Emission of 2-phenylethanol from its β-D-glucopyranoside and the biogenesis of these compounds from [2H8] L-phenylalanine in rose flowers

Article abstract of DOI:10.1016/j.tet.2003.10.130

The hydrolysis of 2-phenylethyl β-D-glucopyranoside (3) was found to be partially inhibited by feeding with 2-phenyl-N-glucosyl-acetamidiumbromide (8), a β-glucosidase inhibitor, resulting in a decrease in the diurnal emission of 2-phenylethanol (2) from

Full text of DOI:10.1016/j.tet.2003.10.130

Tetrahedron 60 (2004) 7005–7013
Emission of 2-phenylethanol from its b-D-glucopyranoside
and the biogenesis of these compounds from [2H8] L-phenylalanine
in rose flowers
Shunsuke Hayashi,a Kensuke Yagi,a Takashi Ishikawa,a Miwa Kawasaki,a Tatsuo Asai,a
Joanne Picone,b Colin Turnbull,b Jun Hiratake,c Kanzo Sakata,c Masayasu Takada,d Koji Ogawad
and Naoharu Watanabea
,
*
aFaculty of Agriculture, Shizuoka University, 836 Ohya, Shizuoka 422-8529, Japan
bDepartment of Agricultural Sciences, Imperial College at Wye, University of London, Wye, Ashford, Kent TN25 5AH, UK
cInstitute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011, Japan
dNihon Shokuhin Kako Co., Ltd., 30 Tajima, Fuji, Shizuoka 417-8530, Japan
Received 5 August 2003; revised 26 October 2003; accepted 26 October 2003
Available online 25 June 2004
Abstract—The hydrolysis of 2-phenylethyl b-D-glucopyranoside (3) was found to be partially inhibited by feeding with 2-phenyl-N-
glucosyl-acetamidiumbromide (8), a b-glucosidase0inhibitor,0resulting in a decrease 0in0the d0iurnal emission of 2-phenylethanol (2) from Rosa
damascena Mill. flowers. Detection of [1,1,2,2 ,30,40,50,6 -2H8]-2 and [1,2,20,3 ,4 ,50,6 -2H7]-2 from R. ‘Hoh-Jun’ flowers fed with
[1,1,2,20,30,40,50,60-2H8]-3 suggested that b-glucosidase, alcohol dehydrogenase, and reductase might be involved in scent emission.
Comprehensive GC-SIM analyses revealed that [1,2,2,20,30,40,50,60-2H8]-2 and [1,2,2,20,30,40,50,60-2H8]-3 must be biosynthesized from
[1,2,2,20,30,40,5060-2H8] L-phenylalanine ([2H8]-1) with a retention of the deuterium atom at a-position of [2H8]-1.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
unfurling.14 In Narcissus flowers, b-glucosidase activity
was found to be involved in the scent emission detected by
gas sensors.15 To date, attempts to correlate concentrations
of glucosides with rhythmic cycles in volatile emission have
been unsuccessful.16,17 Synchronous diurnal rhythms have
been detected in the concentration of both 2 and its acetate,
but not in the concentration of 3, in the Trifolium repens
flowers.17
Floral scent compounds, such as monoterpenoids, shikimate
derived compounds, and fatty acids-derived compounds, are
synthesized in some flowers. These compounds are emitted
to attract pollinators,1–7 to attract or to deter herbivores,8
and/or to prevent bacterial and fungal infections.
2-Phenylethanol (2) is one of the dominant floral scent
compounds emitted from roses such as Rosa damascena spp
and R. ‘Hoh-Jun’. Rhythmic cycles in the emission of 2 are
observed in R. ‘Hoh-Jun’ as shown in Figure 2, although
R. damascena Mill. transiently emits 2 (cf. Fig. 5). It has been
proposed that glycosidically bound volatiles act as aroma
precursors,9–12 and their hydrolysis may be responsible for
the rhythmic emission pattern observed in R. ‘Hoh-Jun’.13
In R. damascena Mill. flowers, the amount of 2-phenylethyl
b-D-glucoside (3) is greater during the stages of early
development and decreases during the unfurling process,
whilst levels of the free form generally increased over the
Although there has been much progress in elucidating
biosynthetic pathways and characterizing enzymes, the
underlying biosynthetic route of 2 in flowers has yet to be
fully understood. We recently reported the biogenesis of 2
from L-phenylalanine (1) in rose flowers.18 In that study,
[1,2,2,20,30,40,50,60-2H8]-1 ([2H8]-1) was fed to R. ‘Hoh-Jun’
and R. damascena Mill. flowers throughout maturation, and
feeding was terminated at full bloom stage. On the basis of
GC-MS analyses, we found that [2H8]-1 was incorporated
into both 2 and 3 when the flowers were fed until full bloom.
In both rose species, the labeling pattern of 2 was almost
identical to that of 3, and indicated the presence of both
[2H7]- and [2H8]-2, and [2H7]- and [2H8]-3. This may be
ascribed to the equilibrium established between 2 and 3.
Although the labeling pattern for 2 and 3 established that
these compounds were produced from 1 via several routes as
Keywords: Rose flower; L-Phenylalanine; Biogenesis; b-Glucosidase; 2-
Phenylethanol; 2-Phenylethyl b-D-glucopyranoside.
*
Corresponding author. Tel.:/fax: þ81-54-238-4870; e-mail address:
acnwata@agr.shizuoka.ac.jp
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.10.130

2
Figure 1. Hypothetical biogenetic pathway for 2-phenylethanol (2) and 2-phenylethyl b-D-glucopyranoside (3) from L-phenylalanine (1). Asterisk denotes the position of a H atom.

S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
7007
shown in Figure 1, the route involving phenylpyruvic acid
(4) was thought to be the major one. However, the detection
of [2H8]-2 and [2H8]-3 strongly suggested the presence of a
biosynthetic route involving the intermediate 2-phenylethy-
lamine (6).
