TEL: +61-3-9731-3261; FAX: +61-3-9731-3250; EMAIL: Chakra.Wijesundera@csiro.au



The effect of the positional distribution of docosahexaenoic acid (DHA) within triacylglycerol (TAG) molecules on the oxidative stability of oil-in-water emulsions was investigated by using a synthetic TAG regio-isomer pair SDS (1,3-dioctadecanoyl-2-[4,7,10,13,16,19-docosahexaenoyl] glycerol) and SSD (1,2–dioctadecanoyl-3-[4,7,10,13,16,19-docosahexaenoyl] glycerol), where S and D represent stearic acid and DHA, respectively. Oil-in-water emulsions (10%, w/w) of each isomer were subjected to accelerated autoxidation by continuous exposure to air at 50C in the absence of light. Oxidation during the exposure (storage) was monitored by measuring a series of volatile compounds characteristic of DHA oxidation. SSD emulsion oxidized faster than SDS emulsion, showing that DHA is more stable to oxidation when located at the sn-2 position of the TAG compared with the sn-1(3) position. This regio-isomeric effect is similar to that previously reported for bulk oil oxidation.


Many of the food products that have been selected for fortification with omega-3 oils such as milk, yoghurt, salad dressings and juices are oil-in-water emulsions. This study, for the first time, demonstrated that the regio-isomeric effects on oxidative stability of docosahexaenoic acid observed for bulk oil also apply to oil-in-water emulsion. Thus, potential exists for enhancing the oxidative stability of omega-3-fortified emulsion foods through modification of triacylglycerol structure.


Several decades of epidemiological and clinical studies have generated a plethora of evidence demonstrating various health benefits of increased dietary intake of omega-3 long-chain polyunsaturated fatty acids (n-3 LC-PUFAs), in particular eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3). The early evidence was related to a reduced risk of cardiovascular disease upon consumption of n-3 LC-PUFA (Schacky and Harris 2007) and their beneficial effects continue to emerge in other areas such as brain and vision development in infants, inflammatory diseases and mental health (Ruxton et al. 2007). The mounting scientific evidence showing beneficial health effects of n-3 LC-PUFAs has prompted leading cardiac and nutrition societies to advocate increased dietary intake of n-3 LC-PUFAs (Harris 2007).

Dietary n-3 LC-PUFAs are almost exclusively obtained from fish and other seafood. Fish consumption, particularly in the Western countries is not adequate to provide the recommended dietary intake (400–600 mg/day, Harris 2007) of n-3 LC-PUFAs. Fortification of food products with n-3 LC-PUFAs is considered one of the viable solutions to this problem, and this has spurred on the food industry around the world to launch thousands of novel omega-3-fortified food products during the past few years. The fortified products cover a wide range of popular foods, including dairy, spreads, bakery products, dressings, juices, confectionery and meat products (Shahidi 2008).

The n-3 LC-PUFAs, including EPA and DHA, are extremely susceptible to oxidative deterioration resulting in rapid development of off-flavor, and consequently low palatability and reduced shelf life of the products in which they are present. This problem has been overcome by protecting the oil from oxidation through microencapsulation before incorporation into food. However, the process of microencapsulation adds to the production cost, and alternative measures of making n-3 LC-PUFAs more resistant to oxidation would be useful.

Recent studies in our laboratory have shown that the positional distribution of DHA within the triacylglycerol (TAG) molecule significantly affects the oxidative stability of DHA-containing TAG in the bulk form (Wijesundera et al. 2008). However, many of the food products targeted for fortification with n-3 LC-PUFAs are emulsions, and here we report on the positional effects of DHA on the oxidative stability of oil-in-water emulsions.



