Comparison of Two Stationary Phases for the Determination of Phytosterols and Tocopherols in Mango and Its By-Products by GC-QTOF-MS

Two different gas chromatography coupled to quadrupole-time of flight mass spectrometry (GC-QTOF-MS) methodologies were carried out for the analysis of phytosterols and tocopherols in the flesh of three mango cultivars and their by-products (pulp, peel, and seed). To that end, a non-polar column ((5%-phenyl)-methylpolysiloxane (HP-5ms)) and a mid-polar column (crossbond trifluoropropylmethyl polysiloxane (RTX-200MS)) were used. The analysis time for RTX-200MS was much lower than the one obtained with HP-5ms. Furthermore, the optimized method for the RTX-200MS column had a higher sensibility and precision of peak area than the HP-5ms methodology. However, RTX-200MS produced an overlapping between β-sitosterol and Δ5-avenasterol. Four phytosterols and two tocopherols were identified in mango samples. As far as we are concerned, this is the first time that phytosterols have been studied in mango peel and that Δ5-avenasterol has been reported in mango pulp. α- and γ-tocopherol were determined in peel, and α-tocopherol was the major tocopherol in this fraction (up to 81.2%); however, only α-tocopherol was determined in the pulp and seed. The peel was the fraction with the highest total concentration of phytosterols followed by seed and pulp, and “Sensación” was the cultivar with the highest concentration of total phytosterols in most cases. There were no significant differences between quantification of tocopherols with both columns. However, in most cases, quantification of phytosterols was higher with RTX-200MS than with HP-5ms column.


Introduction
Mango (Mangifera indica L.), a member of Anacardiaceae family in the order of Sapindales is native to Southeast of Asia and is widely cultivated at both tropical and subtropical latitudes [1]. There are more than 70 genera, and 1000 varieties, and it is considered one of the most important tropical fruits [2]. Presently, mango is cultivated on an area of approximately 3.7 million ha worldwide. Besides, mango fruit is in the second position as a tropical crop, only behind bananas in terms of production and acreage used [3].
Mango fruits provide energy, dietary fibre, carbohydrates, proteins, fats, and phenolic compounds, which are vital to normal human growth, development, and health. Moreover, it has been well documented that mango fruits are an important source of micronutrients, vitamins such as tocopherols, and other phytochemicals such as phytosterols [4][5][6]. Tocopherols play a role as an in vivo antioxidant

Optimization of GC-QTOF-MS Methods
The gas chromatography (GC) methods were optimized using a real sample extract of mango peel in full scan mode.
Temperature of the injection port was checked to establish the considered optimal one. It was ranged between 200 and 300 • C and 280 • C was selected as optimal for both columns. Furthermore, the gradient of temperatures of the oven and the gas flow were tested to choose the best compromise between resolution and analysis time. The initial column temperature was checked for both columns (150-250 • C) and 210 • C was selected for RTX-200MS, whereas 250 • C was chosen for HP-5ms. The higher polarity of RTX-200MS column provided a lower retention of compounds in the stationary phase and a lower resolution among them. Thus, a lower initial temperature was needed to obtain a good separation. After that, an increase of temperature was required until an isothermal step was reached: 260 • C for RTX-200MS and 270 • C for HP-5ms. The elution of the compounds of interest was produced in this step; the analysis time for RTX-200MS was much lower than the one obtained with HP-5ms; however, a proper resolution between peaks was achieved with HP-5ms whereas RTX-200MS produced an overlapping between β-sitosterol and ∆ 5 -avenasterol. An isothermal step below 260 • C produced an excessive retention of the peaks, increasing their width and the resolution was not significantly improved. Then, the final column temperature was optimized and 310 • C was selected for both columns.
The parameters related to the mass detection method were optimized, taking into account the area of the mass spectrometry (MS) signal of the standards compounds available (α-tocopherol, β-sitosterol, stigmasterol, campesterol and dihydrocholesterol). These parameters were chosen to get a compromise solution to obtain the maximum signal for most of the peaks detected in the samples under study.

