Isolation and identification of 2-methoxy-3,5-dimethylpyrazine, a potent musty compound from wine corks

Article abstract of DOI:10.1021/jf049484z

The compound responsible for a fungal must taint evident in industry assessments of wine corks was identified as 2-methoxy-3,5- dimethylpyrazine. The identification was made on the basis of gas chromatography/odor analyses, collection of material using micropreparative techniques, determination of chemical properties of collected material, and comparison by gas chromatography/mass spectrometry with an authentic sample, synthesized from 2-hydroxy-3,5-dimethylpyrazine. 2-Methoxy-3,5-dimethylpyrazine is an extremely potent compound with an unpleasant, musty, moldy aroma and an aroma threshold in a white wine of 2.1 ng/L. While its contribution to the frequency and intensity of cork taint in bottled wine is yet to be established, it has been assessed by some wine industry personnel as second only to 2,4,6-trichloroanisole as a cause of cork taint in Australian wine.

Full text of DOI:10.1021/jf049484z

J. Agric. Food Chem. 2004, 52, 5425−5430 5425
Isolation and Identification of 2-Methoxy-3,5-dimethylpyrazine, a
Potent Musty Compound from Wine Corks
ROBERT F. SIMPSON, DIMITRA L. CAPONE, AND MARK A. SEFTON*
The Australian Wine Research Institute, P.O. Box 197, Glen Osmond SA 5064, Australia
The compound responsible for a “fungal must” taint evident in industry assessments of wine corks
was identified as 2-methoxy-3,5-dimethylpyrazine. The identification was made on the basis of gas
chromatography/odor analyses, collection of material using micropreparative techniques, determination
of chemical properties of collected material, and comparison by gas chromatography/mass
spectrometry with an authentic sample, synthesized from 2-hydroxy-3,5-dimethylpyrazine. 2-Methoxy-
3,5-dimethylpyrazine is an extremely potent compound with an unpleasant, musty, moldy aroma and
an aroma threshold in a white wine of 2.1 ng/L. While its contribution to the frequency and intensity
of cork taint in bottled wine is yet to be established, it has been assessed by some wine industry
personnel as second only to 2,4,6-trichloroanisole as a cause of cork taint in Australian wine.
KEYWORDS: Cork taint; 2-methoxy-3,5-dimethylpyrazine; 2-hydroxy-3,5-dimethylpyrazine; 7-methyl-1,6-
dioxaspiro[4.5]decane; mustiness; cork; wine
INTRODUCTION
tinguish fungal must from other forms of cork taint was
confirmed by our gas chromatography/odor (GC/O) assessments
of extracts of corks. Such extracts invariably gave a discrete,
intensely musty, odorous zone in the GC/aromagram that did
not correspond to any previously described compound associated
with cork taint. However, no peak could be detected in this
region of the chromatogram and the causative agent remained
unidentified.
This paper reports the isolation and identification of 2-meth-
oxy-3,5-dimethylpyrazine (1, Figure 1) as the compound re-
sponsible for the fungal must taint, a determination of its aroma
threshold in a white wine, and the measurement by stable isotope
dilution analysis of its concentration in wine in which contami-
nated corks were soaked.
Cork taint in bottled wine adversely affects consumer
acceptability and decreases the commercial value of the wine.
Most large wineries in Australia conduct “in house” assessments
of the occurrence of taint in corks to determine whether they
will accept each consignment (1).
Although 2,4,6-trichloroanisole (TCA) is considered to be
the primary cause of cork taint (2-4), other compounds
identified in wine corks also have been implicated in contribut-
ing to taint in some bottled wines (3). These compounds include
geosmin and 1-octen-3-one (4). Consequently, TCA cannot be
considered the sole cause of cork taint, and this is supported
by observations made by winery personnel and in our laboratory.
The largest of the Australian wine companies routinely con-
ducts assessments of wine corks, and these assessments have
led to a relatively comprehensive listing of types of taint
observed in natural and agglomerated wine corks (5). The causes
of some of these taints have not been identified, and some of
the taints are considered unlikely to cause a problem in bottled
wine because the intensity of the taint is low or the taint is an
artifact of the assessment procedure. A taint the winery panelists
described as “fungal must” was considered by the panelists to
be the second most important form of cork taint after TCA.
Fungal must taint was claimed on occasions to be present at an
intensity that would make the wine undrinkable and, where this
taint was assessed as strong, could lead to rejection of batches
of cork.
MATERIALS AND METHODS
Materials. All compounds were obtained from Sigma-Aldrich and
were of the highest purity available. All solvents were Mallinckrodt
Nanopure grade and were verified for purity by gas chromatography/
mass spectrometry (GC/MS) prior to use. Wine samples were obtained
from retail outlets. Samples of 7-methyl-1,6-dioxaspiro[4.5]decane were
generously donated by Dr. M. T. Fletcher of the University of
Queensland.
NMR Analysis. All NMR analyses were conducted using a Varian
Unity Inova spectrometer operating at 600 MHz for proton and 150
MHz for carbon spectra, respectively. Spectra are of deuteriochloroform
solutions, and assignments were established by HMQC and HMBC
experiments.
