The C-4 stereochemistry of leucocyanidin substrates for anthocyanidin synthase affects product selectivity

Article abstract of DOI:10.1016/S0960-894X(03)00711-X

Anthocyanidin synthase (ANS), an iron(II) and 2-oxoglutarate (2OG) dependent oxygenase, catalyses the penultimate step in anthocyanin biosynthesis by oxidation of the 2R,3S,4S-cis-leucoanthocyanidins. It has been believed that in vivo the products of ANS are the anthocyanidins. However, in vitro studies on ANS using optically active cis- and trans-leucocyanidin substrates identified cyanidin as only a minor product; instead both quercetin and dihydroquercetin are products with the distribution being dependent on the C-4 stereochemistry of the leucocyanidin substrates.

Full text of DOI:10.1016/S0960-894X(03)00711-X

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3853–3857
The C-4 Stereochemistry of Leucocyanidin Substrates for
Anthocyanidin Synthase Affects Product Selectivity
Jonathan J. Turnbull,a Michael J. Nagle,b Jurgen F. Seibel,a Richard W. D. Welford,a
Guy H. Grantb and Christopher J. Schofielda,*
aThe Oxford Centre for Molecular Sciences and The Dyson Perrins Laboratory, The Department of Chemistry,
South Parks Road, Oxford OX13QY, UK
bThe Physical and Theoretical Chemistry Laboratory, The Department of Chemistry, South Parks Road, Oxford OX13QZ, UK
Received 24 February 2003; revised 3 July 2003; accepted 4 July 2003
Abstract—Anthocyanidin synthase (ANS), an iron(II) and 2-oxoglutarate (2OG) dependent oxygenase, catalyses the penultimate
step in anthocyanin biosynthesis by oxidation of the 2R,3S,4S-cis-leucoanthocyanidins. It has been believed that in vivo the pro-
ducts of ANS are the anthocyanidins. However, in vitro studies on ANS using optically active cis- and trans-leucocyanidin sub-
strates identified cyanidin as only a minor product; instead both quercetin and dihydroquercetin are products with the distribution
being dependent on the C-4 stereochemistry of the leucocyanidin substrates.
# 2003 Elsevier Ltd. All rights reserved.
The anthocyanins 1 are a large class of plant flavonoids
that display a variety of biological roles.1 Their bio-
synthesis proceeds via the formation of a 2S-flavanone 2
from 4p-coumaroyl-CoA 3 and malonyl-CoA 4 in a
reaction catalysed by a polyketide synthase, chalcone
synthase (CHS) (Fig. 1).2,3 Ring closure to the ‘closed-
flavonoid’ nucleus is then catalysed by chalcone iso-
merase (CHI).2,4 A subsequent sequence of oxidation–
reduction–oxidation reactionsiscatalysed by flavanone
3b-hydroxylase (F3H), dihydroflavonol reductase
(DFR) and anthocyanidin synthase (ANS), respec-
tively.2 Finally, glycosylation is thought to be mediated
The direct productsof ANS catalysishave been widely
assumed to be the anthocyanidins 5.2,6,9ꢀ11 However, we
recently reported that cyanidin 6 represented only a
minor product of in vitro incubationsof ANS with
commercial leucocyanidin (LCD) (a mixture of
2R,3S,4S-cis-7 and 2R,3S,4R-trans-8 epimers).12 The
major product, after quenching, wasthe four-electron
oxidation product, quercetin 9. Of the observed two-
electron oxidation products, cis-DHQ 10 (only one
enantiomer observed, >90% ee) predominated over the
more thermodynamically stable trans-DHQ isomer 11
and cyanidin 6.
by
a UDP-glucose:flavonoid 3O-glycosyltransferase
(FGT) yielding the anthocyanin products 1.2,5,6 ANS,
and several other oxygenases of flavonoid biosynthesis
including F3H, belong to the iron(II) and 2OG depen-
dent oxygenase family.2,6 These enzymes typically cou-
ple the two-electron oxidation of substrates with the
reaction of 2OG and dioxygen to produce succinate and
carbon dioxide. The oxidation of substrates is believed
to be mediated by a reactive iron oxidising species
Here, we report in vitro studies with ANS using opti-
cally active LCD substrates epimeric at the C-4 posi-
tion. The results demonstrate that the stereochemistry
at the C-4 position of the LCD substrate directly affects
product selectivity (Figs. 2 and 3).
