Rationalizing the formation of damascenone: Synthesis and hydrolysis of damascenone precursors and their analogues, in both aglycone and glycoconjugate forms

Article abstract of DOI:10.1021/jf8018134

Storage of megastigma-4,6,7-trien-3,9-diol (5), and megastigma-3,4-dien-7- yn-9-ol (6) in aqueous ethanol solution at pH 3.0 and 3.2 gave exclusively damascenone (1) and damascenone adducts at room temperature. The diol (5) had half-lives for the conversion of 32 and 48 h at pH 3.0 and pH 3.2, respectively. The acetylenic alcohol (6) had half-lives of 40 and 65 h at the same pH levels. In order to study the reactivity of the C-9 hydroxyl function in 5 and in the previously investigated allenic triol 2, two model compounds, megastigma-4,6,7-trien-9-ol (7) and megastigma-6,7-dien-9-ol (8) were synthesized. No 1,3-transposition of oxygen to form analogues of damascenone was observed when 7 and 8 were subjected to mild acidic conditions. Such transposition takes place only with highly conjugated acetylenic precursors such as 6 or tertiary allenic alcohols such as 2. The placement of glucose at C-3 of 5 and at C-9 of 6 gave the glycosides 9 and 10, respectively. The effect of such glucoconjugation was to increase the observed half-lives by a factor of only 1.6-1.7 for the allenic glucoside 9, and by 2.1-2.2 for the acetylenic glucoside 10. These studies indicate that the effect of glycosylation on damascenone formation is probably not important on the time scale of wine making and maturation.

Full text of DOI:10.1021/jf8018134

J. Agric. Food Chem. 2008, 56, 9183–9189 9183
Rationalizing the Formation of Damascenone:
Synthesis and Hydrolysis of Damascenone
Precursors and Their Analogues, in both Aglycone
and Glycoconjugate Forms
MERRAN A. DANIEL,†,‡,§ CAROLYN J. PUGLISI,†,‡,§ DIMITRA L. CAPONE,‡,§
‡,§,|
GORDON M. ELSEY,*,†,‡,§,| AND MARK A. SEFTON
School of Chemistry, Physics and Earth Sciences, Flinders University, P.O. Box 2100,
Adelaide 5001, Australia, The Australian Wine Research Institute, P.O. Box 197,
Glen Osmond, Adelaide 5064, Australia, and Cooperative Research Centre for Viticulture,
P.O. Box 154, Glen Osmond, Adelaide 5064, Australia
Storage of megastigma-4,6,7-trien-3,9-diol (5), and megastigma-3,4-dien-7-yn-9-ol (6) in aqueous
ethanol solution at pH 3.0 and 3.2 gave exclusively damascenone (1) and damascenone adducts at
room temperature. The diol (5) had half-lives for the conversion of 32 and 48 h at pH 3.0 and pH 3.2,
respectively. The acetylenic alcohol (6) had half-lives of 40 and 65 h at the same pH levels. In order
to study the reactivity of the C-9 hydroxyl function in 5 and in the previously investigated allenic triol
2, two model compounds, megastigma-4,6,7-trien-9-ol (7) and megastigma-6,7-dien-9-ol (8) were
synthesized. No 1,3-transposition of oxygen to form analogues of damascenone was observed when
7 and 8 were subjected to mild acidic conditions. Such transposition takes place only with highly
conjugated acetylenic precursors such as 6 or tertiary allenic alcohols such as 2. The placement of
glucose at C-3 of 5 and at C-9 of 6 gave the glycosides 9 and 10, respectively. The effect of
such glucoconjugation was to increase the observed half-lives by a factor of only 1.6-1.7 for the
allenic glucoside 9, and by 2.1-2.2 for the acetylenic glucoside 10. These studies indicate that the
effect of glycosylation on damascenone formation is probably not important on the time scale of
wine making and maturation.
KEYWORDS: Damascenone; allenic alcohols; acetylenic alcohols; glycoconjugate; hydrolysis; noriso-
prenoid; carotenoid metabolites; oxygen transposition.
INTRODUCTION
megastigmanes such as damascenone (1). As 1 can be formed
by mild acid hydrolysis of crude plant isolates, much research
has focused on chemical mechanisms for transposition of oxy-
gen from C-9 to C-7 to form damascenone, resulting in the
identification of the allenic triol 2 as a damascenone precursor
(7). At room temperature and under mild acid conditions, 2 is
rapidly converted to 3-hydroxydamascone (3) and the enyne
diol 4, as major products, along with small amounts of
damascenone (1) (Figure 1A) (7). The enyne diol 4 was also
found to produce both 1 and 3, but at a rate several orders of
magnitude lower than that observed for 2 (5). The lower
reactivity of 4 was considered to be consistent with the
observation by Olsson et al. (8) that, under acidic conditions,
R-hydroxy allenes rearrange much faster than do R-hydroxy
acetylenes.
