Acid-catalyzed hydrolysis of alcohols and their β-D-glucopyranosides

Article abstract of DOI:10.1021/jf9904970

The hydrolysis, in model wine at pH 3, of the allylic, homoallylic, and propargylic glycosides, geranylβ-D-glucopyranoside, [3'-(1'-cyclohexenyl)- 1'-methyl-2'-propynyl]-β-D-glucopyranoside, (3'RS,9'SR)(3'-hydroxy-5'- megastigmen-7-yn-9-yl)-β-D-glucopyranoside, (3',5',5'-trimethyl-3'- cyclohexenyl)-β-D-glucopyranoside, E-(7'-oxo-5',8'-megastigmadien-3'-yl)-β- D-glucopyranoside (3-hydroxy-β-damascone-β-D-glucopyranoside), and their corresponding aglycons has been studied. In general, aglycons were more rapidly converted to transformation products than were the corresponding glucosides. Glycoconjugation of geraniol in grapes is a process that reduces the flavor impact of this compound in wine, not only because geraniol is an important flavor component of some wines but also because the rate of formation of other flavor compounds from geraniol during bottle-aging is reduced. However, when flavor compounds such as β-damascenone are formed in competition with flavorless byproducts, such as 3-hydroxy-β-damascone, by acid-catalyzed hydrolytic reactions of polyols, then glycoconjugation is a process that could enhance as well as suppress the formation of flavor, depending on the position of glycosylation. (3'RS,9'SR)-(3'-Hydroxy-5'- megastigmen-7'-yn-9'-yl)-β-D-glucopyranoside hydrolyzed more slowly but gave a higher proportion of β-damascenone in the products than did the aglycon at 50 °C. Reaction temperature also effected the relative proportion of the hydrolysis products. Accelerated studies do not parallel natural processes precisely but only approximate them.

Full text of DOI:10.1021/jf9904970

J. Agric. Food Chem. 2000, 48, 2033−2039
2033
Acid -Ca ta lyzed Hyd r olysis of Alcoh ols a n d Th eir
â-D-Glu cop yr a n osid es†
George K. Skouroumounis*,‡,# and Mark A. Sefton#
Department of Chemistry, The University of Adelaide, Adelaide, Australia 5005, and The Australian Wine
Research Institute, P.O. Box 197, Glen Osmond, SA, Australia 5064
The hydrolysis, in model wine at pH 3, of the allylic, homoallylic, and propargylic glycosides, geranyl-
â-D-glucopyranoside, [3′-(1′′-cyclohexenyl)-1′-methyl-2′-propynyl]-â-D-glucopyranoside, (3′RS,9′SR)-
(3′-hydroxy-5′-megastigmen-7-yn-9-yl)-â-D-glucopyranoside, (3′,5′,5′-trimethyl-3′-cyclohexenyl)-â-D-
glucopyranoside, E-(7′-oxo-5′,8′-megastigmadien-3′-yl)-â-D-glucopyranoside (3-hydroxy-â-damascone-
â-D-glucopyranoside), and their corresponding aglycons has been studied. In general, aglycons were
more rapidly converted to transformation products than were the corresponding glucosides.
Glycoconjugation of geraniol in grapes is a process that reduces the flavor impact of this compound
in wine, not only because geraniol is an important flavor component of some wines but also because
the rate of formation of other flavor compounds from geraniol during bottle-aging is reduced.
However, when flavor compounds such as â-damascenone are formed in competition with flavorless
byproducts, such as 3-hydroxy-â-damascone, by acid-catalyzed hydrolytic reactions of polyols, then
glycoconjugation is a process that could enhance as well as suppress the formation of flavor,
depending on the position of glycosylation. (3′RS,9′SR)-(3′-Hydroxy-5′-megastigmen-7′-yn-9′-yl)-â-
D-glucopyranoside hydrolyzed more slowly but gave a higher proportion of â-damascenone in the
products than did the aglycon at 50 °C. Reaction temperature also effected the relative proportion
of the hydrolysis products. Accelerated studies do not parallel natural processes precisely but only
approximate them.
Keyw or d s: â-Damascenone; geraniol; monoterpenes; norisoprenoids; hydrolysis; glycosides; precur-
sors
INTRODUCTION
Some of the hydrolytic data have also been reported in
a preliminary form (Skouroumounis et al., 1992, 1993).
In wines and other beverages, the development of
flavor is of great importance to the consumer. Flavor is
derived from a range of compounds, differing in chemical
functionality. These compounds, which originate from
primary or secondary metabolic pathways in grape
berries or other fruits, can be formed by many routes
such as the mevalonic acid, shikimate, and polyketide
pathways as well as from carotenoid breakdown prod-
ucts (Ohloff et al., 1973; Sefton et al., 1989; Williams et
al., 1992; Razungles et al., 1993). Many compounds
apparently important to wine flavor such as â-dama-
scenone (1) are thought to be derived from the hydrolytic
breakdown of grape-derived conjugates (mainly glyco-
conjugates) of complex secondary metabolites (Francis
et al., 1992, 1999,; Sefton et al., 1993, 1994; Winter-
halter and Skouroumounis, 1996), yet little is known
about the comparative hydrolytic behavior of such
conjugates and their corresponding aglycons.
EXPERIMENTAL PROCEDURES
P r ep a r a tion of Su bstr a tes. The alcohols (2-6) and the
corresponding â-D-glucopyranosides (2a -6a ) (Figure 1) were
obtained or synthesized as described previously by Skourou-
mounis et al. (1995) except that 3-hydroxy-â-damascone (6)
was synthesized by refluxing the acetylenic triol (26), (Massy-
Westropp et al., 1992) in 0.1 N HCl for 1 h. The reaction
mixture was cooled and extracted with dichloromethane. The
aqueous layer was then salted out (sodium chloride), unreacted
acetylenic triol was extracted with ethyl acetate and dried, and
the above procedure was repeated until only trace amounts of
acetylenic triol (26) were left. The dichloromethane extracts
were combined and evaporated, and the oil was directly
chromatographed under standard conditions (Still et al., 1978)
using diethyl ether as solvent, giving 3-hydroxy-â-damascone
(6) as a clear oil (70%) and â-damascenone (1) (16%). 1-(1′-
Cyclohexenyl)-2-buten-1-one (21) was synthesized according
to the method of Richter and Otto (1990). All chiral substrates
used in the hydrolytic studies were racemates.
