L., known as the “apricot”, belongs to the genus Prunus
of the sub-family Prunoideae and the family Rosaceae [1
]. It is native to China, and was later introduced around the Mediterranean basin [2
]. Apricot is one of the oldest known oil seed crops, and it plays an important role in the health and vitality of humans. Oil extracts from the kernel of the plant P. armeniaca
L. have shown a remarkable pharmacological effect, including high free radical scavenging capacity (antioxidant) [3
]; inhibitory activity against several enzymes in an tumor development experiment [5
]; and antinociceptive [6
], antimicrobial [4
], anticancer [7
], anti-inflammatory [8
], hepato-protective [9
], and cardioprotective activities [10
]. The large amount health benefits of P. armeniaca
L. kernel begs the investigation of its chemical composition, thus leading to the identification of polyphenols, lipids, carotenoids, organic acids, amygdalin, and mineral elements. In particular, different classes of polyphenols, including phenolic acids, flavonoids, and antocyanins, have been positively identified.
This seed oil, as in the case of other vegetable oils, is mainly constituted of lipids, including a high proportion of triacylglycerols (TAGs) [11
]. TAGs, the most abundant lipids in nature, are triesters of fatty acids (FAs) with glycerol. In particular, each of the three positions of glycerol may be occupied by different FAs. The sum of all possible combinations of FAs makes the oil a particularly complex mixture of TAGs, which asks for advanced analytical techniques for a detailed elucidation. Until now, several studies have investigated the FA profile of P. armeniaca
L. kernel oil [12
] after a trans-esterification procedure [12
], wherein only two of them reported the native TAG composition [14
], as they are effectively assumed by humans. The most abundant identified FAs were oleic and linoleic acids, followed by palmitic and stearic acids, and then the major TAGs derived from the combination of these FAs, such as triolein, dioleyl-linolein, dioleyl-palmitin, and dilinoleyl-olein. Moreover, only few studies on the characterization of chemical composition of P. armeniaca
L. cultivated in Tunisia were published. The fruit of Tunisian P. armeniaca
L. has been studied as a source of carotenoid compounds [21
], with the kernel flour being recommended as a protein source [22
]; however, studies on the chemical composition of Tunisian P. armeniaca
L. kernel extract have not been reported until now.
The extraction of bioactive compounds from plant material have been increasingly undertaken in the last decade in order to better understand their beneficial properties [23
]. Typically, bioactive compounds, such as carotenoids, polyphenols, and lipids are extracted by using a mixture of two or three solvents: polyphenolic compounds are commonly extracted through the well-known Montedoro method, employing methanol/water [24
]; carotenoids are isolated by using more apolar solvents, such as hexane, ethyl ether, ethyl acetate, and acetone [25
]; while lipids are commonly obtained by the well-known Folch method [26
], which recommends chloroform/methanol as an extraction mixture. Specifically, for lipids and carotenoids, the use of different polarity solvents is mandatory to maximize the recovery for both polar (phospholipids and xantophills) and apolar compounds (TAGs and carotenes).
In the present research, we used successively four different pure solvents (dichloromethane, chloroform, ethyl acetate, and ethanol) in order to obtain four extracts with different chemical composition; specifically, ethanol extract was expected to be the most concentrated in polyphenols, immediately followed by ethyl acetate extract, while the dichloromethane and chloroform extracts were expected to contain almost solely apolar compounds. High-performance liquid chromatography (HPLC) coupled with mass spectrometry (MS) was used to elucidate the chemical composition of the four extracts, allowing for both qualitative and quantitative considerations. The reversed phase (RP) separation mechanism was selected for the analysis of both polar (polyphenols) and apolar compounds (lipids and carotenoids). In particular, polyphenols were separated on RP-HPLC by using a previously developed chromatographic method [27
] in gradient elution with acidified water and acetonitrile as mobile phases, while a new chromatographic approach was investigated for lipid separation. In fact, a RP-HPLC method, based on the use of a C30 stationary phase, commonly used for carotenoid analysis [28
], was applied, leading to the separation and identification of only acylglycerol compounds, whereas carotenoids were definitely not detected. Focusing on lipid separation, they are normally eluted according to the increasing partition number (PN) or equivalent carbon number (ECN), related to carbon (CN) and double bond (DB) numbers according to the following relationship: PN = ECN = CN − 2DB. Taking into account the high complexity of a TAG mixture, several co-elutions normally occur in the same PN region. Within this context, the additional aim of the present work was to investigate the retention behavior of a C30 column for TAG separation. Interestingly, the more hydrophobic nature of this column, compared to a more conventional C18, positively affected the chromatographic resolution as an effect of the increased retention, especially for low-PN compounds and positional isomers, similarly to the good resolution achieved for carotene isomers in previous works [28
In addition, gas chromatography (GC) coupled to both mass spectrometry (MS) and flame ionization detector (FID) was used for the qualitative and quantitative determination of FAs obtained after transesterification of intact lipids. The quali-quantitative FA profile was helpful to support the identification of TAGs, which most likely contained the most abundant FAs (at least > 0.1% of the total content).