ꢀ-Damascenone (more recently referred to as simply dama-
scenone, 1) is one of the most important natural aroma
compounds known; it has been identified in many different types
of plant material and is also one of the mainstays of the
international perfume industry (1). Thirteen-carbon secondary
metabolites with the megastigmane skeleton, including dama-
scenone (1), are widely found in plants and plant products (2-6).
Most of these contain oxygen at the C-9 position, which is
usually presumed to have resulted directly from oxidative
cleavage of a carotenoid precursor. Simple carotenoid degrada-
tion alone, however, cannot directly generate 7-oxygenated
* Corresponding author. Tel: +61 8 8303 7295. Fax: +61 8 8303
7116. E-mail: Gordon.Elsey@adelaide.edu.au.
† Flinders University.
More recently, both the dienyne species 6 (9) and the allenic
diol 5 (10) were shown to be intermediates in the formation of
damascenone from 2, and the former was observed as an in-
termediate in the hydrolysis of the latter (Figure 1B). No trace
of 3-hydroxydamascone (3) was found from hydrolysis of either
‡ The Australian Wine Research Institute.
§ Cooperative Research Centre for Viticulture.
| Present address School of Agriculture, Food and Wine, The
University of Adelaide, Waite Campus, PMB 1, Glen Osmond, SA
5064, Australia.
10.1021/jf8018134 CCC: $40.75  2008 American Chemical Society
Published on Web 09/04/2008

9184 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Daniel et al.
Figure 1. A: Major hydrolysis products of allenic triol 2. B: Route of formation of damascenone 1 from allene triol 2.
reaching the required value. Damascenone was quantified by the method
reported in Daniel et al. (12), which utilizes GC/MS and d4-dam-
ascenone as internal standard.
5 or 6. This indicated the possibility that the formation of 3
from 2 might proceed directly, via loss of the C-9 R-allenic
hydroxyl in 2. Puglisi et al. (10) observed that, under the reaction
conditions (pH 3.0, model wine), the diol 5 initially reacted
exclusively at C-3, presumably because this generates the more
stable cation. However, placement of a glycoconjugating unit
on the C-3 hydroxyl group could conceivably reduce the
reactivity at this site to the point where reaction at C-9 becomes
competitive (5).
Methods. (9R)-9-Hydroxymegastigma-3,5-dien-7-yne (6). To a
solution of (R)-but-3-yn-2-ol (2.1 g, 0.03 mol) in cold (-78 °C) ether
(40 mL) was added n-BuLi (2.5 M in hexanes, 23 mL, 0.058 mol)
dropwise. The reaction mixture was allowed to warm to room tem-
perature before enone 20 (9) (2.78 g, 0.02 mol) was added dropwise.
The reaction was stirred at room temperature for 2 days before being
quenched with saturated NH4Cl. The organic layer was diluted with
ethyl acetate and washed with brine, before being dried (Na2SO4) and
purified by silica column chromatography to give (9R)-6,9-dihy-
droxymegastigm-4-en-7-yne (3.63 g, 87%). To this was added pyridine
(4.15 g, 52.5 mmol, 4.24 mL) and acetic anhydride (3.57 g, 35 mmol,
3.3 mL). The mixture was heated at reflux overnight, cooled, and diluted
with ethyl acetate. The organic layer was washed with 10% HCl
solution, aqueous Na2SO4 and brine before being dried and concentrated
in Vacuo. The crude material was purified by column chromatography
(15% ethyl acetate in hexane) to produce (9R)-6-hydroxy-9-acetoxyme-
gastigm-4-en-7-yne (3.2 g).
To a predried flask containing P2O5 (16 g) and celite (8 g), under an
inert atmosphere, was added toluene (50 mL) and pyridine (10 mL).
The mixture was stirred to form a slurry, and a solution of the previously
prepared acetate (1.0 g, 4.8 mmol) in toluene (5 mL) was then added.
The mixture was stirred at reflux for 2 days, cooled, filtered, and washed
with diethyl ether followed by ethyl acetate. The residue was con-
centrated in vacuo and purified by column chromatography to yield
(9R)-9-acetoxymegastigma-3,5-dien-7-yne (0.14 g, 22%).
To a 50% aqueous ethanol solution (2 mL) was added KOH (0.015
g, 0.14 mmol) and (9R)-9-acetoxymegastigma-3,5-dien-7-yne (0.03 g,
0.13 mmol). The mixture was stirred for 2 h at room temperature, and
the residue was then extracted with dichloromethane. The organic layer
was washed with water and brine, dried (Na2SO4) and concentrated in
Vacuo to yield (9R)-(6) (0.0247 g, 99%).
(9S)-9-Hydroxymegastigma-3,5-dien-7-yne (6). The (9S) enantiomer
of 6 was prepared in a manner identical to that described above for the
(9R) enantiomer, with the only difference being the use of (S)-but-3-
yn-2-ol to introduce the stereochemistry on the side-chain. Except for
the sign of the optical rotation measured, the pure (9S)-enantiomer was
indistinguishable from the (9R)-enantiomer.
