Precursors to oak lactone. Part 2: Synthesis, separation and cleavage of several β-D-glucopyranosides of 3-methyl-4-hydroxyoctanoic acid

Article abstract of DOI:10.1016/j.tet.2004.05.070

The β-D-glucopyranosides of all four stereoisomers of 3-methyl-4-hydroxyoctanoic acid have been prepared. The (3S,4S) and (3R,4R) species were prepared from cis-5-n-butyl-4-methyl-4,5-dihydro-2(3H)-furanone (cis-oak lactone) by a process involving ring-opening with base and protection of the carboxyl function as its benzyl ester. The glucose unit was introduced by a modified Koenigs-Knorr procedure. A different strategy was necessary for synthesis of the (3S,4R) and (3R,4S) compounds. This was based on the reductive ring-opening of trans-oak lactone and subsequent protection of the primary alcohol as its t-butyldiphenylsilyl ether. Separation of the individual glucosides was effected by preparative thin layer chromatography. Those corresponding to the nature-identical isomers of oak lactone have been shown to produce oak lactone under both acidic hydrolysis and pyrolysis conditions. The galloyl-β-D-glucoside of the cis-species, obtained as a natural isolate from the wood of Platycarya strobilacea, was also found to produce cis-oak lactone upon both acid hydrolysis and pyrolysis. Both the nature-identical (4S,5S) cis-oak lactone and its non-natural (4R,5R) enantiomer have been prepared from their corresponding glycosides and their aroma thresholds in white wine were determined to be 23 and 82 μg/L (ppb) respectively. The aroma threshold of the nature-identical isomer in a red wine was 46 μg/L.

Full text of DOI:10.1016/j.tet.2004.05.070

Tetrahedron 60 (2004) 6091–6100
Precursors to oak lactone. Part 2: Synthesis, separation
and cleavage of several b-D-glucopyranosides
of 3-methyl-4-hydroxyoctanoic acidq
b
c
Kerry L. Wilkinson,a,b Gordon M. Elsey,a,b Rolf H. Prager, Takashi Tanaka
,
*
and Mark A. Seftona
aAustralian Wine Research Institute, PO Box 197, Glen Osmond, SA 5064, Australia
bSchool of Chemistry, Physics and Earth Sciences, Flinders University, PO Box 2100, Adelaide, SA 5001, Australia
cGraduate School of Biomedical Sciences, Nagasaki University, 1-14 Bunkyo-machi, Nagasaki 852, Japan
Received 9 February 2004; revised 6 May 2004; accepted 20 May 2004
Abstract—The b-D-glucopyranosides of all four stereoisomers of 3-methyl-4-hydroxyoctanoic acid have been prepared. The (3S,4S) and
(3R,4R) species were prepared from cis-5-n-butyl-4-methyl-4,5-dihydro-2(3H)-furanone (cis-oak lactone) by a process involving ring-
opening with base and protection of the carboxyl function as its benzyl ester. The glucose unit was introduced by a modified Koenigs–Knorr
procedure. A different strategy was necessary for synthesis of the (3S,4R) and (3R,4S) compounds. This was based on the reductive ring-
opening of trans-oak lactone and subsequent protection of the primary alcohol as its t-butyldiphenylsilyl ether. Separation of the individual
glucosides was effected by preparative thin layer chromatography. Those corresponding to the nature-identical isomers of oak lactone have
been shown to produce oak lactone under both acidic hydrolysis and pyrolysis conditions. The galloyl-b-D-glucoside of the cis-species,
obtained as a natural isolate from the wood of Platycarya strobilacea, was also found to produce cis-oak lactone upon both acid hydrolysis
and pyrolysis. Both the nature-identical (4S,5S) cis-oak lactone and its non-natural (4R,5R) enantiomer have been prepared from their
corresponding glycosides and their aroma thresholds in white wine were determined to be 23 and 82 mg/L (ppb) respectively. The aroma
threshold of the nature-identical isomer in a red wine was 46 mg/L.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
impart favourable sensory characteristics to the beverage.
Some 200 oak derived volatile compounds have been
identified in wines or spirits that have been so treated,1
and there are doubtless others awaiting isolation and
It has long been recognised that the use of oak barrels as
vessels for the fermentation and/or maturation of wine can
q For Part 1, see Ref. 12.
Keywords: Oak lactone; Glycosides; Hydrolysis; Pyrolysis; b-Glucosidase; Aroma thresholds.
*
Corresponding author. Tel.: þ61-8-8201-3071; fax: þ61-8-8201-2905; e-mail address: gordon.elsey@flinders.edu.au
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.05.070

6092
K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
identification. Of these, the most important are considered to
be the (4S,5S) cis- and (4S,5R) trans-isomers of 5-n-butyl-4-
methyl-4,5-dihydro-2(3H)-furanone 1, also known as ‘oak
degradation studies with analysis by tlc, and while the
proposed structure was consistent with the observed data,
the evidence was not conclusive. Recently, we synthesised
an authentic sample of 2 from cis-oak lactone and
demonstrated that this original structural assignment was
erroneous.12 The second and third potential precursors, 3
and 4, were isolated from the wood of P. strobilacea, a
member of the walnut (Juglandaceae) family of trees by
Tanaka and Kuono.13 Compound 4 has subsequently been
found in a sample of oakwood.14 Both 3 and 4 are
characterised by the presence of the b-D-glucopyranosyl
moiety at C4 of the open chain form of cis-oak lactone, with
4 being further substituted at the 60 position with the galloyl
group. Hydrolysis of both compounds in strong acid
produced the expected (4S,5S) cis-oak lactone. To date, no
potential precursor to trans-oak lactone has been isolated,
nor indeed proposed.
lactone’ or ‘whisky lactone’.2 These were first identified by
3,4
¨
Suomalainen and Nykanen after an earlier study had
incorrectly assigned their structure (tentatively) as a
branched d-nonalactone.5 Of the two, the cis-isomer is
considered to be the more important in sensory terms,
having a reported aroma threshold for the racemate of
92 mg/L (ppb) in white wine, whereas the aroma threshold
for the racemic trans-isomer has been reported as 460 ppb in
the same medium.6 Up to now, no thresholds of the naturally
occurring isomers in wine have been reported. Aroma
descriptors for both isomers include ‘coconut’, ‘citrus’ and
‘vanilla’.2,7 A study by a Japanese group has shown a
moderate correlation between the presence of oak lactones
and the perceived quality of whiskey samples.8 A recent
sensory study showed a positive correlation between the
concentration of the cis-isomer and the aroma intensity of
the ’coconut’ descriptor in Chardonnay wine.9 Similar
correlations between this isomer and the aroma intensity of
coconut, ’vanilla’ and ‘berry’ were found in Cabernet
Sauvignon wines.
