Evaluation of allergens in propolis by ultra-performance liquid chromatography/tandem mass spectrometry

Article abstract of DOI:10.1002/rcm.5025

The purified extract of propolis is used as a traditional remedy for the treatment of several diseases. Its beneficial activities are mainly attributed to the polyphenolic fraction. Nevertheless, propolis can cause allergic dermatitis and the sensitization rate in humans is increasing significantly mainly in younger subjects. The aim of this study was to develop and validate a selective and sensitive ultra-performance liquid chromatography tandem mass spectrometry analysis (UPLC/MS/MS) for the evaluation of the amount of caffeic acid and its esters with allergenic action in raw propolis samples and commercial formulations. The separation was carried out on a 1.7 μm C ₁₈ BEH Shield column and the detection performed by means of electrospray ionization in negative ion mode with multiple reaction monitoring. The confirmation of formulae of the precursor and product ions was accomplished by injection into a high-resolution system (FTICR-MS) using accurate mass measurements. The error was below 1.4 ppm.The range of the standard curves was 0.5-10 μg/mL and dihydrocaffeic acid was used as internal standard (IS). The lower limit of detection (LLOD) for 3-methyl-2-butenyl- (3M2B), 3-methyl-3-butenyl- (3M3B), 2-methyl-2-butenyl- (2M2B), benzyl- (CABE), phenylethylcaffeic acid (CAPE) and for caffeic acid (CA) and the IS was 0.1 and 0.3 μg/mL, respectively. The recoveries were in the range 96-104% and the intra- and inter-day precisions were within 6.2%. In the European (n = 8) and Asiatic (n = 3) propolis the most abundant allergens were CABE>3M2B> CAPE>3M3B>CA>2M2B. These compounds were not found in the red (n = 1) and green (n = 1) Brazilian propolis. Hydroalcoholic extracts (n = 6) and tablets (n = 6) were analyzed by the proposed UPLC/MS/MS method. The results showed that in the commercial products CABE, 3M2B, CAPE and 3M3B were also the most abundant.

Full text of DOI:10.1002/rcm.5025

Research Article
Received: 27 December 2010
Revised: 18 March 2011
Accepted: 18 March 2011
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682
(wileyonlinelibrary.com) DOI: 10.1002/rcm.5025
Evaluation of allergens in propolis by ultra‐performance liquid
chromatography/tandem mass spectrometry
*
Claudio Gardana and Paolo Simonetti
Università degli Studi di Milano, DiSTAM – Dipartimento di Scienze e Tecnologie Alimentari e Microbiologiche, Via Celoria 2,
20133 Milan, Italy
The purified extract of propolis is used as a traditional remedy for the treatment of several diseases. Its beneficial
activities are mainly attributed to the polyphenolic fraction. Nevertheless, propolis can cause allergic dermatitis and
the sensitization rate in humans is increasing significantly mainly in younger subjects. The aim of this study was to
develop and validate a selective and sensitive ultra‐performance liquid chromatography tandem mass spectrometry
analysis (UPLC/MS/MS) for the evaluation of the amount of caffeic acid and its esters with allergenic action in raw
propolis samples and commercial formulations. The separation was carried out on a 1.7 μm C18 BEH Shield column
and the detection performed by means of electrospray ionization in negative ion mode with multiple reaction
monitoring. The confirmation of formulae of the precursor and product ions was accomplished by injection into a
high‐resolution system (FTICR‐MS) using accurate mass measurements. The error was below 1.4 ppm.The range of
the standard curves was 0.5–10 μg/mL and dihydrocaffeic acid was used as internal standard (IS). The lower limit of
detection (LLOD) for 3‐methyl‐2‐butenyl‐ (3M2B), 3‐methyl‐3‐butenyl‐ (3M3B), 2‐methyl‐2‐butenyl‐ (2M2B), benzyl‐
(CABE), phenylethylcaffeic acid (CAPE) and for caffeic acid (CA) and the IS was 0.1 and 0.3 μg/mL, respectively. The
recoveries were in the range 96–104% and the intra‐ and inter‐day precisions were within 6.2%. In the European (n = 8)
and Asiatic (n = 3) propolis the most abundant allergens were CABE>3M2B>CAPE>3M3B>CA>2M2B. These
compounds were not found in the red (n = 1) and green (n = 1) Brazilian propolis. Hydroalcoholic extracts (n = 6) and
tablets (n = 6) were analyzed by the proposed UPLC/MS/MS method. The results showed that in the commercial
products CABE, 3M2B, CAPE and 3M3B were also the most abundant. Copyright # 2011 John Wiley & Sons, Ltd.
Propolis is a complex natural product containing material
collected from buds or resin and beeswax produced by
honeybees. Its polyphenolic profile depends on the region of
collection[1] and the chemical composition differs substantially
between propolis samples from tropical and temperate
regions. In fact, in Europe, Asia, North America and some
South American regions the composition of propolis is very
similar, the main constituents being flavonoids, aromatic acids
and their esters.[2] In contrast, green, red and brown Brazilian
propolis contains less flavonoids and more quinic acid esters.
