Abstract
The objective of the present study was to investigate the marine n-3 phospholipids present in commercial New Zealand fish co-products (king salmon head, roe and skin; blue mackerel head, roe, skin and male gonad; hoki roe; gurnard skin and head; ribaldo skin and head; snapper skin and head), and to investigate the effect of pulsed electric fields (PEF) and protease pre-treatments on total lipid yield and quality. Total lipid was extracted from the co-product samples using a newly developed hexane/ethanol based solvent system called ETHEX, which was found to be better for phospholipid extraction, compared to the FOLCH and PALC methods in terms of processing time (ETHEX: 1 hour ; PALC: 15 h; FOLCH : 12 h), total lipid extraction (ETHEX:14.3%; PALC:14.6%; FOLCH : 12.9% of wet tissue), and extraction of crude phospholipid in lecithin (ETHEX:4.95% ; PALC:4.89%; FOLCH : 3.15% of wet tissue) from hoki roe. Fatty acid composition and phospholipid profile were determined using gas chromatography with flame ionization detection (GC-FID) and nuclear magnetic resonance spectroscopy (NMR), respectively. The effect of various NMR parameters (NMR solvent, buffer pH, relaxation delay and scan number) were investigated using commercial phospholipid standards and the NMR analysis was validated by quantification of phospholipids in some common natural matrix materials (chicken egg, soybean lecithin and Krill oil), for which values are reported in the literature. The optimised NMR conditions (solvent: D2O, pH: 7.4, scan number: 192 and relaxation delay: 3.5 s) were used to investigate the phospholipid profile of the fish co-product samples. The co-product samples were divided into three different groups, comprising fatty fish (salmon), moderately fatty fish (blue mackerel), and lean fish (gurnard, ribaldo and snapper). A comprehensive lipidomic profile (total lipid, phospholipid yield and lecithin recovery, lipid fractionation content (NL, neutral lipid; PL, phospholipid), fatty acid composition in total lipid, NL and PL, and the positional distribution (sn-1,3 and sn-2) of n-3 fatty acids (EPA, DPA and DHA) in all co-product samples were analysed.
On the basis of g wet tissue, blue mackerel roe was found to be a better source of phospholipids (38.6 µmol), compared to head (9.89 µmol), skin (13.5 µmol), and male gonad (10.0 µmol) of the same species. In general, across all fish species co-products investigated, the total lipid extracted from roe was found to have a higher proportion of n-3 fatty acids (44.4%), including EPA (11.3%) and DHA (27.5%), compared to head (total n-3 = 36.6%; EPA, 9.08%: DHA, 21.9%), skin (total n-3 = 34.8%; EPA, 9.63%; DHA, 19.5%) and male gonad (total n-3 = 42.5%; EPA, 12.1%; DHA, 24.7%). Similar to blue mackerel, king salmon roe was found to contain the highest amount of phospholipid (26.53 µmol/g) and n-3 fatty acids (43.32%), followed by head (PL = 10.76 µmol/g; n-3 = 7.21%) and skin (PL = 4.98 µmol/g; n-3 = 8.23%). Among the three lean fish species (gurnard, ribaldo and snapper), gurnard head (GH) and snapper head (SnH) were found to contain a higher amount of total lipid (5.9-6.3%) than other samples (1.2-3.9%), including a considerable amount of bioactive n-3 fatty acids such as EPA (GH = 9.05%; SnH = 5.06%), DPA (GH = 2.78%; SnH = 2.93%) and DHA (GH = 12.8%; SnH = 7.72%) in the polar lipid fraction. Among all of the studied co-products, hoki roe was found to be the best source for n-3 phospholipid, and hence hoki roe was used to investigate the effect of PEF pre-treatment, with three levels of PEF field strength (0.625, 1.25, and 1.875 kV) and frequency (25, 50, and 100 Hz), as well as protease pre-treatment, with three different protease preparations (Alcalase, FP-II and HT), on total lipid yield and quality. The roe treated with high field strength (1.875 kV/cm) and frequency (100 Hz) PEF was found to improve the total lipid yield without compromising the lipid quality, based on a TBARS value of <5 mg MDA/kg lipid and peroxide value (PV) of <5 mEq peroxide/kg lipid, and that there was no significant variation in total n-3 fatty acids, EPA, DPA, or DHA, compared to the control. However, this PEF treatment caused sn-2 phospholipid EPA and DHA to be relocated to the sn-1,3 positions. Despite the good yield of n-3 fatty acids and PL, the highest PEF intensity was found to result in negative structural changes in hoki roe lipids.
Total lipid yield was also found to be improved by protease pre-treatment with Alcalase (16-20%), FP-II (16-18%) and HT (16-18%), compared to the control (14-15%). Alcalase was found to exhibit the highest proteolytic capabilities with both casein and hoki roe protein as substrates, followed by HT and FP-II, hence Alcalase generated more extensively hydrolysed protein in hoki roe. Although Alcalase was found to result in the extraction of the highest total lipid, compared to that of HT and FP-II, it was found, however, that the fatty acid content and yield of phospholipids were decreased with an increase of Alcalase concentration used in the pre-treatment hydrolysis. There is the possibility that hydrophobic peptides in the Alcalase hydrolysate may have partitioned into the lipid fraction, and it may be that a higher proportion of bioactive peptides were generated in the hydrolysate that helped to limit lipid oxidation. Overall, the study presented in this thesis concludes that, while all of the fish co-products investigated contain substantial amounts of health-promoting lipids, fish roe is the best source of bioactive n-3 phospholipids, and the extracted yield could potentially be improved by using PEF and protease pre-treatments. The results obtained should be of interest to the fish processing and pharmaceutical industries, and also to food chemists with interests in bioactive fatty acids and nutraceuticals derived from marine sources.