2. Results and discussion
2.1. Synthesis of glycosides 3
The two diastereomeric cis-glucosides were prepared as
shown in Scheme 1. Ring opening of racemic cis-oak
lactone with potassium hydroxide and subsequent trapping
of the carboxylate with benzyl bromide gave the benzyl
esters 5, which were successfully converted into their
corresponding b-D-glucopyranosides 6 via a modified
Koenigs–Knorr procedure.15 We chose to employ the
tetrapivaloylated bromoglucose as the reagent for the
introduction of the carbohydrate unit as we have found, in
keeping with earlier reports,16 that this species produces
essentially none of the a-anomer. In contrast, the
corresponding tetraacetate often yields significant
amounts (ca. 10%) of this isomer. After separation of the
protected glucosides 6 was effected via preparative tlc,
Despite their importance to the aroma and flavour of
alcoholic beverages, the origin of these compounds remains
unclear. cis- and trans-Oak lactone are already present in
green oakwood, but additional quantities of these com-
pounds can be generated in the wood during the drying
(seasoning) and coopering processes,2 in model wine oak
extracts heated to 50 8C,10 and even in the injector block of a
gas chromatograph during analysis of oak extracts.7 Such
observations suggest the presence of at least one precursor
form of oak lactone in oakwood; to date the literature has
provided only three possible candidates. Otsuka et al.11
isolated, from oakwood powder, a compound to which they
assigned the structure 2. This assignment was based on
Scheme 1.

K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
6093
Table 1. Measured and reported specific rotations for all four isomers of
oak lactone
which had a measured specific rotation (in methanol) of
[a]D¼274. The other lower-eluting isomer produced oak
lactone whose specific rotation was [a]D¼þ79. Based on
the data collected in Table 1, the higher-eluting isomer of 6
is assigned the (3S,4S) absolute stereochemistry, with the
lower-eluting isomer assigned the (3R,4R) absolute stereo-
chemistry. Quite independently, the stereochemistry of each
was also assigned based on the results of chiral GC–MS
(Cyclosil B column). A solution of racemic cis-oak lactone
showed two peaks, with retention times of 14.57 and
14.62 min (Table 2). That the first of these peaks was due to
the nature-identical (4S,5S) isomer was confirmed by
analysis of a sample of cis-oak lactone obtained from a
natural oak extract. Finally, the major synthetic oak lactones
produced by hydrolysis of the glucosides 3 also revealed the
higher eluting isomer of 6 to have the (3S,4S) configuration.
Further details of the minor oak lactone isomers in the
hydrolysates are discussed in the following section.
Isomer
[a]D,a this study
[a]D,a lit.b
cis (4S,5S)c
cis (4R,5R)
trans (4S,5R)c
trans (4R,5S)
274
þ79
278
þ76
þ96
295
þ100
297
a
Measured as a solution in methanol.
Ref. 23.
Nature-identical isomer.
b
c
depivaloylation gave the (3R,4R) and (3S,4S) isomers of 3.
The stereochemistry of each isomer was assigned based on
the results of either of two experiments: strong acid
hydrolysis of the glycosylated acid 3 derived from the
higher-eluting isomer (on tlc) of 6 produced oak lactone
Table 2. Retention times on chiral GC (Cyclosil B column) of racemic,
natural and synthetic isomers of oak lactone
In the case of the trans species, an alternative strategy was
required due to problems in both the isolation and handling
of the analogous benzyl ester. Although the desired ester
could be prepared (as evidenced by NMR), attempts at
isolation by chromatography resulted in recovery of only
relactonised trans-oak lactone. When an attempt was made
to glycosylate the crude benzyl ester, the only recovered
products were trans-oak lactone and the O-b-D-gluco-
pyranoside of benzyl alcohol.16 This rapid lactonisation of
the trans-isomer, relative to the cis-isomer has been found
to be a general feature of these two compounds, and a
Isomer
Retention timesa (Cyclosil B)
Racemic cis
Natural cis
1st cis isomer
2nd cis isomer
14.57 14.62
14.57
14.57
—
—
—
14.62
Racemic trans
Natural trans
1st trans isomer
2nd trans isomer
13.93
13.93
—
14.12
—
14.12
—
13.93
a
Minutes.
Scheme 2.

6094
K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
comprehensive kinetic investigation into the chemistry of
the aglycone components of both cis- and trans-3 has been
completed.17
stereochemically pure; this was confirmed by both NMR
spectroscopy and, importantly, chiral GC–MS analysis of
the enzyme hydrolysate. However, chiral GC–MS analysis
of oak lactone from the acid hydrolysate indicated only 92%
purity, with 8% of the (4S,5R) trans isomer present.
Therefore, we conclude that the newly formed (4S,5R)
trans-oak lactone is an artefact of strong acid hydrolysis,
indicating the potential for acid-catalysed epimerisation at
C4 under these conditions.
The alternative strategy (Scheme 2) employed was based on
an earlier synthesis we had reported of the putative oak
lactone precursor 2.12 Reduction of racemic trans-oak
lactone with LiAlH4 followed by selective protection of the
primary alcohol function gave 8 which, as before, was
successfully glycosylated and separated into the two
diastereomers of 9. Removal of the silyl group followed
by oxidation and depivaloylation furnished the two trans-
diastereomers of 3. As was the case for the two cis-
glycosides, the stereochemistry of the two trans-glycosides
was assigned based on both the specific rotations of the two
isomers of trans-oak lactone produced after cleavage of the
sugar moiety in 3, as well as chiral GC–MS analysis of the
derived lactones (Tables 1 and 2, respectively). In this case,
the lower-eluting isomer (on tlc) of 9 ultimately provided
the nature-identical trans-oak lactone, and is therefore
assigned (3S,4R) absolute stereochemistry, while the first-
eluting isomer of 9 is assigned (3R,4S) absolute
stereochemistry.
1H NMR spectroscopy of the (3S,4S) cis-b-D-glucoside 3
indicated it to be 95% pure, with 5% of the (3R,4S) trans
isomer present, but neither the (3S,4R) trans nor (3R,4R) cis
isomers were present in detectable quantities. This last point
was confirmed by enzymatic cleavage of the sugar unit and
inspection of the product mixture. The proportion of (4R,5S)
trans-oak lactone in the enzyme hydrolysate (14%) is higher
than that detected by NMR in the original glucoside, and
may reflect some stereoselectivity of the b-glucosidase
towards the (3R,4S) glucoside 3. Analysis of the strong acid
hydrolysate confirmed the absence of the (4R,5R) cis isomer
of oak lactone, as well as the presence of 6% of the (4R,5S)
trans isomer. However, it also showed the presence of the
(4S,5R) trans isomer (4%). As with the hydrolysis of 4, this
isomer of trans-oak lactone is presumed to arise from
epimerisation at C4 of the (3S,4S) glucoside. The other
component of this hydrolysate, namely the (4R,5S) isomer
of trans-oak lactone appears to be the expected hydrolysis
product of the minor component in the original glucoside
sample. Given that epimerisation takes place to convert (a
small proportion of) cis-glucoside into trans-oak lactone,
one might expect to see the converse taking place. The fact
that this hydrolysate contained no detectable RR cis-oak
lactone may simply reflect the detection limits of the
instrument, or alternatively, it may be a manifestation of the
thermodynamic preference for formation of the trans isomer
relative to the cis.17
2.2. Chiral GC–MS analysis of oak lactones in
hydrolysates
Although NMR spectroscopy indicated that each of the
synthetic b-D-glucosides was of high purity, chiral GC–MS
analysis gave a more detailed indication of the stereo-
chemical purity of the glycoconjugates, and the oak lactones
formed by their hydrolyses. Furthermore, before determin-
ing the aroma impact of the individual oak lactone
stereoisomers, it was desirable to know the precise
composition of the solutions under investigation. Addition-
ally, these data were important in interpreting the results of
the hydrolysis and pyrolysis experiments.