The typical compound of green Brazialian propolis is a
prenylated p‐coumaric acid derivative, named artepillin C.[3]
Propolis seems relatively safe, with a calculated safe dose of
about 1.4mg/kg body weight (BW)/day in humans,[4] and the
purified extract of propolis is used as a traditional remedy for its
antibacterial, antiviral, antifungal, anti‐inflammatory, antiox-
idant, immunomodulatory, cariostatic, antitumor[5] and anti‐
Helicobacter pylori[6] activities. These beneficial activities are
mainly attributed to flavonoids and phenolic acids with their
esters such as caffeic acid phenethyl ester (CAPE).[7] Never-
theless, propolis can cause occupational allergic eczematous
contact dermatitis mostly in apiarists,[8] but also in consumers
with an allergic predisposition.[9] The sensitization rate to
propolis in patients suffering from dermatitis has been reported
to vary between 1.2 and 6.6%[10–17] and a significant increasing
trend from 1995 (2%) to 2002 (13.7%) in young subjects was
reported.[18] The allergenic action seems to be due to some
caffeic acid derivatives[16] (Fig. 1), and 3‐methyl‐2‐butenyl
caffeate (3M2B) seems the most active compound.[19]
Propolis composition has been evaluated using different
analytical methods including micellar electrokinetic chroma-
tography (MEKC),[20] gas chromatography/mass spectrometry
(GC/MS),[21] nuclear magnetic resonance (NMR),[22] liquid
chromatography (LC)/MS,[2] electrospray ionization (ESI)‐
MS[23] and, recently, a semi‐qualitative determination was
performed by ambient sonic‐spray ionization MS.[24]
Despite the great deal of available analytical information
regarding propolis composition, data concerning the allergenic
fraction are scarce. Recently, Aliboni et al.[25] reported the
content of two non‐prenylated esters with low allergenic
activity, benzyl cinnamate and benzyl salicylate, in raw propolis
samples. Medana et al.[26] reported the determination of some
flavonoids and phenolic acid esters by LC/ESI‐MS/MS in
propolis samples of different origin, but they evaluated only
one prenyl caffeate and its chromatographic peak overlapped
benzyl caffeate. The discrimination between different prenyl
caffeates and their quantitative determination are important to
assess the allergenic potential of propolis. Thus, for the first time
we developed and validated a sensitive and simple method
using ultra‐performance liquid chromatography coupled with
* Correspondence to: C. Gardana, DiSTAM – Division of
Human Nutrition, University of Milan, Via Celoria 2, 20133
Milan, Italy.
E‐mail: claudio.gardana@unimi.it
Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682
Copyright # 2011 John Wiley & Sons, Ltd.

C. Gardana and P. Simonetti
O
HO
R
O
HO
Compound
CV Elab
Product ions
R
H2C
Benzyl caffeate
24
21
21
21
20
21
134(100), 133(16),
161(11), 178(8), 179(3),
269(1)
CABE, (m/z)- 269
CH2
CH3
3-methyl-2-butenyl caffeate
134(100), 178(9), 179(5),
135(4), 161(4), 247(2)
3M2B, (m/z)- 247
H3C
CH3
2-methyl-2-butenyl caffeate
21 134(100), 178(15), 135(4),
CH2
2M2B, (m/z)- 247
179(4), (161), 247(3)
CH3
Phenylethyl caffeate
18
135(100), 161(31),
134(17), 179(13), 283(1)
H2C CH2
CAPE, (m/z)- 283
CH2
H3C
CH2
3-methyl-3-butenyl caffeate
24
15
21
15
135(100), 179(42),
134(19), 161(26), 247(5)
3M3B, (m/z)- 247
Caffeic acid
135(100), 134(12), 179(2)
H
CA, (m/z)- 179
O
HO
15
15
137(100), 121(38),
109(21), 135(13), 181(5)
OH
HO
dihydrocaffeic acid (IS)
DHCA, (m/z)- 181
CV: cone voltage (V), Elab: energy applied during collision (eV)
Figure 1. Chemical structure of IS, caffeic acid and its esters determined in propolis raw
samples and commercial preparations. The main product ions obtained in ESI‐MS/MS
negative ion mode are reported in order of percent relative intensities (in brackets).
tandem mass spectrometry (UPLC/MS/MS) for the determi-
Cruz Tech. (Santa Cruz, CA, USA). Methanol, acetonitrile,
ethyl acetate, tetrahydrofuran (THF), HCl, NaCl, NaHCO3 and
formic acid were supplied by Merck (Darmstadt, Germany).
Ultrapure water was obtained from a MilliQ apparatus
(Millipore, Milford, MA, USA). The raw propolis samples were
collected from the following areas: Italy (1–3), Macedonia (4–5),
Poland (6), Uruguay (7), France (8), Nepal (9), China (10–11),
Brazil (12–13). The propolis samples were stored at −20°C.
The commercial products containing propolis were purchased
in a local drugstore.
nation of caffeic acid and its esters (CAEs) with allergenic
action. The proposed method was then subsequently applied
to determine these compounds in raw propolis samples
from different ethnogeographical origin and commercial
preparations containing propolis.
