Establishing the Phenolic Composition of Olea europaea L. Leaves from Cultivars Grown in Morocco as a Crucial Step Towards Their Subsequent Exploitation

In Morocco, the recovery of olive agro-industrial by-products as potential sources of high-added value substances has been underestimated so far. A comprehensive quantitative characterization of olive leaves’ bioactive compounds is crucial for any attempt to change this situation and to implement the valorization concept in emerging countries. Thus, the phenolic fraction of olive leaves of 11 varieties (‘Arbequina’, ‘Hojiblanca’, ‘Frantoio’, ‘Koroneiki’, ‘Lechín’, ‘Lucque’, ‘Manzanilla’, ‘Picholine de Languedoc’, ‘Picholine Marocaine’, ‘Picual’ and ‘Verdal’), cultivated in the Moroccan Meknès region, was investigated. Thirty eight phenolic or related compounds (including 16 secoiridoids, nine flavonoids in their aglycone form, seven flavonoids in glycosylated form, four simple phenols, one phenolic acid and one lignan) were determined in a total of 55 samples by using ultrasonic-assisted extraction and liquid chromatography coupled to electrospray ionization-ion trap mass spectrometry (LC-ESI-IT MS). Very remarkable quantitative differences were observed among the profiles of the studied cultivars. ‘Picholine Marocaine’ variety exhibited the highest total phenolic content (around 44 g/kg dry weight (DW)), and logically showed the highest concentration in terms of various individual compounds. In addition, chemometrics (principal components analysis (PCA) and stepwise-linear discriminant analysis (s-LDA)) were applied to the quantitative phenolic compound data, allowing good discrimination of the selected samples according to their varietal origin.


Introduction
Global production of virgin olive oil has steadily increased over the past decades, reaching 3.1 million tons during the 2017/2018 crop season [1,2], which makes olive tree the sixth most relevant oil crop in the world [3]. Furthermore, its undeniable economic importance has

Profiling and Qualitative Characterization of the Phenolic Fraction of Olive Leaves from the Selected Eleven Cultivars
The first stage of this work was designed to carry out a comprehensive characterization of the phenolic profiles of the leaves from different olive varieties, trying to identify as many compounds as possible. Tentative identifications were achieved by considering the information provided by the two detectors (DAD (UV-vis spectra) and MS (m/z spectral data)), the data achieved for the commercial standards (when available), as well as by comparing the information regarding retention time and elution order with the previously published reports [26][27][28][29][30]. Accurate mass data obtained in full-scan mode in a Q-TOF MS was processed with the SmartFormula™ Editor tool included in DataAnalysis 4.0 (Bruker Daltonik, Bremen, Germany), which provides a list of possible elemental formulas. Table 1 lists (according to their elution order) the 38 phenolic compounds tentatively identified in the studied leaves samples and presents the calculated molecular formula for each compound, together with the error (difference between experimental and theoretical m/z of the detected [M − H] − ion) and mSigma™ (value showing the concordance with the theoretical isotopic pattern of the compound). Figure 1 shows the Extracted Ion Chromatograms (EICs) of the main identified phenolic compounds found in a sample of 'Picholine Marocaine' leaves.

Profiling and Qualitative Characterization of the Phenolic Fraction of Olive Leaves from the Selected Eleven Cultivars
The first stage of this work was designed to carry out a comprehensive characterization of the phenolic profiles of the leaves from different olive varieties, trying to identify as many compounds as possible. Tentative identifications were achieved by considering the information provided by the two detectors (DAD (UV-vis spectra) and MS (m/z spectral data)), the data achieved for the commercial standards (when available), as well as by comparing the information regarding retention time and elution order with the previously published reports [26][27][28][29][30]. Accurate mass data obtained in full-scan mode in a Q-TOF MS was processed with the SmartFormula™ Editor tool included in DataAnalysis 4.0 (Bruker Daltonik, Bremen, Germany), which provides a list of possible elemental formulas. Table 1 lists (according to their elution order) the 38 phenolic compounds tentatively identified in the studied leaves samples and presents the calculated molecular formula for each compound, together with the error (difference between experimental and theoretical m/z of the detected [M − H] − ion) and mSigma™ (value showing the concordance with the theoretical isotopic pattern of the compound). Figure 1 shows the Extracted Ion Chromatograms (EICs) of the main identified phenolic compounds found in a sample of 'Picholine Marocaine' leaves.  Table 1.

