Rapid discrimination of bergamot essential oil ... - Wiley Online Library

Journal of

MASS SPECTROMETRY

Research article Received: 18 January 2016

Revised: 8 July 2016

Accepted: 19 July 2016

Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jms.3820

Rapid discrimination of bergamot essential oil by paper spray mass spectrometry and chemometric analysis Domenico Taverna,* Leonardo Di Donna, Fabio Mazzotti, Antonio Tagarelli,* Anna Napoli, Emilia Furia and Giovanni Sindona A novel approach for the rapid discrimination of bergamot essential oil from other citrus fruits oils is presented. The method was developed using paper spray mass spectrometry (PS-MS) allowing for a rapid molecular profiling coupled with a statistic tool for a precise and reliable discrimination between the bergamot complex matrix and other similar matrices, commonly used for its reconstitution. Ambient mass spectrometry possesses the ability to record mass spectra of ordinary samples, in their native environment, without sample preparation or pre-separation by creating ions outside the instrument. The present study reports a PS-MS method for the determination of oxygen heterocyclic compounds such as furocoumarins, psoralens and flavonoids present in the non-volatile fraction of citrus fruits essential oils followed by chemometric analysis. The volatile fraction of Bergamot is one of the most known and fashionable natural products, which found applications in flavoring industry as ingredient in beverages and flavored foodstuff. The development of the presented method employed bergamot, sweet orange, orange, cedar, grapefruit and mandarin essential oils. PS-MS measurements were carried out in full scan mode for a total run time of 2 min. The capability of PS-MS profiling to act as marker for the classification of bergamot essential oils was evaluated by using multivariate statistical analysis. Two pattern recognition techniques, linear discriminant analysis and soft independent modeling of class analogy, were applied to MS data. The cross-validation procedure has shown excellent results in terms of the prediction ability because both models have correctly classified all samples for each category. Copyright © 2016 John Wiley & Sons, Ltd. Additional supporting information may be found in the online version of this article at the publisher’s web site. Keywords: ambient mass spectrometry; paper spray; bergamot essential oil; chemometric analysis

Introduction

J. Mass Spectrom. 2016, 51, 761–767

* Correspondence to: Domenico Taverna, Department of Experimental and Clinical Medicine ‘Magna Græcia’ University of Catanzaro, Campus ‘S. Venuta’, Viale Europa, Loc. Germaneto, 88100 Catanzaro, Italy. E-mail: [email protected] Antonio Tagarelli, Department of Chemistry and Chemical Technologies—CTC, University of Calabria, via P. Bucci, cubo 12/C, Arcavacata di Rende (CS), 87036, Italy. E-mail: [email protected] University of Calabria, Department of Chemistry and Chemical Technologies— CTC, Arcavacata di Rende, Cosenza 87036, Italy

Copyright © 2016 John Wiley & Sons, Ltd.

761

Bergamot essential oil is a valuable product having multiple uses. Most of the bergamot (citrus bergamia) production is located in Italy, in the Calabria region, as well as most of the essential oil really extracted from the fruit. Other fruit parts such as pulp or juice are still considered either as waste materials or used for minor commercial aims: fruit juices or confectionery products such as cake and others.[1] Within the last decade, several products based on bergamot waste material extracts have been introduced to the market as food supplements, because of the presence of statin-like molecules.[2] Besides that, bergamot essential oil is largely used in cosmetics, representing one of the main ingredients of perfumery products. Furthermore, bergamot essential oil has been proven to exert important antioxidant, antimicrobial biological activities.[3,4] Recently, bergamot as well as other citrus fruits essential oils has found increasing appreciation within flavoring industry as an ingredient in beverages and flavored foodstuff for its aromatic properties. The essential oil taste allows for the intake of salt, fat and sugar reduction, which are usually added to foods to increase their flavor, thus creating a great alternative to artificial flavoring processes. It has become, therefore, of primary importance the quality and safety of the product. Bergamot essential oil can, in fact, be easily found in a wide range of variability in terms of price and, of course, of quality. The commercial value of essential oils is also subject to sensible shifts as

