Flavonol 3-O-glycoside hydroxycinnamoyltransferases
from Scots pine (Pinus sylvestris L.)
Florian Kaffarnik
1,
*, Werner Heller
1
, Norbert Hertkorn
2
and Heinrich Sandermann Jr
1
1 Institute of Biochemical Plant Pathology, GSF-Research Center for Environment and Health, Neuherberg, Germany
2 Institute of Ecological Chemistry, GSF-Research Center for Environment and Health, Neuherberg, Germany
Several plant hydroxycinnamoyltransferases (HCTs)
have been described for the biosynthesis of function-
ally important secondary metabolites, e.g. phytoalexins
[1–3] or flower pigments [4–7]. They most commonly
use CoA esters as activated donor substrates [8] and
transfer the hydroxycinnamoyl moiety to a hydroxyl
or amino group of acceptor substrates. Other donor
substrates such as glucosyl esters are occasionally
observed [9]. Acceptor substrates include anthocyanin
glycosides [4–7,10], a flavonol 3-O-glycoside [11],
amines [1–3,12–15], meso-tartrate, shikimate and qui-
nate [16,17], fatty acids [18] or alkaloids [19] (for an
overview see [20]). The biochemistry of HCTs using
anthocyanin glycosides, or amines such as agmatine,
tyramine or anthranilate, as acceptor substrates has
been investigated in more detail [1,3,4,10,12,14,21].
Genes encoding N-HCTs acting on amines such as tyr-
amine, noradrenaline and serotonin [22,23] as well as
O-HCTs acting on anthocyanins [6,7,10] have recently

thesis of these compounds are catalyzed by enzymes that transfer the acyl
part of hydroxycinnamic acid CoA esters to flavonol 3-O-glucosides. A
newly developed enzyme assay revealed three flavonol 3-O -glucoside
hydroxycinnamoyltransferases (HCTs) in Scots pine needles with specifici-
ties for positions 3¢¢,4¢¢ or 6¢¢. The positions of the acyl groups were identi-
fied by cochromatography with reference compounds and by NMR
spectroscopy. The enzymes were characterized by molecular mass, isoelec-
tric point, and also pH and temperature optima. Substrate specificities for
flavonol glycosides and hydroxycinnamic acid CoA esters as well as kinetic
properties of 3¢¢- and 6¢¢HCT suggested that acylation preferably occurs
with glucosides and p-coumaroyl-CoA. In addition, acylation takes place in
a well-defined order, beginning at position 6¢¢ followed by acylation at posi-
tion 3¢¢. These results give the first detailed characterization of flavonol
3-O-glycoside HCTs involved in the protection of plant tissues against
UV-B (280–315 nm) radiation.
Abbreviations
HCT, hydroxycinnamoyl-CoA flavonol 3-O-glucoside hydroxycinnamoyltransferase (EC 2.3.1 ); I3G, isorhamnetin 3-O-glucoside; K3G,
kaempferol 3-O-glucoside; Q3G, quercetin 3-O-glucoside.
FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS 1415
is mediated by hydroxycinnamoyltransferase enzymes.
These steps would introduce p-coumaric and ferulic
acid residues at position 3¢¢ and ⁄ or 6¢¢, respectively,
of flavonol 3-O-glucosides. In the Scots pine,
3-O-glycosides of three different flavonol types, namely
kaempferol, isorhamnetin and quercetin, have been
detected. Interestingly, similar diacylated flavonol 3-O-
glycosides are not only found in coniferous leaves but
also in the leaves of broadleaf trees, such as oak spe-
cies [27,29]. This suggests that these metabolites may
play an important role in UV-B screening in a variety