The model TAGs 1,3-dioctadecanoyl-2-(4,7,10,13,16,19-docosahexaenoyl) glycerol (SDS) and 1,2–dioctadecanoyl-3-(4,7,10,13,16,19-docosahexaenoyl) glycerol (SSD) were synthesized in greater than 95% regio-isomeric purity according to the chemo-enzymatic procedure reported previously (Fraser et al. 2007). Tween 40 was purchased from Sigma-Aldrich (Sydney, NSW, Australia). The solid phase extraction cartridges and carboxen/polydimethylsiloxane solid phase microextraction (SPME) fibers (75 µm) were purchased from Phenomenex (Sydney, NSW, Australia) and Supelco (Sydney, NSW, Australia), respectively.

Purification of Model TAG

The synthetic model TAG regio-isomers were purified immediately before use to remove any oxidation products and impurities, which could potentially influence oxidation rates. For this purpose, the TAG (5 g) was completely dissolved in warm hexane/diethyl ether (95:5 v/v, 50 mL) and carefully loaded onto an amino solid phase extraction cartridge (Strata SI-1 Silica, 55 µm, 70 A, 20 g/60 mL Giga Tubes; Phenomenex, Lane Cove, NSW, Australia), which had been preconditioned with hexane/diethyl ether (95:5 v/v, 80 mL). Pure TAG without any contaminants (as verified by thin-layer chromatography and high-performance liquid chromatography) was obtained by eluting the cartridge with the above solvent (350 mL) and subsequent removal of the solvent under vacuum.

Preparation of Emulsions

Emulsions of SSD and SDS in water were prepared by homogenizing each TAG (5 g) with an aqueous solution of Tween 40 (1%, w/w, 95 g). The Tween solutions were prepared in high-purity (Millipore) water. Each TAG was melted and homogenized with the Tween solution by using an Ultraturax mixer (8,000 rpm, 3 min) to first obtain a coarse emulsion, which was then passed through an Emulsiflex C5 high homogenizer (Avestin, Ottawa, ON, Canada). Two passes at 400 bar for 10 min furnished the final emulsion. The loop of the latter homogenizer was maintained at 55C by using a water bath.

Particle Size Measurement

The particle size distribution of the freshly prepared SSD/SDS emulsions was determined by using a Malvern Mastersizer 2000 (Malvern Instruments Ltd., Malvern, U.K.). A particle refractive index of 1.456 and absorption of 0 was assumed for all emulsions. Samples were stirred and added to circulating distilled water to obtain an obscuration of approximately 15%. All measurements were carried out in triplicate at room temperature (approximately 22C). Typical particle size distribution values showed that both SSD and SDS emulsions were unimodal with narrow distributions and had small particles with particle diameters (D[3,2]) of 0.120 and 0.128 µm for SDS and SSD, respectively. These values were within the 0.1–100 µm range applicable to general food systems (McClements and Decker 2000).

Accelerated Oxidation

The SSD and SDS emulsions were each placed in open round-bottom flasks (50 mL) and heated inside a dark oven at 50C. Samples of the emulsions (1 mL, in triplicate) were withdrawn and transferred into 10-mL amber-colored glass headspace vials after 1.25, 2.75, 4.25, 6.25, 7.75, 9.25, 10.75 and 12.25 h, and subjected to headspace analysis for volatile lipid oxidation products.

Headspace Analysis

Volatile compounds generated during oxidation were determined by headspace SPME (carboxen/polydimethylsiloxane fiber) combined with gas chromatography-mass spectrometry (GC-MS). The fiber was inserted into the sample headspace and the vials incubated at 50C for 15 min. The fiber was then withdrawn and transferred to the GC injector (operated in the splitless mode) and held for 7 min to desorb the extracted volatile compounds into the GC column. The entire series of events was performed by using a CombiPAL Auto Injector (CTC Analytics, Zwingen, Switzerland). GC-MS was performed using an Agilent Model 6,890 GC and Model 5973 MSD (Palo Alto, CA) fitted with a VOC fused silica capillary column (60 mm, 0.32 mm i.d., 0.18 µm film thickness; Agilent, Melbourne, Vic., Australia). The GC oven was programed from 40C increasing to 220C at the rate of 22C/min and held at that temperature for further 14 min. Helium was used as the carrier gas at a constant flow rate of 2.0 mL/min. The injector was initially operated in splitless mode and then switched to split mode (1:20) 2 min after sample injection. The temperature of the injector and the MS detector were both held at 230C. The MS was operated in scan mode (29–250 amu). Data analyses were performed using Chemstation software, and compounds were identified by reference to a library of spectra (Wiley 275). Volatile compounds were quantified using the abundance of specific target ions.