Accordingly, solutions of each of the nature-identical oak
lactone stereoisomers produced by both strong acid and
enzyme hydrolysis of the corresponding b-D-glucosides 3
were examined by chiral GC–MS (Cyclosil B column) and
the results are collected in Table 3. That the acid hydrolysis
conditions could influence the relative proportion of oak
lactone isomers was established by analysis of the acid
hydrolysate of the galloyl-b-D-glucoside 4. This compound
was a natural isolate and is therefore expected to be
Similar considerations of the strong acid and enzyme
hydrolysates of the (3S,4R) trans-b-D-glucoside 3 suggested
that this diastereomer was of slightly lower purity,
approximately 85–90%, with 9% of the (3R,4S) isomer,
as well as smaller amounts (,2–3%) of each of the (3S,4S)
and (3R,4R) isomers. Although neither the (3S,4S) nor the
(3R,4R) isomers could be detected by NMR spectroscopy, it
is probable that the limited sensitivity of NMR spectroscopy
could have impeded their detection at these low concen-
trations. Their presence in the enzyme hydrolysate is
strongly supportive of this.
Table 3. Purity (%) of glycosides 3 or 4, and of oak lactones obtained from
hydrolysis of 3 or 4, as determined by NMR spectrosopy or chiral GC–MS
analysis
2.3. Hydrolytic and pyrolytic behaviour of glucosides 3
Sample
Analysis SR-trans RS-trans SS-cis RR-cis
Although the glucosides 3 present themselves prima facie as
candidates for the generation of natural oak lactone, their
provenance in oak wood has yet to be established. Their
detection in sub-mg/kg amounts by HPLC, unlike that of 4,
is likely to be hampered by the absence of a suitable
chromophore.14 However, it is worth reiterating that
(3S,4S)-3 has previously been isolated and identified in
walnut. Thus, we were keen to investigate the behaviour
towards both acidic media (approximating barrel matu-
ration) and pyrolysis (approximating the toasting undergone
during cooperage) of the isomers of 3 with absolute
3S,4S-(4)a
Enzyme hydrolysate GC–MS
NMR
—
—
8
—
—
—
100
100
92
—
—
—
Acid hydrolysate
GC–MS
3S,4S-(3) NMR
Enzyme hydrolysate GC–MS
—
—
4
5
14
6
95
86
90
—
—
—
Acid hydrolysate
GC–MS
3S,4R-(3) NMR
Enzyme hydrolysate GC–MS
94
87
85
6
9
9
—
2
3
—
2
3
Acid hydrolysate
GC–MS
a
Natural isolate obtained from the wood of P. strobilacea.

K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
6095
stereochemistry corresponding to the natural isomers of oak
lactone. In addition to the glucosides prepared for this study,
we also investigated the behaviour of the galloyl-b-D-
glucoside 4.
Table 5. Quantified amounts of oak lactone produced in the pyrolysis of 3
and 4
Isomer
cis-1a
trans-1a
Yieldb
Control untoasted
Control toasted
n.d.
0.2
n.d.
0.1
—
–
Table 4 contains quantified amounts of oak lactone
produced hydrolytically, and shows that after 48 days at
pH 3.0 and 100 8C, the hydrolyses of both the (3S,4S) and
(3S,4R) isomers of 3 were greater than 90% complete. At the
lower temperature (45 8C), no more than trace quantities of
oak lactones were produced by either isomer over this time.
This is perhaps not surprising as previous studies within our
laboratories have shown that hydrolysis of glycosides under
mild conditions is extremely slow except where the
glycoside is attached at an activated hydroxyl position.18
Under the same higher temperature conditions, the hydroly-
sis of the galloyl substituted glycoside 4 was only
approximately 20% complete after 35 days. In the case
of 4, the expected cis-oak lactone is accompanied by
approximately 8% of the trans-isomer, produced by
epimerisation under the acidic conditions employed, as
discussed above. This minor epimerisation is also observed
in the hydrolysis of the (3S,4S) isomer of 3. Of the 10%
trans produced, approximately half can be attributed to an
impurity in the original glucoside, and the remainder to
epimerisation. In contrast, the small proportion of cis-oak
lactone in the hydrolysate of (3S,4R)-3 is likely to arise
mostly from impurities in the glucoside, rather than by
epimerisation (see Table 3). In this case, lack of epimerisa-
tion can be attributed to the thermodynamic stability of the
trans compared to the cis isomer.
(3S,4S)-3 toasted
cis:trans
60.0
89
7.1
11
29
26
15
(3S,4R)-3 toasted
cis:trans
4.0
6
55.4
94
(3S,4S)-4 toasted
cis:trans
36.9
5
2.0
95
a
(mg/g wood); mean value from three replicates. Values were in agreement
to ca. 2%.
Yield defined as the percentage of total oak lactone produced relative to
the theoretical maximum.
b
amounts of the oak lactones, (20–30% conversion, Table
5) again with epimerisation of cis to trans being more
important than the converse.
2.4. Sensory evaluation of (4S,5S)-cis-1 and (4R,5R)-cis-1
In order to satisfactorily assess the aroma impact of the
individual stereoisomers of 1, it is necessary to know the
precise composition of the solution under investigation.
Accordingly, solutions of each of the four isomers of 1
produced were examined by chiral GC–MS (Cyclosil B
column) with the results collected in Table 6. While no one
of the solutions is ‘pure’ in the literal meaning of the word,
they are each strongly enriched in their respective isomer.
Given that the reported threshold for racemic cis-oak
lactone (92 ppb in white wine) is much lower than the
reported threshold for racemic trans-oak lactone (460 ppb)
in the same medium, it is apparent that small amounts of
trans-isomer in the solution of predominantly cis-oak
lactone would be expected to have little impact on the
sensory properties of the solution. Conversely, however,
small amounts of the more potent cis-isomer in the
predominantly trans-solution would be problematic. Conse-
quently, sensory analysis was limited to the cis-enriched
solutions.
Table 4. Quantified amounts of oak lactone produced by the hydrolysis of 3
and 4, at pH 3.0
Isomer
Temperature (8C)
Time (days)
Total 1a
(3S,4S)-3
45
8
48
1.1 (0.2)
1.6 (0.3)
(3S,4S)-3
100
8
48
cis:trans
171.2 (31)
516.2 (93)
90: 10
(3S,4R)-3
(3S,4R)-3
45
8
48
n.d.
n.d.
Table 6. Chiral GC–MS analysis of the composition of oak lactones
produced by strong acid hydrolysis of 3
100
8
48
cis:trans
158.0 (28)
521.6 (94)
6: 94
Sample
SR-trans
RS-trans
SS-cis
RR-cis
(3S,4S)-4
(3S,4S)-4
45
5
35
n.d.
n.d.
SS-cis
RR-cis
SR-trans
RS-trans
4
5
85
2
6
7
9
96
90
4
3
—
84
3
100
5
35
cis:trans
17.6 (3)
106.8 (20)
92: 8
1
1
a
A duo–trio test20 was conducted to establish whether or not
there was a perceptible difference in aroma impact between
the two cis-enriched solutions; a young (,12 months old)
neutral, dry white wine (2002 South Australian Chenin
Blanc) was spiked with either (4S,5S)-cis-1 or (4R,5R)-cis-1
(161.7 mg/L) and presented to a panel of 25 judges, 20 of
whom correctly identified the sample which differed from
the reference. These data show that the two cis-isomers are
significantly different (at the 99% confidence level).
Furthermore, they bring into question the wisdom of
conducting sensory impact studies using racemic cis-1.
mg oak lactone; mean value from two replicates. Values were in
agreement to ca. 2%; values in parentheses represent the percentage of
oak lactone formed relative to the theoretical maximum.
