Development of a candidate reference method for the simultaneous quantitation of the boar taint compounds androstenone, 3α-androstenol, 3β-androstenol, skatole, and indole in pig fat by means of stable isotope dilution analysis-headspace solid-phase micro

Article abstract of DOI:10.1021/ac201465q

The steroidal pig pheromones androstenone (5α-androst-16-en-3-one), 3α-androstenol (5α-androst-16-en-3α-ol), and 3Β-androstenol (5α-androst-16-en-3Β-ol) as well as the heterocyclic aromatic amines skatole and indole, originating from microbial degradation

Full text of DOI:10.1021/ac201465q

ARTICLE
pubs.acs.org/ac
Development of a Candidate Reference Method for the Simultaneous
Quantitation of the Boar Taint Compounds Androstenone,
3r-Androstenol, 3β-Androstenol, Skatole, and Indole in Pig Fat by
Means of Stable Isotope Dilution AnalysisÀHeadspace Solid-Phase
MicroextractionÀGas Chromatography/Mass Spectrometry
Jochen Fischer,† Paul W. Elsinghorst,‡,z Mark B€ucking,§ Ernst Tholen,|| Brigitte Petersen,^ and
Matthias W€ust*,†
†Department of Nutrition and Food Sciences, Bioanalytics, University of Bonn, Endenicher Allee 11-13, D-53115 Bonn, Germany
‡Pharmaceutical Institute, Pharmaceutical Chemistry I, University of Bonn, An der Immenburg 4, D-53121 Bonn, Germany
§Divison of Applied Ecology, Fraunhofer Institute, Auf dem Aberg 1, D-57392 Schmallenberg, Germany
Institute of Animal Science, Animal Breeding and Husbandry Group, University of Bonn, Endenicher Allee 15, D-53115 Bonn,
Germany
^Institute of Animal Science, Preventive Health Management, University of Bonn, Katzenburgweg 7-9, D-53115 Bonn, Germany
S Supporting Information
b
ABSTRACT: The steroidal pig pheromones androstenone (5R-androst-16-
en-3-one), 3R-androstenol (5R-androst-16-en-3R-ol), and 3β-androstenol
(5R-androst-16-en-3β-ol) as well as the heterocyclic aromatic amines
skatole and indole, originating from microbial degradation of tryptophan
in the intestine of pigs, are frequently recognized as the major compounds
responsible for boar taint. A new procedure, applying stable isotope dilution
analysis (SIDA) and headspace solid-phase microextractionÀgas chroma-
tography/mass spectrometry (HS-SPMEÀGC/MS) for the simultaneous
quantitation of these boar taint compounds in pig fat was developed and
validated. The deuterated compounds androstenone-d3, 3β-androstenol-d3,
skatole-d3, and indole-d6 were synthesized and successfully employed as internal standards for SIDA. The new procedure is
characterized by a fast, simple, and economic sample preparation: methanolic extraction of the melted fat followed by a freezing and
an evaporation step allows for extraction and enrichment of all five analytes. Additional time-consuming cleanup steps were not
necessary, as HS-SPME sampling overcomes fat-associated injector and column contamination. The method has been validated by
determining intra- and interday precision and accuracy as well as the limit of detection (LOD) and limit of quantitation (LOQ).
Additionally, a cross-validation for androstenone, skatole, and indole was carried out comparing the results of 25 back fat samples
obtained simultaneously by the new SIDAÀHS-SPMEÀGC/MS procedure with those obtained in separate GC/MS and high-
performance liquid chromatography fluorescence detection (HPLC-FD) measurements. The cross-validation revealed comparable
results and confirms the feasibility of the new SIDAÀHS-SPMEÀGC/MS procedure.
n order to avoid the incidence of boar taint in pork meat
surgical castration of male piglets without anesthesia is applied
Boar taint is an undesired off-flavor, which is often described as
sweaty, musky, urine- or faecal-like and is perceived when fresh
meat or meat products are heated prior to consumption.11À13
The major compound associated with boar taint is the steroidal
pig pheromone 5R-androst-16-en-3-one (ANON) that has an
unpleasant urine-like and sweaty odor.11 Being synthesized in the
testes, ANON is released into the bloodstream via the testicular
vein and subsequently accumulated in adipose tissue due to its
I
routinely in many European countries.1 Since it is evident that
piglets sustain serious distress when castrated without anesthesia,
castration is generally not accepted, particularly by animal welfare
organizations.1À3 Hence, several European countries, e.g., Germany,
Norway, and The Netherlands, voluntarily ban the painful
castration and start rearing entire male pigs instead, which
remains the most practicable alternative.3À9 Yet, the abandon-
ment of castration has a major drawback influencing the meat
quality: the incidence of boar taint in a significant number of
carcasses, which is estimated to exceed 10%.10
Received: June 10, 2011
Accepted: July 28, 2011
Published: July 29, 2011
r
2011 American Chemical Society
6785
dx.doi.org/10.1021/ac201465q Anal. Chem. 2011, 83, 6785–6791
|

Analytical Chemistry
ARTICLE
lipophilic character.12,13 Next to ANON, the musky-smelling
steroidal pig pheromones 5R-androst-16-en-3R-ol (3R-OL) and
5R-androst-16-en-3β-ol (3β-OL), all three belonging to the
family of C19-Δ16 steroids, contribute to boar taint.13À18 Pub-
lished data concerning the concentrations of 3R-OL and 3β-OL
in back fat are rare but are reported to be present in significantly
lower concentrations than ANON (ANON/3R-OL/3β-OL =
10/1/0.5).14,19À21 Nevertheless, the odor thresholds of 3R-OL
and 3β-OL are comparable with the threshold of ANON and
may have an influence in samples with high concentrations,
especially when heated before consumption.13 Furthermore,
skatole (SK) and indole (IND), both microbial degradation
products of tryptophan, formed in the intestine by specific
bacteria, contribute to boar taint.22,23 Their odor perceptions
are often described as musty and faecal-like.11 SK and IND are
partly absorbed by the intestinal mucosa, distributed via the
bloodstream, and finally enriched in adipose tissue and thus also
contribute to boar taint.24
As the abandonment of castration increases a growing amount
of smelling carcasses will potentially end up in the food chain.