Statistical Analyses

Analysis of variance (ANOVA) was performed using the statistical package MINITAB release 14.


The hydroperoxides formed by autoxidation of DHA undergo decomposition to form a range of volatile compounds, which can be used to monitor the progress of its oxidation. There were no qualitative differences in the profiles of the volatile products from the autoxidation of the TAG regio-isomers SDS and SSD. The ion chromatogram shown in Fig. 1 is a typical representation of the volatile compounds observed in the headspace of the model emulsions after accelerated oxidation. The most abundant and consistently occurring volatiles as detected by headspace SPME analysis were propanal, 2-propenal, 2-butenal, 1-penten-3-ol, 1-penten-3-one, 2-ethylfuran, trans-2-pentenal, cis,trans-2,4-heptadienal and trans,trans-2,4-heptadienal. As DHA was the only unsaturated fatty acid present in the TAG used for making the emulsions, all of the volatile oxidation products named above can be regarded as products of DHA oxidation. Thus, any one of them could potentially be used to monitor oxidation of DHA and other n-3 PUFAs. Propanal is the most widely used for this purpose (Boyd et al. 1992; Min et al. 2003; Venkateshwarlu et al. 2004; Senanayake and Shahidi 2007), probably because of its relatively easy detection, although the precision of its measurement is not always satisfactory (Iglesias et al. 2007). Other volatile products that have been suggested as markers for oxidation of DHA and similar n-3 LC-PUFAs include 1-penten-3-ol (Let et al. 2007), 1-penten-3-one (Aidos et al. 2002; Lee et al. 2003; Let et al. 2007) and trans,trans-2,4-heptadienal (Aidos et al. 2002; Lee et al. 2003; Lyberg and Adlercreutz 2006; Let et al. 2007). In the present study, we used propanal, 1-penten-3-ol, 1-penten-3-one, cis,trans-2,4-heptadienal and trans,trans-2,4-heptadienal to monitor oxidation of the SDS and SSD emulsions.

Figure 1.

The volatile compounds were analyzed by headspace solid phase microextraction combined with gas chromatography-mass spectrometry. Identification: peak 1: propenal and propanal, peak 2: 2-butenal, peak 3: 1-penten-3-ol, peak 4: 1-penten-3-one, peak 5: 2-ethylfuran, peak 6: trans-2-pentenal, peak 7: cis,trans-2,4-heptadienal and peak 8: trans,trans-2,4-heptadienal.

Figures 2 and 3 show the development of propanal and trans,trans-2,4-heptadienal, respectively, in SDS and SSD emulsions subjected to autoxidation at 50C for up to 12 h. Similar graphs were observed for 1-penten-3-ol, 1-penten-3-one and cis,trans-2,4-heptadienal (not shown). All of these secondary oxidation products appeared in the headspace of the emulsion as soon as the oxidation began without any apparent induction period. Oxidation without an apparent induction period has been observed previously for DHA-rich fish oils oxidized at 50C (Frankel et al. 2002). The concentrations of the volatile oxidation markers were significantly lower (P < 0.001) in the SDS emulsion than in the SSD emulsion throughout the oxidation. After 12 h oxidation when sampling was stopped, propanal (Fig. 2) and 1-penten-3-ol (not shown) were still developing in the SDS emulsion, whereas their concentrations appeared to have peaked for SSD. The measured concentration of the secondary oxidation products is dependent on the amount of primary oxidation products (monohydroperoxides) generated and the amount converted to products such as free fatty acids. Hence, in the normal course of oxidation, the concentrations of secondary oxidation products increase to a maximum before falling off. In general, the TAGs most susceptible to oxidation are the quickest to reach the concentration maxima for the secondary oxidation products. The results of the accelerated autoxidation study clearly demonstrated that the SDS emulsion oxidized more slowly than the SSD emulsion. This implies that DHA is more stable to oxidation when located at the sn-2 position of the TAG compared with the sn-1(3) position.

