Anal. Chem. 2005, 77, 5961-5964
Absolute Method for the Assay of Oleuropein in Olive Oils by Atmospheric Pressure Chemical Ionization Tandem Mass Spectrometry Antonio De Nino,† Leonardo Di Donna,† Fabio Mazzotti,† Enzo Muzzalupo,‡ Enzo Perri,‡ Giovanni Sindona,†,* and Antonio Tagarelli†
Dipartimento di Chimica, Universita` della Calabria, via P. Bucci, cubo 12/C, I-87030 Arcavacata di Rende (CS), Italy, and CRA, Istituto Sperimentale per l’Olivicoltura, c.da Li Rocchi, I-87036 Arcavacata di Rende (CS), Italy
Oleuropein (OLP, 1), the active ingredient present (i) in food integrators extracted from olive leaves, (ii) in table olives, and (iii) in extra virgin olive oils is a nutraceutical whose health benefits have been widely documented. A new analytical method for its assay, which is based on the utilization of atmospheric pressure chemical ionization tandem mass spectrometry and on the use of a synthetic labeled analogue, the 4-trideuteriocarboxyoleuropein (2), as an internal standard, is presented. The results obtained with extra virgin olive oils from different cultivars and different Italian regions are discussed. Oleuropein (OLP, 1, Scheme 1), a secondary metabolite of terpenoid origin, is the main iridoid of Olea europaea and represents one of the 600 species of the Oleaceae family.1 This secoiridoid glucoside is responsible for the bitter taste of olive drupes and leaves and is associated with many healing effects in humans and animals.2-6 OLP is the most important active ingredient present in olive leaf extracts, available in the marketplace as a food integrator, and is considered the panacea of many important syndromes.7 Only recently has the presence of oleuropein in virgin olive oil been unambiguously ascertained by mass spectrometric methods.8 The same method was applied to detect its presence in olive leaves.9 The chemical properties of 1, such * To whom correspondence should be addressed. Phone: +39-0984-492046. Fax: +39-0984-493307. E-mail:
[email protected]. † Universita` della Calabria. ‡ CRA. (1) (a) Panizzi, L. M.; Scarpati, J. M.; Oriente, E. G. Gazz. Chim. Ital. 1960, 90, 1449-1485; (b) Jensen, S. R.; Franzyk, H.; Wallander, E. Phytochemistry 2002, 60, 213-231. (2) Manna, C.; Migliardi, V.; Golino, P.; Scognamiglio, A.; Galletti, P.; Chiariello, M.; Zappia, V. J. Nutr. Biochem. 2004, 15, 461-466. (3) Evangelista, C. M. W.; Antunes, L. M. G.; Francescano, H. D. C.; Bianchi, M. L. P. Food Chem. Toxicol. 2004, 42, 1291-1297. (4) Somova, L. I.; Shode, F. O.; Ramnanan, P.; Nadar, A. J. Ethopharmacol. 2003, 84, 299-305. (5) Markin, D.; Duek, L.; Berdicevsky, I. Mycoses 2003, 46, 132-136. (6) Visioli, F.; Caruso, D.; Galli, C.; Viappiani, S.; Galli, G.; Sala, A. Biochem. Biophys. Res. Commun. 2000, 278, 797-799. (7) http://www.curezone.com/foods/oliveleaf.asp. (8) Perri, E.; Raffaelli, A.; Sindona, G. J. Agric. Food Chem. 1999, 47, 41564160. (9) (a) De Nino, A.; Lombardo, A.; Perri, E.; Procopio, A.; Raffaelli, A.; Sindona, G. J. Mass Spectrom. 1997, 32, 533-541. (b) De Nino, A.; Mazzotti, F.; Morrone, S. P.; Perri, E.; Raffaelli, A.; Sindona G. J. Mass Spectrom. 1999, 34, 10-16. 10.1021/ac050545h CCC: $30.25 Published on Web 08/10/2005
© 2005 American Chemical Society
as its lack of solubility in lipophilic matrixes, pose important recovery problems when it has to be assayed in fatty matrixes, such as olive oil.8,10 Availability tests in biological fluids have been recently introduced,11 owing to the widespread distribution of food integrators based on oleuropein and to the many beneficial effects reported for this nutraceutical compound. Finally, the Mediterranean tradition of feeding goats and sheep olive leaves12 has been recently extended to cows. The active ingredient present in olive leaves13 can be, in fact, transferred without any damage, to the pellets used for animal nutrition.14 It should be considered, however, that the positive action of any phytochemical active ingredient might be counterbalanced by some negative side