1. Introduction
There is consistent evidence that
n-3 long-chain polyunsaturated fatty acids
(n-3 LCPUFA) produce important health benefits, evidencing their role as mediating bioactive lipids within the protective action exerted by diets rich in these compounds [
1]. Marine oils are widely used due to their high content of
n-3 LCPUFA, such as EPA and DHA, and their consumption reports a low prevalence of heart, circulatory and inflammatory diseases as they reduce the blood pressure in patients with systolic hypertension [
2]. A low
n-6/
n-3 ratio, together with the daily consumption of
n-3 LCPUFA, exerts a great influence on the prevention of non-communicable diseases [
3], as well as on the decrease in mortality and cardiac risk [
4]. The
n-3 LCPUFA are therefore considered as determining factors in the prevention and treatment of cardiovascular and neurodegenerative diseases [
5]. A diet rich in
n-3 LCPUFA plays an important role in the last trimester of pregnancy and in child nutrition, since DHA especially influences the development and functioning of the nervous system and visual organs of fetuses and newborns [
6]. According to the FAO [
7], the daily adequate consumption of EPA + DHA should be 250 mg, while in the case of children, daily 150 mg consumption would be necessary for optimal brain development. In addition to the consumption of fish, other products containing
n-3 LCPUFA are currently used, such as capsules and foods declared safe from a chemical, organoleptic and toxicological point of view.
In the digestive process, an important hydrolysis product of triacylglycerols (TG) by lipases is 2-monoacylglycerol (2-MAG), which is absorbed through the wall of the intestine. The formation of 2-MAG is due to the fact that lipases hydrolyze mainly the
sn-1 and
sn-3 positions of the glycerol backbone [
8]. Therefore, the specific distribution of fatty acids (FAs) in the glycerol molecule can play a key role in the digestion and absorption of lipids together with the physical state and their phase behavior in water [
9]. Therefore, the location of EPA and DHA in the
sn-2 position is essential to maximize their bioavailability. Commonly, sAG having their
sn-2 group substituted with a long-chain FA are characterized by their unique nutritional and physiological activity. Notably, the
n-3 LCPUFA at the
sn-2 position of TG with medium-chain FAs at the
sn-1 and
sn-3 positions are easily absorbed [
10]. This is due to the fact that pancreatic lipase at the digestive level releases medium-chain FAs from the extreme positions, which are quickly absorbed and used as an energy source, while 2-MAG is rapidly absorbed in the enterocyte [
11]. For this reason, diacylglycerols (DG) with a low energy intake are also considered adequate for the control of obesity and for people with fat malabsorption and other metabolic problems [
12].
Structured acylglycerols (sAG) or structured lipids (SL), whose composition and positional distribution of FAs in the glycerol skeleton have been modified, can be obtained by enzymatic catalysis reactions and/or genetic engineering. As a result, “tailor-made” lipids can be obtained with favorable physical characteristics and chemical properties, as well as nutritional benefits [
10]. In order to maximize the FA bioavailability, sAG should present a particular FA distribution at the glycerol skeleton [
13,
14]. Among biocatalysts, lipases have shown to be the most versatile on the basis of their regiospecificity for modifying oils and fats to obtain high-added-value products, such as EPA + DHA supplements, among other sAG. Lipases are defined as enzymes that hydrolyze FAs and alcohol ester bonds and catalyze esterification and transesterification reactions (acidolysis, alcoholysis and interesterification). Additionally, intra–interesterification including interchange of FAs in the same TG molecule backbone can occur. In general, hydrolysis takes place in media with a high water content, esterification takes place in media with a minimal water content and transesterification is efficiently catalyzed in a waterless mixture using immobilized enzymes. Lipases are widely distributed in animals, plants and microorganisms. Those of animal origin are obtained from gastric, intestinal and pancreatic tissues. Vegetable lipases abound in oilseeds (soybeans and peanuts), while microbial ones are produced mainly by fungi and yeasts such as
Aspergillus niger,
Mucor miehei,
Rhizopus delemar,
Geotrichum candidum,
Candida rugosa and
Candida antarctica. Lipases differ in their properties according to the organisms that originate them. The commercial preparations that are used for the modification of oils and fats mostly come from microorganisms [
15]. Lipases act on the carboxylic ester bonds present in TG to release FAs and TG. Interestingly, lipases have positional specificity for FAs in TG, which is particularly important in lipid modification [
16,
17]. Thus, their specificity towards different FAs bound to TG is affected by many factors such as the distribution of FAs in the three positions of the TG molecule [
18]. Notably, immobilized lipase B from
Candida antarctica (CALB) has shown to be nonspecific in the intra–interesterification process of TG, meaning that acyl groups are not introduced in specific positions of the glycerol since the reaction occurs randomly with respect to the position.