We report, here, a study of the rate of formation and yields
of damascenone (1) from the aglycones 5 and 6, and their
corresponding glucoconjugates 9 and 10. This was conducted
in order to determine whether glycoconjugation had important
effects on the rates of reaction of, and product distribution from,
the aglycone forms. We have also studied the reactivity of the
model compounds 7 and 8 in order to assess whether transfor-
mations of secondary R-hydroxy allenes can take place under
physiological or mild food processing conditions when compet-
ing reactions are absent or slow.
MATERIALS AND METHODS
Materials. Chemicals were purchased from Sigma-Aldrich. (S)-
Phorenol (16) (ee >96%) was provided by Dr. K. Puntener, Hoffmann-
La Roche International. All solvents used were pesticide grade from
OmniSolv (Darmstadt, Germany). X4 is a mixed hydrocarbon solvent,
with n-hexane as the major component. All organic solvent solutions
were dried over anhydrous sodium sulfate before being filtered. pH
measurements were made with a EcoScan pH 5/6 m (Eutech Instru-
ments, Singapore), which was calibrated before use. Column chroma-
tography was performed using silica gel 60 (230-400 mesh) from
Merck. Optical rotations were measured on a Polaar 21 polarimeter
1
operating at 20 °C. Routine H and 13C NMR spectra were recorded
(in CDCl3) with a Varian Gemini spectrometer at operating frequencies
of 300 and 75.5 MHz, respectively. All compounds gave spectroscopic
data that were consistent with the expected structures. Mass spectra
were recorded on a Hewlett-Packard (HP) 6890 gas chromatograph
fitted with a liquid HP 6890 series injector and coupled to a HP 5973
mass spectrometer, with the instrument set up as described by Janusz
et al. (11); specifically, the oven temperature was started at 80 °C, held
at this temperature for 1 min, increased to 170 at 4 °C/min, then to
250 at 50 °C/min, and held at this upper temperature for 10 min. The
injector was held at 220 °C and the transfer line at 260 °C. Buffer
solutions were prepared by saturating a 10% ethanol solution with
potassium hydrogen tartrate and adding 10% tartaric acid solution until
(9R)-9-O-(ꢀ-D-glucopyranosyl)-megastigma-3,5-dien-7-yne (10). To
a solution of (9R)-6 (0.12 g, 0.65 mmol) in dry dichloromethane (3
mL) was added silver triflate (0.03 g, 1.1 mmol, 1.6 eq), S-collidine
(0.14 mL, 1.1 mmol, 1.6 eq) and R-bromo-2,3,4,6-tetrapivaloylglucose
19 (0.57 g, 0.98 mmol, 1.5 eq). The reaction was stirred overnight at
room temperature with the exclusion of light. The reaction was
quenched with NaHCO3 solution, and the organic layer was diluted
with dichloromethane. The organic layer was washed with brine, dried

Synthesis and Hydrolysis of Damascenone Precursors
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9185
(MgSO4), and concentrated in Vacuo to yield 0.66 g of a brown toffee-
like substance. The crude material was purified by column chroma-
tography (5% ethyl acetate in X4) to yield 0.32 g (72%) of the
tetrapivaloyl-protected glucoside.
purified by column chromatography (10% EtOAc/X4) to yield 262 mg
(62%) of a white crystalline compound (mpt 163-165 °C).
(3S,9R)-3-(O-ꢀ-D-Glucopyranosyl)-9-hydroxymegastigma-4,6,7-
triene (9). (R)-But-3-yn-2-ol (141 mg, 2.0 mmol) in anhydrous ether
(10 mL) at -78 °C was treated with n-BuLi (2.5 M, 1.61 mL, 4.02
mmol) and the mixture stirred at room temperature for 1 h. It was then
recooled to -50 °C and ketone 17 (262 mg, 4.0 mmol) in ether (10
mL) added dropwise. The reaction was warmed to room temperature
and stirred for 2 days, then quenched with a saturated solution of NH4Cl,
diluted with ether, washed with brine, dried (Na2SO4), and the solvent
evaporated. The product was purified by column chromatography (40%
EtOAc/X4) to yield 151 mg (52%) of (3S,9R)-18 as a white solid. To
a solution of this alkyne 18 (151 mg, 0.21 mmol) in anhydrous ether
(10 mL) at 0 °C was added LiAlH4 (56 mg, 1.48 mmol). The mixture
was heated under reflux for 5 h, quenched with EtOAc and a small
amount of water, and the solvent removed. The residue was dissolved
in MeOH, the solution filtered, and solvent removed. The product was
purified by column chromatography (15% MeOH/CH2Cl2) to yield 40
mg (52%) of a sticky solid.
General Hydrolysis Procedure for Aglycones 5,6 and glucosides 9,10.