effects.15 Accordingly, tocopherols, typical phytochemicals known for their antioxidant property, can exhibit a pro-oxidant activity in vitro at given experimental conditions.16 The need for an absolute analytical method for the assay of oleuropein in different environments has therefore led to the utilization of a mass spectrometric MS/MS approach, based on the use of the labeled internal standard 2, aimed at avoiding any chromatographic separation in the assaying of the analyte from different matrixes. A case study is represented by the quantitative determination of 1 in experimental virgin Italian olive oils in the framework of the national “VATIPICI” project. EXPERIMENTAL SECTION Chemicals. Solvents and reagents were obtained commercially (Sigma-Aldrich, St. Louis, MO). Demethyloleuropein was extracted from ripened drupes of the Leccino cultivar by a slightly modified literature procedure.17 A mixture of 5 kg of crushed (10) (a) Montedoro, G.; Servili, M.; Baldioli, M.; Miniati, E. J. Agric. Food Chem. 1992, 40, 1571-1576. (b) Coni, E.; Di Benedetto, R.; Di Pasquale, M.; Masella, R.; Modesti, D.; Mattei, R.; Carlini, E. A. Lipids 2000, 35, 45-54. (11) (a) Del Boccio, P.; Di Deo, A.; De Curtis, A.; Celli, N.; Iacoviello, L.; Rotilio, D. J. Chromatogr., B 2003, 785, 47-56. (b) Visioli, F.; Galli, C.; Galli, G.; Caruso, D. Eur. J. Lipid Sci. Technol. 2002, 104, 677-684. (12) Delgado-Pertinez, M.; Chesson, A.; Provan, G. J.; Garrido, A.; GomezCabrera, A. Ann. Zootech. 1998, 47, 141-150. (13) Aruoma, O. I.; Deiana, M.; Jenner, A.; Halliwell, B.; Kaur, H.; Banni, S.; Corongiu, F. P.; Dessi, M. A.; Aeschbach, R. J. Agric. Food Chem. 1998, 46, 6, 5181-5187. (14) Regione Calabria project POR Mis. 3.16-2004. (15) Zhou, S. F.; Koh, H. L.; Gao, Y. H.; Gong, Z. Y.; Lee, E. J. D. Life Sci. 2004, 74, 935-968. (16) Kontush, A.; Finckh, B.; Karten, B.; Kohlschutter, A.; Beisiegel, U. J. Lipid Res. 1996, 37, 1436-1448. (17) Ragazzi, E.; Veronese, G.; Guiotto, A. Ann. Chim. 1973, 63, 13-27.
Analytical Chemistry, Vol. 77, No. 18, September 15, 2005 5961
Scheme 1
Table 1. 500 1H NMR of Wild-type (1) and Deuterated (2) Oleuropein
proton 1 3 5 6a 6b 8 COOMe 10 1′ 2′ 4′ 7′ 8′
OLP multi[R1dCH3] plicity 5.86 7.51 3.86 2.40 2.40 5.95 3.65 1.65 4.08 2.68 6.48 6.60 6.64
s s dd dd dd q s d m t dd d d
J, Hz
9.2; 4.09 14.64; 9.2 14.64; 9.2 6.82 7.85 7.49 7.82; 2.04 1.7 8.18
OLP-d3 multi[R1dCD3] plicity 5.86 7.51 3.86 2.40 2.40 5.95 1.65 4.08 2.68 6.48 6.60 6.64
s s dd dd dd q s d m t dd d d
J, Hz
9.2; 4.43 14.31; 9.2 14.31; 9.2 6.82 8.2 7.16 8.17; 2.04 2.04 8.17
drupes with 1.5 L of CH3OH was filtered after standing for 60 min at room temperature. The solvent was removed under reduced pressure, and the residue was dissolved in water (0.5 L) and washed with n-hexane (3 × 250 mL). The water solution was then extracted with ethyl acetate (5 × 250 mL), and the organic layers were dried over Na2SO4 and evaporated to dryness. The residue was purified by FPLC using a Pharmacia LKB (Uppsala, Sweden) apparatus equipped with a C18 (4.6 × 35 cm) reversedphase column, a peristaltic pump, and a fraction collector. The analyte was eluted with a linear water/methanol gradient, starting from 0% methanol, up to 30% methanol in 2 h, at a flow rate of 2 mL/min. The collected fractions were monitored by HPLC. Synthesis of Oleuropein-d3 (OLP-d3). The labeled internal standard was prepared by conventional chemistry18 using an ethereal solution of diazometane-d2. Oleuropein-d3 was obtained at quantitative yield and purified by HPLC. 1H NMR data are listed in Table 1. The elemental composition of 2, obtained by highresolution electrospray mass spectrometry, is [M + H]+, calculated 544.2109, measured 544.2097; isotopic distribution: d2 ) 3.12%, d3 ) 96.88%. (18) De Nino, A.; Di Donna, L.; Mazzotti, F.; Perri, E.; Raffaelli, A.; Sindona, G.; Urso, E. Biologically active Phytochemicals in Food; Pfannhauser, W., Fenwick, G. R., Khokhar, S., Eds.; Royal Society of Chemistry: Cambridge, UK, 2001; pp 131-133.