Scots pine needles were assayed with isorhamnetin 3-O-glucoside
(A, upper panel; I3G) and 6¢¢-p-coumaroyl-kaempferol 3-O-glucoside
(A, lower panel; tiliroside) as the acceptor and p-coumaroyl-CoA as
the donor substrates. I3G was chosen as the nonacylated substrate
with crude cell extracts because the respective product 1 with K3G
comigrated with a minor nonflavonoid hydroxycinnamoyl by-product
which prevented quantification of 6¢¢-p-coumaroyl-kaempferol
3-O-glucoside. S, substrate; IS, internal standard; pinosylvin methyl
ether.The peaks marked 1–6 were identified as acylated products
by their diode array spectra (B). Characteristics of the UV spectra
of acylated compounds are the absorption maximum at 315 nm
due to the hydroxycinnamic acid moieties and shoulders at 270 and
350 nm originating from the flavonol 3-O-glycoside [26]. Differences
between monoacylated (1–3) and diacylated compounds (4–6) con-
sist in a higher proportion of the absorbances at 315 nm and
350 nm for diacylated compared to monoacylated compounds.
Spectra shown are normalized at 315 nm.
Hydroxycinnamoyltransferases from Scots pine F. Kaffarnik et al.
1416 FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS
for simple nonflavonoid p-coumaric acid derivatives.
K3G and I3G gave similar results (data not shown)
but one of the p-coumaric acid related by-products
detected in the assays with I3G as substrate comigrat-
ed with tiliroside, one of the reaction products of K3G
(Fig. 2A). Using tiliroside as the acceptor substrate
one minor and one major diacylated product were
detected (Fig. 2A, lower panel, peak 5 and 6). In con-
trol assays with heat-inactivated enzyme preparations,
or omitting either donor or acceptor substrate, none of
the expected acylated products was obtained. Diode

different monoacylated and two diacylated products of
flavonol 3-O-glucosides were formed by HCT activities
in crude cell extracts from Scots pine needles. Coinjec-
tion experiments of authentic 6¢¢-p-coumaroyl K3G
standard (tiliroside) allowed the identification of the
respective enzyme product observed in the chromato-
grams.
Separation of HCT activities
The formation of several products in assays of crude
cell extracts raised the question of whether different
enzymes were involved. The separation of enzyme
activities was successfully carried out by anion
exchange chromatography of protein on Q-Sepharose
after ammonium sulfate precipitation. The fractions
eluting from the ion exchange column were tested with
both K3G and tiliroside as substrates. Three separate
activities were detected with K3G (Fig. 3A; peaks
I–III). Peak I represents the protein fraction not
retained by the column and gave a product corres-
A
B
Fig. 3. Separation of hydroxycinnamoyltransferase activities by anion exchange chromatography on Q-Sepharose. Protein extracted from
Scots pine needles was chromatographed on Q-Sepharose after ammonium sulfate precipitation. HCT activities of collected fractions were
determined with K3G (A) and tiliroside (B) as substrates. Using K3G three different HCT activities (A, I–III) were separated giving products
that corresponded to compound 1 (d), compound 2 (
) and compound 3 (.) in Fig. 2A. In contrast, only two activities were detected with
tiliroside (B, IV and V) giving compounds 5 (
)and6(.), respectively, in Fig. 2A. The solid line represents protein concentration, measured
as absorption at 280 nm. The dotted line shows changes in conductivity, caused by the sodium chloride gradient applied.
F. Kaffarnik et al. Hydroxycinnamoyltransferases from Scots pine

H-
COSY-NMR spectroscopy.
The product corresponding to compound 2 of Fig. 2
showed a chemical shift for H-4¢¢ of 4.79 p.p.m. com-
pared to 3.23 p.p.m. of the nonacylated K3G, indica-
ting that the acyl group was at position 4 of the
glucose molecule. On the other hand, the product cor-
responding to compound 3 of Fig. 2 showed a chem-
ical shift for H-3¢¢ of 5.02 p.p.m. compared to
3.34 p.p.m. of K3G, and was therefore acylated at
position 3 of the glucose molecule. The NMR data
(see Experimental procedures for details) combined
with the results of cochromatography thus proved the
existence of three separate position-specific enzymes,
i.e. 3¢¢-, 4¢¢- and 6¢¢HCT, in Scots pine needles. Both
3¢¢- and 4¢¢HCT convert nonacylated flavonol
3-O-glucosides in addition to the 6¢¢-monoacylated
tiliroside, giving the respective monoacylated 3¢¢- and
4¢¢-p-coumaroyl flavonol 3-O-glucoside (Fig. 2; com-
pounds 3 and 2), and diacylated 3¢¢,6¢¢- and 4¢¢,6¢¢-di-p-
coumaroyl K3G (Fig. 2; compounds 6 and 5). The
simultaneous presence of 3¢¢HCT and 6¢¢HCT in crude
cell extracts directly gave rise to diacylated products of
flavonol 3-O-glucosides, e.g. compound 4 in Fig. 2.
Consistently, this is in agreement with the acylation
pattern found in Scots pine, where p-coumaric and
ferulic acids were identified at positions 3¢¢ and 6¢¢ of
flavonol 3-O-glucosides [28]. The discovery of products
acylated at position 4¢¢ was somewhat surprising,
because no corresponding metabolites have been des-