Thus, the quantitation of boar taint compounds will be the major
challenge in order to keep smelling carcasses out of the food
chain and thereby implement boar mast successfully. In the past,
several analytical methods have been developed to determine
boar taint compounds in pig back fat samples ranging from a
number of immunoassays, e.g., radioimmunoassay (RIA),25
enzyme-linked immunosorbent assay (ELISA),26 or fluoroim-
munoassay (FIA),27 to several chromatographic applications like
high-performance liquid chromatography (HPLC),28,29 gas
chromatography (GC),19,30 or gas chromatography/mass spec-
trometry (GC/MS).31,32 More recently a liquid chromatogra-
phyÀtandem mass spectrometry method for the simultaneous
detection of SK, IND, and ANON has been published,33 but as
pointed out by the authors, matrix interferences were encoun-
tered when determining androstenone in fat. In addition, an
HPLC fluorescence detection (HPLC-FD) procedure for simul-
taneous detection of SK, IND, and ANON was published but
requires a derivatization step for fluorescence detection of
ANON.34 All other aforementioned methods separately deter-
mine either the indolic compounds or ANON. The quantitation
of 3R-OL and 3β-OL as a cumulative value has been previously
achieved by applying a GC flame ionization detection (GC-FID)
method.20 In addition, a more selective GC/MS procedure was
developed which allows for separate quantitation of both 3R-OL
and 3β-OL.19
As the lipophilic boar taint compounds are embedded in an
interfering fat matrix, a special focus has to be put on the sample
preparation to overcome fat-associated matrix effects like insuffi-
cient analyte extraction.33 The most reliable method to eliminate
such matrix effects is the application of a stable isotope dilution
assay (SIDA). Because isotopically labeled internal standards and
the corresponding analyte show almost identical physicochem-
ical properties, SIDA achieves superior accuracy and precision
and thereby delivers reliable results. Hence, SIDA is the method
of choice, especially when the quantitation of trace odorants
is required.35À37
Figure 1. Structures of the four synthesized deuterium-labeled internal
standards for boar taint analysis by SIDAÀHS-SPMEÀGC/MS: (1)
androstenone-d3, (2) 3β-androstenol-d3, (3) skatole-d3, (4) indole-d6
(ref 38).
fast, simple, and economic sample preparation and is highly
valuable for routine analysis of boar taint.
’ EXPERIMENTAL SECTION
Materials. All chemicals were obtained from Sigma-Aldrich
(Steinheim, Germany), Roth (Karlsruhe, Germany), VWR
(Darmstadt, Germany), or Alfa Aesar (Karlsruhe, Germany) in
analytical grade and were used without further purification unless
stated otherwise. The deuterium-labeled internal standards
(Figure 1) androstenone-d3 (1, ANON-d3), 3β-androstenol-d3
(2, 3β-OL-d3), and skatole-d3 (3, SK-d3) were prepared and
characterized in our laboratory as previously reported.38 In brief,
ANON-d3 was prepared from commercially available 3β-hydro-
xyandrost-5-en-17-one by selective deuteration of its B-ring
double bond using Pd/C and a Mg0/D2O/dioxane system,
followed by a Shapiro reaction of the ketone to provide 3β-
OL-d3, which was finally oxidized using PDC to obtain ANON-
d3 in 32% overall yield. SK-d3 was easily obtained by reduction of
methyl 1H-indole-3-carboxylate using LiAlD4. Aside of these,
indole-d6 (4, IND-d6) was obtained as follows: commercially
available perdeuterated indole-d7 (0.4 mmol, 50 mg, Isotec,
Miamisburg, Canada) was refluxed in a methanolic sodium
methoxide solution (10 mL, 1%, w/v). After complete exchange
(ND to NH) within 24 h as monitored by GC/MS, the solution
was evaporated to dryness in vacuo. The remaining residue was
redissolved in petroleum ether (10 mL) and subsequently
washed with water (5 mL). The organic layer was separated,
and the aqueous phase was extracted with petroleum ether (2 Â
10 mL). The combined organic layers were dried with anhydrous
sodium sulfate, filtered off, and finally evaporated in vacuo to
afford IND-d6 as a slightly red solid (44.1 mg, 88.2%); mp
49À51 ꢀC, lit.39 51À52 ꢀC. NMR spectra were recorded on a
The present work describes the development of a headspace
solid-phase microextractionÀgas chromatography/mass spec-
trometry (HS-SPMEÀGC/MS) procedure for the simultaneous
determination of the boar taint compounds SK, IND, ANON,
3R-OL, and 3β-OL in pig back fat samples using deuterium-
labeled internal standards for SIDA. The new method applies a
6786
dx.doi.org/10.1021/ac201465q |Anal. Chem. 2011, 83, 6785–6791

Analytical Chemistry
ARTICLE
Bruker Avance 300 DPX spectrometer operating at 300 MHz for
1H NMR and at 75 MHz for 13C NMR. Chemical shifts are given
in δ values (ppm) referring to the signal center using the solvent
peaks for reference (chloroform-d1: 7.26/77.0). Coupling con-
stants J are reported in hertz (Hz). To characterize the spin
multiplicity the following abbreviations are used: bs broad
singlet, d doublet. H NMR (CDCl3) δ: 6.58 (d, 0.2H, J
(H, H) = 2.0 Hz), 8.13 (bs, 1H, NH). 13C NMR (CDCl3) δ:
102.34 (C-3), 127.60 (C-3a), 135.61 (C-7a). GC/MS: purity
99%. MS CI (m/z, rel int %): 123([M-d5 + H]+, 28); 124 ([M-d6 +
H]+, 100). Isotopic purity: d5 22%, d6 78%.