2. Materials and Methods
2.1. Chemicals and Reagents
Reagent grade quality ethanol, chloroform, ethyl acetate, dichloromethane, methanol, n-heptane, sodium methoxylated, and boron trifluoride in methanol (14% w/v) were purchased from Merck Life Science (Merck KGaA, Darmstadt, Germany).
LC–MS grade methanol, acetonitrile acetic acid, water, and HPLC grade methyl tert-butyl ether were also acquired from Merck Life Science (Merck KGaA, Darmstadt, Germany).
Standard of gallic acid, protocatechuic acid, coumarin, chlorogenic acid, catechin, epicatechin, and ferulic acid were purchase form Merck Life Science (Merck KGaA, Darmstadt, Germany).
2.2. Plant Seed Materials
Apricot kernels were purchased from a local market in Kondar (latitude 35°49′34″ N, longitude 10°38′24″ E), a rural region in the Tunisian Sahel, situated about 30 km from the northwest of Sousse governorate. They were milled using an electric grinder (Moulinex AR1100, France) and sieved using sieves with pore sizes of 710 μm. The powder was stored in sealed plastic bags at 4 °C until used.
2.3. Oil Extraction
Prunus armeniaca L. kernels were subjected to subsequent reflux extractions with 300 mL of four different solvents. One hundred grams of seed powder were mixed with 300 mL of dichloromethane; after 90 min under reflux, the solid particles were filtered using a filter paper, thus obtaining the residue I and dichloromethane extract. From the filtered extract, dichloromethane was evaporated, yielding a pure oil extract. Residue I was subjected to the extraction with chloroform solvent to obtain residue II and the chloroformic extract. The latter was evaporated to obtain the pure oil, while residue II was extracted with ethyl acetate solvent to obtain residue III and the ethyl acetate extract oil. Finally, residue III was mixed with 300 mL of ethanol, and the ethanol extract was obtained and evaporated to dryness. Extracts were stored at 4 °C until used. Extractions were performed in triplicate.
2.4. Fatty Acid Methyl Ester (FAME) Preparation
Twenty milligrams of each kernel extract were added to 500 µL of sodium methoxylated in methanol (0.5% w/v) and mixed for 120 s by using a digital shaker (IKA-Werke GmbH and Co. KG, Staufen, Germany) at 2000 rpm. The solution was heated for 15 min at 95 °C. Then, 500 µL of boron trifluoride diluted in methanol (14% w/v) was added to the reaction mixture, which was shaken for 120 s at 2000 rpm and heated for 15 min at 95 °C. Afterwards, 350 µL of n-heptane and 250 µL of saturated NaCl solution were added to the mixture; after 120 s of vortex mixing and 5 min of incubation, the upper heptanic phase was injected into the GC systems.
2.5. Sample Preparation for TAG Analysis
Thirty milligrams of each kernel extract were dissolved in 1 mL of a methanol/methyl tert-butyl ether (v/v) solution. The resulting solutions were filtered through a 0.45 μm Acrodisc nylon membrane (Merck Life Science, Merck KGaA, Darmstadt, Germany), prior of the injection into the HPLC- MS system via atmospheric pressure chemical ionization (APCI) interface.
2.6. Sample Preparation for Polyphenol Analysis
Dichloromethane and chloroform extracts were dissolved in chloroform (10 mg mL−1), while ethyl acetate and ethanol extracts were dissolved in methanol (10 mg mL−1). The resulting solutions were filtered through a 0.45 μm Acrodisc nylon membrane (Merck Life Science, Merck KGaA, Darmstadt, Germany) prior to the HPLC–MS analysis for the determination of polyphenolic compounds.