2R,3S,4R-trans-LCD 8, prepared by sodium borohy-
dride mediated reduction of the C-4 ketone function of
2R,3R-trans-DHQ 11, wasepimeriesd at C-4 under
mildly acidic conditions, according to Stafford et al.13
Separation of the C-4 LCD epimers 7/8 hasbeen
reported on a small scale using paper chromatography
and HPLC. Separation of the LCD epimersfrom the
phosphate HPLC buffer was achieved by addition of a
C-18E solid-phase extraction purification step.13 The
7,8
.
[Fe(IV)=O $ Fe(III)-O ].
*Corresponding author. Tel.:+44-1865-275677; fax: +44-1865-
275625; e-mail: christopher.schofield@chem.ox.ac.uk
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00711-X

3854
J. J. Turnbull et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3853–3857
Figure 1. Anthocyanin 1 biosynthesis. Both the previously proposed in vivo products of ANS, anthocyanidins 5, and the nascent in vivo product
proposed in this work, a 4S-flav-2-en-3,4-diol 19, are shown.
Figure 2. Quercetin 9, isthe major product formed on incubation of
Figure 3. Productsformed on incubation of 2 R,3S,4R-trans-LCD 8
2R,3S,4S-cis-LCD 7 with ANS (minor products, include 2R,3S-cis-10
with ANS. It cannot be ruled out that 2S,3R,4S-trans-LCD 13 isa
DHQ and 2R,3R-trans-11 DHQ and cyanidin 6). It cannot be ruled
poor substrate for ANS and may form DHQ products enantiomeric
out that 2S,3R,4R-cis-LCD 12 is a poor substrate for ANS.
with those shown above.
HPLC analysis of incubations of ANS with either opti-
purified LCDswere characteriesd by 1-D and 2-D 1H
NMR. The identity of the separated C-4 epimers was
demonstrated, in part, by analysis of the C-3,C-4 H
coupling constants (J=3.5 Hz for cis-LCD 7 and J=7.0
Hz for trans-LCD 8).13 In addition to preparing opti-
cally active 2R,3S,4S-cis-LCD 7 and 2R,3S,4R-trans-
LCD 8, racemic cis-LCD 7/12 and trans-LCD 8/13
substrates were also synthesised.
cally active 2R,3S,4R-trans-LCD 8 or racemic trans-LCD
8/13 gave a more even product distribution, with cis-DHQ
clearly being the major product (Table 1 and Fig. 3). These
results imply that most (if not all) of the cis-DHQ 10,
trans-DHQ 11 and cyanidin 6 produced in ANS incu-
bationsof commercial LCD 7/8 arises from 2R,3S,4R-
trans-LCD 8.
1
A 2OG decarboxylation assay was used to measure the
specific activity of ANS with the different LCD sub-
strates (Fig. 4). A reduced rate of 14CO2 formation was
observed in the presence of racemic cis-LCD 7/12 and
trans-LCD 8/13 substrates relative to the chirally pure
substrates 7 and 8 (Table 2).
HPLC analysis of incubations of ANS with either opti-
cally active 2R,3S,4S-cis-LCD 7 or racemic cis-LCD 7/12
demonstrated that quercetin 9 wasthe major product
(Table 1 and Fig. 2). Minor productsobserved were cis-
DHQ 10, trans-DHQ 11 and cyanidin 6.