EXPERIMENTAL
Chemicals and reagents
Synthesis of CABE, 2M3B and 3M3B
Caffeic acid (CA), 3‐methyl‐2‐butenyl‐CA (3M2B), internal
standard (dihydrocaffeic acid, DHCA), caffeic acid phenyl
ethyl ester (CAPE), dimethylaminopyridine (DMAP) and 1,3‐
dicyclochexylcarbodiimide (DCC) were purchased from
Sigma‐Aldrich (St. Louis, MO, USA). Methyl caffeate was
from Chromadex (Milan, Italy), ethyl caffeate from Water-
stone Tech. (Carmel, IN, USA) and ethyl ferulate from Santa
2‐Methyl‐3‐butenyl‐, benzyl‐ and 3‐methyl‐3‐butenyl caffeate
were synthesized as described by Sestili et al.[27] with slight
modifications. Briefly, 0.11 mmoL DMAP, 0.11 mmoL DCC
and 1.7 mmol of the alcohol were added to a solution of
1.1 mmoL CA in 15 mL of THF at 4°C. The mixture was stirred
at room temperature and after 24 h the reaction mixture was
wileyonlinelibrary.com/journal/rcm Copyright # 2011 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682

Determination of propolis allergens by UPLC/MS/MS
filtered using a Gooch funnel. The solution was dried, the
residue dissolved in ethyl acetate and the sample sequentially
washed with 0.1 N HCl, NaCl and then NaHCO3 saturated
solutions. The organic solution was evaporated in vacuum, the
dry residue suspended in methanol and then centrifuged
at 1000 g for 10 min before injecting 200 μL in the semi‐
preparative LC system. The chromatographic system con-
sisted of an Alliance model 2695 (Waters, Milford, MA, USA)
equipped with a diode‐array detector model 2998 (Waters)
and a 10 μm ODS‐2 Partisil 10 (Whatman, Clifton, NJ, USA)
column was used for the separation at a flow rate of 2.5 mL/
min. The eluents were 0.1% HCOOH (A) and CH3CN (B) and
the linear gradient was as follows: 35% to 70% B in 10 min and
then 70% B for 5 min. The peak of the analyte was collected by
a WFC II fraction collector (Waters), evaporated to dryness
under nitrogen and the residue suspended in methanol.
The fragmentation transitions for the multiple reaction
monitoring (MRM) were m/z 179→135 for CA, m/z
247→135 for 3M3B, m/z 247→134 for 3M2B and 2M2B, m/z
269→134 for CABE, m/z 283→135 for CAPE and m/z 181→137
for DHCA, with a dwell time of 0.2 s per transition.
FTICR‐MS conditions
All analyses were performed on an APEX II Fourier transform
ion cyclotron resonance (FTICR) mass spectrometer equipped
with a 4.7 Tesla magnet (Bruker Daltonics, Billerica, MA, USA)
and an Apollo ESI source (Bruker Daltonics). ESI and ion
transfer conditions were optimized in the range 112–250 Da
using a standard solution of NaTFA in acetonitrile.
Samples were introduced into the ESI source via a Cole‐
Parmer Series 74900 syringe pump at a rate of 120μL/h, and all
mass spectra were acquired in negative ion mode. The
electrospray capillary voltage was 3.5kV, the capillary skimmer
was set to 20V, a countercurrent flow of nitrogen gas (150°C)
was employed for the desolvation and argon was used to
improve the fragmentation. The scan range was from m/z 100
to 1000 at a resolution of 50.000–120.000.
Sample preparation
Raw propolis
Frozen propolis samples (200 g) were firstly finely ground in a
mill, passed through a 500 μm (35 mesh) sieve, and then 1 g of
the powder was sonicated with 1 mL IS (2 mg/mL) and 70 mL
of ethyl acetate for 10 min. The mixture was centrifuged at
1500 g for 5 min, and the supernatant transferred into a flask,
while the solid residue was extracted once as described above.
The extracts were dried under nitrogen and the residues
dissolved in 100 mL methanol. The solutions obtained were
serially diluted and centrifuged at 4000 g for 1 min before the
UPLC/MS/MS analysis.
All experimental event sequences were controlled using
Bruker Daltonics XMASS 7.0.8 software.
Preparation of working solutions
The primary stock solutions of CA esters (0.1 mg/mL), CA and
DHCA (1 mg/mL) were prepared in methanol and diluted to
give working solutions in the range of 0.5–10 μg/mL. All stock
solutions and the working solutions were stored at −80°C
and −20°C, respectively.
Commercial preparations
Method validation
The IS methanolic solution (1 mL, 1 mg/mL) was added to the
alcoholic extracts (1 mL) and the volume adjusted to 10 mL
with methanol. Tablet samples (5 g) were finely ground, and
then about 0.5 g of the powder was sonicated with 1 mL IS
(1 mg/mL) and 8 mL of methanol for 10 min. The mixture was
centrifuged at 1500 g for 5 min, the supernatant transferred
into a flask and the volume adjusted to 10 mL with methanol.
The solutions obtained were serially diluted and centrifuged at
4000 g for 1 min before the UPLC/MS/MS analysis.
Calibration curves were constructed for each standard and
internal standard at six concentration levels; four independent
determinations were performed at each concentration and
regression analysis was carried out to determine the linearity of
the calibration graphs. Moreover, dihydrocaffeic acid was used
as IS to evaluate the loss of analytes during sample preparation.