2.7. GC–MS Analysis of FAMEs
GC–MS analyses were carried out on a GCMS-QP2010 (Shimadzu, Duisburg, Germany) equipped with a split/splitless injector and an AOC-20i autosampler. The chromatographic column was a SLB-Il60i (30 m × 0.25 mm id, 0.20 μm film thickness) column (Merck Life Science). The temperature program was as follows: 50 °C to 280 °C at 3.0 °C/min. Injector was kept at 280 °C; injection volume was 0.2 µL with a split ratio of 1:20. Helium was used as a carrier gas at 30 cm/s linear velocity and an initial inlet pressure of 31.7 KPa (50 °C). MS parameters were as follows: mass range 40–550 amu, with an event time of 0.20 s; ion source temperature 200 °C, interface temperature 220 °C; ionization mode EI (70 eV), detector voltage 0.98 kV. The GCMS solution software (Shimadzu) was used for data collection and handling. The C4–C24 FAMEs standard solution was used for linear retention indices (LRIs) calculation to support identification of analytes. Moreover, peaks assignment was carried out on the basis of a double filter, namely, the MS similarity spectra (over 80%) and a LRIs ±10 range compared to the value reported in the commercial database used (LIPIDS Mass Spectral Library (Shimadzu)).
2.8. GC–FID Analysis of FAMEs
GC–FID analyses were carried out on a GC-2010 (Shimadzu) equipped with a split/splitless injector (280 °C), an AOC-20i autosampler, and an FID detector. GC column, temperature program, and carrier gas were the same as previously described for GC–MS analyses, apart from the inlet initial pressure of 99.5 kPa (constant linear velocity equal to 30 cm/s). The FID temperature was set at 280 °C, and gas flows were 40 mL/min for hydrogen and make-up gas (nitrogen) and 400 mL/min for air. Data were collected by using LabSolution software (Shimadzu). A relative quantification was also carried out. Analyses were performed in triplicate.
2.9. HPLC–APCI/MS Analysis of Lipid
HPLC–MS analyses were carried out by using a Nexera X2 system (Shimadzu, Kyoto, Japan) coupled to an LCMS-2020 detector equipped with an APCI source. The Nexera X2 system consists of a CBM-20A controller, two LC-30AD dual-plunger parallel-flow pumps (120.0 MPa maximum pressure), a DGU-20A5R degasser, a CTO-20AC column oven, and a SIL-30AC autosampler.
Separations were carried out on a C30 column (250 mm length × 4.6 mm inner diameter, 5 µm particle size) provided by YMC Europe (Schermbeck, Germany). Mobile phases were (A) methanol/methyl tert-butyl ether/water (81:15:4 v:v:v) and (B) methanol/methyl tert-butyl ether/water (15:81:4 v:v:v) under the following gradient program: 0–20 min, 0% B, 20–110 min, 0–100% B. The flow rate was set at 800 µL/min with oven temperature of 35 °C; injection volume was 20 µL.
MS detection was performed in full scan mode and in positive polarity with the following APCI parameters: interface temperature, 350 °C; DL (desolvation line) temperature, 300 °C; heat block temperature, 300 °C; and nebulizing gas (N2) and drying gas (N2) flows were 1.5 and 5 L/min, respectively. The range of acquisition was 200–1200 m/z, with an event time of 2 s. Data were collected by using LabSolution software (Shimadzu). A semi-quantification on the basis of peak area percentages was also carried out. Analyses were performed in triplicate.
