Phenols, Volatile Compounds, Organic Acids and Antioxidant Activity of Strawberry Tree (Arbutus Unedo L.) Fruits Belonging to Five Genotypes Growing in Morocco

ABSTRACT This study aims to identify the individual phenolics and volatile compounds, as well as the organic acids of strawberry tree (Arbutus unedo L.) genotype fruits. The antioxidant activities were also assessed using three methods (DPPH, ABTS and βeta carotene bleaching assays) significant differences (p˂0.05) were observed among all the genotypes. Total phenols varied from 25.37 to 39.06 mg GAE/g dried weight (DW), total flavonoids ranged between 3.30 and 7.07 mg RE/g DW, and anthocyanins varied from 0.15 to 0.64 mg cya-3-glu/100 g DW. Moreover, the antioxidant activities were in the range of 3.33–21.08, 2.25–19.58, and 1.08–13 mg ascorbic acid equivalent/g DW for DPPH, ABTS and βeta carotene bleaching assays, respectively. Seventeen phenolics compounds were identified by HPLC in A. unedo fruits. Gallocatechol and catechin were the most abundant compounds. Among the volatile compounds identified, hexadecanoic acid was the most abundant in all the genotype fruits. The principal component analysis revealed that the first two components formed 66.47% of the total inertia.


Introduction
The strawberry Strawberry tree (Arbutus unedo L.), is evergreen shrub belonging to Ericaceae family endemic to Mediterranean region and North Africa (Sulusoglu and Cavusoglu, 2011). A. unedo is a Medicinal plant naturally grown as population or solitary tree in countries, such as Morocco, Tunisia, Algeria, Turkey, Syria, Greece, Croatia, France, Portugal, and Spain (Serçe et al., 2010). It is considered as an important source of molecules with high antioxidant potential, due mainly to polyphenols concentrated in its fruit, which play a major role in safeguarding health, because of their biological functions, such as antimutagenicity, anticarcinogenicity, and antiaging (Rodríguez et al., 2013). The A. unedo fruit is suitable for the production of alcoholic beverages, jams, jellies, and marmalades (Pallauf et al., 2008) but also for medicinal purposes (Ruiz-Rodriquez et al., 2011). In Morocco, it is known as "Sasnou" and it is widely used in traditional medicine, such as antiseptics, site, random samples of fruits were harvested at their fully ripened stage, and transferred to the laboratory for physicochemical and phytochemical analysis. Fruits were frozen at −80°C, freeze-dried, and ground, then kept in appropriate conditions for subsequent use.

Physico-Chemical Analyzes
Total soluble solids (TSS) were assessed according to (AOAC, 2002) by triplicate with a digital refractometer (Atago N1; Atago Co. Ltd., Tokyo, Japan) at 20°C and expressed as %. Total titratable acidity (TA) was also determined according to (AOAC, 2002) by triplicate using an automatic titration device (877 Titrino plus, Metrohm ion analyses CH9101, Herisau, Switzerland) with 0.1 N NaOH up to pH 8.1, using 1 mL diluted juice in 25 mL distilled H 2 O, and the results were expressed as g malic acid per 100 g fw (Celikel et al., 2008). The pH was measured using a pH meter according to the method described by (AOAC, 2002). Weigh 10 g of the fruit cut into small pieces, add 100 ml of distilled water, and mix for 5 min until juice was obtained. The measurement was made by immersing the pH meter electrode in the solution.