Analysis by GC/MS/O. Gas chromatography/mass spectrometry/
olfactometry (GC/MS/O) was carried out with a Hewlett-Packard (HP)
6890N gas chromatograph fitted with a Gerstel MPS2 autosampler.
The gas chromatograph was coupled to an HP 5973N mass spectrometer
and a Gerstel olfactory detector port (ODP2). The Gerstel MPS2 was
operated in fast liquid injection mode with a 10 µL syringe (SGE,
Melbourne, Australia) fitted.
Duncan and colleagues were the first to recognize fungal must
as distinct from taints caused by other, known, musty cork
contaminants (5). The ability of these winery panelists to dis-
* To whom correspondence should be addressed. Tel: +61 8 8303 6600.
Fax: +61 8 8303 6601. E-mail: mark.sefton@awri.com.au.
10.1021/jf049484z CCC: $27.50 © 2004 American Chemical Society
Published on Web 07/27/2004

5426 J. Agric. Food Chem., Vol. 52, No. 17, 2004
Simpson et al.
Figure 1. 2-Methoxy-3,5-dimethylpyrazine (1), 2-hydroxy-3,5-dimethylpyrazine (2), 1,3,5-trimethyl-2(1H)-pyrazinone (3), (5S,7S)-7-methyl-1,6-dioxaspiro-
[4.5]decane (4), and 2-hydroxy-3,6-dimethylpyrazine (5).
The gas chromatograph was fitted with either a 30 m × 0.25 mm
Phenomenex fused silica capillary column ZB-Wax, 0.25 µm film
thickness, or a J&W fused silica capillary column DB-1701 (same
dimensions and film thickness). A crosspiece was installed at the end
of the analytical column, 2.2 m of 220 µm i.d./320 µm o.d. deactivated
fused silica tubing (SGE) was connected from the crosspiece to the
ODP, and 2 m of 110 µm i.d./320 µm o.d. deactivated fused silica
tubing was connected from the crosspiece to the mass spectrometer,
giving a split ratio of 2:1 in favor of the ODP. The carrier gas was
helium (BOC Gases, Ultra High Purity) with a flow rate of 1.8 mL/
min. A deactivated borosilicate glass liner, 4 mm i.d., with a plug of
silanized glass wool (2-4 mm) tapered at the column end was installed
in the inlet, and the splitter, at 28:1, was opened after 60 s. Two µL
injections were done in pulsed splitless mode with an inlet pressure of
25.0 psi maintained until splitting. The oven temperature was started
Figure 2. Mass spectrum of synthetic 2-methoxy-3,5-dimethylpyrazine
(1).
at 50 °C, held at this temperature for 1 min, increased to 220 °C (to
250 °C for the DB-1701 column) at 10 °C/min, and held at this upper
temperature for 10 min. The injector was held at 200 °C (220 °C for
the DB-1701 column), and the transfer line at 250 °C (275 °C for the
DB-1701 column). Positive ion electron impact spectra at 70 eV were
recorded in the range m/z 35-350 (cycle time, 1.01 s). For determi-
nation of 1, analyses were carried out as described above, except that
the GC column (DB-1701) was connected to the mass spectrometer
only, the mass spectrometer was operated in the selected ion monitoring
(SIM) mode, and the helium flow rate was 1.2 mL/min. The oven
temperature was started at 50 °C, held at this temperature for 1 min,
increased to 240 °C at 10 °C/min, and held at this upper temperature
for 10 min. The injector was held at 200 °C, and the transfer line at
240 °C. Chiral analyses were carried out using the conditions reported
previously (6).
Syntheses. 2-Hydroxy-3,5-dimethylpyrazine (2) was prepared from
the condensation of alaninamide with methylglyoxal according to the
method of Jones (7), modified to suit present-day laboratory equipment
and techniques. Alaninamide was generated from the commercially
available hydrochloride by the addition of sodium hydroxide (15 M)
in the first step of the preparation. Sufficient alkali was added to give
a pH > 7, which was monitored by moistened universal indicator strips.
Compound 2 was purified by dissolving the crude product in 10% v/v
methanol in chloroform and pouring the solution through a 30 mm ×
20 mm i.d. column of Woelm neutral alumina, activity 2. The column
was eluted with additional solvent, and the product, after removal of
solvent under a vacuum, was purified by crystallization as described
by Jones (7).
and the solution was extracted with pentane (3 × 20 mL). The solvent
was removed, and the product, 2-methoxy-3,5-dimethylpyrazine (1, ca.