Phenolic Contents in Different Olive Leaves Cultivars
Prior to quantifying the identified phenolic compounds, the analytical method was properly validated in terms of linearity, precision (intra-and interday repeatability), limit of detection (LOD) and limit of quantification (LOQ). Thus, as reported in Section 3.2.1, dilutions of the standard solution mixture were prepared and injected into the LC-IT MS system (which was the instrument used for quantifying). Method linearity was evaluated by plotting the peak areas versus the corresponding concentrations (mg/L) of each standard analyte using the least squares method. Calibration curves were built using the values from three replicates of each concentration level analyzed within the same day (n = 3). LODs and LOQs of the individual compounds in the standard solutions were calculated as the lowest concentration at which a signal-to-noise (S/N) ratio was greater than 3 and 10, respectively. Intra-and interday repeatability were also estimated; to do it so, we calculated the relative standard deviation (RSD (%)) of peak area for 4 injections of 4 different extracts of the quality control (QC) sample carried out within the same sequence (intraday) or over 4 days (interday). Obtained results for the evaluated analytical parameters are summarized in Table S1 (Supplementary materials).
As shown in the table, linearity of the method was satisfactory over the assayed range with correlation coefficient (r 2 ) higher than 0.9918 in all cases. The LODs ranged from 3 to 97 µg/L and the LOQs ranged from 11 to 325 µg/L, for apigenin and rutin, apiece. The method led to excellent precision values (RSD (%)) always lower than 9.4% (values ranged from 1.8% to 7.5% for the intra-day repeatability and from 2.1% to 9.4% for the inter-day repeatability). Consequently, the proposed analytical method could be successfully applied for the determination of 38 phenolic compounds in the selected 55 olive leaves samples.
Quantification in MS was done using external calibration curves of the corresponding pure standard analytes for: Oleuropein, apigenin, apigenin-7-glucoside, hydroxytyrosol, luteolin, luteolin-7-glucoside, pinoresinol, rutin, tyrosol and vanillic acid, whereas for those identified compounds for which reference pure standards were not available, a calibration curve from structurally related substances was used. Thus, tyrosol glucoside, elenolic acid glucoside isomers (1, 2 and 3), secologanoside isomers (1 and 2) and ligstroside aglycon were quantified using tyrosol calibration curve; hydroxytyrosol glucoside and oleuropein aglycon isomers (1 and 2) were quantified in terms of hydroxytyrosol; apigenin rutinoside and luteolin diglucoside in terms of rutin; chrysoeriol-7-glucoside and luteolin-glucoside isomers (1 and 2) by using luteolin-7-glucoside calibration curve; to quantify oleuropein diglucoside, 2"-methoxyoleoropein isomers (1 and 2), hydroxyoleuropein and ligstroside, the standard of oleuropein was employed; naringenin was determined in terms of apigenin; and finally, quercetin, diosmetin, and the unknown isomers of C 15 H 8 O 7 were quantified by using luteolin as reference standard. It is important to bear in mind that the response of the standards can differ from the response of the analytes present in the olive leave extract samples, and consequently, the quantification of these compounds (both in terms of total amount and individual contents) is only an estimation of their occurrence in the analyzed samples.
The total phenolic compounds content (sum of the content of individual phenolic compounds determined) and the total phenolic content per chemical class (sum of the content of individual phenolic compounds belonging to the same chemical family) of the olive leaves from the different studied cultivars are given in Figure 2. Results are expressed as mean ± standard deviation. As can be seen, on average terms, total phenolic content ranged from around 11 g/kg DW to 44 g/kg DW; 'Picual' was the poorest variety of the studied selection and 'Picholine Marocaine' was the richest one. Secoiridoids were by far the most abundant group of phenols in all the analyzed samples regardless of the variety, excepting 'Arbequina' and 'Picual' samples for which flavonoids (in glycosylated form) were predominant.  Among the studied cultivars, the highest secoiridoids content (34 g/kg DW) was found in 'Picholine Marocaine' leaves extracts, whilst 'Picual' samples presented the lowest concentration level (5 g/kg DW). The highest level of total flavonoids in glycosylated form was observed in 'Picholine de Languedoc' samples (10 g/kg DW) and the lowest one (6 g/kg DW) in 'Verdal' leaves; however, regarding this group of analytes, the differences found among the cultivars were not as noticeable as for others. As far as the other sub-category of flavonoids is concerned, it is possible to highlight that flavonoids in aglycon form were found within the range 165-532 mg/kg DW, defined by 'Picholine Marocaine' and 'Arbequina', respectively. The content in terms of simple phenols and, in particular, the amounts of vanillic acid and pinoresinol were negligible-in all the cultivars-when compared with secoiridoids levels. In this regard, the concentrations of simple phenols ranged between 218 mg/kg DW and 2124 mg/kg DW, for 'Frantoio' and 'Picholine Marocaine' leaves extracts, respectively. The content of the quantified lignan was found between 8.7 mg/kg DW (Lucque) and 16 mg/kg DW ('Frantoio'). Finally, the amount of the phenolic acid fluctuated from 7 mg/kg DW to 19 mg/kg DW; 'Picholine Marocaine' and 'Picual' exhibited the extreme concentration levels.
After getting the quantitative results, the existence of significant variations (both regarding total phenolic content and chemical class content) was investigated. One-way ANOVA revealed statistically significant differences among the concentration of phenolic compounds in leaves from different cultivars. Our results support those found in literature with regard to the intervariety variability of the total phenolic content in olive leaves [26,27,30,31]. In general, our quantitative data are also similar to those included in previous reports, even though the comparison in this regard is not very straightforward; it is necessary to check whether the results from other authors are given as DW (or maybe without drying), and also to have a look at the compounds used as pure standards for the quantification and the methodology applied (extraction protocol and determination conditions). In addition, there are other obvious factors influencing the possible quantitative results, such as the cultivar, the pedoclimatic conditions, the harvesting time, etc.
In this work, for instance, the adaptability of an olive variety to the pedoclimatic conditions of the site of cultivation could largely condition its leaves metabolites. That could explain the divergence between our results regarding 'Picual' and 'Arbequina' cv. and those achieved by Talhaoui et al. [26,27]; generally the concentration levels found for some phenolic compounds were higher for the varieties which were cultivated in their country of origin (Spain, in this case). The same is applicable to underline that 'Picholine Marocaine' proved to be the cultivar (from the 11 selected herewith) with the highest quantity of phenolic compounds, possibly due to the fact that it is a Moroccan autochthonous variety with verified high adaptability to Moroccan environmental conditions.
Considering the simple phenols content, the selected varieties could be clustered in two groups: those with hydroxytyrosol as the most abundant simple phenol ('Arbequina', 'Frantoio', 'Lucque', 'Manzanilla', 'Picholine de Languedoc', 'Picual' and 'Verdal'), and those cultivars with hydroxytyrosol glucoside as the predominant substance within this category ('Hojiblanca', 'Koroneiki', 'Lucque', and 'Picholine Marocaine'). Hydroxytyrosol levels varied from 119 to 323 mg/kg DW, in 'Frantoio' and 'Picholine Marocaine', respectively. The latter variety was also the richest regarding hydroxytyrosol glucoside (1510 mg/kg DW), whilst 'Arbequina' was the poorest one (10 mg/kg DW). Tyrosol (23-61 mg/kg DW) and tyrosol glucoside (48-237 mg/kg DW) were also found in the samples under study. Vanillic acid and pinoresinol were quantified in the studied olive leaves too. Their concentration levels were relatively low in every sample (<19 mg/kg DW for vanillic acid, and <15 mg/kg DW for pinoresinol) (Tables 2-4). Table 4. Found content (average values and standard deviation, mg/kg DW) of the determined phenolic compounds in the evaluated olive leaves cultivars. ANOVA results are included; significant differences in the same row are indicated with different superscript letters (comparison among the 11 cultivars investigated in this study, p < 0.05).