Organic Acids and Ascorbic Acid Profiles
The samples (0.5 g) were extracted with 5 mL of Milli-Q water by incubation for 30 min under ultrasonication at 25 and kHz 20°C as described by Hernández et al. (2016). Next, the slurry was centrifuged at 15,000 g for 20 min (Sigma 3-18 K; Sigma, Osterode am Harz, Germany), and the supernatant was filtered through a 0.45 μm Millipore filter and used for analysis. All extractions were carried out in triplicate. The chromatographic analysis was carried out according to Hernández et al. (2016). Thus, 10 μL of extract were injected into a Hewlett-Packard HPLC Series 1100 (Wilmington DE, USA) with an autosampler and an UV detector, set at 210 nm and coupled with a refractive index detector (HP 1100, G1362A). A column (Supelcogel TM C-610 H column 30 cm × 7.8 mm) and apre-column (Supelguard 5 cm × 4.6 mm; Supelco, Bellefonte, PA) were used for the analyses of both organic acids and ascorbic acid. The elution buffer consisted of 0.1% phosphoric (V/V) at a flow rate of 0.5 mL min− 1 , and organic acid absorbance was measured at 210 nm using a diode-array detector (DAD). Calibration curves were used for the quantification of organic acids and ascorbic acid showing good linearity (r 2 ≥ 0.999). The results were expressed as g 100 g −1 of dry weight (DW).

Extraction Procedure
One gram of powder from each sample was mixed with 25 mL of ethanol (1:25, w/v) at 25°C for 15 min using an IKA T-18 digital Ultra-Turrax homogenizer. The homogenate was then centrifuged for 10 min at 6,000 rpm and the supernatant was removed from the residue. The latter was homogenized with ethanol and the supernatant removed as above. The supernatants were then combined and filtered.

Total Phenols
Total phenols content (TPC) of A. unedo was determined by the reduction of phosphotungsticphosphomolybdic acid (Folin-Ciocalteu's reagent) to blue pigments, in alkaline solution according to Folin as described by Ben Salem et al. (2018). Briefly, 100 µL of diluted sample (1/100) with ethanol was added to 400 µL of 1/10 diluted Folin Ciocalteu reagent. After 5 min, 500 µL of 10% (w/v) sodium carbonate solution was added. After 1 h of incubation at room temperature, absorbance at 765 nm (spectrophotometer Spectraphysic Jasco V-630, Japan) was measured in triplicate. Total polyphenols content was expressed as mg gallic acid equivalents per g dry weight of A. unedo fruit (mg GAE/ g DW).

Total Flavonoids
Total flavonoids content (TFC) was measured using the colorimetric method with aluminum chloride (Lamaison and Carnat, 1990). One mL of the sample was diluted separately then mixed with 1 mL of a 2% aluminum chloride solution. The mixture was incubated at room temperature for 15 min. Rutin was used to develop the calibration curve. The absorbance was measured at 430 nm (spectrophotometer Spectraphysic Jasco V-630, Japan). The results were expressed as mg rutin equivalents per dry weight of A. unedo fruit (mg RE/g DW).

Determination of Antioxidant Activities
The antioxidant activity (AA) was evaluated using three different assays: (i) DPPH assay, (ii) ABTS assay, and (iii) the βeta carotene bleaching test. The antioxidant activity was determined in triplicate and the results were presented as a mean ± standard deviation.

DPPH Free Radical Scavenging Capacity
The DPPH (1,1-diphenyl-2-picrylhydrazyl) radical scavenging activity of the samples was determined according to Ben Salem et al. (2018). Thus, DPPH solution was prepared by dissolving 0.1 g of DPPH in 1 L methanol (0.1 g L −1 ). Then, one mL of this solution was added to 125 µL of each extract. The mixture was stirred thoroughly and incubated in the dark at room temperature for 10 min. The absorbance of both sample and control was measured at 517 nm using a Lambda EZ 150 (spectrophotometer Spectraphysic Jasco V-630, Japan), and the DPPH radical scavenging activity was calculated using the following equation (2): DPPH scavenged (%) = {(Ac -As)/Ac} * 100 (2) where, Ac and AS refer to the control and sample absorbances, respectively. IC50 value (mg equivalent to ascorbic acid/g dry weight) defines the inhibitory concentration at which tested radicals were scavenged by 50%. It was calculated by plotting inhibition percentage of each test against the sample extract dilutions.