5962
Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
Caution: diazometane-d2 is harmful and explosive and should be handled with all safety precaution. Instrumentation. All experiments were carried out using an MDS Sciex API 2000 triple quadrupole mass spectrometer equipped with an APCI source (Applied Biosystems, Foster City, CA). The mass spectrometer was interfaced with an Agilent Technologies 1100 HPLC system (Waldbronn, Germany). All data were acquired and analyzed using Analyst software, version 1.2. Aqueous parts-per-million solutions of the analyte were delivered to the heated nebulizer by flow injection analysis (FIA); the flow rate was 400 µL min-1 of CH3CN/5 mM ammonium acetate aqueous solution (50/50). The experiments were performed at a source temperature (TEM) of 450 °C and at curtain gas (CUR) and source gas (GS1, GS2) pressures of 45, 70, and 50 psi, respectively, while the nebulizer current (NC), the declustering potential (DP), the focusing potential (FP), and the entrance potential (EP) were set to 4, 70, 250, and 5 V, respectively. The collision energy (CE) value was 40 eV, and the collision gas pressure (CAD) was set to 2 (9.38 × 1014 molecules cm-2 gas thickness). The MRM experiments were performed using a dwell time of 250 ms. The spectra were acquired at unit resolution. The HPLC apparatus used to monitor the collected fractions of demethyloleuropein was a Hewlett-Packard 1090 (Waldbronn, Germany) system equipped with a Rheodyne injector and fitted with a 20-µL loop and a UV detector set at 280 nm. A 250 × 4.6 mm, 5-µm, reversed-phase Alltima C18 column (Alltech, USA), heated at 30 °C, was eluted with a linear gradient of watermethanol for a run time of 30 min, starting from 10% methanol, at a flow rate of 1 mL/min. The 1H NMR spectra were recorded from an Avance-500 spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany) in DMSO-d6; chemical shifts (δ) were measured in parts per million and coupling constants (J), in Hertz. High-resolution electrospray ionization (ESI) experiments were carried out in a hybrid API Q-Star Pulsar-i (MSD Sciex Applied Biosystem, Toronto, Canada) mass spectrometer equipped with an ion spray ionization source at 7000 fwhm resolution. Samples were introduced by direct infusion (5 µL/min) of a solution containing the analyte (10 pmol/µL, dissolved in a solution of 0.1% ammonium acetate, acetonitrile/water 50/50) at the optimum ion spray voltage (IS) of 4800 V. The source nitrogen (GS1) and the curtain gas flow were set at 20 and 25 psi respectively, while the first declustering potential (DP1), the focusing potential (FP), and the second declustering potential (DP2) were kept at 50, 220, and 10 V relative to ground, respectively. Sample Preparation. A 20-µL portion of a stock solution (5 ppm) of internal standard was added to 1 g of olive oil and
RESULT AND DISCUSSION Tandem mass spectrometry represents the method of choice in the analysis and assay of microcomponents present in food matrixes. The accuracy and precision of the assay is enhanced when an absolute method, based on the availability of a labeled internal standard with the same chemical properties, is used. The APCI-MS/MS spectrum of the [M + NH4]+ species of OLP (1, Scheme 1) at m/z 558 is characterized by the presence of few
specific fragments, one of which is represented by the species at m/z 137, due to the fragmentation of the catecholic moiety of the molecules (Scheme 1).9 Whatever the mechanism of ion formation, it was reasonable to consider that the methyl group (Scheme 1) of the ester moiety at position 4 of OLP (1) should not be involved in the formation of the fragment at m/z 137. The trideuteromethyl oleuropein (OLPd3, 2, Scheme 1) was therefore selected as a suitable internal standard for the quantitative analysis of oleuropein in various matrixes. A low-cost synthesis of 2 was devised starting from the free acid form of oleuropein, conventionally indicated as demethyloleuropein, which is particularly abundant in olive drupes of some particular cultivars, such as Leccino; the extraction of this active ingredient and its chemical transformation into 2 was performed in a straightforward and affordable way.17 Both 500MHz, 1H NMR (Table 1) and high-resolution