J. J. Turnbull et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3853–3857
3855
Table 1. Observed product distributions from ANS incubations (20 min) with 2R,3S,4S-cis-LCD 7 and 2R,3S,4R-trans-LCD 8 after quenching
with MeOH (percentage product distributions calculated by analysing HPLC peak area against standards)13,14
Substrate
% cis-DHQ
% trans-DHQ
% Cyanidin
% Quercetin
10
11
6
9
2R,3S,4S-cis-LCD 7
2R,3S,4R-trans-LCD 8
10
55
3
11
2
4
85
30
LCD 7/8 are prone to epimerisation (at C-2 and C-4) and it is possible that some of the minor products arise from ‘contaminating’ stereoisomers.
However, no evidence was obtained by HPLC analysis to indicate that such epimerisations were occurring under the incubation conditions.
Table 2. Rate of 14CO2 formation from 1-[14C]-2OG in the presence of ANS
Substrate
Specific activity
of 14CO2 formation
from 1-[14C]-2OG
Specific activity of
14CO2 formation from
1-[14C]-2OG corrected
for uncoupled turnover
(mkat kgꢀ1
)
(mkat kgꢀ1
)
Racemic cis-LCD 7/12
2R,3S,4S-cis-LCD 7
Racemic trans-LCD 8/13
2R,3S,4R-trans-LCD 8
84.1 (ꢁ6.9)
181.5 (ꢁ28.7)
188.4 (ꢁ19.3)
385.0 (ꢁ31.4)
41
138
145
342
Values are means of three experiments, standard deviation is given in parentheses.
Figure 4. 2OG turnover by ANS.
These results imply that the active site of ANS is selec-
tive for LCD substrates with the natural 2R,3S-stereo-
chemistry; this conclusion is supported by
crystallographic evidence.15 The studies also infer that
2R,3S,4R-trans-LCD 8 ismore catalytically active
under in vitro conditionsthan the natural 2 R,3S,4S-cis-
LCD 7 stereoisomer.
Figure 5. Possible initial oxidation products of 2R,3S-LCDs 7/8.
In the formation of cis-DHQ 10 from either 2R,3S,4S-
cis-LCD 7 or 2R,3S,4R-trans-LCD 8 there isan altera-
tion in the relative C-2 and C-3 stereochemistry. Possi-
ble intermediatesenabling thisare a flav-3-on-4-ol 15/16
(from 2R,3S,4S-cis-LCD 7 and 2R,3S,4R-trans-LCD 8,
respectively), a flavan-3,3,4-triol 17/18, a flav-2-en-3,4-
diol 19/20 or a 2R-flav-3-en-3,4-diol 21 (Fig. 5).
Crystallographic analyses and studies demonstrating
C-3 hydroxylation of the unnatural substrate naringenin
14 (Fig. 1) by ANS, support a mechanism for ANS cat-
alysis involving initial C-3 oxidation of the flavonoid
nucleus.15,16 Furthermore, when an LCD substrate
labelled with deuterium at C-4 wasincubated with
ANS, the label wasconserved in the cyanidin product.
Thisdemonstratesthat C-4 oxidation of LCDsdoesnot
occur during ANS catalysed cyanidin formation.16
Crystallographic work has implied that direct oxidation
15
at C-2 isunlikely.
Thus, oxidation of 2R,3S,4S-cis-
LCD 7 or 2R,3S,4R-trans-LCD 8 at C-3 to give either
flav-3-on-4-ols 15/16, respectively, or flavan-3,3,4-triols
17/18, respectively, is the most probable mechanism of
reaction.
For the formation of trans-DHQ 11 from incubationsof
ANS with 2R,3S-LCD 7/8 substrates the relative ste-
reochemistry at the C-2 and C-3 positions is conserved.
Formation of thisproduct could thusoccur via C-4
oxidation. However, since formation of cyanidin 6 and
cis-DHQ 10 (see below) almost certainly occurs via
reaction at C-3, formation of trans-DHQ 11 via reaction
at C-3 seems likely.