P r ep a r a tion of Hyd r olysa tes. Hydrolytic studies of sub-
strates were carried out with 10% aqueous ethanolic solutions
(model wines). The hydrolyses were conducted at pH 1.0, 1.1,
or 3.0 in sealed ampules. Typical wine pH varies from 2.8 to
4.0. Model wine hydrolyses were conducted at pH 3.0. Before
the ampules were sealed, the solutions were deoxygenated
with a stream of oxygen-free dry nitrogen (Shriver, 1969) for
15 s. All hydrolyses were done in duplicate. A constant-
temperature oven (Contherm Incubator) was used for experi-
ments conducted at 50 °C over extended periods (>2 days).
The temperature was monitored with a standard thermometer.
Constant-temperature oil baths were used at 50, 80, or 100
This paper reports the mild acid-catalyzed hydrolyses
of some model and naturally occurring allylic, homoal-
lylic, and propargylic alcohols and their â-D-glucopyran-
osides. The synthesis of these â-D-glucopyranosides has
been described previously (Skouroumounis et al., 1995).
† The authors dedicate the manuscript to the memory of the
late Evangelos Cotsaris, a fellow scientist.
* Corresponding author (e-mail gskourou@awri.adelaide.
edu.au).
‡ The University of Adelaide.
#
The Australian Wine Research Institute.
10.1021/jf9904970 CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/13/2000

2034 J. Agric. Food Chem., Vol. 48, No. 6, 2000
Skouroumounis and Sefton
Ta ble 1. P r od u cts of th e Hyd r olysis of Ger a n iol (2) a t 50
a n d 80 °C, p H 3, in 10% Aqu eou s Eth a n ol
50 °C yields (%)
80 °C yields (%)
producta
1 h 4 h 8 h 24 h 1 h 4 h 8 h 24 h
2,2,6-trimethyl-2-vinyl-
tetrahydropyran (7)b
myrcene (8)b
-
-
-
+
0.1 0.3 0.9 3.7
+
-
-
-
-
0.1 0.1 0.2 0.2 0.5 0.8 1.7
limonene (9)b
-
+
-
+
0.1 0.3 0.9 2.2
0.1 0.2 0.3 0.6 1.1 1.7
1.8 2.7
0.1 0.3 1.1 1.8
(Z)-ocimene (10)b
(E)-ocimene (11)b
terpinolene (12)b
linalool (13)
0.1 0.1 0.2 0.3
1
-
-
+
2.5 6.2 13 28 25 46 43 13
0.3 0.4 0.5 1.4 1.1 7.5 20 44
R-terpineol (14)
nerol (15)b
+
0.1 0.4 0.7 0.8
2
3.3 4.2
geraniol (2)
97 93 85 67 71 37 17 2.7
2,6-dimethyl-7-octene-
+
0.1 0.2 0.6 0.3 2.5 6.2 11
2,6-diol (16)b
cis-1,8-terpin (18)b
trans-1,8-terpin (17)b
(Z)-3,7-dimethyl-2-octene-
1,7-diol (19)b
-
-
+
-
-
+
-
-
-
-
-
-
+
-
-
0.5 4.3
0.5
0.2 0.7 2.9
-
0.1 0.1
(E)-3,7-dimethyl-2-octene-
-
-
0.7 1.6 0.6 2.3 2.5
3
1,7-diol (20)b
F igu r e 1. Aglycons and their â-D-glucosides. Relative stere-
ochemistry only.
a
Assignments of all compounds have been established previ-
b
ously in this laboratory by Strauss (1983). Concentration based
on the assumption that components give the same GC-FID
peakarea/mg as the internal standard. -, not detected; +, <0.1%.
Coefficients of variation for all analyses were <12%.
°C for hydrolyses carried out over shorter periods. The
hydrolytic media were made up as follows: model wines at
pH 1.0; (+)-tartaric acid was dissolved in 10% aqueous ethanol,
and the pH was adjusted with sulfuric acid (1 M); pH 1.1, 0.1
N HCl solution; pH 3.0, 10% aqueous ethanol was saturated
with potassium hydrogen tartrate, and the pH was adjusted
with sulfuric acid (1 M); pH was determined with a pH meter,
standardized with a reference solution (Weast, 1979-1980).
For D2O experiments, pH 1.0 or 3.0, tartaric acid was dissolved
in D2O and the water was evaporated under reduced pressure.
This was repeated twice. Finally, the tartaric acid (d4) was
dissolved in D2O and the pH adjusted with D2SO4 for pH 1.0
and with anhydrous potassium carbonate for pH 3.0. The
aglycon (6) used for hydrolytic studies in D2O was dissolved
in deuterated ethanol (EtOD), and solvent was evaporated.
This was repeated three times to ensure complete deuterium
exchange. Finally, hydrolysis was carried out under model
wine conditions.
The following amounts of aglycon or glycoside (milligrams
per milliliter of hydrolytic medium) and internal standard (I.S.)
were used: geraniol(2) or its â-D-glucopyranoside (2a ) (each 1
mg/mL) with 2-phenylethanol (I.S., 1 mg/mL); enyne alcohol
(3), the â-D-glucopyranoside (3a ), 3,5,5-trimethyl-3-cyclohexen-
1-ol (5), or the â-D-glucopyranoside (5a ) (each 1 mg/mL, with
n-butyl benzene [I.S., 1 mg/mL]); enynediol (4) (0.5 mg/mL),
the â-D-glucopyranoside (4a ) (2 mg/mL), or the â-D-glucopy-
ranoside (6a ) (1 mg/mL), each with octadecanol (I.S., 1 mg/
mL). All internal standards were added as dichloromethane
solutions after hydrolysis but before workup.