acyltransferases [3].
Isoelectric points of the partially purified proteins
were determined by chromatofocusing on a Mono-P
column. Both 3¢¢- and 4¢¢HCT had a pI of 4.7, whereas
6¢¢HCT appeared at pI 7.9. Maximal activities were
determined for both 4¢¢- and 6¢¢HCT at pH 8 and
44 °C. Half maximal values for 4¢¢HCT were obtained
at pH 6.8 and 8.5, and at 36 and 50 °C. For 6¢¢HCT,
half maximal values were at pH 6.5 and 9.2, and at 36
and 52 °C. Maximal activity for 3¢¢HCT was at pH 7
and 40 °C and half maximal values were at pH 6.2
and 8.0, and at 28 and 47 °C.
Kinetic parameters of 3¢¢- and 6¢¢HCT
Partially purified 3¢¢- and 6¢¢ HCT, the two major
HCT activities in Scots pine needles, were tested for
their kinetic parameters with p-coumaroyl- and feru-
loyl-CoA as donor substrates and K3G, tiliroside
Hydroxycinnamoyltransferases from Scots pine F. Kaffarnik et al.
1418 FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS
and 3¢¢-p-coumaroyl K3G as acceptor substrates
(Table 1). Using enzyme preparations after anion
exchange chromatography on Q-Sepharose 3¢¢HCT
showed a distinctly lower apparent K
m
value with
p-coumaroyl-CoA than with feruloyl-CoA, whereas
6¢¢HCT has comparable apparent K
m
values for both
CoA esters. This is consistent with the observation

[31].
Substrate specificity
To test the substrate specificity of 3¢¢- and 6¢¢HCT,
a number of flavonol 3-O-glycosides were analysed
(Table 2). Variation of the B-ring substitution pattern
of the flavonol had minor but distinct influence on the
transferase activities. 3¢¢HCT showed higher activity
with kaempferol and isorhamnetin 3-O-glucosides
with a more lipophilic B-ring compared to quercetin
Table 1. Apparent Michaelis–Menten parameters of 3¢¢-and
6¢¢HCT. The apparent Michaelis–Menten parameters were deter-
mined using enzyme preparations from anion exchange chromato-
graphy on Q-Sepharose which fully separated the HCT activities
(Fig. 3). n.d., not detectable.
Enzyme Substrate
K
m
(lM)
V
max
(lkatÆkg
)1
)
V
max
⁄ K
m
(katÆM
)1
Ækg

fixed substrate.
c
100 lM K3G as fixed substrate.
Fig. 4. Suggested sequential acylation of
flavonol 3-O-glucosides. The K
m
values of
3¢¢HCT indicated a higher affinity to tiliroside
(16 l
M) than to K3G (47 lM). While 6¢¢HCT
did not acylate 3¢¢-monocoumaroylated K3G
at position 6¢¢, the K
m
value for K3G (22 lM)
was in the same range as that of 3¢¢HCT for
tiliroside. This indicates a sequential acyla-
tion of flavonol 3-O-glucosides, first at posi-
tion 6¢ followed by acylation at position 3¢¢.
C, p-coumaroyl; F, feruloyl.
F. Kaffarnik et al. Hydroxycinnamoyltransferases from Scots pine
FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS 1419
3-O-glucoside. In contrast, 6¢¢HCT preferred a more
polar B-ring of the substrate showing the highest activ-
ity with quercetin 3-O-glucoside.
Comparing different quercetin 3-O-glycosides
revealed high specificity towards glucose for both
enzymes (Table 2). For 3¢¢HCT the hydroxyl group at
position 4¢¢ clearly influences activity. The 3-O-b-d-gal-
actoside with axial configuration exhibited only 18%
activity under standard assay conditions compared to