Individual solutions of 5, 6, and 9 and both (9R)-10 and (9S)-10 (1
mg/L and 1.7 mg/L for aglycones and glycosides, respectively) in model
wine (buffered 10% aqueous ethanol) at pH 3.0 and 3.2 were sealed in
ampoules in triplicate and heated at 25 °C in a water bath. Ampoules
were removed periodically, opened, and to 1 mL was added internal
standard (d4-damascenone), and the solution was extracted with a 2:1
mix pentane/ethyl acetate and subjected to GC-MS analysis (ZB-Wax,
30 m × 0.25 mm × 0.25 µm column) to determine the formation rates
of ꢀ-damascenone. The damascenone content was quantified as
described in Daniel et al. (12).
To a solution of a portion of this protected glucoside (0.02 g, 0.36
mmol) in methanol (2 mL) was added sodium methoxide (4 eq) in
methanol (2 mL). The mixture was stirred at room temperature
overnight (monitored by TLC) and purified using column chromatog-
raphy (Merck silica 60 HF254) (10% MeOH in dichloromethane.) to
yield pure material (9R)-(10) (9.7 mg, 75%).
(9S)-9-O-(ꢀ-D-glucopyranosyl)-megastigma-3,5-dien-7-yne (10). (9S)-
(6) was converted into the corresponding glycoside by a method
identical to that described above for the (9R) enantiomer.
(9R)-9-Hydroxymegastigma-4,6,7-triene (7). (R)-But-3-yn-2-ol (762
mg, 0.01 mol) was added to ketone 20 (500 mg, 3.6 mmol) by a
procedure identical to that described above for compound 6. The product
was recrystallized from X4 to yield diol (9R)-21 as a white solid (485
mg, 64%). To a solution of (9R)-21 (465 mg, 2.24 mmol) in dry ether
(25 mL) was added LiAlH4 (594 mg, 15.6 mmol) at 0 °C. The mixture
was heated under reflux for 5 h, then quenched with acetone, diluted
(ether), washed (10% NaOH solution, brine), dried (Na2SO4), and the
solvent evaporated. The product was purified by column chromatog-
raphy (10% EtOAc/X4) to give a colorless oil (313 mg, 73%).
(9S)-9-Hydroxymegastigma-4,6,7-triene (7). The (9S) analogues of
7 were prepared by methods identical to those described above using
(S)-But-3-yn-2-ol.
(9R)-9-Hydroxymegastigma-6,7-diene (8). (R)-But-3-yn-2-ol (151
mg, 2.16 mmol) in ether (10 mL) at -78 °C was treated with n-BuLi
(2.5 M in hexanes, 1.7 mL, 4.25 mmol) and the reaction stirred for 1 h
at room temperature. The solution was recooled to -78 °C and added
dropwise to a solution of 22 (100 mg, 0.71 mmol) in ether (5 mL) at
-78 °C. The mixture was warmed to room temperature and stirred for
16 h, then quenched with NH4Cl, washed with water and brine, dried
(Na2SO4), and the solvent evaporated. The product was purified by
column chromatography to yield (9R)-23 as a colorless oil (127 mg,
84%). To a solution of alkyne (9R)-23 (127 mg, 0.61 mmol) in Et2O
(10 mL) at 0 °C was added LiAlH4 (161 mg, 4.23 mmol). The reaction
was heated under reflux for 5 h, then quenched with acetone, diluted
with Et2O, washed with an aqueous solution of 10% NaOH, brine, dried
(Na2SO4), and the solvent evaporated. Purification by column chro-
matography (10% EtOAc/X4) yielded 33 mg (28%) of (9R)-8 as a
colorless oil.
RESULTS
Synthesis of the Glucosides 9 and 10. The diol 5 (10) and
both the (9R) and (9S) isomers of the alcohol 6 (9) were prepared
as described previously. Glycosylation of each of the dienyne
alcohols 6 proceeded smoothly using a modified Koenigs-Knorr
procedure (13) to give both (9R)-10 and (9S)-10. We chose to
use the tetra-pivaloylated bromoglucose (19) as the reagent for
the introduction of the carbohydrate unit as we have found, in
keeping with earlier reports (14), that this species produces only
the ꢀ-anomer. Synthesis of the C-3 glycoside of allene 5 was
not so straightforward; attempts to introduce the glucose unit
at the C-3 position in 5 itself proved futile, with the aglycone
being too unreactive under neutral conditions and prone to
undergoing side reactions under acidic conditions. The desired
glycoside was eventually prepared successfully via the
method shown in Figure 3. Glucosylation of (S)-phorenol
(S)-16 (15) was successfully achieved (16) in 62% yield.
Addition of (2R)-3-butyn-2-ol proceeded smoothly, with no
evidence of addition of the acetylide species across the
pivaloyl carbonyls. Finally, reduction with LAH accom-
plished both rearrangement of the side chain, as well as
deprotection of the glycoside unit to give 9.