ABTS Assay
The ABTS• [2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid] radical scavenging assay were determined according to Dorman and Hiltunen (2004). Thus, 990 µL of each extract was incubated in 10 µL ABTS (7 mM)-ETOH and 2.45 mM potassium persulfate solution after sonicated at 20°C for 15 min during 16 h in the dark. The mixtures were incubated for 18 h in the darkness at room temperature. The ethanol was used to dilute the stock solution of ABTS until absorbance of 0.70 ± 0.05 was reached at a wavelength of 734 nm. The antioxidant activity results were expressed as mg equivalent ascorbic acid per g dry weight (DW).

βeta Carotene Bleaching Assay
The βeta carotene blanching assay was determined according to Barros et al. (2010). βeta carotene (0.5 mg) in 1 mL of chloroform was taken in a amber bottle and mixed with 200 mg of linolenic acid and 600 mg of Tween 80 (polyoxyethelene sorbitan monopalmitate). The chloroform was removed under nitrogen, and the resulting solution was immediately diluted with 30 mL of triple distilled water and the emulsion was mixed well for 1 min. The emulsion was further diluted with 120 mL of oxygenated water and used for assay. To each sample extract (0.5 mL), 2.5 mL of the prepared emulsion mixture was added and then vigorously mixed. A control consisting 0.5 mL of ethanol and 2.5 mL of emulsion was also analyzed. The absorbance of reaction mixture was read immediately (t = 0) at 470 nm against blank, consisting of emulsion mixture, except β-carotene, and at the 60 min interval for 2 h (t = 120). The tubes were incubated in a water bath at a temperature of 50 •C between measurements. Color measurement was monitored until the β-carotene color disappeared. The linoleic acid peroxidation inhibition uses the following Equation (3): where, Ao and Aoo refer to the absorbance measured at the beginning of samples and control incubation, respectively. At and Aot are the final absorbance of samples and control, respectively.

Extraction Method
Samples (1 g) were mixed with 10 mL of methanol: water (80:20, v/v) and then, the mixtures were sonicated during 30 min, and macerated one hour in refrigeration (4°C). After the time, the samples were centrifuged for 10 min, 8000 g at 4°C. The supernatants were collected and the pellets were mixed with 10 mL of acetone: water (70:30, v/v) and the same steps were repeated (sonication, maceration, and centrifugation). Then, the supernatants were combined and evaporated to dryness using a rotary evaporator R-205 (Büchi, Flawil, Switzerland) under reduced pressure, at 40°C. 5 mL of methanol were added to the residue, and the mixture was well shaken in a stirrer for 2 min. Due to the high sugar content present in the samples, which could interfere with the HPLC column, the samples were loaded onto a C18 Sep-Pak cartridge, previously conditioned with 5 mL of methanol, 5 mL of pure water, and then with 5 mL of 0.01 mol L −1 HCl. The cartridge was washed with 5 mL of pure water and then eluted with acidified methanol (0.1 g L −1 HCl). The collected fractions were stored at −20°C until further use.

Determination of Polyphenolic Compounds
Polyphenolic profiles of all samples were determined by High-Performance Liquid Chromatography (HPLC) according to Genskowsky et al. (2016). A volume of 20 µL of the samples were injected into a Hewlett-Packard HPLC series 1200 instrument (Woldbronn, Germany) equipped with a diode array detector (DAD) and a C18 column (Mediterranea sea 18, 25 × 0.4 cm, 5 micrometers particle size) from Teknokroma, (Barcelona, Spain). Polyphenolic compounds were analyzed in standard and sample solutions using a gradient elution at 1 mL min-1 . The mobile phases were composed by formic acid in water (1:99, v/v) as solvent A and acetonitrile as solvent B. The chromatograms were recorded at 280, 320, 360, and 520 nm. Polyphenolic compounds identification was carried out by comparing UV absorption spectra and retention times of each compound with those of pure standards injected under the same conditions. The compounds were quantified through calibration curves of standard compounds injected under the same conditions. Phenolic acid standards were dissolved in methanol at different concentrations between 10 and 200 μg mL −1 ; flavonoids standards were dissolved in methanol at different concentrations between 1 and 250 μg mL −1 . Quantification of anthocyanins was carried out based on linear curves of authentic standards. A cyanidin 3-glucoside calibration (concentration between 1 and 250 μg mL −1 ) was used for cyanidin derivatives.

