MoO was extracted from dried M. oleifera kernels that had been ground into a fine powder using a Waring blender Model 32BL 80 (Dinamic Corpration of America, New Hartford, Connecticut, USA). The ground kernel (500 g) was placed in a cellulose filter paper (Whatman No. 1) cone and the oil was extracted using light petroleum ether (b.p 40–60 °C) in a 5 L Soxhlet extractor for 8 hours (AOAC, 1984). The solvent was then evaporated off using a rotary evaporator Model N-1 (Eyela, Tokyo Rikakikal Co., Ltd., Japan) and the recovered oil was heated in a convection oven at 60 °C for 1 hour to remove residual solvent. The oil was then flushed with 99.9% nitrogen and stored at −20 °C until further use (Abdulkarim et al., 2005). The crude oil was used without further refining.

VCO was produced using the natural fermentation method described by Bawalan and Chapman (2006). One kilogram of fresh coconut milk was placed in a 2 L beaker, covered with aluminum foil and left standing at room temperature (25–27 °C) for 24 hours for natural fermentation to take place. The upper layer (oil layer) was then filtered through a filter paper (Fisher QC100, Fisher Scientific, Leicestershire, UK) and heated at 60 °C in a convection oven overnight to remove extra moisture in the oil. The oil was kept in −20 °C for analysis.

Sixty grams (60 g) of melted M. oleifera seed oil/palm olein (MoO/PO) and M. oleifera seed oil/virgin coconut oil (MoO/VCO) blends at ratios of 70:30, 50:50 and 30:70 (w/w), each dissolved separately in 100 mL of water-saturated n-hexane, were placed in 250 mL conical flasks. Lipozyme RMIM (10% w/w of oil mixture) was then added to initiate catalysis. The reaction mixtures were agitated at 60 °C at 200 rpm for 0, 3, 6, 12 and 24 h. The blend ratios were selected based on preliminary experiments that show desired changes after blending and IE. The reaction mixture at 0 h acted as the control. After the reaction, each oil mixture was filtered through a Whatman No. 1 filter paper to remove the enzyme and the oil mixtures were kept at −20 °C for further analysis. The reaction was carried out in duplicate. The stored mixtures were thawed at room temperature and then heated at 60 °C for 30 min to ensure complete melting before analysis.

Free fatty acids (FFAs) were removed from the reacted oil using the method described by Long et al. (2003). The reacted oil (30 g) was placed in a 250 mL conical flask and 50 mL of acetone: ethanol (1:1 v/v) were added and mixed. The mixture was then titrated against 0.1 N NaOH to a phenolphthalein end-point. The titrated sample was diluted with 10 mL of boiling distilled water and transferred to a 250 mL separating funnel. The mixture was shaken and allowed to settle for several minutes. The aqueous phase which contained the FFA, was separated from the organic phase and discarded. The organic phase was then dried at 60 °C overnight.

The triacylglyceride (TAG) profiles of the oil samples were obtained by injecting 10 µL of oil samples (10% in acetone, v/v) into a reversed-phase high performance liquid chromatograph (HPLC) comprising a Waters Alliance HPLC (Waters Corporation, Milford, Massachusetts, USA) equipped with a separation module (Waters Model 2695), an auto-injector and a refractive index detector (Waters Model 2414). The TAG were separated using a commercially packed RP-18 column (250×4 mm) with particle size 5 mm (Merck, Darmstadt, Germany) held in a column oven set at 35 °C. Elution of TAG from the column was achieved using a mixture of acetone/acetonitrile (63.5:36.5) as the mobile phase at a flow rate of 1 mL·min⁻¹. The total run time was 100 min and TAG peaks were identified based on the retention time of available TAG standards and the results of Chen et al. (2007) and Marina et al. (2009a). Peak areas produced by the data integrator (Empower-Pro Version 2002, Waters Corporation, Milford, Massachusetts, USA) were used to quantify the components based on relative percentages.