The elimination of water from flavan-3,3,4-triols 17/18 is
favoured in a stereo-electronic sense by a diaxial arrange-
ment of the newly introduced C-3 hydroxyl group and the
hydrogen involved in elimination. Tautomerisation of
flav-3-on-4-ols 15/16 isismilarly favoured by involve-
ment of a pseudo-axial hydrogen by allowing maximal

3856
J. J. Turnbull et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3853–3857
overlap with the adjacent p orbital of the carbonyl
function.
The preferred formation of quercetin 9 and low levelsof
two-electron oxidation productsobserved in vitro from
incubationsof 2 R,3S,4S-cis-LCD 7 may reflect the lack
of release (or very tight binding) of a 4S-flav-2-en-3,4-
diol intermediate 19. The presence of a trapped ‘pro-
duct’ in the active site may also be related to the lower
specific activity for 2R,3S,4S-cis-LCD 7 compared to
2R,3S,4R-trans-LCD 8 in vitro with the 2OG turnover
assay. The possibility of two substrate binding sites is
supported by crystallographic analyses.15 In vivo, how-
ever, it isclear that ANS together with the final enzyme
of anthocyanin 1 biosynthesis, a UDP-glucose:flavonoid
3O-glycosyltransferase (FGT) mediate anthocyanin 1
formation. It may be that ANS and FGT interact to
allow release of a 4S-flav-2-en-3,4-diol 19, which isthe
substrate for FGT (Fig. 1). This is supported by studies
Analysis of the structures of the flav-3-on-4-ols 15/16
and the flavan-3,3,4-triols 17/18 modelled into the ANS
active site implied that the C-2 hydrogen of all these
intermediatesis pseudo-axial (Fig. 6). However, the C-4
hydrogen is pseudo-equatorial for the intermediates
from 2R,3S,4S-cis-LCD 7, and pseudo-axial for the
intermediatesfrom 2 R,3S,4R-trans-LCD 8. Thus, for-
mation of a 4S-flav-2-en-3,4-diol 19 from the initial C-3
oxidation product of 2R,3S,4S-cis-LCD 7 islikely to be
kinetically favoured, over formation of a 2R-flav-3-en-
3,4-diol 21 (Fig. 7). In contrast, on C-3 oxidation of
2R,3S,4R-trans-LCD 8 formation of both a 4R-flav-2-
en-3,4-diol 20 and a 2R-flav-3-en-3,4-diol 21 may be
considered favourable (Fig. 8).
Formation of cyanidin 6 from 2R,3S-LCDsinvolving
loss of the C-4 hydroxyl group from a flav-2-en-3,4-diol
19/20 intermediate has been suggested.5,6,10 The current
work is consistent with this proposal. However, in vitro,
quercetin 9 isthe major product from 2 R,3S,4S-cis-
LCD 7. A further catalytic cycle, involving C-3 or C-4
oxidation of the intermediate 19 could lead to the major
quercetin 9 product (Fig. 7). Modelling studies on the
active site of ANS with flav-2-en-3,4-diols (both 4S-19
and 4R-20) indicated that the C-4 hydrogen of the 4S-
flav-2-en-3,4-diol 19 ismore accesible to C-4 oxidation
than that of the 4R-flav-2-en-3,4-diol 20. Thisdifference
may rationalise the increased levels of quercetin 9
observed in incubations of 2R,3S,4S-cis-LCD 7 relative
to 2R,3S,4R-trans-LCD 8 (see below).
The increased likelihood of 2R-flav-3-en-3,4-diol 21
formation from oxidation of trans-LCD 8/13 substrates
can explain why these substrates give more DHQ 10/11
productsthan observed with 2 R,3S,4S-cis-LCD 7 (Fig. 8).
