Toluene, acetonitrile, methanol, dichloromethane, and hexane were supplied by Carlo Erba (Val de Reuil, France). Petroleum ether was obtained from Fisher Scientific SA (Loughborough, UK). Fatty acid methyl ester (FAME) standards were acquired from Nu-Chek-Prep (Elysian, MN, USA). Tocopherol standards, ammonium acetate, sodium bicarbonate, and 7% boron trifluoride in methanol (BF₃-MeOH) were provided by Sigma-Aldrich (Saint-Quentin Fallavier, France). Lutein and Zeaxanthin standards were generous gifts from Roche (Neuilly-sur-Seine, France).

Pecan nuts from Moore and Mahan cultivars were harvested at complete maturity from a restricted zone on the INRGREF experimental farm (National Institute for Research in rural engineering, water and forest), in Mateur, Northern Tunisia (Longitude: 09°48′ E; Latitude: 37°15′ N; Altitude: 5 m above sea level).

Husks were removed And the kernels were detached manually from the shell and dried in an oven at 60 °C. Using a mortar, the dry samples were finely ground for the extraction process.

Oil was extracted from dry milled pecan kernels (5 g) using a soxhlet apparatus with light petroleum ether for 6 h at 42 °C. The solvent was removed from the resulting extract by rotary vacuum evaporation at 50 °C. The obtained oil was stored in dark glass bottles at 4 °C until analysis.

The fatty acid composition was determined as fatty acid methyl esters (FAME) prepared as described previously by Bouali et al. (2013). Twenty milligrams of oil were mixed with 0.7 mL boron trifluoride reagent (BF3/methanol, 7 %) and 0.3 mL toluene under nitrogen and the tube was heated at 95 °C for 2 h and then cooled to room temperature. FAMEs were extracted by adding 5 mL of sodium bicarbonate and 3 x 2 mL hexane. The obtained mixture was centrifuged for 3 min. The solvent was removed under a nitrogen stream. Then, the FAMEs were dissolved in hexane and analyzed on a Hewlett Packard Model 4890 gas-chromatograph (Palo Alto, CA, USA) with a CP-Sil 88 (100 m × 0.25 mm I.D. capillary column, 0.2 µm film, Varian, Les Ulis France). The injector and detector temperatures were 250 and 280 °C, respectively. The initial oven temperature was held at 60 °C for 5 min, increased to 165 °C at 15 °C/min and held for 1 min at this temperature; then increased to 225 °C at 2 °C/min, before being maintained at this temperature for 17 min. The peaks were identified based on their retention times using synthetic and commercial standards. Analyses were performed in triplicate.

The TAG analysis was performed by gas chromatography (GC) using the same chromatographic conditions as reported by Bootello et al. (2016). About 5 mg of pecan kernel oil were dissolved in 1.8 mL of heptane. One μL of this solution was injected into the GC, an Agilent 7890 gas chromatograph (Palo Alto, CA) equipped with a flame ionization detector (FID) and a Quadrex Aluminum-Clad 400-65HT chromatography column (30 m length; 0.25 mm diameter; 0.20 µm film thickness: Bellefonte, PA, USA). Hydrogen was used as the carrier gas at a linear gas rate of 50 cm·s^-1 and a split ratio of 1:50. The detector and injector temperatures were 370 and 350 °C, respectively; the oven temperature was 335 °C, and a pressure gradient from 100 to 200 kPa was applied. Peak identification was performed by comparing their retention times with those of a standard mixture and known samples. Quantification of the peaks was made by internal normalization of the chromatographic peak area, and the results are expressed in relative percentage.

The tocopherol and xanthophylls analysis was performed by reverse-phase high performance liquid chromatography (RP-HPLC) using the procedure previously described by Bouali et al. (2013). Samples (40 mg) were diluted in 1 mL of mixed HPLC mobile phase: acetonitrile/methanol containing 50 mM ammonium acetate/water/dichloromethane (700:150:50:100, V/V/V/V). Then, the mixture was vortexed.