Synthesis of the Allenes 7 and 8. Two forms of the target
compound, 7, differing in their stereochemistry at C-9, (Figure
4) were prepared by addition of the dilithio derivative of either
(2R)- or (2S)-3-butyn-2-ol to 2,6,6-trimethylcyclohex-2-enone
(20) (9) followed by reaction with LAH to give 7, whose
spectroscopic data were in accord with those previously
published (17). In each case, the product was a pair of
diastereomers that differed in the stereochemistry at the allenic
position, C-6. Separation of the diastereomers was not consid-
ered to be crucial and was not attempted in either case. The
second target compound, (8, Figure 4) was obtained, in an
analogous manner, from 2,6,6-trimethylcyclohexanone (22). As
was the case with 7, the use of either (2R)- or (2S)-3-butyn-2-
ol gave a mixture of diastereomers.
(9S)-9-Hydroxymegastigma-6,7-diene (8). The (9S) analogues of 8
were prepared by methods identical to those described above using
(S)-but-3-yn-2-ol.
General Hydrolysis Procedure for Allenes 7 and 8. Each of (9R)-7,
(9S)-7, (9R)-8, and (9S)-8 (50 mg/L) was dissolved in model wine (5
mL, pH 3.0) with naphthalene (50 mg/L) as internal standard and sealed
in glass ampoules under a nitrogen atmosphere. Triplicate ampoules
were heated at 25 °C for 12, 24, 72, and 168 h or at 45 °C for 12, 24,
and 72 h. The contents were then extracted with pentane/EtOAc (2:1)
and analyzed by GC-MS. Chiral GC was conducted using a cyclosil-B
column, 30 m × 0.25 mm × 0.25 µm.
(9S)-9-Ethoxymegastigma-6,7-diene (11). KH (60 wt. % in mineral
oil, 20 mg, 0.29 mmol) was washed with X4 and added to a stirred
solution of (9S)-8 (25 mg, 0.13 mmol) in DMF (2 mL) at 0 °C. The
mixture was stirred cold for 30 min, then ethyl iodide (200 µL, 2.5
mmol) was added and the mixture warmed to room temperature and
stirred for 1 h. The reaction was diluted with Et2O, dried, and the solvent
evaporated to give a crude mixture of the product, which was purified
by column chromatography (5% EtOAc/X4) to yield 17 mg (59%) of
a colorless oil.
(4S)-2,6,6-Trimethyl-4-O-(2′,3′,4′,6′-tetrapiValoyl-ꢀ-D-glucopyrano-
syl)-cyclohex-2-en-1-one (17). To a solution of alcohol (S)-16 (100 mg,
0.65 mmol) in dry dichloromethane (2 mL) was added silver triflate
(133 mg, 0.52 mmol), s-collidine (138 µL, 1.04 mmol), and 19 (540
mg, 0.93 mmol). A second portion of silver triflate (133 mg, 0.52 mmol)
was added and the mixture diluted with CH2Cl2 (3 mL), then stirred at
room temperature in the absence of light for 16 h. The reaction was
quenched with a saturated solution of sodium bicarbonate, washed with
brine, dried (Na2SO4), and the solvent evaporated. The product was

9186 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Daniel et al.
Figure 2. Structures of compounds studied.
Figure 3. Synthesis of C-3 glycoside 9.
Figure 4. Synthesis of allenes 7 and 8.
Hydrolysis of Aglycones 5 and 6 and Their Respective
Glucosides 9 and 10. The hydrolyses of 5, 6, 9, and 10 were
all conducted in two model wine solutions (10% ethanol in
water, pH 3.0 and pH 3.2) at 25 °C in sealed ampoules, which
were removed at selected time intervals and analyzed for their
damascenone content, by the procedure reported by Daniel et

Synthesis and Hydrolysis of Damascenone Precursors
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9187
Figure 5. Formation of damascenone (1) over time during hydrolysis of aglycone 6 at pH 3.0 (left) and pH 3.2 (right). Both hydrolyses were conducted
at 25 °C.
Table 1. Half-Lives and Glycoside/Aglycone Half-Life Ratios for the
Formation of Damascenone (1) from Compounds 5, 6, 9, and 10
c,d
t1/2
compounda
pHb
ratioe
5
3.0
3.2
3.0
3.2
32
48
50
80
9
∼1.6
∼1.7
6
3.0
3.2
3.0
3.2
3.0
3.2
40
65
84
144
84
144
(9R)-10
(9S)-10
∼2.1
∼2.2
∼2.1
∼2.2
Figure 6. Stereochemical relationship between the starting allenes 7 and
the 7,8-didehydrotheaspirane products obtained.
a All hydrolyses were conducted at 25 °C and at starting concentrations of 1
mg/L (aglycones) or 1.7 mg/L (glycosides). b Solutions were buffered with potassium
6) on the basis of their spectroscopic data (17). When the 7,8-
didehydrotheaspiranes produced from the hydrolysis of (9R)-7
were examined by chiral phase GC, only two isomers were
observed, at retention times of 13.28 and 14.53 min. Similarly,
the hydrolysis of (9S)-7 produced two isomeric 7,8-didehy-
drotheaspiranes, with retention times of 14.38 and 14.87 min.