LLOQ was defined by the lowest injected inter‐day concentra-
tion whose RSD% resulted to be <20%,[28] while LLOD was
defined by the lowest concentration the assay can differentiate
from background levels (signal‐to‐noise (S/N) ratio >3). The
accuracy (matrix effect) of the procedure was determined by a
recovery test according to the published method.[29] Briefly,
three raw propolis samples (1g) were spiked with different
amounts of IS, CA and CAEs (1–5–10mg). The spiked samples
were extracted under optimized conditions, and the recovery
rates calculated. Each sample was extracted in triplicate and
analyzed in duplicate. The peak identity related to the IS, CA
and CAEs was confirmed by co‐chromatography with an
authentic standard and by MS/MS. Intra‐ and inter‐day
precisions of the assay were verified by analyzing spiked
samples three times for five consecutive days. Precision was
determined by evaluating standard deviations of the amounts
and of the retention times. Two analysts evaluating the amounts
of the analytes in a raw propolis sample estimated the
ruggedness of the proposed UPLC/MS/MS method. Each
analyst performed six tests, and standard and extract solutions
were injectedin triplicate. Robustness was estimated byvarying
several chromatographic conditions such as flow rate
UPLC/MS/MS conditions
The chromatographic system consisted of an Acquity UPLC
system (Waters) coupled to a Quattromicro triple quadrupole
mass spectrometer (Waters). A 1.7μm C18 BEH Shield column
(150 × 2.1mm, Waters) was used for the separation at a flow rate
of 0.45 mL/min. The column was maintained at 40°C and the
separation was performed by means of a linear gradient elution
(eluent A, 0.05% HCOOH; eluent B, 0.05% HCOOH in
acetonitrile). The gradient was as follows: 20 to 30% B in
4 min, 30 to 40% B in 5 min, 40 to 60% B in 3 min, 60 to 90% B
in 1 min and then 90% B for 2 min. The capillary voltage was set
to 2.7 kV, the cone voltage and the energy applied during
fragmentation (Elab) was specific for each compound (see
Fig. 1). All mass data were obtained in the negative ion mode.
Thesource temperature was 130°C, the desolvation temperature
was 380°C and argon was used at 2.1× 10−3 mbar to improve
fragmentation in the collision cell. Masslinx 4.0 acquired data
with the Quan‐Optimize option for the fragmentation study.
Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682 Copyright # 2011 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm

C. Gardana and P. Simonetti
(n = 5) were evaluated analyzing spiked samples in triplicate
and the resulting RSD% values were lower than 5.5 and 6.2%,
respectively. The RSD% of the retention times was lower than
0.22%. The results on the effect of the external factor on the
degree of reproducibility of the UPLC/MS/MS method were
compared statistically; there was no significant difference in the
evaluation of the amount of CA and CAEs in the analyzed
propolis extracts (p = 0.29). Regarding robustness, slight varia-
tions in buffer composition and pH, flow rate and column
temperature did not change the peak shape and resolution.
Moreover, moderate variations in cone ( 3V) and capillary
( 0.5 kV) voltage did not influence significantly the quantifica-
tion of the analytes. It should be pointed out that increasing
cone voltage values drastically reduced the quantitative data,
mainly for caffeic acid and the IS. Thus, the proposed methods
were found to be very robust and rugged even if careful
attention had to be paid to cone voltage values.
0.05 mL/min, column temperature 5°C, organic strength
50%, cone voltage 5V and capillary voltage 0.3kV. Data
were analyzed by one‐way analysis of variance (ANOVA) for
independent samples considering significant a level of p < 0.05.
Stability studies
The powder and extracts of propolis and the standards
solutions were stored at 4°C and −20°C for up to 90 days to
evaluate their stability. Moreover, standard solutions and
propolis extracts were placed in the autosampler at 20°C and
their stability evaluated overnight.
UPLC/MS/MS optimization
In order to optimize ESI conditions for IS, CA and CAEs,
quadrupole full scans were carried out both in positive and
negative ion mode. Precursor ion, capillary, extractor and RF lens
voltages were optimized by direct infusion experiments, while
product ions, cone voltage and the energy applied during frag-
mentation (Elab) were evaluated by the Quan‐optimize option.
Stability studies
The means of CAEs recovered from raw propolis samples and
commercial formulations stored at 4°C for up to 90 days were
consistent (>94%) and caffeic acid did not increase indicating
that its esters were not hydrolyzed. Standard solutions stored
at −20°C for up to 90 days showed a reduction lower than
2.4% for all the analytes. Standard solutions and propolis
extracts in methanol were stable in the autosampler at 20°C
overnight (RSD < 1.9%).
RESULTS AND DISCUSSION
Method validation
The calibration curves for all the analytes were linear in the
range of 0.5–10 μg/mL, and the LLOD was 0.3 μg/mL for the
IS and CA and 0.1 μg/mL for CAEs. The mean percentage
recovery of the analytes from spiked raw samples of propolis
was in the range of 96–104% (Table 1). Selectivity was
assessed by co‐chromatography with an authentic standard
and MS/MS evaluation. The intra‐ and inter‐day precisions
UPLC/MS/MS optimization
The negative ESI ion mode was selected as it was more sensitive
and the IS, CA and CAEs gave a better fragmentation. Figure 1
shows the chemical structures of these compounds, their main
Table 1. Recovery test of the UPLC/MS/MS assay
Added
(mg)
Detected
(mg)
s.d.