The pyrolysis of both 3 and 4 was investigated by adsorbing
the desired compound onto oakwood powder, and then
subjecting the whole to conditions which were expected to
closely mimic barrel toasting temperatures.19 The oakwood
for this experiment was chosen for its intrinsically low
levels of oak lactone, even after toasting. In contrast to their
hydrolytic behaviour, the pyrolysis of all three glycosides
resulted in reasonably rapid formation of significant

6096
K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
The aroma detection thresholds21 of solutions of each of the
cis isomers of 1 were determined in the same neutral white
wine. The aroma detection threshold for the solution
enriched in (4S,5S)-cis-1 (90% pure) was calculated to be
23 mg/L. The best estimate threshold for each panellist
was the geometric mean of the highest concentration
missed and the next higher concentration tested. The
group threshold was calculated as the geometric mean of
the individual best estimate thresholds. The distributions
of best-estimate thresholds for individual panellists are
shown in Figure 1. Informal descriptors used by the panel
to describe this isomer included: ‘fruity’, ‘coconut’,
‘woody’, ‘caramel’, ‘buttery’, and ‘vanilla’. The aroma
detection threshold for the solution rich in the non-natural
isomer, (4R,5R)-cis-1 (84% pure) was similarly calculated
to be 82 mg/L, with informal descriptors including ‘lime’,
citrus, ‘honey’, ‘coconut’, ‘vanilla’, ‘burnt apple’ and
‘cinnamon’. Finally, the aroma detection threshold of the
solution rich in (4S,5S)-cis-1 was determined in red wine to
be 46 mg/L, with similar informal descriptors to those from
the white wine study.
3. Conclusions
Hydrolytic studies at wine pH have shown that the simple or
substituted glucosides 3 and 4 are not likely to be significant
sources of oak lactones during barrel maturation of wines
unless microorganisms with b-glucosidase activity are
present. Such derivatives cannot explain the observation10
that oak lactone concentration in pH 3 oak extracts can
increase after the oak is removed from soaking. In contrast,
such compounds are likely to be important sources of
additional oak lactone formed during barrel manufacture
and toasting.
4. Experimental
4.1. General
Chemicals were purchased from Sigma-Aldrich. Commer-
cial oak lactone (50:50) was fractionated by spinning band
column distillation to give pure fractions of both the cis- and
trans-isomers. All solvents used were HPLC grade from
OmniSolv and HiPerSolv. Column chromatography was
performed using silica gel 60 (230–400 mesh) from Merck.
Preparative thin layer chromatography was performed using
glass-backed silica gel 60 plates (20£20 cm) from Merck.
All organic solvent solutions were dried over anhydrous
sodium sulfate before being filtered. 1H and 13C NMR
spectra were recorded with a Varian Gemini spectrometer at
operating frequencies of 300 and 75.5 MHz, respectively.
Mass spectra were recorded on a Hewlett–Packard (HP)
6890 gas chromatograph fitted with liquid HP 6890 series
injector and coupled to a HP 5973 mass spectrometer. Optical
rotations were measured with a PolAAr 21 polarimeter.
Microanalyses were performed at Microanalytical Services,
University of Otago, New Zealand. Oak lactone was
quantified by the SIDA method reported previously.7
4.1.1. (3R,4R) and (3S,4S) Benzyl 3-methyl-4-hydroxyoc-
tanoate (65). To a solution of cis-oak lactone (1.10 g,
7.05 mmol) in methanol (30 mL) was added potassium
hydroxide (400 mg, 7.05 mmol) and the reaction mixture
was stirred at room temperature for 16 h. The solvent was
evaporated and the residue dissolved in DMF (30 mL).
To this solution was added benzyl bromide (840 mL,
7.07 mmol) and the reaction mixture stirred at room
temperature for 16 h. The solution was diluted with water
(10 mL) and extracted with ether (3£20 mL). The combined
organics were washed with water (2£20 mL), dried and
concentrated. The resulting oil was purified by column
chromatography (20% ethyl acetate in hexane) to give a
colourless oil (1.59 g, 85%). [Found: C, 72.7; H 9.2.
C16H24O3 requires C, 72.69; H 9.15%]; dH (CDCl3) 7.40–
7.25 (5H, m, ArH), 5.12 (2H, s, CH2Ar), 3.55 (1H, m, H4),
2.53 (1H, dd, J¼15.2, 6.6 Hz, H2a), 2.29 (1H, dd, J¼15.2,
7.6 Hz, H2b), 2.28 (1H, m, H3), 1.44–1.24 (6H, m, H5,6,7),
0.92 (3H, d, J¼6.9 Hz, H9), 0.89 (3H, t, J¼7.2 Hz, H8); dC
(CDCl3) 173.4, 135.9, 128.5, 128.2, 126.9, 74.2, 66.2, 38.4,
35.4, 33.8, 28.4, 22.7, 14.0 13.6; ESI-MS (80% MeOH)
287.2 (MþNaþ).
Figure 1. Histograms showing best estimate threshold distributions for
(a) (4S,5S)-cis-1 in white wine, (b) (4R,5R)-cis-1 in white wine, and
(c) (4S,5S)-cis-1 in red wine.
4.1.2. (3R,4R) and (3S,4S) Benzyl 3-methyl-4-O-(20,30,40,
60-tetrapivaloyl-b-D-glucopyranosyl) octanoate (6).

K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
6097
2,3,4,5-Tetra-O-pivaloyl-a-D-glucopyranosyl bromide, pre-
pared according to Kunz et al.16 (3.5 g, 6.04 mmol) was
added to a solution of (^5) (1.6 g, 6.01 mmol) in
dichloromethane (25 mL) containing silver triflate (1.6 g,
6.23 mmol) and 2,6-lutidine (700 mL, 6.04 mmol). The
reaction mixture was stirred at room temperature in the dark
for 16 h before being quenched with saturated sodium
bicarbonate solution and extracted with dichloromethane
(2£15 mL). The combined organic extracts were washed
with brine (25 mL), dried and concentrated. The crude
product was purified by column chromatography eluting
with 10% ethyl acetate in hexane to give a colourless resin
(3.2 g, 70%). [Found: C, 66.3; H 9.0. C42H66O12 requires C,
66.12; H 8.72%]; A solution of this diastereomeric mixture
in dichloromethane was loaded onto preparative tlc plates
(20 cm£20 cm, Silica 60) and eluted two or three times (as
required) with dichloromethane. The plates were visualised
under UV light, the fractions of interest removed and
extracted with 80% dichloromethane in methanol
(2£50 mL) and methanol (2£50 mL).
Potassium hydroxide (194.6 mg, 3.48 mmol) was added
and the reaction mixture stirred at room temperature for
16 h. The reaction mixture was acidified to pH 3 with 10%
hydrochloric acid solution and extracted with ether (10 mL).
The crude aqueous glycoside solution was then purified by
chromatography on XAD-2 resin to give a colourless resin
(106.1 mg, 71%).