(10 mM) was fed to R. damascena Mill. flowers at stage 3.
As shown in Figure 4b, the level of 3 gradually decreased by
7 mmol/flower in the control group, whereas the level in the
fed group was almost constant (16–14 mmol/flower) during
the dark period (6–18 h after the feeding commenced).
During the day-time (18–30 h after the feeding com-
menced) the level of 3 remained fairly constant (13–
17 mmol/flower) in the fed group, but decreased by 5 mmol/
flower in the control group. The concentration of 2 in the
tissue increased gradually to reach its maximal level
(17 mmol/flower) 24 h after the feeding commenced in the
control group, but the concentration did not exceed 5 mmol/
flower during the unfurling process in the fed group
(Fig. 4a).
In this paper, we describe the involvement of b-glucosidase
in the emission of 2, based on experime0nts using a b-
glucosidase inhibitor and feeding with [1,1,2,2 ,30,40,50,60-2H8]-
3, as well as details of the biosynthetic pathways of 2,
deduced from feeding with [2H8]-1, in R. damascena Mill.
and R. ‘Hoh-Jun’ flowers.
2. Results and discussion
In R. damascena Mill. flowers, the amount of 3 is greater in
early development and declines over the unfurling process,
whereas levels of the free form generally increased over the
unfurling process under natural conditions.14 As shown in
Figures 2 and 5, we found that 2 was emitted in higher
concentrations when exposed to light than when kept in the
dark at a constant temperature. Loughrin16 has suggested
that changes in the rate of the glycoside hydrolysis may be
responsible for the emission of scent compounds. We
therefore attempted to verify the involvement of b-
glucosidase and its substrate 3 in the production of 2 in R.
damascena Mill. flowers.
Figure 4. Effect of 2-phenyl-N-b-glucopyranosyl-acetamidiumbromide (8)
on the accumulation of 2-phenylethanol (2) (a) and 2-phenylethyl b-D-
glucopyranoside (3) (b) accumulated in the flowers of Rosa damascena
Mill. during the unfurling process. W, Control group; X, fed 8. Intact plants
were exposed to the same conditions as in Figure 2. Shaded and non-shaded
areas correspond to exposure to dark and light, respectively. p p p P,0.01,
p p P,0.05, pP,0.1 for paired comparisons of the control versus the fed
group. Stages of flowers are defined in the text.
Figure 2. 2-Phenylethanol (2) emission during the unfurling process in
Rosa ‘Hoh-Jun’ flowers of intact plants maintained under a 12-h photo
period at constant temperature (22 8C) and relative humidity (75%). Shaded
and non-shaded areas correspond to exposure to dark and light,
respectively. Stages of flowers are defined in the text.
We examined the effect of 2-phenyl-N-b-glucosyl-acetami-
diniumbromide (8, Fig. 3), a glucosidase inhibitor,19,20 on
the hydrolytic activity of crude enzymes prepared from
flower petals. The crude enzymes (prepared from 100 to
400 mg of flower petals) released 1.3 mmol of 2 in 1 h from
5 mmol of 3, and more than 95% of the hydrolytic activity
was inhibited by 0.6 mmol of 8. As each flower accumulated
15 mmol of 3 at stage 3, it was expected that more than
2.0 mmol of 8 would inhibit the hydrolysis of 3 in each
flower. Therefore, 300 ml of an aqueous solution of 8
As shown in Figure 5, diurnal emission reached its
maximum level at 21 h after feeding commenced in the
control group, whereas a peak was not observed in the fed
group. b-Glucosidase activity increased between stages 4
and 6 (Fig. 5). The increase in b-glucosidase activity was
observed prior to the increase in the concentration of 2 in the
control flowers. As 8 does not bind to b-glucosidase via a
covalent bond,20 similar level of b-glucosidase activity was
detected in the crude extract even in the fed group. Although
the high b-glucosidase activity at stage 9 cannot be
rationally explained, these observations strongly suggest
that b-glucosidase activity is involved in both the
production and the emission of 2. We therefore, focus our
research on the effect of the inhibitor, 8, on the rhythmic
emission of 2 observed in R. ‘Hoh-Jun’ and R. damascena
semperflorens quatra saisons21 flowers
Figure 3. 2-Phenyl-N-b-glucopyranosyl-acetamidiumbromide (8), a b-
glucosidase inhibitor.

7008
S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
Figure 5. Effect of 2-phenyl-N-b-glucopyranosyl-acetamidiumbromeide (8) on 2-phenylethanol (2) emission from Rosa damascena Mill. flowers, and changes
in b-glucosidase activity during the unfurling process R. damascena Mill. flowers. W, Control group; X, fed 8. Bar indicates b-glucosidase activity. Intact
plants were exposed to the same conditions as in Figure 2. Shaded and non-shaded areas correspond to exposure to dark and light, respectively. p p P,0.05,
pP,0.1 for paired comparisons of the control versus the fed group. Stages of flowers are defined in the text.
During the study, [1,1,2,20,30,40,50,60-2H8]-3 was fed to R.