Ta ble 2. P r od u cts of Hyd r olysis of
Ger a n iol-â-D-glu cop yr a n osid e (2a ) a t 50 a n d 80 °C, p H 3,
in 10% Aqu eou s Eth a n ol
50 °C yields (%)
80 °C yields (%)
producta
1 h 4 h 8 h 24 h 1 h 4 h 8 h 24 h
2,2,6-trimethyl-2-vinyl-
tetrahydropyran (7)b
myrcene (8)b
-
-
-
-
-
-
0.1 0.6
0.1
0.1 0.3
0.1 0.2
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
limonene (9)b
(Z)-ocimene (10)b
(E)-ocimene (11)b
terpinolene (12)b
linalool (13)
0.1 0.1 0.3
0.3
-
-
0.1 0.2 0.5 0.5 1.5 2.9 4.6 3.8
R-terpineol (14)
0.4 0.4 0.5 0.5
1
-
0.8 1.8 5.3
0.1 0.3 0.4
nerol (15)b
-
-
-
-
geraniol (2)
0.15 0.15 0.1
-
0.2
-
-
-
-
-
2,6-dimethyl-7-octene-
-
0.4 0.3 0.2
0.6 2.2
2,6-diol (16)b
cis-1,8-terpin (18)b
trans-1,8-terpin (17)b
(Z)-3,7-dimethyl-2-octene-
1,7-diol (19)b
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
0.3
-
0.3
(E)-3,7-dimethyl-2-octene-
-
-
-
-
-
-
-
-
1,7-diol (20)b
a
Assignments of all compounds have been established previ-
b
ously in this laboratory by Strauss (1983). Concentration based
An a lysis of Hyd r olysa tes. Hydrolysates were worked up
by cooling the ampules, adding the internal standard, and then
extracting with dichloromethane (2 × 1 mL). The dichlo-
romethane layer was washed with saturated sodium hydrogen
carbonate (1 mL), dried (MgSO4), filtered, and injected directly
into the GC for analysis. For the hydrolyses of enynediol (4)
the dichloromethane extract was concentrated by distillation
through a column of Fenske’s helices prior to analysis.
Analyses were performed with a Varian 3300 gas chromato-
graph coupled to either a Finnigan TSQ 70 mass spectrometer
or a flame ionization detector. The column used was a 30 m
J &W DB-1701 fused silica column (0.25 mm i.d. and 0.25 µm
film thickness) with helium carrier gas at a linear velocity of
40 cm/s. Injections were made with the injector at 200 °C, and
the splitter, with a split ratio of 1:10, was opened after 0.3
min. Flame ionization detection was used for compound quan-
tification with a Hewlett-Packard autointegration system. For
identification, hydrolysates were analyzed by GC-MS under
the following conditions: The column was held for 1 min at
50 °C or at 100 °C for compounds 4, 4a , 6 and 6a , then pro-
on the assumption that components give the same GC-FID
peakarea/mg as the internal standard. -, not detected. Coefficients
of variation for all analyses were <12%.
grammed at 4 °C/min to 250 °C, and held at that temperature
for 20 min. Electron impact spectra were taken at 70 eV.
Id en tifica tion of Hyd r olysis P r od u cts. 1H NMR spectra
were recorded with a J EOL J NM-PMX (60 MHz) or a Bruker
CXP300 (300 MHz) spectrometer. All compounds in Tables 1
and 2 (Figure 2) have been identified previously by Strauss
(1983). 1-(1′-Cyclohexenyl)-2-buten-1-one (21) was compared
with the synthetic sample: 1H NMR (60 MHz/CDCl3) δ 7.1-
6.6 (m, 3H), 2.4-2.2 (m, 4H), 1.9 (d, J ) 5 Hz, 3H), 1.8-1.6
(m, 4H); m/z 151 (8), 150 (46), 149 (8), 136 (9), 135 (100), 122
(6), 121 (11), 117 (11), 109 (9), 108 (6), 107 (17), 93 (6), 91 (7),
81 (19), 79 (24), 77 (6), 69 (45), 53 (5), 41 (38), 39 (14).
Compounds 29 and 30 were identified by comparison with
authentic materials (Sefton et al., 1989). Compounds 22, 24a ,
25a , 27a -d , and 28 were tentatively assigned on the basis of
interpretation of their mass spectral data. Compound 22: m/z

Acid Hydrolysis Glucopyranosides
J. Agric. Food Chem., Vol. 48, No. 6, 2000 2035
F igu r e 2. Products from hydrolysis of geraniol. Relative
stereochemistry only.
F igu r e 3. Compound numbers as referred to in text. Relative
stereochemistry only.
168 (23), 153 (5), 150 (8), 126 (4), 125 (36), 124 (7), 99 (7), 98
(100), 83 (25), 81 (3), 79 (5), 77 (3), 71 (3), 70 (37), 69 (9), 67
(4), 56 (3), 55 (15), 53 (3). Ethyl ether 24a (diastereoisomer 1)
m/z 254 (8), 239 (8), 208 (13), 195 (11), 193 (6), 175 (5), 168
(13), 167 (100), 166 (8), 153 (6), 151 (18), 149 (10), 123 (8), 122
(9), 121 (34), 113 (7), 95 (7), 93 (9), 86 (8), 79 (7), 73 (35), 69
(14), 67 (7), 55 (14), 45 (51), 43 (20), 41 (14); (diastereoisomer
2) m/z 254 (7), 239 (8), 209, (9), 208 (20), 195 (10), 193 (13),
175 (8), 168 (13), 167 (100), 166 (9), 153 (8), 151 (20), 149(11),
140 (12), 139 (11), 138 (7), 126 (15), 125 (9), 123 (14), 122 (10),
121 (39), 113 (7), 107 (6), 95 (10), 93 (10), 91 (6), 81 (6), 79
(11), 77 (8), 73 (33), 69 (13), 67 (9), 55 (17), 45 (48), 43 (26), 41
(23). Ethyl ether 25a : m/z 236 (7), 221 (7), 192 (3), 175 (5),
150 (5), 149 (51), 134 (3), 133 (36), 131 (3), 123 (11), 122 (100),
121 (59), 120 (8), 119 (9), 113 (6), 107 (25), 106 (4), 105 (19),
103 (5), 93 (4), 91 (14), 83 (7), 79 (10), 78 (3), 77 (9), 73 (44), 69
(5), 65 (4), 55 (3), 49 (4), 45 (57), 43 (8), 41 (9). Compound 27a :
m/z 122 (37), 107 (100), 91 (50), 79 (14), 49 (17). 27b: m/z 122
(38), 107 (100), 105 (18), 91 (52), 84 (25), 79 (17), 77 (12), 51
(12). 27c: m/z 122 (40), 107 (100), 93 (47), 91 (58), 86 (45), 79
(85), 77 (35), 66 (30), 49 (58). 27d : m/z 122 (15), 107 (100),
105 (33), 92 (16), 91 (65), 86 (18), 84 (33), 79 (16), 77 (15), 68
(17), 49 (25). 28 m/z 140 (8), 125 (32), 109 (8), 107 (18), 97
(25), 82 (100), 69 (22), 55 (25). Compounds 24 and 25 have
been tentatively identified previously (Sefton et al., 1989).
