Comparison of Oxidative Stability of Monogalactosyl Diacylglycerol , Digalactosyl Diacylglycerol , and Triacylglycerol Containing Polyunsaturated Fatty Acids

Oxidative stability of three different lipid classes, namely, monogalactosyl diacylglycerol (MGDG) and digalactosyl diacylglycerol (DGDG) from spinach and edible brown seaweed (Akamoku) and triacylglycerol (TAG) of linseed oil was compared. Analysis of oxygen consumption and polyunsaturated fatty acid (PUFA) composition demonstrated that spinach DGDG had the highest oxidative stability, followed by Akamoku DGDG, Akamoku MGDG, spinach MGDG, and linseed TAG. These results disagree with the order of oxidative stability expected from the average number of bis-allylic positions of each lipid. Additionally, DGDG constituents of both spinach and Akamoku showed higher oxidative stability than their MGDG constituents. The unusual oxidative stability of MGDG and DGDG could be conferred by the protection of bis-allylic positions of the PUFA against oxidative attack by the galactosyl moiety of the GL.


Introduction
Monogalactosyl diacylglycerol (MGDG), digalactosyl diacylglycerol (DGDG), and sulfoquinovosyl diacylglycerol (SQDG) are the main lipid constituents of plant leaves (Figure 1) [1].In addition, these polar lipids are known to be the major constituents of seaweeds [2] [3].Together, MGDG and DGDG account for more than half of the total membrane lipids in plant leaves [4].Therefore, these glyceroglycolipids (GLs) are thought to play important roles in the photosynthetic membranes of higher plants, algae, and bacteria [5].
It is generally accepted that autoxidation of PUFA proceeds in a free radical chain reaction through three stages: i) initiation, ii) propagation, and iii) termination.The free radical theory of autoxidation was defined as the attack of oxygen at the allylic position leading to the formation of unsaturated hydroperoxides.These hydroperoxides can then decompose into peroxyl and alkoxyl radicals, which followed by secondary oxidation products including aldehydes, ketones, alcohols, and acids.Impaired taste, flavor, and texture of foods are among the adverse effects of secondary oxidation products, which may be toxic compounds.Thus, lipid oxidation is a major concern that can shorten the shelf life of foods.
Plants possess intrinsic antioxidant compounds for defense against oxidation, including ascorbate, glutathione, phenolic compounds, tocopherols, and carotenoids.Innate enzymes such as superoxide dismutase, catalase, ascorbate peroxidase, and glutathione reductase are capable of scavenging reactive oxygen species and eventually protecting plants from oxidative stress [10] [11] [12].Another defense system against oxidative stress in photosynthetic tissues may be related to the presence of PUFA in the form of GL.However, few studies have examined the resistance of GL to oxidation [4].Therefore, it is very important to determine the role of PUFA in the form of GL and its characteristics response.
This study evaluated the oxidative stability of GL by comparing the oxidative stability of 5 types of lipids: spinach MGDG, spinach DGDG, brown seaweed MGDG, brown seaweed DGDG, and linseed oil triacylglycerol (TAG).

Referenced Compound and Reagent
Silica gel (BW-60F) for column chromatography was purchased from Fuji Sylysia Chem.Ltd. (Kasugai, Aichi, Japan).Activated Carbon and Celite (545 RVS) were purchased from Nacalai Tesque, Inc. (Kyoto, Japan).MGDG, DGDG, and SQDG standards were purchased from Lipid Products (Redhill, UK), while triolein and tricaprylin as medium chain TAG (MCT) were obtained from Wako Pure Chemical.All other chemicals and solvents used in this study were of analytical grade and high-performance liquid chromatography (HPLC)-grade solvents were used for HPLC analysis.