The thermal properties of the oil samples were investigated by differential scanning calorimetry (DSC) using a Perkin-Elmer Diamond DSC (Shelton, Connecticut, USA) based on the procedure described by Abdulkarim et al. (2005). The instrument was calibrated using indium and zinc. The purge gas used was 99.9% nitrogen with a flow rate of 100 mL·min⁻¹ and a pressure of 20 psi. Samples (5–8 mg that had been completely melted at 60 °C) were sealed in aluminum volatile DSC pans and placed in the instrument's sample chamber. An empty DSC pan was used as the reference. The sample chamber was heated to 70 °C and held for 2 min to erase crystal memory in the oil. It was then cooled to −60 °C at the rate of 10 °C·min⁻¹, and held at −60 °C isothermally for 2 min and heated from −60 °C to 70 °C at the rate of 10 °C/minute. The heating and cooling thermograms were recorded and the values for onset, peak and end set temperatures tabulated. The onset temperature indicates the temperature at which the melting process starts and the end set temperature indicates the complete melting temperature of the oil.

The solid fat contents (SFC) of all samples at different temperatures were calculated from their thermal profiles according to Adhikari et al. (2010), as obtained above. Each DSC melting thermogram was divided into different temperature sectors (0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50 and 55 °C) and the total crystallization energy (J/g) at each temperature was then converted into percentage (%) fat based on the following equation: (%) Fat =[ Partial ( J / G )/ Total ( J / G )]×[ 100 ]

A statistical analysis was carried out using MINITAB 14 (Minitab Inc., State College, PA, USA) by subjecting all the data obtained to a univariate analysis of variance (ANOVA) to determine the significant differences among samples, monitored by Tukey's pairwise test, and significance was defined at p<0.05. All measurements were carried out in triplicate and reported as the mean and standard deviation of independent trials.

Interesterification (IE) requires the hydrolysis of oil by lipase in order to produce free fatty acids (FFAs) and a mixture of mono- and diacylglycerols. These products of hydrolysis will then be synthesized to generate more of either existing or new kinds of TAG. Since not all FFAs participated in the IE process, they may accumulate in the reaction mixture. The original amount of FFA in extracted MoO and VCO found to be 0.56% and 0.23% respectively where no FFA was recorded in refined bleached deodorized (RBD) PO. In this study, after the reaction, the FFA contents were found to be increased significantly (p<0.05) with the time of reaction compared to the unreacted blends for both MoO/PO and MoO/VCO in which the highest FFA levels were produced after 24 h reaction time. Post-IE FFA levels for MoO/PO blends ranged between 1.85 and 2.14%, whereas the control values (pre-IE) ranged from 0.37 to 0.60%. For MoO/VCO blends, post-IE and pre-IE values of FFA content ranged from 1.75 to 3.24% and 0.23 to 0.42%, respectively. Prior to the analysis of other physicochemical properties, as many FFAs as possible were removed by solvent-solvent fractionation as they might act as an interference. The FFA levels of interesterified MoO/PO and MoO/VCO blends after their removal were between 0.15 and 0.18% and 0.12 and 0.18%, respectively.

Figure 1 shows the triacylglyceride (TAG) profiles of MoO/PO and MoO/VCO blends at the ratio of 50:50 w/w and their interesterified blends at zero (control) and 24 h reaction time. After Lypozyme RMIM IE, several TAG peaks increased in height indicating synthesis, while others decreased indicating hydrolysis. By comparing with the control (pre-IE), TAG groups (S3, S2U, U2S and U3, where S and U represent saturated and unsaturated fatty acids, respectively) of MoO/PO and MoO/VCO interesterified blends shown in Table 1 provide evidence of changes in the TAG composition based on the degree of unsaturation.