Extraction and GC-MS Analysis
Static headspace extraction of volatile compounds was performed by using solid-phase microextraction (SPME) with a 65 µm Divinylbenzene/Carboxen/Polydimethylsiloxane (DVB/CAR/PDMS) fiber. The analysis of A. unedo components was carried out by gas chromatography-mass spectrometry (GC-MS) using a gas chromatography Agilent 7890 A with masse selective detector 5975 Network MSD and coupled to an automatic sampling system MPS (Gerstel), a polyethylenglycol capillary column VF-WAXms (30 m × 0.25 mm i.d. × 0.25 μm film thickness) and a split/splitless injector, and the Library pal 600 k. About 1 g of the investigated sample was placed into a 20 mL vial closed with a screw and heated to 60°C for 20 min and the fiber was then exposed to strawberry headspace. After 20 min, the SPME fiber was automatically with drawn from the vial and introduced into the GC injector. Working conditions were: splitless mode with injector temperature at 250°C, the oven temperature program was 50°C for 4 min, rising at 5°C/min to 230°C (held for 10 min); then rising at 10°C/min to 250°C; and finally, 3 min at 250°C, a constant flow of 1 ml/min (helium) was set up. Mass spectra were recorded in EI mode at 70 eV, scanning the 35-395 m/z range. The interface and source temperatures were 230 and 250°C, respectively.

Statistical Analysis
Prior to the statistical analyses, data were tested for normality and homogeneity of variance using SPSS software v22 (IBM, SPSS Statistics, Armonk, New York, United States). The means were evaluated according to descriptive statistics represented as Mean ± SE. Data analysis was performed using IBM SPSS v22. Analysis of variance (One-way ANOVA) was performed to test significant differences among the samples. The differences among means were estimated with Duncan new multiple range test (DMRT). Correlation coefficients and their levels of significance were calculated using Pearson correlation. Principal component analysis was carried out using correlation matrix. In addition, a scatter plot was created according to the first two principal components (PC1 and PC2).

Physicochemical Parameters
The results for titratable acidity, pH and total soluble solids for all genotypes were summarized in Table 2. Significant variations were observed among genotypes (p < .001). The titratable acidity ranged from 0.65 to 1.01 g malic acid/100 g FW with an average of 0.83 g malic acid/100 g FW. The highest value was recorded in "TAH" (1.01 g malic acid/100 g FW) while the lowest value was observed in "MDZ" (0.65 g malic acid/100 g FW). The titratable acidity of A. unedo fruits reported in this study was higher than those found by other authors; Özcan and Hacıseferoğulları (2007) and Vidrih et al. (2013). They reported values of 0.51% and 0.4%, respectively. However, the results were lower than the ones recorded by Doukani and Hadjer (2015) (2.14%) in Algerian A. unedo genotypes. Also, Celikel et al. (2008) recorded titratable acidity value ranged between 0.80 and 1.59%. The significant difference in acidity could most probably be due to the climate factor and the process of fruit ripening (Messaid, 2008).
The pH values varied from 2.44 "KSB" to 3.92 "LAN" with an average of 3.36. These values were approximately similar with those revealed by Ruiz-Rodriquez et al. (2011) and González et al. (2011). They recorded 3.47 and 3.50, respectively. However, the results of this study were lower than those found by Serçe et al. (2010) and Özcan and Hacıseferoğulları (2007) who reported 5.57 and 4.6, respectively. The low pH value can have a big advantage in manufacturing. The differences depend on many factors including the climate, region, and ripeness of the fruit (Huberson, 2008;Messaid, 2008) The total soluble solids of A. unedo fruits varied from 14.83% "LAN" to 18.53% "KSB" with an average of 16.87%. Similar results were reported by Doukani and Tabak (2015). They recorded values ranged from 16.66 to 17.66%. The values of this study were higher than those found by Müller et al. (2010) and Serçe et al. (2010) who reported (8.1%) and (11.9%), respectively. The fruit of A. unedo L. has higher soluble solids content than that of Arbutus andrachnae (14%), blackberries (9.5%), and raspberries (6.2%) (Seker and Toplu, 2010). These differences can be related to the climate, soil type, and the process of fruit ripening (Serçe et al., 2010).