The employment of supercritical fluids (FSC) has shown many advantages to produce sAG on the basis that such fluids can be defined as a compressible matter, they behave like gases but have the density of a liquid (0.1–1 g/mL) and, therefore, they have solvent power [
19,
20,
21,
22]. Among supercritical fluids, carbon dioxide (CO
2SC) is found to be an ideal solvent for food applications. It is considered a Generally Recognized as Safe (GRAS) compound [
23], has a variable density and great solvent power and can be easily separated from the reaction medium by depressurization. The greatest advantages of CO
2SC over other liquid solvents are its high diffusivity, and low viscosity and surface tension, which allow it to accelerate mass transfer in enzymatic reactions [
20,
24]. When CO
2 is compressed at a temperature of 31.1 °C and a pressure of 72.9 atm above its critical point, it does not liquefy but reaches a dense gaseous state and behaves like a solvent. The use of CO
2SC for biologically active lipid synthesis significantly reduces the use of organic solvents, avoids waste removal problems, eliminates the use of potentially toxic and flammable solvents and reduces the reaction time. Notably, the use of lipases in CO
2SC to produce SL is to increase the solubility of lipid and hydrophobic substrates in nonpolar media, producing reverse reactions to hydrolysis and favoring synthesis such as esterification and transesterification [
25].
The main aim of the current study was the synthesis of sAG starting from commercial salmon oil. For carrying out the intra–interesterification, immobilized lipase B from
Candida antarctica (nonspecific) (CALB) Novozyme
R 435 under CO
2SC conditions was used. This commercial enzyme was chosen for being nonspecific, as well as for its avalaiblity and efficiency in this type of process. Furthermore, identification and checking of EPA and DHA presence at the
sn-2 position during oil fractionation were carried out by mass spectrometry (MS, MALDI-TOF). This analytical technique has shown to be an accurate tool for identifying the FA presence in oils as well as their locations in TG molecules [
26,
27,
28]. The novelty of the current study can be explained on the basis of using a combination of advanced processing conditions and analytical tools to obtain sAG compounds from commercial salmon oil susceptible to be employed in the human diet.
3. Discussion
In this study, sAG synthesis with EPA and DHA at position
sn-2 (EPA/DHA sAG) by enzymatic intra–interesterification was prepared starting from DRCSO. The characterization of DRCSO through oxidative stability analyses such as FFA, PV and
p-AV, among others, showed experimental values within the limits established by RSA, CODEX and GOED; this indicates that the current starting salmon oil is fit to be used. Commercial refined salmon oil evaluated by GLC presented an FA composition characteristic of marine oils. Regarding EPA and DHA, both acids are at the lowest limit of the ranges established for this type of oil with 3.92 ± 0.04 EPA (within a range of 2–6) and 3.83 ± 0.04 DHA (within a range of 3–10) (g/100 g TFA) [
32]. This is directly related to the fact that the level of these FAs will depend on the feeding conditions previously applied. One possibility for assessing the diet previously provided is the level of linoleic acid (9c, 12c-18:2) present in the oil; thus, a normal wild individual will have values of 1.5–2.5% of this FA, while cultured individuals can reach values as high as 8–15%. The oil presented 15.77 ± 0.07 g linoleic acid/100 g TFA, which indicates that an important part of the fish’s diet was a source of vegetable oil [
44]. This in turn would influence the low level of EPA and DHA, since
n-3 LCPUFA contents are higher in diets whose energy intake comes from fish oil [
45]. Thus, DRCSO presented physicochemical characteristics that allow its use as a raw material for fractionation in a CO
2SC medium. The effect of the intra–interesterification process under supercritical CO
2 conditions was also analyzed and was compared with the initial DRCSO. The DRCSO presented EPA in the
sn-2 position according to the analysis carried out by mass spectrometry. In addition, it was found that the presence of EPA and DHA and their position in the DG and TG of DRCSO can be identified with greater probability using the CHCA1 matrix by mass spectrometry. It was observed that EPA is more likely to be found in the
sn-2 position, while DHA is mostly located in the
sn-3 position in TG of DRCSO. The low probability of DHA in position
sn-2 may be due to the larger size of this LCPUFA with respect to EPA. The EPA/DHA sAG synthesis obtained by intra–interesterification caused changes in the sAG FA composition and content of the fractionated samples obtained with respect to the initial oil when the reaction was catalyzed by the immobilized lipase B enzyme obtained from
Candida antarctica (Novozyme
R 435). The fractionated samples obtained at different times in CO
2SC media also showed changes in the position of EPA and DHA when analyzed by mass spectrometry (MALDI-TOF). Notably, the third fractionation, whose CO
2SC conditions were 120 min, 40 °C and 140 bar, presented the highest yield of EPA or DHA in the
sn-2 position in sAG during the intra–interesterification process.