4.2.1. (3S,4S) 3-Methyl-4-O-b-D-glucopyranosyloctanoic
acid (3). The free glycoside (33.5 mg, 56%) was obtained
from (3S,4S)-6 (135.5 mg, 0.18 mmol) as described above,
[a]D¼229.8 (c 0.67, CH3OH), lit.13 [a]D¼223; dH
0
(CD3OD) 4.31 (1H, d, J¼7.8 Hz, H1 ), 3.82 (1H, dd,
0
0
J¼11.8, 2.2 Hz, H6a ), 3.67 (1H, dd, J¼11.8, 5.2 Hz, H6b ),
0
3.63 (1H, m, H4), 3.40–3.18 (3H, m, H3 ,4 ,5 ), 3.16 (1H, dd,
0
0
0
J¼9.0, 7.8 Hz, H2 ), 2.62 (1H, dd, J¼14.6, 4.8 Hz, H2a),
2.27 (1H, m, H3), 2.15 (1H, dd, J¼14.6, 8.6 Hz, H2b),
1.62–1.24 (6H, m, H5,6,7), 0.95 (3H, d, J¼6.6 Hz, H9), 0.91
(3H, t, J¼7.1 Hz, H8); dC (CD3OD) 178.5, 104.9, 84.3, 79.0,
78.5, 76.3, 72.6, 63.7, 38.9, 35.1, 32.9, 30.0, 24.6, 16.0,
15.3; ESI-MS (80% MeOH) 359.4 (MþNaþ).
Compound (3S,4S)-6. dH (CDCl3) 7.45–7.29 (5H, m, ArH),
0
5.25 (1H, dd, J¼9.3, 9.3 Hz, H3 ), 5.15 (1H, d, J¼12.0 Hz,
4.2.2. (3R,4R) 3-Methyl-4-O-b-D-glucopyranosylocta-
noic acid (3). The free glycoside (22.7 mg, 41%) was
obtained from (3R,4R)-6 (154.4 mg, 0.20 mmol) as
described above, [a]D¼216.4 (c 0.55, CH3OH); dH
0
CH2Ar), 5.08 (1H, dd, J¼9.3, 9.3 Hz, H4 ), 5.06 (1H, d,
0
J¼12.0 Hz, CH2Ar), 5.01 (1H, dd, J¼9.3, 7.8 Hz, H2 ), 4.52
0
0
(1H, d, J¼7.8 Hz, H1 ), 4.18 (1H, dd, J¼12.3, 1.8 Hz, H6a ),
0
0
3.92 (1H, dd, J¼12.3, 5.4 Hz, H6b ), 3.56–3.34 (2H, m,
(CD3OD) 4.26 (1H, d, J¼7.7 Hz, H1 ), 3.85 (1H, dd,
0
H4,5 ), 2.52–2.14 (3H, m, H2,3), 1.54–1.20 (6H, m, H5,6,7),
0
0
J¼12.6, 1.9 Hz, H6a ), 3.72–3.58 (2H, m, H6b ,4), 3.40–3.20
0
0
0
0
1.21, 1.15, 1.11, 1.10 (36H, 4s, CMe3), 0.90–0.80 (6H, m,
H8,9); dC (CDCl3) 178.0, 177.2, 176.5, 176.4, 172.7, 135.9,
128.6, 128.4, 128.3, 99.8, 80.7, 72.8, 71.9, 71.8, 68.2, 66.1,
61.9, 38.8, 38.7, 38.7, 38.6, 37.2, 33.0, 31.2, 27.8, 27.3,
27.2, 27.1, 27.1, 22.6, 14.4, 13.9; ESI-MS (80% MeOH)
785.6 (MþNaþ).
(3H, m, H3 ,4 ,5 ), 3.15 (1H, dd, J¼9.1, 7.7 Hz, H2 ), 2.66
(1H, dd, J¼15.1, 6.8 Hz, H2a), 2.26 (1H, m, H3), 2.12 (1H,
dd, J¼15.1, 7.6 Hz, H2b), 1.62–1.22 (6H, m, H5,6,7), 0.95–
0.89 (6H, m, H8,9); dC (CD3OD) 178.8, 104.9, 83.9, 79.0,
78.6, 76.2, 72.6, 63.9, 39.5, 35.5, 33.0, 30.1, 24.7, 15.4,
15.3; ESI-MS (80% MeOH) 359.4 (MþNaþ).
Compound (3R,4R)-6. dH (CDCl3) 7.42–7.28 (5H, m, ArH),
0
4.2.3. (3R,4S) and (3S,4R) 3-Methyloctan-1,4-diol (67).
LAH (540 mg, 14.2 mmol) was added to a cooled solution
of trans-oak-lactone (2.0 g, 12.82 mmol) in anhydrous THF
(40 mL). The reaction mixture was then heated at reflux
overnight before being quenched by addition of acetone
(5.0 mL) followed by addition of a solution of sodium
hydroxide (1.0 M, 4.0 mL) and saturated sodium sulfate.
The resulting solids were removed by filtration and the
filtrate concentrated and coevaporated with CH3CN
(3£10 mL) to give the product (1.91 g, 93%) as a colourless
oil. dH (CDCl3) 3.80–3.58 (2H, m, H1), 3.43 (1H, m, H4),
1.75–1.25 (9H, m, H2,3,5,6,7), 0.94 (3H, d, J¼6.6 Hz, H9),
0.90 (3H, d, J¼7.2 Hz, H8); dC (CDCl3) 75.9, 60.5, 36.3,
35.2, 34.1, 28.0, 22.8, 16.5, 14.0.
5.29 (1H, dd, J¼9.2, 9.5 Hz, H3 ), 5.16 (1H, d, J¼12.3 Hz,
0
CH2Ar), 5.11 (1H, dd, J¼9.5, 10.0 Hz, H4 ), 5.02 (1H, d,
0
J¼12.3 Hz, CH2Ar), 4.99 (1H, dd, J¼9.2, 7.9 Hz, H2 ), 4.55
0
0
(1H, d, J¼7.9 Hz, H1 ), 4.18 (1H, dd, J¼12.1, 1.8 Hz, H6a ),
0
3.94 (1H, dd, J¼12.1, 4.9 Hz, H6b ), 3.68–3.56 (2H, m,
0
H4,5 ), 2.57 (1H, dd, J¼15.3, 4.8 Hz, H2a), 2.22 (1H, dd,
J¼15.3, 8.4 Hz, H2b), 2.16 (1H, m, H3), 1.54–1.20 (6H, m,
H5,6,7), 1.17, 1.13, 1.13, 1.09 (36H, 4s, CMe3), 0.88 (3H, t,
J¼6.8 Hz, H8), 0.82 (3H, d, J¼6.6 Hz, H9); dC (CDCl3)
177.9, 177.1, 176.3, 176.3, 173.2, 136.1, 128.4, 128.1,
128.0, 98.9, 80.3, 72.4, 71.9, 71.5, 68.0, 65.9, 61.7, 38.8,
38.7, 38.7, 38.7, 38.0, 32.7, 31.2, 28.1, 27.1, 27.1, 27.0,
27.0, 22.8, 13.9, 13.1.; ESI-MS (80% MeOH) 785.6
(MþNaþ).