'Picholine Marocaine' 'Picual' 'Verdal'
Hydroxytyrosol glucoside 1510 d ± 67 11 a ± 6 15 a ± 9 Secologanoside is. The results of the current study demonstrate that content of individual phenolic compounds in olive leaves is, as expected, closely related to the variety. Indeed, when compared by one-way ANOVA, the contents of the determined compounds were significantly different among the cultivars. Since all the varieties investigated in the current work were grown in the same experimental field using similar agronomic practices, the observed differences regarding the biosynthesis of secondary metabolites can be attributed to the genetic variability. These findings are in good agreement with those reported in literature, as reviewed in detail by Talhaoui and co-workers [24].
Besides, the results of Tukey's test indicated that individual contents of olive leaves from different cultivars had their own features. Focusing, for instance, on 'Picholine Marocaine' traits (Table 4), some specific characteristics can be pointed out. These leaves showed, on average, the highest total phenolic compounds content. This variety is the richest one in terms of secoiridoids (presenting the highest amount of various of these compounds); it presents low concentrations levels of flavonoids in aglycon form, lignans and phenolic acids; however, it contains considerable amounts of simple phenols (in particular, hydroxytyrosol glucoside) and flavonoids in glycosylated form. Thus, it appears that this variety presents, among the other studied cultivars, the greatest potential to be used as plausible source of bioactive compounds, what means that it could be a very promising choice in a future strategy of recycling and valorization of olive leaves from Moroccan olive agro-industry.