As complete identification of TAG in MoO/PO is limited by the unavailability of standards, only relative changes in the concentrations of MMP, POO+POL, MPL, PPL, OOO and PPO (where M, P, O and L indicate myristic, palmitic, oleic and linoleic acids respectively) were tracked. POO and POL were quantified together due to both having the same retention time under the conditions used in the analysis. The interesterified (24 h) MoO/PO blends showed major increases in POO+SOL (U2S) due to the incorporation of oleic acid from MoO fatty acids. The levels of U2S were found to be 30.57%, 25.88% and 22.48% with enhancement values of 4.51%, 3.66% and 4.34% for 30:70, 50:50 and 70:30 w/w, respectively. On the other hand, PPO and PPL levels (S2U) decreased considerably with the values of 2.58% and 1.59% in 50:50 w/w respectively and less saturated TAGs (U2S) were increased. As a result, a gradual increase in U2S and a decrease in S2U were observed in all blend ratios. Palla et al. (2012) reported that an increase occurred in U2S after sunflower oil was reacted with palmitic-stearic acid mixtures using Lipozyme RMIM as a result of a reaction between U3 and S2U groups. Debnath et al. (2011) also reported on the increase of U2S and the decrease in S2U in interesterified PO and rice bran oil blends compared to the unreacted PO. As shown in Table 1, the level of U3 TAGs (OOO) remained unchanged except in 50:50 w/w with a 1.23% decrease (from 21.55% to 20.32%) however, S3 TAGs were increased in all interesterified MoO/PO binary blends with the values of 1.13% in 70:30, 1.83% in 50:50 and 2.55% in 30:70. A similar observation was reported by Ghazali et al. (1995a) where an increase of S3 in all interesterified PO which was catalyszd using microbial lipases from different sources. The MoO/PO blends that showed the highest increase s in U2S after being interesterified by Lipozyme RMIM for 24 h were 30:70 and 50:50 ratios. The decrease in the level of S2U and increase in the level of U2S are desired (da Silva et al., 2010; Ribeiro et al., 2009) as such; an increase in TAG containing unsaturated fatty acids may contribute positively to the stability to the oil as well as to its health benefits. Unexpectedly, S3 enhancement resulted in interesterified MoO/PO, which is also desirable in this study, as MMP (myristic myristic palmitic) would be the representative of this type of TAG group (Figure 1a). MMP, which is a medium-medium-long chain (MML) TAG, is generally called MLCT (medium long chain TAG) and when considered as a structured lipid (SL), would be a beneficial and nutritional fat (Lee et al., 2012) which was generated due to the rearrangements of the TAG occurring after IE .

The same trend with MoO/VCO blends following Lipozyme RMIM IE occurred in the case of U2S and S2U TAGs. As shown in Table 1, as IE progressed, the level of U2S increased while, S2U decreased gradually in all ratios of the MoO/VCO blends. However, the trend for changes in U3 (OOO) was completely different from the MoO/PO results. The rate of U3 hydrolysis was much lower (0–1.23%) in MoO/PO IE compared to MoO/VCO IE (3.15–6.73%) where OOO showed the greatest decrease in concentration over the time of IE (Table 1). The highest amount of TAG hydrolysis (OOO decreased 6.73%) occurred after 24 h IE of MoO/VCO 70:30 w/w, while the lowest amount (no change in OOO value) was detected for the 24 h interesterified MoO/PO 70:30 w/w. in the case of MoO/VCO IE, other TAGs that showed corresponding decreases proportional to the reaction time belong to S3 TAGs consisting of CCLa, CLaLa, LaLaLa, LaLaM, and LaPM.

As shown in Table 2, the function of Lipozyme RMIM on MoO/VCO IE resulted in the increase in MLCT species (LaLaO, LaMO, LaPM and LaPO) concentration as well as the synthesis of a new TAG species (LaOO). In fact, the major MLCT group generated following IE was MLL (medium-long-long) comprising LaOO in all MoO/VCO ratios with a maximum value of 5.43% in 24 h interesterified MoO/VCO 70:30 w/w (Table 2). The most noteworthy increase in TAG concentration was observed for MML (medium-medium-long chain), the TAG group comprising LaLaO and LaMO in all MoO/VCO ratios (Table 2). The total amount of generated MLCT with incorporated oleic acid (LaLaO, LaMO, LaPO and LaOO) was maximized in 24 h interesterified MoO/VCO 50:50 with the value of 17.86%. It has been well established that substrate ratio is an important parameter to obtain the desirable yield in IE (Fomuso et al., 1998).