4.2.4. (3R,4S) and (3S,4R) 1-t-Butyldiphenylsilyloxy-3-
methyloctan-4-ol (68). TBDPSCl (4.22 g, 15 mmol) and
(^7) (2.2 g, 13.75 mmol) in pyridine (40 mL) were stirred
at room temperature for 72 h, after which time the solvent
was removed. The residue was diluted with dichloro-
methane (100 mL) and washed with saturated copper sulfate
solution (100 mL), saturated sodium bicarbonate solution
(100 mL) and water (100 mL). The organic phase was
then dried and concentrated to give a pale yellow oil, which
was purified by column chromatography (100% dichloro-
methane) to give a colourless oil (5.17 g, 94%). [Found: C,
75.6; H, 9.6. C25H38O2Si requires C, 75.32; H 9.61%]; dH
(CDCl3) 7.75–7.65 (4H, m, ArH), 7.47–7.36 (6H, m, ArH),
4.2. General procedure for depivaloylation of cis-
glycosides (6)
Sodium metal (130 mg, 5.6 mmol) was dissolved in
methanol (5 mL) and the resulting solution was added to a
solution of (3R,4R and 3S,4S) 6 (337.5 mg, 0.44 mmol) in
methanol (10 mL). The reaction mixture was stirred at room
temperature for 16 h. The mixture was then stirred for a
further 30 min in the presence of acidified Amberlite IRC-
50 (H) ion exchange resin. The reaction mixture was
filtered, concentrated in vacuo to remove methyl pivalate,
and the residue obtained dissolved in water (10 mL).

6098
K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
3.80–3.62 (2H, m, H1), 3.45 (1H, m, H4), 1.80–1.25 (9H,
m, H2,3,5,6,7), 1.09 (9H, s, tBu), 0.93 (3H, t, J¼6.0 Hz, H8),
0.89 (3H, d, J¼6.6 Hz, H9); dC (CDCl3) 135.6, 133.5, 129.6,
127.6, 75.7, 62.0, 35.9, 34.5, 33.8, 28.2, 26.8, 22.8, 19.1,
16.2, 14.1.
9.48%]. To this mixture (159 mg, 0.241 mmol) was added
TEMPO (50 mg, 0.32 mmol) and BAIB (200 mg,
0.6 mmol) in acetonitrile (2 mL) and water (3 mL).22
After stirring at room temperature for 48 h, the reaction
mixture was extracted with ethyl acetate (2£30 mL) and the
extracts washed with 5% citric acid solution (15 mL) and
water (15 mL), before being dried and concentrated to give
an orange oil. The crude product was purified by column
chromatography (2% methanol in dichloromethane) to give
(3R,4S) and (3S,4R) 3-methyl-4-O-(20,30,40,60-tetrapivaloyl-
b-D-glucopyranosyl)octanoic acid (10) as a pale yellow
resin (132 mg, 81%).
4.2.5. (3R,4S) and (3S,4R) 1-t-Butyldiphenylsilyloxy-3-
methyl-4-O-(20,30,40,60-tetrapivaloyl-b-D-glucopyrano-
syl)octane (9). The alcohol (^8) (2.83 g, 7.1 mmol) was
glycosylated in an identical manner to that described above
for the cis-species, using 2,3,4,5-tetra-O-pivaloyl-a-D-
glucopyranosyl bromide (4.11 g, 7.1 mmol), silver triflate
(1.824 g, 7.1 mmol) and 2,6-lutidine (820 mL, 7.1 mmol).
The crude product was purified by column chromatography
(10% ethyl acetate in hexane) to give a colourless resin
(4.2 g, 66%). [Found: C, 68.5; H 9.2. C51H80O11Si requires
C, 68.27; H 8.99%]; The individual glycoside diastereomers
were separated in an identical manner to that described for 6.
4.3.1. (3R,4S) 3-Methyl-4-O-(20,30,40,60-tetrapivaloyl-b-D-
glucopyranosyl)octanoic acid (10). Conversion of (3R,4S)-
9 (144.0 mg, 0.16 mmol) into the acid (3R,4S)-10 was
accomplished as outlined above to give a colourless oil
(87.0 mg, 80%); dH (CDCl3) 5.31 (1H, dd, J¼9.4, 9.0 Hz,
0
0
H3 ), 5.11 (1H, dd, J¼9.8, 9.4 Hz, H4 ), 5.03 (1H, dd, J¼9.4,
0
0
Compound (3R,4S)-9. dH (CDCl3) 7.68–7.62 (4H, m, ArH),
0
7.8 Hz, H2 ), 4.61 (1H, d, J¼7.8 Hz, H1 ), 4.22 (1H, br d,
0
0
7.48–7.35 (6H, m, ArH), 5.26 (1H, dd, J¼9.3, 9.3 Hz, H3 ),
J¼12.2 Hz, H6a ), 3.97 (1H, dd, J¼12.2, 5.3 Hz, H6b ), 3.66
0
0
(1H, m, H5 ), 3.50 (1H, m, H4), 2.6–2.46 (1H, m, H2a),
5.10 (1H, dd, J¼9.3, 9.3 Hz, H4 ), 5.01 (1H, dd, J¼9.3,
0
0
7.8 Hz, H2 ), 4.50 (1H, d, J¼7.8 Hz, H1 ), 4.18 (1H, dd,
2.18–1.96 (2H, m, H2b,3), 1.50–1.10 (6H, m, H5,6,7) 1.20,
1.15, 1.14, 1.10 (36H, 4s, CMe3), 0.91 (3H, d, J¼6.3 Hz,
H9), 0.85 (3H, t, J¼7.1 Hz, H8); ESI-MS (80% MeOH)
695.4 (MþNaþ).
4.3.2. (3S,4R) 3-Methyl-4-O-(20,30,40,60-tetrapivaloyl-b-D-
glucopyranosyl)octanoic acid (10). (3S,4R)-9 (202.2 mg,
0.23 mmol) was treated as outlined above to give the acid
(3S,4R)-10 as a colourless oil (113.9 mg, 75%); dH (CDCl3):
0
0
J¼12.0, 1.8 Hz, H6a ), 3.93 (1H, dd, J¼12.0, 5.1 Hz, H6b ),
0
3.72–3.44 (4H, m, H1,4,5 ), 1.58–1.48 (1H, m, H3), 1.48–
1.20 (8H, m, H2,5,6,7), 1.20, 1.14, 1.14, 1.11 (36H, 4s,
COCMe3), 1.06 (9H, s, SiCMe3), 0.84 (3H, t, J¼6.9 Hz, H8),
0.79 (3H, d, J¼6.9 Hz, H9). dC (CDCl3) 178.0, 177.2, 176.4,
176.3, 135.5, 134.0, 129.7, 127.7, 99.4, 82.1, 72.9, 71.8,
71.6, 68.1, 61.8, 61.8, 38.8, 38.7, 38.7, 35.2, 32.0, 29.2,
27.9, 27.3, 27.2, 27.0, 27.0, 26.9, 22.6, 14.5, 14.0; ESI-MS
(80% MeOH) 919.8 (MþNaþ).
0
5.29 (1H, dd, J¼9.5, 9.5 Hz, H3 ), 5.09 (1H, dd, J¼9.5,
0
0
9.3 Hz, H4 ), 4.98 (1H, dd, J¼9.5, 7.9 Hz, H2 ), 4.57 (1H, d,
0
0
Compound (3S,4R)-9. dH (CDCl3) 7.68–7.62 (4H, m, ArH),
0
J¼7.9 Hz, H1 ), 4.23 (1H, br d, J¼12.0, H6a ), 3.94 (1H, dd,
0
0
7.44–7.33 (6H, m, ArH), 5.29 (1H, dd, J¼9.6, 9.3 Hz, H3 ),
J¼12.0, 5.5 Hz, H6b ), 3.66 (1H, m, H5 ), 3.47 (1H, m, H4),
2.42–2.32 (1H, m, H2a), 2.24–1.95 (2H, m, H2b,3), 1.50–
1.10 (6H, m, H5,6,7), 1.19, 1.12, 1.12, 1.08 (36H, 4s, CMe3),
0.94 (3H, d, J¼6.5 Hz, H9), 0.87 (3H, t, J¼6.7 Hz, H8); ESI-
MS (80% MeOH) 695.4 (MþNaþ).