mass spectrometry have provide evidence on the extent of labeling of 2 (see Experimental Section). As expected, the APCI-MS/MS spectra of the [M + NH4]+ ions of compounds 1 and 2 displayed a common fragment at m/z 137, which was formed independently of the label position (Figure 1). The transitions m/z 558 f m/z 137 for the analyte and m/z 561 f m/z 137 for the labeled internal standard, respectively, were selected for the quantitative assay by tandem mass spectrometry. Oleuropein, a molecule soluble in polar solvents, is suspended or emulsified in the residual water traces present in virgin olive oil; therefore, its recovery is quite low and unreproducible.8 The assay of 1 by an “absolute” method of analysis could confer to virgin olive oil the right to be considered a functional food.20 A spin-off effect for the marketing of this peculiar foodstuff could be represented, in fact, by the indication of the content of pharmacologically active principles, determined by a high-tech approach. The presence of oleuropein could also be related to food processing, thus, to the quality and safety of the total olive oil chain. We have therefore validated the method of analysis here proposed by applying APCI-tandem mass spectrometry, in the MRM mode, to samples of refined olive oil spiked with both 1 and 2 (see Experimental Section). The method was then applied to a survey of virgin olive oils produced from different cultivars grown in different Italian regions as a function of different agronomical and processing parameters within a joint national project (VATIPICI) of the Italian Ministries of Research (MIUR) and Agriculture (MIPAF). Table 3 outlines the dependence of the amount of oleuropein in virgin olive oils from two different cultivars, Coratina and Carolea, grown in different places with substantially different climate, whose drupes were harvested at similar ripening stages. Analytical data related to the food processing have been achieved, as well. The data obtained shows that the content of OLP strictly depends on the above-mentioned variables. Table 3 shows, in fact, that there is a general tendency in the enhancement of OLP content with the latitude of the harvested drupes, independent of the observed cultivar. Values in the range of 0.357 ppm for the
(19) Ruiz-Gutierrez, V.; Perez-Camino, M. C. J. Chromatogr., A 2000, 885, 321341.
(20) Diplock, A. T.; Aggett, P. J.; Ashwell, M.; Bornet, F.; Fern, E. B.; Roberfroid, M. B. Brit. J. Nutr. 1999, 81, S1-S27.
Table 2. Analytical Data for the Calibration Curvea std solns (ppm) 0.050 0.100 0.200 0.400 0.800
area ratio 0.562 0.590 0.616 1.080 1.041 1.013 1.982 1.985 2.002 4.212 4.297 4.261 8.313 8.260 8.212
av area ratio
SD area ratio
RSD % (av)
0.5893
0.0270
4.58
1.0447
0.0337
3.22
1.9897
0.0108
0.54
4.2567
0.0427
1.00
8.2617
0.0505
0.61
a A fixed amount of internal standard (0.100 ppm) was added to each sample of the calibration curve (y ) 1.0322x - 0.0285, R2 ) 0.9993).
homogenized; the mixture was dissolved in 5 mL of n-hexane and loaded into the SPE cartridge (C18 1 g, 6 mL, J. T. Baker, Phillipsburg, NJ). The cartridge was washed with n-hexane (3 × 5 mL), and then CH3OH was added (2 × 5 mL) to elute the analyte.19 The collected methanolic fraction was then evaporated under reduced pressure, and the residue was dissolved in 1 mL of CH3OH and injected into the mass spectrometer. The assay by mass spectrometry was performed through a fitting line (r2 ) 0.9993) built from five solutions containing different concentrations of labeled and unlabeled oleuropein. Each sample was injected three times, and the average value with the standard deviation was used to calculate the RSD% (Table 2). Analytical Parameters. The limit of detection (LOD) and the limit of quantitation (LOQ) for the olive oil were calculated following the directives of IUPAC and the American Chemical Society’s Committee on Environmental Analytical Chemistry, that is, as follows:
SLOD ) SRB + 3σRB SLOQ ) SRB + 10σRB
where SLOD is the signal at the limit of detection, SLOQ is the signal at the limit of quantitation, SRB is the signal of the reagent blank, and σRB is the standard deviation for the reagent blank. The concentrations were calculated by the standard curve.
Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
5963
Figure 1. APCI-MS/MS spectrum of (A) oleuropein and (B) d3-labeled oleuropein. Table 3. Analytical Data for Filtered and Not Filtered Virgin Olive Oils cultivar
Italian region (latitude)
Carolea
Calabria (38° 57′ N)
Carolea
Abruzzo (42° 28′ N)
Carolea
Apulia (40° 37′ N)
Coratina
Calabria (38° 57′ N)
Coratina
Abruzzo (42° 28′ N)
Coratina
Apulia (40° 37′ N)
Frantoio
Calabria (38° 57′ N)
Frantoio
Abruzzo (42° 28′ N)
treatment
ppm value
RSD %
filtered not-filtered filtered not-filtered filtered not-filtered filtered not-filtered filtered not-filtered filtered not-filtered filtered not-filtered filtered not-filtered
0.357 ( 0.014 0.263 ( 0.004 0.245 ( 0.014 0.130 ( 0.002 0.296 ( 0.002 0.225 ( 0.004 0.175 ( 0.002 0.132 ( 0.005 0.116 ( 0.001 0.093 ( 0.003 0.222 ( 0.002 0.118 ( 0.001 0.344 ( 0.008 0.214 ( 0.002 0.203 ( 0.001 0.133 ( 0.001
4.04 1.38 5.64 1.47 0.68 1.62 1.10 4.06 0.81 3.48 1.06 0.80 2.26 0.94 0.71 0.82
southern Carolea, to 0.093 for the northern Coratina have been detected. The repeatability of the measurements is supported by the RSD % values that are in all cases under 5.70%. Moreover, a definite difference in OLP content was found between filtered and unfiltered oils. In general, the filtered oils preserve the natural antioxidant, and the nonfiltered ones show a lower amount of oleuropein. The relative amount of 1 in filtered and unfiltered oils could be related to the presence of water, assuming that the lower the amount of residual water in oil, the less the availability of hydrolytic enzymes, such as β-glucosidase, which metabolize the active ingredient. The accuracy value, close to 100% in both fortified samples; the recovery test, higher than 80%; the detection and quantitation limits, in the range of 25 ÷ 30 ppb (Table 4); and the reproducibility (RSD%), always lower than 7%, checked on different days from independent extraction of the same analyte, highlight the uniqueness of the proposed method (Table 4). 5964 Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
Table 4. Accuracy, Recovery, LOQ, and LOD Measurements; Reproducibility of the Method fortified sample (ppm) 0.700 0.150
ppm value
RSD accuracy LOD LOQ recoverya % % (ppm) (ppm) %
0.677 ( 0.025 3.71 0.155 ( 0.008 4.92
96.77 103.40
0.027
0.029
85
cultivar
Italian region
treatment
ppm value
RSD %
Carolea Coratina Frantoio
Calabria Abruzzo Abruzzo
filtered filtered filtered
0.361 ( 0.022 0.115 ( 0.008 0.204 ( 0.011
6.17 6.57 5.41
a
The recovery was estimated by using an external calibration curve.
CONCLUSIONS The increasing intake of food additives containing oleuropein and the emerging role of the so-called Mediterranean diet rich in antioxidants, such 1, has prompted the evaluation of an analytical procedure suitable for the assay of this pharmacologically active ingredient, even in difficult matrixes, such as olive oil. The examined experimental olive oils are the best source of information for the evaluation of any specific analyte present in the foodstuff, since they are properly produced and stored. The data thus obtained allow, for the first time, a proper evaluation of the nutraceutical properties of virgin olive oil. Oleuropein is present at the parts-per-billion level in the foodstuff, provided that best manufacturing practices are strictly followed in all the steps of the olive oil production. ACKNOWLEDGMENT This research has been carried out within the Italian national project VATIPICI funded by MiPAF and MIUR.
Received for review March 31, 2005. Accepted July 8, 2005. AC050545H