Organic Acids
The results obtained for organic acids content were reported in Table 3. Significant differences (P < .001) were observed among the genotypes. Four organic acids were identified by HPLC for all A. unedo genotypes. Citric acid, malic acid, ascorbic acid, and succinic acid were identified in the  investigated samples. Citric acid was determined as the major organic acid, followed by malic acid. The citric acid content ranged between 1.74 g/100 g "KSB" and 5.32 g/100 g "LAN" with an average of 3.17 g/100 g. The results of citric acid content in this study were higher than those reported by Serçe et al. (2010) and Doukani and Hadjer (2015) who recorded 0.03 g/100 g and 8.56 mg/100 g, respectively. However, Ruiz-Rodriquez et al. (2011) reported a total absence of citric acid. The malic acid content ranged from 1.53 g/100 g "KSB" to 2.87 "TAH" g/100 g with an average of 2.19 g/100 g. The results of malic acid content in this study were higher than those reported by Serçe et al. (2010) and Doukani and Hadjer (2015) who recorded 0.34 g/100 g and 282.3 mg/100 g values, respectively. However, the results obtained in this study were lower than those recorded by Alarcão-E-Silva et al. (2001). They reported 5.99 g/100 g in A. unedo fruits from Portugal. The ascorbic acid content varied from 0.28 g/100 g "KSB" to 1.00 g/100 g "TAH" with an average of 0.72 g/100 g. The results obtained were higher than those recorded by other authors; Pallauf et al. (2008) recorded 6.03 mg/100 g in Spanish A. unedo fruits. Ascorbic acid values recorded in this study were also higher than those reported by Alarcão-E-Silva et al. (2001); Pimpão et al. (2013) and Morales et al. (2013). They recorded 346 mg/100 g, 89 mg/100 g and 182 mg/100 g, respectively. The succinic acid content ranged from 0.49 g/100 g "CHF" to 4.66 g/100 g "LAN" with an average of 1.52 g/100 g. In another study, Doukani and Hadjer (2015) recorded traces of succinic acid in Algerian A. unedo fruits. Comparing our results with those of other authors, some organic acids were absents in our fruits, notably: oxalic, fumaric, lactic, suberic, and quinic acids. Fumaric (0.15 g/100 g), lactic (0.05 g/100 g), suberic (0.023 g/100 g), and quinic (7.35 g/100 g) acids were detected and quantified by Ayaz et al. (2000) in Turkish A. unedo fruits. In other studies, Ruiz-Rodriquez et al. (2011) and Morales et al. (2013) recorded values of oxalic acid 0.05-0.15 g/100 g and 0.09 g/100 g, respectively. The presence and composition of organic acids can be affected by various factors, such as: growing conditions, maturity, season, geographical origin, and soil type.

Total Phenols
The total phenols content (TPC) of A. unedo fruits were reported in Table 4. Significant differences (p = .044) were observed among the genotypes. The total phenols content ranged between 25.37 mg/g DW "KSB" and 39.06 mg GAE/g DW "LAN," with an average of 30.98 mg/g DW. Previous studies indicated a wide variation on total phenolic content among A. unedo genotypes, grown in diverse agro climatic conditions including Spain, Croatia, Algeria and Turkey, The TPC of A. unedo fruits reported in this study was higher than those reported by other authors; Doukani and Tabak (2015) and Isbilir et al. (2012. They recorded a range of 7.02 to 14.74 mg GAE/g and 14.29 mg GAE/g in Algerian and Turkish A. unedo genotypes, respectively. In another study; Seker and Toplu (2010) reported a TPC ranging from 17.7 to 25.8 mg GAE/g). According to these results, and despite natural variations, total phenols content in fruits of A. unedo grown in Morocco fruits was always over 39.06 mg GAE/g DW, indicating that it could be considered as an excellent source of polyphenols content which is of great importance in light of the fact that modern diets are often lacking of bioactive compounds.