0
5.07 (1H, dd, J¼9.9, 9.6 Hz, H4 ), 4.99 (1H, dd, J¼9.3,
0
0
7.8 Hz, H2 ), 4.56 (1H, d, J¼7.8 Hz, H1 ), 4.21 (1H, dd,
0
0
J¼12.3, 1.8 Hz, H6a ), 3.92 (1H, dd, J¼12.3, 6.0 Hz, H6b ),
0
3.76–3.58 (3H, m, H1,5 ), 3.48 (1H, m, H4), 1.90–1.20 (9H,
m, H2,3,5,6,7), 1.18, 1.14, 1.14, 1.11 (36H, 4s, COCMe3),
1.03 (9H, s, SiCMe3), 0.88 (3H, t, J¼6.6 Hz, H8), 0.82 (3H,
d, J¼6.6 Hz, H9); dC (CDCl3) 178.0, 177.2, 176.5, 176.3,
135.5, 134.0, 129.5, 127.6, 99.0, 82.6, 72.6, 71.9, 71.6, 68.3,
62.2, 62.0, 38.8, 38.7, 38.7, 33.7, 32.3, 30.4, 28.2, 27.2,
27.2, 27.1, 27.0, 26.9, 22.9, 15.8, 14.0; ESI-MS (80%
MeOH) 919.8 (MþNaþ).
4.3.3. (3R,4S) 3-Methyl-4-O-b-D-glucopyranosyloctanoic
acid (3). Protected glycoside (3R,4S)-10 (87.0 mg,
0.13 mmol) was converted into the free glycoside according
to the procedure outlined above, but omitting the second
KOH step, to give (3R,4S)-3 (43.5 mg, 100%); dH
0
(CD3OD): 4.33 (1H, d, J¼7.7 Hz, H1 ), 3.85 (1H, dd,
0
0
)
J¼11.7, 2.4 Hz, H6a ), 3.69 (1H, dd, J¼11.7, 5.2 Hz, H6b
0
0
0
0
4.3. General procedure for deprotection and oxidation
of 9
3.57 (1H, app. q, H4), 3.39–3.16 (4H, m, H2 ,3 ,4 ,5 ), 2.54
(1H, dd, J¼14.6, 4.6 Hz, H2a), 2.26 (1H, m, H3), 2.14 (1H,
dd, J¼14.6, 8.8 Hz, H2b), 1.62–1.24 (6H, m, H5,6,7), 0.99
(3H, d, J¼6.7 Hz, H9), 0.92 (3H, t, J¼7.2 Hz, H8); dC
(CD3OD): 178.4, 105.0, 85.0, 79.0, 78.5, 76.2, 72.6, 63.7,
39.3, 35.4, 32.9, 29.1, 24.6, 16.8, 15.3; ESI-MS (80%
MeOH) 359.4 (MþNaþ); [a]D¼215.5 (c 0.52, CH3OH).
TBAF (230 mg, 0.72 mmol) was added to a solution of
(3R,4S) and (3S,4R)-9 (540 mg, 0.6 mmol) in THF (20 mL)
and stirred at room temperature for 16 h during which time
the solution became yellow in colour. The solvent was
evaporated and the residue dissolved in ethyl acetate
(30 mL) and washed with saturated sodium bicarbonate
solution (10 mL), 5% citric acid solution (10 mL) and water
(2£10 mL). The organic phases were dried and concentrated
and the crude product was purified by column chroma-
tography (10% ethyl acetate0in0hexane) to give (3R,4S) and
(3S,4R) 3-methyl-4-O-(20,3 ,4 ,60-tetrapivaloyl-b-D-gluco-
pyranosyl)octan-1-ol as a colourless oil (230 mg, 87%).
[Found: C 63.9; H 9.3. C35H62O11 requires C, 63.80; H
4.3.4. (3S,4R) 3-Methyl-4-O-b-D-glucopyranosyloctanoic
acid (3). Protected glycoside (3S,4R)-10 (113.9 mg,
0.17 mmol) was converted into the free glycoside according
to the procedure outlined above, to give (3S,4R)-3 (55.9 mg,
0
98%); dH (CD3OD): 4.30 (1H, d, J¼7.7 Hz, H1 ), 3.86 (1H,
0
dd, J¼11.8, 2.2 Hz, H6a ), 3.68 (1H, dd, J¼11.8, 5.3 Hz,
0
0
0
0
0
H6b ) 3.58 (1H, app. q, H4), 3.40–3.14 (4H, m, H2 ,3 ,4 ,5 ),
2.59 (1H, dd, J¼15.3, 5.0 Hz, H2a), 2.24 (1H, m, H3), 2.10

K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
6099
(1H, dd, J¼15.3, 8.5 Hz, H2b), 1.60–1.24 (6H, m, H5,6,7),
0.97 (3H, d, J¼6.8 Hz, H9), 0.94 (3H, t, J¼7.2 Hz, H8); dC
(CD3OD): 178.8, 104.3, 83.8, 78.9, 78.6, 76.1, 72.6, 63.8,
39.7, 35.9, 32.3, 28.8, 24.9, 17.9, 15.3; ESI-MS (80%
MeOH) 359.4 (MþNaþ); [a]D¼221.3 (c 0.7, CH3OH).
4.7. Enzyme hydrolysis of glycosides (3S,4S)-3, (3S,4R)-3
and (3S,4S)-4
To a solution of the required glycoside (,20 mg) in pH 5.0
buffer solution (9 mL) was added either AR2000 enzyme
(,30 mg) or almond emulsion b-glucosidase enzyme
(,30 mg), and warmed at 30 8C for 48 h. The reaction
mixture was acidified to pH 1.0 with 10% hydrochloric acid,
allowed to stand overnight, and then extracted with ether
(2£25 mL). The combined ether extracts were dried and
concentrated. A small portion was dissolved in dichloro-
methane (approx. 1 ppm) and analysed by chiral GC–MS.
4.4. General procedure for strong acid hydrolysis of
glycosides 3 to oak lactone (1)
(3R,4S)-3 (36.3 mg, 0.11 mmol) was dissolved in water
(9 mL), concentrated sulfuric acid (1 mL) and dioxane
(1 mL) and refluxed for 16 h. The cooled mixture was
extracted with ether (2£10 mL), washed with water
(10 mL), dried and concentrated. The resulting oil was
purified by column chromatography (20% ether in pentane)
to give (4R,5S)-trans-1 as a colourless oil (8.5 mg, 62%):
[a]D¼297 (c 0.34, CH3OH), lit.23 [a]D¼295.
4.7.1. Pyrolysis of glycosides (3S,4S)-3, (3S,4R)-3 and
(3S,4S)-4. To a sample of powdered oakwood (1 g) in a
large glass ampoule (50 mL) was added a solution of either
glycoside 3 (500 mg in EtOH, 100 mL) or glycoside 413
(800 mg in EtOH, 100 mL). The ampoules were sealed and
heated at 235 8C for 30 min (equivalent to a heavy toasting).
The ampoules were then opened and extracted with model
wine (20 mL) for 72 h, prior to analysis for oak lactone
content.7
(4S,5R)-trans-1 was prepared as above (17.1 mg, 78%)
[a]D¼þ100 (c 0.48, CH3OH), lit.23 [a]D¼þ96.