40% yield), was purified by microdistillation using a Kugelrohr
1
apparatus. H NMR 2.40 (3H, d, J ) 0.5 Hz, C5-CH3), 2.43 (3H, d,
J ) 0.5 Hz, C3-CH3), 3.92 (3H, s, OCH3), 7.75 (1H, br. s, C6-H);
13C NMR 19.1 (C3-CH3), 20.1 (C5-CH3), 53.4 (OCH3), 136.7 (C1),
143.2 (C5), 143.6 (C3), 157.1 (C2); m/z 138 (100), 137 (43), 123 (10),
120 (31), 109 (47), 107 (27), 95 (16), 82 (34), 68 (16), 54 (35), 52
(11), 42 (38), 39 (16). The mass spectrum (Figure 2) was essentially
the same as that reported by Mottram et al. (8), and the 1H NMR
spectrum was virtually identical to that reported by Czerny and
Grosch (9). The purity of 1 was greater than 99%, based on GC/MS
analysis. The aqueous solution was further extracted with dichlo-
romethane (3 × 20 mL), and the solvent was evaporated, yielding 1,3,5-
trimethyl-2(1H)-pyrazinone (3, ca. 60% yield) as a white crystalline
solid. 1H NMR 2.13 (3H, d, J ) 0.5 Hz, C5-CH3), 2.39 (3H, s,
C3-CH3), 3.41 (3H, s, NCH3), 6.76 (1H, s, C6-H); 13C NMR 19.5
(C5-CH3), 20.75 (C3-CH3), 37.0 (NCH3), 124.6 (C6), 130.4(C3), 155.7
(C2), 156.9 (C5); m/z 138 (91), 110 (41), 109 (100), 95 (30), 69 (11),
68 (68), 54 (11), 42 (51), 41 (19), 39 (15). Kovats GC retention index
2083 (ZB-Wax).
[2H3]-2-Methoxy-3,5-dimethylpyrazine (d-1) was prepared from 1
according to the procedure of Kotseridis et al. (10): m/z 142 (21), 141
(100), 140 (46), 123 (10), 122 (18), 112 (47), 110 (24), 82 (24), 58
(10), 55 (10), 54 (35), 45 (34), 43 (12), 42 (11), 39 (13).
Compound 2 (1.0 g) was dissolved in a small volume of methanol
(about 30 mL), diethyl ether was added, and the solution was cooled
in ice. An excess of diazomethane in ether was added (caution, see
safety notes below!). The reaction was complete after 16 h at room
temperature; that is, no starting material remained. GC/MS analysis
showed only two products in the ratio of 40:60, based on peak area.
After quenching the excess diazomethane with acetic acid, the bulk of
the ether was removed under a vacuum, water (30 mL) was added,
Enrichment of 2-Methoxy-3,5-dimethylpyrazine (1) from Corks.
Wine corks (140, 24 mm × 20 mm diameter) that had been classified
in the Southcorp Wines assessments as having a moderate incidence
of fungal must taint were placed in a 5 L flask with water (1.2 L)
purified through a Milli-Q purification system (Millipore Corp.,
Bedford, MA). Steam distillate (200 mL) was collected and extracted
with pentane (2 × 50 mL). The pentane extract was washed with
saturated sodium hydrogen carbonate and water, dried over anhydrous

“Fungal Must” Cork Taint
J. Agric. Food Chem., Vol. 52, No. 17, 2004 5427
sodium sulfate, and concentrated to ca. 100 µL using a Vigreux column
packed with glass Fenske helices. Isolation of a basic fraction of the
cork components was obtained in a similar manner, except that sodium
tetraborate (36 g) was added prior to distillation to give a pH of 8-9.
Sodium hydroxide (5 M) was added to the distillate to give pH 12,
and the basified distillate was extracted with pentane (2 × 50 mL).
The pentane solution was then extracted with sulfuric acid (5 M, 2 ×
10 mL); the acid extract was basified with sodium hydroxide and then
re-extracted with pentane. This pentane extract was washed with
saturated sodium hydrogen carbonate, dried over anhydrous sodium
sulfate, and concentrated to ca. 100 µL with a Vigreux column.
Microscale Preparative GC. The components of the concentrated
pentane extract of the steam distillate were chromatographed on a
DB-1701 GC column. Those components corresponding to a section
of the chromatogram (7-8 min) in which the compound giving fungal
must taint eluted (odor detected at 7.53 min) were trapped using a 700
mm length of 0.25 mm deactivated fused silica tubing (“collection
loop”) connected by a glass connector (SGE) to the tubing normally
leading to the ODP. A second glass connector was placed at the exit
end of the collection loop so that the loop could be sealed after each
collection and placed in a freezer at -18 °C between collections. The
tubing was shaped into a coil that was held in place with a cotton thread
with slip-knots so that the coils could be released after the final
collection. The collection loop was placed in liquid nitrogen a few
seconds before connecting to the olfactory detector outlet before each
collection. After 12 collections, the exit end of the collection loop was
pushed into the inside of a disposable glass capillary tube containing
chloroform (5-10 µL). The chloroform was taken up by the capillary
tubing (due to capillary action), and the end of the tubing was then
lifted so that the solvent slowly flowed under gravity through the
collection loop. The other end of the collection loop (“entry end”) was
placed in a vial insert, and the residual solvent in the collection loop
was expelled by gentle pressure. The chloroform solution was partially
evaporated by taking the solvent up in a 5 µL syringe and expelling
several times. The solution was then injected into the GC/MS system
and analyzed using a ZB-Wax column.