Sample Preparation
Spinach powder (ca. 2 kg) was soaked in methanol (12,000 mL) at room temperature, which was incubated in the dark overnight (approximately 16 h) [13].The extract was filtered through a ceramic filter funnel lined with filter paper (No. 2, Qualitative Filter Paper; 150 mm; Advantec®; Tokyo, Japan).The filtrates were pooled and the solvent was removed using a pilot scale rotary evaporator at less than 40˚C.Traces of solvent remaining in the extract were completely removed in the dark under vacuum, leaving a dark green viscous liquid.This dark green viscous liquid was collected by dissolution in an equivalent volume of methanol, which was designated as crude spinach lipids.
Crude spinach lipids were further dissolved in chloroform-methanol-water (10:5:3, v/v/v) [13].The solution was placed into a separatory funnel for liquid-liquid distribution.After shaking, the funnel was allowed to stand overnight.
The lower layer, a mixed lipid layer with methanol and chloroform, was collected and again dissolved in water using a new separatory funnel.In this case, the same volume of water as that in the first separation, was added; after shaking, the funnel was allowed to stand for overnight separation.The lower layer Food and Nutrition Sciences was then concentrated under a vacuum in a rotary evaporator.Remaining traces of organic solvents and water were removed in a desiccator (approximately 3 days) under a high vacuum, leaving the sample in an amber-colored rotary flask.
Total lipids of spinach (ca.20 g) were eventually collected by dissolving in an equivalent volume of chloroform and subjected to consequent analysis or stored at −30˚C in an equivalent volume of ethanol.The same procedure was used for Akamoku powder.

Lipid Class Analysis
Total lipids of spinach and Akamoku were subjected to lipid class analysis by preparative thin layer chromatography (TLC) [14].The lipid fraction was dissolved in chloroform-methanol-water (65:25:4, v/v/v) and spotted onto a 0.25-mm silica gel plate (Silica gel 60G; Merck, Darmstadt, Germany).The plate was developed with the same cocktail mix solvents of chloroform-methanol-water (65:25:4, v/v/v) and the spots were visualized by spraying the plate with orcinol-sulfuric acid or Dittmer reagent, followed by charring.
The lipid sample was also analyzed by preparative TLC with the same specification of silica gel plate mentioned previously but using n-hexane-diethyl ether-acetic acid (80:20:1, v/v/v) as the developing solvent.The spots were detected using 60% aqueous sulfuric acid charring.
The chromatogram was photographed with a digital camera and the image of the silica gel plate was acquired and transferred to a computer.The image was cropped and saved in bitmap format.The percentage ratio of each lipid fraction (as compared to standards) in the sample was expressed as the bitmap percentage of the total bitmap intensities [14].

Purification of GL from Spinach and Akamoku
Total lipids of spinach and Aakamoku (ca.20 g) were first passed through a column (70 × 6 cm i.d.) packed with a chloroform slurry mixture of silica gel.
The whole column was wrapped with aluminum foil to protect the GL from light-induced degradation.The elution was first conducted with chloroform (approximately 3000 mL) and then with acetone (approximately 13,000 mL).
Fractions eluted with acetone were used as GL (through continuous elution, with appropriately adjusting the flow rate manually, for more than a week).The first whole dark green-black fraction was designated as MGDG and the consequent fractions (approximately 24 fractions) of each 500 mL collected with acetone (clear light green solution) were designated as the DGDG.
Although the absence of chlorophyll was confirmed in the DGDG, a trace amount of chlorophyll was detected in the MGDG on TLC.Therefore, MGDG (ca.
To confirm the absence of impurities, the TAG fraction was subjected to preparative TLC.The lipid fraction was spotted onto a 0.25-mm silica gel plate.The plate was again developed with n-hexane-diethyl ether (60:40, v/v) and spots were detected with iodine vapor or 60% aqueous sulfuric acid charring as compared to standard triolein.

Tocopherol Analysis of Spinach GL, Akamoku GL, and Linseed TAG
Tocopherol analysis was performed for all 5 types of purified lipids, Spinach MGDG and DGDG, Akamoku MGDG and DGDG, and Linseed TAG, with a Hitachi HPLC system equipped with a pump (Hitachi L-2130, Hitachi Seisakusho, Co., Tokyo, Japan) and fluorescence detector (Hitachi L-2485) [4].Analysis was conducted on a silica column (Si 60, 250 × 4.6 mm i.d.; Kanto Chemical Co., Tokyo, Japan) protected with a guard column (15 × 3.2 mm) with the same stationary phase.The mobile phase was n-hexane-2-propanol (99.2:0.8,v/v) at a flow rate of 1.0 mL/min.The fluorescence detector was set at Ex. 298 nm and Em.325 nm.