3-Hydroxy-â-damascone (6) at pH 1.0 for 8 h at 100 °C in D2O/
10% EtOD: 1H NMR (300 MHz/CDCl3) δ 6.75-6.65 (b, 1H),
6.14 (d, J ) 15.8 Hz, 1H), 4.1 (m, 1H), 2.34 (ddd, J ) 16.6,
5.9, 1 Hz, 1H), 2.0 (dd, J ) 16.6, 9.9 Hz, 1H), 1.72 (m, 1H),
1.55 (s, 3H), 1.48-1.44 (m, 1H), 1.14 (s, 3H), 0.97 (s, 3H). The
C10 methyl doublet at δ 1.72 was replaced by a complex
multiplet resulting from partial deuterium incorporation at
C10: m/z 211 (7), 210 (27), 209 (38), 208 (16), 195 (15), 194
(32), 193 (42), 177 (29), 176 (54), 175 (39), 150 (25), 149 (27),
122 (29), 121 (100), 107 (19), 105 (30), 93 (27), 91 (20), 79 (19),
72 (13), 71 (62), 70 (95), 69 (54), 55 (31). The above compounds
are represented in Figure 3.
RESULTS AND DISCUSSION
Hyd r olysis of Ger a n iol (2) a n d Its Glu cop yr a n -
osid e (2a ). Hydrolysis of the simple allylic alcohol
geraniol (2) and its glucopyranoside (2a ) was carried out
at a typical wine pH (3.0) at 50 and 80 °C. The results
are shown in Tables 1 and 2.
The hydrolysis of geraniol led ultimately to 14 prod-
ucts observed by gas chromatography (Figure 2). These
all increased in concentration with increasing temper-
ature and reaction time except for linalool (13), which
is an intermediate in the formation of R-terpeniol (14)
and other products from geraniol (2) (Baxter et al.,
1978). No ethyl ethers, including geraniol ethyl ether,
were observed in the hydrolysates, even though the
hydrolysis medium contained ethanol. The hydrolysis
of many monoterpenes under a variety of acid conditions
has previously been studied and reviewed by Baxter et
al. (1978) and Strauss and Williams (1983). The pattern
of products shown in Table 1 is consistent with results
previously reported (Baxter et al., 1978; Ohta et al.,
1991).
As a control for purity of starting material and for
possible artifacts of analysis, geranyl-â-D-glucopyrano-
side (2a ) was dissolved in 10% ethanol and the solution
extracted with dichloromethane (2 mL). Analyses by
GC-MS showed that 0.16% free geraniol (2) was present.
Hydrolysis of the glucoside (2a ) at 50 °C gave 3 products
only (Table 2), linalool (13), a larger amount of R-ter-
pineol (14), and 2,6-dimethyl-7-octene-2,6-diol (16),
compared to the 13 compounds identified in the aglycon
hydrolysis at this temperature. No additional geraniol
(2) was observed other than that present in the control

2036 J. Agric. Food Chem., Vol. 48, No. 6, 2000
Skouroumounis and Sefton
Ta ble 3. Hyd r olysis of Aglycon (3) a n d â-D-Glu cosid e (3a )
a t p H 3 a n d 100 °C
aglycon (3)
â-D-glucoside (3a )
product (% yield)
product (% yield)
time
3
21
22
3
21
22
4 h
8 h
24 h
46 h
98.5
94.8
83.5
73.7
1
3.9
13.6
21.5
0.5
1.3
2.9
4.8
4 h
8 h
24 h
46 h
0.9
1.9
5.3
5.9
0.5
1.4
5.6
7.4
trace
0.2
0.7
F igu r e 4. Proposed ionization of geraniol-â-D-glucopyranoside
0.9
during acid hydrolysis at pH 3.0.
a
Coefficients of variation for all analyses were <12%.
run. At 80 °C additional products were seen, along with
linalool (13) and R-terpineol (14) as the major products.
Again, no additional geraniol (2) other than that in the
control was observed, suggesting that under these
conditions primary allylic glucosides cleave at the ether
linkage and not at the glycosidic bond (Figure 4).
Sch em e 1. In ter p r eta tion of th e Ma ss Sp ectr u m of
Com p ou n d 22
Earlier studies reported by Williams et al. (1982) on
the hydrolysis of geranyl-â-D-glucopyranoside (2a ), un-
der more vigorous conditions, suggested that a substan-
tial amount of geraniol (2) was formed, although no
quantitative data were reported in that study. Their
results indicate that breakage of both the C1-O and
the glycosidic bond may have occurred at the higher
temperature. Strauss and Williams (1983) noted that
when glucopyranoside (2a ) was hydrolyzed (for 30 min
at an unspecified temperature), the proportion of total
monoterpene ethers to alcohols was approximately equal
to the molar ratio of ethanol to water present in the
reaction mixture. No ethyl ethers, in particular geranyl
ethyl ether, were detected in our study.