Chemical Tests. Material was trapped (7-8 min) from a single
injection of the concentrated pentane extract of the steam distillate,
described in the preceding section, using the DB-1701 column. In this
case, the effluent from the tubing normally connected to the ODP was
bubbled through water (500 µL) contained in a 2.5 mL glass vial. A
small volume (4 drops, ca. 75 µL) of this solution was added to 1 mL
vials, each containing one of the following (1 drop): 30% phosphoric
acid, KOH (1 M), sodium borohydride (5 mg) in KOH (1 M), copper
sulfate (0.3 M), and 2,4-dinitrophenylhydrazine (DNP, 0.05 M) in 30%
perchloric acid. The aroma of the solutions was then evaluated by four
assessors familiar with the fungal must taint.
Soaking of Corks in Wine. Wine corks that had been classified as
having a sporadic incidence of fungal must taint were obtained from
Southcorp Wines. A random sample of 100 corks was treated as follows.
Each cork was placed in a 150 mL glass jar, a neutral, white wine
(100 mL) was added, and an end of the cork was pierced with a
hypodermic needle that submerged the cork in the wine when the glass
lid was placed on the jar. The cork was held in the wine for 48 h at
room temperature (ca. 25 °C). Several wines were selected that were
judged as exhibiting some fungal must taint.
was added, and the pentane extract was concentrated using a Vigreux
column to ca. 100 µL. The concentrated extract was then transferred
to a vial insert, sealed in a 1 mL sample vial, and analyzed by GC/MS
in SIM mode. The ions monitored in SIM runs were m/z 112, 122,
140, and 141 for d-1 and 109, 120, 137, and 138 for 1. Selected
fragment ions were monitored for 30 ms each. The italicized ion for
each compound was the ion typically used for quantitation, having the
best signal-to-noise ratio and the least interference from other wine
components. The other ions were used as qualifiers. The analytical
method was validated by a series of standard additions of 1 (0, 2 to
100 ng/L, n ) 6 for the analyte) to a dry white wine (11.5% ethanol,
pH 3.2). The standard addition curves obtained were linear throughout
the concentration range, with a coefficient of determination (r2) of 0.998
and the linear regression equation y ) 1.68x - 0.0105.
For the analysis of 1 in cork, the cork was ground in a coffee blender
and the ground cork was placed in a 100 mL glass jar with a glass
stopper. Pentane (40 mL) was added, and the cork was left to soak for
24 h at room temperature (ca. 25 °C). The pentane was decanted through
a glass funnel containing a plug of silanized glass wool to remove any
cork fragments from the solution. The procedure was repeated, and
the combined pentane extract was dried over anhydrous MgSO4. The
pentane extract was transferred to a 100 mL separatory funnel, internal
standard (10 ng of d-1 in ethanol, 1 mL) was added, and the pentane
was extracted with cold sulfuric acid (2 × 10 mL) and so forth,
following the same procedure as for the wine.
Sensory Assessment. The aroma threshold of 1 in a young (<12
months old), neutral, dry white wine was determined according to
the American Society for Testing and Materials (ASTM) method
E 679-79, using 33 judges, as described by Meilgaard et al. (11). The
judges were of European origin, aged between 20 and 55, with similar
numbers of males and females. The white wine (11.5% v/v, pH 3.2)
had a free and total sulfur dioxide content of 35 and 215 mg/L,
respectively. Wines were presented in ascending order of concentration
of 1, at 0.3, 0.9, 2.7, 8.1, 24.3, and 72.9 ng/L. Panelists assessed the
aroma but did not taste the samples. Those who could detect the spiked
wines at all of these concentrations were then tested at lower
concentrations of 1. Panelists were also asked to describe the aroma
associated with the wine. The best-estimate threshold for each panelist
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.
RESULTS AND DISCUSSION
Isolation and Identification of the Fungal Must Taint
Compound. In our investigations, GC/O assessments of pentane
extracts of whole single corks on both a relatively nonpolar and
a polar column indicated that the taint was associated with only
one part of the chromatogram (although no peak was evident
at this retention time), suggesting that the aroma was due to a
single component. The observation of this odorous component
by GC/O in the room-temperature pentane extracts of corks
showed that it was not just an artifact of the distillation process
described below.
Steam distillation was chosen as the first step in the isolation
and concentration process to reduce the quantity of silicone and
paraffin components originating from the cork coatings. Pen-
tane extracts of the steam distillate were then analyzed by
GC/MS/O.
The fungal must taint eluted on a Carbowax column at
approximately the same retention time as (E)-2-octenal and
1-octen-3-ol. However, authentic reference compounds at con-
centrations similar to those in the extracts did not have a sig-
nificant aroma and their aroma at higher concentration did not
resemble fungal must.
