We collected ripe legumes from the L. leucocephala plant (University Campus, King Saud University) and ginning seeds. Seeds were powdered, and the oil was extracted from powdered seeds through a soxhlet apparatus using petroleum ether as solvent. After extraction, petroleum ether was removed by a vacuum rotatory evaporator and yield was found as 5.67%. Petroleum ether, sulphuric acid (BDH, Poole, England), and methanol (Sigma-Aldrich, Steinheim, Germany) were used without any further purification.

For the experiments, an Accu TOF LC plus (JMS-T 100LP) system consisting of a DART ion source (IonSense, DART), an Accu TOF LC plus a higher solution time-of-flight mass spectrometer (JEOL, Japan) and an auto injection valve were used. The operating conditions of the DART ion source were as follows: positive ion mode; discharge needle voltage: 3.0kV; perforated and grid electrode potentials: +100 and -100 V, respectively. Conditions of TOF-MS: cone voltage: 2000 V, monitored mass range: m/z 50–10000; acquisition rate: 1600 spectra min⁻¹; resolving power: approx. 6000 FWHM (full width at half maximum). The distance between the DART gun exit and mass spectrometer inlet was 7mm. The sampling glass rod was immersed for 1s into the sample containing approx. 500μL of the respective sample and was transferred to the optimized position in front of the DART gun exit and desorbed for 5s, while the spectral data were recorded. For each sample/PEG standard, a minimum of three repeated measurements were carried out. At the end of each run, the mass spectra of polyethylene glycol (PEG, average relative molecular weight 600/1000, Sigma–Aldrich, USA) methanol solution (200μL/mL⁻¹) were acquired to perform mass drift compensation (Easter J L et al., 2014; Wang Wang Y et al., 2014). The influence of gas beam temperature on signal intensity, LUCO and FAMEs were analyzed at 200 °C. The final spectra were obtained after the background was subtracted. The operating conditions were as follows: Gas Chromatography, Clarus 500 GC with FID, Perkin Elmer, USA, Column: Supelco Wax, 30 m long, 0.32 mm id, phase thickness 0.25μm, injection mode: split; split ratio, 1:100, 1.0μL injection used, diluted (1/10 v/v, FAMEs/hexane), Column temperature 140–240 °C, 4°C/min, an initial temperature hold at 140 °C for 5 min, FID temperature 260 °C, Helium carrier gas flow 1.0 mL/min, Hydrogen 30 mL/min, Air 300 mL/min. The FAMEs mix standard was analyzed under the same operating conditions to determine the peak identity. The FAMEs were expressed as relative area percentage (Dodds et al., 2005; Fardin-Kia et al., 2013).

LUCO spectra were taken on a FT IR spectrometer, Spectrum 100, Perkin Elmer, USA. ¹H and ¹³C NMR spectra were recorded on Jeol DPX400 MHz using deuterated chloroform as solvent and tetra methyl silane as an internal standard. The thermal analysis of LUCO was carried out by simultaneous thermogravimetric analysis-differential scanning calorimetry (TGA-DSC) (Model TGA/DSC1, Mettler Toledo AG, Analytical CH-8603, Schwerzenbach, Switzerland) in nitrogen atmosphere (99.997 purity). The metal analysis of acid digested LUCO oil was analyzed with Inductively Coupled Plasma-Mass Spectroscopy (ICP-MS); ThermoScientific, Instrument, USA.

A certain amount (250mg) of LUCO sample was put into a digestive vessel of 60 mL capacity, with 8 mL of 65% nitric acid and the mixture was shaken carefully with a clean polytetrafluroethylene (PTFE) bar. After 30 minutes, the vessel was closed. The sample and blank (without oil) vessels were placed in the Topwave Analytik Jena microwave for digestion and the program temperature was set at 140 °C, 50 bar pressure for 8 minutes (ramp). The vessels were then left to cool at room temperature. The digested aliquots were filtered with a whatman no. 42 filter paper and the filtrate was diluted with 50 mL deionized water (Pehlivana et al., 2008; Hsu W H et al., 2013). ICP-MS calibration was carried out by external calibration using 10 ppm of multi element standard solutions (Aristar grade, BDH laboratory supplies, England) for the trace elements. The sample and blank were analyzed in triplicate. The results are summarized in Table 1.