Fatty Acid Composition of Spinach GLs, Akamoku GLs, and Linseed TAG
The fatty acid composition of all 5 purified lipids was determined by gas chromatography (GC) after conversion of fatty acyl groups in the lipid to their methyl esters as described by Prevot and Modret [16], with slight modifications.
Briefly, to an aliquot of total lipid (ca.20 mg for GLs and 10 mg for TAG), 1 mL of n-hexane and 0.2 mL of 2N NaOH in methanol solution were added, vortexed for 10 s, and incubated at 50˚C for 30 s. Next, 0.2 mL of 2N HCL in methanol Food and Nutrition Sciences was added to the solution followed by vortexing for 1 min.The mixture was separated by centrifugation at 1000 ×g for 5 min.The upper hexane layer containing fatty acid methyl esters was recovered and subjected to GC. GC analysis was performed using Shimadzu GC-2014 (Shimadzu Corporation, Kyoto, Japan) equipped with a flame-ionization detector and capillary column (Supelco™ Column; Omegawax-320; 30 m × 0.32 mm i.d.; Sigma-Aldrich, St. Louis, MO, USA).The detector, injector, and column temperatures were 260˚C, 250˚C, and 200˚C, respectively.The carrier gas was helium at a flow rate of 50 kPa.The fatty acid content was expressed as a weight percentage of total fatty acids.

Oxidation Analysis of Purified Lipids
Each purified lipid (100 mg) was placed in a 2-mL aluminum sealed vial with a butyl gum septum (GL Science; Tokyo, Japan) and then incubated at 50˚C in the dark.The level of oxygen and nitrogen in the headspace gas of the vial was estimated using a GC system (Shimadzu GC-14B) equipped with a thermal conductivity detector and stainless steel column (3 m × 3.0 mm i.d.) packed with a molecular sieve 5A (GL Science) as described by Cho et al. [9].The temperatures at the injection port, detector port, and column oven were 120˚C, 120˚C, and 70˚C, respectively.The helium flow was 50 kPa.Three separate vials containing similar samples were prepared and incubated.A small portion (20 μL) of the headspace gas was taken from each vial using a microsyringe through the butyl gum septum at selected times during oxidation.Before the incubation, the ratio of peak area of oxygen to nitrogen was around 0.26.The ratio decreased with the progress of lipid oxidation.The percentage decrease (%) in oxygen could be calculated from the changes in the oxygen to nitrogen ratio compared to the ratio before incubation.In Three replicate measurements of all data and value at different oxidation times of the stored samples were expressed as the mean ± SD (n = 3).
After oxidation, PUFA contents were again analyzed by GC.The preparation of fatty acid methyl esters and GC analysis was performed as described above.

Lipid Class Analysis
When each lipid composition of both Spinach and Akamoku GLs was visually analyzed based on the spot intensities of TLC (Figure 2), the overall distribution of their main lipid classes was determined and is shown in Table 1.Only MGDG and DGDG from both Spinach and Akamoku were analyzed, while SQDG was not analyzed because it generally showed low levels in Spinach and thus was not the focus of this research.

Purification of GL and Linseed TAG
Purified MGDG and DGDG from Spinach and Akamoku showed spots corresponding to MGDG and DGDG standards on TLC, respectively.In addition,  visible spots with small streaks of chlorophyll were observed during TLC analysis with Spinach MGDG and Akamoku MGDG, while both Spinach DGDG and Akamoku DGDG were free from spots containing small streaks of chlorophyll after the first purification.However, after both MGDGs were subjected to a second purification with a carbon column, only a single spot corresponding to MGDG standard was detected on the TLC for each type of sample.
Purified Linseed TAG, in contrast, showed only a single spot corresponding to the lipid standard triolein on analytical TLC.No other impurities such as free fatty acids, monoacylglycerol, or diacylglycerol were detected.

