The results of this study have shown (Figs. 2 and 3) that the concentration of hydroperoxides reaches a plateau before there is any significant increase in the CL intensity for both LA and ML at 60°C. This period of constant hydroperoxide concentration indicates that the decomposition of the hydroperoxides is occurring at a similar rate to their formation and the low level of CL would be due to only a small proportion of the hydroperoxides decomposing. It would thus be expected that as the concentration of hydroperoxides eventually decrease, due to the depletion of LA or ML, the CL would reach a peak.
The increased rate of lipid peroxidation at elevated temperatures (60°C in the present study) is most likely due to a thermal homolysis reaction (25). The activation energy of this process is the bond dissociation energy, so that the rate of homolysis would increase as the thermal energy of the system increases. The C-H bond adjacent to the two double bonds in LA and ML is relatively weak, increasing the likelihood of homolysis (26). The temperature also has a marked effect on the propagation of lipid peroxidation, where there was found to be almost an absence of chain reactions in LA at 0°C increasing to an average chain length of 7 at 48°C.
Initially the formation of hydroperoxides in these experiments should be due to the same initiation reactions that would occur at elevated temperatures. Radiolysis of the aqueous environment does produce highly reactive oxygen species such as OH· and H2O2 which are capable of initiating lipid peroxidation (25). However the production and lifetimes of radicals such as OH· are so small that in a dilute solution of micelles the chances of reaction with LA or ML would be very low. Thus the initial increase in hydroperoxide concentration is slower than for the experiments at 60°C, as would be expected at a lower temperature (Figs. 4-6). However, as the concentration of hydroperoxides increases, the effect of the glucose catalyst becomes apparent. The enol form of glucose attacks the hydroperoxides already in solution producing radicals which can accelerate the formation of further hydroperoxides. Hence the concentration of hydroperoxides increases more rapidly than in a similar system with no glucose (Figs. 4-6).
The results with glucose (Figs. 5 and 6) show that the hydroperoxide concentration does not plateau as when oxidation occurs at 60°C, but rather reaches a peak, corresponding (16) to a concentration of 4.2 x 10-5 mol dm-3, and then decreases. The peak concentration is the same for both ML and LA. At this concentration the rate of decomposition of the hydroperoxides must overtake the rate of formation. The CL increases most rapidly when the concentration of hydroperoxides are falling, which is in agreement with the accepted mechanism of excited state formation during the breakdown of hydroperoxides (3).
Metal catalysis of lipid oxidation has been recognised and studied for many years and the reaction mechanisms are often complex (3, 28-31). In this study, Fe2+-NTA was used as a catalyst for the peroxidation of LA, and similar results were obtained to those using glucose (Fig. 7 compared to Fig. 5). The main difference was that both the hydroperoxide and the CL increase occurred earlier for LA catalysed by iron than by glucose. In addition, the CL intensity increased more rapidly and reached a higher level using iron as the catalyst. These results indicate that the Fe2+-NTA is a more efficient catalyst of the oxidation of LA than glucose and that, unlike glucose, it is capable of catalysing the initiation process as well as the decomposition process. The concentration of hydroperoxides also peaks at a slightly lower level than with glucose, which may also be due to the ability of the iron complexes to initiate and catalyse lipid peroxidation more effectively than glucose.
Fe2+-NTA showed a marked acceleration of the peroxidation of LA compared to the system without a catalyst. The presence of iron accelerates the formation of hydroperoxides by a number of initiation mechanisms. The effect of chelating Fe2+ with NTA is to lower the reduction potential of iron and accelerate the oxidation of Fe2+ to Fe3+ in chelated form (23). In the current work, oxygen was readily available and thus some of the Fe2+ would have been partially oxidised to Fe3+. The autoxidation of Fe2+ to Fe3+ can initiate lipid peroxidation via a Fenton reaction (29) and Fe3+ alone, can initiate lipid peroxidation by direct electron transfer. The availability of both Fe2+ and Fe3+ allows redox cycling (29) of iron to occur, increasing the rate of lipid peroxidation. The Fe2+ is also capable of decomposing the hydroperoxides already formed to produce radicals (29,30) capable of reacting with other fatty acid molecules producing more hydroperoxides.
The results of the current work show that the CL increase is associated with the decomposition of hydroperoxides. In every case, the concentration of hydroperoxides had peaked and were level or falling before the CL rose most rapidly. This is in agreement with Cadenas (3) who proposed that the decomposition of lipid hydroperoxides lead to peroxyl radicals which recombine to produce excited species, which are responsible for the light emission. A close relationship between CL and malondialdehyde formation has been established (4,14,15). In the current study, the formation of aldehydes and ketones(including malondialdehyde) measured as absorption of light at 270-280nm appeared as CL began to increase most rapidly, but did not show a similar rate of increase. Noll et al. (15) suggest that the decomposition of membrane lipid peroxyl radicals to malondialdehyde and the formation of electronically excited species proceed via different pathways. The hydroperoxide precursors to malondialdehyde decompose early in lipid peroxidation while the remaining hydroperoxides accumulate before decomposing to produce mainly photoemissive species. The current results show that the concentrations of secondary products remain level while the CL is rising, and thus the CL is more directly related to the breakdown of hydroperoxides than to the formation of secondary products such as malondialdehyde.
Throughout the current work, the indicators associated with the degradation of fats and oils, such as the development of smell and colourisation, became apparent as the levels of hydroperoxides increased. These occurred before the increase in CL. It has been reported (8,9) that the CL intensity of food oils is closely correlated to their deterioration. These workers found that the emission intensity was closely correlated to oxidative deterioration and concluded that CL would be a good measure of the quality of foods. However, their results show that initially, CL drops or remains level while oxidative deterioration increases. In addition, for some foods and oils studied the CL actually decreased with deterioration. The results obtained in the present study show that degradation is not related to CL directly, but occurs before emission begins to increase. In fact, the degradation is related more to the formation of hydroperoxides as initial products of oxidation, and therefore, quantification of the level of hydroperoxides in the system would be a better indication of the oxidative deterioration of food products.
Cash et al. (7) estimated the shelf life of various vegetable oils as the length of the CL induction period. The induction period being the period before CL begins to increase non- linearly. They suggested that traditional techniques for measuring shelf lives, such as monitoring the levels of secondary products of oxidation, underestimate shelf lives at low temperatures. However, the smell and colourisation of the model system used in this study appeared well before the end of the CL induction period. These results suggests that the CL measurement would, in fact, be an overestimate of the shelf life.
Biological systems of lipids are a great deal more complex than those found in vegetable oils. The process of lipid peroxidation has been associated with human disease and toxicology (3). The CL from a wide range of biological systems, including tissue and blood plasma has been studied (32). CL measurement could provide a non invasive technique for diagnosing various medical conditions. The source of emission and link to the actual mechanism of excited species formation and hence the relationship to disease metabolism would need to be determined. Model systems such as that used in this study may provide useful information on the actual chemical processes involved. In this study glucose and Fe2+-NTA were found to catalyse lipid peroxidation and CL. Iron toxicity in biological systems, for example, could possibly be determined by CL measurement.
The simultaneous monitoring of lipid hydroperoxide concentration and light emission in this study revealed that the CL, arising from the generation of electronically excited species during lipid peroxidation, was closely associated with the decomposition of these hydroperoxides. These results support the generally accepted source of chemiluminescent species, which is through the breakdown of lipid hydroperoxides into peroxyl radicals which recombine to produce excited species (3).
Acknowledgements. - Financial support for this project from the Internal Grants Committee of Victoria University of Wellington is gratefully acknowledged.
21 June 1999