Figure 2.

Concentration of propanal is expressed in arbitrary units as abundance of ion 58 amu. The error bars represent standard deviations for triplicate analyses.

Figure 3.

Concentration of trans,trans-2,4-heptadienal is expressed in arbitrary units as abundance of ion 81 amu. The error bars represent standard deviations for triplicate analyses.

Oxidation in oil-in-water emulsion is considerably more complex than oxidation in bulk oil. A multitude of factors influence oxidation of a given lipid in emulsion, including droplet size, emulsifier type, interfacial properties and the presence of metal ions, prooxidants, metal chelators and antioxidants, among others (McClements and Decker 2000). In the present study, the effects of these factors on oxidation were kept constant by preparing the two emulsions in identical fashion. The same emulsifier and emulsification conditions were employed for the preparation of the emulsions, and the droplet size was not significantly different. The oxidation of the emulsions was also performed under identical conditions. Hence, the observed differences in oxidative stability can only be attributed to the difference in DHA location in TAG.

To the authors' knowledge, this is the first time that TAG regio-isomerism has been shown to be a factor influencing lipid oxidation in emulsions. This effect has previously been demonstrated for bulk oil oxidation, where PUFA was observed to oxidize more slowly when attached to the sn-2 position compared with the sn-1(3) position in TAG (Wada and Koizumi 1983; Endo et al. 1997; Wijesundera 2008; Wijesundera et al. 2008). The present study has shown that this regio-isomeric influence on oxidative stability extends to oil-in-water emulsions stabilized with Tween 40. Further studies are required to determine whether the same regio-isomeric effect occurs with other emulsifiers used by the food industry. Furthermore, the present study was conducted with model TAG in which DHA occurred in combination with two saturated fatty acids. In natural fats and oils, PUFA occur in combination with monounsaturated fatty acids as well as saturated fatty acids. Natural oils also contain TAG molecular species, which are entirely made up of PUFA. For bulk oils, it has been shown that the superior oxidative stability of sn-2-bound PUFA also applies for TAG in which PUFA occurs in combination with two monounsaturated acids (Hoffmann et al. 1973; Wijesundera et al. 2008). Further studies are needed to determine regio-isomeric effects on the oxidative stability of oil-in-water emulsions of other TAG species abundant in natural oils.

Although experimental evidence has clearly shown that TAG regio-isomerism influences lipid oxidation in both oil-in-water emulsion and bulk oil, the reason for this effect is not known. The reaction mechanism and factors influencing lipid oxidation are substantially different between emulsified and bulk lipids (McClements and Decker 2000), and the mechanisms by which regio-specificity influences oxidation in the two systems need not be the same. In oil-in-water emulsions, lipid oxidation occurs at the interface, and it is possible that sn-2-bound DHA are able to orient a greater portion of its unsaturated groups inside the interior of the oil droplets rather than at the interface compared with sn-1(3)-bound DHA, thus making them less available for the oxidation reaction. However, there is no direct experimental evidence to support such a hypothesis.


In oil-in-water emulsions made with Tween 40 as the emulsifier, TAGs containing one chain of DHA and two chains of saturated fatty acid are more stable to oxidation when DHA is located at the sn-2 position compared with the sn-1(3) position. Therefore, modification of TAG structure is a potential means of enhancing oxidative stability of oil-in-water emulsion foods fortified with omega-3 oils.


The authors thank Claudio Ceccato for assistance with preparation of emulsions and headspace analysis, Peter Fagan for assistance with purification of synthetic TAGs and Simon Gladman for statistical analysis.