‘Hoh-Jun’ 0flowers at stage 3 to confirm the emission of
[1,1,2,20,3 ,40,50,60-2H8]-2. As shown 0in Figure 6,
[1,1,2,20,30,40,50,60-2H8]-2 and [1,2,20,30,40,5 ,60-2H7]-2 were
detected0in0sa0mple0of the head-space gas, 0and the0ratio of the
[1,1,2,2 ,3 ,4 ,50,6 -2H8]-2 to [1,2,2 ,30,40,5 ,60-2H7]-2
decreased during the unfurling process until senescence
stage (between stages 3 and 8). This observation strongly
suggests that [1,1,2,20,30,40,50,60-2H8]-2, once produced
by endogenous b-glucosidase, is transformed to
[1,2,20,30,40,50,60-2H7]-phenylacetaldehyde (5) by alcohol
dehydrogenase, and subsequently interconverted into
[1,2,20,30,40,50,60-2H7]-2. In our preliminary experiments,
we detected alcohol dehydrogenase activities toward 2 in
the crude enzymes prepared from R. ‘Hoh-Jun’ petal tissues
(data not shown). This indicates that the oxidation–
reduction process is a significant factor involved in
regulating the production and emission of 2.
compounds, but also [2H8]-2 and [2H8]-3 were detected as
minor components in the flowers. This suggests that the
flowers can produce 2 and 3 from 1 via several pathways
from 1.0 As shown in Figure 6, one deuterium atom of
[1,2,20,3 ,40,50,60-2H8]-2 was lost during0 the oxidation–
reduction reaction yielding [2,20,30,40,5 ,60-2H7]-2. This
prompted us to reconsider the biosynthetic route from 1 to
2 during feeding with [2H8]-1, because [2H7]-2 and [2H7]-3
are likely to be derived from [2H8]-2 and [2H8]-3,
respectively. Below, we examined the effects of [2H8]-1
feeding over a short unfurling process in order to elucidate
the actual biosynthetic route of 2 by tracing changes in the
labeling patterns of [2Hn]-2 and [2Hn]-3.
Initially, we attempted to confirm the transformation of
[2H8]-1 to [2Hn]-2 and [2Hn]-3 by increasing the concen-
tration of [2H8]-1 that was fed. Compound [2H8]-1 was fed
to the flowers for 48 h after unfurling initiated (stage 3), and
the flower petals were collected and extracted with pentane
in a microwave oven. In the previous paper,18 all [2Hn]-2
were detected at the same tR, but in this study, each [2Hn]-2
isotopomers was detected at different tRs (18.23, 18.30, and
18.32 min for [2H8]-2, [2H7]-2, and [2H6]-2, respectively,
due to the isotope effects under the conditions used (see
Section 3). Therefore, the amount of each isotopomer of
[2Hn]-2 was estimated from the GC-SIM traces at m/z 130
for [2H8]-2, m/z 129 for [2H7]-2, and m/z 128 for [2H6]-2 at
each tR. The ratio of [2Hn]-2 plus [2Hn]-3 to 2 plus 3
increased as the concentration of [2H8]-1 increased
(measured ratio 0.00033, 0.0010, 0.0032, 0.0036, 0.0073,
and 0.013 at a concentration of 2.1, 4.3, 8.7, 17.3, 34.5, and
69.0 mM, respectively). This indicated that 1 is an actual
precursor of 2 in the flowers. The overall yield of [2Hn]-2
plus [2Hn]-3 was ca. 1.1% when all of the [2H8]-1 fed
was absorbed in the flowers. As [2H8]-1 contains ca. 15% of
Figure 6. Changes in the ratio of [1,1,2,20,30,40,50,60-2H8]-2 to
[1,2,20,30,40,50,60-2H7]-2 emitted from the flowers of Rosa ‘Hoh-Jun’ after
feeding with [1,1,2,20,30,40,50,60-2H8]-3. W, [1,1,2,20,30,40,50,60-2H8]-2; X,
[1,2,20,30,40,50,60-2H7]-2. Intact plants were exposed to the same conditions
as in Figure 2. Stages of flowers are defined in the text.
1
[2H7]- and/or [2H6]-1 on the basis of H NMR analyses
We recently found that [2H8]-1 is transformed to 2 and its b-
D-glucopyranoside 3 via 4, and we proposed that this is the
main biosynthetic pathway of 2 and 3 in R. damascena Mill.
and R. ‘Hoh-Jun’ flowers.18 The proposed pathway and the
labeling pattern are illustrated in Figure 1;18 this pathway
has been already observed in yeasts and various plant
tissues.22–24 In our experiments, not only [2H7]-2 and [2H7]-
3, the primary isotopomers of the deuterium labeled
of the methyl ester, the total empirical peak height for
[2H7]-2 must be subtracted by 15% of each value. The
amount of [2H6]-2 was not considered because the peak
(which accounted for less than 5% of the total amount
of [2Hn]-2) for [2H6]-2 might have originated from
[2H6]-1 at the initial feeding stage. On the basis of the
prerequisite condition, the ratio ([2H8]-/[2H7]-2¼50–58/
40–42, [2H8]-/[2H7]-3¼50–58/40–42) remained almost

S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
7009
As shown in Figure 7(a) and (b), changes in the ratio of
[2H8]-2/[2H7]-2 change corresponded to those in the ratio of
[2H8]-3/[2H7]-3 throughout the unfurling process (between
stages 3 and 6). The ratio of [2H8]-2/[2H7]-2 and [2H8]-3/
[2H7]-3 was 63/37, and 62/38, respectively, at the initial
sampling point, but had decreased to reach 42/58 and 47/53,
respectively, by 72 h after the feeding commenced. This
correlation indicates that an equilibrium exists between 2
and 3 in the petal tissue, and abstracting one of the
deuterium atoms from [2H8]-2 by the action of alcohol
dehydrogenase caused a decline in the ratio. The difference
between the current value and the previously reported ratio
[2H8]-2/[2H7]-2¼[2H8]-3/[2H7]-3¼10–20/90–80)18 must
be due to this abstraction of a deuterium atom from the
initial isotopomer, [2H8]-2, during longer duration of the
previous study. Therefore, 2 must be converted from 1
without the loss of a hydrogen atom at the a-position.
Interestingly, the postulated intermediates, [2H8]-5 and
[2H8]-6,25 were not detected in R. ‘Hoh-Jun’ flowers upon
feeding with [2H8]-1. This might be due to the rapid
metabolism of [2H8]-1 to [2H8]-2 in the0 flower petals.
