P. pedunculatus seeds were collected in September 2015, from the desert area in Baotou city of the Inner Mongolia Autonomous Region, China. The average length, width, thickness and weight of seed were 11.52 mm, 8.72 mm, 7.65 mm and 0.39 g, respectively. The samples were stored at 4 °C prior to analysis. All chemicals used were of analytical grade. The reference standards for fatty acid, tocopherols (α, β, γ, and δ isomers) were purchased from Sigma-Aldrich.

Raw P. pedunculatus seeds (49% water content) were pre-treated employing four different processing methods before extracting oil: microwaving (M): 200 g seeds were put on a tray as a single layer in a microwave oven at 2450 MHz for 10 min; roasting (R) (NATIONAL ENG CO., LTD) at 200 °C for 20 min; steaming (S): 200 g seeds were put on a tray as a single layer, the tray was on the steamer’s perforated strainer, and the seeds were steamed by the bottom boiling water for 20 min; combined roasting (at 100 °C for 15 min) and steaming (at 100 °C for 7 min) (RS). Raw seeds (RW) were used to obtain the unprocessed oil to be used as control.

All pre-treated P. pedunculatus seeds (1.5 kg) (samples obtained from four different processing methods) were wrapped inside four layers of filtration cloth and pressed using a laboratory hydraulic press (dimension: 650 mm (L) × 800 mm (D) × 1370 mm (H)) (National Eng. Co., Ltd., Korea). The maximum pressure applied for the mechanical hydraulic press extraction was 60 MPa (for 20 min). The oils collected were centrifuged to remove any particles and then stored at −70 °C prior to analysis. The oil collected from RW (unprocessed) was used as the control.

The color of all extracted oil samples (RW, M, R, S, and RS) was analyzed spectrophotometrically following the method by Yoshida et al. (1999) with a slight modification. As an index of color development, the absorbance at 420 nm of 5% (w/v) solutions of oils in isooctane was used.

Oil samples (25 mL) were placed in 100 mL beakers and stored in the dark in an oven at 65 °C for 168 h. Samples were taken out at regular intervals of 24 h, i.e., after 0 h, 24 h, 48 h, 72 h, 96 h, 120 h, 144 h, and 168 h, and were then stored at 65 °C until used for further analysis.

The contents of conjugated dienes (CD) were determined using the IUPAC 2.505 method (1987). The oil (10–20 mg) was weighed and dissolved in isooctane in a 25 mL volumetric flask. The absorbance at 233 nm was recorded by a UV-visible spectrophotometer (Shimadzu, Japan). Pure isooctane was used as a reference. CD values were calculated based on the following formula: CD = 0.84 × (Absorbance of solution at 233 nm/bc−k₀), where c is the concentration of oil in g per 1L, b is the length of the cuvette in cm, and k₀ is the absorption of acid (0.03).

The 2-thiobarbituric acid value was determined using the AOCS Cd 19–90 official method (1998). The 2-TBA (200 mg) was dissolved in 1-butanol in a 100 mL volumetric flask and stored in the dark at room temperature for 12 h. Before evaluating the 2-TBA value in the oil samples, an appropriate volume of 2-TBA solution was collected from the stock solution using suction filtration and again dissolved in 1-butanol in a 100 mL volumetric flask. The oil sample (200 mg) was dissolved in 1-butanol in a 25 mL volumetric flask. Next, 5 mL P. pedunculatus oil solution and 5 mL 2-TBA solution were transferred into a screw-capped test tube. The mixed solution was incubated at 95 °C in a thermostated water bath for 2 h, followed by cooling in an ice bath for 10 min. 1-butanol (5 mL) and 2-TBA solution (5 mL) were used as references. The absorbance at 530 nm was recorded using a UV-visible spectrophotometer (Shimadzu, Japan). Results were expressed as mg malondialdehyde (MAD) equivalents per g of oil. 2-TBA values were calculated based on the following formula: 2- TBA = 50 × Absorbance of solution at 530 nm/ the weight of P. pedunculatus oil.

Fatty acid composition was determined according to the method described by Bimakr et al. (2013). The oil sample (60 mg) was dissolved in 4 mL isooctane in a test tube and 200 µL 2 mol/L KOH-methanol were added. The mixed solution was then stirred for 30 s and given a standing time of 1–2 min during which 1 g NaHSO₄ was added, and again the mixture was stirred for 1 min followed by a standing time of 10 min. The supernatant was finally collected for gas chromatography (GC) analysis. The GC (6890N, Agilent, USA) was equipped with a capillary column (VF-23ms 30 m × 0.25 mm × 0.25 μm, Agilent, USA) and a flame ionization detector (FID). The oven temperature had the following program: 0–40 min at 180 °C followed by 40–60 min at 220 °C, with a rate of 4 °C /min. The FID temperature was maintained at 250 °C. The flow rates of helium, air and hydrogen were 1.0 mL/min, 300 mL/min, and 30 mL/min, respectively.

OSI values (AOCS Official Method Cd 12b-92) were obtained in duplicate for all oil treatments and the control using the Oxidative Stability Instrument (Omnion, Inc.). The instrument was run using a 3.000 ± 0.0001 g oil sample and a heating block temperature of 110 °C.