The MTT viability assay was performed with slight modifications as previously described (Mosmann T, 1983). Human hepatocarcinoma (HepG2) cells (2 ×10³) were seeded in 96 well plates and incubated in a CO₂ incubator at 37 °C for 24 hour. The cells were exposed to the culture medium and LUCO (0, 50, 100, 300, 700 μL/mL) for 24 h. After exposure, 100 μL of a MTT solution was added to each well and incubated for 4 hour at 37 °C. After incubation the formed formazon crystal was dissolved in DMSO (100 μL). Absorbance was taken at 530 nm using the enzyme-linked immune sorbent assay (ElISA) reader and the cell viability was calculated as:

The representative TOF-MS spectra of LUCO are shown in Figure 1. The spectra were divided into three categories according to ion fragmentations. The elemental (C, H, O) compositions of the major peaks were calculated by a software based on the exact mass number of the elements. The mass differences of identified compounds differed from their theoretical mass numbers by less than 5.0 mmu (Table 2). Our target focused only on triglycerides, diglycerides and monoglycerides, and no other molecules. We tentatively identified triacylglycerol ion peaks of palmitoyldilinolein (PLnLn, m/z 854.736), trilinolein (LnLnLn, m/z 874.780), linolyldilinolein (LLnLn, m/z 875.767), dipalmitoylbehenene (PPB, m/z 892.783), trilinolein (LLL, m/z 879.743), linolinediolein (LOO, m/z 883.773), triolein (OOO, m/z 886.767), stearinedilinolien (SLnLn, m/z 879.743), gadolyldilinolein (GLnLn, m/z 906.782), archadoyldilinolien (A LnLn, m/z 907.748), gadoyldiolein (GLL, m/z 910.758), archedoyldilinolein (ALL,m/z 912.770), gadoyldiolein (GOO, m/z 913.778), behenendilinolein (BLnLn, m/z 936.806), behenendiolein (BOO, m/z 936.806), lenolyldistearine (LSS, m/z 884.749), linolene digadolein (LnGG, m/z 938.803), olyldigadolein (OGG, m/z 942.821), linolyldiarchidein (LAA, m/z 944.827), and olyldiarchidiene (OAA, m/z 946.816). However, the differentiation of positional isomers of TAGs with two or three different fatty acids is not possible by TOF-MS, due to the absence of separation of the isobaric molecules prior to ionization. The relative intensities of molecular adducts, as well as molecular ions are present in TOF-MS spectra (Vaclavik et al., 2009; Hajslova et al., 2011; Bosque-Sendra et al., 2012; Xu et al., 2015).

In the TOF-MS spectra (Figure 1), the extensive fragmentation of TAG acquisition resulted in the occurrence of numerous types of ions under the experimental conditions. The most abundant were diacylglycerol fragment ions created by the loss of one fatty acid molecule from the glycerol backbone.

Ions at m/z 593 to 709 were identified as markers for diacylglycerols fragment ions: palmitolinolein (PL, m/z 593.493), oleyllinolein (OL, m/z 620.530), diolein (OO, m/z 623.530), palmitoylgadolein (PG, m/z 624.551), linolylarchidin (LnA, m/z 647.540), oleinecosadien (OS, m/z 623.532), oleylgadolein (OG, m/z 651.559), Linolylbehenine (LnB, m/z 677.669), Linolylbehenene (LB, m/z 680.5545), archideinstearine (AS, m/z 653.0447), gadolein stearine (GS, m/z 651.559), and archadeinbehenein (AH, m/z 709.6644). The main fragment ion (m/z 601.51) was the base peak in all spectra of LUCO; the least intensive was the monoglycerol ion of its corresponding fatty acid ion (Wang et al., 2014). It is, therefore, TOF-MS followed by an appropriate method for the differentiation of TAGs for the identification of compounds.

These observations confirmed that the results obtained from TOF-MS data were good and it was the choice of instrument for the screening of vegetable oil. Therefore, it is necessary to resort to multivariate techniques, which are important and proven techniques for complex data analysis (Vaclavik et al., 2009). The fatty acids involved may all be the same or different number of acids in many possible combinations. It is precisely this variation which gives rise to the wide spectrum of vegetable oils. It is not possible here to write the exact fatty acid sequence of the triglycerides. TOF-MS provides us a new way for vegetable oil analysis with speed of analysis, simplicity of sample introduction, and high sensitivity (Lesiak A D et al., 2014). The TOF-MS technique has been applied for the first time for the profiling of LUCO and was done successfully.