This clearly shows that the C-9 hydroxyl did not undergo
epimerization prior to cyclization.
Hydrolysis of the hydroxy allene 8 proceeded more slowly;
the substrate had a half-life of approximately 4 days at 45 °C
and had undergone approximately 25% conversion after two
weeks at 25 °C. A single product (albeit as mixtures of isomers
that were not resolvable by chiral GC) was produced and was
identified as the allenic ether 11. This structural assignment was
confirmed by synthesis of the allenic ether by treatment of 8
with potassium hydride followed by quenching with ethyl iodide.
As was the case in the hydrolysis of 7, no rearrangement to
produce the enone side chain was observed, that is, compound
12 was not detected as a product.
c
hydrogen tartrate and tartaric acid; t1/2 (hours) defined as the time taken for
damascenone concentration to reach half its final value. d Half-lives determined
by interpolation (e.g., Figure 6); errors estimated as ( 5%. e Defined as t1/2 glyc
/
t1/2 aglyc
.
al. (12). In every case, the final products of reaction, observable
by GC/MS, were damascenone (1) plus small amounts of both
the C-9 hydrate 14 and the C-9 ethanol adduct 15. In all
hydrolysates, the final quantified amount of damascenone
corresponded to approximately 70% of the theoretical maximum.
The evolution of damascenone from the propargyl alcohol 6
at both pH 3.0 and pH 3.2 over time is shown in Figure 5.
From these data, the times for the concentration of damascenone
to reach half of the maximum value (t1/2) were determined as
40 and 65 h, respectively. Similar graphs were constructed for
all hydrolyses. Table 1 lists the t1/2 values for the formation of
damascenone from 6 and from the allenic precursor 5, as well
as the corresponding glycosides, 9 and (9R)- and (9S)-10 at both
pH values. These provide ratios (t1/2glyc/t1/2aglyc) of 1.6 at pH
3.0 and 1.7 at pH 3.2 for 9:5 and 2.1 at pH 3.0 and 2.2 at pH
3.2 for 10:6. Both isomers of 10 were converted to damascenone
at an identical rate.
DISCUSSION
Hydrolysis of the Allenes 7 and 8. The hydrolyses of 7 and
8 were conducted in model wine solution (10% ethanol in water,
pH 3.0) at both 25 and 45 °C. The half-life of 7 at 25 °C was
approximately 10 days. For both reaction temperatures, each
pair of isomers of the hydroxy allene 7 produced product
mixtures that were identical to one another by NMR and achiral
phase GC. These comprised >90% of the total product and were
identified as the isomeric 7,8-didehydrotheaspiranes (24, Figure
We had earlier shown that the allene diol 5 and the acetylenic
alcohol 6 were both intermediates in the hydrolytic conversion,
at room temperature and pH 3.0, of allenic triol 2 into
damascenone (1). They were not, however, intermediates in the
formation of the two major hydrolysis products of 2, that is,
3-hydroxydamascone (3) and the enyne diol 4 (Figure
1A) (9, 10). The alcohol 6 was also observed as an intermediate
in the hydrolytic conversion of 5 into 1 (Figure 1B) (10). The

9188 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Daniel et al.
Figure 7. Proposed route for the formation of all of the different products of the hydrolysis of allene 2.
hydrolytic data reported here show that the rates of formation
of 1 from 6 (half-lives (t1/2) of 32 and 48 h at pH 3.0 and 3.2,
respectively) are lower than those from 5 (half-lives of 40 and
65 h, respectively at the same pH values) (Table 1). This
indicates that at least two separate pathways might be involved
in the conversion of 5 to damascenone and that acetylene 6 is
not an obligatory intermediate in this conversion (18).
ygen from secondary allenic alcohols to form R,ꢀ-unsaturated
ketones does not take place and only occurs with highly
conjugated acetylenic precursors such as 6 or tertiary allenic
alcohols such as 2 (8). Small amounts of the hydrate 14 and
ether 15 observed in some hydrolysates appear to be end
products rather than intermediates in these transformations as
their reactivity at pH 3 is relatively low.
Presumably, the delocalized C-3 cation 27 (Figure 7) formed
by loss of the C-3 hydroxyl in 5 subsequently either loses the
C-8 proton to yield the acetylenic alcohol 6, or undergoes
hydration at C-7 followed by tautomerism (with concomitant
loss of water) to directly give damascenone (1).