se
(Lyon, France). CoA esters of p-coumaric and ferulic acids
were essentially synthesized according to a published method
[32]. The products (0.12 mmol) were purified using a Fracto-
gel EMD DEAE 650 (S) column (gel bed 12 mL) (Merck,
Darmstadt, Germany) and an A
¨
KTA Explorer system
(Amersham Biosciences, Freiburg, Germany). The solvents
used were 0.1 m formic acid (A) and 1.5 m sodium formate
(B). After application of the crude reaction product
( 0.25 mmol in 10 mL) the column was washed with 50 mL
of solvent A. A gradient from 0 to 100% B in a total volume
of 110 mL was then applied, followed by 320 mL solvent B.
Fractions showing appropriate UV spectra (maxima at 259
and 334 nm for p-coumaroyl-CoA, 256 and 346 nm for feru-
loyl-CoA) were collected, pooled and desalted on a Dowex
50 WX 8 column (Aldrich, Steinheim, Germany). Other
chemicals used were of highest available purity and were pur-
chased from Sigma (Steinheim, Germany).
Protein determination
Protein concentration was measured according to the
method of Bradford [33] using bovine serum albumin
(BSA) as standard.
Protein extraction
Analytical scale
Approximately 100 mg of needle material from seedlings
or pine trees, shock frozen in liquid nitrogen, was coarsely
homogenized with pestle and mortar. Fifty milligrams
poly(vinylpolypyrrolidone) (PVPP) and 3 mg Celite were

–OH 67 100
Isorhamnetin 3-O-Glc –OMe 101 52
Quercetin 3-O-glycosides
Quercetin 3-O-b-
D-glucopyranoside
(isoquercitrin)
–OH 100 100
Quercetin 3-O-b-
D-galactopyranoside
(hyperoside)
–OH 18 52
Quercetin 3-O-a-
L-arabinopyranoside
(guaijaverin)
–OH 15 0
Quercetin 3-O-a-
L-rhamnopyranoside
(quercitrin)
–OH 0 0
Hydroxycinnamoyltransferases from Scots pine F. Kaffarnik et al.
1420 FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS
Preparative scale
Approximately 1700 g of needle material was harvested
from field-grown trees at the time of highest specific activity
(June and July), immediately frozen in liquid nitrogen and
ground with a pestle and mortar. After lyophilization for
48 h, the dried material was ground for 3 min at 4 °Cinan
analysis mill A10 (IKA Labortechnik, Staufen, Germany)
and stored at )80 °C. Cell extracts were prepared on ice
from 25 to 30 g needle powder, 60 g PVPP and 4 g Celite

20 000 g for 5 min.
Partially purified fractions
The total assay volume was 100 lL in 100 mm sodium
phosphate, 5 mm DTE, pH 6.8. The final substrate concen-
trations and test procedure were as described above.
Protein concentration and desalting
All steps were carried out at 4 °C or on ice. The crude
cell extract was fractionated by ammonium sulfate pre-
cipitation (25–60% saturation). After centrifugation at
30 000 g for 30 min, an upper layer was formed, contain-
ing the protein and PEG 1450 [35]. The protein–PEG-
phase was separated by filtration through Miracloth and
dilution into buffer A [20 mm Tris ⁄ HCl buffer, pH 7.5
containing 10% (v ⁄ v) glycerol, 1 mm DTE and 1 mm
EDTA]. Desalting was performed using Sephadex G-25
(Amersham Biosciences).
Anion exchange chromatography
A 64 mL Q-Sepharose fast flow column (Amersham Bio-
sciences) was pre-equilibrated with buffer A. The concentra-
ted extract (190 mL) was loaded onto the column, and
after washing with two column volumes of the same buffer,
the enzyme was eluted with a gradient from 0 to 0.5 m
NaCl in five column volumes at a flow rate of 7.5
mLÆmin
)1
. Fractions of 10 mL were collected and assayed
for HCT activity and protein concentration.
Gel filtration chromatography
A Superose 6 HR 10 ⁄ 30 column (Amersham Biosciences)
was pre-equilibrated with a buffer containing 100 mm