The taint compound eluted at approximately the same time
as (E)-7-methyl-1,6-dioxaspiro[4.5]decane on a DB-1701 col-
umn. Both (E)- and (Z)-7-methyl-1,6-dioxaspiro[4.5]decane
were detected in the cork distillates but had weak aroma only,
Analysis of Corks and Wine. Sodium tetraborate (3 g) was added
to the wine (100 mL) and dissolved by stirring. The wine was
transferred to a 150 mL separatory funnel, and internal standard (10
ng of d-1 in ethanol, 1 mL) was added. The wine was extracted with
minimal shaking with pentane (2 × 10 mL). The slight emulsion that
formed was broken by adding a few drops of ethanol to the contents
of the separatory funnel. The pentane extract was washed with water
(5 mL) and then extracted with cold sulfuric acid (1 M, 2 × 10 mL).
The flask containing the acid was kept cold by storing in ice. The acid
extract was washed with pentane (5 mL) and then basified by the careful
addition of saturated sodium hydrogen carbonate (40 mL). The pH of
the solution was checked with universal indicator strips to ensure the
final pH was greater than 7. The basified solution was extracted with
pentane (2 × 8 mL). The pentane extract was washed with water (5
mL) and dried over anhydrous MgSO4. Isooctane (2 drops, ca. 30 µL)

5428 J. Agric. Food Chem., Vol. 52, No. 17, 2004
Simpson et al.
at the concentration found in the extracts. The absolute
stereochemistry of these compounds was determined by chiral
GC/MS analysis of the cork extract and of authentic reference
compounds. The dominant isomer was the 5S,7S form (4, 86%),
and this was accompanied by minor amounts of the 5R,7R
enantiomer of 4 (12%) and the 5R,7S diastereoisomer (2%).
To our knowledge, these compounds have not been reported
previously as components of cork bark. They have been
identified in the bark of certain angiosperm trees (12), and they
are important in the insect communication of several conif-
erophagous bark beetles (13). The bark beetles are able to
distinguish between host and nonhost trees by the presence of
these compounds, and also the males of certain species of beetle
synthesize these compounds which are used to repel rival males.
demonstrates that both the amine and amide nitrogen of the
alaninamide can react rapidly with the ketone function of the
methylglyoxal and that the regioselectivity of the reaction
depends on the availability of the amine nitrogen during the
initial few minutes.
Methylation of 2 with diazomethane gave two products. The
desired product (1) was extracted with pentane after the removal
of ether and the addition of water to the reaction mixture and
was obtained in near-pure form. It was further purified by
microdistillation using a Kugelrohr apparatus to give material
that was >99% pure. The major product, 1,3,5-trimethyl-2(1H)-
pyrazinone (3), was obtained from the reaction mixture in near-
pure form by further extraction of the aqueous reaction mixture
with dichloromethane. The relative proportion of the methylated
products 1 and 3 indicates that the starting material (2) exists
predominantly in the keto-form in the ether-methanol solution
to which the diazomethane was added. This is unsurprising since
hydroxypyrazines are known to exist as keto-enol tautomers
(14).
Fungal must material was trapped in water using the
GC-ODP tubing and the chemical properties determined, based
on the effect on aroma. Mustiness was retained in the presence
of potassium hydroxide, sodium borohydride, and copper sulfate
but was lost on addition to phosphoric acid. These tests indicated
that the compound was basic and was not an aldehyde, ketone,
or thiol.
Overall, the preparation of 2 and subsequent methylation with
diazomethane provided a sound approach for the synthesis of
1. Compound 2 was prepared in high yield and was readily
purified by recrystallization to give very pure material, free from
5. This simplified the next step, where there was no need to
separate similar components. The methylation step proceeded
well to give the required product that was isolated in near-pure
form by extraction with a nonpolar solvent, even though it was
the minor product. Attempts to convert 2 into 1, using methyl
iodide and bases such as potassium or sodium carbonate, gave
low yields and led to considerable decomposition. Other
approaches to the synthesis of 1 (9) were investigated earlier in
our work but found to be less satisfactory.
Knowing that the fungal must compound was likely to be
basic, the acid soluble components were isolated from a second
steam distillation of 140 contaminated corks. This approach
provided the means for removal of the acidic and most of the
neutral components and gave considerable concentration of the
fungal must compound. A significant peak with a retention time
corresponding to the fungal must taint was obtained on the ZB-
Wax column. The peak contained only one component whose
MS and retention time matched those of synthetic 2-methoxy-
3,5-dimethylpyrazine (1). Compound 1 had a Kovat’s retention
index of 1417 on the DB-1701 and 1446 on the ZB-Wax
column. As reported by Mottram et al. (8) in their investigation
of an off-odor in machine cutting emulsion, only one isomer of
methoxydimethylpyrazine was detected in our investigations of
cork volatiles. The two isomers coeluted on the ZB-Wax column
but were completely resolved on the DB-1701 column, and as
little as 1% of the second isomer (2-methoxy-3,6-dimethylpyra-
zine) as a proportion of total methoxydimethylpyrazines would
have been detected, if present. Compound 1 was also observed
when material collected from multiple injections of a pentane
extract of the steam distillate of contaminated corks from a
DB-1701 column was re-injected onto a Carbowax column.