The hydrolytic experiments with the model allenic alcohols
7 and 8, which are 3-deoxy analogues of the polyols 5 and 2,
respectively, indicate that the conversion of simple secondary
R-hydroxy allenes to R,ꢀ-unsaturated ketones under mild acid
conditions is much slower than the reaction of the damascenone
precursors 2, 5, and 6. Even when reaction at C-9 took place,
as in the formation of the ethyl ether 11 from the alcohol 8, no
transposition of oxygen from C-9 to C-7 was observed. It seems
unlikely, therefore, that such transpositions account for dama-
scenone formation in nature.
These observations, together with the results of the earlier
hydrolytic experiments (7, 9, 10) give a clearer picture of the
formation of the various end products from the allenic triol 2.
The formation of all such products is presumably initiated
exclusively by acid-catalyzed breakage of the tertiary C-5
carbon-oxygen bond (Figure 7) to give cation 25. Loss of the
C-8 proton would lead to formation of the enynediol 4, while
hydration at C-7 would ultimately give 3-hydroxydamascone
(3). This mechanism for the formation of these two main
hydrolysis products is analogous to the two mechanisms
proposed above for the formation of damascenone from diol 5.
The formation of diol 5 from triol 2 is relatively minor and
results from the loss of the C-4 proton. The hydrolytic studies
with the model compounds 7 and 8 indicate that, under mild
acid conditions and room temperature, 1,3-transposition of ox-
Effects of Glycoconjugation. We have previously compared
the rates of hydrolysis of alcohols and their corresponding
glycosides (5). In the case of geraniol and its O-ꢀ-D-glucoside,
introduction of the sugar unit produced an approximate 10-fold
reduction in the rate of linalool formation. Similarly, the rate
of formation of damascenone from the enyne diol 4 was some
8 times more rapid than from its C-9 O-ꢀ-D-glucoside 13.
However, direct comparison of the these rate ratios between
geraniol/geranyl glucoside and 4/13 is not strictly appropriate,
as in the latter case damascenone is the result of a multistep
sequence of transformations, and without direct knowledge of
the kinetics of the individual intervening steps, the actual effect
of the sugar on the ionization of the C-9 alcohol function is
impossible to determine accurately.
The effect of glycosylation on the rate of damascenone
formation from the acetylenic alcohol 6 (an approximately 2-fold
reduction) was small. Again, as it is not known whether cleavage
of the C-9-oxygen bond is rate determining, it is not possible
to estimate the effect of glucosylation on this step. The other
damascenone precursor studied, 5, showed a rate reduction of
1.6 upon introduction of glucose at C-3. This ratio was also
determined by monitoring the production of damascenone,
which again, is the product of several interconversions. In our
previous study concerning the hydrolysis of the aglycones 5
only (10), it was observed that when diastereomerically pure
samples of 5 were hydrolyzed in model wine (pH 3.0) and the
reaction was interrupted early (i.e., before one half-life had
elapsed), the starting material was accompanied by a second
diastereomer of 5. Chiral GC analysis showed that this new
diastereomer had arisen from exclusive epimerization at C-3.

Synthesis and Hydrolysis of Damascenone Precursors
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9189
from carotenoids. HelV. Chim. Acta 1973, 56, 1514–1516.
(5) Skouroumounis, G. K.; Sefton, M. A. Acid-catalyzed hydrolysis
of alcohols and their ꢀ-D-Glucopyranosides. J. Agric. Food Chem.
2000, 48, 2033–2039.
(6) Sefton, M. A. Hydrolytically-released volatile secondary metabo-
lites from a juice sample of Vitis Vinifera grape cvs Merlot and
Cabernet Sauvignon. Aust. J. Grape Wine Res. 1998, 4, 30–38.
(7) Skouroumounis, G. K.; Massy-Westropp, R. A.; Sefton, M. A.;
Williams, P. J. Precursors of damascenone in fruit juices.
Tetrahedron Lett. 1992, 33, 3533–3536.
The fact that the proportions of the original diastereomer and
the new epimer were equal indicated that complete epimerization
had occurred and that epimerization was occurring much more
rapidly than conversion into damascenone (1). Thus, despite
the latter step(s) being rate-determining, the fact that glycosy-
lation at C-3 influenced the rate of formation of damascenone
indicates that the true effect of introduction of the sugar on
reactivity at this position could be much larger than the observed
rate ratio of 1.6 would suggest.
Most polyol precursors to volatile aroma compounds ac-
cumulate in plants as glycoconjugates, and while chemical
studies on aglycones are useful in understanding precursor-
product relationships, it is the behavior of these glycoconjugates
that will determine the rate and extent of formation of important
odorants such as damascenone. Our studies indicate that the
effect of glycosylation on the behavior of the allenic triol in
mild aqueous acid might not be important on the time scale of
wine making and maturation. The effect of glycosylation at C-3
or C-9 of the triol 2 on product distribution remains undeter-
mined. Conceivably, either steric or inductive effects could
influence the degree to which glycosylation of the triol affects
the competing steps shown in Figure 7. However, the evidence
to hand suggests that glycosylation is unlikely to affect product
distribution by influencing the relative reactivity of the three
carbon-oxygen bonds in 2 as we had previously specu-
lated (10).