In addition, the presence of MG and DG with DHA and EPA in position
sn-2 in EPA/DHA sAG was identified, a figure higher than that observed in the same matrix (CHCA1) of the initial oil sample, which shows the effect of fractionation during the intra–interesterification process of the last extraction. However, MG, DG and TG of EPA/DHA sAG in Extraction 2 (60 min, 40 °C and 140 bar) in the CHCA1 matrix could not be identified, showing a total disappearance of these compounds in the fractionated samples when analyzed by mass spectrometry. Thus, the mass spectrometry results did not show any probability of TG formation being EPA/DHA sAG compounds formed under the extraction conditions used, although the contrary was observed in the TLC results. However, this may be due to the efficiency of the matrix used for the SL. Thus, they show a different behavior from that observed in the initial oil, according to the fact that visualization of the compounds requires the use of a matrix, which would be one of the factors that favor the fragmentation of the compounds [
14,
46]. Additionally, the ionization process, although corresponding to a soft type, can induce the fragmentation of TG to originate DG or FFA, artificially increasing the content of the samples [
47,
48]; this fact would explain the difference between values observed in TLC of EPA/DHA sAG and Extraction 3. However, the possibility of greater compounding with DHA and EPA in the
sn-2 position is an advantage with respect to the initial oil, since it shows that the extraction conditions in a supercritical fluid (temperature, time and pressure) can be manipulated to exert changes in the positioning of EPA and DHA.
4. Materials and Methods
Deodorized and refined commercial salmon oil (DRCSO) was provided by Fiordo Austral S. A. (Santiago, Chile). This oil corresponds to the production line authorized to export to the European Community in accordance with the current regulations of the program of quality assurance for human consumption of SERNAPESCA based on NCh 2861:2004. Fatty acid methyl ester (FAME) standards, fatty acid (FA) standards and C23:0-methyl ester (2COT N-23M-A29-4 NU-CHECK-PREP-INC) were purchased from NU-CHEK PREP, INC (Elysian, MN, USA). All solvents and chemicals used (including urea, ethanol, α-tocopherol and n-hexane) were of analytical grade (Merck, Santiago, Chile). DRCSO was stored at −70 °C under nitrogen atmosphere until being employed.
4.1. DRCSO Characterization
DRCSO was characterized by means of different chemical analyses (
Table 1). The standard AOCS official method [
49] procedure was employed for the following assessments: free fatty acids (FFA) content (official method Ca 5a-40), peroxide value (PV; official method Cd 8b-90),
p-anisidine value (AV; official method Cd 18–19), TOTOX value (official method Cg 3–91), insoluble impurities content (official method Ca 3a-46), unsaponifiable matter (UM) content (official method Ca 6b-53) and moisture and volatile matter contents (official method Ca 2d-25). Additionally, conjugated diene (CD) and triene (CT) formation was measured at 233 nm and 268 nm, respectively [
50], the results being expressed in agreement with the following formula: CD (or CT) = B × V/w, where B is the absorbance reading at 233 (or 268) nm, V is the volume (mL) and w is the mass (mg) of oil measured.
4.2. DRCSO Fatty Acid Composition
The DRCSO FA profile and EPA/DHA quantification were assessed in an HP 5890 series II GLC with a flame ionization detector (FID) with the injection system split. A fused silica capillary column (100 m length × 0.25 mm × 0.2 μm film thickness) coated with SPTM-2560 (Supelco, Bellefonte, PA, USA) was used. DataApex ClarityTM software (DataApex Ltd., Prague, Czech Republic) for chromatogram analysis was applied. A methylation process was performed to obtain FAMEs. For this, a two-step process was performed according to previous research. The reference standard NU-CHEK GLC463 was used to identify the FA profiles by comparing the retention times [
51]. The integration of the chromatographic peaks was carried out from baseline to baseline. The concentration of the different FAMEs was determined from the calibration curves by assessment of the peak/area ratio. The quantification of all the individual FAs (g/100 g TFA) was achieved by employing C23:0 methyl ester as the internal standard according to the AOCS method [
52].