Expected
(mg)
Recovery
(%)
Analyte
CA
(mg)
RSD %
Error %
0
1
5
10
0
1
5
10
0
1
5
10
0
1
5
10
0
1
5
10
0
1
5
10
3.7
4.5
8.9
13.4
9.7
0.2
0.2
0.2
0.3
0.3
0.3
0.4
0.3
0.5
0.5
0.5
0.6
0.4
0.3
0.4
0.3
0.6
0.6
0.8
0.8
0.3
0.3
0.4
0.4
5.9
4.9
5.1
5.2
2.9
3.0
3.2
2.6
3.9
4.0
4.1
4.3
7.7
6.3
6.6
4.9
4.1
4.1
4.9
4.9
3.1
3.1
3.2
3.5
4.7
8.7
13.7
96
102
98
−4.3
2.3
−2.2
3M3B
3M2B
2M2B
CABE
CAPE
10.9
15.1
19.2
11.8
12.5
16.3
21.1
4.7
10.7
14.7
19.7
102
103
97
2.0
2.7
−2.5
12.8
16.8
21.8
97
97
97
−2.6
−3.2
−3.4
5.5
5.7
9.7
14.7
96
104
98
−3.5
4.1
−2.0
10.1
14.4
15.1
16.3
20.7
24.5
10.8
11.3
16.1
20.2
16.1
20.1
25.1
101
103
98
1.2
3.0
−2.4
11.8
15.8
20.8
96
102
97
−4.3
1.8
−2.9
wileyonlinelibrary.com/journal/rcm Copyright # 2011 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682

Determination of propolis allergens by UPLC/MS/MS
product ions and the cone voltage value and the energy applied
during collision for their production. The collision‐induced
dissociation (CID) spectrum of CA gave a typical product ion
with m/z 135 corresponding to [M–COO]–, while its 3‐methyl‐3‐
butenyl ester gave product ions with m/z 179 and 135
corresponding to [M–prenyl]– and [M–prenyl–COO]–, respec-
tively. In the same way, the CAPE fragmentation gave product
ions with m/z 179 and 135 corresponding to [M–phenylethyl]–
and [M–phenylethyl–COO]–, respectively. In contrast, in MS/
MS 3M2B and 2M2B yielded typical product ions with m/z 178
and 134 corresponding to [M–prenyl]‐• and [M–prenyl–COO]‐•,
respectively. The same product ions were also obtained from
CABE fragmentation, and in Figs. 2 and 3 MS2 spectra of the
different CAEs and their hypothetical fragmentation pattern are
reported, respectively. Thus, 3M2B, 2M2B and CABE gave
common product ions resulting from homolytic cleavage with
formation of a radical product with m/z 134, and this behavior
was presumably related to the stability of the derived prenyl
and benzyl radicals. In particular, the benzyl radical derived
from CABE fragmentation can rearrange to give a more stable
radical ion (Fig. 3). In contrast, CAPE and 3M3B gave a
heterolytic cleavage and the main product ion was the ion with
m/z 135. Thereby, the different fragmentations depend on the
alcohol moiety structure. Specifically, it seems that when the C
atom linked to the carboxyl group is allylic the MS/MS
fragmentation gives negative radical products (m/z 134 and
135
135
100
100
A
C
3M3B
CAPE
%
%
179
133
161
161
134
133
179
247
283
0
0
134
134
100
100
B
D
3M2B
2M2B
CABE
%
%
0
133
161
178
178
247
269
179
110 120 130 140 150 160 170 180 190 200 210 220 230 240 m/z
0
110
m/z
130
150
170
190
210
230
250
270
Figure 2. Product ions of 3M3B (A), 3M2B and 2M2B (B), CAPE (C), and CABE (D) obtained
in ESI negative mode at lower energy applied during collision.
CH
2
-
CH
-
2
H C
2
CH
3
CH
2
O
HO
HO
HO
HO
CAPE
3M3B
-CO
O
OH
2
(m/z)
179
(m/z)
135
C
C
CH
HO
HO
HO
HO
CH
C
CH
3
H C
2
C
CH
3
(m/z)
161
(m/z)
133
CH
3
H C
2
O
CH
3
HO
HO
HO
HO
OH
3M2B
CABE
CH
2
-CO
2
(m/z)
178
(m/z) 134
CH
CH
2
2
Figure 3. Hypothetical fragmentation pattern of the different caffeic acid esters.
Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682 Copyright # 2011 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm

C. Gardana and P. Simonetti
178), while negative product ions (m/z 135, 179) are obtained if
the allyl group is not present.
comppounds identified through FTICR‐MS/MS experiments
along with accurate mass, molecular formula, error (in ppm)
between the mass found and the accurate mass of CA and
CAEs, and the MS/MS ions.
The MS/MS spectra of CAEs, obtained at lower Elab, showed
also the presence of product ions with m/z 161, which could be
the result of H2O elimination from the deprotonated caffeoyl
group. On the contrary, the product ion with m/z 161 was not
present in the MS/MS spectrum of CA. Thus, it seems that this
fragment is derived from the cleavage of the ester bond (Fig. 3).
Methyl‐, ethyl caffeate and ethyl ferulate were evaluated for
their potential use as IS but for different reasons they were not
used. Methyl and ethyl caffeate showed a chromatographic
behavior, a deprotonated molecule and a fragmentation pattern
similar to ferulic‐ and 3,4‐dimethylcaffeic acid, respectively, two
compounds present in several propolis samples.[2] Regarding
ethyl ferulate, although it gave typical product ions, it was not
possible to employ it because it is a compound naturally found
in some European propolis samples (data not shown).