Therefore, we fed 34.5 mmol of [1,1,2,2,20,3 ,40,50,60-2H9]-6,
to the R. ‘Hon-Jun’ flowers at stage 4 for 96 h to verify
whether 2 and/or 3 would be produced via 6. The flowers
were extracted successively with pentane and ethyl acetate
to test for [2H8]-2 and/or [2H8]-3. [2H8]-20w0as0not detected
in the volatile fraction, although [1,2,2,2 ,3 ,4 ,50,60-2H8]-3
Figure 7. Changes in the ratio of [2H8]-2 to [2H7]-2 detected in the volatile
fraction (a) and as the aglycone of 3 (b) in the flowers of Rosa ‘Hoh-Jun’
after feeding with [2H8]-1. W, [2H8]-2; X, [2H7]-2. Intact plants were
exposed to the same conditions as in Figure 2.
was detected in
a trace amounts (0.0072 mmol,
yield¼0.021%). This yield was significantly lower than
that (11%) from [2H8]-1, indicating that 6 might not be
involved in the biosynthesis of 2 from 1.
identical, regardless of the increase in the concentration
of the fed [2H8]-1.
As previously reported26, 5, a key intermediate in the
biosynthetic pathway, has been synthesized from 1 by a
myeloperoxidase/hydrogen peroxide/chloride system
derived from neutrophils. In a cell-free system, 5 is also
non-enzymatically produced from 1 in the presence of the
ortho-quinone of catechin, during which Schiff-base
formation with 1 has been proposed as an intermediate.27
We next compared changes in the 2H8]-2/[2H7]-2 and [2H8]-
3/[2H7]-3 ratios over time. Feeding of [2H8]-1 was started at
stage 3 of the R. ‘Hoh-Jun’ flower, and the flowers were
subsequently collected at 6, 12, 24, 48, and 72 h after
feeding began. Most of the [2H8]-1 solution was absorbed by
72 h. The flowers were successively extracted by pentane
and ethyl acetate in a microwave oven yielding volatile and
glycosidic fractions as previously reported by Oka et al.14
Recently,
a cytochrome P450 dependent oxidase,
Figure 8. Possible biogenetic pathways for 2 from 1 in the rose flowers.

7010
S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
CYP79A2, has been characterized in Arabidopsis and has
been shown to oxidize 1, yielding phenylacetaldoxime (7).28
Compound 7 can be enzymatically or non-enzymatically
oxidized yielding 5. In these cases, a deuterium atom at the
a-position of [2H8]-1 must be retained to yield [2H8]-5,
which then subsequently yields [2H8]-2. Furthermore, 7 has
been identified among volatile compounds in the flower
concrete of Michelia champaca L.29 We therefore,
examined whether 7 is an intermediate in the biosynthetic
pathway 0 in 0R. ‘Hoh-Jun’ flowers. We injected
[1,2,20,30,4 ,50,6 -2H7]-7 (14 mmol/ml of DMSO) into R.
‘Hoh-Jun’ 0flowers. The overall yield (0.82 mmol0, 5.8%) of
[1,2,20,30,4 ,50,60-2H7]0-20(0.4%) and [1,2,20,30,40,5 ,60-2H7]-3
(5.4%) from [1,2,20,3 ,4 ,50,60-2H7]-7 was determined on the
basis of the GC-MS analyses of the pentane and ethyl
acetate fractions extracted from flowers detached 24 h after
injection, suggesting that 7 is involved in the biosynthesis of
2. Although the transformation of deuterium labeled 5 has
not been confirmed yet, we can propose a plausible
biosynthetic pathway for 2 as shown in Figure 8, on the
basis of the results so far.
The fraction (1.00 g, 42% of the total) was dissolved in
MeCN, and stirred for 24 h at ambient temperature in the
presence of 10.4 g (25.0 mmol) of 2,3,4,6-tetra-O-acetyl-
1-bromo a-glucose (Sigma), Hg(CN)2 (2.80 g, Sigma),
˚
and m0olecular sieves (4 A). After purification,
[1,1,2,2 ,30,40,50,60-2H8]-2-phenylethyl 2,3,4,6-tetra-O-
acetyl b-D-glucolyranoside, it was treated with 5 ml of
MeONa (1 M) in MeOH–CHCl3 (1:1), followed by Dowex
50W-£4 (Hþ from), and the product was purified by
silica gel0 chromatography (CHCl3 –MeOH) to yield
[1,1,2,20,3 ,400,500,600-2H8]-3 (889 mg, 3.00 mmol, 40.1%
from [1,1,2,2 ,3 ,4 ,50,60-2H8] styrene). FABMS (pos.) ana-
lyses gave an ion peak at m/z 293 [MþH]þ. 1H NMR
(500 MHz, CD3OD) dH: 2.93 (1H, s, H-2), 3.22 (1H, dd,
J¼8.7, 8.9 Hz, H-200), 3.35 (1H, 0t0, J¼8.9 Hz, H-400), 3.39
(1H, ddd, J¼8.9, 5.4, 2.2 Hz, H-5 ), 3.45 0(01H, t, J¼8.9 Hz,
H-300), 3.69 (1H, dd,00J¼11.7, 5.4 Hz, H-6 a), 3.99 (1H, dd,
J¼11.7, 2.2 Hz, H-6 b), 4.45 (1H, d, J¼8.7 Hz, H-100); 13
C
NMR (125 MHz, CD3OD) dC: 37.4 (t, J¼19.2 Hz, C-2),
63.600(C-600), 72.4 (C-400), 72.7 (quin. J¼21.6 Hz, C-1), 75.9
(C-2 ), 78.6 (C0-300), 78.6 (C-500), 105.1 (C-100)0, 128.8 (t,
J¼25.8 Hz, C-6 ), 131.0 (2C, t, J¼23.8 Hz, C-5 ,70), 131.4
(2C, t, J¼24.8 Hz, C-40,80), 141.2 (C-30). GC-MS analyses
gave ion peaks at m/z 130 [M]þ0 (100%) and m/z 129
[M21]þ (19%) for [1,1,2,20,30,40,5 ,60-2H8]-2, and peaks at
m/z 122 [M]þ (100%) and m/z 123 [Mþ1]þ (8.3%) for 2,
respectively. On the basis of t0he0se0 data, the ratio of the
[2H7]-isotopomer in [1,1,2,20,3 ,4 ,5 ,60-2H8]-2 was calcu-
lated to b0e 17%.0 The ratio of [2H7]-isotopomer (17%) in
[1,1,2,20,3 ,40,50,6 -2H8]-3 was also estimated by GC-MS of
the products obtained from [1,1,2,20,30,40,50,60-2H8]-3 after
hydrolysis with b-glucosidase (from almond).