happens for every natural product strictly related to crop production. To meet the market demand, essential oils are often reconstituted, starting from cheaper oils, lowering the price and compromising their quality as well.[5] The production processes,[6,7] of citrus essential oils can significantly affect the relative amounts of non-volatiles, mainly oxygen heterocyclic compounds such as psoralens, coumarins and flavonoids as minor constituents.[8–10] These are secondary metabolites usually present within citrus fruits and highlighting a wide variety of beneficial effects on human health.[11,12] Because the presence of most of these compounds is characteristic for the citrus species, their determination is usually performed for quality control. Different analytical techniques are currently used for the determination of the molecular composition of bergamot essential oil as well as for its qualitative and quantitative characterization. The

Journal of

MASS SPECTROMETRY

D. Taverna et al.

development of analytical procedures often employs the use of mass spectrometry coupled with chromatography and diode-array detectors for the non-volatiles species as well as the use of gas chromatography-MS methods.[13] Moreover, further analysis is usually needed to assess the capability of essential oils, such as antimicrobial activity measurements by determining the inhibition of bacterial growth or bioactivity assays.[14,15] The limited availability of real bergamot essential oil has prompted its illicit adulteration with other natural or synthetic products. Thus, a methodology allowing bergamot essential oil to be easily, rapidly and accurately discriminated from the other citrus oils is needed. A novel approach for a rapid molecular profiling, obtained bypassing chromatography, in citrus fruits essential oils was therefore developed. This spectrometric methodology has been coupled with chemometric analysis for a reliable and molecularly based comprehensive discrimination of essential oils. Oxygen heterocyclic compounds such as psoralens, coumarins and some flavonoids were determined using paper spray mass spectrometry (PS-MS), an outstanding implementation of ambient mass spectrometry techniques which has already found applications in food chemistry.[16–18] In the last decades, a family of MS sources has been developed in order to produce ions under ambient conditions.[19,20] Ambient mass spectrometry possesses the ability to record mass spectra of solid and liquid samples, in their native environment, without sample pre-treatment by creating ions outside the instrument at atmospheric pressure and room temperature. Qualitative and quantitative analysis performed using ambient MS experiments typically requires few seconds. The present study reports a PS-MS method for the estimation of the characteristic molecules of the non-volatile fraction within some of the most commonly used citrus fruits essential oils. The ambient MS method was coupled with two multivariate chemometric approaches, allowing for a rapid and reliable determination of fraud within bergamot essential oil based on the characteristic molecular signatures. The development of the presented method employed six citrus fruits essential oils: sweet orange, bergamot, sour orange, mandarin, cedar and grapefruit. Furthermore, essential oil mixtures were prepared and MS analyzed with the aim to simulate the case of a bergamot essential oil adulteration. Two supervised pattern recognition procedures, linear discriminant analysis (LDA) and soft independent modeling of class analogy (SIMCA), were applied to MS data, and the discrimination and classification ability of these multivariate techniques was evaluated by the cross-validation approach.

Experimental

762

Samples: Bergamot essential oils were purchased from ‘Consorzio del bergamotto’ of Reggio Calabria, Italy. Citrus fruits essential oils (mandarin, sour orange, grapefruit, sweet orange, cedar) were purchased from a local store, 100% pure and natural. In particular, three samples of sour orange, three of sweet orange, three of cedar, four of mandarin, three of grapefruit and 20 of bergamot essential oils were collected for a total of N = 37 raw samples. Oils were stored at 4 °C in brown bottles. Four mixtures for each citrus fruit essential oil (sour orange, sweet orange, cedar, mandarin and grapefruit) were prepared with 70 parts by weight of bergamot essential oil and 30 parts by weight of the other essential oil for a total of 20 mixture samples. Each mixture was prepared by randomly selecting essential oils from the available samples. Mass spectrometry and ambient MS source: essential oils were analyzed using a triple-