In this study, IE resulted in the increase and synthesis of new TAGs, which was revealed by the emergence of several lower molecular weight TAGs, namely MLCT. The first commercially available MLCT sold in the market by Nisshin Ollio Ltd. Group, Ltd, uses the enzymatic IE process for its production involving coconut or palm kernel oil and an edible vegetable oil (rapeseed, cottonseed and soybean oil) (Lee et al., 2012). Khatoon et al. (2005) reported on interesterified palm stearin with coconut oil, where LaLaLa decreased tremendously while that of LaLaO and LaOO experienced major increases. In other studies, the incorporation of medium chain saturated FAs occurred in safflower oil (Utsugi et al., 2009) and anhydrous milk fat (Liew et al., 2001). From the health benefit point of view, MLCT not only can provide us with nutritional properties from the essential fatty acids incorporated, but most importantly it can also help to reduce body weight and body fat accumulation in the body (Lee et al., 2012).

During IE, Lipozyme RMIM catalyzes the cleavage of FAs from the glycerol backbone of some TAG species at the 1, 3 position (those that decreased in concentration after the reaction) and re-esterifies randomly at the free hydroxyl positions on the glycerol backbone of the existing acylglycerols. The difference in lipozyme RMIM functionality toward PO and VCO interesterified blends with MoO could be due to the processing methods undergone. As mentioned by O'Brien (2003), the PO produced after the refining, bleaching and deodorizing (RBD) procedure with heating is applied especially during the deodorization process between 204 and 245 °C. However, VCO, known as a functional oil or mineral oil was produced in mild conditions (Chen et al., 2007) in order to maintain the bioactive compounds. As indicated by khoramnia et al. (2011) and Ebrahimpour et al. (2011), the mineral contents of the cultures play an essential role in the function of lipases.

Generally, as indicated by Ribeiro et al. (2009), the increased U2S content in the MoO/PO and MoO/VCO blends produced by enzymatic interesterification lead to greater potential for applying these interesterified blends in foods. S2U TAGs are associated with enhanced technological functionality and improved sensorial properties while S3 TAGs are responsible for the structure of the fat (Rodrigues and Gioielli, 2003). In addition, the contribution of oleic acid from MoO into the interesterified blends made them a potential candidate to be used as frying oils. As demonstrated by Abdulkarim et al. (2007a), MoO revealed great resistance during the frying procedure due to its high oleic acid content. Therefore, it could be proposed that the frying quality of PO enhanced after IE with MoO due to a lower degree of saturation and higher resistance. It could be also indicated that interesterified MoO/VCO fryng quality improved due to the MLCT generation and oleic acid contribution as the original VCO is not suitable for frying due to the high médium chain fatty acids and foam formation (Lee et al., 2012).

Figures 2 and 3 show the melting and crystallization thermograms of the interesterified MoO/PO and MoO/VCO blends, respectively. Generally, by increasing the proportion of MoO, the melting profile of both IE blends is supposed to shift towards lower temperature regions indicating a generation of TAGs with lower melting points. However, as a result of the S3 increase in interesterified MoO/PO, the corresponding melting profile shifted towards higher temperature regions (Table 1).