The assays were carried out on a Jasco PU-1580 apparatus plus intelligent pump equipped with an automatic injector system AS300 (Thermo Finnigan, Les Ulis, France) and a Jasco MD-1510 plus multi wavelength detector (JASCO International Co., Ltd., Japan). Chromatographic separation was performed on a VIDAK C18 column (25 mm x 4.6 mm id, 5 μm particle size) connected with a Nucleosil C18 column (25 mm x 4.6 mm id, 5 μm particle size). The mobile phase was a mixture of acetonitrile/ methanol at 50 mM ammonium acetate/ water/ dichloromethane (700:150:50:100, V/V/V/V) at a flow rate of 2 mL/min. Samples of 18 µL were injected, the wavelengths were set at 298 nm for tocopherols and 450 nm for lutein and zeaxanthin. The identification was carried out by comparing the retention times and absorption spectra of unknown peaks to authentic standards and by adding α-, δ-, γ-tocopherol, lutein, and zeaxanthin standards to the sample for co-chromatography. The quantification was based on the external standard method.

Thermal properties of pecan kernel oils were carried out on a diferential scanning calorimetry (DSC) using a Q2000 V23.5 scanner (TA instruments, New Castle, DE, USA). Dry nitrogen was used as the purge gas at a rate of 20 mL/min. Oil samples (6-8 mg) were weighed directly into aluminium pans. An empty sealed pan was employed as control.

Samples were subjected to the following temperature program: 90 °C isotherm for 5 min, and then cooled to -80 °C at a rate of 10 °C/min, held for 20 min. Oil samples were then heated from -80 °C to 90 °C at a rate of 5 °C/min to acquire the data of the melting thermogram. The crystallization was achieved by heating the oils at 90 °C for 5 min and then decreasing the temperature to -80 °C at a rate of 10 °C/min.

Crystallization and melting parameters: the melting enthalpy (ΔH_m), the temperature of the major peak of melting (P_m), the initial and end temperature of the melting phase (T onset _m and T end _m), the crystallization enthalpy (ΔH_c), the temperature of the major peak of crystallization (P_c), the initial temperature of the crystallization phase (T onset _c), the range of the transition phase (R: the temperature difference between T onset and T end) were determined.

The results were expressed as mean values and standard deviation and were analyzed using the statistical analysis system XLSTAT (version 2015). The differences between cultivars were analyzed using one-way ANOVA followed by Duncan’s test (P < 0.05), and principal component analysis (PCA).

Oil contents, expressed in percentage of dry weight, are reported in Table 1. It can be deduced that the amount of lipids differs among pecan nut cultivars as has been observed previously (Ribeiro et al., 2020). Moore kernels had lipid content significantly higher (72.57 %) than those of Mahan (69.64 %). The kernels investigated in this study exhibited a substantial fount of lipids for food with more than 69% of this macronutrient (Table 1). In the present study, the oil content of both cultivars was significantly higher than that of the Burkett cultivar (68.42% dry weight) as previously studied (Bouali et al., 2013).

These values are similar to those published by Miraliakbari and Shahidi, (2008) but higher than those reported by do Prado et al. (2013) and Ribeiro et al. (2020). Differences in lipid content in pecans could be attributed to cultivar, geographic factors, ripening stage, and variations in extraction methods (Miraliakbari and Shahidi, 2008; Bouali et al., 2013; Ribeiro et al., 2020). Most nuts contain high amounts of fat (almond 60.10%, pistachio nut 22.09%, and walnut 32.95%) (Rabadán et al., 2019).

Table 1 shows the fatty acid composition of Mahan and Moore cultivars. Fourteen fatty acids were detected. Oleic acid was the predominant one (55.16-56.81% of total fatty acids). Linoleic acid was the second in order of importance (31.60-32.73%). Among the remaining fatty acids only palmitic (6.20-6.37%) and stearic (2.27-2.61%) acids showed considerable amounts. The major fatty acids found in the present study are, in general, comparable to those reported in the literature (do Prado et al., 2013; Ribeiro et al., 2020).