2.10. HPLC–PDA–ESI/MS Analysis of Polyphenols
Analyses were carried out on a Shimadzu Prominence LC-20A system (Shimadzu, Kyoto, Italy), including a CBM-20A controller, two LC-20 AD dual-plunger parallel flow pumps, and a DGU-20A3 on-line degasser. The LC system was coupled to a photodiode array (PDA) serially connected to an LC–MS 2020 mass spectrometer by an electrospray (ESI) interface (Shimadzu, Milan, Italy). HPLC separation was performed on an Ascentis Express RP C18 column (2.7 µm, 150 mm, and 4.6 mm; Merck Life Science, Merck KGaA, Darmstadt, Germany). The mobile phase consisted of water/acetic acid (99.85/0.15 v/v, solvent A) and acetonitrile (solvent B), under the following gradient elution program: 0–5 min, 5% B; 15 min, 10% B; 30 min, 20% B; 60 min, 50% B; 70 min, 100% B. LC flow rate was 1 mL min−1 and injection volume was 10 μL. PDA detector was applied in the range of λ = 200–400 nm, and the polyphenol chromatograms were extracted at λ = 280 nm (sampling frequency: 40 Hz, time constant: 0.08 s). MS analysis was performed in negative and positive mode in the mass range m/z 100–800 with an event time of 0.3 sec; nebulizing gas (N2) and drying gas (N2) flow rate were 1.5 L min−1 and 15 L min−1, respectively; interface temperature was 350 °C; heat block and DL (desolvation line) temperatures were 300 °C. Data were collected by using LabSolution software (Shimadzu, Kyoto, Japan).
2.11. Statistical Analyses
In order to evaluate variability within the different assays, we applied descriptive statistic to our outcomes, and the software IBM SPSS Statistics (version 20.0) was used to calculate the means, confidence intervals (CI 95%), and standard deviations (SD).
Assuredly, the hot extraction method using four organic solvent of increasing polarity represented a high efficiency extraction method. A total yield of about 20% (w
) of oil per gram of kernel was obtained, comparable or even more so than that reported in literature studies from Iran and Turkey [12
]. The present work represents the first report on Tunisian oil extracts from P. armeniaca
L. kernel. Extraction by solvents of different polarity allowed us to obtain different extracts containing different polarity chemical constituents. Specifically, polar solvents, such as ethanol and ethyl acetate solvent, were efficient for the extraction of lower molecular weight components, compared with the non-polar solvents, such as dichloromethane and chloroform.
As for apolar compounds, the use of a very retentive column, such as the one based on a C30 stationary phase, appeared to be the most appropriate choice to detect and satisfactorily separate both acylglycerols and eventually prenol lipids, such as carotenoids and carotenoid esters. In this specific case, we did not find prenol lipids.
The GC–FID/MS analyses of FAMEs revealed intense peaks in the chromatograms of all the samples, independently from the employed solvent. This finding is related to the wide range of lipid classes, from polar phospholipids and free FAs, soluble in polar and medium-polarity solvents, compared to apolar sterols and acylglycerols, more soluble in apolar and medium-polarity solvents, strongly depending on the unsaturation degree of the FA bound to the glycerol backbone. As a consequence, through looking at the absolute area values, rather than their percentages, the dichloromethane, ethanol, and ethyl acetate extracts were found to be quite similar, while the chloroform extract was the poorest sample. Such a finding is partially in contrast with the acylglycerol results that showed the different trend of ethyl acetate > chloroform > dichloromethane >> ethanol, which could indicate the presence of a high amount of polar lipids, such as phospholipids, which could be the object of further investigation. To this regard, a recent study regarded the determination of glycerol-phospholipids in three North African apricot (P. armeniaca
L.) seed varieties, whose oils were extracted by the common Bligh and Dyer procedure [41
]. Generally, phospholipids and glycolipids represent a minor fraction in vegetable oils commonly obtained by different extraction methods. Hence, in the current study, the analysis of the four different oils could provide unexpected results about these lipid families.