We are now focusing on the identification of the
hypothesized intermediates, such as [2H8]-5 and [2H8]-7,
in order to clarify the exact pathway that yields [2H8]-2 and
[2H8]-3 from [2H8]-1 in rose petals.
In conclusion, b-glucosidase is thought to be partly
involved in controlling diurnal emission of 2 in the R.
damascena Mill. flowers. In addition, our analyses of the
flower tissue after feeding have enabled us to propose a
novel biosynthetic pathway for 2 and 3 via [2H8]-5 from
[2H8]-1 in rose flowers.
3.1.2. Synthesis of [1,1,2,2,20,30,40,50,60-2H9]-2-phenyletyl-
amine (6). LiAlD4 (207 mg, 4.95 mmol, Aldrich) and
anhydrous AlCl3 (264 mg, 1.98 mmol, Sigma) were added
to die0thyl ether (6 ml), and stirred for 1 h under ice-cooling.
[1,1,2 ,30,40,50,60-2H7]-Phenylacetonitrile (204 mg, 1.65 mmol,
Aldrich) was added and stirred for 1 h at ambient tem-
perature. [1,1,2,2,20,30,40,50,60-2H9]-6 was obtained after
treatment with 1 M NaOH and extraction with diethyl
ether (197 mg, 1.50 mmol, yield 91%). EIMS analyses gave
a peak at m/z 131 [M]þ; 13C NMR (125 MHz, CD3OD) dC:
39.1 (quin., J¼19.1 Hz, C-2), 43.3 (quin., J¼21.0 Hz, C-1),
126.7 (t, J¼24.3 Hz, C-40), 129.00 (2C, t, J¼24.3 Hz, C-20,
60), 129.3 (2C, t, J¼23.8 Hz, C-3 ,50), 140.7 (s, C-10). The
deuterium ratio was not estimated.
3. Experimental
3.1. Chemicals and biochemicals
2,3,3,20,30,40,50,60-[2H8]
L-Phenylalanine
1
([2H8]-1,
98 atom% H, Aldrich) was used. In the H NMR spectral
2
analyses of the methyl ester of [2H8]-1, the H/H ratio was
2
evaluated to be 17/83, 3/97, and 4/96 for H-2,3,3, and 20-60,
respectively. 2-Phenyl-N-glucosyl-acetamidiumbromide (8)
was synthesized by the method of Hiratake et al.19,20 b-
Glucosidase (EC 3.2.121, 500 units/mg from almond,
Sigma) and naringinase (300 units/mg from Penicillium
decumbens, Sigma) were used for enzymatic hydrolysis of
glycoconjugated volatile compounds.
3.1.3. Synthesis of [1,2,200,300,40,50,60-2H7]-phenylacetaldox-
ime (7). [1,2,20,30,40,5 ,60-2H7]-phenylacetaldoxime (7)
was prepared from [1,2,2 ,30,40,50,60-2H7]-1-nitrostyrene30
according to the method of Sera et al.31 [1,2,20,30,40,50,60-2H7]-
1-Nitrostyrene (500 mg, 3.3 mmol) was treated by lead
powder (5.5 g, 33 mmol) in acetic acid-DMF (1.5:20 v/v,
21.5 ml) for 2 h at ambient temperature. The reaction
mixture was added to water (20 ml) and extracted with
diethyl ether (20 ml£3 times) yielding crude extract, which
was then purified by column chromatography 0on silica gel
(hexane–diethyl ether) to yield [1,2,20,30,4 ,50,60-2H7]-7
(340 mg, 2.4 mmol, yield 73%). EIMS m/z 142 [M]þ; 13C
NMR (125 MHz, CD3OD) dC: 31.1 (t, J¼19.5 Hz, C-2 of Z-
isomer), 35.4 (t, 0J¼19.5 Hz, C-2 of E-isomer), 129.0 (02C,0t,
J¼24.4 Hz, C-2 , 60), 129.1 (2C, t, J¼23.9 Hz, C-3 , 5 ),
3.1.1. Synthesis of [1,1,2,20,30,40,50,60-2H8]-2-phenyl-
ethyl b-D-glu0co0pyranoside ([1,1,2,20,30,40,50,60-2H8]-3)
[1,1,2,20,30,40,5 ,6 -2H8]-2 was synthesized by reduction
with 9-BBN (030 mmol) in THF (10 ml) from
[1,1,2,20,30,40,50,6 -2H8] styrene (2.00 g, 17.8 mmol,
98 atom% 2H, Cambridge Isotope Laboratories, MA,
USA), although the labeling pattern is not exactly the
same as one of the isotopomers, [1,2,2,20,30,40,50,60-2H8]-2,
produced from [2H8]-1. The reaction mixture was treated
with 6 M aq. NaOH (30 ml) and 35% H2O2 (30 ml). The
desired compound was obtained after silica gel chromato-
graphy (hexane–EtOAc) yielding 2.38 g of a fraction
containing both [1,1,2,20,30,40,50,60-2H8]-2 and the solvent.