Varietal Discrimination
The genetic diversity of olive trees cultivated all around the world has been explored to identify their varietal origin. Discrimination of the varietal origin of olive trees based on their leaves traits is frequently carried out studying morphological characteristics and genetic markers. Certainly, great advances have been made to explore and prove the usefulness of various olive leaf's molecular markers, such as amplified fragment length polymorphism, random amplified polymorphic DNA and genomic simple sequence repeat, as reliable tools to differentiate and characterize the genetic diversity of olive cultivars [33,34]. Although these techniques are very valuable, they also have some drawbacks such as complicated pretreatment and DNA extraction procedures, high cost and special requirements for operators. Consequently, there is a need to explore the effectiveness of other analytical approaches to deal with these limitations. The combined application of profiling of olive leaves and chemometrics could be an effective alternative. Hence, in this study, beyond our interest on evaluating the phenolic composition of leaves from different cultivars, we also explored the ability of these compounds to trace the samples varietal origin.
A first attempt to differentiate among the studied varieties was carried out by applying principal components analysis (PCA) to a standardized and centered matrix data, which was constructed with the 38 measured variables (phenolic compounds) and the 55 leaves samples (three extraction replicates). PCA was logically employed as unsupervised method to examine natural grouping of the samples according to their varietal origin in two-dimensional principal components (PCs) plans where each PC is a linear correlation of the original variables (latent variable), and each PC is orthogonal to any other. In this manner, this method studies data structure in a reduced dimension, covering the maximum amount of the information present in the original dataset.
The results of s-LDA classification and prediction are summarized in the confusion matrices shown in Table 5, displaying re-allocation of samples coming from a given cultivar (corresponding to a matrix row) into the possible categories (the columns). As can be seen from this table, the s-LDA discriminant functions achieved very satisfactory recognition and prediction abilities, being the overall correct rate in both cases 100%. Accordingly, it is possible to assert that the olive leaves phenolic content could be useful for olive cultivars differentiation. Subsequently, the potential of applying a supervised multivariate method (stepwise linear discriminant analysis (s-LDA)) was tested. The applicability of the method was cross-validated by using the leave-one-out procedure. The Wilks λ value (0.000) showed that the model was very discriminating, and, in addition, revealed that the probability of correct classification was very high, considering that the p value was very low (p < 0.0001). Moreover, the forward stepwise statistics, with F-to-enter equal to 1.0 and F-to-remove equal to 0.5, selected 20 variables to be used in the relevant final models: hydroxytyrosol glucoside, 2"-methoxyoleuropein isomer 2, apigenin-7-glucoside, unknown isomer 1, unknown isomer 2, unknown isomer 3, elenolic acid glucoside isomer 1, elenolic acid glucoside isomer 2, ligstroside, ligstroside aglycon, luteolin, luteolin diglucoside, luteolin-glucoside isomer 1, oleuropein aglycon isomer 1, oleuropein isomer 2, oleuropein isomer 3, rutin, secologanoside isomer 1, secologanoside isomer 2 and tyrosol glucoside.
The results of s-LDA classification and prediction are summarized in the confusion matrices shown in Table 5, displaying re-allocation of samples coming from a given cultivar (corresponding to a matrix row) into the possible categories (the columns). As can be seen from this table, the s-LDA discriminant functions achieved very satisfactory recognition and prediction abilities, being the overall correct rate in both cases 100%. Accordingly, it is possible to assert that the olive leaves phenolic content could be useful for olive cultivars differentiation.