(4S,5S)-cis-1 was prepared as above (10.5 mg, 60%) [a]D¼
274 (c 0.42, CH3OH), lit.23 [a]D¼278.
4.7.2. GC–MS chiral analysis. Samples of oak lactone
produced by strong acid hydrolysis of the various
glycosides, 3 or 4, dissolved in dichloromethane were
analysed with a Hewlett–Packard (HP) 6890 gas chromato-
graph fitted with liquid HP 6890 series injector and coupled
to a HP 5973 mass spectrometer. The liquid injector was
operated in fast liquid injection mode with a 10 mL syringe
(SGE, Australia) fitted. The gas chromatograph was fitted
with an approx. 30 m£0.25 mm J and W fused silica
capillary column Cyclosil-B, 0.25 mm film thickness. The
carrier gas was helium (BOC gases, Ultra High Purity), flow
rate 1.2 mL/min. The oven temperature was started at 50 8C,
held at this temperature for 1 min, then increased to 220 8C
at 10 8C/min and held at this temperature for 10 min. The
injector was held at 220 8C and the transfer line at 240 8C.
The sample volume injected was 2 mL and the splitter, at
42:1, was opened after 36 s. Fast injection was done in pulse
splitless mode with an inlet pressure of 25.0 psi maintained
until splitting. The glass liner (Agilent Technologies) was
borosilicate glass with a plug of resilanised glass wool
(2–4 mm) at the tapered end to the column. Positive ion
electron impact spectra at 70 eV were recorded in the range
m/z 35–350 for scan runs. All solvents were Mallinckrodt
nanopure grade, and verified for purity by GC–MS prior to
use.
(4R,5R)-cis-1 was prepared as above (12.1 mg, 63%) [a]D¼
þ79 (c 0.50, CH3OH), lit.23 [a]D¼þ76.
4.5. Aroma detection thresholds of (4S,5S)-cis-1 and
(4R,5R)-cis-1
The aroma threshold of (4S,5S)-cis-1 in a young (,12
months old) neutral dry white wine (2002 South Australian
Chenin blanc) was determined according to the American
Society for Testing and Materials (ASTM) method E 679,
using 24 judges. The judges were of European origin, aged
between 20 and 50, with similar numbers of males and
females. The white wine had a free sulfur dioxide content of
24 mg/L. Wines were presented (as part of a triangle test) in
ascending order of (4S,5S)-cis-1 concentration, at 2.0, 6.1,
18.5, 53.9, 161.7 and 485.1 mg/L. Panellists smelt, but did
not taste the samples. Those who could detect the spiked
wines at all of these concentrations were then tested at lower
concentrations; conversely, those who could not detect the
spike at any of the concentrations were tested at higher
concentrations. The aroma threshold of (4R,5R)-cis-oak
lactone was determined in the above manner. The aroma
threshold of (4S,5S)-cis-1 was also determined in a young
(2002 South Australian Shiraz) red wine, as described
above.
4.6. Mild acid hydrolysis of glycosides (3S,4S)-3, (3S,4R)-
3 and (3S,4S)-4
Acknowledgements
The authors wish to acknowledge the assistance of: Heather
Smyth, Dimi Capone, Kate Lattey and Dr. Leigh Francis in
helping to organise the sensory determinations; staff and
students of the AWRI for participating in these sessions;
Dimi Capone and Dr. Alan Pollnitz for assistance with the
GC–MS and chiral GC–MS analyses; Profs. Peter Høj and
Isak Pretorius and Drs. George Skouroumounis and Mike
Perkins for helpful discussions and suggestions. This project
was supported by Australia’s grapegrowers and winemakers
through their investment body, the Grape and Wine
Research and Development Corporation (GWRDC), with
A solution of the required glycoside was prepared by
dissolving either 3 (1.2 mg) or 413 (1.7 mg) in water
(1000 mL containing 10% EtOH) buffered to pH 3.0 by
saturating a solution with potassium hydrogen tartarate, and
adjusting the pH with 10% aqueous tartaric acid solution.
Portions (7 mL) of the solutions were sealed in glass
ampoules and heated at the temperatures and times
indicated in Table 2. The ampoules were then opened and
analysed for oak lactone content as described by Pollnitz
et al.7

6100
K. L. Wilkinson et al. / Tetrahedron 60 (2004) 6091–6100
11. Otsuka, K.; Sato, K.; Yamashita, T. J. Ferment. Technol. 1980,
58, 395–398.
matching funds from the Federal Government. This work
was conducted by The Australian Wine Research Institute
and Flinders University as part of their collaborative
research program. One of us (K.L.W.) would like to thank
the GWRDC for financial assistance in the form of a
research scholarship.
12. Raunkjær, M.; Sejer Pederson, D.; Elsey, G. M.; Sefton, M. A.;
Skouroumounis, G. K. Tetrahedron Lett. 2001, 42,
8717–8719.
13. Tanaka, T.; Kouno, I. J. Nat. Prod. 1996, 59, 997–999.
´
14. Masson, E.; Baumes, R.; Le Guerneve, C.; Puech, J.-L.
J. Agric. Food Chem. 2000, 48, 4306–4309.
15. Koenigs, W.; Knorr, E. Ber. 1901, 34, 957–981.
16. Kunz, H.; Harreus, A. Liebigs Ann. Chem. 1982, 41–48.
17. Wilkinson, K. L.; Elsey, G. M.; Prager, R. H.; Pollnitz, A. P.;
Sefton, M. A. J. Agric. Food Chem. 2004, 52, in press.
18. Skouroumounis, G. K.; Sefton, M. A. J. Agric. Food Chem.
2000, 48, 2033–2039.
References and notes
1. Maga, J. A. Food Rev. Int. 1989, 5, 39–99.
2. Maga, J. A. Food Rev. Int. 1996, 12, 105–130.
¨
3. Suomalainen, H.; Nykanen, L. Process Biochem. 1970, 5,
13–18.
4. Suomalainen, H.; Nykanen, L. Naeringsmiddelindindustrein.
¨
19. Lebrun, L. Aust. Grapegrower Winemaker. 1991, 331,
142–148.
1970, 23, 15–30.
5. Kahn, J. H.; Shipley, P. A.; LaRoe, E. G.; Conner, H. A.
J. Food Sci. 1969, 34, 587–591.
20. Meilgaard, M.; Civille, G. V.; Carr, B. T. Sensory Evaluation
Techniques; 3rd ed. CRC: New York, 1999.
6. Chatonnet, P. PhD Thesis; University of Bordeaux II, France,
1991.
21. Lawless, H. T.; Heymann, H. Sensory Evaluation of Food,
Principles and Practices; 1st ed. Chapman & Hall: New York,
1998.
7. Pollnitz, A. P.; Jones, G. P.; Sefton, M. A. J. Chromatogr. A
1999, 857, 239–246.
22. De Mico, A.; Margarita, R.; Parlanti, L.; Vescovi, A.;
Piancatelli, G. J. Org. Chem. 1997, 62, 6974–6977.
23. Gu¨nther, C.; Mosandl, A. Liebigs Ann. Chem. 1986,
2112–2122.
8. Otsuka, K.; Zenibayashi, Y.; Itoh, M.; Totsuka, A. Agric. Biol.
Chem. 1974, 38, 485–490.
9. Spillman, P. J. PhD Thesis; University of Adelaide, 1998.
10. Pollnitz, A. P. PhD Thesis; University of Adelaide, 2000.