Total Flavonoids
The results of the total flavonoids content were summarized in Table 4. A significant variation was observed (p = .002) among genotypes. The total flavonoids content ranged between 3.30 "KSB" and 7.07 mg GAE/g DW "TAH," with an average of 5.20 mg GAE/g DW. These concentrations were higher than those recorded by Pallauf et al. (2008) (0.32 mg/100 g), Bouzid et al. (2014) (2.18-6.54 mg EC/g) and by Jurica et al. (2017) (0.23-0.28 mg EQ/g). These differences could be attributed to the used methods and experimental conditions.

Total Anthocyanins
The total anthocyanins content was reported in Table 4. A statistically significant variation (p = .024) was observed among the genotypes studied. The anthocyanins values varied from 0.15 mg equivalent cya-3-glu/100 g DW "KSB" to 0.64 mg equivalent cya-3-glu/100 g DW "MDZ" with an average of 0.34 mg equivalent cya-3-glu/100 g DW. These values were lower than the ones recorded by Pallauf et al. (2008) (3.77 mg equivalent cya-3-glu/100 g DW).

Antioxidant Activities
The results obtained for antioxidant activity based on the radical scavenging capacity DPPH, ABTS, and βeta carotene were reported in Table 5. Significant differences (p˂0.001) were observed among the genotypes. The average antioxidant activity values were 8.93, 7.82, and 5.58 mg ascorbic acid equivalents/g dry weight as determined by DPPH, ABTS, and βeta carotene assays, respectively. All genotypes presented scavenging effects against DPPH radical ranging from 3.33 to 21.08 mg ascorbic acid equivalent/g DW. The fruits collected from "LAN"presented the lowest IC 50 value, revealing the highest radical scavenging activity among the samples and, therefore, the highest antioxidant activity. These results were higher than those recorded by other authors. They reported that the value of scavenging activity (DPPH) of A. unedo fruit grown in Tunisia was 3.2 mg BHT equivalent/g DW (Ben Salem et al., 2018). Fonseca et al. (2015) reported also, a value of IC 50 ranging from 1.87 to 3.93 mg trolox equivalent/g DW in Portuguese A. unedo fruit. However, the results obtained in this study were lower than the values reported by Barros et al. (2010). They analyzed the antioxidant activity of three wild fruits, and they recorded values of scavenging activity (DPPH) 22.35, 29.85, and 21.4 mg trolox equivalent/g DW for A. unedo, Prunus spinosa and Rosa canina sl. respectively. The antioxidant activity determined by βeta carotene assay ranged between 1.08 and 13 mg ascorbic acid equivalent/g DW. The fruits of genotype "LAN" had significantly the lowest ABTS value, 1.08 mg ascorbic acid equivalent/g DW and, therefore, the highest antioxidant activity. The results obtained in this study were lower than those reported by other authors; Isbilir et al. (2012) analyzed the bleaching activity of βeta carotene. They recorded IC 50 values varied from 9.25 to 15.85 mg/g DW in Turkish fruits. In another study, Barros et al. (2010) analyzed the antioxidant activity through βeta carotene bleaching method of three wild fruits (A. unedo, Prunus spinosa, and Rosa canina sl.) and they recorded values 38.7, 49.3, and 19.8 mg trolox equivalent/g DW, respectively. Free radical scavenging activity of samples was determined by ABTS radical cation decolorization assay (Table 5). The value of ABTS assay ranged between 2.25 and 19.58 mg ascorbic acid equivalent/g DW. The fruits of genotype "LAN" revealed also the lowest ABTS value, 2.25 mg ascorbic acid equivalent/g DW and, therefore, the highest antioxidant activity. The antioxidant capacity of A. unedo fruits determined in this study was higher than the amount presented by Ben Salem et al. (2018) who recorded (5.1 mg trolox/g DW) in Tunisian A. unedo fruits. The A. unedo fruits had strong antioxidant activity for the βeta carotene assay. The different antioxidant levels observed in this study may reflect a relative difference in the ability of antioxidant compounds in extracts to reduce the free radical DPPH, ABTS, and oxidative bleaching of βeta carotene in vitro systems. Antioxidant activity was widely studied on A. unedo fruits by using different antioxidant determining methods such as ABTS, TEAC, FRAP, DPPH, etc. The studies indicated that type of extraction of phenols present in fruits of A. unedo influenced the antioxidant activity (Barros et al., 2010;Fortalezas et al., 2010;Isbilir et al., 2012;Mendes et al., 2011;Morales et al., 2013;Pallauf et al., 2008;Ruiz-Rodriquez et al., 2011;Seker and Toplu, 2010). In addition, several studies reported that A. unedo fruit was found to be a powerful antioxidant plant more than other fruit, such as pomegranate (Gil et al., 2000), red and green grape, and apple juices, (Santini et al., 2014), pomace (Maragò et al., 2015), grape (Schempp et al., 2015;Liu et al., 2018) which can be explained by the higher composition of strawberry, pomegranate, grape, and apple in polyphenols.