Polyphenol content was determined using the Folin-Ciocalteu assay, following the method by Parry et al. (2005) with a slight modification. The oil sample (2.50 g) was dissolved in 2.5 mL n-hexane in a test tube, and 7.50 mL 80% methanol were added to extract polyphenols. The mixed solution was stirred for 1 min and given a standing time of 9 min, extracted three times at room temperature for 10 min and then centrifuged at 3000 rpm for 10 min to collect the solution; excess methyl alcohol was removed by rotary evaporation at 66 °C. Next, 2.5 mL of 10% Folin-Ciocalteu reagent and 10 mL of 7.5% sodium carbonate solution were added to the solution taken in a 25 mL volumetric flask and made up to volume with distilled water. After 1 h of incubation, the absorbances were measured at 765 nm using a UV-visible spectrophotometer (Shimadzu, Japan). The polyphenol content was expressed in mg of polyphenols per 1000 g of extracted oil.

Tocopherol content was determined according to AOCS Method (AOCS, 1998) using a HPLC equipped with a fluorescence detector. In brief, 0.25 g oil sample was taken in a 25 mL volumetric flask and diluted in n-heptane. Next, the mixture was filtered with a PTFE filter (pore size 0.45 µm), and 10 µL of the filtrate was injected into the HPLC. The HPLC system was equipped with a LiChroCART @ 250-4 column (250 mm × 4.0 mm), 2695 pump, and 2475 Multi λ Fluorescence Detector (Waters Corporation, USA). Excitation and emission wavelengths were set at 295 nm and 330 nm, respectively. The mobile phase consisted of n-hexane/ tetrahydrofuran (1000/40 by vol.), and a flow rate of 1.0 mL/min was used. The tocopherol content was expressed in milligram of tocopherols per 1000 g of extracted oil.

The data are expressed as mean ± standard deviation. The analyses were performed using SPSS (version 17.0 for Windows 2007, SPSS Inc.); results with p ≤ 0.05 were considered as statistically significant

We found the crude fat content as high as 52.19 %, accounting for more than half of P. pedunculatus seeds. The oil yields obtained from P. pedunculatus seeds obtained employing different processing treatments ranged from 74.87% to 85.21% (Figure 1). The oils obtained using RS, R, and M methods exhibited significantly higher percentages of oil yield compared with that of unprocessed raw seeds (RW) 72.35%. This might be because of the de-naturation of proteins in P. pedunculatus seeds at higher temperature during process treatments, leading to a decrease in solubility. Due to de-naturation, hydrophobic groups get diverted outwardly, thus exposing and accumulating oil in the globulin on the seeds’ surface; this, in fact, favors oil extraction.

The different processing treatments of P. pedunculatus seeds resulted in changes in oil color (Figure 2). The color of P. pedunculatus oil obtained from R seeds did not differ from that of RW seeds at 420 nm; however, the color of the oil obtained from M seeds increased and became significantly darker (p ≤ 0.05) compared with RW, R, S, and RS. This could be due to the products formed in the Maillard reaction during M. treatment (Wu Huiling et al., 2010), or the cell structure of the seeds changed during the M. treatment; the pigment and phospholipids were more likely to be extracted together with the oil (Wang, 2011).

Table 1. summarizes the fatty acid composition of the oils extracted from P. pedunculatus unprocessed (RW) and processed (M, R, S, and RS) seeds. As evident from Table 1, the fatty acid compositions of five oil samples (RW, M, R, S, and RS) remained unchanged; however, the difference in content was significant. Oleic acid (C18:1) and linoleic (C18:2) acid were the main fatty acids in all samples, respectively. The main reason for oil rancidity was the oxidation of highly unsaturated fatty acids (USFA); lower content of USFA increases the oil’s stability. Regardless of processing techniques, the USFA content of P. pedunculatus oil was high with the oil sample obtained from RW showing 97.67% USFA. The oil obtained from S seeds contained a high USFA of 98.08%, followed by RS with 97.78%. The oil obtained from RS had a 47.93 ratio of USFA to SFA and the calculated oxidizability (Cox) value of 3.45. The Cox value of the oils was calculated by percentage of unsaturated C18 fatty acids, applying the formula proposed by Fatemi and Hammond (1980):

(Cox=[1×(C18:1%)+10.3×(C18:2%)+21.6×(C18:3%)]/100)

As the Cox value of oil is lower, its antioxidant ability was higher. In addition, the oil obtained from RS seeds showed the highest stability with an oil stability index (OSI) of 19.13 h. The Oil Stability Index (OSI; AOCS Official Method Cd 12b-92) had largely replaced the AOM as a means to measure an oil’s susceptibility to oxidation (Broadbent et al., 2003). The USFA in P. pedunculatus oil was mainly composed of MUFA (70.91%). Therefore, in the calculation of Cox values, the content of oleic acid was lower than linoleic acid and linolenic acid, suggesting a lower oxidant rate of oleic acid than other C18 in PUFA (Rossi et al., 2007). This indicated better stability of the P. pedunculatus oils extracted from different processing treatments; the Cox values for processed P. pedunculatus oils were less than that of the RW oil, while the OSI value of the oil from RW seeds was the lowest among them all.