S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
7011
135.0 (s, C-10 of E-isomer), 135.2 (s, C-10 of Z-isomer),
150.5 (t, J¼26.8 Hz, C-1 of E-isomer), 150.7 (t, J¼26.8 Hz,
C-1 of Z-isomer). The deuterium ratio of [1,2,20,30,40,50,60-2H7]-
7 was not estimated. The ratio (3:1) of Z- to E-isomers was
the flower. The calyx was removed and the petal parts of
each flower head were crushed in liquid nitrogen and stored
at 280 8C until use.
3.4. [1,1,2,20,30,40,50,60-2H8]-2-Phenylethyl b-D-gluco-
pyranoside (3) feeding in R. ‘Hoh-Jun’ flowers
1
determined on the basis of the H NMR spectrum of 7,
prepared from 1-nitrostyrene30 according to the method
1
mentioned above. H NMR (500 MHz, CD3OD) of 7; dH:
3.54 (0.25H, d, J¼12.0 Hz H-2 of E-isomer), 3.75 (0.75H, d,
J¼10.0 Hz, H-2 of Z-isomer), 6.89 (0.75H, t, J¼10.0 Hz,
H-1 of Z-isomer), 7.54 (0.25H, t, J¼12.0 Hz, H-1 of E-
isomer), 7.20–7.36 (5H, H-20-60).
A solution of [1,1,2,20,30,40,50,60-2H8]-3 (3.4 mM, 1500 ml)
was fed to the flowers at stage 3 in intact R. ‘Hoh-Jun’ plants
as described above. The floral scent compounds emitted
from the flowers were trapped by the dynamic head space
method using automatic sampling system equipped with a
time controller and six air-valves developed by the authors.
A Tenax-TA column (150 mg, 3 mm i.d. £100 mm) was
used as an absorbent with an air at flow rate of 500 ml/min.
Scent compounds were eluted from the column with
CH2Cl2 –diethyl ether (1:1, v/v) and concentrated at
40 8C. The concentrated materials were directly analyzed by
GC-SIM.
3.2. Plant materials
Rosa damascena Mill. and Rosa ‘Hoh-Jun’ plants were
grown on the University Farm, Faculty of Agriculture,
Shizuoka University, Japan under natural conditions in 5-l
pots, watered daily and fed weekly with a nutrient solution.
Each plant bore 30–50 (R. damascena) and 1–4 (R. ‘Hoh-
Jun’) flowers at various stages. R. ‘Hoh-Jun’ is a cultivar
that produces flowers throughout the year, whereas
R. damascena produces flowers from the end of
April until mid-May in Shizuoka, Japan. Plants were
acclimatized for 7 days to a constant temperature of
22^2 8C and a 12 h/12 h light/dark (LD) cycle in a
controlled environment growth room with a natural light
source (approximately 300 mmol m22 s21). In subsequent
experiments, plants were acclimatized to 22^2 8C, 75^5%
humidity and 12 h photoperiods (300 mmol m22 s21, pro-
vided by fluorescent lamps) in a controlled environment
growth room.
3.5. [1,2,2,20,30,40,50,60-2H8] L-Phenylalanine (1), [2H8]-1
feeding in R. ‘Hoh-Jun’ flowers
Feeding studies done in October and November 2002
were conducted on flowers at stage 3 in intact R. ‘Hoh-Jun’
plants as0 descr0ibed above. An aqueous solution of
[1,2,2,20,3 ,40,50,6 -2H8]-1 at each concentration, 1.1, 2.1,
4.3, 8.7 and 17.3 mM, 1.5 ml was fed as described above.
The total amount of [2H8]-1 absorbed by the plant was not
evaluated. Three of the treated flower heads were detached
just below the ovary of the flower at 48 h after feeding
started. The petal parts were crushed in liquid nitrogen and
stored 280 8C. [2Hn]-2/2 and [2Hn]-3/3 were extracted by
pentane and ethyl acetate, respectively. The ratio of [2Hn]-2/
2 and [2Hn]-3/3 in the tissues were estimated from the
GC-SIM data.
The stages of floral growth are defined as follows: stage 1,
immature flowers, sepals tightly closed; stage 2, sepals
retracting, petal whorl tightly closed, petals beginning to
lighten in color; stage 3, commencement of petal
unfurling (4 days after stage 1), sepals fully retracted,
outer petal whorl beginning to loosen; stage 4, outer
petal whorl opened, inner petal whorl beginning to
loosen; stage 5, inner petal whorl partly opened but
reproductive organs not yet visible; stage 6, inner and
outer whorls open, reproductive organs not visible;
stage 7, full bloom flower, inner and outer whorls open,
reproductive organs visible; stage 8, inner and outer
whorls completely open, 30 h (R. damascena) or 5 days
(R. ‘Hoh-Jun’) after stage 3; stage 9, senescence stage,
anthers dried and darkened, petals pale.
In May, 2003, [2H8]-1 (17.3 mM, 1.5 ml) was fed to five
flowers at stage 3 in intact R. ‘Hoh-Jun’ plants as described
above. The flowers were detached at 6, 12, 24, 48, and 72 h
after feeding started and were treated as described above.
Changes in the ratio of [2H8]-2/[2H7]-2 and [2H8]-3/[2H7]-3
during the unfurling process were estimated from the
GC-SIM data.
3.6. [1,1,2,2,20,30,40,50,60-2H9]-2-Phenylethylamine (6)
feeding in R. ‘Hoh-Jun’ flowers
3.3. Phenyl-N-glucosyl-acetamidiumbromide (8) feeding
in R. damascena Mill. flowers
In May, 2003, [1,1,2,2,20,30,40,50,60-2H9]-6 (23 mM solution,
pH 6.8 as HCl salt, 1500 ml) was fed to five immature
flowers at stage 3 for 96 h until the full bloom stage (stage 7)
of R. ‘Hoh-Jun’ by the method just described above. The
flowers were detached at 96 h after feeding started and were
treated as described in Section 3.3.