Tocopherol Analysis
HPLC analysis either showed complete removal of tocopherols for nearly all types of lipids or only a very small amount of this antioxidant compound remained for the remaining few types of lipids.Complete removal was confirmed after double purification of all GL samples.Minute and negligible amounts of tocopherol, however, were detected for samples of Linseed TAG, as reported by Shimajiri et al. [17], which do not affect comparisons of the oxida-Food and Nutrition Sciences tive stability results.
Table 3 shows the number of bis-allylic positions per molecule of each lipid.The number of bis-allylic positions per molecule or gram of each lipid was determined from the molar concentration of each PUFA and the mean molecular   3).

Stability of Purified Lipids in Bulk Phase
When the oxidative stability of different types of lipids was compared by measuring the decrease in oxygen concentration in the bulk phase (Figure 3), the stability was the highest for Spinach DGDG, followed by Akamoku DGDG and MGDG, and Spinach MGDG.Linseed TAG was shown to have the lowest oxidative stability.The oxidative stability of polyunsaturated lipids is decreased with an increasing number of bis-allylic positions [18].Based on the number of bis-allylic positions of each lipid (Table 3), Spinach DGDG was expected to be easily oxidized as compared with Akamoku DGDG, Linseed TAG, and Akamoku MGDG.However, the oxidative stability results, shown in Figure 3, were quite different from the expected average number of bis-allylic positions.In addition, both GLs with the DGDG constituent were shown to be more stable than their MGDG constituent when they showed a mild decrease in oxygen concentration over time (Figure 3).Although Spinach and Akamoku MGDG were In phosphatidylcholine liposomes, the model system of biological membranes, DHA showed higher oxidative stability than LA [27] [28].These results can be explained by the protective conformation of the bis-allylic positions of DHA against hydrogen abstraction.The characteristic structure of the DHA (22:6n-3) moiety has been observed in several membrane models.A molecular modeling approach showed that DHA of diacylglycerol may uniquely influence acyl chain packing arrangements in cell membranes [29] [30] [31].When in these conformations, DHA hexanes can form tightly packed arrangements to protect the bis-allylic positions of DHA from oxidative attack.
Therefore, the characteristic oxidative stability of GL found in the present study did not result from contamination, but rather from differences in the chemical structure of each lipid molecule.MGDG and DGDG contain both polar (galactosyl moiety) and non-polar (fatty acyl group) regions on the same molecule, which can form a self-assembly structure [32].In this structure, galactosyl moieties may protect the bis-allylic positions of PUFA in GL through various interactions.The present study revealed the higher oxidative stability of DGDG than that of MGDG for the first time.This may be because of the stronger protective effect of di-galactosyl moieties on PUFA against oxidative attack than that of mono-galactosyl moieties.

Conclusion
In the present study, we found that PUFAs in the form of GLs were oxidatively more stable than those of TAG.In addition, when the oxidative stability of MGDG and DGDG from Spinach and Akamoku was compared, DGDG showed higher oxidative stability than its corresponding MGDG.These results improve the understanding of the oxidation of GL containing PUFA in biological systems and may be useful for preventing PUFA from undergoing oxidative deterioration.Further studies are required to elucidate the detailed mechanism of the unusually high oxidative stability of GLs.

Figure 2 .
Figure 2. Representative TLC of spinach lipids.Sample solution was spotted on a 0.25-mm silica gel plate.The plate was developed with chloroform-methanol-water (65:25:4, v/v/v) and the spots were visualized by charring.

Figure 3 .
Figure 3. Oxidative stability of Spinach MGDG (open square), Spinach DGDG (closed square), Akamoku MGDG (open triangle), Akamoku DGDG (closed triangle), and Linseed TAG (open circle).Stability was analyzed by measuring the decrease in oxygen in the headspace gas of the vial.Oxidation was performed at 50˚C in the dark.The data are expressed as the mean ± SD of three independent experiments.

Table 1 .
Lipid class composition of Spinach GL and Akamoku GL.
GL class (% of total GL)

Table 2 .
Major fatty acid composition of oxidation substrates before and after the oxidation.

Table 3 .
Average number of bis-allylic positions of substrate lipids.Food and Nutrition Sciences weight (MW) of each lipid.The molar concentration of PUFA was calculated based on the weight percentage of PUFA before oxidation (Table2).The mean MW of each lipid was obtained from the mean MW of all fatty acyl moieties and MW of the galactosyl and glycerol moieties.MWs were as follows: Spinach