Samples. An overall population of 1054 crossbred intact boars
of type Piꢀetrain  Baden-W€urttemberg hybrid sows were either
group penned or single penned and feed ad libitum until reaching
a slaughter weight between 85 and 95 kg. Back fat samples from
the neck region were taken at slaughter, wrapped in aluminum
foil, vacuum-packaged, and stored at À20 ꢀC until analysis.
Androstenone and skatole levels in all samples were determined
by conventional methods.19,29 For method development and
cross-validation 25 samples were selected with respect to their
back fat androstenone and skatole levels in order to cover the
whole physiologically relevant concentration range including
low, medium, and high levels of both compounds.
Sample Preparation and Headspace Solid-Phase Micro-
extraction. The back fat samples were thawed and separated
from skin. Subsequently, the fat tissue was diced and heated for
2 min at 180 W in a microwave. The occurring connective tissue
was separated from the liquid fat by decanting. An aliquot of
500 mg of the connective-tissue-free fat was transferred into a
2 mL plastic cap and spiked with 250 ng of ANON-d3, 250 ng of
3β-OL-d3, 50 ng of SK-d3, and 50 ng of IND-d6 to achieve final
concentrations of 500 ng/g ANON-d3 and 3β-OL-d3 as well as
100 ng/g SK-d3 and IND-d6. To allow for equilibration, the
sealed cap was thoroughly shaken for 30 s, stored for 10 min at
55 ꢀC, and again mixed for 30 s. Subsequently 1 mL of methanol
was added to the liquid fat. Extraction was achieved by a single
repetition of the above-mentioned mixing procedure (mix 30 s,
store 10 min at 55 ꢀC, mix 30 s). In order to separate the fat
phase, a freezing step was carried out by centrifuging the samples
(10 min, 6500 rpm, À15 ꢀC). The methanolic supernatant was
transferred into a 10 mL headspace vial and evaporated to
dryness at 40 ꢀC by a gentle stream of air. The headspace vial
was sealed and placed in an autosampler device (Varian Combi
Pal, Darmstadt, Germany), operating with a heated agitator and
an SPME assembly, using a fused-silica fiber coated with 65 μm
poly(dimethylsiloxane)/divinylbenzene (PDMS/DVB) (Supelco,
Bellefonte, PA). Automated HS-SPME sampling was carried out
as follows: equilibration for 5 min at 100 ꢀC; extraction for
30 min at 100 ꢀC; desorption for 20 min within the injector.
Gas Chromatography/Mass Spectrometry Conditions.
GC/MS analysis was carried out using a Varian GC-450,
equipped with a Varian VF-5 ms capillary column (30 m Â
0.25 mm  0.25 μm), coupled to a Varian MS-240 ion-trap (EI,
scan range 50À300 m/z, Varian, Darmstadt, Germany). The
flow was set at 1.0 mL/min using helium as carrier gas. The
temperature program was set as follows: start at 50 ꢀC, hold for
3 min, then raise to 160 ꢀC at a rate of 10 ꢀC/min, followed by a
rate of 5 ꢀC/min up to 240 ꢀC, hold for 1 min. The injector
temperature was kept at 270 ꢀC. Splitless injection was carried
out for 3 min, then the split valve was opened to result in a split
ratio of 1:100. To allow for quantitation, the mass spectra were
recorded in full scan mode using electron impact ionization (EI).
Subsequently, the peak area ratios of analyte and internal
standard (IS) were determined by displaying the specific mass
traces of each analyte and each corresponding IS in selected ion
monitoring (SIM) mode. The selected mass traces (m/z) were as
follows: SK m/z 130, SK-d3 m/z 133 + 134, IND m/z 117, IND-
d5/6 m/z 122 + 123, ANON m/z 257 + 272, ANON-d3 m/z
260 + 275, 3R-OL m/z 241 + 259 + 274, 3β-OL m/z 241 + 259 +
274, and 3β-OL-d3 m/z 244 + 262 + 277.
1
4
Calibration. Except for 3R-OL, calibration curves were de-
termined for each of the boar taint compounds by applying their
previously synthesized isotopomers as internal standards. In the
case of 3R-OL no corresponding isotopomer was synthesized as
internal standard, but 3β-OL-d3 was used instead for calibration
due to its structural similarity and similar retention time. A seven-
point matrix calibration was performed in duplicate by spiking
melted sow fat with ANON, 3R-OL, and 3β-OL and palm oil
with SK and IND, since sow fat contains small genuine amounts
of SK and IND. Defined quantities of deuterium-labeled internal
standards and analytes were added to the melted fat. The
concentrations of deuterium-labeled internal standards were
set constant for each calibration level and were 100 ng/g for
SK-d3/IND-d6 and 500 ng/g for ANON-d3/3β-OL-d3. The
amounts of analyte added to the calibration levels were as
follows: 0.5, 1, 10, 100, 250, 500, and 1000 ng/g for SK and
IND and 50, 250, 500, 750, 1000, 2500, and 5000 ng/g for
ANON, 3R-OL, and 3β-OL. The spiked samples were subse-
quently extracted as described before and analyzed by
SIDAÀHS-SPMEÀGC/MS. Linear calibration curves were ob-
tained by plotting the peak area ratios (analyte/IS) versus the
concentration ratios (analyte/IS).