The result of compositional determination based on peak area of LUCO is given in Table 3. Vegetable oils have a characteristic fatty acid composition that is useful for product authentication (Tavassoli-Kafrani et al., 2016). On the other hand, fatty acid composition not only depends on species but on region, climate, degree of ripeness, harvesting and processing conditions. LUCO is rich in Linoleic acid (52.08%); the fraction analyzed contains palmitic, stearic, oleic, linolenic, arachidic, gadoleic, and behenic acid (Nehdi et al., 2013, 2014).

FTIR(cm^-1): 3472.20, 3008.54 (-CCH=CH-C); 2926.07 (CH₂.symmetrical); 2854.82, asymmetrical); 1746.51 (C=O,ester); 1661.00 (-HC=CH-); 1465.33 (vibration of deformation–CH-), 1377.63 (deformation vibration of methylene group), 1238.14 (deformation vibration in the plane of =CH from the unconjugated cis double bonds), 1163.72 (-CCOO-C); 1099.93,722.79 (vibration of C-C) (Figure 2) (Barison et al., 2010).

¹H NMR, CDCl₃, δ, ppm: 0.869–0.878 (-CH ₃), 1.241–1.291 (-CH ₂-), 1.597 (-O-CO-CH₂-CH₂-CH ₂), 2.011–2.046 (-CH₂-CH=CH)-,2.280–2.299 (-OCO-CH2-), 2.742–2.774 (=CH-CCH ₂-CCH=), 4.141–4.156 (CCH ₂-COCO-CH₂-C), 4.264–4.273 (>CHOCO-CH₂-C-), 5.330–5.362 (-CCH=CH-C-) (Figure 3).

¹³C NMR, CDCl₃, δ, ppm: 14.158–14.204 (-CH₃), 25.704, 27.279, 29.206–29.787, 31.607 (-CH₂-), 62.177 (>CHOCO-CH₂-), 68.944 (-OCH₂-CH<), 127.966–130.298 (-CH=CH-), 172.911–173.332 (C=O) (Figure 4) (Alam et al., 2014).

FTIR shows –CH₂, -CH₃ symmetrical and asymmetrical absorption bands at 2926.07cm^-1 and at 2854.82cm^-1. The absorption bands for -CH= CH- str appear at 3008.54cm^-1 and also the carbonyl ester band at 1746.51cm^-1. The presence of unsaturation is also supported by characteristic peaks in the ¹H NMR and ¹³C NMR spectra at 5.330–5.362 ppm and 127.966–130.298ppm, respectively. The rest of the characteristic peaks for typical functional groups in the LUCO backbone have been mentioned above.

DSC (Figure 5) thermogram of L. leucocephala oil shows two endotherms, first at -30.25 to -13.53 °C, centered at -24.51 °C, the second endotherm starting at -11.50 to 8.57 °C, centered at 3.33 °C. The first endotherm belongs to phase transition and the second to melting. The TGA-DTG (Figure 6) thermogram shows the onset of degradation at 350 °C, no weight loss was observed at this temperature; only 4% weight loss occurs at 380 °C due to moisture and volatile impurities. 50% weight loss occurs at 460 °C. The TGA thermogram shows single step degradation. At 525 °C, a complete decomposition of oil occurs. This decomposition is assigned to the degradation of the hydrocarbon chain of the oil. In the DTG curve, an endothermic transition between 375 °C and 525 °C assigned to the hydrocarbons was also observed. The maximum weight loss of the sample according to the DTG curve was observed at 375 °C to 525 °C.

Figure 7 shows the growth inhibition graph of LUCO at different doses (0, 50, 100, 300, 700 μL/mL). Taking the results of all concentrations into account, the cell death in HepG2 cells was found to be higher at (700 μL/mL) by the MTT assay. This result shows that the cytotoxicity of HepG2 cells was oil dose dependent. In this study, we quantified the cytotoxicity of oil in human hepatocarcinoma cells for 24 hour by employing the MTT test. Finally, IC₅₀ values were calculated for the oil according to the MTT test.