In the current study, the fatty acid composition of both cultivars was significantly different from the Burkett cultivar previously investigated, which showed a particular composition with very high oleic acid (69.04%) and low linoleic acid (20.69%) contents (Bouali et al., 2013). This variation could be attributed to genetic differences.

In pecan kernel oil, saturated fatty acids (SFA) were the minor group (8.75-9.28% of total fatty acids): whereas monounsaturated fatty acids (MUFA) were the highest (56.55 - 58.30%) (Table 1). It is important to mention that the pecan nut oil is exceeded in monounsaturated fatty acids level by olive oil (58.07-71.13%) (Ouni et al., 2011) but followed by other nut oils like pistachio nut (54.28%), cashew (57.54%) and walnut (19.16%) (Al Juhaimi et al., 2018). Polyunsaturated fatty acids (PUFA) were present at high levels (31.63 - 34.17% of total fatty acids).

Overall, the obtained results (MUFA > PUFA > SFA) are similar to those observed in Brazilian pecan nuts (Ribeiro et al., 2020) and other geographical origins such as Mexico (Rábago-Panduro et al., 2020).

In a previous study, Miraliakbari and Shahidi (2008) reported that pecan nut oil contains 96.4% of triacylglycerols. Therefore, it was interesting to investigate the TAG compositions among pecan nut cultivars. The TAGs were identified and quantified by GC. Significant differences between the two cultivars were detected. As shown in Table 2, thirteen compounds were identified and quantified: POP, PLP, POS, POO, PLS, POL, PLL, SOO, OOO, SOL, OOL, OLL, LLL (where P, palmitoyl; S, stearoyl; O, oleoyl; and L; linoleoyl). The predominant TAG was OOL (28.05 - 28.54%) in both cultivars, followed by OOO (25.96%) and OLL (15.53%) in the Moore cultivar. In the Mahan cultivar, OLL (20.62%) was the second major TAG and OOO was the third (17.77%). The remaining TAGs had smaller percentages (Table 2). The amount of POS was the lowest in both cultivars. The variation, observed among cultivars, may be attributed to genetic causes, mostly acyltransferase activities during the TAG assembly. Several authors have emphasized the effect of genotype on TAG composition in other nuts (Bouabdallah et al., 2014; Taş and Gökmen, 2015). The TAG composition is considered the most effective parameter for distinguishing cultivars (Bouabdallah et al., 2014), which is confirmed by our findings.

Our results are different from those reported by Fernandes et al. (2017) for pecan nuts obtained from local grocery stores in Brazil. These same authors reported the predominance of PLP + OOO + PoPP. This difference could be attributed to cultivar and geographical variations, ripeness stages, storage conditions and differences in extraction methods and analytical procedures.

For walnut (Juglans regia), a member of the same family as the pecan nut (the Juglandaceae family), 19 TAG compounds were detected in this oil, with LLL as the major one (Bouabdallah et al., 2014).

The amount of tocopherols is a fundamental parameter for assessing the quality of oils. The tocopherol composition is reported in Table 3. γ-Tocopherol was predominant in both analyzed oils, representing about 98% of total tocopherols. The highest amount was observed in the Moore cultivar (259.82 mg/kg of oil); whereas the lowest was detected in the Mahan cultivar (197.30 mg/kg of oil). γ-Tocopherol is considered the most frequent isoform of vitamin E in several nuts (Kornsteiner et al., 2006). α-Tocopherol was the second predominant isoform, accounting for 2.36% and 2.29% of the total tocopherols in Moore and Mahan cultivars, respectively. Several investigations have revealed that γ-tocopherol possesses beneficial activities which are not common to α-tocopherol. γ-Tocopherol has, notably, stronger anti-cancer, anti-inflammatory, antioxidant, cardioprotective properties (Smolarek and Suh, 2011; Mathur et al., 2015), greater detoxification of nitrogen oxides and higher inhibition of cyclooxygenase-2 (COX-2) (Mathur et al., 2015) than α-tocopherol. In addition, γ-tocopherol, but not α-tocopherol, exerts an anti-proliferative effect on lung and prostate cancer cells (Jiang et al., 2004).