Sensitivity. The limit of detection (LOD) and the limit of
quantitation (LOQ) were calculated as the concentration level
with a signal-to-noise ratio of 3:1 and 10:1, respectively.40
Accuracy and Precision. To determine intraday accuracy and
intraday precision, four replicates of a low calibration level
(50 ng/g SK/IND; 500 ng/g ANON/3R-OL/3β-OL) and four
replicates of a high calibration level (500 ng/g SK/IND; 2500
ng/g ANON/3R-OL/3β-OL) were prepared and subsequently
analyzed within 1 day. The interday accuracy and interday
precision were evaluated in the same manner, but analyzing only
one replicate of each calibration level per day (within 4 days).
The coefficient of variation (CV) and the relative error (RE)
were used as a measure for accuracy and precision of the method.
Cross-Validation. A set of 25 back fat samples of entire male
pigs was prepared as described above and subsequently ana-
lyzed by SIDAÀHS-SPMEÀGC/MS. For comparison, the
concentrations of ANON, SK, and IND in the same samples
were additionally determined by adapting a previously pub-
lished reversed-phase (RP)-HPLC procedure29 with fluores-
cence detection (FD) for the measurement of SK and IND as
well as a previously published GC/MS procedure for the
measurement of ANON.19 The experimental conditions were
as follows: RP-HPLC-FD measurement was carried out on a
Dionex Ultimate 3000 HPLC system coupled to a Dionex RF
2000 fluorescence detector (Dionex, Idstein, Germany) using a
C18 column (Hypersil ODS C18, 5 μm, 150 mm  4.6 mm, MZ
Analysentechnik, Mainz, Germany) as well as a C18 precolumn
(Hypersil ODS C18, 5 μm, 10 mm  4.6 mm, MZ Analysen-
technik, Mainz, Germany) applying isocratic elution (60%
0.02 M acetic acid, 25% acetonitrile, and 15% isopropyl alcohol).
The column thermostat temperature was set at 40 ꢀC, and the
injection volume was 10 μL. The detection parameters were set
6787
dx.doi.org/10.1021/ac201465q |Anal. Chem. 2011, 83, 6785–6791

Analytical Chemistry
ARTICLE
as follows: excitation wavelength, 270 nm; emission wavelength,
350 nm; emission gain, 4.0; emission response, 0.5 s; emission
sensitivity, medium. 2-Methylindole was used as internal
standard.
GC/MS analysis was carried out using an Agilent GC 6890
(Agilent Technology, Waldbronn, Germany), equipped with a
Restek Rxi-5HT capillary column (30 m  0.25 mm Â0.25 μm;
flow, 1 mL/min using helium as carrier gas; Restek GmbH, Bad
Homburg, Germany), coupled to an Agilent MSD 5973 mass
spectrometer (Agilent Technology, Waldbronn, Germany) op-
erating with a fast scan upgrade (software version E 02.00 SP1,
Agilent Technology, Waldbronn, Germany). The temperature
program was as follows: start 130 ꢀC, hold for 1 min, then raise to
265 at 30 ꢀC/min, subsequently raise to 300 at 5 ꢀC/min, hold
for 1 min, finally raise to 360 at 20 ꢀC/min, hold for 10 min.
Androstanone was used as internal standard.
’ RESULTS AND DISCUSSION
Internal Standards. A SIDAÀHS-SPMEÀGC/MS proce-
dure for the simultaneous quantitation of boar taint compounds
in pig back fat samples is presented for the first time. As the
application of SIDA requires a set of deuterium-labeled internal
standards, ANON-d3, β-OL-d3, and SK-d3 were synthesized in
our laboratory as previously reported (Figure 1).38 In addition,
IND-d6 was obtained from commercially available indole-d7 by
an H/D exchange in methanolic sodium methoxide solution.
This step was necessary since the deuterium atom attached to the
nitrogen atom of indole-d7 is labile to H/D-exchange whenever
indole-d7 is dissolved in a protic solvent, e.g., methanol (data not
shown). When the recorded NMR spectra of the synthesized
IND-d6 were analyzed, an additional but tolerable H/D-ex-
change was observed, as the spectra show signals at δ 6.58 (1H
NMR) and at δ 102.34 (13C NMR), indicating the presence of
hydrogen at carbon 3. The extent of the d5 species, possessing
hydrogens at carbon 3 and attached to the nitrogen atom, was
determined to be 22% as calculated from the fragmentation
patterns of the labeled and the corresponding unlabeled com-
pound obtained from chemical ionization (CI) MS spectra.