In MoO/PO interesterified blends, by increasing MoO, a new peak (peak 3) was generated in 50:50 w/w and 70:30 w/w. As IE progressed, the enthalpy of the major peak 1 in MoO/PO 30:70 and 50:50 w/w blends increased, while peak 2 decreased followed by an overlapping with peak 1. This indicates a reduction in TAGs as represented by peak 2 and a possible formation of more TAGs, which was represented by peak 1. However, in the MoO/PO 70:30 blend, peak 1 gradually disappeared and the TAG represented by peak 2 became the dominant TAG. This is as expected as the blend contained a higher quantity of MoO which, when compared to PO, has a higher degree of unsaturated FA. These findings indicated that the blends were composed mainly of low melting TAGs, and the TAGs representing peak 3 were contributed by MoO. This result is expected as the main FA of the blends is oleic acid and OOO is the main TAG. da Silva et al. (2010) reported that triunsaturated TAGs have melting points from −14 °C to 1 °C, and triolein has a melting point of −4 °C (Hynes, 2010).

In the crystallization thermograms of unreacted MoO/PO blends (Figure 2), three crystallization peaks (peaks 1, 2 and 3) were observed. However, peak 2 gradually disappeared when the proportion of MoO was increased over the time of IE. Concurrently, it was observed that the size of peak 3 increased and as IE progressed it shifted to a lower temperature region, indicating the synthesis of lower melting TAGs. From the cooling thermograms of MoO/PO 70:30 blends, two obvious crystallization peaks were observed: TAGs representing Peak 1, which crystallized at 3.36 °C, and those representing Peak 2, which crystallized at −49.97 °C. After 24 h IE, the TAGs in Peak 1 crystallized at a higher temperature (6.69 °C), whereas there was little change in the peak temperature for the TAGs represented by Peak 2.

The above results show that the IE of different blends of MoO/PO would produce TAG species with wide ranging crystallization and melting temperatures. The resulting blends were found to have higher end set temperatures than the melting temperatures before IE. However, the melting temperatures were still lower than the original MoO, except for the blends containing 30% MoO which melted at 22.97 °C. Abdulkarim et al. (2007b) reported the increase in melting temperature from 18.9 °C (control) to 20.5 °C after 24 h self-IE of MoO using Lipozyme IM 60. This was reflected in the crystallization temperature increase from 1.6 °C to 12.9 °C. Long et al. (2003) reported that the slip melting point of palm olein/flaxseed oil blends (90:10 w/w) also increased from 14.1 °C to 19.9 °C after 6 h IE using Lipozyme IM, indicating that the synthesis of higher melting point TAGs took place during the rearrangement of the FA moiety.

The results for the IE of the MoO/VCO blends (Figure 3) provide several important observations concerning the melting and crystallization profiles. In the melting profile of the 30:70 blend, a low melting Peak 2 represented unstable TAG crystals of low melting points that melted earlier and this peak then became broader and finally overlapped with the major TAG peak (peak 1) after 6–24 h of IE, resulting in peak 1 shifting to a lower melting region indicating the reduction of high melting TAGs due to oleic acid incorporation and a higher degree of unsaturation. The melting thermogram of the interesterified MoO/VCO 70:30 blend (Figure 3) displayed a well-defined peak which is related to the low, medium and high melting points of TAGs which melt in a wide temperature range. The major melting peak (peak 1) of 0 h blend occurred at 7.47 °C with shoulder peaks (peak 2) at −0.32 °C and a minor peak (peak 3) at −12.87 °C. After the IE of MoO/VCO 70:30 blend, the minor and major melting peaks shifted towards the low temperature region, resulting in a reduction in the melting temperature of the interesterified blend. This indicated that the rearrangement of FAs responsible for these peaks leading to the formation of lower melting TAGs after IE were most likely to be influenced by unsaturated FA (oleic acid incorporation) from MoO. For blends with lower proportions of MoO, the thermal properties of the interesterified oil revealed an influence of saturated FA rich in VCO.

Generally, in the MoO/VCO crystallization thermogram, the major crystallization peak (peak 1) was observed with a minor peak (peak B) for the 30:70 blend ratio. During the IE of MoO/VCO 30:70, the major crystallization peak decreased and the minor peak increased. Both results from the melting and crystallization thermograms indicate that a softer mixture has been produced.