Amount of CA and CAEs in propolis samples
A good separation of the analytes was achieved with a C18
BEH Shield sub‐2 μm column and gradient mode elution, and
Fig. 4 shows the chromatograms of a raw propolis extract
acquired in MRM mode. Benzyl caffeate (0.85 0.39%), 3M2B
(0.64 0.33%), CAPE (0.61 0.29%) and 3M3B (0.54 0.38%)
were the most abundant CAEs found in the analysis of
European and Asiatic raw propolis samples (Table 3), respec-
tively. In these samples lower amounts of 2M2B (0.2 0.1%)
were also present. On the contrary, these compounds were not
found in the red and green Brazilian raw propolis samples.
Commercial hydroalcoholic extracts and tablets were also
analyzed by the proposed UPLC/MS/MS method. The results
obtained showed that CABE, 3M2B, CAPE and 3M3B were the
most abundant CAEs in all of them, indicating that the tested
commercial products were prepared using European or
Asiatic raw propolis samples.
FTICR‐MS
High‐resolution collision‐induced dissociation (CID) was
employed to confirm the identity of the analytes and their
product ions in the negative mode using ESI‐FTICR‐MS. The
resolution over all fragment ions was from 60.000 to 90.000. In
the MS2 spectra of 3M2B and 3M2B, the [M–H]– ion at m/z 247
gave mainly product ions at m/z 178 and 134, and a lower
amount of product ions at m/z 179 and 135. The MS2 spectra of
CABE showed similar fragmentation features, confirming that
these CAEs underwent homolytic cleavage with formation of
a radical product. Thus, the FTICR‐MS data confirm the results
obtained using the triple quadrupole system.
The FTICR‐MS/MS spectra of 3M3B and CAPE showed
ions corresponding to the deprotonated molecule [M–H]–, the
loss of the acyl moiety (m/z 179) and the loss of acyl and CO2
(m/z 135) as characteristic fragment ions. These CAEs gave also
a deprotonated molecule at m/z 161, resulting from water
elimination from the deprotonated caffeoyl group. On the
contrary, the product ion at m/z 161 was not found in the MS2
spectrum of 3M2B, 2M2B, CABE and CA. Table 2 shows the
CONCLUSIONS
A sensitive, specific and reproducible UPLC/MS/MS method
for the quantitative determination of caffeic acid and CAEs with
allergenic action in propolis samples has been developed and
validated. The amounts of these compounds was successfully
determined in raw propolis and commercial preparations. The
results obtained showed that in all the analyzed extracts the
most abundant caffeates were benzyl‐, 3‐methyl‐2‐butenyl‐,
phenylethyl‐ and 3‐methyl‐3‐butenyl ester, respectively. De-
spite being bioactive compounds with both toxicological and
pharmacological effects, they are not actually exploited as
markers of propolis quality – the latter is normally based only
on the content of flavonoids. Thus, it is advisable that
Table 2. Precursor ion and fragments of high‐resolution ESI CID for caffeic acid and its esters
Product ion
Calculated
Error
Compound
3M2B
[M–H]–
(m/z)
(m/z)
(ppm)
MF
247.09758
178.02719
134.03744
178.02726
134.03739
178.02731
134.03736
179.03515
135.04529
161.02453
133.02969
179.03498
135.04523
161.02459
133.02968
135.04522
178.02716
134.03732
178.02716
134.03732
178.02716
134.03732
179.03498
135.04515
161.02442
133.02950
179.03493
135.04515
161.02442
133.02950
135.04515
0.2
0.9
0.6
0.5
0.8
0.3
0.9
1.0
0.7
1.4
0.3
0.6
1.1
1.4
0.5
C9H6O4
C8H6O2
C9H6O4
C8H6O2
C9H6O4
C8H6O2
C9H7O4
C8H7O2
C9H5O3
C8H5O2
C9H7O4
C8H7O2
C9H5O3
C8H5O2
C8H7O2
3M2B
CABE
3M3B
247.09751
269.08193
247.09764
CAPE
283.09758
CA
179.03504
MF: molecular formula [M–H]–.
wileyonlinelibrary.com/journal/rcm Copyright # 2011 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682

Determination of propolis allergens by UPLC/MS/MS
CAPE
CABE
3M2B
2M2B
3M3B
CA
DHCA
Figure 4. Typical UPLC/MS/MS chromatogram (MRM mode) of a raw propolis extract
(sample 1). m/z 181→137 for IS (DHCA); 179→135 for CA; 247→135 for 3M2B; 247→134 for
3 M2 and 2M2B; 283→135 for CAPE; and 269→134 for CABE.
Table 3. Amountsa of caffeic acid and CAEs with allergenic action found in raw propolis samples from different geographic
areas (1–13), commercial hydroalcoholic extracts (A–F) and tablets (I–VI). The raw propolis samples were from Italy (1–3),
Macedonia (4–5), Poland (6), Uruguay (7), France (8), Nepal (9) and China (10–11). The samples 12 and 13 were red and
green Brazilian propolis, respectively
mg/g
Sample
CA
3M3B
3M2B
2M2B
CABE
CAPE
Total %
RSD %
1
5.50
6.10
2.40
7.20
5.30
2.80
1.70
2.90
0.12
7.63
2.10
N.F.
N.F.
0.09
0.79
0.13
0.14
0.17
0.08
6.70
8.10
5.40
8.40
8.10
2.20
2.80
5.50
0.32
9.30
2.40
N.F.
N.F.
0.10
0.97
0.14
0.15
0.21
0.09
7.60
9.20
7.90
9.80
9.10
3.70
2.80
6.30
0.44
10.54
2.70
N.F.