wileyonlinelibrary.com/journal/jms

quadrupole mass spectrometer MS 320 from Varian (Varian Inc., Palo Alto, Ca), originally equipped with an ESI source. The ESI source was removed and in house implemented with a high voltage clip and a metallic stage for paper spray capability. Whatman paper (Whatman® qualitative filter paper, Grade 1 circles, diam. 110 mm.) triangles were cut at the following dimensions: 150-mm base, 250-mm height and then used as solid support for sample manual spotting allowing the essential oils to be directly analyzed bypassing sample pre-purification/preparation as well as the chromatography usually employed for sample purification and analyte separation before mass spectrometry. Sample handling: essential oils were first diluted by ethanol (100 mg, 3-ml ethanol), added of a solution of coumarin (1 ppm) used as internal standard and then spotted (15 μl) onto a paper triangle. The latter was allowed to dry for 1.5 min then placed on the high voltage clip in front of the spectrometer. PS-MS analysis procedure: measurements were carried out in full scan mode (mass range 100–500), using positive polarity because of a needle voltage of 4000 volts. The total run time was 2 min: 15 μl of methanol was manually spotted each 15 s onto the paper triangle embedded with the citrus peel extracts. PS-MS analyses were replicated 3 times. Data processing: total ion current of each analysis was averaged, and the relative intensities of each ion were normalized with the relative intensity of the internal standard ion. The peak detection parameters were set as follow: initial S/N ratio 5; initial peak width 4 sec; initial tangent height 10%. Statistical analysis: classification was carried out by two multivariate chemometric techniques: LDA and SIMCA. These procedures were performed by means of V-Parvus 2009. [21] Before LDA construction, principal component analysis (PCA) was performed by Parvus software. Spectra recorded in centroid mode were manually aligned, and the lists containing all m/z values and relative abundances of all ions in spectra were exported into a Microsoft Excel file. Then, the m/z values of interest were selected and uploaded into the Parvus software package.

Results and discussion PS-MS has been developed as a direct, fast and low-cost sampling and ionization method for mass spectrometric analysis of complex mixtures. The method is now applied for the first time to focus on a rapid and reliable molecularly based discrimination of the identification of true bergamot essential oil within citrus essential oils. The development of the presented method employed six citrus fruits essential oils: sour orange, bergamot, sweet orange, mandarin, cedar and grapefruit. All the samples were analyzed by PS-MS using the equipment displayed in Fig. 1. In particular, the instrument used for mass spectrometric analyses was in house implemented to allow the essential oil raw material ionization within ‘ambient’ conditions. The adopted procedure, as previously described, was based on the manual deposition of the essential oil matrix onto the paper center, and the dried triangle was installed onto the clip placed in front of the spectrometer. PS-MS measurements were carried out in full scan mode, using positive polarity allowing the ionization of molecules such as psoralens, furocoumarins and some flavones typically present within the non-volatile fraction of such essential oils. The total run time was 2 min, using a few microliters of solvent. Thus, the presented approach appears not only as fast compared with other analytical methods reported in literature but also ‘green’ because the use of solvents for both sample preparation and sample purification (e.g. by chromatography) is almost avoided.


J. Mass Spectrom. 2016, 51, 761–767

Journal of

MASS SPECTROMETRY

PS-MS and chemometrics of bergamot essential oil

Figure 1. a) Scheme of ambient MS ionization by paper spray reaction and b) picture of a in house implemented open source mass spectrometer adapted for PS-MS analysis.