Volatile Compounds Characterization
In this study, 25 volatile compounds were identified in A. unedo fruits using HS-SPME method combined to GC-MS analysis. Results of volatile compounds were reported in Table 8. The volatile compounds present in all the genotypes were hexadecanoid acid, tetradecanoid acid, hexadecanoic acid, methyl ester, dodecanoic acid, and phenol. Hexadecanoid acid was the most abundant in  Eigenvalues higher than |0.5| are marked in bold. A. unedo fruits, ranging from 27.68% in the "CHF" genotype to 52.18% in the "MDZ" genotype. Moreover, 9-octadecenoic acid (Z)-was the second most abundant compound, ranging from 1.18% for the "CHF" genotype to 37.60% for the "LAN" genotype followed by tetradecanoid acid which varied from 6.90% in the "LAN" genotype to 18.04% in the "TAH" genotype. Other minor compounds, such as octadecanoic acid, methyl ester, hexadecanoic acid, ethyl ester, 3-dodecene, (E) and benzene (2 Methyl-2-propeny), were also identified and the content was not exceeded 1%. Benzene (2 Methyl-2-propeny) was only presented in the "KSB" genotype and in very low amount (0.46%). Additionally, hexadecanoic acid, ethyl ester, and octadecanoic acid methyl ester were only identified in "TAH" genotype. According to the results of Oliveira et al. (2011), alcohols are the main component of the volatile fraction of Turkish A. unedo fruits and the main volatile compound identified was (Z)-3-hexen-1-ol. This volatile compound was also identified in strawberries and their products (Barron and Etiévant, 1990;Hakala et al., 2001;Hamilton-Kemp et al., 1996).