Olive Leaves Sampling and Preparation
In order to avoid any possible influence of the environmental and agricultural management practices on the obtained results, all olive leaves samples were collected at an experimental orchard in the National School of Agriculture of Meknès in Northern Morocco. Sampling was performed in December 2015, coinciding with the harvesting season in Meknès region, when olive leaves are available as an olive oil processing by-product. This region has a Mediterranean climate type with an average pluviometry of 660 mm/year, and hot and dry summers (maximum temperature up to 40 • C). All necessary agronomic practices (pruning, irrigation, fertilization and pest management) were done according to current olive orchards management standards. Olive trees were vase-trained at a spacing of 7 × 5 m.
Eleven different cultivars were included in this study: a Moroccan autochthonous and predominant variety so-called 'Picholine Marocaine', and ten Mediterranean cultivars recently introduced in Morocco ('Arbequina', 'Hojiblanca', 'Frantoio', 'Koroneiki', 'Lechín', 'Lucque', 'Manzanilla', 'Picholine de Languedoc', 'Picual' and 'Verdal'). Five olive leaves samples per cultivar were randomly collected from cardinally-oriented branches with different directions around the tree's canopy. Accordingly, a total of 55 olive leaves samples were considered in this work. The leaves were dried at room temperature to constant weight during several days. Once their water content was less than 3%, samples were finely ground in a kind of coffee grinder (but controlling the temperature). Average moisture was calculated after drying different samples in a desiccation oven for 12 h at 100 • C (these tests were just valid to assess the olive leaves moisture; the extraction protocol was obviously not applied to the resulting dried olive leaves). Pre-treated samples were stored in sealed containers and kept below −20 • C in the absence of light till analyzed.
A QC sample was prepared by mixing an equivalent amount of each one of the studied samples; it was used for different purposes: To optimize the extraction procedure, to ensure the proper performance of the analytical system, and to evaluate the analytical parameters of the method.

Chemical and Reagents
All the chemicals used in this study were of analytical grade. Water was daily deionized by using a Milli-Q system from Millipore (Bedford, MA, USA). Ethanol was supplied by J.T. Baker (Deventer, The Netherlands). Methanol and acetonitrile, both of LC-MS grade, were purchased from Prolabo (Paris, France). Acetic acid and pure standards of apigenin, apigenin-7-glucoside, hydroxytyrosol, luteolin, luteolin-7-glucoside, pinoresinol, rutin, tyrosol and vanillic acid were acquired from Sigma-Aldrich (St. Louis, MO, USA); whereas oleuropein was purchased from Extrasynthese (Lyon, France).
A stock standard solution was prepared by dissolving the appropriate amount of each compound in methanol. Then, diluted working solutions were obtained at nine different concentrations (0.5 mg/L; 1 mg/L; 2.5 mg/L; 5 mg/L; 12.5 mg/L; 25 mg/L; 50 mg/L; 100 mg/L and 200 mg/L) and were stored at −20 • C. If any other concentration level was required for a particular sample or to establish the analytical parameters of the method, it was logically prepared.

Phenolic Compounds Extraction
Pre-treated olive leaves were taken from the freezer and sieved through a 0.5 mm metal sieve, to obtain a standard particle size. 0.1 g of each powdered sample were accurately weighed into a centrifuge tube with a screw cap, and 10 mL of ethanol-water (80:20, v/v) were added. Then, the mixture was vortexed for 45 s and sonicated for 30 min in an ultrasonic bath from J.P. Selecta (Barcelona, Spain). The resulting extract was centrifuged for 5 min at 5974 g, the supernatant was collected and the residue was re-extracted again following the same procedure as above. Both supernatants were pooled and evaporated to dryness under reduced pressure at 35 • C in a rotavap R-210 (Buchi Labortechnik AG, Flawil, Switzerland). Next, the residue was reconstituted with 5 mL methanol, filtered through a 0.22 µm Nylaflo™ nylon membrane filter from Pall Corporation (Ann Arbor, MI, USA) and subsequently analyzed (or stored in a freezer below −20 • C prior to analysis). Each sample was prepared in triplicate. Every sample was extracted and analyzed by LC-MS on the same day (or within 48-72 h approx.).