Tautomerisation in the active site results in preferential
(at least) formation of cis-DHQ 10, possibly due to liga-
tion of the C-3 hydroxyl group to the Fe(II) centre (the
cis/trans equilibrium position is ca. 1:9 in solution).14
Figure 6. Modelsof the flav-3-on-4-ols 15/16 and 2R,4S-flavan-3,3,4-
triol 17 of LCD in the active site of ANS. Torsion angles for the flav-3-
on-4-olsare from the appropriate H to C-2/C-4, C-3, O.
Figure 7. In vitro formation of quercetin 9 by ANS and in vivo for-
mation of cyanin by ANS and FGT.

J. J. Turnbull et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3853–3857
3857
showing that the enzymes of flavonoid biosynthesis
form a membrane bound multi-enzyme complex.17ꢀ21
Acknowledgements
We would like to thank The Biotechnology and Biolo-
gical Sciences Research Council for financial support.
References and Notes
1. Bohm, B. A. Introduction to Flavonoids; Harwood Aca-
demic, 1998.
2. Forkmann, G., Heller, W. In Comprehensive Natural Pro-
duct Chemistry, 1999, p 713.
3. Jez, J. M.; Noel, J. P. J. Biol. Chem. 2000, 275, 39640.
4. Jez, J. M.; Noel, J. P. J. Biol. Chem. 2002, 277, 1361.
5. Nakajima, J.; Tanaka, Y.; Yakushiji, H.; Saito, K. J. Biol.
Chem. 2001.
6. Saito, K.; Yamazaki, M. New Phytol. 2002, 155, 9.
7. Schofield, C. J.; Zhang, Z.-H. Curr. Opin. Struct. Biol.
1999, 9, 722.
8. Prescott, A. G.; Lloyd, M. D. Nat. Prod. Rep. 2000, 17, 367.
9. Heller, W.; Forkmann, G. In The Flavonoids; Harborne, J.
B. Ed.; Chapmand and Hall: London, 1988, p 400.
10. Heller, W.; Forkmann, G. In The Flavonoids, Advances in
Research since 1986, Harborne, J. B., Ed.; Chapman and Hall:
London, 1993.
11. Saito, K.; Kobayshi, M.; Gong, Z. Z.; Tanaka, Y.;
Yamazaki, M. Plant J. 1999, 17, 181.
12. Turnbull, J. J.; Sobey, W. J.; Aplin, R. T.; Hassan, A.;
Firmin, J. L.; Schofield, C. J.; Prescott, A. G. Chem. Commun.
2000, 2473.
13. Stafford, H. A.; Lester, H. H.; Porter, L. J. Phytochem-
istry 1985, 24, 333.
14. Kiehlmann, E.; Li, E. P. M. J. Nat. Prod. 1995, 58, 450.
15. Wilmouth, R. C.; Turnbull, J. J.; Welford, R. W. D.;
Clifton, I. J.; Prescott, A. G.; Schofield, C. J. Structure 2002,
10, 93.
16. Welford, R. W. D.; Turnbull, J. J.; Claridge, T. D. W.;
Prescott, A. G.; Schofield, C. J. Chem. Commun. 2001, 1828.
17. Stafford, H. A. Rec. Adv. Phytochem 1974, 8, 53.
18. Stafford, H. A. Plant Physiol. 1991, 96, 680.
19. Winkel-Shirley, B. Physiologia Planatarium 1999, 107,
142.
20. Winkel-Shirley, B. Plant Physiol. 2001, 126, 485.
21. Burbulis, I. E.; Winkel-Shirley, B. P. Natl. Acad. Sci. USA
1999, 96, 12929.
Figure 8. Mechanism of product formation of ANS mediated oxida-
tion of 2R,3S,4R-trans-LCD 8. The 4R-flav-2-en-3,4-diol 20 may also
be released from the active site of ANS and tautomerise to give a ther-
modynamic distribution of DHQ, with trans-DHQ predominating.