EI-MS spectra of the internal standards and the corresponding
boar taint compounds are presented in Figure S-1 (see the
Supporting Information). When the mass cluster of ANON-d3
is inspected, the ions m/z 242, 260, and 275 show an m/z value
that is three mass units higher than the respective ions m/z 239,
257, and 272 of genuine ANON. The same differences were
evident between 3β-OL-d3 (m/z 244, 262, 277) and the genuine
analytes 3R-OL and 3β-OL (m/z 241, 259, 274). When the EI-
MS spectra of SK and SK-d3 are analyzed, the ions m/z 130 and
m/z 132 are the most abundant. As the ion m/z 130 is assigned to
the [M À H]+ fragment of SK, the corresponding ion m/z 132
originates from the [M À D]+ fragment of SK-d3 and thereby
reveals the methyl group as the preferred place of fragmen-
tation.41 Nevertheless, the less abundant ions m/z 133 and m/z
134 of SK-d3 were selected and summed for quantitation, since a
significant overlap of mass signals between SK and SK-d3 was
detected, when using the ion m/z 132 for quantitation. Because
of the sixfold labeling of IND-d6, the most abundant ion m/z 123
of IND-d6 has a six mass units higher m/z value than the
corresponding ion m/z 117 of the unlabeled isotopomer IND.
Hence, all synthesized internal standards were found suitable for
SIDA, when selecting the appropriate mass traces for quantita-
tion in SIM mode.
Figure 2. HS-SPMEÀGC/MS analysis of a back fat sample of an entire
male pig containing 1273 ng/g androstenone, 142 ng/g 3β-androstenol,
295 ng/g 3R-androstenol, 181 ng/g skatole, and 60 ng/g indole.
Selective ion extraction of the full scan chromatogram (A) is exemplarily
displayed for skatole and androstenone determination. The arrows
indicate the peaks of skatole (B), androstenone (D) and their corre-
sponding isotopomers skatole-d3 (C) and androstenone-d3 (E).
Sample Preparation, SPME Sampling, and GC/MS Anal-
ysis. Sample preparation was carried out adapting a previously
reported procedure.34 After methanolic extraction of the boar
taint compounds and subsequent separation of interfering fat
components by a freezing step, the procedure was extended by an
evaporation step: the methanolic supernatant was decanted into
a headspace vial and finally evaporated to dryness in order to
increase the vapor pressure of the analyte within the headspace
vial. No further cleanup was necessary, as HS-SPME sampling
avoids injector and column contamination, which is often
observed with direct injection of fat containing extracts. In order
to find optimum extraction parameters for HS-SPME sampling,
various sampling conditions were tested. For this purpose,
different extraction temperatures (50, 70, 90, 100, 110 ꢀC) and
extraction times (10, 20, 30, 40, 50 min), as well as different fiber
coatings (DVB/PDMS 65 μm, polyacrylate (PA) 85 μm, PDMS
7 μm, DVB/carboxen (CAR)/PDMS 50/30 μm) were applied.
In addition, different amounts of fat (250, 500, 750, 1000 mg)
and different methanol volumes (1, 2, 3, 5 mL) were tested. The
decision whether a parameter was adequate or not was based on
the respective peak areas applying a univariate optimization
design. In the beginning, the optimal fiber coating was selected
by analyzing replicates with the four above-mentioned fibers.
From this experiment it was evident that the PDMS fiber was the
6788
dx.doi.org/10.1021/ac201465q |Anal. Chem. 2011, 83, 6785–6791

Analytical Chemistry
ARTICLE
Table 1. Calibration Data of the Developed SIDAÀ
HS-SPMEÀGC/MS Method
Table 2. Concentrations of Androstenone, 3r-Androstenol,
3β-Androstenol in 25 Back Fat Samples of Entire Male Pigs
Determined by SIDAÀHS-SPMEÀGC/MSa
concn range
(ng/g)a
LOD
(ng/g)a,b
LOQ
(ng/g)a,c
linearity (R2)
androstenone
3R-androstenol
3β-androstenol
analyte
skatole
sample
(ng/g)
(ng/g)
(ng/g)
0.9996
0.9998
0.9993
0.9929
0.9932
0.5À800
1.0À800
50À5000
70À5000
65À5000
0.1
0.5
35
0.5
1.0
60
1
90
91
<50
<50
87
<45
<45
<65
195
<65
<65
68
indole
2
androstenone
3R-androstenol
3β-androstenol
3
298
50
70
4
339
117
113
118
226
295
90
45
65
a ng analyte per gram melted fat. b LOD = limit of detection (special
samples were further diluted for LOD determination). c LOQ = limit of
quantitation.
5
366
6
388
7
452
8
461
76
most suitable for the pig pheromones ANON, 3R-OL, and 3β-
OL, but not suitable for the indolic compounds SK and IND. For
simultaneous determination of all five boar taint compounds,
both PDMS/DVB and PA fiber were suitable, but PDMS/DVB
gave slightly higher peak areas for SK and IND and thus was
selected as the fiber of choice. Nevertheless the PDMS fiber is
recommended, when the exclusive detection of ANON, 3R-OL,
and 3β-OL is required. When the extraction times (10, 20, 30, 40,
50 min) are considered, the pheromones and the indolic
compounds showed opposite tendencies: prolonged extraction
times lead to increased pheromone peak areas, whereas the peak
areas of the indolic compounds were decreasing. As the resulting
peak areas of the pheromones are generally less intensive than the
resulting peak areas of the indolic compounds, an extraction
temperature of 30 min was selected to get sufficient pheromone
peak areas. For the same reason, the extraction temperature (50,
70, 90, 100 ꢀC) was set at 100 ꢀC, even though SK and IND
showed higher peak areas when temperatures were set at 70 or
90 ꢀC. No significant differences were found when adding
different volumes of methanol (1, 2, 3, 5 mL). Hence, 1 mL of
methanol was added to all samples. When the sample amount
used for extraction is considered, it can be generally stated that an
increased amount of fat lowers the pig pheromone peak areas.