Methods Validation
Four calibration curves were prepared using tocopherol, campesterol, stigmasterol, and β-sitosterol standards. Table 1 summarizes the different parameters of each standard compound for both columns. All calibration curves showed good linearity between different concentrations depending on the analytes in question. The calibration plots reveal good correlation between peak areas and analyte concentrations, and the regression coefficients were always higher than 0.99. Limit of detection (LOD) and limit of quantitation (LOQ) were calculated according to S/N ratio. LOD was found to be within the range 0.883-1.304 µg/mL whereas LOQ was within 2.942-4.342 µg/mL for HP-5ms column, whilst for RTX-200MS column LOD was within 0.578-0.852 µg/mL and LOQ was within 1.925-2.841 µg/mL. As it can be seen, LOD and LOQ of the method optimized using the RTX-200MS column were lower than those obtained for HP-5ms column. Therefore, it can be said that the optimized method for the analysis of tocopherols and phytosterols using RTX-200MS column has a higher sensibility than the HP-5ms methodology.
Intraday and interday repeatability were developed besides the relative standard deviations (% RSD) ( Table 2). Intra-and interday precision of analysis time were very similar for both columns, whereas the intra-and interday precision of peak area were lower for HP-5ms compared to RTX-200MS.

Identification of Phytosterols and Tocopherols of Mango Fractions
Both optimized methodologies were applied to mango samples and peak identification was done (Table 3 and Figure 1).    Table 3 for identification numbers. IS (internal standard).  Table 3 for identification numbers. IS (internal standard).
Six compounds: two tocopherols and four phytosterols were detected in mango samples. The identification of every compound was corroborated with its standard as exception of γ-tocopherol and ∆ 5 -avenasterol. γ and α-tocopherols eluted in the first place. γ-tocopherol (peak 1) at m/z 488, presented a fragment at m/z 223 [M-255] due to the loss of the side chain (C 16 H 33 ) and another fragment at m/z 263 [M-(225+40)], because of the cleavage of the side chain accomplished by the breakdown of chroman structure with hydrogen rearrangement and loss of a methyl acetylene fragment [25]. This compound was only detected in mango peel.
γand β-tocopherol are isomers; therefore the obtained ions by gas chromatography-mass spectrometry (GC-MS) are the same for both of tocopherol forms [26]. However, in previous studies, the presence of γ-tocopherol and lack of β-tocopherol in mango seed kernel was reported [5]. Indeed, β-tocopherol is the rarest tocopherol homologue present in the plant world with some exceptions, for instance, coffee beans [27] or various types of bran and germ of the wheat [28,29], the presence of γ-tocopherol instead of β-tocopherol can be rather stated in mango.
α-Tocopherol (peak 2), at m/z 502, showed fragments at m/z 237 and 277 due to a similar fragmentation pattern previously indicated for γ-tocopherol. α-Tocopherol was detected in mango pulp, peel, and seed. Previous research reports the presence of α-tocopherol in pulp [6].
All of the identified phytosterols were detected in pulp, peel, and seed of mango. The first phytosterol (peak 3) at m/z 472, was identified as campesterol. It showed three typical fragments: at  + . This fragmentation pattern is in agreement with previous literature and also with the spectra of its standard [30,31].
Peak 4 at m/z 484 and fragments at m/z 469, 394, 379, 355, 343, 255, 253, 213, and 129, that coincide with the same fragmentation pattern as campesterol, was identified as stigmasterol. This identification was corroborated by the injection of its standard and it is in agreement with Pelillo et al. [30].
Last phytosterol determined in mango was ∆ 5 -avenasterol (peak 6) at m/z 484 and with the typical fragments at 469, 394, 379, 355, 343, 255, 253, 213, and 129. Besides, it showed an important fragment at m/z 386 corresponding to [M-98] + . This compound was previously reported in seed [5] and, as far as we are concerned, this is the first time that ∆ 5 -avenasterol has been identified in mango peel and pulp.
On the one hand, it is interesting to highlight that RTX-200MS column was not able to clearly separate ∆ 5 -avenasterol in any fraction of mango. On the other hand, total time of analysis was almost 5 min shorter using RTX-200MS than HP-5ms.
To our knowledge, this is the first time phytosterols have been analyzed in mango peel; therefore, this is the first time that these compounds have been identified and quantified in mango peel. Furthermore, as far as we are concerned, ∆ 5 -avenasterol has been reported in mango pulp for the first time in this work.