Analytical Procedure and MS Conditions
For chromatographic analysis, an Agilent 1200 Series HPLC system (Agilent Technologies, Santa Clara, CA, USA) operated by Windows NT based ChemStation software and equipped with a binary solvent pump, a degasser, an autosampler, a column oven and a diode array detector (DAD) was used. Separation was performed on a Zorbax C18 analytical column (4.6 × 150 mm, 1.8 µm particle size) from Agilent Technologies (Santa Clara, CA, USA) protected by a guard cartridge and maintained at 25 • C. Injection volume was set at 5 µL. Phenolic compounds elution was achieved with 0.5% acetic acid in water (Phase A) and acetonitrile (Phase B) at a flow rate of 0.8 mL/min and the following gradient program: 0 to 25 min, 5-50% B; 25 to 27 min, 50-95% B; 27 to 27.5 min, 95-100% B; finally, the B content was decreased to the initial conditions (5%) in 1 min and the column was re-equilibrated for 0.5 min prior to the next injection. Double on-line detection was carried out using a DAD (with 240 nm, 254 nm, 280 nm and 330 nm as selected wavelengths) and a mass spectrometer.
MS analyses were made using two mass spectrometers (both running in negative ionization mode). The first one, a micrOTOF-Q II TM (Bruker Daltonik, Bremen, Germany) equipped with a quadrupole-time-of-flight (Q-TOF) analyzer and an electrospray ionization interface (ESI), was used to investigate the phenolic extracts of the studied olive leaves and to identify as many compounds as possible within the profiles. For this purpose, mixtures of all the extracts coming from the same variety (prepared by mixing an equivalent volume of each one) and the QC sample were analyzed by using this platform. External MS calibration was performed using a 74900-00-05 Cole Palmer syringe pump (manufactory, Vernon Hills, ID, USA) directly connected to the interface, equipped with a Hamilton (Reno, NV, USA) syringe. The calibration solution (sodium formate cluster containing 5 mM sodium hydroxide in the sheath liquid of 0.2% formic acid in water/isopropanol 1:1 v/v) was injected at the beginning of the run, and all the spectra were calibrated prior to compound identification. The other MS platform was a Bruker Daltonic Esquire 2000™ Ion Trap (IT) mass spectrometer (Bruker Daltonik), which was also coupled to the LC system through an ESI source. This coupling was used to carry out the quantification of the identified substances in all the samples under study.
For both MS detectors, the flow eluting from the LC column was split using a flow divisor 1:4, so that the flow rate entering into the MS detector was approximately 0.2 mL/min. The following source parameters were adopted for IT MS (and equivalent ones for Q-TOF MS): Capillary voltage, 3200 V; drying gas (N 2 ) flow and temperature, 9 L/min and 300 • C, respectively; nebulizer pressure, 30 psi. In IT MS, Ion Charge Control (ICC) was set at 10,000 and 50-1000 m/z was the selected scan range. Instrument control and data processing were carried out using the software Esquire Control and Data Analysis 4.0, respectively (Bruker Daltonik).
Quantitative determinations were carried out using the calibration curves obtained from commercially available pure standards. The results were expressed as mg of analyte/kg of olive leaves dry weight (DW).

Statistical Analysis
All data were reported as mean ± standard deviation (n = 5, corresponding to the number of samples per studied cultivar). Comparisons between means were performed by applying One-way Analysis of Variance (ANOVA) with Tukey's post-hoc test, using IBM SPSS Statistics 20 (SPSS Inc., Chicago, IL, USA). The differences between studied varieties were considered significant with p < 0.05. Furthermore, PCA and s-LDA were performed on phenolic compounds quantitative data to assess the potential of these substances to discriminate the studied samples according to their varietal origin. Multivariate data analysis was performed with the Microsoft Office Excel 2016 software (Microsoft Corporation, Redmon, WA, USA) and the statistical software XLSTAT version 2015.04.1 (Addinsoft, Paris, France).

Conclusions
The achieved results demonstrated-in the Moroccan context-the potential of the olive leaves as an underexploited natural source of interesting substances with inherent applications in different fields; their recovery could be a valuable alternative for the sustainable and environmentally friendly management of olive leaves mills by-products.
In Morocco, olive orchards are predominantly planted with 'Picholine Marocaine' variety. In 2015 about 1.15 million tons of olive fruits were harvested; olive leaves represented on average 6% of harvested olive fruits, which means about 27.6-34.5 thousand tons of dry olive leaves. Considering our results (for the autochthonous Moroccan cv. in particular), they could potentially contain around 650-825 tons of oleuropein, which are actually wasted. It is time to establish an integrated approach for the sustainable extraction of high value-added molecules from olive leaves in Morocco.
Apart from the clear future practical application of this work (isolation of the bioactive compounds of interest such as oleuropein), it is important to highlight that the comprehensive methodology used, combining LC-MS data on phenolic compounds and related substances with chemometrics, resulted to be a very effective tool for achieving an adequate discrimination among the olive leaves from different cultivars.
Supplementary Materials: The following are available online, Table S1: Analytical parameters of the developed LC-MS method, including calibration curves equations and r 2 , LOD and LOQ, linear ranges and repeatability (expressed as %RSD). Funding: This research was funded by the Spanish Government (Ministerio de Educación, Cultura y Deporte) with a FPU fellowship (FPU13/06438), the Vice-Rector's Office for International Relations and Development Cooperation of the University of Granada, and the contract 30C0366700 (OTRI, University of Granada, Spain).