From the amounts of linalool (13) plus R-terpineol (14)
formed at 80 °C after 1 h from geraniol (2) or its
glucoside (2a ), the aglycon geraniol (2) reacts at a rate
∼10 times faster than its glucoside (2a ) (Tables 1 and
2). At the lower temperature (50 °C), the differences
were even greater. These results are in good agreement
with the recently published results of Ohta et al. (1991),
who reported that the conversion of geraniol-â-D-gluco-
side (2a ) to linalool (13) at 100 °C was ∼8 times slower
than the corresponding hydrolytic conversion of geraniol
(2). The reactivity of the glucoside (2a ) is contrary to
the belief expressed earlier (Williams et al., 1982, 1989)
that glycosides of allylic alcohols (e.g., geraniol) form
the allylic cation more readily than do the corresponding
aglycons and that geraniol (2) is a major product of the
hydrolysis of geranyl glucoside (2a ).
Hyd r olysis of Alcoh ols a n d Glycocon ju ga tes
Rela ted to â-Da m a scen on e (1) F or m a tion . Ohloff et
al. (1973) have shown that the acetylenic triol (26) gave
â-damascenone (1) and 3-hydroxy-â-damascone (6) upon
strong acid hydrolysis and suggested that the enyne diol
(4) was an intermediate in this conversion. Subse-
quently, the enyne diol (4) was confirmed as a grape/
wine constituent both in free and in glycoconjugated
forms (Sefton et al., 1989; Baderschneider et al., 1997).
Preliminary studies by us showed that both the aglycon
(4) and the glucoside (4a ) give â-damascenone (1) and
3-hydroxy-â-damascone (6) on mild hydrolysis (Skour-
oumounis et al., 1993). In the study reported here, the
role of glycoconjugation in modifying the hydrolytic
behavior of the enyne diol (4) was first examined by
studying the hydrolysis of the model propargyl alcohol
(3) and the corresponding glucoside (3a ). The results are
shown in Table 3.
give two major products, a ketone (21), formally derived
via a Meyer-Schuster rearrangement (Swaminathan
and Narayanan, 1971; Edens et al., 1977), and a second
product, tentatively identified from its mass spectrum
as the hydrate (22). The mass spectrum of compound
22 included a molecular ion at m/z 168 and no charac-
teristic m/z 109 ion peak via loss of 59 (C3H7O•),
expected for the C9 hydroxy ketone (23), but it did have
a base peak of 98, a strong m/z 70 (37%), and a
significant m/z 69 (9%) ion (Scheme 1).
The hydrolysis of the C9 glucoside (3a ) was found to
be slower than that of the aglycon (3). After 46 h, the
products were detected in 14% yield only, compared to
26% for the aglycon. The major products were the
aglycon (3), ketone (21), and â-hydroxyketone (22). It
is uncertain whether the formation of the aglycon (3)
arose via hydrolysis of the glycosidic or the ether
linkage. Thus, the magnitude of the reduction in the
rate of breakage of the ether bond, associated with
glycosylation of aglycon (3), could not be determined.
However, the formation of rearrangement products 21
and 22 was ∼3times faster with the aglycon (3) than
with the glucoside (3a ) (Table 3). The hydrolytic studies
clearly indicate that the model acetylenic alcohol (3) and
its â-glucoside (3a ) undergo relatively slow hydrolysis
and that the propargylic glucoside reacts more slowly
than its corresponding aglycon.
At pH 3 and 100 °C the reactivity of compound 3 was
low. After 46 h, ∼26% of compound 3 had reacted to

Acid Hydrolysis Glucopyranosides
J. Agric. Food Chem., Vol. 48, No. 6, 2000 2037
Ta ble 4. P r od u cts of th e Hyd r olysis of En yn ed iol (4) a n d Its C-9 Glu cop yr a n osid e (4a ) a t p H 3
aglycon (4)
â-D-glucoside (4a )a
product (% yield)
product (% yield)
% conversion
1
6
25
24
1
6
4
100 °C
4 h
1 day
2 days
50 °C
100 °C
4 h
23
81
98
2.8
10
14
20
71
83
-
trace
trace
-
trace
trace
0.6
5.0
2.1
33
1.9
1.8
2 days
50 °C
1 day
7 days
28 days
2
13
44
0.1
0.8
2.3
1.8
12
41
-
-
-
-
-
-
7 days
28 days
0.14
1.0
1.6
11
1.1
3.8
20 °C
28 days
90 days
1 year
2.6
6.4
22
trace
0.2
0.9
2.6
6.2
21
-
-
trace
1.8
-
-
trace
12
6.3 yearsb
82
2.6
61
a
Some of these data have been reported previously Skouroumounis (1993). -, not detected. Coefficients of variation for all analyses
b
were <12%. Trace quantities of ethyl ether of compounds 24 and 25 at C-9 were also tentatively identified.
Ta ble 5. Hyd r olysis of th e Aglycon (5), Glu cosid e (5a ),
a n d Glu cosid e (6a ) a t Va r iou s p H, Tim e, a n d
Tem p er a tu r e Con d ition s
The yields of products of the hydrolyses of the enyne
diol (4) and its glucoside (4a ) under various conditions
are shown in Table 4.
yielda (%)
Hydrolysis of the enyne diol (4) at pH 3 and 100 °C
gave â-damascenone (1) and 3-hydroxy-â-damascone (6)
as the major products. The latter is not converted to the
former under these conditions (see below), and so at any
given pH and reaction temperature the ratio of com-
pound 1 to compound 6 remained constant, regardless
of reaction time. Small quantities of the hydrates 24 and
25 (tentatively identified only) were observed with
prolonged hydrolysis. This indicates that, in the absence
of significant generation of â-damascenone (1) from one
or more precursors during aging, â-damascenone (1)
[and 3-hydroxy-â-damascone (6)] already present in
wine can decrease in concentration through formation
of this hydrate.
â-Damascenone (1) and 3-hydroxy-â-damascone (6)
were formed ∼8 times more slowly from the glucoside
(4a ) than from the aglycon (4) (Table 4). However, at
the lower temperature (50 °C) the proportion of â-dama-
scenone (1) formed was greater from the glucoside (4a )
than from the aglycon (4). It appears that by slowing
the acid-catalyzed ionization of the propargyl alcohol
function, the presence of the glycopyranosyl moiety
promotes other competing transformations within the
molecule. The side chain of 4a can react either directly
to give Meyer-Schuster rearrangement products or
indirectly via the aglycon (4).