Although 1 overlapped with a similar amount of 5-methyl-2-fur-
fural, a good spectrum of 1 was obtained. The presence of 1 in
this pentane extract, prior to partitioning into strong acid solu-
tion, showed that it was not an artifact of strong acid treatment.
Aroma Threshold of 2-Methoxy-3,5-dimethylpyrazine (1).
The aroma threshold of 1 in a neutral white wine was determined
to be 2.1 ng/L. The distribution of best-estimate thresholds is
shown in Figure 3. The threshold of 1 is unlikely to have been
influenced by the possible presence of traces of 2-methoxy-
3,6-dimethylpyrazine as Czerny and Grosch (9) have shown that
the latter has a much higher threshold in water than the former.
The most common descriptors used by the assessors were
“dirty”, “dusty”, “musty”, and “moldy”. A few assessors also
described the higher concentrations as “chocolate” or “coffee”.
The aroma threshold of 1 compares closely with that of TCA,
which has an aroma threshold in wine of 1-4 ng/L (4, 15, 16)
and is recognized as one of the most potent taint compounds
affecting a wide range of foods and beverages.
Analysis of Cork and Wine in Which the Cork Was
Soaked. The internal standard prepared for the analysis of 1 in
wine and corks was 2-methoxy-3,5-dimethylpyrazine-d3 (d-1).
This material was prepared in high yield by treating 1 with D2O
in the presence of D2SO4, using a procedure described by
Kotseridis et al. (10) for preparing deuterated 2-methoxy-3-
isobutylpyrazine. Only three hydrogen atoms in 1 were readily
exchanged, and more forcing conditions, such as higher tem-
perature and stronger acid, led to decomposition. Kotseridis et
al. (10) had indicated that the benzylic protons in 2-methoxy-
3-isobutylpyrazine were readily exchanged. The exchange of
deuterium for benzylic protons may be promoted in these
molecules by the adjacent methoxyl group. No back-exchange
of deuterium was observed during the analysis of wines, based
on the unaffected isotopic ratio of the major fragment ions m/z
139-142.
Synthesis of 2-Methoxy-3,5-dimethylpyrazine (1). Com-
pound 2 was prepared by a modification of the method of Jones
(7). Three syntheses were carried out during the course of our
investigations, and it was found that the yield was strongly
influenced by reaction conditions.
When alaninamide was regenerated from its hydrochloride
by addition of strong alkali in methanol solution prior to mixing
with methylglyoxal, a high yield of 2 was obtained. GC/MS
analysis of the reaction product showed the presence of only
two components, 2 and 2-hydroxy-3,6-dimethylpyrazine (5) (7),
present in a ratio of 90:10. The addition of water to the reaction
mixture (to give a homogeneous solution) gave less 2 (with a
ratio of 55:45). When the hydrochloride was used without
pretreatment with alkali, the major product was 5 (with a ratio
of 5:95) even though the alaninamide hydrochloride and
methylglyoxal were quickly added together at low temperature
(-20 °C) and alkali was then added within a few minutes. This
The wine chosen for the analysis of 1 in corks was a “bag-
in-box” wine that, therefore, had no contact with a cork prior

“Fungal Must” Cork Taint
J. Agric. Food Chem., Vol. 52, No. 17, 2004 5429
Figure 3. Best-estimate aroma thresholds of 2-methoxy-3,5-dimethylpyrazine (1) in a neutral dry white wine.
possess an intense odor; its odor threshold in water was deter-
mined as 0.4 ng/L. These authors also suggested that the meth-
oxypyrazines they detected in raw coffee have a bacterial origin.
Table 1. Content of 1 in Corks and Concentration in Wine in Which
the Corks Were Soakeda
content (ng) of 1
The origin of 1 in corks was not examined during our
investigations. The cutting punches that produce the cork
cylinders could be one possible source of the taint. This would
explain the apparent total extraction of 1 from two of the corks
listed in Table 1; that is, the material would be contained on or
near the surface of the cork. Also, bacteria capable of producing
1 could be present in areas where the cork is processed or stored
and could grow on the cork at some stage of the processing.
However, the origin of 1 may not necessarily be bacterial. It
has been recognized for many years that certain aroma-intense
microbial metabolites occurring in foods and beverages can be
produced by different types of microflora. For example, geosmin
can be produced by Actinomycetes (bacteria) (17), cyanobacteria
(algae) (18), and a number of Penicillium spp. (yeasts) (19).
Davis et al. (20) isolated the microflora from wine corks
imported into Australia and showed that numerous molds,
bacteria, and yeast were present on the corks, but the metabolites
produced by these microorganisms were not investigated.
remaining in the cork
after soaking in wine
concn (ng/L) of
1 in the wine
cork no.
1
2
3
ndb
nd
52
37
22
61
nd
blank (i.e., no cork)
a Each cork was soaked in wine (100 mL) for 48 h at room temperature. b nd
) not detected (limit of detection: 0.5 ng in cork and 1 ng/L in wine).
to the analysis. The content of 1 in each cork and the wine in
which the cork was soaked is shown in Table 1. Only cork 3
was shown to still contain 1 after soaking, and the associated
wine contained the highest concentration of 1. This wine also
had the strongest musty, moldy aroma. The absence of 1 in the
blank (i.e., wine only) showed that this compound was cork-
derived.