Quantification of Phytosterols and Tocopherols of Mango Fractions
The methods performed were used to quantify phytosterols and tocopherols in mango pulp, peel, and seed. The results of the phytosterols and tocopherols quantification in different mango fractions and cultivars are showed in Table 4.
In relation to tocopherols, they were present in peel, pulp, and seed; however, peel exhibited higher total concentration of tocopherols than pulp and seed. The concentration of tocopherols was 74.4-79.8% higher in peel than in pulp for the different cultivars when HP-5ms was used and 75.5-79.4% when RTX-200MS was used, whereas tocopherol content was 94.69-97.99% higher in peel than in seed for the different cultivars when HP-5ms was used and 95.48-98.13%. Regarding cultivars, "Sensación" presented the highest content in total tocopherols in all the fractions studied, followed by "Keitt" and "Osteen". α and γ-tocopherol were simultaneously determined in peel, whereas α-tocopherol was only identified in pulp and seed. α-tocopherol was the major tocopherol in peel. Its percentage with HP-5ms column was 77.39%, 72.28%, and 75.10% of the total tocopherols, in "Keitt", "Osteen", and "Sensación", respectively, and with RTX-200MS column was 81.19%, 72.75%, and 73.77% of the total tocopherols, in "Keitt", "Osteen", and "Sensación", respectively. Table 4. Phytosterols and tocopherols quantification in mango pulp, peel and seed of "Keitt", "Osteen", and "Sensación" cultivars measured with HP5 and RTX columns by GC-QTOF-MS expressed as µg/g dry matter. Different letters in the same raw indicate significant differences (p < 0.05). Three replicates of each sample were analysed.
Concentrations of tocopherols and phytosterols in mango are in the same range of concentrations for results found by Vilela et al. [6] in pulp and by Jin et al. [32] in seed. However, differences observed could mainly be attributed to agronomical and environmental factors. The cultivar can also highly influence the appearance and content of tocopherols in fruit seeds [33][34][35].
It is important to notice that there was no difference between quantification of tocopherols with both columns. However, in most cases, quantification of phytosterols was higher with RTX-200MS column than with HP-5ms column.
Statistical analyses were used to determine the correlation between quantifications carried out with HP-5ms and RTX-200MS columns. The results showed that there was a high correlation between total tocopherol and total phytosterol concentration found with HP-5ms and RTX-200MS columns, respectively (r = 0.9931; p < 0.01 and r = 0.9707; p < 0.01, respectively).

Samples
Mangoes were provided by Miguel García Sánchez e Hijos, S.A. (Motril, Spain) from the October 2015 harvest. About 10 kg of mango fruits of the cultivars "Keitt", "Osteen", and "Sensación" were manually separated in peel, pulp, and seed. The samples were freeze-dried in a lyophilizer (Advantage Plus EL-85 freeze dryer, SP Scientific, Ipswich, Suffolk, UK) and, then, milled and kept at −18 • C until use.

Isolation of Unsaponifiable Fraction by Hot Saponification
Hot saponification was performed according to Caligiani et al. [21] with some modifications. Briefly, 1.0 mL of dihydrocholesterol (c = 1.0 mg/mL) was added to 15 g of sample powder and saponification was carried out by boiling and stirring for 1 h with 100 mL of 1.0 N potassium hydroxide in ethanol-water solution (4/1 v/v). The sample was transferred to a separating funnel and 100 mL of distilled water was added. Then, an extraction was carried out with 50 mL of diethyl ether which was repeated four times. The ether extracts were pooled into a separating funnel and washed four times with 50 mL of distilled water. The organic phase was dried with anhydrous sodium sulphate, filtered, dried in a rotary evaporator, and then the residue was weighed and kept at −18 • C until use.

