aFLS enzyme, 2-OG) and prod-
ucts (quercetin 6, succinate) of the enzyme assay with con-
trols (GEB, water) included. The optimum wavelength of
375 nm was used for measuring quercetin 6 in the enzyme
assays.

74
C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
Clifton, I.J., McDonough, M.A., Ehrismann, D., Kershaw, N.J., Grana-
Each set of experiments was performed with duplicates
tino, N., Schofield, C.J., 2006. Structural studies on 2-oxoglutarate
oxygenases and related double-stranded b-helix fold proteins. J. Inorg.
Biochem. 100, 644–669.
and the specific activities and kinetic parameters reported
were based on at least three sets of independent
experiments.
Cushnie, T.P.T., Lamb, A.J., 2005. Antimicrobial activity of flavonoids.
Int. J. Antimicrob. Agents 26, 343–356.
Cushnie, T.P.T., Hamilton, V.E.S., Lamb, A.J., 2003. Assessment of
the antibacterial activity of selected flavonoids and consideration
of discrepancies between previous reports. Microbiol. Res. 158, 281–
289.
Doan, L.X., Hassan, A., Lipscomb, S.J., Dhanda, A., Zhang, Z.H.,
Schofield, C.J., 2000. Mutagenesis studies on the iron binding ligands
of clavaminic acid synthase. Biochem. Biophys. Res. Commun. 279,
240–244.
Elkins, J.M., Hewitson, K.S., McNeill, L.A., Seibel, J.F., Schlemminger,
I., Pugh, C.W., Ratcliffe, P.J., Schofield, C.J., 2003. Structure of
factor-inhibiting hypoxia-inducible factor (HIF) reveals mechanism of
oxidative modification of HIF-1alpha. J. Biol. Chem. 278, 1802–1806.
Forkmann, G., deVlaming, P., Spribille, R., Wiering, H., Schramm, A.W.,
1986. Genetic and biochemical studies on the conversion of dihydrofl-
avonols to flavonols in flowers of Petunia hybrida. Z. Naturforsch. 41,
179–186.
3.8. Verification of the product detected at 375 nm using high
performance liquid chromatography
HPLC was used to verify the compound that was detected
spectroscopically at 375 nm. The presence of additional
compounds which could contribute to the absorbance read-
ing at 375 nm other than the product, quercetin 6, would be
detected using the sensitive reversed-phase HPLC analysis.
¨
The liquid chromatography system used was AKTA purifier
(Amersham Biosciences) consisting of a UV-900 detector,
Autosampler A900, and a 5 lm 4.6 · 150 mm SunFiree
C18 column from Waters (USA). The UNICORN 5.0 pro-
gram (Amersham Biosciences) was used to coordinate the
operation of the chromatographic and data acquisition
processes.
Guex, N., Diemand, A., Peitsch, M.C., 1999. Protein modeling for all.
Trends Biochem. Sci. 24, 364–367.
Harborne, J.B., Williams, C.A., 2000. Advances in flavonoid research
since 1992. Phytochemistry 55, 481–504.
Havsteen, B., 1983. Flavonoids, a class of natural products of high
pharmacological potency. Biochem. Pharmacol. 32, 1141–1148.
Havsteen, B.H., 2002. The biochemistry and medical significance of the
flavonoids. Pharmacol. Ther. 96, 67–202.
Jones, D.T., 1999. Protein secondary structure prediction based on
position-specific scoring matrices. J. Mol. Biol. 292, 195–202.
Koehntop, K.D., Emerson, J.P., Que Jr., L., 2005. The 2-His-1-carbox-
ylate facial triad: a versatile platform for dioxygen activation by
mononuclear non-heme iron (II) enzymes. J. Biol. Inorg. Chem. 10,
87–93.
Kondo, A., Blanchard, J.S., Englard, S., 1981. Purification and properties
of calf liver c-butyrobetaine hydroxylase. Arch. Biochem. Biophys.
212, 338–346.
The enzyme assay was performed as described earlier and
MeOH (70 lL) was added to stop the reaction after 7 min.
The reaction mixture was centrifuged at 12,000 rpm for
5 min to pellet any precipitate. Subsequently, the superna-
tant was injected into the HPLC system. Solvent A (80%
(v/v) of 10 mM KH2PO4, pH 2.0, 0.06 mM TBAB; 20%
(v/v) MeOH) was used to equilibrate the column and the
compound was subsequently eluted in a linear gradient of
solvent A and B (100% MeOH). The elution step was carried
out at 0.5 mL minꢀ1 over 25 min. The absorption profiles of
the compounds were monitored at 290 nm and 375 nm to
detect for the presence of substrate 4 (DHQ) and product
6 (quercetin), respectively. Pure DHQ 4 and quercetin 6
(Sigma) were used as controls to verify the identity of the
product formed by its retention time.
Lange, S.J., Que Jr., L., 1998. Oxygen activating nonheme iron enzymes.
Curr. Opin. Chem. Biol. 2, 159–172.
Loke, P., Sim, T.S., 2000. Mutational analysis of tyrosine191 in the
catalysis of Cephalosporium acremonium isopenicillin N synthase. J.
Biochem. 127, 585–589.
Acknowledgements
Loke, P., Sim, T.S., 2001. A comparison of three site-directed mutagenesis
kits. Z. Naturforsch. 56, 810–813.
Lukacin, R., Groning, I., Pieper, U., Matern, U., 2000. Site-directed
mutagenesis of the active site serine290 in flavanone 3b-hydroxylase
from Petunia hybrida. Eur. J. Biochem. 267, 853–860.