Feeding studies done in April to May of 2002 and 2003 were
conducted on opening flowers at stage 3 in intact R. ‘Hoh-
Jun’ plants as described above. A thin needle with a cotton
thread (50 mm in length) was inserted at 2 p.m. on day 0
through the top of the ovary, just above the ovules. This
thread acted as a wick enabling the absorption of an aqueous
solution of 8 (10 mM, 300 ml) from an Eppendorf tube
attached to the stem. A control group was evaluated in
parallel in which water was fed instead of 8. The total
amount of 8 absorbed by the plant was not evaluated. Three
treated flower heads at 6, 12, 16, 18, 20, 24, 30, 36, and 42 h
after feeding startted were detached just below the ovary of
3.7. [1,2,20,30,40,50,60-2H7]-2-Phenylacetaldoxime (7)
feeding in R. ‘Hoh-Jun’ flowers
In June, 2003, 1 ml of [1,2,20,30,40,50,60-2H7]-7 (14 mM of
DMSO solution) was injected to the top of the ovary, just
above the ovules of three immature flowers at stage 3. The
flowers were detached at 24 h after injection and treated as
described in Section 3.3.

7012
S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
3.8. Extraction of the volatile compounds and
glycoconjugates
3.11. Estimation of b-glucosidase activity in the petal
tissues during unfurling process of R. damascena Mill
A portion (equivalent to one flower head) of each type of
frozen flower petal was extracted twice with pentane (20 ml
each) and then twice with EtOAc (20 ml each) in a
microwave oven according to the method of Oka et al.11
Ethyl octanoate (20 mg) and phenyl b-D-glucopyranoside
(20 mg) were used as internal standards in the GC, GC-MS
and GC-SIM analyses. The pentane extract was concentrated
and directly analyzed by GC-MS and GC-SIM revealing the
presence of 2 and [2Hn]-2. The EtOAc extract was evaporated
in vacuo and the resulting residue was re-dissolved in a citrate
buffer (10 mM, pH 6.0) and hydrolyzed by a mixture of b-
glucosidase (2500 units) and naringinase (3000 units) yield-
ing the volatile compounds. After extraction with an
azeotropic mixture (bp 38 8C) of pentane–CH2Cl2 (2:1),12
the resulting extract was analyzed by GC-MS and GC-SIM to
evaluate the amount of 2 and [2Hn]-2 as an aglycone part of 3
and [2Hn]-3, respectively. To quantify 3 in the flower tissues,
the ethyl acetate extract was directly analyzed for pertri-
fluoroacetates (TFA) derivatives by GC.
2 ml of citrate buffer (50 mM, pH 6.0, containing 5 mM
EDTA 2 Na, 2 mM dithiothreitol, 10 mM ascorbic acid, and
1% Triton X-100) was added to 0.1 g of liquid nitrogen
powder (prepared from flower petals at the onset of
unfurling) and stirred for 30 min at 4 8C, and was then
centrifuged at 2000g for 10 min at 4 8C. The supernatant
was passed through a membrane filter (0.45 mm) to give a
crude enzyme solution. The b-glucosidase activity of the
crude enzyme solution was evaluated as follows: (1) 50 mM
of 3, 100 ml in citrate buffer, plus crude enzyme, 200 ml; (2)
citrate buffer, 100 ml, plus crude enzyme, 200 ml; (3)
50 mM of 3, 100 ml in the citrate buffer plus citrate buffer,
200 ml. The three mixtures were incubated for 360 min at
37 8C, and the 2 released from 3 was extracted 5 times with
pentane–CH2Cl2 (2:1) in the presence of ethyl decanoate
(10 mg). The combined extract was concentrated and
analyzed by GC-MS. The conditions were the same as
described in Section 3.10. The activity was expressed as the
amount of 2 (mmol) released from 3 in the initial 60 min of
the incubation period. The inhibitory effect of 8 (6.0£1029
,
3.9. GC analyses
1.0£1028, 6.0£1028, 1.0£1027, and 6.0£1027 mol) toward
the crude enzymes prepared from flower petals (0.1, 0.2, and
0.4 g fresh weight of the petals) at stage 6 was examined as
described above. The hydrolysis of 3 was inhibited by 95%
at a dose of 6.0£1027 mol of 8.
Ethyl acetate extract (from 100 to 500 mg of fresh flower
petals) was concentrated and analyzed by GC after con-
version into TFA-derivatives by treatment with N-methyl-
bis(trifluoroacetamide-pyridine. A HITACHI G-3000 gas
chromatograph was used for the analyses. A TC-1 column
(df¼0.25 mm), dimensions 0.25 mm i.d.£30 m, was used;
the column temperature was raised from 150 to 280 8C
(increment 2 8C/min); the injection temperature was 230 8C.
For identification and quantitative analysis, 3 was analyzed
as its TFA-derivative, and 0.1 ml of the reaction mixture was
injected onto the GC column. Identification and quantifi-
cation were carried out based on the peak area and the
relative retention time (tR0 ¼2.957 for tetra O-TFA-3)
relative to phenyl b-D-glucopyranoside-tetra-O-TFA,
which was used as an internal standard (tR¼11.32 min).
Acknowledgements
This work was supported in part by a grant-in-aid to N. W.
for scientific research on priority areas (A)(2) from the
Ministry of Education, Science, Sports and Culture of Japan,
and by grant-aid to N.W. for the development of innovative
plants and animals using transformation and Cloning
from the Ministry of Forestry Fisheries and Agricultural
Sciences of Japan. We also express our many thanks to
Prof. Y. Ebizuka of the University of Tokyo for his
informative discussions on the biogenesis of floral scent
compounds.
3.10. GC-MS and GC-SIM analyses
The GC-MS analysis was conducted with a SHIMADZU
QP505A gas chromatograph-mass spectrometer equipped
with a SHIMADZU GC-17A. The column was a 30 m TC-
WAX type with 0.25 mm i.d.£30 m; the column temperature
was elevated from 100 to 170 8C (2 8C/min); the injector
temperature was 250 8C; the ionizing voltage was 70 eV; the
scanning speed was 0.5 scan/s with the range of m/z 40–250.