(v ⁄ v) glycerol and 1 mm DTE or 25 mm diethanolam-
ine ⁄ HCl (pH 9.5), 10% (v ⁄ v) glycerol and 1 mm DTE for
determination of the isoelectric point of 3¢¢- and 4¢¢HCT or
6¢¢HCT, respectively. The pH gradient was generated in the
column during the passage of a solution of Polybuffer 74
(1 : 10, pH 4.0) or Polybuffer 96 (1 : 10, pH 6.0) with 10%
(v ⁄ v) glycerol and 1 mm DTE. The flow rate was 0.5
mLÆmin
)1
, and fractions of 0.5 or 0.8 mL were collected
and assayed for HCT activity and protein concentration.
F. Kaffarnik et al. Hydroxycinnamoyltransferases from Scots pine
FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS 1421
Enzyme characterization
The characterization of HCT activities was performed with
partially purified enzyme preparations after anion
exchange chromatography. All measurements were per-
formed as triplicates. For determination of pH-depend-
ence, enzyme preparations were buffer-exchanged with
NAP-5 columns (Amersham Biosciences) in 100 mm
sodium phosphate, pH 6.5–8.5 (3¢¢- and 4¢¢HCT) or
100 mm sodium phosphate, pH 6.0–8.0 and 50 mm
Tris ⁄ HCl, pH 7.0–9.5 (6¢¢HCT). For determination of the
kinetic parameters K
m
and V
max
the following substrate
concentrations were used: p-coumaroyl- and feruloyl-CoA
10–450 lm with fixed acceptor concentrations of 100 lm,

at 314 nm. Appropriate peaks were manually collected and
identified by analytical HPLC. For comparison, K3G, til-
iroside and 2¢¢,6¢¢p-di-coumaroyl kaempferol 3-O-glucoside
were measured as reference substances.
1
H NMR spectra
were acquired with a Bruker DMX 500 NMR spectrometer
(Rheinstetten, Germany) operating at 500.13 MHz proton
frequency from a few mg of sample in 750 lLCD
3
CN
(d
1
H ¼ 1.93 p.p.m.) usually at 303 K with 90 deg pulses
[90°(
1
H) ¼ 9.3 ls], acquisition time of 3.2 s and a relaxa-
tion delay of 7 s. Gradient enhanced (length, 1 ms; recov-
ery, 450 ls), absolute value 2Q-COSY NMR spectra were
acquired with aq ¼ 234 ms and 470 increments in F1 at a
sweep width of 4370 Hz.
4¢¢-p-coumaroyl kaempferol 3-O-glucoside (analogue
to compound 2 in Fig. 2)
1
H-NMR (500 MHz, CD
3
CN, 273 K, c. 150 lg): d ¼ 8.09
(2H, AA¢; H-2¢⁄6¢), d ¼ 7.64 (H, d; H-7¢¢¢), d ¼ 7.50 (2H,
AA¢; H-2¢¢¢ ⁄ 6¢¢¢), d ¼ 6.94 (2H, XX¢; H-3¢⁄5¢), d ¼ 6.82
(2H, XX¢; H-3¢¢¢ ⁄ 5¢¢¢), d ¼ 6.47 (H, d; H-8), d ¼ 6.32 (H, d;

ance and Giovanni Romussi, Genova, for providing a
sample of 2¢¢,6¢¢-di-p-coumaroyl kaempferol 3-O-glu-
coside.
References
1 Hohlfeld H, Schu
¨
rmann W, Scheel D & Strack D
(1995) Partial purification and characterization of
hydroxycinnamoyl-Coenzyme A: tyramine hydroxycin-
namoyltransferase from cell suspension cultures of Sola-
num tuberosum. Plant Physiol 107, 545–552.
2 Matsukawa T, Isobe T, Ishihara A & Iwamura H
(2000) Occurence of avenanthramides and hydroxycin-
namoyl-CoA: hydroxyanthranilate N-hydroxycinna-
Hydroxycinnamoyltransferases from Scots pine F. Kaffarnik et al.
1422 FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS
moyltransferase activity in oat seeds. Z Naturforsch 55c,
30–36.
3 Yang Q, Reinhard K, Schiltz E & Matern U (1997)
Characterization and heterologous expression of
hydroxycinnamoyl ⁄ benzoyl-CoA: anthranilate N-hydro-
xycinnamoyl ⁄ benzoyltransferase from elicited cell
cultures of carnation, Dianthus caryophyllus L. Plant
Mol Biol 35, 777–789.
4 Callebaut A, Terahara N & Decleire M (1996) Antho-
cyanin acyltransferases in cell cultures of Ajuga reptans.
Plant Sci 118, 109–118.
5 Kamsteeg J, van Brederode J, Hommels CH & van Nig-
tevecht G (1980) Identification, properties and genetic
control of hydroxycinnamoyl-Coenzyme A: anthocyani-