J. Mass Spectrom. 2016, 51, 761–767

(Supplemental Fig. 2). Tandem mass spectrometry experiments were also carried out onto bergamot essential oil collected fractions as well as onto the fractions of the other citrus fruits essential oils involved in this study (Supplemental Fig. 3). So, the data matched PS-MS findings, confirming the reliability and the sensitivity of the PS-MS approach. Statistical analysis. Although several molecules were uniquely identified within a specific sample, it is important to find a systematic, reproducible and reliable way to differentiate them based on their chemical composition. This goal can be achieved by using multivariate statistical analysis. In particular, pattern recognition analysis was carried out by means of two supervised chemometric techniques (LDA and SIMCA) in order to obtain classification rules for distinguishing bergamot essential oil from other similar and cheaper essential oils commonly used for its reconstitution. The two chemometric techniques were applied to the data matrix of the relative abundances of all ions acquired in the mass range m/z 100–500, constituted by 37 rows (20 bergamot essential oils and 17 other essential oils) and 401 columns (variables). LDA and SIMCA were performed using two classes as input a priori: bergamot (bergamot essential oil samples) and non-bergamot (all the other essential oil samples). The reliability of the classification needed to be validated, and, therefore, a cross-validation procedure was carried out. The sample set was randomly divided into a training set and a validation set, the latter containing 1/k of the samples (k is called the cancelation group). Such a division allows for a sufficient number of samples in the training set, and a representative number of members in the validation set considered as unknown. The process was reiterated k times with different random constitutions of both sets to ensure that all of the samples were included in the validation set at least once. The goodness of the classification for both LDA and SIMCA was evaluated in terms of prediction ability, which is equal to the percentage of the validation set members correctly classified. LDA is certainly one of the most widely used classification techniques.[22–25] The data in the training set is used to defines n 1 delimiters (where n is the number of classes) so as the multivariate space of the objects is divided in as many subspaces as the number of categories. Discriminant functions (canonical roots) are obtained as a linear combination of descriptor that maximizes the ratio of variance between categories to variance within classes. An important feature of LDA is that more robust model can be constructed by using a total number of samples equal at least three times the number of variables. Indeed, discriminant analysis using a low ratio of samples to variables generates an unstable model, and



763

Furocumarins as well as some flavonoids and psoralens were profiled directly from the raw essential oils for all the citrus samples (Fig. 2). PS-MS molecular profiling within 100–500 mass range produced spectra highlighting similarities and differences among citrus fruits essential oil samples, mainly because of the unique presence of some coumarins and flavonoids; this type of spectrometric measurement allows for a rapid and reliable discrimination of bergamot essential oil from all the other citrus essential oils. Bergamot essential oil displays an almost unique molecular profile in the investigated mass range even if some ions were also recorded within the sweet orange essential oil spectra, for example the ion at m/z 339 corresponding to bergamottin (Supplemental Table 1). Other essential oils, such as orange and mandarin, highlight instead a lot of similarities; for instance, both orange and mandarin profiled spectra have recorded two relatively abundant ions at m/z 395 and 425, respectively, assigned as sodium adducts of two flavonoids tangeretin and nobiletin (Supplemental Table 1). Further, also cedar and grapefruit essential oils, once analyzed via ambient mass spectrometry, showed many similarities within the considered mass range (e.g. the ion at m/z 301 which is present in both samples). Thus, the complexity of the essential oils raw material as well as the belonging to the same biological family, for example citrus, can both make it difficult to distinguish an essential oil from another. Furthermore, PS-MS/MS experiments were used for identification. Furocoumarins such as 7-hydroxycoumarin (m/z 163), 6,7-dimethoxycoumarin (m/z 207) bergapten (m/z 217) and bergamottin (m/z 339) were identified within bergamot; others such as dihydrocoumarin (m/z 149) also in sweet orange and cedar. Furthermore, flavonoids such as tangeretin (m/z 373) and nobiletin (m/z 403) were found within orange and mandarin samples. Figure 3 displays a representative MS/MS spectrum of the m/z ion 339 and the related structure allowing for molecular identification of bergamottin (5-gernoxypsoralen) within bergamot essential oil. However, the MS/MS spectrum displays also the fragment at m/z 253.6 which is likely because of the presence of an isobaric species. The raw sample was not treated, so molecules were directly fragmented after paper spray ionization of the 15-μl essential oil deposited onto the paper triangle. It is noteworthy that paper spray technique has the capability to allow for both a rapid molecular screening and the possibility for a ‘on line’ identification using the tandem mass spectrometry from a droplet of raw material. The same samples were submitted to liquid chromatographyultraviolet (LC-UV) analysis (Supplemental Fig. 1), and collected fractions were then analyzed by mass spectrometry by flow injection

Journal of

MASS SPECTROMETRY

D. Taverna et al.

Figure 2. PS-MS molecular profiling within 100–500 mass range; spectra highlight similarities and differences among citrus fruits essential oil samples, in particular for furocumarines and some flavonoids allowing for a rapid (2 min) discrimination of bergamot essential oil.