The peaks for the four major ions (m/z 138, 137, 120, and
109) monitored in the analysis of 1 in the wine extract of cork
3 were symmetrically enhanced on coinjection with authentic
1, without any change in their ratio, giving further evidence
that the compound quantified by selected ion monitoring for
these extracts was identified correctly.
Origin and Occurrence of 2-Methoxy-3,5-dimethylpyra-
zine (1). Mottram et al. (8) had reported that 1 was responsible
for an obnoxious odor present in certain machine cutting
emulsions used in engineering workshops. These authors
described the odor as “musty, foul drains, or sour dishcloths”.
They isolated an aerobic, Gram-negative bacterium which, when
grown, gave 1 as the only major component of the volatile
fraction isolated from the culture broth. Mottram et al. (8) were
unable to characterize the organism with known bacterial species
but provided a culture to the National Collection of Industrial
Bacteria in Aberdeen (accession number 118020). Certain
bacteria (Pseudomonas spp.) are known to produce pyrazine
off-odors, and a mold (Aspergillus flaVus) produces compounds
in the same chemical class (8 and references therein).
Compound 1 has also been identified in coffee (9), where it
was described as having an “earthy” aroma. It was shown to be
important to the aroma of both raw and roasted coffee and to
Mottram et al. (8) concluded that 1 was likely to be a
relatively common cause of off-odor in the environment.
However, there has been no further report of 1 as a cause of
off-odor in the published literature since then and only one report
of its occurrence in a food product (9). Perhaps, this is because
of the difficulty in analyzing for this compound and the
exceedingly low concentration at which it can cause an off-
odor. We also think 1 could be a cause of many instances of
off-aroma in the environment and in foods and beverages and
anticipate that the greater recognition now given to 1 by Czerny
and Grosch (9) and by this publication and the development of
improved techniques to synthesize and identify this compound
will lead to more reports of it being a cause of off-odor.
Currently, we have little data on the content of 1 in corks or
the extent to which 1 is leached into the wine after bottling.
However, indications are that 1 may be the second most
important taint affecting wine corks and, like TCA, may be a
serious problem for the wine industry by adversely affecting
the quality and commercial value of bottled wine.
The presence of 1 in bottled wine as a cause of cork taint is
the subject of ongoing investigations at our Institute. In the

5430 J. Agric. Food Chem., Vol. 52, No. 17, 2004
Simpson et al.
(7) Jones, R. G. Pyrazines and related compounds. I. A new synthesis
of hydroxypyrazines. J. Am. Chem. Soc. 1949, 71, 78-81.
(8) Mottram, D. S.; Patterson, R. L. S.; Warrilow, E. 2,6-Dimethyl-
3-methoxypyrazine: a microbiologically-produced compound
with an obnoxious musty odour. Chem. Ind. 1984, 448-449.
(9) Czerny, M.; Grosch, W. Potent odorants of raw Arabica coffee.
Their changes during roasting. J. Agric. Food Chem. 2000, 48,
868-872.
(10) Kotseridis, Y.; Baumes, R.; Skouroumounis, G. K. Synthesis of
labeled [2H4]â-damascenone, [2H2]2-methoxy-3-isobutylpyrazine,
[2H3]R-ionone, and [2H3]â-ionone for quantification in grapes,
juices and wines. J. Chromatogr., A 1998, 71-78.
(11) Meilgaard, M.; Civille, G. V.; Coerr, B. T. Sensory eValuation
techniques, 2nd ed.; CRC Press: Boca Raton, FL, 1991; pp 124-
133.
(12) Huber, D. P. W.; Gries, R.; Borden, J. H.; Pierce, H. D., Jr. A
survey of antennal responses by five species of coniferophagous
bark beetles (Coleoptera: Scolytidae) to bark volatiles of six
species of angiosperm trees. Chemoecology 2000, 10, 103-113.
(13) Franke, W.; Kitching, W. Spiroacetals in insects. Curr. Org.
Chem. 2001, 5, 233-251.
meantime, recent work in our laboratory has shown that wine
corks have a very low affinity for 3-iso-butyl-2-methoxypyrazine
(21), in stark contrast to their absorptive capacity for TCA (22).
The greater volatility of 1 (as evidenced by its GC behavior)
compared to TCA, coupled with a greater hydrophilicity of the
pyrazine, suggests that methods, such as hot water washing or
steam cleaning, currently being investigated by cork producers
to remove taint from punched corks should have a good chance
of reducing or even eliminating this form of cork taint.
SAFETY
Diazomethane is a toxic, explosive gas. It should be handled
at all times in a fume hood fitted with a strong exhaust system.
The use of a safety shield is also strongly recommended. Users
should acquaint themselves with standard chemistry texts on
this substance before proceeding with its use.