Besides γ- and α-tocopherols, δ-tocopherol was also present in pecan kernel oil, with low amounts representing 0.3 and 0.25 % of the total tocopherols in Moore and Mahan cultivars, respectively.

These results indicated that pecan kernel oil provides natural antioxidants, with the Moore cultivar exhibiting the greatest tocopherol content (266.93 mg/kg of oil). In this study, the total tocopherol content of both cultivars was significantly higher than that of the Burkett cultivar previously investigated (Bouali et al., 2013). Similar results have been found by other authors (do Prado et al., 2013) implying that tocopherol content is highly cultivar-dependent.

The total tocopherol contents found in this study were higher than those reported by Kornsteiner et al., (2006) and Fernandes et al,. (2017). In contrast, they were lower than those reported by Miraliakbari and Shahidi (2008). It is well-known that tocopherol content is strongly influenced by genetic differences, maturity (Bouali et al., 2013), geographic and climatic conditions (do Prado et al., 2013), as well as differences in extraction methods and analytical procedures (Miraliakbari and Shahidi, 2008).

Miraliakbari and Shahidi (2008) studied the tocopherol composition of oils extracted from several nuts by different solvents and found that the tocopherol content in pecan kernel oil ranges from 453.9 to 491.1 mg/kg oil, exceeded by the total tocopherol content in hazelnut oil. It has been pointed out that pecan kernel oil contains the greatest amount of γ-tocopherol among tree nut oils (Miraliakbari and Shahidi, 2008).

Both cultivars were also investigated for their levels of lutein and zeaxanthin. Among xanthophylls, lutein and zeaxanthin prevent the risk of several diseases (Ma and Lin, 2010). Thus, it is beneficial to assess the levels of these compounds in pecan nuts in order to select the cultivar which contains the greatest amount of xanthophylls.

The amounts of lutein and zeaxanthin as well as their sum are reported in Table 3. Concerning total xanthophylls, a similarity was found between the cultivars. In the present study, the total xanthophylls contents in both cultivars (0.22-0.23 mg/kg of oil) was significantly higher than that of the Burkett cultivar (0.14 mg/kg of oil) already studied (Bouali et al., 2013). It is important to highlight that Carya illinoinensis kernel oil is not a substantial source of dietary xanthophylls.

In pecan kernel oil, lutein was the predominant xanthophyll. Our findings are in disagreement with the previous study of Kornsteiner et al., (2006) who did not detect any lutein. The same authors revealed that pecan nut oil was surpassed in lutein level by pistachio nut oil (15-96 mg/kg of oil) (Kornsteiner et al., 2006). Indeed, pistachio nut oil contains the highest level of lutein among all nut oils; whereas almond, Brazil nut, cashew, hazelnut, macadamia and walnut oils do not contain any lutein (Kornsteiner et al., 2006). Nonetheless, zeaxanthin was not detected in either cultivar of Carya illinoinensis.