764

analysis is also weakened by the presence of highly correlated variables (redundant information), which yields a less reliable model. Accordingly, in the present study, PCA has been initially applied to reduce the original dataset into a new highly representative subspace, which attempts to maintain most of the variability of the data. Afterwards, LDA was performed by considering from the first three to the first twelve (62–90% of the total variance) principal


components (PCs) as variables. The best mean percentage of prediction ability in the cross-validation procedure was achieved by taking into account the first four PCs (68.8% of the total variance). In this case, all the samples of bergamot essential oils were correctly classified, and only one sample belonging to non-bergamot category was classified as bergamot oil (mean prediction ability 97.3%, Table 1).


J. Mass Spectrom. 2016, 51, 761–767

Journal of

MASS SPECTROMETRY

PS-MS and chemometrics of bergamot essential oil

Table 2. Variables selected by stepwise linear discriminant analysis (descending order) Variable 114 366 301 477 402 270 146 140 236 389

Figure 3. MS/MS spectrum of the m/z ion 339 and the related structure allowing for molecular identification of bergamottin (5-gernoxypsoralen) within bergamot essential oil. The raw sample was not treated, so molecules were directly fragmented after paper spray ionization of the 15μl essential oil deposited onto the paper triangle.

Another approach to reduce the dimensionality of data is stepwise LDA. In this technique, the most significant variables involved in sample differentiation were selected using a Wilks’ λ as a selection criterion. An F statistic was used to determine the significance of the changes in Wilks’ λ when the influence of each new variable was evaluated. Accordingly, selection of the variables was performed by stepwise LDA to discard redundant information and to select only those variables that actually contributed to increase of classification ability. The stepwise analysis has retained 10 variables shown in Table 2. The latter 10 variables were not identified nor assigned by tandem mass spectrometry, but they can be ascribed to the essential oil molecular composition because these ions were not detected within the blank paper experiment, carried out using just solvent. Again, the cross-validation procedure was performed in order to verify the goodness of the method in terms of prediction ability. The proposed model has correctly predicted all samples for Table 1. Number of correctly classified samples (prediction matrices) for LDA performed on PCs, stepwise LDA and SIMCA techniques (rows represent the true class, columns report the assigned class) LDA performed on principal components

Bergamot Non-bergamot Mean prediction ability 97.3% Stepwise LDA Bergamot Non-bergamot Mean prediction ability 100% SIMCA

Non-bergamot 0 16

Bergamot 20 0

Non-bergamot 0 17

Bergamot 20 0

Non-bergamot 0 17

J. Mass Spectrom. 2016, 51, 761–767

SENS (%) 100 100

SPEC (%) 100 100

0.2804 0.086756 0.052133 0.039576 0.02923 0.024421 0.018307 0.013417 0.009584 0.00761

each category, and this suggests that the presented method may be a potential choice for discriminating bergamot essential oils from other similar ones (Table 1). LDA belongs to ‘hard’ classification techniques. This means that at least two classes have to be defined, and the discriminant delimiter is determined using the contribution by all of the categories specified in the problem. This entails the correct definition of the non-compliant category (in this study non-bergamot category) by a thorough, representative sampling because this class has a critical influence on the decision rules. Unfortunately, such a requirement is not always realizable, and, therefore, if the non-compliant category is not properly defined, a more appropriate chemometric pattern recognition strategy is required. The class modeling techniques analyze each category separately and permit the construction of a multivariate enclosed space of a single class of interest to verify whether a sample is compatible or not with the characteristic of that class. SIMCA is a class modeling technique that builds a class model based on the significant PCs of the category. [26–28] Then, in this technique the models (one for each classes) can overlap and/or leave some regions of the multivariate space unassigned. Therefore, two parameters to validate the classification can be defined: sensitivity (SENS) and specificity (SPEC).[29] SENS of a class is referred to the percentage of objects belonging to the class correctly accepted by the class model. SPEC of a class corresponds to the percentage of objects not belonging to the class correctly rejected by the class model. SIMCA was applied to the data matrix constituted by 37 rows and 10 columns (corresponding to the ten variables previously selected), and models obtained were based on the first nine PCs for each class. Validation of the model was carried out by a fivefold cross-validation procedure. The SIMCA model has allowed the correct classification of all samples for each category and, moreover, has provided excellent results also in terms of mean sensitivity (100%) and mean specificity (100%, Table 1). In order to check the addition of essences with lower commercial value to pure bergamot essential oil, as would more likely be the case in adulteration situation, the same chemometric approaches were applied to the data matrix constituted by 40 rows (20 bergamot essential oils and 20 mixtures of bergamot oil with the other essential oils) and 401 columns (ions acquired in the mass range m/z 100–500). Also, in this case, LDA and SIMCA were carried out using two classes as input a priori: bergamot and mixture. Initially, LDA was performed by considering from the first four to the first 13 PCs (72–90% of total variance), and the best result in terms of mean prediction ability (95%) was obtained by taking into account the first five PCs (76% of total variance). This multivariate model