As previously reported (Skouroumounis et al., 1993),
the ratio of â-damascenone (1):3-hydroxy-â-damascone
(6) in the products of the hydrolysis of compounds 4 or
4a is dependent on the reaction temperature. The ratio
is a reflection of the selectivity of ionization exhibited
at the different temperatures, with the proportion of
â-damascenone (1) in the products from both the aglycon
(4) and the glycoside (4a ) being greater at a higher
temperature. Thus, heating wine (or juice) is not only
likely to accelerate the generation of â-damascenone (1)
but could ultimately lead to a higher yield of this
compound than could be obtained by natural wine aging.
The reactivity of compounds 4 and 4a is significantly
greater than that of the analogues 3 and 3a , and this
is attributed to the extra methyl substituents which can
stabilize incipient carbocations via the inductive effect.
Hyd r olysis of 3,5,5-Tr im eth ylcyclo-3-h exen -1-ol
(5), 3-Hyd r oxy-â-d a m a scon e (6), a n d Th eir â-D-
Glu cop yr a n osid es (5a a n d 6a ). It has been suggested
conditions
aglycon (5)
pH 3, 100 °C, 24 h
pH 1, 50 °C, 24 h
pH 1, 80 °C, 24 h
pH 1, 100 °C, 8 h
glucoside (5a )
pH 3, 100 °C, 24 h
pH 1, 50 °C, 24 h
pH 1, 80 °C, 24 h
pH 1, 100 °C, 8 h
glucoside (6a )
27a -d
5
6
28 29 30
+, +, +, +
+, +, +, -
14, 15, 8.0, 11 22
38, 13, 12, 19 3.2
83
79
1.5 6.5 1.0
2.0 8.4 1.0
1.0 12 2.0
2.1 5.9 2.7
- - - -
- - - -
13
8.5
24
-
-
-
-
-
-
-
-
-
-
+, +, +, +
+
39, 10, 12, 19 10
4.5
pH 3, 100 °C, 24 h
pH 1.1, 100 °C, 8 h
16
35
a
-, not detected; +, trace amount, <0.1%. Coefficients of
variation for all analyses were <12%.
that 3-hydroxy-â-damascone (6) acts as a precursor of
â-damascenone (1) in nature (Ohloff, 1978). However,
Ohloff et al. (1973) have also demonstrated that 3-hy-
droxy-â-damascone (6) does not give â-damascenone (1)
under strongly acidic conditions but rather products
derived by Nazarov type cyclization (Rouvier and San-
telli, 1983). Nevertheless, nothing is known about the
behavior of the corresponding glucoside (6a ) under mild
acid conditions. Before studying this hydrolysis, we
examined the hydrolysis of 3,5,5-trimethylcyclo-3-hexen-
1-ol (5) and its glucoside (5a ), as a model for homoallylic
cyclohexenols at wine pH, as these functional groups
are common to many C13 norisoprenoids. Recently, the
glucoside (5a ) has been identified as a natural product
in saffron (Straubinger et al., 1998).
At pH 3 the aglycon (5) was relatively unreactive, and
even after 24 h at 100 °C, it was found that only 17%
had reacted. The main products, identified from their
mass spectra, were trace amounts of the isomeric dienes
of undetermined structure (27a -d ), the isomeric alcohol
(28), and small quantities of oxidation products 29 and
30 (Table 5). The â-glucoside (5a ) gave ∼13% of the
aglycon (5) under these conditions. Under more forcing
conditions (pH 1.0), various products were observed
(Table 5).
At 80 °C, 24 h, the aglycon (5) underwent dehydra-
tion, yielding a total of 48% diene formation, whereas
the glycoside (5a ) yielded trace amounts of dienes and
aglycon (5) (24%).

2038 J. Agric. Food Chem., Vol. 48, No. 6, 2000
Skouroumounis and Sefton
Sch em e 2. Hyp oth etica l P a th w a ys for th e Decom p o-
sition of a C-9 Glycocon ju ga te of th e Allen ic Tr iol 31
At pH 1 and 100 °C, virtually all of the aglycon (5)
had reacted after 8 h with multiple product formation.
Only 3.2% of the aglycon remained, and 27a -d ac-
counted for 82% of the products. The glycoside (5a ),
under the same conditions, yielded 80% dienes (27a -
d ) and 10% aglycon (5). It is evident that at the lower
temperatures there are differences between the reactiv-
ity of the glycoside and aglycon but at the highest
temperature (100 °C) the products were formed in
similar yields after 8 h, with dehydration products
predominant.
Although these results indicate that simple homoal-
lylic cyclohexenols and their glucosides are likely to have
low reactivity at wine pH, it is conceivable that the
additional conjugation with the enone side chain of â-D-
glucopyranoside (6a ) could enhance the reactivity of the
homoallylic glycosidic function. However, when the
hydrolysis of â-D-glucopyranoside (6a ) was carried out
at pH 3.0 and 100 °C for 24 h, 16% of aglycon,
3-hydroxy-â-damascone (6), was obtained but no trace
of â-damascenone (1) could be detected. At pH 1.1 and
100 °C for 8 h, only 3-hydroxy-â-damascone (6) was
obtained in 35% yield (Table 5). No elimination products
were seen with the hydrolysis of the glycoside (6a ), even
under forcing conditions. The reason for these differ-
ences is unclear. It is possible that the side chain is
easily enolized and that this favors the formation of
glycoconjugates of Nazarov cyclization products, which
are known to prevail under even more forcing conditions
(Rouvier and Santelli, 1983). To check this hypothesis,
hydrolysis of 3-hydroxy-â-damascone (6) was carried out
in deuterated solvents. After 4 h at 100 °C and pH 3.0,
3-hydroxy-â-damascone (6) was found to be stable and
there was no deuterium incorporation. However, at pH
1.0 and 100 °C for 80 min, Nazarov cyclization products
were evident in trace amounts and the 3-hydroxy-â-
damascone (6) incorporated deuterium, albeit only in
the side chain methyl and not in the ring. This was
alkyne (33) were tentatively identified as intermediates
in this transformation (Skouroumounis et al., 1992).