The thermal parameters supplied by the heating and cooling thermograms are given in Table 4. The melting thermograms of both cultivars displayed profiles with two not well-separated endothermic phase transitions. As shown in Table 4, the temperature of the first major peak of melting (P_m) was -29.2 ºC for Mahan cultivar and -27.4 ºC for Moore cultivar with an onset temperature (T onset _m) ranging between -41.7 °C (Mahan) and -37.5 °C (Moore) and an end temperature (T end_m) varying between -26.3 and -24.7 °C for Mahan and Moore, respectively. The temperature of the second P_m was -21.2 ºC for Mahan and -19.4 ºC for Moore with a T onset _m ranging between -26.3 °C (Mahan) and -24.7 °C (Moore) and a T end _m varying between -7.6 °C (Mahan) and -4.5 °C (Moore). Mahan showed the lowest P_m. The non-significant difference in P_m values is attributed to the fact that the Mahan cultivar presents more unsaturated TAGs and fatty acids, and consequently, this leads to the appearance of a lower melting point (Mansor et al., 2012) than that of the Moore cultivar. As explained in the literature, the first peak detected in both samples at a low-temperature region of the heating curve might be linked to the melting of the lowest stability polymorphic form of TAGs (Rezig et al., 2012). The range of transition phase (R) varied between 33 and 34.1 °C for Moore and Mahan, respectively. The melting enthalpy (ΔH_m) varied significantly among cultivars. The Moore cultivar had the highest value (70.2 J/g). The melting behavior of both cultivars revealed identical patterns but different DSC parameter values which confirm that the melting behavior was affected by cultivar. This difference is in concordance with the variation in TAG composition detected between the two cultivars.

Tan and Che Man (2002) investigated the melting behavior of walnut (Juglans regia) oil and reported a broad endothermic peak with smaller ones. These peaks were attributed to the melting of the predominant TAGs detected in this oil, namely LLL, LLLn, and OLL. The same authors also studied the melting behavior of hazelnut oil and found a large peak.

The crystallization profiles showed an exothermic phase transition with one well-defined peak for both cultivars. The temperature of the major peak of crystallization (P_c) was -63.3 for Mahan cultivar and -58.9 ºC for Moore cultivar (Table 4). As explained previously, here too, the Pc observed for both samples at a low-temperature zone was mainly due to the high contents in OOL, OLL, and OOO, since these unsaturated TAGs are known to crystallize at lower temperatures (Yanty et al., 2013). Both cultivars showed similar values for initial temperature of crystallization (T onset _c) (-24.8 - -24.6 ºC). The crystallization enthalpy (ΔH_c) varied significantly among cultivars indicating that the genotype significantly influences the ΔH_c value.

The crystallization behavior of both cultivars exhibited identical patterns. However, different values for DSC parameters demonstrated that it was linked to cultivar. This significant variability (Table 4) may be explained by the dissimilarity of fatty acid and TAG compositions detected between cultivars.

The crystallization behavior in pecan nut oil differs remarkably from that of walnut oil. Indeed, a wide exothermic peak and two smaller ones were detected in walnut oil with a ΔH_c value of -57.4 J/g (Che Man and Tan, 2002). The major peak was attributed to the crystallization of the main TAGs found in this oil (Che Man and Tan, 2002).

The differences found in the thermal analysis supply a basic comprehension of the thermodynamic variations linked to the chemical composition. As indicated by the melting and crystallization parameters, the lowering of the transition temperatures is directly related to the high level of unsaturated TAGs present in pecan kernel oil. To our knowledge, no previous studies have investigated the thermal properties of pecan nut oil and hence, it was not possible to compare our findings with prior works.

For a deeper investigation of the differences and similarities between the two studied cultivars, a principal component analysis was performed (Figure 1).

According to the PCA results, the first two components explained 90.19% of the total variance with the first one at 82.14% and the second one 8.05% of the total variance.

Moore and Mahan were different, located at the positive and negative sides of component 1. Moore was typified by great levels of oil content, C16:1n-9, C18:1n-9, C18:1n-7, C20:1n-9, MUFA, γ-tocopherol, α-tocopherol, δ-tocopherol, OOO, POO, POL, PLS, POP, POS, SOO, OOL, PLP, and high values for ΔH_m, ΔH_c, P_m, P_c, T end _m, T onset _m. Mahan was characterized by substantial levels of C14:0, C15:0, C16:0, C17:0, C18:0, C20:0, C22:0, SFA, C16:1n-7, C18:2n-6, C18:3n-3, PUFA, UFA, SOL, OLL, LLL, PLL, lutein, and high values for T onset _c, R.

These findings confirm the great variability in chemical composition and thermal behavior between the two pecan cultivars.