765

Bergamot Non-bergamot Mean prediction ability 100%

Bergamot 20 1

Wilks lambda

Journal of

MASS SPECTROMETRY

D. Taverna et al.

Table 3. Number of correctly classified samples (prediction matrices) for LDA performed on PCs, stepwise LDA and SIMCA techniques (rows represent the true class, columns report the assigned class). Mixtures were prepared with 70 parts by weight of bergamot essential oil and 30 parts by weight of the other essential oil LDA performed on principal components

Bergamot Mixture Mean prediction ability 97.3% Stepwise LDA Bergamot Mixture Mean prediction ability 100% SIMCA Bergamot Mixture Mean prediction ability 100%

Bergamot 20 2

Mixture 0 18

Bergamot 20 0

Mixture 0 20

Bergamot 20 0

Mixture 0 20

classified correctly all the bergamot oils, whereas two mixture samples were included in the pure bergamot oil category (Table 3). The same data matrix was subjected to stepwise LDA. This analysis has retained seven variables, and the following cross-calibration procedure has correctly predicted all samples for each category (Table 3). Finally, SIMCA was applied to the data by considering the seven variables selected by the stepwise LDA model. Also in this case, all samples were correctly classified, and excellent sensitivity and specificity values were obtained (Table 3).

Conclusion and perspectives

766

Ambient mass spectrometry is a suitable technique allowing for a rapid molecular profiling of a raw material at room temperature and at atmospheric pressure. So, this mass spectrometry-based technique, here used for bergamot essential oil discrimination from other similar and cheaper essential oils commonly used for its reconstitution, is probably one of the most rapid, economic and enough sensitive for quality control. Besides that, the PS-MS approach offers a host of advantages: rapid analysis (2 min) can be carried out; minimum sample pre-treatment is needed, thus saving time and laboratory costs; the data processing required for a molecular profiling is easy and fast as well. The capability of PS-MS profiling to act as marker for the classification of bergamot essential oils was evaluated by using multivariate statistical analysis. Two pattern recognition techniques (LDA and SIMCA), each with different characteristics, were applied to MS data, and prediction ability of these chemometric techniques was tested using cross-validation procedure. Both approaches allowed us to obtain excellent results because the constructed models have correctly classified all samples for each category. In order to check the applicability of the proposed protocol to adulteration cases, the same chemometric approaches were applied to data matrix constituted by pure bergamot oils and mixtures of bergamot oil with the other essential oils. Also, in this case, models allowed to correctly classify all samples for each class. This suggests that the presented method may be a potential choice for discriminating


SENS (%) 100 100

SPEC (%) 100 100

bergamot essential oils from other similar ones commonly used for its marketing. Finally, the PS-MS methods could be implemented directly at the facilities of the Calabrian bergamot oil producers.

Acknowledgements Authors would like to thank Commission European Union, European Social Fund, PSR 2007–2013, Misura 1.2.4. for fundings and Dr Ezio Pizzi, President of ‘Consorzio del bergamotto’ of Reggio Calabria, Italy, for bergamot essential oil samples.