Lukacin, R., Wellmann, F., Britsch, L., Martens, S., Matern, U., 2003.
Flavonol synthase from Citrus unshiu is a bifunctional dioxygenase.
Phytochemistry 62, 287–292.
We thank Dr. Liu Xianan from Colorado State Univer-
sity for providing A. thaliana total RNA extract.
References
Moriguchi, T., Kita, M., Ogawa, K., Tomono, Y., Endo, T., Omura, M.,
2002. Flavonol synthase gene expression during citrus fruit develop-
ment. Physiol. Plant. 114, 251–258.
Narayana, K.R., Reddy, M.S., Chaluvadi, M.R., Krishna, D.R., 2001.
Bioflavonoids classification, pharmacological, biochemical effects and
therapeutic potential. Indian J. Pharm. 33, 2–16.
Ormrod, D.P., Landry, L.G., Conklin, P.L., 1995. Short term UV-B
radiation and ozone exposure effects on aromatic secondary metabolite
accumulation and shoot growth of flavonoid-deficient Arabidopsis
mutants. Physiol. Plant. 93, 602–610.
Anzellotti, D., Ibrahim, R.K., 2000. Novel flavonol 2-oxoglutarate
dependent dioxygenase: affinity purification, characterization, and
kinetic properties. Arch. Biochem. Biophys. 382, 161–172.
Anzellotti, D., Ibrahim, R.K., 2004. Molecular characterization and
functional expression of flavonol 6-hydroxylase. BMC Plant Biol. 4,
20.
Blanchard, J.S., Englard, S., Kondo, A., 1982. c-Butyrobetaine hydrox-
ylase: a unique protective effect of catalase. Arch. Biochem. Biophys.
219, 327–334.
Britsch, L., Grisebach, H., 1986. Purification and characterization of (2S)-
flavanone 3-hydroxylase from Petunia hybrida. Eur. J. Biochem. 156,
569–577.
Britsch, L., Heller, W., Grisebach, H., 1981. Conversion of flavanone to
flavone, dihydroflavonol and flavonol with an enzyme system from cell
cultures of parsley. Z. Naturforsch. 36, 742–750.
Pelletier, M.K., Murrell, J.R., Shirley, B.W., 1997. Characterization of
flavonol synthase and leucoanthocyanidin dioxygenase genes in
Arabidopsis: further evidence for differential regulation of early and
late genes. Plant Physiol. 113, 1437–1445.
Pietta, P.G., 2000. Flavonoids as antioxidants. J. Nat. Prod. 63, 1035–
1042.

C.S. Chua et al. / Phytochemistry 69 (2008) 66–75
75
Schwede, T., Kopp, J., Guex, N., Peitsch, M.C., 2003. SWISS-MODEL:
an automated protein homology-modeling server. Nucleic Acids Res.
31, 3381–3385.
Sim, J., Sim, T.S., 1999. Mutational evidence supporting the involvement
of tripartite residues His183, Asp185, and His243 in Streptomyces
clavuligerus deacetoxycephalosporin C synthase for catalysis. Biosci.
Biotechnol. Biochem. 64, 828–832.
dependent oxygenases of flavonoid biosynthesis. J. Biol. Chem. 279,
1206–1216.
Welford, R.W.D., Clifton, I.J., Turnbull, J.J., Wilson, S.C., Schofield,
C.J., 2005. Structural and mechanistic studies on anthocyanidin
synthase catalysed oxidation of flavanone substrates: the effect of C-
2 stereochemistry on product selectivity and mechanism. Org. Biomol.
Chem. 3, 3117–3126.
Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W:
improving the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties and weight
matrix choice. Nucleic Acids Res. 22, 4673–4680.
Turnbull, J.J., Sobey, W.J., Aplin, R.T., Hassan, A., Firmin, J.L.,
Schofield, C.J., Prescott, A.G., 2000. Are anthocyanidins the imme-
diate products of anthocyanidin synthase? Chem. Commun., 2473–
2474.
Turnbull, J.J., Nagle, M.J., Seibel, J.F., Welford, R.W.D., Grant, G.,
Schofield, C.J., 2003. The C-4 stereochemistry of leucocyanidin
substrates for anthocyanidin synthase affects product selectivity.
Bioorg. Med. Chem. Lett. 13, 3853–3857.
Turnbull, J.J., Nakajima, J., Welford, R.W.D., Yamazaki, M., Saito, K.,
Schofield, C.J., 2004. Mechanistic studies on three 2-oxoglutarate-
Wellmann, F., Lukacin, R., Moriguchi, T., Britsch, L., Schiltz, E.,
Matern, U., 2002. Functional expression and mutational analysis of
flavonol synthase from Citrus unshiu. Eur. J. Biochem. 269, 4134–4142.
Wilmouth, R.C., Turnbull, J.J., Welford, R.W.D., Clifton, I.J., Prescott,
A.G., Schofield, C.J., 2002. Structure and mechanism of anthocyanidin
synthase from Arabidopsis thaliana. Structure 10, 93–103.
Zhang, Z.H., Ren, J.S., Stammers, D.K., Baldwin, J.E., Harlos, K.,
Schofield, C.J., 2000. Structural origins of the selectivity of the
trifunctional oxygenase clavaminic acid synthase. Nat. Struct. Biol. 7,
127–133.
Zhang, Z.H., Ren, J.S., Harlos, K., McKinnon, C.H., Clifton, I.J.,
Schofield, C.J., 2002. Crystal structure of a clavaminate synthase-
Fe(II)-2-oxoglutarate-substrate-NO complex: evidence for metal cen-
tered rearrangements. FEBS Lett. 517, 7–12.