The opposite effect was found for the indolic compounds.
Finally, 500 mg of fat was applied in all analysis.
9
469
<65
129
120
123
245
174
160
171
185
341
159
275
263
375
351
331
453
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
658
179
376
225
194
441
352
253
488
820
289
453
362
644
562
655
815
699
724
853
862
921
1608
1713
1773
2060
2506
2870
3126
3172
3790
4978
mean
range
1410
330
181
90À4978
50À820
45À453
a Odor thresholds in cottonseed oil (μg/g) (ref 13): androstenone, 0.6;
3R-androstenol, 0.9; 3β-androstenol, 1.2. Frequently mentioned con-
sumer acceptance limit for androstenone (ref 46) in back fat: 500 ng/g.
Data not available for 3R-androstenol and 3β-androstenol.
All GC/MS chromatograms were recorded in full scan mode
and subsequently displayed in SIM mode for quantitation. With
respect to sensitivity, quantitation of ANON, 3β-OL, and 3R-OL
was carried out by summing peak areas of either two or three
specific ions. For the same reason, the selective ions m/z 133 and
m/z 134 of SK-d3 were used, when quantifying SK. In the case of
IND-d6, the ions m/z 122 and m/z 123 were suitable for
quantitation. A full scan GC/MS chromatogram of a boar back
fat sample, containing 1273 ng/g ANON, 142 ng/g 3β-OL, 295
ng/g 3R-OL, 181 ng/g SK, and 60 ng/g IND is presented in
Figure 2. Exemplarily, the specific mass traces of SK (m/z 130),
SK-d3 (m/z 133 + 134), ANON (m/z 257 + 272), and ANON-d3
(m/z 260 + 275) are additionally displayed in SIM mode in
Figure 2.
Calibration. Matrix calibration was performed by spiking
either sow fat (ANON, 3R-OL, and 3β-OL) or palm oil (SK,
IND) with defined quantities of deuterium-labeled internal
standards and analytes. Calibration curves were constructed by
plotting the peak area ratios (analyte/IS) versus the concentra-
tion ratios (analyte/IS). As illustrated in Table 1 the obtained
regressions show excellent linearity, possessing coefficients of
determination of >0.99 (n = 7) within the given concentration
ranges. Concentrations exceeding the highest calibration level of
each compound were not tested, since the applied concentra-
tions cover almost all reported physiological concentrations. The
LOQs of SK and IND were found to be 0.5 and 1.0 ng/g. The
corresponding LOD was 0.1 ng/g for SK and 0.5 ng/g for IND.
Higher LOQ (60 ng/g) and LOD (35 ng/g) values were found
for ANON, but are still satisfactory for boar taint analysis, when
taking into account that the frequently mentioned ANON limit
of consumer acceptance in back fat samples was found at either
500 or 1000 ng/g.7,42,43 The determined LOQ and LOD values
for 3R-OL and 3β-OL were as follows: LOQ (3R-OL), 70 ng/g;
LOD (3R-OL), 50 ng/g; LOQ (3β-OL), 65 ng/g; LOD (3β-
OL), 45 ng/g (Table 1). The values of precision (CV) and
accuracy (RE) were determined for each boar taint constituent
by measuring four replicate samples at a low and a high
concentration level within 1 day (intraday) and within 4 days
(interday), respectively. Satisfactory RE values were obtained for
all analytes, ranging between À8.6% and 7.0%. When the precision
of the method is considered, it can be stated that none of
the CV values exceeded 8.4% and hence were satisfactory as
6789
dx.doi.org/10.1021/ac201465q |Anal. Chem. 2011, 83, 6785–6791

Analytical Chemistry
ARTICLE
Table 3. Intra- and Interday Accuracy (RE) and Precision
(CV) of SIDAÀHS-SPMEÀGC/MS Determination for Sam-
ples Spiked with Low and High Concentrations of Each Boar
Taint Compound
3R-OL, and 3β-OL are considered in detail, it is obvious that the
ANON values exceed the 3R-OL and 3β-OL values in every case.
Figure S-3 (see the Supporting Information) shows the correla-
tion between the found concentrations of 3R-OL (Supporting
Information, Figure S-3A) and 3β-OL (Supporting Information,
Figure S-3B), respectively, and the concentrations of ANON.
The coefficients of determination, 0.68 and 0.80, in both linear
regression plots indicate that increased ANON levels are accom-
panied by increased 3R-OL and 3β-OL levels. This finding is in
good agreement with their biosynthesis, as 3R-OL and 3β-OL
originate from enzymatic reduction of ANON in the boar
testes.45 When only the determined androstenol concentrations
are compared, 3R-OL was found in higher concentrations than
3β-OL in most samples. The mean concentration ratio of
3R-OL/3β-OL was found to be 1.7 (Supporting Information,
Figure S-3C).
added
found (ng/g)
analyte
(ng/g)
(mean ( SD)
CV (%)a
RE (%)b
intraday (n = 4)
54 ( 1
skatole
50
500
50
3.0
5.0
2.8
1.0
4.0
5.3
5.4
7.5
6.7
5.5
À8.1
1.8
490 ( 24
48 ( 1
indole
3.5
500
500
2500
500
2500
500
2500
488 ( 5
2.2
androstenone
3R-androstenol
3β-androstenol
476 ( 19
2482 ( 132
518 ( 27
2289 ( 173
505 ( 33
2381 ( 132
4.4
0.3
À5.2
7.0
’ CONCLUSION
À2.5
3.3
A novel and reliable SIDAÀHS-SPMEÀGC/MS procedure
for the simultaneous quantitation of ANON, 3R-OL, 3β-OL, SK,
and IND in pig fat has been developed. The method validation,
including a cross-validation experiment, demonstrates excellent
performance, accuracy, and precision within the concentration
ranges of interest when investigating boar taint. Hence, the novel
SIDAÀHS-SPMEÀGC/MS procedure is recommended as a
candidate reference method for future boar taint studies in order
to ensure reliable results.