References [1] L. Di Donna, G. Gallucci, N. Malaj, E. Romano, A. Tagarelli, G. Sindona. Recycling of industrial essential oil waste: Brutieridin and Melitidin, two anticholestrolaemic active principles from bergamot albedo. Food Chem. 2011, 125(2), 438–441. [2] L. Di Donna, G. De Luca, F. Mazzotti, A. Napoli, R. Salerno, D. Taverna, G. Sindona. Statin-like principles of bergamot fruit (Citrus bergamia): isolation of 3-hydroxymethylglutaryl flavonoid glycosides. J. Nat. Prod. 2009, 72, 7, 1352. [3] M. Leopoldini, N. Malaj, M. Toscano, G. Sindona, N. Russo. On the inhibitor effects of bergamot juice flavonoids binding to the 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) enzyme. J. Agric. Food Chem. 2010, 58, 10768. [4] G. Mandalari, R. N. Bennett, G. Bisignano, D. Trombetta, A. Saija, C. B. Faulds, M. J. Gasson, A. Narbad. Title: Antimicrobial activity of flavonoids extracted from bergamot (Citrus bergamia Risso) peel, a byproduct of the essential oil industry. J. Appl. Microbiol. 2007, 103, 2056. [5] M. Russo, I. Bonaccorsi, R. Costa, A. Trozzi, P. Dugo, L. Mondello. Reduced time HPLC analyses for fast quality control of citrus essential oils. J. Essential Oil Res. 2015, 27, 307. [6] P. Q. Tranchida, I. Bonaccorsi, P. Dugo, L. Mondello, G. Dugo. Analysis of Citrus essential oils: state of the art and future perspectives. A review. Flavour Fragr. J. 2012, 27, 2. [7] P. Dugo, M. Russo. The oxygen heterocyclic components of citrus essential oils. In Citrus Oils: Composition, Advanced Analytical Techniques, Contaminants, Biological Activity(G. Dugo, L. Mondello Eds.). Taylor & Francis: London, 2011 405–443.


J. Mass Spectrom. 2016, 51, 761–767

Journal of

MASS SPECTROMETRY

PS-MS and chemometrics of bergamot essential oil [8] A. Dahan, H. Altman. Food–drug interaction: grape-fruit juice augments drug bioavailability—mechanism, extent and relevance. Eur. J. Clin. Nutr. 2004, 58, 1–9. [9] G. Dugo, P. Q. Tranchida, A. Cotroneo, P. Dugo, I. Bonaccorsi, P. Marriott, R. Shellie, L. Mondello. Advanced and innovative chromatographic techniques for the study of citrus essential oils. Flavour Fragr. J. 2005, 20, 249–264. [10] G. Dugo, I. Bonaccorsi, D. Sciarrone, L. Schipilliti, M. Russo, V. Raymo, A. Cotroneo, P. Dugo, L. Mon-dello. Characterization of cold-pressed and processed bergamot oils by using GCFID, GC–MS, GC–C-IRMS, enantio-GC, MDGC, HPLC, and HPLC-MS-ITTOF. J. Essent. Oil Res. 2012, 24, 93–117. [11] L. Di Donna, D. Iacopetta, A. R. Cappello, G. Gallucci, E. Martello, M. Fiorillo, V. Dolce, G. Sindona. Hypocholesterolaemic activity of 3hydroxy-3-methyl-glutaryl flavanones enriched fraction from bergamot fruit (Citrus bergamia): “in vivo” studies. J. Func. Foods. 2014, 7, 558–568. [12] A. Wei, T. Shibamoto. Antioxidant/Lipoxygenase inhibitory activities and chemical compositions of selected essential oils. J. Agric. Food Chem. 2010, 58, 7218–7225. [13] E. L. Belsito, C. Carbone, M. L. Di Gioia, A. Leggio, A. Liguori, F. Perri, C. Siciliano, M. C. Viscomi. Comparison of the volatile constituents in cold-pressed bergamot oil and a volatile oil isolated by vacuum distillation. J. Agric. Food Chem. 2007, 55, 7847–7851. [14] W. Si, X. Ni, J. Gong, H. Yu, R. Tsao, Y. Han, J. R. Chambers. Antimicrobial activity of essential oils and structurally related synthetic food additives towards Clostridium perfringens. J. Appl. Microbiol. 2009, 106, 213–220. [15] H. F. Yen, C. T. Hsieh, T. J. Hsieh, F. R. Chang, C. K. Wang. In vitro antidiabetic effect and chemical component analysis of 29 essential oils products. J. Food and Drug Analysis. 2015, 23, 124–129. [16] J. Liu, H. Wang, N. E. Manicke, J. Lin, R. G. Cooks, Z. Ouyang. Development, characterization, and application of paper spray ionization. Anal. Chem. 2010, 82, 2463–2471. [17] D. Taverna, L. Di Donna, F. Mazzotti, B. Policicchio, G. Sindona. Highthroughput determination of Sudan Azo-dyes within powdered chili pepper by paper spray mass spectrometry. J. Mass Spectrom. 2013, 48, 544–547. [18] F. Mazzotti, L. Di Donna, D. Taverna, M. Nardi, D. Aiello, A. Napoli, G. Sindona. Evaluation of dialdehydic anti-inflammatory active principles in extra-virgin olive oil by reactive paper spray mass spectrometry. International Journal of Mass Spectrometry 2013, 352, 87–91.