Correlation Among Variables
In order to identify the relations between biochemical traits, all variables were subjected to bivariate correlation using the Pearson coefficient. Significant correlations at the level of 0.05 or 0.01 are summarized in the Table 9 and 10 In the current study, the correlation value was found between  DPPH and total anthocyanins (r = 0.931; p < .05) as well as between malic acid and titrable acidity (r = 0.763; p < .01). Citric acid was also correlated to titrable acidity (r = 0.522; p < .05), pH (r = 0.751; p < .01) and soluble solids (r = −0.949; p < .01). The results reported also, correlations between DPPH and anthocyanins (r = 0.645; p < .01). Moreover, this study revealed correlations between ABTS and ascorbic acid (r = 0.526; p < .05), anthocyanins (r = 0.748; p < .01), and DPPH (r = 0.883; p < .01). In addition, it conveyed correlations between βeta carotene and ascorbic acid (r = 0.514; p < .05), anthocyanins (r = 0.554; p < .05), DPPH (r = 0.950; p < .01), and ABTS (r = 0.864; p < .01). Also, cyanidin-3-glucoside was correlated to anthocyanins (r = 0.680; p < .01). In this study, no correlation was observed between the antioxidant activity and total phenols. These results must be interpreted with caution as the Folin-Ciocalteu method used over estimates the concentration of phenolic containing compounds, such as ascorbic acids and vitamins, could interfere during total phenols evaluation and that do not give significant correlation. Furthermore, the synergism between the antioxidants in the mixture makes the antioxidant capacity not only dependent on the concentration, but also on the structure and the interaction between the antioxidants. However, different works have reported good linear correlations between antioxidant activity test and total phenols (Anastasiadi et al., 2010;Liu et al., 2008;Serçe et al., 2010;Su and Chien, 2007). The correlation coefficients may provide information on the parameters that are potentially important in assessing A. unedo genotypes (Norman et al., 2011). Significant and strong correlated traits can be used to predict other ones, and could be considered of importance for genotypes characterization and discrimination (Podgornik et al., 2010).

Principal Component Analysis
The aim of this analysis was to identify the main factors to reduce the number of effective parameters to use in classification of the A. unedo genotypes based on their biochemical, antioxidant capacity and volatile compounds. In our study, only a principal component loading of more than | 0.5| was considered as being significant for each factor. Total variance was explained by four components. The first two components was explained 66.47% of the total variability observed ( Table 11). The first component accounted for 39.85% of the total variance, which is strongly influenced by the gallic acid derivative (0.89), chlorogenic acid (0.89), ellagic acid derivative I (0.96), ellagic acid derivative II (0.86), ellagic acid (0.95), rutin (−0.93), cyanidin-3,5-diglucoside (0.87), tetradecanoic acid (0.88), limonene (0.97), and pentadecanoic acid (0.90). The second component accounted for 26.63% of the total variance and is mainly influenced by gallic acid (0.91), gallocatechol (0.91), and 9-octadecenoic acid (Z) (0.88). Generally, these results were in accordance with those reported in previous A. unedo biochemical studies (Colak, 2019;Gündoğdu et al., 2018). They have reported that the biochemical attributes are important in order to evaluate the variation in traits of A. unedo genotypes. Scatter plot was prepared according to the first two principal components: PC1 and PC2 (respectively, 39.85 and 26.63% of total variance) that discriminate between the genotypes according to their volatile compounds and biochemical characteristics (Figure 1 and 3). Starting from negative to positive values of PC1, the distribution of genotypes indicated an increased in the succinic acid and the most of phenolic compounds. Whereas, starting from negative to positive values of PC2, total soluble solids, malic acid and the most of volatile compounds decreased in their values. However, the distribution of genotypes indicated an increase in the titratable acidity, pH and citric acid. Our results are in agreement with several studies (Colak, 2019;Gündoğdu et al., 2018).

Conclusion
This study revealed that A. unedo fruits can be considered an important source of polyphenols (25-39 mg GAE/g DW). Among the 17 phenolic compounds identified by HPLC, gallocatechol and catechin were the most abundant compounds. Moreover, four organic acids were identified in A. unedo fruits which citric acid was the most dominant. Results showed also that hexadecanoic acid was the most abundant volatile compound in all the studied genotypes. According to results obtained in this study, A. unedo fruits are strong radical scavengers that can be considered as good sources of natural antioxidants, the fact that may encourage their daily intakes as an alternative source of bioactive compounds in the local population diet. In view of its biochemical composition, the use of A. unedo fruits in some food and medicinal products may be also suggested. This study contributes not only to a better knowledge of these wild fruits but also to their valorization.