Coenzyme A transferase involved in the biosynthesis of
kaempferol-3-(p-coumaroyl triglucoside) in Pisum sati-
vum. Z Naturforsch 32c, 765–768.
12 Burhenne K, Kristensen BK & Rasmussen SK (2003) A
new class of N-hydroxycinnamoyltransferases. Purifica-
tion, cloning, and expression of a barley agmatine cou-
maroyltransferase (EC 2.3.1.64). J Biol Chem 278,
13919–13927.
13 Ishihara A, Miyagawa H, Matsukawa T, Ueno T,
Mayama S & Iwamura H (1998) Induction of
hydroxyanthranilate hydroxycinnamoyl transferase
activity by oligo-N-acetylchitooligosaccharides in oates.
Phytochemistry 47, 969–974.
14 Yu M & Facchini PJ (1999) Purification, characteriza-
tion, and immunolocalisation of hydroxycinnamoyl-
CoA: tyramine N-(hydroxycinnamoyl) transferase from
opium poppy. Planta 209, 33–44.
15 Hedberg C, Hesse M & Werner C (1996) Spermine and
spermidine hydroxycinnamoyl transferase in Aphelandra
tetragona. Plant Sci 113, 149–156.
16 Hoffmann L, Maury S, Martz F, Geoffroy P & Legrand
M (2003) Purification, cloning, and properties of an
acyltransferase controlling shikimate and quinate ester
intermediates in phenylpropanoid metabolism. J Biol
Chem 278, 95–103.
17 Hohlfeld M, Veit M & Strack D (1996) Hydroxycinna-
moyltransferases involved in the accumulation of caffeic
acid esters in gametophytes and sporophytes of Equise-
tum arvense. Plant Physiol 111, 1153–1159.
18 Lotfy S, Javelle F & Negrel J (1996) Purification and

24 Schnitzler J-P, Jungblut TP, Heller W, Ko
¨
fferlein M,
Hutzler P, Heinzmann U, Schmelzer E, Ernst D, Lange-
bartels C & Sandermann H (1996) Tissue localization of
UV-B-screening pigments and of chalcone synthase
mRNA in needles of Scots pine seedlings. New Phytol
132, 247–258.
F. Kaffarnik et al. Hydroxycinnamoyltransferases from Scots pine
FEBS Journal 272 (2005) 1415–1424 ª 2005 FEBS 1423
25 Fischbach R, Kossmann B, Panten H, Steinbrecher R,
Heller W, Seidlitz HK, Sandermann H, Hertkorn N &
Schnitzler J-P (1999) Seasonal accumulation of ultravio-
let-B screening pigments in needles of Norway spruce
(Picea abies (L.) Karst.). Plant Cell Environ 22, 27–37.
26 Jungblut TP, Schnitzler J-P, Heller W, Hertkorn N,
Metzger JW, Szymczak W & Sandermann H (1995)
Structures of UV-B induced sunscreen pigments of the
Scots pine (Pinus sylvestris L.). Angew Chem Int Ed
English 34, 312–314.
27 Romussi G, Parodi B & Caviglioli G (1991) Flavonoid-
glycoside aus Quercus pubescens Willd., Quercus cerris
L. und Quercus ilex L. Pharmazie 46, 679.
28 Jungblut TP (1996) Wirkung von UV-B Strahlung und
Ozon auf den Sekunda
¨
rstoffwechsel der Kiefer (Pinus
sylvestris L.). PhD Thesis, Ludwig-Maximilians-Univer-
sity, Munich, Germany.
29 Romussi G, Cafaggi S & Ciarallo G (1983) Ein neues