[19] Z. Takàts, J. M. Wiseman, B. Gologan, R. G. Cooks. Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science 2004, 306, 471–473. [20] J. H. Gross. Direct analysis in real time: a critical review on DART MS. Anal. Bioanal. Chem. 2014, 406, 63–80. [21] M. Forina, S. Lanteri, C. Armanino, C. Casolino, M. Casale, P. Oliveri. 2009. V-PARVUS 2009, Univerity of Genoa, Genoa. [22] C. Benincasa, J. Lewis, G. Sindona, A. Tagarelli. The use of multi element profiling to differentiate between cow and buffalo milk. Food Chem. 2008, 110, 257–262. [23] C. Benincasa, J. Lewis, E. Perri, G. Sindona, A. Tagarelli. Determination of trace element in Italian virgin olive oils and their characterization according to geographical origin by statistical analysis. Anal. Chim. Acta 2007, 585, 366–370. [24] G. Lo Feudo, A. Naccarato, G. Sindona, A. Tagarelli. Investigating the origin of tomatoes and triple concentrated tomato pastes through multielement determination by inductively coupled plasma mass spectrometry and statistical analysis. J. Agric. Food Chem. 2010, 58, 3801–3807. [25] L. Di Donna, F. Mazzotti, A. Naccarato, R. Salerno, A. Tagarelli, D. Taverna, G. Sindona. Secondary metabolites of Olea europaea leaves as markers for the discrimination of cultivars and cultivation zones by multivariate analysis. Food Chem. 2010, 121, 492–496. [26] H. Benabdelkamel, L. Di Donna, F. Mazzotti, A. Naccarato, G. Sindona, A. Tagarelli, D. Taverna. Authenticity of PGI “Clementine of Calabria” by multielement fingerprint. J. Agric. Food Chem. 2012, 60, 3717–3726. [27] E. Furia, A. Naccarato, G. Sindona, G. Stabile, A. Tagarelli. Multielement fingerprinting as a tool in origin authentication of PGI food products: tropea red onion. J. Agric. Food Chem. 2011, 59, 8450–8457. [28] G. Lo Feudo, B. Macchione, A. Naccarato, G. Sindona, A. Tagarelli. The volatile fraction profiling of fresh tomatoes and triple concentrate tomato pastes as parameter for the determination of geographical origin. Food Res. Int. 2011, 44, 781–788. [29] M. Forina, C. Armanino, R. Leardi, G. Drava. A class-modelling technique based on potential functions. J. Chemometrics. 1991, 5, 435 453.

Supporting information Additional supporting information may be found in the online version of this article at the publisher’s web site.

767

J. Mass Spectrom. 2016, 51, 761–767