interday (n = 4)
54 ( 1
skatole
50
500
50
1.3
2.1
5.0
0.8
4.1
4.0
7.4
7.0
8.4
4.8
À8.0
5.4
470 ( 10
48 ( 2
indole
3.8
500
500
2500
500
2500
500
2500
485 ( 4
2.9
androstenone
3R-androstenol
3β-androstenol
493 ( 20
2477 ( 96
527 ( 39
2355 ( 166
535 ( 45
2417 ( 118
0.9
0.5
À7.1
3.9
’ ASSOCIATED CONTENT
À8.6
1.3
S
Supporting Information. Figures S-1ÀS-3. This material
b
a CV = coefficient of variation. b RE = relative error.
is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
well (Table 3). Moreover, all determined values are within the
guideline ranges recommended by Commission Decision concern-
ing the performance of analytical methods and the interpretation
of results.44
Corresponding Author
*Phone: ++49-228-732361. Fax: ++49-228-733499. E-mail:
matthias.wuest@uni-bonn.de.
Cross-Validation and Pheromone Levels. To further vali-
date the novel SIDAÀHS-SPMEÀGC/MS procedure a cross-
validation experiment for androstenone, skatole, and indole was
carried out. Therefore, the results of 25 back fat samples obtained
by SIDAÀHS-SPMEÀGC/MS were compared with the results
obtained from the same samples by RP-HPLC-FD measurement
of SK and IND29 and GC/MS measurement of ANON.19 For
comparison, the SIDA results were plotted against their corre-
sponding GC and HPLC results and were analyzed by orthogo-
nal distance regression.40 The obtained correlation plots are
shown in Figure S-2 (see the Supporting Information). For
androstenone and skatole no proportional systematic error (P =
95%) and no constant systematic error (P = 99%) were observed.
However, for indole a small, but statistically significant, constant
systematic error (P = 99%) was observed. Hence, the SIDA
method gave lower results when compared with those obtained
by HPLC-FD. Nevertheless, it is evident that the SIDAÀGC/
MS method presented here delivers reliable results.
Present Addresses
zCentral Institute of the Bundeswehr Medical Service Munich,
Ingolst€adter Landstrasse 102, D-85748 Garching, Germany.
’ ACKNOWLEDGMENT
The authors thank S. Kehraus for the recording of NMR
spectra. This work was financially supported by the Ministry for
Climate Protection, Environment, Agriculture, Nature Conser-
vation and Consumer Protection of North Rhine-Westphalia,
Germany.
’ REFERENCES
(1) Puppe, B.; Sch€on, P. C.; Tuchscherer, A.; Manteuffel, G. Appl.
Anim. Behav. Sci. 2005, 95, 67–78.
(2) Taylor, A.; Weary, D. M.; Lessard, M.; Braithwaite, L. Appl. Anim.
Behav. Sci. 2001, 73, 35–43.
(3) Leidig, M.; Hertrampf, B.; Failing, K.; Schumann, A.; Reiner, G.
Appl. Anim. Behav. Sci. 2009, 116, 174–178.
(4) Fredriksen, B.; Johnsen, A. M. S.; Skuterud, E. Res. Vet. Sci. 2010,
90, 352–357.
All measured concentrations for ANON, 3R-OL, and 3β-OL
in 25 real samples using the novel SIDAÀHS-SPMEÀGC/MS
are presented in Table 2. As can be seen, the concentrations of
the pheromones are highly variable with maximum values of
4978 ng/g ANON, 820 ng/g 3R-OL, and 453 ng/g 3β-OL.
When the determined pheromone concentrations of ANON,
(5) von Borell, E.; Baumgartner, J.; Giersing, M.; J€aggin, N.; Prunier,
A.; Tuyttens, F. A. M.; Edwards, S. A. Animal 2009, 3, 1488–1496.
6790
dx.doi.org/10.1021/ac201465q |Anal. Chem. 2011, 83, 6785–6791

Analytical Chemistry
ARTICLE
(6) D€usseldorfer Erkl€arung zur Ferkelkastration. http://www.zds-
bonn.de/duesseldorfererklaerung_zur_ferkelkastration.html (accessed
February 22, 2010).
(7) Babol, J.; Squires, J. Food Res. Int. 1995, 28, 201–212.
(8) Claus, R.; L€osel, D.; Lacorn, M.; Mentschel, J.; Schenkel, H.
J. Anim. Sci. 2003, 81, 239–248.
(42) Desmoulin, B.; Bonneau, M.; Frouin, A.; Bidard, J. Livest. Prod.
Sci. 1982, 9, 707–715.
(43) Claus, R.; Weiler, U.; Herzog, A. Meat Sci. 1994, 38, 289–305.
(44) Commission Decision of 12 August 2002 implementing Council
Directive 96/23/EC concerning the performance of analytical methods
and the interpretation of results.