ACKNOWLEDGMENT
We are indebted to Blair Duncan, formerly of Southcorp Wines,
for bringing to our attention this hitherto unrecognized form of
cork taint and Alan Pollnitz who conducted the early GC/O
assessments of extracts of contaminated corks. We thank Jodie
Peters of Southcorp Wines and Andrew Kleinig formerly of
Southcorp Wines for providing us with samples of corks and
for current observations made during their cork quality assess-
ments. We also thank Mary Fletcher of the Department of
Chemistry, University of Queensland, for providing us with
samples of the 7-methyl-1,6-dioxaspiro[4.5]decane stereoiso-
mers. We are grateful for the assistance of Gordon Elsey and
George Skouroumounis of the AWRI for advice on aspects of
the synthetic work, to the staff and students of the AWRI for
participating in the sensory panel, and to Peter Høj and Sakkie
Pretorius for their continued advice and encouragement.
(14) Brown, D. J. Chemistry of Heterocyclic Compounds. Vol. 58:
The Pyrazines, Supplement I; John Wiley and Sons: New York,
2002.
(15) Duerr, P. Wine quality evaluation. In Proceedings of the
International Symposium on Cool Climate Viticulture and
Enology, CorVallis, OR; Oregon State University: Corvallis, OR,
1985; pp 257-266.
(16) Liacopoulos, D.; Barker, D.; Howland, P. R.; Alcorso, D. C.;
Pollnitz, A. P.; Skouroumounis, G. K.; Pardon, K. H.; McLean,
H. J.; Gawel, R.; Sefton, M. A. Chloroanisole taint in wines. In
Proceedings of the Tenth Australian Wine Industry Technical
Conference, Sydney, New South Wales, 2-5 August 1998; Blair,
R. J., Sas, A. N., Hayes, P. F., Høj, P. B., Eds.; Australian Wine
Industry Technical Conference Inc.: Adelaide, SA, 1999; pp
224-226.
(17) Gerber, N. N.; Lechevalier, H. A. Geosmin, an earthy-smelling
substance isolated from Actinomycetes. Appl. Microbiol. 1965,
13, 935-938.
LITERATURE CITED
(18) Safferman, R. S.; Rosen, A. A.; Mashni, C. I.; Morris, M. E.
Earthy-smelling substance from a blue-green alga. EnViron. Sci.
Technol. 1967, 1, 429-430.
(19) Larsen, T. O.; Frisvad, J. C. Characterization of volatile
metabolites from 47 Penicillium taxa. Mycol. Res. 1995, 99,
1153-1166.
(20) Davis, C. R.; Fleet, G. H.; Lee, T. H. The microflora of wine
corks. Aust. Grapegrower Winemaker 1981, 208, 42, 44.
(21) Capone, D. L.; Sefton, M. A.; Pretorius, I. S.; Høj, P. B. Flavour
“scalping” by wine bottle closures. Aust. New Zealand Wine Ind.
J. 2003, 18 (5), 16-20.
(22) Capone, D. L.; Skouroumounis, G. K.; Barker, D. A.; McLean,
H. J.; Pollnitz, A. P.; Sefton, M. A. Absorption of chloroanisoles
from wine by corks and by other materials. Aust. J. Grape Wine
Res. 1999, 5, 91-98.
(1) Leske, P.; Bruer, N.; Sefton, M. A review of cork sensory
assessment methods. In Proceedings of the ASVO Oenology
Seminar on Corks and Closures; Leske, P A., Eglington, J. M.,
Eds.; Australian Society of Viticulture and Oenology: Adelaide,
1995; pp 24-26.
(2) Capone, D. L.; Skouroumounis, G. K.; Sefton, M. A. Permeation
of 2,4,6-trichloroanisole through cork closures in wine bottles.
Aust. J. Grape Wine Res. 2002, 8, 196-199 and references
therein.
(3) Simpson, R. F. Cork taint in wine: a review of the causes. Aust.
New Zealand Wine Ind. J. 1990, 5, 286-288, 289, 291, 293-
296.
(4) Amon, J. M.; Vandepeer, J. M.; Simpson, R. F. Compounds
responsible for cork taint in wine. Aust. New Zealand Wine Ind.
J. 1989, 4, 62-69.
(5) Duncan, B. Closure quality control at Southcorp Wines. In
Proceedings of the ASVO Oenology Seminar on Corks and
Closures; Leske, P A., Eglington, J. M., Eds.; Australian Society
of Viticulture and Oenology: Adelaide, 1995; pp 29-30.
(6) Wilkinson, K. L.; Elsey, G. M.; Prager, R. H.; Tanaka, T.; Sefton,
M. A. Precursors to oak lactone II: Synthesis, separation and
cleavage of several â-D-glucopyranosides of 3-methyl-4-hy-
droxyoctanoic acid. Tetrahedron 2004, 60, 6091-6100.
Received for review March 30, 2004. Revised manuscript received May
27, 2004. Accepted June 6, 2004. This project was supported by
Australian grapegrowers and winemakers through their investment
body, the Grape and Wine Research and Development Corporation,
with matching funds from the Australian Government.
JF049484Z