It was evident from these studies that the three end
products of allene triol (31) hydrolysis were formed from
several competing pathways. Glycosylation (which is
thought to be the terminal step in the biotransforma-
tions of secondary metabolites) is likely to influence the
relative importance of these pathways (e.g., Scheme 2).
Thus, the tentatively identified bifunctional intermedi-
ate (32) might eliminate a hydroxyl function, under acid
conditions, to form either the C3 or C9 carbocation
(megastigmane numbering; Aasen et al., 1972), the
former leading to the dienealkyne (33) and ultimately
to â-damascenone (1), and the latter process leading
mostly to 3-hydroxy-â-damascone (6) (Skouroumounis
et al., 1992). If either the C3 or C9 position is stabilized
by glycosylation, then ionization of the underivatized
hydroxyl can be favored. In the case of C9 glycosylation,
this would be expected to enhance â-damascenone (1)
formation.
Con clu sion . Allylic, propargylic, and homoallylic
alcohols are more reactive than their corresponding â-D-
glucopyranosides. Although the rates of hydrolysis of
such glucosides are significantly lower than for the
aglycons (possibly orders of magnitudes lower in some
cases), this may nevertheless have an important impact
on the direction of flavor release. In the case of floral
grape varieties, where geraniol (2) is concerned, glyco-
sylation is a process that lowers the flavor of the grape
juice/wine, not only because geraniol (2) itself is an
1
demonstrated by high-field H NMR spectroscopy, which
showed the collapse of both the side chain vinyl methyl
doublet and the vinyl proton at C8 to complex multip-
lets. The C4 multiplets remained unchanged. This
suggests that the enone side chain is rotated out of
conjugation with the ring double bond, presumably to
relieve steric interaction with the geminal dimethyl
groups. Thus, the hydrolytic behavior of 3-hydroxy-â-
damascone glucoside (6a ) was unlikely to be influenced
by enolization in the cyclohexenol ring.
Im p lica tion s for â-Da m a scen on e (1) F or m a tion
fr om th e Allen ic Tr iol (31). The multifunctional
allenic triol (31) (Scheme 2) is known to have a half-life
of <24 h at pH 3 and room temperature (Skouroumounis
et al., 1992, 1993), yet a glycoconjugate has been
islolated from Lyium halimifolium Mil. (Naf et al., 1990)
at the natural plant pH of 4.0 as well as from apples
Malus sylvestris Mill. cv. J onathan (Winterhalter et al.,
1993). The allenic triol (31) has also been observed in
enzyme hydrolysates of glycosidic fractions obtained
from Merlot grapes (Sefton, 1998). The allenic triol (31)
has been postulated as the natural precursor to â-dam-
ascenone (1) (Ohloff et al., 1973; Isoe et al., 1973), and
recent studies in our laboratories (Skouroumounis et al.,
1992) have shown the allenic triol (31) does give
â-damascenone (1) in small amounts, <1%, after 24 h
at 24 °C and pH 3.0. The two major products, however,
were 3-hydroxy-â-damascone (6) and enynediol (4).
Trace amounts of the allenic diol (32) and the diene-

Acid Hydrolysis Glucopyranosides
J. Agric. Food Chem., Vol. 48, No. 6, 2000 2039
Sefton, M. A.; Skouroumounis, G. K.; Massy-Westropp, R. A.;
Williams, P. J . Norisoprenoids in Vitis vinifera White Wine
Grapes and the Identification of a Precursor of Damascenone
in these Fruits. Aust. J . Chem. 1989, 42, 2071-2084.
Sefton, M. A.; Francis, I. L.; Williams, P. J . The volatile
composition of Chardonnay juices: a study by flavor precur-
sor analysis. Am. J . Enol. Vitic. 1993, 44, 359-370.
Sefton, M. A.; Francis, I. L.; Williams, P. J . Free and bound
volatile secondary metabolites of Vitis vinifera grape cv.
Sauvignon Blanc. J . Food Sci. 1994, 59, 142-147.
Shriver, D. F. The Manipulation of Air-Sensitive Compounds;
McGraw-Hill Series in Advanced Chemistry; McGraw-Hill:
New York, 1969; p 199.
important flavorant in some wines but because the rate
of formation of other flavor compounds from geraniol
(2) is also reduced. However, when flavor compounds
in wines are formed from polyfunctional aglycons in
competition with flavorless byproducts by hydrolytic
reactions, then glycoconjugation through its influence
of the reactivity of alcoholic functional groups can
influence the fate of the aglycon and can either enhance
or suppress the formation of flavor.
ACKNOWLEDGMENT
We thank Drs. R. A. Massy Westropp and P. J .
Williams and Professors P. Winterhalter and P. Høj for
their advice and encouragement.
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.
Skouroumounis, G. K.; Massy-Westropp, R. A.; Sefton, M. A.;
Williams, P. J . â-Damascenone Formation in J uices and
Wines. In Progress in Flavour Precursor Studies; Schreier,
P., Winterhalter, P., Eds.; Allured Publishing: Carol Stream,
IL, 1993; pp 275-278.
Skouroumounis, G. K.; Massy-Westropp, R. A.; Sefton, M. A.;
Williams, P. J . Synthesis of Glucosides Related To Grape
And Wine Aroma Precursors. J . Agric. Food Chem. 1995,
43, 974-980.
Still, W. C.; Kahn, M.; Mitra, A. A rapid chromatographic
technique for preparative separations with moderate resolu-
tion. J . Org. Chem. 1978, 43, 2923-2925.
Straubinger, M.; Bau, B.; Eckstein, S.; Fink, M.; Winterhalter,
P. Identification of Novel Glycosidic Aroma Precursors in
Saffron (Crocus sativus L.). J . Agric. Food Chem. 1998, 46,
3238-3243.
LITERATURE CITED
Aasen, A. J .; Kimland, B.; Almqvist, S.-O.; Enzell, C. R.
Tobacco chemistry 15. New tobacco constituentssthe struc-
tures of five isomeric megastigmatrienones. Acta Chem.
Scand. 1972, 26, 2573.
Baderschneider, B.; Skouroumounis, G. K.; Winterhalter, P.
Isolation of Two Glucosidic Precursors of â-Damascenone
From Riesling Wine. Nat. Prod. Lett. 1997, 10, 111-114.