4.3. Preparation of EPA/DHA sAG by Enzymatic Intra–Interesterification Process under CO2-Supercritical Conditions (CO2SC) Using Different Extraction Times
EPA/DHA sAG synthesis by enzymatic intra–interesterification was prepared starting from DRCSO. For the enzymatic intra–interesterification, a supercritical CO
2 reactor Speed SFE system model 7071 (Applied Separation, Allentown, PA, USA) was used, with the following conditions: 20 mL of substrate; supercritical temperature of 40 °C; times of 30, 60 and 120 min; and supercritical pressure of 140 bar. The immobilized lipase B from
Candida antarctica Novozyme
R 435 was employed for the current study for being nonspecific and for its availability and efficiency. This enzyme was maintained at 20% of the substrate [
22,
24,
53,
54]. According to the reaction time employed, three different fractions were obtained.
4.4. Purification of EPA/DHA sAG by Neutralization with NaOH
The EPA/DHA sAG obtained by the enzymatic intra–interesterification process under CO2SC were purified by FA neutralization with NaOH to remove the remnant FFA of the reaction and then collected in hexane for GLC analysis. For this, mixtures with ethanol and phenolphthalein, titration with sodium hydroxide and washes with hexane were carried out. The purification state of each sample was followed by TLC.
4.5. Fatty Acid Composition and Quantification of Fractionated EPA/DHA sAG
Fractionation was carried out maintaining the process conditions (40 °C and 140 bar) for three periods of time (30, 60 and 120 min), and the EPA/DHA sAG samples were extracted to be evaluated by GLC, similar to
Section 4.2.
4.6. Identification of EPA/DHA sAG and DRCSO by TLC
EPA/DHA sAG samples and DRCSO were identified by TLC on Silica gel 60 F254 plates (Merck, Santiago, Chile). A solution of hexane, diethyl ether and glacial acetic acid (80/20/2,
v/
v/
v, respectively) was used as eluent. The order of elution on the chromatographic plate from bottom to top was evaluated according to the compounds’ decreasing polarity: MG, DG and TG [
55]. The plates were stained with iodine solution so that unsaturated fatty acids may be visualized.
4.7. Positional Analysis of EPA and DHA by Mass Spectrometry (MALDI-TOF)
The location of EPA and DHA in the resulting glycerol backbone was detected by mass spectrometry (MALDI-TOF) analysis. For the determination of EPA or DHA in the
sn-2 position of the EPA/DHA sAG, samples were analyzed before and after fractionation. After purification, aliquots were taken to be analyzed by mass spectrometry in a “Matrix-Assisted Laser Desorption/Ionization-Time-Of-Flight” (MALDI-TOF) Microflex (Bruker Daltonics Inc., Billerica, MA, USA) instrument in positive ion mode by reflection detection. For the analysis of the spectra, the mMass version 5.5.0 program was used according to the protocols of Strohalm et al. [
26], Strohalm et al. [
27] and Niedermeyer and Strohalm [
28].
Working solutions of the oils with a concentration of 1.0 mg/mL were prepared in chloroform/isopropanol 1:1. The 5-chloro-2-mercaptobenzothiazole (CMBT1) matrix was prepared at a concentration of 10.0 mg/mL in methanol. The working solutions and the matrix were mixed in a 1:1 ratio; then, 0.7 μL was deposited on a micro-scout sample plate (Bruker Daltonics Inc., Billerica, MA, USA). For the detection of the monoisotopic m/z signals of the acquired spectra, the MALDI-TOF peptides algorithm was used (signal/noise ratio of 3.0 and a relative intensity limit of 0.1%). For the identification, the detected monoisotopic m/z signals were analyzed and assigned through the Match & Annotate option of the Compound Search tool by comparing them with the theoretical monoisotopic masses of different types of glycerolipids (GL) contained in the base of LIPID MAPS (data version 11/16/2013). Subsequently, the coincidences observed were examined manually, and in each case, the experimental isotope distribution was compared with the theoretical one through the Show Isotopic Pattern option. The same procedure was performed by mixing the samples with the α-cyano-4-hydroxycynamic acid (CHCA1), 2,5-dihydroxybenzoic acid in methanol (DHB1) and 2,5-dihydroxybenzoic acid in methanol and trifluoroacetic acid 0.10% v/v (DHB2) matrices.
4.8. Statistical Analysis
An ANOVA of the parameters with a significance level of p ≤ 0.05 was performed according to Fisher’s method (LSD). The 95% confidence intervals of each quality parameter were calculated, considering the number of replicates and the standard deviation of each sample. The statistical program Statgraphics Centurion XVI-2011 (Stat Point Technologies, Inc., Rockville, MD, USA) was used.