Determination of Phytosterols and Tocopherols by GC-QTOF-MS
Phytosterols and tocopherols were analyzed in a gas chromatograph coupled to a quadrupole-time of a flight mass spectrometer (GC-QTOF-MS) (Agilent Technologies, Santa Clara, CA, USA) consisting of a gas chromatograph (7890B Agilent Technologies), a QTOF mass spectrometer (7200 Agilent Technologies), and an autosampler (GC Sampler 120 Agilent Technologies).
Two methodologies were developed to separate the phytosterols and tocopherols from the different parts of mango by GC-QTOF-MS. To that end, two GC columns were used: HP-5ms (Agilent Technologies, Santa Clara, CA, USA) and RTX-200MS (Restek Corporation, Bellefonte, PA, USA). HP-5ms is a non-polar column with very low bleed characteristics; this column has been widely used for the determination of phytosterols and tocopherols. RTX-200MS is a mid-polar column that offers exceptional thermal stability, low bleed, and superior inertness, even for active compounds. This column has also been used for the analysis of phytosterols in recent years [36].
Quantification was carried out with both HP-5ms and RTX-200MS columns. Four calibration curves were prepared using the following standards: tocopherol, campesterol, stigmasterol, and β-sitosterol. Dihydrocholesterol was used as internal standard (IS). Calibration curves were built using the standard/IS peak area ratio versus standard concentration. Standards were injected under the optimum conditions of the method found for both columns.
LOD and LOQ were calculated based on S/N (Signal to Noise) ratio. They were established injecting the standard solutions by the injection of the smallest amounts which provide S/N = 3 for LOD and S/N = 10 for LOQ. Method accuracy was determined by the closeness of the test value to the nominal value and was evaluated with separately prepared individual primary stock solutions, mixtures, and working solutions of all standards. It was calculated over the linear dynamic range at three concentration levels: low (0.1 µg/mL), medium (50 µg/mL), and high (100 µg/mL) via three assays per concentration on different days. The analyte concentrations were calculated from calibration curves and accuracy was calculated by the ratio of this calculated concentration versus the theoretical one.
Moreover, intraday and interday repeatability were developed to assess the performances of the method, taking into account the standards at the three levels of concentration. Because of that, one mango extract was injected (n = 3) three times a day (intraday precision), and for seven consecutive days (interday precision, n = 21). The relative standard deviations (% RSD) of analysis time and peak area were determined for RTX-200MS and HP-5ms columns.
Before the analysis, the samples were silylated according to Sweeley et al. [37]. Briefly, silylation was performed adding 200 µL of a mixture containing pyridine/hexamethyldisilazane/ trimethylchlorosilane 5/2/1 v/v/v to 5 mg of unsaponifiable fraction. The sample was agitated and incubated at 40 • C for 20 min. Afterwards, it was dried and reconstituted in 2 mL of n-hexane. Finally, the solution was centrifuged at 3500 rpm and 4 • C for 10 min. Separations were carried out with different methods depending on the capillary column used.

HP-5ms Column
The first column used was an HP-5ms (30 m × 0.25 mm inner diameter, 0.25 µm film thickness) from Agilent Technologies. The chromatographic conditions of the method were as follows: initial temperature 250 • C; a temperature gradient of 10 • C·min −1 until reaching 270 • C and holding that temperature for 25 min; a temperature gradient of 10 • C·min −1 until reaching 310 • C and holding that temperature for 5 min;. The injector temperature was 280 • C and the transfer line temperature was 300 • C. The injection volume was 1 µL in split mode with a 1:10 split ratio. The carrier gas was helium at a flow rate 1.2 mL/min.

RTX-200MS Column
The second column was an RTX-200MS (30 m × 0.25 mm inner diameter, 0.25 µm film thickness) from Restek Corporation. The chromatographic conditions of the method were as follows: initial temperature 210 • C; a temperature gradient of 15 • C·min −1 until reaching 260 • C and holding that temperature for 10 min; a temperature gradient of 15 • C·min −1 until reaching 310 • C and holding that temperature for 2 min. The injector temperature was 280 • C and the transfer line temperature was 300 • C. The injection volume was 1 µL in split mode with 1:50 split ratio. The carrier gas was helium at a flow rate 1.2 mL/min. Instrumental conditions used for the MS detector were as follows: acquisition mode, total ion current; ion source temperature, 230 • C; detector voltage, 70 eV; scan range, from 50 to 700 m/z; scan speed, 1250 µ/s; solvent delay time, 6 min. Data were filed and processed by the software, MassHunter Workstation (Agilent Technologies, Santa Clara, CA, USA).

Conclusions
Two methodologies have been developed for the determination of tocopherols and phytosterols in pulp, peel, and seed of three mango cultivars ("Keitt", "Osteen", and "Sensación") by GC-QTOF-MS using two columns with different stationary phases. The analysis time was shorter using the RTX-200MS column than with the HP-5ms. Moreover, the optimized RTX-200MS method could obtain a higher sensibility and precision of peak area than the HP-5ms methodology. However, the RTX-200MS column gave cause to an overlapping between β-sitosterol and ∆ 5 -avenasterol.
Four phytosterols and two tocopherols were identified in the different mango fractions studied and no significant differences were found for the quantification results of tocopherols obtained with the two columns. Nevertheless, in most cases, quantification of phytosterols was higher with the RTX-200MS column than with the HP-5ms column. Indeed, the results showed that there was a high correlation between total tocopherol and total phytosterol concentrations found with the HP-5ms and RTX-200MS columns.
Mango fat has demonstrated similar characteristics to cacao butter; however, there is a lack of information about the distribution of some of the families of mango fat in mango fruit. Thus, this work provides evidence about the distribution of phytosterols and tocopherols in mango pulp, peel, and seed. Besides, this is the first time that phytosterols have been studied in mango peel and that ∆ 5 -avenasterol has been reported in mango pulp.