The diversity in composition of FAs is a good indicator of the stability and quality of oils. In our study, 15 different FAMEs were detected in all extracts. Conversely, a minor number of FAMEs, from 5 to 10, were identified in previous reports [10
], which normally limit their attention to C16 and C18 FA families. The monounsaturated C20:1 FA was previously detected only by Amiran et al. [12
], who performed a soxhlet extraction with hexane; on the other hand, they did not identify the ω-3 linolenic acid. The latter was reported at trace level in other works [17
], with the exception of Orhan et al. [10
], who reported a relative concentration around 10% of the total FA composition in some Turkish apricot oils obtained by maceration with hexane. In the present work, 5% of linolenic acid was found in the ethanol kernel extract, confirming the importance of investigating different extraction solvents to produce oils with different chemical compositions. Moreover, the C17 FAs were not reported previously; specifically, the monounsaturated C17:1 that contributed to the total composition in the same percentage than C20:1. Altogether, the dichloromethane extract was the most similar to previous reports, where oleic acid C18:1 was contained at a percentage larger than 60%, and linoleic acid C18:2 represented more than 20%, while the saturated palmitic acid was present at low content (around 5%). Such results were quite similar to those reported in previous works dealing with the determination of FAs in seed oil, obtained from apricots cultivated in Turkey [10
], India [18
], and Moorpark [17
], thus concluding that the geographical origin has only a minimal influence on the lipid composition. Indeed, the levels of oleic and linoleic acids became more and more similar by increasing the extraction solvent polarity (both between 40% and 50% in ethyl acetate extract) up to the ethanol extract, in which, as already pinpointed, a significant amount of linolenic acid appeared. These results corroborate different literature reports about the beneficial properties of apricot seed oil [10
]. Oleic acid, the major fatty acid identified in vegetable oils produced in the in Mediterranean countries, and at the basis of the Mediterranean diet, e.g., olive oil, presents different medicinal properties, such as reduction of inflammation and blood pressure, inhibition of cancer proliferation, and enhancement of fungicidal and bactericidal actions, while moreover exerting a prominent role in drug absorption [42
]. Linoleic acid (omega-6) is an essential FA that prevents cancer and cardiovascular diseases since it is the precursor of important signaling molecules [12
], and it also produces a serum cholesterol reduction [43
]. However, an optimal ratio of ω-6/ω-3 FAs should always be maintained to guarantee a proper healthy status. In this regard, the role of linolenic acid as an essential ω-3 FA involved in many metabolic pathways should be taken into account [44
The way in which all these FAs are combined in DAG and TAG structures also plays a central function in terms of oil stability [45
] and biological activities, and lipolytic enzymes will have a key role to increase the FA availability in tissues [46
In our extracts, 24 acylglycerols were identified, and OOL, OLL, OOO, and LLL were the major TAGs, in contrast to previous reports [14
], which found OOO and OOL as the major TAGs and quantified OLL and LLL at levels less than 15% and 3%, respectively. In the present work, OLL largely overcame 20% of the total non-polar lipid composition and reached 33% in the ethanol extract; in a similar way, LLL was near 10% of the total fraction and increased to 13% in the ethanol residue. These differences could be explained by the different extraction conditions of the oil from apricot kernel. In fact, in previous works, a soxhlet extraction by petroleum ether [14
] and the conventional extraction mixture chloroform/methanol [15
] were employed. In fact, some differences could be related to the differences in geographical origin of the samples under investigation. Hence, future perspectives could regard the comparison between apricot kernels of different provenances or cultivars, or the application of a conventional extraction method on the same sample. It should be specified that the geographical origin could have a negligible effect on the total FA composition, while it could significantly affect the way in which FAs are combined in glycerol-phospholipids and acylglycerols, due to a different activity of specific enzymes.
As for polar compounds, 11 polyphenols consisting of 5 phenolic acids; amygdalin; coumarin; and 4 flavonoids, including 2 flavan-3-ols (catechin and epicatechin), the isoflavonoid acetylgenistin, and the flavone dimethoxylflavone, were reliably identified in this study, thanks to the combined use of PDA and MS detection. Our results showed differences compared to literature data from China [39
], relative to apricot kernels of the same botanical species (Prunus armeniaca L.
), in which amygdalin and a chlorogenic acid derivative were the major compounds, but flavanols, isoflavonoids, coumarins, and flavones were not detected. This could be related to the different extraction technologies that were based on the use of microwaves, which indeed were more efficient in the extraction of antocyanins and tannins. In another study on the characterization of kernel microwave extracts of a different Prunus
species (Prunus sibirica
L.), antocyanins again resulted in being the major compounds, but phenolic acids were also present at significant levels [37
]. Some compounds detected at a high amount in the fruit of the same species cultivated in Croatia [47
], such as catechin and epicatechin, were also detected at low levels in the seed extract of the present Tunisian cultivar, while coumarin has previously been reported in the kernel extract of a different Prunus
species (Prunus mahaleb
L.) cultivated in Italy [38
]. All these data suggest that, as in the case of lipids, the extraction procedure plays an important role in the obtained polyphenol composition, in which, as it is already well-known, they are minor components whose biosynthesis is strongly affected by pedoclimatic factors. Hence, the comparison between different environments, extraction conditions, and apricot varieties could represent the future perspective of the present study that reported for the first time the polyphenol composition of Tunisian apricot seeds.