Baxter, R. L.; Laurie, W. A.; McHale, D. Transformations Of
Monoterpenoids In Aqueous Acids. The Reactions Of Lin-
alool, Geraniol, Nerol and Their Acetates In Aqueous Citric
Acid. Tetrahedron 1978, 34, 2195-2199.
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.
Francis, I. L.; Sefton, M. A.; Williams, P. J . Sensory Descriptive
Analysis of Hydrolyzed Precursor Fractions from Semillon,
Chardonnay, and Sauvignon Blanc Grape J uices. J . Sci.
Food Agric. 1992, 59, 511-520.
Francis, I. L.; Kassara, S.; Noble, A. C.; Williams, P. J . The
Contribution of Glycosidic Precursors to Cabernet Sauvignon
and Merlot Aroma; Waterhouse, A. L., Ebeler, S. E., Eds.;
ACS Symposium Series 714; American Chemical Society:
Washington, DC, 1999; pp 13-30.
Isoe, S.; Katsumura, S.; Sakan, T. Neoxanthin breakdown.
Helv. Chim. Acta 1973, 56, 1514-1516.
Massy-Westropp, R. A.; Skouroumounis, G.; Snow, M. R.;
Tiekink, E. R. T. Crystal structure of (3′R*,1R*,4S*,6S*)-
1-(3′-hydroxybut-1′-ynyl)-2,2,6-trimethylcyclohexane-1,4-
diol. Z. Kristogr. 1992, 198, 151-154.
Naf, R.; Velluz, A.; Thommen, W. Isolation Of A Glucosidic
Precursor Of Damascenone From Lycium halimifolium Mil.
Tetrahedron Lett. 1990, 45, 6521-6522.
Ohloff, G. Importance of minor components in flavors and
fragrances. Perfum. Flavor. 1978, 3, 11-22.
Ohloff, G.; Rautenstrauch, V.; Schulte-Elte, K. H. Model
Reactions In The Biosynthesis Of Compounds In The
Damascone Series, And Their Preparative Application. Helv.
Chim. Acta 1973, 56, 1503-1513.
Ohta, T.; Morimitsu, Y.; Sameshima, Y.; Samuta, T.; Ohba,
T. Transformation from Geraniol, Nerol and Their Gluco-
sides into Linalool and R-Terpineol during Shochu Distil-
lation. J . Ferment. Bioeng. 1991, 72, 347-351.
Razungles, A.; Gunata, Z.; Pinatel, S.; Baumes, R.; Bayonove,
C. EÄ tude quantitative de compose´s terpe´niques, noriso-
pre´no¨ıds et de leurs pre´curseurs dans diverses varie´te´s de
raisins. Sci. Aliments 1993, 13, 59-72.
Richter, F.; Otto, H.-H. 1,4-Pentadien-3-ones. 31. 1-(1-Cyclo-
hexenyl)-2-propen-1-ones: synthesis, properties, and ste-
reoselective formation of 3-aryl-4-(1-cyclohexenylcarbonyl)-
perhydro-1-naphthalenones by sequential Michael reaction.
Liebigs Ann. Chem. 1990, 7-14.
Rouvier, C. S.; Santelli, M. The Nazarov Cyclization. Synthesis
1983, 429-442.
Sefton, M. A. Hydrolytically-released volatile secondary me-
tabolites from a juice sample of Vitis vinifera grape cv.
Merlot and Cabernet Sauvignon. Aust. J . Grape Wine Res.
1998, 4, 30-38.
Strauss, C. R. Structure And Chemistry Of Flavour Precursors
In Grapes. Ph.D. Thesis, University of Adelaide, Australia,
1983.
Strauss, C. R.; Williams, P. J . The effect of distillation on grape
flavour components. In Flavour of Distilled Beverages:
Origin and Development; Piggott, J . R., Ed.; Chichester, Ellis
Horwood: Stirling, Scotland, 1983; pp 120-133.
Swaminathan, S.; Narayanan, K. V. The Rupe And Meyer-
Schuster Rearrangements. Chem. Rev. 1971, 71, 429-438.
Weast, R. C. CRC Handbook of Chemistry and Physics, 60th
ed.; CRC Press: Boca Raton, FL, 1979-1980; D 148.
Williams, P. J .; Strauss, C. R.; Wilson, B.; Massy-Westropp,
R. A. Studies on the hydrolysis of Vitis vinifera monoterpene
precursor compounds and model monoterpene â-D-glucosides
rationalizing the monoterpene composition of grapes. J .
Agric. Food Chem. 1982, 30, 1219-1223.
Williams, P. J .; Sefton, M. A.; Wilson, B. Nonvolatile Conju-
gates of Secondary Metabolites as Precursors of Varietal
Grape Flavor Components. In Flavor Chemistry, Trends and
Developments; Teranishi, R., Buttery, R. G., Shahidi, R.,
Eds.; American Chemical Society: Washington DC, 1989;
pp 35-48.
Williams, P. J .; Sefton, M. A.; Francis, I. L. Glycosidic
Precursors of Varietal Grape and Wine Flavor. In Flavor
Precursors: Thermal and Enzymatic Conversions of Precur-
sors; ACS Symposium Series; Teranishi, R., Takeoka, G.,
Gu¨ntert, M., Eds.; American Chemical Society: Washington,
DC, 1992; pp 74-86.
Winterhalter, P.; Skouroumounis, G. K. Glycoconjugated
Aroma Compounds: Occurrence, Role and Biotechnological
Transformation. Adv. Biochem. Eng./ Biotechnol. 1996, 55,
73-105.
Winterhalter, P.; Full, G.; Herderich, M.; Schreier, P. Applica-
tion of on-line coupled chromatographic techniques in flavor
precursor studies. In Progress In Flavour Precursor Studies;
Schreier, P., Winterhalter, P., Eds.; Allured Publishing:
Carol Stream, IL, 1993; pp 97-101.
Received for review May 12, 1999. Revised manuscript received
J anuary 21, 2000. Accepted J anuary 31, 2000. We thank the
Grape and Wine Research and Development Corp. for funding
this research.
J F9904970