(8) Olsson, L-I.; Claesson, A.; Bogentoft, C. Allenes and acetylenes.
II. Acid catalyzed reactions of R-allenic and R-acetylenic tertiary
alcohols. Acta Chem. Scand. 1973, 27, 1629–1636.
(9) Puglisi, C. J.; Elsey, G. M.; Prager, R. H.; Skouroumounis, G. K.;
Sefton, M. A. Identification of a precursor to naturally occuring
ꢀ-damascenone. Tetrahedron Lett. 2001, 42, 6937–6939.
(10) Puglisi, C. J.; Daniel, M. A.; Capone, D. L.; Elsey, G. M.; Prager,
R. H.; Sefton, M. A. Precursors to damascenone: synthesis and
hydrolysis of four isomeric 9-dihydroxymegastigma-4,6,7-trienes.
J. Agric. Food Chem. 2005, 53, 4895–4900.
(11) Janusz, A.; Capone, D. L.; Puglisi, C. J.; Perkins, M. V.; Elsey,
G. M.; Sefton, M. A. (E)-1-(2,3,6-Trimethylphenyl)buta-1,3-diene:
a potent grape-derived odorant in wine. J. Agric. Food Chem.
2003, 51, 7759–7763.
(12) Daniel, M. A.; Elsey, G. M.; Capone, D. L.; Perkins, M. V.;
Sefton, M. A. Fate of damascenone in wine: the role of SO2. J.
Agric. Food Chem. 2004, 52, 8127–8131.
(13) Koenigs, W.; Knorr, E. Derivatives of grape sugar and galactose.
Ber. Dtsch. Chem. Ges. 1901, 34, 957–981.
(14) Kunz, H.; Harreus, A. Glycosidsynthese mit 2,3,4,6-tetra-O-
pivaloyl-R-D-glucopyranosylbromid. Liebigs Ann. Chem. 1982,
41–48.
(15) Tanake, A.; Yamamoto, H.; Oritani, T. Lipase-catalyzed asym-
metric synthesis of (R) and (S)-4-tert-butyldimethylsilyloxy-2,6,6-
trimethyl-2-cyclohexenone and their dihydro derivatives. Tetra-
hedron: Asymmetry 1995, 6, 1273–1278.
(16) Desmares, G.; Lefebvre, D.; Renevret, G.; Le Drian, C. Selective
formation of ꢀ-D-glucosides of hindered alcohols. HelV. Chim.
Acta 2001, 84, 880–889.
(17) Heilmann, W.; Azizur-Rahman, Bauml, E.; Mayr, H. Carboca-
tionic cyclizations and rearrangements in the damascone series.
Tetrahedron 1988, 44, 6047–6054.
ACKNOWLEDGMENT
We are grateful to Dr. K. Puntener, Hoffmann-La Roche
International, for providing a sample of authentic (S)-phorenol
(ee >96%). We wish to thank colleagues from the The
Australian Wine Research Institute, particularly Drs. G. K.
Skouroumounis and M. J. Herderich for helpful discussions and
feedback.
Supporting Information Available: 1H NMR, HRMS (ESI),
and 13C NMR data for the compounds studied. This material is
available free of charge via the Internet at http://pubs.acs.org.
(18) Edens, M.; Boerner, D.; Chase, C. R.; Nass, D.; Schiavelli, M. D.
Mechanism of the Meyer-Schuster rearrangement. J. Org. Chem.
1977, 42, 3403–3408.
LITERATURE CITED
(1) Pickenhagen, W. In FlaVor Chemistry - Thirty Years of Progress;
Teranishi, R., Wick, E. L., Hornstein, I., Eds.; Kluwer Academic/
Plenum Publishers: New York, 1999; pp 75-87.
(2) Winterhalter, P., Rouseff, R. Carotenoid-DeriVed Aroma Com-
pounds: An Introduction; Winterhalter, P., Rouseff, R., Eds.; ACS
Symposium Series 802; American Chemical Society: Washington,
DC, 2002; pp 1-17.
(3) Ohloff, G.; Rautenstrauch, V.; Schulte-Elte, K. H. Model reactions
for the biosynthesis of damascone compounds and their preparative
application. HelV. Chim. Acta 1973, 56, 1503–1513.
(4) Isoe, S.; Katsumura, S.; Sakan, T. The synthesis of damascenone
and ꢀ-damascone and the possible mechanism of their formation
Received for review June 12, 2008. Revised manuscript received August
5, 2008. Accepted August 6, 2008. This project was supported by
Australia’s grapegrowers and winemakers through their investment
body, the Grape and Wine Research and Development Corporation,
with matching funds from the Australian Government, and by the
Commonwealth Cooperative Research Centres Program. The work was
conducted by The Australian Wine Research Institute and Flinders
University as part of the research program of the Cooperative Research
Centre for Viticulture.
JF8018134