1 ml of each concentrate of solvent extract from petal tissues
and of eluate from a TENAX columns was injected to the
GC-MS. The identification of 2 and [2Hn]-2 was established
by comparing their MS spectra with authentic samples as
previously described by Watanabe et al.18 Quantitative
analyses of 2 and [2Hn]-2 were carried out from the SIM
traces at m/z 130 [Mþ] of [2H8]-2, m/z 129 [Mþ] of [2H7]-2,
m/z 128 [Mþ] of [2H6]-2, m/z 98 [C7Dþ7 ], m/z 97 [C7D6Hþ],
m/z 96 [C7D5H2þ] and at m/z 122 [Mþ] and m/z 91 [C7Hþ7 ]
for 2. GC-SIM conditions were 0 the 0same as
mentioned above. Authentic [1,1,2,2 ,30,40,5 ,60-2H8]-2
(MW 130) and 2 were detected at tR¼18.23 and
18.48 min, respectively.
References and notes
¨
1. Knudsen, J. T.; Tollsten, L.; Bergstron, L. G. Phytochemistry
1993, 33, 253–280.
2. Schiestl, F. P.; Ayasse, M.; Paulus, H. F.; Erdmann, D.;
Francke, W. J. Chem. Ecol. 1997, 23, 2881–2895.
3. Borg-Karlson, A. K.; Valterova, I.; Nilsson, L. A. Phytochem-
istry 1994, 46, 1169–1172.
4. Ervik, F.; Tollsten, L.; Knudsen, J. T. Plant Syst. Evol. 1999,
217, 279–297.
5. Odell, E.; Raguso, R. A.; Jones, K. N. Am. Midl. Nat. 1999,
142, 257–265.
6. Jurgens, A.; Webber, A. C.; Gottsberger, G. Phytochemistry
2000, 55, 551–558.
7. Miyake, T.; Yamaoka, R.; Yahara, T. J Plant Res. 1998, 111,
199–205.
8. Omura, H.; Honda, K.; Hayashi, N. J. Chem. Ecol. 2000, 26,
655–666.

S. Hayashi et al. / Tetrahedron 60 (2004) 7005–7013
7013
9. Francis, M. O.; Banthorpe, D. V.; Le Patourel, G. N. J. Nature
1970, 228, 1005–1006.
20. Inoue, K.; Hiratake, J.; Mizutani, M.; Takada, M.; Yamamoto,
M.; Sakata, K. Carbohydr. Res. 2003, 338, 1477–1490.
21. Picone, M. J.; Clery, R. A.; Watanabe, N.; MacTavish, H. S.;
Turnball, C. G. N. Planta, in press.
10. Francis, M. O. J.; O’Connell, M. Phytochemistry 1969, 8,
1705–1708.
11. Banthorpe, D. V.; Le Patourel, G. N. J.; Francis, M. J. O.
Biochem. J. 1972, 130, 1045–1054.
22. Albertazzi, E.; Cardillo, R.; Servi, S.; Zucchi, G. Biotech. Lett.
1994, 16, 491–496.
12. Akermann, I. E.; Banthorpe, D. V.; Fordham, W. D.; Kinder,
J. P.; Poots, I. J. Plant Physiol. 1989, 134, 567–572.
13. Loughrin, J. H.; Hamilton-Kemp, T. R.; Andersen, R. A.;
Hildebrand, D. F. Physiol. Plant 1991, 83, 492–496.
14. Oka, N.; Ohishi, H.; Hatano, T.; Hornberger, M.; Sakata, K.;
Watanabe, N. Z. Naturforsch. 1999, 54c, 889–895.
15. Reuveni, M.; Sagi, Z.; Evnor, D.; Hetzroni, A. Plant Sci. 1999,
147, 19–24.
23. Robins, R. J.; Woolley, J. G.; Ansarin, M.; Eagles, J.;
Goodfellow, B. J. Planta 1994, 194, 86–94.
24. Dickinson, J.; Salgado, L. E.; Hewlins, J. E. J. Biol. Chem.
2003, 278, 8028–8034.
25. Walsch, C. Enzymatic Reaction Mechanisms; W.H. Freeman
and Co: New York, 1978; Chapter 14, pp 451–454.
26. Hazen, S.; d’Avignon, A.; Anderson, M. M.; Hsu, F. F.;
Heinecke, J. W. J. Biol. Chem. 1998, 273, 4997–5005.
27. Saijo, R. Jpn. Agric. Res. Q. 1973, 7, 202–207.
28. Wittstock, U.; Halkier, B. A. J. Biol. Chem. 2000, 275,
14659–14666.
16. Loughrin, J. H.; Hamilton-Kemp, T. R.; Burton, H. R.;
Andersen, R. A.; Hildebrand, D. F. Phytochemistry 1992, 31,
1537–1540.
17. Jakobsen, H. B.; Christensen, L. P. Plant Cell Env. 2002, 25,
773–781.
29. Kaiser, R. J. Ess. Oil Res. 1991, 3, 129–146.
30. Campos, P. J.; Garcia, B.; Rodriguez, M. A. Tetrahedron Lett.
2000, 41, 979–982.
18. Watanabe, S.; Hayashi, K.; Yagi, K.; Asai, T.; MacTavish, H.;
Picone, J.; Turnbull, C.; Watanabe, N. Biosci. Biotechnol.
Biochem. 2002, 66, 943–947.
31. Sera, A.; Yamauchi, H.; Yamada, H.; Itoh, K. Synlett 1990,
477–478.
19. Guo, W.; Hiratake, J.; Ogawa, K.; Yamamoto, M.; Ma, S. J.;
Sakata, K. Bioorg. Med. Chem. Lett. 2001, 11, 467–470.