N.F.
0.12
1.10
0.17
0.17
0.24
0.12
2.40
2.30
1.10
3.40
2.10
1.20
1.10
2.60
N.F.
3.33
0.40
N.F.
N.F.
0.04
0.35
0.02
0.02
0.04
0.01
9.00
9.60
10.90
13.00
8.70
3.70
10.50
11.20
1.00
12.49
3.50
N.F.
N.F.
0.14
1.30
0.23
0.26
0.38
0.15
7.00
8.60
4.70
8.10
9.00
2.90
4.70
8.90
0.80
9.71
2.60
N.F.
N.F.
0.11
1.01
0.15
0.17
0.23
0.11
3.82
4.39
3.24
4.99
4.23
1.65
2.36
3.74
0.27
5.30
1.37
2.88
3.19
3.40
4.21
3.30
4.85
4.66
3.48
7.40
3.02
5.84
2
3
4
5
6
7
8
9
10
11
12
13
I
0.06
0.55
8.38
5.44
8.10
8.72
7.06
8.74
RSD %
5.08
7.84
4.98
6.01
5.26
7.87
II
III
IV
V
0.08
0.09
0.13
VI
Sample
A
0.06
mg/mL
Total %
0.39
0.57
0.41
1.45
0.66
0.54
0.11
0.69
0.41
1.61
0.82
0.62
0.17
0.78
0.52
2.15
0.89
0.73
0.18
0.25
0.11
0.29
0.11
0.12
N.F.
0.93
0.71
2.71
1.41
1.11
0.24
0.72
0.52
1.84
1.10
0.69
0.19
B
0.27
C
1.01
D
0.50
E
0.38
F
0.09
aValues are expressed as mean of four determinations. N.F. not found.
Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682 Copyright # 2011 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm

C. Gardana and P. Simonetti
and propolis as screening tools in the detection of fragrance
allergy. Br. J. Dermatol. 2001, 145, 268.
[14] S. Rajpara, M. Wilkinson, C. King, D. Gawkrodger,
J. S. English, B. N. Statham, C. Green, J. E. Sansom, M. M.
Chowdhury, H. L. Horne, A. D. Ormerod. The importance of
propolis in patch testing‐a multicentre survey. Contact
Dermatitis. 2009, 61, 287.
[15] U. Jappe, A. Schnuch, W. Uter. Rosacea and contact allergy to
cosmetics and topical medicaments‐retrospective analysis of
multicentre surveillance data 1995–2002. Contact Dermatitis.
2005, 52, 96.
analytical control of propolis should be based on determina-
tion of CAEs as a further marker, not only of quality, but also
of the safety of the derived products. In particular, the
3‐methyl‐2‐butenyl caffeate (3M2B) quantification should be
taken into account since it seems to be the most allergenic
compound and is available as a pure reference standard. We
strongly recommend that its amount should be made easily
visible on the external packaging so that consumers can choose
the product with the lowest content of allergenic compound.
[16] B. M. Hausen. Evaluation of the main contact allergens in
propolis (1995 to 2005). Dermatitis 2005, 16, 127.
Acknowledgements
[17] T. Hasan, T. Rantanen, K. Alanko, R. J. Harvima, R. Jolanki,
K. Kalimo, A. Lahti, K. Lammintausta, A. I. Lauerma,
A. Laukkanen, T. Luukkaala, R. Riekki, K. Turjanmaa,
E. Varjonen, A. M. Vuorela. Patch test reactions to cosmetic
allergens in 1995–1997 and 2000–2002 in Finland – a
multicentre study. Contact Dermatitis. 2005, 53, 40.
[18] F. Giusti, R. Maglietta, P. Pepe, S. Seidenari. Sensitization to
propolis in 1255 children undergoing patch testing. Contact
Dermatitis. 2004, 51, 255.
[19] M. C. Acciai, M. Ginanneschi, S. Bracci, A. Sertoli. Studies of
the sensitizing properties of propolis. Contact Dermatitis. 1990,
23, 274.
[20] L. Jiang, G. Fang, Y. Zhang, G. Cao, S. Wang. Analysis of
flavonoids in propolis and Ginkgo biloba by micellar
electrokinetic capillary chromatography. J. Agric. Food Chem.
2008, 56, 11571.
The authors are grateful to Specchiasol S.r.l (Bussolengo, VR,
Italy) for providing the raw propolis sample and express their
gratitude to Dr. Martina Scaglianti for synthesis of the caffeic
acid esters. The authors are grateful to Dr. Marco Pappini
(CIGA, Centro Interdipartimentale Grandi Attrezzature) for the
FTICR‐MS analysis.
REFERENCES
[1] V. S. Bankova, S. L. De Castro, M. C. Marcucci. Propolis: recent
advances in chemistry and plant origin. Apidologie 2000, 31, 3.
[2] C. Gardana, M. Scaglianti, P. G. Pietta, P. Simonetti.
Analysis of the polyphenolic fraction of propolis from
different sources by liquid chromatography‐tandem mass
spectrometry. J. Pharm. Biomed. Anal. 2007, 45, 390.
[3] M. C. Marcucci, F. Ferreres, A. R. Custodio, M. Ferreira,
V. S. Bankova, C. Garcia‐Viguera, W. A. Bretz. Evaluation of
phenolic compounds in Brazilian propolis from different
geographic regions. Z. Naturforsch. 2000, 55, 76.