(9) Bonneau, M.; Walstra, P.; Claudi-Magnussen, C.; Kempster, A. J.;
Tornberg, E.; Fischer, K.; Diestre, A.; Siret, F.; Chevillon, P.; Claus, R.;
Dijksterhuis, G.; Punter, P.; Matthews, K. R.; Agerhem, H.; Bꢀeague, M. P.;
Oliver, M. A.; Gispert, M.; Weiler, U.; von Seth, G.; Leask, H.; Font
i Furnols, M.; Homer, D. B.; Cook, G. L. Meat Sci. 2000, 54, 285–295.
(10) Chen, G.; Zamaratskaia, G.; Andersson, H.; Lundstrom, K.
Food Chem. 2007, 101, 439–448.
(45) Brophy, P. J.; Gower, B. D. Biochem. J. 1972, 128, 945–952.
(46) Walstra, P.; Claudi-Magnussen, C.; Chevillon, P.; von Seth, G.;
Diestre, K.; Matthews, K. R.; M.; Homer, D. B.; Bonneau, M. Livest. Prod.
Sci. 1999, 62, 15–28.
(11) Claus, R.; Weiler, U.; Herzog, A. Meat Sci. 1994, 38, 289–305.
(12) Patterson, R. L. S. J. Sci. Food Agric. 1968, 19, 31–38.
(13) Brooks, R. I.; Pearson, M. Meat Sci. 1989, 25, 11–19.
(14) Ohloff, G.; Maurer, B.; Winter, B.; Giersch, W. Helv. Chim. Acta
1983, 66, 192–217.
(15) Thompson, R. H.; Pearson, A. M.; Banks, K. A. J. Agric. Food
Chem. 1972, 20, 185–189.
(16) Patterson, R. L. S. J. Sci. Food Agric. 1968, 19, 434–438.
(17) Beery, K. E.; Sink, J. D.; Patton, S.; Ziegler, J. H. Food Sci. 2008,
36, 1086–1090.
(18) Prelog, V.; Ruzicka, L. Helv. Chim. Acta 1944, 27, 61–66.
(19) García-Regueiro, J.; Diaz, I. Meat Sci. 1989, 25, 307–316.
(20) Brennan, J. J.; Shand, P. J.; Fenton, M.; Nicholls, L. L.; Aherne,
F. X. Can. J. Anim. Sci. 1986, 66, 615–625.
(21) Bonneau, M. Livest. Prod. Sci. 1982, 9, 687–705.
(22) Claus, R.; Dehnhard, M.; Herzog, A.; Bernal-Barragan, H.;
Gimꢀenez, T. Livest. Prod. Sci. 1993, 34, 115–126.
(23) Deslandes, B.; Gariꢀepy, C.; Houde, A. Livest. Prod. Sci. 2001,
71, 193–200.
(24) Chen, G.; Zamaratskaia, G.; Andersson, H.; Lundstrom, K.
Food Chem. 2007, 101, 439–448.
(25) Andresen, Ø. Eur. J. Endocrinol. 1975, 79, 619–624.
(26) Claus, R.; Lacorn, M.; Ostertag, C. Meat Sci. 2008, 80,
934–938.
(27) Tuomola, M.; Harpio, R.; Knuuttila, P.; Mikola, H.; L€ovgren, T.
J. Agric. Food Chem. 1997, 45, 3529–3534.
(28) García Regueiro, J. A.; Rius, M. A. J. Chromatogr., A 1998,
809, 246–251.
(29) Dehnhard, M.; Claus, R.; Hillenbrand, M.; Herzog, A.
J. Chromatogr., B 1993, 616, 205–209.
(30) de Brabander, H. F.; Verbeke, R. J. Chromatogr., A 1986,
363, 293–3ꢀ02.
ꢀ
(31) Magard, M. Å.; Berg, H. E. B.; Tagesson, V.; J€aremo, M. L. G.;
Karlsson, L. L. H.; Mathiasson, L. J. E.; Bonneau, M.; Hansen-Møller, J.
J. Agric. Food Chem. 1995, 43, 114–120.
(32) Thompson, R.; Pearson, A. M. J. Agric. Food Chem. 1977,
25, 1241–1245.
(33) Verheyden, K.; Noppe, H.; Aluwe, M.; Millet, S.; Vandenbussche,
J.; Debrabander, H. J. Chromatogr., A 2007, 1174, 132–137.
(34) Hansen-Møller, J. J. Chromatogr., B 1994, 661, 219–230.
(35) Schieberle, P. In Characterization of Food: Emerging Methods;
Gaonkar, A. G., Ed.; Elsevier Science: Amsterdam, The Netherlands,
1995; pp 403À431.
(36) Cervino, C.; Asam, S.; Knopp, D.; Rychlik, M.; Niessner, R.
J. Agric. Food Chem. 2008, 56, 1873–1879.
(37) Chetschik, I.; Granvogl, M.; Schieberle, P. J. Agric. Food Chem.
2010, 58, 11018–11026.
(38) Fischer, J.; Elsinghorst, P. W.; W€ust, M. J. Labelled Compd.
Radiopharm. 2011in press.
(39) Motoyama, Y.; Kamo, K.; Nagashima, H. Org. Lett. 2009,
11, 1345–1348.
(40) Reichenb€acher, M.; Einax, J. W. Challenges in Analytical Quality
Assurance; Springer: Berlin, Heidelberg, Germany, 2011.
(41) Whitehead, T. R.; Price, N. P.; Drake, H. L.; Cotta, M. A. Appl.
Environ. Microbiol. 2008, 74, 1950–1953.
6791
dx.doi.org/10.1021/ac201465q |Anal. Chem. 2011, 83, 6785–6791