The solid fat content (SFC) profile as a function of temperature for the MoO/PO and MoO/VCO blends, interesterified at various times and ratios, is shown in Table 3. The SFC of the blends at 0 °C is related to the degree of saturation: the higher the amount of MoO in the blends, the lower the SFC is at 0 °C. Likewise, the greater the quantity of either PO or VCO in the blends, the higher the SFC is at 0 °C. The obtained result indicates that unlike MoO/VCO, IE caused the formation of more saturated TAGs which then contributed to an increase in melting temperature in the MoO/PO blends which is in accordance with the results obtained in previous sections. In this agreement, Long et al. (2003) and Abdulkarim et al. (2007b) demonstrated elevated SFC after IE indicating the ability of the enzyme to form higher-melting TAGs.

The findings show that for each MoO/PO blend ratio, the temperature at which SFC is zero increased after IE, compared to the control (0 h). However, it still remained below ambient temperature (25–27 °C) or body temperature (35–37 °C). After 24 h IE, the SFC% of all MoO/PO ratios increased where 30:70 w/w blend revealed more obvious changes compared to other ratios. The major difference was observed at 10 °C and 15 °C with SFC values of 4.6% and 0% in unreacted (0 h) MoO/PO 30:70 w/w blend which increased to 11.1% and 6% after 24 h IE, respectively. As stated in the TAG section, due to the TAG rearrangements after the IE of MoO/PO, the level of S3 (MML) increased by up to 2.55% in the 30:70 blend. It could be concluded that 2.55% of S3 enhancement brings about 6.5% and 6% more SFC% in 10 and 15 °C, respectively. According to Rodrigues and Gioielli (2003), S3 TAGs are responsible for the structure of fatty foods.

Consequently, interesterified MoO/PO revealed a harder structured fat blend produced from two liquid oils due to the function of Lipozyme RMIM. Changes in the amount of TAG species after IE are normally accompanied by a change in the SFC, depending on the type of TAG formed (Debnath et al., 2011; Koh et al., 2010).

In interesterified MoO/VCO blends the level of S3 decreased following the increase in theU2S level resulting in the production of softer fat. Soares et al. (2012) also reported SFC reduction in a coconut oil-canola oil interesterified blend. At 20 °C, the interesterified MoO/VCO 50:50 blend showed 0% SFC while it was higher for the 30:70 blend and lower for 70:30 due to the difference in the oleic acid incorporation associated with the MoO level. At low temperatures (4–10 °C), where SFC gives an indication of the fat's spreadability, VCO originally do not offer satisfactory characteristics with SFC values of 89.8%. Ribeiro et al. (2009), Wassell and Young (2007) and Lida and Ali (1998) demonstrated that for spreadability at refrigeration temperatures, the SFC should not exceed 32% at 10 °C in order to meet the minimum plasticity requirement. The binary blends of 70:30 w/w MoO/VCO can be regarded as offering good spreadability at low temperatures (10 °C) with SFC values of 23.1 to 11.2% after 0 to 24 h IE, respectively. The product's stability and its resistance to oiling-out could be characterized in fats with a minimum of 10% SFC between 20 and 22 °C (Lida et al., 1998; Wassell and Young, 2007). The unreacted binary blend of MoO/VCO 30:70 w/w with 20.1% and 10.9% SFC at 20 and 22 °C, respectively and the interesterified blend after 6 h with 11.3% SFC fall within this group of fats.

Blends with zero SFC at 25 °C including MoO/PO 50:50 and 70:30 w/w after 12 h IE and MoO/VCO 30:70 are suitable for incorporation into the fat phase in ice-cream formulations as demonstrated by Liew et al. (2001). In addition, the MoO/VCO interesterified blends with elevated levels of MLCT may improve the quality of pastries as Ariffin et al. (2010) reported that cookies made from medium- and long-chain triacylglycerol-enriched margarines have good texture and sensory perception as well as excellent nutritional benefits.