[21] I. Márquez Hernández, O. Cuesta‐Rubio, M. Campo Fernández,
A. Rosado Pérez, R. Montes de Oca Porto, A. L. Piccinelli,
L. Rastrelli. Studies on the constituents of yellow Cuban
propolis: GC‐MS determination of triterpenoids and
flavonoids. J. Agric. Food Chem. 2010, 58, 4725.
[4] G. A. Burdock. Review of the biological properties and toxicity
of bee propolis (propolis). Food Chem. Toxicol. 1998, 36, 347.
[5] J. M. Sforcin, V. S. Bankova. Propolis: is there a potential for the
development of new drugs? J. Ethnopharmacol. 2011, 133, 253.
[6] M. M. Abdel‐Latif, H. J. Windle, B. S. Homasany, K. Sabra,
D. Kelleher. Caffeic acid phenethyl ester modulates Helico-
bacter pylori‐induced nuclear factor‐kappa B and activator
protein‐1 expression in gastric epithelial cells. Br. J. Pharm.
2005, 146, 1139.
[7] I. Jasprica, A. Mornar, Z. Debeljak, A. Smolcić‐Bubalo,
M. Medić‐Sarić, L. Mayer, Z. Romić, K. Bućan, T. Balog,
S. Sobocanec, V. Sverko. In vivo study of propolis supple-
mentation effects on antioxidative status and red blood cells.
J. Ethnopharmacol. 2007, 110, 548.
[8] O. Gulbahar, G. Ozturk, N. Erdem, A. C. Kazandi,
A. Kokuludag. In vivo study of propolis supplementation
effects on antioxidative status and red blood cells. Ann.
Allergy Asthma Immunol. 2005, 94, 509.
[9] S. G. Fernandez, E. A. Aleman, B. E. García Figueroa,
E. G. Fagoaga, J. M. O. Rivera, A. I. Purroy. Direct and
airborne contact dermatitis from propolis in beekeepers.
Contact Dermatitis. 2004, 50, 320.
[22] M. L. Castro, A. M. do Nascimento, M. Ikegaki, C. M.
Costa‐Neto, S. M. Alencar, P. L. Rosalen. Identification of a
bioactive compound isolated from Brazilian propolis type 6.
Bioorg. Med. Chem. 2009, 17, 5332.
[23] A. C. Sawaya, D. M. Tomazela, I. B. S. Cunha, V. S. Bankova,
M. C. Marcucci, A. R. Custodio, M. N. Eberlin. Electrospray
ionization mass spectrometry fingerprinting of propolis.
Analyst 2004, 129, 739.
[24] A. C. Sawaya, P. V. Abdelnur, M. N. Eberlin, S. Kumazawa,
M. R. Ahn, K. S. Bang, N. Nagaraja, V. S. Bankova,
H Afrouzan. Fingerprinting of propolis by easy ambient sonic‐
spray ionization mass spectrometry. Talanta 2010, 81, 100.
[25] A. Aliboni, A. D’andrea, P. Massanisso. Propolis specimens
from different locations of central Italy: chemical profiling
and gas chromatography‐mass spectrometry (GC‐MS)
quantitative analysis of the allergenic esters benzyl cinna-
mate and benzyl salicylate. J. Agric. Food Chem. 2011, 59, 282.
[26] C. Medana, F. Carbone, R. Aigotti, G. Appendino, C. Baiocchi.
Selective analysis of phenolic compounds in propolis by
HPLC‐MS/MS. Phytochem. Anal. 2008, 19, 32.
[27] P. Sestili, G. Diamantini, A. Bedini, L. Cerioni, I. Tommasini,
G. Tarzia, O. Cantoni. Plant‐derived phenolic compounds
prevent the DNA single‐strand breakage and cytotoxicity
induced by tert‐butylhydroperoxide via an iron‐chelating
mechanism. Biochem. J. 2002, 364, 121.
[10] E. Rudzki, P. Rebandel, E. Jaworski. Comparison of the
eliciting properties of 3 different propolis samples. Dermatitis
1998, 39, 142.
[11] B. M. Hausen, P. Evers, H. Stuwe, W. Konig, E. Wollenweber.
Propolis allergy (IV). Studies with further sensitizers from
propolis and constituents common to propolis, poplar buds
and balsam of Peru. Contact Dermatitis. 1992, 26, 34.
[12] J. Machackova. The incidence of allergy to propolis in
605 consecutive patients patch tested in Prague. Contact
Dermatitis. 1988, 18, 210.
[28] V. P. Shah, K. K. Midha, J. W. Findlay, H. M. Hill, J. D. Hulse,
I. J. McGilveray, G. McKay, K. J. Miller, R. N. Patnaik,
M. L. Powell, A. Tonelli, C. T. Viswanathan, A. Yacobi.
Bioanalytical method validation – a revisit with a decade of
progress. Pharm. Res. 2000, 17, 1551.
[29] B. K. Matuszewski, M. L. Constanzer, C. M. Chavez‐Eng.
Strategies for the assessment of matrix effect in quantitative
bioanalytical methods based on HPLC‐MS/MS. Anal. Chem.
2003, 75, 3019.
[13] S. Wohrl, W. Hemmer, M. Focke, M. Gotz, R. Jarisch. The
significance of fragrance mix, balsam of Peru, colophony
wileyonlinelibrary.com/journal/rcm Copyright # 2011 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2011, 25, 1675–1682